Space in Language and Linguistics: Geographical, Interactional, and Cognitive Perspectives 9783110312027, 9783110311969

Citation preview

Space in Language and Linguistics linguae & litterae

24

linguae & litterae Publications of the School of Language & Literature Freiburg Institute for Advanced Studies

Edited by

Peter Auer · Gesa von Essen · Werner Frick Editorial Board Michel Espagne (Paris) · Marino Freschi (Rom) Ekkehard König (Berlin) Michael Lackner (Erlangen-Nürnberg) Per Linell (Linköping) · Angelika Linke (Zürich) Christine Maillard (Strasbourg) · Lorenza Mondada (Basel) Pieter Muysken (Nijmegen) · Wolfgang Raible (Freiburg) Monika Schmitz-Emans (Bochum)

24

De Gruyter

Space in Language and Linguistics Geographical, Interactional, and Cognitive Perspectives

Edited by Peter Auer, Martin Hilpert, Anja Stukenbrock and Benedikt Szmrecsanyi

De Gruyter

ISBN 978-3-11-031196-9 e-ISBN 978-3-11-031202-7 ISSN 1869-7054 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. 쑔 2013 Walter de Gruyter GmbH, Berlin/Boston Printing: Hubert & Co. GmbH & Co. KG, Göttingen 앝 Printed on acid-free paper 앪 Printed in Germany www.degruyter.com

Contents

V

Contents

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi Integrating the perspectives on language and space . . . . . . . . . .

1

Section 1: Geography and variation across languages Michael Cysouw Disentangling geography from genealogy . . . . . . . . . . . . . . .

21

Johanna Nichols The vertical archipelago: Adding the third dimension to linguistic geography . . . . . . . . . . . . . . . . . . . . . . . . . .

38

Walter Bisang Language contact between geographic and mental space . . . . . . .

61

Commentary: The notion of space in linguistic typology (Ekkehard König) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Section 2: Geography and variation within languages Barbara Johnstone Ideology and discourse in the enregisterment of regional variation . . 107 Paul Kerswill Identity, ethnicity and place: The construction of youth language in London . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Bernd Kortmann How powerful is geography as an explanatory factor in morphosyntactic variation? Areal features in the Anglophone world . 165 Elvira Glaser Area formation in morphosyntax . . . . . . . . . . . . . . . . . . . 195 John Nerbonne How much does geography influence language variation? . . . . . . . 222

VI

Contents

Commentary: Lost in space? The many geographies and methodologies in research on variation within languages (Benedikt Szmrecsanyi) . . . . 240

Section 3: Interactional spaces Lorenza Mondada Interactional space and the study of embodied talk-in-interaction . . 247 Heiko Hausendorf On the interactive achievement of space – and its possible meanings . 276 Jürgen Streeck Plaza: Space or place? . . . . . . . . . . . . . . . . . . . . . . . . . 304 John B. Haviland Xi to vi: “Over that way, look!”: (Meta)spatial representation in an emerging (Mayan?) sign language . . . . . . . . . . . . . . . . . . . 334 Commentary: What difference does space make for interaction and interaction for space? (Anja Stukenbrock) . . . . . . . . . . . . . 401

Section 4: Mobile spaces Pentti Haddington Action and space: Navigation as a social and spatial task . . . . . . . 411 Elwys De Stefani Rearranging (in) space: On mobility and its relevance for the study of face-to-face interaction . . . . . . . . . . . . . . . . . . . . . . . 434 Commentary: Being mobile, talking on the move: Conceptual, analytical and methodological challenges of mobility (Lorenza Mondada) . . . . 464

Section 5: Mediated spaces Marco Jacquemet Language, media, and digital landscapes . . . . . . . . . . . . . . . . 473

Contents

VII

Michael Beißwenger Space in computer-mediated communication: Corpus-based investigations on the use of local deictics in chats . . . . . . . . . . . 494 Christian Mair and Stefan Pfänder Vernacular and multilingual writing in mediated spaces: Web-forums for post-colonial communities of practice . . . . . . . . 529 Susanne Uhmann Pointing in the abdomen . . . . . . . . . . . . . . . . . . . . . . . . 557 Commentary: Making space (Monica Heller) . . . . . . . . . . . . . . . 601

Section 6: Typology and spatial reasoning Gisela Fehrmann Exploiting space in German Sign Language: Linguistic and topographic reference in signed discourse . . . . . . . . . . . . . . . 607 Jürgen Bohnemeyer and Randi Tucker Space in semantic typology: Object-centered geometries . . . . . . . 637 Alan Cienki Gesture, space, grammar, and cognition . . . . . . . . . . . . . . . . 667 Commentary: Is there a deictic of frame of reference? (Holger Diessel) . 687

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 693

VIII

Contents

Integrating the perspectives on language and space

1

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

Integrating the perspectives on language and space

This book brings together different perspectives on the relation of language and space. The contributions in this volume were originally presented at a conference series held at the Freiburg Institute for Advanced Studies in the fall of 2009. These meetings focused on three aspects of language and space that are quite well-researched within themselves, but which so far are lacking productive interconnections. Specifically, the Freiburg workshops on language and space addressed the following issues: Language, space, and geography Grammar, space, and cognition Language and interactional spaces It was the chief purpose of these workshops, and by extension of this volume, to create a common forum for research traditions that show substantial overlap in their subject matter but which differ in theoretical outlook and research methodology. It is our contention that a rapprochement of these largely separate research traditions is necessary for further progress in the study of language and space and that a more inclusive view holds mutual benefits for each of the participating fields. But what exactly are the traditions that we aim to bring closer together? Under the rubric of “language and space”, linguists have been researching several seemingly unrelated areas of language. Most straight-forwardly perhaps, “language and space” refers to the areal distribution of linguistic forms. Here, the objects of investigation are linguistic expressions that vary across space, so that it is possible to map linguistic forms onto geographical areas. In another interpretation, “language and space” focuses on the grammatical expressions of a language which are used to locate entities or describe movements in space. This notably includes deictic elements, but also spatial adpositions and lexical elements that denote spatial relations. Yet another view of “language and space” refers to the use of space in linguistic interaction, i.e. how groups of speakers arrange themselves in space and how they make use of co-speech gesture. The use of space as a communicative resource has also been studied extensively in research on signed languages. When one compares these three traditions – areal linguistics, research on spatial expressions, and the study of space in linguistic interaction – it appears that these fields of enquiry are not only concerned with entirely different phex x x

2

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

nomena, but that they furthermore embrace opposing ways of conceptualising space that have been competing in Western philosophical thinking for a long time. Areal linguists regularly refer to absolute spaces (“containers” existing independently of the linguistic forms that are located “within them”), while grammatical markers of spatial relations often encode relative spaces, so that for instance a deictic expression locates an object in space relative to a point of reference, such as the speaker or a salient landmark. It seems that language lends itself to both views of space – the absolute and the relative – but with very different resources. In order to motivate a rapprochement of the research traditions that are represented in this volume, we want to argue that the seemingly mutually exclusive opposition between the pragmatics of space in the narrow, deictic sense and the areal distribution of linguistic forms is less compelling than is often thought. The commonalities between the two become apparent as soon as they are seen as two instances of linguistic indexicality that are both rooted in interaction. It is this idea that motivated the tripartite structure of the FRIAS workshops on language and space and that is continued in the different sections of this volume, which are meant to mutually illuminate one another. When studies on different aspects of language and space are arranged side by side, one can begin to see unity in the diversity that characterizes the study of language and space. In the remainder of this introductory chapter, we will first offer a more indepth argument for the unification of perspectives that we envision; after that, we present a short explanation of how the volume is structured and what topics are covered by the individual contributions.

1.

The traditional view

Since antiquity, Western philosophical thinking about space has been characterised by two different approaches. The Aristotelian view is that of space being absolute, i.e. existing before and beyond individuals’ acting in space, and even before and beyond the existence of any matter in it. In this tradition, space was first (in antiquity) thought to be a container, the limits of which were the stars. Later (in Newton’s view), the container lost its limitations and was conceived as endless; nonetheless, it was thought to exist independently of the objects it contained and of their relationship to each other – an empty space (cf. von Weizsäcker 1986). Since modernity, the Euclidian notion of space as a three-dimensional, metric space has also dominated common-sense thinking. The popularity of this view is due in many ways to the development of techniques of mapmaking in early modernity, a technology of reducing three-dimensional to two-dimensional space by which every point on the

Integrating the perspectives on language and space

3

earth can be defined exactly by its longitudinal and latitudinal parameters. GPS and the applications of geographical computer technology are perhaps further strengthening the idea of an absolute space, since this technology provides ways of locating all sorts of things and experiences independently of the spatial location or perspective of the human user. There is, however, a different tradition of theorising space, according to which it is always relative. The phenomenological version of relative space introduces the human experiencer and the body (in the sense of Leib – ‘lived body’; Merleau-Ponty 1945) as the reference point from which all spatial relations originate. For phenomenologists, space is always oriented towards the lived body, its movements and actions, so that human beings live space, rather than live in space. Space is thus actively construed by human agents, be it as fleeting interactional spaces which are formed and dissolved with every social encounter (Goffman 1963), or as more permanent places – socially meaningful spaces that are relevant because of the activities taking place in them, the values ascribed to them, or the social conventions that are associated with them. Although it was claimed above that common-sense thinking about space since modernity has been dominated by the absolute notion of space, relative notions of space also continue to play an important role today in everyday practices. For instance, it continues to be more important to find the way from ‘here’ to ‘there’ incrementally (i.e. by being able to decide correctly at each junction where to turn), than to know the location of the origin and the target in absolute space. Modern technologies of wayfinding (GPS systems in cars) are adapted to this preference: They are not restricted to indicating the way from ‘here’ to ‘there’ on a map, but they also translate this knowledge into pragmatically contextualized route directions that guide the car driver from one location to the next until the target has been reached. The relative conception of space is firmly established in the linguistic tradition of studying deictic expressions. One universal design feature of human languages is the existence of linguistic signs that deal with space in a relative, perspectival way (Bühler 1934). These signs have been the object of detailed study (cf. Jarvella & Klein eds. 1982; Weissenborn & Klein eds. 1982; Fillmore 1975[1997]), and the recent cognitive turn in linguistics has further contributed to this field of research (cf. Habel & von Stutterheim 2000) by linking the pragmatics of spatial deixis to research on human (collaborative) orientation and movement in space. Most of this research is concerned with spaces that are visually accessible. The relative nature of spatial deictics is obvious, since their interpretation is dependent on the establishment of an origo, most often the speaker. This origo can be transferred to

4

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

non-speaking, even imaginary humans (displaced origines, deixis in the imagination), and some types of objects can even be conceptualised as having an intrinsic perspectivity (Bühler 1934; Levinson 2003), yet the nature of spatial reference of this kind remains perspectival. Recent research on pointing activities and spatial deixis in non-European cultures/languages has also demonstrated how the choice of deictic elements, rather than being determined by spatial parameters, construes spaces (including social ones; cf. Hanks 1990, 2004; Enfield 2003). Independently of this research tradition, linguists have used absolute notions of space to investigate the diversity of language(s) in areal terms. In this tradition (sometimes called geolinguistics, or areal linguistics), geography is used to locate linguistic forms in space. Its origins go back to nineteenth-century dialectology, the field of linguistics in which arguably the highest degree of geolinguistic sophistication has been reached. As shown in Auer (2005), its origins coincide with the heydays of spatial thinking as a constitutive feature of nation building, i.e. the political idea that nations have their own language spaces which they occupy exclusively (and vice versa: language spaces are at the basis of the existence of nation states). However, linguistic cartography has survived the national(istic) preoccupation with space and its political connotations and has flourished in recent years, after a period of dominantly non-spatial sociolinguistic theorising about language variation. Again, the availability of new technological tools seems to have played its part in the emergence of a new generation of map-based, partly georeferenced linguistic studies, which are now concerned not only with the distribution of dialect forms in national or transnational spaces, but also with the distribution of linguistic forms in the languages of the world (see, as a particularly impressive example, the WALS).

2.

A reappraisal: the field of spatial indexicalities

At first sight, then, it seems that language is linked to space in two different ways; first, spatial deictics provide languages with a specific technique for constructing spaces which are perspectival and orientated. Second, language features are distributed in a way which can be mapped onto geographical space. However, a closer look into how space is made relevant in interaction will reveal that this distinction is less clear than it initially appears. We would like to argue that the way in which interaction is organised necessitates a reappraisal of the seemingly simple distinction between spatial pragmatics on the one hand and the areal distribution of linguistic forms on the other. The reappraisal suggests a shift of perspective particularly on the areal side, but partly also on the side of spatial

Integrating the perspectives on language and space

5

deixis. The basic argument is that both spatial deictics and linguistic forms indexing geographical position overlap in their function to emplace individuals and social communities. The term spatial indexicality is suggested as an overarching term which covers both ways of referring to and making relevant space through language. Despite the arguments presented in this section for a rapprochement of the two conceptions of space, it goes without saying that there are fundamental differences regarding the nature of deictic indexicality on the one hand and geographic, variation-based indexicality on the other. Both are locational devices, but they locate in different ways, with different limitations and on different cognitive and social foundations: (1) Spatial deictics function on the basis of shared spatial representations that are contingent on the interactional “formation” in the sense of Kendon (1976) (i.e. the roles and bodily positions of speaker and hearer). This representation of space is highly adaptive and temporally unstable, since the shared space is recreated with each new interactional formation. By contrast, linguistic variation can be used to locate speakers in space only on the basis of relatively stable knowledge about how people speak in different places, i.e. on the basis of ethnolinguistic knowledge. Matching speakers with places on the basis of language is a procedure in which the experience of difference between the way ‘we’ speak and the way ‘they’ speak is the index (cf. Irvine’s notion of style as difference, 2001), and knowledge about who supposedly speaks in which way is the indexed. This may of course be second- or thirdhand knowledge and not derived from participants’ personal experience. (2) Spatial deictics are a way of referring to objects tout court, while variationbased indexicality is obviously restricted to locating speakers; non-human referents usually do not speak and therefore do not have a linguistic style on the basis of which they could be ‘emplaced’. But by the same token, variation-based indexicality carries more than geographical meaning; it almost inevitably implies a social evaluation of the speaker. It shares this feature with non-deictic spatial expressions. Deictic indexicals of space, by contrast, do not usually carry social meaning. (3) Variation-based indexicality is relatively stable across different situations, but it is also subject to interpretative mistakes (somebody is wrongly assigned a place because the hearer does not have sufficient ethnolinguistic knowledge) and fakes (somebody may pretend to be from a place from which s/he actually is not). Above all, over historical time the link between

6

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

geography and language may be weakened, entirely dissolved, or redefined, even though stereotypical knowledge about it may persist in a community. In general, speakers are able to accommodate their interlocutors, and they may have more than one “authentic” linguistic variety or style at their disposal. It is one of the consequences of modern and postmodern social as well as geographical mobility that the most telling indices of spatial belonging, the traditional dialects, have disappeared more or less completely, and that speakers are either multidialectal (and multilingual) or only use varieties which are relatively difficult to locate in space. (4) Spatial deictics constitute a well-defined field of linguistic expressions whose primary function it is to locate referents in space. Variation-based indexicality is a secondary kind of indexicality in the sense that any linguistic expression, in addition to whatever function it may have, can carry an additional layer of information which emplaces the speaker. We can now come back to our initial claim that the distinction between absolute and relative notions of space cannot be associated with the two types of spatial indexicality in a one-to-one way. First, it should be noted that nondeictic spatial expressions (such as place names) often occur in combination with deictic elements in interaction and that deictic and non-deictic expressions are often mutually paraphrased. More important for our argument is however another group of spatial deictics, which are built on an absolute notion of space, i.e. so-called environment-centred frames for reference (cf. the contributions by Bohnemeyer & Tucker and Diessel in this volume). In the modern European languages, they survive in two formats which still compete in referential expressions for locations in geographical space relative to each other. The older format is based on natural topography, while the newer format is based on cardinal directions. For instance, the location of, say, the city of Munich relative to the city of Ratisbon can be described as ‘up(-hill)’, and accordingly, one would go from Ratisbon nach München rauf ‘up to Munich’ based on the fact that Munich is closer to the Alps than Ratisbon; but competing with this absolute spatial reference is another one in which Ratisbon is ‘up’, and a speaker who follows this system would go from Ratisbon nach München runter ‘down to Munich’, based on the fact that Munich is south of Ratisbon, and ‘south’ equals ‘down’ on a map. In both reference systems, the spatial description of the two cities (and movements from one to the other) remains the same, regardless of whether the speaker is in Munich or Ratisbon or in neither of the two cities. The verbal expression is not dependent on the speaker-origo.

Integrating the perspectives on language and space

7

Absolute frames of reference of this kind have recently been the object of a huge amount of anthropological research, mainly on languages in Northern Australia and Central America in which they are not only a marginal possibility like in German, but the dominant, unmarked system (see Levinson 2003, Bennardo 2009). In these languages, all objects are placed relative to a frame of reference which is determined by (usually culturally salient) distinctions such as ‘towards the sea’ vs. ‘towards the mountain’, or ‘up-mountain’/‘down-mountain’. The typological distinction between languages that favour origo-centred, and those that favour absolute1 frames of reference shows that the placement of objects in space is not universally based on a relative notion of space. There is reason to believe that absolute systems are typical of traditional, pre-modern societies. However, as the example of the cardinal directions shows, modern societies have specialised uses of absolute systems as well, even though their use is limited – above all, absolute frames of reference would not be used when the objects to be described spatially are within the visual reach of the participants or within buildings. While the typologically somewhat marginal case of languages that prefer absolute frames of reference for spatial expressions shows that spatial pragmatics and absolute notions of space are not incompatible, we now turn towards the more important relationship between variation-based spatial indexicality and relative notions of space. The decisive point to be made here is that although the areal distribution of language features can be plotted onto maps that are the technical counterpart of an absolute, two-dimensional conception of geography, these maps do not do justice to the ways in which linguistic variation functions as an index to a speaker’s spatial belonging in interaction. They disregard the sociolinguistic function of linguistic variability as a way to tell where somebody “comes from” and where s/he “belongs”, both in the concrete, spatial and in the more derived, social sense. We would like to propose that the categorisation of interactional co-participants – particularly of strangers – is a fundamental function of language, and that “placing” a speaker plays a central part in this categorisation, minimally consisting of the distinction between “one of our kind” and “one of a different kind”, which is perhaps a version of “someone I can understand (easily)” and “someone I cannot understand (easily)”. As soon as a community shares 1

Sometimes the term “relative frame of reference” is used to refer to ego-centred frames of reference; this however, is misleading since intrinsically organised frames are equally relative. Origo-centred frames of reference are therefore one type of relative frames only.

8

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

some ethnolinguistic knowledge about how language features are areally distributed, this knowledge can enter into a process of social categorisation in which social attributes are derived from spatial locations. Linguistic differences are exquisitely suited for this purpose; they can be extremely finegrained and thereby convey information that is more subtle than, for instance, bodily attributes such as skin colour. Furthermore, they are less subject to voluntary change than, for instance, clothing styles: languages can be learned, and accents lost, but this takes effort and time and is subject to many restrictions. Linguistic features therefore are a relatively reliable and at the same time fine-tuned instrument to display and recognise local and thereby social differences between speakers. The claim put forward here is that linguistic differences index spatial parameters (belonging, place, provenance) in terms of social parameters (inclusion/exclusion, superior/inferior, known/unknown). The exact nature of the link between spatial and social categorisation is complex and subject to historical change. Positing such links is contingent both on the structure and size of (to-be-categorised) speakers’ repertoires (their multilinguality, multidialectality, multistylistic resources) and the ethnolinguistic knowledge of the recipient-interpreters. In Europe, as well as elsewhere, both aspects have been substantially restructured over the last 500 years or so: many dialectal forms have been lost, regional and national varieties have developed. Regardless of these changes, the ethnolinguistic resources recipients use to place a speaker areally and socially remain perspectival, i.e. they are structured differently in the vicinity of the speakers’ own location and in faraway places (cf. Preston 1989, Niedzielski & Preston 2000). Building on Silverstein’s (2003) orders of indexicality and Labov’s (1972) distinction between indicators, markers and stereotypes, Johnstone et al. (2006) identify three types of variation-based indexicalities which can be defined by their different alignments of geography and social structure. Very roughly, they can be approximated to traditional, modern and late modern models of spatial-social indexicalities, although a precise historical account would need to differentiate between groups of speakers along the social hierarchy with their specific linguistic repertoires. The first (“traditional”) model of spatialsocial indexicality is based on a strict one-to-one relationship between spoken language and geographical location. In this model, spatial parameters can be read from language more or less directly; the way in which somebody speaks betrays his or her place of origin (which equals the place of living), and very often, this place of origin/living also makes some kind of social categorisation possible. The “others” are those who live in another place. As a rule of thumb in ethnolinguistic categorisation, the relative degree of incom-

Integrating the perspectives on language and space

9

prehensibility indicates the distance of the place to which the speaker belongs (with all the fallacies such a rule of thumb may imply). In modernity and late modernity, this strict “language–body–place connection” (Quist 2010) becomes untied in two steps. In Johnstone’s second-order indexicality, space remains the most important source of linguistic heterogeneity, but its interpretation assumes a more complex structure. Since one symbol of and justification for modern nation building is language (Stukenbrock 2005), sameness or difference of language is now relevant not only on the level of the traditional dialects, but also on the level of the nation-state, which has “its own language”. The national language space, unlike dialectal space, cannot be grasped by experience alone: it is an imagined space with borders that mainly exist on paper, above all in the format of maps,2 and which is defined by a new type of language variety, the standard. The tension between the local vernacular (dialect) and a normative, standardised language associated with the nation-state becomes the motor and the new symbol of social differentiation, i.e. it defines a vertical structure on top of the existing horizontal one. This leads to a reevaluation of all forms of speaking: what used to be nothing more than the ‘natural’ way of speaking in a given location now becomes, in the worst case, the language of the underprivileged classes who have no access to education, a variety that needs to be avoided in out-group situations. The second social transformation is the late modern one which leads to a third model of spatial-social indexicality based on linguistic heterogeneity. It occurs in the context of the processes called globalisation which set in during the last quarter of the last century, and which Bauman (1998: 8) calls the “Great War of Independence from space”. But globalisation not only means dissolving the language-space connection. The age of globalisation has also seen a countertendency – a new interest in symbolising belonging in spatial terms, in turning abstract space into places (Cresswell 2004) which are impregnated with meaning, and which symbolise belonging. The ‘local community’ is no longer brought about by space itself (if it ever was), since the former limits on travel and communication beyond the local community have largely disappeared. Rather, people who live in a location – born there or (more often) not – may choose to construe a local identity for themselves. Place-making activities abound. The less speakers “naturally” give off information about “their” spatial grounding by the way they talk (based on the assumption that they have no choice), the more important talk about place seems to be; and there is surely more of it – about places of birth, places of 2

See Anderson 2006: 170–178 as well as Dipper & Schneider (eds.) 2006 on the relationship between modern cartography and the nation state.

10

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

living, differences between places – than 100 years ago. In the tension between standardisation and destandardisation which characterises the present situation, the new media play an important role. For one thing, they provide an instrument of communication by which diasporic language spaces resulting from migration can be held together and tied back to their “home country”. Additionally, they enable non-standard, vernacular forms to spread around the globe (Meyerhoff & Niedzielski 2003), as well as make available a repertoire of linguistic forms which are not those the speaker/ writer “owns” (crossing in Rampton’s 1995 sense): these linguistic variables appear in new functional contexts and assume multiple meanings (as prototypically shown by the use of fragments of Afro-American English or Jamaican Creole by speakers of completely different ethnic belonging). localization of referents through spatial deixis on the basis of deictic systems, shared frames of reference, bodily coordination

localization of speakers through the space of the interactional encounter

geolinguistic spatial indexicality on the basis of shared knowledge: who speaks where how?

Fig. 1: Dimensions of “language and space”: the double spatial indexicality of language

In sum, the spatial-social indexicality of linguistic variation is deeply linked to the ongoing social transformations of the world; however, it always becomes relevant and is activated within situated encounters – the same encounters in which participants establish reference by spatial deictics. The double spatial indexicality of language (cf. Fig. 1) which results from these employments of very different linguistic resources can become intertwined in many ways. Although this argument does not preclude the possibility and justification of treating the two types of spatial indexicality in different domains of linguistic research (spatial pragmatics on the one hand, areal linguistics on the other), it also speaks in favour of a reappraisal of the theoretical and methodological embeddings of the two fields. In particular, it is not the case that spatial pragmatics is built on relative notions of space, while areal linguistics is intrinsically bound to absolute notions of space. In the field of person reference and person categorisation, both overlap and are often used together.

Integrating the perspectives on language and space

3.

11

Structure of this volume

The present volume is structured into six sections. The first two sections address language and geography, the first focusing on variation across languages and the second on variation within languages. Section three and four present different analyses of interactional spaces, both in stationary and mobile settings. Section five extends the discussion to mediated spaces, which include digital landscapes and computer-based channels of communication. Section six wraps up the volume with contributions on typology and spatial reasoning. In each of the sections, the contributions are concluded by a commentary that highlights common threads or offers pointers that go beyond the specific case studies. The following paragraphs give a brief overview of the individual contributions.

3.1

Geography and variation across languages

Michael Cysouw addresses the problem that typological similarity between languages can be due to two potentially confounding factors: a close genealogical relation or close geographical proximity. Quite often these two factors go hand in hand so that it is difficult to disentangle their respective effects. Cysouw uses data from the WALS to develop a solution to this problem. Johanna Nichols argues that areal linguistics has paid insufficient attention to the variable of altitude, which often disappears in the making of two-dimensional maps. Nichols shows that altitude has tangible sociolinguistic effects, notably isolation, but also others, which in turn leave their mark on language structure. Focusing on mountain areas with a central crest, Nichols explores what generalizations can be made about how language responds to verticality relations. Walter Bisang explores two aspects of language contact and shows that contact-induced structural convergence is not limited to geographic contiguity. He compares situations of contact through speaker mobility with contact situations in which speakers and their respective speech communities are in contact only by way of writing. As Bisang argues, the latter type is not limited to modern electronic communication (cf. section 5 on mediated spaces); a far longer tradition can be claimed for them. In discussing these phenomena, Bisang shows that language contact matters even beyond geographically contiguous areas. Ekkehard König concludes this section by reflecting on the notion of space in linguistic typology.

12 3.2

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

Geography and variation within languages

Barbara Johnstone elaborates on the concept of place, which critically goes beyond a geographic localization, but which is revealed to be an ideological construct that is created in human interaction. Varieties of speech that speakers perceive as being tied to places emerge through the process of enregisterment. Encounters with linguistic differences or emblematic linguistic forms can enregister these forms, that is, turn them into a mark of place and place-identity. Paul Kerswill discusses identity, ethnicity and place in the construction of youth language in London. His paper reports on two research projects, namely Linguistic innovators: the English of adolescents in London, and Multicultural London English: the emergence, acquisition and diffusion of a new variety. In these projects, Kerswill and collaborators collected speech data from the inner London boroughs of Hackney, Haringey and Islington, as well as the outer borough of Havering. The projects investigated not only the production of inner- and outer-city speech but also the social construction of innerand outer-city varieties in a complex interplay of factors such as ethnicity, social class, gender, age and place. Bernd Kortmann poses the question how well geography can explain degrees of morphosyntactic variation. By comparing a large catalogue of morphosyntactic features across a global set of varieties of English, Kortmann shows that geography discriminates well at the micro-level, particularly at the level of traditional British dialects. At the same time, the global comparison shows that variety type, i.e. the question whether a variety of English is for instance a traditional dialect, a learner variety, or a creole, largely outranks geography as a predictor of morphosyntactic structure. Elvira Glaser makes the case that the geographical distribution of forms, although well-researched, is not as properly understood as would be desirable, specifically so with regard to morphosyntactic forms. On the basis of data from Swiss German, she develops an approach towards a firmer grasp of this idea. She also argues that the geographical distribution of morphosyntactic variants is prevailing, even though the dissolution of traditional dialects is a commonplace in popular discussions of language. John Nerbonne shares his opening observation with Johanna Nichols: Most geolinguistic studies have reduced geography to the mere distance between two points. He goes on to develop an alternative operationalization of geography that includes the notion of area. In a study based on German dialect atlas data it is shown that this operationalization is a useful complement to the simpler, distance-based implementation. Nerbonne suggests that

Integrating the perspectives on language and space

13

geography can in fact be modeled in a more complex way that includes several mixed forms of geographical influence on language. Concluding this section, Benedikt Szmrecsanyi comments on the many geographies and methodologies in research on variation within languages, asks whether research in this area may be lost in space, but identifies the social dimension of language variation as a common denominator towards which all contributions are oriented. 3.3

Interactional spaces

Heiko Hausendorf explains that space in linguistic interaction is not an invariant medium that pre-exists the speakers, but rather that space is interactively achieved in the speech situation. Hausendorf breaks down this process into the sub-problems of co-orientation (referring to perception), co-ordination (referring to movement) and co-operation (referring to action). A close inspection of these sub-problems reveals that the speech situation emerges as the participants’ solution to these problems. Jürgen Streeck investigates the concepts of space and place in a study of social interaction on a South American plaza in the city of Cartagena de Indias, Colombia. The study shows how the plaza can be understood as a self-organizing and self-sustaining system of face-to-face interactions, where repeated types of social interaction lead to an emplacement of their participants. This for instance includes the social apprenticeship of younger participants, who observe, and eventually participate in, the age-graded social practices that the plaza affords. John Haviland reports on spatial expressions in a nascent sign-language that was created by Zinacantec Indians from the Mexican Chiapas Highlands. Haviland compares the spatial frames of reference of the surrounding spoken lanugage, Tzotzil, with the devices that are used in this emerging sign language. This is done through an elaboration of Jakobson’s (1957) distinction between a narrated event and a speech event, which is used here to distinguish a narrated space, a speech event space and an interactional space. Lorenza Mondada further develops the notion of interactional space, defining it as the situated, mutually adjusted changing arrangements of the participants’ bodies within space. These arrangements depend on the activity the participants are engaged in, the objects that are involved in this activity, and the participants’ mutual focus of attention. This interactional space is dynamic, it is constantly updated within the activity. Mondada illustrates these concepts with examples of conversational openings and acts of giving directions. The section ends with a commentary by Anja Stukenbrock.

14 3.4

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

Mobile spaces

Pentti Haddington focuses on mobile interactional spaces in a study of how speakers handle the navigation of a car as a collaborative social and spatial task. On the basis of video-recorded interactions in cars, it is shown how, on the one hand, action shapes space, i.e. how participants construct a mutual understanding of space and, on the other hand, how space shapes action, i.e. how participants adapt to the pre-existing features of space by adjusting their verbal and embodied behavior to the spatial context. Elwys De Stefani also addresses aspects of linguistic interaction in mobile settings, emphasizing that participants may modify their relative positions in space as they interact. Certain activities, such as giving directions or pointing out topographic features, require bodily repositionings that allow participants to use the affordances that are provided by the spatial environment. From this point of view, the surrounding space of an interaction is a resource that speakers can exploit in shifting ways for their communicative purposes. In her commentary, Lorenza Mondada sketches the conceptual, analytical and methodological challenges posed by the new mobility paradigm as it “mobilizes” our notions of language, space, (inter)action and context. 3.5

Mediated spaces

Marco Jacquemet describes how contemporary electronic media create a communicative space that is characterized by language mixing, hybridization, and syncretic communicative practices. Whereas mediated communication has in the past been approached as a unidirectional, centralized, and production-centered phenomenon, these defaults no longer apply in contemporary digital environments. Jacquemet’s paper explores the communicative practices of social groups in various media technologies, exposing how different languages and communicative codes are simultaneously in use. Michael Beißwenger investigates the use of spatial deictics in computer-mediated written communication, specifically chats. Using spatial deictics in environments where the communicating parties do not share the same physical space and neither see nor hear each other is bound to be problematic. Beißwenger shows how these problems are dealt with in the practice of interaction, which can draw on commonalities in the physical computer workspace (keyboard, mouse) and the chatroom interface on the screen to anchor deictic expressions. Christian Mair and Stefan Pfänder study the ways in which specific social groups use their real-world vernacular and multilingual resources in

Integrating the perspectives on language and space

15

order to “perform” online identities. Mair and Pfänder observe that digital communication technologies are instrumental in the current global spread of ex-colonial languages such as Cyber-Jamaican or Cyber-Peruvian. On the basis of web forum data, Mair and Pfänder show that these new digital vernaculars challenge conventional assignations of prestige and stigma familiar from face-to-face interaction, and that it is necessary to acknowledge them as medium-specific vernacular styles. Susanne Uhmann’s study analyzes the use of spatial deictics by surgeons who perform laparoscopic surgery. Deictics such as here and there thus refer to coordinates in a space that the interacting parties can only see on a computer screen. Uhmann identifies three important aspects that characterize use of deictics in such a mediated space: First, surgeons can only use their laparoscopic instruments, not their hands, for pointing. Second, the manoeuvrability of these instruments is highly restricted. Third, all team members can observe the instruments and their actions on a monitor. In her commentary, Monica Heller discusses the social practice of constructing space in mediated contexts as a phenomenon that can be conventionalized to different degrees. 3.6

Typology and spatial reasoning

Gisela Fehrmann examines the use of space in several signed languages of the world, outlining their functions and constructing a typology of space usage in signed languages. The use of space in signed languages includes for instance the association of a sign with a point in the signing space, a referential index. Another important aspect of the use of space is inherent in the concepts of reversed space and mirrored space, which are alternative strategies to engage with the deictic signs used by an interlocutor. By discussing these strategies, Fehrmann shows that signers are not merely representing a pre-existing space, but that they are actually making space. Jürgen Bohnemeyer and Randi Tucker discuss the hypothesis that speakers of languages in which shape-based meronymy is a prevalent resource for the expression of spatial relations may show a bias against the use of relative frames of reference. Data from Yucatec Maya is brought to bear on this hypothesis. In Yucatec Maya, human body meronyms (foot, head, arm, etc.) are routinely used to express spatial relations between non-human objects. Bohnemeyer and Tucker offer experimental results that speakers of Yucatec Maya indeed show a preference for intrinsic frames of reference over relative frames of reference.

16

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

Alan Cienki examines the relation of language and space with regard to co-speech gesture. He offers an overview of research on gesture that takes into account the spatial dimensions of gestures, their possible connections to grammar, and the consequent cognitive implications. In doing so, Cienki distinguishes pointing gestures and referring gestures, which in turn may either represent an object, trace an imagined outline of an object, or re-enact a manual movement. In addition, there are also discourse-related gestures that function as metanarrative deictic elements. Cienki concedes that current research does not support a clear one-to-one correspondence between co-verbal gesture and grammar, but the evidence does point to substantial systematicity in gesture use. In the final commentary, Holger Diessel connects the contributions of this section to the typological debate surrounding relative and absolute frames of references in the world’s languages, showing how the grammatical class of demonstratives can shed new light on this issue.

References Anderson, Benedikt 2006: Imagined Communities (2nd enlarged edition). London: Verso. Auer, Peter 2005: The construction of linguistic borders and the linguistic construction of borders. In: Markku Filppula, Juhani Klemola, Marjatta Palander & Esa Penttilä (eds.), Dialects Across Borders. Selected Papers from the 11th International Conference on Methods in Dialectology (Methods XI), Joensuu, August 2002, 3–30. Amsterdam: Benjamins. Bauman, Zygmunt 1998: Globalization: The Human Consequences. New York: Columbia University Press. Bennardo, Giovanni 2009: Language, Space, and Social Relationships. Cambridge: University Press. Bühler, Karl 1934: Sprachtheorie. Jena: Gustav Fischer. Cresswell, Tim 2004: Place. A Short Introduction. Malden, MA: Blackwell. Dipper, Christof & Ute Schneider (eds.) 2006: Kartenwelten – Der Raum und seine Repräsentation in der Neuzeit. Darmstadt: Primus. Enfield, Nick 2003: Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language 79(1): 82–117. Fillmore, Charles 1975/1971 [1997]: Lectures on Deixis. Stanford: CSLI Publications. (originally distributed as: Fillmore 1975/1971: Santa Cruz Lectures on Deixis. Indiana University Linguistics Club). Goffman, Erving 1963: Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press of Glencoe. Günzel, Stefan 2006: Einleitung zu Teil II/Phänomenologie der Räumlichkeit. In: Jörg Dunne & Stephan Günzel (eds.), Raumtheorie: Grundlagen aus Philosophie und Kulturwissenschaften, 105–128. Frankfurt am Main: Suhrkamp. Habel, Christopher & Christiane von Stutterheim (eds.) 2000: Räumliche Konzepte und sprachliche Strukturen. Tübingen: Niemeyer.

Integrating the perspectives on language and space

17

Hanks, William 1990: Referential Practice: Language and Lived Space among the Maya. Chicago, Ill.: The Univ. of Chicago Press. Hanks, William 2004: Explorations in the deictic field. Current Anthropology 46(2): 191–220. Jakobson, Roman 1957 (1971): Shifters, verbal categories and the Russian verb. In: Selected Writings, Vol. II, 130–142. The Hague: Mouton. Irvine, Judith 2001: Style as distinctiveness: The culture and ideology of linguistic differentiation. In: Penelope Eckert & John Rickford (eds.), Stylistic Variation in Language, 21–43. Cambridge: Cambridge University Press. Jarvella, Robert J. & Wolfgang Klein (eds.) 1982: Speech, Place and Action. Chichester: Wiley and Sons Ltd. Johnstone, Barbara, Jennifer Andrus & Andrew E. Denielson 2006: Mobility, indexicality, and the enregisterment of ‘Pittsburghese’. Journal of English Linguistics 34(2): 77–104. Kendon, Adam 1976 [1990]: The F-formation system: spatial-orientational reelations in face to face interaction. Man Environment Systems 1976 (6): 291–296. [Expanded version in Kendon, Adam. 1990. Conducting Interaction. Cambridge: CUP]. Labov, William 1972: Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Levinson, Stephen C. 2003: Space in Language and Cognition. Cambridge: Cambridge University Press. Merleau-Ponty, Maurice 1945: Phénoménologie de la perception. Paris: Gallimard. Meyerhoff, Miriaj & Niedzielski, Nancy 2003: The globalisation of vernacular variation. Journal of Sociolinguistics 7(4): 534–555. Niedzielski, Nancy A. & Dennis R. Preston 2000: Folk Linguistics. Berlin: Mouton de Gruyter. Preston, Dennis R. 1989: Perceptual Dialectology: Nonlinguists’ Views of Areal Linguistics. Dordrecht, Holland /Providence, RI: Foris. Quist, Pia 2010: Untying the language-body-place connection: A study on linguistic variation and social style in a Copenhagen community of practice. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space. An International Handbook of Linguistic Variation. Vol. 1: Theories and Methods, 632–648. Berlin/New York: Mouton de Gruyter. Rampton, Ben 1995: Crossing. London: Longman. Stukenbrock, Anja 2005: Sprachnationalismus. Sprachreflexion als Medium kollektiver Identitätsstiftung in Deutschland (1617–1945). Berlin/New York: de Gruyter. WALS = World Atlas of Linguistic Structures (Online) http://wals.info/index. Weissenborn, Jürgen & Wolfgang Klein (eds.) 1982: Here and There. Cross-linguistics studies on Deixis and Demonstration. Amsterdam: Benjamins. Von Weizsäcker, Carl Friedrich 1986: Aufbau der Physik (2nd edition). München/ Wien: Hanser.

18

Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi

Section 1: Geography and variation across languages

Disentangling geography from genealogy

21

Michael Cysouw

Disentangling geography from genealogy

1.

Introduction

There are various reasons why two languages might be typologically similar. Basically, there are four different possible causes of similarity that can be classified into two groups of two, as shown in (1). (1) Languages can be typologically similar because of: A. Historical factors, being either: i. Genealogical descent (“vertical transfer”) ii. Borrowing (“horizontal transfer”) B. A-historical factors, being either: i. Inherent characteristics of human language (“universals”) ii. Coincidence (“chance”) One of the major challenges for current typological-comparative linguistics is to find methods to disentangle which of these reasons apply in any given situation of typological similarity. Traditionally, there has been great interest in developing methods to separate historical factors from a-historical factors through various kinds of sampling. The basic idea behind such methods is that it is possible to control for the influence of historical factors by carefully selecting languages across known genealogical and areal groupings. The remaining question then is how to distinguish universals from chance. Both questions – how to remove historical influences from the data, and how to subsequently distinguish universals of human language from chance effects – are much-debated questions in the field of linguistic typology (cf. Cysouw 2005 for a survey), and much more could be said about these topics. Yet, the current paper will not concentrate on such a-historical characteristics of human language, but focus on historical factors that result in typological similarity. When investigating historical factors leading to typological similarity, it might seem as if there are likewise two questions to be dealt with. First, how to factor out a-historical factors, and, second, how to distinguish similarities caused by genealogical descent from similarities caused by borrowing. However, on closer inspection the first question turns out to be easily answerable.

22

Michael Cysouw

A-historical factors should apply equally to all languages, so their effects should be statistically equal for all languages. The influence of universals and chance thus amounts to a constant factor in the diversity of languages, which can simply be ignored when investigating historical factors. Still, there is a potential problem when including many different characteristics in the comparison (as will be the case in this paper). Ideally, to investigate historical factors throughout many characteristics, these characteristics should be independent of each other. A group of, say, ten characteristics that are all definitionally similar in a collection of, say, hundred characteristics in total will introduce an a-historical bias into the comparison, favoring the linguistic similarities as found in the cluster of those ten definitionally similar characteristics. The most glaring of such dependencies will have to be removed (e.g. by weighting the characteristics). The remaining question, which will be the main topic of this paper, is how to disentangle typological similarities caused by genealogical descent from similarities caused by borrowing. In this paper I will not seek to settle this question for individual cases of shared characteristics between two specific languages (e.g. why do French and German have no distance contrast in demonstratives?; see Diessel 2008; Cysouw 2011), because individual historical developments cannot be predicted by a general theory of human language. Specific historical events can only be reconstructed by an in-depth investigation of the actual history of a specific situation. However, I propose that the influence of borrowing visà-vis genealogical descent can be investigated in the aggregate (cf. Nerbonne and Siedle 2005; Nerbonne 2009 on the notion “aggregate”). To investigate the relationship between typological structure, genealogical descent, and borrowing, I will use data from the World Atlas of Language Structures (WALS, Haspelmath et al. 2005). This resource provides information about typological structure and genealogical descent, but not about possible contact or the probability of borrowing. To approach the probability of borrowing, I will use the present-day geographical distribution of languages, assuming that the probability of borrowing is inversely correlated with geographical distance. Specifically, geographically close languages will have a higher probability of contact, and likewise a higher probability of borrowing.

2.

The Eurasian data set

As a concrete example to discuss the current proposals I will use a dataset, drawn from WALS, featuring a selection of Eurasian languages. The Eurasian macro-area is chosen because most languages and their approximate geographical location will be familiar to most readers. WALS includes data

Disentangling geography from genealogy

23

on 391 different Eurasian languages, but far from all are covered in any detail (missing data is a general problem when dealing with WALS quantitatively; cf. Cysouw 2008). To obtain a data set with sufficient coverage, only languages that appear in at least 70 WALS maps are included here, resulting in a sample of 32 Eurasian languages (see Appendix A). Given the data from WALS, I will define a notion of pairwise structural similarity. There are various different aspects that can be included in such a definition of similarity, only a selection of which I will be using here. Also note that in practice I will define a notion of distance, which is of course just a trivial transformation of any notion of similarity. To define a notion of (dis)similarity, the following principles can be used (see Appendix B for the details): x

x

x

x

x

Basically, the distance between two languages is related to the number of characteristics that are different between the two languages. Because of the many missing data points, this value has to be normalized to the number of comparisons made for each language pairing, i.e. the sum of the number of similarities and the number of differences (cf. the “relativer Identitätswert” [RIW] presented in Goebl 1984). Weightings can be used to balance the impact of the characteristics, making some characteristics more important than others. This mechanism will be used here to remove some of the most glaring definitional redundancies in the WALS data. Further, similarities are not necessarily all equal. Languages that share a rare characteristic can be seen as more similar than two languages that share a common characteristic (because the sharing of a rare characteristic is a more telling similarity, cf. the “gewichteter Identitätswert” [GIW] in Goebl 1984). Finally, differences are not necessarily all equally different. For example, a language A with a small vowel inventory is less different from a language B with an average vowel inventory than from a language C with a large vowel inventory. Both the pairing (A, B) and (A, C) are different, but to different degrees. Such specification of internal structure of WALS characteristics will not be used here, because it is far from obvious how such specification of differences should be determined, and how they should be combined with specifications of similarities. Therefore, exploring this issue is reserved for another occasion.

For this paper I will start with the basic fraction of the number of differences divided by the number of comparisons made as a measure of dissimilarity be-

24

Michael Cysouw

tween two languages (this yields Goebl’s “relativer Identitätswert”). In addition, each feature (i.e. each ‘map’ in WALS) can be weighted to remove definitional redundancies in the WALS data. This is necessary because, on closer inspection, the features included in WALS are not independent of each other. The relationship between the various features in WALS turns out to be a highly complex topic, with various overt and covert dependencies between them (cf. Cysouw 2008). For the sake of this paper, I adopted the following solution to remove the most glaring redundancies. I grouped the features into sets of (definitionally) related ones (see Appendix B), and every feature in such a set is weighted by the inverse of the number of features in the set. For example, the WALS features 30, 31, 32, and 44 all deal with gender marking, and without correction, two languages without gender marking will be counted as being similar four times on all four features, though it is the same underlying similarity that is counted each time. To correct for this implicit weighting, each of the features will be explicitly weighted as counting only 1/4th (because there are four features in the set). Such weighting could also be used to emphasize typologically stable features. Further, similarities between two languages can be weighted. Following Goebl’s (1984) basic insight that sharing rare features is more telling than sharing common features, a weighting of similarities can be introduced. Such a weighting is specified for each value in each feature. So, for example, there is a difference when two languages both have tone, and when they both lack tone. Goebl proposed to weight each similarity by the fraction of occurrences in the sample. For example, in WALS there are 307 out of 527 languages that do not have tone (Maddieson 2005b). For two languages that both do not have tone, instead of counting one similarity, Goebl proposed to count 0.417 = 1 – (307/527) similarities. Languages that share a complex tone system, which is much rarer, are assigned a higher similarity of 0.833 = 1 – (88/527). Another variant of this principle would be to interpret the frequency of a characteristic as typological information, and the rarer a characteristic, the more informative it is. From an information-theoretic perspective one would calculate the similarity for not having tone as 0.235 = –log(307/527) and for having complex tone as 0.777 = –log(88/527). In general, the weights based on the logarithm will give similar, but slightly more extreme weights compared to the weights based on the fraction, especially for characteristics that occur in less than 15 % of all languages in WALS. Based on these different possible notions of dissimilarity, various distance matrices were compiled for the 32 Eurasian languages selected from WALS. These matrices represent different ways to define pairwise aggregate dissimilarities. It turns out that the various ways to define typological dissimi-

Disentangling geography from genealogy

25

larity only differ in detail, and all matrices are strongly correlated (Pearson’s r ranges between 0.89 and 0.99; see Appendix B). Given any such notion of typological dissimilarity, the main question addressed in this paper is whether it is possible to say something about how much of this dissimilarity is caused by genealogical descent, and how much by geographical proximity.

3.

Genealogy

Two languages that are both descendants from one and the same proto-language will share typological characteristics that have not changed since they split from their last common predecessor. Given that changes will accumulate over time, it is to be expected that closely related languages share more similarities than languages that separated earlier. Such a trend is clearly visible in typological data, and this observation has even led to the proposal that typological profiles might be used for the reconstruction of historical descent (cf. Nichols 1992; Dunn et al. 2005). Such an approach is of course only viable when the influence of genealogical descent on the typological profile is stronger than the influence of any subsequent areal convergence. The impact of genealogical descent is also clearly visible in the current data selection, as shown in Figure 1. Taking the genealogical classifications as specified in WALS, I have separated all language pairings into three groups, being either (1) of the same genus, (2) different genus within the same family, or (3) non-related given currently accepted views (see Appendix A for a survey of the genealogical classification of the current selection of languages). The same-genus-pairings involve only Germanic, Romance and Slavic languages. Pairs of languages from the same family, but not from the same genus include pairings from Altaic (which does not include Korean and Japanese according to WALS), Indo-European, Nakh-Dagestanian and Uralic. To quantitatively assess the strength of the correlation between typological distance and genealogical distance, I performed Mantel tests (Mantel 1967) to correlate the various typological distance measures with genealogical distance. Genealogical distance was simply defined as an approximate linear scale, being ‘1’ when two languages were from the same genus, as ‘2’ when they came from the same family, but from different genera, and as ‘3’ otherwise (note that these numbers are of course ranks, but I do not know of a Mantel test that can deal with ranks). All Mantel tests for the different kinds of typological distance were highly significant, with only slight differences in the statistics, as shown in Table 1. Logarithmic value weightings (to favor rare similarities) combined with the feature weightings (to reduce the impact

26

Michael Cysouw

Fig. 1. Languages from the same genus are on average more similar typologically than languages from different genera, but from the same family, which are in turn more similar than unrelated languages.

of groups of definitely similar features) yielded the highest correlation scores. For the remainder of this paper I will restrict attention to this typological distance measurement, as it seems to be the one most closely matching genealogical descent.

No value weighting Value weighting by fraction Value weighting by logarithm

No feature weighting

Feature weighting

0.574 0.634 0.616

0.603 0.616 0.642

Table 1. Correlations between genealogical distance and different definitions of typological distance. All correlations are highly significant at p < 0.001 (according to a Mantel test), though there are slight differences in the strength of the correlation scores.

4.

Geographical proximity

In line with Tobler’s first law of geography that “Everything is related to everything else, but near things are more related than distant things” (Tobler 1970: 237), typological (dis)similarities are strongly related to geographical distance. Geographically close languages are in general typologically similar, while geographically distant languages are generally typological different.

Disentangling geography from genealogy

27

Fig. 2. There is a strong correlation between the geographical distance between two languages and their typological dissimilarity. Every observation in this figure is a language pairing, shown here as a box plot grouping language pairings depending on their geographical distance in bins of 1000 kilometers.

This correlation is immediately obvious for the current selection of 32 Eurasian languages. Shown in Figure 2 are all 496 language pairings (i.e. 32 × 31/2 pairings) grouped by geographical distance. Note that as a measure of geographical distance I have here simply taken the linear distance on a perfect sphere between two coordinates on the surface (see Appendix A for the coordinates used). To aid visual interpretation, all language pairings are grouped into ›bins‹ of thousand kilometers, i.e. all language pairings with a distance between 0 and 1000 form one group, and all language with a distance between 1000 and 2000 form another group, and so on. For each of these groups a box is shown in Figure 2, the medians of which show almost a linear relationship between geographic distance and typological distance, surprisingly in this case without any sign of the expected flattening at extreme geographical distances (cf. Nerbonne, this volume, on an indepth discussion of the nature of this relationship). Using a Mantel test to evaluate this correlation gives again a highly significant result (r = 0.616, p < 0.001). A visually more impressive way to show the strength of the correlation between geographical distance and typological distance is shown in Figure 3. This figure plots the first two dimensions of a non-metric multidimensional scaling (MDS) of the typological distances. Multidimensional scaling is a method to mathematically derive abstract dimensions of variation from a matrix of distances. The first dimension is defined such that as much of the

28

Michael Cysouw

Fig. 3. The first two dimensions of non-metric multidimensional scaling of language distances based on WALS data for 32 Eurasian languages (the precise locations of the names in the MDS has been slightly tweaked manually to prevent overlapping). The first dimension (shown horizontally) strongly correlates with geographical longitude. The second dimension (shown vertically) separates the Caucasian languages with Burushaski, Basque and Chukchi from the rest.

variation as possible is captured. Subsequent dimensions account for the leftover variation in diminishing order. In the current example the first dimension accounts for 67 % of the variance. In Figure 3 this first dimension is shown horizontally and shows an astonishing overlap with geographical longitude. The west European languages are shown to the left, the far Eastern and Siberian language (except Chukchi) to the right. The second dimension, shown vertically, captures another 12 % of the variance. This dimension basically separates the outliers from the Eurasian mainstream, namely the Caucasian languages together with Burushaski, Basque and Chukchi. The central result of the MDS analysis, though, is that the most important dimension (the first dimension) is almost perfectly related to geographical longitude. Correlating this first dimension (i.e. the left to right order of the languages in Figure 3) with longitude reaches extremely high significance (r = 0.88, p < 0.001). This implies that the longitudinal location of the current set of languages can be predicted from the value on the first dimension of the MDS, which is in itself just a derivative of typological similarity. It is thus possible to predict typology from geography (or vice versa). A linear in-

Disentangling geography from genealogy

29

terpolation of the first MDS dimension with longitude results in the following formula: Longitude (in degrees) = 57.8 + 199 × dimension 1 of the MDS. The constant 57.8 in this formula is the average longitude of the sampled languages. This formula can be used to predict geographical location. For example, Spanish has a value of –0.31 on the first MDS dimension, so the formula predicts a longitude of –3.89 degrees (= 57.8 – 199 × 0.31). This perfectly matches the data from WALS, which situates Spanish geographically in Madrid at a longitude of –4 degrees. So, when investigating the typological distances through multidimensional scaling, the most important dimension correlates strongly with geographical longitude, indicating that geographical longitude is one of the central factors determining typological variation.

5.

Separating genealogy from geography

Typological distance between languages is strongly correlated with genealogical distance, but also with geographical distance. Further, genealogically related languages are in general located geographically close to each other, so one would also expect a correlation between genealogy and geography. And indeed, for the current test case this correlation is statistically significant (r = 0.367, p < 0.001), though the correlation is clearly less strong than the correlations discussed previously for genealogy and geography. Yet, we are left with a tangle of significant correlations between typology and geography, typology and genealogy, and geography and genealogy. There are various statistical approaches that might help to tear apart the interaction between these factors. In general, the issue could be approached using basic regression modeling, were it not for the fact that we are dealing with distances here. Different from the normal situation in which one would like to use regression modeling, the ‘observations’ here are pairwise measures of distance for which it is not clear whether they can simply be treated as ‘atomic’ observations. The problem is that dissimilarity measures are not independent observations, because each language is compared with all others, so duplication of information is introduced. However, we can of course just pretend that the dissimilarities are real observations and perform a linear regression. The results are shown in Table 2. All factors turn out to be highly significant, so it still remains unclear which factor is more important for the explanation of typological distances (and the significance values are probably meaningless anyway, for the reasons discussed above).

30

Intercept Genealogy Geography

Michael Cysouw Estimate

Std. Error

t-value

Pr(>|t|)

0.3481 0.1316 1.822·10–5

1.505·10–2

23.14 15.31 14.03

< 0.001 *** < 0.001 *** < 0.001 ***

8.594·10–3 1.299·10–6

Table 2. Linear regression of typological distances to genealogical and geographical distance (r2 = 0.58). All factors are highly significant, but this is probably meaningless because the ‘observations’ are distances.

Assessing significance of correlations between distances is normally done using the Mantel test (as we did in the previous sections). There is a variant of the Mantel test that can assess the significance of the correlation between two distances matrices while keeping a third distance matrix constant. Such a test is called a partial Mantel test (Legendre 2000). This approach seems to be ideal to address the current problem, but unfortunately it does not decide on the question which of the correlations is most important. The correlation between typology with genealogy, while keeping geography constant, is significant (partial Mantel r = 0.568, p < 0.001). However, the correlation between typology with geography, while keeping genealogy constant, is similarly significant (partial Mantel r = 0.534, p < 0.001). There are other recent developments in statistical methods to deal with this problem. Specifically, there is a proposal for multivariate regression modeling with distance matrices as dependent variable and factors or continuous variables as independent variables (Zapala and Schork 2006). In this approach, we can use the typological distances as dependent variable, but we have to use separate categorical predictor variables for family and genus, and continuous predictor variables for longitude and latitude. From the results, as shown in Table 3, it seems that the genealogical factors are the strongest factors and the only significant ones. The results thus indicate that the basic correlation is between typology and genealogy, and that the correlation with geography is only a secondary effect. However, note that the results differ rather radically when the order of the predictor variables is changed, which casts doubt on the proper interpretation of these results.

31

Disentangling geography from genealogy Df

F Model

R2

Pr(>F)

Family

14

4.8814

0.69641

0.001 ***

Genus

13

1.9653

0.26035

0.011 *

Latitude

1

1.2761

0.01300

0.237

Longitude

1

0.9670

0.00985

0.512

Residuals

2

0.02038

Total

31

Table 3. Regression model using typological distance as dependent variable. Genealogical factors (Family and Genus) are the only significant factors. Geographical factors (Latitude and Longitude) are not significant.

Finally, it is highly informative to look at the typological residuals after regression with genealogy or geography. I will use here statistics calculated in a similar fashion as the linear regression reported in Table 2, though I shall ignore the significance values. The basic idea is to remove the impact of genealogical relatedness from the typological distances, and look at the residuals of the typological distances (and do so likewise for geographical distances). The interpretation of such residuals seems to be linguistically interesting. First, if there is a correlation between genealogical distance and typological distance, then it should be interesting to spot language pairings that are more similar typologically than expected from genealogy. Such excess similarity might be indicative of areal convergence. Second, given that there is a correlation between geographical distance and typological distance, I will look for language pairings that are more similar than expected given their geographical distance. Such language pairings – which are, in a sense, ‘too far away’ for their typological distance – might be indicative of (relatively) recent population movement. So I linearly regressed typological distance against genealogical distance, and then ordered all language pairings according to their residuals. The languages pairings with the lowest residual typological distance are all geographically close, and many pairings indeed seem to be readily interpretable as cases of language contact. High on the list are the pairings Korean-Japanese, Khalkha-Japanese and Khalkha-Korean. These are languages that are known to have been in close contact over centuries, up to the point that they are sometimes claimed to be genealogically related. Also, the classic European Sprachbünde are represented in the top of the list: the Baltic Sprachbund (RussianFinnish, Latvian-Finnish), the Balkan Sprachbund (Bulgarian-Greek, Albanian-Greek), and the Charlemagne Sprachbund (German-French). Also on top of the list are the pairings Armenian-Georgian and Armenian-Turkish, which

32

Michael Cysouw

are also clear examples of language contact. The remaining top pairings are less clear: Burushaski-Georgian, Burushaski-Lezgian and Nenets-Evenki. Whether the observed surplus of typological similarity in these cases is the result of contact, or caused by other factors, is unclear to me. The waters are muddier when we attempt to remove geographical influence from typological distance, but there still is some indication of an influence of population movement. When looking at the residuals of typological distance after regression by geographical distance, the language pairing with the lowest residual typological distance is Khalkha-Turkish, with Evenki-Turkish following a bit further down. Turkish is clearly an example of a language subject to a relatively recent population movement over a long distance, leading to a situation in which the Turkish language is still relatively similar typologically to its Altaic kin, though geographically it is too distant. The other language pairings in the top twenty of ‘too similar’ languages relative to their geographical distance are all pairings of Indo-European languages.

5.

Conclusion

Investigating the typological diversity of the world’s languages has been an active field of research over the last few decades. However, the basic premise has been that this kind of research is worthwhile because it will help unravel universal properties of human language. The strong correlations with historical factors, both genealogical and geographical, as discussed in this paper cast doubt on the allegedly important role of universal properties on the currently observable typological diversity. The entanglement between typological diversity and genealogical relationship has been acknowledged for a long time in the literature. In contrast, the similarly intricate entanglement between typological diversity and geographical proximity has not sparked similarly in-depth investigations. The existence of correlations between genealogy and geography should not merely be seen as a nuisance factor in the investigation of universal properties of human languages – it can also be taken as a possible starting point to unravel the dynamics of typological change and language history. Instead of building samples that from the start prevent genealogical or geographical bias, I think we should deliberately collect data from samples with such ‘biases’. Only by including many related languages and/or geographically close languages will it be possible to investigate the impact of genealogy and geography on typological diversity. And any correlations attested can then be accounted for statistically. As argued in this paper, it does even seem to be possible to infer situations of contact or population movement from ‘biased’ typological samples.

33

Disentangling geography from genealogy

Appendices Appendix A: Languages selected from WALS Name Abkhaz

Longitude 41

Latitude 43.08

Genus

Family Northwest Caucasian Ainu Indo-European Indo-European Basque Indo-European Burushaski ChukotkoKamchatkan

52 56 62 48 42 52 39 47 42.17

Northwest Caucasian Ainu Albanian Armenian Basque Slavic Burushaski Northern ChukotkoKamchatkan Germanic Tungusic Finnic Romance Kartvelian Germanic Greek Ugric Avar-Andic-Tsezic

Ainu Albanian Armenian Basque Bulgarian Burushaski Chukchi

143 20 45 -3 25 74.5 187

43 41 40 43 42.5 36.5 67

English Evenki Finnish French Georgian German Greek Hungarian Hunzib

0 125 25 2 44 10 22 20 46.25

Ingush

45.08

43.17

Nakh

Irish Japanese Ket Khalkha Korean Latvian Lezgian

–8 140 87 105 128 24 47.83

53 37 64 47 37.5 57 41.67

Celtic Japanese Yeniseian Mongolic Korean Baltic Lezgic

Nenets Nivkh Polish Russian Spanish Turkish Yukaghir

72 142 20 38 –4 35 150.83

69 53.33 52 56 40 39 65.75

Samoyedic Nivkh Slavic Slavic Romance Turkic Yukaghir

Indo-European Altaic Uralic Indo-European Kartvelian Indo-European Indo-European Uralic NakhDaghestanian NakhDaghestanian Indo-European Japanese Yeniseian Altaic Korean Indo-European NakhDaghestanian Uralic Nivkh Indo-European Indo-European Indo-European Altaic Yukaghir

34

Michael Cysouw

Appendix B: Defining typological dissimilarity The basic formula to establish the typological dissimilarity (or: distance) between two languages L1 and L2 is based on the number of similar characteristics s and the number of different characteristics d. When the basic “languages by characteristics” data matrix is complete for n characteristics, then of course s = n – d, but in typological data there will probably always be many missing data points, so it will be necessary to establish s and d independently. Given s and d, the basic unweighted dissimilarity D between L1 and L2 is defined as: Dunweighted(L1, L2) = d / (d + s) = 1 – s / (d + s) = 1 – RIW (RIW: Goebl’s Relativer Identitätswert) To introduce a stronger influence of rare characteristics, a weighting function v(s) for similar characteristics can be included. Instead of simply counting each similarity as ‘1’, each similarity si is counted as v(si), and the formula then includes a summation over all these v(si): Dvalue-weighted(L1, L2) = d / (d + ∑ v(si)) The basic idea is that this function v should rate those similarities higher that are only rarely attested. Goebl originally proposed to take for each value si the fraction of occurrence psi in the data, and then define v(si) = 1 – psi. This is referred to in the paper as ‘weighting by fraction’ and the resulting typological distance between two languages is then identical to 1 – GIW (Goebl’s Gewichteter Identitätswert). Another possible approach is to define v(si) = – log(psi), which can be seen as a measure of information content: the rarer the shared characteristic, the more informative it is for the similarity between the languages. This weighting is referred to in the paper as “weighting by logarithm”. A similar principle of weighting can also be applied to differences. Instead of counting each difference as ‘1’ it is possible to explicitly specify the precise value for each of the various differences. For example, for consonant inventories (Maddieson 2005a) there is a much larger difference between language pairings with a small versus large consonant inventory than between language pairings with a small versus average consonant inventory. There are two practical problems preventing me from adding such a weighting here. First, it is unclear how such weights should be determined, other than by ad-

Disentangling geography from genealogy

35

ding intuitively specified numerical values. Second, it is unclear how such a specification of differences interacts with specification of similarities. Specifically, language pairings with many differences might in special cases become more similar than language pairings with many similarities. However, these problems should be surmountable given more research. Further, the features as a whole can be weighted, so instead of counting any similarity or difference in a feature F (i.e. in a specific ‘map’ in WALS) equally as 1, a function w(F) can be defined to selectively change the impact of complete features. The resulting typological distance will then be defined as: Dfeature+value-weighted(L1, L2) = ∑ w(dj) / (∑ w(dj) + ∑ v(si) · w(si)) This feature-weighting function has been used in the current paper to remove some obvious definitional dependencies between features in WALS. Specifically, the features 3, 25, 95, 96, and 97 have been weighted as zero (i.e. they have been removed from the data) because they are combinations of other features in WALS. Similarly, the features 139, 140, 141, and 142 have been weighted as zero because the set of languages discussed in these features is incompatible with the other features. Moreover, the following groups of definitionally related features have been weighted by the inverse of the number of features in the group (i.e. a feature in a group of four is weighted as 1/4): – – – – – – – – – – – – – – – –

14, 15, 16 (stress system) 30, 31, 32, 44 (gender marking) 37, 38 (articles) 39, 40 (clusivity, also known as “inclusive/exclusive distnctions”) 49, 50 (case marking) 26, 51, 69 (affixation) 81, 82, 83, 85, 86, 93 (sentence word order) 87, 88, 89, 91 (nominal word order) 84, 90, 94 (complex sentence order) 77, 78 (evidentiality) 98, 99 (alignment) 40, 29, 100, 101, 102, 103 (verbal person inflection) 113, 114 (negation) 122, 123 (relativization) 125, 126, 127, 128 (clause conjunction) 132, 133, 134, 135 (color terms)

36

Michael Cysouw

The result of all these weightings are six different measures of typological distance. These measures are all strongly correlated, as shown in Table 4, so the effects of using one over the others are minimal. When the paper simply refers to ‘the’ typological distance, this will imply the distance determined by weighting features and weighting the values by logarithm. Table 4. Different ways to measure typological dissimilarity are highly correlated (Pearson r values for all measurement pairings).

A) No weighting B) Only value weighting (by logarithm) C) Only value weighting (by fraction) D) Only feature weighting E) Weighting feature + value (by logarithm) F) Weighting feature + value (by fraction)

A

B

C

D

E

F

1.000 0.979

0.979 1.000

0.985 0.994

0.892 0.910

0.943 0.969

0.984 0.988

0.985

0.994

1.000

0.912

0.966

0.992

0.892 0.943

0.910 0.969

0.912 0.966

1.000 0.942

0.942 1.000

0.903 0.966

0.984

0.988

0.992

0.903

0.966

1.000

References Cysouw, Michael 2005: Quantitative methods in typology. In: Gabriel Altmann, Reinhard Köhler & Rajmund Piotrowski (eds.), Quantitative Linguistics: An International Handbook, 554–578. Berlin/New York: Mouton de Gruyter. Cysouw, Michael 2008: Generalizing scales. In: Marc Richards & Andrej Malchukov (eds.), Scales, 379–396. Leipzig: Institut für Linguistik, Universität Leipzig. Cysouw, Michael 2011: Quantitative explorations of the world-wide distribution of rare characteristics, or: the exceptionality of north-western European languages. In: Horst Simon & Heike Wiese (eds.), Expecting the Unexpected, 411–431. Berlin/ New York: Mouton de Gruyter. Diessel, Holger 2008: Distance constrasts in demonstratives. In: Martin Haspelmath, Matthew M. Dryer, David Gil & Bernard Comrie (eds.), The World Atlas of Language Structures Online, Chapter 41. Munich: Max Planck Digital Library. Dunn, Michael, Angela Terrill, Ger Reesink, Robert A. Foley & Steve C. Levinson 2005: Structural phylogenetics and the reconstruction of ancient language history. Science 309(5743): 2072–2075. Goebl, Hans 1984: Dialektometrische Studien: anhand italoromanischer, rätoromanischer und galloromanischer Sprachmaterialien aus AIS und AFL. (Beihefte zur Zeitschrift für Romanische Philologie 191). Tübingen: Niemeyer. Legendre, Pierre 2000: Comparison of permutation methods for the partial correlation and partial Mantel tests. Journal of Statistical Computation and Simulation 67(1): 37–73.

Disentangling geography from genealogy

37

Maddieson, Ian 2005a: Consonant inventories. In: Martin Haspelmath, Matthew S. Dryer, David Gil & Bernard Comrie (eds.), World Atlas of Language Structures, 10–13. Oxford: Oxford University Press. Maddieson, Ian 2005b: Tone. In: Martin Haspelmath, Matthew S. Dryer, David Gil & Bernard Comrie (eds.), World Atlas of Language Structures, 58–61. Oxford: Oxford University Press. Mantel, Nathan 1967: The detection of disease clustering and a generalized regression approach. Cancer Research 27(2): 209–220. Haspelmath, Martin, Matthew S. Dryer, Bernard Comrie & David Gil (eds.) 2005: The World Atlas of Language Structures. Oxford: Oxford University Press. Nerbonne, John 2009: Data-driven dialectology. Language and Linguistics Compass 3(1): 175–198. Nerbonne, John & Christine Siedle 2005: Dialektklassifikation auf der Grundlage Aggregierter Ausspracheunterschiede. Zeitschrift für Dialektologie und Linguistik 72(2): 129–147. Nichols, Johanna 1992: Linguistic Diversity in Space and Time. Chicago: University of Chicago Press. Tobler, Waldo R. 1970: A computer movie simulating urban growth in the Detroit region. Economic Geography 46: 234–240. Zapala, Matthew A. & Nicholas J. Schork 2006: Multivariate regression analysis of distance matrices for testing associations between gene expression patterns and related variables. Proceedings of the National Academy of Sciences of the United States of America 103(51): 19430–19435.

38

Johanna Nichols

Johanna Nichols

The vertical archipelago: Adding the third dimension to linguistic geography

1.

Introduction

Typology, areal linguistics, and linguistic geography have had relatively little to say about the impact of space, particularly landscape, on linguistic type and language spreads. The goal of this paper is to use language spread and typology in mountain languages to illustrate what impact landscape can have on language. Aspects of the ecological and geological landscape such as vegetation, rivers, soil, length of growing season, climate, altitude, and barriers shape the directionality (literal and sociolinguistic), rate, and frequency of language spreads and constrain the sociolinguistic variables that make for greater vs. lesser complexity. Space in this approach is a matter of physical reality, and it is three-dimensional. Altitude is an important causal factor determining vegetation, stream flow, length of growing season, availability of water, and temperatures. In the survey done here, altitude figures as an independent variable, and such things as the distribution of genealogical lines, typological features, linguistic complexity, and sociolinguistic status are dependent variables which react to altitude. The patterns I describe are traditional ones and apply equally to huntergatherers and food producers and to pedestrian and transport-based interactions and spreads. For the last few centuries they have been outweighed by the sociolinguistic status unique to state and imperial languages, and they are probably all moribund now. Geography shapes language, in this case in the specific sense that altitude predicts aspects of sociolinguistics, grammatical complexity, and areality, and for sets of languages it predicts structural diversity and family tree structures. Of course, altitude is not the direct and proximate cause of these things. For instance, isolation favors (or at least does not disfavor) complexity, mountain geography favors isolation, and complexity of sound systems necessarily entails expansion along certain dimensions. Thus ejectives and uvulars can be found in mountain areas – not because harsh mountain geography deterministically causes languages to add harsh consonant series ( ! ), but because isolation favors complexity.

The vertical archipelago: Adding the third dimension to linguistic geography

39

Geography does work more or less directly on sociolinguistics. Peripheral locations like mountain highlands preclude large open networks and often entail precarious or at least dependent economies (thus, more complexity and higher individual variation), while central locations favor development of large networks, economic dominance, and expansion of speech communities by absorption of adult second language learners (thus, less complexity and less openness to individual variation). Why does sociolinguistics work this way? It is known that languages whose speech communities have absorbed significant numbers of adult second language learners have lower complexity than close sisters that have not absorbed second language learners (Trudgill 2009, 2011; Szmrecsanyi and Kortmann 2009; Szmrecsanyi 2009). Another factor may be that, in larger and more complex societies, emergent social structures take on some of the tasks of communication, identification, stance, pragmatics, etc. that are borne only by individual use of language in simpler societies, where grammatical elaboration bears these functions (I infer this possibility from work in applied complex systems theory, e.g. Yoffee 2001, 2005). The term “vertical archipelago” was coined in Andean studies to describe the geography of ancient Andean political-economic systems: ethnic groups, power structures, and market and production systems had discontinuous distributions, vertically arranged, so as to take advantage of the ecological variety in the Andes (see Murra 1956, 1985; Stanish 1992). Though the Caucasus had no comparable political organization, at least in the central and eastern Caucasus languages and economic systems had similarly discontinuous and vertically arranged distributions, and for the same ecological reasons. In what follows I will deal only with mountain areas having a central crest, like the Caucasus, the Alps, and the Andes. The sociolinguistic dynamic is very different in areas like the Andean altiplano – enclosed high plateaus with interior drainage and distinctive ecologies requiring specialized economic adaptation – and their linguistic geography is also very different. To my knowledge there are only three large altiplano areas traditionally inhabited by food producers: the Andes, Tibet, and the New Guinea central highlands. The linguistic geography I describe for central crest areas, in contrast, recurs in numerous highland areas and can be taken as the general response of languages to verticality relations.

2.

Language areas

A language area (or Sprachbund or linguistic union) is a geographical area within which unrelated or only distantly related languages have come to resemble each other markedly as a result of language contact. The classic lan-

40

Johanna Nichols

guage area is the Balkan Sprachbund, the part of the southern Balkan peninsula where Bulgarian, Macedonian, Albanian, southern Romanian, Greek, and Romany have long been in close contact and share a distinctive grammatical profile that has not been inherited from the ancestor of any of them (Sandfelt 1930; Lindstedt 2000; Joseph 2001; Vermeer 2005; and many other sources). Other well-known linguistic areas are greater India (Emeneau 1956; Masica 1976), Mesoamerica (Campbell, Kaufman & Smith-Stark 1986), and Southeast Asia (Enfield 2005). Apart from their localization, geography per se plays almost no role in the work done so far on these areas. The works just cited employ what can be called categorical Sprachbund theory: the area is identified on linguistic grounds and precisely defined by finding that (usually small) set of structural properties that are present in all the languages of the area and in no adjacent or nearby languages. A semi-categorical approach is that of Masica (1976), who seeks structural features typical of the area and traces each out as far as it reaches, so that the area in question is at the center of a number of isoglosses that surround it more or less widely. Dryer (1989) describes very large areas (continent-sized or larger) with significantly different preferences for different values of typological variables. Geography plays a larger role in the predictive areality theory of Bickel & Nichols (2003, 2006): areas are determined on non-linguistic grounds, chiefly geography and secondarily archaeology and other cultural factors, and significantly different frequencies of variables are expected and sought between languages in the area and languages outside of it. Areas so defined include trans-continental ones like the Pacific Rim area (coastal and near-coastal Australasia, Asia, and North, Central, and in part South America) and the Caucasus-Himalayas enclave zone (preserving probably formerly more widespread features that were submerged in the great inner Eurasian language spreads). Donohue 2004, Donohue, Wichmann, Albu 2008 (for Oceania), and Güldemann 2010 (for Africa) are other works in this spirit. Nichols (1992, 1997) found correlations between particular geographical properties (latitude, coast, mountains, etc.) and the degree of structural and genealogical diversity in the languages of those geographies. Nettle (1999) argued that diversity is caused not by geography but only by time (it is said to decrease with duration of human habitation) but conceded that, in areas which do not conform to this claim, it is geography that predicts diversity. Adding in factors such as size of continent and ecological variety that he did not consider, geography predicts diversity in the conforming areas also. Thus some modest steps have been taken toward looking for correlations between geography and linguistic properties. The present paper is a close study of just one small region, the eastern half of the Great Caucasus range

The vertical archipelago: Adding the third dimension to linguistic geography

41

(southern Russia, northern Georgia and Azerbaijan) and the Nakh-Daghestanian language family that occupies it. I show that geography and specifically altitude correlates with linguistic structural properties and also with the sociolinguistic status of languages.

3.

The eastern Caucasus

The vertical layout of their speakers’ physical environment has profoundly influenced the lexicons, grammars, and pragmatics of the Nakh-Daghestanian languages as well as cultural phenomena such as architecture, land use, economic relations, and mythology. It has also shaped the genealogical and typological stratification of languages in the central and eastern Caucasus. In a nutshell, overall grammatical complexity of the languages is strongly influenced by altitude and by sociolinguistics, which is itself contingent on altitude. In addition, there are localized contact effects that are not particularly influenced by altitude (though their spread appears to proceed along vertical lines). The Great Caucasus range runs west by northwest to east by southeast between the Black and Caspian Seas and on a map has a linear configuration with a single main crestline. This linear geography has had some influence on the languages of the Nakh-Daghestanian (or East Caucasian) language family which occupies its eastern half. The languages evidently dispersed from a proto-homeland in the southeast of the present range, spreading along the foothills from southeast to northwest and from the foothills into the highlands. Consistent with this likely history, the family breaks down into a series of major branches each coinciding with a position along the northwest-tosoutheast line and each further subdivided for the most part north to south, i.e. vertically (see Figure 1 for the family tree). The family is very old1 and has 1

Probably about 8000 years old, one of the world’s oldest demonstrated families. The figures for cognate retention, adjusting for the rates of inter-branch borrowing that are demonstrable among the younger branches, are on this order, and the reconstructable vocabulary is consistent with an early farming culture, with terms for the earliest domesticates and none for ‘horse’, ‘wheel’, etc. The archaeological site of Chokh in highland Daghestan, about 8000 years old, is one of the earliest farming sites outside of Mesopotamia (Zohary and Hopf 2000: 221, 244–245), and there is essential cultural continuity from that site to historical Daghestanian cultures (Amirkhanov 1987). Thus the compact distribution of NakhDaghestanian languages may be due to an early farming spread. If so, this is possibly the only case in all of western Eurasia where the modern languages descend from those of the first farmers in the area. Elsewhere, whatever languages first spread with farming have been erased by the subsequent great spreads of the Indo-European, Semitic, and Turkic families.

42

Johanna Nichols

Fig. 1. The Nakh-Daghestanian family tree. The diagram shows the branches down to about the middle ages. Languages and dialects are shown underneath. Dotted lines are branches whose exact attachment are not yet certain. Figures at the right are approximate time depths (years ago). Asterisk = languages with mutually unintelligible dialects that could be considered distinct languages.

what I estimate to be about 47 daughter languages (including some “dialects” that are mutually unintelligible) in five to seven branches (depending on how two questions of higher-level classification are resolved). (For a language list and maps see Korjakov 2006; online maps: Dahl & Veselinova 2006, Gippert 1999–2003.) The major phylogenetic division within the family is that between the Nakh branch in the far west of the range and the Daghestanian branch, which comprises all the others. Some structural properties of the languages also follow a linear northwest-to-southeast distribution: for example, the number of verbs taking gender agreement is highest in the southeast and lowest in the northwest; preverbs were affixed to verb stems earliest in the east (where they are least transparent both semantically and formally) and latest or not at all in the west. But the focus of this paper is the vertical distribution of elements of language structure and sociolinguistics. The major phylogenetic division in the family coincides with an ecological divide: languages of the Nakh branch (Chechen varieties, Ingush, Batsbi) are spoken in the better-watered, less rocky central Caucasus, where travel and communication are easier and direct contact with grain-farming lowlands is

The vertical archipelago: Adding the third dimension to linguistic geography

43

generally available. Languages of the Daghestanian branch are spoken in the Russian province of Daghestan and nearby Georgia and Azerbaijan, where the climate is drier, the land steeper and rockier, towns relatively isolated, and access to fertile lowlands limited (to the north is the North Caspian desert, to the south the extremely steep southeastern face of the Great Caucasus). In the central (and western) Caucasus, speech communities are relatively large and span several to many towns and many villages. In Daghestan many speech communities are small and limited to a single town and its outlying hamlets. In the west, clan exogamy is traditional; in Daghestan, clan endogamy is possible (and, in highland communities, even favored). Cross-cutting the ecological divide and the language families, and running the entire length of the Caucasus, is a social and economic division between lowlanders and highlanders. Lowlanders inhabit the plains and low foothills, and have traditionally been engaged in grain farming and trade (major bazaars were in the lowlands, and Silk Road routes passed nearby). Highlanders inhabit the high foothills and true highlands, where growing seasons are short, crops at risk as a result, and trade and communications are primarily with the lower foothills and lowlands. The traditional economic basis of the highlands is sheepbreeding, supplemented in Daghestan by crafts (pottery, rugs, metalwork, etc.) and in the central Caucasus formerly by stonemasonry. Not only trade communications but also connections with winter pastures were vertically aligned, as markets, dormant fields for winter pasturing, and longer growing seasons were found in the lowlands. As a result there was a traditional standing pattern of asymmetrical vertical bilingualism: highlanders learned lowland languages for market communication and also because a good portion of the adult working-age male population was transhumant, spending the winter months in lowland pastures; but lowlanders had no need to learn highland languages. Because of the economic prestige and general usefulness of lowland languages, lowland isoglosses and entire languages tended to drift uphill, reducing highland languages to enclaves and eventually displacing them entirely, while highland languages and isoglosses rarely spread downhill (for this dynamic see Volkova 1974; Wixman 1980; Nichols 2005). Consequently, major inter-ethnic languages generally had a large vertical extent, ranging from lowlands to highlands, while highland languages could acquire lowland outposts only if highlanders physically moved to the lowlands. This pattern was disrupted during the 19th and 20th centuries by Russian (czarist, Soviet, and post-Soviet) highland clearances, deportations, importation of Cossack and other Russian populations, and economic disruption, so that by now most of the languages are endangered to some degree.

44

4.

Johanna Nichols

Survey

This paper surveys two subtypes of complexity: inventory size and opacity. I measure inventory size as the number of contrastive elements in a system or subsystem (e.g. consonants in the phoneme inventory, argument roles indexed on the verb; see Nichols 2009). Opacity is the number of steps, processes, mergers, syncretisms, allomorphies, etc. standing between the underlying category or form and its surface exponent (e.g. syncretic, zero, and non-segmental marking of genders are opaque, as is arbitrary semantics of genders). (Low opacity will sometimes also be called transparency below.) Both inventory size and opacity complicate learning of the language as nonnative by adult learners, and hence are linked in the literature to sociolinguistic functions such as interethnic language, written language, local ethnic language, and so on – inventory size and opacity are expected to be low for an inter-ethnic language. Only inventory size seems directly connected to the delicacy of marking of interpersonal pragmatics that is linked in the literature to population size of the speech community. The whole question is complicated by the fact that population size is often used as a proxy for social complexity and sociolinguistic function. Trudgill (2009, 2011) provides a summary of expected correlations between population size, network density, and second-language learning on the one hand and inventory size, redundancy, and opacity on the other.2 The survey reported here supports the importance of second-language learning in reducing inventory size, but only as a within-area or within-society effect. At this writing 26 languages have been surveyed for all grammatical variables or all but one. Published information, dissertations, and ongoing fieldwork will eventually make it possible to survey another 6 to 10 of the languages. Since ecological conditions and population sizes are so different between Daghestan and the central Caucasus (where the two Nakh languages in the survey are spoken), some counts below are done for both the entire 26 languages and just the 24 Daghestanian ones; a few leave out the three big inter-ethnic languages of Daghestan and nearby (for a total of 21 languages). Population sizes are from Korjakov (2006) in most cases; calculations below use not the raw population figures but their orders of magni2

Trudgill maintains that second-language learning of languages became important only after the Neolithic and therefore higher complexity was the usual earlier situation. Large-scale language spreads at least partly involving language shift are the rule in seasonal continental interiors and at high latitudes, however (Nichols 1992), so that the effects of second-language learning are likely to have existed in all early prehistory as well.

The vertical archipelago: Adding the third dimension to linguistic geography

45

tude (measured as 4 figures, 5 figures, etc.). Altitudes and coordinates are from the Fallingrain Global Gazetteer (http://www.fallingrain.com/world/).

5.

Inventory size

By the definition used here, measuring inventory size is a matter of counting up the number of phonemes, of cases, of genders, etc. The parts of the grammar surveyed here are the following, based on those used in Nichols (2009) with some family-specific additions (asterisked): Phonology: number of contrastive manners of stop and affricate articulation number of contrastive points of obstruent articulation number of contrastive short vowel qualities number of types of pharyngealization* complexity of syllable structure (based on Maddieson 2005 but more finely scaled) x x x x x

Inflectional synthesis of the verb: number of generic inflectional categories marked on the verb (tense, aspect, pluractionality, argument number, evidentiality, mirative, negation, interrogative) number of roles marked on the typical finite ditransitive verb

x

x

Classification: number of possessive classes number of target genders (Corbett 1991) in the singular number of target genders in the plural number of controller genders (Corbett 1991) number of oblique case stem extensions* number of plural stem extensions*

x x x x x x

Syntax: number of distinct alignments in finite clauses number of major word orders number of places of marking for each of the three arguments in ditransitive clauses (on head, on dependent, both, neither) number of different categories in argument agreement marking (gender, person-number, assertor involvement marking [a.k.a. conjunct/disjunct; Creissels 2008], hierarchical)

x x x

x

46

Johanna Nichols

Lexicon: inclusive/exclusive opposition in pronouns number of different derivational types in the 18-pair verb list of Nichols, Peterson & Barnes (2004) (survey not yet complete) number of different verb prefix slots (deictic, directional)* x x

x

Note that some of these characters are distinctive (e.g. phonemes, different genders) and some are redundant (e.g. oblique and plural stem extensions, which mark declension classes). The overall inventory size is simply the sum of the individual totals for all these features: the number of consonant manner distinctions, plus the number of verbal inflectional categories, plus the number of possessive classes, plus the number of alignments … and so on, except that the numbers of oblique and plural stem extensions were lumped as low, medium, and high. Proper weighting of all these figures is important but has not really been attempted here. In the survey so far, inventory size shows no correlation with population size, altitude, or longitude, no matter what the set of languages surveyed (all 26 languages, just the Daghestanian languages, all but the inter-ethnic languages, just the languages having gender). There is, however, a correlation with geography and sociolinguistics. Three languages are distinctly low in overall inventory size: Avar, Lezgi, and Udi. These are the three important inter-ethnic languages of the eastern and southern Caucasus. Avar, the language of the large and powerful Avar Khanate and its predecessor, the Sarir kingdom (discussed again below), was and is a major trade language and lingua franca in the northern Caucasus and continues to add speakers by language shift. Udi is now reduced to three enclaves (in the villages of Vartashen and Nidzh in Azerbaijan and the offshoot Zinobiani (formerly Okt’omberi) in Georgia), but was a major language of the Transcaucasus in the first millennium CE.3 There is little good information on the sociolinguistic history of Lezgi, but it has evidently been an 3

It was known as Aghwan or Alwan in historical sources of the time, and is called Caucasian Albanian in Western historical sources. It was one of three Transcaucasus languages deemed sufficiently important by the early Christian church that missionaries were sent to devise scripts for the three languages and translate essential materials such as Gospels, mass texts, important sermons, etc. into them. The other two languages were Georgian and Armenian, which still use the scripts devised in the early first millennium. Udi lost most of its speakers in shifts to Armenian and Azeri, and the writing system died out and was laboriously analyzed (and the identity of Caucasian Albanian established) by 20th- and 21st-century scholarship. See Gippert et al. 2009.

The vertical archipelago: Adding the third dimension to linguistic geography

47

important language for some time, as shown by (i) its large number of speakers, (ii) its location,4 and (iii) the fact that it is now an inter-ethnic language in several market towns and is influencing and absorbing speakers from neighboring Aghul and Rutul. The notable simplicity of these three inter-ethnic languages is strong support for the contention that a language that spreads by language shift to populations that learn it as adults becomes simplified. The outcome of this survey indicates that inventory size can bea good sociolinguistic marker in general: in Nakh-Daghestanian, reduction of inventory size indicates a history of language spread by shift, while high inventory size is associated with non-spreading languages and especially those isolated in the highlands. The comparison works well within Daghestan; Chechen and Ingush are large but complex, and their demography reflects the central Caucasian ecology. (All the other languages of the central and western Caucasus are large, or were large historically: Kabardian, Ossetic, Adyghe, Abkhaz, Mingrelian, Svan.) They are highland languages in origin; in recent centuries lowland Chechen has locally absorbed some lowland speakers (probably originally of Kumyk), but Ingush has absorbed almost no lowland speakers. But once the ecology is held constant, as can be done by surveying only Daghestan, a good correlation of inventory size with sociolinguistic status emerges.

6.

Opacity

My survey of opacity is in its early stages, and only three domains are included here: Gender agreement (of verbs with arguments) One opacity point is scored for each of the following properties: non-overt agreement markers: one or more of the gender categories has zero exponence syncretism within the singular gender agreement paradigm (for some morphological context) syncretism within the plural gender agreement paradigm x

x

x

4

It centers on a sheltered fertile southwest Caspian plain and formerly no doubt included the important Derbent pass, where the Caspian plain narrows and all caravans could be easily taxed. There is a long history of states and kingdoms in this area, known from Persian and Arabic sources, though we have no information about their languages. Coastal cities in the area have long had Tat (Iranian) populations, and the region was under Azerbaijani Iranian control for perhaps two millennia prior to the Turkic entry that brought Azeri and Turkish west of the Caspian, but Iranian speech seems not to have spread beyond the cities.

48

Johanna Nichols

Fig. 2. Altitude (x-axis) by opacity (y-axis). Number of languages: N = 26. Shared variance: R2 = 0.27 x

x x

any non-discrete markers (i.e. non-segmentable markers, such as ablaut or coexponence with some other category) allomorphy of any markers (other than mechanical allophonic variants) agreement marking in more than one morphological position (e.g. initial in some verbs but medial in others)

Semantic arbitrariness of gender classification One opacity point for each of the following: number of genders with semantic arbitrariness number of genders with semantic arbitrariness among animate nouns

x x

Opacity of argument coding (on verbs or on the arguments) One point for each of the following: the marking pattern variously known as egophoric, assertor involvement marking (Creissels 2008), or conjunct/disjunct hierarchical alignment, where referential hierarchies determine access of arguments to agreement slots (called hierarchical alignment by Nichols 1992, and relative-hierarchical alignment by Blake and Mallinson; a recent work is Haude 2009)

x

x

There is no correlation of opacity with either population size or longitude (for all 26 languages, just Daghestan, or just languages with gender). There is

The vertical archipelago: Adding the third dimension to linguistic geography

49

Fig. 3. Inventory size (x-axis) by opacity (y-axis). Number of languages: N = 28. Shared variance: R2 = 0.23

a good correlation with altitude, especially for just Daghestan (see Figure 2). This correlation is strengthened if only languages with gender are counted (understandably, since gender figures in the opacity calculation), but it is weakened if the three inter-ethnic languages Avar, Lezgi, and Udi are removed, though those languages are among the least opaque (and Lezgi and Udi have lost gender entirely). There is also a good positive correlation of overall opacity with overall inventory size (see Figure 3). Though the big inter-ethnic languages are all low-altitude languages, the correlation with altitude is not an artifact of the fact that the inter-ethnic languages are low in opacity. The correlation survives removal of these languages (though it is statistically weakened by the loss of these three extreme cases). Of course it cannot be that sheer elevation, or the climate of higher elevations, directly causes linguistic opacity. Rather, it is the isolation of the high-altitude languages – their distance from each other and from the lowland centers – that favors the development of opacity. Inspection of the more opaque gender paradigms makes it clear that they have developed from natural phonological changes, chiefly assimilation, as well as entrapment of prefixes when former adverbs become fused to the verb as preverbs. Semantically arbitrary gender classes arise in the larger gender systems (those with four or five genders) in several ways. The initial or final consonant of a noun root can be reanalyzed as an overt inherent gender marker and the gender of the noun adjusted accordingly; for example, in Ingush as in most Nakh-Dag-

50

Johanna Nichols

hestanian languages there are nouns like butt ‘moon’, bolx ‘work, job, deal’, bwi ‘army’ that belong to the B gender class and begin with b-. Such nouns are significantly more frequent than expected (Nichols 2007). In such words phonology, rather than semantics, determines gender, so that the gender cannot be predicted from the semantics. Final labials also echo gender, so that e.g. loam ‘mountain’, tq’aam ‘wing’, t’om ‘war; weaponry’ and suffixed deverbal nouns in -am such as biezam ‘love’, bielam ‘laughter’, loarham ‘account’ and many others belong to B gender. The ultimately Arabic loanword q’oalam ‘pencil, pen’ also belongs to B gender, probably because of its final consonant. That is, very natural simplifying processes operating in one domain of grammar (phonology, morphology) increase opacity and arbitrariness in another domain of grammar (the semantics of gender). Such processes go on at all times in all languages, and it must be that in the better-connected lowland languages and especially in the inter-ethnic languages they are eliminated by analogy (and not that they occur faster or more often in isolated languages).

7.

Geographical and typological distance: inclusive-exclusive oppositions

Proto-Nakh and Proto-Daghestanian both made an inclusive/exclusive distinction in first person plural pronouns, though since the Nakh exclusive is cognate to the Daghestanian inclusive (Nichols 2003) it is not obvious that the opposition can be reconstructed to Proto-Nakh-Daghestanian. A number of daughter languages, and even whole branches, have lost the opposition (see Figures 4 and 5). The pattern of loss can be summarized as follows: Nakh and Avar-Andic form a cluster where all the languages have an inclusive/exclusive opposition. These languages are known to have interacted closely with each other going back at least to the middle ages: Avar was the lingua franca for the Avar-Andic area and beyond, and is still more important than Russian as a lingua franca in the immediate area; Avar and Chechen interacted in the lowlands and Chechen has a number of Avar loans; historically, the Andi collected tribute from the other Andic peoples and the highland Chechens; Andic men hired out as shepherds for the highland Ingush and Chechen. In addition to the inclusive-exclusive retention, these same languages resemble each other in the extent to which they make use of causativization in verbal derivation; and they exclusively share a rare Type 5 clitic with identical and very unusual behavioral properties (though the clitics are not cognate) (Nichols & Peterson 2010). At the Dargi-Lezgian boundary, some southernmost Dargi languages and the two northernmost Lezgian languages preserve the opposition. In some

The vertical archipelago: Adding the third dimension to linguistic geography

51

Fig. 4. Retention and loss of inclusive/exclusive oppositions in the daughter languages. Heavy line = retained, light line = lost, heavy dashed line = daughters differ. Italicized language and branch names are those that have lost the opposition.

Fig. 5. The geography of inclusive/exclusive oppositions in the Nakh-Daghestanian languages. Plot symbols are cultural and economic centers of language ranges (for small highland languages the range is often coterminous with the town where the language is spoken). Black circles = languages retaining the opposition, white = languages lacking it.

52

Johanna Nichols

cases it is clear that there has been contact: notably, Chirag Dargi is uphill of Aghul (Lezgian) and on the same river (the Chiragchaj), so its main economic connection must be directly downhill with Aghul. These languages are on the high geographical periphery of their respective branches, so the combination of their isolation from lowland centers and their adjacency to each other has evidently delayed the loss of the opposition. In addition, the isolated Archi (Lezgian outlier) and Khinalug (separate branch or divergent Lezgian) retain the opposition. In all the other branches, there has been a tendency to lose the opposition: the tendency is total in Tsezic and Lak, neartotal in Dargi, and predominant in Lezgian. Thus, retention of the inclusive/ exclusive opposition can be a marker of contact and also of isolation. The contact relations that seem to be telling are the vertically oriented ones along watercourses.

8.

Sociolinguistics and verticality

As was shown above, language spreads involving shift and adult second language learning favor transparency and reduced inventory size, as in Avar, Lezgi, and Udi. These are all lowland to foothill languages in origin. The Andic languages, a set of very closely related languages somewhat less closely related to Avar, are also quite non-complex and transparent. Far from acquiring speakers by language shift, these are small highland languages that have been undergoing shift to Avar in recent centuries. Now, the Avar Khanate, known as the Kingdom of Sarir before its conversion to Islam, existed in its historical location for nearly 3000 years, since the early first millennium BCE. Avar spread along the Avar Koisu canyon and across the crest into southern Georgia and northern Azerbaijan, and also across the northern foothills and up the Andi Koisu, beginning to displace Andic languages and influencing both the Andic languages and the still higher Tsezic languages. There is a good deal of dialect diversity in Avar, showing that it spread as a politically and economically important language some time ago, but this is not remotely close to 3000 years’ worth of diversification. I would estimate about 500 years, enough to produce clear phonological differences that cause borderline non-intelligibility. The cluster of small Andic languages to the west of Avar is very close; impressionistically, these languages differ from each other about as much as Spanish and Portuguese, or Russian and Ukrainian, do, so they reflect perhaps 500–1000 years of divergence. I suggest the following explanation of the Andic reduced inventory size and high transparency. Early in the history of the Sarir Kingdom, ProtoAvar-Andic began spreading from the vicinity of the capital, Khunzakh,

The vertical archipelago: Adding the third dimension to linguistic geography

53

across the foothills and up the Avar Koisu and then the Andi Koisu canyons. Language and dialect spreads and influence continued to emanate from the important market and urban center Khunzakh over a long time, moving up the river canyons and into the highlands through vertical bilingualism. The first descendants of Proto-Avar-Andic to take root in the upper Avar Koisu have probably vanished, replaced in later spreads of their close sisters. At some point Proto-Andic spread up the Andi Koisu, absorbing unknown languages (sisters of the still-higher Tsezic languages?), and Avar later began to spread along the same route. Thus both the Andic languages and Avar owe their non-complexity to language spread from the Sarir/Avar capital area, and the more recent spread of Avar proper has wiped out what otherwise would have been intermediate varieties, creating a sharp distinction between Avar and Andic. How long does it take a language to become less transparent after a contact episode fostering transparency? The Andic spread must have been early medieval to late pre-medieval, yet the languages remain quite transparent. Udi spread even earlier, in the first millennium CE, yet remains quite transparent. Thus contact-induced transparency can evidently persist for centuries. How long does it take a language to become less complex? I have no clear evidence, but the process could be rapid, as shown by the close sisters and neighbors Lezgi and Tabassaran. Lezgi, as noted above, is a large inter-ethnic language; Tabassaran is not an inter-ethnic language and is not known to have been spreading recently. Lezgi has a low overall inventory size and has no gender at all (the ultimate in gender transparency); Tabassaran is fairly high in inventory size and has gender at least dialectally. The phylogenetic closeness of the two languages suggests that Lezgi may have undergone inventory size reduction and become more transparent relatively recently. Another case is Ingush and Chechen, highland languages which spread to the lowlands in the middle ages. Chechen spread by language shift to lowland speakers of other languages probably including Kumyk, and Ingush has acquired some speakers from Ossetic and probably Kabardian in recent centuries. Their gender systems are formally quite transparent (though not semantically transparent) while their inventory size is high, suggesting that pressures to restore transparency operate more quickly than reduction of inventory size. It is a basic principle of dialect geography that in a dialect area innovations emanate from the center and archaisms are retained longer at the periphery, and it is especially telling when the same archaism recurs at distant points on the periphery. In a self-standing mountain region like the Caucasus, the literal

54

Johanna Nichols

geography of the dialect-geographical notions of center and periphery are reversed. The highlands, as we have seen, are the isolated areas where archaisms are preserved (such as the inclusive/exclusive opposition in Archi, Khinalug, and the northern Lezgian-southern Dargi contact zone), though they are at the literal center. Highland languages sharing retentions can be geographically discontinuous, as Archi, the northern Lezgian languages, and Khinalug are; this is analogous to recurrent peripheral archaisms in a non-mountainous linguistic area. Meanwhile, in mountain regions the lowlands are centers of innovation, though they are at the literal periphery of the area: innovations such as loanwords from Avar and Kumyk (in the north) or Lezgi, Persian, and Armenian (in the south) spread into the mountains from the lowlands, phonological innovations such as front rounded vowels spread in from the Turkic languages of the lowlands, lowland isoglosses spread uphill, influence of inter-ethnic languages spreads uphill, and the languages themselves spread uphill. Thus the literal outer edge of the Caucasus functions as a set of centers of innovation, while the inner highlands function as the dialect-geographical periphery ordinarily does, preserving shared archaisms. The outer edge has, for geographical reasons, had fairly uniform linguistic input. The nature of language spreading on and near the Eurasian steppe (Nichols 1998, 2011) has meant that members of the same language family have tended to be the important lowland languages on both the north and south sides of the Caucasus. Ossetic in the north (now a mountain remnant of the once powerful and widespread Alanic language(s) and its/their Scythian ancestor) and Persian and its relatives in the south represent the western extent of the Indo-Iranian spread, and later Khazar and then Kumyk and Karachay-Balkar to the north and Azerbaijani to the south represent the Turkic spread. Thus linguistically similar contact effects have shaped the reduction of inventory size and opacity around much of the lowland area, so that to some extent the lowland periphery functions as a single center as in the usual linguistic geographical area.

9.

The archipelago: Discontinuous verticality

As was discussed in section 3, the central and eastern Caucasus was dominated by asymmetrical vertical bilingualism, whereby highlanders knew lowland languages but not vice versa. This created a standing pattern of uphill spread of isoglosses and languages. There are many aspects of the linguistic geography of mountain highlands that can be explained by this dynamic. If languages move uphill but not downhill, we should expect to see remnants and isolates in highlands, and that is in fact fairly common. Language isolates

The vertical archipelago: Adding the third dimension to linguistic geography

55

and isolate branches in highlands include the isolates Burushaski and Kusunda in the western Himalayas and a number of unique Sino-Tibetan branches in the eastern Himalayas. In the Caucasus, Lak is an isolate branch found only in the highlands, and the Tsezic and Andic branches are found only in the highlands. Indeed, the three indigenous language families of the Caucasus – Nakh-Daghestanian, West Caucasian, and Kartvelian – have no kin anywhere else and illustrate the same principle. Furthermore, in the Caucasus highlands we find stable one-village and very small languages, while in the lowlands small and one-village languages are usually receding enclaves. Important lingua francas such as Avar originate in the lowlands or foothills. There are examples of what I call “Burushaski distributions”, named after the western Himalayan language isolate Burushaski, which is spoken in two separate enclaves in the highest inhabitable reaches of two different mountain valleys but has no continuous lowland territory; the Indo-Aryan language Shina has spread in the lowlands between the two valleys, stranding the Burushaski communities. In the Caucasus, both the Andic and the Tsezic languages are found in the highlands of both the Andi Koisu and the left tributaries of the Avar Koisu, but are cut off in the lowlands by Avar. Distributions like these indicate that languages spread into the mountains and are succeeded by other spreading languages; one after the other, languages move uphill and end up in the highlands, but they never originate there and undergo large downhill spreads. However, this standing directionality is not the entire story, and it does not always result in regular, monotonic uphill creep. If it accounts for some distributions, there are facts it cannot account for. Burushaski-like distributions are not as common as they should be. The only possible example in the Caucasus is the split of both the Andic and the Tsezic branches of NakhDaghestanian between the Andi Koisu and the Avar Koisu (Andic languages occupy the middle stretches of both rivers and Tsezic ones the upper reaches); but it is possible that, rather than spreading uphill along both major rivers and then being stranded by the Avar spread, they both moved up the Andi Koisu and, when already in the highlands, from there across the intervening range from the upper tributaries of the Andi Koisu to the upper tributaries of the Avar Koisu. Locally, one can find Burushaski distributions of Vlach Romanian (highlands) and Slavic (lowlands) in the Balkan Peninsula, and between Basque and Spanish in the Pyrenees, and other such, but cases where an isolate language or branch has its entire distribution split between two highland areas and cut off in the lowlands are rare. In addition, highland languages in the high mountain ranges often appear to have been very long in situ. Both Nakh-Daghestanian (see again note 1)

56

Johanna Nichols

and Himalayan Tibeto-Burman may well have spread with the first settled farming in their ranges. It is not known how long Basque has been in its historical range, but it has a typical southwestern Eurasian structural profile (Nichols 2007) and so could be long pre-Indo-European. Of the major Eurasian mountain ranges, only in the Alps have the relatively recent spreads of Romance, Germanic, and Slavic completely effaced the earlier languages (which are known from historical and inscriptional evidence to have included Celtic, itself the product of the Indo-European spread and not the first farming population in the long-settled Alps). If asymmetrical vertical bilingualism produced uphill spreads with measurable regularity and in measurable time frames, there should be more such cases. Though they are not common, downhill spreads have occurred. Ingush and Chechen (of the Nakh branch of Nakh-Daghestanian) now have large ranges extending from highlands to lowlands. The case of Ingushetia is particularly clear: in the middle ages, the highest highlands were densely populated, wealthy, and culturally prestigious, and in the later middle ages the Ingush demographic and economic center shifted to the foothills (Krupnov 1971 [map: pp. 26f.], Nichols 2005). Linguistically it is clear that isoglosses spread downhill in Chechen and Ingush until relatively recently; ethnographically, highland towns have ancient origins while lowland towns derive from highland ones; all clans derive from highland towns; and there are ancient ethnonyms for large Nakh ethnic and tribal groupings only in the highlands. Finally, in the Caucasus there are a number of discontinuous distributions where lowland languages appear to have leapfrogged into the upper highlands, bypassing the lower highlands. Karachay-Balkar, a Turkic language of the western Caucasus, has lowland and highland ranges but no mid-altitude range. On the south slope, the Khevsur dialect of Georgian overhangs Kisti, the south-slope dialect of Chechen. Avar has several such discontinuities: it has spread along the lower Andi Koisu and in the uppermost highlands, leapfrogging over the Andic languages along the middle Andi Koisu, and it has similarly bypassed the Tsezic languages closer to the Avar Koisu to spread to the upper highlands. Thus the linguistic geography of Avar is a true vertical archipelago. Wixman (1980: 59) explains how this has happened in the case of the Andic language Botlikh. The eponymous town of Botlikh is a market town in the lower Botlikh ranges, and this fact made its language locally important and able to withstand Avar pressure, while in the formerly Botlikh-speaking highlands where there was no Botlikh-speaking market Avar proved more powerful, and that part of the former Botlikh population has shifted to Avar. Since the market towns are usually in the lower reaches of each Andic or

The vertical archipelago: Adding the third dimension to linguistic geography

57

Tsezic language’s territory, those languages remain strongest in their lower ranges and yield to Avar in the hinterlands. There are other factors as well that may have favored leapfrogging of lowland languages to highlands. Recall that the highland male population was traditionally transhumant, moving downhill to slaughter and sell surplus livestock in the markets and spending the winter months in foothill or lowland winter pastures or in seasonal work. These men were bilingual in their highland language and the lowland one. The higher the highland village, the shorter its growing season and the longer the time spent in the winter pastures. The winter season, usually spent with a few male relatives (but without one’s wife, children, and parents) in the vicinity of larger market towns than existed in the highlands, must have been a period of male bonding in those kin groups and much interaction with lowlanders – in the lowland language, which would have been associated with economic productivity as well as male company. This package of profit, bonding, and lowland language then moved uphill for the summer months and was doubtless appealing to the younger males in the household. There were also cases of migration from lowlands to highlands. Ingush highland town foundation traditions often trace the clan back to an ancestor migrating in from the lowlands. Xaradze & Robakidze (1968) suggest that, for Ingushetia, such cases may actually have been settlement or resettlement of an economically dependent population by powerful highland landlords. In Ingushetia this process did not create highland enclaves of lowland languages, but in regions with more and smaller languages it could have. In general, highland economies were precarious: the growing season was too short for grain crops, and households depended on the sale of livestock in the lowlands. The land was populated up to full carrying capacity and there were disputes over access to pasture and boundaries of hayfields. Hill (2001) argues that a contingent or secondary claim to essential economic resources makes an individual more accepting of variation and more prone to shift dialects than an individual with more secure claims on resources, and this model predicts that the highest highlanders will have less stable dialect norms and will be more susceptible to dialect or even language shift. (Information on variation in highland Nakh-Daghestanian communities can probably still be collected so as to test this prediction.)

10. Conclusions To summarize, the vertical dimension is very important in shaping linguistic geography and sociolinguistics in mountain areas. The highlands are iso-

58

Johanna Nichols

lated, very conservative, yet prone to dialect and language shift. One-way vertical bilingualism makes upward spread of isoglosses, dialects, and languages the rule, though as we have seen this does not entail regular, gradual uphill creep; in the cases where we can reconstruct some prehistory, the uphill movement seems to have been saltatory. The resulting language and dialect map can well be called a vertical archipelago.

References Amirkhanov, Khizri Amirkhanovich 1987: Chokhskoe poselenie: Chelovek i ego kul’tura v mezolite i neolite gornogo Dagestana. Moscow: Nauka. Bickel, Balthasar & Johanna Nichols 2003: Typological enclaves. 5th biannual conference, Association for Linguistic Typology. http:www.uni-leipzig.de/~autotyp/ download/enclaves@ALT5–2003BB-JN.pdf. Bickel, Balthasar & Johanna Nichols 2006: Oceania, the Pacific Rim, and the theory of linguistic areas. Berkeley Linguistics Society 32S: Languages and Linguistics of Oceania. Blake, Barry J. & Graham Mallinson 1981: Language Typology: Cross-linguistic Studies in Syntax. Amsterdam: North-Holland. Campbell, Lyle , Terrence Kaufman & Thomas C. Smith-Stark 1986: Mesoamerica as a linguistic area. Language 62: 530–570. Corbett, Greville 1991: Gender. Cambridge: Cambridge University Press. Creissels, Denis 2008: Remarks on so-called “conjunct/disjunct” systems. Presented at Syntax of the World’s Languages III, Berlin. www.deniscreissels.fr/public/Creisselsconj.disj.pdf. Dahl, Östen & Ljuba Veselinova 2006: Language map server. http://www.esri.com/ news/arcuser/0206/language_ms1of2.html. Donohue, Mark 2004: Typology and linguistic areas. Oceanic Linguistics 43: 214–232. Donohue, Mark, Søren Wichmann & Mihai Albu 2008: Typology, areality, and relatedness. Oceanic Linguistics 48: 223–232. Dryer, Matthew 1989: Large linguistic areas and language sampling. Studies in Language 13: 257–292. Emeneau, Murray B. 1956: India as a linguistic area. Language 32: 3–16. Enfield, Nicholas J. 2005: Areal linguistics and Mainland Southeast Asia. Annual Review of Anthropology 34: 181–206. Gippert, Jost 1999: Languages of the Caucasus. TITUS (Thesaurus Indogermanischer Text- und Sprachmaterialien) Didactica. http://titus.fkidg1.uni-frankfurt.de/didact/ karten/kauk/kaukasn.htm. Gippert, Jost, Wolfgang Schulze, Zaza Aleksidze & Jean-Pierre Mahé 2009: The Caucasian Albanian Palimpsests of Mount Sinai. Turnhout, Belgium: Brepols. Güldemann, Tom 2010: Sprachraum and geography: Linguistic macroareas in Africa. In: Alfred Lameli, Roland Kehrein & Stefan Rabanus (eds), Language and Space: An International Handbook of Language Variation, Vol. 2: Language Mapping, 561–585. Berlin/New York: Mouton de Gruyter. Haude, Katharina 2009: Hierarchical alignment in Movima. International Journal of American Linguistics 75(4): 513–532.

The vertical archipelago: Adding the third dimension to linguistic geography

59

Joseph, Brian 2001: Is a Balkan comparative syntax possible? In: María Luisa Rivero & Angela Ralli (eds.), Comparative Syntax of Balkan Languages, 17–43. Oxford: Oxford University Press. Korjakov, Jurij. B. 2006: Atlas kavkazskix jazykov. Moscow: Piligrim. Krupnov, Evgenij I. 1971: Srednevekovaja Ingushetija. Moscow: Nauka. Lindstedt, Jouko 2000: Linguistic Balkanization: Contact-induced change by mutual reinforcement. In: Dicky G. Gilbers, John Nerbonne & Jos Schaeken (eds.), Languages in Contact, 231–246. Amsterdam, Atlanta: Rodopi. Maddieson, Ian 2005: Syllable structure. In: Martin Haspelmath, Matthew Dryer, David Gil & Bernard Comrie (eds.), World Atlas of Language Structures, 54–57. Oxford: Oxford University Press. Masica, Colin P. 1976: Defining a Linguistic Area: South Asia. Chicago: University of Chicago Press. Nettle, Daniel 1999: Linguistic diversity of the Americas can be reconciled with a recent colonization. Proceedings of the National Academy of Science of the U.S.A. 96: 3325–3329. Nichols, Johanna 1992: Linguistic Diversity in Space and Time. Chicago: University of Chicago Press. Nichols, Johanna 1997: Modeling ancient population structures and movement in linguistics. Annual Review of Anthropology 26: 359–384. Nichols, Johanna 1998: The Eurasian spread zone and the Indo-European dispersal. In: Roger Blench & Matthew Spriggs (eds.), Archaeology and Language II: Archaeological Data and Linguistics Hypotheses, 220–266. London: Routledge. Nichols, Johanna 2003: The Nakh-Daghestanian consonant correspondences. In: Dee Ann Holisky & Kevin Tuite (eds.), Current Trends in Caucasian, East European, and Inner Asian Linguistics: Papers in Honor of Howard I. Aronson, 207–251. Philadelphia: John Benjamins. Nichols, Johanna 2005: The origin of the Chechen and Ingush: A study in alpine linguistic and ethnic geography. Anthropological Linguistics 46: 129–155. Nichols, Johanna 2007: A typological geography for Indo-European. In: Karlene Jones-Bley, Martin E. Huld, Angela Della Volpe & Miriam Robbins Dexter (eds.), Proceedings of the 18th Annual UCLA Indo-European Conference, 191–211. Washington, DC: Institute for the Study of Man. Nichols, Johanna 2009: Linguistic complexity: A comprehensive definition and survey. In: Geoffrey Sampson, David Gil & Peter Trudgill (eds.), Language Complexity as an Evolving Variable, 110–125. Oxford: Oxford University Press. Nichols, Johanna 2011: Forerunners to globalization: The Eurasian steppe and its periphery. In: Cornelius Hasselblatt, Peter Houtzagers & Remco van Pareren (eds.), Language Contact in Times of Globalization 2, 177–195. (Studies in Slavic and General Linguistics.) Amsterdam: Rodopi. Nichols, Johanna, & David A. Peterson 2010: Contact-induced spread of the rare Type 5 clitic. PAGES Linguistic Society of America Annual Meeting, Baltimore. Nichols, Johanna, David A. Peterson & Jonathan Barnes 2004: Transitivizing and detransitivizing languages. Linguistic Typology 8:149–211. Sandfelt, Kristian 1930: Linguistique balkanique: Problèmes et résultats. Paris: Klincksieck. Szmrecsanyi, Benedikt 2009: Typological parameters of intralingual variability: Grammatical analyticity versus syntheticity in varieties of English. Language Variation and Change 21(3): 319–354.

60

Johanna Nichols

Szmrecsanyi, Benedikt & Bernd Kortmann 2009: Between simplification and complexification: Non-standard varieties of English around the world. In: Geoffrey Sampson, David Gil & Peter Trudgill (eds.), Language Complexity as an Evolving Variable, 65–79. Oxford: Oxford University Press. Trudgill, Peter 2009: Sociolinguistic typology and complexification. In: Geoffrey Sampson, Peter Trudgill & David Gil (eds.), Language Complexity as an Evolving Variable, 98–109. Oxford: Oxford University Press. Trudgill, Peter 2011: Sociolinguistic Typology: Social determinants of linguistic structure and complexity. Oxford: Oxford University Press. Vermeer, Willem 2005: The rise of the Balkan linguistic type: Preliminary considerations for private consumption only. MS, University of Leiden. Volkova, N. G. 1974: Etnicheskij sostav naselenija Severnogo Kavkaza v XVII-nachalo XX veka. Moscow: Nauka. Wixman, Ronald 1980: Language Aspects of Ethnic Patterns and Processes in the North Caucasus. Chicago: University of Chicago. Yoffee, Norman 2001: The evolution of simplicity. Current Anthropology 42(5): 767–769. Yoffee, Norman 2005: Myths of the Archaic State: Evolution of the Earliest Cities, States, and Civiliations. Cambridge: Cambridge University Press. Zohary, Daniel & Maria Hopf 2000: Domestication of Plants in the Old World. Oxford: Oxford University Press.

Language contact between geographic and mental space

61

Walter Bisang

Language contact between geographic and mental space

1.

Introduction

The geographic or areal factor is of crucial importance for linguistic typology. Only when geographic contiguity and language contact are properly integrated into the statistical methodology of linguistic typology is it possible to find reliable universal generalizations (for different methodological approaches, cf. Dryer 1992; Rijkhoff and Bakker 1998; Janssen, Bickel, and Zuñiga 2006; Wichmann and Holman 2009; Bakker 2011). This can be illustrated by a classical example discussed in Dryer (1989), who pointed out that the mere statistical frequency of word-order patterns across the globe cannot be taken at face value for drawing typological generalizations. About eight to nine percent of the world’s languages have verb-subject-object order (VSO; 9.20 % in Tomlin 1986; 8.05 % or 85 out of 1056 languages in Dryer 2005). However, this percentage is heavily biased by the sociohistorical factor of the successful geographic/areal diffusion of Austronesian languages, since 71 % of the VSO languages belong to that family. Without Austronesian, VSO order would only be attested in some three percent of the world’s languages. Even though Dryer’s (1989) example clearly shows the relevance of the geographic/areal factor for linguistic typology, the question of the present paper is: To what extent is the plotting of linguistic structures onto a map of the world sufficient to assess the impact of contact on typological generalizations? This paper will show that contact-induced structural convergence is not limited to geographic contiguity. For that purpose, it will look at two types of contact spaces beyond simple geographic contiguity: (i) situations of contact through speaker mobility, i.e. situations in which the contact takes place outside of the geographic space in which at least some of the languages involved are spoken, and (ii) contact situations based on virtual concepts of space in which speakers and their respective speech communities are in contact only indirectly by way of writing. The phenomena described in the present paper also belong to two types. The first type is a single feature and thus reflects standard typological methodology. The second type is hidden complexity (cf. Bisang 2009 and

62

Walter Bisang

section 2) – a far more general property which is reflected across many different features and is rather stable in situations of language contact in the sense that languages which do not have it cannot acquire it simply through contact. Creole languages are a good example of contact through speaker mobility. The majority of them are geographically separated either from the speakers of the superstrate languages (most often lexifier languages such as English, Dutch, French, Portuguese and Spanish) or from both the superstrate and the substrate languages. As a consequence, the contact situation cannot be depicted on a map by mere contiguity. The second type of contact needs more explanation. Speakers may have internalized a mental geography of places they want to belong to and cultures they want to be associated with, even if they are not in direct contact with them. These imagined landscapes of belonging have their linguistic impact even though it is difficult to plot them onto a map. Thus, identity is created by claiming differences from and similarities with other cultures, and this generates a more abstract conceptual map and linguistic consequences. Such situations do not only arise in modern electronic communication; they have a long tradition across cultures through processes of written exchange and the need to translate texts from one language into another. The situations to be analysed will be Vietnam and its contact with China and the influence of Western languages like English on Chinese (Modern Standard Chinese, i.e. Mandarin). The prototypical phenomena used in areal linguistics are clearly defined individual features such as the presence/absence of a certain grammatical category (e.g. the existence of plural marking, tense marking, etc.) or a certain value within an attribute (e.g. VO vs. OV as two values of basic word order). Probably the best-known example in which individual grammatical features are plotted onto a map of the world to show their geographic diffusion is the World Atlas of Language Structures (Haspelmath et al. 2005). Many areas are characterized by specific clusters of features (cf. Sprachbünde or zones of convergence, Bisang 2006b; spread zones and residual zones, Nichols 1992). In this paper, I will look at hidden complexity, a more general property which is based on the way in which the two competing motivations of explicitness and economy (Haiman 1983) interact in grammatical systems. Individual languages adopt different solutions – some of them tend more to the economy-side, while others are characterized by more explicit structures. The distribution of these solutions across the globe shows an interesting areal clustering in East and mainland Southeast Asian (EMSEA) languages, in which economy and pragmatic inference (Bisang 1996, 2008) take a par-

Language contact between geographic and mental space

63

ticularly dominant position. EMSEA languages are thus characterized by a high degree of hidden complexity. The two types of geographically non-contiguous contact-induced structural influence and their reflections on individual typological features and on hidden complexity will be discussed as follows: Contact through speaker mobility will be illustrated with regard to its consequences for hidden complexity. Contact situations based on virtual concepts of space will be studied from both perspectives: those of individual typological features as well as those of hidden complexity. Since typological studies generally focus on individual typological features, one example which clearly illustrates how typological change happens with almost no direct contact should be sufficient. The less-known phenomenon of hidden complexity will be discussed for both types of non-contiguous contacts. The structure of the paper will therefore be as follows: Hidden complexity as a result of the two competing motivations of explicitness and economy will be introduced in section 2. Sections 3 to 5 will deal with contact through speaker mobility and hidden complexity. Section 3 will illustrate the degree of hidden complexity as it is manifested in the two phenomena of pro-drop and relative clause formation in Chinese as a representative of EMSEA languages. Sections 4 and 5 will do the same for West African languages (Yoruba, Fongbe) and five Creole languages, respectively. As the comparison across sections 3 to 5 will show, EMSEA languages are characterized by the highest degree of hidden complexity, which is not even reached by Creole languages (in spite of McWhorter’s 2001, 2005 claim of their highest degree of simplicity, cf. section 2). Sections 6 and 7 address instances of language change through virtual space. Section 6 illustrates how written contact with Western languages leads to changes in the discourse structure of Chinese (use of the bèi-passive) and increases the potential of lowering the degree of hidden complexity (reduction of pro-drop), a possibility that is taken up by certain contemporary Chinese authors. Section 7 will illustrate how contact through writing also affects the emergence of new grammatical rules. Vietnamese compounding changed from exclusively head-initial to both head-initial and head-final through heavy borrowing from Chinese. The paper will end with a short conclusion in section 8. The following table summarizes the phenomena to be described and their assignment to the two types of noncontiguous contact:

64

Walter Bisang

Contact through speaker mobility

Contact based on virtual concepts of space

Hidden complexity

Individual typological feature

§§ 3–5, comparison of pro-drop and relative clauses in Chinese, West African languages and Creole languages § 6, impact of Western languages on discourse structures in Chinese

---

§ 7, emergence of new word-formation rules in Vietnamese through heavy borrowing from Chinese

Table 1: Types of non-contiguous contact and linguistic phenomena

The explanations given in the present paper are based on the assumption that grammar-external sociocultural factors have an impact on grammar by processes of selection and diffusion, and the paper thus adopts Croft’s (1995, 2000; also cf. Bisang 2004, 2006a) “integrative functionalism”. In a situation of linguistic variation, speakers select a variant they will use in a concrete speech situation according to grammar-internal and sociolinguistic criteria. Thus, a particular construction belonging to a set of variants may be selected because it optimally follows certain cognitive properties such as economy, explicitness, analyticity or parsing, or because it is associated with certain social features that matter in a given speech situation. If this is true at the level of individual speakers it is also true at the level of geographic/areal diffusion, since the distributional patterns of linguistic types across the world’s languages are ultimately the result of the successful diffusion of particular linguistic structures as they are selected by individual members of speech communities.

2.

Explicitness vs. economy, overt vs. hidden complexity

The definition of complexity is notoriously difficult and has been the subject of controversial discussions in linguistics. There are basically two different lines of argumentation. One takes complexity at its face or surface value, while the other one understands complexity as the result of underlying simplicity. The first approach defines complexity in terms of grammatical categories and rules as they are manifest in overt morphological and syntactic structures. It has been discussed in various recent publications on linguistic typology such as McWhorter (2001, 2005), Dahl (2004) and Miestamo, Sin-

Language contact between geographic and mental space

65

nemäki, and Karlsson (2007). The second approach understands seemingly complex structures as epiphenomena of the recursive application of a small number of simple universal principles which are specifically human and constitute the faculty of language. This approach is prominently taken by Hauser, Chomsky, and Fitch (2002) and Fitch, Hauser, and Chomsky (2005). It is also discussed in linguistic typology in a volume edited by Givón and Shibatani (2009). The present paper discusses complexity from yet another perspective which is based on the competition between explicitness and economy in grammars of individual languages (Haiman 1983; Prince and Smolensky 2004). This competition is motivated by what Levinson (2000: 6, 27–30) calls the “articulatory bottleneck”. Since articulation and overt encoding of information are much more “expensive” than inference, language structure has to opt for a certain degree of economy to the detriment of explicitness. The degree to which grammars of individual languages allow economy varies cross-linguistically. In many instances of obligatory marking, the structure of a language simply forces the speaker to explicitly encode a certain grammatical category even if it could easily be inferred from context. In other instances, the structure of a language does not force the speaker to address that category if its content is contextually available. Current approaches to complexity have a strong tendency to concentrate on the explicitness side of grammar. Since they look at what is overtly expressed in individual languages, they deal with what I call “overt complexity” (Bisang 2009). McWhorter’s (2001, 2005) surface-based approach is a straightforward case in point. Degrees of complexity are measured on the basis of markedness patterns and the number of rules that are needed for the production of grammatically acceptable utterances. In the case of recursionbased approaches, what matters is the overt linguistic manifestation of a small number of syntactic rules as they are used recursively. The fact that simple rules do not necessarily produce simple behaviour is impressively illustrated in computer sciences by Wolfram (2002), who shows how a simple cellular automaton rule can produce a fascinating pattern of black and white cells after a few hundred recursive steps – a pattern that seems impenetrable to the human mind. However, as pointed out in the introduction (section 1), complexity does not only depend on explicitness; it also has an economic side. Economy is characterized by the lack of obligatory grammatical categories and by the comparatively high importance of pragmatic inferencing. Both of these properties generate surface simplicity which requires pragmatic enrichment. Since the surface structure of an utterance alone does not provide enough

66

Walter Bisang

grammatical information, its adequate interpretation depends on covert or hidden information which needs to be pragmatically inferred. Thus, economy is associated with what I call “hidden complexity” (Bisang 2009). Hidden complexity adds to the complexity of seemingly simple surface structure through rather complex processes of inference. Before hidden complexity is described in more detail, I will introduce overt complexity as described by McWhorter (2001, 2005) against the background of his claim that Creoles represent the world’s least complex type of languages. As will be shown, the integration of the economic side of complexity does not support his hypothesis. Since it turns out that the area of EMSEA languages is characterized by a comparatively high degree of hidden complexity, it will also be shown that the extent to which a grammatical system is influenced by hidden complexity is of geographic importance. McWhorter’s (2001, 2005) starting point is based on the observation that Creole languages are the result of an extreme contact situation in which speakers of different mutually unfamiliar languages are forced to talk to each other. In such a situation, speakers tend to avoid categories and structures that go beyond what is needed for basic communication: “Creoles, in being recently borne of communication vehicles deliberately designed to eschew all but the functionally central (pidgins), are unique examples of natural languages with much less contingent accumulation of “ornamental” elaboration than older grammars drag along with them” (McWhorter 2005: 43). In McWhorter’s (2001, 2005) view, grammars of Creole languages display less complexity than languages with a longer past because of their lack of historically accumulated “‘ornamental’ elaboration”. Many grammatical categories and grammatical distinctions attested in the world’s languages are the result of a long period of development within a more or less stable speech community. Creole languages simply did not have the time to develop many overt distinctions that are the result of millennia of grammaticalization and reanalysis in older languages. To corroborate his statement, McWhorter’s (2005: 45) approach is based on the intuition that “an area of grammar is more complex than the same area in another grammar to the extent that it encompasses more overt distinctions and/or rules than another grammar”.

3.

The extent of hidden complexity in EMSEA languages

EMSEA languages are well known for their optional expression of grammatical categories such as definiteness, number, tense-aspect, verbal arguments and person (Sapir 1921: 92; Bisang 1996, 2008). Pro-drop is one in-

Language contact between geographic and mental space

67

stance of optionality which is discussed very prominently in the literature (for a good survey, cf. Neeleman and Szendröi 2007). In radical pro-drop languages, verbal arguments can be omitted completely, even without crossreference marking on the verb, if they are inferrable from context. The following sentence illustrates radical pro-drop in Chinese: (1) a. Nˇı zuóti¯an kàn-le diànyˇıngi ma? you yesterday see-PFV film Q ‘Did you see a film yesterday?’ b. ø kàn-le

øi.

see-PFV ‘[ I ] saw [one].’ The argument nominals which take the ø-positions in (1b) must be inferred from the context. Since the question in (1a) is directed at the addressee who replies to it, the first-person subject interpretation is a straightforward option for (1b). The noun diànyˇıng ‘film’ as the object of the verb kàn ‘see’ in (1b) has been mentioned in the previous question in (1a) and is thus easily retrievable, too. Pragmatic inference is of central importance in many other domains of Chinese grammar as well, including relative clause formation. Chinese relative clauses always end with the attributive marker de and precede their head nouns. The syntactic function and the semantic role of the head noun within the relative clause can be inferred from the fact that arguments of intransitive and monotransitive verbs must be zero-marked. Since pro-drop is generally possible in relative clauses, constructions consisting merely of a transitive verb are ambiguous – the head noun may be coreferent with the subject (actor argument) or the object (undergoer argument). Given our knowledge of the world, actor coreference is the expected interpretation in (2a), while undergoer coreference corresponds to what is expected in (2b): (2) a. Actor coreference (preferred): [øi ch¯ı ø de] réni eat ATTR man ‘the man who eats’

68

Walter Bisang

b. Undergoer coreference (preferred): [ø ch¯ı øi de] píngguˇoi eat ATTR apple ‘the apple which [X] is eating’ The interpretations as they are reflected in the English translations are not strictly due to grammar. Grammar itself leaves a lot of leeway. It is only our knowledge of the roles that men and apples usually have with a verb like ch¯ı ‘eat’ that triggers these interpretations as default solutions. Given the right context, the head noun in (2a) can be interpreted as an undergoer and the head noun in (2b) can be understood as an actor. With certain head nouns and within certain contexts (cf. Ning 1993), the head noun can also be interpreted in terms of a non-argument role as a locative (2c) or an instrumental (2d): (2) c. Locative coreference: [t¯a ch¯ı ch¯unjuˇan de] fànguˇan s/he eat spring.roll ATTR restaurant ‘the restaurant where s/he ate spring rolls’ d. Instrumental coreference: [w˘o xiˇe xìn de] máobˇı I write letter attr pencil ‘the pencil I write a letter with’

4.

