Syntactic Variation and Genre 9783110226485, 9783110226478

Citation preview

Syntactic Variation and Genre

Topics in English Linguistics 70

Editors

Bernd Kortmann Elizabeth Closs Traugott

De Gruyter Mouton

Syntactic Variation and Genre Edited by

Heidrun Dorgeloh Anja Wanner

De Gruyter Mouton

ISBN 978-3-11-022647-8 e-ISBN 978-3-11-022648-5 ISSN 1434-3452 Library of Congress Cataloging-in-Publication Data Syntactic variation and genre / edited by Heidrun Dorgeloh, Anja Wanner. p. cm. ⫺ (Topics in english linguistics; 70) Includes bibliographical references and index. ISBN 978-3-11-022647-8 (alk. paper) 1. Grammar, Comparative and general ⫺ Syntax. 2. Language and languages ⫺ Variation. I. Dorgeloh, Heidrun. II. Wanner, Anja. P291.S9545 2010 415⫺dc22 2010039817

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.d-nb.de. ” 2010 Walter de Gruyter GmbH & Co. KG, 10785 Berlin/New York Cover image: Brian Stablyk/Photographer’s Choice RF/Getty Images Typesetting: PTP-Berlin Protago TEX-Production GmbH, Berlin Printing: Hubert & Co. GmbH & Co. KG, Göttingen ⬁ Printed on acid-free paper Printed in Germany www.degruyter.com

Preface This volume is based on a workshop we organized as part of the Annual Meeting of the German Linguistic Society (DGfS) in 2007. The workshop focused on the role of syntax in the emergence of new genres and brought together insights from research that uses the concept of genre as a reference-point for the description (and, possibly, explanation) of patterns of morphosyntactic variation. Our thanks go to everyone who made the workshop a success: the selection committee of the DGfS for the opportunity to offer the workshop, the Deutsche Forschungsgemeinschaft for providing travel support for invited speakers, and, most importantly, the workshop participants, who were willing to listen to and learn about approaches to the intersection of grammar and genre that were different from their own. Organizing a workshop is one thing, editing a book another. While this collection grew out of the DGfS workshop, it is not a proceedings volume. Most of the papers selected from the workshop were rewritten, revised, or newly contextualized. In addition to the articles based on the workshop, this volume also includes papers by scholars whom we specifically invited to contribute. We thank all authors for their commitment to the consideration of the term “genre” as a multi-faceted concept and to facilitating communication among different approaches. We also warmly thank the editors of the series Topics in English Linguistics, Elizabeth Closs Traugott and Bernd Kortmann, for their enthusiasm for this project, and our editor at Mouton, Birgit Sievert, for her patience and advice, and for making book-editing look effortless. We are grateful to the Graduate School at the University of WisconsinMadison for providing financial support through travel and research grants. Special thanks go to Katie Gilbert and Ania Hutnik for their scrupulous copyediting at various stages of the manuscript, and to Katie Gilbert also for applying her expertise at editing manuscripts written by non-native speakers of English. Additionally, we are very much indebted to Gero Brümmer for his assistance with compiling the index. This book is also the result of a long-standing transatlantic collaboration and friendship. We thank Dieter Stein for introducing us to each other at a DGfS summer school many years ago and for supporting our work ever since. Alas, nothing would have been accomplished without the unequivocal support of our families, who put up with many inconveniences arising from lengthy phone calls at odd hours or work meetings that were squeezed into precious vacation time. You are the best. Heidrun Dorgeloh and Anja Wanner

Contents

Preface

v

Introduction Heidrun Dorgeloh and Anja Wanner

1

Part 1: Genre: Form, constitution, and change Genre as difference: The sociality of linguistic variation Janet Giltrow Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre Tuija Virtanen A new genre for a specialized community Maurizio Gotti Do some genres or text types become more complex than others? Javier P´erez-Guerra and Ana E. Mart´ınez-Insua Mein Problem ist/mein Thema ist (‘My problem is/my topic is’): How syntactic patterns and genres interact Wolfgang Imo Thank you for thinking we could: Use and function of interpersonal pronouns in corporate web logs Cornelius Puschmann