The extent of hidden complexity in two West African languages

In West African languages pragmatic inferences in the two grammatical domains discussed in section 3 are extremely restricted. This will be illustrated by data from the two West African languages of Yoruba (Niger-Congo: Benue-Congo: Defoid) and Fongbe (Niger-Congo: Kwa). In Yoruba, there are no pro-drop subjects and objects must be overt in simple independent clauses with no focus and wh-questions. The only instances in which subject pro-drop is possible are related to the subject pronoun of the third-person singular ó, which must be omitted in front of the negation kò or with the future marker yó (cf. Bamgbos.e 1967: 42). Examples 3a–d are grammatical because they show overt expressions in the subject

Language contact between geographic and mental space

69

position as well as in the object position. Thus, there is a noun (3a, c) or a pronoun (3b, d) in the subject position, and the object position is either filled with a noun as in 3a, b or with a pronominal-object marker as in 3c, d. If one or both of these positions are empty as in 3e–g the sentence is ungrammatical: (3) Yoruba: On pro-drop with the verb rà ‘buy’:1 a. Ayò. ´ ras.o.. Ayò. ´ rà as.o Ayo HTS buy clothes ‘Ayo bought clothes.’

b. ó ras.o.. ó rà as.o.. he buy clothes ‘He bought clothes.’

c. Ayò. ´ ràá. Ayò. ´ rà-á. Ayo HTS buy-3.OBJ ‘Ayo bought it.’

d. ó ràá. ó rà-á. he buy-3.OBJ ‘He bought it.’

e. *ras.o.. ø rà as.o buy clothes

f. *Ayò. ´ rà. Ayò. ´ rà ø. Ayo HTS buy

g. *rà. ø rà ø. buy

Similarly, the function of the head noun in Yoruba relative clauses is not left to pragmatic inference. There are special constructions for subject coreference, object coreference and adjunct coreference. In each case, the relative clause follows the head noun and is introduced by the relative marker tí. In 1

Some explanations: (i) HTS = High Tone Syllable. This tonal marker is a floating tone which occurs at the final syllable of subject NPs under certain semantic conditions. Bisang and Sonaiya (1999) define these conditions against the syntactic analysis of Déchaine (1993) in terms of validation: The High Tone marks that the relation between a subject X and a predicate p in a state of affairs is positively realized. (ii) VO constructions: Verbs have the phonological structure of CV, while nouns have VCV. In VO constructions, there are processes of elision between the vowel of the verb and the initial vowel of the noun (for details, cf. Sonaiya 1990). Some examples: • kà ‘read’ + ìwé ‘book’ f kàwé ‘read a book’ • fo. ‘wash’ + as..o ‘clothes’ f fo..s.o ‘wash clothes’ • rú ‘violate’ + òfin ‘law’ f rúfin ‘violate the law’ • ko. ‘write’ + ìwé ‘book’ f kò. wé ‘write a book, write’ • je. ‘eat’ + ohun ‘thing’ f je.un ‘eat’

70

Walter Bisang

the case of subject coreference, the subject position must be filled by a subject pronoun (4a). In the case of object coreference, the object marker, which is obligatory in simple independent clauses if no overt noun is mentioned, must be dropped (4b). Finally, locative coreference entails the presence of the locative particle ti (4c): (4) Yoruba: Relative clauses: a. Subject coreference: obìnrin [t’ó maa ràá]. obìnrin tí ó maa rà-á woman REL 3.SG.SUBJ TA buy-3.OBJ ‘the woman who bought it’ b. Object coreference: mo rà øi lánàá] is.ui [tí yam REL 1.SG buy yesterday ‘the yam I bought yesterday’ c. Locative coreference: ó mò. o. jà [tí mo ti ràá]. he know market REL I PART:LOC buy:3.OBJ ‘He knows the market where I bought it.’ Fongbe, as described by Lefebvre and Brousseau (2002), is remarkably similar to Yoruba. Simple independent clauses with a transitive verb have an obligatory subject and an obligatory object. There are, however, a few verbs with optional subjects which Lefebvre and Brousseau (2002: 246, 276–277) call “verbs licensing expletive subjects”. Two examples are given in (5) and (6): (5) Lefebvre and Brousseau (2002: 276): ɔ` Kɔ`kú jε` àzɔ`n. (É) cí 3.SG seem COMP Koku be.sick ‘It seems that Koku is sick.’ (6) Lefebvre and Brousseau (2002: 277): nú Kɔ`kú ní yì. É vε`-wú 3.SG be-difficult COMP Koku SUBORD leave ‘It is difficult that Koku leaves.’

Language contact between geographic and mental space

71

With simple transitive verbs, the object position is not allowed to be empty as long as the object is not extracted to the focus position or is a wh-word which must take that position. In the case of topic extraction, the object position is filled facultatively by a resumptive pronoun. There is a special class of “inherent object verbs”, i.e. verbs that cannot appear without an object of some kind. In example 7, the verb nyà ‘hunt’ occurs with a normal object. Even if no particular object is meant it is still necessary to add an overt object. Lefebvre and Brousseau (2002: 247–249) discuss three strategies. In the first strategy illustrated in 8, the verb appears with an object which is typically associated with that verb. Another option is the use of the noun nú ‘thing’ (9). Finally, the verb can be combined with its nominalized form, derived by the class prefix à- (10): (7)

Fongbe (Lefebvre and Brousseau 2002: 247): nyà gbé hunt animal ‘to hunt’

(8)

Fongbe: Typical object (Lefebvre and Brousseau 2002: 248): kùn hún drive vehicle ‘to drive’

(9)

Fongbe: Object meaning ‘thing’ (Lefebvre and Brousseau 2002: 248): nú eat thing ‘to eat’ ù

(10) Fongbe: cognate object (Lefebvre and Brousseau 2002: 249): ì àì believe belief ‘to believe’ In Fongbe, there is again not much room to pragmatically infer the function of the head noun in the relative clause. In general, the head noun “is linked to a position within the relative clause through the lexical nominal operator, é ” (Lefebvre and Brousseau 2002: 161). This operator attracts lexical material. In the case of relative clauses, it attracts resumptive pronouns by moving them from their initial position in the relative clause to the position immediately after é. This process generates surface structures which clearly differ

72

Walter Bisang

for each type of coreference. In example 11a on subject coreference, the resumptive subject pronoun é is attracted to the nominal operator. In the case of object coreference, the operator é attracts the object pronoun è, while the object position itself is empty (11b). If the function of the head noun in the relative clause is that of an adjunct the resumptive pronoun is combined with a post-position and both markers together are attracted to the nominal operator e (11b). The extraction site is empty as in the case of coreference with the direct object in 11b. (11) a. Subject coreference (Lefebvre and Brousseau 2002: 161): wá] ɔ´ súnû [é-é man OP-3.SG.SBJ come DEF ‘the man who came’ b. Object coreference (Lefebvre and Brousseau 2002: 161): àsɔ´ni [é-èi Kɔ`kú ú øi] ɔ´. crab OP-3.SG.OBJ Koku eat DEF ‘the crab that Koku ate’ c. Locative coreference (Lefebvre and Brousseau 2002: 161): Kɔ`kú sɔ´ àwíì ó øi] ɔ´. xàsùni [é-é-m`εi basket OP-3.SG.OBJ-in Koku take cat put DEF ‘the basket in which Koku put the cat’ The above description shows that Yoruba and Fongbe are clearly non-prodrop. Subject and object arguments cannot be zero in simple independent clauses. Relative clauses are characterized by obligatory structural patterns which clearly distinguish between subject coreference, object coreference and non-argument coreference. These obligatory patterns leave no room for pragmatically inferring the role of the head noun in the relative clause. Thus, the grammatical system excludes hidden complexity.

5.

The extent of hidden complexity in Creoles

This section will show that Creoles show less hidden complexity than EMSEA languages. It will be argued that this is due to the stability of the degree of hidden complexity which is hard to change once speakers have acquired the solution adopted by their language (also cf. section 1). For that purpose, pro-drop and relative clause formation will be analysed in five Creole languages, which are all described in the very useful syntactic survey

Language contact between geographic and mental space

73

of Holm and Patrick (2007). Three of them are Atlantic Creoles, two of them Pacific Creoles. The Atlantic Creoles are Angolar (Portuguese-based Creole spoken on São Tomé, an island in the Gulf of Guinea, Africa), Berbice Dutch (Dutch-based Creole spoken in Guyana, South America) and Haitian Creole or Kreyòl Ayisyen (French-based Creole spoken in Haiti). The Pacific Creoles are Tok Pisin (English-based Creole, a variety of Melanesian Pidgin spoken in Papua New Guinea) and Zamboangueño (Spanish-based Creole, a variety of Chabacano or Philippine Creole Spanish, spoken in and around Zamboanga City on the southern tip of Mindanao Island). In the Atlantic Creoles, pro-drop seems to be rather limited. Thus, Neeleman and Szendröi’s (2007) conclusion that Creoles such as Haitian, Jamaican and Papiamento are not radical pro-drop languages is correct. In spite of this, DeGraff (1993) argues that Haitian is a pro-drop language, and in fact there are constructions with no overt subject marking. Example 12 illustrates the raising verb genle ‘seem’, which cannot take an expletive subject. Other constructions without expletives are existential predicates and weather predicates. In other contexts, however, overt arguments are necessary (13). Thus, the claim of Déprez (1994) that Haitian is not a pro-drop language is very convincing on the whole. She shows that person markers are not clitics because they are not necessarily adjacent to the verb and can be coordinated. In addition, pronouns are obligatory in extraposition (14): (12) Haitian Creole (DeGraff 1993: 72): a. ø genle Jak damou. b. *li genle Jak damou. seem Jack be.in.love 3.SG seem Jack be.in.love ‘Jack seems to be in love.’ (13) Haitian Creole (DeGraff 1993: 72): a. Li pati. b. *ø pati 3.SG leave leave ‘S/He left.’ (14) Haitian Creole (Déprez 1994): *(Li) difisil pou pale ak Jan. 3.SG difficult to speak with John ‘It is difficult to speak with John.’ In Berbice Dutch, the subject position of a simple independent clause must be filled. As Kouwenberg (1994: 176) points out, this even applies to instances in which expletive it is used in English: “Where expletive ‘it’ appears

74

Walter Bisang

in English, a referential pronoun or noun appears in BD [BD = Berbice Dutch, W.B.]. The weather verb koro ‘fall’ (of rain) requires as its subject εnε ‘rain’ and BD has no raising verbs such as ‘seem’, ‘appear’” (Kouwenberg 1994: 176). The following example illustrates the use of the third-singular pronoun o in an impersonal construction of reported knowledge: (15) Berbice Dutch (Kouwenberg 1994: 177): o bi wa da twe sosərapu. 3.SG say PST be two sisters ‘It is said (they) were two sisters.’ In certain discourse environments, however, subjects are not realized. The most important context in which this happens is in sequences of events. In this case, only the first verb of a sequence of chronologically concatenated events takes the subject marker (and other preverbal material) – in all the subsequent predications, the subject remains unexpressed: (16) Berbice Dutch (Kouwenberg 1994: 180): en ø plagi just mu mui, di potman ʃi 3.PL PAST.HAB go PURP trouble DEF old.man 3.POSS junggwa:, ø jef ʃi nanaʃi. gutwap, kap ʃi thing: PL cut 3.POSS sugarcane eat 3.POSS pineapple ‘They would go to trouble the father’s things, cut his sugarcane, eat his pineapples.’ Angolar seems to be one of the languages in which the subject position is not necessarily filled by an overt element. Even though the subject position is filled very frequently at least by a pronoun, Maurer (1995: 62) points out that subject pronouns can be omitted if they can be inferred from the context. The following passage from Maurer’s (1995) text collection is particularly interesting because after two instances of zero-marking with the second and the third verbs, the subject is mentioned again with the fourth and the fifth verbs which repeat the same state of affairs that has been used with no overt subject before:

Language contact between geographic and mental space

75

(17) Angolar (Maurer 1995: 160): atê Ben. A2 pê sitaka, ø dhuma mbambu têlu OK 3.PL put beam drive.in bamboo completely up.to ponta kampu e loke fêtu. I minhu rê, u˜ a limit/end field DEM finish/completely and corn their one kobo. u˜ a taminha a ka pê kobo. u˜ a taminha baga ø pê plate put hole one plate 3.PL TA put hole. one ka pê kobo taminha a plate 3.PL TA put hole. ‘OK. They put the beams [to the ground] and drove [their] bamboo sticks [into the ground] all over the field up to its limits. As for their corn, they put one plate [of it] into a hole [produced by the bamboo stick]. They put one plate into a hole. They put one plate into a hole. [The repetition is used to say that they put a plate of corn into each of the holes.]’ In the two Pacific Creoles of Zamboangueño and Tok Pisin, pro-drop behaviour is generally comparable to the substrate languages involved. In Zamboangueño, null pronouns generally occur very frequently. Thus, there is no overt subject in the following two examples (since Zamboangueño is a VSO language, the ø-markers are put after the verb): (18) Zamboangueño (Lipski and Santoro 2007: 376; initially from Forman 1972: 168): Andá ø alyì na réyno. go there to kingdom ‘He goes there to the kingdom.’ (19) Zamboangueño (Lipski and Santoro 2007: 376): Kwándo sale ø afwéra ya murí ø. when leave outside PST die ‘When he went outside, he died.’

2

The pronoun a, which is called pronom indéfini ‘undefined pronoun’ by Maurer 1995: 59, can either refer to an undefined subject or to the third plural. In the latter case, it stands for the pronoun ane/ene ‘they’ (Maurer 1995: 61).

76

Walter Bisang

Lipski and Santoro (2007: 376, footnote 2) mention a number of zero pronouns in various contexts such as non-statives with past reference, infinitive constructions marked by para ‘for’, relative clauses, passives, serial verbs and reflexive constructions. The two examples above additionally indicate that pro-drop is also possible in simple independent clauses. Given the frequency of radical pro-drop in Philippine languages, the pro-drop behaviour of Zamboangueño seems to be in line with its substrate languages. In Tok Pisin, the predicate is marked by the predicate marker i, which immediately precedes the verb. It occurs with nominal subjects and with thirdperson elements in general, and also with some other person markers (for details, cf. Verhaar 1995: 70–73). Example 20 illustrates person markers which have to co-occur with the predicate marker i; example 21 illustrates instances in which i is not possible if it immediately follows the subject pronoun: (20) Tok Pisin (Verhaar 1995: 71–72): Em / ol / mipela / yupela / yumitupela i amamas. 3.SG 3.PL 1.PL.EXCL 2.PL 1.DU.INCL PM be.happy ‘S/He / they / we (excl.) / you (pl.) / the two of us (incl.) is/are happy.’ (21) Tok Pisin (Verhaar 1995: 71): Mi / yu / yumi amamas. 1.SG 2.SG 2.SG.INCL be.happy ‘I am happy. / You are happy. / We (incl.) are happy.’ The subject must be marked overtly in simple independent clauses. Even in the case of the third-person singular, at least a pronoun is needed (22). In addition, instances in which an overt noun phrase is followed by the third-person pronoun are quite frequent (23): (22) Tok Pisin (Verhaar 1995: 35): Em i kam insait gen. 3.SG PM come in again ‘He came in once again.’ (23) Tok Pisin (Verhaar 1995: 29): Bos bilong pilai em i makim wanpela boi. leader GEN game 3.SG PM mark/appoint one boy ‘The leader of the game appoints one boy.’

Language contact between geographic and mental space

77

With weather predicates, the element expressing the weather quality takes the subject position as in ren i kamdaun [rain PM come.down] ‘it rains’. The only instances in which the subject position remains unfilled are impersonal constructions as in i gat ‘there is’, i tru ‘it is true’, i luk ‘it looks’, i tambu ‘it is forbidden’, etc.: (24) Tok Pisin (Verhaar 1995: 35): I gat wanpela rot tasol bilong kisim mani. PM have/there.is one way only for get/catch money ‘There is only one way to get income.’ The marginality of pro-drop in simple independent main clauses parallels the situation in the Austronesian substrate languages to a considerable extent. Tolai (Austronesian: Eastern Malayo-Polynesian: Oceanic; spoken in Papua New Guinea, East New Britain Province), which is one of the relevant substrate languages (cf. Mosel 1980), has an obligatory agreement pronoun which is coreferent with the subject noun phrase. The subject noun phrase itself is optional. In example 25, the agreement pronoun of the thirdperson plural is dia, while the corresponding subject pronoun is diat ‘they’ as in 25b: (25) Tolai (Mosel 1980: 121): a. A tarai dia vana. DEF men SM:3.PL go ‘The men went.’ b. Diat dia vana. 3.PL SM:3.PL go ‘They went.’ The agreement pronoun of the third-person singular is i in Tolai. If Tok Pisin i is derived from the English pronoun he it is very plausible that its use in the preverbal position was enhanced by the Tolai agreement pronoun i even though its functional range is not limited to the third-person singular as in Tolai. As illustrated by the following example from Mosel (1980), the structure of a simple sentence with a nominal subject shows remarkable similarities between Tolai (26a) and Tok Pisin (27b):

78

Walter Bisang

(26) Simple sentence with intransitive verb (Mosel 1980: 130): a. Tolai: A pap i pot. DEF dog 3.SG came ‘The dog came.’ b. Tok Pisin: Dok i kam. dog PM come ‘The dog came.’ The presentation of the above five Creole languages shows that the Atlantic Creoles are not radical pro-drop languages. The only language which allows pro-drop to a certain extent is Angolar. If the pro-drop behaviour of Atlantic Creoles is compared to the superstrate and substrate languages involved, it turns out that neither their superstrate languages (French, Dutch and Portuguese) nor their West African substrate languages (cf. section 3) are radically pro-drop. The option of dropping subjects in Angolar may be due to the fact that Portuguese does not need to fill its subject position, while French and Dutch do. The pro-drop behaviour of the two Pacific Creoles clearly reflects the properties of the contact languages. In the case of Tok Pisin, neither the grammar of English nor the grammar of the relevant Austronesian languages allow for the subject to be retrieved exclusively by inference from the context. In Zanboangueño, the situation is different. The Philippine substrate languages are radically pro-drop, i.e. subjects can be dropped without concomitant person-marking on the verb. In Spanish, pro-drop is possible but the subject is still marked by way of agreement on the verb. Thus, the relative freedom with which arguments can be omitted in Zamboangueño seems to be well founded in the pro-drop behaviour of its contact languages. Prodrop in Zamboangueño seems to come closest to EMSEA languages. In relative clause formation, Creole languages generally leave less room for pragmatic inference and hidden complexity than EMSEA languages as well. In Haitian, there is a clear asymmetry between subject and non-subject coreference. In the case of subject coreference, relative clauses are marked by the complementizer ki (27) (cf. the analysis of DeGraff 1992, 2007: 110), while there is no marker for object coreference (28). Non-arguments are taken up by a preposition followed by a resumptive pronoun in the relative clause (29):

Language contact between geographic and mental space

79

(27) Haitian: Relative clause with subject coreference (Muysken and Veenstra 1995: 154): poul [ki kouvri pitit li] chicken REL cover little 3.SG ‘a hen covering her chicks’ (28) Haitian: Relative clause with object coreference (DeGraff 2007: 111): Annou vote pou kandida [nou vle] a. let.1.PL vote for candidate 1.PL want DEF ‘Let’s vote for the candidate we want.’ (29) Haitian: Relative clause with non-argument coreference (DeGraff 2007: 111): Demokrasi bay pèp la prezidan [yo te vote pou li] democracy give people DEF president 3.PL PST vote for 3.SG a. DEF ‘Democracy gave the people the president they voted for.’ Angolar also has two different markers (cf. Maurer 1995: 55–58). The marker ki is used with subject coreference (30), while ma occurs with nonsubject coreference (31). In the case of non-argument coreference, the head noun is represented be a zero-pronoun preceded by a serial verb (32a) or a resumptive pronoun preceded by a preposition (32b): (30) Angolar: Relative clause with subject coreference (Maurer 1995: 55): ome si [ki ba tamba] man DEF REL go catch.fish ‘the man who left to catch fish’ (31) Angolar: Relative clause with object coreference (Maurer 1995: 55): ome si [ma m bê] man DEF REL 1.SG see ‘the man I saw’

80

Walter Bisang

(32) Angolar: Relative clause with non-argument coreference (Maurer 1995: 56): a. ome si [ma n ga taba ra ø] man DEF REL 1.SG TA work give ‘the man for whom I work’ b. ome si [ma n ga taba ku ê] man DEF REL 1.SG TA work with 3.SG ‘the man with whom I work’ Berbice Dutch marks its relative clauses with the all-purpose relativizer wati ‘what’ or the locative relativizer wanga ‘where’. The relativizer wati is a complementizer in Kouwenberg’s (1994: 371) analysis. It is optional in restricted relative clauses with subject and object coreference and in some other constructions described by Kouwenberg (1994: 373–380). In contrast, wanga is obligatory. Examples (33) and (34) illustrate relative clauses with subject and object coreference with and without overt coreference marking in the relative clause. The next two examples, 35 and 36, illustrate non-argument coreference. Locative coreference is marked by wanga (35), while any other non-argument coreference is expressed with an adposition plus a resumptive pronoun (36): (33) Berbice Dutch: Relative clause with subject coreference: a. With no overt subject marking in the relative clause (Kouwenberg 2007: 42): eni ha kεnε [wati ø das dil mεtε eni]. 3.PL have people REL HAB deal with 3.PL ‘There are people [lit. ‘they have people’] who deal with them.’ b. With overt subject marking in the relative clause (Kouwenberg 1994: 371): en jungman [wato mu-te fan ando fεntε and:3.SG find:PF one young.man REL:3.SG go-PF from diskanZi w εr]. this.side again ‘She met a young man who (he) came from over here too.’

Language contact between geographic and mental space

81

(34) Berbice Dutch: Relative clause with object coreference: a. With overt object marking in the relative clause (Kouwenberg 1994: 375): draki [wat ju kan jefjo ka]. o drak difrεn 3.SG bear different bear REL 2.SG can eat:3.SG NEG ‘It bears another fruit which you can eat (it).’ b. With no overt object marking in the relative clause (Kouwenberg 1994: 378): grasa [wato jefja ø]. das pruf kεk di HAB taste how DEF grass REL:3.SG eat:PF ‘ … (it) tastes like the grass that it feeds on.’ (35) Berbice Dutch: Relative clause with locative coreference (Kouwenberg 1994: 380): stima kum]. di gu steling jʎn birbiʃanga, [wang di DEF big jetty be river:LOC REL DEF steamer come ‘The large jetty is at the river, where the steamer comes.’ (36) Berbice Dutch: Relative clause with non-locative adpositional coreference (Kouwenberg 1994: 371): [watεkε pantε ju bot ø] di sem jεrmatoko DEF same woman:child REL:1.SG tell:PF 2.SG about ‘the same daughter that I told you about’ In Zamboangueño, relative clause formation seems to be less complex than in many other Creoles according to Lipski and Santoro (2007: 383–384). Relative clauses are usually (but not obligatorily, cf. 38) introduced by the relativizer ke or kyén. In the case of argument coreference, the coreferent noun phrase within the relative clause is represented by a zero-pronoun. Thus, there is no overt subject in 37 and 38, while there is no overt object in 39. In the case of non-argument coreference, there are various superstrateinfluenced constructions. The one illustrated in 40 is based on pied-piping (preposition/case marker kon ‘with’ plus relativizer): (37) Zamboangueño: Relative clause with subject coreference (Lipski and Santoro 2007: 383): øi na gargánta] el mana héntei [kyén ya man tunúk DEF PL people REL PST DRV be.prick. by.thorn in throat ‘people who have gotten fish spines caught in their throat’

82

Walter Bisang

(38) Zamboangueño: Relative clause with no relativizer (Lipski and Santoro 2007: 384): muntá øi na bisikléta]. Ya enkontrá yo uno polísi [ta PST meet 1.SG INDEF police PROG ride on bike ‘I met a policeman who was riding on a bike.’ (39) Zamboangueño: Relative clause with object coreference (Lipski and Santoro 2007: 383): øi], mi hermano. El hombrei , [ke ya man enkontrá tu DEF man REL PST DRV meet 2.SG my brother ‘The man whom you met is my brother.’ (40) Zamboangueño: Relative clause with non-argument coreference (Lipski and Santoro 2007: 383): El persona [kon-kyen ta kombersá tu] byen bwéno gayót DEF person with-REL PROG talk 2.SG very good EMPH ‘The person you are talking to is very nice indeed.’ In Tok Pisin, relative clauses follow their head noun. They can be marked in three different ways (Faraclas 2007: 362; also see Dutton 1985; Verhaar 1995: 215–233): (i) They immediately follow their head noun with no relativeclause marker, (ii) they are introduced by we ‘where’ or husat ‘who’ and/or (iii) they are marked by the demonstrative particle ia (from English here), which introduces and concludes the relative clause. In the case of subject coreference, the subject of the relative clause may remain unmarked or take the form of a resumptive pronoun. Example 41 on subject coreference also illustrates the different types of relative-clause marking (ø, we, ia). With object coreference, the head noun can be represented by a resumptive pronoun optionally (42). The verbal suffix -im, which marks transitivity and is related to English him, seems to have “a secondary and anaphoric function” (Faraclas 2007: 362) in this case. With adjuncts, the resumptive pronoun must follow a preposition that marks the semantic role (43). (41) Tok Pisin: Relative clause with subject coreference (Dutton 1985: 140–142): a. Mi lukim dok [(em) i ranim pik bilong mi]. 1.SG see dog 3.SG PM chase pig POSS 1.SG ‘I saw the dogs that chased my pig.’

Language contact between geographic and mental space

83

ranim pik bilong mi]. b. Mi lukim dok [we i 1.SG see dog REL PM chase pig POSS 1.SG ‘I saw the dogs that chased my pig.’ ranim pik bilong c. Mi lukim dok [ya (em / we) i 1.SG see dog REL 3.SG REL PM chase pig POSS ya] mi 1.SG REL

‘I saw the dogs that chased my pig.’ (42) Tok Pisin: Object coreference with no resumptive pronoun (Verhaar 1995: 224): Ol i mas wasim olgeta klos [ol i putim asde]. 3.PL PM must wash all/every clothes 3.PL PM wear yesterday ‘They must wash the clothes they wore the day before.’ (43) Tok Pisin: Non-argument coreference (Verhaar 1995: 228): Mi no laik kisim dispela olgeta pasin kranki [yupela i 1.SG NEG like accept DEM all customs strange 2.PL PM tok long en]. talk about 3.SG ‘I don’t accept all these strange customs you are talking about.’ Relative clauses in Tok Pisin with no overt relativizer are remarkably similar to those in Tolai. Example 44 illustrates subject coreference, while example 45 is on object coreference: (44) Tolai: Subject coreference (Siegel 2008: 187): i ga mut-kutu pa ra ul=i ra luluai 3.SG TA cut-sever COMPL DEF head=POSS DEF chief [i ga ubu na na=na]. 3.SG TA kill DEM mother=POSS:3.SG ‘he cut off the head of the chief [who had killed his mother].’

84

Walter Bisang

(45) Tolai: Object coreference (Mosel 1980: 113): To Karvuvu i gire ra en [To Kabinana i vut meme]. To Karvuvu PM see DEF fish To Kabinana PM come with:it ‘To Karvuvu saw the fish which To Kabinana brought.’ Based on the similarity of Tolai relative clauses with Tok Pisin, Siegel (2008: 187–188) claims that the relativizer we in Tok Pisin is not due to influence from the substrate languages. As he argues, it is rather the consequence of “renewed contact with English or language-internal expansion” (Siegel 2008: 188). More generally, the basic structure of the relative clause in Tok Pisin is related to the substrate, while additional marking strategies (the use of we and/or ia) emerged later. The degree to which the function of the head noun in the relative clause must be inferred from context is lower in each of the five Creoles discussed above than in EMSEA languages. Non-argument coreference is clearly expressed by a construction that differs from the one that is employed with argument coreference. In Haitian and Angolar, there is an additional subject/object asymmetry which determines the selection of the relative-clause marker (ki vs. ø in Haitian, ki vs. ma in Angolar). This reflects the properties of the superstrate and the substrate languages in which the surface structure of the relative clause clearly differs between subject coreference and object coreference. In Berbice Dutch, the markedness patterns do not necessarily support subject/object asymmetry even though it is possible to make a clear-cut distinction if explicit marking is needed. This also applies to the Pacific Creoles of Zamboangueño and Tok Pisin. The following table summarizes the findings on pro-drop and coreference marking in relative clause formation: Creole

Pro-drop

Relative clause formation and distinctions in coreference marking

Haitian

–

Argument/non-argument distinction, subject/object asymmetry (ki vs. ø)

Angolar

+/–

Argument/non-argument distinction, subject/object asymmetry (ki vs. ma)

Berbice Dutch

–

Argument/non-argument distinction, of subject vs. object coreference

Tok Pisin

–

Argument/non-argument distinction, optional distinction of subject vs. object coreference

Language contact between geographic and mental space Zamboangueño

+

85

Argument/non-argument distinction, optional distinction of subject vs. object coreference

Table 2: Summary of hidden complexity in five Creole languages

As can be seen in table 2, none of the languages reaches the same degree of hidden complexity as it is found in Chinese. Chinese shows more hidden complexity than Creoles and thus allows even simpler surface structures than these languages. The Creole that comes closest to Chinese is Zamboangueño with its pro-drop properties and its facultative marking of the subject-vs.-object distinction in relative clauses. As pointed out above, these properties are in harmony with the properties of the Philippine substrate languages (and to a certain extent even with Spanish as a weak pro-drop language). It thus seems that Creoles are the product of contact among languages which belong to a certain type of overt marking (e.g. English/ French/etc. and West African languages). Speakers may reduce the complexity of their language to a considerable extent in extreme situations of contact, but they do not seem to go beyond the limits of hidden complexity that exists in the contact languages involved. Creoles may thus represent the limits of what can be reduced/omitted from the perspective of the contact languages involved, but there are languages such as the ones of EMSEA whose grammatical systems can produce even less marked, simpler surface structures which potentially leave more to pragmatic inference than Creoles do.

6.

Contact in mental space and the change of discourse structures

The gradual functional approximation of the Chinese bèi-passive to passives of Western languages like English or German as well as the more frequent use of overt nominal elements in the subject and object positions are wellknown instances of the convergence of discourse structures through translation. Both phenomena will be discussed in this section. It will also be shown that there is no full convergence. The situation of virtual contact only leads to a certain degree of convergence and creates many additional innovations at the same time. Chinese passives are formed with the marker bèi (Chao 1968; Li and Thompson 1981: 492–508, Huang, Li, and Li 2009: 112–152). They differ from passives as they are known in English or German in at least the following two ways: (i) They occur in two forms called the canonical passive and

86

Walter Bisang

the indirect passive and (ii) they imply that the events they mark are adversative to the subject NP. In the canonical passive, the patient NP takes the subject position and the actor NP follows bèi if there is one (46a vs. 46b). Indirect passives are characterized by an additional experiencer NP which takes the subject position, while the actor NP follows bèi and the patient NP remains in the postverbal position (46c): (46) a. T¯a bèi mˇuq¯ın mà-le. s/he PASS mother scold-PFV ‘S/He was scolded by mother.’ b. T¯a bèi mà-le. s/he PASS scold-PFV ‘S/He was scolded.’ c. Indirect passive (Huang, Li, and Li 2009: 140): Zh¯angs¯an bèi tˇufˇei dˇasˇı-le bàba. Zhangsan PASS bandit kill-PFV father ‘Zhangsan had his father killed by bandits.’ As can be seen in example 46c, the bèi-construction conveys adversative meaning, i.e. Zhangsan is negatively affected by the fact that bandits killed his father. This does not only apply to indirect passives but also to direct passives: (47) Chinese (Li and Thompson 1981: 494): g¯ongs¯ı chèzhí-le. T¯a bèi he PASS company fire-PFV ‘He was fired by (his) company.’ Due to their adversative meaning, the use of Chinese passives tends to be much more restricted than in English. In the last fifty years, however, the bèiconstruction is no longer strictly associated with adversative meaning. As Chao (1968) points out, the increasing number of translations from European languages generated a kind of “translatese” in which adversative meaning has become obsolete. “But recently, from translating foreign passive verbs, ‘by’, or some equivalent in a Western language, is mechanically

Language contact between geographic and mental space

87

equated to bey and applied to verbs of favorable meanings …” (Chao 1968: 703). In the following example, the event of being elected is certainly positive for the subject and has no negative connotations: (48) Li and Thompson (1981: 497): rénmín xuˇan zuò dàibiˇao le. Zh¯angs¯an bèi Zhangsan PASS people elect serve.as representative PF ‘Zhangsan has been elected by the people to be (their) representative.’ The loss of adversative meaning is due to contact by writing, as translations from European languages neutralized the meaning of the bèi-construction. Whether its meaning is fully neutralized to the extent that any event irrespective of its emotional content can be passivized is still debatable. In spite of this, it is quite remarkable that the bèi-construction is nowadays even used as a syntactic test in many linguistic publications by linguists with native competence in Chinese. In the following example from Li (1990), the bèi-construction is used for proving the existence of subject/object asymmetry in Chinese. While the actor of the predicate jiˇancha´ ‘examine’ can undergo tough-movement (49a), the patient needs to be passivized first by the bèi-construction (49b): (49) Chinese: (Li 1990: 129): a. Zhè ge y¯ısh¯eng hˇen nán/ róngyì jiˇanchá-wán Lˇısì ma? this CL doctor very difficult/ easy examine-finish Lisi Q ‘Is this doctor difficult/easy to examine Lisi?’ (Is it difficult for the doctor to examine Lisi?) b. bèi zhè ge y¯ısh¯eng jiˇanchá-wán ma? Lˇısì hˇen nán/róngyì Lisi very difficult/easy BEI this CL doctor examine-finish Q ‘Is Lisi difficult/easy to be examined by the doctor?’ (Is it difficult/easy to be examined by the doctor for Lisi?)

88

Walter Bisang

Even though the loss of the adversative meaning moves the bèi-construction closer to the passive, there are still a number of differences. One of them is the existence of an indirect passive which is incompatible with the passive construction as one finds it in English or German. Another one is long-distance passivization, in which the dependency between the subject (Zha ng Sa n) and its trace is unbounded: (50) Chinese: Long-distance passive (Huang, Li, and Li 2009: 132): Zh¯angs¯an bèi Lˇısì pài jˇınggchá zhu¯azˇou-le. Zhangsan Pass Li Si send police arrest-PFV ‘Zhang San was “sent-police-to-arrest” by Li Si.’ The absence of obligatory arguments is a second property of Chinese which is undermined to a certain extent by written contact. A constructed example of how inferrable arguments are omitted was given in example 1 above. The following example is from a short story written in 1934 by Lu Xun (1881–1936), an eminent author of modern Chinese literature. It starts with the introduction of the two protagonists (Gongsun Gao and Mozi, a famous philosopher of the 5th century BC), which remain unmarked in the subject and object positions of the subsequent events. In the case of the predicate zài jia ‘be at home’, both protagonists are potential arguments. Since the reader knows that Gongsun Gao wants to see Mozi, it is reasonable to infer that it is Mozi who stays at home. The subject of the next predicate jiàn bu zháo ‘be unable to meet’, however, is Gongsun Gao, since the reader knows that he is the one who wants to meet Mozi. Thus, not even a change of subject is a sufficient reason for the overt marking of arguments as long as the situation is clear from the context. Finally, the same constellation of Gongsun Gao meeting Mozi is reported in the next event, again with overt marking of subject and object. (50) Chinese: Lack of overt arguments (from Lu Xun, F¯eig¯ong ‘Opposing aggression’):

zhˇao Mò-zˇıj, Zˇıxià de túdì G¯ongs¯un G¯aoi lái Zixia POSS disciple Gongsun Gao come seek Mo-Master

Language contact between geographic and mental space

89

øi jiàn zài ji¯a, yˇıjing hˇaojˇı huí le, øj zˇongshi bú already several times PF never NEG be.at home see

huòzhˇe dì-wˇu huí ba, zhè bu zháo øj. Dàyu¯e shì dì-sì NEG reach about be fourth or fifth time EXCL this cái qiàqiˇao zài ménkˇou øi yùjiàn øj , … finally by.chance at doorway meet ‘Gongsun Gaoi, a disciple of Zixia, was looking for master Moj for several times and [hej] was never at home, so [hei] was unable to meet [himj]. It was at about the fourth or the fifth time that [hei] met [himj] in the doorway …’ From the perspective of pragmatic implicatures (Levinson 2000; Huang 1994), the use of a pronoun in instances where zero-marking is possible triggers different interpretations. If zero-marking is associated with coreferent interpretation by default, the use of a more prolix pronominal form induces the complementary M-implicature of disjoint interpretation (cf. Huang 1994: 128–146). In the case of the sentence pair in 51, one would expect coreference between Xiao Ming and the ø-pronoun in 51a as the most informative default interpretation (I-implicature), while the use of the more marked pronoun in 51b induces a disjoint interpretation through an M-implicature. Since pragmatic rules can always be violated, this is not the whole story. Other interpretations are indicated in the translation of 51. This is due to various factors. Huang (1994: 137) mentions background assumptions, non-natural meaning, semantic entailments, context and priority pragmatic inferences. (51) Chinese (Huang 1994: 134–135): a. Xiˇao Míng y¯ı jìn w¯u, ø Xiao Ming as.soon enter room

jiù bˇa mén then take door

gu¯anshàng-le. close-PFV ‘As soon as Xiao Ming1 entered the room, he1/2/I/you/they closed the door.’

90

Walter Bisang

b. jiù bˇa mén Xiˇao Míng y¯ı jìn w¯u, t¯a Xiao Ming as.soon enter room 3.SG then take door gu¯anshàng-le. close-PFV ‘As soon as Xiao Ming1 entered the room, he2 closed the door.’ In more recent literary texts, some authors tend to use overt arguments more often even in instances in which the inference of the relevant referent is unproblematic. One example is Yu Hua (*1960). The following passage from his famous novel Xi¯ongdì ‘Brothers’, published in 2005, is a representative example of the frequency with which he explicitly marks the subject. Needless to say, the use of the pronoun t¯a ‘s/he’ does not M-imply disjunctive reading in 52 – the referent of this pronoun is always Li Lan, who is mentioned at the beginning of the passage: (52) The use of overt pronouns (Yu Hua 2005, Xi¯ongdì ‘Brothers’. Vol. I, chapter 25):

Lˇı Lán huí

dào ji¯a zh¯ong, zài jìngzi qián zˇıxì ø Li Lan return to house inside at mirror front carefully kàn-le zìjˇı de t¯urán zìjˇı, t¯a yˇe bèi watch/examine-PFV self she also PASS self MOD sudden t¯a yˇou-le yí ge c¯anglˇao xià-le yí tiào. ránhòu oldness scare-PFV one jump after.that she have-PFV one CL t¯a juéde zìjˇı zhùjìn-le y¯ıyuàn yˇıhòu, búyàng yùgˇan, bad premonition she feel self enter-PFV hospital after t¯a yˇıjing xˇı-diào-le kˇenéng ø ch¯u-bu-lái-le. maybe get.out-NEG-come-PFV she already wash-wipe.off-PFV

mˇashàng mˇan-tóu de su¯an chòuwèi, t¯a méiyˇou entire-head MOD acid/sour smell she NEG:PFV immediately yòu zhù-le jˇı ti¯an. qù y¯ıyuàn, t¯a zài ji¯a-li go hospital she at house-in again/still stay-PFV some day

Language contact between geographic and mental space

91

‘When Li Lan was back in her home and carefully examined herself in the mirror she was scared by the sudden oldness [of her features]. After that, she had a bad premonition and she had the feeling that she may not get out of the hospital once she got in there. She had already washed the sour smell out of her hair [but] she did not go to the hospital immediately and she still stayed in [her] house for a few days.’ The increasing frequency of overt arguments seems to be due to the influence of languages like English. Most likely, the process which initiated the overt expression of subjects was once more the translation of written texts. The more frequent use of overt arguments in Chinese texts produces the effect of “translatese” and of openness to new ways of expression as they are found in Western languages. However, this does not mean that subject marking has become obligatory in Chinese. Even in Yu Hua’s texts, there are passages with many hidden subjects. The following example from the same novel starts with the pronoun t¯a ‘he’, while all the subsequent events have no overt subject: (53) Chinese (Yu Hua 2005, Xi¯ongdì ‘Brothers’. Vol. I, chapter 22):

T¯ai bù

zh¯ıdào zìjˇı yào gàn shénme, øi zhˇı zhı¯dào zìjˇı he NEG know SELF FUT do what only know SELF

le, øi jiù zhˇao ge zài zˇou-lái zˇou-qù, øi lèi PROG walk-come walk-go tired PF then look.for CL øi kˇe le, øi jiù qù øi dìf¯ang øi zuò-xià-lái, place sit-go.down-come be.thirsty PF then go

le, øi jiù nˇıng-k¯ai mˇou ge shuˇılóngtóu, øi è turn.on-open any CL water.tap hungry PF then øi ch¯ı jˇı kˇou lˇeng fàn shèng cài. huí-ji¯a return-home eat a.few mouthful cold rice leftover vegetables

‘He (Li Guangtou) didn’t know what he would have to do, [he] only knew that he was moving this way and that way. When [he] got tired [he] looked for a place to sit down, when [he] got thirsty, [he] turned on the next available water tap, when [he] got hungry [he] went home and ate a few bites of cold rice and leftover vegetables.’

92

Walter Bisang

It is hard to say why there are so many overt subjects in example 52, while they remain mostly unexpressed in example 53. A lot depends on discourse and style. The lack of overt marking in 53 may have to do with the creation of a discourse paragraph. The activities of Li Guangtou are all seen as different consequences of his disorientation and unrest. This overarching notion provides the coherence between the individual events which are linked by the specific emotional state of a single person in the function of the subject. In contrast, the events in example 52 are combined in a less homogenous way. There is an explicit temporal distance between Li Lan’s look into the mirror and her subsequent bad premonition. The next action of not returning to the hospital immediately stands in some causal relation to that premonition. In spite of this, the frequency with which pronouns are used in this context is rather remarkable and can be seen as a contact-induced innovation. In conclusion, one can say that the loss of the adversative meaning of the bèi-construction and the increasing frequency of overt argument marking are due to contact by writing. Given the differences between the linguistic systems of languages like English and Chinese and the sociocultural distance connected to contact by writing, the total structural adaptation of Chinese to the model structures in English and other European languages is unlikely. What happens is the emergence of new structures with certain affinities to the properties of the respective model structures plus completely different properties. In the case of the bèi-construction, the differences mentioned are the existence of the indirect passive and long-distance passivization. In the case of the increasing overt marking of arguments, individual authors have a large field of individual freedom for establishing narrative coherence and for experimenting with new stylistic forms which do not exist in English and did not exist in that form in older Chinese texts either.

Language contact between geographic and mental space

7.

93

The role of mental space in the emergence of morphological rules in Vietnamese

Even though Vietnamese has the reputation of being a prototypical isolating language, its very rich derivational morphology is characterized by specific rules which do not operate in syntax. As will be shown in this section, this is due to a special type of contact which is based on “contact by writing” (Bisang 2001), i.e. a type of contact that takes place in a mental space which is independent of concrete geographic contact. To understand this, it is necessary to look back at the history of Vietnam before dealing with the linguistic facts. Vietnam was ruled by China between 179 BC and 938 AD. There are two periods within that rather long time span in which Chinese words were copied extensively into Vietnamese: • Period of Old Sino-Vietnamese (pronunciation of Nanbeichao, 420–589 AD) • Period of Sino-Vietnamese (7–10th centuries AD) The number of Chinese loan words that made their way into Vietnamese under Chinese rule was remarkable, but it was even more remarkable that a considerably greater number of Chinese words were introduced at the time of indirect contact, i.e. after the 10th century AD when Vietnam was governed by independent dynasties. The reason for this was that the Chinese government system still formed the model for Vietnam and that Confucianism replaced Buddhism in the 13th century. Thus, the cultural prestige of Chinese civilization supported its linguistic influence even without intensive contact among large numbers of speakers. Proof of this indirect written contact comes from the pronunciation of the Chinese characters. The Sino-Vietnamese pronunciation of Chinese characters was established at the time of direct contact, but it was also used with those words which came into Vietnamese later, at a time when the Chinese pronounced those words differently. The massive influx of Chinese loan words did not only enrich the lexicon of Vietnamese, it also introduced word-formation patterns of Chinese which did not exist in Vietnamese before. The remainder of this section will briefly describe how the asymmetry between morphological and syntactic rules developed from contact by writing in Vietnamese. Vietnamese and Chinese noun phrases are characterized by different word order. While Vietnamese modifiers follow their head nouns (54a), Chinese modifiers precede them (54b):

94

Walter Bisang

(54) a. Vietnamese: b. Chinese: [MODIFIER – HEAD ] [HEAD – MODIFIER ] xe mói xı¯n ch¯e [ ] car new new car ‘a new car’ ‘a new car’ Chinese applies the same order of elements in word formation, i.e. the modifier precedes its head. In example 55, the lexeme ji¯a, which also occurs as an independent word with the meaning of ‘house, family’, takes the word final position. In word formation, it was first used to denote collectives, then to denote individuals belonging to that collective. Nowadays it is used mostly as a class noun (CN) for different types of scholars and intellectuals (the difference between 55a and 55b will be relevant later when Chinese is compared with Vietnamese): (55) a.

b.

zuò-ji¯a [make/write-CN] ‘writer’, zhu¯an-ji¯a [special-CN] ‘specialist’, k¯exué-ji¯a [science-CN] ‘scientist’ f¯amíng-ji¯a [invent-CN] ‘inventor’, jánji¯u-ji¯a [research-CN] ‘researcher’

Even morphemes whose original Chinese meaning is no longer present in their contemporary derivational function take the word-final position. Thus, the morpheme -zhˇuyì, which corresponds to English -ism, consists of the two elements zhˇu ‘owner, head, leader/leading’ and yì ‘meaning’. It follows its morphological basis: (56)

shèhuì-zhˇuyì [society-ism] ‘socialism’ xiànshí-zhˇuyì [reality/real-ism] ‘realism’

In Vietnamese, both morpheme orders are possible today. Heads with a Vietnamese lexical item get postnominal modifiers in most instances, while heads borrowed from Chinese get prenominal modifiers. This is very nicely illustrated by the two heads with the meaning of ‘house’: nhà is Vietnamese, and gia is Chinese and corresponds to ji¯a in 55. Some of the Chinese words in 57 have retained their Sino-Vietnamese element gia when borrowed into Vietnamese (57a). These words correspond to 55a with the class noun gia in the word-final position. Some other words that correspond to 55b take the Vietnamese head nhà, which occurs word-initially (57b):

Language contact between geographic and mental space

95

(57) a. With Chinese head gia: cf. (55a): tác gia [make/write-CN] ‘creative person, artist’ chuyên gia [special-CN] ‘specialist’ khoa ho.c gia [science-CN] ‘scientist’ b. With Vietnamese head nhà: cf. (55b): nhà phát minh [CN-invent] ‘inventor’ nhà nghiên cúu [CN-research] ‘researcher’ Some Sino-Vietnamese elements of word formation do not follow the same order as in Chinese. This is the case with the marker chu ngh˜ıa for -ism. A comparison of 56 with 58 shows that chu ngh˜ıa is word initial in Vietnamese even though its corresponding Chinese form is word final: (58) Vietnamese: chu ngh˜ıa xã hô.i ‘socialism’ chu ngh˜ıa hiê.n thu. c ‘realism’ Before language contact with Chinese, the sequence [MODIFIER -HEAD ] was unknown in Vietnamese. It was introduced through copying of wordformation processes as illustrated in 55 and 56 from Chinese. Before that time, word formation was as homogeneously [HEAD -MODIFIER ] in Vietnamese as it was [MODIFIER -HEAD ] in Chinese, running parallel with syntax in both languages. It was only by contact with the Chinese pattern of word formation that a new type of morphological rule was introduced which contradicted the [HEAD -MODIFIER ] serialization of syntax. In most recent times, a few Vietnamese words seem to allow both options without any difference in meaning. This is the case with the morpheme truong ‘group leader’, which corresponds to Chinese zhˇang: (59) Words with both morpheme orders: a. truong lóp / lóp truong [leader-group / group-leader] ‘group leader’ b. truong viê.n / viê.n truong [leader-institute / institute-leader] ‘director of an institute’ Finally, in the case of chu ngh˜ıa, the two morpheme-order options are sometimes used to distinguish between the nominal and the modifying function of a word. If the concept is used nominally, we get [chu ngh˜ıa-modifier] as in 60a, while if the concept itself modifies a noun, morpheme order is in-

96

Walter Bisang

versed and we get [modifier-chu ngh˜ıa] as in 60b. The same applies to tính ‘nature’ in 61a vs. 61b. (60) a. Nominal function: chu ngh˜ıa xã hô.i ISM-society ‘socialism’

b. Modificational function: xã hô.i chu ngh˜ıa society-ISM ‘socialistic’

(61) a. Nominal function: tính thiê.n NATURE -good ‘goodness’

b. Modificational function: thiê.n tính good-NATURE ‘of a good character’

c. Nominal function: tính ác NATURE -bad ‘badness, evil’

d. Modificational function: ác tính bad-NATURE ‘of a bad/evil character’

8.

Conclusion

The aim of the present paper was to show that language contact matters even beyond geographically contiguous areas. Language change and structural convergence through contact take place in the mind/brains of individuals. Their ability to use linguistic structures from two or more languages remains the same, regardless of whether they know them from concrete situations of face-to-face contact, from distant contact through written sources or – more recently – through new electronic media. The effects of non-contiguous contact can sometimes be described in terms of individual typological features, as in the case of Vietnamese compounding and the emergence of head-final structures in addition to the older head-initial structures (section 7). In many instances, however, contact goes beyond individual features and concerns more basic properties. As this paper has shown, hidden complexity is a good example of such a deep-rooted property which is characterized by a high degree of stability in the sense that it is hard to be acquired through contact. In the case of contact through speaker mobility as represented by Creoles (sections 3 to 5), it was shown that they preserve their hidden complexity to a large extent despite of great geographical distances between the substrate and the superstrate languages involved. Creoles are the product of contact among languages which belong to a certain type of overt marking (e.g. English/

Language contact between geographic and mental space

97

French/etc. and West African languages in the case of most Atlantic creoles) and they may represent the limits of what can be reduced or omitted from these languages in socially extreme contact situations. Thus, Creoles are just normal languages that evolved under certain contact situations – their structures do not necessarily lead us to an earlier stage of language evolution. In the case of contact based on virtual space as discussed regarding the influence of Western languages on contemporary Chinese, it was shown that hidden complexity can be reduced in certain contexts (section 6). This contact situation is completely different from that of Creoles. The speakers of Chinese still speak their language but they make use of certain properties of Western languages for stylistic purposes. Since the use of structures showing more overt complexity has not become an obligatory part of the language system, hidden complexity is not abandoned – there are only additional options which enrich the structural inventory of the language as a whole. At a further stage, this leads to structural innovations whose detailled properties are probably cross-linguistically unique. This is illustrated by the bèi-construction in Chinese. Contact matters, even beyond areal contiguity. Another question discussed in this paper was to what extent the effects of language contact can be quantified on the basis of individual features. The quantification of pragmatic effects as they show up in hidden complexity may well be a challenge to linguistic typology which equals the challenge of non-contiguous contact.

References Bakker, Dik 2011: Language sampling. In: Jae-Jung Song (ed.), The Oxford Handbook of Linguistic Typology, 100–127. Oxford: Oxford University Press. Bamgbos.e, Ayò. 1967: A Short Yoruba Grammar. Ibadan: Heinemann. Bisang, Walter 1996: Areal typology and grammaticalization: Processes of grammaticalization based on nouns and verbs in East and mainland South East Asian languages. Studies in Language 20 (3): 519–597. Bisang, Walter 2001: Syntax/morphology asymmetry in Vietnamese – a consequence of contact by writing between Vietnamese, Chinese and Standard Average European Languages. In: Birgit Igla & Thomas Stolz (eds.), “Was ich noch sagen wollte …” A Multilingual Festschrift for Norbert Boretzky on Occasion of his 65th Birthday, 189–201. Berlin: Akademie-Verlag. Bisang, Walter 2004: Dialectology and typology – an integrative perspective. In: Bernd Kortmann (ed.), Dialectology Meets Typology. Dialect Grammar from a Cross-Linguistic Perspective, 11–45. Berlin/New York Mouton de Gruyter. Bisang, Walter 2006a: Contact-induced convergence: typology and areality. In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, Vol. 3, 88–101. Oxford: Elsevier.

98

Walter Bisang

Bisang, Walter 2006b: Linguistic areas, language contact and typology: some implications from the case of Ethiopia as a linguistic area. In: Yaron Matras, April McMahon & Nigel Vincent (eds.), Linguistic Areas. Convergence in Historical and Typological Perspective, 75–98. Hampshire: Palgrave MacMillan. Bisang, Walter 2008: Grammaticalization and the areal factor: the perspective of East and mainland Southeast Asian languages. In: María José López-Couso & Elena Seoane (eds.), Rethinking Grammaticalization. New Perspectives, 15–35. Amsterdam/ Philadelphia: John Benjamins. Bisang, Walter 2009: On the evolution of complexity – sometimes less is more in East and mainland Southeast Asia. In: Geoffrey Sampson, David Gil & Peter Trudgill (eds.), Language Complexity as an Evolving Variable, 34–49. Oxford: Oxford University Press. Bisang, Walter & Remi Sonaiya 1999: The function of the high tone syllable in Yoruba. The Journal of African Languages and Linguistics 20: 1–19. Chao, Yüan Ren 1968: Grammar of Spoken Chinese. Berkeley/Los Angeles: University of California Press. Croft, William A. 1995: Autonomy and functionalist linguistics. Language 71: 490–532. Croft, William A. 2000: Explaining Language Change. An Evolutionary Approach. Essex: Pearson Education. Dahl, Östen 2004: The Growth and Maintenance of Linguistic Complexity. Amsterdam/ Philadelphia: Benjamins. Déchaine, Rose-Marie 1993: Predicates across categories: Towards a category-neutral syntax. Doctoral dissertation, University of Massachusetts. Amherst, Mass: GLSA. DeGraff, Michel Frederic 1992: Haitian null subjects revisited. In: Claire Lefebvre (ed.), Travaux de recherche sur le Créole Haitien 11, 59–74. Québec: Université du Québec à Montréal, Département de Linguistique. DeGraff, Michel Frederic 1993: Is Haitian a pro-drop language? In: Rancis Byrne & John Holm (eds.), Atlantic Meets Pacific: A Global View of Pidginization and Creolization, 71–90. Amsterdam/Philadelphia: Benjamins. DeGraff, Michel Frederic 2007: Kreyòl Ayisyen, or Haitian Creole. In: John Holm & Peter L. Patrick (eds.), Comparative Creole Syntax. Parallel Outline of 18 Creole Grammars, 101–126. London: Battlebridge. Déprez, Viviane 1994: Haitian Creole: a pro-drop language? Journal of Pidgin and Creole Languages 9: 1–24. Dryer, Matthew S. 1989: Large linguistic areas and language sampling. Studies in Language 13: 257–292. Dryer, Matthew S. 1992: The Greenbergian word order correlations. Language 68: 81–138. Dryer, Matthew S. 2005: Chapter 81. Order of subject, object and verb. In: Martin Haspelmath, Matthew S. Dryer, David Gil & Bernard Comrie (eds.), The World Atlas of Language Structures. Oxford: Oxford University Press. Dutton, Thomas 1985: A New Course in Tok Pisin. Canberra: Australian National University. [Pacific Linguistics, D.67]. Faraclas, Nicholas 2007: Tok Pisin (Creole English). In: John Holm & Peter L. Patrick (eds.), Comparative Creole Syntax. Parallel Outline of 18 Creole Grammars, 355–372. London: Battlebridge.

Language contact between geographic and mental space

99

Fitch, W. Tecumseh, Marc D. Hauser & Noam Chomsky 2005: The evolution of the language faculty: clarifications and implications. Cognition 9: 179–210. Forman, Michael 1972: Zamboangueño with grammatical analysis. Ph.D. dissertation, Cornell University. Givón, Talmy & Masayoshi Shibatani (eds.) 2009: Syntactic Complexity. Diachrony, Acquisition, Neuro-Cognition, Evolution. Amsterdam/Philadelphia: Benjamins. Haiman, John 1983: Iconic and economic motivation. Language 59: 781–819. Haspelmath, Martin, Matthew S. Dryer, David Gil & Bernard Comrie (eds.) 2005: The World Atlas of Language Structures (WALS). Oxford: Oxford University Press. Hauser, Marc D., Noam Chomsky & W. Tecumseh Fitch 2002: The faculty of language: What is it, who has it, and how did it evolve? Science 298: 1569–1579. Holm, John & Peter L. Patrick (eds.) 2007: Comparative Creole Syntax. Parallel Outlines of 18 Creole Grammars. London: Battlebridge. Huang, C.-T. James, Yen-Hui Audrey Li & Yafei Li 2009: The Syntax of Chinese. Cambridge: Cambridge University Press. Huang, Yan 1994: The Syntax and Pragmatics of Anaphora. A Study with Special Reference to Anaphora. Cambridge: Cambridge University Press. Janssen, Dirk P., Balthasar Bickel & Fernando Zuñiga 2006: Randomization tests in language typology. Linguistic Typology 10: 419–440. Kouwenberg, Silvia 1994: A Grammar of Berbice Dutch Creole. Berlin/New York: Mouton de Gruyter. Kouwenberg, Silvia 2007: Berbice Dutch (Creole Dutch). In: John Holm & Peter L. Patrick (eds.), Comparative Creole Syntax. Parallel Outline of 18 Creole Grammars, 25–52. London: Battlebridge. Lefebvre, Claire & Anne-Marie Brousseau 2002: A Grammar of Fongbe. Berlin/New York: Mouton de Gruyter. Levinson, Stephen C. 2000: Presumptive Meanings. The Theory of Generalized Conversational Implicatures. Cambridge, Mass.: The MIT Press. Li, Audrey Yen-Hui 1990: Order and Constituency in Mandarin Chinese. Dordrecht: Kluwer. Li, Charles N. & Sandra A. Thompson 1981: Mandarin Chinese. A Functional Reference Grammar. Berkeley, Los Angeles and London: University of California Press. Lipski, John M. & Maurizio Santoro 2007: Zamboangueño Creole Spanish. In: John Holm & Peter L. Patrick (eds.), Comparative Creole Syntax. Parallel Outline of 18 Creole Grammars, 373–396. London: Battlebridge. Maurer, Philippe 1995: L’angolar. Un créole afro-portugais parlé à São tomé. Hamburg: Buske. McWhorter, John H. 2001: The world’s simplest grammars are creole grammars. Linguistic Typology 5: 125–166. McWhorter, John H. 2005: Defining Creole. Oxford: Oxford University Press. Miestamo, Matti, Kaius Sinnemäki & Fred Karlsson (eds.) 2007: Language Complexity. Typology, Contact, Change. Amsterdam/Philadelphia: John Benjamins. Muysken, Pieter & Tonjes Veenstra 1995: Haitian. In: Jacques Arends, Pieter Muysken & Norval Smith (eds.), Pidgins and Creoles. An Introduction, 153–164. Amsterdam/Philadelphia: Benjamins. Neeleman, Ad & Kriszta Szendröi 2007: Radical pro drop and the morphology of pronouns. Linguistic Inquiry 38: 671–714. Nichols, Johanna 1992: Linguistic Diversity in Space and Time. Chicago and London: The University of Chicago Press.

100

Walter Bisang

Ning, Chunyan 1993: The overt syntax of relativization and topicalization. Ph.D. dissertation, University of California at Irvine. Prince, Alan & Paul Smolensky 2004: Optimality Theory: Constraint Interaction in Generative Grammar. Oxford: Blackwell. Rijkhoff, Jan & Dik Bakker 1998: Language sampling. Linguistic Typology 2: 263–314. Sapir, Edward 1921: Language. An Introduction to the Study of Speech. New York: Harcourt, Brace and Company. Siegel, Jeff 2008: The Emergence of Pidgin and Creole Languages. Oxford: Oxford University Press. Sonaiya, Remi 1990: Yoruba vowel deletion. Linguistique Africaine 5: 77–93. Tomlin, Russell S. 1986: Basic Word Order: Functional Principles. London: Croom Helm. Verhaar, John W. M. 1995: Toward a Reference Grammar of Tok Pisin. An Experiment in Corpus Linguistics. Honolulu: University of Hawaii Press. Wichmann, Søren & Eric W. Holman 2009: Assessing Temporal Stability for Linguistic Typological Features. Munich: Lincom Europa. Wolfram, Stephen 2002: A New Kind of Science. Champaign, IL: Wolfram Media Inc.

The notion of space in linguistic typology

101

Ekkehard König

Commentary: The notion of space in linguistic typology

1.

The space of variation

The most elementary use of the notion of space in linguistic typology is a metaphorical one: The goal of typology is generally defined as the study of unity within diversity or, more specifically, as mapping out the space and the limits of variation between languages. This “space of variation” is generally represented by a one-dimensional diagram, describing a two-term opposition (prepositions vs. postpositions; tonal contrasts vs. no tonal contrasts) or a scale with different instances of a category (number of genders/noun classes, of cases, etc.). A more sophisticated way of representing the space of variation, however, is a hierarchy in which the options found in the interaction of variant properties found in the world’s languages are ordered on the basis of implicational relations. A diagram of type (1) is used to express a chain of implications in the interaction of two categories: If a language has the option X in some position of this scale it also has all the options further to the left: (1) A > B > C > D > E The hierarchies in (2) and (3) are two well-known examples of representing the space of variation in this way: (2) is a statement about the interaction of reflexive markers with the category “person” and states that if a language has a special reflexive marker distinct from personal pronouns for some person it will also encode reflexivity for all persons further to the left. English, Russian and Yiddish are well-known examples with reflexive markers for all persons (e. g. myself, yourself, him-/herself), whereas most Germanic and Romance languages only have a special reflexive marker for the 3rd person (e.g. German sich): (2) Interaction of person and reflexive marking: 3 > 2 > 1

102

Ekkehard König

The (extended) Hierarchy of Animacy (Individuation) in (3), which reflects the anthropocentric nature of language, is a more important example of characterizing the space of variation in this way: (3) The (Extended) Animacy Hierarchy (Silverstein 1976) 1st, 2nd pronouns > 3rd pronouns > proper names > human common nouns > non-human animate common nouns > inanimate common nouns > place > time There is a large number of variant properties that can be characterized with the help of this scale: number distinctions, for example, tend to occur with expressions on the higher end of the hierarchy in those languages where they are not obligatory (Japanese, Mandarin, etc.). Furthermore, important aspects or ordering nominal arguments and of special case marking can be stated in terms of this scale in a wide variety of languages. In the three articles of this section this notion of “space”, which always relates to categories or constructions and their interactions rather than to languages as wholes, is taken for granted as important background but not raised for detailed discussion.

2.

Geographic space and language contact

Languages can be similar as a result of genealogical descent (deriving from the same proto-language) but also as a result of geographic proximity and extensive language contact. In order to capture the similar and invariant properties of languages which are due to general (functional, cognitive) principles of language organization the historical causes of such similarity have to be clearly identified and separated from the a-historical factors. This is the topic of Michael Cysouw’s contribution, whose main concern it is to disentangle geographic factors from genealogical ones on the basis of sophisticated and innovative statistical methods. In these analyses the spatial notion of distance plays an important role, which by way of metaphorical transfer is not only applied to geography, but also to genealogy and typological similarity or difference. With the help of precisely defined notions of genealogical distance (same genus, same family or unrelated) and typological distance (a function of the number of characteristics that are different between two languages), Cysouw is able to demonstrate clear correlations between genealogical distance and typological distance, on the one hand, and between geographical distance and typological distance, on the other. In this way corre-

The notion of space in linguistic typology

103

lations between different types of distance (or similarity) and the space of variation between languages as wholes can be visualized impressively in the form of two-dimensional diagrams (cf. Fig. 1 and 2). Geographic space and its impact on the distribution of linguistics features and languages also figures prominently in Johanna Nichols’ article, but here the innovation resides in the fact that physical space is seen as three-dimensional: For the very first time, the impact of altitude on the economic and sociolinguistic situation of speakers as well as the complexity, transparency and areality of their languages is examined in great detail by a typologist. Her main point is that, in addition to the horizontal dimension, the vertical dimension is of crucial importance in shaping linguistic geography and sociolinguistics in mountain areas. The highlands are isolated and equipped with very limited economic resources. Highlanders learn lowland languages for market communication, lowlanders, by contrast, have no need to learn highland languages. This means that languages spread or even shift in the vertical direction with significant effects on the complexity and transparency of the languages concerned. Altitude thus correlates significantly with linguistic structural properties and with sociolinguistic status of speakers.

3.

Contact through migration and cultural space

Walter Bisang’s contribution also deals with language contact, though not as a result of geographic contiguity, but as a result of (enforced) speaker mobility and as a result of what he calls “mental spaces”, i.e. contact by writing. Given that the term mental spaces is used as a technical term in Cognitive Linguistics (= small conceptual arrays of referents constructed for use in thought and action, cf. Fauconnier 1994), this term is probably somewhat unfortunate for the phenomenon in question and I have therefore replaced it by “cultural spaces”. The examples are mainly drawn from Creole languages and their substrate as well as superstrate languages. The phenomena discussed are Pro-Drop (the omission of contextually given arguments), relative clause formation, (adversative) passive constructions and rules of word formation. The main focus of the article, however, is a type of complexity in which various meanings and categories are not overtly expressed but have to be inferred by the hearer. Bisang opposes this type of “hidden complexity”, found especially in languages like Mandarin Chinese, to the more widely used concepts, which are typically based on numbers of categories or rules overtly encoded. By looking at Creole languages from the point of view of McWhorter’s (2001) hypothesis that these are the least complex languages and by looking at recent innovations in Chinese and Vietnamese as a result of

104

Ekkehard König

translation and cultural contact with European languages (in the case of Chinese) and Chinese (in the case of Vietnamese), the author is able to formulate important insights concerning the genesis, transmission and loss of such “hidden complexity”. In how far do the notions of “space” used in this article differ from the ones discussed above? Creole languages are certainly the result of enforced migration of speakers and of an extreme contact situation, but the space in which the contact occurs is physical or geographic like in some of the cases discussed above. The space relevant for “contact by writing” is certainly different, but is probably better characterized as “cultural” rather than “virtual” or “mental”. Several different, though related, notions of space are used in the contributions of this section. Some are clearly physical, others are metaphorical. Another differential aspect concerns the number of dimensions. We find one-, two- and three-dimensional concepts of space. What is called “mental space” (> cultural space) by Bisang does not involve geographic contiguity and is based on cultural interaction and exchange. Overall, the uses made of the notion “space” found in these articles exhausts the spectrum of what is relevant and illuminating in the study of cross-linguistic variation. Moreover, the descriptions of the impact of space and the formal explications of this notion found in these contributions are largely original and highly illuminating.

References Fauconnier, Gilles 1994: Mental Spaces: Aspects of Meaning Construction in Natural Language. Cambridge: Cambridge University Press. McWhorter, John H. 2001: The world’s simplest grammars are creole grammars. Linguistic Typology 5: 125–166. Silverstein, Michael 1976: Hierarchy of features and ergativity. In: Robert M.W. Dixon (ed.), Grammatical Categories in Australian Languages, 112–171. Canberra: Australian Institute of Aboriginal Studies.

Section 2: Geography and variation within languages

Ideology and discourse in the enregisterment of regional variation

107

Barbara Johnstone

Ideology and discourse in the enregisterment of regional variation1

1.

Introduction

This chapter adopts a social constructivist, discourse analytic approach to the question of how geography and language variation are related. Sociolinguists have typically thought of place in physical terms, as the location of speakers or varieties in space, on the globe or on a map. In order to understand the role of place in the sociolinguistic processes I am interested in, however, we need to conceptualize place not just as a demographic fact about speakers but as an ideological construct, created in human interaction. I am interested how places, and ways of doing things associated with places, emerge in social interactions that are enabled and constrained by particular material, historical, and cultural factors. In this chapter, I explore the idea of the regional variety. Since the earliest days of dialectology, linguists have always been skeptical about the possibility of actually finding ways of speaking that have clear geographical boundaries. At the same time, however, we are also often drawn to talk as if there were such things as Southern speech or Pittsburghese. In this, we follow the lead of laypeople, who often have no trouble labeling and characterizing speech varieties in such a way as to suggest that they have sharp boundaries. If we simply rule out the lay view as uninteresting and wrong, we risk missing the ways in which laypeople’s ideas about regional variation can be consequential in the study of patterns of linguistic variation and change. 1

I am grateful to my fellow participants in the workshop on Language and Geographical Space held at the Freiburg Institute for Advanced Research (FRIAS) in November, 2009, most particularly to organizers Peter Auer, Martin Hilpert, Anja Stukenbrock, and Benedikt Szmrecsanyi, for providing the venue and audience for the first draft of this chapter and introducing me to some new ways of conceptualizing dialectology. A generous research fellowship from FRIAS from January to July, 2011, allowed me to deepen my understanding of some of these issues, and comments from Peter Auer helped greatly to sharpen the argument in the chapter. I am also grateful to my fellow workers on the Pittsburgh Speech and Society Project, which was partially funded by U.S. National Science Foundation Award No. BCS-0417684.

108

Barbara Johnstone

If we want to understand why people think that there are clearly bounded regional varieties of language, we need to understand the social and semiotic practices that give rise to this idea. One useful framework for doing this is the idea of “enregisterment” suggested by Asif Agha (2003, 2006). Enregisterment is the process by which linguistic forms become linked to sociolinguistic “registers”, or ways of speaking, acting, and being. Using Pittsburghese, the imagined dialect of the Pittsburgh, Pennsylvania area, as my example, I show how the plots of personal-experience narratives about encounters with linguistic difference can enregister linguistic forms with place and place-identity. Theoretically, then, the chapter argues for the usefulness of enregisterment as a way of understanding how language ideology arises and circulates. Methodologically, it argues for the value of discourse analysis in the study of language and space.

2.

Regional varieties are not empirical objects

The idea that languages can be divided up into varieties has always had an ambiguous status in linguistics. Dialectologists and sociolinguists have known since the earliest attempts to map variant forms that the world does not present itself to us with neat linguistic boundaries waiting to be discovered. Dialects and dialect boundaries are idealizations. W. N. Francis, a fieldworker on the Survey of English Dialects, points this out on the first page of the introductory chapter of his textbook Dialectology: The truth is that dialect boundaries are usually elusive to the point of non-existence. Very seldom does a traveler cross an imaginary line and suddenly find the people using a new and quite different dialect from that used on the other side of the line (Francis 1983: 2–3).

To support the point, Francis quotes 19th century dialectologists like Gaston Paris and Louis Gauchat. In their textbook, J.K. Chambers and Peter Trudgill similarly point out that “the labels ‘dialect’ and ‘accent’ […] are used by linguists in an essentially ad hoc manner” (Chambers & Trudgill 1998: 5), since what can actually be observed are continua of variation rather than discrete varieties. Current dialectology suggests that rather than starting with relatively small sets of pre-chosen variable features (different ways of saying the same word, different words for the same thing, different grammatical or discourse patterns for the same purpose) and looking for dialect boundaries by aggregating the dividing lines between variants of them, we need to look at how much larger numbers of variant tokens cluster together, trying to remain ag-

Ideology and discourse in the enregisterment of regional variation

109

nostic about what the clusters represent and to pay attention not just to the centers of clusters but to their peripheries (Goebl 1982; Kretzschmar 1996; Nerbonne, this volume; Szmrecsanyi 2012; Viereck 1985). Once our aim is no longer to find and describe varieties or languages, but rather to find and describe the spatial patterns that emerge in large corpora of speech, we are led to reimagine varieties and languages not as empirical objects but as ideological ones, concepts that come into being in particular historical and material contexts, via particular sets of discursive practices. To put it another way, the only thing we can actually observe are specific facts about how specific people pronounce specific sounds and which words and structures people adopt in particular situations in which they have multiple options. If we make enough such observations (and computers now allow us to assemble and analyze sets of millions of linguistic data points), we can detect patterns in the spatial distribution of particular forms. In the U.S., for example, we could find more instances of /ay/ pronounced as [a:] (as in mah for my) in the southeastern states than elsewhere. We could find other features as well whose distribution was different in the southeast than elsewhere. But the patterns are never exactly the same from one feature to another, and there are always exceptions to the rule, forms that occur where we do not expect them to. In order to claim that there is a southern “variety” or “dialect”2 of U.S. English, we must abstract away from the facts we can actually observe, generalize over them, and ignore the exceptions. We may want to do this, for one reason or another, but when we do we are not simply describing something but creating it, for some particular set of reasons. This is what it means to say that language varieties are socially constructed. Contemporary dialectologists thus stress the political, historically contingent reasons for dialectology’s mapping of boundaries (Auer 2005) and the ways in which people’s perceptions of regional linguistic boundaries and regionally variable features are ideologically shaped (Niedzielski 1999; Niedzielski & Preston 1999; Preston 1989). What regional dialectologists study are not varieties, but the distribution of regionally variable features, and what they find are not boundaries, but continua; not neat bundles of isoglosses but fuzziness. Modern variationist sociolinguistics has some of its roots in and still draws on the data of the European and North American dialect atlas projects 2

The term dialect is used differently in continental Europe than in the in the U.S. and the U.K., where dialect is often synonymous with regional variety. In this chapter, I use dialect only when the people I am quoting or paraphrasing might use it.

110

Barbara Johnstone

(particularly later ones in which social differentiation was acknowledged and described). For example, in his seminal paper “The social motivation of a sound change”, William Labov (1963: notes 11, 12) cites the Linguistic Atlas of New England. Yet sociolinguists’ objects of study are, likewise, speech communities or communities of practice, not regions; distributions and correlates of variation, not varieties. In general, the sociolinguistics of the later 20th century moved away from explicit concern with the linguistic effects of space, turning instead to the study of other “sociolinguistic patterns” (Labov 1972a), accounting for patterns of variation with reference to class, gender, and style. Early work in sociolinguistics was precisely about how and why speech patterns varied within areas like Martha’s Vineyard or the Lower East Side neighborhood of New York, because spatially defined areas are rarely linguistically homogeneous. Language differences are not automatic reflexes of speakers’ physical locations or places of origin, nor do speakers’ locations or places of origin always explain very much of the variation among them (Szmrecsanyi 2010). Social constructivist and phenomenological approaches to place from cultural/critical and humanistic geographers also suggest the need to reexamine how we have been conceptualizing explanatory variables connected with place (Johnstone 2004). Rather than asking about how speech co-varies with physical location, we are beginning to ask how meaningful places are constructed in speech and other forms of interaction, how individuals experience place, and how the use of one set of linguistic variants versus another can result from and contribute to these processes. For example, Carmen Llamas (2007) has shown that people in an English town on the border of two vernacular regions differ in how they pronounce certain words depending on whether they feel that the town belongs in one region or the other. Penelope Eckert (2004) correlates the use of urban-sounding speech variants with how teenagers in Detroit experience the city. Scott Kiesling and the rest of our Pittsburgh research team have started to explore the role of various kinds of local orientation in shaping patterns of use (Kiesling et al. 2005), and Kiesling and I have explored how people’s experiences of the city affect how they hear and evaluate local speech forms (Johnstone & Kiesling, 2008). And yet the term dialect, often used in the English-speaking world as a synonym for variety, is still very much in evidence in sociolinguistics, and it is often used to label place-linked patterns of speech. The term appears in textbooks: Ronald Wardhaugh’s An Introduction to Sociolinguistics (2006) and Janet Holmes’ Introduction to Sociolinguistics (2008) both have sections or chapters on “regional dialects” as well as “social dialects”. And it appears in research ar-

Ideology and discourse in the enregisterment of regional variation

111

ticles. A search for dialect in the journal Language in Society, founded well after the heyday of the large-scale regional dialectology projects, yields 820 hits, at least 16 of which are articles with dialect in the title. Some of these are very recent. Edinburgh University Press published a series of books about place-defined ways of speaking called “Dialects of English”; the series’ first volume appeared in 2007. The term dialect appears 18 times in the program for the 2009 New Ways of Analyzing Variation conference, the North American gathering of sociolinguists which was founded specifically to move the field beyond the kind of theoretically naïve dialectology that was seen as the old way of analyzing variation. Even though sociolinguists are in theory more interested in patterns of variation and change within communities than in differences between them, and despite our well-founded skepticism about the ontological status of language varieties and dialects (not to mention languages), our results are often used to make claims about areal varieties like the Ocracoke brogue or Southern speech. In doing this, we implicitly adopt a different view of language and of the reasons for variation, a view much more like that of laypeople.

3.

Regional varieties are social constructs

Although linguistic variation may be audible to someone listening for it, a variety is not. What linguists and laypeople alike encounter in lived experience are particular speakers, writers, or signers, saying particular things in particular ways. The variation between one speaker and another, or between an individual’s speech in one situation as opposed to another, is often unnoticeable to a particular hearer. In order to become noticeable, a particular variant must be linked with an ideological schema in terms of which it can be evaluated in contrast to another variant. Associations between particular features of pronunciation, grammar, and vocabulary, on one hand, and imagined languages, varieties, and speech communities, on the other, arise in social practices that are enabled and constrained by larger-scale political and economic conditions. As Michael Silverstein (1998: 204) puts it, “users of languages in essence construct culturally particular concepts of denotational normativity that bind subsets of them into ‘language’-bearing groups”. The links between social groups and languages, although not entirely arbitrary, are unstable, especially when social and geographic mobility create situations in which people come into contact with other ways of speaking. As a result of such contact, argues Silverstein (1998: 415), communities become increasingly aware of the peculiarities of their ways of speaking, and this in turn encourages them to think of themselves as groups. At the same time,

112

Barbara Johnstone

the concept of locality is unsteadied and the link between language and place is problematized. According to Silverstein (1998: 404), global-scale processes such as colonization, de-colonization, global economies in which communication and information are commodified, and diasporic flows of people have created a situation in which locality no longer has an essential link to language. Linguistic locality is produced, as “particular, geopolitically conceptualized, bounded swatches of the earth [are] attached to particular labels for ‘languages’ – and their bearers” (Silverstein 1998: 405). For many non-linguists, however, regional varieties seem very real. There is evidence of this in how people talk about regional variation. As Chambers and Trudgill put it, “we are used to talking of accents and dialects as if they were well-defined, separate entitles: ‘a southern accent’, ‘the Somerset dialect’” (Chambers & Trudgill 1998: 5). There is also evidence from Dennis Preston’s (1989) well-known dialect mapping experiments. Asked to draw dialect boundaries on maps and label the dialects, people have no trouble. They may not agree on where the boundaries are or on the names, but they do not question the possibility of performing the task. In my sociolinguistic interviews in the Pittsburgh area, I asked people to describe “Pittsburghese”. No one challenged the question’s presupposition that there is such an entity, and all were willing to provide examples. The answers paint a picture of a set of speech forms that is far more sharply distinct from neighboring varieties than is actually the case. Typically, Pittsburghese is thought to be a set of words, phrases, or “sayings” that are unique to Pittsburgh, different from what can be heard anywhere else, and incomprehensible to outsiders. One or two people called it an “accent”. Some labeled Pittsburghese as “slang”, “colloquialisms”, or “dialect” (in the sense of a nonstandard variety), but others called it a “language”. Whether or not it exists from a dialectologist’s point of view, Pittsburghese is central to the way Pittsburghers imagine themselves and their city. As a set of highly standardized representations of pronunciations, words, and phrases that people think are local, Pittsburghese is visible throughout the city as well as in places where ex-Pittsburghers congregate. Pittsburghese can be seen in folk dictionaries and online glossaries, on souvenir beer mugs and t-shirts, even on graffiti tags and in the names of rock bands, literary magazines, and restaurants. Pittsburghese forms can be used in casual written or spoken interaction to make playful reference to one’s Pittsburgh identity or that of one’s interlocutors. Five of the most commonly used forms, represented here as they usually are in writing but also common in oral performances of Pittsburghese, are listed in Figure 1 together with some characteristic contexts of use

Ideology and discourse in the enregisterment of regional variation

113

form

gloss

examples

yinz, yunz

‘you’, pl.

Enemies of Yinz is a rock band; a yinzer is a person with a strong local identity; Yinzplay was the name of a museum exhibit for children; a linguist might say See yinz tomorrow in a playful email to fellow linguists.

dahntahn

‘downtown’ (with monophthongized /aw/)

In a political cartoon, the destination of a city bus is Dahntahn.

Stillers, Stillerz

the Pittsburgh Steelers (Pittsburgh’s American football team, with laxed /i/ before /l/)

Decorative license plates and bumper stickers for cars say Stillers in the team’s colors; on a t-shirt, the term is defined as four time Super Bowl champs with unfinished business.

jagoff

an irritating or stupid person (Although its origin is different, jagoff sounds similar enough to jack off, ‘masturbate’, to be useable in strategically ambiguous ways.)

A t-shirt says I’m surrounded by jagoffs. Another t-shirt represents its pronunciation the way a dictionary would ( j˘ag'off ) and defines it using other colloquial terms: “1. idiot 2. jerk 3. butthead”. A chalked sign in a bar in Washington, DC that caters to Pittsburghers says Stillers ain’t no jagoffs.

n’at, n’nat

‘and that; et cetera, and so on, and stuff like that’

A magazine advertisement for Iron City beer depicts a Thanksgiving bird made from chipped ham (a Pittsburgh specialty, and also an often-represented term in Pittsburghese) with pierogies (eastern European dumplings thought of as the typical local food) for wings and beer bottles for legs. The caption is Who needs turkey n’at?

Fig. 1. Some Pittsburghese words and their uses

Almost every linguistic feature that is included in written representations and oral imitations of Pittsburghese is a feature that some Pittsburgh-area speakers actually produce in everyday, unselfconscious speech, although the representations almost always suggest that the range of the feature’s use is

114

Barbara Johnstone

narrower than it is. For example, the monophthongal pronunciation of /aw/ that is almost inevitably represented in the word dahntahn is, for people whose phonological systems dictate a monophthongal realization of /aw/, not confined to the word dahntahn. Rather, it occurs whenever this phoneme occurs in a closed syllable (Johnstone, Bhasin & Wittkofski, 2002). Pittsburgh speech and Pittsburghese are related not as an accurate and an inaccurate depiction of the same thing, but as interacting forces, one linguistic and one ideological, that shape Pittsburgh speech today and its trajectory of change. Like any other regional variety Pittsburghese is thus both not real – there is no set of linguistic forms with tightly bundled isoglosses surrounding the city – and very real – consequential for how Pittsburghers imagine themselves and the city and for how they use language. We are led, then, to ask when and why a regional variety like Pittsburghese might emerge as a useful category for making sense of the world. To answer this question, we need to think about how people experience linguistic variation and how ideas about language, space, and place shape their experience, linking linguistic forms with experienced places.

4.

Enregistering variation

Geographic mobility associated with economic change has historically resulted at the same time in the collapse of linguistic distinctions among people from different places (Milroy 2002; Trudgill 1986) and, at least in some places, in increased popular attention to regional variation (Beal 2009; Johnstone & Baumgardt 2004). This is because the social and economic conditions that cause people to speak more alike are the same as those that give rise to the activities in which dialects or regional varieties are constructed as shared representations of ways of talking linked to place and other aspects of social identity. Linguistic anthropologist Asif Agha provides a framework for understanding this apparent paradox. In a study of the history of Received Pronunciation (RP) in Britain, Agha (2003, 2006) points to some of the mechanisms involved in the identification of a set of linguistic forms as a “linguistic repertoire differentiable within a language as a socially recognized register” which has come to index “speaker status linked to a specific scheme of cultural values” (Agha 2003: 231). What became RP was once a set of phonological features the use of which was more or less automatic consequence of growing up in southeastern England, where everyone spoke that way. Accordingly, the features had no social meaning to the people who used them. Over the past three centuries, however, via a variety of prescriptivist ideas and discursive and metadiscursive ac-

Ideology and discourse in the enregisterment of regional variation

115

tivities that circulated these ideas, a set of these features were “enregistered” as an accent. That is to say that a set of pronunciation features that were once heard only in part of southern England came to be represented collectively in the public imagination as a prestige variety, RP, and the value of this variety was maintained across time and region via practices like schooling that reiterated its value and its link to social status. To reiterate, a form that is enregistered is one that is linked with a way of speaking (or “register”) associated with a personal or social identity. RP is a set of enregistered forms linked with social prestige. Regional varieties are sets of forms that are enregistered according to a different ideological schema (or set of cultural values), one which links variation in speech with place. This ideological schema comes to the fore when people encounter people who are from elsewhere and who sound different. Being from Pittsburgh and using particular speech features become linked by virtue of a variety of discursive practices: people argue about “our local dialect” on an internet discussion forum; people who have moved away share nostalgic memories of how people talk at home; someone realizes that t-shirts imprinted with local words and phrases will sell. In the remainder of this chapter, I explore the role of personal-experience narrative in the enregisterment of Pittsburghese. I show that personal narrative is one of the discursive practices through which certain speech features that can be heard locally are typified as signals of localness, and normative instructions about how to hear and use this (imagined) vernacular variety are disseminated.

5.

Narrating encounters with linguistic difference3

My colleagues and I have described elsewhere how geographic, linguistic, and historical facts came together after 1950 or so to set the stage in Pittsburgh for the enregisterment of Pittsburghese as a variety linked ideologically with place (Johnstone, Andrus & Danielson 2006). Onto this stage have emerged discursive practices and artifacts that serve to enregister Pittsburghese in the local imagination as unique and unchanging. One of these practices is the telling of stories about encounters with linguistic difference, either as an outsider coming to the city or as a Pittsburgher going elsewhere. To highlight the similarities and differences among these personal experience stories, I draw on Vladimir Propp’s (1968) structural analysis of folktales. To account for their ideological force, I turn to Labov & Waletzky’s (1997) well-known work on the structure and function of personal-experience narrative. Thus in the pro3

This section is adapted from Johnstone 2007a.

116

Barbara Johnstone

cess of illustrating how regional-variety enregisterment can work, I also point to the value of discourse analysis in the study of dialectology. In narratives by Pittsburghers about linguistic encounters elsewhere, local speech is evoked in one of two ways. In some, the Pittsburgher is told that some word or bit of grammar he or she uses is nonstandard, or at least different from what someone from elsewhere would say. The narrative in example (1) arose as Molly G., a woman in her 30s, answered an interview question, “So, have you ever heard of Pittsburghese?”.4 Example (1) Molly G. Well, I was in college- ((Laughing)) This is so embarrassing and this is going to be on the tape. ((Intake of breath)) My roommate said, “You know, that isn’t proper English”. I said, “What?” And she said, “You- you said your ‘shirt needs ironed.’” I’m like, “Well, it does”. BJ ((Laughs)) Molly G. She said, “Well, it either ‘needs ironing,’ or it ‘needs to be ironed.’” And it never occurred to me … I had never been corrected [all] the way through school BJ [Mm-hmm.] Molly G. even though we stuttered- studied grammar and everything else that … BJ Mm-hmm. Molly G. And that’s a Pittsburgh thing. ((1 second pause)) I- I think. This narrative enregisters “needs ironed” in two ways. The roommate’s reaction, as Molly represents it, enregisters the form as non-standard (not “proper English”). Molly reinforces this link between “needs ironed” and non-standardness in saying that she “had never been corrected” for using it, 4

The interviewees’ names are pseudonyms. In the transcribed extracts, I use normal orthography and punctuation as much as possible, for readability. x [ ] Single square brackets enclose simultaneous speech, which is left-aligned. x [[ ]] Double square brackets enclose phonetic transcription, when it is necessary for following the extract. x = Equals signs indicate that the second utterance follows immediately on the first. x ( ) Empty single parentheses indicate the presence of verbal material that could not be made out. x (( )) Double parentheses enclose transcriber comments about voice quality, gaze, or nonverbal sounds.

Ideology and discourse in the enregisterment of regional variation

117

framing it as incorrect. In the final line, however, Molly further enregisters the form with Pittsburgh: “that’s a Pittsburgh thing”. Kristi G. was a student in her early 20s, talking about an experience at her university in another state. Here is her answer to a question about whether anyone had ever told her she spoke like a Pittsburgher. Example (2) Kristi G. Um, the only people that, I’ve really noticed, like, get on me because of my accent is, I called, I asked for a gumband from some kid from Ohio, and he didn’t know what that was, and I was like, “A gumband, I need it for this”, and so, ’bout five minutes later hhe, like, figured out it was a rubber band, or something like that, so, they got me on that one. Here, “gumband” is enregistered with place both by virtue of the fact that Kristi cited the word in connection with a conversation about place and by virtue of how she identifies her interlocutor: “some kid from Ohio”, making place relevant by using it as a membership categorization device (Sacks 1972). In other stories, someone recognizes the Pittsburgher’s origin on the basis of his or her accent. Jen R., a woman in her 40s, and her daughter, Donna R., who was 13, co-narrate such a story. They were talking about local speech. Example (3) Jen R. Well when I’ve been in [different] states, in different cities, BJ [Mm hmm] Donna R. They’ll- they’ll say “You’re from Pittsburgh”. Jen R. Yeah, they’ll immediately [say,] Donna R. [When we were in] South Carolina, right? Jen R. “You’re from Pennsylvania=”, Donna R. “=Yeah. you’re- you’re [definite-]” Jen R. [“Are you-], are you from the Pittsburgh area?” Donna R. Yeah. ((laughing)) BJ Does that happen to you, too, or=? Donna R. =Yeah. I mean, I remember one time, we were in South Carolina visiting my, my uncle and my two cousins and my aunt. And we went to (s- some) store, and we were talking about how like the South kind of moves slow,

118

Barbara Johnstone

[You know?] Jen R. [Yeah, ] [God, it drives you crazy.] Donna R. [And then she’s like] she’s like, “You guys from Pennsylvania?” We’re like “Yeah”. And she’s like “You guys wouldn’t happen to be from Pittsburgh, right?” And we’re like, “We’re from Pittsburgh”. And she’s like, “Oh, okay. I can tell by your accent”. In this narrative, what is enregistered with Pittsburgh is not a particular word or structure but an “accent”. This creates (or, more likely in this situation, reinforces) the idea that there is a Pittsburgh variety rather than any particular idea of what it consists of. Outsiders’ stories have to do with communicative difficulties they encounter in Pittsburgh. An example comes from a radio talk show on which I was interviewed. The interviewer, Lynn Cullen, moved to Pittsburgh as an adult, to take a new job. She introduced the topic of the interview with a personal-experience narrative. Example (4) Lynn The first night I ever spent in Pittsburgh, uhm, I had come in to Cullen: look for an apartment, or a home, someplace to live because I was going to be moving here to live, and there was a horrible blizzard, that night, and I found myself snowed in at a Holiday Inn on the Parkway East, and, I just tuned on the TV, thought I might as well watch Channel 4 where I was going to be employed, and the first interview was with, a guy who owned a gas station, right off the Parkway. His name was Peewee. ((1.5 sec)) And Peewee was talkin’ to the reporter about how he was knee deep in people coming in off the Parkway and he couldn’t help ’em any more and there wasn’t any gas, and there wasn’t any help, and people were stuck, and his tow truck couldn’t this and that. ((breath intake)) ((2 sec)) I think that’s what he was talking about. I did not really understand a word the man said. And I remember sitting there and thinking, “Is he talking English?” wondering why, this wasn’t being subtitled.

Ideology and discourse in the enregisterment of regional variation

119

As with extract (3), the “I can tell by your accent” story, what Cullen does here is to link Pittsburgh with an accent (arguably even a language, given the fact that she represents herself as wondering whether the garage owner was “talking English”). Another such story has to do with an outsider’s more specific communicative difficulty having to do with a pair of words, towel and tile, that are homophonous in some Pittsburghers’ speech. I am the narrator in this example; my interlocutor, Raymond T., is a native Pittsburgher. Example (5) BJ In fact, one of the first, encounters I had when I moved to Pittsburgh was a, walking, the dog in the, in Frick Park in the morning and there was a, gentleman who also walks his dog who’s a real estate agent, from Squirrel Hill, he’s, we were talking about what we were going to do during the day, he said, that he was gonna have some workmen come in and replace the [[talz]] in his bathroom. Raymond T. [((laughing))] BJ [I thought “Why would you need workmen ((laugh voice)) to replace the towels in your bathroom, why couldn’t you do that yourself ?”] Raymond T. [((laughs))] BJ Turned out he meant the tiles, (in the bathroom). Raymond T. [((laughing))] They were heavy [[talz]]. BJ They were heavy towels, yeah, ((laughing voice)) yeah. Here, as in examples (1) and (2) above, it is a particular feature of speech that is enregistered with Pittsburgh, represented as something that happened “when I moved to Pittsburgh” and narrated in the context of a discussion of Pittsburgh speech.

6.

Plot types and ideological work

No two personal narratives of linguistic encounter are identical, and all are based in some way in personal experience, which is necessarily idiosyncratic. Thus what circulates as a model for the discursive practice of telling stories like these are not actual stories but rather plots, or semantic scaffolds on which stories can be built. To describe these plots, we need a way of abstracting away from the particular details of stories. Proppian “morphological” analysis is useful for this. Vladimir Propp’s (1968) system of functional analy-

120

Barbara Johnstone

sis was intended to aid in the classification and comparison of fairytales.5 A “function”, for Propp (1968: 21), “is understood as an act of a character, defined from the point of view of its significance for the course of the action”. Functions are repeatable from tale to tale, no matter which character fulfills them and how, and in a class of stories with the same functions, their sequence is always identical. The structure of a given fairytale type is described as a series of Roman-numbered clauses, each encapsulated in a noun such as ‘interdiction’, ‘flight’, or ‘departure’. Annotations following each clause provide descriptive detail and examples. Using Propp’s method to describe the plot types exemplified above, we might arrive at something like Figures 2 and 3.

I. THE TELLER /PROTAGONIST IS A PITTSBURGHER (initial situation) II . THE TELLER /PROTAGONIST LEAVES HOME (move) 1. The move may be that of a student going to college, someone moving for work, someone going on vacation. III . THE TELLER /PROTAGONIST ENCOUNTERS AN OUTSIDER (encounter) IV. THE TELLER /PROTAGONIST SAYS SOMETHING (utterance) 1. Typically, the teller/protagonist says very little, a word or a phrase. V. THE OUTSIDER REACTS (reaction) 1. The reaction orients to or comments on some aspect of the form of the teller/protagonist’s speech. 2. The reaction can take the form of a correction, indication of failure to understand, or recognition of the Pittsburgher’s provenance. VI . THE REACTION CAUSES THE TELLER /PROTAGONIST TO MAKE A GENERALIZATION ABOUT PITTSBURGH SPEECH (generalization) 1. Usually, this generalization is explicit and functions as the point of the narrative. Fig. 2. Encounter with linguistic difference: Type 1

5

Propp’s structuralism can be interpreted as implying that there is a universal deep structure underlying the fairy tale genre that is drawn on to generate particular tales. I do not adopt this view. I see Propp’s formulation as a useful way of making ex post facto generalizations about the structure of narrative, not as a generative mechanism.

Ideology and discourse in the enregisterment of regional variation

121

I. THE TELLER /PROTAGONIST IS NOT A PITTSBURGHER (initial situation) II . THE TELLER /PROTAGONIST COMES TO PITTSBURGH (move) 1. The move may be that of a student coming to college, someone moving for work, someone on vacation. III . THE TELLER /PROTAGONIST ENCOUNTERS A PITTSBURGHER (encounter) IV. THE PITTSBURGHER SAYS SOMETHING (utterance) 1. The Pittsburgher may say very little, a word or a phrase. V. THE TELLER /PROTAGONIST MISUNDERSTANDS OR FAILS TO UNDERSTAND (reaction) 1. The misunderstanding has to do with some aspect of the form of the Pittsburgher’s speech. VI . THE REACTION CAUSES THE TELLER /PROTAGONIST TO MAKE A GENERALIZATION ABOUT PITTSBURGH SPEECH (generalization) 1. Usually, this generalization is explicit and functions as the point of the narrative. Fig. 3. Encounter with linguistic difference: Type 2

Thinking about linguistic encounter narratives in this abstract way highlights similarities and differences among them and points to how they work ideologically to circulate claims about what counts as Pittsburghese and who speaks it, to link this variety with place, and to differentiate it sharply from other varieties. For one thing, both plot types require geographic mobility (Function II) and an encounter between a Pittsburgher and a non-Pittsburgher that is usually face-to-face. Encounters by Pittsburghers with outsiders happen outside of Pittsburgh, and encounters by outsiders with Pittsburghers happen in Pittsburgh. Thus speech is ideologically linked with place not just in the details in the stories, but on the more abstract level of their plots. The two story types are different with regard to the social identity of the encounteree (Function III). Pittsburghers’ stories (Type 1) tend to name and describe the non-Pittsburgher they encounter in ways that link linguistic variation with place, while non-Pittsburghers’ stories (Type 2) tend to name and describe the Pittsburghers they encounter in ways that link linguistic variation with class and ethnicity. In the stories in which the teller/protagonists are Pittsburghers narrating encounters elsewhere, the encounter is often with someone identified as a social peer: students’ encounters are with fellow students, for example. Social identities are suggested only indirectly if at all (“my roommate” (1), “she” (3)), with the exception of identities connected

122

Barbara Johnstone

with place: “this kid from Ohio” (2). In these stories, then, linguistic difference is correlated with place, and not with such social identities as class or ethnicity. In the Type 2 stories, by contrast, non-standard speakers are linked with class identities. Cullen’s encounter in (4) is with a gas station owner whose identity is further linked with class via his name, “Peewee”. In my encounter in (5), I identify my interlocutor as “gentleman” and as “a real estate agent”, and further by naming his upper-middle-class neighborhood, “from Squirrel Hill”. Although the class differential between outsider and Pittsburgher is maximized in Cullen’s story and minimized in mine, both stories link local speech with social class. Both plot types include a “reaction” function (Function V). In both subtypes, communicative difficulty is sometimes represented as a complete failure to communicate: In (4), Cullen’s character “did not really understand a word [Peewee] said”, despite the fact that she is able to paraphrase him at length, and wonders whether he is actually “talking English”. In (2), it took the kid from Ohio “five minutes” to figure out what the narrator was asking for. In (5), I represent my character as having failed to understand what my neighbor was talking about, wondering aloud why he would need a workman to replace his towels, rather than, as actually happened, figuring out immediately and silently that he must be talking about tiles. In stories like this, the non-Pittsburgher is represented as failing to do the kind of interpretive work that would be expected from people communicating across (fairly minor) linguistic difference. Rather than drawing on contextual factors to figure out what could be going on, the outsider runs into an interpretive wall. In (1), for example the roommate reacts to the narrator’s saying her shirt needs ironed not with the sort of second-assessment move (something like “Yes, it sure does” or “Oh, no, it doesn’t”) that would be expected as the second part of this adjacency pair but by pointing out that the phrase “isn’t proper English”. These narrative representations of communicative difficulty as communicative failure work ideologically to differentiate Pittsburgh speech far more sharply from other ways of speaking than is justified by the empirical facts. When the reaction function involves recognition of fellow Pittsburghers, as in example (3), the recognition is immediate and unambiguous “they’ll immediately say ‘You’re from Pennsylvania’;” “You’re definitely [from Pennsylvania]”. Even the wording of Donna R.’s representation of the South Carolina woman’s question gets edited in production from a hedged yes/no question to a confirmation-seeking tag question that projects a much more certain stance: “You guys wouldn’t happen to be from Pittsburgh, right?” Representing the recognition and what is recognized in this way also works

Ideology and discourse in the enregisterment of regional variation

123

to link speech with place and differentiate Pittsburgh speech from other varieties. The generalization (Function VI) that follows the reaction continues this ideological work. Before a difference between a Pittsburgher’s speech and someone else’s is noticed, it is completely unnoticeable: “It never occurred to me”, says Molly G., that needs ironed wasn’t “proper English”. When the difference is noticed, it is linked with place as “a Pittsburgh thing”.

7.

Narrative evaluation and enregisterment

Personal experience narrative is not the only discursive practice in which the enregisterment of linguistic variation takes place. Among many other such practices are more dialogic conversation (Johnstone 2007b), online talk (Johnstone & Baumgardt 2004), and the production and consumption of t-shirts (Johnstone 2009), as well as labels for ways of speaking, public sphere metadiscourses such as newspaper cartons and educational policy, and elicited judgments such as the results of opinion polls (Agha 2003). But personal narrative is perhaps a uniquely effective genre for purposes of enregisterment, because there are interactional reasons for boundaries between varieties to be drawn more sharply in this genre than in others, more so than the facts on the ground might justify. The plot types I have described are realized as conversational narratives, told in real time in face-to-face interaction. Thus, as Labov (1972b: 345–396) showed, they need to be highly evaluated. Evaluative material states or highlights the point of the story, why the audience should keep listening and allow the teller to keep talking. Evaluation may occur in clauses that comment on the story from outside: “I did not really understand a word the man said” (4); “This is so embarrassing”, or in clauses that attribute evaluative commentary to characters in the story: “I remember sitting there and thinking, ‘Is he talking English?’” (4). Alternatively or in addition, evaluation can be embedded in the narrative, in the form of such things as extra detail about characters (“some kid from Ohio” [2]), suspension of the action via paraphrase or repetition (“They’ll say, ‘You’re from Pittsburgh.’ / Yeah, they’ll immediately say, ‘You’re from Pennsylvania, are you from the Pittsburgh area?’” [3], and intensifiers (“’bout five minutes later” [2]; it never occurred to me” [1]). “I did not really understand a word the man said” makes a stronger bid for interlocutors’ continued attention than would “He was a little hard to understand”, or “He sounded a little different”, even though the latter formulations might be truer. In these stories, in other words, the interactional demand for evaluation pushes narrators to exaggerate the differences between

124

Barbara Johnstone

their speech and that of the others they encounter and the scale of the interactional difficulty to which these differences give rise. By the same token, Pittsburghers narrating encounters with fellow Pittsburghers elsewhere are interactionally constrained to exaggerate the recognizability of their accent by fellow Pittsburghers. This means that narratives like these are particularly well suited for producing and circulating ideological differentiation (Irvine 2001) among (imagined) regional varieties.

8.

Discussion

Representations and celebrations of regional linguistic variety often arise in the context of economic and social changes that lead to linguistic leveling. As Newcastle speech levels to a regional standard in the wake of outmigration (Watt 2002), people start to talk about the “Toon” (Beal 1999). As islanddwellers in the eastern U.S. encounter more and more outsiders, and their regional varieties die, they cling to one or two local forms (Schilling-Estes 1998, 2002). According globalization theorist Stuart Hall, “[t]he return to the local is often a response to globalization” (Hall 1991: 33). People symbolically readopt local practices and identities, claims Hall, when they are overwhelmed by political and economic processes larger than they can control. But economic globalization is not a new phenomenon. Globalization began in the 16th century with the European voyages of discovery, and the early 21st century is not the first time that economic change and widespread geographical mobility have led people to notice the loss of linguistic diversity. Joan Beal (2009: 139) shows how contemporary popular discourse about the loss of regional speech differences in Britain echoes the discourse of the late 18th and early 19th centuries, when “the enclosure of common land, the mechanisation of agriculture, and the Industrial Revolution … caused people to move from the countryside into rapidly-expanding industrial towns and cities”. Beal points out that many of the regional varieties of English that are now considered endangered were themselves the result of leveling processes sparked by geographic mobility. In England, a boom in dialect dictionaries in the 19th and early 20th centuries was accompanied by a surge in dialect literature and the development of regional dialect societies. In the U.S., 19th-century ‘local color’ fiction featured respelled representations of regional varieties, and actors performing stereotypical regional characters were popular on the entertainment circuit. The American Dialect Society was founded in 1889, at the height of the Gilded Age of industrialization and the accompanying immigration from Europe and geographical mobility in the U.S.

Ideology and discourse in the enregisterment of regional variation

125

Regional dialectology, along with efforts to document and/or preserve linguistic diversity, might thus be seen as attempting to push back against the homogenizing forces of industrialization and widespread geographic mobility and linguistic leveling that industrialization sparked. The dialect atlases of the 19th and 20th centuries recorded distinctive regional forms just as, or sometimes well after, the geographical and social isolation that maintained these differences was abating. Current efforts to document endangered languages emerged after many of the languages in question were already past reviving. It is tempting to adopt Hall’s account, assuming that linguistic and metalinguistic practices like these represent (belated) reactions to the economic and social processes and pressures of globalization. I argue against such accounts. At least when it comes to language, renewed attention to the local is not a nostalgic or desperate response to globalization but an inevitable concomitant of globalization. This is because changes attendant on globalization – geographic mobility, the increased heterogeneity of local demography, and economic change that forces people to re-imagine themselves – are precisely the conditions that most effectively foster language awareness. As we have seen in this chapter, narratives about encountering people from other places who speak differently serve to de-link nonstandard forms from social class and link them instead with place. Such narratives arise only when people actually encounter people from other places, and this happens only when people are geographically mobile.

References Agha, Asif 2003: The social life of a cultural value. Language and Communication 23: 231–273. Agha, Asif 2006: Language and Social Relations. New York: Cambridge University Press. Auer, Peter 2005: The construction of linguistic borders and the linguistic construction of borders. In: Markku Filppula, Juhani Kemola, Marjatta Palander & Esa Penttila (eds.), Dialects across Borders, 3–30. Amsterdam/Philadelphia: John Benjamins Publishing Company. Beal, Joan C. 1999: ‘Geordie Nation’: language and regional identity in the north-east of England. Lore and Language 17: 33–48. Beal, Joan C. 2009: Enregisterment, commodification, and historical context: ‘Geordie’ versus ‘Sheffieldish.’ American Speech 84(2): 138–156. Chambers, Jack K. & Peter Trudgill 1998: Dialectology. Cambridge: Cambridge University Press. Eckert, Penelope 2004: Variation and a sense of place. In: C. Fought (ed.), Sociolinguistic Variation: Critical Reflections, 107–118. New York, Oxford: Oxford University Press. Francis, Winthrop N. 1983: Dialectology: An Introduction. Essex, UK: Longman.

126

Barbara Johnstone

Goebl, Hans 1982: Dialektometrie. Vienna: Oesterreichische Akademie der Wissenschaften. Hall, Stuart 1991: Old and new ethnicities. In: Anthony D. King (ed.), Culture, Globalization, and the World-system, 19–39. Basingstoke, UK: MacMillan. Holmes, Janet 2008: An Introduction to Sociolinguistics. 3rd edition. Essex, UK: Pearson Education Limited. Irvine, Judith T. 2001: ‘Style’ as distinctiveness: The culture and ideology of linguistic differentiation. In: Penelope Eckert & John R. Rickford (eds.), Style and Sociolinguistic Variation, 21–43. Cambridge: Cambridge University Press. Johnstone, Barbara 2004: Place, globalization, and linguistic variation. In: Carmen Fought (ed.), Sociolinguistic Variation: Critical Reflections, 65–8. New York: Oxford University Press. Johnstone, Barbara 2007a: A new role for narrative in variationist sociolinguistics. In: Michael Bamberg (ed.), Narrative: State of the Art, 57–67. Amsterdam/Philadelphia: John Benjamins. Johnstone, Barbara 2007b: Linking identity and dialect through stancetaking. In: Robert. Englebretson (ed.), Stancetaking in Discourse: Subjectivity in Interaction, 49–68. Amsterdam/Philadelphia: John Benjamins Publishing Company. Johnstone, Barbara 2009: Pittsburghese shirts: Commodification and the enregisterment of an urban dialect. American Speech 84(2): 157–175. Johnstone, Barbara & Dan Baumgardt 2004: “Pittsburghese” online: vernacular norming in conversation. American Speech 79: 115–145. Johnstone, Barbara & Scott F. Kiesling 2008: Indexicality and experience: exploring the meanings of /aw/-monophthongization in Pittsburgh. Journal of Sociolinguistics 12(1): 5–33. Johnstone, Barbara, Jennifer Andrus & Andrew E. Danielson 2006: Mobility, indexicality, and the enregisterment of “Pittsburghese”. Journal of English Linguistics 34(2): 77–104. Johnstone, Barbara, Neeta Bhasin & Denise Wittkofski 2002: “Dahntahn Pittsburgh”: Monophthongal /aw/ and representations of localness in southwestern Pennsylvania. American Speech 77: 148–166. Kiesling, Scott F., Jennifer Andrus, Neeta Bhasin & Barbara Johnstone 2005: Local orientation and local-sounding speech in Pittsburgh: Complicating the picture (poster). Presented at New Ways of Analyzing Variation, 34, New York, NY. Kretzschmar, William A. 1996: Quantitative areal analysis of dialect features. Language Variation and Change 8: 13–39. Labov, William 1963: The social motivation of a sound change. Word 19: 237–309. Labov, William 1972a: Sociolinguistic Patterns. Philadelphia, USA: University of Pennsylvania Press. Labov, William 1972b: The transformation of experience in narrative syntax. Language in the Inner City, 354–396. Philadelphia: University of Pennsylvania Press. Labov, William & Joshua Waletzky 1997: Narrative analysis: Oral versions of personal experience. Narrative Inquiry 7: 3–38. Llamas, Carmen 2007: ‘A place between places’: Language and identity in a border town. Language in Society 36(4): 579–604. Milroy, Lesley 2002: Mobility, contact and language change – working with contemporary speech communities. Journal of Sociolinguistics 6: 3–15.

Ideology and discourse in the enregisterment of regional variation

127

Niedzielski, Nancy 1999: The effect of social information on the perception of socio-linguistic variables. Journal of Language and Social Psychology 18: 62–85. Niedzielski, Nancy & Dennis R. Preston 1999: Folk Linguistics. Berlin/New York: Mouton de Gruyter. Preston, Dennis R. 1989: Perceptual Dialectology. Dordrecht: Foris. Propp, Vladimiar 1968: Morphology of the Folktale. Austin: University of Texas Press. Sacks, Harvey 1972: On the analysability of stories by children. In: John J. Gumperz & Dell H. Hymes (eds.), Directions in Sociolinguistics: The Ethnography of Communication, 325–345. New York: Rinehart and Winston. Schilling-Estes, Natalie 1998: Investigating ‘self-conscious’ speech: The performance register in Ocracoke English. Language in Society 27(1): 53–83. Schilling-Estes, Natalie 2002: On the nature of isolated and post-isolated dialects: innovation, variation, and differentiation. Journal of Sociolinguistics 6: 64–85. Silverstein, Michael 1998: Contemporary transformation of linguistic communities. Annual Review of Anthropology 27: 410–426. Szmrecsanyi, Benedikt 2012: Geography is overrated. In: Sandra Hansen, Christian Schwarz, Philipp Stoeckle & Tobias Streck (eds.), Dialectological and Folk Dialectological Concepts of Space, 215–231. Berlin: Walter de Gruyter. Trudgill, Peter 1986: Dialects in Contact. Oxford/NY: Basil Blackwell. Viereck, Wolfgang 1985: Lingistic atlases and dialectometry: the Survey of English Dialects. In: John M. Kirk, Stewart Sanderson & J. D. A. Widdowson (eds.), Studies in Linguistic Geography, 94–112. London: Croon Helm. Wardhaugh, Ronald 2006: An Introduction to Sociolinguistics. 5th edition. Oxford, New York: Blackwell. Watt, Dominic 2002: ‘I don’t speak with a Geordie accent, I speak, like, the Northern accent’: contact-induced levelling in the Tyneside vowel system. Journal of Sociolinguistics 6: 44–63.

128

Paul Kerswill

Paul Kerswill

Identity, ethnicity and place: The construction of youth language in London

1.

Introduction: Youth language, indexicality and place

“Youth language” is labile, contingent and transient, subject to fashion, serving as a badge of nonconformity, but above all constructing for its speakers a set of complex identities for deployment as markers of different stances in relation to varying interlocutors and shifting conversational contexts. The term “youth language” is best placed in inverted commas because “languagehood” is contested by both researchers and users. A major strand of youth language research concentrates on these new urban ways of speaking as markers of identity: they are seen primarily as registers, or styles, reflecting young people’s particular communicative choices. Irvine (2001: 23) refers to linguists’ “broad conception of style” as a “social semiosis of distinctiveness”. A style, for Irvine (p. 22), becomes possible because linguistic characteristics (or ‘indexes’) of a particular social group become interpreted as carrying social meaning, and this in turn informs people’s interpretations of their own social world and their position within it. For the argument to be developed in this chapter, the implication of this interpretation of style is that it does not entail a cohesive, distinct language system – a “language” by some definitions – but rather a collection of features belonging to any linguistic component which co-occur more or less strongly. Clearly, any co-occurring features have the capacity to become stylised in this sense. At one extreme, the young person who occasionally uses the address term bruv and the pragmatic marker man with his friends is indexing a stance towards those friends at that moment, while the speaker who codeswitches between two available languages in her community may be doing so to reach a particular conversational outcome, achieving this through her knowledge of the indexicality of those languages in her community. Few linguists would want to call the behaviour of the youth in the first case the alternation of two “codes”, while in the second any criteria one might set for the alternation of “languages” are much easier to fulfil. The crunch comes when linguists’ and speakers’ views are compared: to what extent does a linguist’s ways of conceptualising language varieties match those of the community? What are the community members’ concepts, and how do they cor-

Identity, ethnicity and place: The construction of youth language in London

129

respond to analytical ‘realities’ arrived at by structural linguists or discourse analysts? How strongly is place implicated in all of these perspectives, lay and ‘expert’? This chapter deals with language and identity among young people in a large multiethnic and multilingual metropolis, from the point of view of place and distance. It does so while also dealing with cross-cutting concepts of race, ethnicity, age, culture and class: it turns out, as we shall see, that these are not simply overlaid as additional factors, but that they interact in a quasi-statistical sense in that they do not “mean” the same thing for people in different places. Naming also becomes a critical issue for this study (cf. Bijvoet and Fraurud 2011): how are the social groups which are identified by the analyst labelled by group members themselves and by other groups with whom they have dealings? How are their language varieties labelled, if at all? What folklinguistic knowledge do people demonstrate about their own and others’ varieties when the issue is raised with them in conversation? What are these language varieties like? And finally, what is the correspondence between these language varieties and folklinguistic knowledge? In our consideration of place, the question arises: what does this tell us about a large city, with its geographical and social divisions, as a “speech community”?

2.

Describing “youth language(s)”: Varieties, styles or repertoires?

As Svendsen and Røyneland (2008) and Quist (2008) point out, at least in Europe urban youth languages have initially been seen as varieties, or lects, distinguishable from other varieties or lects by applying descriptive linguistic techniques. The new way of speaking may be seen as a new dialect, as Svendsen and Røyneland (2008: 80) claim: “multiethnolectal [i.e. collective, cross-ethnic – PK] speech style contributes to a further increase of dialect diversity in Norway”, while cautioning: “It is often taken as a manifestation of lack of competence rather than as a new Norwegian dialect”. If one takes this “variety” approach to urban youth language, one often ends up resorting to tabulated lists of highly disparate features, which may be syntactic, morphological, phonological/phonetic, suprasegmental, lexical, or discourse-pragmatic. This is true of both North European and Sub-Saharan African urban youth languages (I mention these two regions if only because the vast bulk of the research has been done here). For example, we find quite detailed lists of this kind for Dutch (Nortier and Dorleijn 2008), Danish (Quist 2008), Swedish (Kotsinas 1994) and German (Wiese 2009), as well as the labelled urban youth languages Nouchi (Ivory Coast), Camfranglais (Cameroon), Sheng

130

Paul Kerswill

(Kenya), and several others (Kiessling and Mous 2004). Lists such as these are never exhaustive. Moreover, quantifying the features following a variationist methodology and rationale is highly problematic. This is because the features are often contingent on the immediate communicative act, taking on direct pragmatic and discourse functions in a way that is only exceptionally the case for variables in (more) stable vernacular varieties. Variationist quantification assumes a concept of “variety” in which variation is inherent to the system (typically, the phonology); as a consequence of this the “variants” have no direct semantic or pragmatic function. This does not stop the variants from acquiring powerful indexical meanings (Johnstone and Kiesling 2008), but the variationist approach usually assumes that the frequency per se is of interest; indeed, in recent variationist work, token frequency is taken as a component of a cognitive explanation of change (e.g. Clark and Trousdale 2009). Variationists are also interested in language internal factors, conceptualised as linguistic constraints. This approach is clearly not appropriate to the study of most manifestations of urban youth language. In fact most authors have problematised it (e.g. Jaspers 2008 from a language-ideology viewpoint) or discussed the “varieties” as practices or styles (Hurst 2009). Cheshire et al. (2011) see London’s youth language as a repertoire drawn from a “variety pool”. This does not, however, absolve us from investigating “varietyhood” from whatever point of view, since linguists, the media and users all willingly engage in a process of naming, each capturing a different aspect of linguistic or social reality. When speakers and, sometimes, the media, name a language variety, this is part of the enregisterment of that language variety: this is the association of linguistic forms with specific social characteristics and specific ideologies, such as social class or correctness (see Johnstone 2011: 34–5). Place is part of the process, in that features of a local dialect become “imbued … with meaning” as local speakers come to notice particular attitudes and reactions “in contexts such as moving into jobs where speaking ‘correctly’ was required” (Johnstone 2011: 34).

3.

The research site: London

London’s population is currently (2011) around 7.7 million (london.gov.uk). It is distributed across 33 boroughs (Figure 1). Data for this chapter is drawn from two large-scale sociolinguistic projects (see Section 4), with speakers drawn from four boroughs: Hackney, Haringey and Islington are all located in the inner city, while Havering, on the eastern outskirts, was formerly part of the county of Essex. All have populations of around 225,000. Hackney

Identity, ethnicity and place: The construction of youth language in London

131

Fig. 1. Map of London, with the boroughs of Hackney and Havering highlighted (from www.cityoflondon.gov.uk/Corporation/maps/london_map.htm). Haringey and Islington are, respectively, to the North and West of Hackney

and Havering have been highlighted on the map because the bulk of the speakers came from these two boroughs. Three sociodemographic factors are likely to impact upon language use and identity formation in the inner city of London: immigration, multilingualism and relative social deprivation. Estimates of the numbers of people born outside the UK are, for Hackney, 35–37 percent, and, for Havering, 6–7 percent (2001 Census; Spence 2008). However, a better measure of the size of the minority ethnic and non-English language heritage population may be the fact that, in Hackney, 54 % of primary school and 44 % of secondary school children have English as an additional language (Inspire 2011), meaning that English is not the main home language. In Havering, the figures are undoubtedly much lower – figures are not available, but out of a sample of 98 pre-school children 14 percent spoke English as an additional language (Havering Council 2011). The number of languages spoken by children in each borough’s schools varies much less, at around 100 for each borough (figures taken from borough websites and McPake 2006). Deprivation curtails people’s ability to participate in social and economic activities, and to have a wide base of social contacts. In terms of a range of indicators of deprivation, Hackney, Islington and Haringey are ranked 1st, 4th and 10th, respectively, out of a total of 355 boroughs in England (2001 Census). Havering is among the

132

Paul Kerswill

less deprived London boroughs with a ranking of 196. We can, therefore, expect considerable differences in language use and social attitudes between the two locations: inner and outer city. This will be a primary focus of the chapter.

4.

The studies

The Linguistic Innovators project1 (2004–2007) recorded 49 adolescents aged 16–19 in the highly multicultural borough of Hackney. We compared the English of the Hackney adolescents with that of 8 older speakers in the borough as well as with 51 adolescents and 8 older speakers in Havering. The latter borough, as we have seen, is predominantly monolingual; in fact, many of the original inhabitants of Hackney were relocated there as part of the London slum clearance that took place after World War II. We grouped our speakers by ethnicity, with a major division between Anglo (white, British origin) and non-Anglo (the remainder, mostly the London-born children or grandchildren of immigrants from developing countries). The main results are reported in Kerswill et al. (2008), Cheshire and Fox (2009) and Gabrielatos et al. (2010). The second project, Multicultural London English2 (2007–2010), focused on a wider range of age groups: 4–5, 8–9, 12–13, 16–19, about 25 and about 40. Its rationale and results are reported in Cheshire et al (2011) and Kerswill et al. (2013); because the current chapter concerns adolescent language practices, I shall be dealing only with the data for the 16–19 year olds.

1

2

Linguistic innovators: the English of adolescents in London 2004–7, funded by the Economic and Social Research Council, Principal Investigator Paul Kerswill, Co-investigator Jenny Cheshire, Research Associates Susan Fox and Eivind Torgersen (ref. RES 000–23–0680). See Kerswill, Torgersen and Fox 2008 and Cheshire and Fox 2009. Multicultural London English: the emergence, acquisition and diffusion of a new variety 2007–10, funded by the Economic and Social Research Council, Principal Investigator Paul Kerswill, Co-investigator Jenny Cheshire, Research Associates Susan Fox, Arfaan Khan and Eivind Torgersen (ref. RES-062–23–0814).

Identity, ethnicity and place: The construction of youth language in London

5.

133

Research questions: Linguistic production and social construction

This chapter concerns the relationship between actual recorded speech in two highly contrasting boroughs of the same city and the speakers’ own perceptions and constructions of speech produced there. There is, therefore, a three-way comparison: (1) speech production in the inner vs. the outer city, (2) inner-city constructions of speech varieties in both the inner and the outer city, and (3) outer-city constructions of the same. As I pointed out earlier, these are overlaid by a number of sociodemographic factors which can be shown to be variably salient for these particular speakers: race or ethnicity, social class, gender, age and place.

6.

Language production in the inner and outer city: Vowels defining a ‘variety’

The speech repertoire of inner-city London can be seen as being composed of a “variety pool” – as already mentioned (Cheshire et al. 2011). The reason for this analysis is the way in which English is being acquired and passed on: a significant proportion of children and adolescents are acquiring the language in an environment where many people are second-language speakers, and this proportion is high enough for the “standard” intergenerational transmission of the local variety to be interrupted (cf. Labov 2007); we have already seen that second-language speakers in Hackney account for around 50 % of the total. The point here is that the acquisition of the local speech repertoire takes place under similar conditions for both those who have English as their home language and those who do not, and from this observation alone we can infer that “Anglo” speakers (whose heritage is English/ British) do not straightforwardly acquire the localised, “Cockney” vernacular, even if their parents might be speakers (see Cheshire et al. 2011, Kerswill et al. 2012 fc). Immediately we can see that this calls into question the everyday notion of a “local” vernacular, perceived as such by local residents: what “local” might mean in the face of high immigration and multilingualism is a crucial question for notions of place and community. The inner-city London feature pool (cf. Siegel 1997; Mufwene 2001; Winford 2003) contains, at the very least, elements from learners’ varieties of English, Englishes from the Indian subcontinent and Africa, Caribbean creoles and Englishes along with their indigenised London versions (Sebba 1993), local London and south-eastern vernacular varieties of English, local and international youth slang, as well as more levelled and standard-like var-

134

Paul Kerswill

Fig. 2. Diphthong system of elderly male speaker from Hackney born 1918 (from Cheshire et al. 2011: 159)

ieties from various sources. Despite the heterogeneity of these sources, it is possible to make generalisations about what is shared across the young working-class speakers, and what differentiates them along the parameters of gender, ethnicity and borough. The most striking single innovation is the radical transformation of the vowel space: while the inventory remains the same, realisations are very different. Figures 2–4 illustrate the changes in the inner city. Figure 2 is a partial vowel system for an elderly white working-class Londoner, showing typical long trajectories for the face, price and goat 3 vowels. Figure 3 shows a relatively extreme version of the transformed vowel space, that of a male adolescent of Afro-Caribbean origin: here, we note the raised and peripheral onsets of face and goat, the lowered, shorter trajectory for price, and a low mouth. Figure 4 displays plots for all groups, though only the onsets of the diphthongs are shown. It is clear that while there are differences involving both gender and ethnicity, particularly involving the raised male nonAnglo face and goat, the most striking contrast by far is with the elderly speaker. Figure 5 shows the partial vowel system a young male speaker from the outer-city, relatively monolingual and monoethnic borough of Havering.

3

I follow the mnemonic system for English vowels introduced by Wells 1982.

Identity, ethnicity and place: The construction of youth language in London

135

Fig. 3. Diphthong system of young male from Hackney, Afro-Caribbean origin, born 1989 (from Cheshire et al. 2011: 160)

(a)

(b)

Fig. 4. London inner city vowels: Multicultural London English project adolescent speakers (aged 16–19). (a) Short monophthongs, (b) diphthongs plus goose and start. (For diphthongs, only onsets are shown; from Cheshire et al. 2011: 163) Key: = Anglo females (N=5) = Anglo males (N=3) = non-Anglo females (N=10) = non-Anglo males (N=8)

136

Paul Kerswill

Fig. 5. Male aged 17, Anglo, Havering (adapted from Kerswill et al. 2008: 479)

This speaker’s diphthongs are very similar to those of the elderly Londoner in Figure 2; however, the vowels of trap and strut are considerably retracted, and in this respect they match both the Anglos and the nonAnglos of Hackney. He is quite representative of young people with his background. Havering, then, is relatively conservative in its diphthongs, but follows what we have called the Southeast anti-clockwise short-vowel shift (Torgersen and Kerswill 2004). We now have a picture of the range of contemporary London working-class vowels, with big differences between age groups and between inner and outer city, and smaller differences between Anglos and non-Anglos, at least in the inner city. Do these systematic differences form part of young Londoners’ knowledge about variation in their city? Part of the social construction of language varieties is their perception by community members. “Perception” in this sense is mediated by the way in which a community attaches certain social values either to particular linguistic features, or indeed to whole language varieties. The key notions here are salience and indexicality. Kerswill and Williams (2002) argue that a feature’s becoming salient is essentially a change in its social evaluation – provided a particular level of psycholinguistically defined noticeability is reached. Social evaluation arises from the indexicality which a feature or variety has acquired (Johnstone 2010). With this in mind, a series of short speech samples was compiled from interviews in both London and Birmingham.4 4

This experiment is described more fully in Torgersen (2012).

Identity, ethnicity and place: The construction of youth language in London

137

These included young inner- and outer-city Londoners of a range of ethnicities, as well as Afro-Caribbean and Anglo Birmingham speakers. These extracts varied systematically in terms of their vowel qualities. The recordings were presented to young inner-city Londoners, who were asked to judge the voices in terms of their owners’ ethnicity and which city or region they came from. Most strikingly, the voices judged to belong to Black or Asian people (out of choices which also included White and “other”) were more likely to be located in London rather than Essex, Birmingham or Manchester (the other choices given). This was true even of the Birmingham voices. Furthermore, voices which were those of Hackney Anglos were more likely to be judged to belong to Black people if the speaker had a multiethnic personal social network. London’s inner city is, then, more strongly associated with non-Anglo speakers than with Anglos, and this is true even when a voice is heard as “Black” but contains phonetic cues which are from another city altogether. Conversely, voices heard as “White” are likely to be placed in the suburbs – in Essex. It can be argued that this outcome is simply the result of listeners’ real-world knowledge: the inner city clearly has proportionally more ethnic minority inhabitants than elsewhere. However, there are clear linguistic correlates to these associations. Listeners seem able to pick up on phonetic features, including – we must assume – particular vowel qualities, and make a positive, and often correct, identification of the speaker’s ethnicity at least when operating with the Anglo/non-Anglo dichotomy. Do listeners first identify ethnicity and then make the association with place, or is it the other way round? It seems that ethnicity comes first: this is suggested by the fact that Black Birmingham speakers are mainly correctly identified as black, but placed in London. This apparently stronger ethnic association is not borne out by the phonetic facts: there are fewer differences between Anglo and non-Anglo speakers in Hackney than there are between either of these groups and Anglo speakers in Havering. Apart from having followed the Southeastern short vowel shift, young Havering speakers’ vowel systems are very conservative in that they continue to have the broad, shifted diphthongs of elderly Londoners. In the following section, we will examine whether lexical variation also patterns along ethnic and place lines. As with the vowels, we ask whether ethnicity or place has priority in production. Finally, using discourse analytic techniques, we will see whether speakers construct their conceptions of locally-relevant language varieties in a way which mirrors the results of the listening test, with the priority apparently given to ethnicity.

138

7.

Paul Kerswill

Lexical variation: Dialect, discourse and style

Inner-city phonological innovation is matched by changes in morphosyntax and quotative expressions (Cheshire et al. 2011). Less has been said, however, about the lexis of young Londoners’ speech: are there also differences in vocabulary, including general lexis, slang and pragmatic markers? Torgersen et al. (2011) have indeed found differences in the use of pragmatic markers between inner and outer city, and between ethnicities. The differences in frequency are small for established pragmatic markers like the tag innit, but the functionally related you get me has a much higher frequency among non-Anglos in both Hackney and Havering, and an overall higher frequency in Hackney than Havering. Major differences exist, then, in pragmatic markers. We will go on to look at lexical variation in other domains: slang and less stylistically marked vocabulary choice. Before we turn to this question, I will consider what kind of information can be obtained through the study of lexical variation. Performing valid studies of lexical variation between sociolinguistically defined groups has, until recently, been considered very difficult because of the problem of representativeness. Innovative elicitation techniques are currently being adopted, for example the Sense Relation Network (Llamas 2007), which has been successfully used to discover regional patterns of lexis across the United Kingdom (BBC Voices Project). Methods where words are directly elicited can tell us about whether a particular group of speakers recognises or uses an item, but not about its actual use in context or about its overall frequency of use. Looking at lexical variation using corpora is a wellknown technique in discourse analysis and critical discourse analysis (see the examples given in Baker 2006), the advantage being that high frequencies of particular words, together with an investigation of the words they co-occur with (collocation), helps in the analysis of discourse prosody, which refers to ways in which the co-occurrence of particular words may reveal certain patterns of evaluation or, more broadly, discourses (see Baker 2006: 87). What would it mean, though, to compare lexis between two or more sociodemographically differentiated corpora of spontaneous speech? Macaulay (1991) attempts this in his study of working-class and middle-class speakers in Ayr. He does not find any difference in the type-token ratios (lexical diversity) between the classes, but does find substantial differences in syllable number (middle class speakers use more polysyllabic words) (Macaulay 1991:114; 116). This appears to correlate with the use of more abstract words by some middle-class speakers. Much more richly differentiated, however, are Macaulay’s results when he focuses on discourse styles, which he in-

Identity, ethnicity and place: The construction of youth language in London

139

vestigates by using frequency counts for particular lexical items (Macaulay 2005: 156–187). His findings for adolescent styles, this time in Glasgow, are obviously relevant to us: he finds that adolescent speakers use taboo words, particularly fuck and shite, but that these are scarcely used by middle-class adolescents. There is a good deal of teasing and verbal challenging, with girls talking about people (particularly other girls) and the working-class boys using address terms or pragmatic markers like mate and man. There are frequent references to violence, working-class boys being in the lead in this. Adult female vs. male styles are also dealt with, but more important for us are Macaulay’s findings for class. Middle-class speakers use evaluative adjectives, as well as the intensifiers quite and very and the hedge sort of, more often than working-class speakers. Middle-class use of evaluative adverbs and adjectives reflects, according to Macaulay, a habitus in which they are confident in expressing their opinions and talking about their feelings, even when these are negative (Macaulay 2005: 184; 186). We have now moved a long way from the idea of a “dialect”, defined in structural terms, or even as a social practice. However, in view of our interest in language production and its relationship with the perception and naming of social groups and language varieties, a discourse analytic approach to vocabulary use would seem to have benefits. The approach taken in this chapter is two-pronged. The first is a quantitative study of word frequency and “keyness”, using the three adolescent corpora and, for comparison, the spoken portions of the British National Corpus Sampler Corpus (Section 8). The second is a concordance analysis of some of the words which turn out to be “key” in the quantitative studies (the notion of “keyness” is explained below), in order to find points in the interviews where there is talk and evaluation of language, ethnicity, place and social groups (Section 9).

140

8.

Paul Kerswill

Lexical variation among inner and outer-London teenagers

The three corpora we will be examining are shown in Table 1, along with the BNC Sampler Corpus, which will be used as a reference.5 Name of Size corpus (words)

600,137

16–19

No. of Genre participants (excluding interviewer) 49

Havering 530,536

16–19

51

257,178

16–19

25

988,819

Teenage c. 660 to elderly

Hackney

MLE (Multicultural London English) BNC Sampler Corpus (spoken texts)

Age of participants

No. of Location Date of files (inrecordterviews ing and recordings) 49 Inner 2005–6 London (Hackney) Socio45 Outer 2005–6 linLondon guistic (Haverintering) view (in 24 Inner 2007–8 pairs) London (mainly Hackney) Range of 148 Great Late spoken Britain 1980s/ text early types 1990s

Table 1. Corpora

When examining the lexis in a corpus, it is important to determine what is distinctive about the words it contains when compared to other corpora. One way of achieving this is to compare the corpus with a reference corpus, which “acts as a good benchmark of what is “normal” in language” (Baker 2006: 43). The software I have used is WordSmith Tools 5.0, which allows the user to carry out a series of analyses, including word frequency, collocations and concordances. Another analysis type is keyness, which is explained by Baker thus: 5

I am grateful to Paul Baker for suggesting the use of this reference corpus.

Identity, ethnicity and place: The construction of youth language in London

141

Using WordSmith, it is possible to compare the frequencies in one wordlist against another in order to determine which words occur statistically more often in wordlist A when compared with wordlist B and vice versa. Then all of the words that occur more often than expected in one file when compared to another are compiled together into another list, called a keyword list. And it is this keyword list which is likely to be more useful in suggesting lexical items that could warrant further examination. A keyword list therefore gives a measure of saliency, whereas a simple word list only provides frequency. (Baker 2006: 125)

The important term here is saliency, which suggests that a word is unusually frequent rather than simply frequent. This means that a keyword may be infrequent in absolute terms, but because its relative frequency is high when compared to a reference corpus it is nonetheless, in some sense, salient. For the most part, lexical words turn up as key, rather than function words. The first analysis was to find keywords in the three London teenager corpora combined when compared to the reference corpus, the BNC. WordSmith found over 1,500 keywords, but automatically limits the list to 500. Very early in the list come a number of personal names, plus three items characteristic of informal conversation: like, yeah and just, as well as the nonstandard forms innit and ain’t. Putting aside these, we next focus on those content words which have the potential to be one or more of the following: an address term, a (component of a) pragmatic marker, slang, taboo, dialectal, an intensifier, an evaluative term, or a word used to label people, places or language varieties. Often, terms are brought up by the interviewer, so in cases where the only use of a word is metalinguistic I have not included it. The first 500 keywords include the following (listed alphabetically): aks (metathesised form of ask), arse, bare, black, blad (transcribed form of blood when used as a pragmatic marker), brother, bruv (clipped form of brother, used as an address term or pragmatic marker), chav, Cockney, cool, crack, cunt, ends (district), friend, fuck, fucked, fucking, gay, geezer, ghetto, girl, guy, hood (of a jacket), language, mad, mate, nang, olders, posh, rude, safe, shit, sister, slang, stoned, weed, white, youse (plural of you). Some words in this list are unsurprising: taboo words figure prominently in the speech of these working-class teenagers (compare Macaulay’s similar finding), as do the kinship terms brother and sister, very much expected given that these are sociolinguistic interviews. Youse probably has its origins in Irish English, is common in Northern British cities, and is now often used in London. But there are items which belong to informal language. These may be general (cool, guy, stoned, weed, bruv, olders), specific to London (geezer – a traditional Cockney term, and ends – probably a recent coinage referring to ‘area where you live’, used by inner-city young people), or else associated with a US or Caribbean-influenced youth culture (aks, ghetto, rude, safe). Two keywords

142

Paul Kerswill

are highly suggestive of subject matter: black and white, which in almost all cases refer to race. Race is a salient topic, as we shall see later. Taken as a whole, these young Londoners’ language is firmly embedded in a range of overlapping repertoires. This suggests that the speakers are using different styles, even within the same interviews. Alternatively, or additionally, it is possible that the different styles are characteristic of particular groups of speakers. To investigate this, we can do a keyword analysis of one corpus against another. There are 243 keywords in an analysis of Hackney versus Havering, and vice versa. Tables 2 and 3 split the analysis into two, showing keywords in the two boroughs separately. Words have been selected from the total of 243 keywords in the same way as in the London–BNC comparison. rank

word

frequency in Hackney

frequency in Havering

58

Hackney

656 (394)

163 (87)

60

guy

413 (248)

62 (33)

62

Bengali

157 (94)

0 (0)

68

man

1286 (772)

761 (404)

70

blad

88 (53)

0 (0)

76

guys

168 (101)

30 (16)

79

okay

100 (60)

6 (3)

80

bruv

150 (90)

24 (13)

84

brothers

416 (250)

192 (102)

86

god

481 (289)

240 (127)

87

Bangladesh

60 (36)

0 (0)

91

blood

167 (100)

43 (23)

92

olders

67 (40)

2 (1)

Table 2. Keywords in Hackney. Frequency per million words (raw frequencies are given in brackets) rank

word

frequency in Havering

frequency in Hackney

21

mates

528 (280)

190 (114)

36

mate’s

96 (51)

8 (5)

51

chav

111 (59)

23 (14)

64

accents

58 (31)

5 (3)

Identity, ethnicity and place: The construction of youth language in London rank

word

frequency in Havering

66

chavs

60 (32)

67 (4)

71

mate

492 (261

299 (179)

143

frequency in Hackney

Table 3. Keywords in Havering. Frequency per million words (raw frequencies are given in brackets)

There are striking differences between the two boroughs. Leaving aside Hackney, Bengali and Bangladesh, several words stand out. In roughly one quarter of all cases, the word man is preceded by a determiner positioned one or (to allow for an attributive adjective) two places to the left, especially the, a, that or her. Most of the remainder, then, are pragmatic markers (or sometimes address terms), as in (1) and (2): (1) (2)

Rufus: It’s cos my brain is dead innit . done too much things man huh6 (Havering) Raymond: When you took off your hat blad I saw your scalp right there man (Hackney)

The word blood and its spelling variant blad (used by the transcriber when it functions as a pragmatic marker) vary greatly in frequency between the two corpora. Blad does not occur at all in Havering, and all the uses of blood in Havering are either metalinguistic or referring to the bodily fluid. Out of the 100 occurrences of blood in Hackney, however, 42 turn out to be pragmatic markers.7 The examples below show these usages: blood/blad as pragmatic marker: (3)

Dean: Apparently my nan had taken me toilet that’s how pissed I was blood I couldn’t even find my way to the toilet (Hackney)

(4)

Chris: I gotta get something to eat in my stomach blad I got . I ain’t ate nothing all day (Hackney)

blad used metalinguistically: (5)

Int.: What about the sound of the you know . people speaking?

black and white

Grant: They talk they talk quite funny they’re like “blad oh blad come here yeah yeah” (Hackney)

6 7

Transciption conventions are given at the end of the paper. This reflects an inconsistency in the transcription, though it does not affect the result or the argument.

144

Paul Kerswill

blood meaning ‘bodily fluid’: (6)

Leon: When he put on his shirt you could see the blood coming out and everything (Hackney)

Bruv is likewise almost exclusive to Hackney, and has a similar function to blad and man. Given that most users are Afro-Caribbean, the form of this word is unexpected: it seems that the Caribbean cognate brada has been replaced by an existing Cockney form: (7) (8)

Chris: This is what she wants us to sit and do bruv . sit and chat (Hackney) Alex: I can’t believe yeah my auntie’s boyfriend opened a window bruv I thought I was gonna get sucked out it was like a hoover bruv (Hackney)

Finally for Hackney, we turn to olders, which is used by these speakers to mean ‘senior member of gang or hip-hop crew’ (though these are not exclusive meanings in youth language more generally): (9)

Alex: then he come down to our boys and joined up with our boys and now I told my olders “I don’t like this boy”. none of my youngers liked him or nothing it was just the olders said “yeah we wanna recruit him into the crew”

Keywords in Havering are far fewer in number: mate and chav are the only two referring to people, as address term or pragmatic marker (mate) or as a designation for a white working-class person with a stereotyped lifestyle and way of dressing (chav). (10) and (11) show the functions of mate, meaning ‘friend’ and as a pragmatic marker, respectively: (10) Dale: then I said I was gonna meet my mate at seven (Havering) (11) Derek: yeah no it’s fucked mate it’s totally fucked cos it’s had two front end smacks (Havering)

With the word chav, we move for the first time to an expression of identity which the speakers either embrace or else reject as belonging to other people. Although the interviewer occasionally brings this term up for discussion, it normally comes up in the course of conversation. A dictionary definition of the word is: British , informal derogatory a young lower-class person typified by brash and loutish behaviour and the wearing of (real or imitation) designer clothes. Origin: 1990s: probably from Romany chavo ‘boy, youth’ or chavvy ‘baby, child’: sometimes said to have originated in Chatham, Kent, and to be a shortening of that name. (Oxford Online Dictionary)

Identity, ethnicity and place: The construction of youth language in London

145

Examples from Havering include the following: (12) Int.: mm is that what everyone’s wearing? Burberry? Mandy: That’s a chav’s clothing Lewis: mm . I have got Burberry but Mandy: I’ll admit . our lot are just chavs ain’t we? Cos we all wear the track … Int.: What do you mean by chavs? Mandy: A chav is like . wears tracksuit bottoms always got the Burberry on the jewellery on (13) Mandy: they get it off Dagenham market . that’s a proper chav place is Dagenham. (14) Martin: I mean you do get some come in the pub that are alright they get looks cos of the way they dress [Int.: mm] cos people think oh it’s a chav but yeah they’re alright some of them

Clearly Mandy and Lewis self-identify as chavs. Martin does not, while admitting that he finds some of them “all right” (see also Jones 2008). So far, we have found large differences in vocabulary use in the inner and outer city boroughs. Although I have not investigated it systematically, most use of the keywords is by male speakers. A second important social variable is ethnicity. As a first approach to this, we do a keyword analysis of the two inner-city corpora, MLE and Hackney. Keywords for MLE, selected in the same way as before, include: bare, bredren, bruv, fuck, sex, skeen (‘OK, understood’), yardie and youths. Taking one word as an example, bredren, we find it used in the sense ‘close friend of either sex’, with a singular meaning (15): (15) Angela: They will kidnap you proper keep you for ages my my bredren got kidnapped Int.: What happened to her? (MLE)

It turns out that all 23 tokens of bredren in MLE are produced by six non-Anglos. These are of a wide range of origins, not just limited to Afro-Caribbeans: in addition to this ethnicity, they are of Portuguese, West African, Mauritian/ Jamaican, Turkish-Cypriot and Anglo/Afro-Caribbean descent. This leads to the question of whether the list of keywords in MLE in relation to Hackney is related to differences in the distribution of Anglo and non-Anglo ethnicities between the two corpora. Of the 25 speakers in MLE, 20 are non-Anglo (80 %), while in Hackney the figure is 28 of 49 (57 %). Moving on to Havering, we find that the number of non-Anglos drops to 14 out of 51 speakers (27 %). However, a closer inspection of the usage of individual words (not all key) by people of different ethnicities reveals a sharp ethnic divide. Significantly, this divide does not appear to be related to the proportions of Anglos vs. nonAnglos in the three corpora, and by extension in the boroughs:

146

Paul Kerswill

Ends in Havering, in the meaning ‘district’: This word occurs 55 times, 50 being produced by 6 of the non-Anglos. Mandem, boysdem, girldem (‘men/boys’, ‘boys’, ‘girls’)8: Of 15 tokens in Hackney and MLE, 14 are uttered by three different non-Anglos – two Afro-Caribbean, the third Anglo/Afro-Caribbean. The one Anglo speaker, Zack, who produced a token, was in conversation with Alex, who contributed 11 tokens. Zack is otherwise a high user of Multicultural London English features, and has a mainly non-Anglo friendship network. There are no examples in Havering. Blad/blud/blood as pragmatic marker: of 58 tokens in Hackney, 55 are produced by non-Anglos. Of these, 26 are produced by two Afro-Caribbeans. Mate: this word occurs 592 times in Havering, and 312 times in Hackney. In MLE, it occurs 119 times. In MLE, almost all tokens are uttered by five of the small number of Anglos.

x

x

x

x

Clearly, there are at least four almost ethnicity-exclusive words in our dataset, three non-Anglo (mainly Afro-Caribbean) and one Anglo. This sharp divide is not reflected in either the phonetic data (vowels) or the morphosyntactic features and quotatives (see Cheshire et al. 2011). Where does this result leave the notion of a “multiethnolect” in London, relatively non-ethnic in nature but with stronger place-based and perhaps class associations instead? This ethnicity-exclusive use of slang does not reflect the patterning of vowel qualities: in Hackney, there was significant, but slight variation in the vowels between the two major ethnic categories, with Anglos aligning themselves more with local non-Anglos than with the outer-city Anglos. We are now in a position to examine what the young speakers actually say about their own linguistic and social identities.

9.

The discursive construction of social and linguistic identities among London working-class young people

9.1

Minority youth speech “reallocated” to a local dialect – but from whose perspective?

A recurring theme in recent studies of multiethnic youth language in north European cities is that there is a shift from its perception as indexing “immigrant” or “minority” status to something more local, associated with a city, 8

-dem is a Creole plural morpheme, used productively in London with the items man, boy and girl. Double plural marking is often found, giving boysdem (as here).

Identity, ethnicity and place: The construction of youth language in London

147

or one or more districts in a city. We are beginning to discern how this “reallocation” applies in London. Before we consider this, we will look at three other European cities. The process of reallocation does not proceed in identical ways. In Oslo, there seems to be a relatively high awareness of a distinctive youth language; linguists describe it as containing vocabulary from immigrant languages, a distinctive rhythm and certain grammatical changes (Svendsen and Røyneland 2008). Aarsæther (2010: 118–9) shows how speakers from the multiethnic “East End” of Oslo associate their language with the place, while youngsters from the affluent West End perceive divisions in terms of ethnicity, not place. Research in Stockholm presents a somewhat different picture (Bijvoet and Fraurud 2010; 2011). In listening tests, young Stockholmers associated multiethnolectal traits with Rinkeby, a multiethnic working-class district whose name has become part of public discourse in the compound rinkebysvenska, or “Rinkeby Swedish”, referring to the multiethnolect. Mirroring Aarsæther’s study, listeners from multiethnic areas did not make an association between language and ethnicity, while speakers of mainstream Swedish did so: instead the multiethnolectal speakers associated the variety with place, while for all listeners the multiethnolect evokes a tough, working-class identity Finally, we consider a case of an originally “ethnic” way of speaking being transformed, over three or more generations, into a multiethnic youth language apparently strongly associated with a particular city. This is the variety of Dutch which is labelled Citétaal as spoken in the city of Genk in the Belgian province of Limburg. Marzo and Ceuleers (2011) argue that a specific variety of Dutch, originally spoken by immigrant Italian coalminers and their families, has now been “re-linked” to a youth identity associated with the city itself. Although the authors do not provide a linguistic analysis, this is a clear example of an ethnolect being re-indexicalised as something authentically representing a place. To sum up: indexicalities associated with ethnolectal and multiethnolectal varieties or styles can be reallocated to a city or a part of a city. Reallocation is not identical for all groups: those who use these varieties are less likely to detect ethnicity as a significant parameter than are people who do not have an ethnic minority background. In all three cases, reallocation is towards working-class, not middle-class identities: as a result, the varieties take over the association with incorrectness and toughness which attaches to working-class varieties generally. In London, there are clearly new working-class identities emerging (Rampton 2010), but we have conflicting evidence as to whether a new, local identity is materialising there. We now approach this question through an analysis of discourses present in the interviews themselves.

148 9.2

Paul Kerswill

The construction of language and social groups in Hackney: Some discourses

Who are the Cockneys? The interviews contained a component in which language and identity were discussed. The vehicle for this discussion was, in most cases, the concept of “Cockney”, which traditionally refers to people born within a limited area in the East End of London. It also refers more generally to the working-class dialect of London, and was the object of disapproval in George Bernard Shaw’s Pygmalion. In Hackney, we find a wide range of responses. (16) shows two speakers, Ryan (Anglo) and Will (AfroCaribbean) demonstrating considerable awareness of the term: (16) Int.: what about Cockneys? do you think [Ryan: uhu] you’re Cockneys? Will: no . Ryan: I hate them . Will: in that Essex sides . or not Hackney not really Hackney in Cockney x Hackney . Int.: okay so [Will: it’s really like] you hate Cockneys what . what how would you describe a Cockney then? … Ryan: white . [Int.: yeah] . er . xxx see this is chav . that’s one word . I hate them … Int.: are there any in this area? Ryan: ah there’s loads . innit? .. loads . /xxx/ Will: there’s black Cockney people as well . Ryan: ah? . Will: there’s black Cockney people as well though .. Int.: where do the Cockney people come from? Ryan: Eastenders Will: I thought it was Essex Ryan: Bethnal Green [Will: xx oh yeah] Bow and places like that Will: yeah Ryan: come from the East End . Int.: mm . always white do you think? Ryan: mainly Will: yeah Ryan: not all the time . [Int.: mhm] depends what upbringing you have Will: it’s really like East Ham and those that area East Ham . Stratford Ryan: see East Ham is dominated by Asians

Several themes arise in this extract: (i) Do I regard myself as a Cockney? (ii) Where is Cockney spoken? (possible answers: here; more traditional Cockney districts; or Essex) (iii) Are they necessarily white? (iv) Do we like them? (v) Are they chavs? In this extract, we probably detect some mutual conver-

Identity, ethnicity and place: The construction of youth language in London

149

gence on the part of both speakers who accept that Cockneys may be either white or black. In (17), Will is this time being interviewed with his friend Raymond, who like him is of Afro-Caribbean origin. (17) Int.:

mm .mm. what about the way people erm speak in London . would you say you were Cockneys? Raymond: no Int.: no . why not? Raymond: cos we use too much slang . there’s slang for everything Int.: yeh . what would you say . erm I mean what does a Cockney mean to you? Will: I don’t even know you know Raymond: to me Cockney like is like “later” . I say “later” [Will: yeh] that would be like [Will: later ] say same . seen as slang but later I would say is Cockney cos you’re dragging the word like later

Further questions arise: (vi) If we speak “slang”, then does that mean we don’t speak Cockney? (vii) What are the phonetic characteristics of Cockney? In answer to (vii), Raymond figures that Cockney contains intervocalic glottal stops. This statement causes confusion later on, because “slang” also contains glottal stops. In fact, slang is one of the few words used to refer to inner-London multiethnic youth language; others are street and ghetto – though all the terms, especially the latter two, are rare in the corpora. In (18), the same people make specific statements about who speaks slang and who speaks Cockney: (18) Int.: Raymond: Int.: Raymond:

mm . what about the programme EastEnders it’s a bunch of bull to me [Will: mm] . it’s boring do . do the people round here speak like that do you think? some people . but that’s what I say they try and pirate it and try and be like EastEnders Int.: who talks like the people on EastEnders round here? Raymond: the people that wanna be Cockney Int.: what’s that mainly older people or younger people Raymond: more older . like in between teens and adulthood would do it . teens more than . nowadays wanna talk slang . or Standard English . more than Cockney . when you’re getting older you wanna . you look back and you think oh I don’t wanna talk slang so they try and be Cockney . that’s yeah

The interviewer here brings in the iconic TV soap EastEnders, set in East London: the boys’ opinion is that EastEnders does represent Cockney. Cockney, in turn, is spoken by older people (“between teens and adult-

150

Paul Kerswill

hood”); interestingly, Raymond is of the opinion that slang is something you put behind you as you grow older. The questions then arise: (viii) Are Cockneys older than us? (ix) Will we stop talking slang when we are adults, and will we shift to Cockney or perhaps Standard English (though it is difficult to impute a particular meaning to the latter term as it is used in the extract)? In (19), the interviewer brings up the topic of whether you can tell the ethnicity (black, white or Asian) of a speaker just by listening to them. (19) Int.:

mm . so do you think all young people are speaking the same in this area or can you tell the difference between a white and a black person [Will: you can tell] or an Asian person? Raymond: when . when . when you see a black person and they’re not talking slang you think oh they’re Cockney cos they’re talking a lot of spoken English . cos most people talk slang so . not everyone talks the same but most people talk slang more than anything else Int.: #1 mm . but would you be able to tell /without looking at a person/ # Will: #2 /a white person . yeah yeah I could/ . I don’t know about Asians though cos Asian some Asian people speak different # Raymond: I donIt know boy Int.: you’re not sure if you could tell no Raymond: I don’t know Will: yeah i can Int.: you think you could . mm . okay . alright .

This is a partial answer to question (iii): Raymond says that when a black person is not talking slang you think they’re Cockney, so it is difficult to tell. Will, however, thinks that he can tell, and thus disagrees with his friend. Particularly revealing for the positioning of Hackney teenagers’ repertoire within a general London sociolinguistic continuum is (20). Mark is mixed heritage Anglo/Afro-Caribbean, and Tina is Anglo/Indian. (20) Int.: Mark: Tina: Int.: Tina:

Mark: Tina: Mark: Tina: Mark: Tina:

not too quickly

alright then right . all of these words as naturally as you can do you know you actually sounded Cockney when you were saying the first words (name=Mark) and then you went into this deeper voice is it? yeh . alright ready? no that’s not really her normal way of speaking alright alright neither is that . you got to say it normal

Identity, ethnicity and place: The construction of youth language in London Mark: Tina:

151

there’s no point if you’re not doing it right I am doing it right (name=Mark) alright

Both Tina and Mark agree that their performance when reading the word list is not their own voice: they are speakers of neither Cockney nor Received Pronunciation (“posh” would be the usual term for the latter), though they are able to imitate these varieties in such a way that listeners can identify what they are doing. In distancing themselves from both Cockney and “posh”, they are partially answering question (x): Where do I situate myself in the repertoire of speakers in the wider community? Extract (21), taken from an interview with Alex (Anglo/Afro-Caribbean) and Zack (an Afro-Caribbean oriented Anglo), thematises race explicitly, while also bringing in both dress and linguistic differences. (21) Int.: Alex:

Zack: Alex: Int.: Alex: Zack: Alex:

Zack: Alex: Zack: Alex:

what do you mean it was racist then? no it was like. you got sweet. which is like the white boys like with collars up like . they don’t wear the clothes we wear like. we got big Airforce trainers. they got like low cut Reeboks and all them like [Zack: mm mm] but like they got Reebok all the sweet mate wearing their Hackett tops /and shit yeah . Hackett tops and all that why do you call them sweet? cos they say sweet they say. cos they’re sweet like we’ll come up and we’ll say safe [Int.: right] cos we’re safe we come from Hackney but they’re from (name of place) [Zack: (name of place)] so they’ll go “sweet sweet bruv cool you alright” you know one of them like Cockney like we’re safe like . you get me they. yeah them Cockney guys they’re like Cockney poshy like they go to the pub on a Friday but we’re all. we’re all cool with them.

Alex begins by referring to dress styles, and then Zack refers to the other people (white adolescents) as “sweet”, referring to what he perceives as a characteristic usage. Alex contrasts this word with his equivalent, “safe”. The discourse presents “Cockneys” as being the “other”: they dress and speak in a certain way, they say “you alright”, and they go to the pub. Unlike the discussion in the previous extracts, Cockneys are clearly presented as a group you need to relate to; from the description here, they are probably in fact “chavs”. However, a little later in the interview, both boys hark back to an old Cockney East End, of the Kray twins (notorious gangsters) and words like

152

Paul Kerswill

“geezer” and “mate”. Three final questions then arise: (xi) What do Cockneys wear, and what do other groups wear? (xii) What words do Cockneys use, and what words do other people use? and (xiii) What sort of social practices do they engage in, and what about other people? In this analysis of the Hackney interviews, a number of discourses, or themes, emerge (here expressed from the point of view of the young Hackney speakers): Othering: Cockneys speak differently from us, and so do posh people Conflict: there are other groups with whom we have to deal, whose behaviour potentially threatens us We are concerned to mark differences between ourselves and Cockneys by alluding to dress, language (words and pronunciations) and social practices Cockneys are older than us x x

x

x

Uncertainty and ambiguity: Group boundaries: in which respects might we be considered Cockneys, or chavs (or members of some other social group)? Cockneys (possibly) live somewhere else, but we don’t all agree

x

x

In terms of language and its associations, the primary discourse is a relatively inclusive one in that race and ethnicity are rarely mentioned. Elsewhere in the interviews, it becomes clear that local area (“ends”), sometimes defined by postcode, is significant in terms of young people’s associations with place, particularly when discussing territorial struggles between youth gangs: they do not usually identify with the whole district, such as “East End” or “Hackney”. In the next section, we examine the Havering interviews, seeking out similarities and differences in the discourses in relation to those found in Hackney. 9.3

The construction of language and social groups in the outer city: Havering

In Hackney, we saw that place was of some relevance to the speakers, in that a number of them believed Cockneys live elsewhere – parts of Hackney are traditionally thought of as lying within the Cockney heartland. In general, however, Hackney speakers did not refer to other parts of London or its suburbs. It will be of interest, therefore, to find out how outer-city speakers

Identity, ethnicity and place: The construction of youth language in London

153

position themselves spatially and linguistically in relation to the inner city which, at least as far as youth culture is concerned, is culturally dominant. Extracts (22) and (23) are from an interview with two boys living in Havering, one Indian, the other Zimbabwean, who are among the very few non-white students in their college. Originally they come from inner London, and the difference between their speech and that of the local students is very salient to them. (22) Int.: Rufus: Talal: Rufus: Int.: Rufus:

Int.: Rufus: Talal: Int.: Talal: Rufus: Talal: Rufus:

Talal:

mm do you find that you use different words to some of the students around here? . yeah yeah you have to in a way mm . cos for someone who is from down these ends I have to like . try and make my English a bit . straight innit . so you /understand/ why you think they wouldn’t understand what you were saying? they don’t understand most of the things yeah right what sort [Rufus: xx] of things? can you think of anything? I dunno really . it depends innit? just comes out innit like xx like if I say “safe” to someone . he thinks that “safe” is like remember that I was telling you about? “safe” oh yeah I I xx . “safe” [Talal: like you s . say if you say “safe” “safe” is] what “safe” basically means like . “see you” like “see you later” yeah “safe” … or /“safe” as in “thanks”/ but you know the I said to one boy yeah? he said that . he thought “safe” is a safe like you put in a house innit . and he said that “I ain’t got a safe in my house”

(23) Talal: yeah … “innit” we all say “innit” as well /innit innit yeah innit/ Rufus: “innit” yeah oh we kind of put like in innit and after everything in Talal: and the they x are “really” they they heh they use the words “really” a lot “really” . like we say “innit innit” like kind of thing … Int.: so they would use “really” er where in places where you would say “innit”? Rufus: no .. I never say “really” that is gay Talal: yeah . like ehm Int.: no we’re talking about the local students maybe would say “really”

154

Paul Kerswill

The discourse here is clearly “London” versus “Essex” (most of the Havering participants consider themselves to be from this county). However, their belief that the invariant tag innit is specific to the inner city is a perception on their part: it is widespread in the south of England. That said, Pichler and Torgersen (2009) have noted new functions for this tag in London, which have not yet diffused to peripheral areas of Great Britain – so the stereotype has a certain observational validity. We turn now to another young person who has recently moved in from London. He is Sean (Anglo/Afro-Caribbean descent), who is with an Anglo friend, Freddy, also from London (Freddy doesn’t speak in this extract). In extract (24), Sean talks about Cockneys and their language. (24) Int.: Sean: Int.: Sean: Int.: Sean:

Int.: Sean: Int.: Sean: Int.: Sean:

mm . so you think the way people talk is different round here then? yeah how do you think people round here talk? . they try to be like our mums and dads . try to speak all Cockney and that . but they can’t . what do you mean they can’t? . like they say they’re Cockney . and {unclear} Cockneys . are brought up like in London . and then they sometimes they take the mickey out of us for being London boys . then they started now they wanna speak like us and that . and when they’re taking . telling us like . taking the mickey out of where we’re from and that . so do you think the people round here talk like your mum and dad? is that what you just said? . like my dad not my mum . like “yeah mate” . like all like that . and that’s how people round here talk? yeah . so are you Cockney then? . nah . I can be if you want me to

Sean explains that Havering young people want to sound like people like his dad, who are from London and are Cockneys. He cites “yeah mate” as a typical Cockney term, used round here. But he himself is a Londoner, though he doesn’t speak like the Cockneys. In Havering, he gets teased for speaking differently, but also admired for it. We infer that “slang” is what he speaks, since he states this later. “Cockneys”, then, typically belong to the adolescents’ parents’ generation, and they originally came from London. In London, the Cockney dialect has, according to the young speakers, been replaced by “slang”. According to Sean, young Londoners’ speech has a tangible influence on Havering (25):

Identity, ethnicity and place: The construction of youth language in London (25) Sean:

Int.:

155

mm . like the slang as well . round our area . you might say something . and then they’ll say “why are you talking .. like that for?” and then three months later they’re talking like it . round here . mm .

He gives the specific example of phat (‘cool’, ‘tempting’), which he claims to have introduced to some Havering friends. Although this is unlikely to be true, since by the time of the recording the word had been in use for a long time, the ideology here is clear: the inner city is innovative, fashionable and cool, while the outer city is behind the times. What of the “native” Havering/Essex young people? Michelle and Rebecca, both Anglo, have been asked whether people “out here” talk differently from people in London. Michelle asks Rebecca if she considers that people in London still speak Cockney (26): (26) Michelle: Rebecca: Michelle: Rebecca:

Michelle: Rebecca: Michelle: Rebecca: Michelle:

they do still talk like in Cockney like? . some of them yeah they say “watch and and chain”9 and . not really that many [Michelle: “apple and pears”10 and] cos there’s not like really many English people up there in that market .. so what part of London is it? just Holloway . Holloway yeah . it’s not that many there ain’t {unclear} a lot of white people down London no more is it

Ethnicity (“English”, “white people”), is now foregrounded, as are highly traditional elements of Cockney, too (“watch and chain” meaning ‘brain’, “apples and pears” meaning ‘stairs’ – both examples of rhyming slang). The girls jointly set up a strong contrast between the old and the new. London is changing, because “there ain’t a lot of white people down London no more”. Later, Michelle talks explicitly about language and ethnicity in the inner city (27): (27) Michelle:

9 10

for the . but . up London you’ve heard them and like they got a complete different language ain’t they to down here like we do we do drop our ts and our hs and that but they

Cockney rhyming slang for ‘brain’. ‘Stairs’.

156

Paul Kerswill use completely different words so (name=Sandra) uses some words sometimes and I’m like . “what?” . I don’t understand what it means or it takes me a minute and I think . “what does that sentence mean?” .. cos I’ve cos [Int.: she mixes more with London people?] yeah yeah yeah . and er a lot of black people go out with black people have got their own like little languages that they talk .. so instead of saying boyfriends {xxx} they say my hubbie or my man or something like that .

She perceives London speech as being a “completely different language”, while “we drop our ts and our hs and that”. The observation that inner city people drop their h’s much less than those in the outer city is entirely correct, as our quantitative studies show (Cheshire et al. 2008). Initially she does not discuss ethnicity in the context of the inner city, but then at the end of the extract goes on to make the point that some words are particularly used by black people. The next extract (28), however, redresses the balance. Here, one speaker argues that inner-city speech is shared by all ethnicities. Amber and Stephanie are both Anglo. (28) Amber:

so other thing us acting black because . it’s not . a black . dress code is it? ..like it’s the area they come from . black people round here wear different clothes to Tottenham boys or Hackney boys .. and my cousin . people say he speaks like a black boy but . he just speaks like a Tottenham boy . he’s not . speaking like he’s only . eight . he don’t speak that cos he wants to be black and he hears black people saying it . that’s just the way he’s talks .. Int.: is that the way everybody talks? Amber: yeah it’s the way he’s been /brought up talking with the accent/ Int.: white and black boys . mm Amber: that’s just like having a northern accent or something cos you were brought up like it Int.: mm Stephanie: in London you’re meant to have like big . cockney accent and it’s all Amber: that’s only a real cockney . Stephanie: yeah but it’s like Amber: in the west and east . Stephanie: it’s not anymore it’s .. black language like talking like them cos they’re all / . around there/ Amber: see I don’t {unclear} as black language though … so I don’t count it as that because not . just how black people talk … /cos everyone does really down there/

Identity, ethnicity and place: The construction of youth language in London

157

Amber is expressing a very strong ideology that language (and clothing style) in London is a matter of place rather than ethnicity: black people in Havering don’t dress like black people in London. People think her young cousin, who lives in London, speaks like a black boy, but it’s simply a matter of where he is being brought up. Stephanie takes a different line by saying that Londoners are supposed to speak Cockney, but they don’t any more – it’s the “black language” now. What identities do the Havering young people express? Rather little is said, and when they do express a view, they do not do so unequivocally (29): (29) Int.: Stephanie: Amber: Stephanie: Int.: Amber: Int.: Amber: Stephanie: Jennifer: Stephanie:

do you think of yourselves as Londoners or or Essex [Stephanie: Essex] people? but [Jennifer: Essex] I know we live in London like outer London yeah London but I’d rather be you would say London would you? bro . like we’re not brought up though . being born in London and . family live in London and that . mm and it’s nah I’d ra I’d . Essex . I’d [Jennifer: yeah] rather be Essex girl than London girl huh . you sure? . nah nah I just saying I I dunno

Havering presents a rather different set of discourses. Discussions of language and dress centre round the contrast between Havering (or “Essex”) and “London” (referring to the inner city, including Hackney). This opposition was largely absent from the Hackney conversations. Young Hackney speakers were at pains to position themselves within the inner city – as not Cockney and as not posh. Probably because young Hackneyites rarely need to travel to the suburbs, Havering and similar districts are not part of their image of the city. For Havering people, the inner city is a place one has experienced and has views about. The inner city has prestige – at least in the eyes of the boys who had moved out from there – and the Havering young people are clearly aware of linguistic differences. The topic of “Cockney” is not strongly picked up by the Havering speakers. Although the Havering accent has more in common with traditional Cockney than do the Hackney equivalents, the speakers do not particularly identify with Cockney. Instead, they explain that their parents might be Cockney if they had migrated out from the city, as many of them had.

158

Paul Kerswill

For Havering, race and ethnicity appear more salient than they are in Hackney. Inner-city talk, regardless of a speaker’s ethnicity, is routinely considered to “sound black”. They ascribe this to the presence of black people there. And yet, even in Havering some people focus more on place than on ethnicity – witness Amber’s statements. Black people tend to be seen by the Havering residents in an undifferentiated fashion: they did not mention the different origins black people have, nor did they mention the high proportions of Londoners from elsewhere in the world. We will return to the AfroCaribbean influence in the final section of this chapter.

10. Comparing constructions and discourses across the inner and outer city The extracts in Section 9 show the range of opinions and types of social constructions found among the young people in the two boroughs as revealed in the recordings. What they do not do is give a full picture of the extent to which these can be generalised across the whole sample. To partially achieve this, a quantitative analysis is needed. Table 4 summarises the numbers of young people who identified themselves in specific ways, as well as the frequency with which certain beliefs about language and people were expressed.

Hackney

Havering

Anglo (n=21)

NonAnglo (n=28)

Anglo (n=37)

NonAnglo (n=14)

Identifies own speech as “Cockney”

4

–

–

–

Positively identifies own speech as “not Cockney”

4

14

2

–

Cockney distinct from language of my area

4

8

–

–

Identifies as a Londoner

4

2

–

3

Identifies own speech as “Essex”

–

–

8

3

Identifies own speech as “slang”

2

5

–

1

Associates local speech with words (e.g. blad, bredren, safe). (Words are mainly Jamaican in origin.)

3

3

–

–

Identity, ethnicity and place: The construction of youth language in London Cockney associated with words (e.g. geezer, all right, mate, sweet, rhyming slang) (Words include Cockney stereotypes.)

5

4

1

–

Cockney associated with particular practices (pubs, fish and chips, tea, beer, being a chav)

2

1

1

2

Cockneys are white

1

1

–

–

Cockneys may be black

–

1

–

–

Cockneys live somewhere other than my local area

3

3

1

–

Black people talk differently from white people

–

–

5

–

Racial conflict mentioned

–

–

–

1

159

Table 4. Self-identifications and beliefs about language and groups

In terms of ascribing oneself an identity, and linking a language variety to it, we can see that non-Anglos in Hackney vociferously rejected “Cockney” as both an identity and as a label for their way of speaking. Far fewer Anglos take this line, while a small number identify as Cockney. In terms of this descriptor, there is a clear ethnic divide, and this is mirrored in the production results for slang. Yet, if the perspective is changed, there are a number of commonalities across the Hackney speakers. Several identify as Londoners. We also see that roughly equal numbers of Anglos and non-Anglos label their speech variety as “slang” or else associate local speech with Jamaicanderived slang expressions. At the same time, a number agree that their language is distinct from Cockney, while asserting that Cockneys themselves live somewhere other than the speakers’ own locality. By contrast, Havering young people do not claim a Cockney identity, even though their language is closer to traditional Cockney than is that of the Hackney residents. A number, both Anglo and non-Anglo, claim an Essex identity; those that claim to be Londoners are those who have migrated out. Only one of the Havering respondents associates their speech with “slang” – in sharp contrast to Hackney. But several Havering young people claim there is a difference between black and white speech, and one mentions racial tensions; neither of these observations are made by the Hackney subjects.

160

Paul Kerswill

11. Place, ethnicity and identity: Multiple variation patterns, multiple perspectives In Section 5, I argued for a three-way comparison: production vs. perception, inner-city views of both inner and outer cities, and outer-city views of the same. We can state some conclusions as follows. In the vowel system, the first production “gap” is between old and young. Secondarily, among the young there is a large gap between inner and outer city, while there is a much smaller difference between Anglos and non-Anglo in the inner city. In terms of perception, our method did not allow for a completely valid comparison, but on phonetic grounds (including vowels) listeners were highly attuned to ethnicity, and appeared to use this as a cue to a speaker’s status as outer or inner city. In some cases, the listeners got the ethnicity “wrong” – this was the case for Anglo speakers with multiethnic networks, who were heard as “Black”. In Hackney, quite small differences mark the ethnic distinctions, and they do so unreliably; yet listeners are able to focus on these in their judgments. In terms of lexical production, there turned out to be big differences in the use of Caribbean-origin slang. Non-Anglos, especially Afro-Caribbeans, used certain items almost exclusively, while there was a smaller number of items which were the preserve of Anglos. This divide seemed to be related to ethnicity, not place. This stands in sharp contrast to vocalic variation, where the distinction was between inner and outer city. I did not investigate the perception of lexis, but instead carried out a discourse analysis of statements about language and identity. To summarise: in Hackney, the strongest reaction is along ethnic lines: non-Anglo people emphatically reject a Cockney identity, probably because “Cockney” is a strong, stereotypable identity with white working-class origins. Yet the ethnic groups find more in common when virtually any other question is put to them. Havering young people, on the other hand, see themselves as distinct from London (they are “Essex”), and people there mention race as a differentiating factor, particularly dividing them from the inner city. In Hackney, the outer city is rarely mentioned, though some people went to shopping centres there. In Havering people seem conscious of the inner city, as a place to visit and as a place which is very different, ethnically and linguistically, from their own direct experience. To what extent do we find the type of “reallocation” described for ethnic or multiethnolectal varieties in Oslo, Stockholm and Genk? In the last 10 years, there has been a considerable media interest in language developments among working-class youth in London. Newspapers have adopted the label

Identity, ethnicity and place: The construction of youth language in London

161

Jafaican, and this has been discussed, in both a positive and a negative light, in a range of publications, including the Sun, the Daily Mail, the Guardian, the Sunday Times and the Economist (Kerswill fc.). Public lectures have also been given on the subject of Multicultural London English. On 12th August 2011, the historian David Starkey was asked, on the TV current affairs programme Newsnight, to comment on the riots and looting which had occurred in English cities that week. He stated that, linguistically and culturally, “the whites have become black”, going on to talk about a “wholly false” “Jamaican patois” which has been “intruded in England”. Predictably, he was roundly criticised for his views. His linguistic comments also drew some criticism: Jamaican patois is not as prominent in London as it was in the 1980s, and what he was almost certainly referring to was Multicultural London English, or Jafaican. This is the variety of English which is often heard as “black” by people who live outside its East End heartland, in places like Havering – as we have seen from the extracts from the interviews. Clearly, this variety is a vernacularised form of speech, and it is characteristic of parts of inner-city London as a new dialect (see also Fox 2007). However, public consciousness of it appears to be much lower than for multiethnolects in other European cities. The term “Jafaican” is beginning to come into wider use, but its frequency in the media is marginal when compared to other linguistic labels, including “Cockney” itself. Users themselves generally do not have a readily available name for it: the only relatively widely used term is “slang”, employed ambiguously to refer to the variety or to the slang items which form part of it. Recent events in London have drawn attention to its presence, in the form of television interviews with users of the variety, while David Starkey’s comments undoubtedly caused a heightened awareness of it, at least among Newsnight viewers. It is very likely that newsworthy events such as the August 2011 riots will help it along the path to enregisterment. Key to symbols used in conversational extracts . short pause … speaker tails off xxx unclear speech (x represents a syllable) /text/ speech between slashes overlaps with next transcription line

162

Paul Kerswill

References Aarsæther, Finn 2010: The use of multiethnic youth language in Oslo. In: Pia Quist & Bente Ailin Svendsen (eds.), Multilingual Urban Scandinavia: New Linguistic Practices, 111–126. Bristol: Multilingual Matters. Baker, Paul 2006: Using Corpora in Discourse Analysis. London: Continuum. BBC Voices Project: http://www.bbc.co.uk/voices. Bijvoet, Ellen & Kari Fraurud 2010: ‘Rinkeby Swedish’ in the mind of the beholder: studying listener perceptions of language variation in multilingual Stockholm. In: Pia Quist & Bente Ailin Svendsen (eds.), Multilingual Urban Scandinavia: New Linguistic Practices, 170–188. Bristol: Multilingual Matters. Bijvoet, Ellen & Kari Fraurud 2011: Language variation and varieties in contemporary multilingual Stockholm: an exploratory pilot study of young peoples’ perceptions. In: Roger Källström & Inger Lindberg (eds.), Young Urban Swedish. Variation and Change in Multilingual Settings, 1–34. Göteborg: University of Gothenburg. Cheshire, Jenny, Susan Fox, Paul Kerswill & Eivind Torgersen 2008: Linguistic innovators: the English of adolescents in London. Final report presented to the Economic and Social Research Council. Cheshire, Jenny & Susan Fox 2009: Was/were variation: A perspective from London. Language Variation and Change 21: 1–38. Cheshire, Jenny, Paul Kerswill, Susan Fox & Eivind Torgersen 2011: Contact, the feature pool and the speech community: The emergence of Multicultural London English. Journal of Sociolinguistics 15(2): 151–196. Clark, Lynn & Graeme Trousdale 2009: The role of frequency in phonological change: evidence from TH-fronting in east-central Scotland. English Language and Linguistics 13(1): 33–55. Fox, Susan 2007: The Demise of Cockneys? Language Change among Adolescents in the “Traditional” East End of London. Unpublished Ph.D. dissertation. Colchester: University of Essex. Gabrielatos, Costas, Eivind Torgersen, Sebastian Hoffmann & Susan Fox 2010: A corpus-based sociolinguistic study of indefinite article forms in London English. Journal of English Linguistics 38: 297–334. Havering Council 2011: London Borough of Havering Childcare Sufficiency Assessment 2011. http://www.havering.gov.uk/CHttpHandler.ashx?id=25804&p =0. Hurst, Ellen 2009: Tsotsitaal, global culture and local style: identity and recontextualisation in twenty-first century South African townships. Social Dynamics 35: 244–257. Inspire 2011: The Education Business Partnership for Hackney. http://www.inspireebp.org.uk/about-hackney.html. Irvine, Judith T. 2001: ‘Style’ as distinctiveness: the culture and ideology of linguistic differentiation. In: Penny Eckert & John R. Rickford (eds.), Style and Sociolinguistic Variation, 21–43. Cambridge: Cambridge University Press. Jaspers, Jürgen 2008: Problematizing ethnolects: Naming linguistic practices in an Antwerp secondary school. International Journal of Bilingualism 12: 85–103. Johnstone, Barbara 2010: Locating language in identity. In: Carmen Watt & Dominic Watt (eds.), Language and Identities, 29–36. Edinburgh: Edinburgh University Press.

Identity, ethnicity and place: The construction of youth language in London

163

Johnstone, Barbara & Scott Kiesling 2008: Indexicality and experience: Exploring the meanings of /aw/-monophthongization in Pittsburgh. Journal of Sociolinguistics 12(1): 5–33. Jones, Owen 2008: Chavs: The Demonization of the Working Class. London: Verso. Kerswill, Paul fc.: The objectification of ‘Jafaican’: the discoursal embedding of Multicultural London English in the British media. In: Jannis Androutsopoulos (ed.), The Media and Sociolinguistic Change. Berlin: de Gruyter.. Kerswill, Paul & Ann Williams 2002: ‘Salience’ as an explanatory factor in language change: evidence from dialect levelling in urban England. In: Mari C. Jones & Edith Esch (eds.), Language Change. The Interplay of Internal, External and Extra-Linguistic Factors, 81–110. Berlin/New York: Mouton de Gruyter. Kerswill, Paul, Eivind Torgersen & Susan Fox 2008: Reversing ‘drift’: Innovation and diffusion in the London diphthong system. Language Variation and Change 20: 451–491. Kerswill, Paul, Jenny Cheshire, Susan Fox & Eivind Torgersen 2013: English as a contact language: the role of children and adolescents. In: Marianne Hundt & Daniel Schreier (eds.), English as a Contact Language, 258–282. Cambridge: Cambridge University Press. Kießling, Roland & Maarten Mous 2004: Urban Youth Languages in Africa. Anthropological Linguistics 46(3): 303–341. Kotsinas, Ulla-Britt 1994: Ungdomsspråk [Youth language]. Uppsala: Hallgren & Fallgren. Labov, William 2007: Transmission and diffusion. Language 83: 344–387. Llamas, Carmen 2007: Field methods. In: Carmen Llamas, Louise Mullany & Peter Stockwell (eds.), The Routledge Companion to Sociolinguistics, 12–17. London: Routledge. Macaulay, Ronald K. S. 1991: Locating Dialect in Discourse: The Language of Honest Men and Bonnie Lasses in Ayr. Oxford: Oxford University Press. Macaulay, Ronald K. S. 2002: Extremely interesting, very interesting, or only quite interesting? Adverbs and social class. Journal of Sociolinguistics 6(3): 398–417. Marzo, Stefania & Evy Ceuleers 2011: The use of Citétaal amoung adolescents in Limburg: the role of space approprition in language variation and change. Journal of Multilingual and Multicultural Development 32; 451–464. Mufwene, Salikoko S. 2001: The Ecology of Language Evolution. Cambridge: Cambridge University Press. Quist, Pia 2008: Sociolinguistic approaches to multiethnolect: language variety and stylistic practice, International Journal of Bilingualism 12: 43–61. Pichler, Heike & Eivind Torgersen 2009: It’s (not) diffusing, innit?: The origins of innit in British English. Paper presented at NWAV 38, University of Ottawa, October 2009. Rampton, Ben 2010: Crossing into class: language, ethnicities and class sensibility in England. In: Carmen Llamas & Dominic Watt (eds.), Language and Identities, 134–143. Edinburgh: Edinburgh University Press. Sebba, Mark 1993: London Jamaican. London: Longman. Siegel, Jeff 1997: Mixing, leveling and pidgin/creole development. In: Arthur K. Spears & Donald Winford (eds.), The Structure and Status of Pidgins and Creoles, 111–149. Amsterdam: Benjamins. Spence, Lorna 2008: A Profile of Londoners by Country of Birth: Estimates from the 2006 Annual Population Survey. London: Greater London Authority.

164

Paul Kerswill

Svendsen, Bente Ailin & Unn Røyneland 2008: Multiethnolectal facts and functions in Oslo, Norway. International Journal of Bilingualism 12: 63–83. Torgersen, Eivind 2012: A perceptual study of ethnicity and geographical location in London and Birmingham. In: Sandra Hansen, Christian Schwarz, Philipp Stoeckle & Tobias Streck (eds.), Dialectological and Folk Dialectological Concepts of Space, 75–95. Berlin/Boston: de Gruyter. Torgersen, Eivind & Paul Kerswill 2004: Internal and external motivation in phonetic change: dialect levelling outcomes for an English vowel shift. Journal of Sociolinguistics 8: 23–53. Torgersen, Eivind, Costas Gabrielatos, Sebastian Hoffmann & Susan Fox 2011: A corpus-based study of pragmatic markers in London English. Corpus Linguistics and Linguistic Theory 7: 93–118. Wells, John C. 1982: Accents of English, Vols. I–III. Cambridge: Cambridge University Press. Wiese, Heike 2009: Grammatical innovation in multiethnic urban Europe: New linguistic practices among adolescents. Lingua 119: 782–806. Winford, Donald 2003: An Introduction to Contact Linguistics. Oxford: Blackwell.

Geography in morphosyntactic variation

165

Bernd Kortmann

How powerful is geography as an explanatory factor in morphosyntactic variation? Areal features in the Anglophone world1

1.

Introduction

Geography as a powerful, in fact the key explanatory factor in language-internal variation is a central premise in traditional dialectology and dialectometry. In both fields it has amply been demonstrated that geographical proximity translates into linguistic similarity and, vice versa, that linguistic dissimilarity increases with an increasing geographical distance of the relevant speech communities. This has been shown for phonological and lexical (including phraseological) variation, as well as in recent years for morphosyntactic variation, always keeping in mind that morphosyntactic variation operates on a considerably larger areal scale than lexical and, especially, phonological variation (for an overview cf. Kortmann 2010a; cf. also Szmrecsanyi 2012 for an innovative application of corpus-based dialectometric methods to morphosyntactic variation in the British Isles). In this paper the strength of the geographical signal will be explored on a global level in 48 largely non-standard varieties of English and 26 English-based Pidgins and Creoles in seven Anglophone world regions: Africa, Asia, Australia, British Isles, Caribbean, North America, and Pacific. (Only for the sake of labelling convenience, one may add to these the South Atlantic. This, however, cannot seriously be argued to constitute a linguistic area in a typo1

This paper is dedicated to the memory of Anna Siewierska. As I was writing this paper in early August 2011, the message came in letting the ALT community know of her death in a car accident. It sent me back for hours thinking of the many occasions over the last 20 years on which we had met, often together with Dik, listening to each other’s talks, but more often chatting, laughing, drinking, eating – simply enjoying each other’s company. It was always nice to know we’d meet again at a conference or a special occasion, and almost nicer even when we met without knowing beforehand and had the opportunity of spending some time together. I will miss the wonderful person and the discussions with the brilliant linguist, as so many of us will, each in their own way. Over the last years, Anna the typologist had developed an interest in variation among dialects of English. The present paper may thus have met with her interest, too.

166

Bernd Kortmann

logical or dialectological sense, given the degree of isolation of the relevant varieties – Falkland Islands English, St Helena English, Tristan da Cunha English – and the enormous distances between them.) The data set drawn upon is the largest one that has ever been available in mapping morphosyntactic variation in the Anglophone world (or beyond), namely the World Atlas of Varieties of English (for the open access electronic resource, short eWAVE, cf. Kortmann and Lunkenheimer 2011, http://ewave-atlas.org). Geography (or: areality) in this paper translates into individual (sets of) morphosyntactic features (out of a set consisting of a total of 235 features) which are pervasively observable in a particular Anglophone world region (e.g. Africa) or a part thereof (e.g. West Africa) against all others. The overall spirit of this paper is to make geography as strong as possible, i.e. to determine how much is left for geography if we take away genealogy, or variety type, i.e. essentially the socio-historical conditions in which a given variety emerged. It will be argued that geography is clearly only of secondary importance to variety type in accounting for overall morphosyntactic profiles; such typological profiles are far better explained in terms of whether we are looking at L1 varieties of English, L2 varieties, or English-based Pidgins and Creoles – regardless in which Anglophone region they are spoken. However, geography can be made considerably stronger if we focus on individual features (rather than entire typological profiles), and could be made even stronger if much more fine-grained morphosyntactic information for the individual varieties, Pidgins and Creoles was available (e.g. complementation patterns for individual verbs, usage constraints in general, pragmatic and stylistic constraints, in particular). The nature and origin of the survey (i.e. WAVE) data used here will be sketched in section 2. In section 3 it will briefly be demonstrated that geography is secondary to variety type as an explanatory factor of morphosyntactic variation when looking at entire typological profiles of varieties of English based on the complete 235 feature set in WAVE. Increasingly narrowed down data sets will be looked at in the three sections to follow. In section 4 the WAVE data set will be used to test an empirical observation variously formulated by Mesthrie over the last years (2006, 2008, 2011), namely that in terms of morphosyntax African varieties of English (key example: Black South African English) are largely of the ‘preserving’ type, whereas (South East) Asian varieties (key example: Singaporean English) are of the ‘deleting’ type. Section 5 will zoom in on one particular variety type, namely the Pidgins and Creoles in WAVE, demonstrating strategies for identifying areal patterns clearly distinguishing, for example, West African from Pacific Pidgins. In section 6 three scenarios will be outlined in which geography can

167

Geography in morphosyntactic variation

be expected to matter in morphosyntax, i.e. where we can expect area-specific patterns and geographical differences to emerge even among varieties belonging to the same variety type. Concluding remarks in light of some of the key issues addressed in this volume will be offered in section 7.

2.

The WAVE data set

The World Atlas of Varieties of English (WAVE) was designed and compiled at the Freiburg Institute for Advanced Studies (FRIAS) and the English Department of the University of Freiburg, Germany, between 2008 and 2011, partly in collaboration with the Max Planck Institute for Evolutionary Anthopology (Leipzig). It was compiled from descriptive materials, naturalistic corpus data, and native speaker knowledge by a team of 70 contributors, all leading experts in their fields. WAVE maps morphosyntactic variation in spontaneous spoken English in 48 varieties of English (10 traditional dialects (L1t), 21 high-contact mother-tongue Englishes (L1c), and 17 indigenized second-language Englishes (L2)), 7 English-based Pidgins (P) and 19 English-based Creoles (C) in seven Anglophone world regions. Table 1 provides an overview of the 74 varieties, Pidgins and Creoles sampled in WAVE and their distribution across variety types and world regions. L1 (30) low-contact L1 high-contact L1 (10) (21) British Orkney and Shet- Irish E (IrE), Welsh E (WelE), Isles (11): land E (O&SE), North of England Manx E (ManxE), Channel Islands E (North), SW of England (SW), SE (ChlsE) of England (SE), East Anglia (EA), Scottish E (ScE) Africa Liberian Settler E (16): (LibSE), White South African E (WhSAfE), White Zimbabwean E (WhZimE)

L2 (18)

P (7) & C (19)

[Maltese E (MltE) British Creole (BrC)

Ghanaian E (GhE), Nigerian E (NigE), Cameroon E (CamE), Kenyan E (KenE), Tanzanian E (TznE), Ugandan E (UgE), Black South African E (BlSAfE), Indian South African E (InSAfE)

Ghanaian Pidgin (GhP), Nigerian Pidgin (NigP), Cameroon Pidgin (CamP), Krio, Vernacular Liberian E (VLibE)

168 America (10):

Bernd Kortmann Newfoundland E (NfldE), Appalachian E (AppE), Ozark E (OzE), Southeast American Enclave dialects (SEAmE)

Colloquial Ameri- Chicano E (ChcE) Gullah can E (CollAmE), Urban African American Vernacular E (UAAVE), Rural African American Vernacular E (RAAVE), Earlier African American Vernacular E (EAAVE)

Caribbean (13):

Bahamian E (BahE)

Jamaican E (JamE) Jamaican C (JamC), Bahamian C (BahC), Barbadian C (Bajan) Belizean C (BelC), Trinidadian C (TrinC), Eastern Maroon C (EMarC), Sranan, Saramaccan (Saram), Guyanese C (GuyC), San Andrés C (SanAC), Vincentian C (VinC)

South and Southeast Asia (7):

Colloquial Singapore E (CollSgE)

Indian E (IndE), Butler E (ButlE) Pakistan E (PakE), Sri Lanka E (SLkE), Hong Kong E (HKE), Malaysian E (MalE)

Australia (5):

Aboriginal E (AbE), Australian E (AusE), Australian Vernacular E (AusVE)

Pacific (8):

New Zealand E (NZE)

Torres Strait C (TorSC), Roper River C (RRC [Kriol])

Colloquial Fiji E (CollFijiE), Acrolectal Fiji E (FijiE)

Hawaiian C (HawC), Bilama (Bisl), Norfolk Island/Pitcairn E (Norf ’k), Palmerston E (PalmE), Tok Pisin (TP)

Geography in morphosyntactic variation South Atlantic (3)

169

St. Helena E (StHE), Tristan da Cunha E (TdCE), Falkland Islands E (FlkE)

Table 1. L1 and L2 varieties, Pidgins and Creoles represented in WAVE

The experts were asked to classify ‘their’ variety as belonging to one of five broad classes of variety types: low-contact L1 dialects, i.e. traditional, regional non-standard mother-tongue varieties like East Anglian English in England or Appalachian English in North America; high-contact L1 varieties, including transplanted L1 Englishes and colonial standards (e.g. Bahamian English, New Zealand English), language shift varieties (e.g. Irish English), and spontaneous spoken standard varieties (e.g. colloquial American English); second language varieties, i.e. indigenized non-native varieties of English that have a certain degree of prestige and normative status in their political communities, like Pakistani English, Jamaican English, Hong Kong English, and Kenyan English, but also non-native varieties that compete with local L1 varieties for prestige and normative status, e.g. Chicano English and Black South African English. The Pidgins in WAVE can mostly be considered expanded Pidgins (e.g. Tok Pisin, Nigerian Pidgin, Ghanaian Pidgin), i.e. in contrast to prototypical pidgins they are less restricted in terms of domains of use, and many of them are spoken as native or primary languages by a considerable proportion of their speakers. This is true, by definition, for all of the Creoles in WAVE, such as Krio, Jamaican Creole, Belizean Creole, Sranan, or Torres Strait Creole. WAVE provides information on a total of 235 features from a dozen domains of grammar (see Table 2) (see p. 170).

170

Bernd Kortmann

Grammatical domain

Sum features in group

% of total features

Pronouns Noun Phrase Tense and aspect Modal verbs Verb morphology Negation Agreement Relativization

47 40 33 7 26 16 15 15

20,0 % 17,0 % 14,0 % 3,0 % 11,0 % 6,8 % 6,4 % 6,4 %

Complementation

11

4,7 %

Adverbial Subordination Adverbs and Prepositions Discourse organization and word order

5 7

2,1 % 3,0 %

13

5,5 %

Table 2. Domains of grammar covered in WAVE

The information in the WAVE database consists of judgements by top experts on the frequency with which each of the 235 features can be encountered in the relevant variety, Pidgin or Creole, complemented by relevant examples. The following ratings were used: A B C D X

feature is pervasive or obligatory feature is neither pervasive nor extremely rare feature exists, but is extremely rare attested absence of feature feature is not applicable (given the structural make-up of the variety/ P/C) ? no information on feature is available WAVE comes in two formats: as an open access interactive database (eWAVE; Kortmann and Lunkenheimer 2011, http://ewave-atlas.org) and as a print volume (Kortmann & Lunkenheimer 2012).2 For details on the informants, individual features, varieties and variety types, the reader is referred to eWAVE. The print volume (pWAVE) offers accounts of individual 2

An updated version of the open-access online tool – eWAVE 20 – will be accessible from autoumn 2013 anwards.

Geography in morphosyntactic variation

171

of the 74 data sets as well as large-scale comparisons and synopses across the individual variety types and Anglophone world regions.

3.

Morphosyntactic profiles in light of Anglophone world regions

In a series of papers based on WAVE and its 2004 forerunner (76 morphosyntactic features in 46 varieties of English and English-based Pidgins and Creoles), Kortmann and Szmrecsanyi have shown that variety type is a far more powerful predictor of the overall morphosyntactic profile than geography. In other words: data sets from the same variety types will account for much more of the observable variation than world region. The same results can also be gained from the much larger 74 varieties set of WAVE, and can easily be read off the two representations of the same basic phenogram in Figures 1 and 2, which represents the structural (dis)similarity of the 74 varieties, Pidgins and Creoles in WAVE based on the entire set of 235 features.3 Figure 1 shows that varieties belonging to the same variety type cluster very nicely together. Going anti-clockwise, just consider the coherent cluster of L1 varieties in the “North by NW” part of the phenogram (from Channel Island English to White Zimbabwean English), with the majority of the traditional, low-contact L1 varieties in the NNW branches (from East Anglian English to Ozarks English), followed by the core L2 branch in the “Western” part (from Pakistan English to Malaysian English) and, most impressively, by the huge coherent cluster of Pidgins (especially in the SW-SSW branches: from Tok Pisin to Cameroon Pidgin) and Creoles (covering the entire Southern and Eastern part of the phenogram: from the Australian Creoles (Torres Strait and River Roper Creole) and especially the Eastern Maroon Creoles to Vernacular Liberian English. Even in the North-NNE part of the phenogram, we find two largely coherent clusters of varieties, namely another set of high-contact L1 varieties of English (from Liberian Settler English to Early African American Vernacular English) and three L2 varieties of English (Indian and Black South African English, Indian English). Even though this is not the place to discuss this phenogram in all its details, three points are worth noting: (a) Phenograms are entirely computer-generated, based in 3

All phenograms in this paper have been produced by Christoph Wolk. The bioinformatic principles and the software underlying these phenograms are explained in Wolk 2009 and Kortmann & Wolk 2013. The latter publication also offers new phenograms, based on a more recent version of eWAVE than was available at the time of writing the present paper.

172

Bernd Kortmann

Fig. 1. WAVE phenogram based on all 235 features: clustering according to variety type

this case on checking the co-presence (A/B/C rating) or co-absence (D/X/? rating) of features across every single of the 235 features and all 74 data sets in WAVE; all the more impressive is the nice clustering of the neighbouring branches in this unrooted tree. (b) The phenogram gives us an absolutely neutral picture of the structural similarity or dissimilarity of the 74 varieties, Pidgins and Creoles among each other; so “outliers” in terms of variety type mean no more than that the overall structural, i.e. the morphosyntactic, make-up of a given variety betrays its socio-history and the socio-cultural conditions in which it is used. More exactly, even a variety classified as a highcontact L1 or an L2 variety may turn out to exhibit the structural profile of a Creole (as in the cases of Aboriginal English and Colloquial Fiji English respectively). (c) The very status of “outlier” may be questioned for varieties where, for example, for a considerable part of the relevant speech community one could also argue for a different variety type classification. The perhaps most prominent case in point is Indian English in the NNE L2 branch, which is clearly a shift variety with a considerable (and growing) absolute number of mother-tongue speakers, thus deserving an L1c classification.

Geography in morphosyntactic variation

173

Fig. 2. WAVE phenogram based on all 235 features: Clustering according to world region

The picture is far more complex and fragmented if one classifies the L1 and L2 varieties, Pidgins and Creoles in the same phenogram according to the Anglophone world region where they are spoken, as done in Figure 2. Even though for some world regions (or parts thereof) we may find miniclusters (e.g. in the Western part of the phenogram there are mini-clusters for South Africa, East Africa, West Africa, and Southeast Asia, and in the Southern part mini-clusters for the Pacific/Australia, West Africa again, and the Caribbean), there are many more outliers than in Figure 1. Besides almost all of these mini-clusters are to be observed within a given variety type, not across variety types. Take, for example, the majority of the West African

174

Bernd Kortmann

data sets, which are split among a “Western” L2 branch (Ghanaian English, Nigerian English, Cameroon English) and a “Southern” Pidgin branch (Ghanaian Pidgin, Nigerian Pidgin, Cameroon Pidgin; plus Krio as a Creole). Only in the case of Liberia do we have two varieties belonging to different variety types which cluster together (Liberian Settler English as a high-contact L1 variety, and Vernacular Liberian English as a Creole). The structural similarities between these two varieties are far greater than between, for example, Nigerian English and Nigerian Pidgin. The basic point following from a comparison of Figures 1 and 2 simply is that, at least when judging on the basis of entire morphosyntactic profiles, geography is of no more than secondary importance compared with variety type. In the next section we shall use the WAVE data set to test a recent and, to my knowledge, the most pointed hypothesis concerning a link between general traits in the morphosyntactic nature of varieties of English and the Anglophone world region where they are spoken.

4.

South East Asia vs. Africa: “Deleters” vs. “preservers”?

Mesthrie & Bhatt (2008: 90–92) suggest a broad dichotomy among World Englishes of “deleters” vs. “preservers”. Varieties qualifying as “deleters” favour the deletion of elements found in Standard (L1) English (e.g. in Colloquial Singapore English deletion of pronominal subjects or objects as in must buy for him ‘We must buy it for him’, deletion of articles I don’t have ticket, deletion of copula (The) house very nice, or absence of inflectional endings as in She always borrow money from me, He eat here yesterday, or She queue up very long to buy ticket for us; all examples taken from Mesthrie in press). “Preservers”, by contrast, are varieties like Black South African English which disfavour the deletion of elements. More exactly, following Mesthrie (2006), they distinguish three types of anti-deletion: (a) undeletion as in BlSAfE Even the teachers at school made us to hate the course, As you know that I am from Ciskei, As I made it clear before, … or Come what may come. Here an element is restored that is often assumed to be deleted or to have an empty node in generative analyses of English; (b) non-deletion, i.e. the presence of a feature of Standard English that is deleted in some non-standard L1 dialects of English, i.e. the relevant non-standard varieties do not exhibit copula deletion, absence of do-support, pro-drop, or gapping; (c) insertion, i.e. grammatical morphemes are inserted which are not found in the underlying structure of Standard English, as in Although I’m not that shy, but it’s hard for me to make friends, or the use of can be able for ‘can’. Geography comes in, and with it the importance of the relevant substrate languages, at the point where Mesthrie & Bhatt (2008) consider (South East)

Geography in morphosyntactic variation

175

Fig. 3. Phenogram based on all 40 “deletion” and 37 “preservation/addition” features in the WAVE feature set: Africa vs. Southeast Asia

Asian varieties of English to be typically of the deleting type (due to the deleting nature of the Chinese substrate) and African varieties of the preserving type (due to the preserving nature of sub-Saharan African languages). This claim is too good not to be tested against the WAVE data set. In a first step, therefore, all those features were identified from the total of 235 features which, according to the definitions by Mesthrie (2006) and Mesthrie & Bhatt (2008), qualify either as features of the deleting type (40 in all) or preserving type (37 in all). The relevant subsets of features are listed in Tables 3 and 4 respectively (see Appendix). In a second step, a phenogram was generated which was run only against the 77 deletion and preservation features in order to see whether the relative

176

Bernd Kortmann

positioning of African and South East Asian varieties, within each of the areal groups, but also relative to each other and the varieties from other world regions would change compared with the phenogram in Figures 1 and 2. This phenogram is presented as Figure 3. The answer that it offers to the question whether Mesthrie & Bhatt’s broad geographical claim is confirmed is, largely, negative. For most of the African and South East varieties of English there are hardly any changes to be observed in the positioning relative to each other and to varieties in other parts of the Anglophone world. In Figure 3 changes can be observed only for Cameroon English, which has moved from the core L2 branch to the margin of the L1c/Creole complex (see thin arrow), and, above all, for Colloquial Singaporean English, which has moved from a position separating the core L2 varieties from the Pacific Pidgins (see the arrow in Figure 1) right into the Creoles (see bold arrow in Figure 3), now assuming a position between Trinidadian Creole and Hawaiian Creole. This is indeed a change fully in line with Mesthrie’s and Mesthrie & Bhatt’s claims concerning Singaporean English as the prototypical instance of a “deleter” variety of English. The more general claim, however, concerning South East Asian “deleters” and African “preservers” is not borne out when tested against the WAVE data. Counterexamples to this generalization can easily be found. For instance, deletion of subject pronouns and/or dummy it is attested for both Indian and Black South African English as well as, for example, East African Englishes; dummy it deletion and object pronoun deletion are attested even for an African L1 variety, namely White South African English. The “deletion” vs. “preservation” criterion promises to be more fruitful when capturing differences between variety types, especially between Pidgins and Creoles (as “deleters”) and L1 varieties (“preservers”). However, the phenogram in Figure 3 does lend support to the recent qualifications which Mesthrie (in press) made with regard to a continuum along the parameter ‘deletion vs. preservation of grammatical formatives’ on which varieties of English can be arranged. At one pole of this continum we find a prototypical deleting variety like Colloquial Singaporean English while a prototypical preserving variety like Black South African English is located at the opposite pole. Mesthrie writes (2012: 100): Speakers with substrates which disfavour deletion fall on the BlSAfE side of the continuum; whereas New Englishes whose substrates are isolating languages like Hakka or Mandarin fall on the SgE end. The continuum makes provisions for varieties falling in between. This continuum can be overridden if there are special circumstances of acquisition – e.g. if there is sufficient access to StE to minimise substrate influence. This appears to be the case with Hong Kong English, though further research is certainly warranted here. The predictions made in this ap-

Geography in morphosyntactic variation

177

proach are that if the EFL varieties in China, Thailand, Laos etc. were to turn into more stable L2s (which is not unlikely in the global village), they will most likely fall on the SgE rather than the BlSAfE end of the continuum.

The stable positions of Hong Kong English and Malaysian English among the L2 varieties in Figure 3 may well be explained by these “special circumstances” overriding the deletion/ preservation continuum.

5.

English-based Pidgins and Creoles: West Africa vs. Caribbean vs. Pacific/Australia

In this section we will zoom in on one particular variety type, Pidgins and Creoles, and will demonstrate useful steps and strategies for identifying areal patterns. The same procedure can be used for the other variety types (and has been used, for example, for identifying a set of truly pan-British Isles features and, within the British Isles, a set of features constituting a large Northern area, and another set for a large Southern area; cf. Kortmann & Langstrof 2012: 1945–1947). 5.1

Steps in identifying areal patterns

Two steps are necessary in order to arrive at the most promising set of candidates among the WAVE features exhibiting the strongest geographical signal. These steps are necessary to ensure that features are not mistaken to form part of a distinctive areal pattern which qualify as either (Step 1) vernacular angloversals, i.e. features found in the vast majority of all non-standard (L1, L2, Pidgin and Creole) varieties of English around the world, or (Step 2) so-called varioversals, in this case features found in the vast majority of all English-based Pidgins and Creoles (for a detailed discussion of (criteria for) vernacular angloversals, varioversals and areoversals, cf. Szmrecsanyi & Kortmann 2009b). WAVE allows us to identify the relevant sets of features very quickly. Table 5 lists those features with the widest distribution in the Anglophone world (found in at least 80 % of all 74 data sets, i.e. with a minimum Mean of 0.80), and Table 6 the next most widely found features (found in at least 70 %, i.e. minimum Mean of 0.70, of the WAVE set of varieties).4 4

For all of the following statistics note that they are based on a prefinal data set. The final WAVE data set as published in Kortmann & Lunkenheimer 2012 was not yet available when this paper was written. There may therefore be minor differences in the percentages and in tue order of tue features in the following tables when compared to tue corresponding tables and statistics in Kortmann & Lunkenheimer 2012.

178

Bernd Kortmann Sum

Mean

221. adverbs other than degree adverbs have the same form as adjectives

68

0.92

229. no inversion/no auxiliaries in main clause yes/no questions

67

0.91

34. forms or phrases for the 2nd person plural pronoun other than you

66

0.89

64

0.86

154. multiple negation

7. me instead of I in coordinate subjects

61

0.82

159. never as preverbal past tense negator

60

0.81

220. degree modifier adverbs have the same form as adjectives

59

0.80

Table 5: The top 7 candidates for vernacular angloversals (found in at least 80 % of all 74 data sets)

147. was for conditional were 78. double comparatives and superlatives

Sum

Mean

58

0.78

54

0.73

172. existential/presentational there’s/there is/there was with plural 53 subjects

0.72

216. omission of StE prepositions (not necessarily with prepositional verbs, but e.g. locative Ps and Ps before temporal expressions)

52

0.70

158. invariant don’t for all persons in the present tense

52

0.70

52

0.70

80. regularized comparison strategies: extension of analytic marking

Table 6. The runners-up for vernacular angloversals (found in at least 70 % of the 74 data sets)

In a second step we need to identify the most widely found features in the variety type we are most interested in, i.e. for the purposes of the current section in the 26 English-based Pidgins and Creoles documented in WAVE. If we settle again on the 80 % threshold, we arrive at the 16 candidates for Pidgin and Creole varioversals in Table 7. Four of these, however, are vernacular angloversals, namely those with a “Yes” or “(Yes)” in the two rightmost columns. This means that they also occur in at least 80 % (“Yes”) or 70 % (“(Yes)”) of all 31 L1 and 17 L2 varieties in WAVE. This leaves us with 12 P/C varioversals.

179

Geography in morphosyntactic variation

Total of % of all also also varieties varieties top in top in with L1 L2 feature 228. no inversion/no aux’s in wh-questions 25

96.2 %

177. deletion of copula be: before AdjPs

25

96.2 %

25

96.2 %

221. adverbs other than degree adverbs have the same form as adj’s

24

92.3 %

Yes

(Yes)

229. no inversion/no aux’s in main clause yes/no questions

24

92.3 %

Yes

Yes

34. forms or phrases for the 2nd person plural pronoun other than you

24

92.3 %

Yes

Yes

216. omission of StE prepositions (not 23 nec’ly with prep. verbs, but e.g. locative Ps and Ps before temporal expressions)

88.5 %

14. no number distinction in reflexives (i.e. 23 PL forms ending in -self)

88.5 %

178. deletion of copula be: before locatives 22

84.6 %

114. go-based future markers

22

84.6 %

77. omission of genitive suffix; possession 22 expressed through bare juxtaposition of nouns

84.6 %

Yes

(Yes)

66. indefinite article one/wan

10. no gender distinction in third person singula 111. past tense/anterior marker been 154. multiple negation 160. no as preverbal negator 5. generalized 3rd person singular pronoun: subject pronouns

22

84.6 %

21 21 21 21

80.8 % 80.8 % 80.8 % 80.8 %

Table 7. 12 top candidates for Pidgin and Creole varioversals (plus 4 vernacular angloversals)

Once these two layers of features are taken away, we can use, in a third step, a simple visualization technique in order to detect areal patterns for the remaining features, i.e. features covering the middle ground between, say, 10 % and 80 % of all Pidgins and Creoles in the WAVE data set. With cheek in tongue, this technique may be said to produce “Grammar Genome Maps”

180

Bernd Kortmann

since the outcome is reminiscent of genome maps (see Figure 4 for a rough visual impression). The simple idea is to introduce a colour code for the WAVE matrix which underlies eWAVE and all statistical analyses of the WAVE data set. Each column represents a Pidgin or Creole, each line a feature. A-rated features are coloured red, B-rated ones green, C-rated ones yellow; the absence of a feature in a variety (i.e. rating D, X, or “?”) is indicated by the absence of colouring. Thick vertical bars separate the major areal divisions among the Pidgins and Creoles in WAVE: Section I represents West Africa, Section II the Caribbean, Section III the Pacific, and Section IV Australia. The Pacific is separated into Hawaiian Creole (IIIa) and the ‘narrow’ Pacific (IIIb). Lines and columns are now arranged in a way to create (a) maximally filled and (colourwise) homogeneous sections and (b) maximally contrasting quadrangles (or: sectors). Thus in Figure 4, the top left quadrangle (Sector A I) contrasts with all quadrangles to the right of it, i.e. these are features found (fairly pervasively even, given the dominantly red and green colouring) in the Pidgins and the Creole (Krio) of West Africa, but only rarely in the Caribbean Creoles (Sector A II), and hardly at all in the Pidgins and Creoles of the Pacific (Sector A III) and Australia (Sector A IV). For the features in the B-sectors, it is West Africa and the Caribbean (Sectors B I and II) which both run against the Pacific and Australia (Sectors B III and IV), where these features are rarely or not at all found. The features in the C-sectors are found rather pervasively in most Caribbean Creoles (Sector C II), but hardly in any of the other Pidgin/Creole world regions. Moving towards the bottom of Figure 4, we end with the F-/G- and H-sectors, all of which have in common that the relevant features are pervasively found (mind the concentration of red and green colouring) in the Pidgins and Creoles of the Pacific and/or Australia, but not nearly as pervasively or not at all in the majority of the Pidgins and Creoles of West Africa and the Caribbean. In F III and F IV the relevant features are found in the “wide” Pacific (i.e. including Hawaiian Creole) and Australia, in G III and G IV in the “narrow” Pacific and Australia, and in H IV pervasively only in the two Australian Creoles (Roper River Creole and Torres Strait Creole). This kind of arrangement in a Grammar Genome Map (short: GGM) can in principle be automatized and serve as input to the researcher in order to make most sense of it. With the help of such a GGM the following generalizations can be made for areal patterns among the 26 English-based Pidgins and Creoles in WAVE. To start with, it is possible to identify clear regional contrasts. Secondly, the Caribbean Creoles boast the largest number of areally distinctive

Geography in morphosyntactic variation

181

Fig. 4. Grammar Genome Map for the 26 Pidgins and Creoles in WAVE

features of all four Pidgin/Creole world regions. Thirdly, where two of these world regions pattern with each other, it is either West Africa which does so with the Caribbean, or (but in a much less pronounced way) the Caribbean with the Pacific and Australia. Only rarely does West Africa pattern with the Pacific and Australia. Let us now zoom in on the individual Pidgin and Creole world regions one by one.

182 5.2

Bernd Kortmann

West Africa

Out of 8 features which are documented in all four West African Pidgins as well as in Krio, there are only three which are exclusively pervasive and virtually absent elsewhere in Pidgins and Creoles and other variety types: F 59 double determiners (elsewhere: only in Saramaccan) Dí yò tú bóy pìkín dèm chóp. (NigP) This 2SG POSS two boy child 3PL eat These your two boys ate. F 116 come-based FUT/INGRESS markers (elsewhere: only in Bislama and Butler E) Ìm kóm yélo wélwél. (NigP) 3sg.SJ become be.pale thoroughly He/she/it became very pale. F 73 existential construction to express possessive (elsewhere: only in Tok Pisin) Mà mòto dé. (NigP) 1SG.POSS vehicle COP My car exists. (I have a car.) 5.3

Caribbean

Out of the most pervasive features in the Caribbean Creoles, only five are exclusively pervasive and virtually absent elsewhere in Pidgins and Creoles and other variety types: F 13 subject PRO forms as bases for reflexives (elsewhere: only in Gullah, BrC, ButlE, RRC) if you see a person call theyself a fisherman, and he can’t swim, ain’t go in no boat! (BahC) F 17 fo/fi + pers.PRO as poss. PRO (elsewhere: only in Norf ’k, BrC) Dat a foyu prablem. (VinC) That cop poss problem ‘That’s your problem.’

Geography in morphosyntactic variation

183

F 18 subject PRO as poss. PRO (elsewhere: only in ButlE, BrC, RRC) We mi ting de? (VinC) Where poss thing loc ‘Where is my thing?’ F 91 do as HAB marker (elsewhere: only in Gullah, BrC) We does pray for when Jesus come – all will be well (BahC) F 109 Perfect marker already (elsewhere: only in GhP, HawC) Were you there already? ‘Have you been there before?’ (BelC) Eastern Maroon Creole clearly is the odd one out among the Caribbean Creoles, lacking many features that the majority has. Moreover, but not surprisingly, British Creole patterns clearly with the Caribbean Creoles (cf. also 5.5 below). 5.4

Pacific and Australia

Hawaiian Creole patterns mostly with the “narrow” Pacific and Australia, far less with the Caribbean Creoles, and not at all exclusively with West Africa (i.e. there is not a single morphosyntactic feature attested in Hawaiian Creole that is found in no other but a West African Pidgin or Creole). Characteristic of the ‘wide’ Pacific in contradistinction to the other three Anglophone Pidgin and Creole world regions are the following three: F 38 specialized plural markers for pronouns (elsewhere only in BelC) us-gang (1PL); as gaiz (1PL), yu gaiz (2PL), dem gaiz (3PL) (all from HawC) F 42 object pronoun drop (elsewhere only in BelC, Gullah, VLibE) That’s why you calls Ø a teacher (referent = ‘me’) (PalmE) F 44 subject pronoun drop: dummy pronouns (elsewhere only in BahC, Gullah, NigP, ButlE) Rienen ‘it is raining’ (Norf ’k) The most distinctive features of the Pidgins and Creoles of the “narrow” Pacific and Australia are the following: F 36 distinct forms for incl/excl 1st person non-singular (elsewhere only in Sranan) Himii ‘you and me’ vs. mienhem ‘me and him’ (Norf ’k)

184

Bernd Kortmann

F 37 more number distinctions in personal pronouns than Sg/Pl Hemii goe nawii ‘let the two of us go swim’ (Norf ’k) F 76 possession: postnominal phrases with bilong (elsewhere only in Krio) Em I haus bilong em ‘It is his house’ (TP) F 87 attributive adjectival modifiers follow head noun (elsewhere only in NigP) man nogud ‘bad person’ (Bisl) F 143 transitive verb suffix -em/-im/-um (elsewhere only in GuyC) Tok ‘to talk, converse’ vs. tokim ‘to say (something), to tell (someone)’ (TP) 5.5

Pidgins and Creoles in other Anglophone world regions

Geographically speaking, there are three Pidgins and Creoles in the WAVE data set which can be considered outliers in the sense that they are spoken in Anglophone world regions which otherwise qualify as almost exclusively L1 or L2 regions. These are Gullah (in North America), British Creole (in the British Isles) and Butler English (in South Asia, more exactly India). Especially for the first two, we receive a very strong geographical signal where they exhibit a morphosyntactic feature characteristic of their world region which is not documented at all for the other Pidgins and Creoles, not even for closely related ones (like Jamaican Creole as a source for British Creole). Thus British Creole has F 155 (ain’t for negated be) and F 156 (ain’t for negated have), both typical of the Southeast of England. British Creole also features the typical Southern English English F 165 (the invariant tag innit), but this case is not as strong since F 165 is also found in Jamaican Creole. That feature which gives away Gullah as a distinctively North American Creole is F 218 (positive anymore ‘nowadays’), as in Colloquial American English Anymore them crows just come and eat all the corn. For Butler English the geographical signal “India” cannot be pinpointed by a particular feature: the one coming closest to such a specifically South Asian feature is F 230 ‘Doubly filled COMP-position with wh-words’, as in What who has eaten? ‘Who ate what?’, which is shared by Butler English, Indian English, and Pakistan English. However, this feature is also documented in three varieties spoken in other Anglophone world regions, namely Irish English, Tristan da Cunha English, and Belizean Creole. The strong Indian signal for Butler English can rather be read off the phenogram in Figures 1 and 2: its overall morphosyntactic

Geography in morphosyntactic variation

185

profile makes it a true outlier among the Pidgins and Creoles in that its position in the phenogram is far away from them, right next to Indian English (and Indian South African English). Thus, although classified by the WAVE informant as a restricted Pidgin given the socio-cultural conditions in which it emerged and is still (though increasingly rarely) used, with regard to its morphosyntax Butler English patterns with the L2 and high-contact L1 varieties in the “Northeastern” part of the phenogram.

6.

Scenarios how geography can matter

Compared with variety type, geography is only of secondary importance as an explanatory and prognostic factor – this has been argued throughout this paper. When looking at the overall morphosyntactic and complexity profiles of varieties of English around the world, there are no geographical patterns which could explain more than the dominant variety type(s) in the relevant Anglophone world region. Geography starts to matter more strongly, though, as soon as we explore the individual features in each variety. At least three relevant scenarios can be distinguished in which we can expect to see area-specific patterns and geographical differences between varieties of the same variety type. Before we discuss these three scenarios of geographical patterning, it should be noted that such patterns on a larger scale contrast with small-scale differences, like specific features occurring exclusively in a few varieties of English spoken in one particular world region or dispersed across different world regions, or special local constraints obtaining with regard to otherwise widely found morphosyntactic features. Nice examples of highly geographically restricted features are progressive was sat and was stood, which are exclusively found in the British Isles (in all Celtic Englishes, the dialects of Northern England and the Southwest of England). Local constraints on otherwise widely found features – even features up to the level of varioversals and vernacular angloversals – may include specific syntactic contexts (including specific constructions, in which a given feature is exclusively found), special discourse texts (as investigated for example in the new research area of variational pragmatics), or a special semantics that goes with the relevant features in a given variety or set of varieties in a given Anglophone world region (for a discussion of a number of such features cf. e.g. Sharma 2009, Tagliamonte 2009, or Davydova et al. 2011). The first scenario in which areal patterns in grammatical variation can be observed concerns the degree of pervasiveness to which a given (set of) morphosyntactic feature(s) is represented in the different world regions. Pervasiveness is to be understood in two ways here: (a) regardless whether it was

186

Bernd Kortmann

rated A (highly frequent), B (neither frequent nor extremely rare) or C (extremely rare), the relevant WAVE feature is observable in at least 75 per cent of the varieties of the relevant world region and is a top feature in no other world region;5 (b) the relevant feature received an A rating in the vast majority, ideally every single, of the varieties of the given world region. For a detailed account of these two types of pervasiveness with many illustrations from the Anglophone world, compare Kortmann (2010b), Kortmann & Szmrecsanyi (2004: 1160–1183). The second scenario which may reveal large-scale areal patterns concerns the degree to which a certain aspect of the overall morphosyntactic profile is borne out in the individual varieties (e.g. the degree of inflectional marking or its loss, or the degree of grammaticity, i.e. overall degree of employing grammatical markers, be they synthetic or analytic; both degrees are measured as token frequencies in natural discourse material). Thus Southeast Asian varieties of English, especially Colloquial Singaporean English and Hong Kong English, seem to be far more radical in getting rid of inflectional endings and grammaticity, in general, than L2 varieties in other parts of the world (for details compare Kortmann & Szmrecsanyi 2009, 2011, Szmrecsanyi & Kortmann 2009c, 2011). The most likely reason for this, as discussed in section 4, is the typological nature of the most influential local substrate language(s). This leads us straight to the heart of the “geographical signal”, as it were, and to the third scenario. Especially for L2 varieties, Pidgins and Creoles, but also for many highcontact L1 varieties, the major source of the geographical signal is language contact between (typically non-standard varieties of) English with one or more indigenous language(s) in the relevant Anglophone world region. This is not simply a historical given, but for many hundreds of millions of speakers part of their everyday linguistic reality, which is characterized by bi- or multilingualism. Geography thus typically translates into substrate effects leaving their imprint on the morphosyntax of varieties of English, a point made very clearly by Sharma (2009: 191): Surface similarities across New Englishes can be skin deep, diverging dramatically upon closer examination, due to substrate systems or substrate-superstrate interaction. […] The degree and distribution of a given feature must be understood in relation to the substrate before any universal claims can be made …

5

For a similar measure used for identifying the top diagnostic features for each area (attestation rate per anglophone world region relative to overall attestation rate elsewhere in the anglophone world), see Kortmann & Wolk 2012: 927–929.

Geography in morphosyntactic variation

187

World region and specific substrate language(s) do matter once it comes to the choice of specific constructions (e.g. special complementation patterns for verbs and adjectives, choice of particles or prepositions in phrasal or (phrasal-) prepositional verbs) and, in particular, to specific lexical material used for coding a given grammatical function (e.g. the choice of tense and aspect markers, of pronouns, or of prepositions and conjunctions). It is exactly properties of this nature which play a central role in Schneider’s (2007) 5-phase model for the evolution of Postcolonial Englishes, more exactly for the step from stage 3 “nativization” to stage 4 “endonormative stabilization” – without wanting to claim that all of the relevant phenomena are due to substrate influence. Schneider writes: Innovations and distinctive structural properties of PCEs [Postcolonial Englishes, B.K.] are frequently positioned at the interface between lexis and grammar, i.e. certain words but not others of the same word class prefer certain grammatical rules or patterns. The patterns as such are not new, nor are the words, but what is novel is the habitual association between them in specific varieties. (Schneider 2007: 83)

However, many of the relevant phenomena documenting the development of endonormative rules are too subtle to be captured by the WAVE survey with its worldwide scope and typological approach (cf. also our remarks on the limitations of WAVE at the very end of section 7). What WAVE is able to capture, on the other hand, is one important process of language change which is responsible for the emergence of specific substrate-induced elements and constructions in the grammars of, especially, L2s, Pidgins and Creoles, namely grammatical replication. This is the term Heine & Kuteva (2006) introduced for structural change which is due to language contact, with restructuring and, above all, contact-induced grammaticalization as its two major types. Kortmann & A. Schneider (2011) identify more than 70 features from the WAVE feature set, i.e. almost a third of the entire set, that can be argued to qualify as instances of various types of grammaticalization, notably including contact-induced grammaticalization, and discuss their distribution across different domains of grammar and the L1 and L2 varieties, Pidgins and Creoles in WAVE. The role of language contact and thus substrate influence is central in all three scenarios of large-scale geographical patterning. (Note for the first scenario, too, that the pervasiveness of a given feature in a given world region may well be a reflex of a shared substrate or a language contact situation of long standing, which may continue into the present.) Ultimately then, the geographical signal in varieties of English (at least in all non-low contact L1s) is a derivative of the geographical signal of their substrate and contact lan-

188

Bernd Kortmann

guages. In a given world region, recurrent structural features of the latter will also find their way into the relevant varieties of English – and likewise, where such varieties exist, into the relevant varieties of other major European superstrate languages, such as Spanish, French or Portuguese. If we take away all contact-/substrate-induced morphosyntactic features, the geographical signal on a large scale, i.e. distinguishing from each other entire Anglophone, Francophone, or Hispanophone world regions when discussing varieties of English, French or Spanish respectively, will largely disappear. So even in scenarios of areal patterning where geography does matter on a larger scale, it is not so much the geography of varieties of English (or French, Spanish, Portuguese, etc.) which matters but that of their contact and substrate languages.

7.

Conclusion

The two most important guiding questions of the entire volume which have informed the present paper were the following: (a) To what extent is geographic distance, as a determinant of linguistic variability, real?, and (b) How powerful is it as an explanatory factor of variation? In a nutshell, these two questions can be answered as follows. Geographic distance is real, but largely on the microlevel, especially (but by no means exclusively) for traditional dialects (L1t) of English in England and North America. As an explanatory factor of variation, geography is far less powerful than variety type at the macrolevel of the forests (recall the comparison of Figures 1 and 2), but it is potentially very powerful on the microlevel, i.e. at the level of the trees down even to the level of the leaves (e.g. the level of local syntactic, semantic or pragmatic constraints, specific constructions, complementation patterns, or grammaticalization processes). Moreover, geography matters in different ways for different variety types. For traditional dialects (L1t) geographical proximity or distance does indeed help account for linguistic similarity or diversity. Stable long-time location in contiguous areas can clearly be shown to result in structural similarity of the varieties in question (for the British Isles compare most recently the results of corpus-based dialectometry in Szmrecsanyi 2012). To a certain extent, this can also be assumed in dialect contact scenarios as they are typical for the formation periods of transplanted (L1c) varieties of English (e.g. in New Zealand). But even for L1c varieties, and above all for L2 varieties of English and English-based Pidgins and Creoles geographical proximity/distance does not matter nearly that much as an explanatory factor of morphosyntactic variation as for L1t varieties.

Geography in morphosyntactic variation

189

It should also be noted that, as known from numerous dialectological studies, the areal patterns emerging from large-scale surveys of morphosyntactic feature sets differ significantly from those of phonetic-phonological feature sets. Such differential behaviour has been shown in two recent studies. Wolk (2009) used the two feature sets in the forerunner version of WAVE on the CD-Rom accompanying Kortmann et al. (2004). Brato & Huber (2012), exploring areality in Anglophone Africa, used the same feature set only for phonetic-phonological features and for morphosyntactic variation in Anglophone sub-Saharan African a preliminary version of WAVE. It is one of their major findings “… that African Englishes exhibit what we would like to call ‘differential areality’: the phylogenetic networks produced from the phoneticphonological and morphosyntactic feature lists are only partially congruent” (Brato & Huber 2012: 180). In both studies, contrasting the morphosyntactic with the relevant phonetic-phonological phenogram yields the following picture: there are different groupings of the varieties, with the morphosyntactic phenograms exhibiting fewer and clearer clusters, whereas the phenograms based on the phonetic-phonological feature set are more diffuse and display much stronger regional (in part, downright local) signals. In other words, at the level of phonetics and phonology, geography accounts for far more variation than it does on the morphosyntactic level. On a global level, this can be interpreted as to reflect the well-known dictum in studies on language-internal variation, namely that syntax unites and accent divides. To round off, two remarks are necessary with regard to the nature of the survey data used in this paper. On the one hand, WAVE is of great value as a neutral source and yardstick for comparative studies on the morphosyntax of varieties of English and English-based Pidgins and Creoles around the world. The focus of the present paper was areality, but WAVE would be equally useful when embarking on a large-scale comparison within the Anglophone world from a different, or more specific perspective. For example, Lunkenheimer (2012) has made an earlier version of WAVE the basis for exploring areality in the Anglophone world for a specific domain of grammar, namely tense and aspect. On the other hand, a typology-style survey like WAVE necessarily glosses over much of variety-internal linguistic and social diversity and is not able to capture all the fascinating variation at the microlevel of, for example, local constraints (be they syntactic, semantic, pragmatic, stylistic, sociolinguistic) on morphosyntactic features. Unfortunately, this is exactly where much of the hidden “power of geography” is to be found. For every single WAVE feature in every single of the 74 L1s, L2s, Pidgins and Creoles in WAVE it is thus worth zooming in on the usage level in natural discourse data. This is where variety-specific corpus-based studies come in, such

190

Bernd Kortmann

as (quantitative as well qualitative) studies based on the corpora forming part of the constantly growing ICE – International Corpus of English – family. Only in tandem will WAVE, as a macrolevel survey, and fine-grained studies exploring individual morphosyntactic properties at the microlevel of individual varieties bring out the full power of geography as an explanatory factor in morphosyntactic variation within the Anglophone world.

Appendix nr

description

nr

description

042 object pronoun drop

160 no as preverbal negator

043 subject pronoun drop: referential pronouns

161 not as a preverbal negator

044 subject pronoun drop: dummy pronouns

170 invariant present tense forms due to zero marking for the third person singular

046 deletion of it in referential it is-construc- 174 deletion of auxiliary be: before progresstions ive 047 deletion of it in non-referential it is-con- 175 deletion of auxiliary be: before gonna structions 056 absence of plural marking only after quantifiers

176 deletion of copula be: before NPs

057 plural marking generally optional (i.e. regardless of the presence of a quantifier): for nouns with human referents

177 deletion of copula be: before AdjPs

058 plural marking generally optional (i.e. regardless of the presence of a quantifier): for nouns with non-human referents

178 deletion of copula be: before locatives

062 use of zero article where StE has definite 179 deletion of auxiliary have article 063 use of zero article where StE has indefi- 193 gapping/zero-relativization in subject nite article position 077 omission of genitive suffix; possession 198 deletion of stranded prepositions in expressed through bare juxtaposition of relative clauses (“preposition chopnouns ping”) 084 comparative marking only with than

199 reduced relative phrases preceding head-noun

086 zero marking of degree

208 deletion of to before infinitives

099 levelling of the difference between pres- 210 non-finite clause complements with bare root form rather than -ing form ent perfect and simple past: simple past for StE present perfect

191

Geography in morphosyntactic variation 101 simple present for continuative or experiential perfect

213 no subordination; chaining construction linking two main verbs (motion and activity)

117 present tense forms for neutral future reference

216 omission of StE prepositions (not necessarily with prepositional verbs, but e.g. locative prepositions and prepositions before temporal expressions)

129 levelling of past tense/past participle verb forms: unmarked forms

220 degree modifier adverbs have the same form as adjectives

132 zero past tense forms of regular verbs

221 other adverbs have the same form as adjectives

158 invariant don’t for all persons in the pres- 228 no inversion/no auxiliaries in wh-quesent tense tions 159 never as preverbal past tense negator

229 no inversion/no auxiliaries in main clause yes/no questions

Table 8. Deletion features in WAVE (40)

nr

description

nr

description

009 benefactive “personal dative” construc- 120 would in if-clauses tion (using the object form of the pronoun) 016 emphatic reflexives with own 017 creation of possessive pronouns with prefix fi- +personal pronoun

121 double modals 133 double marking of past tense

038 specialized plural markers for pronouns 134 a-prefixing on ing-forms 039 plural forms of interrogative pronouns: 135 a-prefixing on elements other than ingusing additional (free or bound) elforms ements 040 plural forms of interrogative pronouns: 143 transitive verb suffix -em/-im/-um reduplication 045 insertion of it where StE favours zero

146 use of verbal suffix -ing with forms other than present participle/gerund

055 different count/mass noun distinctions 154 multiple negation / negative concord resulting in use of plural for StE singular 059 double determiners (e.g. demonstrative/ 191 relativizer doubling article + possessive pronoun, with possessive pronoun preposed or postposed) 064 use of definite article where StE favours 194 resumptive/shadow pronouns zero 065 use of indefinite article where StE favours zero

196 correlative constructions

192 nr

Bernd Kortmann description

nr

description

070 proximal and distal demonstratives with 202 unsplit for to in infinitival purpose clauses ‘here’ and ‘there’: this here, that there; dis-ya, dis-de; dem-ya, dem-de; dis ya/dis de, dem ya/ dem de, etc. 074 phrases with for + noun to express pos- 203 for (to) as infinitive marker session: for-phrase following possessed NP; the for-phrase may precede or follow the possessed NP 075 phrases with for + noun to express pos- 204 as what / than what in comparative clauses session: for-phrase preceding possessed NP 076 postnominal phrases with bilong/blong/ long/blo to express possession

207 substitution of that-clause for infinitival subclause

078 double comparatives and superlatives

209 addition of to where StE has bare infinitive

100 levelling of the difference between pres- 214 conjunction doubling: clause + conj. + ent perfect and simple past: present perconj. + clause fect for StE simple past 103 do as unstressed tense marker (without habitual or other aspectual meanings)

215 conjunction doubling: correlative conj.s 233 presence of subject in imperatives

Table 9. Preservation and addition features in WAVE (37)

References Brato, Thorsten & Magnus Huber 2012: English in Africa. In: Raymond Hickey (ed.), Areal Features of the Anglophone World, 161–185. Berlin/Boston: de Gruyter Mouton. Davydova, Julia, Michaela Hilbert, Lukas Pietsch & Peter Siemund 2011: Comparing varieties of English: problems and perspectives. In: Peter Siemund (ed.), Linguistic Universals and Language Variation, 291–323. Berlin/Boston: de Gruyter Mouton. Kortmann, Bernd 2010a: Areal variation in syntax. In: Peter Auer & Jürgen Erich Schmidt (eds.), Language and Space: An International Handbook of Linguistic Variation. Vol 1: Theories and Methods, 837–864. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd 2010b: Variation across Englishes. In: Andrew Kirkpatrick (ed.), Routledge Handbook of World Englishes, 400–424. London: Routledge. Kortmann, Bernd & Kerstin Lunkenheimer (eds.) 2012: The Mouton World Atlas of Variation of English. Berlin/Boston: de Gruyter Mouton. Kortmann, Bernd, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie & Clive Upton (eds.) 2004: A Handbook of Varieties of English. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd & Christian Langstrof 2012: Regional varieties of British English. In: Alexander Bergs & Laurel Brinton (eds.), English Historical Linguistics: An International Handbook, 1929–1950. Berlin/Boston: Mouton de Gruyter.

Geography in morphosyntactic variation

193

Kortmann, Bernd & Kerstin Lunkenheimer (eds.) 2011: The Electronic World Atlas of Variation in English. Leipzig: Max Planck Institute for Evolutionary Anthropology. http://ewave-atlas.org. Kortmann, Bernd & Agnes Schneider 2011: Grammaticalization in non-standard varieties of English. In: Bernd Heine Heiko Narrog (eds.), The Oxford Handbook of Grammaticalization, 263–278. Oxford/New York: Oxford University Press. Kortmann, Bernd & Benedikt Szmrecsanyi 2004: Global synopsis: morphological and syntactic variation in English. In: Bernd Kortmann, Edgar W. Schneider, Kate Burridge, Rajend Mesthrie & Clive Upton (eds.), A Handbook of Varieties of English, Vol. 2, 1142–1202. Berlin/New York: Mouton de Gruyter. Kortmann, Bernd & Benedikt Szmrecsanyi 2009: World Englishes between simplification and complexification. In: Lucia Siebers & Thomas Hoffmann (eds.), World Englishes –– Problems, Properties and Prospects: Selected Papers from the 13th IAWE Conference, 265–285. Amsterdam/Philadelphia: Benjamins. Kortmann, Bernd & Benedikt Szmrecsanyi 2011: Parameters of morphosyntactic variation in World Englishes: prospects and limitations of searching for universals. In: Peter Siemund (ed.), Linguistic Universals and Language Variation, 264–290. Berlin/Boston: de Gruyter Mouton. Lunkenheimer, Kerstin 2012: Tense and aspect. In: Raymond Hickey (ed.), Areal Features in the Anglophone World, 329–358. Berlin/Boston: de Gruyter Mouton. Mesthrie, Rajend 2006: Anti-deletions in an L2 grammar: A study of Black South African English mesolect. English World-Wide 27: 111–145. Mesthrie, Rajend 2012: Deletions, antideletions and complexity theory, with special reference to Black South African and Singaporean Englishes. In: Bernd Kortmann & Benedikt Szmrecsanyi (eds.), Linguistic Complexity: Second Language Acquisition, Indigenization, Contact, 90–100. Berlin: De Gruyter. Mesthrie, Rajend & Rakesh Mohan Bhatt 2008: World Englishes: The Study of New Linguistic Varieties. Cambridge: Cambridge University Press. Schneider, Edgar W. 2007: Postcolonial English: Varieties around the World. Cambridge: Cambridge University Press. Sharma, Devyani 2009: Typological diversity in New Englishes. English World-Wide 30: 170–195. Szmrecsanyi, Benedikt 2012: The Geolinguistics of Grammatical Variability in Traditional British English Dialects: A Large-Scale Frequency-Based Study. Cambridge: Cambridge University Press. Szmrecsanyi, Benedikt & Bernd Kortmann 2009a: The morphosyntax of varieties of English worldwide: a quantitative perspective. Lingua 119(11): 1643–1663. Szmrecsanyi, Benedikt & Bernd Kortmann 2009b: Vernacular universals and angloversals in a typological perspective. In: Markku Filppula, Juhani Klemola, Heli Paulasto (eds.), Vernacular Universals and Language Contacts: Evidence from Varieties of English and Beyond, 33–53. London/New York: Routledge. Szmrecsanyi, Benedikt & Bernd Kortmann 2009c: Between simplification and complexification: Non-standard varieties of English around the world. In: Geoffrey Sampson, David Gil & Peter Trudgill (eds.), Language Complexity as an Evolving Variable, 64–79. Oxford: Oxford University Press. Szmrecsanyi, Benedikt & Bernd Kortmann 2011: Typological profiling: learner Englishes versus indigenized L2 varieties of English. In: Joybrato Mukherjee & Marianne Hundt (eds.), Exploring Second-Language Varieties of English and Lear-

194

Bernd Kortmann

ner Englishes: Bridging a Paradigm Gap, 167–187. Amsterdam/Philadelphia: Benjamins. Trudgill, Peter 2006: New-dialect Formation: The Inevitability of Colonial Englishes. Edinburgh: Edinburgh University Press. Wolk, Christoph 2009: Classifying geographic variation: Morphosyntax and phonology. M.A.Thesis, University of Freiburg, Germany.

Area formation in morphosyntax

195

Elvira Glaser

Area formation in morphosyntax

1.

Introduction

In the following contribution, I proceed from the fact that the spatial distribution of linguistic phenomena in general as well as in dialectology is not yet fully understood. Spatial distributions are, however, a fundamental characteristic of language, and they play an important role in recent typological discussions. Up until now, reasoning about linguistic borders and about the diffusion of variants was primarily based on phonological data. The role and behavior of syntactic variation (in a wide sense, including morphosyntax)1 were nearly completely outside the mainstream. Before dealing with specific problems in (morpho)syntax-centered dialectology (Section 3), I will therefore review the literature on variant distributions (Section 2). Section 4 is concerned with modern dialectological research avenues, and Section 5 discusses basic questions in dialect geography, as well as in contact linguistics and typology. Section 6 concludes and re-emphasizes the fundamental issue of feature distributions as a common challenge in geolinguistics.

2.

The geographic distribution of linguistic features: General considerations

2.1

Terminological clarification

The quest for linguistic differences between neighboring communities has informed dialectological scholarship since its inception. Nerbonne and Heeringa maintain that it is “axiomatic in dialectology that language variety is structured geographically” (2007: 287). In this contribution, I take the undisputable spatial distribution of linguistic features as my starting point. By “spatial distribution” I refer to the fact that linguistic features are neither evenly nor randomly distributed in a speech community or within closely re1

In this contribution syntax is referred to in a rather wide, pre-theoretical sense so as not to exclude interesting grammatical phenomena. This does not mean that the analyst cannot subsequently draw on narrower definitions of morphology and syntax if different spatial distributions are suspected.

196

Elvira Glaser

lated neighboring varieties; instead, they are clustered in certain regions and absent in others. Of course, this does not mean that all linguistic variables show this behavior. There are, however, linguistic features that exhibit an uneven distribution within a given area, a fact that oftentimes puts the native speaker in a position to locate other speakers – an outgrowth of the spatial indexicality (Oesterreicher 2007: 62) of language. In what follows, the notions space and spatial refer to a physical extension in general, i.e. to a confined area where speakers live and use common linguistic features. With respect to a concrete geographical extension, the more or less interchangeable notions area and region are preferred. “Region”, however, is reserved for reference to traditional geographical spaces or in order to differentiate between local places and larger, supra-local regions. So, area is the most neutral expression in this context.2 The differentiation of place versus space in a more social or geographical sense will not concern us in this paper. 2.2

Linguistic boundaries in various linguistic approaches

The relevant theories about the spatial character of dialects largely date back to the early 20th century. Dialect borders were no longer seen as reflecting the settlement borders of old tribal groups. This was a consequence of a change affecting the status of tribes. They were no longer considered fixed entities but had changed into historically and culturally shaped subjects without precise inherited settlement places. The historical and cultural context was seen as determining the position of isoglosses, because often isoglosses were found overlapping with historical boundaries or other aspects of local culture. Thus, it seemed obvious that there must be a correlation between these phenomena, although to this day the exact nature of this correlation is not clear. Despite much criticism of the Kulturraum-paradigm developed by the German kulturmorphologische Schule (cf. Knobloch 2010: 108, 122), the idea that ancient communicative boundaries are mirrored in dialect boundaries survives in modern sociolinguistic research (Britain 2010: 154). After the rejection of 19th century determinism about the influence of geographical features on language, physical boundaries are nowadays widely seen as less important. As only one factor impeding communication among many, their effect can at most be considered an indirect one. Auer (2005) has argued that “imagined” borders, as cognitive constructs, may sometimes explain lin2

In German, the term Raum, which is widely used in this context, translates into both area and space. For a critical discussion of the linguistic concept of Raum and its – sometimes problematic – implications see Oesterreicher 2007.

Area formation in morphosyntax

197

guistic areality better than “real” communicative barriers. In modernity (where real, physical or political, barriers are rare) in particular, the concept of communicative boundaries seems to be more convincing than the traditional one. We are still left, however, with the problem of how to explain the genesis of the imagined borders. A similar issue is discussed in Britain (2010), according to whom dialect boundaries are “often found in breaks between the socio-spatial networks” (2010: 151). A completely different approach to explaining the location of dialect boundaries is Moulton’s (1961) structuralistic attempt to define limits of variation. Labov (2006: 4) cites Moulton’s work as a rare example of work in dialect geography that is relevant to general linguistics. According to Moulton, the diffusion of certain dialect features in a given area can be explained by investigating pre-existing language systems. Thus, the split of back vowels in certain Swiss German dialects is explained as being conditioned by a certain type of front vowel system. Though Moulton’s attempt was much admired and often cited,3 it did not attract a lot of subsequent work. There are, however, more recent models of language change (cf. Haas 2010) that are broadly in line with Moulton’s idea that there are language-internal motivations for isoglosses. Schmidt (2010), for example, has presented a linguistic dynamics approach that incorporates Moulton’s view of dialect atlases as linguistic laboratories. The comparison of 19th century atlas data to recent investigations makes possible a real-time analysis of changes in the spatial limitation of linguistic features, and thus enables an assessment of the role of internal and external factors for the construction of dialect boundaries. This is compatible with Heap’s stance (Heap 2006: 615) that “a measure of preexisting linguistic similarity” between neighboring varieties should be included in geolinguistic models. Oesterreicher’s (2001) concept of “Übernahmebereitschaft” (‘borrowing readiness’) works along similar lines, and addresses the fact that many sociolinguistic models of feature diffusion are focused too narrowly, according to Oesterreicher, on external features. In historical linguistics, too, isoglosses matter. Consider a recent workshop that raised the question whether “related isoglosses spread with similar speed and/or steepness.4” Whatever “related” exactly means, this formulation insinuates that there is a deeper linguistic connection between the isoglosses 3 4

See the discussion in Gilles and Siebenhaar 2010: 772–773. International Conference of Historical Linguistics, convenors: Roeland van Hout (Nijmegen), Gertjan Postma (Meertens Instituut), Giuseppe Longobardi (Trieste).

198

Elvira Glaser

of different features. Proposals to include the factor “space” in quantitative models of language variation likewise attest to a broad consensus in the community that “space” is important. 2.3

Diffusion of innovations in various approaches

Obviously, a research question closely connected to the explanation of dialect boundaries is how to model the spread of linguistic innovations. Innovation diffusion has indeed been a major subject during the last 50 years of sociolinguistic scholarship. In early American and British variationism, the spatial aspect of diffusion was more or less ignored, while the social embedding of new variants at individual locations was in the focus of sociolinguistic interest. Later on, with the development of various models of diffusion such as wave diffusion, cascade diffusion, or cultural hearth diffusion5, space reentered the variationist sociolinguistics paradigm – in particular in the form of distance measures, along with gravity models, adopted from economic geography. The basic idea behind all these models is that linguistic innovations spread thanks to varietal contact, just as other behavioral innovations are transmitted thanks to social contact. The models differ in the conception of the hierarchies of the varieties in contact which the features follow during their diffusion. The cascade model, for example, attributes more importance to larger population centers. In the end, however, diffusion models suggest that we should be dealing with a more or less homogeneous distribution of the new feature within a larger area. Extant work on the diffusion of variation has several shortcomings. For one thing, while previous studies may explain the general process of innovation diffusion, they fail to shed light on the related problem of dialect boundaries as discussed above.6 And even the basic idea of innovation diffusion by contact faces a paradox of sorts: it is commonly assumed that neighboring speakers tend to accommodate. Accordingly, neighboring dialects are usually more similar than distant ones, and they should converge over time. Yet, as Chambers and Trudgill (1998: 93) state, patterns of “criss-crossing isoglosses separating even contiguous villages from one another” are characteristic of regions with “a long settlement history”. When the settlement history does not go far back in time, “dialect features tend to be shared over relatively great distances”. Obviously, there are two forces at work. One is 5 6

For a short description of the different models, see Britain 2010: 148–151. For a discussion of critical points based on Gregory 1985, see Britain 2009: 149–151.

Area formation in morphosyntax

199

the tendency towards accommodation in discourse. The other force derives from the fact that language is always subject to change; as a consequence, the propagation of the changed features creates new differences between neighboring speech communities. Dialectometry is one of the geolinguistic disciplines that investigate the link between geography and linguistic features. Recent research indicates that the neighborhood effect was perhaps overestimated; cf. Nerbonne and Heeringa’s statement “[W]e interprete the role of geographic proximity and increased social contact not as forces promoting linguistic similarity but rather as forces promoting linguistic differentiation” (2007: 293). Nonetheless, based on the investigation of a great number of dialect variants Nerbonne and Heeringa judge the influence of geography as massive (2007: 274), and they offer that “the strength we have shown geography to have as an explanatory variable makes it nearly inconceivable that social variables […] could ever be stronger” (291). Nerbonne (this volume) discusses various concepts of geography, some relating to different operationalizations of distance (physical, linear; or logarithmic distance; travel distance; and so on), and one relating to locations’ affiliations with dialect regions. On the basis of German dialect data Nerbonne shows that distance as well as dialect area membership can account for pronunciation differences to a certain degree. The two notions in conjunction account for about 47 % of the total variation. In any case, the results of the study support the concept of areal clustering rather than that of a dialect continuum. Note that Nerbonne conceives of dialect regions as a kind of social network, a concept that could explain their influence. The Augsburg-Ulm dialectometry project (Rumpf et al. 2010) has yet another focus: analyzing the distribution of individual linguistic variants in order to provide “insight into factors that determine these distributions” (2010: 74). Based on lexical data from the Sprachatlas von Bayerisch-Schwaben (1996–2009), the project seeks to test the hypothesis “that linguistic similarity entails similar spatial structures” (2010: 75). First results reveal a certain role of close semantic relationships in shaping similar spatial distributions (2010: 92). Future research must quantify the extent to which language-external or language-internal factors contribute to spatial structures; at present the Augsburg-Ulm dialectometry project refrains from giving far-reaching explanations of how the structures could have evolved. Schmidt’s (2010) dialectology approach relies on the concept of linguistic synchronization – a kind of calibration of the language competence of individuals through linguistic interaction. In his view, it is not only external factors that determine the ultimate propagation of linguistic features. Rather, there

200

Elvira Glaser

are also cases of clear interaction of internal factors limiting the spread of features. This is quite in line with Haas (2010), who sketches a model of dialect geography originally developed in the 1970s and updated since against the backdrop of the recent (sociolinguistic) models of feature diffusion. Taking the example of short vowel lowering and long vowel raising in Swiss German dialects, Haas presents a sophisticated diffusion model which distinguishes between extensive diffusion of a new feature (relative to speaker and space) and intensive diffusion (relative to the language system) (2010: 657) within a linguistic continuum. According to Haas, the different dialect areas reflect different stages in the diffusion process, which exhibits a certain “internal and implicative coherence” (2010: 653). The role of phonemic systems functioning “as supportive, impeding, or formative factors” (Haas 2010: 657) in the adoption of the innovations can be studied on dialect maps. Thus, Haas (2010) argues for language-internal constraints on the diffusion process. He points out that “the linguistic coherence of a region is more important than its boundary” (2010: 664). 2.4

Interim summary: Spatial distributions and related questions

In summarizing the different approaches to the explanation of spatial distributions, we note that there are three factors that are recurrent in the literature: (1) settlement histories, (2) internal (i.e. structural) factors, and (3) patterns of communication.7 Labov (2006: 9) defines three corresponding questions, dubbing them “the broader questions of causation that are specific to dialect geography” and subsequently considers the role of these factors concerning the expansion and location of dialect boundaries. On the basis of the Atlas of North American English (ANAE), Labov shows that there are correspondences in the location of different phonological isoglosses when they are structurally related (Labov 2006: 305). He also concedes, however, that the structural considerations “do not fully explain why the North/Midland boundary remains so firmly in place” (Labov 2006: 305), and calls for “further explorations into the communicative patterns and cultural geography of the area” (Labov 2006: 305). Whereas settlement history is no longer considered important in recent research, it is the exact role of communication networks that is in the focus of attention. A further research question, which is closely related to the issue of feature diffusion and language boundaries, concerns the grouping of local dialects 7

Siebenhaar 2010 mentions frequency as an additional factor but considers external factors the most important ones.

Area formation in morphosyntax

201

into dialect regions. Nevertheless, it is important to point out that the problems are not identical. The determination of dialect regions is typically connected to the quest for isogloss bundles. It is, however, often difficult to obtain clear cut boundaries with nicely coinciding isoglosses. This is why some researchers became convinced that the notion of dialect areas has to be given up in favor of the concept of dialect continua (vgl. Bloomfield 1984 [1933]: 341). When dealing with variation, it is also possible to define boundaries within a continuum quantitatively, as demonstrated by Horvath and Horvath (2001) on the basis of phonological data. With Haas I consider regions “the outcome of diffusion” (2010: 664). Therefore, the issue of dialect groupings is a problem that is subordinate to the quest for general principles underlying the geographic diffusion of linguistic features. Dialect regions, which are never homogeneous, are defined both by speakers and linguists on the basis of certain features, and it is of course interesting to study the extent to which the two concepts of regions, those defined by speakers themselves and those defined by linguists, overlap. However, while it is clear that dialect groupings constitute a major topic both in traditional dialectology and in modern dialectometry (see Heeringa and Nerbonne 2001; Kelle 2001), a more detailed discussion of these concepts is beyond the scope of the present contribution.8

3.

Dialect geography and morphosyntax: The research questions

We conclude that there are still plenty of open questions concerning the spread of dialectal features, the location of isoglosses, and the identification of dialect areas. In the past, both dialect-geographical and sociolinguistic-variationist research has been primarily concerned with phonology and – sometimes – the lexicon. Grammatical variation has rarely been investigated. With regard to German dialects some scholars have maintained that syntax does not show geographical variation at all. According to Löffler (2003), for example, syntactic variation across dialects is to be explained by principles of oral language production and does not differ from the syntax of the standard language (2003: 109, 116).9 Nonetheless, most dialectologists have come to agree that syntactic variants, too, can show a non-random areal distribution. 8

9

Questions of feature diffusion and its motivations are also discussed with respect to the related issue of dialect convergence (cf. Siebenhaar 2010), another concept that cannot be dealt with here in any depth. B. Szmrecsanyi pointed out to me that in English linguistics similar opinions are to be found, cf. Lass 2004: 374: “English regional phonology and lexis […] are generally more salient and defining than regional morphosyntax.”

202

Elvira Glaser

Although dialects do exhibit features of spoken language (such as repetition, ellipsis, or incoherence) it is also true that there are systematic differences in the syntax of various (spoken) dialects.10 We find robust evidence for this assumption in different language groups, but suffice it to mention here recent results concerning Dutch and German dialects, especially Swiss German (Barbiers et al. 2005, 2008; De Vogelaer and Devos 2008; Seiler 2004; Glaser 2003; Fleischer 2011). Even in nonstandard language, which is commonly considered to be less structured geographically, we do find variants with a clear difference in their spatial scope. This is true, for example, with respect to the use of the definite article with proper nouns in German.11 As far as Swiss German dialects are concerned, the different preferences in the serialization of auxiliary and past participle in present perfect formation were already noticed by the authors of the Swiss German dialect atlas (SDS). A pertinent map showing the ordering of the auxiliary bi ‘I am’ and the past participle gsi ‘been’ in a subordinate clause was consequently included in the atlas (Hotzenköcherle et al. 1975, 261; see Map 1). An area with a nearly exclusive use of the ordering of bi gsi is clearly discernable in the southwest of German-speaking Switzerland. In addition to word order in the verbal group, investigations carried out in Zurich have identified various morphosyntactic variants that show a spatial distribution in the Swiss German area. These cover various grammatical domains, such as – for example – clause linkage with infinitival purposive clauses as in (1), congruency in passive and resultative constructions as in (2), declension of proper names as in (3), prepositional dative marking as in (4), verb doubling as in (5), depictive marking as in (6), and converb marking as in (7).12

10

11

12

See the discussion in Krefeld 2008 concerning different sources of linguistic variation which should ideally all be taken in consideration, at least when a dialect location is taken as a starting point. Map 76 in Eichhoff 2000: 76 depicts the results of a survey investigating the use of a definite article along with the female proper name Ruth. Although one might have the impression that the use of the definite article with proper names is quite widespread, it is actually only in the southern part of the German-speaking area that the article is omnipresent. For maps and details concerning (i) purposive clauses see Seiler 2005, (ii) verb doubling: Frey and Glaser 2007, (iii) resultative and depictive marking as well as the declension of proper names: Bucheli Berger 2005, 2006, Bucheli Berger and Glaser 2004, (iii) differential object marking concerning dative case: Seiler 2003. For general information on the SADS project see http://www.ds.uzh.ch/dialektsyntax/ phaenomene.html.

Area formation in morphosyntax

203

Map 1. Word order variation in Swiss German (cf. Hotzenköcherle et al. 1975, map 261)

(1) für es billet z löse/zum es billet löse ‘in order to buy a ticket’ (2) d-strooss is immer no ufgrissni/ufgrisse ‘the street is still torn up’ (3) i ha (de) Fritz(en) gseh ‘I have seen Fritz’ (4) daas ghöört (i/a) miinere Schweschter ‘this belongs to my sister’

204

Elvira Glaser

(5) er laat de Schriiner (la) cho ‘he has the carpenter come’ (6) si trinkt d-Milch warm(e) ‘she drinks the milk hot’ (7) er isch hinketse heigloffe ‘he went home limping’ We notice quite different geographical patterns in the distribution of the above variants. Whereas resultative marking, as in (2), exhibits a south-north split well-known in Swiss German dialectology for phonology and the lexicon,13 there are also several small-sized areas such as those exhibiting depictive marking (Map 2) and converb marking (Map 3). The west-east division known from other linguistic levels is also to be found, e.g. with respect to the construction of the infinitival purposive clause and verb doubling. There are, moreover, phenomena, such as for instance prepositional dative marking, whose regional distribution does not seem familiar (cf. Seiler 2003). One may conclude from this that the distribution of morphosyntactic variants is quite inconspicuous in comparison with other linguistic levels. That geographically conditioned syntactic variation indisputably exists does not entail, however, that the distribution of syntactic variants is identical to the distribution of phonological or lexical variants. It seems, for example, quite common that a syntactic variant is restricted to a specific area, but nevertheless competes with another more widely used variant, as is the case with depictive marking in Swiss German (see above). As a tendency, this situation seems to be more common in syntax than in phonology – but as yet we do not know enough about the specifics of dialect grammars to conclusively judge the case. Be that as it may, the fundamental questions in dialect syntax are the classical ones well-known from the study of dialect phonology or lexis: What is the distribution of the variants, and how can it be described? Are there specific geographical patterns? Are we dealing with a continuum, or do we find clear-cut boundaries? How can we explain the distribution of the variants, and the location of the isoglosses? x

x x

13

The southern part is more or less the alpine region known as Höchstalemannisch (‘highest Alemannic’) in traditional dialectological terms.

Area formation in morphosyntax

Map 2. Depictive marking in north-east Swiss German (SADS data, multiple choice question II.13)

Map 3. Converb marking in central Swiss German (SADS data, multiple choice question II.23)

205

206

Elvira Glaser

An additional question that is specific to the study of dialect syntax is the following: Is there a basic difference between syntactically defined areas and areas defined by other linguistic levels? Indeed, area formation in syntax begs some fundamental questions concerning the nature of grammar. The status of syntactic variants differs across different theoretical frameworks: syntactic variation is either seen – for example, in formal syntax – as deeply rooted in grammar and linked to other (grammatical) phenomena, or else it is considered the outcome of more or less deliberate speakers’ choices, as in sociolinguistics. In the face of these differing theoretical views, the distribution of syntactic variants in space is certainly a major topic thanks to competing explanations as to how the distributions come about. And so it is worthwhile to take a closer empirical look at the spatial variation of (morpho-)syntactic phenomena. We note in this connection that in the past few decades, theoretical linguists have become interested in dialectal variation. In the Generative Grammar framework, dialect variants are considered to be indicative of microvariation, i.e. minimal differences between language systems, and as such they are of interest when investigating “Universal Grammar” (cf. Weiss and Brandner 2009). Minimal differences in grammar can help to uncover the boundaries of variation. As a consequence, it is now no longer exclusively the grammatical system of the individual speaker that is in the theoretical focus, but also geographically conditioned variation. The distribution of the features in space is not believed to be arbitrary, at least as long as it shows clear patterns. The authors of the manual of the Dutch Dialect Syntax Project put it as follows: “If two maps depicting distinct morpho-syntactic variables show a considerable overlap, one might want to see if these phenomena are correlated in a meaningful way.”14 It is the explanation of these correlations which is considered of theoretical interest, quite similar to Moulton’s phonological program or Haas’ model of sound change (see above). So far, attempts to model morphosyntactic diffusion in space have been quite rare. Barbiers (2009), after having discussed the empirical evidence in several cases of Dutch dialect variation, concludes “that the syntactic module allows for some optionality” (2009: 1621), but there are limits to syntactic variation. These limits are seen in general syntactic conditions. Barbiers claims, for example, “that the base structures of the language varieties under discussion do not vary” (Barbiers 2009: 1621). In the context of the underlying idea of a common gramx

14

See the manual of the Dutch Dialect Syntax Project: http://www.dialectsyntax.org/index.php/manual-mainmenu-67/introduction-mainmenu-70.