29

53 85 111

141

167

viii

Contents

Part 2: Syntactic variation based on genre Grammatical constructions and communicative genres Susanne G¨unthner

195

Genre effects in the replacement of reflexives by particles Britta Mondorf

219

Clause linkage techniques as a symptom of discourse traditions: Methodological issues and evidence from Romance languages Johannes Kabatek, Philipp Obrist, and Valentina Vincis

247

Syntactic constructions as a means of spatial representation in fictional prose Rolf Kreyer

277

Agreement in educated Jamaican English: A corpus-based study of spoken usage in ICE-Jamaica Susanne Jantos

305

I know you guys hate forwards: Address pronouns in digital folklore Theresa Heyd

333

List of contributors

359

Index

361

Introduction Heidrun Dorgeloh and Anja Wanner

1. Goals and motivations This collection of studies on syntactic variation and genre originated in a workshop held at the annual meeting of the German Linguistic Society in 2007 at the University of Siegen. The workshop brought together linguists with very different specializations, united for that occasion by a willingness to integrate the concept of genre – present mostly in theory, but much less systematically so in empirical linguistics – into their own work. The idea was to see how far theories and applications of genre had been and could be operationalized for the study of grammatical variation, i.e. to take a former discourse category beyond discourse and literary studies, and thereby to restitute the relevance of its form. While the workshop focused on the emergence of genres, it is the aim of this book to explore the question of how syntactic variation is motivated by the context of genre in general and to integrate insights from attempts for text classification and rhetorical views on genre, as well as aspects of specialized discourse and the interactional dimension of spoken discourse. The study of morphosyntactic variation (and change) has become a substantial field of study in linguistics, including the study of English in its history and varieties (e.g., Nevalainen et al. 2008). The focus of these studies has traditionally been more on the classical sociolinguistic variables, while more recent work highlights the contextual, functional, and processual parameters that determine grammatical variation (e.g., Rohdenburg and Mondorf 2003). By contrast, work that takes into account variables explicitly referring to text and discourse is usually confronted with the problem that there is a lack of accepted criteria to distinguish between categories such as register, text or discourse type, and style. The situation usually becomes worse when one adds the term genre to that list: Usage of this term ranges from working definitions such as “categories distinguished by mature speakers of a language” (Biber 1989: 5) to complex rhetorical theories of “genre as social action” (Miller 1984: 151). That means that approaches to genre either take the concept for granted, or they tend to focus on its socio-pragmatic context; in both cases, the relevance of (morphosyntactic) form is reduced or remains

2

Heidrun Dorgeloh and Anja Wanner

undefined. As will be seen throughout this volume, there is no singular or even just dominant tradition for approaching the study of syntactic variation and genre. In this volume, therefore, we aim at bringing together theoretical work on text types and genres with studies of morphosyntactic form and change. What all contributions share is that the choice of the syntactic phenomena under investigation is motivated by an interest in genres and their linguistic reflections. That means that the focus of interest lies more on textual varieties, rather than on language variation according to use in general, i.e. styles and registers. Nevertheless, depending on the research traditions that the authors affiliate with, varieties as well as variables differ substantially, as do the methods that are employed for studying them. Doing justice to the concept of genre in a study of grammatical variation can never be more than an approximation. This introduction is structured as follows: Since there is more than one understanding of what exactly falls under syntactic variation, section 2 deals with the various traditions within this field and with the range of syntactic phenomena studied under this heading. Section 3 situates the concept of genre within the study of language use: Section 3.1 is dedicated to the relationship between register and genre and to the question of what is gained by postulating genre as a category. Section 3.2 brings back into focus the aspect of form-function relationships within genre, thus relating it to the more form-based varieties style and text type. Along the lines of Biber and Conrad (2009), who use all three concepts – register, genre, and style – “to refer to three different perspectives on text varieties” (Biber and Contrad 2009: 2, our emphasis), we argue in favor of a co-existence of the concept of genre with its competitors and thus for the justification of a variety of approaches and methods, such as are present in this volume. Finally, in section 4 we deal with the major traditions in the study of syntactic variation and its interrelationship with genre: We start out by discussing three major paradigms of linguistic inquiry in section 4.1. Since the studies in our volume are all concerned with patterns of variation based on data from language use, it will become clear that they are all associated with the descriptive-interpretive paradigm (Sankoff 1988). In section 4.2, we conclude our introduction with an overview of the papers contained in this volume. Articles in the first part are mainly concerned with the concept of genre, either from a theoretical point of view or in the context of the emergence of a genre, while articles in the second part of the book take a specific phenomenon of syntactic variation as their starting point and examine how its patterns of occurrence can be explained in relation to genre.