Area formation in morphosyntax

207

mar, the question is why some dialects make some choice while others make another one. These choices could be due to sociolinguistic circumstances, but it is of course also possible “that the grammatical properties determining such choices have not been discovered yet” (Barbiers 2009: 1622). This last point, according to Barbiers, is the reason why more data on dialectal syntactic variation are needed. Though I agree that we need more data, the notion of a common grammar underpinning dialects is somewhat unpalatable from a dialectological point of view. In sum, then, as far as I can see some mere first steps have been made to devise a theory of how geographical patterns of grammatical features come about. For example, exploring correlations between various linguistic levels with respect to distance Spruit et al. (2009: 1640) concede “that extralinguistic, but clearly non-geographic explanations are equally plausible as candidates to explain the correlation”. As mentioned in the previous section, in historical linguistics spatial distributions have always been considered an important source for the reconstruction of language change, as synchronic diversity can be interpreted diachronically. Yet, in this connection the status of syntax is unclear. Longobardi and Guardiano (2009) marshal a quantitative analysis using the parametric comparison method on the basis of 15 standard languages, and offer “that syntactic differentiation proceeds more slowly than lexical differentiation” (2009: 1694) – a conclusion which is not uncontroversial and stands in contradiction to classical comparative principles which for reconstruction hardly ever take syntax into consideration. Our literature review would seem to have suggested that at present there is no agreed framework for analyzing the geographical distribution of grammatical features and constructions.

4.

Dialect geography and morphosyntax: Some research pathways

The fact of the matter is, then, that there seems to be a renewed interest in the description and explanation of geographically conditioned linguistic variation, which for a long time has been of marginal interest in general linguistics. There are recent handbooks and new periodicals that deal with the topic15, a growing number of dialect atlases introducing new methods or subjects,16 and – in particular – there is the prospering field of dialectometry. 15

16

Suffice it to mention the recently published first volume of Language and Space (Auer and Schmidt 2010) and the periodical Dialectologia et Geolinguistica founded in 1993. Let us mention here Labov et al. 2006 based on telephone interviews in North America and Barbiers et al. (SAND 2005; 2008) for Dutch dialect syntax.

208

Elvira Glaser

Dialectometry essentially counts linguistic differences between locations in a dialect grid (cf. Goebl 2010 for an overview). The calculation of dialect differences can be visually depicted in colorful maps, usually based on aggregate measures of linguistic distance (or similarity). In this fashion more general spatial patterns can be identified, which can then be explained intra-linguistically or extra-linguistically. This classic form of dialectometry has quite rarely been used in German dialectology, and dialectometric work on aggregate morphosyntactic variability is non-existent. In any event, dialectometricians may also correlate linguistic distances with geographic distances (or other distance measures), which approximate the likelihood of contact and communication. Interestingly, previous research in this spirit has identified quite different correlation strengths. Exploring correlations between linguistic levels and geography, Spruit (2008) found, among other things, a correlation value of 0.55 between syntactic distances and geographic distances, based on Dutch dialect atlas data (Barbiers et al. 2005). This means that only about 30 % of the differences can be accounted for by geographic distances. On the basis of a reduced data set of Dutch dialects, however, Spruit et al. (2009) report a stronger correlation, such that 45 % of the variation is explained. They consider geography a valuable predictor of linguistic distances. On the other hand, drawing on English corpus data, Szmrecsanyi (2012) reports that less than 10 % of the syntactic variation can be explained by geographic distances, and so he concludes that geography plays a weaker role then one might think. A similarly weak role of geography one would also expect in the case of Swiss German syntax – Zurich-based research indicates that there are many more or less homogeneous syntactic areas, and few phenomena that are distributed in a continuum-like fashion. This is why a simple correlation with geographic distance would actually contradict the impression of a geographically unbalanced distribution of the variants. This impression is rather compatible with Szmrecsanyi’s finding that factoring in dialect groupings explains one third of the linguistic variability, which is more than mere geographical distance can account for. Indeed, geography would be “overrated” (Szmrecsanyi 2012) if one would seek to explain syntactic variation mainly by geographical distance. Szmrecsanyi (2012) also submits that atlas-based dialectometry may exaggerate the role of geography. Now, it is true that dialect atlases focus on features that seem geographically interesting and leave out variation found all over the place. This procedure follows from general cartographic principles. As a matter of fact, an unbalanced geographical distribution is the raison d’être for a map. An atlas is not intended to present all the relevant data of a local dialect. Mapping requires some degree of significant geographic distribu-

Area formation in morphosyntax

209

tion, and this does not exclude the existence of other types of linguistic variation. Certainly, a syntactic peculiarity of low frequency will find its way into an atlas, but it is probably not present in corpus data and does not play an important role in every-day communication. Nevertheless and all the more, its specific geographical range should be subject to dialectological analysis. Whether or not to ignore it in a big-picture analysis of variation depends on what we are interested in. A low-frequency variant can be of structural interest17, and it may be considered characteristic by locals and their neighbors. In any case, even if we concentrate on features with a significant geographic distribution, we cannot expect to find a simple correlation between geographic distance and syntactic distance as long as there is a certain amount of features clustering in an area. Let us put the problem the other way round. It is because geolinguistic variation is neither random nor a mere correlate of geographic distance that it is worth being investigated. The relevant diagram in Spruit (2008: 55) shows that neighboring (Dutch) dialects can be linguistically similar or not, whereas dialects far away from each other are in all likelihood less similar. Within this range (proximity vs. distance), we find various ways in which syntactic variants are geographically distributed. It is the study of these different geographic patterns that will advance our knowledge of the general principles underlying the geographic distribution of linguistic variation. Rumpf et al. (2010) define and analyze clusters of visually similar maps in order to find structural similarities concerning, for instance, semantic relationships. The atlas data they use does not cover morphosyntactic phenomena, yet the methodology presented is suitable for various research questions. Clusters of maps with similar spatial structures may be analyzed with respect to similarities of the grammatical features involved. Thus, research in dialect grammar requires not only a quantity-based comparison of the geographic patterns, but also a qualitative type of areal syntactic comparison. In this perspective, it is crucial to know not only about quantitative similarity but also about cooccurrence patterns, an issue that dialectometrical research has only recently begun to investigate (Spruit 2008: ch. 5). An ambitious objective like this brings about the problem of the classification and assessment of the syntactic variants as the basis for the calculation. Spruit (2008: ch. 5) presents a first sketch of how to manage the measuring of differences in word form and word order.

17

This holds true, for instance, for depictive marking in the North-East of Switzerland; see Bucheli Berger and Glaser 2004.

210

Elvira Glaser

Against this backdrop, work is underway in Zurich to explore new dialectsyntactic datasets, for the sake of addressing gaps in our knowledge about dialect syntax. Concerning Swiss German we can rely on data collected from 2000 through 2004 and stored in a database18 (cf. http://www.ds.uzh.ch/dialektsyntax/). Preliminary analyses have revealed many morphosyntactic variants which were mapped with the help of GIS software. The mapping technique we have used so far is the one customary in German dialect geography: symbol maps, which plot symbols for different variants at the points of investigation. Such interpretive maps enable the perception of areas where a certain variant occurs, thus giving a visual impression of the linguistic diversity in space. They are similar to maps explicitly presenting areas marked off by lines (isoglosses), but they avoid the lines’ often criticized arbitrary character. Visually depicting the distribution of variants in map space is not, however, a straightforward task. It is especially the co-occurrence of several variants at a measuring point – for example, in transition zones – which creates problems. Dialectometrically generated maps draw on color coding, which helps to display the varying pervasiveness of a variant. This sort of information is difficult to integrate in symbol maps. On the contrary, the existence of a specific repertoire of variants at a certain location or in a small area is not substantially more recognizable on a colored map (compare Map 4, which plots symbols, to the choropleth Map 5, based on kernel density estimation (KDE) of dominant variants, both on purposive clause linkage). As a consequence, if we are interested in the actual co-occurrence of variants at a certain location, blended color maps as in Map 5 are per se not helpful. The color maps showing the distribution of a single variant are, however, very useful in order to obtain an impression of an east-west continuum, respectively (see Map 6 and Map 7, taken from Sibler [2011, 29], also based on KDE of the dominant variants). The discussion of the advantages and disadvantages of choropleth maps show that we need to go beyond customary dialect cartography, with its exact reference to dialect variants and measuring points, to explore new quantitative cartographic methods which are typically based on the aggregation or blending of data. In this context, it is clear that the colored maps used in dialectometry are also subject to interpretation with respect to the spatial distributions mapped. They are particularly apt to provide an overall impression of the similarity of the dialects. But the adequacy of a projection to geography depends on its desired function. We note in particular that the quantitative basis of dialectometrical maps makes possible further numerical 18

For more information see http://www.ds.uzh.ch/dialektsyntax/.

Map 4. Purposive clause linkage in Swiss German (SADS data, translation task I.1, point symbol map)

Area formation in morphosyntax

211

212

Elvira Glaser

Map 5. Purposive clause linkage in Swiss German (SADS-data, translation task I.1; interpolation based on KDE of dominant variants, see Sibler (2011: 29)

Map 6. Purposive clause linkage in Swiss German (SADS data, translation task I.1; interpolation based on KDE, Sibler 2011: 29): zum-variant

Area formation in morphosyntax

213

Map 7. Purposive clause linkage in Swiss German (SADS data, translation task I.1; interpolation based on KDE, Sibler 2011: 29): für … z-variant

approaches to the similarities and dissimilarities underpinning the geographical patterns, as is shown by Rumpf et al. (2010) with respect to lexical maps. Promising research avenues include the simulation of diffusion processes (see Nerbonne 2010), as well as the implementation of topological and cultural information in the modeling. Dialectologists working in qualitative paradigms have recently started investigating correlations between various syntactic phenomena. Seiler (2005) shows that the spatial co-occurrence of the syntactic variants of infinitival purpose-clause formation in Swiss German dialects (basically e.g. für es billet z löse vs. zum es billet löse ‘in order to buy a ticket’) is not random, but rather grammatically organized. He argues for the notion of an inclined plane instead of the classical isogloss. The western variant für z reaches the eastern parts of the Swiss German area, but the further east it goes the more it is restricted in terms of grammatical context and preference.19 The same is true, conversely, for the eastern variant zum, which is attested in the most western parts only in a specific grammatical context. Analyzing the three-verb19

The questionnaire used in the SADS project asked for an evaluation of the accepted variants as to which one was the preferred one.

214

Elvira Glaser

cluster arrangement in Swiss German (e.g. i han wele gaa ‘I have wanted to go’), Seiler (2004) likewise concludes that transition areas are defined not simply by a random mix of two (or more) “consistent” neighboring grammars, but attest grammars of their own. The next step would consist, then, of systematically comparing the areal distribution of the variants and the (non)correspondence of the respective geographic areas in a comprehensive manner. As mentioned above, we already know that grammatical phenomena do not behave in a uniform manner. That means that there will be phenomena with quite clear boundaries and restricted regional occurrence, such as converbs (see above) and other phenomena – for example, the doubled indefinite article in expanded nominal attributes (ä ganz ä gueti arbet (‘(a) really a good work’), where variation covers a large area (Steiner 2006). The identification of specific geographic patterns begs the question if phenomena showing similar patterns share something in common. Dialectometrical analysis techniques may help us to spot similar patterns, while the analysis and interpretation of communalities is reserved for qualitative dialectological analysis. Another important research agenda is informed by the following question: “Is syntax different?”20 Or, in more general terms: “Is there a difference between grammatical areas and areas defined by other linguistic levels?” The crucial issues are co-occurrence patterns and the geographical scope of the variants. As mentioned above, in syntax we often meet the situation that a particular feature characterizes one area, but in a neighboring area we find variation between two features instead – and not another area defined by one typical variant. Isoglosses may thus mark off variation zones and not homogeneous areas. Another hypothesis posits that the range of syntactic features, in general, is larger than that of phonological ones. If this is true, the question arises how we can explain the differential effect of, say, accommodation in phonology and syntax. Discussing the different views on the role of geography, Szmrecsanyi (2012) refers to the assumption “that morphosyntax is less amenable to geographic diffusion than e.g. pronunciational variability.” There is indeed a controversy as to the character of syntactic transfer. Whereas some scholars consider syntax to be highly variable, others regard it as rather stable (cf. Spruit 2008: 52). In any case, we know very little about the differential behavior of the various linguistic levels (phonetics, morphology, lexicon and syntax) in geographic space, and specifically we do not currently know whether in fact syntactic constructions vary more freely than e.g. pronunciation features, as is commonly assumed. So, the comparison of syn20

This was the title of a talk on Swiss German dialects by Guido Seiler and Claudia Bucheli Berger at the Berkeley Germanic Linguistics Roundtable 2006.

Area formation in morphosyntax

215

tactic areas with areas defined by other linguistic levels will help to elucidate the nature of syntactic variation. Observe moreover that the comparison of historical with modern data may tell us how stable the areal distribution of syntactic variants is, also in comparison to other linguistic levels. The observation of ongoing changes or the discovery of a completed change in the geographical distribution of syntactic variants is informative about the stability of syntactic phenomena. At present we can, for instance, observe how the formation of a recipient passive (the so-called kriegen-passive) infiltrates more and more Swiss German dialects.21 In light of the assumption of the specificity of syntactic change, the fact that regional syntactic variants are often not recognized as such even by linguistically sensitive speakers (cf. Spruit 2008: 53) – a finding we have corroborated in our work on Swiss German syntax – is quite interesting. There are, however, important differences across grammatical phenomena. We need further research into possible connections between stability, the salience of morphosyntactic variation, and the exact nature of the grammatical variant (cf. word form, word order, congruency, and so on). Additionally, we need to know about the extent to which the range of variants and the existence of boundaries are dependent on the design of the inquiry. We need to explore, for example, how areas and boundaries would change if we only took into account one informant per measuring point in accordance with the traditional dialectological criteria. Moreover, we need to specifically study how the number of informants impacts the resulting variability at selected locations, and we should consider scaling effects as a result of basing the analysis on a larger or smaller number of measuring points. This is a topic that would profit from an interdisciplinary approach, as scaling effects are a well-known phenomenon in geographic information science.

5.

Dialect geography and areal typology: Common questions

The quest for the grammatical conditioning of syntactic variation within dialects is fully in line with the approach outlined in Haspelmath (2008), who calls for fine-grained analyses of hierarchies within restricted grammatical domains of various languages. This line of analysis assumes, however, the existence of linguistically motivated co-occurrence patterns, and of grammatical limits of the transferability of variants from one language system to another. This is not uncontroversial. To some degree, this assumption is incompatible with the notion that variants spread subject to sociolinguistic cir21

For more details see Glaser 2005.

216

Elvira Glaser

cumstances and stop at real-world or imagined borders. If we accept this notion we have to wonder about the extent to which the co-occurrence of variants is interesting at all to grammatical research. For according to this model co-occurrence patterns would have been created by sociological factors rather than by language-internal factors. Accordingly, typologists are faced with the same puzzle as dialectologists interested in the diffusion of syntactic features. The central dialectological issue of area formation through morphosyntax closely resembles the typological core issue as stated recently in Bickel (2007): “What’s where why?” Our concern is to determine if it is possible to base the answer to this question on linguistic facts, or not. The existence of syntactic areas where a certain variable is present prompts various questions with respect to the diffusion of syntactic constructions. The areal clustering of language structures is likely due to contact between speakers. Accordingly, area formation is linked to contact linguistics. Both in language contact and in dialect contact we observe transfer processes of syntactic variants. The transfer of grammar in contact situations has become a major subject in contact linguistics research in the last decennia (cf. Heine and Kuteva 2005; Matras and Sakel 2007). Nevertheless, there is no agreement concerning the linguistic or social conditioning of transfer and the limits of transferability. So, contact linguistics is confronted with the same questions as dialectology as regards the impact of social factors vis-à-vis language-internal constraints guiding the results of language contact. A comparative approach to contact linguistics combined with cross-linguistic typological work (see Siemund and Kintana 2008) seems to be a promising way to probe the mechanisms of linguistic convergence, which is a fundamental process in the evolution of a linguistic area (Sprachbund). As for areal linguistics, Bickel and Nichols (2006) attempt to redefine linguistic areas in order to solve the notorious problem of nonconformity and fuzzy boundaries, a problem that besets dialectology as well. Thus, cross-disciplinary reflection on questions of area formation, boundary formation, language contact (including dialect contact), feature diffusion and finally the association of language and geography is imperative. Holman et al. (2007) explicitly associate their concept of ‘isolation by distance’ among the world’s languages, which is inspired by population genetics,22 with the situation faced by dialectologists. Indeed, their intention to investigate “empirical correlations between linguistic differences and geographical distance” (395) re22

Holman et al. 2007: 394 mention spatial autocorrelation as an equivalent term used by ecologists to refer to distribution in geographical space. It is also used in Geographic Information Science; see Sibler 2011: 62, 64.

Area formation in morphosyntax

217

sembles very much the dialectological research question addressed in dialectometry. The common goal of dialect geography, dialectometry and areal typology is to better understand the mechanisms behind the geographical distribution of linguistic features. As it is often grammatical phenomena that are in the focus of areal typology, there is substantial potential for cross-fertilization between dialect syntax and areal typology.

6.

Concluding remarks

Although the quest for the motivations of the distribution of dialectal variants – in other words, the location of isoglosses – is an old one, we still know very little about these issues, at least as far as morphosyntax is concerned. In the end, we aim to come closer to solving the problem of explaining the distribution of variants and the location and nature of syntactic isoglosses. There is a very fundamental question concerning feature diffusion that is still unanswered: Is it really true that linguistic innovations diffuse thanks to (social) contact, just as other behavioral innovations do – or is there a difference? Interdisciplinary approaches are needed to address this question. Because typologists are likewise interested in finding correlations between linguistic features, the question of how the coexistence of features comes about is highly relevant in typology as well. Pure coincidence cannot be ruled out, but since the distribution of features across language areas is commonly attributed to language contact, there is a link to contact linguistics. In this context, we may wonder about the extent to which the diffusion of variants can be reduced to the effect of social contact, and about the extent to which there are structural limits. In sum, the question of area formation is, to this day, a central question for the understanding of language and language change. It is in any event not the case that the issue of geographically distributed linguistic phenomena is disappearing in modern times. The idea that dialect areas were stable in the past and are only now becoming fuzzy seems to be a figment of 19th century thinking. It is a natural development in feature diffusion that geographical distributions emerge in real time, as is shown by Labov (2006) concerning 20th century phonological change. Christen (2003), too, points to the areal character of expressive particles which have come up only recently in modern Swiss German youth language. It may be that the size of linguistic areas will change to become larger in the future – but geographical distributions as such are here to stay. As a consequence, they merit the linguist’s attention.

218

Elvira Glaser

References Auer, Peter 2005: The construction of linguistic borders and the linguistic construction of borders. In: Markku Filppula, Juhani Klemola, Marjatta Palander & Esa Penttila (eds.), Dialects across Borders, 1–30. Amsterdam/Philadelphia: Benjamins. Auer, Peter & Jürgen E. Schmidt (eds.) 2010: Language and Space. Berlin/New York: Mouton de Gruyter. Barbiers, Sjef 2009: Locus and limits of syntactic microvariation. Lingua 119: 1607–1623. Barbiers, Sjef, Hans Bennis, Gunther De Vogelaer, Magda Devos & Margreet van der Ham 2005: Syntactic Atlas of Dutch Dialects. Volume I: Pronouns, Agreement, and Dependencies. Amsterdam: Amsterdam University Press. Barbiers, Sjef, Hans Bennis, Gunther De Vogelaer, Johan van der Auwera & Margreet van der Ham 2008: Syntactic Atlas of Dutch Dialects. Part II: Auxiliaries, Verb Clusters and Negation. Amsterdam: Amsterdam University Press. Bickel, Balthasar 2007: Typology in the 21st century: Major current developments. Linguistic Typology 11: 239–251. Bickel, Balthasar & Johanna Nichols 2006: Oceania, the Pacific Rim and the theory of linguistic areas. In: Proceedings of the 32nd Annual Meeting of the Berkeley Linguistics Society. Available online at: http://www.uni-leipzig.de/~bickel/research/papers/bickel_nichols_bls32.pdf. Bloomfield, Leonard 1984 [1933]: Language. Chicago: University of Chicago Press. Brandner, Ellen & Helmut Weiss (eds.) 2009: Microvariation in the quantificational domain. Special Issue of Journal of Comparative Germanic Linguistics 12. Britain, David 2010: The variationist approach. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space, 142–163. Berlin/New York: Mouton de Gruyter. Bucheli Berger, Claudia 2005: Passiv im Schweizerdeutschen. In: Helen Christen (ed.), Dialekt/ologie an der Jahrtausendwende. Linguistik online 24 (3): 49–77. Bucheli Berger, Claudia 2006: Syntaktische Raumbilder im Höchstalemannischen. In: Hubert Klausmann (ed.), Raumstrukturen im Alemannischen. Beiträge der 15. Arbeitstagung zur alemannischen Dialektologie, 91–96. Schloss Hofen, Vorarlberg (Schriften der Vorarlberger Landesbibliothek, Band 15). Graz/Feldkirch. Bucheli Berger, Claudia and Elvira Glaser 2004: Zur Morphologie des (ko)prädikativen Adjektivs und Partizips II im Alemannischen und Bairischen. In: Franz Patocka & Peter Wiesinger (eds.), Morphologie und Syntax deutscher Dialekte und Historische Dialektologie des Deutschen. Beiträge zum 1. Kongress der Internationalen Gesellschaft für Dialektologie des Deutschen, Marburg/Lahn, 5.–8. März 2003, 189–226. Wien: Edition Praesens. Chambers, Jack K. & Peter Trudgill 1998: Dialectology (2nd edition). Cambridge: Cambridge University Press. Christen, Helen 2003: Uu fein, welts guet und rüüdig schöön. Überlegungen zu lexikalischen Aspekten eines SchweizerDeutsch der Regionen. In: Beat Dittli, Annelies Häcki Buhofer & Walter Haas (eds.), Gömmer MiGro? Veränderungen und Entwicklungen im heutigen SchweizerDeutschen, 25–38. Freiburg/Schweiz: Universitätsverlag Freiburg. De Vogelaer, Gunther & Magda Devos 2008: On geographical adequacy, or: how many types of subject doubling in Dutch. In: Sjef Barbiers, Margreet van der Ham,

Area formation in morphosyntax

219

Olaf Koeneman & Marika Lekakou (eds.), Microvariations in Syntactic Doubling, 249–274. Bingley: Emerald. Eichhoff, Jürgen 2000: Wortatlas der deutschen Umgangssprachen, Volume 4. München/ Bern: K.G.Saur. Fleischer, Jürg 2011: … und habe es ihr gesagt: Zur dialektalen Abfolge pronominaler Objekte (eine Auswertung von Wenkersatz 9). In: Elvira Glaser, Jürgen E. Schmidt & Natascha Frey (eds.), Dynamik des Dialekts – Wandel und Variation. Akten des 3. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD), 77–100. [Beiheft der Zeitschrift für Dialektologie und Linguistik 144] Stuttgart: Steiner. Frey, Natascha & Elvira Glaser 2007: Doubling phenomena in Swiss German dialects. In: Sjef Barbiers, Margreet van der Ham, Olaf Koeneman & Marika Lekakou (eds.), European Dialect Syntax Project. Papers presented at the Workshop on Syntactic Doubling, Amsterdam, March 2006. Available online at http://www.meertens.knaw.nl/projecten/edisyn/Online_proceedings/Paper_Glaser-Frey.pdf. Gilles, Peter & Beat Siebenhaar 2010: Areal variation in segmental phonetics and phonology. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space, 760–786. Berlin/New York: Mouton de Gruyter. Glaser, Elvira 2003: Schweizerdeutsche Syntax: Phänomene und Entwicklungen. In: Beat Dittli, Annelies Häcki Buhofer & Walter Haas (eds.), Gömmer MiGro? Veränderungen und Entwicklungen im heutigen SchweizerDeutschen, 39–66. Freiburg/Schweiz: Universitätsverlag Freiburg. Glaser, Elvira 2005: Krieg und kriegen: zur Arealität der BEKOMMEN-Periphrasen. In: Annelies Häcki Buhofer, Ulla Kleinberger Günther & Elisabeth Piirainen (eds.), Krieg und Frieden. Auseinandersetzung und Versöhnung in Diskursen, 43–64. Tübingen: Francke Verlag. Goebl, Hans 2010: Dialectometry and quantitative mapping. In: Alfred Lameli, Roland Kehrein & Stefan Rabanus (eds.): Language and Space. Vol. 2: Language Mapping, 433–457. Berlin/New York: Mouton de Gruyter. Haas, Walte 2010: A study on areal diffusion. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space, 649–667. Berlin/New York: Mouton de Gruyter. Haspelmath, Martin 2008: Parametric versus functional explanations of syntactic universals. In: Theresa Biberauer & Anders Holmberg (eds.), The Limits of Syntactic Variation, 75–107. Amsterdam: Benjamins. Heap, David 2006: Spatial variation (geolinguistics). In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, 2nd edition, 612–616. Oxford: Elsevier. Heeringa, Wilbert & John Nerbonne 2001: Dialect areas and dialect continua. Language Variation and Change 13: 375–400. Heine, Bernd & Tania Kuteva 2005: Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Holman, Eric W., Christian Schulze, Dietrich Stauffer & Søren Wichman 2007: On the relation between structural diversity and geographical distance among languages: Observations and computer simulations. Linguistic Typology 11: 393–421. Horvath, Barbara M. & Ronald J. Horvath 2001: A multilocality study of a sound change in progress: The case of /l/ vocalization in New Zealand and Australian English. Language Variation and Change 13: 37–57. Hotzenköcherle, Rudolf, Robert Schläpfer, Rudolf Trüb & Paul Zinsli (eds.) 1975: Sprachatlas der Deutschen Schweiz. Vol. III: Formengeographie. Tübingen: Francke.

220

Elvira Glaser

Kelle, Bernhard 2001: Zur Typologie der Dialekte in der deutschsprachigen Schweiz: Ein dialektometrischer Versuch. Dialectologia et Geolinguistica 9: 9–34. Knobloch, Clemens 2010: The kulturmorphologische Ansatz in dialectology and the German language space ideology. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space, 107–125. Berlin/New York: Mouton de Gruyter. König, Werner (ed.) 1996–2009: Sprachatlas von Bayerisch Schwaben. Heidelberg: Winter. Kortmann, Bernd (ed.) 2004: Dialectology Meets Typology. Dialect Grammar from a CrossLinguistic Perspective. Berlin/New York: Mouton de Gruyter. Krefeld, Thomas 2008: Räumlich fundierte Varietätenlinguistik – eine Option und ihre Geschichte. In: Gerald Bernhard and Heidi Siller-Runggaldier (eds.), Sprache im Raum – Raum in der Sprache, 91–105. (Akten der sprachwissenschaftlichen Sektion des Deutschen Italianistentages in Bochum, 23.–25. März 2006. Frankfurt et al.) Labov, William, Sharon Ash & Charles Boberg 2006: The Atlas of North American English. Berlin/New York: Mouton de Gruyter. Lass, Roger 2004: South African English. In: Raymond Hickey (ed.), Legacies of Colonial English: Studies in Transported Dialects, 363–386. Cambridge: Cambridge University Press. Löffler, Heinrich 2003: Dialektologie. Eine Einführung. Tübingen: Narr. Longobardi, Giuseppe & Cristina Guardiano 2009: Evidence for syntax as a signal of historical relatedness. Lingua 119: 1679–1706. Matras, Yaron & Jeanette Sakel (eds.) 2007: Grammatical Borrowing in Cross-linguistic Perspective. Berlin/New York: Mouton de Gruyter. Moulton, William G. 1961: Lautwandel durch innere Kausalität: Die ostschweizerische Vokalspaltung. Zeitschrift für Mundartforschung 28: 227–251. Nerbonne, John 2010: Measuring the Diffusion of Linguistic Change. Philosophical Transactions of the Royal Society B: Biological Sciences 365: 3821–3828. Nerbonne, John & Wilbert Heeringa 2007: Geographic Distributions of Linguistic Variation Reflect Dynamics of Differentiation. In: Sam Featherston & Wolfgang Sternefeld (eds.), Roots: Linguistics in Search of its Evidential Base, 267–297. Berlin/ New York: Mouton de Gruyter. Oesterreicher, Wulf 2001: Historizität – Sprachvariation, Sprachverschiedenheit, Sprachwandel. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher & Wolfgang Raible (eds.), Language Typology and Language Universals. An International Handbook, HSK 20.2, 1554–1595. Berlin/New York: Mouton de Gruyter. Oesterreicher, Wulf 2007: Raumkonzepte in der Sprachwissenschaft: Abstraktionen – Metonymien – Metaphern. Romanistisches Jahrbuch 58: 51–91. Rumpf, Jonas, Simon Pickl, Stephan Elspaß, Werner König & Volker Schmidt 2010: Quantification and statistical analysis of structural similarities in dialectological area-class maps. Dialectologia et Geolinguistica 18: 73–100. Schmidt, Jürgen E. 2010: The linguistic dynamics approach. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space, 201–225. Berlin/New York: Mouton de Gruyter. Seiler, Guido 2003: Präpositionale Dativmarkierung im Oberdeutschen. Stuttgart: Steiner. (Zeitschrift für Dialektologie und Linguistik, Beiheft 124). Seiler, Guido 2004: On three types of dialect variation, and their implications for linguistic theory. Evidence from verb clusters in Swiss German dialects. In: Bernd Kortmann (ed.), Dialectology Meets Typology. Dialect Grammar from a Cross-Linguistic Perspective, 367–399. Berlin/New York: Mouton de Gruyter.

Area formation in morphosyntax

221

Seiler, Guido 2005: Wie verlaufen syntaktische Isoglossen, und welche Konsequenzen sind daraus zu ziehen? In: Eckard Eggers, Jürgen Erich & Dieter Stellmacher (eds.), Moderne Dialekte – neue Dialektologie. Akten des 1. Kongresses der Internationalen Gesellschaft für Dialektologie des Deutschen (IGDD), 313–341. Stuttgart: Steiner. Sibler, Pius 2011: Visualisierung und geostatistische Analyse mit Daten des Syntaktischen Atlas der Deutschen Schweiz (SADS). Masterarbeit, Geographisches Institut der Universität Zürich. http://www.geo.uzh.ch/de/lehrstuehle-und-abteilungen/geographische-informationssysteme/forschung/msc-theses. Siebenhaar, Beat 2010: Horizontal convergence of linguistic varieties in a language space. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space, 241–258. Berlin/New York: Mouton de Gruyter. Siemund, Peter & Noemie Kintana (eds.) 2008: Language Contact and Contact Languages. Amsterdam: Benjamins. Spruit, Marco René 2008: Quantitative perspectives on syntactic variation in Dutch dialects. Ph.D. thesis, University of Amsterdam, LOT Dissertation Series 174. Utrecht: LOT. Spruit, Marco René, Wilbert Heeringa & John Nerbonne 2009: Associations among linguistic levels. Lingua 119: 1624–1642. Steiner, Janine 2006: Syntaktische Variation in der Nominalphrase – ein Fall für die Dialektgeographin oder den Soziolinguisten. In: Hubert Klausmann (ed.), Raumstrukturen im Alemannischen. Beiträge der 15. Arbeitstagung zur alemannischen Dialektologie, 109–115. Schloss Hofen, Vorarlberg. 19.–21. 9. 2005. Graz/Feldkirch. (Schriften der VLB, Band 15). Szmrecsanyi, Benedikt 2012: Geography is overrated. In: Sandra Hansen, Christian Schwarz, Philipp Stoeckle & Tobias Streck (eds.), Dialectological and Folk Dialectological Concepts of Space, 215–231. Berlin: de Gruyter.

222

John Nerbonne

John Nerbonne

How much does geography influence language variation?

1.

Introduction

Two insights about language variation are standard and uncontroversial, first that languages may vary in many ways and second, that nearby language varieties are generally – but not always – more similar than distant ones. Dialectometry supports the measurement of linguistic similarity and its inverse, linguistic difference or distance, in various forms and therefore satisfies a prerequisite for determining the degree to which geography influences language variation, a line of investigation that is effectively closed to research traditions that shun measurement in favor of cataloguing differences. The present paper shows how dialectometry arrives at measurements and proceeding from these, how the influence of geography on language variation may be ascertained through regression designs. In addition to this, two notions of geography are contrasted, one in which the linguistic differences between varieties is compared to the geographic distance between them (including derivatives of distance such as travel time), and another, in which sites are partitioned into areas. We can measure the contribution of areas in the same sort of regression design used to gauge the influence of geographical distance, and, finally we may examine a combination of the influence of distance and areas. We close with some reflections on the ways in which we have conceptualized geography and some speculation on the effect of modern communication technology, asking whether space as we have conceptualized it here is likely to continue to be as influential in the future as it has been in the past.

2.

Quantitative dialectology

Dialectometry, or quantitative dialectology, has arisen to solve several problems that existed in traditional dialectology (Goebl 1986) – in particular, the concrete problem of “non-overlapping isoglosses”, or more generally, the problem of identifying geographical structures underlying the distribution of linguistic features. The crux of the dialectometric solution has been to aggregate over many linguistic features before seeking geographic interpretation.

How much does geography influence language variation?

223

Goebl (1986) speaks suggestively here of the dialectometer “condensing” (verdichten) linguistic atlas data. In addition to opening new avenues in which to seek dialect areas (or dialect continua), dialectometry likewise introduces replicable procedures into the study of dialects and provides a basis on which the research may seek more abstract regularities or “laws” (Nerbonne 2009). The present essay is concerned primarily with the latter task, that of formulating more general principles in language variation. Some simple ways of measuring language similarities are remarkably effective. Séguy (1971) simply examined the lexical realization of a set of concepts and counted how often there was agreement and how often disagreement. In later work he extended this technique by examining linguistic differences not only in word choices, but also in pronunciation, morphology and syntax. A slightly more complex analysis is useful to gauge differences in the pronunciation of words as these are recorded in dialect atlases, i.e. in the form of phonetic transcriptions. Levenshtein distance counts both the number of substitutions needed to transform one transcription into another, as well as the numbers of insertions and deletions, always seeking the minimum number necessary (Nerbonne and Heeringa 2010). We illustrate the effect of the procedure on two Dutch pronunciations of melk (the word for ‘milk’), namely Frisian [mɔəlkə] (Grouw) and standard Dutch [mεlək] (Delft). On the one hand one may examine the operations needed to map one pronunciation to another: mɔəlkə mɔlkə mεlkə mεlk mεlək

delete ə replace ɔ by ε delete ə insert ə

Or, on the other hand, we may equivalently inspect the alignment induced by this process: k ə m ɔ ə l m ε l ə k 1 1 1 1 There are several advantages to using Levenshtein distance to measure the distance between transcriptions as opposed to collecting categorical features from dialect data collections (Nerbonne et al. 2010: 42–43). First, using Le-

224

John Nerbonne

venshtein distance automates a larger part of the process of analyzing dialect atlas data, obviating in particular the need for the researcher to extract categorical differences by hand. Second, we increase the amount of data compared by nearly an order of magnitude, as we effectively compare each word at each segment position. A list of one hundred words thus typically yields about 500 segment comparisons. Third, and related to the second, using all of the segments of transcribed words means that fewer words suffice for reliable assessments of the dialect distances between sites. Typically, thirty to forty transcribed words are sufficient. Fourth, and related to the last two, the use of Levenshtein distance involves comparing all the segments in the data collected, and therefore a large number of segments that happen to be in words that were chosen to illustrate dialect differences. This means the data that is analyzed was effectively collected in a manner closer to the sort of random data selection that is customary in corpus linguistics and generally recommended in statistical procedures. About 80 % of the segments subject to analysis were collected only because they happen to occur in a word with an ‘interesting’ segment. So 20 % was selected intentionally, and 80 % unintentionally. In all of the data sets we examine below, we measure the pronunciation difference of every pair of corresponding words (phonetic transcriptions), typically about one hundred words, in each pair of sites, often thousands (of pairs of sites), in the collection using Levenshtein distance. We take the site difference then to be the average pronunciation distance per word pair, ignoring in this way the occasional cases in which data is missing. We then follow Heeringa and Nerbonne (2001) in applying a simple regression design to the aggregate distances thus assayed, using the geographic distance between sites as the independent variable and the aggregate pronunciation distances as the dependent variable. Given that the present contribution is intended for a volume on language and space, it is worth pausing to reflect explicitly on why as simple a spatial or geographic concept as distance plays so central a role in the theorizing. To begin, let us forswear any interest in the physical properties of space, which we imagine having no influence on speech patterns or perceptions (not even in virtue of the transmission of the sound waves). But as Bloomfield (1933, Ch. 3.4 et passim) notes, speakers adjust their speech regularly for their speech partners, leading him to hypothesize that linguistic habits would likely follow the lines of the densest communication. Space or geography, as we theorize about it here, is therefore a social concept interesting for the indications it suggests about where people are communicating with one another. We suppose that people are also less likely to communicate if they live further away from each other than if they live closer to each other, that is, in the usual

How much does geography influence language variation?

225

case. This is the basis for examining the relation between simple distance and linguistic similarity. We do not suppose at all that distance is the only spatial or geographic influence on the similarity of speech habits (see further), but we examine it first, as it is very basic to the influence geography has on speakers. The above, Bloomfieldian (1933) view is adopted here for its simplicity. It may seem to imply that variation inevitably diffuses only in order to facilitate communication with others, but I do not attribute this view to Bloomfield. As we know, in fact, variation also arises as a means for a speaker to differentiate himself or herself from others, and this sort of variation also diffuses. We shall not distinguish the two different sorts of changes in speech habits, those motivated essentially by accommodation and those motivated by a wish to assert differentiation. This perspective on linguistic variation will be ignored in the rest of this essay as I do not see how it would figure in a spatial or geographic model. We pursue a second step only in the case of one data set, the German set (see below). In this second step we complicate the simple regression design by including several more independent variables, in addition to the simple distances just described, namely a series of “dummy” binary variables which take the value one (1) whenever two sites under comparison originate from two specific, different, a priori-defined dialect areas (e.g. the Bavarian area in Germany and the Alemannic German Southwest), and zero (0) otherwise. This procedure follows a suggestion by Shackleton (2007, 2010) and is motivated by the wish to investigate whether dialect areas independently contribute to pronunciation distance – over and above the sheer distance between the sites that would be accounted for in the simpler design.

3.

Data and analysis

In this section we describe the data on which our analyses are based, reporting at the same time on the regression analyses seeking to explain aggregate pronunciation distance based on geographic distance. We turn to the potentially different contribution that might be made by areal distinctions in a following section. We note here that our use of aggregate pronunciation difference (or equivalently, average differences) naturally tends to inflate the correlation coefficients reported below. If we examined the individual word pronunciations, the correlation coefficients would drop substantially. The focus on the aggregate difference has long been standard in quantitative dialectology (see the remarks on “condensing” data above), but the deeper justification

226

John Nerbonne

for using aggregate measures is that we are interested in the properties of linguistic varieties, i.e. the overall speech habits in a community, and not merely the properties of individual words. 3.1

Data and distance analysis

We compare six different dialect data collections, using the same material presented in Nerbonne (2010). That paper focused on the sub-linear (normally logarithmic) growth of average phonetic distance as a function of geographic distance, while the present paper focuses on the degree to which variation is explained via geography, including geographical distance but also areal partition. Because we have published focused papers on each of the data sets we use, the descriptions below are somewhat summary; interested readers are referred to the original papers. Alewijnse, Nerbonne, van der Veen & Manni (2007) used pronunciation data from Bantu data collected in Gabon by researchers from the Dynamique du Langage project in Lyon.1 Since the Gabon Bantu population consisted of migratory farmers until recently, it might be expected to show a different influence of geography on linguistic variation. The data involve broad phonetic transcriptions of 160 concepts taken from 53 sampling sites. Tone was not analyzed as the Bantu experts were skeptical about how reliably it had been recorded and transcribed. The geographic locations recorded were those provided by native speaker respondents, but they should be regarded in some cases as “best guesses” considering how mobile the population has been (over long periods of time). The pronunciation differences were analyzed using the procedure sketched in Section 2 above, and these correlate moderately with logarithmic geographic distances (r = 0.469). Houtzagers, Nerbonne & Proki´c (2010) obtained data on Bulgarian dialects from Prof. Vladimir Zhobov’s group at the St. Clement of Ohrid’s University of Sofia. The research analyzed broad phonetic transcriptions of 156 words from 197 sampling sites in Bulgaria. Palatalized consonants, which are phonemic in Bulgarian, are represented in the data, but stress is not. The pronunciation difference measurement described above was applied, where alignments were constrained to respect syllabicity, meaning that vowels were only allowed to align with vowels, and consonants only with consonants. Because of the long Ottoman occupation of Turkey (until 1872), its patterns of variation may be atypical, but the correlation of pronunciation and logarithmic geographic distance was measured at r = 0.488. 1

http://www.ddl.ish-lyon.cnrs.fr/.

How much does geography influence language variation?

227

Nerbonne & Siedle (2005) obtained data from the Deutscher Sprachatlas in Marburg.2 The pronunciations of 186 words had been collected at 201 sampling sites for the project Kleiner Deutscher Lautatlas. A team of phoneticians transcribed the data narrowly; each word was transcribed twice independently and disagreements were settled in consultation so that there was consensus about the results. The pronunciation differences were measured using Levenshtein distance, where alignments were constrained as above to respect syllabicity. Logarithmic geographic distance correlates strongly with pronunciation in this data set (r = 0.566). Kretzschmar (1994) reports on the LAMSAS (Linguistic Atlas of the Middle and South Atlantic States) project, conceived and carried out mainly by Hans Kurath, Guy Lowman and Raven McDavid in the 1930s and again in the 1950s and 1960s.3 Due to differences in fieldworker/transcriber practices, we analyze only the 826 interviews which Guy Lowman conducted in the 1930s involving 151 different response items. LAMSAS used its own transcription system, which we converted automatically to X-SAMPA for the purpose of this analysis. This analysis we conducted using a variant of the measurements above. Nerbonne (to appear) describes some aspects of the analysis in more detail, in particular the degree to which phonological structure is present. Since the area of the present U.S. has only been English speaking for the last several centuries, it may retain traces of migration disturbance in the geographic distribution of linguistic variation. We nonetheless measured a strong correlation between pronunciation and geographic distance after applying a logarithmic correction to the latter (r = 0.511). Wieling, Heeringa & Nerbonne (2007) analyze the data of the projects Morphologische Atlas van Nederlandse Dialecten (MAND) and Fonologische Atlas van Nederlandse Dialecten (FAND) (Goeman and Taeldeman 1996). In order to avoid a potential confound due to transcription differences, Wieling et al. analyze only the data from the Netherlands, and not that of Flanders. The former included 562 linguistic items from 424 varieties. Since the Netherlands comprises only 40 000 km2, the MAND/FAND is one of the densest dialect samplings ever. The pronunciation differences were measured using the technique described above, where alignments were constrained to respected syllabicity. Pronunciation distance correlates strongly with the logarithm of geographic distance (r = 0.622). Gooskens & Heeringa (2004) analyze the variation in 15 Norwegian versions of the fable of the International Phonetic Association, “The North 2 3

http://www.uni-marburg.de/fb09/dsa/. The data is publicly available at http://hyde.park.uga.edu/lamsas/.

228

John Nerbonne

Wind and the Sun”.4 The material was again analyzed using the pronunciation difference measurements explained above. Britain (2002) urges dialectologists to examine critically the underlying conceptions of geography they employ. Interestingly from this point of view, Gooskens (2004) compares two geographic explanations of the linguistic differences, one based on “asthe-crow-flies” distances, and another based on the (logarithmic) travel time estimates of the late nineteenth century, showing an improvement in correlation (from r = 0.41 to r = 0.54). The motivation for examining the two operationalizations was naturally that travel time is expected to be the better reflection of the chance of social contact, i.e. Bloomfield’s (1933) “density of communication”. We conclude from this section that there is a simple and measurable influence which geographic distance exerts on aggregate linguistic differences. Figure 1 summarizes the six data sets discussed above and sketches the regression line in each case. It is an empirical finding, not a theoretical prediction, that geography accounts for 16 % to about 37 % of the linguistic variation in these data sets (100 × r2). We note that the potential disturbances caused by migration, occupation, and recent settlement appear insubstantial enough in the cases examined so as not to disturb the overall tendency. For readers of this volume who may not be as familiar with regression analyses as they would like, it is worth emphasizing some relevant, wellknown facts about regression analyses in order to fend off misunderstandings. First, what counts as a “strong correlation” varies from one field of investigation to another, but many research fields use r > 0.5 as a cutoff. Second, the square of the correlation coefficient is the fraction of variance (variation) accounted for by the independent variables in the regression analysis. Given that language variation is highly complex, and that we have examined a substantial amount of it (and not merely some selected variables), one cannot expect extremely high levels of correlation (or of the derived percentage of explained variance). Third, obtaining a correlation coefficient of r = 0.6 (and thereby accounting for 36 % of the variation in a data set) means that a great deal of variation is still unaccounted for, and so there is lots of work to be done. Fourth, we want to keep in mind that saying, e.g., that distance accounts for 36 % of the variation in a data set does not mean that another, alternative explanation could not account for more than 64 %. This is eminently possible, in particular, where distance might correlate with an alternative explanation to some degree, for example if there were a way to measure the ease of communication between settlements, or the strength of 4

They are making use of material from http://www.ling.hf.ntnu.no/nos/

How much does geography influence language variation?

229

Fig. 1. Pronunciation differences in Bantu, Bulgaria, Germany, the US Eastern Seaboard, Netherlands and Norway as functions of geographic difference. In each case a logarithmic regression line is drawn. Please note that the vertical y-axes have not been calibrated.

230

John Nerbonne

the ties that bind the communities. Fifth, we advisedly use the mean pronunciation distance between varieties as our dependent variable, and not, e.g. the pronunciation distance between the pronunciations of a particular word for the simple reason that we are interested in the entire variety spoken at a given site. As we noted above, correlations involving a mean are inevitably inflated, as they remove a significant amount of noise from the data. Statisticians have introduced the term “ecological fallacy” for steps in reasoning that infer properties of individuals (words) from the properties of aggregates (sites) (Agresti & Franklin 2009). This study is not guilty of the ecological fallacy because we focus on the distances between entire varieties, and not on the distances of individual words. 3.2

Areal analysis

There are several reasons to look beyond distance as explicans in variationist linguistics. Reflecting on its introduction above, we recall that we were inspired by Bloomfield’s (1933) notion of “density of communication”. But if the density of communication influences language variation, then there is good reason to think that we might gauge this influence not only via distance but perhaps altso through other phenomena that might provide some purchase on communicative density. One such phenomenon is dialect area, the basis of undoubtedly the most popular visualization of geographic influence on language variation – that of dialect maps showing a partition of a language area into non-overlapping dialect areas. In Map 1 we present Wrede’s division of Germany into six major areas.5 Given this map it is natural to ask whether language variation is less a matter of continuous variation along the dimension of distance between sites, and rather more a matter of sites participating in a partition of relatively homogeneous areas. The quantitative perspective allows us to compare the influence of distance (see above) with that of area.

4.

Analysis

We focus in this section on the German data, examining the effect of geographic distance through a regression design in which both bare distance but also areal differences are used as independent variables. Naturally they are both examined for their effect on the dependent variable, linguistic distance. 5

Niebaum & Macha 2006: 38 explain that Wrede’s map was published posthumously in 1937 by Bernhard Martin as Deutscher Sprachatlas, 9th Lieferung, Karte 56.

How much does geography influence language variation?

231

Map 1. Wrede’s (1937) atlas of German dialects has been taken here in order to use an authoritative source as a hypothesis. We shall refer to the areas (clockwise, from top left) as the northwest, the northeast, Saxony, Bavaria, the southwest, and the Palatinate. The dark lines separate Wrede’s areas while the divisions in coloring depend on the automatic “tiling” done around data collection sites. There is a slight mismatch when sites practically lie on borders.

In order to evaluate the influence of area quantitatively, we shall introduce variables which indicate when two sites are in two specific, different areas. For example, we shall introduce a variable “Southwest-vs-Bavarian” which has the value “1” if, and only if, one site in question is from the southwest in

232

John Nerbonne

Wrede’s map and the other from Bavaria. It has the value zero in every other case, i.e. when both sites are in the southwest, when they are both in Bavaria, when one site is in the southwest or Bavaria and the other is somewhere else, and also when both sites are outside of both areas. Note that there are six areas, meaning that we shall need to introduce fifteen such variables, one for each pair of areas. Shackleton (2007: 61ff) was the first to suggest analyzing the effects of dialect areas in this way, using it to show the significance of dialect areas he obtained by clustering distance measures. Note that cluster differences are derived from linguistic distance measures, meaning that they clearly may not be regarded as statistically independent. Without criticizing Shackleton’s work, let us note that some circularity may be lurking in the regression analysis that uses cluster differences derived from linguistic differences to explain the same set of linguistic differences. Let us therefore examine here the areal differences shown in Wrede’s (1937) independently derived map. The analysis here differs only in that we have chosen not to use a derivative of the linguistic distance measure as its own explicans. Instead, we examine the effect of an independently proposed set of areal differences, Wrede’s.6 Ignoring that difference, our regression analysis follows Shackleton’s in technical design. The fifteen areal differences variables are added to the multiple regression design. The comparison of bare distance versus areas as explanatory variables may be regarded as the quantitative form of the old question of whether dialects should be regarded as organized spatially as continua or as areas, i.e., discrete partitions of relatively uniform sets of sites. If the continuum view is correct, areal differences should explain nothing in models in which bare distance is included. But if on the other hand the areal view is sufficient, then the areal variables should turn out to be significant, and bare distance should play no explanatory role. Naturally there is a “third” view which acknowledges that gradual variation is consistent with both constant rates of additional variation as well as rates that may change rather abruptly. If linguistic variation is constantly increased, we obtain a perfect continuum. If, on the other hand, the rate of variation changes abruptly, we obtain a situation in which both the continuum and the areal views have some validity.7 6 7

But see note 7. One comment on our use of Wrede’s 1937 areas has been the charge that the circularity we note in the previous paragraph is still present, albeit indirectly, since Wrede derived his areas from an analysis of linguistic differences – even if he used a different set of data (from a century earlier) and different, less formalized methods. In the eyes of this criticism our construction has merely replaced our data and analysis with Wrede’s, but in both cases one is trying to explain linguistic dif-

How much does geography influence language variation?

233

As we introduce such variables indicating areal difference, we nonetheless need to exercise some caution. The more areas we distinguish (and therefore the more variables we introduce), the better the chance of seeing a statistical significant effect for at least some areal differences. In the most extreme (and uninteresting) case, in which a variable were introduced for each pair of sites, the linguistic distance could be predicted perfectly. But as long as we examine only a relatively small number of areas, the question of whether bare distance or areal distinctions are the better predictors is genuine. As long as the number of areas distinguished is low, we expect to see that bare distance remains significant – at least within the large areas, which, after all, are not completely uniform.

5.

Results

Linear distance, corrected logarithmically (see above), correlates moderately with linguistic distance (r = 0.54), thus accounting for a bit more than 29 % of the variation in the data. This is slightly lower than the figure in Nerbonne & Siedle (2005), due to slight differences in measurement. 5.1

(Binary) areal differences

We turn then to the binary areal differences, examining each pair of differences in turn. We create fifteen pairs of models in which we examine the set of sites in two areas at a time, for example the sites in the northwest together with the sites in the north east. We then examine first just the effect of distinguishing a given pair of areas, but we also then examine the effect of including linear distance for just those pairs of sites distinguished in the areas. While all pairs turn out to predict linguistic difference to a significant degree, some are only fairly weak predictors, and significantly worse predictors than distance (the continuum model). We provide the results in Table 1 below. ferences via linguistic differences. We might directly agree to the suggestion to try to explain the aggregate linguistic distances using areas derived from other sorts of human contact, say travel, trade, or migration, which could definitely be intriguing. But note that, however the areas are derived, the quantitative comparison of the explanation based only on distance versus the explanation based on areas remains interesting (just as Shackleton’s analysis also is). Wherever we can conclude that area adds little to distance explanatorily, that conclusion is not challenged by the pedigree of the areas. We note further, anticipating the results below, that we predict that the analysis of genuine dialect continua, such as the Swedish continuum, would lead to very different results (Leinonen 2010).

234

John Nerbonne areal difference

1.

Northwest-Northeast

2.

explained areally

explained by distance

2.9 %

17.9 %

Northeast-Saxony

48.3 %

47.6 %

3.

Saxony-Bavaria

51.7 %

55.7 %

4.

Bavaria-Southwest

5.9 %

29.4 %

5.

Southwest-Palatinate

5.5 %

33.7 %

6.

Palatinate-Northwest

46.4 %

29.4 %

7.

Northwest-Saxony

54.8 %

42.7 %

8.

Northwest-Bavaria

51.8 %

50.4 %

9.

Northwest-Southwest

51.4 %

41.0 %

10

Northeast-Bavaria

67.9 %

64.6 %

11

Northeast-Southwest

52.1 %

56.6 %

12

Northeast-Palatinate

44.2 %

53.0 %

13

Saxony-Southwest

30.0 %

43.6 %

14

Saxony-Palatinate

22.7 %

40.4 %

15

Bavaria-Palatinate

18.0 %

39.6 %

Table 1. Explanatory strength of areal distinctions. The central columns shows the amount of pronunciation variation explained by a given binary areal distinction; the rightmost column the amount explained by distance.

We should add that our measurements cannot be interpreted to mean, e.g., that there is no Bavaria-southwest distinction (Table 1, line 4), but only that Wrede’s (1937) border is not a worthwhile distinction to be drawn for this dataset, which, of course, was collected nearly a century after Wrede’s. This conclusion also assumes that we are examining the sites together with the distances between them, and that we are focusing on the question of whether Wrede’s border should also be drawn to partition the set of sites in the south of Germany. Other researchers are free to attempt to show that another border distinguishing the west and the east in the south is a more explanatory one. Unless they draw the border in a very different fashion, they are unlikely to be successful in this endeavor of course, as they can only obtain different results to the degree that they distinguish sites differently from the way we have here. We return to a further discussion of this issue at the end of this section. If we restrict our attention to areas sharing a border (lines 1–7, 13–14 in Table 1; compare Wrede’s map in Map 1) then we see that the northeast-

How much does geography influence language variation?

235

Saxon border is slightly more predictive than distance (line 2), and that the border between the northwest and the more southern areas are massively explanatory (lines 6–7). No improvement in explained variation is associated with either the east-west border in the north, nor with any of the borders in the south. Since this is a paper on methods in dialectology, and not on German dialectology in particular, we shall not pursue the potentially shocking conclusion here that, assuming continuum effects due to geographic distance, there are really only two important German dialect areas: the north (Plattdeutsch) involving the northeast and northwest on the one hand, and the south on the other, including all the others (Saxony, Bavaria, the southwest and the Palatinate). What would be shocking to the Germanist is not that we detect traces of this major division, nor even that it appears to be the most significant division, but rather that other areas explain rather little variation. To draw that conclusion with confidence we should need to examine a range of perturbations of Wrede’s partition of sites, but the conclusion is definitely suggested by Table 1, where the only distinctions that explain much variation involve areas from either side of the north-south dividing line. All the other areal divisions either explain less than distance alone, or only marginally more. If we assume, as seems reasonable, that distance is the more primary notion of geography (compared to area), and then ask what further notions are explanatory, then only some of Wrede’s areas would cross the threshold of utility in explanation. 5.2

Combined models

If we ignore linear distance but include all the binary areal differences, then we find a somewhat stronger correlation (r = 0.58) than we did for distance alone (r = 0.54), accounting for 33.8 % of the variation in the data. This difference is statistically significant due to the large number of site pairs involved (20 100 pairs involving the 201 sites), but we shall not dwell on that. It is a surprising result that the 15 binary variables together explain more of the variation than the simple distance, vindicating the traditional view in German dialectology, which is visualized in Wrede’s (1937) map, that dialectal space is not a continuum, but rather influences pronunciations discretely, in virtue of the individual sub-spaces (areas) to which sites belong. We also note that not every areal distinction influences linguistic differences significantly in this large model. In particular, Wrede’s distinction between the northwest and the northeast correlates only insignificantly with linguistic differences (p ≈ 0.36), as does the distinction between the Palatinate and the

236

John Nerbonne

southwest (p ≈ 0.24), while the distinction between the southwest and Bavarian is only borderline significant (p ≈ 0.044). Given our examination of the pairs of areas above, these results showing insignificance are hardly surprising. We also have the option of examining models which combine linear distance and areal differences, however, and here we see that the traditional view, which we vindicated above, was also incomplete. Taken together, the two notions of geography correlate much more strongly with geography (r = 0.69), accounting therefore for 47.2 % of the pronunciation variation in the data. It is straightforward to interpret this result as indicating that a good deal of variation is not explained by areal differences, and that some continuum effects persist even with fifteen binary areal distinctions.

6.

Conclusions

The focus of this paper has been the demonstration that quantitative dialectology may contribute new perspectives to discussions of how language variation and space interact. First, we may calculate a measure of the degree to which language variation depends on space – this is the percentage of explained variance in regression models such as those presented above. Second, quantitative dialectology is neutral with respect to the geographical notions brought to the table, assuming that they may be incorporated into regression models. This paper has demonstrated how two different conceptions of geography – linear distance on the one hand versus areal divisions on the other – may be compared in a quantitative fashion. It is clear that the division into geographic areas is only one of the organizing geographic concepts that might be evaluated quantitatively. A second candidate might be political borders, e.g. the case of the U.S.-Canadian border, which Boberg (2000) has argued to define a linguistic barrier. A third might be the radially shaped diffusion from centers of population, trade or government, which are related to the existence of dispersed “relic areas” (Chambers & Trudgill 1998: 94, 119). Other candidates are the “staggered” patterns (Niebaum & Macha 2006: 105) of distributions resulting from diffusions of similar, but not identical dynamics, or perhaps the ribbon-shaped diffusion along important trade and travel routes. Some such patterns will not be analyzed successfully without some clever additions to the methodology I have sketched above, since it is not always clear how to characterize these structures in a way amenable to quantitative analysis, but the effort will be worthwhile.

How much does geography influence language variation?

237

Third, we have not just suggested the feasibility of mixed forms of geographic influence, but more importantly we have demonstrated that geography in fact is influential in a mixed form, involving both distance and area in a very well studied language area such as Germany.

7.

Reflections on geography

In all of what is written above, geography – whether geographic distance or as the basis of an areal division among varieties – certainly should not be understood as a physical influence on language variation, but rather as a useful reification of the chance of social contact. Accordingly, we regard space as something which promotes or discourages social contact, and which does so by virtue first of the distances which settlements may be from one another, and second, by the regions which settlements may co-occupy (or fail to co-occupy). With respect to the first, we note that a space which defines travel times, rather than kilometers of remoteness of positions, may function better than space as it is normally conceived, as the temporal notion influences the chance of social contact even more directly than the geographical notion. Recall Gooskens (2004), discussed above. With respect to the second, we hypothesize that regions are influential to the degree to which they represent relatively closed social networks. In speculating this way I do not mean to suggest that the borders of the German areas sketched by Wrede (1937) were ever impermeable for the purposes of communication, only that communication tended to involve people within a single area more than people from two different areas. This, at any rate, would be the Bloomfieldian (1933) line. Space proves to be massively influential from this perspective, but one should not imagine that its influence is inescapable. Where modern, interactive means of communication support intimate, extended exchange, where interlocutors have the opportunity to appreciate not only what each other is saying, but also how it is being said, and where occasionally adopting one another’s speech habits can be appreciated, one might expect such interactions to influence patterns of language variation, at least in the short term. In this case one might also expect notions like the difficulty of communicative interaction to play a role in predicting linguistic dissimilarity much like that now played by physical remoteness. But let us note at the same time that communities that are defined by occasional communication (lots of blogging, net lists, twitter, etc.), and which are difficult to gauge with respect to their language variety might be characterized by changes that reflect brief accommodation (linguistic registers) rather than by longer term changes in lan-

238

John Nerbonne

guage habits. The jury is still out on whether the newest generation of communication technology will influence language variation more permanently.

References Agresti, Alan & Christine Franklin 2009: Statistics. The Art and Science of Learning from Data. 2nd edition. Upper Saddle River, New Jersey: Pearson. Alewijnse, Bart, John Nerbonne, Lolke van der Veen & Franz Manni 2007: A computational analysis of Gabon varieties. In: Petya Osenova et al. (eds.), Proceedings of the RANLP Workshop on Computational Phonology, 3–12. Workshop at Recent Advances in Natural Language Processing (RANLP), Borovetz. Bloomfield, Leonard 1933: Language. New York: Holt, Rinehart and Winton. Boberg, Charles 2000: Geolinguistic diffusion and the U.S.-Canada border. Language Variation and Change 12(1):1–24. Britain, David 2002: Space and spatial diffusion. In: Jack K. Chambers, Peter Trudgill & Natalie Schilling-Estes (eds.), The Handbook of Language Variation and Change, 603–637. Oxford: Blackwell. Chambers, Jack K. & Peter Trudgill 1998: Dialectology. Cambridge: Cambridge University Press. Goebl, Hans 1986: Muster, Strukturen und Systeme in der Sprachgeographie. Mondo Ladino 10: 41–70. Goeman, Anton & Johan Taeldeman 1996: Fonologie en morfologie van de Nederlandse dialecten. Een nieuwe materiaalverzameling en twee nieuwe atlasprojecten. Taal en Tongval, 48: 38–59 Gooskens, Charlotte 2004: Norwegian dialect distances geographically explained. In: Britt-Louise Gunnarson, Lena Bergström, Gerd Eklund, Staffan Fridella, Lise H. Hansen, Angela Karstadt, Bengt Nordberg, Eva Sundgren & Mats Thelander (eds.), Language Variation in Europe: Papers from ICLaVE 2, 195–206. Uppsala: Uppsala University. Heeringa, Wilbert & John Nerbonne 2001: Dialect areas and dialect continua. Language Variation and Change 13: 375–400. Houtzagers, Peter, John Nerbonne & Jelena Prokic 2010: Quantitative and traditional classifications of Bulgarian dialects compared. Scando-Slavica 59(2): 29–54. Kretzschmar, William A. (ed.) 1994: Handbook of the Linguistic Atlas of the Middle and South Atlantic States. Chicago: The University of Chicago Press. Leinonen, Therese 2010: An Acoustic Analysis of Vowel Pronunciation in Swedish Dialects. Ph.D. thesis, University of Groningen. Nerbonne, John 2009: Data-driven dialectology. Language and Linguistics Compass 3(1): 175–198. Nerbonne, John 2010: Measuring the diffusion of linguistic change Philosophical Transactions of the Royal Society B: Biological Sciences, 365: 3821–3828. Nerbonne, John to appear: Various variation aggregates in the LAMSAS south. In: Catherine Davis & Michael Picone (eds.), Language Variety in the South III, Tuscaloosa: University of Alabama Press Nerbonne, John & Wilbert Heeringa 2010: Measuring dialect differences. In: Peter Auer & Jürgen Erich Schmidt (eds.), Theories and Methods, 550–567 (Language and Space Vol. 1). Berlin: de Gruyter Mouton.

How much does geography influence language variation?

239

Nerbonne, John, Jelena Proki´c, Martijn Wieling & Charlotte Gooskens 2010: Some further dialectometrical steps. In: Gotzon Aurrekoetxea & Jose L. Ormaetxea (eds.), Tools for Linguistic Variation Bilbao. 41–56. Supplements of the Anuario de Filología Vasca “Julio Urquijo”, LIII. Nerbonne, John & Christine Siedle 2005: Dialektklassifikation auf der Grundlage aggregierter Ausspracheunterschiede. Zeitschrift für Dialektologie und Linguistik 72(2): 129–147. Niebaum, Hermann & Jürgen Macha 2006: Einführung in die Dialektologie des Deutschen. 2nd, revised ed. Tübingen: Niemeyer. Shackleton, Robert G., Jr. 2005: English-American speech relationships: A quantitative approach. Journal of English Linguistics, 33: 99–160. Shackleton, Robert G., Jr. 2010: A quantitative examination of English-American speech relationships. Ph.D. thesis, University of Groningen. Séguy, Jean 1971: La relation entre la distance spatiale et la distance lexicale. Revue de Linguistique Romane 35(138): 335–357. Wieling, Martijn, Wilbert Heeringa & John Nerbonne 2007: An aggregate analysis of pronunciation in the Goeman-Taeldeman-van Reenen project data. Taal en Tongval 59(1): 84–116.

240

Benedikt Szmrecsanyi

Benedikt Szmrecsanyi

Commentary: Lost in space? The many geographies and methodologies in research on variation within languages1

1.

Introduction

The five papers in this section differ markedly in terms of (i) how geography is conceptualized, (ii) the type of data that is being drawn upon, (iii) the way in which the interplay between language and geography is approached analytically, and (iv) how the findings are interpreted. This commentary aims to scrutinize these differences in turn.

2.

‘Geography’

By and large, the papers feature three different notions of how geography interacts with language structure, use and perception: place, labeling, and social identity: Barbara Johnstone is interested in how “place” is an ideological construct that emerges in social interaction. She thus focuses on the enregisterment of “Pittsburghese” as a regional variety that appears to be more real folk-linguistically than dialectologically (in the traditional sense). Along similar lines, Paul Kerswill explores geographical, social, and ethnic divisions in London to establish how young people label urban language varieties – either their own variety or those of others. areality and boundaries: Bernd Kortmann draws on a conceptualization of geography that is inspired by the customary way of thinking in areal linguistics: to what extent are particular structural features characteristic of varieties of English in particular anglophone world regions? Elvira Glaser’s contribution is likewise – though on a much smaller geographic scale – interested in area formation and in the diffusion of linguistic features in Swiss German dialects; her paper puts particular emphasis on borders, boundaries, and isoglosses. x

x

1

I am grateful to David Britain for inspiring this title.

Lost in space?

241

geographic distance: Areality and dialect areas also play an important role in John Nerbonne’s contribution. Nerbonne, however, is additionally concerned with the extent to which geographic distance (in km) between German dialect locations predicts linguistic distance, or difference, between these locations. The predictive power of geographic distance, a measure which is seen as a proxy for the likelihood of social contact, is subsequently compared to that of areality. Notice that the contributions by Johnstone and Kerswill highlight the social nature of space, which is why the contributions exhibit a certain affinity to the Perceptual Dialectology Paradigm (e.g. Niedzielski and Preston 1999): what matters is the ideologies and map labels that language users have in their mind, and not so much “objective” maps as created by geographers. By contrast, Kortmann’s and Glaser’s interest in areality and boundaries follows a venerable tradition in typology and dialectology that goes back at least to Schleicher (1863). Their approach is beholden to non-linguistic notions of geography and geographic map space, as is Nerbonne’s take on geographic distance as a gradient exploratory factor, which ultimately goes back to Schmidt’s (1872) “wave theory” (Wellentheorie) of linguistic change and diffusion. We note, finally, that all of the geography- and map-based papers in the section (Kortmann, Glaser, Nerbonne) consider space essentially a two-dimensional plane. That is to say, unlike Johanna Nichols (this volume), none of the authors in the section takes altitude, in addition to latitude or longitude, into account.

x

3.

Data

In terms of data sources, the papers in this section fall into two groups: Johnstone and Kerswill analyze naturalistic discourse – narratives (Johnstone) and interviews (Kerswill) – to address issues of (regional) identity construction by exploring how speakers talk about language and space, rather than how speakers talk as a function of geographic space. Kortmann, Nerbonne, and Glaser explore survey and atlas material, which maps the occurrence or non-occurrence of linguistic features in geographic space: The World Atlas of Varieties of English (Kortmann), the Kleiner Deutscher Lautatlas – Phonetik (Nerbonne), and e.g. the Syntaktischer Atlas der deutschen Schweiz (Glaser). This is another way of saying that Johnstone and Kerswill rely on usagebased data (although they of course also address matters of perception and attitudes), while Kortmann, Nerbonne, and Glaser draw on structured data (tabular data) created by atlas/survey compilers, fieldworkers and/or x

x

242

Benedikt Szmrecsanyi

expert informants. In this endeavor, Kortmann and Glaser are interested in morphosyntactic variation, while Nerbonne explores accent variation.

4.

Analysis

It is clear from the foregoing discussion that the name of the game in the contributions by Johnstone and Kerswill is discourse analysis. The paper by Johnstone, which is the most qualitative contribution in the section, offers a genuinely qualitative analysis, while Kerswill also marshals some quantitative techniques (such as keyword analysis). The contribution by Glaser is fairly programmatic in nature: to the extent that empirical issues are addressed, the paper focuses on feature selection, dataset creation, and mapping. Kortmann marshals frequency analyses and various exploratory analysis techniques (such as phenograms and Grammar Genome maps) to gauge the relative importance of areality in the dataset under analysis. Nerbonne’s contribution is the most quantitative paper in the section, which uses regression designs to disentangle the explanatory power of geographic distance and areality.

5.

Interpretation

Given that the contributions ask different research questions, operate on different notions of space, and use different data sources and analysis techniques, it is hardly surprising that the ways in which findings are interpreted likewise differ in many ways. It seems to me that the overall most crucial distinction on the interpretational plane is that between rather dynamic approaches (Johnstone, Kerswill), which focus on the emergence of regional identities and naming preferences, and more static approaches (Kortmann, Nerbonne, Glaser), which analyze synchronic snapshots of dialect and variety landscapes, although matters of diffusion are of course also addressed.

6.

Conclusion

In summary, the five contributions are characterized by the following contrasts: Geography: less geographic, more social and identity-based (Johnstone, Kerswill) versus less social, more geographic and map-based (Kortmann, Nerbonne, Glaser). Data: usage-based (Johnstone, Kerswill) versus atlas/survey material (Kortmann, Nerbonne, Glaser). x

x

Lost in space?

243

Analysis: qualitative (Johnstone, Glaser) versus quantitative (Kortmann, Nerbonne); Kerswill covers the middle ground. Interpretation: dynamic (Johnstone, Kerswill) versus static (Kortmann, Nerbonne, Glaser). Given these substantial differences, are we truly lost in space then? I believe we are not so much lost but rather faced with an exciting tangle of language, society, and space. The papers in the section explore different but equally ‘real’ facets of this entanglement. The lowest common denominator is that all authors are in agreement, implicitly or explicitly, that geographically conditioned variation and variationally conditioned perceived geography are ultimately social phenomena.

x

x

References Niedzielski, Nancy A. & Dennis Richard Preston 1999: Folk Linguistics. Berlin/New York: Mouton de Gruyter. Schleicher, August 1863: Die Darwinsche Theorie und die Sprachwissenschaft. Offenes Sendschreiben an Herrn Dr. Ernst Haeckel, o. Professor der Zoologie und Direktor des zoologischen Museums der Universität Jena. Weimar: Böhlau. Schmidt, Johannes 1872: Die Verwandtschaftsverhältnisse der indogermanischen Sprachen. Weimar: Böhlau.

244

Benedikt Szmrecsanyi

Section 3: Interactional spaces

Interactional space and the study of embodied talk-in-interaction

247

Lorenza Mondada

Interactional space and the study of embodied talk-in-interaction

1.

Introduction

This paper highlights some of the ways in which the concept of interactional space can contribute to conversation analysis, ethnomethodology and interactional linguistics. As I use it (Mondada 2005, 2007a, 2009a), the concept draws on the work of Erving Goffman, Adam Kendon and Charles Goodwin. The paper takes into consideration, within the detailed description of turns at talk and of talk-and-other-conducts-in-interaction, the relevance of participants’ bodies as arranged in the material surroundings in which their social activities take place. Interactional space contributes to our understanding of phenomena studied both by Conversation Analysis (such as the sequential organization of action, the organization of participation and the organization of turns) and by Interactional Linguistics (such as linguistic projections, the emergent construction of turns, grammar as a situated set of resources).

2.

The notion of ‘interactional space’

Within linguistics, space has been extensively studied in terms of how local reference can be established in various ways within a diversity of language systems, highlighting the nexus between speaker, grammar and cognition (see Bohnemeyer & Tucker, in this volume). Grammatical and phonological variation as anchored in specific places has been studied within areal typology and dialectology, and represented in linguistic atlases (see Cysouw, in this volume, Nichols, in this volume, Glaser, in this volume, Kortmann, in this volume). In these cases, space is often seen from a cartographic perspective, and tends to be reduced to a point or an area characterized by a homogeneous linguistic system or usage pattern. Other, more sociolinguistic, studies have focused on specific ways of speaking within particular communities in given (often urban) places – with place being thought of more as a socio-geographical context than as a structured environment for practices (see Streeck, in this volume). Thus, in these traditions within linguistics, space has been considered either (a) as a referent or a cognitive represen-

248

Lorenza Mondada

tation, coded by specific grammatical and lexical forms, or (b) as a place, dealt with on different (macro, meso, micro) levels as a setting where shared uses of language are observable (for an extensive review of the ways in which space and language have been conceptualized, see Mondada 2000; Auer & Schmidt 2010). What these paradigms have not considered, however, is space as the material surroundings in which embodied talk-in-interaction, and more generally social action, takes place. Within Conversation Analysis, the spatiality of talk-in-interaction was first taken into account in studies dealing with place formulations (Schegloff 1972). These studies have also considered the place in which/from where the participants are talking. Place formulations have been analysed in everyday conversation (Schegloff 1972), and also in specific settings in which they have an important role in the achievement of specific tasks and activities. For example, in emergency calls, place formulations are crucial for successful help dispatch (Bergmann 1993; Fele 2008; Mondada 2010, 2011b; Zimmerman 1992). Place reference here plays a key role in the planning, coordination and organization of action and mobility within space. Location, deictic reference and place descriptions are organized not only by reference to an object, a person or an event to be localized within space: they are also formatted to take into consideration the position of the speaker and that of co-participants – i.e. by taking into account the space in which speakers/recipients are located in addition to the referred-to space. This is particularly visible in direction-giving, which involves practical formulations taking into account a starting-point as well as a target, and which are produced for and within mobile actions. Direction-giving has been studied from a variety of perspectives in linguistics (see Klein 1979; Wunderlich 1976: 363–366 for early works) and psychology (see for example Allen 2000; Denis et al. 1999; Mark and Gould 1995; Taylor and Tversky 1992), as well as in Conversation Analysis (for early works see Auer 1979; Psathas 1979). Instead of privileging linguistic and cognitive representations of space which are activated and actualized while delivering the route description, the latter has emphasized the coordinated way in which the itinerary is generated in interaction and through situated practical reasoning. Itineraries are provided in response to an inquiry, within the formats of various “direction giving sequences” which are specifically organized depending on how the initial question is formulated (cf. “how to get there” vs. “where are you?” sequences, Psathas 1986) and in such a way as to monitor the recipient’s understanding (Psathas 1991; Mondada 2007b). This monitoring implies adjustments to the position of the recipients which can be either fixed or mobile, as in cars (Haddington and Keisanen 2009; Haddington 2010; Mondada 2005)

Interactional space and the study of embodied talk-in-interaction

249

or aeroplanes (Nevile 2004). Route-giving can mobilize a variety of linguistic, gestural and bodily resources (Mondada 2007c) as well as material artefacts such as maps and internet resources (Psathas 1979; Brown and Laurier 2005; Mondada 2010, 2011b). Place descriptions and direction-giving thus involve in a crucial way the position of the participants of the interaction. This position can be seen as a point within space, formulated in more or less selective and abstract ways. It can also be conceived as being achieved through the finely-tuned coordination of an arrangement of bodies in face-to-face interaction, i.e. as an interactional space The idea of interactional space emerged very early in the interactionist tradition: Goffman, in Asylums (1961), Behavior in Public Places (1963), and through the notion of “territories of the self ” (1971) and Kendon, through his notion of “F-formation” (1990), have articulated the organization of action and the organization of space – the latter conceived as both constraining and reshaping the former. Goffman and Kendon have shown that body arrangements in space create temporary and changing, bounded territories, recognized both by participants involved in the encounter and by bystanders. The positions of the bodies delimit a temporary “ecological huddle” (Goffman 1964) which materializes the “situated activity system” (Goffman 1961). These arrangements constitute “focused gatherings” (Goffman 1963), which are defined by mutual orientation and shared attention as displayed by body position, posture, gaze and addressed gestures. This interest in temporary territories and in their effectiveness was shared by Ashcraft and Scheflen, observing on the basis of video-taped encounters in private and public settings that “the unoccupied space in the center of the group nevertheless becomes a claimed territory. Others outside the circle customarily recognize the territory” (1976: 7). Kendon (1990: 248–249) conceptualized this territory by introducing the notion of F-formation: body positions and orientations build an arrangement favouring a common focus of attention and the engagement in a joint activity. F-formations can take different shapes: faceto-face is the one usually studied but other configurations exist as well, such as side-by-side or L-formations. Kendon comments on the fact that the particular disposition of the bodies creates a particular interpretative framework which configures the social interaction going on in that setting: “there is a systematic relationship between spatial arrangement and mode of interaction” (1990: 251). F-formations can be studied by emphasizing their lability or their stability: whereas Kendon points out that F-formations, once they are established, have a durable nature, it is possible to highlight their dynamic and ever-changing nature.

250

Lorenza Mondada

Goodwin (2000, 2002, 2003, 2007a) has insisted on the mutual relationships between embodied actions and material environment, defining what he calls “contextual configurations”. Contextual configurations frame and make visible action as it unfolds and as it is organized by a multiplicity of semiotic resources (2000: 1490). If the analysis of talk has to take into consideration the embodied actions of the participants, the study of gesture or body postures cannot be developed in isolation, but has to describe the way in which the structure of the environment contributes to the organization of gesture (2002): the arrangement of multiple bodies provides for the contextualization of gesture. Drawing from these inspirations, Mondada (2005, 2007a, 2009a, 2011a) suggests that interactional space is constituted through the situated, mutually adjusted changing arrangements of the participants’ bodies within space, as they are made relevant by the activity they are engaged in, their mutual attention and their common focus of attention, the objects they manipulate and the way in which they coordinate in joint action. This interactional space is constantly being (re-)established and transformed within the activity (De Stefani and Mondada 2007; LeBaron and Streeck 1997; Mondada 2009a, 2011a; De Stefani 2010). In this sense, the bodily arrangements of the participants constitute mobile configurations and mobile formations. Defined in this way, the notion of interactional space is intended to contribute both to an interactional conceptualization of space and to a spatial conceptualization of interaction. As regards the former, interactional space implies a praxeological perspective on space, insisting on the reflexive elaboration of situated action and of relevant spatial features. The challenge is to consider the material surroundings of action both as constraints and as resources – conceiving space as action-shaping and action-shaped, in a way that avoids any spatial determinism (Hausendorf, Mondada & Schmitt 2012). The challenge is also to relate this approach of space to the sequential and temporal organization of talk and embodied action, in order to show how interactional space unfolds moment by moment within the coordinated adjustment of various simultaneous streams of action and sets of multimodal resources. As regards the latter, interactional space approaches the context of the interaction by considering the material and spatial environment of talk; it enlarges the vision of multimodality by taking into consideration the entire body. Likewise, the focus on interactional space permits us to consider the spatial dimension of participation framework by taking into consideration its embodied nature (embodied participation, see Goodwin 2000; Goodwin and Goodwin 2004) together with the relevance of the spatial arrangement of

Interactional space and the study of embodied talk-in-interaction

251

bodies and body orientations (see Schmitt 2012). Also, it reveals that interaction often takes place in mobile settings, with the bodies of the participants constantly moving instead of remaining static (Haddington, Mondada & Nevile 2013). In this paper, I will employ this notion of interactional space as a way to cast new light on different levels of the organization of the interaction: at the level of its overall organization, I show how interactional space is actively constituted by the participants in the opening of the exchange (section 3); I then show how it is transformed in the transition from one activity to the other (section 4) and how participation frameworks dynamically evolve (section 5); finally, I show that it also affects the gradually emerging organization of turns at talk (section 6).

3.

The establishment of interactional space in conversational openings

Openings have been traditionally described on the basis of telephone conversations as the sequential moment where the joined entry of all of participants in the interaction is achieved in a finely coordinated way. Summonsanswer sequences (Schegloff 1968) provide the basic device through which participants display their mutual availability and readiness to engage in further talk. Interestingly, although there is a large amount of literature on telephone openings, openings of face-to-face interactions are still largely underrepresented. Naturalistic, video-recorded data show that face-to-face openings share a number of features with telephone openings (cf. PilletShore 2008) but also that long preparatory pre-opening” activities occur before the participants actually talk to each other (see Mondada & Schmitt 2010 for various empirical studies). These preparatory activities are largely based on multimodal resources and practices, such as gaze, body orientation, walking, entering a room, taking a seat, etc., which are significantly anchored in space and which prepare and configure the emergent interactional space. Thus, openings are a perspicuous setting from which to observe the way in which interactional space progressively emerges and is eventually stabilized. More particularly, the analysis of openings of encounters in public places has revealed how interactional space is actively established by the participants, through the organization of their walking trajectories and the mutual adjustments and convergent movements, transforming various “vehicular units” into a unique one (Goffman 1963), through “sighting” (Kendon 1990) and the first exchange of gaze, as well as progressive mutual identification and recognition (see Mondada 2009a for a systematic analysis). Thus,

252

Lorenza Mondada

a multiplicity of embodied resources is mobilized for the achievement of interactional space. In order to explore the various ways in which the interactional space is shaped we focus on the opening of a face-to-face encounter. The fragment not only shows the multiple multimodal resources which participants exploit in this sequential environment, but also the possible alternative spatial configurations that can be created by them. The fragment also points to the fact that the video-recording device also contributes to the way in which the interactional space is designed – since participants orient here to the fact that the interaction is being recorded – and the ethical concerns relative to their informed consent. In excerpt 1, Rita comes to pick up her boyfriend, Guy, in order to drive him to an open air concert. We join the action as Rita, waiting in the car, sees him approaching. The opening of their encounter takes place outside the car, orienting to the fact that there are cameras inside: Rita gets out of the car in order to ask Guy if he agrees to be recorded, and it is only after some negotiations that they sit together in the car and begin the ‘how-are-you’ sequence. As we can see, the interactional space of the encounter is actively shaped by the participants in a variety of ways. For transcript conventions – especially for those concerning multimodality – see the note at the end of this paper. Extract 1.1 (EMIC 1707R306c)

Interactional space and the study of embodied talk-in-interaction

253

As Guy is walking towards the car, Rita suspends the imminent opening of the encounter by producing three “attends” (‘wait’) (2, 4, 6), while she exits from the car (1–6, images 1–5). In this way she delays the greetings, and she positions herself in a way that creates an other interactional space than the one given by the passenger cell of the car in which Guy is usually supposed to sit. The fact that this usual positioning of the participants is delayed is explicitly formulated by Rita’s temporal expression (‘before you enter the car’, 6). Thus, both meet outside the car, establishing a face-to-face frontal configuration across the roof (5–6, image 6). As this first interactional space is established, Rita informs Guy about the recording device:

254

Lorenza Mondada

Extract 1.2

Speaking over the roof of the car, Rita tells Guy about the recording device (7). Her subsequent turn-constructional units are produced in an incremental way: confronted with the silence, the reluctant responses and the facial expressions of her partner, she adds several increments and expansions to her explicative turn, as they maintain their body positions – Guy inspecting the interior of the car from outside (9–10, 16–17). In fact, this sequence is a request for authorization (formulated as an informing or even as a warning – “faut que j’avertisse” 7 ‘I have to tell you’), which is only granted by Guy in line 18. Again, it is interesting that the issue of interactional space is formulated by Rita herself (‘since it’s not a space as usual’, 16). An alternative interactional space is thus created for the request, escaping from the camera, and is maintained until the request is granted and the sequence is completed. As soon as Guy accepts, Rita enters the car (19), initiating the adoption of a more common body positioning of the participants. Guy also begins to enter the car, but this movement is delayed by several questions:

Interactional space and the study of embodied talk-in-interaction

255

Extract 1.3

When Guy begins to enter the car (21), he delays his movement in several ways: first, he asks more questions about the study (21, 25, 28), re-occasioning Rita’s proposal to stop the recording (32–25); second, his movement is very slow, suspended several times (beginning in line 21 and ending in line 35). These verbal and embodied delays clearly display a dispreferred response to Rita’s request to authorize the recording.

256

Lorenza Mondada

As Guy finally sits in the car (34), and as Rita again suggests they stop the camera (35), they perform another action which achieves a new interactional space: Guy grasps the authorization form and uses it to obstruct the camera angle (35, images 7 and 8) while they kiss each other (36–40). In this way, they perform the greeting sequence in a special, intimate interactional space, escaping from the camera, which is created in an ad hoc fashion for this particular action. As soon as Guy puts away the forms and as they sit side by side, in front of the camera, the how-are-you sequence is produced, which completes the opening sequence. The interactional space for the journey and the conversation is achieved at this point. The first excerpt has shown how various interactional spaces are established successively by the participants in the course the opening. This reveals that: – the establishment of the interactional space is a process and an achievement of the participants, which can take some time, can be delayed, can be transformed and repaired; – the specificities of the interactional space for a given activity are normatively and socially oriented to by the participants, and can be negotiated and actively configured; – even in a constrained (and apparently pre-structured or architectured) space, such as the car’s passenger cell, the participants use different practices for shaping the adequate interactional space for their actions; – the recording device reflexively contributes to the shaping of the interactional space, inter alia by the way in which the camera delimits the visible and documentable frame of the publicly recorded actions, and by its possible ethical implications for the participants (Mondada 2009b). In sum, our analysis of the fragment shows the intrinsically dynamic and flexible nature of interactional space: even if its establishment in the opening is consequential for the rest of the conversation, its configuration constantly changes as it unfolds.

4.

Transitions and reconfigurations of interactional space

Interactional space is often radically redesigned in transitions from one activity, task, or episode to another, thus contributing to the accountable and recognizable character of the transition itself. The organization of transitions is a practical problem for the participants, who mobilize various multimodal resources to achieve it and to make it publicly visible and intelligible (Bruxelles, Greco & Mondada 2009; Beach 1993; Modaff 2003; Mondada &

Interactional space and the study of embodied talk-in-interaction

257

Traverso 2005; Mondada 2006; Robinson & Stivers 2001; Deppermann, Schmitt & Mondada 2010). The reconfiguration of the interactional space contributes to this achievement through a change of orientation and a new distribution of the bodies of the participants. Guided visits (De Stefani 2010; Kesselheim 2010; Mondada 2005, 2012; Stukenbrock & Birkner 2010) are a perspicuous setting for observing these transformations, since the sequential format of the visit is achieved by walking from one explained/described scene to the next. Each movement from one place to the next corresponds to a change of spatial organization,. In fragment (2), taken from a guided visit to a garden, we focus on an occurrence of such a transition. We join the action as the four participants are walking through the garden, following a path in the middle of a variety of trees and flowers. The fragment begins at the end of a previous explanation given by Jean, the cultural manager of the site, about a piece of art – a mirror (which is on their left, visible on the right margin of image 1) – which is part of an ongoing exhibition in the garden. This explanation is given within a participation framework in which Jean has the floor and is the centre of attention. After the completion of Jean’s explanation, Luc, the gardener, explains another object, a tree. The transition from one to the other is bodily projected both by Jean himself and by Luc. Excerpt 2 (jardivis – 5/1.40)

258

Lorenza Mondada

At the beginning of the fragment, Jean explains the piece of art. Yan and Elise are his recipients. They are looking at the object, whereas Luc, the gardener, is looking at them, staying at a slight distance (image 9), This embodied participation framework characterizing the initial interactional space shows who is the main speaker holding the floor, who are the instructed recipients and who is a peripheral participant. In lines 1–2, Jean completes his explanation. His turn construction, uttered in a lower voice and closing intonation, provides for several possible completion points: after the title of the art work, after “voilà”, and at the end of the increment, added in line 2. The imminent completion is also bodily projected by his making a first step forward at the first possible completion (on “portes”, line 1) and by Luc, who begins even sooner to walk towards the next object, while Jean is mentioning the title of the last one. Leaving the static position focused on the previous object (the mirror) and walking towards the next (a tree) constitutes an embodied way of achieving completions and transitions. Thus, both Jean and Luc orient to the imminent completion of the current episode. Luc additionally prepares the orientation towards the next object of the tour by extending his right hand and touching the fruit of a tree in front of him. When he first does this (image 10), nobody is looking at him. The main recipients of the guided visit, Yan and Elise, are still bodily oriented towards Jean and the object he is explaining, the mirror. Interestingly, Luc retracts his hand while Yan is still producing some kind of response to the explanation (4). As soon as Yan’s turn is finished, Luc stretches out his arm again while

Interactional space and the study of embodied talk-in-interaction

259

breathing in (line 5) – both work as projectors of a next turn (and initiators of the next sequence). During the pause which follows Yan turns to Luc (6). At this point, Luc points to the fruit and initiates the description of the next object, “les amélan:ches” (7–8). Elise turns to him as well as soon as he has uttered the turn-initial “voyez”, (7), a discourse marker which works as an attention-getting device, and as he inserts a parenthetical remark about the fact that he had already mentioned these plants before (7). The construction of the turn which initiates the new sequence – responded to by Yan with a change-of-state token (“ah”, 10) and with a tactile response (touching the fruit) (image 11) – is coordinated with securing the attention of the main recipients and is finely adjusted to the delays in the establishment of this joint attention. It also corresponds to the establishment of a new interactional space, in which all co-participants form a new interactional space around the “amélan:ches”. In sum, transitions between one stage of the tour through the garden and the next are achieved by the participants dissolving the previous interactional space and by progressively establishing another one. This transformation is prepared, projected and negotiated by the participants, who closely monitor the state of attention of their co-participants and adjust to them in a stepwise and finely-tuned way.

5.

Interactional space and the transformation of embodied participation frameworks

The transformation of interactional space is strongly related to the dynamic changes of the participation framework which evolve during the course of the interaction. Interactional space is sensitive to these transformations and reflexively contributes to their achievement. The notion of participation was introduced by Goffman (1979); its reception in linguistics has led to a rather abstract typology of participants (Levinson 1988) which has been useful for distinguishing various “voices” within the same speech event. It has failed, however, to address the actions and practices constituting participation as a dynamic process, and the embodied dimension of participation. The notion has been revisited by Goodwin and Goodwin (2004), focusing on participation in situated courses of action. They have foregrounded the role of multimodal resources, including the use of objects and the structure of the environment. In addition, they have investigated their temporal coordination within emergent and switching participation modes. Multimodality is not only involved in the “animation” of various voices within a narrative (2004: 231–237), but also in the embodied

260

Lorenza Mondada

constitution of participation frameworks. An example is “by-play” (as investigated by Goodwin 1997), which occasions changes in footing, partitions the audience and its serious or playful stances towards the on-going talk, and disturbs the official line of talk as well as the bodily posture of the person holding – and ceding – the floor (1997: 84). As soon as we analyse participation frameworks as embodied and situated, we can describe them as interactional spaces achieved through mutual gaze, common foci of attention, reciprocal body orientations, disposition of the bodies within the environment and alignment within the same activity (cf. Goodwin: “participation frameworks create an embodied, multi-party environment”, 2007b: 15). Conversely, instances of disengagement (Szymanski 1999; Goodwin 1981: 95–125) are visible in participants’ gaze and bodily retractions from the interactional space (see for other examples LeBaron and Streeck 1997; McIllvenny 1996; Rae 2001; Deppermann, Schmitt, and Mondada 2010). How changes in participation are implemented within the interactional space is shown in an exemplary way in the next excerpt, in which two passers-by ask a couple the way. The couple first responds together, constituting a first interactional space; but then the husband literally leaves the space, disengaging from the activity on the grounds of ‘not being from here’ and therefore ‘not knowing’: Excerpt 3 (MTP/Etuves9)

Interactional space and the study of embodied talk-in-interaction

261

The participants walk from two opposite directions towards each other and begin to stop walking at the beginning of line 1: at the end of line 1, the interactional space for the encounter has been established; the four participants have reached a stable face-to-face formation (image 12). After a lapse (2), A begins to answer the request (3), displaying her recognition of the activity and her willingness to provide a route description. This is also shown by her turning towards possibly relevant directions within the local environment (image 14). Before A’s projected description is given, B announces (5) his incompetence and claims membership in the category “stranger”, which makes him irrelevant to the task at hand. This announcement is followed by a change in posture and in the participation framework: he moves behind A (images 15 and 16).When A turns her head again towards F and E, and begins to deliver the route description and to point, B has moved behind her (image 5), and literally left the interactional space. We here observe a reconfiguration of the interactional space, from a faceto-face formation between two couples to one with three people. In sum, interactional space is constantly changing. It is transformed by orienting to the ongoing changes in the activity and in the participation and thereby contributes to their public visibility, and achievement. Inclusions and exclusions, modifications of participation statuses, recipient design and mutual orientations are made visible by the bodily positions of the participants, shaping the evolving interactional space.

262

6.

Lorenza Mondada

Interactional space and the situated adjustments of turns at talk

The configuration and transformation of interactional space are not only sensitive to global interactional structures, like opening sequences, transitions and changes of participation frameworks. Interactional space is also reflexively oriented to the emergent and incremental formatting of turns at talk and to their situated and embodied details. The online formatting of turns is obviously related to their temporal unfolding (Goodwin 1979; Auer 2005, 2009): turns are produced step by step by the speaker, reflexively oriented towards the actions of the co-participants, which reshape their projections and trajectories. This emergence of talk is not only related to temporal constraints, however, but also more broadly to situated and contingent constraints – among others to the ongoing achievement of the interactional space. Turns adjust to the arrangements of the bodies within space, to the foci of attention of the participants, and to their mutual attention. In the previous excerpts, we have shown that the beginning of the interaction (excerpt 1) can be delayed if an adequate interactional space has not been yet constituted. Similar delays can occur in transitions; alternatively, transitions can be prepared and projected such that they can be done smoothly (for example, by pre-arranging the interactional space, anticipating the next task, etc.). Delays and anticipations characterize the organization of turns at talk, too: when a turn is already on its way, it can be delayed or accelerated for adjustment to the ongoing changes of the interactional space (cf. Relieu 1999 for an early account). In the following fragments, turns are formatted while they emerge as the speaker and/or the recipients are walking. Participants progressively configure the interactional space which is adequate for the micro-activity of pointing at a referent. Thus, these fragments show the importance of interactional space for syntax and deixis (for other examples, see Mondada 2005, 2007a). Excerpts 4 and 5 are again taken from a corpus of route-descriptions. E and F ask a passer-by (P) where a touristic landmark is located. Here, I focus less on the configuration of the participants in the opening of the interaction (cf. supra), but more on the moment when the route description is provided by the passer-by: Excerpt 4 (MTP-E11, 33’45’’)

Interactional space and the study of embodied talk-in-interaction

263

P responds to the request at line 6. She mentions the target and begins to point (image 17), projecting more to come. At this point, she retracts her pointing gesture and delays her turn completion by a pause (7). The end of the turn is delivered later, with a deictic (“là:” 8) and a new pointing gesture (image 18–19–20), after P has initiated a rearrangement of the participants’ body positions. Already in line 6 P has begun to walk, inviting F to follow her; she stops when she has reached an adequate position for her pointing, with which F bodily aligns bodily. Only when this new interactional space has been created does she deliver her route description, supplemented by an extended pointing gesture (image 20).

264

Lorenza Mondada

Delays and movements which reconfigure the interactional space are common in deictic reference and pointing to an object not directly accessible, as can also be seen in the next fragment, taken from the same corpus: Excerpt 5 (MTP, 31.00)

In this excerpt, too, P begins his route description by repeating the target in the syntactic format of a left dislocation; but, instead of completing the clause, he inserts various materials, side commentaries and formulations. He also inspects the environment (image 21), and walks from one place to another (image 22) (3–4, 6). The active establishment of a new interactional space is explicitly formulated by him (4), as well as the delays they occasion (“attendez”, line 3). P finally stops (9) at the end of a long pause, during which

Interactional space and the study of embodied talk-in-interaction

265

the participants have walked together in silence. When E and F have stopped too, and after the insertion of further delaying material (10), P begins to point and finally completes the spatial description with a deictic element and a pointing gesture (“un peu à droite là”, line 12, image 23). Thus, in excerpts 4 and 5, a description is launched, then suspended, while some walking is going on. This transforms the initial interactional space and adapts it to the next action, a spatial reference based on deictics and pointing. We can observe a specific sequential distribution of linguistic resources: whereas the NP referring to the target is left-dislocated to the beginning of the route description (“l’église saint-roch” in exc. 4, line 6; exc. 5, line 3) and responds to the query at a moment where the adequate context for reference is not yet available, the pointing gesture and the deictic element are positioned at the end, when the reorganization of the interactional space is complete. Pointing and indexical reference are achieved only after the new interactional space has been established. Moreover, whereas during the beginning of the description and the walk the recipients look at P, they look in the direction indicated by the pointing gesture when the description is available and the pointing occurs: they bodily orient towards the target, assuming a position within the interactional space which is no longer face-to-face but rather sideby-side. In the previous excerpts, the name of the referent was given in the request, initiating a search for it within the environment. As we have seen, this is taken up early in the response – typically through a left dislocation – and projects more to come. Its elaboration comes later, and the rearrangement of the bodies of the participants is the precondition for the right indexical practices. In other cases, the referent is not shared by all participants. Its introduction orients to the possible dispersed attention of the participants and to the task that has to be achieved in order to organize their joint attention, even before the referent can be accessed. In this case, the interactional space is reorganized before the specification of the referent and other projective resources are used – typically pseudo-clefts, which have a first part characterized by a strong projective potential and by an underspecified argument (Günthner and Hopper 2010), followed by a second part which can be delayed. The following excerpt is taken from a visit made by four people to a famous building, in which Yan, an architect, points at and describes a detail of the roof for his co-participants. The excerpt is based on a video-recording using two cameras: camera glasses filming more or less precisely what Yan is looking at, and a mobile camera following the participants:

266

Lorenza Mondada

Excerpt 6 (Cep1/56.27–Lunc52.12)

Yan self-selects after a long silence, during which the co-participants walk down a ramp from the first floor to the ground floor. The group moves in a dispersed order (image 1): Yan walks behind them, still in the middle of the ramp, when he begins to talk, and the others walk one after the other and look in different directions. In this context, Yan is confronted with a practical problem: the current interactional space lacks mutual attention as well as a common focus of attention, and is characterized by walking as individual actions rather than walking together. The way in which he multimodally formats his turn addresses these peculiarities and contributes to a reconfiguration of the interactional space in a way that re-establishes mutual coordination and joint attention. In order to do so, Yan, having spotted a detail in the ceiling (1, image 24A), actively monitors the position of the co-participants (1, image 25B) before he begins his turn. Excerpt 6a

Interactional space and the study of embodied talk-in-interaction

267

Yan gazes at the co-participants during the first part of his turn (“c’est vraiment c’que:”, line 2). Still monitoring their position, he produces the first part of a construction, which is indeterminate from a semantic perspective and projects more to come. This first part is followed by something which can work as a second part but which can also be interpreted as another first part of a pseudo-cleft (in a pivot construction: [c’est vraiment + [[c’que: jmoi j’adore,] (.) c’est p]], line 2). This pseudo-cleft also projects more to come. Two strongly projective constructions are used one after the other, without yet specifying the referential or the semantic gist of the construction. Consequently they work as an attention-getting device, sensitive to the peculiar distribution of the participants’ bodies within space. His recipients respond to these projections: progressively they look at him and then at the ceiling he is pointing at, as shown in the ELAN version of the transcript: Excerpt 6b

Yan begins his second pseudo-cleft construction when Elise, who is the last one to turn to him, begins to gaze at him. Having secured their gaze, Yan points vertically and looks up. These two movements constitute instructions to look in the same direction (Hindmarsh and Heath 2000).

268

Lorenza Mondada

As shown in excerpt 6c, the participants progressively reorient their gaze from Yan to the ceiling he is pointing at (images 26, 27, 28). It is only after a pause of 0.6 seconds – as Yan looks down again, and sees that a first co-participant, Elise, begins to look up – that he introduces the referent, with a complex indexical expression (“ces plans là”). Excerpt 6c

During the production of the demonstrative NP, Yan continues to monitor his recipients. At this point, he sees that Sophie and Elise are looking up at the ceiling, but that Jean is still looking at him (image 27). It is only at the

Interactional space and the study of embodied talk-in-interaction

269

completion of the NP that everybody looks up towards the object referred to (image 28). The multiple preparations and projections of Yan’s turn display its recipient-oriented character; this orientation does not concern the co-participants individually but as a group – a group to be monitored, coordinated, and directed for the successful achievement of reference. The excerpts of this section share a number of features: (a) the emergent organization of the turn is finely adjusted to an emergent spatial reconfiguration; (b) deictic reference is produced, both with pointing gesture and deictic expressions, only when the interactional space has been adequately reshaped for the action being carried out within the turn; (c) as the reconfiguration of the interactional space is under way, the organization of the turn is characterized by constructions projecting more to come and by delays of the referential focus. Thus, the emergent turn format is finely adjusted to the stepwise reorganization of the interactional space. These features seem to be particularly important for the grammar of deictic expressions in use (see Mondada, 2005, 2007a for a systematic analysis across a variety of settings). It shows that deixis is not so much a “context-dependent” but more a “context-shaping” phenomenon (see, for example, Hanks 1992: 70: “verbal deixis is a central aspect of the social matrix of orientation and perception through which speakers produce context”) where the context is a reflexive accomplishment of the participants’ actions and the (re-)arrangements of their bodies within space. These features, however, also highlight the situated and contingent nature of the organization of turns at talk. They suggest an embodied conception of recipient design, governed by bodies’ dispositions and arrangements, coordinated action and joint attention, as well as mutual orientations, all spatially organized.

7.

Conclusion

In this paper, I have discussed some of the phenomena studied in conversation analysis and interactional linguistics at different levels of organization (overall structure, activity organization, sequence organization, turn organization) to which the notion of interactional space can contribute. Taking into account space in the analysis of social interaction has various consequences, both for a spatialized vision of language in interaction and for an interactional view on space. For the study of space it allows us to develop a praxeological approach to space, which takes it to be reflexively shaped by social action and not just a pre-

270

Lorenza Mondada

existing deterministic constraint on it. This approach is important for the way in which we conceive context in general, as being both structuring for and structured by social action, and also for the way in which we conceive of the materiality of the environment. This materiality has structured features and constraints, but these are effective only when they are enacted and made relevant within a course of action. In this sense action contributes to configuring space, by selectively discriminating and highlighting relevant features. Interactional space is the result of these practices. For the study of language in interaction, the notion of interactional space invites us to take into consideration not only the embodied aspects of action but also their spatial distribution and arrangement within the environment, using that environment as a structuring resource. It is important not to forget that these arrangements are temporally deployed in a finely coordinated way. The temporality of talk, the simultaneities of gesture and other embodied dimensions of action, and the movements of the body are organized within single gestalts that have to be studied as a whole, taking into consideration the entire body and its mobility. For the description of linguistic phenomena, this means that an interaction-based grammar should integrate not only the sequential position of turns-at-talk, but also the related sequential positions in which the “environmentally coupled gesture” (Goodwin, 2007a) and spatially arranged embodied postures are concomitantly and systematically produced.

Transcript conventions Data have been transcribed according to conventions developed by Gail Jefferson (2004). An indicative translation is provided line by line. Multimodal details have been transcribed according to the following conventions: * * delimitate a speaker’s gestures and action descriptions. + + delimitate another speaker’s gestures and action descriptions. *---> gesture or action described continues across subsequent lines. *--->> gesture or action described continues until and after excerpt’s end. –--->* gesture or action described continues until the same symbol is reached. >>-gesture or action described begins before the excerpt’s beginning. .... gesture’s preparation. –--gesture’s apex is reached and maintained. ,,,,, gesture’s retraction.

Interactional space and the study of embodied talk-in-interaction

yan im #

271

participant doing gesture is identified in the margin when (s)he is not the speaker. the exact point where screen shot (figures) has been taken is indicated, with a specific sign showing its position within turn-at-talk.

References Allen, Gary L. 2000: Principles and practices for communicating route knowledge. Applied Cognitive Psychology 14(4): 333–359. Ashcraft, Norman & Albert E. Scheflen 1976: People Space: The Making and Breaking of Human Boundaries. New York: Anchor. Auer, Peter 1979: Referenzierungssequenzen in Konversationen: Das Beispiel “Ortsangaben”. Linguistische Berichte 62: 94–196. Auer, Peter 2005: Projection in interaction and projection in grammar. Text 25(1): 7–36. Auer, Peter 2009: Online syntax. Thoughts on the temporality of spoken language. Language Sciences 31: 1–13. Auer, Peter & Jürgen E. Schmidt (eds.) 2010: Language and Space: An International Handbook of Linguistic Variation, Vol. 1, Theories and Methods. Berlin/New York: De Gruyter. Beach, Whyne A. 1993: Transitional regularities for ‘casual’ “okay” usages. Journal of Pragmatics 19 (4): 325–352. Bergmann, Jörg R. 1993: Alarmiertes Verstehen: Kommunikation in Feuerwehrnotrufen. [Alarmed understanding: Communication in emergency calls to the Fire Department] In: Thomas Jung & Stefan Müller-Doohm (eds.), “Wirklichkeit” im Deutungsprozess. Verstehen und Methoden in den Kultur- und Sozialwissenschaften [“Reality” in sense-making processes. Understanding and methods in cultural studies and social sciences], 283–328. Frankfurt am Main: Suhrkamp. Brown, Barry & Eric Laurier 2005: Maps and journeys: an ethnomethodological investigation. Carthographica 40(3): 17–33. Bruxelles, Sylvie, Luca Greco & Lorenza Mondada 2009: Pratiques de transition: Ressources multimodales pour la structuration de l’activité [Practices of transition: multimodal resources for structuring the activity]. In: Françoise Détienne & Véronique Traverso (eds.), Méthodologies d’analyse de situations coopératives de conception, 221–302. Nancy: Presses Universitaires de Nancy. Denis, Michel, Francesca Pazzaglia, Cesare Cornoldi & Laura Bertolo 1999: Spatial discourse and navigation: an analysis of route directions in the city of Venice. Applied Cognitive Psychology 13(2): 145–174. Deppermann, Arnulf, Reinhold Schmitt & Lorenza Mondada 2010: Agenda and emergence: Contingent and planned activities in a meeting. Journal of Pragmatics 42(6): 1700–1718. De Stefani, Elwys 2010: Reference as an interactively and multimodally accomplished practice: Organizing spatial reorientation in guided tours. In: Massimo Pettorino, Antonella Giannini, Isabella Chiari & Francesca Dovetto (eds.), Spoken Communication, 137–170. Newcastle upon Tyne: Cambridge Scholars Publishing.

272

Lorenza Mondada

De Stefani, Elwys & Lorenza Mondada 2007: L’organizzazione multimodale e interazionale dell’orientamento spaziale in movimento [The multimodal and interactional organization of spatial orientation in movement]. Bulletin VALS ASLA 85: 131–159. Fele, Giolo 2008: The collaborative production of responses and dispatching on the radio: video analysis in a medical emergency call center. FQS / Forum: Qualitative Social Research 9(3), article 40: http://www.qualitative-research.net/index.php/ fqs/article/view/1175/2617. Goffman, Erving 1961: Encounters: Two Studies in the Sociology of Interaction. Indianapolis: Bobbs-Merrill. Goffman, Erving 1963: Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press. Goffman, Erving 1964: The neglected situation. American Anthropologist 66(6): 133–136. Goffman, Erving 1971: Relations in Public: Microstudies of the Public Order. New York: Harper and Row. Goffman, Erving 1979: Footing. Semiotica 25(1–2): 1–29. Goodwin, Charles 1979: The interactive construction of a sentence in natural conversation. In: George Psathas (ed.), Everyday Language: Studies in Ethnomethodology, 97–121. New York: Irvington Publishers. Goodwin, Charles 1981: Conversational Organization: Interaction Between Speakers and Hearers. New York: Academic Press. Goodwin, Charles 2000: Action and embodiment within situated human interaction. Journal of Pragmatics 32(10): 1489–1522. Goodwin, Charles 2003: The body in action. In: Justine Coupland & Richard Gwyn (eds.), Discourse, the Body and Identity, 19–42. New York: Palgrave/Macmillan. Goodwin, Charles 2007a: Environmentally coupled gestures. In: Susan D. Duncan, Justine Cassell & Elena T. Levy (eds.), Gesture and the Dynamic Dimension of Language, 195–212. Amsterdam/Philadelphia: Benjamins. Goodwin, Charles 2007b: Participation, stance and affect in the organization of activities. Discourse and Society 18(1): 53–73. Goodwin, Charles & Marjorie H. Goodwin 2004: Participation. In: Alessandro Duranti (ed.), A Companion to Linguistic Anthropology, 222–244. Malden, MA: Blackwell. Goodwin, Marjorie H. 1997: By-play: negotiating evaluation in storytelling. In: John Baugh, Crawford Feagin, Gregory R. Guy & Deborah Schiffrin (eds.), Towards a Social Science of Language: Papers in Honour of William Labov, Volume 2, 77–102. Philadelphia: John Benjamins. Günthner, Susanne & Paul Hopper 2010: Zeitlichkeit and sprachliche Strukturen: Pseudoclefts im Englischen und Deutschen [Temporality and linguistic structures: Pseudo-clefts in English and German]. Gesprächsforschung – Online-Zeitschrift zur verbalen Interaktion 11: 1–28 (www.gespraechsforschung-ozs.de). Haddington, Pentti 2010: Turn-taking for turntaking: mobility, time, and action in the sequential organization of junction negotiations in cars. Research on Language and Social Interaction 43(4): 372–400. Haddington, Pentti & Tiina Keisanen 2009: Location, mobility and the body as resources in selecting a route. Journal of Pragmatics 41(10): 1938–1961. Haddington, Pentti, Lorenza Mondada & Maurice Nevile (eds.) 2013: Mobility and Interaction. Berlin/Boston: De Gruyter.

Interactional space and the study of embodied talk-in-interaction

273

Hanks, William F. 1992: The indexical ground of deictic reference. In Alessandro Duranti & Charles Goodwin (eds), Rethinking Context, Language as an Interactive Phenomenon, 43–77. Cambridge UK: Cambridge University Press. Hausendorf, Heiko, Lorenza Mondada & Reinhold Schmitt (eds.) 2012: Raum als interaktive Ressource [Space as an Interactive Resource]. Tübingen: Narr. Hindmarsh, Jon & Christian Heath 2000: Embodied reference: a study of deixis in workplace interaction. Journal of Pragmatics 32: 1855–1878. Jefferson, Gail 2004: Glossary of transcript symbols with an introduction. In Gene H. Lerner (ed.), Conversation Analysis: Studies from the First Generation, 13–31. Amsterdam: Benjamins. Kendon, Adam 1990: Conducting Interaction: Patterns of Behavior in Focused Encounters. Cambridge, UK: Cambridge University Press. Kesselheim, Wolfgang 2010: Zeigen, erzählen und dazu gehen: Die Stadtführung als raumbasierte kommunikative Gattung [Showing, narrating and walking: city guided tours as a communication genre based on space]. In: Marcella Costa & Bernd Müller-Jacquier (eds.), Deutschland als fremde Kultur: Vermittlungsverfahren in Touristenführungen, 244–271. München: Iudicium. Klein, Wolfgang 1979: Wegauskünfte [Itinerary descriptions]. Zeitschrift für Literaturwissenschaft und Linguistik 9(33): 9–57. LeBaron, Curtis D. & Jürgen Streeck 1997: Built space and the interactional framing of experience during a murder interrogation. Human Studies 20: 1–25. Levinson, Steven 1988: Putting linguistics in a proper footing: Explorations in Goffman’s concepts of participation. In: Paul Drew & Anthony Wootton (eds.), Erving Goffman: Exploring the Interaction Order, 161–227. Cambridge: Polity Press. Mark, David M. & Michael D. Gould 1995: Wayfinding directions as discourse: verbal directions in English and Spanish. In: Judith F. Duchan, Gail A. Bruder & Lynne E. Hewitt (eds.), Deixis in Narrative: A Cognitive Science Perspective, 387–405. Hillsdale, NJ: Lawrence Erlbaum. McIlvenny, Paul 1996: Heckling in Hyde Park: Verbal audience participation in popular public discourse. Language in Society 25: 27–60. Modaff, Daniel P. 2003: Body movement in the transition from opening to task in doctor-patient interviews. In: Philip Glenn, Curtis D. LeBaron & Jenny Mandelbaum (eds.), Studies in Language and Social Interaction: In Honor of Robert Hopper, 411–422. Mahwah, NJ: Lawrence Erlbaum. Mondada, Lorenza 2000: Décrire la ville. La construction des savoirs urbains dans l’interaction et dans le texte [Describing the City. The Construction of Urban Knowledge in Interaction and in Texts]. Paris: Anthropos. Mondada, Lorenza 2005: La constitution de l’origo déictique comme travail interactionnel des participants: une approche praxéologique de la spatialité [The constitution of deictic origo as an interactional achievement of participants: a praxeological approach of spatiality]. Intellectica 2/3: 41–42, 75–100. Mondada, Lorenza 2006: Participants’ online analysis and multimodal practices: projecting the end of the turn and the closing of the sequence. Discourse Studies 8: 117–129. Mondada, Lorenza 2007a: Interaktionsraum und Koordinierung [Interactional space and coordination]. In: Arnulf Depperman & Reinhold Schmitt (eds.), Koordination. Analysen zur multimodalen Interaktion, 55–94. Tübingen: Narr. Mondada, Lorenza 2007b: Deixis spatiale, gestes de pointage et formes de coordination de l’action [Spatial deixis, pointing gesture, and forms of action coordina-

274

Lorenza Mondada

tion]. In J.-M. Barbéris & M. C. Manes Gallo (eds.), Parcours dans la ville: la description d’itinéraires piétons, 261–285). Paris: L’Harmattan. Mondada, Lorenza 2007c: Multimodal resources for turn-taking: pointing and the emergence of possible next speakers. Discourse Studies 9(2): 195–226. Mondada, Lorenza 2009a: Emergent focused interactions in public places: a systematic analysis of the multimodal achievement of a common interactional space. Journal of Pragmatics 41: 1977–1997. Mondada, Lorenza 2009b: Video recording practices and the reflexive constitution of the interactional order: some systematic uses of the split-screen technique. Human Studies 32: 67–99. Mondada, Lorenza 2010: Reassembling fragmented geographies. In: Monika Büscher, John Urry & Katian Witchger (eds.), Mobile Methods, 138–163. London: Routledge. Mondada, Lorenza 2011a: The interactional production of multiple spatialities within a participatory democracy meeting. Social Semiotics 21(2): 283–308. Mondada, Lorenza 2011b: Géographies mobiles et divergentes: l’établissement interactionnel des lieux d’intervention dans des appels au secours [Mobile and divergent geographies: the interactive establishment of dispatch places in emergency calls]. Revue d’Anthropologie des Connaissances 5(2): 390–436. Mondada, Lorenza 2012: Garden lessons: embodied action and joint attention in extended sequences. In: Hisashi Nasu & France C. Waksler (eds.), Interaction and Everyday Life: Phenomenological and Ethnomethodological Essays in Honor of George Psathas. Lanham MD: Lexington Books, 293–311. Mondada, Lorenza & Reinhold Schmitt (eds.) 2010: Situationseröffnungen: Zur multimodalen Herstellung fokussierter Interaktion [Openings: on the multimodal constitution of focused interaction]. Tübingen: Narr. Mondada, Lorenza & Véronique Traverso 2005: (Dés)alignements en clôture. Lidil 31: 34–59. Nevile, Maurice 2004: Beyond the Black Box: Talk-in-interaction in the Airline Cockpit. Aldershot: Ashgate. Pillet-Shore, Danielle 2008: Coming together: Creating and maintaining social relationships through the opening of face-to-face interactions. Ph.D. dissertation, Department of Sociology, UCLA. Psathas, George 1979: Organizational features of direction maps. In: George Psathas (ed.), Everyday Language: Studies in Ethnomethodology, 203–225. New York: Irvington Press. Psathas, George 1986: The organization of directions in interaction. Word 37(1–2): 83–91. Psathas, George 1991: The structure of direction-giving in interaction. In: Deirdre Boden & Don H. Zimmerman (eds.), Talk and Social Structure: Studies in Ethnomethodology and Conversation Analysis, 195–216. Cambridge: Polity Press. Rae, John 2001: Organizing participation in interaction: Doing participation framework. Research on Language and Social Interaction 34(2): 253–278. Relieu, Marc 1999: Parler en marchant: Pour une écologie dynamique des échanges de paroles [Talking while walking: towards a dynamic ecology of verbal exchanges]. Langage and Société 89: 37–67. Robinson, Jeffrey D. & Tanya Stivers 2001 Achieving activity transitions in primarycare encounters: From history taking to physical examination. Human Communication Research 27(2): 253–298. Schegloff, Emanuel A. 1968: Sequencing in conversational openings. American Anthropologist 70: 1075–1095.

Interactional space and the study of embodied talk-in-interaction

275

Schegloff, Emanuel A. 1972: Notes on conversational practice: formulating place. In: David N. Sudnow (ed.), Studies in Social Interaction, 75–119: New York: Free Press. Schmitt, Reinhold 2012: Körperlich-räumliche Grundlagen interaktiver Beteiligung am Filmset [Embodied and spatial foundations of interactional participation on a filmset]. In: Heiko Hausendorf, Lorenza Mondada & Reinhold Schmitt (eds.), Raum als interaktive Ressource. Tübingen: Narr. Stukenbrock, Anja & Karin Birkner 2010: Multimodale Ressourcen für Stadtführungen [Multimodal Resources for City Tours]. In: Marcella Costa & Bernd MüllerJacquier (eds.), Deutschland als fremde Kultur: Vermittlungsverfahren in Touristenführungen. 214–243. München: Iudicium. Szymanski, Margaret 1999: Re-engaging and dis-engaging talk in activity. Language in Society 28: 1–23. Taylor, Holly A. & Barbara Tversky 1992: Spatial mental models derived from survey and route descriptions. Journal of Memory and Language 31(2): 261–292. Wunderlich, Dieter 1976: Studien zur Sprechakttheorie. Frankfurt am Main: Suhrkamp. Zimmerman, Don 1992: Achieving context: openings in emergency calls. In: Graham Watson & Robert M. Seiler (eds.), Text in Context: Contributions to Ethnomethodology, 35–51. Newbury Park, CA: Sage.

276

Heiko Hausendorf

Heiko Hausendorf

On the interactive achievement of space – and its possible meanings1

“Your social situation is not your country cousin” (Goffman 1964: 134).

1.

Preliminaries

The broad issue of language and space, which is intertwined with the issue of interactional space but goes, at the same time, beyond, will not be dealt with systematically in this paper. Instead, I will confine myself to the role of space in face-to-face interaction at the conceptual level of conversational analysis (“CA” in what follows to refer to approaches that have been inspired by the classic CA studies and that have been further developed according to different analytic needs within sociology and linguistics, inside and outside the Anglophone world). While there is a rapidly growing number of empirical studies dealing with interactional space in one way or another (see below), an overall theoretical framework for integrating the different approaches and straightening out the dizzying array of terminology is still lacking. The present paper makes some progress in this direction. It is a position paper rather than an empirically based account. As a starting point, I shall take an approach to space that is typical of CA. Instead of accounting for space in terms of the spatial parameters of a speech situation existing somehow a priori to interaction, space and the speech situation itself are assumed to be interactively achieved. Otherwise, space remains a subject beyond the reach of CA – a subject relevant only to, say, physicists, architects and landscape designers. As such, the idea of a given speech situation disappears (Hausendorf 1995), and along with it the idea of hard-core parameters such as space. “The interactive achievement of space” has accordingly become a sort of slogan implicitly and explicitly stated in a number of concrete empirical analyses that have emerged in recent years. These are connected with catchphrases such as “interactional space”, “multimodality”, “embodiment” or “situatedness” (for recent reviews of the Anglophone literature cf. the relevant handbook articles in 1

Many thanks to Peter Auer for many very helpful comments – and for advising me on the “theory of affordances” by Gibson (1977).

On the interactive achievement of space – and its possible meanings

277

D’hondt, Östman, and Verschueren 2009; Ziemke, Zlatek, and Frank 2007; Frank, Dirven, and Ziemke 2008; Enfield and Levinson 2006; Kecskés and Mey 2008; Norris and Jones 2005; and for further references including the German-speaking literature the contributions in Schmitt 2007; Mondada and Schmitt 2010; Deppermann and Linke 2010). In this paper I raise some of the questions lying behind the interactive achievement of space motto instead of repeating, rephrasing and reemphasizing the point that space does not determine what is going on in face-toface interaction. Take this as a given – and all the interesting questions are still left open. To name but a few: What is it that is interactively achieved when we talk about “space”? What do we mean by “interactive achievement”? Whatever space may be in the end, might we not be well advised to accept that space is already, in a manner of speaking, present at the beginning of interaction? But what could “interactive achievement” then mean? And what kinds of problems do we encounter when we speak about interaction and its “achievements”? How, in detail, does the interactive achievement of space occur? Although there is a language of space (grammar and lexis), there is, obviously, no need to talk explicitly about space in every case. What are the interactive devices and forms (see below) we are looking for when talking about the interactive achievement of space? x x

x

Far from suggesting an overall solution to these questions, I will try out the idea of a broad and open concept of interactional space that is bound to the co-participants’ emplacement (Streeck, Goodwin and LeBaron 2011) within a concrete speech situation. I believe it is the concrete “here” rather than the abstract space that matters. The kind of problem that we are faced with when we talk about interactional space is then, first of all, one of achieving a mutually shared “here” for perception, movement and action. I will introduce this problem as one of situational anchoring that face-to-face interaction is in principle confronted with: interaction has to situate itself. It is one of the genuine problems of face-to-face interaction and it can be split up into the sub-problems of co-orientation (referring to perception), co-ordination (referring to movement) and co-operation (referring to action). The often naively assumed “speech situation” can then be reconstructed as the participants’ solution to these problems. According to CA methodology, problems can be taken as interactional jobs or tasks, the solution to which can be reconstructed in terms of devices, i.e. the different ways or methods of doing the job, and forms, i.e. the concrete audible and visible manifestations of devices at the surface level

278

Heiko Hausendorf

of discourse. This framework of jobs, devices and forms (Hausendorf and Quasthoff 1992, 1996) will be introduced to avoid conceptual confusion when talking about interactional space, namely the kind of confusion that arises from the idea that space is not only an interactional achievement but also an important interactional resource. To elaborate on this idea – space as resource and achievement – I suggest replacing the space-as-achievement motto (see above) with the terminology of co-orientation, co-ordination and co-operation jobs and their devices and forms. We can then systematically account for space as a resource that is used when fulfilling the task of situational anchoring with its subtasks of co-orientation, co-ordination and co-operation. In contrast to bodily and verbal resources that have been largely dealt with in recent research (on embodiment and deixis, for instance, see above), the spatial resources have often been more or less neglected. The general point is to come to terms convincingly with details of theory, methodology and empirical data while taking seriously the matter of interactive achievement. Of course this holds true for more than just space or situation. It generally concerns the acknowledgement of phenomena such as natural language, the participants’ bodily and cognitive facilities or social “macro” structures (to name only what is obvious) as resources for face-toface interaction. Note that it hardly seems possible to imagine anything that has not been claimed to be “interactively achieved” in discourse. Nearly all aspects of the “world” can gain the status of interactive achievement (hierarchy and culture, gender and ethnicity, institutions and social belongings, contexts and situations, etc.). CA research of the last three decades or so is – among other things – an attempt to widen our scope of phenomena that are interactively achieved – and to somehow disregard the scope of phenomena interactively made use of. The statement of interactive achievement is, therefore, by no means exclusively relevant to space. But space – in opposition to time (“sequentiality”), which has been our predominant concern in CA – seems to be an elusive concept, given that our CA-trained point of view remains a language-oriented one, at least for those of us applying it for linguistic purposes. I hope this will become less abstract as we proceed. Before turning to possible meanings of interactional space in the next section and going into the details of the problem of situational anchoring thereafter, let me include a side note on previous research traditions. Ever since early CA research discovered the analytic prospects of recording and transcribing spoken discourse (telephone conversation, strictly speaking) which is of great value in reconstructing the details of verbal sequentiality (cf., for instance, the contributions in Hausendorf 2007), the visual manifestations of face-to-face interaction that had been previously studied based on video re-

On the interactive achievement of space – and its possible meanings

279

cordings within so-called context analysis (Kendon 1990a, first 1973) have to some extent been lost from sight. It is not by chance that this analytical preference for verbal data brought with it a long-lasting data preference for audio instead of video recordings – which has only recently been reversed. Nevertheless, we do not have to start from scratch when turning to the issue of situational anchoring. There is a great deal of research within this “context analysis” tradition (Kendon 1990c) that has already provided evidence for co-orientation, co-ordination and co-operation. An outstanding example is Kendon’s impressive study of greeting sequences at a private outdoor party which provides strong evidence for bodily interaction (partly described in terms of “co-orientation” and “co-ordination”) that starts long before the first verbal greeting pair (“close salutation”) appears (Kendon 1990a). Apart from details and terminology, the question is: Are we actually beyond these early studies? And if so, what is new and what is the substantial progress we have made or could at least claim to have made when returning to old wine in new skins? My answer is threefold: To begin with, there is new technology, i.e. electronic data of digitalized video recordings that allow for computer aided simulations and presentations of face-to-face interaction. It was the then-innovative technology of videotape recorders that were available to everybody that triggered early interaction studies and allowed for new insights. Today, it is the innovative technology of electronic video recordings that will surely trigger new types of data and, along with new forms of (re)presentation, also fresh concepts of situation. This process has already started, but, interestingly enough, it has started outside our research (namely within AI-research: cf., for instance, Pfeiffer 2010a, 2010b and other research projects conducted at the Bielefeld Center of Excellence on Cognitive Interaction Technology (CITEC): http://www.cit-ec.de/). Second, there is methodology. What we have learned about the structural properties of sequentiality on the basis of audio recordings has to be adopted when studying multimodality based on video recordings (Schmitt 2006). In contrast to what at first seems evident, the speech situation is a strictly dynamic phenomenon emerging step by step from manifestations of co-orientation, co-ordination and co-operation that have to be accounted for empirically. Accordingly, the simultaneity of what is visible has to be transformed in terms of what has been made visible among the participants sequentially (Schmitt 2006). Furthermore, there are spatial resources which co-orientation, co-ordination and co-operation can attach to and whose methodological implications have been scarcely taken x

x

280

x

Heiko Hausendorf

into account up to now (cf., for instance, Emmison and Smith 2000; Kissmann 2009). Last but not least, there is theory, i.e. a theory of face-to-face interaction as a genuine social reality. In the end, interactive achievement depends on the assumption of interaction as a system of problem solving and multitasking. What we are heading for in research on interactional space is no longer some sort of human ethology (which context analysis might have had in mind: cf., for instance, Kendon 1990a: 201) but a theory of faceto-face interaction as a particular manifestation of communication.

The last argument is the only one I will return to at any length in this paper, while the other two arguments (technology and methodology) will not be followed up here (except for spatial resources). I have chosen to resist the temptation of providing empirical evidence and picking out some potentially illustrative pieces of data, as this would remove me too far from the theoretical paper I am aiming at.

2.

What does space represent in CA research?

What exactly does the notion of space refer to when the interactive achievement of space is discussed? It is striking what CA and related research includes under “interactional space” or “spatial” even at first glance. In addition, there are manifold overlaps. Incidentally, the notion of interactional space is sometimes also used in a metaphorical way to refer to the participants’ rights and obligations (for instance, with regard to the regularities of courtroom interaction: Adelswärd, Aronsson, Jönsson, and Linell 1987). In contrast, our concept of interactional space is to be taken literally. Space then can be what is directly accessible to the participants’ sensory perception, i.e. what is visible, audible, can be touched … (perceived space: something just here, over there …) what is available to the participants’ body movements, i.e. what is within reach, “stand-on-able”, “walk-on-able” (Gibson 1977, 1968), go-throughable, pass-by-able … (used space: a line of seats, a passage, a pedestrian area …) what is known as a social group’s territory (ideological or imagined space: a nation, a state, a principality …) what is prepared for socially organized (institutionalized) and highly specialized use: built and furnished space (buildings like a sport stadium, a hospital, a court, a university and their interiors) x

x

x

x

On the interactive achievement of space – and its possible meanings x

x

x

x

x

281

what is topographically defined as a distinct space and/or place (named space: towns, cities, a continent, a street, a place) what is geographically outlined by means of cartography (measured space: borders between territories, parts of the earth) what has naturally emerged during the last ice age or thereabouts (formed space: mountains, valleys, deserts, plains, lakes, glaciers …) what is linguistically defined through ways of speaking different languages or dialects (spoken space: a dialect region, a linguistic area) etc.

It goes without saying that this list is not exhaustive in any sense. But it illustrates the range of phenomena being talked about when we use a term like space. Presented in this way, our topic is obviously an interdisciplinary one relevant to many disciplines and approaches: geology (formed space) and geography (measured space), cognitive, social and ecological psychology (perceived space), ethology (used space), political sciences and sociology (ideological and imagined space), architecture (built and furnished space), and, last but not least, linguistics (named and spoken space). According to the interdisciplinary variety of perspectives, space has been dealt with rather differently as far as theory and methodology are concerned. Take, for instance, the ways in which space has become a topic within linguistics when the relation between space and language is discussed. Roughly speaking, there are three different ways to account for this relation:

Fig. 1. Space as a topic in linguistics

To begin with, one could argue that language is bound to space and that it is a phenomenon that exclusively appears in space. We then talk about speaking and listening, spoken (and heard) discourse, and orality. Space in spoken discourse can be something very small (a corner within a room, i.e. a part of built and furnished space) and something very big (a linguistic area as a part of spoken space). As a part of spoken discourse, it has traditionally been

282

Heiko Hausendorf

studied within area and/or contact linguistics and dialectology, often as a matter of linguistic variation (cf. the recent handbook by Auer and Schmidt 2010 and other studies indicating a renewed interest in concepts of space within this research: Muysken 2008; Bickel and Nichols 2006). Conversely, one could argue that space is something that appears within language. We then talk about spatial language (Hayward and Tarr 1995): grammar and lexis, grammatical and lexical structures of natural languages, grammaticalization and lexicalization pathways in which spoken spaces have found their fixed and solidified forms, so to speak. Named space is an obvious outcome of this kind of space within language. Different “grammars of space” illustrated by cross-linguistic research (cf. Levinson and Wilkins 2006) provide further evidence of this kind of space within language. A lot of research on deixis has followed this direction (cf. the recent review by Sidnell 2009). Finally, there is a third way of dealing with language and space which states that space is, first of all, a social construct emerging from discourse. It is from this point of view that the language and space issue could benefit from what has been called a “spatial” and “topographic turn” in social sciences or a “sociology of space” (cf. Löw 2001; Schroer 2007; BachmannMedick 2006; Döring & Thielmann 2008). We then talk about language in terms of its contribution to discourse strategies of constructing certain views of space and spatiality. Take, for instance, political and regulatory discourses of landscaping and town planning and how they shape our view of urban and rural environments, of built and furnished space. Space often becomes a contested concept within such discourses, a topic of negotiation and deliberation (cf., for instance, Backhaus, Reichler & Stremlow 2007; Richardson & Jensen 2003 and the review by Sidnell 2009). At a theoretical level, there is no reason to restrict our concept of interactional space to one of these approaches and therefore it is better not to exclude certain types of spaces or restrict ourselves to selected types of spaces by means of extant definition (stating, for instance, that there are “physical” spaces e.g. formed and built spaces as opposed to “interactional” ones). Otherwise, we would miss some interesting connections between the different approaches and between the disciplines. We would, furthermore, skip over many ways in which all of these spaces are interactively achieved. But as a consequence of this, we will need a stand-alone concept of space suited to containing all these possible (perceived, used, built, formed, spoken, named …) spaces that, at the same time, allows for the elaboration of the connection between space(s) and face-to-face interaction. Perhaps the wisdom hidden within ordinary language can help us. Consider the deictic expression here and its use in everyday conversation. It is

On the interactive achievement of space – and its possible meanings

283

easy to imagine (and to illustrate empirically) that it could refer to each of our spaces: If we think of the unlimited variety of locations here can refer to in spoken (and written) discourse, we end up with a list of spaces like the one given above. Space, then, stands for the possible location of an interactional episode: Where did it happen? I would like to suggest that this is the place where “spatiality” makes its appearance, in whatever dimension you like, or more precisely, the participants like. Spatiality is then a part of the origo (to use Bühler’s term) of the speech situation, or, in line with Goffman, it is a crucial aspect of co-presence. Co-presence, i.e. an interactive “we”, implies space and spatiality, i.e. an interactive “here”. As such it must be interactively achieved. It is in no way predetermined where “here” is, but is up to the participants to make clear in one way or another which kind of space is to be relevant for their interactive “here”. As an aspect of co-presence emerging within and by means of face-to-face interaction and comprising what could possibly be referred to by using the word here, space, indeed, appears to be an interactive achievement. That is at least a start, but not one that gets us very far – unless we turn to the meaning of the matter of interactive achievement itself. This involves the relation between space and interaction while most previous linguistic research on space and language has focussed on the relation between cognition and space (“spatial cognition”: Levinson 1996: 356; Peterson, Nadel, Bloom & Garrett 1996; Levinson 2003; Hickmann & Robert 2006), which is a subject in its own right.

3.

What do we mean by interactive achievement of space?

What exactly are we referring to by interactive achievement: a creatio ex nihilo, some sort of reproduction, a communicative construction? Who or what lies behind the achievement: Who is achieving something? Questions like these make clear reference to the general theoretical background of CA. But if one wants to make inroads into the matter of interactive achievement, there is no way to avoid entering into some sort of theory of interaction. Let us start by considering the way a sociologist like Goffman goes about studying the issue of face-to-face interaction. For Goffman, as is well known, face-to-face interaction is “a little system of mutually ratified and ritually governed face-to-face action” arising “whenever two or more individuals find themselves in one another’s immediate presence” (Goffman 1964: 135). From this it follows that co-presence is not a condition that exists externally or preceding the social situation, but is achieved through the perception of being perceived by others (Hausendorf 2003) – an important insight

284

Heiko Hausendorf

that I will return to later on. This is the moment when something starts that goes beyond the participants displaying meaning and understanding: a little system of its own, a social reality sui generis. We must keep in mind that when speaking about interactive achievement, we are referring to a social system (Luhmann [1972] 2005). Much of Goffman’s work is devoted to the systematics of this social system, which has been dismissed to some extent in more recent research. Consequently, there is a lack of a well-grounded theoretical framework of interaction as a genuine social reality – even when “situated human interaction” is explicitly set on the agenda. Take, for instance, Goodwin (2000), who emphasizes “situated activity systems” as a relevant “environment” to investigate “human action, cognition and talk-in-interaction” (2000: 1519). What is then the proper subject of such research: “human action”, “cognition” or “talk-in-interaction”? Or the “situated activity system” itself ? For this reason, we need some sort of theory of interaction. One particularly interesting aspect is the idea of face-to-face interaction as a multitasking system, triggered by a set of interactive problems. I argue that for the study of interactive achievement, these problems are best understood as tasks that face-to-face interaction is confronted with. We can, then, return to our space-as-here statement (see above, section 2), in order to explore this and to understand how space (in all its facets) can make a difference in interaction. A long tradition of empirically fruitful CA and related research makes it easy to give a picture of what can be counted as essential interactive tasks. The following figure uses questions to indicate the kinds of problems that have to be “solved” simultaneously and, correspondingly, the kinds of tasks that have to be completed simultaneously (see Fig. 2, the “it” in some of the questions refers to interaction). The system is meant to hold true for every episode of face-to-face interaction. Strictly speaking, it is by solving these problems that interaction actually comes into being, which is related to the status of interaction as a self-referential, “autopoietic” social system (Luhmann 1984). This is not to say that the empirical manifestation is in each and every case immediately obvious. This manner of thinking about interaction in terms of problem solving will be familiar to CA-trained readers. Conversation as problem solving is a concept introduced right from the beginning of CA research. For instance, Schegloff & Sacks (1973) discuss the closing problem. Cf., furthermore, the conceptual framework offered by Kallmeyer and Schütze (1976). The general idea of interaction as multitasking can be found in Goffman (1977), although Goffman did not come up with a list of concrete tasks. The empirical phenomena behind these tasks can easily be related to the standard themes

On the interactive achievement of space – and its possible meanings

285

Fig. 2. Interaction as a system of multitasking

of CA research and have often been noticed (cf., for instance, Auer 1986: 27 with a similar list) but scarcely developed as genuine problems of face-toface interaction: opening up and closing of interaction; technically speaking: the achievement, maintenance and termination of co-presence (When does it start? When does it end?), turn taking (Who comes next?), the organization of contributions to conversational topics (What comes next?), contextualization and framing (What is going on?), presentation of self and others and social categorization (Who are “we”?) and, last but not least, situational anchoring (Where is “here”?). x

x x

x x

x

I suggest that these tasks refer to genuine interactive problems that must be solved in whatever way possible when interaction comes into being. My theoretical argument is: It is precisely the handling of these problems within and through interaction that allows us to talk about co-presence of participants, turn taking, coherence of contributions,

x x x

286

Heiko Hausendorf

context of utterances, situation of speech, and self and others as interactive achievements. Much more could and perhaps should be added. But the issue I wish to address does not only concern these interactive tasks in general. Since our interest is in interactional space, we had better move on to the task of situational anchoring (“Where does it happen?”). I would like to propose replacing the ‘interactive achievement of space’ with the task of situational anchoring. The benefit is: Situational anchoring is a genuinely interactive task which allows us to study the different resources participants can make use of when doing this job. We will see that space, in all of its relevant aspects (see above, section 2), will indeed re-enter the scene as a powerful and complex resource of coorientation, co-ordination and co-operation (see section 5 below). The task of situational anchoring has been much less investigated than the other tasks. Take, for instance, the who/what-comes-next-problems, and consider the way in which these time-in-interaction problems have been successfully conceptualized in terms of sequentiality. Although there is the honourable tradition of “context analysis” (see above, section 1), there can be no doubt that we are still searching for an equally powerful conceptualization of the problems of space-in-interaction. Along with the task of situational anchoring, space-in-interaction has long been neglected. Before exploring the task of situational anchoring and its subtasks in some detail, it is important to be aware that space is intertwined with at least some of the other problems, too. Here are just a few examples: Clearly, space can become a topic of discourse, since it is explicitly talked about (by means of spatial lexis and spatial semantics: “space within language”, see Fig. 1). It is, as such, a part of the organization of topical talk (what comes next?), which may – but not necessarily – overlap with the task of situational anchoring. Space can of course become a cue and a resource for the participants’ sense of social belonging and identity (e.g. in terms of origin, country and home). It is, then, a part of the presentation of self and others (who is participating?), perhaps (but not necessarily) overlapping with the task of situational anchoring. Space can become a powerful resource to localize an interactive episode within a certain institutional setting (think of a purpose-built space such as a court). It is, then, a part of the task of framing and contextualization (what’s going on?), possibly overlapping with the task of situational anchoring.

x x x

x

x

x

On the interactive achievement of space – and its possible meanings

4.

287

What are the devices and forms of situational anchoring?

Having transformed the interactive achievement of space into the task of situational anchoring, the question still remains: How is situational anchoring fulfilled empirically? If we want to get more than an impressionistic picture, we need to explore the task of situational anchoring in more detail. I would therefore like to suggest that it is in itself a complex task consisting of three different subtasks (Fig. 3). It follows that the concrete devices and forms (see above, section 1) of situational anchoring must be related to these subtasks. The threefold distinction between the subtasks separates sensory perception, bodily movement and social action. These are to be understood as the relevant dimensions of situational anchoring. A common here must be achieved for all three. There is a participants’ here of perception (Wahrnehmungsraum: Kruse & Graumann 1978: 179) that could be studied in terms of different modes of sensation (cf. the analysis of classroom interaction by Breidenstein 2004 showing that there are different spaces according to visibility, audibility and touchability partly overlapping and partly dispersing). There is a participants’ “here” of locomotion (treated as being part of the Wahrnehmungsraum but nevertheless accounted for separately by Kruse & Graumann 1978). Furthermore, there is a participants’ “here” of action (Handlungsraum in Kruse and Graumann’s terms). Corresponding to these three types of “here”, there are three types of subtasks: A task of co-orientation, a task of co-ordination and a task of co-operation. Particularly the terms “co-orientation” and “co-ordination” are widespread in both earlier and more recent studies (Kendon 1990a: 160 talks about an “orientational frame”; cf. Deppermann & Schmitt 2007 for an explicit account of the notion of “co-ordination”). But they generally do not refer to interactional tasks in the strict sense introduced here (section 3) and they are generally not related to the distinction of perception, movement and action. It is not by chance that “co-” prefixes each task: We are talking about genuine interactive problems, which are different from the types of problems a single participant might be confronted with who is, of course, steadily perceiving, moving and acting. To start with, the “co-” implies that the individual must do this together with others doing the same: That calls for joint attention, joint movements and joint action. The “joint” in joint attention, joint movements and joint action implies the emergence of interaction as a social reality sui generis. The “co-”, therefore, means that perceiving, moving and acting become interactive forms, i.e. manifestations of interaction itself. Again, the question is: How does it happen? Perhaps it is best to start with perception and co-orientation. There can be no doubt that co-orientation is of central concern for interaction. It refers to

288

Heiko Hausendorf

Fig. 3. Problems of situation

sensory perception as perhaps the most important medium of interaction. It is the moment “when two or more individuals find themselves in one another’s immediate presence” that makes up that little system of interaction (Goffman 1964). And both finding each other in their immediate presence, obviously, depends on mutual perception, on indicators by means of which participants know that they have seen each other and, in doing so, become “present” (cf. Kendon 1990a: 153). More precisely, it is a mechanism of perceived perception that lies behind interaction (Hausendorf 2003). This was Goffman’s initial argument. It is theoretically worked out in the framework of the social-systems approach of Luhmann (1984: 560). In this approach, interaction is taken to be a particular type of social system, i.e. a particular manifestation of communication, namely communication under the terms of co-presence (cf. Hausendorf 1992; Kieserling 1999). Independent of the sociology-of-interaction tradition that Goffman has made popular, the matter of perceived perception has recently become an issue for developmental and cultural psychologists as well as primatologists under the heading of “joint attention” taken to be a manifestation of social cognition (cf., for instance, Bruner 1995; Dunham & Moore 1995; Tomasello 1995; see below, this section). Another recently discussed issue is that of “presence” in virtual or online and hybrid environments (cf., for instance, Davenport & Buckner 2005; Turner & Davenport 2005). What becomes very clear from these approaches is that presence is not a physical fact but a social construct, interactively achieved by perceiving that one has been perceived. This holds for the opening of interaction which starts (sometimes long) before the first hello is spoken and heard (as shown, for instance, by Kendon 1990a and taken up in some of the contributions in Mondada & Schmitt 2010). The most obvious manifestation of perceived

On the interactive achievement of space – and its possible meanings

289

perception is, of course, eye contact. This is the reason why looking each other in the eye is risky: as soon as eye contact goes beyond the fleeting fragments of seconds of random gaze, interaction as a genuinely social phenomenon starts and can no longer be ignored (whatever might be done afterwards). Perceived perception is the basic mechanism for achieving presence. It must be maintained throughout the exchange although the participants, of course, need not maintain eye contact all the time. In many cases, there is a minimum of perceived perception (Hausendorf 2003: 258; sometimes also referred to as “common ground”: LeVine 2007; Enfield 2008) that is taken for granted unless there are clear indicators of a participant’s denial of attention (for instance, someone falling asleep). Another type of problem with perceived perception emerges when there is reason to assume that the visà-vis is not an intentional agent so that attention cannot be considered intentional perception (maybe in the case of mental disease or young infants). From this point of view, it becomes clear that co-orientation and joint attention do not necessarily coincide with each other, as stressed in research on joint attention during child development (cf. the contributions in Moore and Dunham 1995; see below). In some cases, co-orientation indeed goes beyond this minimum of taken-for-granted attention. These are the cases in which the co-orientation task and, along with it, the situational anchoring task come to the fore. Note that sensory perception is not an interactive phenomenon per se. Accordingly, not everything that participants can and do perceive while engaged in face-to-face interaction becomes relevant for the ongoing interaction. On the contrary, it is a rather small selection of what is perceived by participants that achieves the status of an interactive relevant phenomenon. Provided that the participants stay within earshot, the perception of what is spoken can be taken for granted (although not really perceived in itself). This is one of the great achievements the evolution of natural language has brought about. Interaction can, in fact, be restricted to the perception of spoken language (as is the case, for instance, in telephone conversation). But there are cases in which the participants’ visual, tactile and even olfactory perception can become a relevant part of what is going on; take, for instance, an exhibition of objects and artefacts. The question then is how perception can become an interactively reliable phenomenon, i.e. a manifestation of interaction. This is exactly the problem of co-orientation. And it also holds for co-ordination and co-operation as far as bodily movement and action are concerned. Co-orientation, co-ordination and co-operation often go together. One could argue that co-operation is the most demanding problem since it

290

Heiko Hausendorf

requires at least co-orientation and, in many cases, co-ordination too. Since co-ordination also requires co-orientation, co-orientation is basic and has to be assumed in all face-to-face interaction – even in “unfocused” interaction (Goffman 1964). Without it, there is no interaction. But the matter is not as simple as it may sound. There is, for instance, good reason to assume that co-orientation, co-ordination and co-operation can fall apart. Goffman’s examples of “unfocused” interaction may be telling cases in this regard. They show how interaction can be restricted to the perception of being perceived – without ensuing manifestations of co-ordination and co-operation. Fleeting encounters between pedestrians (“an exchange of glances between strangers as they pass on the street”; Kendon 1990a: 153; for a detailed account see Ciolek & Kendon 1980) or gatherings in waiting rooms or elevators show that there is interaction consisting of a minimum of reduced shared attention – a potential for co-ordination and co-operation that can but need not be exploited by those present. One could argue that even these reduced encounters are manifestations of social practice, i.e. achievements of co-operation. But the co-operation in this case would consist solely of avoiding further co-operation. Another highly telling case is interaction between adults and young children, i.e. cases in which co-operation is still being developing and learned. There is evidence of some reduced forms of co-orientation (namely in terms of following an adult’s gaze) before joint attention (Tomasello 1995: 105). In other words, co-orientation could be regarded as a vehicle for developing co-ordination and co-operation. Additionally, there seems to be “pure” co-orientation and co-ordination among non-human primates (Tomasello, George, Kruger, Farrar & Evans 1985), which raises the question of what further assumptions are implied in joint attention. Tomasello (2000) argues that it is, first of all, the assumption of intentionality. This argument goes with our understanding of face-to-face interaction as communication, i.e. participants make a distinction between the information that there is something interesting enough to be looked at and the message that there is something alter wants to show ego. This argument follows Luhmann’s approach (1984: 193): Drawing a distinction between information and message (“Mitteilung”) is the sense of understanding (“Verstehen”) and the beginning of communication (starting from alter, so to speak). Linguists typically have an issue with this definition since they are trained to think of communication from the beginning, namely a speaker’s (ego’s) intention to communicate. Luhmann’s point is: It is alter who has to make an assumption of intentionality, i.e. draw a distinction between information and message. Whenever s/he can be expected to do so, communication starts whether there has been some foregoing intentionality or not. Perhaps this is an

On the interactive achievement of space – and its possible meanings

291

achievement that depends on conditions we have not yet sufficiently examined in our research. So much for the relation between the three subtasks. Some words, then, about the devices and forms (see above, section 1) of co-orientation, co-ordination, and co-operation are in order. The following cases make clear the direction in which we have to look. To begin with, co-orientation is primarily achieved in terms of mutual perception – strictly speaking, in terms of perceived perception. It is, in principle, the task of making perception perceptible in itself. A prototypical solution for this requirement is a verbal deictic expression like this accompanied by finger pointing. In line with recent approaches to deixis from very different starting points (cf., for instance, Kleiber 1983; de Mulder 1996; Hausendorf 2003; Weinrich 2005: 444), one could argue that the forefinger does not point but displays an act of seeing. The forefinger of the outstretched arm can accordingly be considered a kind of visualized gaze (Hausendorf 2010). It therefore does not come as a surprise that finger pointing is essentially improper for identifying objects, as recently suggested by experimental studies in AI (Pfeiffer 2010a). Both finger pointing and a simultaneous verbal deictic expression contribute to co-operation in transforming an individual act of visual perception into an interactive move. It is an intention of letting the other take part in perception that is communicated by finger pointing as well as verbal deixis (“joint attention” in the sense of Tomasello 1995). Deixis is therefore a device of co-operation which can take verbal as well as nonverbal forms. Another less striking nonverbal device of co-operation is the participants’ joint alignment of the direction of gaze: Ostentatiously orienting one’s own attention in the same direction as the attention of others, perhaps towards something in front and below, might suffice to establish co-orientation. The message is then: There must be a relevant something in front and below – otherwise there would be a small interactive scandal. As mentioned above, co-ordination usually implies co-orientation. But it goes beyond it. Note that participants are sensory systems which are locomotive, i.e. free to move and change places and positions. Participants’ locomotion is a potential problem since it principally endangers co-presence and co-orientation: For those sitting around a table and bound to stay fixed on a chair it is easier to maintain shared attention than for those in motion (e.g. walking around in an exhibition hall). Generally speaking, the problem is to stay together, to keep alive the mutuality of co-presence when in motion. The requirement is one of pack behavior as Goffman was not afraid of putting it (1974: 19), and it has to be learned. Moving around takes a lot of attention and perceptive capacities so that those in motion cannot maintain eye con-

292

Heiko Hausendorf

tact all the time and therefore cannot easily take for granted their being together. Accordingly, it takes more than perception to move around together with others, it requires locomotion that demonstrates its mutuality. Obviously, distance matters in this regard. We therefore consider “proxemics” (Hall 1976; cf. also Wegner 1985) a solution to our problem of co-ordination: Keeping up a distance of reach and staying within each other’s range of perception (for instance, staying in earshot) in a stable and lasting, non-random way is an effective way of showing that a group is moving together. Holding hands or having arms around each other while walking is also a clear, but rather ambitious, case since it depends on intimacy (albeit depending on cultural norms). The synchronization of movements is another phenomenon that could be mentioned here: Participants “doing the same” (walking in the same direction, turning round in the same direction, staying at the same place) are often considered to be “together”. Synchronization, however, is not only an effective but also a very demanding answer to the co-ordination problem. It requires intense training in some cases, e.g. a dancing couple (Loenhoff 2003; Müller & Bohle 2007) or a team at work. Furthermore, it is often supported by verbal directioning or by reassuring eye contact. All these phenomena are well known and my only point here is to say that they can be related to a problem of co-ordination as a relevant part of the situational anchoring task. While the problem of co-orientation is related to perception, the problem of co-ordination is related to locomotion. Co-operation, finally, depends on co-ordination. It goes beyond it in that co-operation entails the job of achieving, maintaining and dissolving formations and configurations best suited for certain types of joint social practice. Staying face-to-face is perhaps the most prominent formation of this kind (Kendon 1990b; Ciolek & Kendon 1980), which gave a name to interaction as a genuinely social phenomenon (“face-to-face interaction”). But it is only one example among a world of possible configurations depending on the type of social action in question (side by side, face-to-back, back-toback …) and depending on the type of environmental circumstances the formation system is embedded in (cf., with a fine-grained differentiation of degrees of intimacy and distance, Ciolek & Kendon 1980). Formations and configurations become more challenging the more participants are engaged in the encounter (formations in circles and semi-circles, formations in queues …). Guided tours provide good examples for studying such formations (cf. Mondada 2007; Kesselheim 2010; Stukenbrock & Birkner 2010). To sum up, formations (or configurations as termed by Vom Lehn und Heath 2007) are manifestations of co-operation, i.e. some sort of social action or Sprachspiel participants jointly contribute to. Often, seeing a

On the interactive achievement of space – and its possible meanings

293

formation allows us to draw conclusions and inferences as to the kind of social activity that is going on – without hearing a single word. Through co-operation with its formations and configurations, sequentiality enters the scene: The simultaneity of co-orientation and co-ordination (synchronicity!) is transformed into the sequential order of turns, verbal as well as bodily, with its well-known implications for “turn taking”. This is the moment when language, too, makes its striking entrance: One might argue, in ontogenesis as well as phylogenesis, that developed forms of sequentiality depend on natural language, or precursors of speaking.

5.

What are the resources co-orientation, co-ordination and co-operation can draw upon?

In this section, I will argue that there are at least three types of resources for situational anchoring: There is co-presence providing bodility and cognition, there is natural language providing highly conventionalized and highly structured linguistic forms, and there is space providing a rich variety of environmental “affordances” (Gibson 1977). While bodily and cognitive resources as well as verbal resources have been described extensively, spatial resources have often been neglected. I will therefore concentrate on the latter and will only briefly turn to bodily and cognitive and verbal resources. Let me start off with co-presence. As Goffman puts it, “the natural home of speech is one in which speech is not always present” (Goffman 1964: 135). Not only turn taking, but particularly the tasks of situational anchoring can be done without speaking (and listening). As was sketched out above, situational anchoring depends, first of all, on the participants’ co-presence in the full sense, including sensory activities, bodily movements and spatial cognition. This does not require speaking and listening (imagine, for example, a dancing couple or a fist fight). In this sense, the participants’ “here” is basically a perceived and moved-through “here”, accessible to the participants’ bodies and minds. Co-presence is obviously a major resource to fulfill the situational anchoring job in all dimensions. Humans are, as Luhmann says, the “sensors” of the interaction system (Luhmann 1984: 558) – and highly developed mobile and intelligent sensors at that. Natural language is another powerful resource for the task of situational anchoring. The grammaticalization of spatial parameters in the world’s natural languages gives an impression of how natural language contributes to coorientation and co-ordination by providing time-tested navigation aids for joint sensory perception concerning basic parameters such as “above” and “below”, “right” and “left”, “in front” and “behind”, “close” and “distant”.

294

Heiko Hausendorf

This point is well described in cross-linguistic research (Levinson & Wilkins 2006). With regard to face-to-face interaction, for instance, LeVine (2007: 264) shows how place reference fulfills “navigational” functions in conversation, i.e. helps “to verify that perceptions are shared by others”. The semantics of the spatial lexicon additionally gives an impression of how known, remembered and imagined spaces can become a part of the participants “here” beyond sensual perception. One could argue that routine answers to everyday challenges of co-orientation, co-ordination and co-operation have become solidified linguistic forms. In this sense, the participants’ “here” is often not only a perceived “here” but also an already known “here” depending on the participants’ familiarity with and knowledge of discourses on space(s) and place(s). Natural language is therefore a powerful resource for situational anchoring. Finally, there is space as another resource that participants can utilize. Notwithstanding the interactive achievement of space, we must not think of the situation as being achieved on some kind of tabula rasa. One could say that prior communicative events (including face-to-face episodes of interaction) have left their traces in spatial environments which, in turn, have become independent. This is anything but new (cf., as a brief review, Keating 2000, 2006). Goodwin (2000) discusses “material structure in the surround” as “semiotic structure without which the constitution of particular kinds of action […] would be impossible” (1492) or, later on, as “semiotic structure provided by the historically built material world” (1517). The question then is how “the mix of semiotic fields” (1517) that is found in the data could be accounted for theoretically (see below). Take, as another example and with regard to the role of materials, Frers (2009) who studies a patient’s file as a “significant factor” in doctor-patient interactions. Further evidence for the role of space, objects, materials and technology can be found within the workplace studies tradition (cf., for instance, Hutchins & Palen 1997; Hindmarsh & Heath 2000) and in approaches towards a “sociology of the seen” (Emmison & Smith 2000: 8; Frers & Meier 2007). Summing up these different approaches to material structures in the environment, space might appear to be a text that can be and will be read by the participants as part of their everyday life competence (Crabtree 2000; Emmison & Smith 2000: 152; Wildgen 2007). Note that the reading of space starts as soon as the participants enter the scene: By taking a seat, following a walkway, opening a door, looking at a picture in an exhibition hall, or resting on a park bench. In some sense, perceived space is, in reality, read space although it may be free of texts. Of course, there is written space as well (“language in the material world”, Scollon & Scollon 2003), i.e. graffiti on walls,

On the interactive achievement of space – and its possible meanings

295

traffic signs or notes posted (“Post-its”). But the reading of space does not depend on texts. Think of built space, such as a city, a church, a court, an auditorium, and how its “understanding” will be displayed by means of walking, sitting, resting and looking, i.e. by means of perception, movement and action. This sort of spatial semiotics makes it clear why situational anchoring and framing can go together in an extremely effective manner. It is important to consider these spatial semiotics when we look for the participants’ “here”. Otherwise, we will not understand how situational anchoring can be fulfilled so effectively, so inconspicuously and so economically as we know it from our familiarity with the everyday routines of face-to-face interaction. Situational anchoring makes use of what is provided in the spatial environment, i.e. what goes beyond the on-board means of face-to-face interaction and lasts independent of co-presence. This holds true for basic affordances of the natural environment as well as for affordances due to humankind’s alteration of the environment, from the invention of tools, objects (such as furniture) and technology to modern architecture of built space. A description of what space as a resource affords the participants can be given in terms of usability: walk-on-ability, stand-on-ability, go-throughability, climb-on-ability, sit-on-ability, look-at-ability, take-hold-of-ability (Gibson 1977: 68). In this sense, usability relates to the participants’ perception, locomotion and action. What space as a resource provides is a complex set of “usability cues”. I want to introduce this term analogously to the term “readability cues”, which are cues that allow readers to identify aspects of textuality while reading (Hausendorf & Kesselheim 2008). According to this framework, a given text consists completely of readability cues. Apart from what the term “cues” might suggest, readability cues cannot be understood as devices added to the “body” of text. The body of text is itself nothing more than a complex set of readability cues. In a way, readability can be understood as a highly specialized (and evolutionarily late) achievement of usability (although the normative notion of “usability” did not become popular before the modern design of “user interfaces”, Norman 2000). Analogous to readability cues, usability cues cannot be restricted to explicit instructions (such as Keep off the grass) but relate to the spatial affordances themselves. They are similar to what gestaltist Kurt Lewin introduced as Valenzen within the social psychology of Raumaneignung (cf. the notion of Aufforderungscharakter räumlicher Gegebenheiten, Kruse & Graumann 1978: 184) and what is the origin of the concept of affordances in ecological psychology (Gibson 1977: 77–78). Note that it is the Aufforderungscharakter that counts and not what is spatially “given”. The former is a social entity, the latter may be a material one.

296

Heiko Hausendorf

Spatial usability cues are a powerful resource for all dimensions of situational anchoring. They provide links for co-orientation, i.e. attractors for looking and hearing. Take, for example, the typical furniture in a classroom which establishes a focal point in terms of a “front”. Or take the typical “white cube” arrangement of works of art (O’Doherty 1996), which strongly supports the perception of objects as works of art (Hausendorf 2010). Spatial usability cues provide links for co-ordination: preferences for walking and for keeping off, for instance, or prepared and fixed ways as in the check-in line at an airport. Spatial usability cues, finally, provide links for cooperation. An example is institutional buildings, such as a court: Its fixed spatial positions are at the same time social positions that provide rights and privileges for turn taking and the type of contributions allowed (as well as those forbidden). Spatial usability cues can be semiotically rich and highly visible, as in the case of written instructions. Usability then results in readability. They can be semiotically poor and inconspicuous, as in the case of properties of substance and surface. Note that usability cues should not be reified, but should be taken with reference to the participants; participants must be familiar (or make themselves familiar) with them in order to notice and realize them (cf. the lines on a sports ground). And, of course, participants are free to ignore them. Unlike package instructions, usability cues, in principle, refer to a broad range of possibilities of use. It is up to the interaction system (and not to ‘space itself ’) whether and which usability cues become interactively relevant. To sum up, there are three different resources employed during situational anchoring: the participants’ sensory and motor skills and spatial cognition, natural languages’ spatial lexis and grammar, and, last but not least, spatial semiotics (referred to here as spatial usability cues); see figure 4.

Fig. 4. Resources for the task of situational anchoring

On the interactive achievement of space – and its possible meanings

297

These resources have been studied more or less independently by different research traditions. Co-presence in all its implications has traditionally been dealt with in micro-sociological and linguistic face-to-face interaction research (see above, section 3). Language of space is of course a topic of linguistics, primarily general linguistics, typology and/or cognitive linguistics (see above, section 2). Until now, the empirical analysis of “spatial semiotics” (Ravelli & Stenglin 2008) or “geosemiotics” (Scollon & Scollon 2003) has primarily been the concern of semiotics and (cultural or social) geography (cf., for instance, Cresswell 2009) – if dealt with at all (cf. the references given in Kesselheim & Hausendorf 2007). The idea of spatial usability cues can be connected with the theory of affordances within “ecological psychology” (Gibson 1986) and – as far as objects are concerned – with research on the “design of everyday things” (Norman 2000). x

x

x

It is suggested here that these resources must be taken into account in order to elaborate on the situational anchoring. Co-orientation, co-ordination and co-operation necessarily exploit all of these resources. They must be analysed for each and every concrete episode of interaction. Otherwise, our understanding of situational anchoring would remain artificially restricted: blind to sensory perception and body movement, blind to language, blind to the spatial semiotics of the material world, and, last but not least, blind to the empirical correlation between the three types of resources. It does not come as a surprise that space proves to be an interactive achievement even in those (telling) cases in which it has become a part of our material world, where it has become visible and tangible. Sociologically or linguistically, there is no such thing as space as such, nor is there interactional space. There is only face-to-face interaction and how it situates itself in a concrete “here” in terms of co-orientation, co-ordination and co-operation. Each of these makes use of the participants’ co-presence (sensory perception, body movement, social cognition), natural language and spatial usability cues.

References Adelswärd, Viveka, Karin Aronsson, Linda Jönsson & Per Linell 1987: The unequal distribution of interactional space. Dominance and control in courtroom interaction. Text – A interdisciplinary Journal for the Study of Discourse 7(4): 313–346. Auer, Peter 1986: Kontextualisierung. Studium Linguistik 19: 22–47.

298

Heiko Hausendorf

Auer, Peter & Jürgen Erich Schmidt (eds.) 2010: Language and Space. An International Handbook of Linguistic Variation. Volume 1: Theories and Methods. (Handbücher zur Sprach- und Kommunikationswissenschaft, 30/1.) Berlin/New York: de Gruyter. Bachmann-Medick, Doris 2006: Spatial Turn. In: Doris Bachmann-Medick (ed.), Cultural Turns. Neuorientierungen in den Kulturwissenschaften, 284–328. Hamburg: Rowohlt. Backhaus, Norman, Claude Reichler & Matthias Stremlow 2007: Alpenlandschaften. Von der Vorstellung zur Handlung: Thematische Synthese zum Forschungsschwerpunkt I “Prozesse der Wahrnehmung und Darstellung von Landschaften und Lebensräumen der Alpen”. Zürich: vdf, Hochschulverlag AG an der ETH. Bickel, Balthasar & Johanna Nichols 2006: Oceania, the Pacific Rim and the theory of linguistic areas. Proceedings of the 32nd annual meeting of the Berkeley Linguistics Society. http://www.uni-leipzig.de/~bickel/research/papers/bickel_nichols_bls32.pdf, last accessed 03. 06. 2009. Breidenstein, Georg 2004: KlassenRäume – eine Analyse räumlicher Bedingungen und Effekte des Schülerhandelns. Zeitschrift für qualitative Bildungs-, Beratungs- und Sozialforschung 5(1): 87–107. Bruner, Jerome S. 1995: From joint attention to the meeting of minds: an introduction. In: Chris Moore & Philip J. Dunham (eds.), Joint Attention. Its Origins and Role in Development, 1–14. Hillsdale, NJ: Erlbaum. Ciolek, Matthew & Adam Kendon 1980: Environment and the spatial arrangement of conversational interaction. Sociological Inquiry 50 (Special issue on Language and Social Interaction, ed. by Don Zimmerman & Candace West): 237–271. Crabtree, Andy 2000: Remarks on the social organisation of space and place. Journal of Mundane Behavior 1: 25–44. Cresswell, Tim 2009: Place. A Short Introduction. [Repr.]. Malden, Mass.: Blackwell. Davenport, Elisabeth & Kathy Buckner 2005: Shifting presence in the classroom. The case of the open plan teaching laboratory. In: Phil Turner & Elisabeth Davenport (eds.), Spaces, Spatiality and Technology, 31–44. (The Kluwer International Series on Computer Supported Cooperative Work.) Dordrecht: Springer. De Mulder, Walter 1996: Demonstratives as locating expressions. In: Martin Pütz & René Dirven (eds.), The Construal of Space in Language and Thought, 29–47. (Cognitive linguistics research.) Berlin: Mouton de Gruyter. Deppermann, Arnulf & Angelika Linke (eds.) 2010: Sprache intermedial. Stimme und Schrift, Bild und Ton. (Jahrbuch des Instituts für Deutsche Sprache 2009.) Berlin: de Gruyter. Deppermann, Arnulf & Reinhold Schmitt 2007: Koordination. Zur Begründung eines neuen Forschungsgegenstandes. In: Reinhold Schmitt (ed.), Koordination. Analysen zur multimodalen Interaktion, 15–54. Tübingen: Narr. D’hondt, Sigurd, Jan-Ola Östman & Jef Verschueren (eds.) 2009: The Pragmatics of Interaction. (Handbook of Pragmatics Highlights, Vol. 4.) Amsterdam: Benjamins. Döring, Jörg & Tristan Thielmann (eds.) 2008: Spatial Turn. Das Raumparadigma in den Kultur- und Sozialwissenschaften. (Sozialtheorie.) Bielefeld: transcript Verlag. Dunham, Philip J. & Chris Moore 1995: Current themes in research on joint attention. In: Chris Moore & Philip J. Dunham (eds.), Joint Attention. Its Origins and Role in Development, 15–28. Hillsdale, NJ: Erlbaum. Emmison, Michael & Philip Smith 2000: Researching the Visual. Images, Objects, Contexts and Interactions in Social and Cultural Inquiry. London: Sage Publications.

On the interactive achievement of space – and its possible meanings

299

Enfield, Nicholas J. 2008: Common ground as a resource for social affiliation. In: István Kecskés & Jacob Mey (eds.), Intention, Common Ground and the Egocentric Speaker-Hearer, 223–254. (Mouton series in pragmatics.) Berlin/New York: Mouton de Gruyter. Enfield, Nicholas J. & Stephen C. Levinson (eds.) 2006: Roots of Human Sociality. Culture, Cognition, Interaction. Oxford: Berg. Frank, Roslyn M., René Dirven, Tom Ziemke & Enrique Bernárdez (eds.) 2008: Body, Language and Mind. Volume 2: Sociocultural Situatedness. (Cognitive linguistics research 35/2.) Berlin/New York: Mouton de Gruyter. Frers, Lars 2009: Space, materiality and the contingency of action. A sequential analysis of the patient’s file in doctor-patient interactions. Discourse Studies 11(3): 285–303. http://dis.sagepub.com/cgi/reprint/11/3/285. Frers, Lars & Lars Meier 2007: Working with the visual. In: Lars Frers & Lars Meier (eds.), Encountering Urban Places. Visual and Material Performances in the City, 171–181. (Re-materialising cultural geography.) Aldershot: Ashgate. Gibson, James J. 1977: The theory of affordances. In: Robert Shaw & John Bransford (eds.), Perceiving, Acting, and Knowing. Toward an Ecological Psychology, 67–82. Hillsdale, N.J/New York: Lawrence Erlbaum Associates. Gibson, James J. 1986: The Ecological Approach to Visual Perception. Hillsdale, N.J./London: Lawrence Erlbaum Associates. Goffman, Erving 1964: The neglected situation. In: John J. Gumperz & Dell Hymes (eds.), The Ethnography of Communication. American Anthropologist, 133–136. (Special issue 6/2.) Menasha: American Anthropological Association. Goffman, Erving 1974: Das Individuum im öffentlichen Austausch. Mikrostudien zur öffentlichen Ordnung. Frankfurt am Main: Suhrkamp. Goffman, Erving 1977: Rahmen-Analyse. Ein Versuch über die Organisation von Alltagserfahrungen. Frankfurt am Main: Suhrkamp. Goodwin, Charles 2000: Action and embodiment within situated human interaction. Journal of Pragmatics 32: 1489–1522. Hall, Edward T. 1976: Die Sprache des Raumes. Düsseldorf: Schwann. First published 1966 The Hidden Dimension. Hausendorf, Heiko 1992: Gespräch als System. Linguistische Aspekte einer Soziologie der Interaktion. Opladen: Westdeutscher Verlag. Hausendorf, Heiko 1995: Deixis and orality: Explaining games in face-to-face interaction. In: Uta Quasthoff (ed.), Aspects of Oral Communication, 181–197. (Research in Text Theory 21.) Berlin/New York: de Gruyter. Hausendorf, Heiko 2003: Deixis and speech situation revisited. The mechanism of perceived perception. In: Friedrich Lenz (ed.), Deictic Conceptualisiation of Space, Time and Person, 249–269. Amsterdam: Benjamins. Hausendorf, Heiko 2010: Interaktion im Raum. Interaktionstheoretische Bemerkungen zu einem vernachlässigten Aspekt von Anwesenheit. In: Arnulf Deppermann & Angelika Linke (eds.), Sprache intermedial. Stimme und Schrift, Bild und Ton, 163–197. Berlin/New York: de Gruyter. Hausendorf, Heiko (ed.) 2007: Gespräch als Prozess. Linguistische Aspekte der Zeitlichkeit verbaler Interaktion. Tübingen: Narr. Hausendorf, Heiko & Uta Quasthoff 1996: Sprachentwicklung und Interaktion. Eine linguistische Studie zum Erwerb von Diskursfähigkeiten. Opladen: Westdeutscher Verlag. Hausendorf, Heiko & Uta Quasthoff 1992: Children’s storytelling in adult-child in-

300

Heiko Hausendorf

teraction: Three dimensions in narrative development. Journal of Narrative and Life History 2(3): 293–306. Hausendorf, Heiko & Wolfgang Kesselheim 2008: Textlinguistik fürs Examen. (Linguistik fürs Examen 5.) Göttingen: Vandenhoeck & Ruprecht. Hayward, William G. & Michael J. Tarr 1995: Spatial language and spatial representation. Cognition 55: 39–84. Hickmann, Maya & Stéphane Robert (eds.) 2006: Space in languages. Linguistic systems and cognitive categories. (Typological studies in language, 66.) Amsterdam: Benjamins. Hindmarsh, Jon & Christian Heath 2000: Embodied reference. A study of deixis in workplace interaction. Journal of Pragmatics 32: 1855–1878. Hutchins, Edwin & Leysia Palen 1997: Constructing meaning from space, gesture, and speech. In: Lauren B. Resnick, Roger Säljö, Clotilde Pontecorvo & Barbara Burge (eds.), Discourse, Tools and Reasoning. Essays on Situated Cognition, 23–40. Berlin, Heidelberg: Springer. Kallmeyer, Werner & Fritz Schütze 1976: Konversationsanalyse. Studium Linguistik 1: 1–28. Keating, Elizabeth 2000: Space. Journal of Linguistic Anthropology 9 (1/2): 234–237. Keating, Elizabeth 2006: Spatiality and Language. In: Keith Brown (ed.), Encyclopedia of Language & Linguistics. 2nd edition, 616–619. Amsterdam: Elsevier. Kecskés, István & Jacob Mey (eds.) 2008: Intention, common ground and the egocentric speaker-hearer. (Mouton series in pragmatics 4.) Berlin/New York: Mouton de Gruyter. Kendon, Adam 1990a: A description of some human greetings. In: Adam Kendon, Conducting Interaction. Patterns of Behavior in Focused Encounters, 153–207. (Studies in Interactional Sociolinguistics 7.) Cambridge: Cambridge University Press. (First published in 1973.) Kendon, Adam 1990b: Conducting Interaction. Patterns of Behavior in Focused Encounters. (Studies in Interactional Sociolinguistics 7.) Cambridge: Cambridge University Press. Kendon, Adam 1990c: Some context for Context Analysis. A view of the origins of structural studies of face-to-face interaction. In: Adam Kendon, Conducting Interaction. Patterns of Behavior in Focused Encounters, 15–49. (Studies in Interactional Sociolinguistics 7.) Cambridge: Cambridge University Press. Kesselheim, Wolfgang 2010: ‘Zeigen, erzählen und dazu gehen’: Die Stadtführung als raumbasierte kommunikative Gattung. In: Marcella Costa & Bernd MüllerJacquier (eds.), Deutschland als fremde Kultur: Vermittlungsverfahren in Touristenführungen, 244–271. (Reihe interkulturelle Kommunikation.) München: Iudicium. Kesselheim, Wolfgang & Heiko Hausendorf 2007: Die Multimodalität der Ausstellungskommunikation. In: Reinhold Schmitt (ed.), Koordination. Analysen zur multimodalen Interaktion, 339–375. Tübingen: Narr. Kieserling, André 1999: Kommunikation unter Anwesenden. Studien über Interaktionssysteme. Frankfurt am Main: Suhrkamp. Kissmann, Ulrike Tikvah (ed.) 2009: Video Interaction Analysis. Methods and Methodology. Frankfurt am Main/Berlin/Bern/Bruxelles/New York/Oxford/Vienna: Lang. Kleiber, Georges 1983: Les démonstratifs démontrent-ils? Sur le sens des adjectifs et pronoms démonstratifs. Le français moderne 51(2): 99–117. Kruse, Lenelis & Carl Friedrich Graumann 1978: Sozialpsychologie des Raumes und der Bewegung. In: Kurt Hammerich & Michael Klein (eds.), Materialien zur Soziologie des Alltags, 177–219. (Kölner Zeitschrift für Soziologie und Sozialpsychologie.) Opladen: Westdeutscher Verlag.

On the interactive achievement of space – and its possible meanings

301

LeVine, Philip 2007: Sharing common ground: The role of place reference in parentchild conversation. In: Deborah Tannen, Shari Kendall & Cynthia Gordon (eds.), Family Talk. Discourse and Identity in Four American Families, 263–282. Oxford: Oxford University Press. Levinson, Stephen C. 1996: Language and Space. Annual Review of Anthropology 25: 353–382. Levinson, Stephen C. 2003: Space in Language and Cognition. Explorations in Cognitive Diversity. (Language, culture and cognition 5.) Cambridge: Cambridge University Press. Levinson, Stephen C. & David Wilkins (eds.) 2006: Grammars of Space. Explorations in Cognitive Diversity. (Language, culture and cognition 6.) Cambridge: Cambridge University Press. Loenhoff, Jens 2003: Grundlagen der kommunikativen Dimension von Körperbewegung und Tanz. In: Antje Klinge & Martina Leeker (eds.), Tanz, Kommunikation, Praxis, 17–31. Münster: LIT Verlag. Löw, Martina 2001: Raumsoziologie. Frankfurt am Main: Suhrkamp. Luhmann, Niklas 1984: Soziale Systeme. Grundriss einer allgemeinen Theorie. Frankfurt am Main: Suhrkamp. Luhmann, Niklas 2005: Einfache Sozialsysteme. In: Niklas Luhmann (ed.), Soziologische Aufklärung 2. Aufsätze zur Theorie der Gesellschaft. 5. Auflage, 25–47. Wiesbaden: VS Verlag für Sozialwissenschaften. Mondada, Lorenza 2007: Interaktionsraum und Koordinierung. In: Reinhold Schmitt (ed.), Koordination. Analysen zur multimodalen Interaktion, 55–93. Tübingen: Narr. Mondada, Lorenza & Reinhold Schmitt (eds.) 2010: Situationseröffnungen. Zur multimodalen Herstellung fokussierter Interaktion. (Studien zur deutschen Sprache 47.) Tübingen: Narr. Moore, Chris & Philip J. Dunham (eds.) 1995: Joint Attention. Its Origins and Role in Development. Hillsdale, NJ: Erlbaum. Müller, Cornelia & Ulrike Bohle 2007: Das Fundament fokussierter Interaktion. Zur Vorbereitung und Herstellung von Interaktionsräumen durch körperliche Koordination. In: Reinhold Schmitt (ed.), Koordination. Analysen zur multimodalen Interaktion, 129–165. Tübingen: Narr. Muysken, Pieter (ed.) 2008: From Linguistic Areas to Areal Linguistics. (Studies in language companion series 90.) Amsterdam: Benjamins. Norman, Donald A. 2000: The Design of Everyday Things. 3rd edition. London: MIT Press. Norris, Sigrid & Rodney H. Jones (eds.) 2005: Discourse in Action. Introducing Mediated Discourse Analysis. Abingdon/Oxon: Routledge. O’Doherty, Brian 1996: In der weißen Zelle. Inside the White Cube. Berlin: Merve. Peterson, Mary A., Lynn Nadel, Paul Peterson Bloom & Merrill F. Garrett 1996: Space and Language. In: Paul Peterson Bloom (ed.), Language and Space, 553–577. (Language, speech, and communication.) Cambridge: MIT Press. Pfeiffer, Thies 2010a: Object deixis: Interaction between verbal expressions and manual pointing gestures. In: Johannes Haack, Heike Wiese, Andreas Abraham & Christian Chiarcos (eds.), Proceedings of KogWis 2010. 10th Biannual Meeting of the German Society for Cognitive Science, 221–222. (Potsdam Cognitive Science Series.) Potsdam: Universitätsverlag Potsdam. Pfeiffer, Thies 2010b: Tracking and visualizing visual attention in real 3D space. In:

302

Heiko Hausendorf

Johannes Haack, Heike Wiese, Andreas Abraham & Christian Chiarcos (eds.), Proceedings of KogWis 2010. 10th Biannual Meeting of the German Society for Cognitive Science, 220–221. (Potsdam Cognitive Science Series.) Potsdam: Universitätsverlag Potsdam. Ravelli, Louise J. & Maree Stenglin 2008: Feeling space: Interpersonal communication and spatial semiotics. In: Gerd Antos (ed.), Handbook of Interpersonal Communication, 355–396. Berlin/New York: Mouton de Gruyter. Richardson, Tim & Ole B.Jensen 2003: Linking discourse and space. Towards a cultural sociology of space in analysing spatial policy discourses. Urban Studies 40(1): 7–22. Schegloff, Emmanuel A. & Harvey Sacks 1973: Opening up closings. Semiotica 8: 289–327. Schmitt, Reinhold 2006: Videoaufzeichnungen als Grundlage für Interaktionsanalysen. Deutsche Sprache 34(1–2): 18–31. Schmitt, Reinhold (ed.) 2007: Koordination. Analysen zur multimodalen Interaktion. Tübingen: Narr. Schroer, Markus 2007: Räume, Orte, Grenzen. Auf dem Weg zu einer Soziologie des Raums. [reprint.]. Frankfurt am Main: Suhrkamp. Scollon, Ronald & Suzanne B. K. Scollon 2003: Discourses in Place. Language in the Material World. London: Routledge. Sidnell, Jack 2009: Deixis. In: Jef Verschueren & Jan-Ola Östman (eds.), Key Notions for Pragmatics, 114–138. (Handbook of pragmatics highlights.) Amsterdam: Benjamins. Streeck, Jürgen, Charles Goodwin & Curtis D. LeBaron 2011: Embodied Interaction. Language and Body in the Material World. New York: Cambridge University Press. Stukenbrock, Anja & Karin Birkner 2010: Multimodale Ressourcen für Stadtführungen. In: Marcella Costa & Bernd Müller-Jacquier (eds.), Deutschland als fremde Kultur: Vermittlungsverfahren in Touristenführungen, 214–243. (Reihe interkulturelle Kommunikation.) München: Iudicium. Tomasello, Michael 1995: Joint attention as social cognition. In: Chris Moore & Philip J. Dunham (eds.), Joint Attention. Its Origins and Role in Development, 103–130. Hillsdale, NJ: Erlbaum. Tomasello, Michael 2000: The Cultural Origins of Human Cognition. 2nd edition. Cambridge, Mass.: Harvard University Press. Tomasello, Michael, Barbara L. George, Ann Cale Kruger, Michael Jeffrey Farrar & Andrea Evans 1985: The development of gestural communication in young chimpanzees. Journal of Human Evolution 14(2): 175–186. Turner, Phil & Elisabeth Davenport 2005: An introduction to spaces, spatiality and technology. In: Phil Turner & Elisabeth Davenport (eds.), Spaces, Spatiality and Technology, 1–4. (The Kluwer International Series on Computer Supported Cooperative Work.) Dordrecht: Springer. Vom Lehn, Dirk & Christian Heath 2007: Perspektiven der Kunst – Kunst der Perspektiven. In: Heiko Hausendorf (ed.), Vor dem Kunstwerk. Interdisziplinäre Aspekte des Sprechens und Schreibens über Kunst, 147–170. München: Fink. Wegner, Dirk 1985: Der persönliche Raum als Modell nonverbaler Proxemik. In: Harro Schweizer (ed.), Sprache und Raum. Psychologische und linguistische Aspekte der Aneignung und Verarbeitung von Räumlichkeit. Ein Arbeitsbuch für das Lehren von Forschung, 162–182. Stuttgart: Metzler.

On the interactive achievement of space – and its possible meanings

303

Weinrich, Harald 2005: Textgrammatik der deutschen Sprache. 3rd edition. Hildesheim: Olms. Wildgen, Wolfgang 2007: Wege in die Stadt oder das Lesen der Stadt als Zeichen. In: Rita Franceschini (ed.), Im Dickicht der Städte I: Sprache und Semiotik, 24–42. (Zeitschrift für Literaturwissenschaft und Linguistik.) Stuttgart/Weimar: Metzler. Ziemke, Tom, Jordan Zlatev & Roslyn M. Frank (eds.) 2007: Body, Language and Mind. Volume 1: Embodiment. (Cognitive linguistics research 35,1.) Berlin/New York: Mouton de Gruyter.

304

Jürgen Streeck

Jürgen Streeck

Plaza: Space or place?

1.

Introduction

In this chapter I report on a study of a plaza, Plaza de la Trinidad, the central square of the working-class neighborhood of Getsemaní in Cartagena de Indias, Colombia. The study is an attempt to understand a social form – the traditional, localized pattern of habitualized communal life that makes up the social entity plaza – as the ongoing product of an emerging and self-sustaining web of face-to-face interactions. Every night, a sizeable sector of the community of Getsemaní assembles in the Plaza, and, as everyone can watch everyone else’s interactions, gains a visceral sense of itself. The community enacts, embodies, and reflects itself in the Plaza. My aim is to find out whether a bottom-up micro-ethnography can help us explain how the traditional social form that bounds these interactions and relationships is re-inscribed or sustained, how the macro-organization of plaza-dwelling is brought about by the totality of person-to-person encounters. What we see in the Plaza in the way of social structure is entirely the result of self-organization; there is no hidden hand that allocates people to one another and to places. Rather, whatever pattern there is has emerged from histories of interaction, which have taken on the form of a complex, self-sustaining system (Streeck and Jordan 2009) – a system that by and large re-instantiates itself: interactions generating just those social experiences that draw everyone back to the scene every evening. Apart from memories of a more distant past that many had retained, the system as we currently observe it has been in existence for about five years. Before, Plaza de la Trinidad was a feared center of the drug trade and the violent crime that accompanied it.1 1

UNESCO made Cartagena a World Heritage Site and provided funds for a complete restoration of parts of the old town. However, this intervention has fundamentally altered the ecology of the old center: many small neighborhood stores and craftsmen’s workshops have given way to expensive boutiques and restaurants, and the old town now thrives on a tourist economy: ciudad museo, critics call it. The adjacent working-class neighborhood of Getsemaní, in contrast, so far has managed to preserve the ecology and economy that characterized all of old Cartagena before. But the gentrification of Getsemaní is a project of the Interameri-

Plaza: Space or place?

305

The Plaza with its clustered inhabitants can also be understood as a form or pattern in Christopher Alexander’s sense (Alexander 1964, Alexander, Ishjikawa & Silverstein 1977): an evolved, traditional structure that has emerged as the result of the self-organization of collective human action, an emerged “transcript” of the structure of lived time; the spatial pattern manifests stages of life and the relationships between the cohorts occupying them. To understand this social form, I have tried to capture with my videocamera how individuals enter, move about, and take positions in the place. I study their interactive emplacement and try to understand how stable forms of social integration emerge from their movements and engagements. While one of my initial interests was in the roles that architectural forms or affordances – the built environment of the Plaza – play in facilitating and sustaining this massive and complex interaction network, I increasingly realized that the diversity of engagements that can take place in a single location requires a different approach, one that reconstructs interactional organization from the ways in which people emplace themselves. The structure of the chapter reflects this shift from space to place: while I begin with a ‘top down’ description of territories and clusters of interactional engagement, I then turn to interaction sequences that reveal how interaction ensembles make places for themselves in the Plaza and its participants acquire senses of place (Feld & Basso 1996). The data for this study are approximately 20 hours of video-recordings that I made, quite visibly to everyone, over a period of three weeks in July and August 2009.2 Plaza de la Trinidad, a small square of not more than 120 by 120 feet, dominated by the church, Iglesia de la Santissima Trinidad, and surrounded by one-and two-story houses, forms the intersection of most of the small streets that cross the neighborhood (see Fig.1). It is almost impossible to get from one point of the neighborhood to another or from Getsemaní to the city center without passing through the Plaza, which thus forms the most likely location for chance encounters with acquaintances, an organic station (Giddens 1984) for social encounters. It is the space where acquaintances are made and where one becomes a publicly known person. The plaza resembles

2

can Development Bank and the local Chamber of Commerce; a marina and apartment buildings would be built, and between 40 and 60 % of the local population who rent their apartments and houses will likely be evicted; Stella Cocchi, pers. comm. Among many others, I want to thank Carmelo Hernandez, leader of Getsemaní’s neighborhood association, and Susan Alvarado for making the field-research possible. Research was conducted while I was a Senior Fellow at Freiburg Institute for Advanced Study (FRIAS), whose support I gratefully acknowledge.

306

Jürgen Streeck

Fig. 1. Map of Getsemaní

an amphitheater with a stage: a circular “stage” surrounded by a ring of “bleachers”, two long curving stone benches, stairs, a church porch, small wooden benches – all providing opportunities to sit down and watch, talk, observe, participate, or contemplate (Figs. 2–5). In this chapter, after summarizing a few key ideas concerning the proper conceptualization of space for interaction research, I outline aspects of the spatial organization of interaction in Plaza de la Trinidad. I then lay out how the approach taken misses the social significance of the place and contrast it with an analysis of social interaction that focuses on how people establish meaningful relations with places in the Plaza, how they acquire shared senses of place, and, in emplacing themselves contribute to the reproduction of the community’s identity over time.

Plaza: Space or place?

307

Fig. 2. East side of Plaza de la Trinidad with the young men’s corner and the stand of a vendedor de minutos

Fig. 3. North side with long bench and a bundled-up inflatable cat

308

Fig. 4. Church portal and porch

Fig. 5. Northwest corner

Jürgen Streeck

Plaza: Space or place?

2.

309

Territories, stations, and clusters

Georg Simmel (1903) is usually credited with being the first sociologist to have realized the important roles that space plays as a medium of social processes and structures. Simmel pointed out that various forms of social organization (Vergesellschaftung), including states and groups, manifest themselves and are “fixated” in spatial forms such as territory, border, and location; that social relationships are externalized in degrees of proximity and distance; and that locations can possess “the character of personal uniqueness” for persons and groups – a dimension that, in English, is associated with the term place. As Löw (2001) has emphasized, to discern spatial aspects of social interaction and organization, it is necessary to conceive space in relativist terms. In Newtonian physics, space was an absolute: a three-dimensional matrix, invariant and independent of human action. It was conceived as a container within which bodies exist and events take place. In contrast, in a relativist universe, as proposed by Einstein, space results from the position and relations among bodies, it is a “relational structure among bodies” (Löw 2001: 34). Space thus conceived is “relational order(ing) of bodies and social goods at a place” (131). Spatial relations and spaces that are synthesized into bounded units are, in other words, the results of spacing activities, which therefore should draw our analytic attention. In The Constitution of Society (1984), Giddens, drawing upon Goffman’s micro-studies of encounters and social order (Goffman 1963, 1971), introduced the concept of stations. A station is the intersection of two or more paths, where individuals encounter each other and engage in focused interaction, thus coming to temporary rest. (The station, we might note, thus becomes a place for them.) Stations interrupt individual mobility: they bundle and center interactions (Schroer 2006: 111). Stations are, in other words, contexts (Giddens 1992: 113), “indices of what takes place or should take place in them” (Schroer 2006: 113). As Schroer (2006: 115) has pointed out, in stations space is made available as a frame of reference for interactions, while these frames of references are reciprocally responsible for specifying the contextuality of space. The space, concretized as locale, acquires its character from the interactions that take place in it.

The analysis of stations thus leads to the investigation of the parties’ positioning, which, as Giddens (1984: 84–5) has pointed out, transcends the context of the here and now:

310

Jürgen Streeck

The positioning of agents in circumstances of co-presence is an elemental feature of the structuration of encounters. Positioning […] involves many subtle modalities of bodily movement and gesture, as well as the general motion of the body through the regional sectors of daily routines. The positioning of actors in the regions of their daily time-space paths, of course, is their simultaneous positioning within the broader regionalization of societal totalities. […] In contemporary societies individuals are positioned within a widening range of zones […], all displaying features of system integration which increasingly relates the minor details of daily life to social phenomena of massive time-space extension.

Prior to sociology’s spatial turn (Warf & Arias 2009), but equally inspired by Goffman (1963, 1971; see Kendon 1988), interaction researchers committed to context analysis (Kendon 1990; Scheflen 1973, 1974) have conducted close, film- and video-based investigations of the spatial organization of encounters in everyday life. Following Scheflen’s seminal studies of postural configurations (Scheflen 1964, 1972) and the territorial organization of social life (Scheflen & Ashcroft 1976), Kendon delineated the spatial maneuvers and positionings by which participants bring about the unity of interactional encounters, distinguish participants from non-participants and by-standers, and mark transitions in the contextual frames of their interaction (Kendon 1970, 1990). He calls the basic spatial-orientational system by which focused interactions are sustained F-formation system and writes: The F-formation system serves as a […] means of maintaining the separate identity and integrity of an interactional situation […] by which the participants can maintain differential access to one another […] It is thus an important part of the means by which behavior is organized (Kendon 1990: 209–210).

F-formations are self-sustaining systems: they frame and maintain focused engagements; they display the engagement and the degree of its accessibility to others, adapt to and express changing activities and participation frameworks, and require interactive work to be suspended. As analysts, we can often infer participation frameworks and patterns of interaction in a cluster from changes in the parties’ relative positioning, orientation, and postures.

3.

Territory and stations

How do these spatial dimensions of social organization express themselves in Plaza de la Trinidad3? I discuss this in four steps, moving from territories, to stations, to interacting clusters and their formations. However, instead of an in-depth analysis I offer the examples only as abbreviated illustrations. 3

For a long-term ethnographic study of a Latin American plaza see Low 2000; for a review of research on social interaction and built form Lawrence & Low 1996.

Plaza: Space or place?

311

Fig. 6. Running game

The open space in the center of the Plaza accommodates much of the pedestrian cross-traffic and, as a consequence, an infinite series of chance encounters and engagements resulting from them, but it is also the field where children’s space-hungry games take place which may involve more than a dozen children. Characteristic games are soccer and baseball, exclusively played by boys, and games of tag, catch, or other running games, which involve younger and older girls along with young boys. Children’s games occupy by far the largest amount of space, but they never fully occupy the center, having a spatial structure that is porous and easily accommodates other engagements in its field (Fig.6). A handful of benches are usually placed at the inner periphery, each accommodating up to four adults. The benches appear to be occupied by sets of people related as peers or by some shared social identity such as age and sex: gaggles of teenagers, sets of middle-aged women, older women and men. It is rare that people who are not somehow related by kind or kin sit on a bench together for an extended period of time. This preference for exclusive engagements distinguishes the small benches from the two long concrete benches along the Northern perimeter that accommodate multiple and diverse interaction clusters. They are the first point of call for tourists who pass through before they get caught up in the attractions of Plaza life, and most groups who come from other parts of town during weekend nights sit and drink beer or soft-drinks here, unless they sit on the stairs leading up to the church. The stairs are flanked by a porch on both sides, which occupies the

312

Jürgen Streeck

entire Southern border of the Plaza and offers two levels of seating. All people who are seated usually face the inside of the Plaza; they are grouped around the center as if on bleachers around a stage. Within this overall framework, some locations serve as stations or docks for recursive activities or as meeting points for established groups of people. I use the term dock to refer to stations that are anchored to a physical location or structure. Docks can be occupied opportunistically or habitually. Docks are “sedimented in time, [so that] sense can be made of them only by considering their routinized, repetitive character” (Giddens 1992: 143). Docks, in sum, are physical structures that anchor sustained interaction systems. On different nights I have counted an average of 23 docks during the busiest times of night in the Plaza. A place can of course be at one hour a foodstand’s dock and at another a posse’s home-base and at yet another time become a station for a an interaction resulting from a chance encounter. One of the most striking docks is provided by the work-station of one vendedor de minutos, a small desk and chair, positioned next to a wooden bench (Fig. 7). A vendedor de minutos is like a living phone booth: owning three or four cellphones, each registered to a different long-distance company, he sells phone service by the minute. But he (and, to a lesser extent, his older female counterpart on the other side of the Plaza) also constitutes the center and switchboard of a nearly unbroken chain of overlapping face-to-face conversations. He opens for business at 8 in the morning and usually does not leave until midnight, but he is hardly ever alone. Sometimes only one person keeps him company, sometimes he is the center of a cluster of 6 or 8, and constantly people arrive to make a call or to return the phone and pay him when they are done. Often they then linger. Another dock is nearby. At the corner of one of the side streets is a small stairway leading to the blind door of an abandoned house. In the evening, this stair is the meeting point for young men, ranging from approximately 2 to 20 years (little ones naturally in the care of someone older), who hang out here regularly. Nobody else seems to claim this location at night, although the space is often occupied by a food-stand before dark. The cluster of young men thus organizes itself around a firm association of place (location) and group. The spatial formations that they adopt during their interactions make use of built space in a recurrent, quasi-conventional fashion: locations index interactional statuses, and their occupancy entails a claim to such status. While some of the lads sit on the stairs and some lean against the wall, others, usually just one or two, stand opposite them, by the curb or just beyond, on the street. The latter position is invariably that of the performer, the narrator, or the performance team. The stairs are the bleachers, the audience’s place (Fig. 8).

Plaza: Space or place?

Fig. 7. Vendedor de minutos

Fig. 8. Young men’s corner

313

314

Jürgen Streeck

Fig. 9. Jugglers performing for the children

4.

Clusters/engagements

What are the features of the interactions that routinely cluster at these docks, and elsewhere in the Plaza? It is apparent that quite a few interacting clusters are customary organizations, i.e. interactions that organize themselves in similar ways day after day, while others frame incidental encounters. Observing clusters over time, one can distinguish different motives or occasions that bring them about, singly or in combination with one another. For example, clusters can form in response to some unusual event, an attraction, such as a pair of jugglers from Argentina who performed occasionally for the children (Fig. 9), before they engaged them in practice sessions (Fig.10), or a man with a large inflatable cat, a jumping castle for children: as he began unfolding it, members of the neighborhood junta quickly objected, policemen appeared, and a prolonged debate ensued. This gathering of authorities in turn attracted the attention of bystanders, including children, who formed a circle around the primary interaction (Fig.11). Quite simply, attractions attract focused attention, embodied in a centripetal configuration of postures. Other clusters emerge as a by-product of shared activities, which may or may not afford the participation of new-comers. Such clusters can be highly dynamic, dispersed, and visible only over time, as the observer registers patterns of action and response across multiple parties. The most typical examples of this are games. Other activities, notably conversation, generate more tightly packed clusters. It is the activity – and its particular stage and participation framework at any moment – that determines the shape of the

Plaza: Space or place?

315

Fig. 10. Juggling lesson

Fig. 11. An attraction

cluster – of any cluster, even those that have not formed for the particular activity. Many clusters form because the Plaza is the meeting place or resting place for some set of community members, a posse, a family, a network of friends, an age cohort (e.g. older Getsemaní women habitually sharing a bench in the early evening). And finally there are what one could think of as space-generated clusters, which form when some set of people find themselves

316

Jürgen Streeck

Fig. 12. Narrative positioning

“pre-clustered” by the space they happen to be in and thus occupying slots in a spatial framework that facilitates or even stimulates interactional engagement. Built spaces – benches, church steps, the church terrace – offer seating opportunities, and clusters with varying degrees of interactional involvement emerge as a result of individuals or teams of people taking these opportunities and finding themselves unable or unwilling to maintain civil inattention (Goffman 1963). Many clusters emerge for multiple, overlapping reasons: groups of friends habitually meeting in the Plaza to engage in conversation, in a particular place that “fits” this type of activity, for example by providing spatial structure to anchor performer and audience-roles, as the corner where the young men congregate. Interacting ensembles make use of external, built spatial features to support their various forms of interaction. In many places we see people configured as narrator/performer and audience, the audience seated on steps, benches, or the porch, the performer standing before them (Fig.12, 13). Far more difficult to observe are interactions between clusters. One can observe minor tacit territorial negotiations between teams using the same or overlapping sectors of space. But if we want to assume that the web of interactions as a whole possesses some measure system of integrity, these occasional cross-cluster engagements will hardly be evidence enough. What we can observe, on the other hand, if we assume a bird’s eye-perspective, is an overall configuration in which mobile, dynamic, and transitory engagements

Plaza: Space or place?

317

Fig. 13. Performance positioning

in the center are surrounded by stationary clusters on the inner and outer peripheries, whose members preferably orient to the center and spend considerable time attending to what is happening there or somewhere else in the Plaza (where by and large everyone is visible to everyone else). In other words, the “higher-level” context of the interacting system “Plaza” replicates, fractal-like, features of the performer-audience configuration often enacted within lower-level interacting clusters. What we gain from this spatial analysis of Plaza interaction is indeed a bird’s eye or top-down picture, an overall image of a seemingly timeless formation. “Zooming in”, we can provide close-up accounts of the spatial maneuvers within and between interacting clusters and with pedestrians passing through. But what is striking about such an account is its abstractness: space is treated as a mere medium of social interaction and the built environment Plaza as a container and affordance-web for infinite social interaction. Analysis in terms of space captures a crucial dimension of social/interactional organization, but it is entirely indifferent to the meanings that people in their daily lives attach to a place due to the experiences that they have made and continue to make here. To them, Plaza de la Trinidad is not a space, but a concrete, specific, irreplaceable place.

318

5.

Jürgen Streeck

Place

“When space feels thoroughly familiar to us, it has become place” (Tuan 1977: 73). We know a place as a result and in the manner of our everyday movements: walking to the mailbox, making a right turn on the way to work, reaching for the plate, turning to face someone sitting at our table. Place is given to us as an “inherent capacity of the body to direct behaviors of the person intelligently, and thus function as a special kind of subject which expresses itself in a ‘preconscious’ way usually described by such words as ‘automatic’, ‘habitual’” (Seaman 1980: 155). That a location has become a place for us reveals itself in the fact that our body knows its way around, “in the readiness with which we move between differentially available reachables [which are] connected and traversed by multiple pathways” (Casey 1999: 53). Such spaces have become lived places (48). Ingold (2000: 101) writes about places: A place owes its character to the experiences it affords to those who spend time there […]. And these, in turn, depend on the kinds of activities in which its inhabitants engage. It is from this relational context of people’s engagement with the world, in the business of dwelling, that each place draws its unique significance. Through dwelling in a landscape, through the incorporation of its features into a pattern of everyday activities […] it becomes home.

While the investigation of space as a dimension of social organization has largely been the province of sociology and interaction studies, place is the domain of humanistic geography (Tuan 1977, 1979; Relph 1976), which draws its main inspiration from phenomenological conceptions of human embodiment (Merleau-Ponty 1992), being-in-the-world (Heidegger 1926), and emplacement (Casey 1999; Malpas 1999). In humanistic geography, ‘place’ means a “focus where we experience the meaningful events of our existence” (Norber-Schulz 1971: 19). While space is abstract, pure form, and spacing is the production of spatial forms and relations, “places have meaning” (Relph 1976: 3). The term emplacement (Casey 1996), emphasizes human agency in the emergence of place. It highlights the fact that places emerge in our experience as we situate and orient ourselves in them in the conduct of our everyday affairs. As we get to know places and incarnate them in our bodies, they become part of our identities: Particular places enter into our self-conception and self-identity inasmuch as it is only in, and through our grasp of, the places in which we are situated that we can encounter objects, other persons or, indeed, ourselves (Casey 1999: 177).

An analysis of place in social interaction is predicated on the equivalence of embodiment and emplacement which Casey (1999: 24) has emphasized:

Plaza: Space or place?

319

Lived bodies belong to places and help to constitute them. […] Places belong to lived bodies and depend on them. If it is true that ‘the body is our general medium for having a world’ (Merleau-Ponty 1962: 146), it ensues that the body is the specific medium for experiencing a place-world.

As we emplace ourselves in locations and get to know them, our bodies incarnate these places. The incarnate knowledge of places – perhaps the form of human knowing on which all other knowledge rests – is manifested in our ability to get around. This is personal knowledge (Polanyi 1958), acquired by and inalienable from the individual body; it is the lived body’s peculiar combination of being at once a ‘general medium for having a world’ and something quite idiosyncratic and personal (as always my body) that enables it to ensure the concreteness of the regions in which we are immersed in emplacements. […] Particular places enter into our self-conception and selfidentity inasmuch as it is only in, and through our grasp of, the places in which we are situated that we can encounter objects, other persons or, indeed, ourselves (Casey 1999: 74, 177).

We must therefore focus our investigation on how individual Plaza dwellers, individually, specifically, for the first time or habitually emplace themselves alone or in concert with others and establish or maintain a concrete personal connection with the Plaza. Of course, whenever they do so in interaction with others, they also organize this interaction – and thus their social experience of the place – in terms of generic spatial organizations, for example the relative positioning and orientation of story-teller and story-recipient, or teller and co-teller.

6.

Emplacement and the acquisition of senses of place

The initial focus of study has therefore been on ways in which children of different ages acquire senses of place and establish social presence, how they make places for themselves in the Plaza. It is not possible to follow individuals’ changing emplacements over a life-time, but we can study how individuals of different ages differentially emplace themselves and how the Plaza accommodates the successive stages of human life. This is a crucial question if we want to maintain that the Plaza indeed figures in the reproduction of the entire community, not just some sectors. (Naturally, small infants and the very old are rarely seen in the Plaza at night.) In the remainder of the chapter I will illustrate in very broad strokes some emplacements of young Plaza dwellers at different stages of life. Almost inadvertently, because the course of life in the Plaza goes from motion to rest, I implement Elias’ proposal to sociologists, that, even though

320

Jürgen Streeck

Fig. 14. Home-base on the church stairs

one tends to think from the state of stillness as the normal state towards motion as a special state, one gets a much better grasp of the states of affairs that are sociology’s subject matter if one does not abstract away from movements, from the character as a process (Elias 1970: 124, my transl.)

Stationary social life, in other words, evolves from social life in motion. Young children begin their Plaza life emplaced in the firm embrace of a parental infrastructure, initially holding them, then as a home-base to which to return (Fig.14). While some children venture on their own to nearby possibilities, they do seem to prefer to embark on joint adventures when a fellow traveler is available. The two little girls in this scene (Figs. 15–19) exemplify an early team-excursion and the joint sense of place that may or may not emanate from it. Having left their adult home-base, they slowly and cautiously wander about. One girl discovers one of the sculptures and explores its affordance as a station, a temporary home away from home. She nudges the other to join the place with her, but her invitation is rejected. A boy of their age arrives, dressed in a striped jumper (thus nicknamed here “Stripes”), tentatively acting on their tacit invitation to join them, but he quickly changes his mind and runs away to join the ball-game of a far older group of children. His attempt to then join a ball game by much older children is an instructive instance of the role of children’s interaction in the reproduction of the public sphere of Getsemaní society. He intervenes in an active, more advanced system, a ball-

Plaza: Space or place?

321

322

Jürgen Streeck

Figs. 15–19. Young children exploring affordances

game beyond his body’s full reach, and makes it adapt to himself (Figs. 20–23). Stripes runs with the big children, sometimes precisely anticipating where a ball will land and positioning himself in relation to the catcher so that he appears as the obvious next recipient. Older children adapt their actions to make his participation feasible, and he appears to favor quasi-participation in older children’s activity system over a more equitable, even superior role in a system into which his female age-mates had invited him. He aims for participation in this system, at his outer reach, rather than the available alternative of interacting with girls his age. Even though a ball game may be indifferent to the identity of a place as long as it meets certain requirements,

Plaza: Space or place?

323

the episode nevertheless demonstrates the equivalence of embodiment and emplacement: “Stripes”’ acquisition of distinct bodily skills makes him an increasingly qualified, locally knowledgeable Plaza dweller, not only in the domain of participating in games, but more generally in negotiating access to the activities of older children. The scene exemplifies that the acquisition of a sense of place is integrated with the acquisition of practices and skills for participation in social activities and relationships; both are intertwined in a person’s experience. The Plaza, in other words, becomes an arena for a developing sense of self. This is also visible in the walking game of two four or five year-old boys (Figs. 24–26), who navigate the terrain backwards, one initially guiding, the

324

Jürgen Streeck

Figs. 20–23. “Stripes” joining a ball game

other following. When there is a glitch – the boy in the follower’s role is unable to walk backwards and turn at the same time and comes out of a 180-degree turn walking forward – the boys quickly recalibrate their motions as the journey continues. Finally, the follower turns another mistaken turn into a twist, a pirouette, which now becomes his pattern which the previous leader strives to follow. A seemingly even more idiosyncratic sense of place is incrementally developed by a sister-brother team, blond children from the U.S. visiting with their mother. The children, through the coordinated exploratory actions of their feet, develop a very specific, ‘culturally other’ sense of the affordances of the Plaza, which they then make available to a local boy. The siblings’ re-

Plaza: Space or place?

Figs. 24–26. Boys walking backwards

325

326

Figs. 27–29. Siblings from abroad exploring local affordances

Jürgen Streeck

Plaza: Space or place?

Figs. 30–32. A new sense of place is disclosed and transmitted

327

328

Jürgen Streeck

lationship with one another and with the place is constituted and mediated by the movements of their feet. Their activity begins as a exploration of the affordances for motor action and experience of different locations, the shared enactive disclosing of what can be done. The exploration is successful for the first time when the boy discovers an outcropping above the porch of the church which he can reach with his hands, and which can be connected via his body to a step that he can reach with his feet (Fig. 27). By holding onto the outcropping, he manages to sustain himself on the step. When his sister, who had walked ahead, looks over and sees him doing this, she returns and also grabs the outcropping, hanging like her brother (Fig. 28). A shared emplacement schema – and a shared experience of the place – are thereby established. At the same time the girl’s actions are somewhat different from her brother’s. She looks down to her feet and examines the ground at the bottom of the church wall with an exploring foot, thus preparing to propel the shared activity further, into new terrain (Fig. 29). Then they step down the stairs and explore the ground together. At some point the boy disentangles himself from the activity and climbs up the church porch and then starts running up the oblique part of the church wall a few steps (Fig.30). His sister follows him (Fig.31). And a little boy who had previously been sitting between them on the terrace, runs up the stairs and to the place where the white boy is doing his routine, and tries the same steps, managing to climb halfway up the church incline (Fig. 32). As children grow older, they begin to respond to the pull of place. Their interactions are more likely to include stationary periods, and sometimes to acquire a home-base, a dock. Their interactions and social experiences, in other words, and to some extent their identities, become coupled with places. But we must be careful to not presumptively attribute general preferences to every individual Plaza dweller, or to do so across contexts. Thus, while stationary emplacement may generally be more characteristic of older children, one episode involves a boy who is not older than four or five. He takes his place at the foot of one of the life-size sculptures that are the landmark of the Plaza and then resides there, receiving visitors (Fig. 33). He quietly sucks a lollipop while observing the goings-on around him. Not only does he claim territory in this fashion, he also creates an idiosyncratic experience of the plaza for himself: he leans against the sculpture’s leg, and he engages with occasional visitors, another boy his age, an older girl, presumably his sister or cousin, who cares for him. A busy space in the Plaza is the large church pedestal, with the stairs leading up to the portal in its center. The pedestal is about 70 centimeters high and can comfortably be leaned against and sat on. The pedestal and stairs,

Plaza: Space or place?

329

Fig. 33. Young boy’s residence

along with the central “plain” of the Plaza, afford the greatest diversity of interactional engagements and formations. People may simply sit on its edge, alone or in teams, and observe the goings on. Parents sit on the church steps while their child or children climb around on them or play in the vicinity, within sight of their parents and returning to them as their human station (see Fig.14). During late afternoons a group of middle-aged men reclines or dozes on the pedestal at the corner that faces Calle del Pozo, because of the cool breeze that blows through this narrow street. Often one sees people sitting side-by-side on the pedestal or on the church steps, engaged in conversation. Frequently, one participant is standing, facing them. The pedestal also serves as an occasional anchor for the complex and diverse interactions of a large cohort of younger teenagers and pre-teens, as an arena for emerging cross-sex interactions, gossip, teasing, horse-play, and quasi-courtship. The physical structure of the church pedestal is almost always implicated in the fluid organization of teenage interaction. Boys may form a seated “with” (Goffman 1963) as they witness the horse-play below; the porch may be the story-recipient slot with the tellers occupying the ground before them; or the row of boys seated on the porch may become the successive recipients of rough affections from a chain of girls passing before them. The versatile affordances of the location appear to be what attracts the group (Figs. 34, 35). (All the while a young-adult couple is engaged in deeply physical and affectionate interaction, she seated, he standing in front of her, offering a model of more mature courtship interaction for the younger ones, Fig. 36.) Making this place a permanent home is impossible for the teenagers,

330

Jürgen Streeck

however, because it is a popular space, often occupied by adults, including strangers. The young men’s corner, as we have seen, is exclusively theirs at night, reflecting their far more assertive presence in the Plaza. In either ensemble, as in many others, one can observe the activities of younger dwellers who seek to gain access to the interaction “from below”, often succeeding in making the interactional system adapt to their special needs and granting them participation, if sometimes only in an “as if ”, ironic mode. Thus, over time, the children of Getsemaní become connected with the Plaza, having incorporated it – or places within it – in their movement repertoire, social practices, and interpersonal habits and stratagems. They do so

Plaza: Space or place?

331

Fig. 34–36. Quasi-courtship and courtship on the church porch

as they participate in broad, but age-graded interaction ensembles until they inevitably reach the point when the system loses interest to them and they move on, having in the meantime, through occasional peripheral participation, gained preliminary access to another cohort. In this fashion, the Plaza becomes a part of their biographical identities and senses of self. Although some children venture out on their own, making their place initially without others, this emplacement overwhelmingly occurs in interaction sequences in which two or more children together disclose and agree upon their sense of some “this” place in the Plaza.

7.

The overall system

We thus begin to discern an overall pattern of integration of routine goings-on in Plaza de la Trinidad, a pattern in which the entire ensemble configures and reproduces itself and thereby contributes to the subjective existence and reproduction of the community of which the Plaza is such an important part. This overall pattern is a drift over each inhabitant’s life-time from positioning him- or herself, first in the center but, as the years pass, towards the periphery, a gradual drift from active, physical, space-hungry play to conversational engagements to quiet observation. A demographic pattern is represented through the successive positions that people seek in the Plaza as they grow older. Children play in the center, and their togetherness takes place in

332

Jürgen Streeck

motion. Once they reach a certain threshold age, they begin to drift into a position on the periphery, which is stationary and provides for the framing of occasions for activities centered around talk. When they reach their twenties, spectator positions become more common. Old people sit side by side on the sidelines and watch. There is a protective quality to this formation: a circle of elders watches over the activities and maturation of the young. Or, to put it differently, in this pattern – which is visible to, though probably unnoticed by, anyone – the community represents itself to itself in a particular formation, one in which adults gather around and monitor the young, and in which growing older means moving from the center to the periphery, from where the action is to a place of contemplation. The structure of the Plaza as a whole, as it is inhabited or enacted by the people, tracks and represents a particular social figuration, in which the community replenishes itself from the center of childhood around which it arranges itself. In this way, the patterns of self-organization in the Plaza are also patterns of self-representation, presenting a particular image of the society to its members and re-inscribing a seemingly natural, invariable order of life both onto the place and the embodied memories of its members.

References Alexander, Christopher 1964: Notes on the Synthesis of Form. Cambridge, MA: Harvard University Press. Alexander, Christopher, Sara Ishjikawa & Murray Silverstein 1977: A Pattern Language. New York: Oxford University Press. Elias, Norbert 1970 [2009]: Was ist Soziologie? Weinheim/München: Juventa. Feld, Steven & Keith H. Basso (eds.) 1996: Senses of Place. Santa Fe, N.M.: School of American Research Press. Giddens, Anthony 1984: The Constitution of Society. Outline of the Theory of Structuration. Berkeley/Los Angeles: University of California Press. Goffman, Erving 1963: Behavior in Public Places. New York: The Free Press. Goffman, Erving 1971: Relations in Public. Microstudies of the Public Order. New York: Basic Books. Heidegger, Martin 1962 [1926]: Being and Time. New York: Harper and Row. Ingold, Tim 2000: The Perception of the Environment: Essays on Livelihood, Dwelling and Skill. London/New York: Routledge. Ingold, Tim & Jo Lee Vergunst 2008: Introduction. In: Tim Ingold & Jo Lee Vergunst (eds.), Ways of Walking. Ethnography and Practice on Foot, 1–20. Aldershot: Ashgate. Kendon, Adam 1970: Movement coordination in social interaction: some examples described. Acta Psychologica 32: 100–125. Kendon, Adam 1982: The study of gesture: some observations on its history. Recherches Semiotique/Semiotic Inquiry 2(1): 25–62.

Plaza: Space or place?

333

Kendon, Adam 1988: Goffman’s approach to face-to-face interaction. In: Paul Drew & Anthony Wootton (eds.), Erving Goffman: Exploring the Interaction Order, 15–40. Cambridge: Polity Press. Kendon, Adam 1990: Conducting Interaction. Cambridge: Cambridge University Press. Lawrence, Denide & Setha M. Low 1990: The built environment and spacial form. Annual Review of Anthropology 19: 453–505. Löw, Martina 2001: Raumsoziologie. Frankfurt am Main: Suhrkamp. Low, Setha M. 2000: On the Plaza: The Politics of Public Space and Culture (1st edition). Austin: University of Texas Press. Malpas, Jeff E. 1999: Place and Experience. A Philosophical Topography. Cambridge: Cambridge University Press. Malpas, Jeff E. 2006: Heidegger’s Topology: Being, Place, World. Cambridge, MA: MIT Press. Merleau-Ponty, Maurice 1945 [1962]: Phenomenology of Perception. Translated by Colin Smith. London: Routledge. Polanyi, Michael 1958: Personal Knowledge. Towards a Post-Critical Philosophy. New York: Harper Torchbooks. Relph, Edward 1976: Place and Placelessness. London: Pion. Scheflen, Albert 1964: The significance of posture in human communication systems. Psychiatry 27: 316–321. Scheflen, Albert 1972: Body Language and Social Order. Englewood Cliffs, N.J.: Prentice Hall. Scheflen, Albert 1973: Communicational Structure. Bloomington: Indiana University Press. Scheflen, Albert 1974: How Behavior Means. Garden City: Anchor Press. Scheflen, Albert 1976: Human Territories. Englewood Cliffs, N.J.: Prentice-Hall. Schroer, Markus 2006: Räume, Orte, Grenzen. Frankfurt am Main: Suhrkamp. Seamon, David 1979: A Geography of the Lifeworld: Movement, Rest, and Encounter. New York: St.Martin’s Press. Simmel, Georg 1903 [1983]: Soziologie des Raumes. In: Heinz-Jürgen Dahme & Otthein Rammstedt (eds.), Georg Simmel. Schriften zur Soziologie, 221–242. Frankfurt am Main: Suhrkamp. Streeck, Jürgen & Jordan J. Scott 2009: Communication as a dynamical self-sustaining system: The importance of time-scales and nested contexts. Communication Theory 19: 448–467. Tuan, Yi-Fu 1974: Topophilia. New York: Columbia University Press. Tuan, Yi-Fu 1979: Space and Place: The Perspective of Experience. Minneapolis: University of Minnesota Press.

334

John B. Haviland

John B. Haviland

Xi to vi: “Over that way, look!”: (Meta)spatial representation in an emerging (Mayan?) sign language 1.

Introduction: language, space, and meta-space

Those aspects of human experience most taken for granted, most widely shared, most seemingly universal and “natural” are for many anthropologists precisely those most in need of conceptual and comparative scrutiny. For those interested in language such scrutiny often begins with the linguistic resources speakers use for talking about apparently shared aspects of human experience: kinds of people and their interrelationships (represented, say, in kinship terminologies or in systems of linguistic gender), elements of the environment (found, for example, in ethnobotanical nomenclature or in color vocabularies), and certain quasi-mathematical aspects of assumed human perceptual experience (for example, numbers, or systems of quantification and classification). Space has recently been a central focus of such comparative conceptual scrutiny.1 Assuming neither a shared conceptualization of physical space, nor some experiential construal of its mathematical or topological properties, the point of departure here is instead the fact that particular languages provide interlocutors with resources for answering questions like “Where is X?” These linguistic resources include language particular devices to indicate such notions as size, distance, shape, position, arrangement, contact, containment, contiguity, alignment, motion, direction, and velocity. An important result is that languages provide interrelated but conceptually incommensurable “frames of reference” for representing spatial relationships and, correlatively, that speakers often give variable priority to different representational devices that reflect such frames of reference. Mayan languages have been important in the typology of spatial language, partly because of the multiple and overlapping frames of reference typically employed by their speakers.2 Typological interest in Mayan spatial conceptualization is very recent, however, when compared to the very long tradition of ethnographic inquiry into Mayan spatial practices – from the physical layout of house, cornfield, or church, to the cosmological significance of 1 2

See, for example, Levinson 2003; Levinson & Wilkins 2006. See, among others, Haviland 1991, 1992, 1993, 200, 2005; de Leon 1992; Brown & Levinson 1993; Brown 1994, 2006; Bohnemeyer & Stolz 2006).

(Meta)spatial representations in an emerging sign language

335

Map 1. The township of Zinacantán, in highland Chiapas, Mexico

spatial orientations, or from the day-by-day calibration of spatial information in interaction,3 to the vast archeological and colonial record of elaborate socio-spatial organization in the Maya area.4 Work on linguistic representations of space in Tzotzil has focused on two striking features of communicative practice in the community of Zinacantán, in highland Chiapas, Mexico (see map 1). The first is the high degree of lexical elaboration in various spatial subsystems in the spoken language. The second is the co-speech gestures that give direct evidence about Tzotzil speakers’ conceptualizations of space even in the absence of corresponding spoken forms. As background to the present chapter I shall, in the next section, briefly sketch results from these two investigations of Tzotzil spoken interaction. Tzotzil, however, is not my principal concern in this chapter. If one’s notion of space is linked in part to the language(s) one habitually speaks, as comparative work on spatial language and conceptualization suggests, then examining the linguistic resources available to different languages for describing space seems a useful initial step in characterizing human spatial conceptualization. Especially compelling in such an enterprise would be a “new” language – one with no apparent direct links to other languages, which evolves over the space of a single generation of initial language users. Even more interesting would be a language which makes direct use of space itself as a semiotic medium. Consider the case of sign languages: here the spatial 3 4

See, for example, Vogt 1992, Gossen 1974a, 1974b, Hanks 1990. For example, W. A. Haviland 1966; Ashmore 1989; Ashmore & Willey 1981, Hanks 1988, 1992; Jones 1989; to cite only a few.

336

John B. Haviland

configurations of visible articulators (the hands, the head, the face, among other) are the primary vehicles for signaling communicative content. Talk about space, in such a language, is itself spatial, so that “spatial language” is necessarily meta-spatial – it uses space to describe space. My recent research has involved just such a “new” language: a nascent sign-language – here called ‘Z’ – created in a single household of Zinacantec Indians from highland Chiapas. The three deaf members of this household have never met other deaf people or been exposed to any other sign language, and the tiny community of signers, both deaf and hearing, has limited contact with any spoken language other than Tzotzil. My primary purpose in this chapter is to show how spatial language works in this first generation sign language, and to compare the spatial frames of reference of spoken Tzotzil with the (necessarily meta-)spatial devices of Z. Are they the same or different? How much of the apparent underlying spatial repertoire of Z either resembles that of Tzotzil or instead represents something invented by the signers?

2.

Space in spoken Tzotzil

In spoken Tzotzil, several linguistic subsystems contribute to spatial descriptions. A hypertrophied set of Tzotzil roots (traditionally called “positionals” in Mesoamerican linguistics) denoting shape, configuration, and anatomy facilitates – indeed, requires for felicitous speech – careful specification of the spatial character of different sorts of objects.5 Much of the topological and geometric specification accomplished in other languages by adpositions or nominal cases (Talmy 1985, Svorou 1994) falls in Tzotzil to the complex anatomical and positional semantic portmanteaux of these positional roots. Tzotzil also elaborates “body part” expressions (Levinson 1994a) which enable descriptions of spatial position via an “intrinsic frame of reference” using as points of locative reference the anatomies of objects construed as virtual bodies. The exact “body-part” distinctions involved thus represent a partially grammaticalized spatial ‘anatomy’ which can be variously applied to different sorts of object. Tzotzil additionally has an elaborate set of grammaticalized auxiliary and directional verbs which permit precise inflection of virtually all predicates with respect to trajectories and motion.6 Finally, the metaphor of an ‘up/down’ opposition, which literally refers to the vertical 5

6

See Laughlin 1975, Haviland 1992, 1994, 1994b for Tzotzil, Brown 1994 for Tseltal. See Aissen 1984, Haviland 1981, 1993b, Zavala 1992.

337

(Meta)spatial representations in an emerging sign language

axis, is conventionally extended to an East/West axis: where the sun rises is thought of as ak’ol or ‘up’, and where it sets as olon ‘down’.7 This opposition allows Tzotzil speakers to apply Levinson’s “absolute frame of reference” which uses a coordinate system conceptually independent of local terrain and landmarks for locating objects and places in relation to one another. These linguistic sub-systems are illustrated in a spontaneous dyadic interaction (videotaped in 1993 in the hamlet of Nabenchauk) in which Peter, a Zinacantec man in his eighties, describes to much a younger compadre the earliest settlement of their village as they stand in the older man’s house compound. Questions of location are naturally prominent throughout this short conversation. Consider first the use of Tzotzil body part words to describe locations. The complete meronymy for Tzotzil is complex8, but a few “part words” exemplify the general principles. The word pat is used to denote a human ‘back’ or, for example, the posterior side of some object which has a distinct anterior sat ‘face, eye’, ni` ‘nose’, or ti` ‘mouth.’9 The corresponding posterior surface is a pat; the posterior end, if there is one, is a chak ‘bottom.’ Pat also denotes the outer surface of an object that is conceived as having a yut ‘interior.’ Describing where his great grandmother settled after her husband cleared the virgin forest, Peter points toward the eastern edge of the valley where the village lies, saying (1) Intrinsic use of pat ‘back’

10

te

s-pat

nakal

yo`-bu

s-na

chikin-p’ine10

THERE residing WHERE 3E-back 3E-house name ‘She lived over in that area behind the house of the Chikin P’in family.’

7

8 9

See Gossen 1974a for an account of some ramifications of this conceptual coincidence in the Tzotzil of neighboring Chamula. Cognate words apply to a parallel distinction in the Tseltal of nearby Tenejapa (Brown 2006, Brown & Levinson 1993) although there the dominant topography seems to have produced a different conventional association: since North is topographically downhill in most of Tenejapa, there up means ‘South’ and down ‘North’. But see Polian & Bohnemeyer (forthcoming) for more details on Tseltal usage more widely. See Laughlin 1998c, Haviland 1992. What defines this anterior extremity is, as the glosses suggest, partly a matter of shape and configuration: a sat is usually a flat surface or point in a flat surface; a ni` is a projection; a ti` is a hole. See Levinson 1994 for related facts about Tseltal.

338

John B. Haviland

He conveys that his grandmother lived in an area that lies on the opposite side from the front (ti` or doorway) of the current Chikin P’in house. The intrinsic orientation of the reference object, the Chikin P’in house with its clearly identifiable parts, fixes the location of the great grandmother’s former house. Later Peter uses pat in a different way to describe the location of an old path that people from lowland villages originally used to make the journey up the mountain to the nearest market town. 11

(2) Relative use of pat ‘back’ Xi

la

ch-jelav li

be

ta

pat11 vitze

THUS EVID ASP-pass ART path PREP back mountain ‘The road used to pass on the far side of the mountain over there.’

Since a mountain, unlike a house, has no clear front side – no sat ‘face’ or ti` ‘mouth’ – the reference to “the mountain’s back” must be calculated relative to the perspective of an observer (here the interlocutors), by a Tzotzil convention that parallels that of English.12 Peter intends to say that the old path ran on the far side of the mountain from where he stands. The location is thus triangulated from (or projected onto) the mountain relative to the observers’ viewpoint. A still different use of a body part word to convey a spatial configuration occurs when Peter reminisces about the deer that once abounded in the forests surrounding the village in the early days before virgin forest was felled to accommodate settlement. (3) Lexicalized body part projecting a spatial layout te`tikil chije, te x-k’ate:t ta x-chak te`-tike wild deer THERE ASP-lying_sideways PREP 3E-bottom tree-PLU ‘Deer would be just be lying about sideways amongst the tree stumps.’ 10

11

12

Tzotzil is written in a Spanish based practical orthography. Abbreviations include 1E = 1st person ergative, 3E = 3rd person ergative, ASP = aspect, ART = article, CL = clitic, DIR = directional, EVID = evidential, PREP = preposition, PLU = plural, There is a further grammatical difference between examples (1) and (2), namely that pat is grammatically possessed in the former intrinsic use, but not in the latter relative use. See de Leon 1994 for further grammatical details. But it differs from that of Hausa (Hill 1982).

(Meta)spatial representations in an emerging sign language

339

The expression x-chak te` ‘lit., bottom of tree’ is lexicalized to mean ‘stump’ (as well as its literal denotation: the cut end of a tree trunk – the end on which the trunk could, in principle, ‘sit’), and it is partly here that Peter conveys the information that the deer are low-lying, on or close to the ground. Example (3) also illustrates the second aspect of Tzotzil spatial language mentioned: how a spatial configuration can be partly encoded via the highly elaborated inventory of Tzotzil positional roots (e.g., Haviland 1994, 1994c). The verb x-k’at-et is based on the positional root k’at ‘sideways, crosswise (predicated of a longish thing)’ which combines information about the shape of the object described with a specific configuration or disposition in space: here that objects which are relatively longish in shape (the deer) are arranged so as to run perpendicular to the reference objects (the tree stumps in the forest). The image resulting from the positional information in the verb plus the body-part modification of the reference object is of deer lounging on the ground partially obscured behind the felled forest trees. Other positional predicates in Peter’s description of the forest are generally evocative of spatial scenes. He describes mushroom hunting, where (4) Positional predicate te lam-al li tajchuch THERE spread_out ART lentinus_mushroom ‘The mushrooms covered the ground.’ Or he describes the scene after the forest was chopped down to accommodate cornfields as (5) mo:l toje, tzel-ajtik large pine heaped_up-PLU ‘Big pine trees all heaped up.’ In both cases, the positional roots give precise spatial indications: lam that the ground was an apparently unbroken blanket of mushrooms; tzel that the heap was composed of longish things in a jumble. Two other lexical systems in spoken Tzotzil systematically encode spatial information. One is the system of motion verbs, which are grammaticalized across the verbal system as both auxiliary verbs and directional particles.13 Peter illustrates the latter as he describes the original clearing of the moun13

See Aissen 1994, Zavala 1992, Haviland 1993b, 1996.

340

John B. Haviland

tainsides in his village. Motioning toward one of the mountain ridges that ring the town he says: (6) Directional particles Tz-boj-ik muyel xi to vi noxtok une ASP-chop-PL DIR:rising THUS CL EVID also CL ‘They chopped the forest all up this way, too Tz-boj-ik tal naka jvaskisetik la une ASP+3E-chop-PL DIR:coming only Vazquez EVID CL And they chopped down this way, only members of the Vazquez family, they say.’ The directional muyel in the first clause is derived from the intransitive root muy ‘ascend’ and allows Peter to add an upward trajectory to the action of chopping trees as he depicts how the early settlers worked their way up a mountain ridge. The directional in the second clause uses the root tal ‘come’ and it incorporates a deictic perspective into the scene: it was on the side of the mountain ridge toward the observers that the colonists continued felling the forest. One final aspect of spatial language in spoken Zinacantec Tzotzil is the conventionalized association between the vertical axis – denoted by the relational nouns ak’ol ‘above’ and olon ‘below’ as well as by verbs of ascending (like muy) and descending – and the East/West axis. Considerable attention is paid to the exact path of the sun, and there are strong symbolic and religious associations with the East/West axis.14 The East, where the sun rises, is ‘high’ and the west, where it sets, is ‘low’. Despite local variations in terrain, it is geographically the case that the lowland cornfields that Zinacantecs frequent, historically, lie largely to the west of the township, and that access to them has been by paths that lead inexorably westward and down. What is called ‘hot country’ in Spanish is olon osil ‘low country’ in Tzotzil; people called j’olonetik ‘lowlanders’ are those from the township’s westernmost settlements.15 And the westernmost and at one time most distant place where Zinacantecs ever used to venture – Mexico City – is still called olontik ‘the low place’ by old timers.

14 15

See Gossen 1974a. They are also called jchobtiketik ‘cornfield people’ because that is where corn grows best.

(Meta)spatial representations in an emerging sign language

341

There is sometimes tension between applying the vertical axis to the actual slope of the landscape as opposed to the east/west axis independent of local inclination. In describing macro-space, however, by ‘up’ and ‘down’ Zinacantecs invariably mean the East/West axis. Peter thus describes the former walking path from his village into San Cristóbal, the closest market center. The path made its way up to a high point just east of the village and continued eastward, descending again into the large village of Nachij. Just before reaching Nachij16 another path branched off to the north, just beyond the house of a well-known person whom his interlocutor mentions. Peter confirms that this is the place he means, placing it directly East of (although, in terms of the local terrain actually lower in elevation than) the point of reference. (7) East/West y-ak’ol s-na konkoron x-k-al-tik 3E-above 3E-house name ASP-1E-say-PLU ‘East of the house of the guy we call Konkorón.’

3.

Space and Zinacantec co-speech gesture

Although all of these spoken Tzotzil forms are frequently used in descriptions of spatial configurations, much of what we know about how Zinacantecs conceive of space comes not from their words but from their gestures. There is indirect but compelling evidence in gesture that space is inherently oriented by cardinal directions. As he pronounced the phrases in all of the examples (1)-(7) above, Peter also produced gestures, in each case supplementing the spoken spatial information with visible representations. Thus, in talking about his great grandmother’s house located (intrinsically) behind the Chikin P’in house, he points in the direction the house would have stood from his current vantage point (Fig. 1a). Placing the old path (relatively) behind the mountain, he also points in the direction he means (Fig. 1b) where the mountain itself blocks the old path on its far side. And as he reminisces about the deer lounging in the forest, he places them demonstratively on a different mountainside (Fig. 1c). 16

Notably, Peter says ta ba Nachij ‘above Nachij’ using a “body-part” word ba ‘top, forehead’ that unambiguously refers to a high point on the vertical axis and is never used to mean “East.”

342

John B. Haviland

Figs. 1a–c. Peter’s line of sight pointing

Fig. 2. Mushrooms covering the forest, trees stacked

Fig. 3. Chopping up one side of the ridge, and back down this way

More iconically he both sketches with his hand and describes with positional roots how the mushrooms carpeted the forest, and how the tree trunks were stacked. The direction of his sketching hand indexes the location involved (Fig. 2). He similarly combines an iconic representation of motion as he traces with his hands the trajectories he captures in words with directional verbs as he describes the felling of the forest, up one side of a mountain ridge and back down the other (Fig. 3). In all of these performances – unsurprisingly, since he is recounting the history of the village where he and his interlocutor are standing – Peter’s pointing gestures directly index visible landmarks. Although he is talking about past events, when the landscape was different, his gestures help his in-

(Meta)spatial representations in an emerging sign language

343

Map 2. The places Peter narrates in and around the village.17

Fig. 4. East of Konkorón’s house

terlocutor to calibrate the past geography (the space Peter is talking about) with the immediate indexical surround (the space in which the current interaction occurs). Mapping the direction of his pointing gestures on a map of 17 the village (Map 2) makes this calibration clear. When he goes on to speak about a more distant location – some 8 to 10 kilometers away, over the mountains to the East and well beyond visible range – Peter also points toward the exact place he describes, showing with a contoured hand how it is located relative to the house of the man his interlocutor mentions. Once again, it is possible to check the directional accuracy of Peter’s gesture by comparing its vector with a map of the area (Map 3), knowing roughly where the spot he mentions stands in relation to the nearby village of 17

All maps were drawn on the basis of arial photographic coverage from the area in and around San Cristóbal de las Casas, Chiapas, Mexico, taken from Google Earth, the sources for which are copyrighted material belonging to Google, INEGI, Cnes/Spot image, and Digital Globe, all copyright 2011.

344

John B. Haviland

Map 3. Map of the wider area between Peter’s village, Nachij, and Zinacantán

Nachij, directly on the old walking path into the town of San Cristóbal. Even at this distance, the direction of Peter’s pointing gesture is surprisingly exact. Zinacantec gestures can be spatially demonstrative in a different way. At another point in Peter’s narrative he describes how people from the village (often Indians who had immigrated there from other Tzotzil communities in search of land and wives) were conscripted into the Mexican army and sent off to fight in distant places. One such man had given a vivid account of battle, and Peter describes how the man was taught to shoot either from a kneeling or prone position. He has recourse to two Tzotzil positional roots: kej ‘kneeling’ and pat ‘lying on the belly with the front of the body raised.’ (8) Shooting in a kneeling or prone position kej-ajtik la ch-ak’ xi toe kneel-PLU EVID ASP+3E-give thus CL ‘They would shoot kneeling like this.’ mo`oje, patal tal ta lum no prone DIR:coming PREP ground ‘Otherwise they would lie down on the ground.’ However, he refines his postural description by acting out the positions via pantomime, showing how the soldiers were trained to kneel on just one knee (as opposed to the standard Zinacantec way of kneeling on both), and to support themselves on their arms when shooting from a prone position (Fig. 5). More striking still is a conceptually more complex gestural specification of absolute direction that Zinacantecs frequently use. Here is a characteristic

(Meta)spatial representations in an emerging sign language

345

Fig. 5. Shooting while kneeling or prone

example, which I unfortunately did not capture on film. Peter was once telling me how I should travel to reach his lowland cornfield, a day’s journey from the village where we sat. He told me to take a truck to the city of Tuxtla, from there to take a 2nd class bus in the direction of a town in central Chiapas, and to ask the bus driver to let me off at a certain named place in the countryside. “When you get off the bus”, he told me, “go that way” – pointing at a spot on the hills rimming the village where we sat, but more than 70 km as the crow flies from where I was heading. How was I to understand his instructions? Apparently what he expected was that I memorize the compass direction in which he had pointed and try to reproduce it when I found myself by the side of the road after the bus let me off. To understand the complexity of such a pointing gesture, consider how another Zinacantec, Martin, who spent many years traveling between the village of Nabenchauk and the distant town of Cancún told me about the route he used to follow. I filmed his original description in 1991 while he sat in my yard in Nabenchauk, and again ten years later, from the quite different vantage point of the town of San Cristóbal, as part of a systematic study of cospeech gesture.18 At that point he had not made the trip in nearly a decade. Given how he was oriented by the compass on both occasions, one can track with precision the directions of Martin’s pointing gestures. A striking feature is the consistent orientation of his pointing gestures, which suggest a highly accurate point-by-point absolute recalibration of the compass directions he took at each major juncture of the trip. 18

See Haviland 2000d, 2005 for more detailed treatment of this route description and its gestures.

346

John B. Haviland

Map 4. Turn-off to Chetumal

Fig. 6. “Chetumal is this way”

Here is one example. At the place on Martin’s route where the road approaches the coastal city of Chetumal, the main highway bypasses the city, which lies to the east and slightly south of the intersection, and it there turns northeast toward Cancún. (See Map 4.) In the 1991 film, Martin describes arriving at the Chetumal turn-off and silently indicates the trajectory of the turn-off road, branching away from the main highway. He then explicitly locates where the city of Chetumal is, flipping his right hand slightly back to the right as he says xi ta xkom chetumal xi toe (‘Chetumal is over that way’). If we interpret this gesture in compass terms, it accurately places Chetumal slightly south of east, at about 100° on a 360° compass with North at 0°. In the 2001 narrative, with a brief turn of his hand off to the southeast, Martin also notes that Chetumal lies off the main trajectory of his route. The corresponding images from the two video recordings are shown in Fig. 6. Both gestures appear to place Chetumal in almost exactly the same compass direction from the turnoff. Understood as ‘absolute’ reckonings of compass directions from the imagined road junction, his gestures reflect a consistent sense of orientation and direction which receives similar expression across the decade-long span (and his different body positions) between the two different narrations.

(Meta)spatial representations in an emerging sign language

4.

347

Narrated and narrating spaces

It is worth reflecting on the conceptual underpinnings these pointing gestures seem to imply. Because gesture about space itself uses space as its communicative medium – it is, in this sense, also “metaspatial” – it seems important to distinguish at least three conceptually different kinds of ‘space’ involved in the Tzotzil gestural practices we have been examining. Jakobson (1957:390), in his classic elaboration of the basic grammatical categories of the verb, distinguished “1. speech itself (s), and its topic, the narrated matter (n); 2. the event itself (E). and any of its participants (P), whether ‘performer’ or ‘undergoer.’” He continues: “[c]onsequently four items are to be distinguished: a narrated event (En), a speech event (Es), a participant of the narrated event (Pn), and a participant of the speech event (Ps), whether addresser or addressee.” Because events generally involve entities arranged in space, one could extend Jakobson’s classification to include both a narrated space (Sn) and a narrating or speech-event space (Ss) within which the narration takes place. The former is the space in which narrated events putatively occur (and which thus may be at least selectively represented in the narration), and the latter is the space of the speech event itself, physically accessible to participants as they talk. As in the case of the other entities Jakobson distinguishes, these spaces are conceptually different: the narrated space is in a clear sense imagined and essentially partial, as it only acquires details as the narration and interlocutors’ own knowledge progressively provide them. The narrating space, within which the speech event occurs, is largely presupposable from the immediate surround of the speech act participants, and partly brought into some kind of correspondence with the narrated space as elements of the narration highlight local places or entities to create such correspondence. For example, Peter locates his long deceased grandmother’s house (part of the narrated space) relative to a contemporary local house known to his interlocutor (in the wider narrating space), to which he can point as shown in Fig. 4. Various mechanisms, including use in narration of immediately perceivable local landmarks, or coincidence of compass directions, can superimpose narrated space on local speech-event space or otherwise calibrate the two conceptually different spaces. In previous work (Haviland 1993a), I have appealed to a further “interactional space” (Si) – related in some ways to what Kendon (1990:211) long ago called “o-space” – which is distinguished from the narrating or speech-event space (Ss) by its centrality not to the narrated events or to the speech-event surround in general but to the specific mutual interaction of the participants

348

John B. Haviland

in the speech event: it is the immediate shared space of the interaction and mutual attention, within which interlocutors usually gesture. (In the case of sign language, to which we shortly turn, it is also the space where signs are generally performed.) A distinguishing conceptual feature of Si is that, just as the narrated space may be independent from the narrating space, the interactional space can also be independent or decoupled from the speech-event space. It is “free”: a space created by and for the immediate interaction. When Peter half kneels to illustrate how the narrated protagonist fired his rifle (Fig. 5), he demonstrates the position in interactional space. However, his spatial use of interactional space is at least partly arbitrary in the sense that exactly where he kneels is irrelevant to the performance (and his interlocutors must understand this). 4.1

Z, an emerging language

Distinguishing different conceptual spaces in this way underlines how space is, first and foremost, a discursively constructed, linguistically structured category of interpersonal interaction. Whatever origins spatial understanding may have in the perceptual capacities and cognitive development of individual human beings, central and important features of a conceptualization of space, on this view, emerge from the way people talk about and otherwise represent spatial relations in their ordinary interactions. As I mentioned at the outset, such a perspective obviously lends special interest, in the comparative study of spatial conceptual systems, to a new language. If a community of speakers relies on its language to structure space, what happens when the linguistic resources for representing space are only beginning to emerge? How, in such a case, does spatial conceptualization come to express itself ? A first-generation sign language, Zinacantec Family Homesign (‘Z’), emerging among five young adults in the township of Zinacantán, Chiapas, México allows a unique view of how spatial language grows out of interactive and social practices. The three deaf and two hearing members of this miniature language community have grown up with no interaction with other deaf people and limited contact with any language other than spoken Tzotzil, in a small and relatively isolated village of peasant Indians. Their communicative system uses a largely visual/manual modality and is the complex result of their interactions with each other, with Tzotzil speakers more widely, and their own processes of invention and innovation. Because of the extensive prior work on spatial representation in both spoken Tzotzil and also cospeech gesture, it is of special interest to see how a sign language emerging in this communicative context provides raw materials for creating linguistic

(Meta)spatial representations in an emerging sign language

349

representations of space, and how those of this manual modality compare with parallel Tzotzil resources. In 1976 a daughter, Jane, was born to my ritual kinsmen Mario and Rose, who already had three older living daughters. Jane never began to speak, although she was sent to school for part of a year, after which she remained at home, like many other Zinacantec girls her age. Six years later a brother, Frank, was born, and he, too, failed to begin to speak. Both children were labeled umaʔ ‘dumb’ – a word which in Tzotzil has the same pejorative polysemy as its English gloss – and raised more or less exclusively by their mother and older siblings. In 1986 another daughter, Terry, was born, and although she also remained silent until she was well over two years old, she suddenly began to speak Tzotzil, as though the silence of her two nearest siblings had until then left her unmotivated to talk. It was only at this point that medical diagnosis revealed that both Jane and Frank were profoundly deaf. Finally, in 1988 – when his older deaf sister was nearly thirteen years old – a youngest sibling, Will, was born, also deaf. What thus presumably began as a typical “homesign” system developed for mutual communication by Jane and the rest of her hearing family was over the span of a decade extended to a medium of communication for the three, and then, four siblings who used it as their only means of interaction, with each other and to a lesser extent with the other hearing members of the family. Added to this mix, five years later, was a niece – Rita – who, although hearing, grew up largely in the company of her signing aunts and uncles and thus became fluent in their emerging sign language as well. I have known all of these children – now young adults – since they were born. Their unique linguistic circumstances cried out for systematic investigation, despite the children’s reluctance to sign in public and the general stigma of their deafness. Mario, the father, my old friend and compadre, was also a major collaborator in research on Tzotzil ritual language and cospeech gesture. When in 2008 the work on an emerging Bedouin sign language by my UCSD colleague Carol Padden and her associates19 inspired me to undertake research on Z, Mario and his children readily agreed.20 By then 19

20

See for example Sandler, Meir, Padden, and Aronoff 2005; Meir, Padden, Aronoff, and Sandler 2007. This material is based upon work supported by the National Science Foundation under Grant No. 0935407, administered by the Center for Research on Language (CRL) at UCSD. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. My principal debts are to the Z signers themselves, acknowledged here by their pseudonyms: Jane, Frank, and Will, as well Terry, Rita, and Victor.

350

John B. Haviland

Fig. 7. Genealogy of the extended household where Z is spoken

Jane had her own hearing son, Victor, now a 5-year-old bilingual signer and Tzotzil speaker, who along with a younger cousin represent the beginning (and perhaps also the end) of the second generation of this miniature Z speech community. (See the genealogical chart in Fig. 7.) Z represents a functionally effective means of communication for the signers in this family, allowing them to participate in apparently all the activities normally facilitated by spoken language in a Zinacantec household. Z signers issue and respond to commands, ask and answer questions about both facts and speculations, recount past events, participate in decisions, plan for the future, tell stories, argue, evaluate, joke, ridicule, criticize, and scold. Despite the very shallow history of their conjointly developed system of signs, despite the unusually high level of presupposable common ground that results from the intimately shared biographies of the tiny Z signing community, and despite the relative isolation in which the Z signers live compared to hearing Zinacantecs, they appear to have no more difficulty than other Tzotzil-speaking Zinacantecs in dealing with (and naming) things and people both familiar and unfamiliar, and generally in negotiating their lives, practical and social. Strikingly, for me as a linguistic anthropologist concerned with the presumed central role of language in cultural transmission, the Z signers seem wholly Zinacantec, in what they know, what they like, what motivates them, how they act, how they move, and how they interact. The central research questions in ongoing work address the structural properties of Z in the face of this evident functional efficacy.

(Meta)spatial representations in an emerging sign language

351

Fig. 8. “at 4:30”

As mentioned, the issue for the present chapter, however, is smaller. How is space represented (and hence, at least as an initial approximation, construed) by the first generation of users of an extremely young language like Z? What resources for communicating about space have the Z signers adopted or invented? 4.2

Space in spontaneous Z conversation

I introduce the Z signers in the context of spontaneous conversation, to illustrate both the general character of the language and some of its spatial resources. Here are Memo and Frank, in a typically competitive interchange for young Zinacantec male siblings. They are talking about which of them will be asked to accompany their brother-in-law who makes periodic trips to a distant market town on the Chiapas Pacific coast to sell flowers, mostly for specific fiestas. These are among the few outings the boys make away from their home village, and they are prized occasions both to escape from quotidian routines and to earn money. The general tenor of the exchange is mutual insult: each brother boasts that he is more likely to be invited on the next trip because the other brother is “useless”. Consider Frank’s first few utterances as he introduces the topic by mentioning that the brother-in-law had just left for the coast earlier that day. He says, “At 4:30 the truck set out and went (to the coast). Dad (will go) tomorrow, he didn’t (go) now”. Different parts of the utterance are illustrated in still frames from the video recording in the following figures In Fig. 8 Frank points to his left wrist (as if at a watch) to refer to the time of day, then displays the number 4 on his right hand, and adds the half hour by drawing his left finger across his right palm. He goes on to say that that was the hour when the truck was loaded and ready to set out (Fig. 9). He signs that the truck went to the coast by performing a ‘go’ verb: he points with his right thumb (see Fig. 10) placing the final destination a long

352

John B. Haviland

Fig. 9. “the truck was loaded and ready to go”

Fig. 10. “it went there”

way away (signaled by the height of the movement arc21) and slightly to the west of south (signaled by the compass direction of the pointing movement). Frank goes on to add that their father (whose proper name in Z is illustrated in Fig. 11: a bent hand with the fingers together held in front of the belly – an iconic reference to the older man’s prominent paunch) did not leave (using a conventionalized negative hand wave seen in Fig. 12), but would be going in the next couple of days (shown with the conventionalized “tomorrow or the day after” sign – rotating the right index finger in several circles oriented away from the body in Fig. 13). At several other points in the conversation the boys make references to trips to the coastal town in question. The form of the signing anticipates a general finding about Z, already evident in the ‘go’ verb illustrated in Fig. 10: signs typically “absolutely” anchor locations in the narrating space. Thus, when Will mocks Frank for being left behind, Frank retorts that frequently Will himself is not invited on the flower selling trips; he illustrates the latter with another pointing gesture, somewhat awkward to perform from his 21

Kendon 1980: 110 describes the same kind of usage in Enga sign. See Calbris 1990, Haviland 1993 for gestural uses of a similar convention to denote distance.

(Meta)spatial representations in an emerging sign language

353

Fig. 11. “Dad”

Fig. 12. “He didn’t go yet.”

Fig. 13. “(He’ll go) tomorrow.”

seated position facing north-northeast: he traces a high backward arc to show the southern trajectory of the trips. (See Fig. 14.) Will continues mocking, by saying to his brother, “Just wait, you’ll see [see Fig. 15] – I WILL be going to the coast”. Will also performs the motion verb with a dramatically exaggerated arc (Fig. 16), ending with a triumphant flourish and grin at his brother. It is worth highlighting the (to many speakers of European languages) almost uncanny directional accuracy of these pointing gestures. In Map 5 I

354

John B. Haviland

Fig. 14. Frank, seated facing north, signs “go to the coast”

Fig. 15. “Just wait!” (Will on the left, Frank on the right)

superimpose on a map of Chiapas the rough directional vector of Frank’s finger point and his and Will’s later more demonstrative renditions of the same journey as they were performed from the signers’ house in Zinacantán (see the inset on Map 5). Even with a very approximate reckoning of the direction indicated, both Frank and Will’s pointing gestures seem to pick out the section of the Chiapas coast that includes the town of Huixtla, where in fact their brother-in-law does go to sell flowers. In my understanding the vector that forms part of the sign for ‘go’ in Z does not depend on some arcane and mysterious directional acuity on the part of the Z signers, who have rarely traveled as far as the Chiapas coast in the course of their lives. Instead, one must consider such dead reckoning of

(Meta)spatial representations in an emerging sign language

355

Fig. 16. Will: “I WILL go (a long way that way).”

Map 5. Map of Chiapas highlands and coast, Zinacantec household inset

location from a given origo, together with a set of visible devices to show relative distance and other aspects of intervening terrain, to be part of the wider convention in Zinacantec co-speech gesture to locate even distant places in terms of absolute compass directions from the current speech

356

John B. Haviland

origo, as illustrated above in Peter’s gesture in Fig. 4 and the corresponding Map 5.22 Such absolute locations (calculated relative to some speech event location) effectively serve as a proxy for Z place names.

5.

Pseudo-experiments about space and place

As a novice student of sign languages, I have borrowed and invented tools for studying Z. I have relied on my previous knowledge of Tzotzil (which is, of course, my conduit to Z through the glosses and interpretations offered by the two hearing signers, Terry and Rita) and of Tzotzil gesture (which offers certain tools for describing Z sign form), trying to make only the most austere assumptions about how Z might work. Most of the Z signing I describe in this paper was elicited in semi-controlled tasks, in which one or two signers describe a photo or short video clip to other signers, who are in turn asked to select a matching photograph or video frame from an array. The descriptions and accompanying clarifying discussion (as well as subsequent critical commentary in Z) are filmed, transcribed, glossed into Tzotzil, and analyzed. Using such pseudo-experimental eliciting techniques has both advantages and defects, obvious in what follows. Matching tasks with individual objects, sometimes distinguished only by color, size, or shape, proved trivial for Z signers. For example, when presented with a signed description of a single object marked from an array like those in Fig. 17, signers had no difficulty picking the corresponding item from a differently arranged array. The result suggests a well-developed conventional lexicon for Z, as well as resources for denoting size, shape, color, etc., and for creating nonce descriptions of novel objects. To elicit spatial descriptions I asked Describers first to describe photographs of specific known local places as well as from unfamiliar sites so that Matchers could pick out the corresponding photos from an array. I then asked all parties involved to tell me where the place was if they knew. These tasks were carried out in different physical locations, sometimes in the signers’ home in the village of Zinacantán, sometimes in my house in nearby San Cristóbal, both places whose locations and orientations are precisely known.

22

From my earliest days in Zinacantán in the mid 1960s, when people ask me where I am from, they routinely request that I show by pointing on the horizon where my homeland lies, or where my current residence is.

(Meta)spatial representations in an emerging sign language

357

Fig. 17. Arrays of objects

5.1

“Absolute” dead-reckoning

Given the apparent use of “absolute” dead-reckoning in Ss as a conventional part of naming known places, it should come as no surprise that when the Z signers want to refer to identifiable places, they do so by pointing in the ‘correct’ direction calculated from their current locations. For example, to describe a picture of the Chamula market from the vantage point of my house in San Cristóbal the signers pointed as in the following illustration (Fig. 18), where Frank, seated on the right, faces almost directly west. (The pointing gestures, though combined in one composite illustration, are all taken from different moments on the video.) Will, Frank, and Jane all appear to indicate a direction a bit north of west. By contrast, while carrying out a similar task while seated in their house compound in the village of Zinacantán, Will and Terry indicated the location of the Chamula church as shown in Fig. 19.23 Drawing the rough vectors thus indicated from these two different vantage points onto a map of the region that includes the Z signers’ house in Zinacantán, the researcher’s house where the first experiments took place, and 23

In the right hand panel shown in Fig. 19, Will is already retracting his hand from the apex of his pointing gesture. It is important to note that there is more to these locational signs than the vector of direction: different hand shapes are involved; the arc of the gesture indicates something about distance and visibility; gaze is sometimes engaged, often ‘sighting’ along the pointing limb; and, crucially, different movements of the hand often seem to suggest something about the intervening terrain. For example, Terry’s gesture on the right frame of Fig. 19 involves a twirling toss of the hand clockwise (from her point of view) and forward, indicating that from where she sits the Chamula church lies on the other side of the high mountain ridge along the northern edge of the valley of Zinacantán. Detailed treatment of these formational details must await another occasion.

358

John B. Haviland

Fig. 18. The Z signers point to identify a picture of the Chamula market.

Fig. 19. Locating Chamula from Zinacantán

the center of Chamula (Map 6) shows that the locating gestures do in fact converge on the intended location. Very similar directional convergence can be observed for other ‘named’ locations in the Z repertoire, including the signer’s natal village of Nabenchauk or the lowland state capital of Tuxtla Gutierrez, both of which lie considerably farther from the immediate horizon. For example, I asked the signers to describe the picture shown in Fig. 20. Jane, sitting in my house in San Cristóbal, describes it as a bucket containing a lemon grass plant (Fig. 21, left panel) which her mother (Fig. 21, right panel) brought from Nabenchauk (Fig. 22 right panel) to their house in Zinacantán (Fig. 22 left panel). The example illustrates clearly three quite different formational principles in Z conventionalized signs: the sign for ‘lemon grass’ – which also means ‘coffee’ or another hot drink – is an arbitrary (though iconic), established, and highly portable convention based pre-

(Meta)spatial representations in an emerging sign language

359

Map 6. Map showing signer’s locations their pointing directions, and Chamula

Fig. 20. A bucket holding a lemon grass plant outside Jane’s house

sumably on both the image of steam rising from the cup and the beverage’s strong smell. (It involves waving a flat ‘5’ hand up and down in front of the nose.) The proper name for the signers’ mother, ‘Mom’, is a somewhat uncomplimentary reference to her prominent belly. The “names” for the towns of Zinacantán and Nabenchauk are based on inferences from a pointed direction which must itself recalculated on every occasion of use from the current speech event origo. (See Map 7 which shows the rough vectors of Jane’s pointing gestures in Fig. 22 from the vantage point of where she was seated.) There seems little doubt that the gestural convention in Zinacantec Tzotzil of locating named locales on the horizon has been incorporated into Z as a formational component of locative signs: both place names for known places, and also locations attached to other sorts of entities. Successful use of such a convention requires both dead reckoning skills and strong inferential intuitions coupled with geographic awareness on the part of interlocutors.

360

John B. Haviland

Fig. 21. Jane signs “hot drink” and “Mom”.

Fig. 22. Jane signs “Zinacantán” and “Nabenchauk” from Haviland house.

Map 7. Haviland house, signers’ home village, and village of Nabenchauk

Maintaining such geographic awareness clearly requires reinforcement and depends on collaborative practices among signers and Tzotzil speakers alike. When the Z signers were unable to identify a pictured place they often indicated their perplexity by pointing in several different directions with an accompanying shrug: “I wonder where that is.” During one of these quasi-experiments in the village of Zinacantán, the signers’ father was also puzzling

(Meta)spatial representations in an emerging sign language

361

Fig. 23. Agua Azul “on the road to Palenque”

Map 8. From Zinacantán to Agua Azul

over such a photograph. When I told him it depicted a place he knew by reputation – the famous waterfalls at Agua Azul – he remarked “That’s on the road to Palenque” and in a seemingly automatic and unconscious way flipped his arm out in a rapid high arc in the correct direction (which, as it happened, lies just clockwise from the angle toward Chamula to the northeast, although Agua Azul is considerably farther away, s. Map 8). Almost certainly related to this use of an absolute spatial frame of reference is a Z convention for talking about time. Frank and Will, who are familiar with watches, tend to name the hours with numbers shown on the fingers. Jane on the other hand often shows the hour with absolute gestures, pointing to an idealized position of the sun in the sky. For example, in a spontaneous conversation about her favorite afternoon soap operas, she once

362

John B. Haviland

Fig. 24. Jane: Will we finish late?

Fig. 25. Noon and one o’clock

asked her sister Terry what time she thought they might return home (after an eliciting session). She asked whether it would be late: first by pointing to her left wrist, and then pointing at the afternoon sky to the west (Fig. 24). She went on to explain that the two television programs she was interested in started at noon and at 1pm, in both cases using a demonstrative pointing gesture at an idealized solar trajectory overhead (see Fig. 25). Using the distinction between Sn and Ss introduced above, it is worth considering in which conceptual space these absolute dead-reckoning pointing gestures operate. Where or at what they are directed? When signers point at a visible landmark or local place, the gesture seems to draw on direction and location in the narrating or speech event space to supply a referent in the narrated event. Pointing at a more distant referent seems essentially similar: it relies on the location of things and places in the narrating space, widely construed, to supply narrated referents. Alternatively, such pointing relies on a convention that the narrated space, which is projected both from these

(Meta)spatial representations in an emerging sign language

363

pointing gestures and from any other narrative elements which allow interlocutors to imagine narrated events, must be superimposed over narrating space. at least with respect to cardinal directions. Both spaces, that is, are identically oriented. (This is not, as far as I can tell, a convention of pointing in my own native English-speaking narrative tradition.) Indicating points along the trajectory of the sun to denote times of day suggests that the latter interpretation – a conventional oriented lamination of Sn on top of Ss – does better conceptual justice to the facts. Pointing to the place where the sun would be at noon – when it is not actually noon – seems to instruct an interlocutor to imagine another time when the sun would actually be where one is pointing in the here-and-now, a transposition that resembles Peter’s superimposing his great grandmother’s now long defunct house onto a location projected form the current Chikin P’in’s house of the moment, in Fig. 1 above. 5.2

Intrinsic, relative, and absolute frames of reference mixed together

Dead reckoning can also be used by the Z signers to argue about location, and it is combined interestingly with other signs that rely on non-absolute frames of spatial reference. Consider the interaction that followed Frank’s description of the church front pictured in Fig. 26. The church has a distinctive set of sculpted and painted arches above its door, which Frank sketches in the air as he begins his description (Fig. 27), adding that it is a church (Fig. 28), and he locates it in direct line of sight from where he sits. He is facing west as he signs, and the Guadalupe church he identifies is directly visible south-southwest from where he sits in my house.

Fig. 26. The church in the Cerrillo square

364

John B. Haviland

Fig. 27. Frank sketches the design of the church front.

Fig. 28. Frank signs ‘church’ (by crossing himself quickly).

In fact, Frank identifies the church several times as the church of Guadalupe, each time by pointing directly at the church, sometimes in a casual unmarked way, sometimes with just a gaze and head tilt (Fig. 29), and later – when he is challenged on his identification – in much more demonstrative ways (Fig. 30), which include facial inflection. Frank’s interlocutors, who are sitting across a table from him (and thus are oriented more or less facing East), recognize that he is describing a church front, but they (rightly) dispute his identification. They claim that the picture shows a different church almost directly west of them in the plaza of the neighborhood called Cerrillo (see Map 9). They signal their disagreement (using negative finger waves – see Fig. 31) and propose their alternate identification with a variety of pointing gestures (Fig.s 32 and 33). Frank ultimately concedes that they are right, in part by pointing at them (Fig. 34). Strikingly, although all the signers here use absolute dead reckoning to indicate location (and as usual their pointing gestures are carefully calibrated from the origo of the speech event), a very different spatial frame of refer-

(Meta)spatial representations in an emerging sign language

Fig. 29. Frank points casually at the Guadalupe church.

Fig. 30. Frank points at the Guadalupe church in more marked ways.

Map 9. Map of the location of the right and wrong churches

365

366

John B. Haviland

Fig. 31. Will and Jane contradict Frank with finger waves.

Fig. 32. Memo points to Cerrillo church.

Fig. 33. Terry points to Cerrillo church.

Fig. 34. Frank concedes that his interlocutors are right.

(Meta)spatial representations in an emerging sign language

367

Fig. 35. Will asks if the church is seen from the front.

Fig. 36. Frank answers that the picture does show the church front.

ence is implied by another aspect of their description of the stimulus picture in Fig. 26. Just after Frank begins to sign, Will asks him whether the picture shows the church from the front. Frank replies that it does. Both Will’s question and Frank’s answer use a distinctive two-handed pushing gesture designed apparently to suggest the perspective of an observer looking at the front surface of an object (Figs. 35 and 36). It is clear that the two men, although using the same signs, have oriented them not absolutely but relative to their own perspectives as observers. Since they face each other, the two ‘pushing’ vectors are in fact performed in exactly opposite directions, as the illustrations show. Even more striking is Will’s use of a similar sign when he turns to Terry to repeat that the picture in question shows the front of the Cerrillo church. He signs that they are looking at the front of the church, but he is now turned toward Terry and his ‘pushing’ gesture now goes from him out toward her. He continues with an oriented finger point in the absolute direction of the church itself (Fig. 37). In terms of Levinson’s typology of frames of reference, the ‘front’ sign seems to rely on a “relative” spatial frame of reference, calculated deictically

368

John B. Haviland

Fig. 37. Will tells Terry that the picture shows the front of the Cerrillo church.

from the perspective of an observer. There seems a clear affinity between this kind of unanchored or interactionally anchored sort of directional vector and the use of what I have called Si, interactional space, illustrated above in Fig. 5, to which we shall turn again at the end of this chapter. The “free” or perspectivally anchored ‘front’ gesture is immediately followed by another pointing gesture which requires, for its interpretation, absolute reckoning of the location of the church in Ss. Z signers’ spatial representations in fact make extensive use of a relative frame of reference that requires projection from an observer’s viewpoint. Although some of the spatial task I asked Z signers to perform were relatively easy for them, certain tasks repeatedly confounded their efforts to achieve a match, apparently because the tasks required certain conceptual transpositions at which the signers are not practiced and which Z provides few ready-made tools to facilitate. Fig. 38 shows a simplified version of the stimuli in one such case which, despite a deeply flawed design, revealed interesting aspects of Z spatial resources. One must first imagine the spatial layout of the task, with the Describer sitting on the right to describe the picture with a thick border. The Matcher, seated on the left, must pick the corresponding picture. If it seems obvious to the reader that the left hand bottom picture on the left is the correct match (“the same picture”), consider a Describer, seated on the right and physically facing West, who uses an absolute frame of reference and describes his picture as (for example) “two animals facing East” or, alternatively who says “the animal directly facing me is on the south”. Which of the Matcher’s pictures would now be the correct match? The actual arrangement of signers for this particular description, shown in Fig. 39, further complicates matters because the Describer’s slide was projected vertically on a computer screen, whereas the Matchers were presented with an array of printed photographs arranged horizontally on the table, requiring a further transposition of perspective.

(Meta)spatial representations in an emerging sign language

369

Fig. 38. Candleholder matching task

Fig. 39. Rita and Terry match, and Jane, holding Vic, describes.

The picture that Jane sees is reproduced the way it appears to her in Fig. 40. The two clay objects have already been identified by the signers as candle holders, in the shape of small animals. The two matchers, confronted with an array of nine different photographs of the same two clay candle holders in distinct configurations, ask Jane to tell them how the animals are oriented. Jane first shows that the figures are oriented straight back from her perspective (i.e., as she sits, on an East-West line) by tracing a straight vector with her flat palm, forward and upwards (Fig. 41). She goes on to sign that both figurines are facing her (Fig. 42). In order to add still more detail, Jane – who does not have access to the

370

John B. Haviland

Fig. 40. The stimulus picture as it appears to Jane.

Fig. 41. Jane signs “that way, straight”.

Fig. 42. “Both facing this way”.

whole array of pictures Terry and Rita are looking at – elaborates a bit further. She notes that while the figurine on the left is facing straight toward her, the figurine on the right is angled slightly outward (Fig. 43), clearly the result of fairly close observation of the original stimulus picture in Fig. 40. Later the signers have recourse to drawing the vectors on the table top with their hands (Fig. 44).

(Meta)spatial representations in an emerging sign language

371

Fig. 43. Jane shows that one figurine is angled slightly.

Fig. 44. Jane sketches the orientation of the figurines on the table top.

Fig. 45. The first (wrong) picture chosen by the Matchers.

How the matchers understood Jane’s description can be inferred from their first (mistaken) choice of a ‘matching’ photograph, shown in Fig. 45. On the basis of Jane’s signs alone it does not seem possible to decide whether her description involves a relative frame of reference, in which she calculates direction relative to her own observer’s perspective, or whether

372

John B. Haviland

Fig. 46. A sample configuration of farm animal toys

she is locating the pictured figurines in an absolute space in which they are facing not toward her but toward the East. (An intrinsic frame of reference is also implied, in the sense that the description appeals to how the figurines are facing which depends on their own intrinsic anatomies.) The Matchers’ misconstrual of Jane’s description is also ambiguous between an interpretation based on absolute directions (which would instruct them to look for a picture in which at least one of the figurines can be understood to face East), or an observer-relative frame of reference in which the matchers fail to recenter their perspective to that of the describer (who is facing in exactly the opposite direction).24 In another type of task, slightly less constrained than the previous one, the Describer was shown a photograph of an array of small plastic farm animals. The Matchers were given the actual toys themselves and were seated behind a screen so that the Describer could sign to them but not see their workspace. Their task was to follow the Describer’s instructions in order to arrange the toys according to the model in the picture. Fig. 46 shows one such stimulus photo which Frank described to Jane, Will, and Terry, also in a face to face configuration. Frank is looking at the photograph, but his view of the actual of toy animals the Matchers must manipulate is obscured by a screen (Fig. 47). In this task is the Matchers must construct a real array of toy animals, directly on the table in front of them, in Ss (and perhaps simultaneously in Si). It seems clear that Frank’s instructions are both intended and interpreted to involve absolutely oriented directions. He begins by specifying two animals: 24

Since I persisted in treating only the ‘same’ photograph as the right answer, the Matchers in this case were very frustrated and resorted to guessing until I showed them the picture that Jane was describing, which allowed them to see how my frame of reference defined the task.

(Meta)spatial representations in an emerging sign language

373

Fig. 47. Frank describing the turkey to the Matchers.

Fig. 48. Frank describes two animals, both facing northeast.

the bluish turkey and the large rooster with a red crest and a blue tail. He then describes how they are to be arranged on the table. Lifting two fingers, representing the two Figures, he turns back over his right shoulder and points both fingers in that direction, showing a northeast vector with his right hand (see Fig. 48). In response to Will’s question about which animal is on which side (Fig. 49), Frank places the turkey on the south (Fig. 50). Frank then uses another striking devices to describe the spatial relationship between the turkey and the rooster. Asked specifically about the turkey’s position, he first points in the same direction to show how the turkey ought to face. He then extends two fingers on his right hand, points them both back in the desired direction, and then indicates (by grabbing it with his other hand) that the southernmost finger corresponds to the turkey (see Fig. 51).

374

John B. Haviland

Fig. 49. Will asks which side is which.

Fig. 50. Frank puts the turkey on the south.

Fig. 51. Frank locates the turkey relative to the rooster.

In this task, anchored in the shared and absolute orientation of local space, Frank’s instructions resulted in an almost perfect match between the Matchers’ toy configuration and how Frank himself wanted the figurines to be arranged. At the end of the task Frank was allowed to rearrange the toys as he wanted, and he made virtually no changes to what his interlocutors had proposed.

(Meta)spatial representations in an emerging sign language

5.3

375

Left and right

Spoken Tzotzil I believe makes virtually no use of a left-right coordinate in describing either location or direction.25 In the texts and transcribed conversations that many researchers have collected over the years I have found only one clear case in which Tzotzil speakers use a left/right expression to describe direction, and this in unique circumstances. In the diaries of two Zinacantecs taken on a visit to the United States by the preeminent Tzotzil lexicographer Robert Laughlin, on one occasion when the travelers are totally lost they describe coming to a crossroad, and for want of any other criterion, choosing the right fork. Then we went out again the next day. We didn’t know which road to take. We came to two roads. We took the one that went to the right.26 Then we saw that we had just come back to the place where we started (Laughlin 1980: 94).

Apparently only in describing such a totally disoriented state would Zinacantecs resort to using a left/right coordinate, perhaps to suggest its total arbitrariness. Nonetheless, in trying to match photographs of places the Z signers clearly do seem to make reference to a deictically centered right-left distinction, although, as in the case of the clay candlestick holders mentioned above, there is also evidence that it is difficult for their interlocutors to adopt the speaker’s point of view rather than sticking to their own. Here are two slightly different sorts of example, one involving an unrecognized place and the other a well-known and absolutely oriented locale. In describing the picture shown in Fig. 52, which shows a San Cristóbal street that leads to a church on a hill, both Frank and Jane mention the church. (See Jane signing ‘church’ in Fig. 53 and Frank signing that it is barely visible at the far end of the street in Fig. 54). Both signers also point out that there are many cars on the street (using both hands as if turning a steering wheel – Fig. 55). To show that the cars are on the left side of the picture, but that they are all parked facing down the street (i.e., on the right hand side of the street coming down from the church) the signers resort to slightly different techniques. 25

26

Brown 2006: 270 writes of a closely related neighboring language that “[t]here is no relative system available in Tenejapan Tzeltal based on oppositions for which the projections from the body provide a coordinate system”. We have already seen above (see example 2) that Tzotzil does employ a projected, deictically construed directional construction using the word pat ‘back.’ The Tzotzil says ta batz’i jk’obtik ‘on our right hand.’

376

John B. Haviland

Fig. 52. Stimulus picture of street with Guadalupe church in background

Fig. 53. Jane signs ‘church’.

Fig. 54. Frank signs ‘far that way’ and ‘small’.

(Meta)spatial representations in an emerging sign language

377

Fig. 55. Jane and Frank both sign ‘car.’

Fig. 56. Jane shows the left side of the road (three times).

Fig. 57. Jane signs “coming this way”.

After signing the street itself with her right hand, Jane singles out its left hand side (from her point of view) by tracing a vector forward with a flat hand, palm inward, thumb perpendicular to the fingers – a movement that she repeats three times (Fig. 56), immediately after mentioning the cars. She goes on to sign that the cars are facing toward her (Fig. 57). In a more demonstrative way, Frank also mentions the cars, then signs the street itself (and both its edges, by using both hands to sketch the street’s vector moving away from him), and then turns his body so that with his right

378

John B. Haviland

Fig. 58. Frank signs ‘street’ and then ‘down the right side’.

hand he can mirror both the side of the street and the direction where the cars are parked (see Fig. 58). It might be possible to think that the signs here still try to preserve the absolute cardinal directions of the scene itself, rather than a body-centric deictic projection. The photograph in Fig. 52 shows a street that travels west to east, with the church of Guadalupe far at the eastern end of it. Although in the end they could not identify which actual church was pictured, both Frank and Jane are surely aware of the convention that generally churches throughout Mexico are built with their main doors to the west and their altars to the East. However, from where they are actually seated in their house, their gestures run in exactly the opposite directions: they place the church slightly north of west, and the cars are portrayed as running in a direction that is actually east southeast. Since we have already seen that the Z signers are scrupulous in placing known locations more or less exactly where they lie on the horizon, more striking still is the signers’ description of the photograph in Fig. 59, which shows a simple flat-roofed house which both signers recognized as being just up the street from the vegetable stand their sister operates in the town of San Cristóbal. The shop is a place they visit frequently, and they have no trouble in dead-reckoning its location from where they sit in their house. Both Frank and Jane begin their descriptions by mentioning the vegetable stand (Fig. 60). Jane continues by tracing a vector corresponding to the road on which both shop and house lie, moving right to left, and placing the target house on the left (Fig. 61). Frank is again more forceful in his signing. Having first mentioned the vegetable stand, he “places” it out in front of his body to the right with a well-defined hand whose fingers are bunched and slightly bent. He then flattens the hand and moves it swiftly to the left (see Fig. 62), where he holds

(Meta)spatial representations in an emerging sign language

379

Fig. 59. Stimulus picture showing a house near the signers’ sister’s shop

Fig. 60. Both Frank and Jane sign ‘shop’ (the vegetable stand).

Fig. 61. Jane signs ‘up the street’ and ‘that side’.

it to signal that from the shop just mentioned the house in question (whose shape he goes on to describe) lies up the street to the left. Once again, to discount the possibility that the signers are tracing the actual cardinal direction involved in going from shop to house, consider Map 10, where I have overlaid over a map of the region an inset of the house

380

John B. Haviland

Fig. 62. Frank signs ‘from here to there’.

Map 10. Map showing orientation of shop and the signers’ vectors

in the village where the signers sit (the directions in which they portray the vector from shop to house are shown with arrows), and also an inset of the configuration of shop and house on the ground. The map shows that whereas the target house is northwest of the shop, both signers portray the vector as roughly southwest, orthogonal to the geographic vector. These examples allow us to conclude that for both known and unknown places, the Z signers are comfortable using a relative, observer centered, horizontal axis, for describing location and motion, in sharp contrast to speakers of Tzotzil who virtually never do so. 5.4

Transpositions and signing spaces

In this last described right-left type usage, the Z signers resemble most speakers of English, who tend to use projections from an observer view-

(Meta)spatial representations in an emerging sign language

381

point to describe horizontal angles in space, and whose interlocutors are practiced in transposing such perspectives to alternative viewpoints to interpret such descriptions (Haviland 1996). By contrast, speakers of Zinacantec Tzotzil appear to prefer, in co-speech gesture at least, to calculate horizontal angles absolutely – using cardinal directions. Although only a single horizontal axis, east-west, is lexicalized, evidence for this absolute frame-of-reference is abundant in the conventions of Zinacantec co-speech gesture. Furthermore, as examples like the description of the Chetumal turn-off (Fig. 6) show, Zinacantec interlocutors are presumably also practiced in transposing indicated cardinal directions onto imagined places, real or otherwise, other than the immediate location of the speech event – otherwise, Martin’s gestures as he describes where Chetumal and Cancún lie from the Chetumal turn-off would be un-interpretable. The different spaces I distinguished above, modeling them on Jakobson’s distinction between narrated and narrating events, were originally postulated to clarify aspects of the speech and gesture of speakers of languages like Tzotzil or Guugu Yimithirr in Australia, who rigorously track places and movements in terms of cardinal directions. In these languages, all locations seem to come with directions attached, so that – as we saw in the first part of this chapter – one automatically projects the orientation of Ss onto Sn, according to language-specific conventions. There are other ways that directions can be projected onto Sn from Ss, notably by transposing the relative perspective of an observer in Ss onto that of some suitable vantage point in Sn, as speakers frequently do in English and as the Z signers appear to expect interlocutors to be able to do in the last few tasks described above. One motivation in earlier work for positing an additional Si, distinct from either the narrated or the narrating space, was to provide for the fact that interactants in both Tzotzil and Guugu Yimithirr sometimes perform gestures in ways not seemingly anchored by cardinal directions. For example, when a Tzotzil narrator like Peter demonstrated how a solider pointed his gun, as it happens, facing east in the direction of his interlocutor (Fig. 63), he seems not to have intended to indicate that the solider actually aimed east. Instead the direction of his gesture is arbitrary, or rather it responds to the interactive conditions of the conversation rather than to the spatial arrangement of things in some world, past or imagined. Spatial relations between entities in this theoretical interactive space Si are thus imagined to be essentially arbitrary, emancipated from any sort of real space, and thus highly abstract. In this sense there is a scale of increasing abstraction from Ss which is constrained by the physical and concrete spatial surround of the speech event, to Sn which is selectively populated by those entities and the spatial relation-

382

John B. Haviland

Fig. 63. Peter mimes how a solider aimed his rifle.

ships between them that a narrator chooses to depict (or an interlocutor chooses to imagine), to Si wherein spatial relationships are absent or only serve as proxies for other kinds of relationships and which respond primarily to interactive needs. A young language like Z whose conventionalized resources for communicating about space are presumably still developing poses an insistent question about the relationship between linguistic spatial practices and speakers’ conceptual resources for thinking about space, including these postulated distinct conceptual spaces. In particular, the contrast between dead-reckoning of location but an alternation between an absolute and a relative right-left representation of narrated horizontal spatial relationships, that I have tried to demonstrate for Z, suggests a series of conventions still in-progress which rely in quite different ways on the theoretical conceptual spaces I have distinguished. Thus, for example, the fact that known locales are absolutely located on the horizon seems to imply that known places, even those that Fig. in narrated events, are always signed in Ss, in local space. To name a known location one does not leave the most concrete, local space of the speech event. Once spatial entities are conceptually implanted in Sn the relationships between them may be denoted with either an absolute or a relative frame of reference. The plastic toys are sketched in local space (perhaps diagrammatically) and one supposes that their absolute orientation is to be reproduced in the narrated space (i.e., the re-constructed array of the actual toys). On the other hand, the house near the sister’s vegetable stand must be understood from the perspective of the speaker in Ss, but transposed to some vantage point in Sn with linguistic devices which can be interpreted in at least two different, and mutually incompatible ways. Deciding between the two frames of reference (as well as the requisite transpositions they imply) requires mental operations and conventions which Z seems not yet to facilitate, judging by the difficulty the signers have in resolving such ambiguities. There remains one last matter to consider about spatial resources in Z, a matter of great theoretical interest which, unfortunately, I can introduce here

(Meta)spatial representations in an emerging sign language

383

Fig. 64. The Z sign for ‘chicken’

in only a preliminary way. This relates to the use of space as a grammatical device. There is evidence that even in a very young sign language like Z, the signers have begun to incorporate space into the grammar of the language in a way that recalls the spatial grammar of, among other things, verb inflection and agreement in established sign languages like ASL. Z thus provides evidence for the potential for Si to serve directly as a morphological medium in the manual modality. One device for turning space into grammar, prevalent in Z, we have so far met only laterally: the use of what I have been calling “haptic” classifiers to show the size, shape, and aspects of the manipulability of objects. We see hints of the phenomenon in the handshapes and configurations the signers use to indicate trajectories (of the road, for example in Fig.s 56 or 58), to show how human beings interact manually with named objects (e.g., cars in Fig. 55, or the vegetable stand in Fig. 60), to show their apparent size (e.g., the distant church in Fig. 54, or even the proper name for the signers’ mother in the right panel of Fig. 21, which portrays her belly in an un-complimentary way) or their shape (the turkey’s tail in Fig. 47 and the left panel of Fig. 50). The same principle is incorporated more directly into grammar, however, when the Z signers combine a common noun for an object with a haptic classifier that shows the size, shape, and manipulability of the object in question. The principle can be illustrated with a favorite example, the Z sign for ‘chicken’ which itself iconically incorporates the standard way of killing a chicken in Zinacantán: a sharp jerk with both hands to break the bird’s neck. Fig. 64 shows Will performing this sign, which, in his rendition, also incorporates a characteristic way of holding the mouth by puckering the lips. I have almost never seen this sign performed alone, however. Instead, Will normally precedes it with a haptic classifier to show the general class of object he is referring to. Thus, to describe a hen he starts with a handshape that indicates the size and characteristic way of holding such an animal

384

John B. Haviland

Fig. 65. Haptic classifier for a hen, and for a full-grown rooster sized animal

Fig. 66. A stimulus photo

(Fig. 65 left side), or to refer to a rooster he indicates a slightly larger bird (Fig. 67 right side) and only then performs the specific noun ‘chicken’. On the other hand, to describe a picture of two chicks (Fig. 66) he first signs a different haptic classifier, then the same neck-breaking chicken sign, immediately followed by the numeral two. The whole performance is illustrated in Fig. 67. It suggests both the conventionalized nature of the noun (since presumably, despite the iconicity, chicks are not so man-handled) and the abstract or grammatical character of the classifier as part of a larger noun-phrase-like construction. Haptic classifiers seem routinely to accompany nouns for commonly handled objects: domestic animals, clothing, tools, utensils, boxes, etc., and they are directly incorporated into the grammar. They rely on the immediately shared interactive bodily space of the signers to convey information that is incorporated into abstract, closedclass, functional elements resembling classifiers in other languages. Probably the most well-known grammatical use of space in established sign languages like ASL is linked to argument structure and to the fact that for certain classes of verbs “verb agreement is marked using spatial positions” (Padden 1990:118), or more specifically “the form of the verb itself makes spatial reference to the subject, object, or both” (Liddell 1990:176). Here is a

(Meta)spatial representations in an emerging sign language

385

Fig. 67. ‘Two chicks’: CLASSIFIER: Tiny, chicken, two

Fig. 68. Frame from video of a woman giving a man a shirt27

rough illustration: the verb ‘give’ in ASL typically involves a specific hand configuration. A signer can sign “I give it to you” by moving the “give” hand from her own body towards that of the addressee, or “you give it to me” by moving it in the opposite direction. For third person arguments, the signer can “place” the giver in one arbitrary position in signing space, the “receiver” in another, and sign “She gives it to him” by moving the “give” hand from the giver’s location to the receiver’s – anaphorically indexing grammatical arguments via 27 previously established spatial positions (i.e., signed “pronouns”). Meir et al. (2007) have shown that a young sign language like ABSL (developed over the last 70 years in a settled Bedouin village in Israel) does not code verb agreement with such a spatial device, although they note that the subject argument of a verb is typically implicit in the fact that a verbal action is performed in a way that iconically treats the signer’s body as the virtual subject. This is, of course, in itself a grammatical use of space, in that the spatial orientation of the signed verb and its relationship to the signer’s body provide essential grammatical information. It could be argued, nonetheless, 27

The stimulus video here was part of a set originally produced by Carol Padden and her associates for their ABSL research.

386

John B. Haviland

Fig. 69. Will signs “give [a shirt]”.

to be less abstract than the ASL convention, which moves agreement (almost) entirely off the signer’s body and into an arbitrarily structured Si. As my last examples will show, Z appears to share features of both ASL and ABSL, suggesting the range of possibilities space affords as a grammatical medium. The spatial affordances made available by the laminated conceptual spaces I have distinguished – Ss, Sn, and Si – are indexed by “inflecting” a verbal sign directly, as well as through body orientation, and gaze. Consider first how the Z signers described a short video which shows a woman passing a shirt to a man (Fig. 68). Describing a small set of such videos was one of the tasks I asked the Z signers to do on the very first day after they agreed, in 2008, to participate in a study of their language. The way they initially represented this video in sign is indicative of the highly telegraphic, largely presupposing style with which they originally approached the tasks I set them. It also illustrates the sign they chose in this case for ‘give’ (or, less contentiously, for denoting the transfer of the shirt from one person to the other). Will’s entire rendering of the video is initially contained in a single action, which he repeats twice. Using an apparently nonce haptic hand configuration that suggests the sort of transferred object involved, namely the shirt, he signs ‘give’ by moving the two grasping hands out away from his own body (Fig. 69). He makes no other apparent attempt to sign explicitly the man, the woman, or the shirt. Frank, who has been asked to match Will’s description against a series of possible still frames, picks one picture and describes it back to Will. He explicitly and opportunistically does sign ‘shirt’ (Fig. 70), and he continues with a mirror image of Will’s sign for ‘give’ (Fig. 71). (The two brothers are sitting on opposite sides of a table.) As they negotiate about which picture Frank should choose, at one point both signers simultaneously sign the ‘give’ verb

(Meta)spatial representations in an emerging sign language

387

Fig. 70. Frank signs ‘shirt’.

Fig. 71. Frank signs ‘give’.

in exactly opposite directions, both using their own bodies as the origo from which the narrated protagonist “gives the shirt” (Fig. 72). In this rendition the recipient is not marked, and the subject is virtually incorporated into the verb by virtue of the action’s being performed from the perspective of the imagined agent. Six months later, when the signers had figured out that both the tasks themselves and my expectations required a much higher level of explicitness in their renditions than they had volunteered that first day, they again described the same series of video vignettes. On this second occasion, both Frank and Jane, seated side-by-side, simultaneously described the video stimulus to the Matchers, and they took advantage of this arrangement to sign explicitly that it was a woman passing the shirt to a man. They used their own bodies as proxies for this gender distinction, an opportunistic device to which they frequently had recourse. Frank started in a somewhat contradictory way. He began by pointing to himself, following with a finger wave to signal negation, and immediately

388

John B. Haviland

Fig. 72. Frank and Will simultaneous sign ‘give’ from opposite vantage points.

Fig. 73. “Not the man, but the woman”.

thereafter pointing to Jane (Fig. 73), as if to say “not the man but the woman”. With a subtle shift of his hand position, he then signaled the transfer of an object – not itself identified – from the woman to the man by drawing his hand from Jane’s position back to his own chest (Fig. 74), a movement he repeated twice. Using his own body as a proxy for the male recipient, and his sister’s for the female giver, he was able to mark grammatical relations in an abstract signing space (that is, in Si) overlaid on top of genders abstracted from Ss. Jane used a variant device to sign “the woman gave the man a shirt”. She performed a sign virtually identical to that used six months previously to show ‘give’ – using both hands in a gripping configuration that suggested that what they held was something like a shirt – and she moved them out-

(Meta)spatial representations in an emerging sign language

389

Fig. 74. Frank signs “The woman gave it to the man.”

Fig. 75. Jane signs “The woman gave it to the man”.

ward from her own body (as if following the “body as subject” convention). However, by demonstratively turning her body toward her brother (see Fig. 75) as she signed ‘give’, Jane was able to exploit their gender difference again to encode “the woman gave it to the man”. Exploiting features of the current local signing or “speech event” space Ss that are not themselves arbitrary (like the actual physical locations of copresent people, Ps – participants of the speech event, in Jakobson’s formulation) but that can be used as at least partially arbitrary proxies for Pn – participants in the narrated event – seems to be one step in the direction of the spatially marked abstract verb agreement of languages like ASL. The Z signers made prolific use of such a device in describing stimulus videos meant to test the marking of presumed arguments in transitive clauses. So, for example, to sign another video clip in which a woman turns to look at a man, Frank signs “look” (twice), and again points first at Jane and then at himself (Fig. 76) to show who was looking at whom. One could liken such a signed construction to an uninflected verb combined with pronominal

390

John B. Haviland

Fig. 76. Frank signs “the woman looks at the man”.

proxies opportunistically extracted from Ss (plus a principle of word order, in this case resembling the order VSO).28 As in Jane’s performance in Fig. 75 above, shifting body orientation can also signal grammatical relations in way that combines a default iconic convention – that the signer’s body stands in for a notional agent – with a different sort of spatialized inflection on a verbal predicate: a further step toward grammaticalization of abstract, arbitrary positions in Si. In fact, Z signers seem to use body orientation in a variety of ways to signal grammatical relations, perhaps least surprisingly in the case of locative arguments. For example, in describing a video clip in which his young nephew was shown walking across a room to stand in front of a television set, Will first made the sign for ‘TV’, placing it slightly to his right, and then demonstratively turned his body before signing (with his feet) that the little boy walked (Fig. 77): “he walked to the TV set.” Will’s reorientation of the body (and the directional arc that he traces in the air afterwards – see Fig. 78 – interpreted by the hearing signers as “he went that way”29) seems to inflect the verb of motion and thus to serve the grammatical function of linking the television to the verb as a kind of allative argument. My final observation about space and grammar in Z links such indexical signs as pointing and bodily orientation to one further such visible device: gaze. Gaze has been argued in ASL to be yet another resource used to mark

28

29

See Haviland 2011 for a preliminary discussion of Z word order patterns. The normal constituent order in spoken Tzotzil is a robust VOS. Just to dispel a different possibility that may have occurred to diligent readers, the actual cardinal direction of the narrated movement here cannot be what Will meant to signal; the actual location of the scene depicted in the video clip is well known to him, close by, and in fact lies directly behind where he is sitting, and not in the direction he indicates.

(Meta)spatial representations in an emerging sign language

391

Fig. 77. Will signs “He walked to the TV”.

Fig. 78. Will traces a directional arc.

agreement.30 Z signers also recruit gaze, in sign, apparently to help signal arguments and distinguish what might be called participant frames; and they do so in ways that preserve distinctions we have already seen between different sorts of conceptual spaces. Unsurprisingly, for example, to sign a verb like ‘see’ the direction of the gaze suggests what is being looked at. Will, talking about a peculiar old man, signs that he saw him yesterday, and the vector he draws from his eye to the object of his vision (Fig. 79) links Sn to the local geography of Ss in a way exactly parallel to dead-reckoning in naming places. Just as one can direct gaze along a pointed vector in local space to indicate what one saw and where, however, one can also emancipate gaze from real space and direct it at an imagined or abstract interactive Si, populated by discursively introduced entities. This appears to be what Frank does when he 30

See Neidle et al. 2000; but compare Thompson et al. 2006.

392

John B. Haviland

Fig. 79. Will signs, “I saw [the old man].”

Fig. 80. Frank: “when I see those guys”

Fig. 81. Frank: “I’ll punch them.”

also sights along a pointed vector (Fig. 80) to sign the threat that when he catches sight of his sister’s boyfriends, he plans to beat them up (Fig. 81). In Fig. 80, Frank’s gaze is directed along an arbitrary vector in Si (and in fact is a kind of ‘fake’ gaze, looking at nothing at all in a neutral middle space – perhaps appropriate to the hypothetical situation he is evoking).

(Meta)spatial representations in an emerging sign language

393

Fig. 82. Terry uses ‘fake’ gaze to ask “was the woman looking?” and then returns a real gaze to her interlocutor.

By the time he signs ‘punch’31 in Fig. 81 his gaze reverts to his interlocutor – that is, it returns to Ss, the site of the speech event. Note that Jane’s gaze did something similar in Fig. 75 above, when she signed ‘give’ with her body oriented toward and apparently looking at or at least in the direction of her proxy recipient, but then returned her gaze to her interlocutors before actually retracting the ‘giving’ hands of her sign. The gaze shift, as it were, brackets off the narrated event (“I see the boyfriends” or “the woman gives the shirt”) from the interactional mutual attention check between the interlocutors in the speech event. The ‘fake’ gaze – apparently directed at some imaginary entity in the abstractly created interactional space – seems to represent a further exploitation of an interactively created or imagined Si within which signed morphology can be abstractly spatialized. In checking that she has properly understood Frank’s description of the video clip of a woman turning to look at a man, illustrated in Fig. 76 above, Terry also presents a ‘blank gaze’ as she mimes the verb ‘look’ – directing it first to her right as she asks Jane if it was the woman who was doing the looking (see Fig. 82), and then directing it to her left (and toward Will seated next to her – see Fig. 83) as she asks Frank (to whom she then shifts her gaze) “was she looking at a man?”.32

31

32

And note that this sign, by analogy with the “body as subject” model described by Meir et al. 2007, seems to illustrate “body as object.” Terry is a hearing signer, and there is a striking parallel between her signing and a standard discursive employment of voice in spoken Tzotzil. To clarify the argument structure of a transitive action, one can ask (with an antipassive) mi chk’elvan li antze ‘Did the woman do the looking?’; or with a full transitive mi isk’el vinik (li antze) ‘Did (the woman) look at the man?’ See Aissen 1990, 1999, Haviland 1981, Ayres 1983, Davies & Sam-Colop 1990, Craig 1979.

394

John B. Haviland

Fig. 83. Terry ‘gazes’ at a proxy protagonist, to ask “was she looking at a man?” and then looks back at her interlocutor.

Just as Z signers represent narrated spaces, Sn, whose geographies are sometimes known and sometimes not, by reference to local space, Ss, populated by presupposable entities with known locations, they also can use a much more abstract, interactively constructed space, Si, with arbitrarily (or opportunistically) created virtual entities, whose locations then can be reprised in the form and direction of manual signs, as well as through posture and gaze. I have presented evidence that these spatial references are recruited, in at least a preliminary way, for grammatical purposes, in particular to mark argument structure in discourse in the emerging sign language.

6.

Summary: Representing space with space

A very young sign language like Z affords special insights about how language construes space. Because it is a poorly documented language (albeit one in the making, and already endangered after a scant generation of existence), its structuring of space in linguistic terms has a compelling typological interest. Because it is young, its speakers can be expected still to be constructing formal resources for communicating about things important to them, including space, a domain they can scarcely avoid talking about. How they do so is thus of immediate diachronic interest. And because the medium is sign, Z necessarily uses space to represent space. It is metaspatial, by design, and thus allows a direct glimpse of the denotative and pragmatic potentials and requisites of space as both a medium and a referential target. This chapter began with a quick review of Tzotzil, the spoken Mayan language that surrounds and overlaps with the tiny Z “language community”, to illustrate typological distinctions that have been proposed for spatial language – most notably different frames of reference – and how spatial notions are realized in different Tzotzil form classes. One striking fact of Tzotzil spa-

(Meta)spatial representations in an emerging sign language

395

tial vocabulary is the elaboration and specificity of the lexical systems involved, especially a developed anatomical meronomy, hypertrophied positional roots, and grammaticalized verbs of motion. Perhaps more striking is the combination of somewhat meager lexical resources – the up/down distinction – with careful and quite precise directional gestures for specifying locations in terms of absolute cardinal directions, transposable between different perspectives or vantage points evident in the utterances of Zinacantec Tzotzil speakers. It is only in the combination of audible and visible features of utterances that the interaction between lexicalized space in speech and the visible manipulation of space in gesture to produce spatial reference are manifested. Co-speech gesture immediately embeds the analyst in a further set of conceptual complications about space, for no matter how we might understand what space is or how it is structured in the abstract, gesturing which may be about space in various ways willy-nilly takes place in space. Elaborating on Jakobson’s distinction between a narrated event En and a speech event Es, I applied to gesture a parallel distinction between a narrated space Sn which interlocutors talk about, and a narrating or speech event space Ss where they are situated when they talk. I also found it useful provisionally to distinguish a further interactional space, Si: a creation – a by-product – of the speech event and the positioning of interlocutors. Si is, in an important sense, unanchored by the wider Ss, and it responds instead to interlocutors’ interactional needs for conjoint attention. It is also where they sometimes gesture. Distinguishing these different spaces involves interlocutors and analysts alike in the problem of how these different spaces are interrelated and coordinated. Thus, when my Zinacantec compadre points to a spot, he may intend his interlocutor to understand that he is pointing in an imagined narrated Sn; that he is pointing to a “real” place somewhere within narrating Ss, the space where he is speaking, construed either locally or more widely; or that he is pointing “arbitrarily” to a locus or entity created by the interaction, that is, in Si. There may, moreover, be interactions between these different spaces, so that a narrated space may be laminated over the top of local space, allowing the absolute directions of one to be transposed onto the other, as in the description of places and directions along my compadre’s route to Cancún. Likewise, the coordinating principle may be relative rather than absolute, if relative or projected relations are involved. Interactive space Si may also be directionally anchored in some way, or it may be free from all orientation other than that imposed by its own conjured entities, whether arbitrarily placed in space or not. A central puzzle for interactants is, then, how to keep these spaces straight.

396

John B. Haviland

Having laid this groundwork, I introduced several striking features of Z spatial practice. First, the Zinacantec signers must keep careful track of the absolute locations of known places, so much so that the standard device for naming them seems conventionally to be pointing to them either line of sight or on the horizon, directly in Ss. This seems one clear example of the direct source in Zinacantec co-speech gesture for a central structural device in the emerging sign language. By contrast, despite the fact that in co-speech gesture Zinacantecs virtually never seem to calculate position or direction on the horizontal plane relative to a speaker’s own body, the Z signers – perhaps because they do not (yet) have conventionalized lexemes for absolute directions (in much the same way that they lack conventional color names, for example) – do appear to apply body-centric relative signs for right and left in descriptions of spatial scenes, a device that clashes with an apparently poorly developed convention for altering perspective or point of view and thus leads to occasional miscommunication. In using such body-relative descriptive devices for spatial relations, the Z signers make extensive use of the directionally ‘unanchored’ interactive space Si, thus building their linguistic devices around an interactively created and manipulated virtual metaspace. The final sections of this chapter expand on the ways that Si can serve as an abstract medium through or upon which spatial diagrams can be constructed to represent a variety of relations, only some of which are literally spatial. The well-known phenomenon labeled “spatial grammar” in developed sign languages allows different spatial devices to assume the functions of grammatical marking, notably argument structure, agreement and anaphora. Even a language like Z, emerging over barely thirty years in a single, tiny language community, can be shown to be using space itself to re-invent anew the abstract notions of grammar.

References Aissen, Judith L. 1984: Control and command in Tzotzil purpose clauses. In: Proceedings of the Tenth Annual Meeting of the Berkeley Linguistics Society, 559–571. Aissen, Judith L. 1990: Una teoría de voz para idiomas mayas. In: Nora England & Stephen R. Elliott (eds.), Lecturas sobre la lingüística maya, 399–419. Plumsock Mesoamerican Studies: South Woodstock. Aissen, Judith L. 1994: Tzotzil auxiliaries. Linguistics 32: 657–690. Aissen, Judith L. 1999: Agent focuses and inverse in Tzotzil. Language 75(3): 451–485. Ashmore, Wendy & Gordon R. Willey 1981: A historical introduction to the study of lowland Maya settlement patterns. In: Wendy Ashmore (ed.), Lowland Maya Settleent Patterns, 3–18. Albuquerque: University of New Mexico Press.

(Meta)spatial representations in an emerging sign language

397

Ashmore, Wendy 1989: Construction and cosmology: politics and ideology in lowland Maya settlement patterns. In: William F. Hanks & Don S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation, 272–286. Salt Lake City: University of Utah Press. Ayres, Glenn T. 1983: Antipassive voice in Ixil. International Journal of American Linguistics 49(1): 20–45. Bohnemeyer, Juergen & Christel Stolz 2006: In: Stephen C. Levinson & David P. Willkins (eds.), Grammars of Space, Explorations in Cognitive Diversity, 273–310. Cambridge: Cambridge University Press. Brown, Penelope 1994: The INS and ONS of Tzeltal locative expressions: the semantics of static descriptions of location. Linguistics 32(4/5): 743–790. Brown, Penelope 2006: A sketch of the grammar of space in Tzeltal. In: Stephen C. Levinson & David P. Willkins (eds.), Grammars of Space, Explorations in Cognitive Diversity, 230–272. Cambridge: Cambridge University Press. Brown, Penelope and Stephen C. Levinson 1993: “Uphill” and “downhill” in Tzeltal. Journal of Linguistic Anthropology 3(1): 46–74. Calbris, Genevieve 1990: The Semiotics of French gesture. Bloomington: Indiana University Press. Craig, Colette G. 1979: Antipassive and Jacaltec. In: Nora C. England (ed.), Papers in Mayan Linguistics, 139–164. Miscellaneous Publications in Anthropology, 6. Columbia, MO: University of Missouri, Department of Anthropology. Davies, William D. & Luis E. Sam-Colop 1990: K’iche’ and the structure of the antipassive. Language 66(3): 522–549. De León, Lourdes 1992c: Locative Body Parts and Geographic Anchoring in Tzotzil Acquisition. Paper presented at the Stanford Child Language Forum, March, 1992. De León, Lourdes 1994a: Exploration in the acquisition of geocentric location by Tzotzil children. Linguistics 32 (4/5): 857–884. Fowler, Catherine S. 2001: Numic cardinal directions. In: Jose Luis Moctezuma Zamarron & Jane Hill (eds.), Avances y balances de lenguas yutoaztecas, 267–292. Instituto Nacional de Antropología e Historia, México. Gossen, Gary H. 1974a: Chamulas in the World of the Sun: Time and Space in a Maya Oral Tradition. Cambridge, MA: Harvard University Press. Gossen, Gary H. 1974b: Another look at world view: aerial photography and Chamula cosmology. In: Evon Z. Vogt (ed.), Aerial Photography in Anthropological Field Research, 112–124. Cambridge, MA: Harvard University Press. Hanks, William F. 1988: Grammar, style, and meaning in a Maya manuscript. International Journal of American Linguistics 54(3): 331–362. Hanks, William F. 1990: Referential Practice: Language and Lived Space Amoung the Maya. Chicago: University of Chicago Press. Hanks, William F. 1992: Intertextualité de l’espace au Yucatán. L’Homme 32(122–124): 53–74. Hanks, William F. 2003: ‘Reduccion’ and the remaking of the social landscape in colonial Yucatan. In: Alain Breton, Aurore Monod Becquelin & Mario Humberto Ruz (eds.), Espacios mayas: usos representaciones creencias, 161–180. Mexico: Universidad Nacional Autónoma de México. Haviland, John B. 1981: Sk’op sotz’leb: el tzotzil de San Lorenzo Zinacantan. Mexico: Universidad Nacional Autónoma de México. Haviland, John B. 1992: ‘Seated and settled.’ Tzotzil verbs of the body. In: Lourdes de

398

John B. Haviland

León & Stephen Levinson (eds.), Space in Mesoamerican Languages, Special Issue of Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung, 45(6): 543–561. Berlin: Akademie Verlag. Haviland, John B. 1993: Anchoring, iconicity, and orientation in Guugu Yimidhirr pointing gestures. Journal of Linguistic Anthropology, Vol. III(1): 3–45. Haviland, John B. 1993b: The syntax of Tzotzil auxiliaries and directionals: the grammaticalization of ‘motion’. Proceedings of the Nineteenth Annual Meeting of the Berkeley Linguistics Society: Special Session on Syntactic Issues in Native American Languages, 35–49. Haviland, John B. 1994: Te xa setel xulem (The buzzards were circling): Categories of verbal roots in (Zinacantec) Tzotzil. Linguistics 32(1994): 691–741. Haviland, John B. 1994b: Verbs and shapes in (Zinacantec) Tzotzil: the case of ‘insert’. Función 15–16 (1994): 83–117. Haviland, John B. 1994c: Special issue: spatial conceptualization in Mayan languages, edited by John B. Haviland & Stephen C. Levinson. Linguistics 32(4/5). Haviland, John B. 1996: Projections, transpositions, and relativity. In: John J. Gumperz & Stephen C. Levinson (eds.), Rethinking Linguistic Relativity, 271–323. Cambridge: Cambridge University Press. Haviland, John B. 1998a: Guugu Yimithirr Cardinal Directions. Ethos 26(1) (March 1998), 25–47. Haviland, John B. 2000a: Early pointing gestures in Zinacantán. Journal of Linguistic Anthropology 8(2), 162–196. Haviland, John B. 2000d: Pointing, gesture spaces, and mental maps. In: David McNeill (ed.), Language and Gesture: Window into Thought and Action, 13–46. Cambridge: Cambridge University Press. Haviland, John B. 2003: Dangerous places in Zinacantec prayer. In: Alain Breton, Aurore Monod Becquelin & Mario Humberto Ruz (eds.), Espacios mayas: usos representaciones creencias, 383–420. Mexico: Universidad Nacional Autonoma de Mexico. Haviland, John B. 2003a: How to point in Zinacantán. In: Sotaro Kita (ed.), Pointing: Where Language, Culture, and Cognition Meet, 139–170. Mahwah, N.J. and London: Lawrence Erlbaum Associates. Haviland, John B. 2005: Directional Precision in Zinacantec Deictic Gestures: (cognitive?) preconditions of talk about space. Intellectica 2005(2–3): 41–42, 25–54. Haviland, John B. 2005b: Dreams of blood: Zinacantecs in Oregon. In Mike Baynham and Anna de Fina (eds.), Dislocations/Relocations: Narratives of Displacement, 91–127. Manchester, UK/Northampton, MA: St. Jerome Pub. Haviland, John B. 2011: Nouns, verbs, and constituents in an emerging ‘Tzotzil’ sign language. In: Rodrigo Gutiérrez-Bravo, Line Mikkelsen & Eric Potsdam (eds.), Representing Language: Essays in Honor of Judith Aissen, 157–171. California Digital Library eScholarship Repository. Linguistic Research Center, University of California, Santa Cruz. (http://escholarship.org/uc/item/0vf4s9tk and http:// escholarship.org/uc/lrc_aissen. Also available from the on-demand publisher BookSurge, ISBNs 0-9836-9380-3, 0-9836938-0-2.) Haviland, Leslie K. & John B. Haviland 1982: Inside the fence: The social basis of privacy in Nabenchauk. Estudios de cultura maya 14: 323–352. Haviland, William A., 1966: Maya settlement patterns: A critical review. In: Archaeological Studies in Middle America, 21–47. Publication, 26. New Orleans: Tulane University, Middle American Research Institute.

(Meta)spatial representations in an emerging sign language

399

Hill, Clifford 1982: Up/down, front/back, left/right: a contrastive study of Hausa and English. In: Jürgen Weissenborn & Wolfgang Klein (eds.), Here and There: Cross-linguistic Studies on Deixis and Demonstration, 13–42. Amsterdam: John Benjamins. Jakobson, Roman 1957: Shifters, verbal categories, and the Russian verb. Mimeo. Russian Language Project, Department of Slavic Languages and Literatures, Harvard University. Jones, Christopher 1989: Builders of Tikal: archaeology and history. In: William F. Hanks & Don S. Rice (eds.), Word and Image in Maya Culture: Explorations in Language, Writing, and Representation, 255–259. Salt Lake City: University of Utah Press. Kendon, Adam 1980: A description of a deaf-mute sign language from the Enga province of Papua New Guinea with some comparative discussion; Part II: The Semiotic functioning of Enga signs. Semiotica 31(1/2): 81–117. Kendon, Adam 1990: Conducting Interaction. Cambridge: Cambridge University Press. Kendon, Adam 2004: Gesture: Visible Action as Utterance. New York: Cambridge University Press. Laughlin, Robert M. 1975: The Great Tzotzil Dictionary of San Lorenzo Zinacantan. Washington, D.C.: Smithsonian Institution. Laughlin, Robert M. 1980: Of Shoes and Ships and Sealing Wax: Sundries from Zinacantan. Washington, D.C.: Smithsonian. Laughlin, Robert M. 1988: What is a Tzotzil? Res 15: 133–156. Levinson, Stephen 1994: Vision, shape, and linguistic description: Tzeltal body-part terminology and object description. Linguistics 32(4/5): 791–856. Levinson, Stephen 2003: Space in Language and Cognition, Explorations in Cognitive Diversity. Cambridge: Cambridge University Press. Levinson, Stephen C. & David P. Wilkins (eds.) 2006: Grammars of Space, Explorations in Cognitive Diversity. Cambridge: Cambridge University Press. Liddell, Scott K. 1990: Four functions of a locus: reexamining the structure of space in ASL. In: Ceil Lucas (ed.), Sign Language Research: Theoretical Issues, 176–200. Washington DC: Gallaudet University Press. Meir, Irit, Carol Padden, Mark Aronoff & Wendy Sandler 2007: Body as subject. Journal of Linguistics 43: 531–563. Neidle, Carol, Judy Kegl, Dawn MacLaughlin, Benjamin Bahan & Robert G. Lee 2000: The Syntax of American Sign Language: Functional Categories and Hierarchical Structure. Cambridge, MA: MIT Press. Padden, Carol A. 1990: The relation between space and grammar in ASL verb morphology. In: Ceil Lucas (ed.), Sign Language Research: Theoretical Issues, 118–132. Washington DC: Gallaudet University Press. Polian, Gilles & Juergen Bohnemeyer 2011: Uniformity and variation in Tseltal reference frame use. Language Sciences 33(6): 868–891. Sandler, Wendy, Irit Meir, Carol Padden & Mark Aronoff (2005): The emergence of a grammar: Systematic structure in a new language. Proceedings of the National Academy of Sciences, 102(7): 2661–2665. Svorou, Soteria 1994: The Grammar of Space. Amsterdam: John Benjamins. Talmy, Leonard 1985: Lexicalization patterns: semantic structure in lexical forms. In: Timothy Shopen (ed.), Language Typology and Syntactic Description, 57–149. London: Cambridge University Press.

400

John B. Haviland

Vogt, Evon 1992: Cardinal directions in Mayan and Southwestern Indian cosmology In: Antropología mesoamericana: homenaje a Alfonso Villa Rojas, 105–128. Gobierno del Estado de Chiapas, Tuxtla Gutiérrez. Zavala Maldonado, Roberto 1992: Mayan Directionals: Patterns of Grammaticalization. Paper presented at the 91st annual Meetings of the American Anthropological Association. San Francisco: Dec. 4, 1992.

What difference does space make for interaction and interaction for space?

401

Anja Stukenbrock

Commentary: What difference does space make for interaction and interaction for space?

1.

Introduction

The four papers in this section differ in terms of theoretical abstraction vs. empirical analysis, in their use of established vs. new concepts, and the type and relevance of the data they use to support their arguments.

2.

Different types and relevance of data

Three of the papers represent empirical studies undertaken in various settings (Haviland, Mondada, Streeck). In contrast, the fourth paper (Hausendorf) offers a theoretical framework designed to integrate different notions of (interactional) space developed in Conversation Analysis (CA) and other closely related paradigms. While Hausendorf does not refer to empirical examples, Streeck’s video-ethnographic analysis of the Plaza de la Trinidad in Cartagena de Indias (Colombia) is based on visual observation of the participants’ bodily behavior, omitting spoken language altogether. Mondada and Haviland both use audiovisual data to analyze the relationship between language, visual bodily modalities, and space. Nonetheless, their data sets also differ significantly. Mondada draws on video-recordings of naturally occurring interactions in which participants speak French and approaches the data from a praxeological perspective, i.e. a reflexive elaboration of situated action and relevant spatial features. Haviland, in contrast, not only uses videorecordings of spontaneous face-to-face interaction between his participants, but also draws on semi-experimental data elicited in various set-ups in order to compare spatial and meta-spatial devices in a nascent sign language developed in a household of Zincantec Indians from Highland Chiapas (Mexico) to the spatial reference frames in the surrounding spoken language, i.e. Tzotzil.

402

3.

Anja Stukenbrock

Shared concepts

What all four papers have in common is their concern for space from an interactional perspective. All four authors approach spatial phenomena on a micro-level of analysis indebted to the tradition of context analysis, ethnomethodology, Conversation Analysis, and research on multimodality. Central to the shared concept of interactional space is the assumption that space is not a given entity, but an interactive achievement of the participants. Not unlike the term “interactional space”, which has become a catchword in a research tradition that goes back to Goffman (1963, 1964, 1971), Kendon & Ferber (1973) and Kendon (1976, 1990), an explicit commitment to an analysis of the interactive achievement of space has become a credo for many interactional studies on space. While both Hausendorf and Mondada explicitly refer to the idea of space as an achievement and further elaborate on the concept of interactional space, Streeck undertakes a shift from space to place and focuses on how interaction ensembles make places for themselves and acquire senses of place in the built environment of the Plaza. His analysis transcends the evanescence of interactional spaces as constituted by the ever-changing bodily configurations of participants in the here-and-now of an ongoing interaction, and includes more enduring or recurring spatial patterns for which he adapts concepts from sociology and cultural geography such as “territories”, “stations”, “docks”, and “clusters”. In Haviland’s paper, interactional space is part of a conceptual trias in which narrated space and narrating space – both owed to Jakobson’s (1957) seminal paper on shifters – figure as the central concepts for an analysis of the referential use of spatial devices in a first-generation sign language, Zinacantec Family Homesign (Z). The term “narrated space” denotes space that is not perceptually accessible, but narrated and imagined. Inversely, “narrating space” refers to the here-and-now, or speech event space in which the participants are located while the narration is being produced. Whereas both narrated and narrating space provide referents for spatial language use, the term “interactional space” specifies the space within which the signers gesture. Haviland’s notion of interactional space is reminiscent of the concept of “o-space” developed by Kendon (1990: 211) to denote the joint space constructed and upheld by the bodily arrangement of two or more interactants. But more than o-space, it also denotes the kind of space that is turned into a grammatical medium during a developmental process in which the morphology in the manual modality or signed morphology becomes abstractly spatialized. We witness how spatial language grows out of communi-

What difference does space make for interaction and interaction for space?

403

cative practices that make use of interactional space and in doing so transform both interactional space and the practices in it: In the course of the emerging conventionalization, space is grammaticized and grammar spatialized. Haviland’s paper is the only one in this section which also examines space that is not perceptually accessible, but narrated and imagined. By analyzing how participants refer to imagined or narrated space, he touches upon a central concern in research on deixis, where the linguistic construction of imagined space is termed Deixis am Phantasma (‘deixis in the imagination’; Bühler 1934) and has only recently been turned into an object of empirical investigation (Haviland 2000; Stukenbrock 2012, in press). If we pay due tribute to the pioneering studies of Goffman and Kendon we need to ask a question also raised by Hausendorf: Do we have anything new to say, and if so, what is it? In order to elaborate on this question, it is enlightening to take a look at the concepts developed in this section and their theoretical background.

4.

So what’s new? Different ways of reformulating and furthering theoretical concepts

4.1

Interactional space and situational anchoring

Apart from discussing new technologies, new types of data, and more advanced methodology, Hausendorf ’s answer to the crunch question of scientific progress starts from the discussion of a central conceptual problem. He reflects upon the possible meanings of interactional space and reformulates the problem of what is meant by interactive achievement of space as a problem of situational anchoring. This problem in turn is defined as the task of achieving a mutually shared, concrete “here” for interaction, and differentiated into three sub-problems or tasks, each of which is linked to what could be called a mode of relating with others to the “world within reach” (Schütz 1945): Thus, the task of co-orientation concerns perception (as well as the perception of perception, Luhmann 1984: 560), that of co-ordination refers to movement, and that of co-operation to action. Although the concepts of co-orientation, co-ordination, and co-operation have been central to CA studies and research on multimodality in recent years, they have not been brought together and differentiated in a systematic way. Hausendorf ’s framework offers a systematic disentanglement of analytically applied but theoretically underdeveloped concepts. It allows for a conceptual clarity in a field which has been blurred by the fact that

404

Anja Stukenbrock

space is conceived of both as an interactional resource and as an interactional achievement (Hausendorf, this volume). By replacing the catchphrase “space as achievement” by the more suitable terminology “situational anchoring” as a genuinely interactive task and by splitting it into three sub-tasks to which interactive devices and forms can be assigned, Hausendorf provides a new heuristic method for empirical analysis. Challenging is his claim that the three sub-tasks imply each other hierarchically. Since the proof of the pudding is in the eating, this claim must stand up to empirical analyses. 4.2

Interactional space and mobility

In her empirical analysis, Mondada, like Hausendorf, refers to the works of Goffman, Kendon, and Goodwin as a point of departure, and further elaborates on the concept of interactional space insofar as she fluidizes it, making it more dynamic, and thus providing a fruitful connection to her section on mobile spaces (this volume; cf. also her commentary in section 4). In her paper on “Interactional space and the study of embodied talk-in-interaction”, in the present section, however, she is not so much concerned with mobile spaces in the sense of spaces on the move, but with a reconceptualization of interactional space as constantly changing, continuously shaped and reshaped by the participants’ bodily configurations and arrangements (Kendon 1990). Participation and participation frameworks (Goffman 1981) can now be described in spatial terms, since they are instantiated as situated, embodied bodily configurations of the participants (Goodwin & Goodwin 2004). Changes in participation are observable in the transformations that those bodily arrangements undergo. Mobility also plays a central role for Streeck who – with reference to Norbert Elias – proposes that “stationary social life […] evolves from social life in motion” and shows how this pattern is laid out in the structured, agegraded interactional ensembles that occupy different regions of the Plaza. Streeck’s various examples of children who jointly explore the surroundings by trying out different movement practices such as running, walking backwards and forwards, climbing and jumping up and down stairs and crevices etc., can also serve to argue that the intimate relationship between movement and perception – their “equiprimordiality” (Gleichursprünglichkeit) – which is self-evident in the individual (e.g. proprioception, kinaesthetics) also manifests itself as a social phenomenon. Moreover, in Streeck’s data, mobility not only appears as the most natural and bodily way of being in the world, of experiencing the world together, and of experiencing togetherness, but also as a generational pattern in which the members of the community move from

What difference does space make for interaction and interaction for space?

405

mobility to rest, from the Plaza’s vivid center of mobile activities to its periphery of stationary observers. If we return to the theoretical framework laid out by Hausendorf, according to whom interactional space is inherently mobile, movement and perception become inseparable as constitutive devices for the construction of an intersubjective, constantly changing “here”. The relationship between movement and perception needs to be conceived of as reciprocal rather than unidirectional: Not only is it true that joint perception allows for joint movement, but joint movement in turn also allows for joint perception of newly emerging, hitherto inaccessible phenomena. It is perception on the move for which co-ordination, or joint movement, constitutes a means and a prerequisite. This becomes more evident if we assume that mobility, mobile bodies, and mobile “withs” (Goffman 1963) rather than static, isolated bodies or talking heads constitute the default case of human sociality. It might be useful to consider “intercorporeality” (cf. Streeck 2009; Streeck, Goodwin & Le Baron 2011) as a concept to further develop our views on “jointness” as a dynamic, multi-facetted phenomenon that is brought about not only by the accomplishment of co-tasks (or by co-tasking), but that can be viewed as an embodied social phenomenon that is deeply grounded in the body’s intersubjective orientedness on all kinds of levels. It seems that the participants’ “here” for sensual perception (joint attention), their “here” for bodily movement (joint movements), and their “here” for social action (joint action) are intricately interwoven. Deixis is a grammatical case in point to illustrate that joint attention (Diessel 2006; Stukenbrock 2009, in print) not only forms a precondition for, but can also turn out to be a result of, co-ordination and – depending on the all-too-fuzzy concept of action – co-operation. Though it is very useful to separate these concepts theoretically, in empirical analyses of embodied practices they merge just as the manifold resources of the participants interlock in the online emergence of interaction. It is this reciprocal dependency that is at the heart of the notion that the “perceivable” and “perceived”, the “usable” and “used” are both action-shaping and action-shaped. 4.3

Interactive emplacement and the role of material surroundings

Among the many ways in which the concept of “space” in face-to-face interaction can be advanced and enriched, a fruitful impulse is derived from reflections on the role of material surrounds. Here, a new balance seems to be underway between conceding the materiality of physical surrounds and preserving a constructionist perspective; this balanced view seems to supersede

406

Anja Stukenbrock

former attempts to negate the givenness of certain spatial properties altogether. However, the degree to which material givenness is theoretically and empirically conceded or integrated into the analysis differs. Hausendorf proposes the concept of spatial “usability cues” to refer to the affordances of formed, built, or furnished space and to emphasize that perceived space must be understood as “read space” (Hausendorf, this volume). Mondada’s praxeological perspective on interactional space and her insistence that the material surrounds should be analyzed as both action-shaped and actionshaping resources and constraints brings her close to Hausendorf ’s concept of affordances and usability cues. These aspects are also relevant for Streeck’s thinking about the way in which community members emplace themselves in the Plaza by making use of perceived, built space and by transforming it into place. But while Hausendorf speaks of the emplacement of the co-participants within a concrete speech situation, Streeck’s concept of emplacement goes beyond the fleeting here-and-now of single interactional encounters and serves to capture more durable aspects – the sedimented spatial patterns of how different interactional ensembles or cohorts habitually occupy certain parts of the Plaza, form clusters, etc. Finally, Haviland’s results suggest that different degrees of familiarity with a place – or possibly a person’s emplacement – may influence the choice of alternative spatial orientational systems encoded in the language(s) of his speakers. He observes that unlike Tzotzil speakers who practically never refer to a relative frame of reference to locate a known place, Z signers, in addition to an absolute frame of reference indebted to the influence of cospeech gesture in Zinacantec Tzotzil, use a relative, observer-centered perspective for referential practices.

5.

Conclusion

To sum up, the papers in this section explore a rich variety of innovative theoretical and empirical issues concerning the relationship between interaction and space ranging from the systematic theoretical reconceptualization of interactional space and related concepts (Hausendorf), to its empirical fluidization and dynamization (Mondada); they are derived both from a phenomenal (Haviland) as well as from a phenomenological widening of perspectives (Streeck) and from an analysis of unique sets of data; and they include microhistorical insights into the development and sedimentation of semiotic resources in nascent communities of practice (Haviland, Streeck) of very different kinds. At this point, temporality is a relevant factor and transcends interactional space in the narrow sense; concepts such as “emplacement”,

What difference does space make for interaction and interaction for space?

407

“usability cues”, and “narrated space” allow us to integrate local, punctual phenomena of interactional encounters and phenomena in which lived time, space, and experience of interactional histories manifest themselves. Apart from an ongoing need for methodological advancement, research on human interaction also requires theoretical replenishment and inspiration. The papers in this section offer insights into how concepts from various disciplines outside linguistics such as human geography, sociology, and phenomenology can be appropriated for research on interaction. I would like to conclude by calling for more courageous theoretical polygamy and, in order to prevent this from drifting into arbitrariness, for greater conceptual and terminological consistency.

References Bühler, Karl 1934: Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena: Fischer. Diessel, Holger 2006: Demonstratives, joint attention, and the emergence of grammar. Cognitive Linguistics 17: 463–489. Goffman, Erving 1963: Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press. Goffman, Erving 1964: The neglected situation. American Anthropologist 66(6): 133–136. Goffman, Erving 1971: Relations in Public: Microstudies of the Public Order. New York: Harper and Row. Goffman, Erving 1981: Forms of Talk. Philadelphia: University of Pennsylvania Press. Goodwin, Charles & Marjorie Harness Goodwin 2004: Participation. In: Alessandro Duranti (ed.), A Companion to Linguistic Anthropology, 222–244. Malden, MA: Blackwell. Haviland, John 2000: Pointing, gesture spaces, and mental maps. In: David McNeill (ed.), Language and Gesture, 113–146. Cambridge: Cambridge University Press. Jakobson, Roman 1957 [1971]: Shifters, verbal categories, and the Russian verb. In: Roman Jakobson: Selected Writings. Vol. 2, 130–147. Berlin/New York: Mouton de Gruyter. Kendon, Adam & Andrew Ferber 1973: A description of some human greetings. In: Richard Phillip Michael & John Hurrell Crook (eds.), Comparative Behaviour and Ecology of Primates, 591–668. London: Academic Press. Kendon, Adam 1976: The F-formation system: Spatial-orientational relations in face to face interaction. Man Environment Systems 1976(6): 291–296. Kendon, Adam 1990: Conducting Interaction: Patterns of Behavior in Focused Encounters. Cambridge: Cambridge University Press. Luhmann, Niklas 1984: Soziale Systeme. Grundriss einer allgemeinen Theorie. Frankfurt am Main: Suhrkamp. Schütz, Alfred 1945: On multiple realities. In: Philosophy and Phenomenological Research 5 (4): 533–576. Streeck, Jürgen 2009: Gesturecraft. The Manu-facture of Meaning. Amsterdam/Philadelphia: John Benjamins.

408

Anja Stukenbrock

Streeck, Jürgen, Charles Goodwin & Curtis LeBaron (eds.) 2011: Embodied Interaction: Language and Body in the Material World. Cambridge: Cambridge University Press. Stukenbrock, Anja 2009: Referenz durch Zeigen. Zur Theorie der Deixis. Deutsche Sprache 37: 289–315. Stukenbrock, Anja 2012: Imagined spaces as a resource in interaction. Bulletin Suisse de Linguistique Appliquée 96: 141–161. Stukenbrock, Anja in press: Deixis in der face-to-face Interaktion. Berlin/New York: de Gruyter.

Section 4: Mobile spaces

Action and space: Navigation as a social and spatial task

411

Pentti Haddington

Action and space: Navigation as a social and spatial task

1.

Introduction

Scholars across theoretical, methodological and disciplinary boundaries continue to be interested in how space, in its many forms, features in human experience, culture, cognition, learning and language (for some overviews, see e.g. Levinson 1996, 2003; Haddington and Keisanen 2009; Mondada 2011). Language and space specifically, as Levinson (1996: 353) notes, have been the targets of abundant analysis in for example linguistics, anthropology and cognitive sciences (see also Levinson 2003). In general, two different approaches can be identified. On the one hand, there is research that investigates language and space by building on the idea that the ways in which humans perceive and sense the world around them creates spatial and cognitive schemata that influence the ways in which space is encoded and represented in grammar and grammatical systems. These analyses for natural reasons focus on language, and these linguistic representations are then seen to reveal how humans perceive, think about and conceptualize spatial concepts or relations. On the other hand, there is research that argues that it is grammar that guides their users to observations of the world and that can restructure cognition. Thus, it is argued, the speakers of different grammars can possibly arrive at different views of the world, for example, as regards space, spatial descriptions and spatial relations (see for example Levinson 1996, 2003). The analysis of language structure provides an important window to the study of space, spatial thinking and spatial cognition. However, in this chapter my aim is to develop a third approach and to argue that it is equally important to analyze and conceptualize how space and spatial features become evident in the ways people act while communicating multimodally and together in real-life social interaction (see also Levinson 2006b). In order to study ‘action and space’, this paper draws on the methodology provided by conversation analysis (e.g. Hutchby and Wooffitt 2008; Schegloff 2007) and multimodal interaction analysis (Goodwin 2000). This methodology represents an empirical analysis of the sequential organization of human action, i.e. how participant actions build on prior actions and events, and through their design display how these actions and events are interpreted. In this approach lan-

412

Pentti Haddington

guage is seen as one possible modality that is used to perform and deliver actions. Since social interactants use language simultaneously with multiple semiotic resources (including gestures, the body and features of the material context) for producing and responding to actions and for interacting with each other, the analysis focuses on the integratedness of these resources, and how interactants use them together for communicating understandings of events and actions as they occur moment by moment in their real-life and everyday contexts (see Goodwin 2000: 1490). Conversation analysis and multimodal interaction analysis use as data audio-video recordings of reallife and naturally-occurring interaction. Data analyses are accompanied with detailed transcriptions and illustrations of the analyzed segment. In this chapter, the analyses are based on video recordings of interaction in cars. These data come from different corpora which together contain approximately 100 hours of real-life and naturally-occurring interaction in cars. The data have been collected in Australia, Finland and the UK.1 The languages spoken in the data are English and Finnish. This chapter focuses on lay practices of navigation and specifically on how space features in human action during navigation in cars. Everyday navigation constitutes an activity that can involve and require the solving of different kinds of spatial puzzles. For example, it can require from the participants orientation to the current location, the destination and landmarks on the way. What is more, navigation can require and it is often accomplished as a joint accomplishment through different kinds of multimodal resources (e.g. talk and gesture). Also knowledge, experience and the participants’ shared interactional history can act as helpful resources in collaborative navigation. Although navigation is a fundamental human activity, we still know little about how people navigate as part of their everyday lives and how space features in this activity (see Haddington 2012; Levinson 1996; Haddington and Keisanen 2009).2 By studying spatial actions in the interactional contexts in which they are used, this chapter tries to conceptualize how space is tied to everyday social conduct, i.e. social actions and activities. The analyses will show that interlocutors do not only express their understandings of space through language, but that they also ‘interact space’. In this chapter, I build on recent interactional research that has criticized linguistic research for treating space as a pre-existing and abstract container for action, objects and events. This 1

2

I want to thank Maurice Nevile and Eric Laurier for kindly sharing their databases with me. For brief overview of previous research on navigation, see Haddington and Keisanen 2009 and Haddington 2012.

Action and space: Navigation as a social and spatial task

413

interactional research has tried to show how, on the one hand, participants in and through their actions structure space as a relevant and consequential feature of the context and how, on the other hand, people orient to and make manifest spatial features around them through linguistic, embodied and other social practices (e.g. Broth 2009; Goodwin 2000, 2003; Kendon 1990a; Laurier 2005, McIlvenny 2009; Mondada 2007b, 2009, 2011).3 This chapter outlines two ways in which space can be seen to feature in human action and shows 1) how action shapes space, i.e. how participants aurally and visually construct an understanding of pre-existing space as an interactional achievement and for the practical purposes of, for example, moving, acting in or describing space (see the analysis in section 3) and 2) how space shapes action, i.e. how participants manifestly display their sensitivity to the pre-existing features of space or understandings of them by adjusting their verbal and embodied behavior on the basis of the spatial context around them (see the analysis in section 4). In sum, this paper analyzes everyday spatial notions – space as a member’s phenomenon. It provides evidence for how participants orient to space as consequential for their actions, how they solve spatial tasks, how spatial features impact on the accomplishment of a task and how spatial meanings are created and achieved collaboratively and in situ, for the practical purposes of an interactional activity. The analyses focus specifically on how linguistic structure, the design of gestures, the exact timing of actions and collaborative participation in an activity reveal how space features in human action.

2.

Cognition, space and action as social and observable phenomena

Linguistic structures are often seen as windows to how humans conceptualize or construe space and spatial meanings (e.g. Lee 2001). In the following, my aim is to study cognition as it is situated in the social organization of action, i.e. to study the human mind in its social environment (see Wilson 2004). The issue of cognition and how cognition functions is largely an unsolved issue for conversation analytic research (see e.g. te Molder and Potter 2005). Some interactional research proposes that the ways in which people 3

See also the references in McIlvenny, Broth and Haddington 2009 and the papers in the special issue “Communicating space, place and mobility” in Journal of Pragmatics 2009, vol. 41/10.

414

Pentti Haddington

go about creating meaning in social interaction, i.e. how actions are produced, oriented to and recognized, are essentially cognitive phenomena (e.g. Antaki 2006; Goodwin 2000; Goodwin and Goodwin 1996; Hutchins 1995; Levinson 2006a; Mondada 2006; Tomasello 1999). It has been highlighted, for example, that human cognition should not be seen as separate from its situated environment.4 In addition to this, it has been argued that cognition must play a significant role in how interlocutors are able to interact with their co-participants: to act together with them, to interpret and respond appropriately to actions made by others, and to anticipate actions and events (see e.g. Goodwin 2003: 36–37; Hutchins 1995: xi-xviii; Mondada 2006: 117; Nevile 2004: 473). As yet, very little is known about these interactional and intersubjective properties of cognition. In the analysis in this chapter, cognition is seen to become real and visible in human conduct in 1) how participants produce actions through the different designs of language, gestures and the body; 2) how participants respond to prior actions and events and thus by visibly connecting to them, display how they identify, recognize and understand them; 3) how participants orient to features of material objects and space, and 4) how participants time their actions with respect to events and actions. Consequently, in this paper I view cognition and cognitive phenomena not as mental capacities and processes, but as social actions and practices that emerge and are dealt with in social interaction (see also Haddington and Keisanen 2009: 1940) and that not only reflect the workings of cognition but which also organize cognition (see Goodwin 2003). The above conceptualization of cognition also leads us to the way in which space is demarcated in this chapter: space is linked to action. Action is located in space and can be connected to its material shapes and features, as well as to objects and other people in it. By acting in particular ways, people display how they experience their spatial lifeworld and at the same time define, shape, create and attach meanings to that space (see also Goodwin 2003). As was mentioned above, the audio-video data that are used in this chapter have been collected in cars. When people travel in cars, the environment around them is continuously changing and transforming. The ‘mobile space’, through which the car is moving, provides an endless resource of spatial features that can be invoked, made relevant and meaningful for action. 4

Different notions have been used to refer to this idea, such as “situated cognition”, “shared cognition”, “extended mind” and “extended cognition” (e.g. Hutchins 1995, Wilson 2004)

Action and space: Navigation as a social and spatial task

415

For in-car navigation, this mobile space is a particularly important resource, and in order to navigate successfully, participants must together create a shared understanding of how the surrounding space or a feature in it is relevant for them and what they can or cannot do in it. Inside the mobile space, any ‘location’ can be invoked, as a kind of snapshot of space, as a relevant and consequential feature of space (Haddington and Keisanen 2009: 1941). Moreover, when drivers and passengers navigate in cars, they have to coordinate their actions in the car’s interactional space5 (i.e. the in-car space, see e.g. Noy 2012) with the events and features in the mobile space. In such situations temporality also starts to play an important role. As will be shown, incar participants communicate through their actions when an action is relevant and required in space. ‘Space’ is therefore an umbrella term used to refer to different kinds of spatialities and spatial configurations that participants orient to and create through their actions and which can be seen to impact the design of their actions.

3.

Action shaping space: Collaborative construction of space for navigation

Interlocutors can frequently be heard to describe spatial features or name places, thereby constructing meaningful spaces and places in ways that are relevant for the purposes of a particular activity (Psathas 1986; Schegloff 1972). In this section, I will analyze an extract in which participants refer to space in different ways and at the same time make it (and locations or referents in it) socially relevant and meaningful for navigation in the sequentially unfolding interaction. The analysis will show that the way in which space is understood, not understood or negotiated becomes evident in the participants’ actions and responses. In Extract 1, two exchange students and friends are driving in a city that they do not know very well. They are using English as a lingua franca. The purpose of their drive is to find a rock club to which both of them are going to see a concert that same evening. They do not have a map and although they do display some familiarity with the area, they do not know the exact route to the destination. In spite of this, they are successful and in the end find the rock club (data not shown). As the extract shows, the participants rely on different kinds of spatial resources and knowledge for their navi-

5

For similar uses of “interaction space” see Mondada 2007a, 2009, Goffman 1963 and Kendon 1990b. See also Goodwin 2000.

416

Pentti Haddington

gation activity. The navigation activity begins when the front-seat passenger (FP) asks the driver if she knows where they are going (line 1).6 (1) Collaborative construction of space: Talk&Drive #003_2 (1:43) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

FP: DRV:

FP: DRV:

DRV: FP:

16 17 DRV: 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

6

FP: DRV:

FP: DRV: DRV: FP: DRV: FP: FP: DRV: FP:

.hh >Do you actually know where we’re going. =I mean, I know where we [are, but do you know where-]< [U:::hm]: I don’t really know where it is, or >where it could beENV = environment, X = unidentified participant. Bold text underneath transcribed talk mark embodied actions and refer to figures that illustrate them.

Action and space: Navigation as a social and spatial task 35 DRV: 36 FP: 37 DRV:

417

&To the right. ((gesture)) Yeah. ((nodding)) &Fig.5 Heh[he]. [Of course.]

Haddington (2010, 2012) shows that in-car navigation is often sequentially organized into three actions: an initiation, a response and a closing. The beginning of Extract 1 (lines 1–7) follows this structure: The passenger initiates navigation in lines 1–2 by asking if the driver knows where the club is. By initiating the activity, the passenger voices her uncertainty of both the exact location of the club and the route to it. At the same time, she establishes her role as an active participant and a resource for driving and navigation. In the response (lines 3–5), it turns out that the driver has only a vague idea of the club’s location, after which the passenger says Okay. (line 7) and closes the navigational sequence. Usually, a navigational problem is solved during such a three-part sequence. However, in this case, the navigational problem – i.e. how to get to the club – is not resolved. This becomes evident in the following turn (lines 9–11), when the driver suggests a practical solution to their unresolved problem: they can stop and ask someone for help, which is not taken up by the passenger. The driver then self-selects and says °hm° .hh Yeah, this is like-. Although the turn is never finished, it projects an utterance that acknowledges that they have not solved the navigational problem. After this, in line 15, the passenger continues the navigation activity by voicing one solution to the navigational problem: So basically we have to go (0.3) south. The passenger thus uses an absolute frame of reference, a cardinal point (“south”) to refer to a direction in space. Indeed, the rock club is, correctly, located almost exactly to the south from their current location (see Fig. 1). It is possible that the passenger knows that the club is located to the south of the city center and that, since she knows where they are right now (see line 2 in the transcript), she is able to position themselves in abstract space relative to the club. However, her knowledge of the club’s location relative to their current location is also manifested in her actions: during her utterance, she turns her gaze to the right, towards the cardinal point that she refers to in talk (see Fig. 2). Her gaze is not directed exactly towards south, but somewhere between south and southwest. Still, by shifting her gaze even this accurately she shows that for her “south” is not an abstract direction on a map, but that she is able to position her current whereabouts with respect to their destination. The use of absolute references for pointing at referents in space is of course not unheard of, as Levinson’s (1996, 2003) research shows. However,

418

Pentti Haddington

Fig. 1. The car’s and the club’s location marked with arrows.

in contrast to the use of relative frames of reference (left, right), their use in Western languages and cultures is not systematic. As for its use in Extract 1, it would seem that there is a practical reason why the passenger uses the absolute reference for referring to the direction of the club’s location. Haddington (2012) has shown that when people initiate navigation for a longer route or to a particular destination, as in Extract 1, they usually name places and streets and thereby identify landmarks that are on their way to the destination. However, in Extract 1, neither the passenger nor the driver seems to know the city well enough to be able to name streets and places in it. Still, the navigational and spatial problem remains to be solved. The pas-

Action and space: Navigation as a social and spatial task

419

Fig. 2. Passenger looking “south”

senger’s use of the cardinal point “south” in space is therefore one resource that she can use for solving that navigational puzzle and which can help them take one step towards the right direction, in spite of the fact that they still do not know the exact way to their destination. This shows that people do not conceptualize or perceive space in a single, pre-determined way, but that they define and install descriptions of space that relevantly respond to the situated demands of a spatial problem (see also Mondada 2007b: 66). For solving this particular navigational problem, the use of absolute references is not ideal. The driver, then, rather than building on the passenger’s spatial description, invokes another strategy. In line 17, in her utterance Basi-

420

Pentti Haddington

Fig. 3. The shape and trajectory of the driver’s pointing gesture and the passenger’s gaze shift toward the gesture.

cally we have to go soon like there. she uses a locational description there. It is accompanied with a pointing gesture that begins from the driver’s lap and continues first upward in front of the driver’s face and then curves slightly to the right (see Fig. 3). The shape and trajectory of the gesture reflect the shape of the road which curves slightly to the right (see Fig. 4) and thereby complements and clarifies the deictic expression “there”. In other words, the driver’s spatial and directional description is produced with respect to and reflects the immediate and visually available location (Goodwin 2007). As the driver produces the gesture, the passenger turns to look at it and is thereby also able to see the driver’s action and the spatial reference. After this, the driver further clarifies the spatial reference by saying I go the same route as bus is going. The driver therefore relies on yet another resource to refer to and describe space: her knowledge of a bus route. In line 19, the passenger says O:kay. and thereby confirms that she has understood the driver’s spatial description and how it relates to their driving.

Action and space: Navigation as a social and spatial task

421

Fig. 4. Shape of the junction and possible routes

At this stage, the driver and the passenger have used different resources and practices in an attempt to resolve a spatial puzzle. Still, they have been able to establish only a general understanding of how to get to their destination. In lines 20–24 they again jointly voice their uncertainty of the exact location of the club and how to get there. But at this stage a new spatial task emerges: where to turn at the next junction. As we can see in Fig. 4, in which the car’s location is marked with an arrow, at this junction it is possible to make an immediate right turn (A), to drive straight for a short distance and then turn right (B) or to drive straight (C). As Haddington (2010) shows, drivers can initiate a junction-negotiation activity by producing for example questions and confirmation requests. This is what the driver does when she says Is it now left or straight. (line 25). The design of her question as an alternative question reflects the physical shape of the junction and the fact that two driving decisions are relevant, one immediately after the other. It is noteworthy that she makes a descriptional error when she invokes “left” as an alternative. In fact, a left-turn is not possible at this junction and as it turns out later, she has made an error and actually means a right-turn. After this she produces a candidate answer in the next turn (line 27) by saying Straight. In line 28, the passenger agrees and then starts but never finishes a response regarding the immediately following

422

Pentti Haddington

Fig. 5. The driver’s pointing gesture as part of a repair

turn-at-a-junction. The driver (in line 29) then agrees and suggests a leftturn. After this the passenger initiates, with high pitch and loud voice quality, a repair in line 32. The repair initiation attempts to correct the driver’s spatial description and the driver indeed complies and produces a repair in line 35. The repair is accompanied again with a pointing gesture (see Fig. 5). The passenger’s repair initiation and the driver’s repair are evident of a discrepancy in how the surrounding space has been socially constructed and understood to be tied to the driving activity.7 Although it is likely that the driver’s space formulation (to the left) is a slip, i.e. she confused left and right, the example shows that such mistakes can be treated as accountable, to be contested and then solved in interaction. Space is not a backdrop to social action. As the above analysis shows, space is made relevant, talked about, referred to, negotiated, and even disputed in and through the participants’ sequentially unfolding actions. We see how the driver and the passenger rely on very different kinds of linguistic and embodied resources to describe and refer to space in order to solve a navigational task (both ‘absolute’ and ‘relative’ frames of reference) and that very different kinds of spatial puzzles have to be solved for navigation (the ‘non-seeable’ location of the destination, the route to the destination and the ‘seeable’ next junction). Moreover, the space that is talked about and referred to8 is produced and understood as part of and relevant for a social activity, namely navigation. The analysis also shows that the driver and the passenger do not have sublime spatial knowledge required to get to their destination without negotiation and navigation. Nevertheless, even a vague idea of the 7 8

For analysis of discrepancies in spatial descriptions, see also Mondada 2010. Mondada 2011 uses the notion “reference space” to refer to the ways in which a description of relevant space is formulated in talk.

Action and space: Navigation as a social and spatial task

423

destination’s possible location is a resource that provides a starting point for step-by-step navigation, and they are finally successful in finding their original destination, the rock club. In sum, when people navigate, very different kinds of spatial puzzles can emerge: Where are we? Where is the destination? How can we get to the destination? Where to turn at the next junction? People can refer to and rely on very different kinds of spatial resources and spatial knowledge to solve these puzzles, such as cardinal points, relative frames of reference (left, right, straight), as well as different features of the location and their knowledge of the area. But what is important to note is that this knowledge of spatial features, concepts and terms is modified by and adapted to the situated and practical requirements of the unfolding activity. Thus, space and spatial meanings are constructed and negotiated in and through social interaction as they are relevant to an unfolding activity. Finally, even though the participants do not know the exact route to their destination, they are able to navigate together, help each other and to correct each other’s mistakes. Consequently, talk and sequential action are resources for creating and organizing a shared understanding of space and how that space is relevant for social action.

4.

Space shaping action: Adjusting actions with temporal and spatial demands

The interconnectedness between space and time in language has been studied broadly, especially in cognitive linguistics. These studies have discussed, for example, how thoughts and perceptions of time and space are reflected in linguistic structure (see e.g. Tenbrink 2010). The aim of this section is to show how space and time can be oriented to as meaningful elements and to shape the design of action. The analyses focus on navigational initiations (see also Haddington 2012) and especially on how different spatial and temporal demands impact on and are communicated through the multimodal design of the initiation, and finally how these understandings are also ratified in the response. In Extract (2) a family of five has just set off after enjoying a weekend lunch in a small British village pub. The family comprises a mother (driver), father (front-seat passenger) and three children (in the backseat). In the extract, the driver and the front-seat passenger are admiring and evaluating some of the buildings in the village. At the same time the car is approaching a junction. In line 1 and in Fig. 6 we can see how just before making an observation about a house in front of the car, the driver turns on the indicator to

424

Pentti Haddington

mark a right-turn (see line 1 in the transcript). Nevertheless, she then immediately requests for confirmation about where to turn at the next junction (Shall I go then-, in line 1) (see also Haddington 2010, to appear). (2) Negotiating the junction: Habitable Cars, Secluded Village (0:44) 1

DRV:

2 3 4 5 6 7 8 9 10 11 12 13 14

FP:

FP: FP:

X: FP: DRV:

15 FP: 16 DRV: 17 FP:

#Well tha%t house is going to fa:l&l=Shall I [(Fgo then-)] #Fig.6a ((DRV turns on the indicator)) %Fig.6b ((DRV points at a house)) &Fig.6c ((DRV turns gaze to right)) [O:h Flo]ok, it’s beautiful, isn’t it. (1.1) Lovely little cottage. (1.7) Lovely. (.) F ((indicator turns off)) (0.7) Da::d? (0.9) >This is a lovely little village isn’t it.< =>(Sha%ll) I go then [this &way?(Shall) I go then this way?(Shall) I go then this way? < -signs in line 14). These prosodic features show that the driver is orienting to the fact that a resolution on where to turn at the next junction is required before the car arrives at the junction. Third, the passenger’s response overlaps with the

428

Pentti Haddington

initiation (see Yeah. in line 14). In other words, the response is temporally positioned to ‘come early’, which shows that the passenger shares the driver’s understanding that a quick resolution, before a location in space, is required. Finally, in-car participants frequently use pointing gestures when they are negotiating junctions (see Fig. 7) and thereby connect their navigational action to the seeable space around them also through an embodied action. In contrast to Extract 2, Extract 3 below shows an example of a navigational initiation that rather than referring to a next junction in the seeable space, refers to a longer route and identifies landmarks in the non-seeable space. As will be shown, the way in which navigational initiations to longer routes are designed and how they are responded to is in marked contrast with navigational initiations to junction-negotiations. Consequently, it shows how space, as it is considered relevant by the participants for their navigation, can shape and determine action in very different ways. In the following extract, four friends are driving to a shopping mall. In line 3, it becomes apparent that there are two alternative routes that they can take: The driver asks the passengers which one they would prefer by asking: Uhm, Do we want to go the Mark Spencer way or the freeway way. (line 3). (3) Alternative routes: Habitable cars, Pilot Cameras (2:25) 1 2 DVR: 3 f 4 5 RRP: 6

…(3.5) Uhm, Do we want to go the Mark Spencer way or the freeway way. …(1.8) (TSK) U=uhm= , …(2.0)

As Haddington (2012) shows, navigational initiations that contain references to place names or road names or other landmarks (“Mark Spencer way” and “The freeway way” in Extract 3) anticipate navigation for the distant future and not for the next junction (cf. the use of deictic forms in Extract 2). In addition, while navigation for the next junction is usually accomplished very quickly (see Extract 2), Haddington (2012) also shows how navigational initiations which locate a spatial reference in the non-seeable space are not treated to require quick decisions. Indeed, in contrast to Extract 2, the driver’s question in Extract 3 does not get a clear and immediate response, but is followed by a pause (line 4), hesitation (line 5) and then another pause (line 6). In sum, the analysis reveals some ways in which space shapes and organizes the participants’ navigational actions. More specifically, it shows that the

Action and space: Navigation as a social and spatial task

429

ways in which participants design their actions display their differential sensitivity to seeable and non-seeable landmarks and locations in space and how these are consequential for, modify or constrain what participants do and how they can do it. Furthermore, it shows that participants adjust their actions temporally with respect to the demands that the mobile space imposes on them. When an action is required before a location in the seeable space, participants also produce their actions quickly. When there is not an immediate need to make a decision, because a relevant location is not in the immediate seeable space, participants can delay their actions. Interestingly, the relationship between a location in space and action is not formulated in meters, miles or kilometres (as GPS navigators do it), nor in seconds or minutes, but through the organisation and design of turns-at-talk. Finally, the analysis shows that navigating in mobile space is often a collaborative task and that the ways in which space impacts on action are not treated as an individual’s concern only but interlocutors construct them in concert with each other.

5.

Conclusions: Interacting space

The link between action and space was already made by Goffman (1964) when he claimed how material settings have an impact on speech behavior. Similar observations about how actions are accomplished in some relation to or are even closely tied to space have also been made by Kendon (1977: 180), Goodwin (2003), LeVine (2007: 266), Mondada (2009: 1995), inter alia. In this chapter, with the help of video recordings made in real-life and naturallyoccurring social interaction and by doing micro-detailed sequential analysis of interlocutors’ linguistic and embodied actions and activities, the general aim has been to further conceptualize the relationship between space and action. First, the aim has been to show how action shapes space. The analysis has shown how interlocutors in and through their social actions refer to, describe, identify and name spatial features and thereby try to establish a shared understanding of space and to solve a practical spatial puzzle for a particular activity. The analysis suggests that space is not just an abstract entity that pre-exists and influences behavior but that it is systematically and dynamically made relevant and constructed for action. Second, the aim has been to show how space shapes action, i.e. how pre-existing space and spatial configurations or understandings of it occasion actions and how social actions and practices are modified and adjusted with regard to the demands of space. The interactional and sequential analysis in this chapter shows that in social interaction participants construct space and make it relevant for the practical purposes of social action. It is a contextual feature that can become

430

Pentti Haddington

an important resource for the production and understanding of social actions. It is not an abstract variable or a static background, but is fluid, emergent, dynamic and consequential and it can be voiced and become manifest in action in different ways. The analysis has focused on navigational initiations and navigational sequences in cars and it has shown that participants define and attach meanings to space and adjust their actions relative to it through their verbal and embodied behavior and for the practical purposes of accomplishing navigation. It has also shown that the use of particular spatial language is contingent on the interactional situation and the kind of navigational or spatial puzzle that the participants are trying to solve and that participants adjust their actions temporally on the basis of the demands imposed by space. These findings suggest that in real-life interaction interlocutors do not treat space as a fixed or pre-existent reality, but as an emergent and contingent feature of context that they form and establish in and through social interaction (see also Mondada 2009: 1995). The analysis also shows that seeing, creating, orienting to and structuring space for the practical purposes of action is a collaborative achievement. Space cannot only be seen as an individual phenomenon, because there often is a need to communicate to others – for example by asking, informing, noticing or requesting – an understanding of space and how space is relevant for action. People also use a multitude of different multimodal resources for interpreting and communicating how space features in an action or an activity. It seems, therefore, that people do not just express their understandings of space through language, but that they interact space by relying on different multimodal resources. One can therefore perhaps ask whether, in naturally-occurring interaction, there really is such a thing as ‘spatial language’, or rather a thing called ‘spatial action’. In a similar vein, one can also ask whether ‘spatial action’ is a linguistic and stored mental phenomenon, or rather an embodied phenomenon that includes language and is made visible and organized interactively in social interaction. Further, one can perhaps ask if, in addition to considering cognitive processes that underlie language structure, it would be relevant to discuss spatial cognition as a situated and embodied process that emerges and is communicated as an interactive practice in semiotically rich environments and through social interaction (cf. Goodwin and Goodwin 1996, Goodwin 2003, Nevile 2004: 473; Haddington and Keisanen 2009: 1957–1958).

Action and space: Navigation as a social and spatial task

6.

431

References

Antaki, Charles 2006: Producing a ‘cognition’ Discourse Studies 8(1): 9–15. Atkinson, John M. & John Heritage 1984: Transcript notation. In: John M. Atkinson & Heritage John (eds.), Structures of Social Action: Studies in Conversation Analysis, ix–xvi. Cambridge: Cambridge University Press. Broth, Mathias 2009: Seeing through screens, hearing through speakers: Managing distant studio space in television control room interaction. Journal of Pragmatics 41: 1998–2016. Goffman, Erving 1963: Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press. Goffman, Erving 1964: The neglected situation. American Anthropologist 66(6): 133–136. Goodwin, Charles 2000: Action and embodiment within situated human interaction. Journal of Pragmatics 32(10): 1489–1522. Goodwin, Charles 2003: Pointing as situated practice. In: Sotaro Kita (ed.), Pointing: Where Language, Culture and Cognition Meet, pp. 217–41. Mahwah, NJ: Lawrence Erlbaum. Goodwin, Charles 2007: Environmentally coupled gestures. In: Susan D. Duncan, Justine Cassell & Elena T. Levy (eds.), Gesture and the Dynamic Dimension of Language, 195–212. Amsterdam/Philadelphia: Benjamins. Goodwin, Charles & Marjorie Harness Goodwin 1996: Seeing as a Situated Activity: Formulating Planes. In: David Middleton& Yrjö Engeström, (eds.), Cognition and Communication at Work, 61–95. Cambridge: Cambridge University Press. Haddington, Pentti 2010: Turntaking for turntaking: mobility, time, and action in sequential organisation of junction negotiations in cars. Research on Language and Social Interaction 43(4): 372–400. Haddington, Pentti 2012: Movement in action: Initiating social navigation in cars. Semiotica 191 (1/4): 145–170. Haddington, Pentti & Tiina Keisanen 2009: Location, mobility and the body as resources in selecting a route. Journal of Pragmatics 41: 1938–1961. Hutchby, Ian & Robin Wooffitt 2008: Conversation Analysis (2nd edition). Cambridge: Polity Press. Hutchins, Edwin 1995: Cognition in the Wild. Cambridge, MA: MIT Press. Jefferson, Gail 2004: Glossary of transcript symbols with an introduction. In: Gene Lerner (ed.), Conversation Analysis: Studies from the First Generation, 13–31. Amsterdam: Benjamins. Keisanen, Tiina 2012: ‘Uh-oh, we were going there’: Environmentally occasioned noticings of trouble in in-car interaction. Semiotica 191: 197–222. Kendon, Adam 1977: Studies in the Behavior of Social Interaction. Bloomington: Indiana University Press. Kendon, Adam 1990a: Behavioral foundations for the process of frame-attunement in face-to-face interaction. In: Adam Kendon (ed.), Conducting Interaction: Patterns of Behavior in Focused Encounters, 239–262. Cambridge: Cambridge University Press. Kendon, Adam 1990b: Spatial coordination in social encounters: The f-formation system. In: Adam Kendon (ed.), Problems of Behavior in Focused Interaction, 209–237. Cambridge: Cambridge University Press.

432

Pentti Haddington

Laurier, Eric 2005: Searching for a parking space. Intellectica 41–42 (2–3): 101–116. Lee, David 2001: Cognitive Linguistics: An Introduction. Oxford: Oxford University Press. LeVine, Philip 2007: Sharing common ground: The role of place reference in parentchild conversation. In: Deborah Tannen, Shari Kendall & Cynthia Gordon (eds.), Family Talk: Discourse and Identity in Four American Families, 263–282. Oxford: Oxford University Press. Levinson, Stephen 1996: Language and space. Annual Review of Anthropology 25: 353–382. Levinson, Stephen 2003: Space in Language and Cognition. Cambridge: Cambridge University Press. Levinson, Stephen 2006a: Cognition at the heart of human interaction. Discourse Studies 8 (1): 85–93. Levinson, Stephen 2006b: On the human ‘interaction engine’. In: Nick Enfield & Stephen Levinson (eds.), Roots of Human Sociality, 39–69. Oxford: Berg. McIlvenny, Paul 2009: Communicating a ‘time-out’ in parent-child conflict: embodied interaction, domestic space and discipline in a reality TV parenting programme. Journal of Pragmatics 41: 2017–2032. McIlvenny, Paul, Mathias Broth& Pentti Haddington 2009: Editorial: Communicating place, space and mobility. Journal of Pragmatics 41: 1879–1886. Mondada, Lorenza 2006: Participants’ online analysis and multimodal practices: projecting the end of the turn and the closing of the sequence. Discourse Studies 8(1): 117–129. Mondada, Lorenza 2007a: Interaktionsraum und Koordinierung. In: Schmitt, Reinhold (ed.), Koordination. Analysen zur multimodalen Organisation, 55–94. Tübingen: Narr. Mondada, Lorenza 2007b: Operating together through videoconference: Members’ procedures for accomplishing a common space of action. In: Stephen Hester & David Francis (eds.), Orders of Ordinary Action, 51–67. Aldershot: Ashgate. Mondada, Lorenza 2009: Emergent focused interactions in public places: A systematic analysis of the multimodal achievement of a common interactional space. Journal of Pragmatics. Special issue on communicating place, space and mobility (41): 1977–1997. Mondada, Lorenza 2010: Reassembling fragmented geographies. In: Monika Büscher, John Urry & Katian Witchger (eds.), Mobile Methods, 138–163. London: Routledge. Mondada, Lorenza 2011: The interactional production of multiple spatialities within participatory democracy meetings. Social Semiotics 21(2): 289–316. Nevile, Maurice 2004: Integrity in the airline cockpit: Embodying claims about progress for the conduct of an approach briefing. Research on Language and Social Interaction 37(4): 447–480. Noy, Chaim 2012: Inhabiting the family-car: Children-passengers and parents-drivers on the school run. Semiotica 191 (1/4): 309–333. Pomerantz, Anita 1984: Agreeing and disagreeing with assessments: some features of preferred/dispreferred turn shapes. In: Atkinson, John M. & Heritage, John (eds.), Structures of Social Action: Studies in Conversation Analysis, 57–101. Cambridge: Cambridge University Press. Psathas, George 1986: Some sequential structures in direction-giving. Human Studies 9: 231–245.

Action and space: Navigation as a social and spatial task

433

Schegloff, Emanuel A. 1972: Notes on conversational practice: Formulating place. In: Sudnow, David N. (ed.), Studies in Social Interaction, 75–119. New York: Free Press. Schegloff, Emanuel A. 2007: Sequence Organization in Interaction: A Primer in Conversation Analysis. Cambridge: Cambridge University Press. Te Molder, Hedwig & Potter, Jonathan (eds.) 2005: Conversation and Cognition. Cambridge: Cambridge University Press. Tenbrink, Thora 2010: The language of space and time. Journal of Pragmatics 43(3): 691–694. Tomasello, Michael 1999: The Cultural Origins of Human Cognition. Cambridge: Harvard University Press. Wilson, Robert A. 2004: Boundaries of the Mind: The Individual in the Fragile Sciences – Cognition. Cambridge: Cambridge University Press.

434

Elwys De Stefani

Elwys De Stefani

Rearranging (in) space On mobility and its relevance for the study of face-to-face interaction

1.

Introduction

Talk has been described as “one of the most pervasive social activities that human beings engage in” (Goodwin & Duranti 1992: 1) and as “the primordial site for sociality” (Schegloff 1987: 208). It is through talk – taking place in a spatial framework – that social relationships are accomplished and maintained. Face-to-face interaction thus emerged as a central object of research within anthropology, social psychology, sociology and linguistics (see Kendon 1990: 15–49). In psychology, the postures that patients assume during therapy sessions have been central to Albert Scheflen’s (1964) work.1 While these early approaches were oriented towards documenting and analyzing human behavior with the help of film recordings, Erving Goffman’s sociological method was based on direct observation and offered descriptions of interactions involving two or more parties. He understood focused interaction to be tightly linked to a face engagement or encounter – defined as “all those instances of two or more participants in a situation joining each other openly in maintaining a single focus of cognitive and visual attention” (Goffman 1963: 89). Thus defined, an encounter can be described at least with regard to two dimensions that are commonly acknowledged as being constitutive for face-to-face interactions, i.e. a) the orientation towards a “single focus of cognitive and visual attention”, and b) the fact that participants sustain their orientation towards a focal event in public, in a visually perceivable way. A detailed analysis of the participants’ body postures during focused interactions is given by Kendon (1990), who calls F-formations the positionings that allow the individuals to dispose their bodies in each other’s proximity and to have

1

In the 1920s and 1930s, Kurt Lewin and Arnold Gesell had addressed similar topics, studying body movements of very young children. Their work was already based on filmed material and led to the definition of a specific methodology termed cinemanalysis (see Gesell 1935).

Rearranging (in) space

435

visual access to each other’s actions.2 Although a substantial amount of research has been published on face-to-face interaction, the notion itself appears to be underspecified, e.g. with regard to the underlying requisite for the participants’ co-presence. In addition, it appears that work on face-to-face interactions is predominantly looking at stationary settings. But participants may modify their position in space as they interact. Until recently, modifications affecting the participants’ positions in the spatial environment have only been scarcely explored by interactionally oriented researchers. Moreover, only a small amount of research has focused on how spatiality is related to the actions that participants are engaged in (see Mondada 2005; McIlvenny, Broth & Haddington 2009; De Stefani 2010; Haddington, Mondada & Nevile 2013). The aim of this paper is thus to reconsider the notion of face-to-face encounter in the light of mobile settings of interaction and to show how participants use space as a resource for interaction.3

2.

Mobility

In this paper I will focus on mobile settings, i.e. on interactions that take place among participants who are moving in space. Mobility has begun to be studied in social science (see Urry’s 2007: 6 description of the mobility turn in sociology) and in the humanities as a central feature of everyday life. Within interactionally oriented explorations, mobility is generally studied with regard to specific research questions, analyzing e.g. route descriptions (Psathas 1986, 1991), navigational practices (Psathas 1976; Hutchins 1995; De Stefani & Mondada 2007), car interactions (Laurier et al. 2008; Haddington & Keisanen 2009; Haddington 2010), museum exhibits (Heath & vom Lehn 2004; Hindmarsh et al. 2005) to name but a few. These studies explore how mobility relates to other dimensions of interaction that have habitually been investigated on the basis of stationary interactions, such as the sequential organization of interaction, turn construction, multimodal resources, etc. Previous research has shown that the label mobility is used to cover a large amount of observable phenomena. This conceptual diversity has rarely been questioned in interactional studies. However, it appears to be useful to introduce a minimal distinction between two sorts of mobility, for all practical pur2

3

See also Clark’s 1973 notion of canonical encounter, describing a situation in which participants are facing each other. Kendon 1990: 251 states that “there is a systematic relationship between spatial arrangement and mode of interaction”: the observation of the participants’ bodily (re)arrangements in mobile settings seems thus to be the next natural step in the analysis of human interaction.

436

Elwys De Stefani

poses: a) mobility as a recurrent feature of face-to-face interaction, as it is observable in local, short-range repositionings and reorientations of the participants’ bodies (that may or may not be achieved through walking etc.); b) mobility as a specific kind of social interaction, such as walking-together (Relieu 1999), running-together (Collinson 2006), driving-together, etc. In the analytic part of this paper, I will focus on both understandings of mobility, concentrating particularly on the ways in which participants constitute a relevant space for their interaction and on their use of spatial features as resources for the interaction at hand. The data analyzed here takes into account settings that are not canonical face-to-face conversations, precisely because the participants are not constantly facing each other (ex. 1 and 2) or because they are sitting side-by-side (ex. 3 and 4).4 In all the excerpts that I will analyze, the participants are engaged in a place description of some kind. One further aim of this paper is thus to show that the participants’ movement in space is both a resource that allows them to deliver an accountable spatial description, as well as a contingency with which they have to cope. Before turning to the analysis of the data, I find it appropriate to quickly recall an early interactional account of a recurrent practice related to space, namely place formulation.

3.

Place formulations in interaction

If referential practices to the spatial5 environment occur in interaction, they are fine-tuned according to the specific action in which the participants are engaged and to the way in which the participants are categorized in and through their interaction (Schegloff 1972; Clark & Wilkes-Gibbs 1986). In his seminal 4

5

The data analyzed in this paper stems from a research project entitled “The constitution of space in interaction: a conversation analytic approach to the study of place names and spacial descriptions”, that was carried out at the University of Bern (Switzerland). It was directed by the author of this paper and profited from the collaboration of two team members, Anne-Danièle Gazin and Anna Claudia Ticca. A former member, Roberta Iacoletti, contributed essentially to collecting and transcribing the data. I thank the Swiss National Science Foundation for the generous support for the period 2008–2012 (project number PP001–119138). I will use the terms place and space indistinctively as both are employed in interactionally oriented work: Schegloff 1972 speaks of place formulations while in recent work we find spatial description (e.g. Mondada 2009: 1194) for roughly the same phenomenon. A theorization of these notions has notably been developed in geography: according to Tuan 1977: 54 “enclosed and humanized space is place. Compared to space, place is a center of established values”. See also Lefebvre 1974 and Casey 1997.

Rearranging (in) space

437

paper, Schegloff (1972) describes five different kinds of place formulations, namely a) geographical formulations (addresses, degrees of latitude and longitude), b) relation to members formulations (such as Chuck’s house), c) relation to landmarks formulations (e.g. nearby the bridge), d) course of action places (like where they put the rubbish) and e) place names. While it appears surprizing that Schegloff would not mention deictic formulations as one further resource that participants employ for describing space, his paper has the merit of demonstrating that participants choose the appropriate place formulation by virtue of a) a location analysis (where are the speakers located with regard to the place that is being described?), a membership analysis (how are the speakers reciprocally categorized?), and c) a topic or activity analysis (which topic is being discussed and/or what is the activity the participants are engaged in?). These analytic procedures of the participants allow them to deliver and to understand place formulations that are appropriate to the task at hand: in other words, these formulations are right for all practical purposes without necessarily being correct from a strictly referential point of view (see also Mondada 2000). What has scarcely been taken into consideration in studies on place formulations is the fact that numerous interactions occur while participants are on the move. The analysis of mobile interactional configurations confronts the researcher with a number of methodological questions, affecting for instance the techniques and practices of data collection in a continuously changing environment with continually varying participation frameworks. Nevertheless, the study of how talk is organized in such settings may offer a better understanding of the reflexive relationship that ties the participants’ interactional practices to the space in which they are moving. There is a large consensus among interactionally oriented scholars on the reflexive relationship that ties a specific social encounter to the (spatial) context in which it is taking place. But while Conversation Analysis has been able to convincingly show how participants continuously reshape, restructure the relevant context through talkin-interaction, it has proven much more difficult to understand how the semiotic fields (Goodwin 2000: 1494) that are made relevant in a social encounter (stationary or mobile) have an impact on the organization of interaction. The following considerations focus on some recurrent observables in mobile episodes of interaction. I will start by illustrating how two participants engaging in a face-to-face interaction constitute the relevant space for their interaction and how specific projected actions (such as indicating someone the way) require a repositioning of the participants’ bodies (§ 4); on the basis of guided tour data, I will then look at how the spatial repositioning of the participants is central for constituting a group that is “identifiable as a proper togethering” (Ryave & Schenkein 1975: 270). In addition, the use of

438

Elwys De Stefani

features inscribed in space for the constitution of a noticeable phenomenon will be analyzed (§ 5); in the last analytic section I will focus on instruction sequences occurring during a driving lesson and analyze the reflexive relationship between the overall mobile setting, the visibility of emerging landmarks and the turn format that is chosen for the instruction (§ 6).

4.

Achieving a face-to-face interaction: Requesting the way

It is commonly assumed that “face-to-face interaction provides the primordial locus for the production of talk” (Goodwin & Duranti 1992: 22) and that it is the “phylogenetic and ontogenetic habitat of natural language” (Ford, Fox, and Thompson 2002: 4). However, the notion of face-to-face is often used in a rather indefinite way, as it tends to be applied to situations of co-presence. Linguistic researchers only rarely question the different ways in which co-presence may be achieved (but see Kendon 1990). Studies on how face-to-face situations are achieved by individuals (who thereby progressively become participants) are still rare (but see Mondada 2009; De Stefani & Mondada 2010). In the first excerpt, I will thus analyze how face-to-face interaction is actively constituted by the participants as they engage in focused interaction. Moreover, I will provide a first reflection on how space is used as a resource for accomplishing the relevant tasks of the ongoing interaction. The excerpt is taken from a corpus of interactions of individuals shopping together (as couples) in a supermarket.6 In this excerpt, Teresa (Ter) is asking a sales assistant (Sas) where she can find particular boxes that she wishes to buy. The interaction can thus be described as an itinerary request followed by a route description: (1a) cons45111 (27:41–27:51) 1

2

3

6

TER

((smack)) ’h signorina mi ˙sa di˙re *dove sono le miss can you tell me where there are ter *.............--> fig ˙2 ˙3 *s˙catole (0.7) per f- (0.2) per met˙ter vi*a the boxes to to store away ter --> *iconic gesture “boxes”--------------------*,,--> fig ˙4 ˙5 *la roba*˙ d’estate o d’inverno ma the things for summer or winter but ter --> *,,,,,,,* ˙6

For a detailed presentation and analysis of the corpus see De Stefani 2011.

439

Rearranging (in) space 4

non in eh ma[terialenot in uh material 5 SAS [le scatole/ the boxes 6 (0.4) 7 TER *le ˙scatole in car*tone *˙ the boxes in cardboard ter *iconic gesture----*,,,,,* fig ˙7 ˙8

The opening of the interaction between Teresa and the sales assistant can be described as a practical problem. In fact, in order to initiate a focused interaction Teresa (the requester) has to stop the employee who is walking from the right to the left side of the figures 1–4 below.

Fig. 1 (0.6s before line 1)

Fig. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

This action is grounded, first of all, on the categorization of the person passing in Teresa’s proximity. Teresa categorizes her visibly as an ‘employee’, very likely on the basis of her clothing, which she recognizably treats as staff dress. Teresa makes this categorization visible by the way in which she acts towards the person who is about to become her interlocutor. She starts her turn with the address term signorina (‘miss’; l. 1) that is followed by the formulation of a request related to a specific (commercial) object: mi sa dire dove sono le scatole (‘can you tell me where the boxes are’; ll. 1–2). This way of addressing the employee is embedded in multimodal actions that are equally oriented towards establishing a reciprocal positioning of the participants’

440

Elwys De Stefani

bodies. While mutual gaze is achieved very early and maintained throughout figs. 2–4, Teresa also employs gestural resources: as she produces her initial turn, she starts performing an iconic gesture representing the shape of a box. This gesture is initiated before the corresponding verbal referent is mentioned (l. 1) and reaches its apex as Teresa says scatole (‘boxes’; Fig. 4), after which she slowly withdraws the gesture (figs. 5–6). Carrying out this gesture is consequential for the achievement of mutual orientation: on the one hand, as she performs the gesture, Teresa turns her upper body further towards her imminent co-participant (see figs. 3–4), on the other hand, the gesture requires visual attention from her co-participant, thus displaying Teresa’s claim for reciprocal orientation of the participants’ bodies. The employee indeed ceases to walk as Teresa describes the referent she is looking for and orients her body towards Teresa (figs. 5–6). So far, I have described the achievement of reciprocal orientation – and thus the constitution of a shared interactional space – as an essential component of opening sequences. While engaging in a social encounter, indivudals accomplish a series of relevant tasks, such as mutual categorization, reciprocal positioning of the bodies, hence displaying disposition to become involved in interaction. It is thanks to these multimodal adjustments that individuals who happen to be present in the same location become co-participants in a focused encounter. Following the opening, Teresa goes on describing the product she is looking for, first with regard to the use she intends to make (‘for storing the summer and winter things’, ll. 2–3), and then by initiating a description of the material the boxes should be made of (l. 4). As the employee requests a clarification (‘the boxes’; l. 5), Teresa delivers a further description of the boxes (‘the cardboard boxes’; l. 7) and at the same time performs another iconic gesture (figs. 7–8).

Fig. 7

Fig. 8

Rearranging (in) space

441

By requesting the location of the ‘boxes’ in the supermarket, Teresa is actually asking for the way to that location. Indeed, as has been shown elsewhere, in stores, commercial products are recurrently treated as landmarks (see De Stefani 2011, 2013). The employee is thus expected to give a route description. Let us see how the way to the boxes is described: (1b) cons45111 (27:51–27:57) 8 9

sas SAS

fig sas 10 TER 11 SAS sas 12 13 TER 14 SAS sas 15 TER

*(0.8)* *.....*--> *quelle˙ per metter dentro i vesti*ti ˙[così/ the ones to put in the clothes and so on ˙9 ˙10 --> *walks----------------------------* [sì yes *sono *s- alla fine dei libri sotto la scala they are at the end of the books underneath the *.....*points--> mobile [(sono proprio lì) escalator (they are right there) [grazie thanks *(di nien[te)* it’s nothing -->*,,,,,,,,,,,,* [grazie thanks

In line 9, the employee can be seen to formulate yet a further description of the boxes (‘the ones for storing the clothes and so on’, l. 9) that is oriented towards obtaining a ratification by Teresa, who produces it in overlap at the end of the employee’s turn (‘yes’, l. 10). At first sight, this turn appears simply to be describable as an attempt of the employee to ensure herself that she has understood correctly what kind of object the customer is seeking. Seen this way, the employee’s turn appears to be a repair initiation of a referential trouble. However, if we take into account the mutimodal actions that the employee is performing while producing her turn, we can see that she is actually already oriented towards the spatial description that she is going to formulate next.

442

Fig. 9

Elwys De Stefani

Fig. 10

As the figures 9–10 show, while she produces her turn, the employee makes a few steps and repositions her body towards the aisle of the supermarket that she is going to use as a relevant semiotic field when describing the itinerary (ll. 11–12). In other words, she is bringing herself into a position that will allow her to deliver the spatial description in a meaningful way, i.e. in a way that is accountable for Teresa. The employee’s repositioning can thus be analyzed as a preparation of the relevant next action, which consists in describing an itinerary. At least three observations seem appropriate on this behalf: a) verbally the employee seems to display uncertainty with regard to the identification of the object Teresa is looking for; however, by repositioning her body at the same time she projects the subsequent relevant action, namely the description of the itinerary. This embodied projection displays that the she has already identified a ‘possible’ object at this point; b) indicating the way and, more generally, describing space in interaction requires preparatory actions: the participants have to accommodate their bodies to the local environment in order to ensure that the description will be intelligible. In other words, they have to reorganize their interactional space and to constitute a common, shared origo (Bühler 1934) that allows them to use verbal and gestural deictic resources. The spatial environment is thus used as a resource for accomplishing social action (see also Mondada 2005); c) with regard to the notion of face-to-face interaction, it becomes apparent that it is patently underspecified: on the surface it appears to describe interactional language used in co-presence. But the ‘nature’ of the participants’ co-presence is rarely questioned, and even less the fact that participants may change their bodily orientations – not in a haphazard way, but with respect to the specific social actions they are engaging in.

Rearranging (in) space

5.

443

Spatial relevances in a collective activity: Guided tours

The organization of joint mobility is not only a topic of research that allows for the description of talk in mobile settings; it is also central for studying the ways in which social relationships are publicly exhibited and perceived. The social sciences have focused on the ways in which sociality is not only displayed by persons engaged in a joint social action, but also recognized by copresent individuals.7 Individuals acting together are publicly recognizable as a with, in Goffman’s (1971) terms.8 A group of persons, or a couple, is recognizable as such not only because the individuals composing it are manifestly (but not necessarily) moving in each other’s proximity. Their belonging together is also exhibited by their interactional engagement: Goffman (1963, 1971) captures this idea by describing couples as participation units. Thus, the questions that arise are related to the ways in which such participation units are organized when the individuals composing it are moving in space. How is togetherness praxeologically achieved and displayed? How are the available physical resources (such as objects, given spatial configurations, etc.) used as referential objects and how does this usage contribute to the maintenance of the participation unit? The following data is taken from a guided tour through the city of Naples. In the excerpt below the group is visiting a castle (Castel dell’Ovo) situated on a peninsula. The guide (Nina) is about to explain certain features of a terrace the group is walking on. The excerpt differs from ex. 1 with regard to two properties. First, face-to-face interaction is initiated by individuals moving on as a group. Therefore, I will look at how the participants’ spatial (re)positioning is contingent with the ongoing activity and how togetherness is achieved. Second, when producing her spatial description, the guide refers to features inscribed in the immediate environment, thus transforming those features into meaningful objects. 7

8

As Goffman’s 1963 notion of civil inattention entails, recognition of others is in general not overtly displayed. Individuals do acknowledge the co-presence of other persons in public space, but they usually exhibit other-recognition in a disattending way. Ryave & Schenkein (1974) have shown that other-recognition is one of the features that people exhibit by moving through public space in specific ways: if a walker’s path is about to cross the trajectory of a couple moving in the opposite direction, the walker will supposedly avoid the couple as a whole, rather then walking through the couple and thereby disrupting the togetherness of the two individuals. Likewise, the members of the couple will most likely try to prevent the two opposite-going parties from colliding. “A with is a party of more that one whose members are perceived to be ‘together’” (Goffman 1971: 19).

444

Elwys De Stefani

I will start the analysis by looking at how the participants position themselves in proximity of the guide as they arrive on the terrace.

Fig. 11 (0:02)

Fig. 12 (0:05)

Fig. 13 (0:08)

Fig. 14 (0.14)

Fig. 15 (0:19)

Fig. 16 (0:31)

The series of stills illustrates how the group members successively arrive on the terrace and how they gather around the guide (Nina). Fig. 11 shows Nina walking ahead of the group, embodying thus her being the group’s guide. In Fig. 12 she adopts a stationary position on the terrace and turns her body in the direction of the visitors who are moving towards her. Nina’s stationary positioning is understood by her co-participants as an invitation to gather around her and thus as the (pre-)beginning of an explanation sequence that the guide is about to deliver. In fact, figs. 13–16 show that the visitors arriving on the platform position themselves in a circle around the guide, embodying thus an F-formation that is habitual for group gatherings (Kendon 1990). As the last of the six visitors arrives, the group is recognizably re-constituted as a participation unit. It is precisely at this moment that the guide starts to deliver an explanation about the terrace the group is standing on (l. 1). Her turn starts with the language unit dunque (translatable as ‘so’) that Italian speakers use frequently to mark transition from one action to another (see De Stefani 2013):

445

Rearranging (in) space

(2a) 9222vgadVP2 (02:10–02:39)

4

5

6 7 8

9 10

11 12 13 14 15 16

17 18

19 20 21 22

e

3

e

2

NINA ’’’h ˙DUNQUE (1) ((smack)) h:: oltre a:˙ (0.2) ahl bellihssimo so uhm apart from the beautiful fig ˙16 ˙17 panorama chhe si gode:^h’˙ su questo: (0.4) fronFte ’h˙ panorama that one enjoys on this front fig ˙18 ˙19 possiamo^anche noGtare bassando- basta che guardiamo (0.2) ai we can also note lowering we just have to look at nostri pie[di˙ our feet fig ˙20 BARB [xxxx[xxx #(sta sopra)=# (is on it) nina #............#walks--> NINA [eh SABI =hm (0.3) (si ve[de) (one can see) NINA [guardate˙ look fig ˙21 (0.5) NINA ’h su questo livelloGquFi:˙ on this level here ˙22 (0.5) BARB?xxx (0.4) NINA hm (0.5) NINA ˙’’h vedete ci sono proprio:˙ il: h:# (0.3) ’’h (0.3)#[°l-°& you see there are really the nina -->#.................# fig ˙23 ˙24 BARB? [xx NINA &è la rasa#tura di strutture mura#rie\˙ it’s the razing of mural structures nina #......................#steps back--> fig ˙25 VICO xxxxx NINA che^è qui that is here VICO xxxxx[xx]xxxxxxxxx[xxx NINA [hm/] [’h ED HA una forma quadrata\# and it has a square form nina --># e

1

446

Elwys De Stefani

The initial state of reciprocal orientation is not maintained for a long time. While at Nina’s turn beginning (l. 1) the visitors are mainly looking in her direction (Fig. 16), the orientations of some members of the group change as Nina mentiones the ‘beautiful panorama’ (l. 1).9 Indeed, Fig. 18 shows four participants gazing towards the sea, and two seconds later one of the group members is walking towards the parapet (Fig. 19). Within a few seconds, the initial orientation of the visitors towards one focus of attention (Nina), has been dissolved, as the comparison between figures 17 and 19 shows.

Fig. 17 (0:35)

Fig. 18 (0:37)

Fig. 19 (0:39)

At this point Nina does no longer have the visual attention of all members of the group. This does not refrain her from delivering information about the terrace on which the group is standing. Note that she continues the initially projected syntactic trajectory at least until l. 3 (‘apart from the beautiful panorama one enjoys on this front, we can also note …’). Subsequently, Nina suspends that syntactic trajectory – as the self-repair bassando- basta (l. 3) makes visible – and formulates what can be described as an insert (basta che guardiamo . ai nostri piedi, ‘we just have to look at our feet’; ll. 3–4).10 Doing so, Nina actually accomplishes two actions that are relevant for the explanation that she is going to deliver. First, she announces that there is something ‘to note’ in the immediate environment (possiamo^anche noGtare, l. 3) – whithout mentioning what the nature of the noticeable phenomenon could possibly be. Second, she invites the visitors to reorient their attention towards ‘their 9

10

Note that Nina marks the information about the panorama as secondary or incidental right from the beginning of her turn, by using the expression oltre a (‘apart from’): oltre a: ahl bellihssimo panorama (‘apart from the beautiful panorama’; l. 1). In the beginning of her turn, Nina used an impersonal verb construction (si gode, ‘one enjoys’; l. 2), whereas following the self-repair she employs the second person plural (guardiamo, ‘we look’; l. 3) and thereby treats the group as an agentive entity (see also nostri piedi, ‘our feet’; ll. 3–4). It is also interesting that verbs of perception are used only as reorientation is explicitly invited – while talking about the panorama Nina uses the verb ‘to enjoy’ (godere).

447

Rearranging (in) space

feet’ and thereby initiates the activity of collectively discovering the noticeable phenomenon. Nina uses verbs of perception such as notare ‘note’ (l. 3), guardiamo ‘we look’ (l. 3) that make the necessity of a riorientation explicit.11 In fact, the object of the guide’s considerations has to be constituted with the help of features that are inscribed in the local environment. Nina orients her gaze towards the floor, at the same time pointing it: as can be seen in Fig. 20, at the end of Nina’s request to look to ‘their feet’, most of the (remaining) participants gaze towards the floor.12

Fig. 20 (0:44)

Fig. 21 (0:46)

Fig. 22 (0:48)

The constitution of the noticeable phenomenon is achieved through spatial descriptions that Nina accomplishes by employing a variety of resources. She starts, as we just have seen, with an invitation addressed to the group to look at the floor (‘at our feet’, ll. 3–4). Two of the visitors – the two ladies facing Nina – deliver comments that unfortunately are not entirely captured by our recording, but that are sufficiently clear to understand that one lady (Barbara) is talking about someone ‘standing on’ (sta sopra, l. 5) the object of Nina’s explanation, while the other one (Sabina) comments that ‘one can see’ (si vede, l. 6) the phenomenon. But the two visitors also embody these comments, on the one hand by orienting their gaze towards the white bricks that form a contour on the floor, on the other hand, in the case of Barbara, by stepping back from the bricks on which she was previously standing (compare figs. 20 and 21). In other words, the visitors reposition themselves so as to maximize the visibility of the current object of interest. It is this repositioning that allows Nina to proceed in the organization of the semiotic field that is relevant for her upcoming explanation. She does so by soliciting the 11 12

See also at the following lines guardate ‘look’ (l. 8), vedete ‘you see’ (l. 16). The two individuals entering the scene from the right lower corner of the stills are not part of the group. However, they will follow for a certain time Nina’s explanations, as can be seen e.g. in Fig. 22, where they orient their gaze in the direction Nina is indicating.

448

Elwys De Stefani

visitors to ‘look’ (guardate, l. 8, now in second person plural), by gazing and pointing towards the floor and by actually stepping onto the white bricks (Fig. 21). Nina now produces a series of turn extensions: at l. 10 she refers to ‘this level here’ (su questo livello GquFi:) while turning to her right with the pointing gesture maintained and her gaze still oriented to the floor (Fig. 22).

Fig. 23 (0:49)

Fig. 24 (0:51)

Fig. 25 (0:55)

The multimodal resources that Nina deploys refer to an area on the floor that is delimited on the surface by white bricks forming a ‘square’, as she will say at l. 22 (una forma quadrata). However, at this point Nina has not yet mentioned the nature of the object of her attention. During her entire walk she rather produces turn extensions that require a strong indexical anchorage. At l. 16, for instance, Nina initiates a TCU that projects the introduction of a referent, as the existential construction ci sono ‘there are’ entails. However, the mention of the referent is delayed by a series of self-repairs and pauses that occur in the second part of Nina’s turn (əh: (0.3) ’’h (0.3) °l-°; l. 16). In contrast, she will deliver the referent without difficulty at l. 18 (è la rasatura di strutture murarie , ‘it’s the razing of mural structures’). How can we account for this delay in the pronunciation of the referent? First, the choral orientation on the area that the guide is about to describe is not yet achieved during the ‘problematic’ turn of l. 16: as Fig. 24 shows, one of the visitors who initially followed the guide’s spatial description (Barbara) is still looking at the bricks on the floor. Second, the hesitations and self-repairs in Nina’s turn appear to be contingent with the missing choral orientation; the temporal extension of Nina’s turn actually allows the participants to reposition themselves, going from the positionings observable in Fig. 23 to an F-formation as represented by Fig. 25. Third, the figures show that some of the visitors (three men in proximity of the parapet) are momentaneously disattending from the tour: in Fig. 24 – which reflects the setting shortly before Nina produces several hesitations and self-repairs – we can see a white-haired man (Vico) pointing towards the sea and, more importantly, Nina who in all likelihood is seeing his momentanous departure from the guided tour (see also Vico’s talk at ll.

449

Rearranging (in) space

19–21). This departure from the group’s visiting activity can be accounted for in at least two ways. On the one hand, Nina’s initial reference to the panorama (l. 2) visibly introduced an alternative object of interest, while on the other hand Nina’s spatial movement allowed for a consolidation of a conversational – and thus actional – schism (Egbert 1997). The participants reconstitute the group after a sequence of about 30 seconds – in which they talk about some islands that can be seen from where they are – that I have omitted in the transcription. We join the group again as that side sequence (Jefferson 1972) comes to an end: (2b) 9222vgadVP2 (02:57–03:22) fig 47 VICO 48 49 NINA

fig

55 56 57

58

59

60 61

(0.2) ???? m[hm VICO [h’ h’ NINA [’h quindi so (0.8)˙ fig ˙28 NINA ’’h (0.5) e eh: quello che infatti stavo spiegando and uhm what I was actually explaining è proprio che se ˙guardate voi per terFra vedete is exactly that if you look to the floor you can see fig ˙29 che c’è ˙ci sono delle strutture murarie that there is there are mural structures fig ˙30 che sono completamente rasate\˙ (0.2)Fhm ’’h e that have been completely razed and fig ˙31 non sono altro che come dicevo i resti delle torri it’s nothing less than as I said the remainings of quadrate ’h di un antico insediamento\ (…) square towers of an ancient settlement e

51 52 53 54

e

50

˙(0.7) ˙26 (>si vede *>vedi< dove c’è’l da*re prece*denza see where the give way (sign) is -->*pp------------------*,,,,,,,,* (7.1)

The episode starts with what Schegloff (1972: 100) calls a relation to landmark formulation (‘where those two persons are’; l. 1) that is followed – after a lengthy pause – by the action request ‘we will again turn left’; ll. 1–2). The landmark ‘two persons’ appears to be well chosen as the student does not display any kind of problem in understanding the instruction nor in identifying the area where she will have to turn left. ‘Those two persons’ is in fact a highly indexical referent, as it has to be identified with regard to the action taking place hic et nunc inside and outside the car. In addition, the instructor executes a pointing gesture accompanying his landmark description that reaches its apex before the referent is named, as is shown in Fig. 32. Beginning a turn by providing a spatial description strongly projects that there is ‘more to come’, that the instructor is going to say something about the location that he has just described.

Fig. 32

While pronouncing the last item of his spatial description (signori), the instructor is already preparing his next pointing gesture. It will reach its maximum extension during the following 0.6 second pause. As the comparison between the Fig. 33 and 34 illustrates, this second pointing is not just pointing forward but also executing a movement from the right to the left that corresponds to the instructor’s subsequent request to ‘turn left’ (ll. 1–2).

454

Elwys De Stefani

Fig. 33

Fig. 34

In this interactionally unproblematic sequence, we observe that the instructor orients his student’s attention to objects (‘two persons’) or actions (‘turn left’) by sequentially organizing different multimodal resources: he executes a pointing gesture and regularly names the relevant object or action after the apex of the gesture has been reached. This holds true for the pointings just analyzed, but also for the subsequent pointing, occurring at ll. 5–6. As for the turn format chosen by the instructor, it is manifestly adjusted to the locally relevant spatial and temporal contingencies: initiating the turn with a relation to landmark formulation is possible because the landmark is visible at that very moment. During the following 0.6 second pause, the instructor executes a pointing gesture that embodies the upcoming instruction to ‘turn left’. Note that he uses the second person plural (svolteremo, ‘we will turn’; l. 1) to describe the requested action, thereby treating it as an action that the participants will accomplish together. Following the request to turn left, the instructor formulates an instruction related to what is presented as the correct execution of a left turn (uno (0.4) due °e° tre (0.3) freccia (0.2) mi rac-

Rearranging (in) space

455

comando , ‘one two and three indicator don’t forget’; l. 4), that comprises specific sequentially structured gaze shifts (rear-view mirror, outside-mirror, left window, referred to as ‘one two and three’) and the activation of the indicator. In the corpus analyzed here, requests for direction change are in fact recurrently followed by instructions concerning the manoeuvring of the car. Let us now turn to an excerpt in which the manoeuvring instruction and the relation to landmark formulation occupy different positions within the turn and where the identification of the landmark is treated as problematic. Before requesting the student to stop the car (ll. 1–2), the instructor had announced that they were going to exercise pulling in backwards in a parking space: (4) 10153sg2BM1 (06:27–06:43) 1 INS ˙adesso\ (0.4)˙metterai la freccia ˙*verso destra dove c’è now you will set the indicator to the right where ˙35 ˙36 ˙37 ins *....................--> 2 *quella*˙ pianta lì ti fer*mi per pia*cere there’s that plant there you stop please ˙38 ins -->*pppppp*,,,,,,,,,,,,,,,,,,* stu *activates indicator 3 INS m: (0.2) o:cch*°ei° okay ins *…--> 4 (0.4) 5 STU *que˙s*ta pian[ta/* this plant ˙39 ins -->*ppppp*,,,,,,,,,,,* 6 INS [sì\ yes 7 (0.2) 8 INS perfetto\ perfect

The instructor’s turn starts with adesso ‘now’, marking the transition from thee previous action to a subsequent action. By publicly displaying that transition, action markers (see De Stefani 2013) like adesso, okay etc. also introduce the subsequent action as being a new action, one that is different from the previous one. The instructor’s turn is thus indexing some new action right from its very beginning. After a short pause, the instructor formulates a request related to the manoeuvring of the car (metterai la freccia verso destra, ‘you will set the indicator to the right’; l. 1), then introduces a landmark (dove c’è quella pianta lì, ‘where there is that plant there’; ll. 1–2) and eventually asks the

456

Elwys De Stefani

student to stop (ti fermi per piacere , ‘you stop please’, l. 2). We may observe several differences in the turn format with respect to the previously analyzed instruction (ex. 3, ll. 1–2). Let us see how the landmark quella pianta lì (‘that plant there’) is introduced in the turn at hand and then treated in the subsequent turns. The landmark formulation is occurring after a first instruction on the use of the indicator (and not at the very beginning of the turn, as it was the case in ex. 3). This different turn shape is contingent on the changing spatial environment. The following stills give an idea of how the world around the participants alters as the instructor produces the initial part of his turn:

Fig. 35

Fig. 36

The stills reveal that the turn is initiated while the student is in the midst of performing a right turn: see the position of the student’s hands on the steering wheel and the participants’ gazes in Fig. 35 just before the instructor says adesso (l. 1). The subsequent 0.4 second pause displays the instructor’s orientation towards the fact that the student is engaged in a specific manoeuvre

Rearranging (in) space

457

that has not yet come to an end. The instructor uses a verb in future tense and in the first person singular (metterai, ‘you will set’) in the following TCU. Such an understanding of the ongoing activity is in agreement with what can be seen in Fig. 36: the student’s hands are still manoeuvring the steering wheel, while the frontal view shows a row of parked cars. This brief analysis shows that in mobile settings, the construction of turns at hand is finely coordinated with the constantly transforming spatial environment and with the temporal dimension of the interaction. It illustrates once again that participants may design their actions as preparatory actions to some relevant next. Therefore, initiating a turn by saying ‘you will set the indicator’ projects that there is more to come. But the instructor could use different resources to project ‘more’: we have seen e.g. in ex. 3 that he initiates his turn with a landmark description (‘where those two persons are’, l. 1), thereby projecting that he is going to say something about the location identified in that way. The question thus arises why in ex. 3 the instructor starts his turn with a spatial description, while in ex. 4 he asks the student to activate the indicator – and only after that relates to a landmark. As the following still shows, the place formulation ‘where there is that plant there’ (ll. 1–2) is produced only once the student is directing the car straight ahead:

Fig. 37

Fig. 37 reproduces the setting just before the instructor starts his pointing gesture: in the front view one can see the street ahead of the car. Only at this point, having accomplished the task of turning right, the instructor is able to localize a referent that can actually be seen by both participants and to formulate a landmark description. The landmark that the instructor chooses is quella pianta lì, ‘that plant there’ (l. 2). The instructor orients to that referent also by performing a pointing gesture, which reaches its apex in correspondence with the deictic quella.

458

Elwys De Stefani

Fig. 38 However, after activating the indicator (l. 2) and after having received an acknowledgment of that action by the instructor (l. 3), the student seems to problematize the referent with the words questa pianta/ ‘this plant’ (l. 5) pronounced with a rising intonation. At the same time, the instructor himself is orienting towards the possible problematicity of the referent – before the student initiates the repair sequence. Indeed, in l. 3 the instructor starts to execute a pointing gesture which will reach its maximum extension as the student pronounces the deictic questa (l. 5; Fig. 39):

Fig. 39

What is at stake here is the identification – and the identifiability – of referents populating the immediate surroundings. Formulating a space description with reference to ‘that plant there’ when the car is passing beneath a hedge is likely to be treated as problematic. But the orientation towards that problematicity appears to be shared. The above analyses have demonstrated that relation to landmark formulations are recurrent in driving lesson instructions. I have shown two pos-

Rearranging (in) space

459

sible placements within the turn, namely at turn beginning (ex. 3) or at a later turn position (ex. 4). These placements are not haphazard: the instructor has to identify the right moment to produce such a description with regard to the continuously changing spatial and temporal environment. The same observation holds true for the initiation of instructions: the right moment is identified through an analysis of the ongoing activity and through a constant examination of the possibilities of navigation that the surrounding spatial configuration offers. In this sense, instruction sequences and spatial descriptions are not only timely positioned and designed (Mondada 2011), but they also adapt to the altering spatial contingencies.

7.

Conclusions

The analyses advanced in this paper call attention to the centrality of mobility in everyday social interactions, both in settings in which mobility occurs as a joint social action (such as driving-together etc.) and in stationary episodes of interaction, when accommodations of the participants’ body orientations take place. The short-range reorientations observed in the participants’ body arrangements are tightly linked to the actions in which the participants are engaged. Certain actions – e.g. showing the way (ex. 1), pointing to topographic features – require bodily repositionings that allow participants to adequately use the semiotic resources provided by the material environment. From this point of view, the physical space in which an interaction takes place is a ready-for-use resource onto which participants may secrete structure (Goodwin 1995). As the analysis of driving school interactions has proven, turn construction is contingent with the continuously changing spatial environment and with the action being accomplished (also through talk). Phenomena such as pauses, hesitations, but also turn-positions that specific elements occupy (such as place formulations in ex. 3 and 4) are sensitive to the spatial and the temporal progress of the interaction. The participants may use space for their specific interactional needs: as a resource for achieving an interactional task (such as describing the way; ex. 1), as a physical environment that is both structured by the presence of available features (ex. 2) and structurable through interactional and multimodal practices transforming it into meaningful space (ex. 2). The participant’s continuous repositionings have proven to be essential in this respect. Mobility, understood as the faculty of being capable of continuously accommodating the body to the interactional tasks at hand, is thus a fundamental resource for

460

Elwys De Stefani

human interaction. With regard to the second understanding of the term proposed in this paper – mobility as a component of joint social action (e.g. driving-together) – the contingent dimension of movement in space has become visible (ex. 3, 4). These two aspects of mobility – mobility as resource and mobility as contingency – are not mutually exclusive, rather, they are observable to varying degrees in mobile settings of interaction. The above considerations also allow for a more precise appreciation of face-to-face interactions. These have to be regarded as an achievement of the participants, who constitute and sustain the interaction at hand – beyond possible mobile episodes. It is not incongruous to state that face-to-face conversations do not always comprise settings in which the participants are literally facing each other. While Kendon’s (1990) notion of F-formation provided us with a series of body arrangements that participants can adopt (such as side-by-side, L-arrangement etc.), this paper has shown how F-formations are actively (re-)constituted by individuals who become co-participants (see De Stefani and Mondada 2010) when they start to display mutual orientation (which is not necessarily displayed by a reciprocal orientation of their bodies) and thus (re-)initiate an instance of focused interaction (Goffman 1963). In line with these considerations, the notion of interactional space (Mondada 2009) pinpoints the praxeological, temporal and sequential organization of face-toface interactions.

Transcription conventions /\ FG

(1.5) [ ] = xxx ( ) ((cough)) video EXtra domani °opera° >< : par-

rising or falling intonation of the preceding segment rising or falling intonation of the following segment (dotted underline) timed pause in seconds and tenths of seconds beginning and end of overlap latching inaudible segment dubious hearing transcriber’s comments stress high volume middle-high volume low volume fast talk stretching of prior syllable cut-off

Rearranging (in) space

461

^ ’h h’ ˘ cahhsahh

liaison inbreath outbreath glottal closure pronounced laughing

*

delimitate the description of a participant’s gesture/hand movement delimitate the description of a participant’s body movement gesture/movement preparation gesture/movement withdrawal gesture/movement maintained gaze/gesture/movement continues across subsequent lines indicates the segment of talk presented in a screen shot

# .... ,,,,. -----> ˙

References Amerine, Ronald & Jack Bilmes 1988: Following instructions. Human Studies 11: 327–339. Bühler, Karl 1934: Sprachtheorie. Die Darstellungsfunktion der Sprache. Jena: Gustav Fischer. Casey, Edward 1997: The Fate of Place. A Philosophical History. Los Angeles: University of California Press. Clark, Herbert H. 1973: Space, time, semantics, and the child. In: Timothy E. Moore (ed.), Cognitive Development and the Acquisition of Language, 27–63. New York: Academic Press. Clark, Herbert H. & Deanna Wilkes-Gibbs 1986: Referring as a collaborative process. Cognition 22: 1–39. Collinson, Jacquelyn A. 2006: Running-together. Some ethnomethodological considerations. Ethnographic Studies 8: 17–29. De Stefani, Elwys 2010: Reference as an interactively and multimodally accomplished practice. Organizing spatial reorientation in guided tours. In: Massimo Pettorino, Antonella Giannini, Isabella Chiari & Francesca M. Dovetto (eds.), Spoken Communication, 137–170. Newcastle upon Tyne: Cambridge Scholars Publishing. De Stefani, Elwys 2011: ‘Ah petta ecco, io prendo questi che mi piacciono’. Agire come coppia al supermercato. Un approccio conversazionale e multimodale allo studio dei processi decisionali. Roma: Aracne. De Stefani, Elwys 2013: The collaborative organization of next actions in a semiotically rich environment. Shopping as a couple. In: Pentti Haddington, Lorenza Mondada & Maurice Nevile (eds.), Interaction and Mobility. 123–151. Berlin/Boston: de Gruyter. De Stefani, Elwys & Anne-Sylvie Horlacher 2008: Topical and sequential backlinking in a French radio phone-in program. Turn shapes and sequential placements. Pragmatics 18 (3): 381–406.

462

Elwys De Stefani

De Stefani, Elwys & Lorenza Mondada 2007: L’organizzazione multimodale e interazionale dell’orientamento spaziale in movimento. Bulletin Suisse de Linguistique Appliquée 85: 131–159. De Stefani, Elwys & Lorenza Mondada 2010: Die Eröffnung von sozialen Begegnungen im öffentlichen Raum. Die emergente Koordination der räumlichen, visuellen und verbalen Handlungen. In: Lorenza Mondada & Reinhold Schmitt (eds.), Situationseröffnungen. Zur multimodalen Herstellung fokussierter Interaktion, 103– 170. Tübingen: Narr. Egbert, Maria M. 1997: Schisming. The collaborative transformation form a single conversation to multiple conversations. Research on Language and Social Interaction 30: 1–51. Ford, Cecilia E., Barbara A. Fox & Sandra A. Thompson (eds.) 2002: The Language of Turn and Sequence. Oxford: Oxford University Press. Garfinkel, Harold 2002: Ethnomethodology’s Program. New York: Rowman and Littlefield. Gesell, Arnold 1935: Cinemanalysis. A method of behavior study. Journal of Genetic Psychology 47(1): 3–16. Goffman, Erving 1963: Behavior in Public Places. New York: The Free Press. Goffman, Erving 1971: Relations in Public. New York: Harper & Row. Goodwin, Charles 1995: Seeing in depth. Social Studies of Science 25: 237–274. Goodwin, Charles 2000: Action and embodiment within situated human interaction. Journal of Pragmatics 32: 1489–1522. Goodwin, Charles & Alessandro Duranti 1992: Rethinking context. An introduction. In: Alessandro Duranti & Charles Goodwin (eds.), Rethinking Context. Language as an Interactive Phenomenon, 1–42. Cambridge: Cambridge University Press. Haddington, Pentti 2010: Turn-taking for turntaking. Mobility, time and action in the sequential organization of junction negotiations in cars. Research on Language and Social Interaction 43(4): 372–400. Haddington, Pentti & Tiina Keisanen 2009: Location, mobility and the body as resources in selecting a route. Journal of Pragmatics 41(10): 1938–1961. Haddington, Pentti, Lorenza Mondada & Maurice Nevile (eds.) 2013: Interaction and Mobility. Berlin/Boston: de Gruyter. Heath, Christian & Dirk vom Lehn 2004: Configuring reception: looking at exhibits in museums and galleries. Theory Culture and Society 21(6): 43–65. Hindmarsh, Jon, Christian Heath, Dirk vom Lehn & Jason Cleverly 2005: Creating assemblies in public environments. Social interaction, interactive exhibits and CSCW. Computer Supported Cooperative Work 14: 1–41. Hutchins, Edwin 1995: Cognition in the Wild. Cambridge MA: MIT Press. Kendon, Adam 1990: Conducting Interaction. Patterns of Behavior in Focused Encounters. Cambridge: Cambridge University Press. Laurier, Eric, Hayden Lorimer, Barry Brown, Owain Jones, Oskar Juhlin, Allyson Noble, Mark Perry, Daniele Pica, Philippe Sormani, Ignaz Strebel, Laurel Swan, Alex S. Taylor, Laura Wattsh & Alexandra Weilenmann 2008: Driving and ‘passengering’. Notes on the ordinary organization of car travel. Mobilities 3: 1–23. Lefebvre, Henri 1974: La production de l’espace. Paris: Anthropos. Local, John 2004: Getting back to prior talk. and-uh(m) as a back-connecting device in British and American English. In: Elizabeth Couper-Kuhlen & Cecilia E. Ford (eds.), Sound Patterns in Interaction. Cross-linguistic Studies from Conversation, 377–401. Amsterdam/Philadelphia: Benjamins.

Rearranging (in) space

463

McIlvenny, Paul, Mathias Broth & Pentti Haddington (eds.) 2009: Communicating Place, Space and Mobility. Special issue of Journal of Pragmatics 41(10). Mondada, Lorenza 2000: Décrire la ville. La construction des savoirs urbains dans l’interaction et dans le texte. Paris: Anthropos. Mondada, Lorenza 2005: La constitution de l’origo déictique comme travail interactionnel des participants. Une approche praxéologique de la spatialité. Intellectica: 41–42: 75–100. Mondada, Lorenza 2009: Emergent focused interactions in public places. A systematic analysis of the multimodal achievement of a common interactional space. Journal of Pragmatics 41: 1977–1997. Mondada, Lorenza 2011: The situated organisation of directives in French: Imperatives and action coordination in video games. Nottingham French Studies 50(2): 19–50. Müller, Gabriele M. 2008: La construction pseudo-clivée dans l’organisation d’activités complexes: Questions de portée. Cahiers de praxématique 48: 213–238. Psathas, George 1976: Mobility, orientation and navigation. Conceptual and theoretical considerations. New Outlook for the Blind 70(9): 385–391. Psathas, George 1986: Some sequential structures in direction-giving. Human Studies 9: 231–246. Psathas, George 1991: The structure of direction-giving in interaction. In Deirdre Boden & Don H. Zimmerman (eds.), Talk and Social Structure. Studies in Ethnomethodology and Conversation Analysis, 195–216. Berkeley/Los Angeles: University of California Press. Relieu, Marc 1999: Parler en marchant. Pour une écologie dynamique des échanges de paroles. Langage et Société 89: 37–67. Ryave, A. Lincoln & James N. Schenkein 1974: Notes on the art of walking. In: Roy Turner (ed.), Ethnomethodology, 265–278. Harmondsworth: Penguin. Sacks, Harvey, Emanuel A. Schegloff & Gail Jefferson 1974: A simplest systematics for the organization of turn-taking for conversation. Language 50: 696–735. Scheflen, Albert E. 1964: The significance of posture in communication systems. Psychiatry 27: 316–331. Schegloff, Emanuel A. 1972: Notes on a conversational practice. Formulating place. In: David Sudnow (ed.), Studies in Social Interaction, 75–119. New York: The Free Press. Schegloff, Emanuel A. 1987: Micro and macro. Contexts and other connections. In: Jeffrey C. Alexander, Bernhard Giesen, Richard Münch & Neil J. Smelser (eds.), The Micro-macro Link, 207–234. Berkeley: University of California Press. Tuan, Yi-Fu 1977: Space and Place. London: Edward Arnold Ltd. Urry, John 2007: Mobilities. Cambridge: Polity Press.

464

Lorenza Mondada

Lorenza Mondada

Commentary: Being mobile, talking on the move Conceptual, analytical and methodological challenges of mobility

1.

Introduction

There has been a mobility turn in the social sciences (Urry 2007, Cresswell 2006 – see the editorial to the new journal Mobilities, created in 2006: Hannam, Sheller, Urry 2006), which has highlighted mobility as one important feature of our contemporary society (as well as of older and other cultures, Clifford 1992). Mobility studies deal with mobility at different scales, their connections and interdependencies: mobilities as geographical and global movements of people (workers, asylum-seekers, scientists, elite experts, terrorists, football champions, etc.), with respect to such phenomena as migration, tourism and international expertise, as well as movement of goods, capitals, information, ideas – not forgetting that the hyper-mobility of some contrasts with the spatial segregation of others. Within linguistics and the social sciences interested in situated social practices and interaction, this mobility turn is raising interesting discussions too – as shown by the special issue of the Journal of Pragmatics on “Communicating place, space and mobility” (edited by McIlvenny, Broth & Haddington 2009) and the edited volume Interaction and Mobility: Language and the Body in Motion (edited by Haddington, Mondada & Nevile 2013). In this framework, the emphasis is on mobile practices, in the detailed analysis of the “methods” (Garfinkel 1967) with which participants engage in situated mobile actions and in talk while moving. Within this field – which is also the focus of this section – the interest in mobility invites to revisit older topics – such as for example deixis and place formulation – and to discover new ones – such as mobile practices –, as well as to invent new methodologies – with a strong focus on video recordings. Looking at talk and social action from a mobile perspective encourages to revisit some key concepts, revealing the strong bias a tacit static view had for a long time on our conceptions of language, action and context.

Being mobile, talking on the move

2.

465

Revisiting old topics, discovering new phenomena

The focus on mobility reveals that most often linguistic and interactional descriptions have favored the observation of static situations as a default case, and hat this has consequences on the way in which phenomena and concepts have been defined and taken for granted. For the study of language and space, the focus on mobility points to new complexities. Deixis has been often described presupposing a static origo (Bühler 1934) and treating displaced origos as an exception. The reference point for deixis, the body of the speaker and the environing space, has been first of all considered as static, building complex systems around the opposition between here and there. In similar ways, spatial reference and place descriptions have been mostly studied as they are produced by a static speaker, for example when locating and referring to a distant place. De Stefani’s and Haddington’s chapters in this section contain several examples of descriptions produced by participants in the move, and of the careful and variegated choice of place formulations they mobilize in these settings. The choice of deictic forms evolves adjusting to a changing context, in which reference points, landmarks, speaker’s location are constantly moving. Mobility affects also, and more generally, the way in which we consider the organization of turns at talk: in mobile situations, talk is not formatted as in static ones, but adjusts to these contextual transformations, with hitches, delays, and hesitations co-occurring while obstacles encountered in space are either transformed, removed or bypassed. The emergent construction of turns-at-talk is sensitive to the details of walking movements in which speakers are engaged and to the spatial ecology they encounter (Relieu 1999, Haddington and Keisanen 2009: 1949, Mondada 2009). Thus, emergent grammar (Hopper 1987) and incremental syntax (Auer, 2007) are specifically shaped in mobile situations – which in turn highlight the malleability of syntax and the indexical character of grammar. Another consequence of this focus on talk in mobile situations is the redefinition of the very notion of context and ecology of talk. The fact that context does not pre-exist as such to talk has been largely discussed within ethnomethodology and conversation analysis: the “reflexivity” of talk points to the fact that it is both context-shaped and context-renewing (Heritage 1984), thus having a transformative effect on context. In mobile situations, this evolutivity of context appears even stronger: mobility invites to look at the way in which context is actually actively achieved by the participants walking, positioning, locating their bodies within in. This leads to an interest in space not so much as a referent to be described but as an emergent con-

466

Lorenza Mondada

figuration, in which the action of the participants and the materiality of the environment are mutually adjusted and transformed (Hausendorf, Mondada, and Schmitt 2012). This has prompted to describe how the interactional space is actively constituted (Mondada 2009, this volume), on the basis of the pioneering work of Ashcraft and Scheflen (1976) on how people meet and form a circle or of Kendon (1977) on the constitution of a face-to-face formation. For example, in the opening of an encounter, a new context of interaction is produced by participants going through a series of mobile steps, such as sighting, walking in a convergent way, approaching, and finally, converging to, hand shaking, hugging, kissing and touching the other’s body (Kendon 1977). Consequently, the very notion of face to face position of the participants is revisited as only one among other bodily configurations and, most importantly, as being the result of positioning and repositioning movements (see De Stefani, this volume). New empirical phenomena for the analysis have emerged from this perspective: mobile practices as walking as a couple, shopping together, approaching an intersection and deciding where to go, searching for a route while driving, slowing down in front of an obstacle, being lost et finding the way, reading maps, searching for a relevant landmark … (see Ryave and Schenkein 1974; Psathas 1976; Laurier et al. 2008; De Stefani 2011 to give just a few older and more recent examples). These mobile practices are often involved in solving “mobility puzzles” such as why-that-here, where-are-wenow, what-is-here-now, what-is-visible-now (Haddington, Mondada, and Nevile 2012).

3.

Methodological challenges

The mobility turn attracts the attention to the fact that participants are not predominantly sitting in front of a table, having a conversation, a meeting or a dinner. The very fact that these situations have been favored for many studies of language and talk relates both to theoretical presuppositions and to methodological choices. Looking at speakers and co-participants while they walk, drive, dance, move around implies a range of methodological challenges for the detailed documentation of their talk and their bodily actions, within evolving contexts. Video recordings seem to be the best methodological solution for capturing them: the focus on mobile contexts and actions invites to definitively abandon a view restricted to the verbal action (audio recordings) but also a view restricted to the upper part of the body (video recordings of talking heads and gesticulating hands). Mobility practices invite to produce video re-

Being mobile, talking on the move

467

cordings of the entire body, considering not only the head (gaze, facial expressions) and hands (gesture), but also the torso (body-torque, Schegloff 1998), body positions (Goodwin 2000), body movements as a whole (Broth and Mondada, forth.). This implies also that new ways of video recording are needed, using multiple cameras, both on the mobile participants and on the evolving environment, mobile cameras following walking “vehicular units” (Goffman 1971) (Relieu 1999; Psathas 1992; Mondada 2009), cameras mounted on vehicles, like cars, bikes, or airplaines (Laurier 2004; Nevile 2004) – as exemplified in Haddington’s and De Stefani’s data in this section and Mondada’s data in this volume. These new forms of data also offer significant challenges for multimodal transcription – given the difficulty of annotating the trajectories and details of movements – often complemented by other forms of visualization, such as images, maps, drawings, and sketches.

4.

Conceptual consequences: Language, action, multimodality and ecology

As sketched in this brief introduction, the focus on mobile actions and talk in the move introduces the possibility of seeing old phenomena in a new light and also of discovering new phenomena for description. This involves also new data and methodological challenges (see also Büscher, Urry, and Witcher 2011). More generally, one can see mobility as literally “mobilizing” the conceptions of language, talk, action and context. The focus on mobility invites to dynamize, temporalize, flexibilize our definitions of what action, linguistic and bodily resources, context and space are. Mobility involves dynamic action, and therefore strengthens the temporal dimension of talk and action: if contemporary linguistics acknowledges that language and grammar cannot be described in a vacuum but in a situated way, the focus on mobility invites to take into account the contingencies characterizing talk in mobile settings. So, for example, time of the walk and time of the turn-at-talk intersect and coordinate in many ways that strengthen the possibility of documenting the malleable character of syntax and other grammatical resources. This concerns language but also other multimodal resources: mobility settings involve the entire body and invite to take into account the way in which different embodied conducts contribute to action formatting as well as turn formatting. Mobile practice also invites to see the ecology of action, the material environment, the space of interaction in a new light, strengthening their praxeological dimensions: space and the environment, as well as material arti-

468

Lorenza Mondada

facts and technologies, are not just pre-existing and determining language and action, but they are actively configured by the very action and talk. This reflexive dimension complexifies our vision of how space is both constrained by language, talk, and action but also shaped by them – through the emergent and variable choice of place formulations, through the positionings and repositionings of the speaker’s and co-participants’ bodies, through the mobilization of artifacts, objects, documents during their action. In this sense, the focus on mobility entails a conceptual “mobilization” of the way in which we think about talk, action, linguistic and embodied resources, as well as context, ecology and space.

References Auer, Peter 2007: Syntax als Prozess. In: H. Hausendorf (ed.), Gespräch als Prozess. Linguistische Aspekte der Zeitlichkeit verbaler Interaktion, 95–124. Tübingen: Narr. Broth, Mathias & Lorenza Mondada forth.: Moving away. The systematic organization of movement in sequence closings. Bühler, Karl 1934: Sprachtheorie: Die Darstellungsfunktion der Sprache. Stuttgart & New York: Fischer. Büscher, Monika, John Urry & Katian Witchger 2011: Introduction: mobile methods. In: Monika Büscher, John Urry & Katian Witchger (eds.), Mobile Methods, 1–19. London/New York: Routledge. Clifford, James 1992: Traveling cultures. In: L. Grossberg, C. Nelson & P. Treichler (eds.), Cultural Studies. London: Routledge. Cresswell, Tim 2006: On the Move: Mobility in the Modern Western World. New York: Routledge. De Stefani, Elwys 2011: ‘Ah petta ecco, io prendo questi che mi piacciono’. Agire come coppia al supermercato. Un approccio conversazionale e multimodale allo studio dei processi decisionali. [“Oh wait here I’ll take these ones which I like”. To act as a couple in the supermarket. A conversational and multimodal approach to the study of decision processes]. Roma: Aracne. Garfinkel, Harold 1967: Studies in Ethnomethodology. Cambridge: Polity Press. Goffman, Erving 1971: Relations in Public: Microstudies of the Public Order. New Brunswich, NJ: Transaction Publishers. Goodwin, Charles 2000: Action and embodiment within situated human interaction. Journal of Pragmatics 32(10): 1489–1522. Haddington, Pentti, Lorenza Mondada & Maurice Nevile (eds.) 2013: Interaction and Mobility: Language and the Body in Motion. Berlin/Boston: de Gruyter. Haddington, Pentti & Tiina Keisanen 2009: Location, mobility and the body as resources in selecting a route. Journal of Pragmatics 41(10): 1938–1961. Hannam, Kevin, Mimi Sheller & John Urry 2006: Editorial: Mobilities, immobilities, and moorings. Mobilities, 1–1, 1–22. Hausendorf, Heiko, Lorenza Mondada & Reinhold Schmitt (eds.) 2012: Raum als interaktive Ressource. Tübingen: Narr. Hopper, Paul 1987: Emergent grammar. Berkeley Linguistics Society 13: 139–157.

Being mobile, talking on the move

469

Kendon, Adam 1977: Spatial organization in social encounters: The F-formation system. In: Adam Kendon (ed.), Studies in the Behavior of Social Interaction, 179–208. Lisse, Holland: Peter DeRidder Press. Laurier, Eric 2004: Doing office work on the motorway. Theory, Culture & Society 21(4/5): 261–277. Laurier, Eric, Hayden Lorimer, Barry Brown, Owain Jones, Oskar Juhlin, Allyson Noble, Mark Perry, Daniele Pica, Philippe Sormani, Ignaz Strebel, Laurel Swan, Alex S. Taylor, Laura Watts & Alexandra Weilenmann 2008: Driving and ‘passengering’: Notes on the ordinary organization of car travel. Mobilities 3(1): 1–31. McIlvenny, Paul, Mathias Broth & Pentti Haddington 2009: Communicating place, space and mobility. Journal of Pragmatics 41: 1879–1886. Mondada, Lorenza 2009: Emergent focused interactions in public places: a systematic analysis of the multimodal achievement of a common interactional space. Journal of Pragmatics 41: 1977–1997. Nevile, Maurice 2004: Beyond the Black Box: Talk-in-interaction in the Airline Cockpit. Aldershot: Ashgate. Psathas, George 1976: Mobility, orientation and navigation: Conceptual and theoretical considerations. New Outlook for the Blind 70(9): 385–391. Psathas, George 1992: The study of extended sequences: The case of the garden lesson. In: Graham Watson & Robert M. Seiler (eds.), Text in Context: Contributions to Ethnomethodology, 99–122. Newbury Park: Sage. Relieu, Marc 1999: Parler en marchant: Pour une écologie dynamique des échanges de paroles [Talking while walking: towards a dynamic ecology of verbal exchanges]. Langage & Société 89: 37–67. Ryave, A. Lincoln & James N. Schenkein 1974: Notes on the art of walking. In: Roy Turner (ed.), Ethnomethodology, 265–274. Harmondsworth: Penguin. Scheflen, Albert E. 1964: The significance of posture in communication systems. Psychiatry 27: 316–331. Schegloff, Emanuel A. 1998: Body torque. Social Research, 65(3): 535–586. Urry, John 2007: Mobilities. Cambridge: Polity.

470

Lorenza Mondada

Section 5: Mediated spaces

Language, media, and digital landscapes

473

Marco Jacquemet

Language, media, and digital landscapes 1.

Mediation

The study of the link between language and ‘the media’ must start with the understanding that all social life is constituted through processes of mediation. In any context, the social organization of everyday life requires an intervening medium for coordination and communication among social actors. This medium is a perceptual framework both enabling and constraining a given set of social practices; it is a reflexive and reifying technology, making society imaginable and intelligible to itself in the form of external representations (Williams 1977; Peters 1999). In this light, there are two definitions of the ‘media’: a broader one, which considers ‘media’ all intervening entities responsible for communication (responsible in particular for operating a separation between phenomenological experience and the referential world) such as speech or ritual performances; and a narrower one, which takes into account particular technologies (from writing to electronic communication), each of which handles the separation between form and content in specific ways. For instance, a signal of distress could be expressed on the telegraph as a continuous Morse code sequence of three-dits/three-dahs/three-dits, Mayday could be used on radio, and SOS could be the visual equivalent of the distress signal – spelled out in individual letters, for example, stamped in a snowbank or formed out of logs on a beach (incidentally, the fact that SOS can be read right side up as well as upside down becomes important for visual recognition if viewed from above). Many mediations came to be perceived as “natural”, “immediate” experience. We don’t usually stop to think that even the spoken word, as Mc Luhan reminded us, was “the first technology by which man [sic] was able to let go of his environment in order to grasp it in a new way” (1994: 57). All communication hides its mediation and presents its product as pristine, whole, and original. In most cases, it is only through the introduction of a new technology that old ones are revealed as mediating processes. For instance speech came to be perceived as a person-to-person activity only with the introduction of telecommunications (especially the telephone), in other words, it became thinkable only “in the shadow of telemediated communication” (Peters 1999: 6).

474

Marco Jacquemet

Dismissive of this process of mediation, most scholarship tends to see a strong separation between language and culture on one side and the media on the other, rather than as co-constitutive. But language and culture are configured as the underlying premises for communication, as always already preexisting the intervention of the media. This perspective does not take into account the fact that language and culture are also effects of social processes of mediation and as such grapple with the internal indeterminacies and external pressures inherent in all processes of mediation. If scholars have been for the most part blind to the mediations through which even culture and language are constituted, it is due to the fact that these mediations have been buried in processes of naturalization. As Mazzarella pointed out: “People may study a ‘traditional’ ritual as performed for TV cameras. But the ritual itself is already a medium, with its own distinctive mechanisms and possibilities of translation and objectification” (2004: 353). Moreover, at times the causal link between culture and media is reversed and the media can be considered the “extension of man” (Mc Luhan 1994) in their power to shape and perhaps even constitute not just experience but also socio-cultural life and its communicative practices. From the role of print capitalism in defining the boundaries of an imagined national community (Anderson 1991), to the capacity of mass-mediated communication to foster linguistic change by spreading unified nationalized standard languages (Gellner 1983), the media have engaged with language and culture a process of mutual shaping that remains largely below the level of awareness. Thus, an investigation into the linkage between language and the media should not focus on the way language (and culture) determine media practices, but on the study of the intersection of two or more systems of mediation. This is particularly salient if we consider the relationship between media technologies and communicative practices. All media technologies have distinct and discernible formal properties that condition how messages are produced, circulated, and received. Because of their hardware, combined with the extent to which content is modified or recreated at different nodes in the social field of circulation, different technologies of communication will have different communicative structures, and in turn these communicative structures will influence the way the technologies are conceived and understood. Moreover, all these technologies are further conditioned by the structures of reception that a particular medium requires and allows. Media technologies invite us to respond to them in particular ways, but do not determine our response. They carry certain scripts that shape our interaction with them (Shove 2003), but the production of meaning still needs to emerge in the context of material encounters between actors and media. The

Language, media, and digital landscapes

475

media offer, in other words, certain “affordances” through which we relate to them (Hutchby 2001), causing communication structures to be shaped in the interaction between the media’s formal properties and actors’ intentionality and communicative skills (for instance the particular participation framework and sequential structure of a TV interview, Schegloff 1988).

2.

Electronic media and digital landscapes

In this light, global communication, and in particular electronic media, is much more than an enabler of people’s interactivity and mobility: it alters the very nature of this interactivity, transforming people’s sense of place, belonging, and social relations. We are now witnessing the emergence of a telemediated cultural field, occupying a space in the everyday flow of experience that is distinct yet integrated with face-to-face interactions of physical proximity. This is the result of an integration of communication technologies into a late modernity able to transform human experience in all its dimensions: from its social field (now globalized and deterritorialized) to a semiocapitalist marketplace (with its shifting methods of production, delivery and consumption) to the production not only of new conveniences and excitements but also new anxieties and pathologies (Tomlinson 2007). In this context, few of the experiences of present-day telemediation have any counterparts beyond the last few decades. Furthermore, electronic media provide people with distinct ways of expressing themselves – “media idioms” in other words. Not only electronic mass mediation creates new genres of entextualization (Eisenlohr 2004), but the sampling and recontextualizing of media content, proper of digital technologies, are becoming basic communicative practices in media cultures characterized by linguistic bricolage and by the tendency to incorporate conversational speech styles into public discourse, giving rise to “dense interpenetration of local performances with style of speech that are reflexively designed, produced, and disseminated though mass mediated institutional and/or electronic communication systems” (Rampton 1999: 423). Language studies have been slow to respond to the challenge of examining media idioms. The study of electronic media has been left to the sociocultural anthropology of media (the product of a serious rethinking of the field of visual anthropology; see Ginsburg et al. 2002) and to the more ‘ethnographic’ approaches to cultural studies (which emerged out of the encounter between sociological analysis of the media and textual analysis, see Gitlin 1983; Baudrillard 1984; Mankekar 1999). However, Spitulnik’s critique (1993) still stands: these studies suffer from inattention to linguistic details

476

Marco Jacquemet

and an overdependence on textual criticism. The few examples of linguistically oriented media analysis (such as Gillespie 1995) which focuses principally on the lexical components of media messages, has neglected to analyze the indexical/pragmatic aspects of media communication and the global spread of media idioms. In particular, we need an analysis of the communicative practices used in the two most significant digital landscapes of late modern communication: telepresence and digital social networks. 2.1

Telepresence

Telepresence refers to a set of digital technologies which allow people to feel as if they were present, give the appearance of being present, or have an effect at a place other than their physical location. In the late 1980s, Paul Virilio translated into philosophical terms Marvin Minski’s idea that we can build a cybernetic machine able to give remote participants the feeling of actually being present to an event. Virilio introduced the concept of telepresence to account for the spacializing role images would have in long-distance video communication (1988; 1990). Interactive, live transmission of video images over great distances becomes in itself a new kind of place. The experience of being in this place is what Virilio called telepresence, which supersedes in real time the real space of objects and sites. Telepresence is the experience of the continuity of real time overcoming the contiguity of real space. The impact of fiber optics, monitors, and video cameras on our vision and on our surroundings is far greater than that of electricity in the nineteenth century. “In order to see,” Virilio observes, “we will no longer be satisfied in dissipating the night, the darkness of the outside world. We will also dissipate time lapses and distances, the outside world itself ” (1990: 72). In consonance with Baudrillard’s understanding of the new informational landscape, Virilio advanced the notion that we no longer inhabit or share a physical public space. Our domain of existence or socialization is now the public image, with its volatile, functional, and spectacular ubiquity. For Virilio, one of the most important aspects of the technologies of digital imaging and of synthetic vision made possible by optoelectronics was the “fusion/ confusion of the factual (or operational) and the virtual,” the predominance of the “effect of the real” (1988: 128) over the physical world. In this light, telepresence is a distinctive existential mode of existing alongside, and in equal terms with, direct embodied relations of presence. Viable forms of interpersonal relations can be established and maintained through the act of establishing and maintaining a digital persona.

Language, media, and digital landscapes

477

The concept of telepresence has been picked up by the computer industry to describe the latest innovation in videoconferencing. In this virtual environment, users scattered around the planet communicate live via integrated voice, video, and data networks, at a speed and quality of transmission paralleling face-to-face interactions. High-definition, multi-panel video screens reproduce the actual size of conference participants. This technology creates the illusion that the two (or more) parties to a conversation are not continents apart but at opposite sides of the same table. Initial reports suggest that participants in such meetings very quickly forget, or at least stop caring, that they are not in the same room. In order to achieve this telepresence effect, rooms in remote locations are equipped with matching conference tables that face wide screens, as well as matching décor, so that the local table blends seamlessly with the remote table projected on the screen. The rooms are equipped with multiple cameras and speakers, supported by enormous computing power, so that meeting participants can make eye contact and overlap or interrupt each other (in order to do so, the delays in audio and video transmission must be negligible, i.e., below 250 milliseconds, the threshold at which the human brain starts to notice the delay, see Cisco 2011). While meetings that make use of telepresence video conference technologies may be the wave of the future, many virtual meetings that take place today combine use of teleconferencing with the Internet to facilitate collaboration on shared projects, but do not simulate face-to-face interactions. Participants are still present to each other, nonetheless, as voices on the phone – and thus experience a form of telepresence. The only ethnography of communication to date that has explored telepresence focused on this sort of teleconferencing. Wasson (2006), in her study of the meetings of virtual teams in a corporate workplace, analyzed how participants in such teleconferences multitasked while meeting in more complex ways than they would during face-to-face interactions. As in face-to-face interactions, a central situational focus was maintained within the interactional space of the virtual meeting, but participants were also simultaneously located in their local spaces. In their local spaces, they were often simultaneously engaged in other activities unconnected to the meeting. These multitasking skills seem to be a common theme of all digitalized environments, pointing at the complex communicative practices occurring simultaneously in the transnational space created by local and virtual environments. While full telepresence remains in 2012 an experimental technology with fewer applications, we can look at digital social networkings, with their combination of text, images, and real-time connectivity as the current frontier for the development of fully inhabited digital landscapes.

478 2.2

Marco Jacquemet

Digital social networkings

Social networking sites are becoming one of the primary communication channels for many young adults, and the numbers are staggering: in summer 2011, Facebook, the jaggernaut of social networks, counted more than 750 million active users (50 % of whom logged on daily, spending over 700 billion minutes per month on the site), was translated in 70 languages (300,000 users helped translate the site through the translations application), and established an entire economy of content production driven by consumers (the average Facebook member produced 90 pieces of content, giving rise to the figure of the produser, Bruns 2008). In these environments, people, mostly young, learn to expand their cultural and communicative capacities by constructing online subjectivities in an open-ended process of becoming. These virtual communities are sites for the expansion of the cultural, communicative, and subjective capacities of their users, who are engaged in an exponential expansion of discrete nodes of both affect and affinity. They do so by allowing users to communicate with people worldwide through the creation of their personal space, where users can personalize their profile pages as they wish with texts, images, music, varying layouts, and various other items (such as links to other websites). According to Gamble (2007), the communicative world of the Internet and other digital domains is doing much more than just developing in a manner that replicated communities of practice based on speech patterns: the communicative practices played out in these virtual communities are advancing into the domain of speech in the physical world. It is now possible to determine which speaker has an active digital life merely by exploring his/ her vocabulary and sentence construction. In other words, the digital domain’s contribution to the evolution of language does much more than just blending the virtual world with the physical one (as suggested by Dubé et al, 2006): digital communication is now a home base for the development of language for a broad section of the general population. Online gaming and its capability to produce real time virtual communities deserves a closer look. Every day millions of pc gamers log in to online worlds to play and interact within a virtual community. Massively multiplayer online role playing games (MMORPGs) – such as Second Life, Ultima Online, World of Warcraft, or Everquest – allow people to don online personalities, or avatars, and duke it out in make-believe environments. These environments typically appear similar to those in the real world, with similar rules for gravity, topography, locomotion, and communication. Communication has, until recently, been in the form of text, but recently real-time voice communi-

Language, media, and digital landscapes

479

cation via the internet has become available. For instance, in the simulation game Second Life, not only can entrepreneurs buy and sell digital real estate, create their own lines of avatar clothing and accessories, and hold virtual concerts, lectures, and sporting events; but in addition many users have come to see these games as an enhanced communications medium for staying in touch with friends and to make new connections. The creation of elective communities spread across continents and able to interact in real time in the virtual confines of cyberspace is pushing the limits of what constitutes social interaction and group communication. While by now the study of computer-mediated communication (CMC) is a well established field (Herring 1996, 2003; Crystal 2006), the indexical-pragmatic functions of CMC in transnational spaces have been largely overlooked. Moreover, existing research on CMC has focused almost exclusively on emergent practices in English whereas, as Danet and Herring (2007) pointed out, roughly two-thirds of Internet users are non-English speakers (one notable exception is the study of code-switching among expatriate South Asians on Usenet in Paolillo 1996, see also the edited volume by Danet and Herring 2007). What are the communicative practices taking place in transnational, digitally-mediated environments, from virtual meetings to social networking sites? Ethnographically-based, linguistically-oriented fieldwork of digital transnational spaces is still quite sketchy and lacks in-depth analysis. As a result we are left to speculate on the complex communicative dynamics people activate when engaging in digital interaction. As Jones (2004) pointed out in his study of the shift from face-to-face to virtual interaction “traditional sociolinguistic conceptualizations of the terms of interaction and the contexts in which it takes place may need to be radically rethought in light of new communication technologies” (Jones 2004: 21). This awareness of a transformed context must inform our investigation of these spaces. The lenses we usually adopt in looking at language must be significantly altered to accommodate for communicative phenomena produced by people present in transnational contexts, whose talk is mediated by deterritorialized technologies, and who interact with both present and distant people. We need to study the communicative practices of these subjects, even if these communicative practices cannot be recognized as part of a single linguistic standard. To elucidate a different approach to the study of language in transnational space, the next section will introduce the concept of transidioma to study precisely this intersection between mobile people and digital communication.

480

3.

Marco Jacquemet

Transidioma and digital networks

One of the most significant breakthroughs in language studies in the late 20th century was the introduction of the notion of communicative practice. Under the influence of European political philosophers such as Foucault and Bourdieu, language and communication scholars adopted the notion of practice to deal not only with communicative codes and ways of speaking (some of the rallying concepts of the first wave of the ethnography of communication), but also with semiotic understanding, power asymmetry, and linguistic ideology. By focusing on the “socially defined relation between agents and the field that ‘produces’ speech forms” (Hanks 1996: 230), a practice-oriented approach can then explore speakers’ orientations, their habitual patterns and schematic understandings, and their indexical strategies. Hanks defined communicative practice as constituted by the triangulation of linguistic activity, the related semiotic code or linguistic forms, and the ideology of social indexicality. He invoked a poetic image of practice as “the point of conversion of the quick of activity, the reflexive gaze of value, and the law of the system” (1996: 11). This triangulation of linguistic activity, semiotic codes, and indexicality needs to be complexified to account for how groups of people that are no longer territorially defined, think about themselves, communicate using an array of both face-to-face and long-distance media, and in so doing produce and reproduce social hierarchies and power asymmetries. I propose to use the term transidioma to describe the communicative practices of deterritorialized groups that interact using different languages and communicative codes simultaneously present in a range of communicative channels, both local and distant. The transidioma is the result of the co-presence of digital media and multilingual talk exercised by deterritorialized/reterritorialized speakers. It operates in contexts heavily structured by social indexicalities and semiotic codes that produced relatively stable power asymmetries and cultural hegemonies. Anyone present in transnational environments, whose talk is produced by both biological and digital means, and who interacts with both present and distant people is engaged in transidiomatic practices. Given the nature of economic globalization, many contemporary work environments – from the offices of international organizations to airport lounges, from international call centers to the board meetings of multinational companies – can be classified as transidiomatic. In addition, a great number of social settings – from living rooms to hospital operating rooms to political meetings – experience a translocal multilingualism interacting with

Language, media, and digital landscapes

481

the electronic technologies of contemporary communication. The world is now full of settings where speakers use a mixture of languages in interacting with friends and co-workers; read English and other ‘global’ languages on their computer screens; watch local, regional, or global broadcasts; and listen to pop music in various languages. Much of the time, they do so simultaneously. Moreover, the transidioma is no longer restricted to areas of colonial and post-colonial contact, but flow through the multiple channels of electronic communication and global transportation, over the entire world, from contact zones, borderlands, and diasporic nets of relationships to the most remote and self-contained areas of the globe. These communicative resources are activated by people needing to operate in multiple, co-present, and overlapping communicative frames. The language they use to communicate depends on the contextual nature of their multi-site interactions, but is necessarily mixed, translated, “creolized” (Hannerz 1992, 1996). Transidiomatic practices usually produce linguistic innovations with heavy borrowing from English (a reminder of the global impact of contemporary English – see Pennycook 2003, 2007; Crystal 1997; De Swaan 2001), but any number of other languages could be involved in these communicative recombinations, depending on specific processes of reterritorialization in which the speakers are engaged. Through transidiomatic practices, diasporic and local groups alike recombine their identities by maintaining simultaneous presence in a multiplicity of sites and by participating in elective networks spread over transnational territories. These recombinant identities are based on multi-presence, multilingualism, and decentered political/social engagements. While individual creativity must be acknowledged in this process of recombination, I do not want to idealize the agency (both social and communicative) of most people involved in transidiomatic practices. These practices are still inserted into a global indexical order which assign superior values to certain systems of communication at the expenses of others. The fact that most aspiring musicians felt until recently compelled to have a presence on Myspace, that managers will soon have to endure interminable videoteleconferencing with long-distant bosses, or that the lack of knowledge on how to compose a proper email message could restrict people’s work opportunities (Shipley and Schwalbe 2007), are all indications that communicative inequalities will continue to shape power relations in the age of globalization. On the linguistic level, we will still encounter a stratified, layered ideology of what to consider a legitimate communicative code; and this legitimacy will be in the hands of the mobile sovereignty discussed above, which is already in-

482

Marco Jacquemet

creasingly preoccupied in assessing what kind of transidiomatic practices are welcomed (such as those of Indian phone operators forced to speak the local English of the area they serve) and what are considered ‘broken English’ or gibberish. Transidiomatic practices seem to be especially at home in the digital environments of diasporic social networks. On the internet, what Karim (2003) called the “media of diaspora”, we find dense conversational code-switchings and mixing across or within turns that are qualitatively different from traditional forms of bilingual written discourse (Androupsopoulos 2008). In the digital communication of deterritorialized speakers multilingualism and linguistic mixings are de rigeur, and their interactions are a combination of membership in both online and offline communities. This is clearly visible in the communicative practices emerging from Facebook’s diasporic communities, and in particular in the two predominant activities activated during Facebook’s interactions to cope with the intrinsic transidiomaticity of these communities: the resort to referential signs (in particular photographs and personal names) and to mechanical translation. 3.1

Referential practices

Facebook’s ability to refer to the offline world by bringing it inside its online domain is proving to be one of its most enduring qualities. Its great impact on the offline world derives from its capacity to offer communicative tools, such as the “wall”, which allows people to link their offline activities to their digital presence. The wall is a space on each user’s profile page that allows friends to post messages for the user to see. One user’s wall is visible to anyone with the ability to see their full profile, and different users’ wall posts show up in an individual’s News Feed. Many users use their friends’ walls for leaving short, quick notes, whereas more private communication is produced in a different environment, called “messages”, where notes are visible only to the sender and recipient(s), much like email. In July 2007, Facebook allowed users to post attachments to the wall, whereas previously the wall was limited to textual content only, thus increasing the multimodality and digital complexity of the communication. Among the most popular attachments available to Facebook members, posting of photographs is undoubtly the most widely used. By overlaying digital data over past face-to-face interactions, online photo posting contributes to the phenomenom of augmented reality, the lifeworld where material and digital items intermigle to create multiple layers of phenomenal experience. Posting photographs and their subsequents comments revisit the ex-

Language, media, and digital landscapes

483

perience captured digitally and extend its event-like presence from the offline to the online world. In addition, in the case of multilingual networks, photo posting allows speakers of various language to be co-present without necessarily having to share one single language. The case below, will illustrate this point. One day after a garden party among a multinational group of friends in California, one of the participants posted a photo of the event, depicting the group of friends sitting or standing around a table. The behavior of one of the standing person (Paolo), who appears to be touching the stomach of a seated woman (Francisca), attracted the attention of another participant of the party, who commented in Tagalog (translation in parenthesis). Case 1

From the album: Added yesterday Share Make Profile PictureTag This PhotoReport This Photo

484

Marco Jacquemet

JQ (Stanford) wrote at 8:39am Anong kinakapa ni Paolo? (What’s Paolo groping?)

JP (San Francisco, CA) wrote at 9:03am kung kumain ng roast si Panchia? she’s vegetarian. (If Panchia [Francisca] has eaten any roast? She is vegetarian.)

JQ (Stanford) wrote at 9:07am Ohhhhhh …

Francisca wrote at 12:32pm it sucks when people comment on your pics and you can’t understand!!! JQ (Stanford) wrote at 12:38pm Oh sorry, Panchia. I was just curious why Paolo’s hand was there. ;-) And Jill responded that Paolo was probably checking to see if you had the porchetta, explaining to me you are vegetarian. Cheers! JQ

Francisca wrote at 12:41pm hahaha … that’s funny, he was actually saying that I’m fat, and was trying to prove it … JP (San Francisco, CA) wrote at 12:47pm ugh! typical! Hahaha The first two turns in Tagalog attracted the attention of the person being talked about (Francisca, referred here by her nickname Panchia). With a metapragmatic move, she comments about not being able to follow the conversation (about herself) because she does not speak Tagalog. The other participants quickly apologize and fill her in, restoring the harmony of the friendly banter. What I want to focus on here is the fact that Francisca was alerted to the conversation by three referential signs: an extrinsic one, when her name was tagged in the posting of the photograph; and two intrinsic ones, the photograph itself and her name as mentioned in an otherwise unintelligible turn. The initial reference of the tag pointed her toward this particular photograph, while the mention of her personal name allowed her to follow the gist of the interaction in Tagalog and to seek clarification. Without the presence of the proper name she would have been blind to the exchange. This means that participants in transidiomatic communication may not understand each other through indexical processes but have to rely on referential and denotation signs to know what is going on. It is not a surprise then

Language, media, and digital landscapes

485

that lexical forms such as proper names, due to their tendency to hold something of their form cross-linguistically, are highly monitored in any transidiomatic environment. This leads us to argue that the shift from single languages or even multilingual repertoires to the transidioma requires a reconceptualization of the connection between communication and shared linguistic and cultural knowledge. We can no longer assume that such shared knowledge exists to provide a common ground from which to negotiate encounters, conflicts, or agendas. The identification and establishment of common ground itself must be understood as a major challenge in the process of transidiomatic communication. 3.2

Mechanical translation

The complexity of online multilingual communication is at time managed through mechanical translation (MT). In the human/computer interaction we are witnessing the rise of an integrated structure of techno-linguistic mechanisms in charge of linguistic exchanges, social interaction, and power relations. These mechanisms are present at different levels: the technical level of computer programming, the linguistic level of the interface between software and users, and the interactive level between users. Machine translation is one such techno-linguistic mechanism. The diffusion of Internet communication foregrounds the problem of translatability. Today, the Internet provides an unprecedented platform for communication across national languages. However, the issue of transnational translatability remains a barrier in global communication via the Internet. We have witnessed a rapid evolution from a time when English seemed destined to become the only language of the Web to the present situation where more and more net-users bring “their own dialects to the online potlatch” (Silberman 2000: 226). In the digital ‘global village’, people still talk in their native tongue, so those interested in global communication must resolve the problem of the intelligibility of their message. They must, in other words, rely on translation. From its genesis at the post-World War II dawn of computing – when ambitious researchers believed it would take only a few years to crack the language problem – until the late 1980s, machine translation consisted almost entirely of what are known as rule-based systems. As the name implies, such translation engines required human linguists to combine grammar and syntax rules with cross-language dictionaries. The simplest

486

Marco Jacquemet

rules might state, for example, that in French, adjectives generally follow nouns, while in English, they typically precede them. But given the ambiguity of language and the vast number of exceptions and often contradictory rules, the resulting systems ranged from marginally useful to comically inept. Systran provides an interesting study of the electronic spread of rulebased translation. This company has been building translation software for more than 30 years. Its initial technology was developed in the early 1960s at Georgetown University within the context of the cold war: it served to monitor Russian communications through a primitive transposition of Russian words into their English equivalents. If the translating filter intercepted a target word, it would flag the message, to be examined later by a human translator. Currently Systran enables its clients – mainly corporations with web sites that must be understood by readers around the world – to “dynamically translate content”. Its software enables users to translate text content from both Mozilla Firefox and Internet Explorer pages into six different languages. Systran provides machine translation software for Google’s Babel Fish (often netsurfers’ first exposure to computer translation) and Go Translator. The competence of rule-governed MT programs is, unsurprisingly, quite limited. The inability to translate ambiguity, irregular syntax, and multiple meanings, and to take context into consideration; make MT the butt of jokes in many linguistics departments.1 The 1959 prediction by Yehoshoa Bar-Hillel, the first U.S. academic researcher to work full time on automatic translation (he was employed by MIT in 1951), that fully automatic, high-quality translation “was an unreachable goal, not only in the near future but altogether” still stands. The flash of intelligence is obviously still missing from MT. Over the past decade, however, machine translation has improved dramatically, propelled by the relentless march of Moore’s law, a spike in federal funding in the wake of 9/11, and, most important, a new idea. The idea dates from the late 1980s and early 1990s, when researchers at IBM stopped relying on grammar rules and began experimenting with sets of already1

See for instance Eco 2001 for the hilarious consequences of multiple translation passages (from language A to language B to language C to language A). This jocular animosity is shared and reciprocated by MT researchers who have even their own jokes against linguists, dating from the times when their methods were under the heavy influence of theoretical linguistics: “Every time we get rid of a linguist, our MT gets better” (quoted in Silberman 2000:228).

Language, media, and digital landscapes

487

translated work known as parallel text. In the most promising method to emerge from the work, called statistical-based MT, algorithms analyze large collections of previous translations, or what are technically called parallel corpora – sessions of the European Union, say, or newswire copy – to divine the statistical probabilities of words and phrases in one language ending up as particular words or phrases in another. A model is then built on those probabilities and used to evaluate new text. A slew of researchers took up IBM’s insights, and by the turn of the 21st century the quality of statistical MT research systems had drawn even with five decades of rule-based work. The success of statistical systems, however, comes with a catch: such algorithms do well only when applied to the same type of text on which they have been trained. Statistical MT software trained on English and Spanish translations of the BBC World Service, for example, excels with other news articles but flops with software manuals. As a result, such systems require large amounts of parallel text for not just every language pair they intend to translate – which may not be available for, say, Pashto – but different genres within those language pairs as well. What is missing from these projects in search of perfect translation is an awareness of the fuzzy nature of all communication and the negotiation of meanings carried out by social groups in the structuration, diffusion, and interpretation of language in context. That does not mean that there is no role for computers in automating the translation process. Even a mechanical translation that fails to render the full significance of the source language may still have value to a reader able to piece together meaning from the lessthan-well-formed text. This point may be made clearer by looking at a recent facebook conversation, prompted by the posting of a modified poster of the movie Avatar by an Italian-speaking member of the multilingual group, composed of English, Spanish, Italian, Tagalog, French speakers, I have been following for the last two years. Paolo used the modified poster to encourage his Italian facebook friends to go to the polling stations to vote for the 2011 popular initiatives against some laws recently approved by the Italian government. He rhetorically asked people where they would go the day after (the day of the elections) and posted the modified image: the bill for the movie had been transformed into A Votar, which could be roughly translated as ‘to the vote’. The clever post elicited a number of responses, some of which simply “liked” the post, while others commented on the voting process. The smooth exchange of Italianonly messages was interrupted by one of Paolo’s friends, Tagalog-speaking Fermin who was in the process of learning Italian, who inserted himself in this

488

Marco Jacquemet

stream with a request for clarification, which turned out to be quite puzzling: “che cosa è tutto questo il circa?” (literally ‘what is all this the approximative?’): Case 2 Paolo

Dove andate domani? (Where are you going tomorrow?)

Giovanna, Lucia and Leide like this.

Paolo

Anche se qui abbiamo già fatto (even if here we already voted) Saturday at 10:24am · Like Unlike ·1 person likes this Lucia quattro ore di aliscafo,un bel mare incazzato e altrettanto bella levataccia … ma ho votato! (four hours on the boat, a quite angry sea and also super early rise … but I voted!) Sunday at 9:06am · Like Unlike ·1 person like this Gabriella Paolo purtroppo i vostri voti non sono validi … (Paolo unfortunately your votes won’t count …) Sunday at 9:34am · Like Unlike Giovanna Un bel batticuore oggi … sono le 18.35 ed ancora i dati di affluenza delle 18 non ci sono (I am very nervous today … it’s 6:35 pm and we don’t have yet the 6 pm report on the number of voters)

Language, media, and digital landscapes

489

Sunday at 9:36am · Like Unlike Paolo Gabriella, però magari riusciamo a far scendere il quorum! (Gabriella, but maybe we’ll cause a lower quorum) Sunday at 5:36pm · Like Unlike Fermin che cosa è tutto questo il circa? (lit. ‘what is all this the approximative?’) Sunday at 5:56pm · Like Unlike Paolo Fermin: In Italia questa domenica e lunedì si votano 4 referendum (contro il nucleare, per l’acqua pubblica, e perchè la legge sia uguale per tutti). (Fermin: In Italy this Sunday and Monday people vote for 4 popular initiatives (against nuclear energy, against water privatization, and against ad personam legislature) Sunday at 6:20pm · Like Unlike Ige I see. A votar = to vote. Ho capito! (I see. A votar = to vote. I understand!) Sunday at 7:20pm · Like Unlike The oddity of Fermin’s turn could have been ignored (or could have even led to sarcastic comments). Paolo however understood his overall meaning and replied appropriately, providing contextual information not easily available to a Filipino living literally on the other side of the globe, mostly insulated from the arcane ways of Italian politics. My impression, confirmed also in ethnographic interviews with both Firmin and Paolo, was that Firmin’s turn was the result of MT. Firmin, as most Filipinos quite at ease in English environments, had used babelfish.com to translate a typical English expression for seeking information: “what is all this about?” became the unfortunate “che cosa è tutto questo il circa” (‘what is all this the approximative?’). Paolo, a multilingual speaker with good English fluency, was able to recognize in Firmin’s query the English expression and he properly attended to it. MT translation made possible a conversation otherwise unattainable. Even if Firmin’s turn was not well formed, it nevertheless achieved its goal of seeking information, thanks to the overall context and Paolo’s familiarity with transidiomatic exchanges.2

2

Another example may clarify my appeal for fuzziness. Recently, while searching for information about tattoo parlors in NYC, I ran a Google search and found an article in a German webzine (magazinUSA.com). Using MT technology, I translated the article, which in part read as:

490

Marco Jacquemet

If we want to translate literary texts such as those by James Joyce (Eco 2001) and attempt to overcome the “restricted permeability” always present in translating such texts (Spivak 1998), we should stay clear of MT technologies for years to come, and maybe forever. If, on the other hand, we are just looking for some pointers (a function, after all, crucial to all processes of indexicality), and are not going to be put off by text that is not ‘standard’3, then MT could still be quite useful. As in face-to-face talk, where people manage to understand each other’s fragmented sentences through continuous feedback and guesswork, computer-mediated communication allows net-users to achieve understanding through rapid back-and-forth exchange (see for instance what is happening during chat room talk, where the high levels of communicative inaccuracy and quick tempo of overlapping themes require a continuous recourse to conversational repairs and redundancy, Jepson 2005). The same sort of interactivity can be imagined taking place between net-users and MT software. Expectations that MT offers a universal, simultaneous operator between different languages (the unicode as unicorn), are based upon the prevailing view of language as a single, well-formed, and self-contained system. This means that the lenses we usually adopt in looking at language must be sigNew York ist die Geburtsstadt des ‘modernen’ Tätowierens mittels mechanischer Hilfmittel. (…) Die Preise variieren extrem, je nach Künstler und Ruf in der Szene – aber Stundensätze in New York zwischen $100-$200 für ein individuelles Design sind normal. Die Top-Künstler der Szene haben Wartelisten bis zu zwei Jahren. Spontane Termine sind allerdings grundsätzlich eher die Seltenheit und meistens wird ein Termin vereinbart. Eine Anzahlung ist üblich. In a few seconds, the MT engine spelled out the following statement: New York is the birth city of the ‘modern’ Taetowierens by means of mechanical helping means. (…) The prices vary extremely, depending upon artist and call in the scene – but hourly rates in New York between $100–$200 for a non-standard Design are normal. The Top artists of the scene have waiting lists up to two years. Spontaneous dates are however basically rather the rarity and a date are mostly agreed upon. A pre-payment is usual.

3

This text is barely comprehensible, requires a good deal of guesswork (“Taetowierens” as ‘tattoos’), and would hardly pass a 6th grade composition exercise. Yet, by applying a little inferential work, I was able to retrieve some valuable information, such as the price of a original design, and the existence of waiting lists for the best artists. Moreover, the article listed tattoo parlors in the city, including links to their websites. All useful information otherwise unattainable. We must also be aware of the ideological process behind all calls for standardization as a strategy of hegemonical pressure (Crowley 1989; Silverstein 1996; Holborow 1999).

Language, media, and digital landscapes

491

nificantly altered to accommodate for communicative phenomena that lack grammatical and syntactical order, or that cannot even be recognized as part of a single standard language.

4.

Conclusions

In this chapter I have argued that our understanding of the link between language and the media must be first problematized as to what to consider a mediation and then expanded to take into account the emerging mediascapes of the 21st century: global medias, online social network websites, multilingual chat rooms, or videoconferencing are all communicative environments which are currently poorly understood. As such, they deserve our full attention. A serious investigation of these communicative environments will necessarily problematize our taken-for-granted, common-sense knowledge of what is a ‘language’ (Jacquemet 2005). It is now time to examine communicative practices based on disorderly recombinations and language mixings until now ordinarily overlooked. We should rethink the concept of communication itself, no longer embedded in national languages and international codes, but in the multiple transidiomatic practices of transnational flows. This will allow our imagination of linguistic exchanges to take shape within the discourse of local cultural becoming, social mutations, and global identities. The synchronic, granular, and multimodal characters of the transidioma allow people to experience the mediation made possible by wider flows and circuits of contemporary mediascapes. While this mediation will have necessarily lost some of the basic communicative characteristics we routinely associate with ‘the media’ (broadcasted, centralized, monolingual, one-way produced), it nevertheless force us to focus our attention on the mutating conditions for communication in the discontinuous space of late modernity.

References Anderson, Benedict 1983: Imagined Communities. London: Verso. Androutsopoulos, Jannis 2008: Discourse-centred online ethnography. In: Jannis Androutsopoulos & Michael Beißwenger (eds.), Data and Methods in Computer-Mediated Discourse Analysis. Special Issue, Language@Internet 5. http://www.languageatinter net.de. Baudrillard, Jean 1984: Simulacra and Simulation. University of Michigan Press. Bruns, Axel 2008: Blogs, Wikipedia, Second Life and Beyond: From Production to Produsage. New York: Peter Lang.

492

Marco Jacquemet

Cisco Corporation 2011: Cisco TelePresence Trends. http://www.cisco.com/application/ pdf/en/us/guest/products/ps7073/c1244/cdccont_0900aecd8054f897.pdf. Crowley, Tony 1989: Standard English and the Politics of Language. Urbana: University of Illinois Press. Crystal, David 1997: English as a Global Language. Cambridge: University Press. Crystal, David 2006: Language and the Internet. Cambridge: University Press. Danet, Brenda & Susan C. Herring 2007: The Multilingual Internet: Language, Culture, and Communication Online. Oxford: University Press. De Swaan, Abram 2001: Words of the World. London: Polity Press. Dubé, Line, Anne Bourhis & Réal Jacob 2006: Towards a typology of virtual communities of practice. Interdisciplinary Journal of Information, Knowledge, and Management 1: 69–93. Eco, Umberto 2001: La Bustina di Minerva. Milan: Bompiani. Eisenlohr, Patrick 2004: Language revitalization and new technologies: Culture of electronic mediation and the reconfiguring of communities. Annual Review of Anthropology 32: 21–45. Gamble, Jennifer M. 2007: Holding environment as home: Maintaining a seamless blend across the virtual/physical divide. M/C Journal, 10(4). http://journal. media-culture.org.au/0708/11-gamble.php. Gellner, Ernest 1983: Nations and Nationalism. Ithaca, NY: Cornell U.P. Gillespie, Marie 1995: Television, Ethnicity and Cultural Change. London: Routledge. Gitlin, Todd 1983: Inside Prime Time. New York: Pantheon Books. Hanks, William F. 1996: Language and Communicative Practice. Boulder, CO: Westview. Hannerz, Ulf 1992: Cultural Complexity: Studies in the Social Organization of Meaning. New York: Columbia University Press. Hannerz, Ulf 1996: Transnational Connections. London: Routledge. Herring, Susan C. 1996: Two variants of an electronic message schema. In: Susan C. Herring (ed.), Computer-Mediated Communication: Linguistic, Social and Cross-Cultural Perspectives, 81–106. Amsterdam: John Benjamins. Holborow, Marnie 1999: The Politics of English: A Marxist View of Language. London: Sage. Hutchby, Ian 2001: Technologies, texts, and affordances. Sociology 35(2): 441–456. Jacquemet, Marco 2005: Transidiomatic practices: Language and power in the age of globalization. Language and Communication 25: 257–277. Jepson, Kevin 2005: Conversations – and negotiated interaction – in text and voice chat rooms. Language Learning and Technology 9(3): 79–98. Jones, Rodney 2004: The problem of context in computer-mediated communication. In: Levine, Philip & Scollon, Ron (eds.), Discourse and Technology, 20–33. Washington, D.C.: Georgetown University Press. Karim, Haiderali Karim 2003: Introduction. In: Karim H. Karim (ed.), The Media of Diaspora, 1–19. London: Routledge. Mankekar, Purnima 1999: Screening Culture, Viewing Politics: An Ethnography of Television, Womanhood, and Nation in Postcolonial India. Durham: Duke University Press. Mazzarella, William 2004: Culture, globalization, mediation. Annual Review of Anthropology 33: 345–67. Mc Luhan, Herbert M. 1994: Understanding Media: The Extension of Man. Cambridge, MA: MIT Press. Paolillo, John C. 1996: Language choice on soc.culture.punjab. Electronic Journal of Communication 6(4).

Language, media, and digital landscapes

493

Pennycook, Alastair 2007: Global Englishes and Transcultural Flows. London: Routledge. Peters, John D. 1999: Speaking into the Air: A History of the Idea of Communication. Chicago: Chicago University Press. Rampton, Ben 1999: Styling the other: Introduction. Journal of Sociolinguistics 3: 421–427. Schegloff, Emanuel A. 1988: From interview to confrontation: observations on the bush/rather encounter. Research on Language and Social Interaction 22: 215–240. Shipley, David & Will Schwalbe 2007: Send: The Essential Guide to Email for Office and Home. New York: Knopf. Shove, Elizabeth 2003: Comfort, Cleanliness and Convenience: The Social Organization of Normality. Oxford/New York: Berg. Silberman, Steve 2000: Talking to strangers. Wired 8. 05. http://www.wired.com/ wired/archive/8.05/translation_pr.html. Silverstein, Michael 1996: Monoglot ‘standard’ in America. In: Donald Brenneis & Ronald K. S. Macaulay (eds.), The Matrix of Language, 284–306. Boulder, CO: Westview. Spitulnik, Debra 1993: Anthropology and mass media. Annual Review of Anthropology 22: 293–315. Spivak, Gayatri Chakravorty 1998: In Other Words: Essays in Cultural Politics. London: Routledge. Tomlinson, John 2007: The Culture of Speed. London: Sage. Virilio, Paul 1988: La machine de vision. Paris: Galilée. [Engl. translation (1994) The Vision Machine. Indiana University Press]. Virilio, Paul 1990: L’inertie polaire. Paris: Christian Bourgois. Wasson, Christina 2006: Being in two spaces at once: virtual meetings and their representation. Journal of Linguistic Anthropology 16(1): 103–130. Williams, Raymond 1977: Marxism and Literature. London: Oxford University Press.

494

Michael Beißwenger

Michael Beißwenger

Space in computer-mediated communication Corpus-based investigations on the use of local deictics in chats*

1.

Introduction

Compared with spoken face-to-face or telephone conversation, the coordination between interlocutors in written dialogic communication via the Internet is more complicated due to the particular technological framework conditions. Interlocutors lack a direct, mutual intersubjective perception and awareness, they have to use writing for the encoding of their dialogue contributions, and their written utterances are not transmitted (and perceived) in a keystroke-by-keystroke mode but at once. Even though the communicative setting is a synchronous one, written interaction in chats or instant messaging systems lacks the property of synchronization between users. Accordingly, coordination is complicated on the level of interactional management as well as on the level of the interactional construction of meaning. The effects of written computer-mediated communication (CMC) on interactional management and the strategies developed by the users of synchronous CMC systems to deal with these effects have been examined in a range of studies on the coordination of actions and the status of turn-taking in synchronous CMC (cf. e.g. Murray 1989; Garcia & Jacobs 1998, 1999; Cherny 1999; Herring 1999; Storrer 2001; Beißwenger 2003, 2007; Schönfeldt & Golato 2003; Zitzen & Stein 2005; Markman 2006). This article makes a contribution to the examination of the construction of meaning in synchronous written CMC. It addresses the use of local deictic expressions in chats. While the meaning of symbols (“Nennwörter”, Bühler 1934 [1982]) can be derived from social conventions, the meaning of deictic or pointing expressions (“Zeigwörtern”) such as the German ich (‘I’), du (‘you’), hier (‘here’), da, dort (‘there’) and jetzt (‘now’) relates to knowledge of the situational (social, temporal, spatial) context in which the utterance is made. * My gratitude goes to Angelika Storrer for valuable hints and discussions on the topic of deixis in CMC as well as to Laura Häckel, Tatjana Kindop, Alla Krasnokutskaya, Alexander Kurek and Bianca Stockrahm for their support in analyzing the corpus data.

Space in computer-mediated communication

495

Deictics are used for pointing within the boundaries of a social, temporal or local coordinate system which is determined and delimited by the respective speaker/writer. In order to understand the deictic utterance, the addressee must be able to reconstruct this coordinate system and identify the origo (Bühler), i.e. the origin of the speaker’s/writer’s deictic pointing (cf. e.g. Bühler 1934 [1982]; Ehlich 1983 [2007]; Hoffmann 1997; Kameyama 2007). According to Ehlich (1983 [2007]: 19), deictic pointing is (a) Ausdruck (Exothese) einer Fokussierung, die [ein Sprecher] S auf Elemente des Verweisraumes selbst vorgenommen hat, (b) Aufforderung an den Hörer H, diese Fokussierung seinerseits zu vollziehen und so für S und H eine gemeinsame Fokussierung innerhalb des Verweisraumes herzustellen” [‘(a) the expression (Exothese) of a focusing operation that [a speaker] S himself has applied to elements of the deictic space, (b) a prompt for the listener H to implement this focusing operation on his part and thus establish for S and H a joint focal orientation within the deictic space’] (translation MB).

Ehlich (1983 [2007]: 20) defines deictic space as a mehrdimensionale Größe, deren Grenzen die möglichen Objekte der deiktischen Prozedur von solchen Objekten trennen, die nicht durch eine einfache Suchaufforderung des Sprechers zur Fokussierung und Findeprozedur des Hörers Gegenstand der kommunikativen Relevanz werden können [‘multidimensional value, the borders of which divide the potential objects of the deictic procedure from such objects that cannot become subject of communicative relevance through a simple search prompt by the speaker’] (translation MB).

Still, deictic pointing is not limited to the physical determinants of face-toface encounters, but can also be expanded to imagined spaces as well as to abstract text or discourse spaces (Ehlich 1983 [2007]). An explorative description of the particularities of local deictic pointing in chats is given in Storrer (2001: 17–21). Storrer shows that in chats at least three different types of spaces are applied as deictic spaces: (a) the interlocutors’ surroundings in the real world, (b) the metaphorical “chat room” and (c) the two-dimensional space of the chat protocol displayed on the computer screen. The present article starts with an outline of the challenges posed by the use and interpretation of local deictics in chats (section 2). Using Bühler’s distinction of types of deictic spaces as a starting point, the article gives an overview of the different concepts of space which serve as deictic spaces when interlocutors use local deictics in chats, including and expanding the spaces differentiated in Storrer (2001) (section 3). Examples are taken from the Dortmund Chat Corpus, a corpus of German chat communication (http://www.chatkorpus.tu-dortmund.de).

496

Michael Beißwenger

The focus of the article is on the use of the German local adverbs hier (‘here’), dort and da (‘there’). Consideration is also given to “quasi-deictic” expressions like oben (‘above/up there’), unten (‘below/down there’), links (‘left’) and rechts (‘right’), which – like the “real” deictics – are related to the origo, but in which “an additional frame of reference comes into play” (Hoffmann 1997: 312; for the difference between deictic and quasi-deictic local adverbs cf. Hoffmann 2007: 239–245). Consideration will also be given to cases in which local deictics are used for the progression of topics (anadeictics, cf. Ehlich 2007). Section 4 discusses why the use and interpretation of local deictics seems to be widely unproblematic even though chat communication is written distance communication in which interlocutors do not share a joint perceptual space. It presents the results of a combined qualitative-quantitative study in which ~1,500 occurrences of proximal hier and distal dort in the Dortmund Chat Corpus have been coded according to the underlying types of deictic spaces. Each occurrence has been evaluated to determine in which cases the relevant deictic space has been explicitly specified in the context and whether the distribution of the relevant spaces differs in different types of chats (free/non-moderated vs. institutional/moderated chats). The results shed light on the achievement of interpersonal understanding in chat as well as what strategies interlocutors develop in order to deal with local deictics even under the conditions of synchronous written communication on the Internet.

2.

Challenges in dealing with local deictics in chat communication

Example 1. Moderated chat question time with Smudo, songwriter and rapper of the German hip hop group Die Fantastischen Vier (Dortmund Chat Corpus, Document 1306017).1 1

anna-sui

2

Smudo

1

holaa smudo! wo bleibt der rest der truppe? ‘hey smudo! where’s the rest of your posse?!’ anna-sui – keine ahnung wo die anderen sind – ich sitze hier in meiner privaten wohnung in hamburg ‘anna-sui – no clue where the others are – i’m sitting here in my private apartment in hamburg’

Jumps in the numbering in the chat contributions occur – in this as well as in all of the following examples – because messages from the original document that are irrelevant to the argumentation have been left out. The sequence can be found in its entirety in the Dortmund Chat Corpus at http://www.chatkorpus.tu-dortmund.de.

Space in computer-mediated communication 3

Brian

4

Smudo

86 zest

87 Smudo

497

hey smudo wo habt ihr den viedoclip gedreht für *mfg* ‘hey smudo where did you guys shoot the video for *mfg*’ brian- in barcelona – guck mal hier http://www.smudo.com/smudo/ mfg_dreh.htm ‘brian- in barcelona – check it out here http://www.smudo.com/ smudo/mfg_dreh.htm’ smudo: was ist eigentlich das besondere an hamburg? was hat die stadt, das andere staedte nicht haben? warum ist hiphop gerade dort so stark? ‘smudo: what’s the best thing about hamburg? what does that city have that others don’t have? why is it that hip hop is so popular there?’ warum hiphop hier so stark ist vermag ich nicht zu sagen. die attraktivitaet von hamburg finde ich hat nichts mit der musik hier zu tun. das wasser ist nah und gut. die stadt ist sehr liberal und bietet sehr viel abwechslung und kulturelles leben ‘i can’t really say why hip hop is so popular here. for me hamburg’s attraction has nothing to do with the music here. the water is close by and good. the city is very liberal and has a lot of different things to offer and lots of culture.’

Example 1 is an excerpt from a moderated chat with a celebrity. The celebrity is Smudo, one of the members of the German hip hop group Die Fantastischen Vier. In the beginning of the chat event, a user named anna-sui asks Smudo about the other members of the group – obviously not knowing that they will not be joining the chat event. The local adverb wo ‘where’ here obviously does not refer to any place in the real world; instead, anna-sui uses it to refer to the online environment where the interlocutors are logged in during the chat and which she imagines as being a sphere of joint presence. Smudo interprets the wo that anna-sui uses as referring to the real world: he replies that he has no idea where his band mates are at the moment. Additionally, he specifies where he is located while logged into the chat using the local deictic hier (‘here’). Since he cannot assume that the other chat participants know where he is physically located during the chat, he provides additional information about where his origo is located: hier in meiner Wohnung (‘here in my apartment’). He also specifies the geographic region where his apartment is situated: hier in meiner Wohnung in Hamburg. Later in the event, in message 86, a user named zest points at the region of Hamburg with the deictic dort ‘there’ – thus implicitly referring to Smudo’s hier from the beginning of the event and creating a deictic space in which his own situatedness and the city or region of Hamburg (which had been introduced as being Smudo’s current location) are subspaces and, thus, belong to the same ‘world’. In his reply, Smudo ratifies the deictic space that zest has introduced by using hier to point at the region of Hamburg (and not – like he did in message 2 – to point at his apartment in Hamburg).

498

Michael Beißwenger

A different use of hier can be found in message 4: After Brian asks Smudo about the real-world location where their recent video had been filmed, Smudo replies with Barcelona, followed by guck mal hier (‘check it out here’) and the URL of a webpage with background information on the video shooting. Pointing at URLs using deictic hier is quite frequent on the Web: The denotation of the deictic in such cases is neither an object in the real world nor an imagined object nor an object on the screen (i.e. the character string representing the URL given after hier). Instead, this hier refers to the content of the webpage which can be found by visiting the webpage connected with the respective URL. Without going into more detail, we can state that in Example 1, the interlocutors constitute three different deictic spaces or concepts of space: the online environment (in Wo bleibt der Rest der Truppe?); two types of environments in the real world: (a) a deictic space including Smudo’s apartment as the proximal subspace and the unspecified rest of the world as the distal subspace; (b) a deictic space including the region of Hamburg as one subspace (at which Smudo points with hier and zest using dort, according to their different locations and, thus, deictic origins) and the unspecified rest of the world (including zest’s current location) as the other subspace; the World Wide Web as a space in which content is organized in documents and which can be metaphorically considered to be “places” to which one can “travel” by entering a certain URL into the browser or by activating a hyperlink. x x

x

The example reveals a lot about how we, in everyday communication, deal with local deictic expressions and local adverbs that have similar functions. It illustrates which challenges and difficulties are connected with using and interpreting pointing expressions in distance communication where the interlocutors do not share a joint perceptual space. According to Klein (1978), the general challenges when interpreting local deictics can be differentiated as follows: the coordination problem (Koordinationsproblem): The listener/reader has to reconstruct what the speaker/writer is pointing at when using hier, dort and da; the identification problem (Identifikationsproblem): The listener/reader has to identify the place where the speaker/writer is located at the time of his utterance or rather – e.g. in the case of “shifted deixis” – the place that serves as the origin for deictic pointing; the problem of the deictic space (Raumproblem): The listener/reader has to determine which concept of space the speaker/writer is using as a basis

x

x

x

Space in computer-mediated communication

x

x

x

499

in his use of local deictic expressions (his real surroundings? An imagined or remembered room/space? A discourse space?); the problem of determining the origo (Problem der Origofestlegung): If the speaker/writer introduces a new or additional origo which cannot be derived from knowledge about the speaker/writer’s current location, then the listener/reader must identify this origo before being able to adequately comprehend pointing at hier, dort and da; the delimitation problem (Abgrenzungsproblem): The listener/reader must determine how the speaker/writer divides the chosen deictic space into subspaces, i.e. how the subspaces created with hier, dort and da can be distinguished from one another; the problem of deictic oppositions (Problem der deiktischen Oppositionen): The listener/reader must determine how the speaker/writer structures the subspaces, which he outlines with the use of dort and da, as opposed to the target of his deictic pointing with hier.

The differentiation proposed by Klein portrays the various reconstruction and coordination efforts necessary to interpret the use of local deictic expressions. Accordingly, Klein states that it is “amazing that such a complex system can work so well because in reality we have relatively few difficulties identifying what is meant when someone says ‘dort’, ‘hier’ and so on” (Klein 1978: 36). Conversely, he asserts that the system must be so effective because, (a) keeping the foundations of human physique and perception in mind, it is a “very obvious” system and (b), in contrast to frames of reference which are independent of the situation and the origo, it is characterized by a great deal of flexibility (Klein 1978: 36). In chat communication, the coordination efforts with which the interlocutors are faced when dealing with local deictics are much more complex than in spoken or face-to-face communication. In addition, the lack of immediate interpersonal perception makes the use of local deictics in chat more ambiguous than in other forms of communication: Unlike in face-to-face communication, interlocutors lack a shared space of perception which would enable them to create joint attention and to make use of “demonstratio ad oculos”. Unlike in face-to-face and telephone conversations, the messages are being transmitted en bloc, i.e. the utterance has to be produced in its entirety, sent and transmitted before it can be received and mentally processed by the addressees. Thus, chat communication is synchronous without synchronization (sensu Auer 2000), i.e., it is not possible for the recipients to simultaneously track the process of verbalization. Instead, only the products of verbaliz-

x

x

500

Michael Beißwenger

ation are perceived – ex retrospect – and thus without the possibility of giving simultaneous feedback (cf. Beißwenger 2007; 2010: 253–261). Moreover, the written nature of the exchange, the use of the Internet as the communication infrastructure as well as the dependency on computers and their peripheral devices all introduce additional topological concepts that become relevant as deictic spaces for linguistic pointing. The problem of dealing with pointing expressions under the conditions of synchronous written distance communication via the Internet has already been addressed in previous research on chat. Storrer (2001) notes that interpreting local deictics is more challenging in chat than in oral discourse, and Haase, Huber, Krumeich & Rehm (1997: 67–68) give an example from a chat log file in which the meaning of hier becomes problematic, since Horst, one of the users, asks for clarification: Example 2 (from Haase, Huber, Krumeich, and Rehm 1997: 67)



Gleich wird Theo herkommen ‘Theo is about to come over’ Hier in den IRC? ‘Here in the IRC?’ Horst: Nein, er kommt mich besuchen für das Wochenende ‘Horst: No, he’s coming to visit me for the weekend’

Surprisingly, cases similar to the one given in Example 2 in which problems of interpreting local deictics are explicitly addressed can hardly be found in the log files included in the Dortmund Chat Corpus.2 That does not necessarily mean that such problems do not occur – still, this finding is surprising at first; it suggests that, even within the framework conditions of synchronous written distance communication, the use and interpretation of local deictics is largely unproblematic. Section 3 describes a set of spatial concepts adopted as deictic spaces in the chat conversations documented in the Dortmund Chat Corpus. Drawing on a corpus-based case study, section 4 subsequently explores why chat participants normally do not have any difficulties dealing with the potential ambiguity of local deictics.

2

I analyzed two subsets of data from the Dortmund Chat Corpus with a total of 66,592 messages (707,132 tokens) which contained 1,472 occurrences of the local deictics hier and dort. The same data sets were used for the quantitative-qualitative investigation described in section 4.

Space in computer-mediated communication

3.

501

Types of deictic spaces in chat communication

According to Bühler’s “modes of pointing” (Modi des Zeigens, Bühler 1934 [1982]: 80) and the distinctions made in Hoffmann (1997: 313–316), one can first differentiate between several types of deixis according to the underlying type of deictic space and the accessibility of its elements: – True (situated) deixis (in Bühler’s terms: demonstratio ad oculos) comprises forms of deictic pointing in real spaces which are accessible for direct perception. If the addressee is not present in the context of the production of the deictic utterance (such as in temporally and/or spatially “displaced” communication sensu Ehlich 1983a), he needs further details to be able to interpret the deictic expression in a way that corresponds with the intention of the speaker/writer. – Imaginative deixis (Bühler: Deixis am Phantasma) comprises forms of pointing at targets in an imagination space. These spaces can be real (remembered) or completely imagined spaces (Bühler 1934 [1982]: 123; cf. also the distinction in Sitta 1991: 14–17). – In discourse deixis the speaker/writer points at statements previously made or to be made by the listener/reader or at preceding or subsequent discourse elements. In these cases, the sequential progression of discourse functions as the deictic space, with the discourse items as targets for deictic pointing. The so-called Lokutive Textdeixis which describes cases of linguistic pointing at elements on the material (textual) surface of the written utterance (Hoffmann 1997: 316) can be counted as a special case of discourse deixis. It is typical of communicating over spatial distance that, in the case of true (situated) deixis, the deictic space relevant for deriving the meaning of local deictics must be created in the imagination of the recipient. The only space in chat communication that is equally accessible for all interlocutors is the written protocol of the chat history shown on the screen. All other types of spaces which can be used as deictic spaces have to be created in the imagination and cannot necessarily be presumed to be accessible to the recipient. Since in constellations of spatial distance without visual contact at least one party has to (re-)construct the relevant deictic spaces in their imagination, I will differentiate the various types of deictic pointing according to the type of underlying deictic spaces. This differentiation is based on the ontological status of the different topological structures: Topological structures can be real, directly perceived or at least potentially experienceable phenomena or creations of the imagination; accordingly, I essentially distinguish between

502

Michael Beißwenger

spaces in the real world and spaces that are construed in the imagination. The notion of ‘space’ in this context can refer to any type of topological concept – threedimensional as well as two- or one-dimensional concepts. In spaces that are construed in the imagination the relevant deictic spaces are accessible for the perception of none of the interlocutors. Their topological properties and structure are established or construed solely by means of the interlocutors’ imagination. They are more or less based on the properties of locations that exist in the real world. In some cases, the properties of the imagined spatial concepts can even be metaphorical and/or their structural properties can, in contrast to real spaces, be extremely reduced. The latter holds, for example, for conceptualizing sets of discourse items as ensembles of objects spatially positioned in relation to one another; the former and the latter both hold for the conceptualization of chat interaction as an encounter in “virtual” places (metaphor of the “chat room”). The essential property of real spaces is that they exist (or have existed) as real, empirical phenomena. If the addressee of the deictic utterance does not know the respective location, he has to imagine it based on non-deictic clues and/or apply other contextual information. Still, it is a space that is or was accessible for empirical verification. This is not the case with spaces that are construed in the imagination. The outcome for the classification outlined here (according to the type of deictic space and not the type of deixis) is first and foremost a clear distinction between the existing and non-existing objects. These can then be subdivided according to further characteristics. The following types of deictic spaces can be empirically substantiated in the data in the Dortmund Chat Corpus: A. Spaces in the real world: A1. Three-dimensional spaces (environments in the real world) A2. Interfaces of objects A3. Interfaces on the screen A4. The chat log file (as a part of the screen interface) B. Spaces construed in the imagination: B1. The discourse space B2. The Internet as a resource space B3. The metaphorical “chat room” B4. Fictitious scenes In the following sections the various deictic spaces and their properties will be characterized more closely by analyzing selected corpus examples.

Space in computer-mediated communication

3.1

503

Spaces in the real world

In the real world, spaces of any size can be constituted as deictic spaces. The anchor point for the proximal subspace (the “here”) coincides either – in the prototypical case – with the spatial position of the speaker/writer or has previously been shifted to a particular (remembered) location in the real world through linguistic or non-linguistic clues (e.g. as in directions that are based on spaces in the real world; cf. Klein 1979). For a correct interpretation of hier (‘here’), the addressee of the deictic utterance must be able to distinguish between the proximal and the distal subspace or at least to locate the origo and the demarcation of hier. Examples of deictic pointing in real spaces have been given in Example 1. Another one is given in Example 3. The sequence shown here is taken from a chat service on the website of a university library which allows library customers with questions about literature research to contact an employee of the library information desk for a one-on-one chat. The chat opens in a pop-up browser window and can only be seen by or made available to the person seeking help (“User”) and the person from the information desk (“Information”). Example 3: Chat-based library help desk at TU Dortmund University (Dortmund Chat Corpus, Document 1203135). 1 User: 2 Information: 3 User: 4 User: 5 Information: 6 Information: 7 User: 8 User: 9 Information: 10 Information: 11 Information:

Wo ist die Webcam zu finden? ‘Where can I find the webcam?’ Die Webcam gibt es nicht mehr ‘We don’t have the webcam anymore’ hallo! :) ‘hello! :)’ hängt also nur noch so rum? ‘just lying around?’ Hallo! ‘Hello!’ Ja, da dort oben keine PCs mehr sind ‘Yes, since there are no more PCs up there’ ok, danke schön! :) ‘ok, thank you! :)’ eigentlich schade ‘too bad’ brauchten wir keine mehr ‘we didn’t need them anymore’ da sie uns nur dazu diente, ‘since we only used them’ dass wir Benutzern unten schon sagen konnten ‘so that we could tell the users down below’

504 12 Information: 13 Information: 14 User: 15 User: 16 Information: 17 User: 18 Information: 19 system:

Michael Beißwenger ob oben noch ein freier Platz ist ‘if space was available up there’ ah versehe ‘ok got it’ danke für die auskunft ‘thank you for the information’ WeBitte sehr;-))3 ‘You’re welcome:-))’ tschööööö ‘byyyyyyye’ Tschüß! ‘Bye!’ *** Benutzer: hat den chat verlassen. *** ‘*** User: has left the chat room. ***’

The sequence in Example 3 shows the entire dialogue, i.e. the webcam is mentioned for the first time in message 1. By using the definite article (Wo ist die Webcam zu finden?) the user indicates that he knows either about the existence of a webcam in the library building or at least about the existence of a video stream from such a webcam on the Web. His use of the deictic interrogative wo (‘where’) in message 1 does not necessarily refer to the spatial properties of the library building, but could also refer to the World Wide Web as a resource space (cf. section 3.4). The response from the help desk employee (message 2 Die Webcam gibt es nicht mehr) also leaves this open. However, the user specifies in message 4 that he is actually referring to the webcam as a device and not the video stream on the Web by referring to physical characteristics (hängt also nur noch so rum?). Subsequently, the messages from the help desk employee not only presume that the user is acquainted with the spatial layout of the (multi-level) library building, but also that he knows where the computer pool and the help desk are located. Pointing at the former location of the webcam in the building with dort oben (‘up there’) would not be helpful for the user without assuming that the user knows about (a) the spatial properties of the building, (b) the location of the computer room in the building, (c) the location of the help desk in the building and (d) the fact that the chat is supervised by the help desk employees. The same goes for the unten/oben (‘up there/down below’) opposition in messages 11 and 12, which also presume the knowledge that when entering the library building one first has to pass the help desk before reaching the floor with the computer room (… damit wir den Benutzern unten schon sagen konnten / ob oben noch ein freier Platz ist). 3

The segment here is probably a typo leftover from rewriting the message before sending.

Space in computer-mediated communication

505

The user seems to possess all of the required knowledge. His reactions – ok, danke schön! (7), ah vers[t]ehe (8), eigentlich schade (14) and danke für die auskunft (15) – suggest that he does not have any problems reconstructing the deictic space presented to him by the help desk employee. The institutional framing of the university library’s chat service as well as the fact that it is a component of the library’s official website suggest that the majority of the users of this service are connected in some way with the institution and its range of services and thus are also acquainted with the premises. 3.2

Interfaces (physical objects, webpages)

Internet-based communication is apparatus-based communication. When communicating through the Internet one has to look at the display of a technical apparatus (screen) and manipulate the displayed objects by using other technical apparatuses (keyboard, mouse). The technical devices used, their surfaces (physical user interfaces) as well as the displayed virtual user interfaces can constitute a matrix for linguistic localizations. Standardized physical user interfaces often have intrinsic coordinate systems. Areas on these interfaces can be referred to in a non-deictic way using quasi-deictic expressions. In Example 4, zora obviously assumes that tränchen’s repeated use of capitalization is not meant as functional but accidental; she thus takes in the perspective of the addressee’s orientation towards the keyboard and uses quasi-deictic rechts (‘right’) and oben (‘up’) to refer to a certain sector and then to a certain key (‘arrow up’) on the keboard: Example 4: Social chat (Dortmund Chat Corpus, Document 2221003). 1226 tränchen 7zIMTEIS HALLO WAS IST MIT DIR HALLO ß ‘7zIMTEIS HELLO WHAT IS WRONG HELLO ß’ 1233 tränchen 7zIMTEIS HALLO WAS IST MIT DIR AHHLO ß ‘7zIMTEIS HELLO WHAT IS WRONG EHHLO ß’ 1235 zora tränchen hallo? ‘tränchen hello?’ 1240 tränchen 7ZIMTEIS HALLO ‘7zIMTEIS HELLO ’ 1242 zora tränchen die taste rechts mit dem pfeil nach oben drücken ‘tränchen press the button on the right with the arrow up’

In the case of deictic pointing within a webpage (i.e. the graphic representation of WWW documents on the computer screen), the dimensions rechts/ links (‘right/left’) and oben/unten (‘above/below’) are fundamental. The matter becomes even more complicated when the target area of deictic or

506

Michael Beißwenger

quasi-deictic pointing lies outside of the area which is currently displayed on the screen so that it first has to be moved onto the screen window by scrolling. The same holds for pointing at parts of the document which are not displayed by default but only after a certain screen object (e.g. the caption of a pull-down menu) has been manipulated with the mouse. In the sequence given in Example 5, Survivor asks kaetzchen (‘little cat’) siehst du auch aus wie die bilder unten? (‘do you also look like the pictures below?’) and, in his next message, posts the URL of a webpage. The webpage contains pictures of a prototype of a smart agent that looks like a cat (iCat) – but not on the part of the page that immediately becomes visible when opening it. Instead, the pictures are further down on the page and cannot be viewed until one has scrolled down (cf. the screenshots in Figure 1). In the upper part of the page, there is a picture of a smart lamp application. After having scrolled down to the part of the page with the pictures of the cat, the part of the page showing the lamp is not visible anymore. In order to display it again, the user has to scroll up to the top of the page again. Example 5: Social chat (IRC) (Dortmund Chat Corpus, Document 2213001) 95

kaetzchen? ‘kaetzchen?’ 96 kaetzchen yo? ‘yo?’ 97 Survivor siehst du auch so aus wie die bilder unten? ‘do you also look like the pictures below?’ 98 Survivor http://www.heise.de/newsticker/meldung/61718 99 Survivor ^^ 101 Survivor da musst ich irgendwie an cat und kaetzchen denken *fg* ‘that somehow made me think of cat and kaetzchen *fg*’ 104 kaetzchen ne, nicht wirklich, mein latexanzug ist schwarz :-P ‘nope, not really, my latex suit is black :-P’ 113 kaetzchen aber dieses Icat teil sieht ja mal richtig hässlich aus … ‘but this Icat thing looks really ugly …’ 114 Survivor hehe ‘haha’ 115 Survivor jo ‘yep’ 117 Survivor dann lieber die nette lampe oben ‘then maybe the nice lamp up top would be better’ Survivor

Survivor’s use of unten (‘below’) in message 97 does not make sense until message 98 when the URL appears. The URL prompts the addressee to shift her origo to the section of the website that appears after activating the URL in a separate browser window.

Space in computer-mediated communication

507

For kaetzchen the deictic space established in messages 97 and 98 as well as the area focused on within this deictic space are initially only imagined spaces. The space does not become real until she has interpreted Survivor’s messages as a prompt (1) to open the mentioned website, then (2) to shift her origo and then (3) to search for the particular area which Survivor is pointing at with die bilder unten. Kaetzchen obviously does not have any problems comprehending the origo shift and complying with the deictic search prompt … die bilder unten, as proven in messages 101 and 104 in which she mentions the iCat prototypes in the target area of the page using the name given there and referring to their color. Survivor can thus conclude that kaetzchen’s current focus is on the lower part of the webpage; accordingly, in message 117 he uses oben

Fig. 1. Screenshots of the webpage referred to in Example 5. Screenshot 1 shows the upper part of the page, which is seen first when the URL is activated; screenshot 2 shows the part with the photos of the iCat prototypes, which is revealed in the lower part of the website after scrolling down.

508

Michael Beißwenger

(‘up top’) to point at the lamp that is shown on the top of the webpage (which kaetzchen, when opening the page, must have seen already but which, after having scrolled down, is not in her current line of sight anymore). 3.3

Deictic pointing at the discourse and its inscription on the screen

The discourse itself can become a deictic space in two ways: as a topological ensemble of discourse objects and as the sequence of written discourse units preserved in the screen protocol. In the first case, the deictic space is constructed through the chronology of the discourse objects, on which one can point back (Examples 6–8) or ahead (Example 9). Example 6: Academic advising on the topic “teaching degrees” (Dortmund Chat Corpus, Document 1202004) 122 14:51

Kahlua

In Germanistik bin ich fast durch, habe aber mein 2. Fach noch nicht gefunden. Dort liegt mein Problem. ‘I am almost finished with German, but still have not figured out what to pick as my second subject. There’s my problem.’

Example 7: Chat-based cross-university seminar on lexicography (Dortmund Chat Corpus, Document 1102013a) 159 10:56 160 10:56

161 10:56

Teacher2 Das Bildwörterbuch schließlich ist ein klassisches onomasiologisches WB … ‘The picture dictionary is a classic onomasiologic dict.’ Teacher2 da sich die Zugriffsstruktur hier völlig von den sprachlichen Zeichen entfernt … ‘Since the access structure here is completely removed from the linguistic signs …’ Teacher2 und die lexikalischen Einheiten rein über Sachgruppen verfügbar gemacht werden … ‘And the lexical units are only made available through subject categories …’

Examle 8: Academic advising on the topic “Einschreibung und Zulassung” (Dortmund Chat Corpus, Document 1202013) 226 22:24:40 Michael

ich studiere im 1. semester Wiwi auf diplom in bochum und wollte jetzt zum wintersemester auf 2 fach bachelor Wiwi und geographie wechseln … also für geographie muss ich mich bewerben aber für Wiwi nur umschreiben oder? ‘i am studying in my 1st Diplom semester Business in bochum and wanted to switch to the 2-subject bachelor

Space in computer-mediated communication

228 22:26:05 B_Pape

509

business and geography in the winter semester … so I would have to apply for geography but in business just transfer over right?’ @michael Nein leider nicht. Auch hier muss man sich bewerben. Und auch noch die Vertiefungsrichtung (BWL oder VWL) angeben. ‘@michael No unfortunately not. Here you will also have to apply. And also declare a major (Business Administration or Economics).’

Example 9: Chat advising on the topic “eBay und Online-Auktionen” (Dortmund Chat Corpus, Document 1204012) 4

12:14

5

12:14

7

12:15

bsommer Hallo und herzlich willkommen zur eBay-Onlineberatung! Ich bin Bianca Sommer und werde versuchen, Ihnen einige Fragen zum Thema eBay zu beantworten, auf die Sie schon immer eine Antwort wissen wollten. ;-) ‘Hello and welcome to eBay online advice! I am Bianca Sommer and will try to answer your questions about eBay that you have always wanted an answer for ;-)’ bsommer Welche Themen möchten Sie heute gerne besprechen? ‘Which topics would you like to talk about today?’ jag Dann hier gleich die erste Frage: Ich habe einen Artikel gekauf der defekt ist, aber als neuwertig beschrieben wurde […] ‘Then here’s the first question: I bought a defective item on eBay that was described as new […]’

When pointing to parts of the written chat history given in the screen protocol, the deictic space is constructed through the linear succession of the displayed messages, which keep scrolling upwards as new messages appear. Contrary to the sequence of the messages, one can point to previous messages with oben (‘above’) as in Examples 10 and 11 (cases of Lokutive Textdeixis, cf. Hoffmann 1997: 353–358): Example 10: Academic advising on the topic “Internationales Studium” (Dortmund Chat Corpus, Document 1202015) 180 22:18:31 Marvin

für stipendien (gerade japan ist ja wohl ziemlich teuer) müßte ich mich dann noch separat kümmern, oder gibt es da auch irgendwo ein verzeichnis mit instituionen die soetwas bieten? ‘for scholarships (japan is especially expensive) I would have to figure out on my own what to do, or is there a directory somewhere with institutions that offer something like that?’ 183 22:19:12 B_Schmid @Marvin Der DAAD hat alle Möglichkeiten in einer Broschüre zusammengefasst, die Sie bei uns kriegen können. Oder im Internet www.daad.de unter Länderinformaitonen schauen.

510

Michael Beißwenger

‘@Marvin The DAAD has put together all of the possibilities in a brochure that you can pick up in our offices. Or check on the Internet at www.daad.de and then under Country Information.’ 184 22:19:34 Marvin ah sorry mein browser hat irgendwie nicht mehr mitgespielt ‘Oh sorry my browser somehow isn’t cooperating’ 185 22:19:58 Marvin war meine frage zu stipendien jetzt angekommen? ‘did you receive my question about scholarships?’ 189 22:20:27 B_Schmid @Marvin Zu den Stipendien habe ich vorhin geschrieben – siehe oben. ‘@Marvin I just wrote about the scholarships – look above.’ Example 11: Academic advising on the topic “Studienkonten: Anrechnung von Auslandssemestern, Restguthaben, Promotion” (Dortmund Chat Corpus, Document 1202008) 443 15:10

Speedy

451 15:11

BJuhre

3.4

Aber im Schreiben steht doch, dass ein Widerspruch keine aufschiebende Wirkung hat und die Zahlungspflicht bestehen bleibt ‘But the letter says that an objection does not delay the process and the payment obligation remains’ @Speedy Lesen Sie weiter oben im Chat auch: Widerspruch ist nicht dasselbe wie ein Härtefallantrag. Widerspruch ist ganz allgemein, wenn der dann abgelehnt wird, können Sie dagegen klagen. ‘@Speedy Read also further up in the chat: An objection is not the same as an application as a hardship case. An objection is very general, if it’s denied, then you can sue.’

The Internet as a resource space

In the Examples 12–14, the deictic dort (‘there’) appears in combination with a URL. In these cases, dort does not function as a prompt for the addressees to shift their origo. Instead, it serves to point at external resources that – unlike the current focus of the interlocutors’ attention, which is fixed on following the chat – are introduced as distal resources that are not visible until the URL is accessed: Example 12: Moderated chat question time with Smudo, text writer and rapper in the German hip hop group “Die Fantastischen Vier” (Dortmund Chat Corpus, Document 1306017) 49

purp:

smudo: “Euer ‘neuestes’ Werk – F4 Unplugged – genial, wie kam es dazu?” ‘smudo: “Your ’new‘ album – F4 Unplugged – is amazing, how did it come about?”’

Space in computer-mediated communication 50

Smudo:

511

kann man alles unter www.smudo.com/smudo/unplugged_proben.htm und www.smudo.com/smudo/unplugged_show.htm anschauen und nachlesen. dort ist auch das gesamte album als 22khz-mp3 anzuhoeren ‘you can check it all out at www.smudo.com/smudo/unplugged_ proben.htm and www.smudo.com/smudo/unplugged_show.htm. there you can also listen to the entire album as a 22khz-mp3.’

Example 13: Moderated chat question time with the Swiss pop singer Francine Jordi (Dortmund Chat Corpus, Document 1306040) 140 Anna2:

Salü Francine wann bist Du wieder mal in der Innerschweiz an einem Konzert zu hören? ‘Hey Francine when will you be coming to Central Switzerland for a concert?’ 141 Francine Jordi: Besuch doch meine Homepage, dort findest Du alle Daten: www.francine-jordi.ch ‘Check out my homepage, there you can find all the information: www.francine-jordi.ch’ Example 14: Chat-based cross-university seminar on lexicography (Dortmund Chat Corpus, Document 1102004a) 224 11:06 225 11:06 226 11:06 227 11:06

Teacher1: OK, bevor ihr in die Grupenm geht die ich gleich nennen werde ‘OK, before you all go into the groups that I am about to name’ Teacher1: ruft in eurem Browser bitte die Folgende Seite / URL auf: ‘enter the following address/URL into your browser:’ Teacher1: http:// … [URL] ‘http:// … [URL]’ Teacher1: *dort* findet ihr die Aufgabe ‘*there* you will find your task’

Cases like these are based on a conceptualization of the World Wide Web as a network of resources, of which one (the one currently shown as the “page” on the screen and thus immediately accessible) depicts the current proximal subspace, while all other resources that are currently out of view lie in an indefinite distal subspace. By following the hyperlinks or opening URLs, those distal resources can be reached. 3.5

The metaphorical “chat room” and the “chat environment”

A unique feature of chat communication is the conceptualization of the mediating technical channel as a location in which people gather. In most chats on the Web, the virtual environment is metaphorically referred to as a “room”

512

Michael Beißwenger

and the communicative availability of a logged-in chat user is regarded as a form of “presence” in that respective location. The metaphorical “chat room” stands for the possibility of coming into contact with other users who are logged in at the same time. Whoever is logged in is available for communication and, thus, perceived as being close; whoever is not logged in is perceived as being absent or distant. The concept of space assigned to the metaphorical “chat room” is rudimentary in that it does not exhibit any dimensions whatsoever: Either one is drin (‘inside’, i.e. available for communicative contact) or one is draußen (‘outside’) or rather nicht drin (‘not inside’) and thus cut off from the possibility of establishing contacts. Drinsein (‘being inside’) or rather (deictic) hier-sein (‘being here’) always refers to being logged into the chat, while draußen-sein (‘being outside’) or nicht-hier-sein (‘not being here’) is not specifically determined in the same way: Whoever is draußen (‘outside’) or nicht hier (‘not here’) could either be somewhere else on the Internet or could have completely gone offline and back to ‘real life’; in any case, that person is no longer available for the ongoing communication in the chat room. The chat room represents both the metaphorical concept of space and the proximal subspace at the same time: With hier (‘here’), the chat room is referred to in its entirety while distal subspaces are not distinguished in it. Accordingly, there are no records in the Dortmund Chat Corpus that document the use of dort (‘there’) for pointing at the chat room or one’s own presence in the chat (cf. Storrer 2001: 18). When deictic dort is used to point at the virtual environment in a chat, it is usually only in cases in which the respective chat environment allows its users to organize their interactions in more than one chat room. In such environments, there is either a predefined set of chat rooms (instead only one room) or the individual users can use a special command to open up individual chat rooms for interacting with a selected group of partners. In such cases the deictic dort is used to point at the other chat rooms, while hier is always used for the chat room in which the producer of the deictic utterance is located at the time. Accordingly, the deictic opposition hier/dort in such cases constitutes the current chat room as a proximal subspace and a certain other chat room as a distal subspace; the deictic space in this case is not the individual (non-dimensional) chat room, but rather the entire chat environment as a (topological) metaphor for the superordinate communication infrastructure. Example 15 shows two uses of hier for pointing at the chat room, and Example 16 shows a sequence in which dort is used to point at another chat room in the same chat environment.

Space in computer-mediated communication

513

Example 15: Academic advising (Dortmund Chat Corpus, Document 1202016) 26

23:20

31

23:20

37

23:21

hallo zusammen ‘hi everyone’ sven: Wollte nur mal kurz spicken, was hier so läuft. ‘Just wanted to check in and see what’s going on here.’ Beraterin: @sven hier ist jetzt echt was los ‘@sven there’s a lot going on here’ sven:

Example 16: Chat-based psycho-social counseling (Dortmund Chat Corpus, Document 1201008) 763 17:08:02 Ratsuchende: ich werde so oft im chat drüben veräppelt. weißt du warum das so ist? ‘i get teased so often over there in the chat. do you know why?’ 778 17:09:48 Berater: du bist leider nicht die einzige, die sich im offenen chat nicht mehr wohlfühlt. das liegt leider an den leuten, die dort stammgäste sind. wir können daran leider zunächst wenig ändern … ‘unfortunately you’re not the only one that doesn’t feel comfortable anymore in an open chat. that has to do with the people who are regularly there. there’s little that we can do about that at the moment …’

3.6

Fictitious scenes

The metaphorical chat room acts as a code for communicative proximity. Even though the metaphor is based on a notion of space that conceives of “rooms” as locations where people gather for social encounters, the concept of the chat room has no spatial dimensions. The non-dimensionality of the chat room and the concept of proximity that is characteristic of chat encounters, however, sometimes seem to inspire chat users to playfully turn the communication environment into virtual scenes by assigning them fictional attributes. Inside these fictitious scenes which are interactively developed and successively arranged, deictic origins can be laid out, proximal and distal subspaces differentiated, deictic oppositions established and local deictics used for forms of deictic pointing at the phantasma. Example 17 show excerpts from a longer sequence in which several chat interlocutors agree to go on a cruise. McMike initiates the crossover into a playful scenario (message 50), and then, after adelheid gives the signal to engage (52: oh ja, in die Wärme … ‘oh yeah, off to where it’s warm …’), opens a new temporary chat room with the fitting name Kreuzfahrt (‘cruise’) (64). Several other interlocutors follow him into this new room, in which the con-

514

Michael Beißwenger

ditions for the cruise game are successively negotiated and the participants assign themselves activities and characteristics that all correspond to the ‘cruise’ scenario: ineli26 greets everyone with ahoi and pretends to get seasick (72), McMike gives orders to seal off the hatches (79), Matrose takes over the post of the skipper (82) and ineli26 the first officer’s position (87), Raebchen – obviously alluding to his nickname (‘little raven’) – dresses up as a seagull (85), adelheid lays out in the sun (88), and McMike, the initiator of the game, asks to be appointed captain (89), which ineli26 promptly does (94). Raebchen, consistently pretending to be the ship’s bird, declares that he is sitting im Krähennest, in the basket at the mast top (108). More than 350 messages later McMike declares – using an “action message” that, through a special command on the chat interface, allows him to attribute a fictitious action to his virtual character – that he will ascend to the top of the mast (466/468). Adelheid pinpoints her own location within the current action as being on deck (470); she identifies McMike’s anticipated new position in the fictitious scene deictically with da oben (‘up there’) and thus as being located in an area that is distal from her own current location (da) and above her own origo (oben). The same point in the imagined room at which adelheid points with da oben (‘up there’), is located by Raebchen with hier oben (‘up here’, 478). He thus pinpoints his own location, as opposed to adelheid, as (still) sitting in the crow’s nest; he expresses the distance to adelheid’s position via the complementary deictic hier, while the quasi-deictic oben is used by both in the same way.4 Example 17: Social chat (Dortmund Chat Corpus, Document 2221001) 50 52 64 65 68

wie wärs mit einer kleinen Kreuzfahrt? ‘how about a little cruise?’ adelheid oh ja, in die Wärme … ‘oh yeah, off to where it’s warm …’ McMike geht in einen anderen Raum: Kreuzfahrt ‘McMike goes to another room: Cruise’ adelheid geht in einen anderen Raum: Kreuzfahrt ‘adelheid goes to another room: Cruise’ Raebchen geht in einen anderen Raum: Kreuzfahrt ‘Raebchen goes to another room: Cruise’ McMike

[…] 71 adelheid

4

wohin geht’s denn? ‘where are we off to then?’

Referring to the same point in an imagined space from two different perspectives while changing the deictic, but retaining the quasi-deictic expression, vividly demonstrates the semantic differences between “real” deictics and quasi-deictic expressions.

Space in computer-mediated communication 72

515

hallooo *ahoi* ‘heyyyy *ahoy*’ 74 McMike wir müssen noch auf den Käptn warten ‘we have to wait for the captain’ 76 adelheid ui, wann kommt denn der? ‘well, when’s he coming?’ 77 ineli26 *schwank* … *seekrank* ‘*sways* … *seasick*’ 79 McMike Luuuuuukeeeen DICHT! ‘Haaaaaaatcheeeeeees CLOSED!’ 82 Matrose damit das klar ist: ich bin der skipper ‘just so everyone knows: i am the skipper’ 85 Raebchen *mövenkostüm anzieh* ‘*putting on seagull costume*’ 87 ineli26 will erste offizierin sein ‘wants to be first officer’ 88 adelheid liegt lieber in der Sonne … ‘prefers laying in the sun …’ 89 McMike könntet Ihr mich bitte zum Käpten ernennen? ‘would you all please appoint me captain?’ 94 ineli26 ernennt McMike zum Kapitaen ‘ineli26 appoints McMike captain’ 97 McMike Danke, dann können wir ja los ‘Thank you, now we can set sail’ 108 Raebchen sitzt im Krähennest *gg* ‘Raebchen is sitting in the crow’s nest *gg*’ […] 466 McMike klettert auf den Mast ‘McMike climbs up the mast 468 McMike *kletterkletter* ‘*climbclimb*’ 470 adelheid McMike, was machst Du da oben??? ‘McMike, what are you doing up there???’ 478 Raebchen ganz schön eng jetzt hier oben ‘pretty tight now up here’ ineli26

A coherent linguistic co-orientation on points and areas in the fictitious space of action in playful sequences like the one in Example 17 highly depends on cooperation. Each new utterance that is designed to contribute to the negotiation of further properties of the fictitious scene is also a ratification of the unbroken individual compliance with the game. Further examples of chat sequences in fictitious scenes are given and analyzed in Beißwenger (2000: 183–200) and Beißwenger (2001). Similar types of sequences have also recently been described for the German Usenet (Bücker 2010: 67).

516

4. 4.1

Michael Beißwenger

A corpus-based study on the use of hier (‘here’) and dort (‘there’) in German chat communication Hypotheses

Despite the variety of possible spatial concepts for deictic spaces in chat communication and despite the fact that interlocutors in chat communicate via distance and perceive each other’s activities output non-simultaneously and only through writing, the use and interpretation of local deictics does not seem to present chat users with any serious problems. At least in the Dortmund Chat Corpus there is hardly any substantial evidence of problems conveying the meaning of local deictics. This leads to the assumption that chat users either always make the relevant deictic space explicit when using a local deictic, or they have strategies for identifying the relevant deictic space even in cases in which the respective spatial concept is not made explicit by the producer of the deictic utterance. Storrer (2001: 18) notes with regard to communication in social chats that “normally […] the metaphorically constituted chat room (serves) as a deictic space”. Following this idea, one can formulate the following hypothesis for social chats: The concept of the metaphorical chat room serves as a default deictic space which is applied to interpret local deictics whenever the deictic space relevant for the comprehension of the local deictic is not made explicit by the producer and cannot be disambiguated through the context (hypothesis 1). In addition, one could assume (1) that in other types of chats (chats in institutional contexts and with a restrictive interaction management, such as moderation) the users also have standard assumptions about what spatial concept usually functions as deictic space in cases in which the relevant deictic space is not specified by the producer or the context and (2) that these ‘standard spaces’ differ for different types of chats (hypothesis 2). 4.2

Corpus and methods

In order to test these hypotheses, a compilation of subcorpora from the Dortmund Chat Corpus has been analyzed. The Dortmund Chat Corpus was collected and built between 2002 and 2008 at the Chair for German Linguistics and Language Didactics at the University of Dortmund. It comprises 478 log files with a total of 140,240 messages or 1.06 million tokens of chat communication from a broad range of application contexts (social chats as well as chats from e-learning contexts, from institutional advising and from media contexts). The corpus is available and documented at http://www.chatkorpus. tu-dortmund.de.

Space in computer-mediated communication

517

The corpus which was compiled for the study (cf. Table 1) comprises approximately 74 % of the total data of the Dortmund Chat Corpus (in tokens). The first subset (“Free Chat”) comprises subcorpora with chats which were not limited to a certain user group, and neither the interaction management nor the topic development were restricted through institutional regulation (in contrast to moderated chats). The prototypical type of chat in this corpus subset are the so-called “social chats” in which users meet for chitchat, making new friends and having fun.5 Other types of chat which are quite similar to these social chats with respect to their organization and the development of topics are open chat-based offerings for psycho-social counseling6, open academic advising chats7 and open user chats following moderated chat sessions with studio guests from a television show on political and economical topics. Even though the latter three types of chats had predefined topics, the topic development was not restricted through moderation and users could also digress from the defined topic. The second subset (“Institutional and Moderated Chat”) comprises a range of subcorpora with data from chats in which the communication is strongly determined by the institution responsible for the chat offer: some feature hierarchical constellations (e.g. teacher–student, information desk– client) as given by the real institutions (e.g. a university seminar, a library), while others are restrictive, in part technically supported forms of interactional management. In the chat-based library information on demand, the communication is limited to purely 1:1 chats and accessible for only one user at a time.8 In the chat-based university seminar, access to the chat environment was granted only to students who had officially enrolled in the course.9 The politician/celebrity chats in media contexts10 were moderated by using a technical protocol that allowed the moderators to control and select the incoming user messages. The chats with invited experts in e-learning seminars were moderated on the basis of a predefined set of conversational rules for 5 6

7

8

9

10

Excerpts from social chats are given in Examples 4, 5 and 17 in section 3. An excerpt from one of the open psycho-social counseling chats is given in Example 16 in section 3. Excerpts from open academic advising chats are given in Examples 6, 8, 10, 11 and 15 in section 3. An excerpt from one of the library information chats is given in Example 3 in section 3. Two excerpts from the chat-based university seminar are given in Examples 7 and 14 in section 3. Three excerpts from moderated celebrity chats are given in Example 1 in section 2 and in Examples 12 and 13 in section 3.

518

Michael Beißwenger

‘taking the turn’ and a moderator role assigning users permission to send new messages. Both subsets comprise of a total of 66,592 user messages. Subcorpora

Messages

Corpus Subset 1 “Free Chat”: SC1.1: “Social chats” outside of media context SC1.2: Open psycho-social counseling SC1.3: Open academic advising chats SC1.4: Open user chats following the TV show “Sabine Christiansen” Subset “Free Chat” total:

46,443

Corpus Subset 2 “Institutional and Moderated Chat”: SC2.1: Moderated politician/celebrity chats SC2.2: Moderated expert chats in e-learning context SC2.3: Chat-based university seminar SC2.4: Chat-based 1:1 library information on demand Subset “Institutional and Moderated Chat” total:

13,314 1,410 3,150 2,275 20,149

Subset 1 + 2 total:

66,592

27,101 9,207 5,771 4,364

Table 1. The corpus used for the investigation (compiled from the Dortmund Chat Corpus, http://www.chatkorpus.tu-dortmund.de); SC = subcorpus

In order to test the hypotheses formulated above, all occurrences of the local deictics hier (‘here’) and dort (‘there’) in the corpus have been identified and counted, and subsequently divided into the two corpus subsets “Free Chat” and “Institutional and Moderated Chat”. In a first step of qualitative analysis each occurrence of hier and dort was assigned to one of the various spatial concepts described in section 3. This step was done through a qualitative analysis of the context for every corpus record. For this purpose, for each record an excerpt of 21 messages was generated from the message sequence in the respective chat log file; the 21-message excerpt comprises the message with the record together with the ten previous and ten subsequent messages. In a second step and also on the basis of the 21-message excerpt, it was determined for each instance of hier and dort whether or not the relevant deictic space was explicitly specified by one of the interlocutors in the deictic message itself or in the context of ten messages before/after the deictic message. If there was no explicit specification of the deictic space (or of a target point/area within the deictic space) in the 21-message excerpt, then the respective record was classified as “deictic space not explicitly specified in the context”.

Space in computer-mediated communication

519

The assumption was that the chat users regard the spatial concept used most often as the deictic space in a particular type of chat to be the default deictic space for interpreting local deictics in that particular type of chat. Thus, the first hypothesis can be considered verified for a particular type of chat (a) when one of the spatial concepts is used as the basis for deictic pointing significantly more often than all other spatial concepts in the respective subcorpus and (b) if the occurrences of hier and dort in the respective subcorpus occur less often together with an explicit specification of the deictic space in the context than those occurrences which refer to other types of deictic spaces. The second hypothesis can be considered verified when (a) the first hypothesis is verified and (b) the findings differ significantly for the two corpus subsets. 4.3

Findings

In the corpus, there were a total of 1,132 occurrences of the deictic hier and 340 occurrences of the deictic dort. Table 2 shows the results for the classification of all 1,132 occurrences of hier distributed among both corpus subsets and classified according to types of deictic spaces (DS). The distribution of the occurrences among the DS types is shown for each subset as an absolute value as well as a percentage of all occurrences in the respective subset (example: subset 1 comprises a total of 879 occurrences of hier; 170 or 19.3 % of those are allotted to DS type ‘environments in the real world’). The last column in the table gives the statistical significance for the differing distribution of occurrences for each DS type in both subsets. The significance was calculated using the x2 test. The result of the classification of the occurrences of hier according to DS types and corpus subsets shows that within the subset “Free Chat” the metaphorical chat room is by far most frequently used as deictic space (65.6 % of all occurrences in the subcorpus). The second most frequent deictic space, “environments in the real world”, comprises 19.3 %; each of the remaining DS types comprises 5 % of occurrences at most. In subset 2 with the institutional/moderated chats the distribution is not quite as clear. Here there are three types of DS that are applied more or less equally often when using the deictic hier: the metaphorical chat room (29.2 %), environments in the real world (26.9 %) and the discourse space (24.9 %). When pointing in a discourse space (DS type B1) the discourse object at which one is deictically pointing is always explicitly named and necessarily close by in the context of the utterance. For the seven other DS types (A1–A4 and B2–B4) the analysis reveals that in the subset “Free Chat” for

520

Michael Beißwenger Corpus Occurrences of subset hier in subset 1 “Free Chat”

Type of deictic space (DS)

Occurrences of hier in subset 2 “Institutional/ Moderated Chat”

Significance (x2)

ALL TYPES:

879 (100 %)

253 (100 %)

DSA 1 = Environments in the real world

170 (19.3 %)

68 (26.9 %)

DSA 2 = Interfaces of objects

8 (0.9 %)

5 (2.0 %)

1.97 (n.s.)

DSA 3 = Interfaces on the screen

2 (0.2 %)

3 (1.2 %)

4.10 (p = 0.05)

DSA 4 = The chat log file

4 (0.5 %)

1 (0.4 %)

0.02 (n.s.)

DSB 1 = The discourse space

46 (5.2 %)

63 (24.9 %)

DSB 2 = The Internet as a resource space

9 (1.0 %)

1 (0.4 %)

DSB 3 = The metaphorical chat room

577 (65.6 %)

74 (29.2 %)

DSB 4 = Fictitious scenes

18 (2.0 %)

Other DS

24 (2.7 %)

32 (12.6 %)

–

Non-conclusive

19 (2.2 %)

5 (2.0 %)

–

0 (0 %)

– 6.72 (p = 0.01)

87.33 (p = 0.001) 0.89 (n.s.)

106.48 (p = 0.001)

5.26 (p = 0.05)

Table 2. Occurrences of hier in corpus subsets 1 and 2, classified according to underlying deictic spaces; DS = deictic space (A1, A2, …, B4 refer to the overview of the types of deictic spaces from the introduction of section 3); n.s. = non-significant distribution

19.1 % (or 151 out of 790) of all occurrences of hier the relevant deictic space is explicitly specified in the context. In subset 2 “Institutional/Moderated Chat” the same holds for 41.5 % (or 63 out of 152) of all occurrences. A detailed overview of the analysis of the context excerpts for the occurrences of hier is given in Table 3: When interlocutors are pointing in the metaphorical chat room, the deictic space is explicitly specified in 10.8 % of the occurrences in the free chats, while in the institutional/moderated chats this is the case for

521

Space in computer-mediated communication

31.1 % of the occurrences. In both subsets the metaphorical chat room is thus explicitly specified less frequently than average when using it as a deictic space. In contrast, when pointing at environments in the real world one finds aboveaverage proportions of explicit specifications of the deictic space: in 39.4 % of the free chats (with an average of 19.1 %) and in 75.0 % of the institutional/ moderated chats (with an average of 41.5 %). The differences between the two subsets – the entirety of the DS types and the two spatial concepts which are most frequently used as deictic spaces – are significant. Corpus Occurrences of subset hier in subset 1 “Free Chat” DS explicit in context Total of all cases, 151 of DSA 1 – A 4 , B 2 – B 4 790

Occurrences of hier in subset 2 “Institutional/ Moderated Chat”

Significance (x2)

(19.1 %) 63 of 152 (41.5 %) 36.21 (p = 0.0001)

Cases for DSA1 67 of 170 (39.4 %) 51 of 68 (75.0 %) 24.61 (p = 0.0001) (real environments) Cases for DSB3 (metaphorical “chat room”)

62 of 577 (10.8 %) 23 of 74 (31.1 %) 23.89 (p = 0.0001)

Table 3. Occurrences of hier in the corpus in which the relevant deictic space (DS) or the target of deictic pointing is explicitly specified in the contextual excerpt (ten messages before or after the message with the deictic). Row 1 gives the total distribution for all occurrences (not differentiated according to DS type and without consideration of cases with DSB1 “discourse space”); rows 2 and 3 give the distribution for the two deictic spaces which are most frequently applied when using hier (DSA1 “real environments” and DSB3 “metaphorical chat room”). The cases assigned to the categories “other” and “non-conclusive” in Table 2 were not included in this evaluation.

The synopsis of the results portrayed in Tables 2 and 3 shows that in free chats the deictic space which is applied the most (the metaphorical chat room) is made explicit significantly less often than the average of occurrences for all types of deictic spaces (p = 0.0001). In the case of the institutional/moderated chats, the metaphorical chat room is also explicitly specified less often than average, but much more often than in the free chats, while real environments in subset 2 were actually explicitly specified in 75 % of all of the cases. Since in the chats documented in subset 2 three types of DS are drawn upon as deictic spaces at about the same rate, it can be assumed that the risk of ambiguity in these types of chats is higher than in the free chats, and accordingly the inter-

522

Michael Beißwenger

locutors explicitly specify the relevant spatial concepts for their uses of deictics more often. On the other hand, in the free chats it seems that dealing with hier is less problematic: In this situation, the metaphorical chat room is applied as a deictic space in 65.6 % of all cases (cf. Table 2). Accordingly, there are proportionally much fewer occurrences of hier with explicit specification of the deictic space than in the institutional/moderated chats; at the same time, the type of deictic space that is drawn upon most often (DSB3, 577 occurences) is explicitly specified fewer times than average (in 10.8 % of all cases). Corpus Occurrences of subset dort in subset 1 “Free Chat”

Type of deictic space (DS)

Occurrences of dort in subset 2 “Institutional/ Moderated Chat”

Significance (x2)

ALL TYPES:

178 (100 %)

162 (100 %)

–

DSA 1 = Environments in the real world

125 (70.2 %)

123 (75.9 %)

1.40 (n.s.)

DSA 2 = Interfaces of objects

2 (1.1 %)

3 (1.9 %)

0.31 (n.s.)

DSA 3 = Interfaces on the screen

12 (6.7 %)

12 (7.4 %)

0.06 (n.s.)

DSA 4 = The chat log file

1 (0.6 %)

0 (0 %)

0.91 (n.s.)

DSB 1 = The discourse space

4 (2.2 %)

8 (4.9 %)

1.80 (n.s.)

DSB 2 = The Internet as a resource space

17 (9.6 %)

14 (8.6 %)

0.08 (n.s.)

DSB 3 = The metaphorical chat room

12 (6.7 %)

0 (0 %)

DSB 4 = Fictitious scenes

0 (0 %)

0 (0 %)

–

Other DS

1 (0.6 %)

2 (1.2 %)

–

Non-conclusive

4 (0 %)

0 (0 %)

–

11.32 (p = 0.001)

Table 4. Occurrences of dort in corpus subsets 1 and 2, classified according to underlying deictic spaces; DS = deictic space (A1, A2, …, B4 refer to the overview of the types of deictic spaces from the introduction of section 3); n.s. = non-significant distribution.

523

Space in computer-mediated communication Corpus Occurrences of subset dort in subset 1 “Free Chat” DS explicit in context Total of all cases, 126 of 169 DSA 1 – A 4 , B 2 – B 4

Significance Occurrences of (x2) dort in subset 2 “Institutional/ Moderated Chat”

(74.6 %) 133 of 152

(87.5 %)

8.60

(p = 0.01)

Cases for DSA1 (= real environments)

88 of 125 (70.4 %) 109 of 123

(88.6 %) 12.60

(p = 0.001)

Cases for DSB3 (= metaphorical ‘chat room’)

10 of 12 (83.3 %) 0 of 0

(0.0 %)

–

Table 5. Occurrences of dort in the corpus, in which the relevant deictic space (DS) or the target of deictic pointing is explicitly specified in the contextual excerpt (ten messages before or after the message with the deictic). Row 1 gives the total distribution for all occurrences (not differentiated according to DS types and without consideration of cases with DSB1 “discourse space”); rows 2 and 3 give the distribution for DSA1 “real environments” and DSB3 “metaphorical chat room”. The cases assigned to the categories “other” and “non-conclusive” in Table 4 were not included in this evaluation.

In the case of the uses of dort, a different picture is presented: 70.2 % and 75.9 % of all occurrences are allotted to pointing at environments in the real world; other deictic spaces play a much smaller role. The distribution in the free chats and in the institutional/moderated chats is to a large extent similar; there are hardly any significant differences. Due to the lack of dimensional properties in the corresponding spatial concept (cf. section 3.5), the metaphorical chat room plays only a marginal role in the use of dort. All in all, the occurrences of dort are accompanied much more often by an explicit specification of the relevant deictic space than is the case with hier (Table 5). Table 6 contrasts the corresponding percentages for hier and dort in both subsets once more: For both deictics, the rate of occurrences with explicit specification of the deictic space in the context is higher in the institutional/moderated chats than in the free chats. In the case of dort the rate is overall significantly higher than in the case of hier: dort is specified almost four times more often in the context of free chats than hier, in the institutional/moderated chats more than twice as often.

524

Corpus subset

Michael Beißwenger Deictic hier: DS explicitly specified (percentage of cases) [f Tab. 3]

Subset 1 “Free Chat”

151 of 790

63 of Subset 2 152 “Institutional/ Moderated Chat”

Significance dort: DS explicitly specified (x2) (percentage of cases) [f Tab. 5]

(19.1 %) 126 of 169

(74.6 %) 208.33 (p = 0.0001)

(41.5 %) 133 of 152

(87.5 %) 70.37

(p = 0.0001)

Table 6. Comparison of occurrences of hier and dort with explicit specification of the deictic space in the context

The results of the corpus investigation can be assessed as follows: At least for uses of the local deictic hier in non-moderated “free” chats, hypothesis 1 can be confirmed: Among the possible spatial concepts there is one that is applied very frequently – that of the metaphorical chat room. The use of this concept as a deictic space is made explicit much less often in the context than the average of all the cases of the use of hier. In contrast, in chats with a strong institutional determination, the concept of the metaphorical chat room is far less dominant than in non-moderated “free” chats. Indeed, cases of pointing at the metaphorical chat room are still made explicit in the context less often than average; however, real environments as well as the discourse space are applied as deictic spaces just as often as the chat room. Overall, the percentage of occurrences with an explicit specification of the relevant deictic space is higher in subset 2 than in subset 1 (41.5 % as opposed to 19.1 % of the cases): Where the risk of ambiguous interpretations is higher, the producers of deictic utterances more often tend to give their addressees explicit linguistic information for the sake of disambiguation. Even though its potential as the default space is obviously much weaker than in the free chats (subset 1), the metaphorical chat room is the type of deictic space that is used often without an explicit specification even in institutional/moderated chats (subset 2); accordingly, hypothesis 2 is falsified: In both corpus subsets there are no differing default deictic spaces, but rather one and the same spatial concept plays the default role in both groups, albeit with differing efficacy. For the uses of the local deictic dort in both subsets, environments in the real world are the type of space which is most frequently used as deictic spaces. However, considering the high proportion of specifications of x

x

x

x

Space in computer-mediated communication

525

deictic spaces in the context (74.6 % in subset 1, 87.5 % in subset 2), it can hardly be deemed a default deictic space. If one considers only the cases in which dort is used to point at real environments, then the proportion of specifications of the deictic space in the context for subset 1 is only slightly lower than average, for subset 2 even slightly above average. Accordingly, for dort both hypotheses must be rejected. The reason may have to do with the specific semantics of dort which requires a higher degree of spatial knowledge than in the case of hier: In order to adequately interpret uses of dort, not only the relevant type of deictic space but also the hierarea within it as well as the delimitation between the proximal and the distal deictic subspaces must be known or at least deducible. Dort is always there where hier is not; thus, for dealing with dort, especially the delimitation problem and the problem of deictic oppositions become relevant (cf. Klein’s (1978) differentiation of challenges for intepreting local deictics; section 2).

5.

Conclusion and outlook

The coordination efforts with which interlocutors are faced when dealing with local deictics seem to be more challenging in chat communication than in other forms of synchronous communication. In contrast to telephone conversations or audio conferences, the special technological framework complicates the possibility to track and quickly adjust to the partners’ behavior and verbal output. Unlike video conferences, the interlocutors cannot use nonverbal behavior (gesture, facial expressions, gaze) to support und disambiguate deictic pointing. In contrast to communication outside the Web, the use of computers and the network infrastructure of the Internet as well as the conceptualization of chat-based interactions as encounters in “virtual spaces” yield additional topological concepts which can be applied as deictic spaces and thus as the basis for linguistic pointing. Some of the spatial concepts that are applied as deictic spaces in chats are directly perceivable for each interlocutor: the real location around him or her, the physical interface of his or her computer workspace (monitor, keyboard, mouse, etc.), the webpage and the chat interface shown on the screen as well as the chat protocol displayed in an area of the chat interface. Further topological structures have to be imaginatively construed so that they can be used as a basis for deictic pointing as well: the discourse space, the resource space ‘Internet’, the ‘chat room’ as a metaphor for the technologically mediated sphere of contact as well as the fictitious scene of playful episodes which are developed and successively arranged through interactive negoti-

526

Michael Beißwenger

ation. Furthermore, the real environments of the other interlocutors have to be imaginatively construed as well as webpages or objects on webpages which are still unknown to the recipient of a deictic utterance. The only spaces which can be assumed to be directly accessible to the perception of all interlocutors are the continuing protocol of chat messages on the screen and the website with the chat interface in which it is embedded. Despite the variety of topological concepts that can be applied as deictic spaces for linguistic pointing, chat users generally have few problems identifying the deictic space which is relevant for the use of a local deictic – if necessary even without having explicitly established the relevant space beforehand. The results of the corpus investigation described in section 4 suggest that, at least for hier in free, non-moderated chats, the chat users have a default assumption which determines the interpretation of occurrences of hier in cases where no other type of deictic space is explicitly specified. This default deictic space is the metaphorical chat room. In other types of chats, the chat room seems to be less dominant – even though still serving as default space in cases where no other spatial concept is explicitly specified as relevant. The results give reason to assume that the system of local deictics can be an efficient device for linguistic localization even in synchronous written communication via the Internet. Default assumptions constitute a strategy to solve referential problems related to the use of local deictics and increased by the communicative framework; a high level of explicitness represents an alternative strategy. When dealing with local deictics in chat – at least for the deictic hier – both strategies are combined: a default assumption for unmarked cases in which the relevant deictic space is not explicitly specified, and explicitness for cases which deviate from the default space. Future investigations into the use of local deictics in other modes of Internet-based communication can contribute to a further understanding of the flexibility of local deixis and its adaptability to new communicative frameworks evolving within and across technically mediated settings.

References Auer, Peter 2000: On line-Syntax – oder: was es bedeuten könnte, die Zeitlichkeit der mündlichen Sprache ernst zu nehmen. Sprache und Literatur 85: 43–56. Beißwenger, Michael 2000: Kommunikation in virtuellen Welten: Sprache, Text und Wirklichkeit. Stuttgart: ibidem. Beißwenger, Michael 2001: Das interaktive Lesespiel. Chat-Kommunikation als mediale Inszenierung. In: Michael Beißwenger (ed.), Chat-Kommunikation. Sprache, Interaktion, Sozialität & Identität in synchroner computervermittelter Kommunikation. Perspektiven auf ein interdisziplinäres Forschungsfeld, 79–138. Stuttgart: ibidem.

Space in computer-mediated communication

527

Beißwenger, Michael 2003: Sprachhandlungskoordination im Chat. Zeitschrift für Germanistische Linguistik 31 (2): 198–231. Beißwenger, Michael 2007: Sprachhandlungskoordination in der Chat-Kommunikation. (Linguistik – Impulse & Tendenzen 26.) Berlin/New York: de Gruyter. Beißwenger, Michael 2010: Chattern unter die Finger geschaut: Formulieren und Revidieren bei der schriftlichen Verbalisierung in synchroner internetbasierter Kommunikation. In: Vilmos Ágel & Mathilde Hennig (eds.), Nähe und Distanz im Kontext variationslinguistischer Forschung, 247–294. (Linguistik – Impulse & Tendenzen 35.) Berlin/New York: de Gruyter. Bücker, Jörg 2010: Sprachhandeln und Sprachwissen. Grammatische Konstruktionen in der kommunikativen Praxis. Ph.D. dissertation, Westfälische WilhelmsUniversität Münster. Bühler, Karl 1934 [1982]: Sprachtheorie. Die Darstellungsfunktion der Sprache. Neudruck mit einem Geleitwort von Friedrich Kainz. (Nachdruck der Originalausgabe Jena 1934.) Stuttgart/New York: Gustav Fischer. Cherny, Lynn 1999: Conversation and Community. Chat in a Virtual World. (CSLI Lecture Notes 94.) Stanford. Ehlich, Konrad 1983 [2007]: Deixis und Anapher. In: Konrad Ehlich (ed.), Sprache und sprachliches Handeln. Band 2: Prozeduren des sprachlichen Handelns, 5–24. Berlin/New York: de Gruyter. Ehlich, Konrad 1983a: Text und sprachliches Handeln. Die Entstehung von Texten aus dem Bedürfnis nach Überlieferung. In: Aleida Assmann, Jan Assmann & Christof Hardmeier (eds.), Schrift und Gedächtnis. Archäologie der literarischen Kommunikation I, 24–43. München: Wilhelm Fink. Ehlich, Konrad 2007: Anadeixis und Anapher. In: Konrad Ehlich (ed.), Sprache und sprachliches Handeln. Band 2: Prozeduren des sprachlichen Handelns, 25–44. Berlin/New York: de Gruyter. Garcia, Angela Cora & Jennifer Baker Jacobs 1998: The interactional organization of computer mediated communication in the college classroom. Qualitative Sociology 21 (3): 299–317. Garcia, Angela Cora & Jennifer Baker Jacobs 1999: The eyes of the beholder: understanding the turn-taking system in quasi-synchronous computer-mediated communication. Research on Language and Social Interaction 32 (4): 337–367. Haase, Martin, Michael Huber, Alexander Krumeich & Georg Rehm 1997: Internetkommunikation und Sprachwandel. In: Rüdiger Weingarten (ed.), Sprachwandel durch Computer, 51–85. Opladen: Westdeutscher Verlag. Herring, Susan C. 1999: Interactional coherence in CMC. Journal of Computer-Mediated Communication 4 (4). http://jcmc.indiana.edu/vol4/issue4/herring.html. Hoffmann, Ludger 1997: Deixis und situative Orientierung. In: Gisela Zifonun, Ludger Hoffmann & Bruno Strecker (eds.), Grammatik der deutschen Sprache (Bd. 1), 310–359. (Schriften des Instituts für deutsche Sprache 7.1.) Berlin/New York: de Gruyter. Hoffmann, Ludger 2007: Adverb. In: Ludger Hoffmann (ed.), Handbuch der deutschen Wortarten, 223–264. Berlin/New York: de Gruyter. Kameyama, Shinichi 2007: Persondeixis, Objektdeixis. In: Ludger Hoffmann (ed.), Handbuch der deutschen Wortarten, 577–600. Berlin. New York: de Gruyter. Klein, Wolfgang 1978: Wo ist hier? Präliminarien zu einer Untersuchung der lokalen Deixis. Linguistische Berichte 58: 18–40.

528

Michael Beißwenger

Klein, Wolfgang 1979: Wegauskünfte. Zeitschrift für Literaturwissenschaft und Linguistik 33: 9–57. Markman, Kris 2006: Computer-mediated Conversation: The Organization of Talk in Chat-based Virtual Team Meetings. Ph.D. dissertation, University Texas at Austin. Murray, Denise E. 1989: When the medium determines turns: turn-taking in computer conversation. In: Hywel Coleman (ed.), Working with Language. A Multidisciplinary onsideration of Language Use in Work Contexts, 319–337. (Contributions to the Sociology of Languages 52.) Berlin/New York: Mouton de Gruyter. Schönfeldt, Juliane & Andrea Golato 2003: Repair in chats: a conversation analytic approach. Research on Language and Social Interaction 36 (3): 241–284. Sitta, Georg 1991: Deixis am Phantasma. Versuch einer Neubestimmung. (Bochumer Beiträge zur Semiotik 31.) Bochum: Brockmeyer. Storrer, Angelika 2001: Sprachliche Besonderheiten getippter Gespräche: Sprecherwechsel und sprachliches Zeigen in der Chat-Kommunikation. In: Michael Beißwenger (ed.), Chat-Kommunikation. Sprache, Interaktion, Sozialität & Identität in synchroner computervermittelter Kommunikation. Perspektiven auf ein interdisziplinäres Forschungsfeld, 3–24. Stuttgart: ibidem. Zitzen, Michaela & Dieter Stein 2005: Chat and conversation: a case of transmedial stability? Linguistics 42 (5): 983–1021.

Vernacular and multilingual writing in mediated spaces

529

Christian Mair & Stefan Pfänder *

Vernacular and multilingual writing in mediated spaces Web-forums for post-colonial communities of practice

1.

Introduction: Research on computer-mediated communication as part of a sociolinguistics of globalisation

By and large, research on computer-mediated communication (CMC) has developed along two different lines of enquiry. A first research tradition, which we would like to refer to as the discourse-pragmatic approach, has focussed on how language structure and discourse conventions are shaped by the new media. Topics investigated include, among other things, compression devices in text messages, the use of emoticons and, more generally, stylistic (in)formality and the position of the new textual genres of CMC vis-à-vis the traditional distinction between speech and writing (cf., e.g., Crystal 2006 or Baron 2008). Discourse-pragmatic analyses of CMC note the presence of informal, and even regional or dialectal, linguistic features in computer-mediated writing, but generally do not follow this up with systematic quantitative analyses of variation. Similarly, code-switching and multilingualism (cf. Danet & Herring 2007) are not very prominent features in empirical data that is typically obtained from Anglophone student populations. These questions, however, move to centre stage in the second tradition, the sociolinguistic approach, which in comparison to the more firmly established discourse-pragmatic one, is still in its early stages (cf., e.g., the papers collected in Androutsopoulos, ed., 2006, a special issue of the International Journal of Sociolinguistics). Taking the cue from constructivist sociolinguistics (e.g. Coupland 2007, 2009; Eckert 2008), research in this tradition is particularly interested in the ways specific social groups use their real-world vernacular and multilingual resources in order to “perform” online identities. In the framework of a joint FRIAS-based project on post-colonial language spread and diversification (“World languages – * The present paper was conceived while the two authors enjoyed the extremely productive and congenial working environment provided by FRIAS, Freiburg University’s Institute for Advanced Studies. It was completed in the framework of a joint project carried out by the two authors and sponsored by Deutsche Forschungsgemeinschaft (DFG grant MA 1652/9, to Mair; DFG PF 699/4, to Pfänder). They are grateful for this support.

530

Christian Mair & Stefan Pfänder

digital languages: Digital monitoring of ongoing change and diversification in English, French and Spanish”), the two authors of the present contribution early on became aware of the importance of new communication technologies in accounting for current trends in the global spread of the ex-colonial languages they were studying. Trivially, the Web has facilitated obtaining large amounts of data documenting the standard varieties of the languages investigated (e.g. online newspaper archives, radio and television news coverage, etc.). More importantly, though, the new media have also raised the public profile of many non-standard and vernacular varieties which previously tended to be “invisible” in the public and media domains, in some instances encouraging the emergence of new communicative styles and opening up new channels for the diffusion of vernacular linguistic resources. Face-to-face interaction in the local speech community continues to be the home base and site of origin of vernacular linguistic usage, but it is clearly no longer the only communicative domain in which vernacular usage plays a significant role – a new challenge which sociolinguists have increasingly responded to over the past decade. Thus, Coupland (2010b: 73) has pointed to: several sociolinguistic assumptions that have remained largely unchallenged since the early years of the discipline, particularly assumptions relating to a fixed meaningful class order, operating through a relatively isolated and intact national framework, where linguistic indexicalities are formed and maintained in warm-bodied social exchanges but under the ideological control of dominant social groups. My conclusion is that this is, nowadays, an account in need of revision.

What has become problematical, in other words, is the link between the vernacular and its territorial base in the community. More and more vernaculars are on the move – be it physically, through currents of migration or diaspora-formation, or metaphorically, through use in CMC or in mediated performances; and this of course challenges traditional notions of sociolinguistic authenticity. The mobile and technologised vernaculars will not remain the same as their territorial correlates for long, but this does not mean that they are necessarily less authentic. What the (socio)linguistics of cyber-space needs to describe is how the new media promote a new mix of vernacular resources and new orders of indexicality which build on the traditional ones but modify and extend them in response to the new contexts in which the vernaculars have come to be used. It is our aim in the present paper to explore the nature and potential of the new vernacular spaces which have been opened up in CMC. In the discourse-pragmatic approach to CMC, space only rarely emerges as problematical, such as when systems of local deixis have to be adapted

Vernacular and multilingual writing in mediated spaces

531

to a new communicative constellation. Such problems are certainly manageable when compared to the profound transformations which vernacular linguistic resources undergo in the CMC environment and which challenge central sociolinguistic notions of community and prestige. Central to an understanding of the new vernacular spaces which have arisen as a result of the domestication of CMC technologies in everyday life is a considerable weakening of the ties between the vernacular, its speech community which is grounded in face-to-face interaction and its usually welldefined territorial home-base. Varieties which historically emerged as regional dialects and until recently survived as vernacular languages of closely circumscribed communities even in diasporic contexts are now being massively “displaced” or “de-territorialised”. Physically, they are displaced because migration and the diaspora may make them accessible globally. Metaphorically, deterritorialisation in CMC and the concomitant detachment from the local community of speakers makes these non-standard resources “free-floating”: implicated in new types of language contact not possible in geographical space, available for crossing into (Rampton 1995, Heller 2006 [1999]) by various outgroups, and in many cases also subject to processes of commodification (Heller 2003), for example if features of a particular variety are associated with globally dispersed sub-cultural practices. Our joint work intends to integrate previous traditions of research on CMC and develop them into an essential component of an emerging new sub-field, the sociolinguistics of globalisation (Blommaert 2010; Coupland, ed. 2010). This is the appropriate frame to handle the tension between a new technology (which allows apparently instant and unbounded deterritorialisation of many vernacular usages) and speakers’ communicative intentions (which remain guided by experience of what these resources mean in the real world). The tension commonly results in a desire to re-localise them in new contexts. As Blommaert (2010: 79) puts it: We need to gain a clear sight of what locality means if we want to understand globalisation. I am not happy with views of globalisation in which the local is only seen as the stable and traditional.

Even if English is the globally dominant language in many ways today, we would like to emphasise that the sociolinguistics of globalisation which we envisage is multilingual. Where English is concerned, it will do full justice to the fact that most varieties of English develop in multilingual settings. Where French and Spanish are concerned, the digital revolution has helped their spread probably even more than English.

532

Christian Mair & Stefan Pfänder

What data from the New Englishes and the overseas varieties of the Romance languages have in common is that they both provide compelling evidence that linguistic and cultural awareness is likely to be particularly keen in post-colonial contexts, where traditional sociolinguistic orders started disintegrating dramatically from the mid to late twentieth century onwards. In this respect, what Bhatt (2010: 520) has said about English is no less true of French in Africa or Spanish in Latin America: One of the defining features of globalisation is the increasingly complex and multifaceted interactions of localism and globalism. The post-colonial contexts present us with a vibrant site where local linguistic forms – inflected by the nexus of activities taking place elsewhere in time and space – are constantly transforming in response to asymmetric exchanges, pluralized histories, power plays, and battles over polysemous signs. The transformation makes available a semiotic space where a repertoire of identities evolves in the inter-animation of the colonial-global and of the indigenous local.

As we will argue on the basis of several case studies, the mediated spaces created by diasporic web forums are particularly poignant examples of the semiotic dynamic Bhatt has in mind. They show mobilisation of vernacular linguistic resources on a massive scale, the emergence of new vernacular styles, new opportunities for contact between vernaculars (including types of contact without equivalent in the “real world” of face-to-face interaction) and, in rare instances, possibly even media-induced structural innovation.

2.

The database: A large multilingual corpus of diasporic web-forum communication

To put the “World Languages – Digital Languages” project on a solid empirical footing, a very large multilingual corpus of diasporic web-forum communication was created by downloading relevant material from the Web in 2008 and processing it for use as a linguistic corpus. As the number and variety of forums which could have been sampled was almost boundless, the selection of donor forums was to some extent determined by subjective factors, such as investigators’ research priorities at the time. This element of arbitrariness notwithstanding, the over-all collection has proved a rich and diverse database which probably exemplifies all important trends in this domain in one way or another. From a World Englishes perspective, Jamaica and Nigeria were chosen as focal regions because both nations are the homelands of far-flung global diasporas. Jamaican/ Caribbean English and Jamaican Creole have additionally gained global media exposure through a variety

533

Vernacular and multilingual writing in mediated spaces (1) Corpus of Cyber-Jamaican (CCJ)

(2) Corpus of Cyber-Nigerian (CCN)

(3) Corpus of Cyber-Cameroonian (CCC)

web-forum used as source

www.nairaland.com www.jamaicans.com www.camerooninfo.net

languages used on forum (in order of frequency)

English, Jamaican English, Pidgin, English/ Jamaican Igbo, Yoruba, Creole other indigenous

time of download 10/2008

French, English, Pidgin, “Camfranglais”, various indigenous

11/2008

10/2008

year of posts

2000–2008

2005–2008

2000–2008

# posts

252,015

244,048

179,563

# posters represented in corpus

2,128

11,718

3,140

ca. 17,3

ca. 22,1

# words (million) ca. 16,9

Table 1. Post-colonial diasporic web forums – web-derived corpora 1–3

of conduits – ranging from post-colonial literature through popular music and globally active sub-cultural movements such as Rastafarianism. While the sedimented lexical results of such linguistic and cultural encounters in the global “mediascape” (Appadurai 1996) have already been documented in authoritative sources such as the Oxford English Dictionary (see, for example, the entries for riddim, overstand or bashment), Nigerian speechways are only beginning to appeal to mass audiences outside the country and its immediate diaspora (compare, for example, the commercial success of Nigeria’s popular “Nollywood” school of action cinema in the Caribbean). Cameroon seemed worth including because of its dual Anglophone and Francophone colonial history (Pfänder 2009). Table 1 summarises the relevant components of the material. Data for Spanish were obtained from the Andean region and Argentina; the latter two corpora (CCA and CCB) document the usage of well comparable communities of practice (soccer fans):

534

Christian Mair & Stefan Pfänder

Corpus

(4) Corpus of CyberPeruvian-4 (CCP4)

(5) Corpus of CyberPeruvian-5 (CCP5)

(6) Corpus of CyberArgentinian (CCA)

web-forum used as source

forosenperu. com

forosperu.net

futbolhinchadas argento.com.ar debolivia. mforos.com

languages used on forum (in order of frequency)

Spanish, Andean Spanish

Spanish, Andean Spanish

Spanish, English

Andean Spanish, Spanish

time of download

1/2009

1/2009

1/2009

1/2009

2005–2008

2007–2008

2006–2008

year of posts 2007–2009

(7) Corpus of CyberBolivian (CCB)

# posts

171,665

374,258

159,315

27,530

# posters represented in corpus

871

8,421

1,017

526

# words (million)

ca. 4,4

ca. 19,9

ca. 7,8

ca. 2,1

Table 2. Post-colonial diasporic web forums – web-derived corpora 4–7

Some usage trends, such as a tendency towards informality and the destandardisation of orthography and written norms, are pervasive in all forums. Other developments are forum-specific. For example, only one of the Andean-Spanish forums seems to be implicated directly in spreading grammatical change in the community (section 5 below). By focussing on vernacular web data, we extend the traditional focus of sociolinguistic analysis, which has prioritised the study of spontaneous vernacular usage in local speech communities (and even viewed literary or otherwise mediated representations of the vernacular with suspicion, as inauthentic or stereotyped representations of the real thing). We are doing so for two reasons. For one thing, we would simply consider it a pity to leave fascinating topics such as the study of vernacular linguistic resources in CMC, pop music or the media to disciplines such as literary studies, media studies or cultural studies. For another, it very soon became clear to us that the mediated vernaculars as we encountered them in our forum data are more than merely truncated, derived or secondary versions of their supposedly

Vernacular and multilingual writing in mediated spaces

535

real and authentic counterparts in the face-to-face community. Rather than deficit, we see difference: in loose and geographically dispersed communities of practice, vernacular resources take on new uses and new significance, but they do so in the service of authentic acts of communication in a new communicative domain. In their use of vernacular resources, forum participants certainly exhibit playfulness, exaggeration and wilfully idiosyncratic behaviour, but all this notwithstanding, the linguistic play we observe is not arbitrary and purely personal whim. At least for the core participants, the real-life sociolinguistic order remains the interpretative baseline against which the conventions of medium-specific usage emerge. In this situation, asking whether the mediated vernacular is a good mimetic representation of the supposedly real thing misses the point; in fact, it would be inauthentic in the new communicative domain if it was no more than that. Online users’ vernaculars are thus characterised by selective representations of non-standard features rather than a mimetic desire to render the whole of the spoken word in the new medium. This selectiveness, as we have pointed out, is not a deficit. If deployed skilfully, even a narrow range of morphosyntactic vernacular features can be used to great effect in styling a vernacular voice suitable for the digital medium. An entirely new source of semiotic potential in the digital vernaculars is the purely visual plane introduced by experimental spelling, which – in the broader understanding of the concept advocated by Blommaert (2010: 180) – can also become part of the medium-specific vernacular “voice”, which “embodies the experiential and practice dimensions of language and refers to the way in which people actually deploy their resources in communicative practice”. On this broader understanding, it is not a paradox to claim that we can hear the vernacular voices of the diaspora in our forum data by reading the messages on the screen. As Coupland (2003: 428) has argued, probing issues of ownership of and legitimate access to languages (or varieties of a language) is part of the authentic “performance space”: ‘The authentic speaker’ will often be the person who is able to occupy this space more productively, reworking traditional symbolic resources in new ways. This may mean playing with the ownership dimension of speech.

The communities studied by us use vernacular linguistic resources as one of their most important strategies to define and occupy the media spaces opened up by CMC. This is a new departure, as the written medium has not traditionally been a hospitable environment for the vernacular. In literature, vernacular writing has tended to be a marginal and minority pursuit. In newspapers, the vernacular is used for humour, folkore and the occasional personal opinion piece.

536

Christian Mair & Stefan Pfänder

This is no longer so in CMC, and this is an empirical fact about contemporary writing practices which a sociolinguistics of the 21st century should account for.

3.

The new media as a means of shrinking distances: Technology supporting mundane interaction in the diaspora

This baseline function of web-forum communication will chiefly be illustrated with examples from the CCJ (the Corpus of Cyber-Jamaican). The name of this resource is not intended to convey a programmatic claim that “Jamaican” is a language, but rather to emphasise that it would be futile to look for strict dividing lines between the several varieties of English used routinely by participants, i.e. international standard varieties of English, such as British or American, Standard Jamaican English (SJE) or Jamaican Creole (JC). In traditional (pre-digital) written usage, JC has never made much headway outside the narrowly circumscribed domains of folklore, humour, direct quotation of folk speakers in the press, and some literary and academic experimentation. The digital media, however, have changed this profoundly in a very short time. Even informal observation suggests that people are quite ready to write JC (spelled in ad hoc orthography) in digitally mediated communication. This is remarkable not least in view of the vast geographical dispersion of the writers who produce “Cyber-Creole”. Among the 2,141 contributors to the forum discussion “captured” at the time of download, 1,318 had indicated a plausible geographical location in their user profiles. Plotting this information on a map, with darker shades indicating higher concentrations of speakers, we get a good visual impression of the world-wide spread but differential degree of entrenchment of Cyber-Creole: What we see is the long shadow of Empire and (post-)colonial migration to Britain and North America. But note that there are also surprisingly strong links to practically the whole of Northern, Western and Central Europe – but rather weak connections to the third pillar of the historic “Black Atlantic” triangle, namely West Africa. CCJ contains texts of different kinds: – passages which read as if spontaneously produced spoken JC was transferred on the screen with little or no complication, – passages with elements of stylisation, due probably to writers adapting to the new medium (at first consciously and experimentally, and then routinely and conventionally), and – passages which are rhetorically crafted and reveal a high degree of metalinguistic awareness regarding the deployment of JC linguistic resources.

537

Fig. 1. CCJ – regional base of 1,318 forum contributors

Vernacular and multilingual writing in mediated spaces

538

Christian Mair & Stefan Pfänder

This is an example of the first kind, a narrative which we can easily imagine being delivered in exactly the same way in face-to-face interaction: (1) [RollinCalf] Satdeh time … braps, beef soup done, seh shi neva go a college fi bwile no yam. One mawnin mi a hat up little mackrel an some pepper an shi tell mi seh mi a tink up di house. [Saturday time – and all of a sudden, the beef soup is ready, and she says she didn’t go to college to cook yams. One morning, I am heating up a little mackerel and some peppers and she tells me that I am stinking up the house] To the extent that this is possible in an English-based ad hoc orthography, the spelling of this passage highlights pronunciation features of JC (the – di, stink – tink). The morphosyntactic profile of the passage is consistent with a position at the basilectal end of the SJE-JC continuum: preverbal progressive marker (a tink, a hat up), uninflected verbal stems with past reference (shi neva go), seh as conjunction (tell mi seh). The second and third type of text, by contrast, deviate to a greater or lesser extent from the conventions of the creole-standard continuum, leading to a mimetic deficit, without, however, becoming inauthentic, as the following passage shows: (2) [Bizi_Q] so what mi can use fi clean gems, stones, diamonds, etc?? mi have de silver/gold cleaning solution but it nuh seem fi do nutten fi de stone dem . . sometimes it look wussa dan when it went in. mi read up pon de net an some sites say use dish washing liquid . . others say that’s a big no no. what unu use clean unu stones? would it be better fimi carry it go a jeweler?? how much dat would cost? [so what can I use to clean gems, stones, diamonds, etc.? I have the silver/gold cleaning solution but it does not seem to do anything for the stones … sometimes it looks worse than when it went in. I read up on the net and some sites say ‘use dish washing liquid’ … others say that’s a big no no. What do you use to clean your stones? Would it be better for me to take it to the jeweller’s?] It is difficult to imagine this being produced spontaneously in face-to-face interaction because in contradiction to continuum conventions a large number of overt-prestige (SJE) and covert-prestige markers (JC) are deployed at the same time. Note, for example, the high frequency of the JC grammatical marker fi, which in many cases corresponds to infinitival to in English translations. In Bizi_Q’s post, fi occurs together with several other basilectal JC features, the 2nd-person plural unu, preverbal negation with no, and the

Vernacular and multilingual writing in mediated spaces

539

optional JC creole plural-marker dem (e.g. in de stone-dem), which is expected. However, its occurrence in phrases with inflected plurals, as in fi clean gems, stones, diamonds, is not. Similarly, unu stones combines a basilectal possessive pronoun with an acrolectal nominal head, which is as incongruous at first sight as the combination of an “English” main clause (would it be better) with a dependent clause that contains a JC serial-verb construction (carry it go). Such deviations from the norms of face-to-face interaction must not, however, be viewed as mere deficits. Rather, they show that, in Blommaert’s terms, the JC features in question have been mobilised to serve new functions in a new medium and in a slightly different new style of interaction. “CyberJamaican” may be a new and different manifestation of JC, but the way it is employed remains grounded in the broad principles of the existing sociolinguistic order. Indeed, it is the shared access to and the shared attitudes towards JC linguistic resources which is one of the identifying characteristics of the forum and makes it a digitally supported community of practice. Any web forum might show somebody with the requisite encyclopaedic knowledge producing a few JC expressions if the topic happens to be Jamaica, its people, or its culture – by way of (pseudo-)quotation or narrative-dramatic enactment.1 In the example quoted above, however, the interactive complexity is deeper. There is no particular connection between the topic, a mundane request for help, and Jamaica. However, by drawing on JC linguistic resources the writer formulates the request in a manner that makes it maximally efficient in its appeal within the community. Just as they would use gesture, intonation etc. to modulate the verbal message and the content of an utterance in face-to-face interaction, contributors to the forum draw on dialect mixing in computer-aided written communication.2 Cultural and sociolinguistic hyperawareness – the “vibrant site” (Bhatt 2010: 520) of the postcolonial politics of language and identity – is apparent in the following passage, illustrating the third kind of vernacularised text: (3) [Blugiant] oww manee peeps inn dem caribbean household versus mzungu oousehold. iff itt tekk two ar more caribbean wage earnas wukkinn more 1

2

This is commonly the case in European forums organised by enthusiasts of Caribbean music, for example www.sunny-music.ru, with its motto “Tolцko pozitivnaѕ muz«ka. Vse o reggi, dab, ska …” [only positive music, all about reggae, dub, ska], in which the technical terminology of the subculture (bashment, riddim, ragga(muffin), Jah, etc.) is variously rendered in Cyrillic and Latin script, indicating different degrees of assimilation and conventionalisation. Similar phenomena have been studied on the basis of e-mail messages, and in a code-switching framework, in Hinrichs 2006.

540

Christian Mair & Stefan Pfänder

owa dan mzungu fi mekk more dan mzungu oousehold widd less peeps wat iss da artikkle seyinn bout da caribbean peeps qualitee aff life [How many people are there in those Caribbean households versus white people’s households? If it takes two or more Caribbean wage earners working more hours than white people to make more than a white household with fewer people what is the article saying about the Caribbean people’s quality of life?] In view of the fact that the externally most striking feature of this passage is visual (“eye-dialect”), namely the sensational spellings in tekk, wukkin and artikkle, asking about the relation between this passage and the use of JC in spoken interaction misses the point here. The double consonant does not have, and is not intended to have, a pronunciation equivalent. In other cases, for example the in aff, the spelling even runs counter to natural pronunciation in SJE and JC. Against this massive orthographic-visual destandardisation, the genuine morphosyntactic creolisms, such as the occasionally absent verbal third-person singular (itt tekk), the unmarked nominal plural (owa) or the basilectal Creolism fi, recede into the background. In fact, as the gloss shows, we are dealing with a re-spelled and partly relexified Standard English utterance displaying the type of syntactic complexity to be expected in a written text. The context of the passage reveals the writer to be an ethnic political activist, and his oppositional stance is emblematically expressed by various non-standard linguistic resources at his disposal: the instances of JC grammar, the emphatically non-standard spellings (only very few of which are in any meaningful correspondence with JC pronunciations), and – interestingly – the word mzungu, which is a mildly derogatory term for whites current not in the Caribbean but in Southern and Eastern Africa.3 To varying extents, contributors to CCJ are thus styling or performing their language, moving it away from a mimetic representation of JC as it would be used in spontaneous face-to-face interaction in local vernacular communities. The game works because, in doing so, they can mobilise real resources shared by the globally dispersed diasporic community, which re-invents itself as a community of practice (Meyerhoff 2002) by participating in the shared space of the web-forum. In the following section, using analysis

3

This example, incidentally, points to the growing importance of post-apartheid South Africa as an exporter of vernacular English features – in the first instance in Africa and the African diaspora (as witnessed here) but ultimately of course in the entire World English market (see the OED entry for mzungu, which shows a development of use from technical-ethnographic writing (19th century) to general usage (20th and 21st centuries).

Vernacular and multilingual writing in mediated spaces

541

of self- and other-stylisation, we shall take a closer look at how people play with claiming or assigning “ownership” of vernacular styles of speech when they post to internet forums.

4.

Self- and other stylisation and the use of multilingual resources on the web

In this section, we describe various processes of stylisation on the web (Coupland 2010a; Lacoste and Mair 2012). A sizeable amount of vernacular forms in our corpora occurs where oral dialogues are re-enacted or performed, especially where a certain manner of speaking is attributed to another “speaker” in an ad hoc, improvised manner (Breyer, Ehmer & Pfänder 2011; cf. also Androutsopoulos 2010). We identify two main types of stylisation practices: self-stylisation (Selbststilisierung, cf. Keim 2007), and other-stylisation (Fremdstilisierung, following Deppermann 2007). In the data we have analysed, multilingual writing practices are almost exclusively found with forum users who have full command of the normative varieties of the locally dominant languages and who thus use multilingual writing as an additional resource. This could be illustrated with examples from all our subcorpora. For reasons of space, discussion will be limited to Bolivian and Cameroonian data here. The analysis of the two relevant forums (focussing on politics in Cameroon and on soccer in Bolivia) shows multilingual styles being deliberately used as rhetorical strategies. Thus, the forums have become spaces for writers who have a variety of local and non-localisable stylistic tools at their disposal and can cross into the varieties of others for all ludic purposes. The first example to be discussed, the Bolivian soccer fan forum (analysed for this purpose from 2006–2009), shows that the members clearly identify with a particular club. They discuss topics such as other clubs and their fans, coaching staff and, most of all, recent games, expressing both disappointment and triumph, as the case may be. Within this community of practice, a common repertoire is formed through characterisations of the typical activities of a group of fans before, during, and after the game. Stereotyped linguistic features are used to create a mocking tone. In the following example, the writer mocks fans of a defeated opposing team by performing an imaginary scene in which the fans try to persuade their coach to improve the club’s chances in the national competition by buying a topnotch player for a high transfer fee. The forum users parody the speech of the opposing team’s fans as they perform them addressing their coach. The

542

Christian Mair & Stefan Pfänder

opponents are characterised as coming from a rural region of the Andes, and dialect features that index rural, uneducated speakers from the highlands are emphasised for comic effect. (4) ya pshh ispanya tenemos plateta comprarime jogadores pshh papeto, puchi caray qui gravi eres ispanya un equipo regio podemos armar jajaja [C’mon Ispanya, we have the cash, buy me some players sweetie, goddam you’re a pain Ispanya, we can put an awesome team together hahahaha] In this example, other-stylisation sensu Deppermann (2007) is obvious. In insulting the team run by president Florencio España, the poster performs a regional stereotype – that of a person from Oruro – through linguistic features. This figure is indexed linguistically through the use of swearwords borrowed from Quechua, or forms based on linguistic transfer (puchi caray), diminutives (plateta/papeto), and globally copied suffixes (e.g. quechua -ri- in compra-ri-me, ‘go buy something’). In these forums, a playful interaction that takes place on a certain day may be picked up many days later by other participants. In doing so, humorous practices for the imitation of accents are copied, but reinterpreted and reformed in the process. What started out as an ad hoc improvisation may give rise to a follow up, and, if successful, to new practices and conventions. Interactions are usually most active within a short time of the original post, but are available for viewing in the archive for several months. We find for example: (5) le dijo al pobre viejito “shabimosh que plata ahy puish papeto, como no wash trayer wenosh jogadoreshhh kirimosh al darwen piña yaaa [I said to the poor old man: we know that cash we have y’see sweetie, how is it you will not bring us good players? We do want the Darwin Piña, go for him!] Here, the same social persona as the one performed in the previous excerpt is characterised through very similar stylisation techniques. If the staging was improvised in the first excerpt (or at some theoretical point in the past when it was first performed), the stereotype is gradually becoming conventionalised as it is revisited and reused by other members of the group. Ultimately, the persona and the features associated with it are stored as part of the community’s sociolinguistic resources. Turning to the following excerpt from our Cameroonian material, one might plausibly assume that there is congruity between the particular African locale, the varieties used, and the general social profile of the speaker:

Vernacular and multilingual writing in mediated spaces

543

(6) Hi Mevio, No mek mi a lap oohhh!!!! U di mimba ce some politik go kam hia? A put ma hand fo faya so dem no go kam. Waiti dem go kam chap ouna non? Dem go go fo lege we dem go gi pagne and mimbo. Franchement noble Mevio ce serait une tres bonne chose, mais je ne vois pas le premier courageux qui viendra se faire bruler les ailes ici. Qui sait! kmer na kmer! [Hi Mevio, don’t make me laugh. Do you think that politicians will come here? I bet they won’t come. What are they really going to eat? They’ll go for the fools who they’ll give cheap local textiles and drinks. Frankly, noble Mevio, this would be a great thing, but I can’t imagine the first courageous one who would come here to have his wings burnt. Who knows? Cameroon is Cameroon!] Just a little later, however, the very same speaker uses an elaborated and near normative style of French, writing in a way that is virtually indistinguishable from the international educated French standard: (7) Je crois avoir rencontré Anicet Ekane du Manidem dans certaines rues de ce cité. Est-il désormais un habitué des lieux? … Je voudrais que nous ayions sur ce site des week-end à thèmes où le Chat devra accueillir un politicien … qui vienne museauter avec les internautes. J’espère que c’est plus clean comme ça. [I think I have met Anicet Ekane of Manidem in (= Mouvement Africain pour la Nouvelle Indépendance et la Démocratie) certain streets of this city. Does he live there now? … I’d like to have on this site weekends with a theme where the Chat welcomes a politician … who comes to chat with the users. I think this is cleaner this way.] Note that the two lexical borrowings from English (week-end and clean) are commonly used, both in France and elsewhere in the francophonie. This group of forum users is part of a particular community of practice, whose main focus is taking part in political discussions on the internet. We have analysed data taken from this community from 2008–2009. Participants in this group typically shun the kinds of communicative connections that would integrate their online community with offline relationships and interactions. The people who post in this forum often write about politicians and politics in Cameroon; however, they are not generally interested in debating them directly on the web or in meeting the politicians personally off-line. Furthermore, forum users characteristically present themselves through their discussion as belonging to diverse geographic and social groups. Explicit shows of support for a particular political group are generally avoided. Rather, we can identify a tendency for users to voice disenchantment with politics in

544

Christian Mair & Stefan Pfänder

general and to criticise mistakes made by politicians from all sides of the political spectrum. To return to our argument, multilingual style is deliberately used as a rhetorical strategy here. It has become a preferred strategy for multilingual writers who have a variety of local and non-localisable stylistic tools at their disposal. Of course, in general terms this strategy is not exclusive to the web. Similar phenomena can be observed in everyday oral use in Cameron as well. It would thus be an interesting research question to find out which segment of the total range of multilingual behaviour commonly observed in Cameroon makes it into the digital communicative space of the Cameroonian forum, or whether there are facets of multilingual behaviour which thrive on the web more than in face-to-face interaction on the ground. We envisage such parallel investigations for all our forums for the future. What is worth noting immediately for the purposes of our present argument is that the young forum activists repeatedly underline that membership in their community of practice is not open to all people, particularly not politicians, “because they cannot write as we do”. In the course of a Frenchlanguage discussion, one of the forum users suggests that the politician the participants are discussing could be invited to participate in the debate. In this interaction, the virtual forum is discursively constructed as a physical space, in which objects such as roads could exist. A forum user claims to have “seen” some of the politicians who might be possible candidates moving along the “roads” of the forum. However, this suggestion is rejected by another user, using a form of the contact variety francanglais (De Féral 2006). Thus, the ad hoc use of diverse linguistic resources allows users to stake out two positions within a single discussion thread: the point of view of politicians and the speech of forum users who style themselves as ‘young’ and ‘trendy’ through, among other things, the competence they display in their multilingual practices. In other words, the multilingual style serves to demarcate the boundary between two groups. Following Keim (2007), we can label this type of linguistic design self-stylisation, which serves to set boundaries between in- and outgroups. In the specific context of present-day Africa, the forum community contributes to an important linguistic and cultural trend, namely the “dissociation from colonial legacies as well as from the postcolonial political elites, impotent administration, and tribalist instrumentalizations of language and language policies” and the move towards the “creation of autonomous African modernities […] that include the city (and the state), brought about by the interplay of both local dynamics and global flows” – as Beck (2010: 11) puts it in her analysis of urban languages in Africa.

Vernacular and multilingual writing in mediated spaces

545

Let us sum up the findings so far (sections 2, 3, and 4). New practices come into being in new communities of practice – mainly because traditional and territorially highly restricted multilingual practices which have been confined to face-to-face interaction are extended to informal writing and given a wider airing in a deterritorialised “mediascape” (Appadurai 1996). Contact varieties such as Andean Spanish or Jamaican Creole in the Americas, Cameroonian Francanglais or Nigerian Pidgin English in West Africa have long been limited to oral use; these vernaculars are being written now. Perhaps the most active among the pioneer writers are members of diaspora communities. This latter observation might be more important than we thought at first sight. The data show clearly that forum users are not so much organised in groups sharing the same, contact-influenced variety of one of the three global European languages English, Spanish or French, no matter where they live. Rather, these varieties become resources for writers whose vernacular used to be another variety – via crossing and stylisation. It is constitutive for the forums that writers use not only local (i.e. dialectal) resources, but also – often highly indexical – elements of the non-prestige contact varieties. Thus, in these mediated spaces, the tie between geographically defined spaces and a specific linguistic repertoires is weakening. The local, regional or national associations of vernaculars vanish as these are used as globalised resources within (frequently diasporic or otherwise transnational) communities of practice which are not so much defined by citizenship or nationhood, but by shared interests (i.e. sports, music, or politics) and mutual engagement.4 Approaching the phenomenon from the other end (i.e. the technologisation of the word in digital communication), we note an interesting effect. As traditional vernaculars are losing some of their local associations through use in the digital media, they become effective in vernacularising the web – in that they help in the creation of particular social and ethnic dialects of “Netspeak” (Crystal 2006) which – through the use of the internationally current conventions of the register (abbreviations and initialisms such as lol, lm(f)ao or emoticons, etc.) and highly vernacular resources – are at once global and local. A question which arises and which we have not answered so far is whether, in appropriate circumstances, the innovative potential on the dis-

4

Of course, it goes without saying that the undermining of national identities is brought about much more easily in forums devoted to music than, say, politics, and that the intense but limited local loyalty of the soccer fan will survive deterritorialisation in the web-forum.

546

Christian Mair & Stefan Pfänder

course level translates into structural innovation in the grammatical repertoire, too. This is what we will discuss in the following section.

5.

Is there ongoing structural change on the web?

In this section, we shall turn to the question of whether the observed multilingual practices may lead to lasting structural change, beyond the discourse moment of ad hoc improvisation. We suggest using the notion of emerging constructions (Auer & Pfänder 2011) to provide a theoretical framework for the uses of multilingual linguistic resources that can be observed in our data. There is a tension in emergence that results from the interaction between given or prefabricated elements and the novel elements arising from the ongoing interaction on the forum. Our question thus is: Can we find direct manifestation of structural changes in language on the internet? And if so, does the innovation start on the internet? A general answer to these questions will, of course, depend on how we define “change” and how we define “language”. If we construe languages as “resources in communicative practice” (Blommaert 2010: 180, cf. section 4 above) and if we understand the virtual global world to be contiguous with, not disconnected from the actual local world (cf. section 1, and in a similar vein, Blommaert 2010: 79), then we will in fact be able to note signs of change in practices (cf. again section 4 for this perspective). If, on the other hand, by “change” we mean the spread of an innovative form throughout the community and its final establishment as part of the obligatory grammatical inventory, things will take longer than the time window opened up by our forum data. There is, however, a promising phenomenon worth further investigation from this perspective in the Peruvian and the Bolivian data. Vernacular constructions can be found in the informal scripturality of the forums. The univerbalised dizque (‘supposingly/the supposed’, literally: ‘saythat’) is one of these constructions. Before 2006, the time of collection of the earliest forum data, this construction was attested in oral language use in functions which do not exactly match the practices documented on the forum. Most linguists assume that the traditional Latin American use of dizque (written also as disque, disq, disk, diske etc.) is a typical example of transfer through contact from Quechua, where nispa (ni-spa, literally ‘say-ing’) just as dizque (literally dice que ‘says that’) expresses non-evidentiality or hearsay, respectively. One good argument for contact-induced change is that the category of hearsay is obligatory and far more grammaticalised in Quechua than in Spanish. Another argument in favour of the contact hypothesis is

Vernacular and multilingual writing in mediated spaces

547

that in Latin America we mainly find dizque as an evidential marker in the Andean countries (cf. Pfänder et al. 2009; Dankel 2013). However, we also find dizque in Mexico (Olbertz 2007) and Colombia (Travis 2006), where Quechua influence cannot have been a factor. In some non-Andean contact varieties (cf. Olbertz 2007), dizque has undergone a further semantic change, away from the notion of hearsay suggested by the etyma to a more general modal meaning of ‘supposed/ly’. This semantic change is a pretty obvious case of what Traugott and Dasher (2001) have termed “subjectivisation”. This – and not the typically oral Andean evidential marker – is by far the most frequent function of the form in Andean Internet writing. In our web corpora, we find roughly 2,000 instances of dizque. In almost half of these cases dizque modifies a noun with a definite or an indefinite article or a possessive pronoun: la dizque bandera (‘the supposed flag’), algun dizque presidente (‘some supposed president’). As far as we can see, this usage has never been analysed in the relevant literature on Andean Spanish. In other words, the most common function of dizque/disque in the web data is the one which is less important in spoken data (e.g. in the Freiburg ANDES corpus of 120 hours of speech recorded in 2000–2012). We do find the typically oral evidential marker. If we go into some detail, we can differentiate two subtypes of this use on the web (here labeled as types 2 and 3, type 1 being the non-univerbated form dice que). We find dizque used as a phrasal adverb, both occurring at the beginning (type 2) or at the very end of a sentence (type 3): Type 2 (8) (a) … ahora Telefónica se limpia y le hecha el pato a los usuarios, dizque un nuevo y misterioso virus está generando abundante tráfico [… now Telefónica washes itself clean and gives the hot potato to the users, it is said that a new and mysterious virus is generating a lot of traffic] (b) JUAJUAJUAJUAJAUA PERO Q TREMENDA ESTUPIUDEZ JAJAJJA TERROR A LOS CHIFLADOS DIZQUE TENEMOS JAJAJAJA POR Q PS ??

[hahahahahaha but what a tremendous stupidity hahahah terrified of the nutters they say we are hahahah why then??]

(c) Cancelados los fuegos artificiales. Dizque hay demasiado viento (más de 100 km/hora) y es peligroso. fireworks cancelled! It is said that there is too much wind (more than 100 km/h)]

548

Christian Mair & Stefan Pfänder

Type 3 (9) jajajajajajajajajajaja 24 horas dizque jajajjaajajajajajajjaajajjajajajaja Sigue respirando aire europeo (dizque) [He is still in Europe (at least this is what we heard)] In Type 3 the syntactic pattern is quite different from type 2, but the function is again evidential. However, this is not the case for the types 4, 5, 6 and 7 which are seldom found in oral, but frequently in the web forum data set. Here, the scope is drastically reduced. In type 4, the marker dizque modifies an adjective: Type 4 (10) … gente con sus discursitos dizque moralistas. [… people with their supposedly moralistic little speeches] … cerca a la frontera dizque chilena [… near the alleged Chilean border ] Dizque also occurs in a pre-nominal slot, modifying a noun and thus functioning as an adjective, marked by a reinforcing element like algun (type 5), a definite article (type 6), or a possessive marker (type 7). Type 5 (11) (a) UN DIZQUE TRAPO DE UNIVERSITARIO [supposed university lecturer’s clothes] (b) un dizque combate en el que el anciano querria ganar [a supposed fight in which the old man wanted to win] Type 6 (12) (a) Los dizque propriotarios del agua [these alleged/supposed/self-appointed owners of the water] (b) todas las dizque democracias … no han sido más que dictaduras [all these alleged democracies … were actually dictatorships] Type 7 (13) (a) pizarro con su dizque look [Pizarro with his supposed look]

Vernacular and multilingual writing in mediated spaces

549

(b) PUES CON TU DIZQUE FUERZA CARGAS 100 KILOS DE CEBOLLA

[Well with your supposed strength you carry 100 kilos of onions] In types 4 to 7, dizque has the meaning of ‘commonly referred to as’, ‘commonly considered’ or ‘supposed’. It is not the origin of information the writer is considering (as in hearsay), but the adequateness of the name is put into doubt. In some cases the use of this marker is commented metalinguistically or highlighted orthographically: digo dizque porque (‘I say dizque because …’) or la “dizque” evidencia (‘the supposed evidence’). This appears as type 8 in our analysis. As regards the distribution of these uses, nationhood still seems to play some role. At least two observations point in that direction. First, we find hardly any instances of dizque in Argentine Web Forums, and secondly, there is a clear tendency of differentiation in writing the diZque in Peru and diSque in Bolivia (Table 3 and Fig. 2): The question arises whether the greater prominence of a usage in the web copora might be due to the specific communicative constraints of the virtual medium. Androutsopoulos (2010) is certainly right in pointing out that we are still lacking reliable comparative and in-depth empirical research on convergence and divergence between the uses of informally spoken and informally written vernacular varieties. CMC approaches to language did show,

Table 3. Dizque in Peruvian, Bolivian and Argentinian web data

550

Christian Mair & Stefan Pfänder

Fig. 2. Dique – regional distribution of orthographical variants

though, that virtual routines tend to build on usage in spoken interaction (Pfänder & Wagner 2010). Thus we would not expect innovations to arise only in the virtual world and remain confined to it. But the specific communicative parameters defining the virtual environment, in particular the need for evaluation of the “other” in very open circumstances, certainly explain why the relevant uses of dizque should have started mushrooming so suddenly. In a functional perspective, the least subjective use is a low assertive movement, and thus, an epistemic use: ‘I say it, since others say so, but I cannot take up the responsibility for this.’ In the more subjective use, the speaker does not highlight an epistemic source of information, but questions or challenges the legitimacy of the assertion. This is one of the highly frequent activity types in mediated spaces and their communities of practice, where people do not know each other well, and where they spend a lot of time evaluating and criticising third parties’ actions, without necessarily being able to evaluate the communicative contexts to the degree of detail and depth which would be possible in face-to-face interaction. If we have a closer look at the spread of dizque, we find that as the form started spreading in the Peruvian web forums from October 2006, a steady increase can be noted during the first twelve months. What also becomes

Vernacular and multilingual writing in mediated spaces

551

Fig. 3. Spread of dizque in Peruvian web data from September 2006 to August 2007

clear from the Figure 3 is a rise of the forms that tend towards a stronger subjectivisation (types 4–7). At the beginning of this section we asked two related questions: Can we find direct manifestation of structural changes in language in web data, and if so, is language use on the internet a likely cause for structural innovation? The answer we can give on the basis of the present case study is that the subjective uses of dizque were present in spoken language before the web. The relaxation of formality conventions and the high intensity of other-evaluation which characterise forum interaction helped introduce dizque into written discourse as well. Thus we do to find spread of dizque to new communicative genres, but no direct manifestation of any functional change as yet.

6.

Conclusions

We have designed our analysis of post-colonial diasporic web-forum communication as a contribution to an emerging sociolinguistics of globalisation, emphasising that this new line of research will only be successful as a multilingual enterprise. Even if languages are disappearing fast, we are not heading towards a monolingual world, and many languages other than English are strongly involved in dynamics of linguistic globalisation. Students of

552

Christian Mair & Stefan Pfänder

varieties of English, French, Spanish or Portuguese5 around the world must not work in ignorance of each other’s findings and theoretical models (as they have often tended to do in the past) and start co-operating. In an attempt to capture the new quality of social, ethnic and linguistic heterogeneity in today’s global cities, Vertovec (2007) has suggested the label “super-diversity”. It is worth asking whether this label might not also be the most fitting description for the growing linguistic diversity in the participatory domains of the World Wide Web. Consider, for example, the self-description of one contributor to www.nairaland.com, a major discussion forum linking Nigerians in Nigeria and the global diaspora, and the source for CCN, the “Corpus of Cyber-Nigerian” described in section 2 above: I’m proud to be what I am. I’m Nigerian, Bahamian and British. When people ask me where I’m from, I don’t say ‘Nigeria’, I say ‘Nigeria and the Bahamas’. That is the truth. I’m not fully Nigerian so why claim to be one? When people on the internet ask me where I’m from, I say ‘I’m originally from Nigeria and the Bahamas but I’m currently living in Belgium’. Again, it’s the truth. (CCN [6986], Retro)

In traditional research on World Englishes, this particular individual would be considered to be an exception or an atypical informant, because she is unlikely to consistently use either Nigerian English or Bahamian English. However, for the sociolinguistics of globalisation, the same individual might well emerge as representative, an individual commanding a broad range of standard and non-standard linguistic resources in their L1, and operating in a multilingual environment in which at least two languages other than English play a significant role. For migrant groups as a whole, communication technology has made a vast difference to their everyday lives and experience (which, incidentally, may explain the energetically experimental use they make of it): The revolution in transport and communications has eased the way for many migrants […] to retain their links ‘back home.’ […] Regular visits, return and remigration, bi-furcated migration, circular migration have all become part of the lexicon of […] migration, while telecommunications and increasingly the Internet have eased the way for more regular and innovative forms of contact and renewal. (Chamberlain 1998: 7)

As we hope to have shown, the study of vernacular and multilingual practices in CMC thus provides one important window on how diasporic and migrant 5

This is a list of the European ex-colonial languages which have come to dominate the contemporary linguistic landscape of the Americas. By mentioning these, we do not wish to exclude transnational languages with a non-European base (e.g. Arabic, Hindi) from consideration in other contexts.

Vernacular and multilingual writing in mediated spaces

553

communities define and position themselves, both with regard to their communities of origin and the social and linguistic mainstream in their new homes. For a full picture, what is needed in addition is, of course, sociolinguistic community studies of Caribbean and African immigrants in New York or Miami, of Andean villagers involved in temporary migration to Lima, Peru, or of Francophone West Africans in Toronto6 – that is some of the important real-world recruiting grounds for contributors to the forums investigated here. On the ground, in the ethnically mixed immigrant quarters of the world’s global cities, super-diversity has led to the emergence of new multiethnic urban vernaculars. In the case of London, for example, Cheshire, Fox, Kerswill & Torgersen (2008) have shown how the traditional urban dialect (Cockney) and more recent ethnic vernaculars, such as London Jamaican/ British Black English, are being replaced by a new dialect, Multi-ethnic London English. Obviously, the new styles which are emerging in diasporic web-forums cannot be compared to such multiethnic vernaculars in all respects, but there are important parallels. In the super-diverse urban neighbourhood as in the digitally supported diasporic forum, the pool of available linguistic resources has expanded and diversified massively, while social ties have become less well defined and sometimes weaker. And yet language and communication remain the only way of ensuring continuing social cohesion. For the present study, we investigated newly emerging multilingual practices on the web, where English, French and Spanish are involved. In the web forums under scrutiny, the participants (or “posters”) explore multiple ways of expressing communicative needs by selecting and (re-)“combining the linguistic resources available to them” (Johnstone 2001: 123). When comparing the English, Spanish and French CMC corpora, we can corroborate recent findings on spoken corpora to the effect that while the francophonie may have lagged behind the Spanish- and English-speaking worlds in terms of liberal attitudes towards diversification (including the conscious and playful use of diverse linguistic resources), things are changing rapidly now (cf. Calvet 2006, Gadet, Ludwig & Pfänder 2008; Pfänder et al. 2009). Our findings on present trends are indeed quite similar for the three world languages under investigation.

6

For the last-named locale, pioneering work has been undertaken by Heller 2006 [1999] in her ethnographic study of language practices in a French-medium school serving a multi-cultural clientele in English-dominant Toronto.

554

Christian Mair & Stefan Pfänder

Most importantly, CMC has emerged as a powerful catalyst for the vernacularisation of writing and, through it, for the vernacularisation and re-territorialisation of that most diffuse, extended and “global” of all communicative domains, the World-Wide Web. In traditional written communication, the vernacular has been quoted rather than used. In the interactive or participatory Web, on the other hand, non-standard linguistic resources and vernacular multilingual practices have become important strategies to carve out and claim a space in an ill-defined and disorientating digital discourse universe. The new digital vernaculars are not deployed arbitrarily and unsystematically, but they challenge conventional assignations of prestige and stigma familiar from face-to-face interaction. In other words, they serve in the creation of new, medium-specific vernacular styles. It is constitutive for the forums that users also use resources of the contact languages. In doing so, the resources are often stylised. We can find self- and other-stylisation. This leads to the adoption of constructions that have so far only been confined to a spatially restricted contact zone. Spanish speaking forum members without an Andean migration background participate from Europe and use – often in a playful and stylising manner – constructions which they would only use in this way on a case-by-case basis in very marked humoristic modalities. Structural innovations directly prompted by CMC, on the other hand, have only rarely figured in our data. A possible case is presented by some extensions of the dizque construction, which have shed the ludic and consciously experimental connotations characterising some of the more striking webrelated innovations and which, as we argued in section 5 above, are on their way to becoming part of a resource pool that is also used in offline discourse. In order to gain a full understanding of such phenomena, we need further studies which systematically compare the forum members’ online practices with their language use in offline interaction.

References Androutsopoulos, Jannis 2006: Introduction: Sociolinguistics and computer-mediated communication. Journal of Sociolinguistics 10 (4): 419–438. Androutsopoulos, Jannis 2010: The study of language and space in media discourse. In: Peter Auer & Jürgen E. Schmidt (eds.), Language and Space: An International Handbook of Linguistic Variation. Volume I: Theory and Methods, 740–758. Berlin/ New York: de Gruyter. Appadurai, Arjun 1996: Modernity at Large: Cultural Dimensions of Globalisation. Minneapolis: University of Minnesota Press.

Vernacular and multilingual writing in mediated spaces

555

Auer, Peter & Pfänder, Stefan 2011: Constructions: Emerging or emergent? In: Peter Auer & Stefan Pfänder: Constructions: Emerging and Emergent, 1–21. Berlin/Boston: de Gruyter. Baron, Naomi S. 2008: Always On: Language in an Online and Mobile World. Oxford: Oxford University Press. Beck, Rose Marie 2010: Urban languages in Africa. Africa Spectrum 54 (3): 11–41. Bhatt, Rakesh M. 2010: Unraveling post-colonial identity through language. In: Nikolas Coupland (ed.), Handbook of Language and Globalisation, 520–539. Malden Mass.: Blackwell. Blommaert, Jan 2010: The Sociolinguistics of Globalization. Cambridge: Cambridge University Press. Breyer, Thiemo, Oliver Ehmer & Stefan Pfänder 2011: Improvisation, temporality and emergent constructions. In: Peter Auer & Stefan Pfänder, Constructions: Emerging and Emergent, 186–216. Berlin/Boston: de Gruyter. Calvet, Louis-Jean 2006: Towards an Ecology of World Languages. Cambridge: Polity. [French original: Pour une écologie des langues du monde, Paris: Plon, 1999] Cheshire, Jenny, Sue Fox, Paul Kerswill & Eivind Torgersen 2008: Ethnicity as the motor of dialect change: Innovation and levelling in London. Sociolinguistica 22: 1–23. Chamberlain, Mary 1998: Introduction. In: Mary Chamberlain (ed.): Caribbean Migration: Globalised Identities, 1–16. London: Routledge. Coupland, Nikolas (ed.) 2010: The Handbook of Language and Globalisation. Oxford: Blackwell. Coupland, Nikolas 2003: Sociolinguistic authenticities. Journal of Sociolinguistics 7(3): 417–431. Coupland, Nikolas 2007: Style: Language, Variation and Identity. Cambridge: Cambridge University Press. Coupland, Nikolas 2009: Response: the mediated performance of vernaculars. Journal of English Linguistics 37: 284–300. Coupland, Nikolas 2010a: The authentic speaker and the speech community. In: Carmen Llamas & Dominic Watts (eds.), Language and Identities, 99–112. Edinburgh: Edinburgh University Press. Coupland, Nikolas 2010b: Language, ideology, media and social change. In: Karen Junod & Didier Maillat (eds.), Performing the Self. (SPELL, Swiss Papers in English Language and Linguistics), 55–78. Tübingen: Narr. Crystal, David 2006: Language and the Internet. 2nd edition. Cambridge: CUP. Danet, Brenda, & Susan C. Herring (eds.) 2007: The Multilingual Internet: Language, Culture, and Communication Online. Oxford: Oxford University Press. Dankel, Philipp 2013: Welche Erfahrung zählt? Kategorien und/oder Frequenzen im Sprachkontakt Spanisch-Quechua. Ph.D. dissertation, University of Freiburg. De Féral, Carole 2006 Décrire un ‘parler jeune’: le cas du camfranglais (Cameroun). Le Français en Afrique 12: 212–240. Deppermann, Arnulf 2007: Stilisiertes Türkendeutsch in Gesprächen deutscher Jugendlicher. Zeitschrift für Literaturwissenschaft und Linguistik 148: 43–62. Eckert, Penelope 2008: Variation and the indexical field. Journal of Sociolinguistics 12: 453–476. Gadet, Françoise, Ralph Ludwig & Stefan Pfänder 2008: Francophonie et typologie des situations. Cahiers de Linguistique: Revue de Sociolinguistique et de Sociologie de la langue française : 143–162.

556

Christian Mair & Stefan Pfänder

Heller, Monica 2003: Globalisation, the new economy, and the commodification of language and identity. Journal of Sociolinguistics 7: 473–492. Heller, Monica 2006 [1999]: Linguistic Minorities and Modernity: A Sociolinguistic Ethnography. 2nd edition. London: Continuum. Hinrichs, Lars 2006: Code-Switching on the Web: English and Jamaican Creole in E-mail Communication. Amsterdam: Benjamins. Johnstone, Barbara 2001: Individual. In: Alessandro Duranti (ed.): Key Terms and Language Culture, 122–125. London: Blackwell. Keim, Inken 2007: Formen und Funktionen von Ethnolekten in multilingualen Lebenswelten – am Beispiel von Mannheim. Zeitschrift für Literaturwissenschaft und Linguistik 148, 89–112. Mair, Christian & Véronique Lacoste 2012: A vernacular on the move: towards a sociolinguistics of mobility for Jamaican Creole. Cahiers de linguistique 38(1): 87–110. Meyerhoff, Miriam 2002: Communities of practice. In: Jack Chambers, Peter Trudgill & Natalie Schilling-Estes (eds.), Handbook of Language Variation and Change, 526–548. Oxford: Blackwell. Olbertz, Hella 2007: Dizque in Mexican Spanish: the subjectification of reportative meaning. Italian Journal of Linguistics 19(1): 151–172. (Special issue on Evidentiality between lexicon and grammar, ed. by Mario Squartini). Pfänder, Stefan 2009: Französisch sprechen heute: Aktuelle Perspektiven auf die europäische, amerikanische und afrikanische Frankophonie. In: Rolf G. Renner & Fernand Hörner (eds.), Deutsch-französische Berührungs- und Wendepunkte, 117–126. Freiburg: Frankreichzentrum. Pfänder, Stefan & Jörg Wagner 2010: Warum wir sprechen wenn wir klicken. In: Rolf Kailuweit & Stefan Pfänder (eds.), FrankoMedia: Französische Sprach- und Medienwissenschaft – ein Aufriss, 95–102. Berlin: Berliner Wissenschafts-Verlag. Pfänder, Stefan, in collaboration with Juan Ennis, Mario Soto & España Villegas 2009: Gramática mestiza: Presencia del quechua en el castellano, 2 Vol. La Paz: Academia Boliviana de la Lengua/Editorial Signo. Rampton, Ben 1995: Crossing: Language and Ethnicity among Adolescents. London: Longman. Traugott, Elizabeth Closs & Richard Dasher 2001: Regularity in Semantic Change. Cambridge: Cambridge University Press. Travis, Catherine E. 2006: Dizque: A Colombian evidentiality strategy. Linguistics 44(6): 1269–1297. Vertovec, Steven 2007: Super-diversity and its implications. Ethnic and Racial Studies 30: 1024–1054.

Pointing within the abdomen: Local deixis under restricted conditions

557

Susanne Uhmann

Pointing within the abdomen: Local deixis under restricted conditions*

0.

Introduction

Within the growing body of empirical research on deixis in situated talk-ininteraction (see, among others, Goodwin 1994, 2003; Koschmann et al. 2001, 2010; Mondada 2002, 2003a, 2003b, 2007; Enfield 2003; Koschmann & LeBaron 2003; Hanks 2005; Fricke 2007; Stukenbrock 2009), this paper aims at explicating the practical, interactional and multimodal circumstances of the production of the local deictic expressions hier (‘here’) and da (‘there’) used by surgeons while carrying out a laparoscopic cholecystectomy. The analysis thus examines a particular setting, which imposes severe restrictions on the practice of pointing accompanying deictic expressions. The laparoscopic technique allows the operation to be performed by making the anatomy inside the patient’s body visible on a TV monitor, but this two-dimensional, enlarged, media-transmitted space has to be re-converted into the three-dimensional abdominal space for performing the surgery and enacting local deixis. The main focus of this paper will be on the “gestural use of deixis” (Fillmore 1982). This means that pointing gestures can only be executed by ‘extensions’ of the body, i.e. by the laparoscopic instruments, whose manoeuvrability is severely constrained (as will be shown in this paper). There is already a considerable amount of research on laparoscopic surgery (cf. especially Koschmann & LeBaron 2003, Koschmann et al. 2001, 2010; Mondada 2002, 2003a, 2003b, 2007), which includes detailed case analyses of deixis and pointing. Some aspects of this research will also play a role in the following discussion, such as, and most importantly “situated practices of reference” (Koschmann et al. 2010), “interactive accomplishment of visibility” (Mondada 2003b) and the intertwinement of surgical activities, and demon* I wish to thank firstly Dr. Rahul Sengupta, the chief visceral surgeon of the Johanniter Hospital at Radevormwald, for allowing ‘his’ operating theatre to become an object of research, and the whole team for supporting me and making me feel highly welcome. I am also very grateful to Colin Foskett for correcting my English and to Peter Auer for helping me see the forest through the trees. Work-in-progress versions of this paper were presented during the FRIAS Workshop “Language and Interactional Spaces” (2009) and at ICCA 2010, Mannheim.

558

Susanne Uhmann

strating or instructing activities (Koschmann and LeBaron 2003, Koschmann et al. 2001, 2010; Mondada 2002, 2003a, 2003b, 2007). However, the main focus of this paper, the multimodal practices of local deixis in conjunction with or without pointing gestures, has not yet been systematically described. The database of this paper consists of four operations performed on four different days by different teams in the course of the most crucial stage of every laparoscopic cholecystectomy: the dissection of the relevant vessels before the team has to come to a decision on whether to start clipping and cutting or whether there is still some doubt about their correct identification. In this stage of the operation two of the three local deictic expressions in German, namely hier (‘here’) and da (‘there’), are frequently used. After giving some research findings about spatial deixis and pointing in section 1, some ethnographic information about laparoscopic cholecystectomy and the surgical field will be provided in section 2. Section 3 starts with the quantitative analysis of in total 79 instances of local deictics. The analysis of the multimodal practice of local deixis in the working space ‘abdomen’ focuses on the description of different pointing practices used by OS and AS1, depending on whether their instrument is free to perform pointing or not. Finally, possible denotational differences of the two local deictics hier (‘here’) and da (‘there’) are investigated.

1.

Local deixis in German

In the introduction to this paper, the German deictic adverbs hier und da were translated as here and there respectively. A detailed discussion of these deictic adverbs is far beyond the scope of this article1, but a few remarks are necessary to highlight the most important findings and some problems that arise out of them for the glossing of my data. Since Bühler (1934/1965), deictic expressions (Zeigwörter) and the field of deixis (Zeigfeld) have been conceptually understood as organized on the basis of a context-dependent, speaker-oriented and egocentric Hier-Jetzt-Ich-Origo (“hic-nunc-ego origo”), which is treated by the co-participants as a kind of anchorage point or “ground zero” (Levinson 2004: 111) in the speech event. The most basic way of pointing is what Bühler calls “objective pointing” (sachliches Zeigen),2 and in this mode speakers refer to entities in their extralin1

2

Klein 2001, Mühlhäusler 2001 and Levinson 2004 provide excellent review articles. The two other deictic modes are anaphoric use and deixis ad phantasma. Other terms in the literature on deixis for objective pointing are “exophoric” or “situational” deixis.

Pointing within the abdomen: Local deixis under restricted conditions

559

guistic environment by “demonstratio ad oculos et ad aures”. The reference space is the shared perceptual field of the speaker and the addressee. In a shared perceptual field, hier functions as a “positional signal” (Positionssignal): “Streng genommen wird mit hier die momentane Position des Sprechers angezeigt und diese Position kann mit jedem Sprecher und mit jedem Sprechakt wechseln” (Bühler 1934/1965: 103).3 According to the literature on the three local (or spatial) deictic expressions hier, da, dort in German (cf. especially Ehrich 1982, 1992), the positional deictic term hier is commonly treated as denoting some space around the origo, a proximal4 – though highly variable – region including (but sometimes even excluding) the speaker. Dort denotes some space not including the origo, a more distal region out of the (visual) reach of the speaker. The denotation of da, the third German local deictic adverbial, is much less clear.5 Ehrich (1982: 62) claims that da is “the least constrained (…) most neutral element within the system of German local deixis”. Da stands – according to Ehrich (1992: 21ff) – in double scalar opposition to hier and dort. On both scales, < dort, da > and < hier, da >, da is the weaker element. Thus via generalized conversational implicatures da signifies “nicht dort” (‘not there’) and “nicht hier” (‘not here’).6 Although Ehrich (1992: 20ff) does not claim that da is in a strict sense a tu-centric term like Latin istic (it has in fact a variable interpretation, namely either position of the speaker, or position of the addressee, or an independent third position), she assumes prototypical denotations or “preferred readings” for the three local German deictics which are closely linked to participant roles: hier : where I (the speaker) am, but not you (the addressee) da : where you are, but I am not dort : where neither I nor you are.

x x x

3

4

5

6

“Strictly speaking here indicates the position of the speaker, and this position can change with every speaker and every speech act” (Bühler 1990: 119). Haase (2001: 762) sees iconicity involved, because cross-linguistic analysis of local deixis reveals that typically “the degree of vowel closeness is inversely proportional with the deictic distance”. Cf. also Bühler’s very tentative gloss “Unser da (…) wird (…) mit Vorliebe für das sofort Erreichbare (…) gebraucht; (…). Das dort (…) hebt sich von dem da als Hinweis auf etwas nicht mehr im augenblicklichen Griffbereich oder Schrittbereich oder Blickbereich (…) des Sprechenden Befindliches ab.” (1934/1965: 100f). Cf. also Rauh’s 1983: 18f distinction between three major regions: “coding place” (= hier), “in connection with the coding place” (= da), and “not in connection with the coding place” (= dort).

560

Susanne Uhmann

According to Ehrich (1992) these local adverbs belong to the system of “positional” deictic terms, which may combine with “dimensional” terms such as horizontal vor/vorn – hinter/hinten (‘front – back’), rechts – links (von) (‘left/right (of)’), and vertical oben/drüber – unten/drunter (‘up/above (that)/ over (that)’ – ‘down/below (that)/underneath (that)’). The use of dimensional terms may depend both on the orientation of participants in the speech event and on the nature and the features of the reference objects. If objects are asymmetrical, i.e. if they have back or front features that remain constant wherever they are located and wherever the interlocuters are located (such as cars or vertebrates do), an “object-intrinsic” rather than the speaker-oriented and egocentric perspective may be used for locating referents.7 Finally, German also has “dynamic” local deictic expressions (cf. Klein 2001: 585, Vater 1996: 45) such as hierhin, hierher, dahin, dorthin (‘hither, thither’). Following Bühler (1934/1965) and Fillmore (1972, 1982), Levinson (1983: 65) distinguishes between a “gestural usage” and a “symbolic usage” of deictic expressions. Symbolic usages of deictic terms require the shared knowledge of the origo, whereas “gestural usages require a moment by moment physical monitoring of the speech event”, therefore they “can only be interpreted with reference to an audio-visual-tactile, and in general a physical, monitoring of the speech event.” Levinson (2004: 111) stresses that “gesture” has to be understood in the widest sense, which includes not only pointing with the hands but also with the lips and eyes and “even vocal intonation can function in a ‘gestural’ way.” In everyday face-to-face conversation adults routinely employ index-finger pointing8 when using gestural deixis in talk about entities in their extra-linguistic environment. Following Kita (2003: 1) “the prototypical pointing gesture is a communicative body movement that projects a vector from a body part. This vector indicates a

7 8

Cf. Vater 1996: 50ff for a detailed discussion of intrinsic orientation. “In pointing, the index finger and arm are extended (…), whereas the remaining fingers are curled under the hand, with the thumb held down and to the side (Butterworth 2003: 9).” Cf. also the G-sign in ASL. Butterworth 2003: 18f also claims that the pointing gesture and the pincer grip (= precision grip, cf. section 3), for holding something between the tip of thumb and the tip of index finger (cf. also the F-sign in ASL), “are coevolved but different aspects of hand function that are specialized, respectively, for precise instrumental action and precise communication. The characteristic hand posture observed in human pointing may be related to the pincer grip but as its ‘antithesis’.” Haviland 2004: 205 however stresses the “pure convention” and Fricke 2007: 155 the formal variability of the pointing gesture. For the anatomical details of the human hand cf. Streeck 2009.

Pointing within the abdomen: Local deixis under restricted conditions

561

certain direction, location or object.” Studies on early communicative development have shown that even preverbal children point (Masataka 2003), and that pointing gestures and language acquisition emerge in close connection and seem to depend on each other (Butterworth 2003, Haviland 2003). Ehrich (1982: 43) notes that da, the most variable of the German local deictics, “belongs to the first words children learn to use when they acquire German as their mother tongue.”9 But the fact that pointing is ubiquitous need not necessarily lead to the assumption that it is also a trivial simple procedure. On the contrary – only detailed multimodal empirical analysis (cf. among others Goodwin 2003, Hindmarsh and Heath 2000, Stukenbrock 2009) has revealed the enormous amount of intra- and interpersonal coordination (cf. Deppermann and Schmitt 2007) in the co-design of deictic gesture and verbal deixis required for the smooth and therefore mostly unnoticed accomplishment of gestural deixis in face-to-face interaction. Maybe not surprisingly, the linguistic literature on gestural deixis is not very explicit when it comes to the pointing gesture itself. Levinson (1983: 80) gives as a gloss for the gestural usage of here “the pragmatically given space, proximal to speaker’s location at CT [coding time, S.U.], that includes the point or location gesturally indicated.”10 The modifying aspect indicated by “pragmatically given” is the crucial part of Levinson’s gloss. An utterance like (1) (1) Place it here. will thus have rather “different implications of precision if said to a crane operator in contrast to a fellow surgeon”. Levinson’s (1983: 80) convincing argument rests upon two crucial assumptions. Firstly, that the fellow surgeon will be able to precisely and unambiguously identify the intended target via vectorial prolongation of the pointing gesture, and secondly, that the surgeon, who gives the instruction, is able to execute a pointing gesture. The former might be an oversimplifying assumption (cf. Stukenbrock 2009: 305ff) and the latter cannot be taken for granted in laparoscopic surgery as we will see in section 3. 9

10

Da is, according to Ehrich 1982, also one of the most frequent words in spoken language. While its frequency in adult speech may be due to its highly polysemous character (cf. Ehrich 1982: 48) its early production may be a result of its phonetic simplicity and its combinability with pointing gestures and various communicative intentions. Cf. also Haviland 2003: 144ff, who notes that taj ‘there’ is the first and so far only deictic in the active lexicon of a 3-year-old Tzotzil-speaking girl. His gloss for the symbolic usage of here: “the pragmatically given unit of space that includes the location of the speaker at CT” (Levinson 1983: 79).

562

Susanne Uhmann

At this point, a short excursus on the cross-linguistic differences between German hier and da and English here and there is considered essential. In the German translation (cf. Levinson 2000: 87) the example (1) above is translated as: (2) Setzen Sie hier an. This seems to be unproblematic if we take as our starting point a system of correspondences between the German threefold and the English twofold system as in Ehrich (1982: 49): (3) German hier da

English here there

dort

But as Lenz (2001: 41) has pointed out, da can be used in contexts where there is not an option. A German-speaking chairperson who wants to express his/ her gratitude to the audience at a conference for staying and listening to a long and utterly boring paper could choose between two options in (4), whereas an English speaker can use only one deictic expression (5) in this context: (4) Danke, dass Sie hier/da geblieben sind. (5) Thank you for staying here/*there. Cross-linguistically, the frequently discussed telephone examples (cf. Fillmore 1972, Ehrich 1992, Lenz 2001), where speaker and addressee only share an acoustic space but not a physical one, also show a relevant difference between German and English: (6) Ist Hans da/*dort? (7) Is John there? These findings assign an intermediate status to da as it corresponds only in some contexts to there but in others to here (cf. Lenz 2001: 41):

Pointing within the abdomen: Local deixis under restricted conditions

(8) German hier

563

English here

da there dort Further evidence for cross-linguistic differences in gestural use, which is the main focus of this paper, can be found in Bühler’s work. In his remarks on case-to-case differences between da and dort Bühler (1934/1965: 100) gives the following description of usage: Wenn ein Kranker am eigenen Körper dem Arzt eine schmerzende Stelle zeigen will, so wird er mit da berühren, was er erreichen kann, und unter Umständen mit dort fortfahren auf eine ihm im Augenblick unerreichbare Körperstelle hinzuweisen. (Bühler 1934/1965: 100).

This is translated into English (cf. Bühler 1990: 115) as follows: When a sick person wants to show the physician where it hurts, he will touch what he can reach, saying da (here) and probably continue with dort (there), indicating another part of his own body that he cannot now reach.(cf. Bühler 1990: 115).

The translator comments on his translation (cf. fn 8, page 115): “There is no single English word that could translate da in most contexts, it varies between here and there.” Thus, two caveats have to be formulated for the English glossing of the transcripts in section 3. Hier will be translated as here, and as dort does not occur in the data, da will always be translated as there just in order to make the use of two different items in German more prominent. The English glosses of German transcripts in section 3 thus do not aim at an exact translation but try to give a rough idea of the ongoing verbal interaction.

2.

The surgical field and laparoscopic cholecystectomy

As this paper will follow the tradition of empirically grounded research on deixis, some ethnographic background information will be necessary to highlight the communicative practices of local deixis in the surgical “field”. In his approach to deictic practice Hanks (2005: 192ff) relies essentially on Bourdieu’s social practice theory and emphasizes the “embedding of language in social fields”. Fields in this sense are distinctive kinds of context, and “embedding” according to Hanks “converts abstract positions like Spr, Adr, Object, and the lived space of utterances into sites in which power,

564

Susanne Uhmann

conflict, controlled access, and the other features of the social fields attach.” Following this approach we will have to show how local deixis is embedded in the surgical field. It was noted in section 1 that for certain referents the use of dimensional local terms such as horizontal vor/vorn – hinter/hinten (‘front – back’), rechts – links (von) (‘left/right (of)’), and vertical oben/drüber – unten/drunter (‘up/ above (that)/over (that)’ – ‘down/below (that)/underneath (that)’) may be problematic because speakers can either choose the speaker-oriented and egocentric “Ich-hier-jetzt-Origo” or an object-intrinsic orientation. In order to avoid ambiguity and misidentification in the surgical field, all medical staff are required to always use the intrinsic perspective based on the patient’s body. The resulting anatomical local expressions are defined according to the intrinsic dimensions of a vertebrate body: A coronal plane runs through the patient’s body from cranial to caudal, a transverse plane passes from dexter to sinister, and a sagittal axis spears from dorsal to ventral. Instead of an ambiguous hier oben (‘up here’) a surgeon would say hier cranial or hier ventral. The data considered in this paper is video data recorded and used by surgeons while carrying out a laparoscopic cholecystectomy (Lap Chole) – the removal of the gallbladder. The gallbladder is a pear-shaped organ of varying size situated in a fossa on the liver undersurface. It is supplied with blood by the arteria cystica (a branch of the arteria hepatica dextra, which supplies the liver with blood) and empties through the ductus cysticus into the ductus

Fig. 1. Calot’s triangle (adapted form Kienzle & Wuchter 1984: 20)

Pointing within the abdomen: Local deixis under restricted conditions

565

choledochus. Arteria cystica and ductus cysticus form – together with the inferior surface of the liver – Calot’s triangle. It is only after the team has a clear view and has agreed on the identification (cf. Koschmann et al. 2010) of the arteria cystica and the ductus cysticus that the two vessels can be clipped and cut, and the cholecystectomy can proceed without major postoperative complications. But because of considerable anatomical variation of the extrahepatic biliary system (cf. Kienzle and Wuchter (1984: 17), even for experienced surgical teams each operation is a new challenge and very little can be taken for granted. On the other hand, Lap Chole is a highly standardized operation. It requires several very small incisions in the abdomen to allow the insertion of operating ports (trocars), small cylindrical tubes through which surgical instruments and the laparoscope are inserted into the abdominal cavity. The number of operating ports depends on the technique used by the surgical team. The laparoscope both illuminates the surgical field and sends a magnified image from inside the body to a video monitor, providing the surgical team with a close-up view of the organs and tissues. The surgical team closely watches the monitor and performs the operation by manipulating the surgical instruments through the operating ports. The whole surgery is stored as a certain Operationsverfahren (‘procedure’), a kind of script knowledge11 shared by all the team members. Its intricate course of distributed tasks and actions can be regarded as a prototypical example of “intra- and interpersonal coordination” (cf. Deppermann & Schmitt 2007) or “center of co-ordination” (cf. Suchman 1997). Lap Chole consists of six major substages. The successful completion of every stage is checked meticulously before the team continues with the surgery: (1) creation of the pneumoperiteneum (inflation of the abdominal cavity with carbon dioxide) (2) insertion of ports, initial inspection of the operating field (3) dissection of Calot’s triangle, clipping and dividing of the arteria cystica and the ductus cysticus (4) detachment of the gallbladder from the liver bed (5) extraction of the gallbladder (6) final inspection, removal of the instruments, closure of the port wounds 11

A detailed description of the procedure and its anchoring as script knowledge as part of the background knowledge shared by the team members as medical experts is given in Uhmann 2010. Cf. Koschmann et al. 2010: 523 for a detailed analysis of performing surgery as “doing a procedure”.

566

Susanne Uhmann

But there is also considerable variation, depending on the operation technique used by the surgical team. The choice of technique has major consequences for team size, positioning and the division of labour between the team members. The procedure recorded here requires four surgical team members, four ports und a set of different surgical instruments depending both on the phase of the procedure and the task performed during the operation. The composition of the team is a result of duty-rosters, which change weekly, and daily schedules, and in case of a Lap Chole four team members act as: OS:

The operating surgeon performs the surgery. He/she holds the dissection instruments (curved dissection forceps or hook) in his/her right hand and operates the laparoscope with his/her left hand. Via a foot switch he/she can put the dissecting instruments under current for dissection and for coagulation. AS1: The first assistant uses grasping forceps to hold the gallbladder under constant tension and moves the grasper slightly backwards and forwards and side to side so that the OS has optimal vision of and access to the anatomical area. AS2: The second assistant holds a double-duty instrument. The suction/irrigation cannula rinses water into the abdomen and removes the blood and water to improve the surgeon’s vision, but it is also used to move intestines around and especially to elevate the liver. SN: The scrub nurse is responsible for the surgical instruments and hands them to the surgeon when needed. OS is what Deppermann and Schmitt (2007: 35), following Goffman (1983), have termed the focal person (Fokusperson) of an interactional ensemble (Interaktionsensemble), because he/she is the centre of attention and the target of continuous monitoring activities. Every member of the interactional ensemble closely watches all his/her verbal and non-verbal actions. This feature of OS will be of particular importance for the empirical analysis of local deixis in the next section. One must bear in mind that there is a certain degree of interdependence between the surgical activities or task roles12 in Lap Chole and the social 12

Cf. Uhmann 2010: 46ff for the analysis of “membership categorization” (cf. Sacks 1992: Lecture 13, vol. 1, Schegloff 2006) in the surgical field via dress code, positioning in the operating theatre, verbal and non-verbal category bound activities, and use of category bound tools.

Pointing within the abdomen: Local deixis under restricted conditions

567

Fig. 2. Aseptic and septic working spaces in the OR.

roles of the participants that depend on their status in the internal hierarchy of the hospital. OS and AS1 must be fully-qualified physicians. But OS may be either an Assistenzarzt (a qualified physician who has not yet been accepted as a full member of the surgical profession), a Facharzt (a fully-qualified surgeon), an Oberarzt (a highly experienced surgeon and senior physician) or the Chefarzt (the head of department). And this applies in exactly the same manner to the surgical role of AS1. SN, however, is always a member of the nursing staff, whereas AS2 may be either an Assistenzarzt or medically trained person substituting a physician. But irrespective of the role allocation in a particular operation, an Oberarzt or the Chefarzt must be present or on call, for he/she bears legal and medical responsibility for proper surgical treatment of the patient.

568

Susanne Uhmann

As the patient receives general anaesthesia, an anaesthetist (AN) and an anaesthetic nurse (ANN) are also present. A circulating nurse (CN) may but need not be present all along the operation. Figure 2 shows not only the positioning of the staff but also the two working spaces of every operating theatre: the septic and the aseptic space. Mobility in the aseptic space is rigidly restricted and the rules of asepsis distinguish two types of team members: those team members who constantly have to avoid coming into contact with anything non-sterile (OS, AS1, AS2, SN), and those who constantly have to avoid coming into contact with anything sterile (AN, ANN, CN) but are at the same time the only links between the outer septic space and the inner aseptic space (cf. also Katz 1981). The working space in which the members of surgical team members coparticipate is not only the septic and aseptic space, however, but most importantly it is a third space, namely the ‘abdomen’. Therefore deixis in the surgical field can be studied not only as a situated multimodal activity within the septic and the aseptic interactional space but also within this third space, which will be the focus of this paper.

3.

Local deixis in the working space ‘abdomen’

3.1

Quantitative distribution of local deictic elements

The database consists of four operations performed on four different days by varying teams. As mentioned in section 2, there is some independence between the surgical task roles (OP, AS1, AS2, SN) and the social roles of the participants. The latter depend largely on experience and knowledge. In the operations that form our database, the allocation of task roles and social roles shows considerable variation: Lap Chole I

Lap Chole II

Lap Chole III

Lap Chole IV

OS

Oberarzt X

Assistenzarzt X

Assistenzarzt X

Chefarzt Y

AS1

Chefarzt X

Chefarzt Y

Chefarzt Y

Oberarzt X

AS2

trainee X

trainee Y

nurse X

nurse X

SN

nurse Y

nurse Z

nurse Z

nurse Y

As can be seen, Lap Choles II and III are instances of apprenticeship training (“scene of instruction”, Koschmann et al. 2010: 525) that includes the performance of gallbladder surgery. This is due to the fact that surgical training for inexperienced young surgeons, Assistenzärzte, is not restricted to the big

Pointing within the abdomen: Local deixis under restricted conditions

569

university hospitals in Germany but forms a part of everyday hospital routine work. The paper focuses on the third stage of laparoscopic cholecystectomy during the dissection of the relevant vessels and before the team has to decide whether to start clipping and cutting or whether there is still some doubt about the correct identification of the arteria cystica and the ductus cysticus. In this phase deictic terms occur in every operation (in total 79 instances)13 and striking accumulations were observed during some operations. But the three deictic terms are not equally represented: (9) Quantitative analysis of hier, da and dort hier

da

dort

25

55

0

Da is more than twice as frequent as hier, and dort does not occur at all. Medical terminology implying a patient-oriented intrinsic perspective such as, dorsal/ventral, cranial/caudal, or dexter/sinister is never used. However, hier and da were frequently combined with dimensional deictics like in hier oben (‘up here’) or da unten (‘down there’) etc. A closer look at the data reveals that the use of local deictics is not evenly distributed across the four operations. As diagram (10) shows, they are massively overrepresented in Lap Chole II and III, but with the interesting difference that in Lap Chole II da is over three times more frequent than hier. (10) Distribution of hier and da

13

hier

da

Lap Chole I:

26. 10. 04

2

1

Lap Chole II:

07. 12. 04

10

30

Lap Chole III:

04. 01. 05

13

19

Lap Chole IV:

17. 01. 05

0

4

I excluded from the analysis all instances of textual deixis involving hier and da and all cases where the accuracy of the transcription is in doubt. The latter which is quite common in the surgical field because speaking in very low voice seems to be a significant part of the professional culture and an important device in the recipient design (cf. Uhmann 2010). Cf. also fn 14 for the absence of talk.

570

Susanne Uhmann

Additionally, diagram (11) shows that only the operating surgeon (OS) and the first assistant (AS1) produce local deictic terms. (11) Speakers of hier and da

OS

hier

da

6

5

Lap Chole I:

26. 10. 04

1

0

Lap Chole II:

07. 12. 04

4

2

Lap Chole III:

04. 01. 05

1

2

Lap Chole IV:

17. 01. 05

0

1

19

49

AS1 Lap Chole I:

26. 10. 04

1

1

Lap Chole II:

07. 12. 04

6

28

Lap Chole III:

04. 01. 05

12

17

Lap Chole IV:

17. 01. 05

0

3

This finding is in accordance with the fact that the other two members of the surgical team, AS2 and SN, hardly speak at all during this stage of the surgery, because a silently working team is a major part of the professional culture.14 And although operations are prototypical examples of what Bühler calls empraktische Rede and Goffman (1981: 134f) “conversation (…) subordinated to an instrumental task at hand”, the right of self-selection for the next speakership is not evenly distributed among the team members. Under the notion of “recipient design” Sacks, Schegloff and Jefferson (1974: 727) have drawn attention to the “multitude of respects in which the talk by a party in a conversation is constructed (…) in ways which display an orientation and sensitivity to the particular other(s) who are co-participants.” It is therefore interesting to have a look at the distribution of the “addressed recipients” (cf. Goffman 1981) of hier and da from this point of view.

14

This is at least partly due to objective needs to consider above all the patient’s safety, because talk increases the risk that the masks will become wet and lose their sterility. So for long periods of time the monotonously beeping noise of the anaesthetic device and the acoustic signal of activated diathermy are the only acoustic manifestations of ongoing interaction.

571

Pointing within the abdomen: Local deixis under restricted conditions

(12) Addressed recipients of hier and da Lap Chole I Lap Chole II Lap Chole III Lap Chole IV hier

da

hier

da

hier

da

hier

da

OS (by AS1)

0

1

4

28

0

4

0

3

AS1 (by OS)

0

0

4

2

1

1

0

1

AS2 (by OS)

1

0

0

0

0

2

0

0

Audience (by AS1) 1

0

2

2

12

11

0

0

Diagram (12) shows that apart from Lap Chole IV, a watching audience is also addressed by AS1. This is done significantly in Lap Chol III and can be explained by the fact that not only I was present as part of a watching audience in Lap Chole III but also the daughter of AS1Y, and we had asked AS1Y to comment on and explain the ongoing surgery. The quantitative analysis represented in diagrams (11) and (12) gives us some very informative findings we can draw upon for further analysis: – AS1 is by far the main producer of hier and da (68). – AS1 uses da more frequently than hier (49:19). – OS uses hier and da with nearly equal distribution (6:5). – In Lap Chole II the main addressee is OS. Da is overwhelmingly more frequent than hier (28:4). – In Lap Chole III the main addressee is a watching audience. Da and hier show nearly equal distribution (11:12) The imbalance in the use of hier and da when OS is the addressed recipient of AS1 is particularly striking and requires explanation. In order to achieve this, we will now turn to the interactional, multimodal analysis of local deixis. 3.2

Multimodal practices of local deixis

3.2.1 The working space “abdomen” The preparation of the working space “abdomen” is a complex team activity, but as it is not the focus of the analysis presented here, the major practices of the surgical team will simply be listed: OS inflates the abdominal cavity with carbon dioxide to create this working space; once all the four ports are in position (see section 2) AS1 grasps the infundibulum of the gallbladder and retracts it in order to expose Calot’s triangle; after the gallbladder has been grasped, AS2 elevates the liver with the suction/irrigation cannula; the oper-

572

Susanne Uhmann

ating table is then moved by a team member working in the septic space (CN, AN or ANN) so that the patient lies with a steep head-up position (AntiTrendelenburg) and left tilt; due to gravity the working space is thus further enlarged. But it is important to bear in mind that – in contrast to the septic and the aseptic space – the working space “abdomen” exists twice: as a rather small three-dimensional physical space inside the patient’s body (a space of about 10 cm between the elevated undersurface of the liver and the colon) and as a magnified two-dimensional, laparoscopically mediated, optical space on the monitor. What is seen on the monitor depends on OS, who establishes the “domain of scrutiny” (Goodwin 1994) by directing the laparoscope. Mondada (2002: 7) has already noted that one of the crucial features of teamwork in this working space is the fact that its participants are visually “present (…) by the instruments they hold”. This of course depends on the laparoscopic technique: In order to share the visually-mediated working space and to achieve mutual orientation inside the patient’s body, all the team members look not at each other or at the operating field but instead closely watch15 the monitor. Unlike in open surgery, all the team members can observe the instruments and their actions on the screen, which thereby become – together with the anatomical space in which they are moved – interpretable resources for the ongoing interaction. Surgical instruments are designed as highly efficient multi-purpose tools. The ones that are used in laparoscopic surgery mirror those used in traditional open surgery except that long shafts serve as “extensions” to the surgeons’ bodies. Thin prolongations of the surgeons’ arms connect to the handle of a working part, which serves as a tiny, highly effective – though anatomically “crippled” – hand. OS’s dissecting forceps function as the surgeon’s thumb and index finger – allowing for the “precision” or “pincer grip” (cf. footnote 8) – and have slightly curved and pointed dual-action jaws, which can be gradually spread to tear tissue apart, closed to grasp tissue, and set under current to burn tissue away and at the same time achieve coagulation. The grasping forceps of AS1 work in a similar way, but its cone points are blunt, only one finger jaw is moveable and it has a locking mechanism. AS2’s ‘hand’ is even more reduced, for the suction/irrigation cannula is just one long hollow rod.

15

Nardi et al. 1993: 331 noticed subtle differences in the way the team members watch the monitor. The members who share the working space watch “intently for long stretches of time”, whereas the scrub nurses tend to throw “quick glances”.

Pointing within the abdomen: Local deixis under restricted conditions

573

Fig. 3a. Dissecting forceps/Maryland (Foto: Aesculap AG)

Fig. 3b. Grasping forceps/crocodile forceps (Foto: Aesculap AG)

Fig. 3c. Sucction/irrigation-cannula (Foto: Aesculap AG)

During the third stage of the surgery the allocation of all these instruments is fixed: OS manipulates the dissecting forceps AS1 holds up the gallbladder with the grasping forceps AS2 elevates the liver with the suction/irrigation cannula Of these three actors, OS stands out as the focal person of the interaction (cf. section 2). As the surgical instruments represent the participants in this visually mediated working space, it can be expected that OS’s instrument, the

x x x

574

Susanne Uhmann

Fig. 4. OS, AS1 and AS2 in the visually mediated working space ‘abdomen’

dissecting forceps, will inherit this attribute and will act and be regarded as the focal instrument. Summing up, three surgical instruments, the dissecting forceps, the grasping forceps und the s/i cannula, partially replace the hands of OS, AS1 and AS2. It can therefore be expected that they will be brought into action not only as surgery tools but will perform additional tasks that have to be taken into account for the analysis of the multimodal practice of deixis in the surgical field. As the dissecting forceps are the focal instrument, we can expect them to be the tool of vital importance in multimodal practice and be monitored continuously by both AS1 and AS2 and made relevant by OS as an interpretable, “semiotic resource” (cf. Mondada 2003a: 68). 3.2.2 Pointing within the abdomen If gestural deixis and pointing have to be executed in the laparoscopic mediated working space ‘abdomen’ it cannot be done with the prototypical pointing gesture using the extended index finger. Nor can it be done with

Pointing within the abdomen: Local deixis under restricted conditions

575

lips, the chin or the viewing direction (cf. Levinson 2004: 111, section 1), because the lips of the surgeons are covered by their sterile masks and looking at the monitor would not be helpful at all for the precise identification of the target. Nevertheless, verbal turns resembling Levinson’s example (cf. section 1) do occur and without an accompanying pointing gesture they would be totally unintelligible: (13) Lap Chole I (8:34) -> 01

OP:

(-)

down

But gesturing and pointing with all sorts of things we hold in our hands (pens, chalk, rulers, laser pointers) is very familiar to us, and literature on the topic has increased our knowledge about the inventory of tools employable in the act of pointing: trowels (cf. Goodwin 2003: 221), hoes, sticks, machetes (cf. Haviland 2004: 208); surgical hooks and scissors (cf. Mondada 2003a: 68ff) and endo-graspers (cf. Koschmann et al. 2010: 530). Not surprisingly, OS’s turn in extract (13) is co-articulated together with a pointing gesture – and the pointing device used here is the substitute for her hand, the dissecting forceps (cf. figure 4): (13) Lap Chole I (8:34)

| -> 01 OS:

(-)

downwards

576

Susanne Uhmann

03 AS2:

[((repositions the irrigation cannula))

04 OS:

ja.

yes.

When OS begins her turn (in line 01), the jaws of her dissecting forceps are closed without containing any tissue and the instrument is positioned near the lower part of the gallbladder. The focal instrument, closely monitored by AS1 and AS2, is thus not dissecting. OS’s next action could be of course dissecting, but she instead keeps the jaws of the instrument closed, quickly moves the forceps, touches an upper part of the gallbladder and gently presses their tip into it. The “stroke” (cf. Kendon (1980) of her gesture is reached with the articulation of hier. The “retraction” is done with the same velocity and completed when the second intonational phrase of her turn in line 02 starts. The design of OS’s turn is exceedingly complex. It combines talk and gesture in a way that makes the movement of the forceps recognizable as a pointing gesture designed for identifying a target through gestural deixis. It is also a recognizable first part (request) of an adjacency pair (cf. Sacks, Schegloff and Jefferson 1974) that makes a non-verbal second part (requested activity) conditionally relevant. And finally, it is recipient-designed in a way that enables AS2 to recognize after the first two syllables that he is the addressed recipient without being addressed by name16, because his hand, the irrigation cannula, is requested to take action. This allows him to display his understanding of OS’s multimodally designed utterance (“action package”, Goodwin 2003) by getting his second part, the repositioning of the irrigation cannula, done in overlap17 with OS’s syntac-

16

17

In contrast to the data analysed by Mondada 2003a, the use of address terms seems to be dispreferred im my data (cf. Uhmann 2010: 61, fn 32), for address terms only occur in critique implicative turns. Fast reaction time of AS2 and SN are highly desired qualities in every surgical team and the lack of it a recurrent topic of complaints by operating surgeons. Optimal recipient design as in example (13) clearly improves the chances of being fast.

Pointing within the abdomen: Local deixis under restricted conditions

577

tically dependent but intonationally independent18 specification of the requested direction runter (‘downwards’). The sequence is closed by OS’s confirmation in line 04. But the pointing gesture in extract (13) is not the prototypical pointing gesture (cf. section 1), because OS touches the gallbladder. In Fillmore’s (1982: 45f) framework for the description of deixis, “touching” is one of the three subtypes of the indexical act of gestural deixis, but “indicating” is the prototypical pointing gesture: “(…) the performing of an act which allows the Hearer to trace, by symbolic extrapolation, a path from the gesture to the thing.” And as an example of the act of “presenting”, Fillmore cites “(…) holding a wine bottle in someone’s view and saying this bottle (…)” (Fillmore 1982: 46). The main reason for separating Touching and Presenting from Indicating is, according to Fillmore (1982: 46), “(…) that in these cases (…) the Hearer has no need to project from the Indexing act to find the indicated place”. Touching could be the preferred pointing gesture in the abdomen because it avoids the “vector-target problem” (cf. Stukenbrock 2009: 305) and ensures the desired amount of precision in the surgical field (cf. Levinson 1983: 80). As already noted by Mondada (2003a: 68) and Koschmann et al. (2010: 530), pointing within the optically mediated working space ‘abdomen’ cannot be done with all the surgical instruments. The distributed task roles of the surgical team make it impossible for the grasping forceps of the AS1 or the suction/irrigation cannula of AS2 to be used for enacting pointing gestures, as they have to sustain the three-dimensional physical working space inside the patient’s body. Only a complex manoeuvre requiring highly precisioned interpersonal coordination, called umgreifen (‘change of hold’), would allow AS1 to let go of the gallbladder after OS has grabbed it. And indeed, there was not a single instance in the data where a “change of hold” was done for the purpose of allowing the grasping forceps to point.19 So it might be expected that the only surgical instrument used for pointing is the dissecting forceps. However, one counterexample to the rule that only the dissecting forceps can be used for pointing was found in the data:

18

19

Cf. Uhmann 1991 and Uhmann 1997: 40ff for the format of intonational phrases in German. It was of course regularly done to ensure better access for dissecting Calot’s triangle.

578

Susanne Uhmann

(14) Lap Chole III (23:20) 01 AS1:

SO! (.)

okay! 02

dAnn (.) wolln=wa jetz: (-) ma richtig opeRIEren; (1.5)

then (.) let’s now start operating properly 03

06

07

=[womit er HIER die gAllenblase ] FASST;> (---)

with that he grabs the gallbladder here [((taps repeatedly the gallbladder with his grasping forceps))] u:nd der Herr SAbisch, ((etc.)) a:nd Mr Sabisch

After phase 1 and 2, the creation of the pneumoperiteneum via inflation of the abdominal cavity with carbon dioxide and the placement of the four trocars (cf. section 2), AS1 both closes the second phase and opens the next one with the particle so (‘okay’). He also formulates the beginning of the actual surgery in line 2. And with increased volume, a prosodic device methodically and regularly used for addressing recipients outside the aseptic space20 (in this case a watching audience standing behind AS2 in the septic space), AS1 comments on the allocation of the surgical instruments to AS1 and AS2. When AS1 starts his explanation in line 03, no instrument is visible yet in the laparoscopically mediated working space, because it has not yet been inserted through the trocar into the abdominal cavity (cf. section 2). The emerging verbal turn of AS1 thus lacks important visual information and this visual information remains absent also during the next intonational phrase in line 04, which mentions AS1 der Erste (‘the first’). Only after a short intra-sentential pause and synchronous with the syntactic completion of this unit of talk does the instrument, the grasping forceps, appear on the monitor 20

Instead of code-switching between French and English (cf. Mondada 2007: 62) low voice is used in my data to address the team in the aseptic space (OS, AS1, AS2 or SN) and increased volume to address everybody else in the septic space. A detailed prosodic analysis is outside the scope of this article, but it can be expected that changes in volume account for membership categorization and space organization (cf. also Mondada 2007: 63f).

Pointing within the abdomen: Local deixis under restricted conditions

579

and is brought – with closed jaws as in extract (13) – into a position from where its tip touches the gallbladder (cf. line 05). But contrary to the gesture analyzed in extract (13), AS1 does not perform retraction immediately after touching but starts to tap the gallbladder (Fig. 14c) repeatedly during his further explanation in line 06: (14) Lap Chole III (23:20)

| 05 AS1:

|

EI:ne (.) ZANge in die hAnd,=

a pair of (.) forceps in his hand

| -> 06

= [womit er HIER die gAllenblase ] FASST;> (---)

with which he grabs the gallbladder here [((taps repeatedly the gallbladder with his grasping forceps))

By means of a most intricate division of labour between talk and gestures, AS1 manages to perform a highly complex multi-duty turn. The main function of the instrument, the grasping of the gallbladder, is explained verbally with the focal accent on the activity FASST (‘grab’); the identification of the gallbladder is done multimodally by means of the local deictic here in conjunction with the touching gesture; and by repeating the gesture, the tapping21, AS1 simultaneously not only recipient-designs his action, taking into 21

Cf. also Hindmarsh & Heath 2000: 1868 on tapping.

580

Susanne Uhmann

account that the addressed recipients – the watching audience – might need some extra time to identify the gallbladder in the anatomical environment displayed on the monitor, but also informs himself and the surgical team about an important condition of the gallbladder, namely its elasticity. (A lack of elasticity would be a highly important feature, because as a result of chronic cholecystitis and inflammatory scarring, gallbladders sometimes calcify, which would have immediate consequences for the ongoing surgery). It is important to note that this instance of pointing with the grasping forceps in extract (14) is only possible because AS1 has not yet brought the forceps into their position (at the infundibilum of the gallbladder) or executed their main function, that is holding up the gallbladder (cf. section 2). Grasping the gallbladder is an indispensable action at the very beginning of stage 3 (when the surgical team collaboratively prepares the working space ‘abdomen’). But if AS1 cannot use the grasping forceps for pointing, how can he be by far the main producer of hier and da (68 instances, cf. 3.1, diagram 11)? How can AS1 enact gestural deixis without being able to point? To answer this puzzle we have to look at those instances in which AS1 is already holding the gallbladder with his grasping forceps and therefore cannot use it for pointing but nevertheless produces local deictics, as in the next extract (15). In extract (15) we now join the surgical team of Lap Chole II in the middle of phase 3 – that is still during their dissection work. A clear view and agreement on identification of the arteria cystica and the ductus cysticus have not yet been achieved. AS1 is holding up the gallbladder, thus he cannot point with the grasping forceps when he suggests in line 08 ich würd mal HIE:R son bißchen probieren (‘I would PART. try a bit here’). Remember that Lap Chole II is one of the two operations in which OS is an inexperienced Assistenzarzt and AS1 is the Chefarzt of the visceral surgery department. So how does OS find out where exactly he should continue to dissect? And how does AS1 make sure that OS does not make any wrong decisions?

Pointing within the abdomen: Local deixis under restricted conditions

(15) Lap Chole II (15:15) 01 OS:

I should be through ((movements of dissector)) (4.0) 02 AS1:

(oben) mal REINgehen und quer

enter Part (from above) and across (“professional vision”)

| 03

da sIEht man [doch;]

there you see Part 04 OS:

[ja

]

yes 05 AS1:

(Is) man doch DURCH. (-)

one is Part through 06

(nur) da DRUNter is noch was

but there underneath is something else (loss of “professional vision”)

| 07

and that I am not (quite) sure

581

582

Susanne Uhmann

| -> 08 AS1:

ich würd mal [HIE:R] son bisschen probieren.=

I would Part try a bit here 09 OS:

[(hm) ]

((camera zoom and pulling over of a vessel with the dissecting forceps))

| 10 AS1

=DA.aS.

that 11 OS

da drunter LÄUft nämlich irgendwas.

12 OS:

hmhm

because there is something running underneath

In line 01, OS says in a very low, hardly perceptible voice that he has reached some point of completion in his dissection (müsst ich eigentlich) durch sein. (‘I should be through’). AS1 does not confirm this guess by a simple acknowledgement token, but instead provides OS in line 02 with a strategy to check again. While OS does as he was told (he drags a vessel away from a ventral position with his closed dissecting forceps and also zooms the camera), AS1 reformulates OS’s guess of durch sein (‘being through’) in lines 03 and 04 as something which can be assumed safely after “professional vision” (Goodwin 1994) has been established: da sIEht man doch; (Is) man doch DURCH. (‘there you see PART. one is PART. through’). Professional vision is achieved when OS produces the acknowledgement token ja (‘yes’) in line 04 (Fig. 15a). But this professional vision has additionally directed AS1’s attention to another feature of the anatomy: (nur) da DRUNter is noch was

Pointing within the abdomen: Local deixis under restricted conditions

583

(‘but there underneath is something else’). He adds in line 07 that he is not absolutely sure what this was (‘something’) might be. This doubt calls for further dissection. But where should it be done? Goodwin (1994: 626) has already declared that “professional vision is (…) unevenly allocated” and has to be learned. And as Lap Chole II is in fact OS’s first laparoscopic cholecystectomy, AS1 has reason to assume that this was (‘something’) might have escaped OS’s inexperienced attention. But AS1 has to cope with additional problems. When he starts to formulate his suggestion in line 08, professional vision is not possible any more, because the hold on the vessel has been lost22, and an accompanying disambiguating pointing gesture with his grasping forceps is – as we know – not possible. Immediately after the enunciation of the local deictic hier in AS1’s suggestion ich würd mal HIE:R son bisschen probieren. (‘I would PART. try a bit here’) and his own acknowledgement token, OS starts to repeat the action that was successful in providing “professional vision” in the first instance. He zooms the camera and pulls the vessel over with the dissecting forceps. This unfolding action, zooming and pulling, is stopped by AS1 in line 10 with the neuter deictic demonstrative pronoun DA.aS. (‘that’), articulated with falling intonation as soon as this was (‘something’) once again comes into view. It might be noteworthy that AS1 still does not name exactly what he claims to see. No definite description is used, but instead the neuter demonstrative pronoun, which is maximally indexical. Thus any kind of categorization is avoided and postponed until after further dissection. We thus re-encounter a division of labour between talk and gesture but this time gestural deixis is jointly accomplished by two different actors: AS1 delivers the verbal turn (lines 8, 10) with the local deictic HIE:R and the object deictic DA.aS. OS provides a kind of pointing gesture, with both his hands working simultaneously. With his left hand he operates the camera like a searchlight, illuminating crucial parts of the anatomical space, and uses the zooming function for close inspection; with his right hand he pulls the vessel over, because the intended referent has to be found somewhere da DRUNter (‘there underneath’) in the three-dimensional space of the patient’s abdomen. Although illuminating, zooming and pulling something over is of course not a prototypical pointing gesture, in this mediated space the action resembles what (Fillmore 1982: 46) has termed “Presenting” (only that OS’s verbal turn does not provide any definite description for categorizing the presented object but a neuter deic22

It is unimportant for the analysis whether this loss of vision was a result of intentional action or due to unskilled handling of the closed and thus slippery dissecting forceps, because there is no explicit reference by any of the participants.

584

Susanne Uhmann

tic pronoun). But the joint action of gestural deixis enacted both by AS1 and OS is not as precise as the pointing-by-touching gestures in extracts (13) and (14). Firstly, the division of labour demands of AS1 the rapid perception of a film-like presentation together with the challenge to stop the screening at that precise moment. Secondly, presenting the picture of just a small anatomical region raises the vector-target problem that has already been mentioned. In extract (15) we can thus observe the limits of performing gestural deixis in the working space ‘abdomen’, when no surgical instrument is available for a pointing/touching gesture. AS1 could not point, because he was holding the gallbladder with his grasping forceps and OS had to use his dissecting forceps to pull a vessel over. But – as I will try to show now – in the majority of instances of gestural deixis the surgical team achieves better and more precise results in multimodal practice, and the key to their success is the use of the focal instrument, the dissecting forceps, as a semiotic resource. In extract (16) we again join the team of Lap Chole II during the beginning of phase 3. OS has been silently dissecting and tearing away connective tissue and fat for nearly four minutes in different places, and the monotonous beeping noise of the anaesthetic device and the acoustic signal of activated diathermy are the only acoustic manifestations of ongoing surgery. (16) Lap Chole II (11: 07) ((08:27–11. 07. OS dissects at several different points; fat keeps getting stuck between the jaws of his dissecting forceps))

((no dissection)) (3.0)

01 OS: -> 02 AS1:

ja DA;

[dA.

yes there there 03 OS:

[(dAs?)]

that -> 04

((touches a possible place with the dissecting forceps)) nee dA DRUNter.=

no there below

Pointing within the abdomen: Local deixis under restricted conditions -> 05

585

=DA: würd ich jetz[versUchen noch=en bißchen FREIzumachen.

there I would now try to expose just a bit more 06 OS:

07 OS:

[

((starts dissection:))

Starting with the affirmation particle ja (‘yes’) AS1’s turn in line 02 is designed as the second part of an adjacency pair. I want to argue that in this adjacency pair the first part is delivered non-verbally. What evidence do we have for this analysis? Immediately before this second pair part (cf. line 01) OS has stopped dissecting, the jaws of the forceps are closed and the instrument hovers nearly motionless above the area of dissection for about three seconds. I want to argue that OS is calling for help (non-verbal first part: where should I continue dissecting), and that AS1 displays his understanding (second part line 02: yes there). How can we account for this analysis? We have established that OS is the focal person, and that the dissecting forceps are the focal instrument and thus watched continuously by AS1, AS2 and SN. In Lap Chole II this becomes particularly evident because not the slightest action or non-action seems to escape AS1’s notice. As OS is an Assistenzarzt (still undergoing training) and AS1 is the Chefarzt (head surgeon), AS1 is responsible both for apprenticeship training and for the safe outcome of the operations in his area of responsibility. Thus, in Lap Chole II, OS is under the permanent supervision of AS1 for at least three reasons: OS’s apprentice role requires maximal enacting of his task role as AS1, as the instructor he has to impart knowledge and technical skill of how to perform laparoscopic dissection, and as the head surgeon his utmost attention is needed to ensure that OS does not make any mistakes that cannot be put right. In extract (16) AS1’s close monitoring of OS’s actions and non-actions has prompted AS1 to advise OS where to continue dissecting. It is important to keep in mind that OS’s task is far from simple, because drawings like Figure 1 (cf. section 2) show marked and separated organs and vessels (it is even common for textbooks to show the cystic vessels in green and the arterial vessels in red). Reality looks totally different and the goal of dissection is what surgeons call the darstellen (‘representing’) or the freimachen (‘exposing’) of the rel-

586

Susanne Uhmann

evant structures, here the arteria cystica and the ductus cysticus, by means of various dissecting techniques like spreading apart, grasping and tearing, cutting and burning away fat and layers of connective tissue. Dissecting therefore means transforming parts of the patient’s body into atlas-like anatomical configurations, and making the relevant structures visible. Especially in laparoscopic surgery, visibility is the prerequisite for identification and classification.23 The fact that OS is both the focal person and a novice supports the analysis, i.e. that the instruction sequence in extract (16) is initiated non-verbally by OS using his dissection forceps not as a surgical instrument but as a semiotic resource. The prolonged absence (about 3 seconds) of any movement that could count as dissecting (jaws closed, no spreading for tearing tissue apart, no grasped tissue, no coagulation) and the motionless hovering of the dissection forceps above the area of dissection shows that OS is non-verbally calling for help in deciding where he should continue dissecting, which is provided by AS1 when he realizes this in line 02 ja DA; dA. (‘yes there; there’). But a surgical novice, like OS in Lap Chole II, has to cope with two different and highly complex tasks. He not only has to learn “professional vision” (cf. also Mondada 2003a: 67f) in order to decide where further dissection is needed, but he also has to develop the technical skill to perform the dissection with his instrument without unintentionally injuring vital parts. Extract (16) shows how AS1 instructs OS to perform the surgery taking these two major educational objectives into account. How is this done? We observe that in extract (16) the local deictic da (‘there’) occurs repeatedly (line 2, line 4, line 5), and this gives the impression that AS1 is piloting OS through rather unfamiliar anatomical territory. If AS1 were holding a moveable instrument, that could perform a pointing gesture by touching (as in extracts 13 and 14), OS would be able to unambiguously identify the pointing target(s). But in Lap Chole II this is not possible because AS1 is holding up the gallbladder with his grasping forceps. However, unlike in extract (15), OS’s dissecting forceps are free, because nothing has to be pulled sideways to ensure clear vision of the anatomical area in question at this moment. In line 03 OS can therefore produce a multimodal turn in overlap with AS1’s syntactically and prosodially non-projectable second da (‘there’). By uttering the neuter deictic demonstrative pronoun with rising intonation dAs? (‘that’), he offers a possible solution, and by means of a touching movement with his closed dissecting forceps he offers a possible pointing target. AS1 can now verbally accept or reject this possible target in the subsequent turn without 23

Cf. Hirschauer 1991: 299ff and Koschman et al. 2010 for a detailed description of the surgeon’s dissecting work.

Pointing within the abdomen: Local deixis under restricted conditions

587

needing to point himself, because verbal acceptance or rejection will be interpretable in relation to the actual position of the dissecting forceps. AS1 not only rejects the possible target in line 04, but also offers a direction – nee dA DRUNter. (‘no there below’) – where the intended location can be found (line 05). Immediately after OS has verbally “claimed understanding” in line 06 he “exhibits understanding” (cf. Sacks 1992: II: 252 and Koschmann et al. 2010: 531), because without hesitating he starts to dissect and tears tissue apart by spreading the jaws of the dissecting forceps (cf. line 07). And this time OS is not corrected, but instead provided with technical dissecting instructions versUchen noch=en bißchen FREIzumachen (‘try to expose just a bit more’). Thus AS1 and OS collaboratively have made sure – in a recursively expandable correction sequence – that OS identifies step-by-step the intended target of AS1’s repeated local deictic da (‘there’), which changes its denotation with the positioning and repositioning of the dissection forceps. In extract (16) we have again encountered a division of labour between talk and gesture done by different actors, who together achieve gestural deixis. AS1 exploits the positioning and repositioning of the focal instrument, the dissecting forceps, for the precise localization of what is meant by the deictic da (‘there’). By treating the position of the dissecting forceps as a nonverbal first part of an adjacency pair, which asks for verbal acceptance or rejection, the participants make sure that if OS needs help, AS1 can successfully pilot him through unfamiliar anatomical territory, even though he is not able to perform a pointing gesture. Summing up, two radically different pointing practices are embodied by OS and AS1, depending on whether their instrument is free for enacting a pointing gesture or not. If their instrument is free for pointing as in extracts (13) and (14), it is used for pointing. The data suggest that in the optically mediated working space ‘abdomen’ the pointing gestures are carried out as “touching” gestures (Fillmore 1982) that ensure the desired amount of precision in the surgical field. The touching gestures themselves are subject to variation. We observe simple touching (extract 13) or tapping (extract 14), because they are locally produced for different recipients and for locating different referents or pointing targets. If AS1’s instrument (the grasping forceps) is not free to enact a pointing gesture, talk and gesture, i.e. the verbal and non-verbal tasks, have to be distributed. AS1 is in this case dependent on OS for enacting a kind of pointing gesture on his behalf. In the data discussed in this section, OS points for AS1 either by using the zooming function of the camera (extract 15) or by positioning (and if necessary re-positioning) the dissecting forceps (extract 16) at the presumed target of AS1’s

588

Susanne Uhmann

gestural deixis. In the case of zooming, the pointing act resembles what Fillmore has (1982) termed “Presenting”; it identifies only certain anatomical area or structure, but it lacks precision. If the focal instrument (the tapered dissecting forceps) is used, the pointing act can be repeated to ensure utmost precision. In this section three instances of gestural deixis with the local deictic hier (‘here’) and one instance of repeated use of da (‘there’) have been analysed. No attempt has been made so far to look for denotational differences between the two deictics. This is the question we finally have to turn to. 3.2.3 Where is hier and where is da? According to the literature on local deixis in German (cf. section 1), the positional deictic term hier (‘here’) denotes a highly variable space around the hicnunc-ego-origo of the speaker (= origo-inclusive). The analysis presented so far is in accordance with Koschmann et al. (2001: 519), who note with respect to engl. here in their analysis of laparoscopic surgery data that in this rather specific setting, its origo is a “virtual origo located in the shared media space” – that is inside the patient’s abdomen – and not the speaker’s physical location in the aseptic space (cf. section 2). Although this might seem to be a possible explanation, it first of all does not apply to all instances of hier in the German data – as we will see in the next extract – and secondly it does not account for the way this rather special denotation is locally, interactively and multimodally accomplished by the participants. Let us first address the problem of how the local, interactive and multimodal accomplishment of reference is achieved when hier (‘here’) does not denote a place inside the patient’s abdomen. We again join the team of Lap Chole I, where local deictics are rarely used, and this is the only instance of hier produced by AS1 (cf. section 3.1). (17) Lap Chole I (09:05) 01 AS1:

sowat schönes möcht=ich auch mal ma[chen dürfen.

something that nice I’d like to do sometime 02 OS:

[der chef möchte gerne

[the boss would like 03

übernEhmen weil das

04 AS1:

ich darf dat NIE. (-)

to take over because it is I’m never allowed to do this 05

(ich) muss immer HÜFten einschlagen (---)

I always have to knock in hips

Pointing within the abdomen: Local deixis under restricted conditions 06 OS:

wir kÖnn mal wEchseln (--)

07 AS1:

(ja) gerne

08 OS:

aber dann SAgense mir wieder=

589

we can swap (yes) fine but then you tell me again =ich hätte die troKAre anders einge(bracht) ((lacht))

I would have put in the trocars differently ((laughing)) -> 09 AS1:

ich dArf hier nicht mehr widerSPREchen. (-)

I am not allowed to contradict here anymore 10 OS:

dO:=och sIcher!

Oh, you are!

A localization of the place indicated by hier (‘here’) in line 09 inside the patient’s body lacks any kind of plausibility. This is of course due to the meaning of AS1’s turn. Unlike the instances of hier (‘here’) discussed in the previous section, AS1 is not talking about any entities that can be located in the working space “abdomen” (the suction cannula, the gallbladder), or giving instructions about surgical actions (dissecting). Instead, he is making a complaint; namely that he is not allowed to object. Objecting is an action that cannot be performed inside a patient’s body. So where is the place indicated by hier (‘here’)? Where is AS1 not allowed to object? Without going into a detailed analysis, which would be based upon laughter and lexical choice, I want to argue that AS1 formulates a mock complaint in line 09, which follows a list of mock complaints in lines 01, 04 and 05. Why does AS1 mock complaint during a surgery that has just started and that is obviously going well (cf. Line 01)? To answer this question we have to draw upon some ethnographic background knowledge. AS1 is the chief physician (Chefarzt) of the trauma surgery section, but he is substituting for the chief surgeon of the visceral surgery during the latter’s vacation. Trauma surgery and visceral surgery are mostly performed in different operating theatres. Thus AS1 is not in ‘his’ operating theatre nor is he performing surgery with one of his trauma surgery apprentices, but with a highly competent visceral surgeon. We can thus conclude that the local deictic hier (‘here’) denotes the operating theatre for visceral surgery, where Lap Chole I is conducted. But due to his status, AS1 is of course entitled to object, as both OS and AS1 know, nevertheless OS explicitly rejects this complaint in line (10).24 Extract (17) thus provides an instance of sym24

Line (10) does not close the mock complaint sequence. There is evidence in the data that OS is not the only recipient addressed but that the recipients included me. As we had not been introduced before the operation started, AS1 could not be

590

Susanne Uhmann

bolic deixis (cf. section 1) and the only instance of hier (‘here’) in the data denoting some space that cannot be located in the patient’s abdomen. In all the other instances of gestural deixis and in all the extracts discussed in the previous section, we encountered a division of labour between talk (local deixis) and gesture (pointing). But we have not yet answered the question, why OS and AS1 use the proximal hier (‘here’) in extracts like (13), (14) and (15), and not dort (‘there’) or da (‘there’), for denoting some referent or for identifying some target in the visually mediated working space ‘abdomen’. In answering this question, we can build upon the assumption that the surgical instruments substitute for the hands of the surgical team members. As the hands of OS, AS1 and AS2 are visually co-present in the working space “abdomen”, their hier (‘here’) can denote a proximal space around their particular instrument. This analysis is convincing for extracts (13) and (14), where OS and AS1 use their instruments, the dissection forceps and the grasping forceps, to perform pointing gestures which even touch the target (cf. also Klein 1978: 26). For extract (15) the argument is not that straightforward, but one might claim that the act of zooming is comparable to the pointing-by-touching gestures and thus explain the use of the proximal deictic hier (‘here’) in extract (15). The quantitative analysis presented in section 3.1 has revealed that da (‘there’) is more than twice as frequent as hier (‘here’) (55:25). Especially in Lap Chole II, when the main addressee is OS, da (‘there’) is overwhelmingly more frequent than hier (‘here’) (28:4).25 So the next two questions to be addressed are where is da and why is da more frequently used than hier? Evidence leading to an answer to the first question can be found in extracts like (18): (18) Lap Chole II (14:13) 01 AS1:

sie [müssen] jetzt lAngsam sich konzenTRIEren.

you have to start to focus now 02 OS: 03 AS1:

[