Introduction

3

2. On syntactic variation The study of syntactic variation has a long-standing tradition, but the understanding of what exactly is meant by the term differs. For some, the term clearly refers to the language system and to the structural availability of variants, “formal alternatives which can be considered optional variants, in the sense that they are nearly equivalent in meaning” (Biber et al. 1999: 14). For example, Standard English as a system provides three options, or variants, for introducing a relative clause: with a wh-phrase (the man who I met. . . ), with that (the man that I met), and with no relativizer at all (the man I met), and one might look at how these three variants can be related to each other. For others, including the contributors in this volume, the term variation is applied in reference to actual language use, and syntactic variation is analyzed as the existence of identifiable patterns of usage in specific communicative settings, with an emphasis on genre-specific form-function correlations. As noted recently by Croft (2010) about a set of data that all verbalize the same kind of content (the famous Pear Stories [Chafe 1980]), morphosyntactic variation, i.e. “variation in verbalization is not occasional; variation is ubiquitous” (Croft 2010: 10, emphasis in the original). Therefore the principled question is how similar two forms have to be to be regarded as variants. Do they have to share a certain number of lexical items, a syntactic core, or a syntactic domain (noun phrase, verb phrase, or clause)? Where does one draw the line? Should sentences like John killed the mouse and My brother’s friend caused the obnoxious rodent to die be considered variants of each other? While it can be argued that these sentences are roughly equivalent in meaning, they use different verbs, nouns, and syntactic structures to express that content. One possibility for limitation is to consider only candidates (to borrow a term of Optimality Theory) of the same structural kind and to regard them as competing constructions rather than variants, or even derivations, of each other. For example, in this volume, Kreyer focuses on six syntactic constructions involving fronted locative phrases and discusses how each type is used to create the effect of immediate observation of a spatial configuration. These would be competing constructions, but would not count as variants of each other. Adding a diachronic perspective to the description of constructional variants often means that variation is framed as competition: One variant may get more popular over time, picking up new functions, while another decreases in use. For example, as discussed by Mondorf (this volume), for selected verb classes the reflexive construction (brace oneself ) and the particle construction (brace up) seem to be on complementary trajectories, with the latter increasing in popularity.

4

Heidrun Dorgeloh and Anja Wanner

Still, a definition of variants like the one given above opens up the question of what exactly it means that two syntactic patterns are “nearly equivalent in meaning.” As pointed out by Rosenbach (2002: 25), “while phonological variants are by definition synonymous, it is not at all clear how ‘sameness’ in syntax should be defined.” How is one to limit the syntactic options a speaker can choose from? In a context-free model of syntax, such as the generative model, the focus is on propositional meaning and, possibly, there is a shared underlying structure among forms that are recognized as variants. For example, active and passive sentences, by and large, have the same truth conditions and express the same propositional meaning and can thus be considered variants at the clause level. By contrast, in a context-bound model of syntax, discourse-dependent variables are taken into account, and passives and actives clearly differ with regard to the distribution of given and new information, among other things. The term variation can also be applied at the level of individual speakers: Situated in a specific context, they need to choose from a menu of options available to them. Rosenbach (2002: 77) prefers the term “choice” for this kind of intraspeaker variation and classifies functional and cognitive approaches to variation as choice-based (vs. variation-based approaches, exemplified by generative grammar and corpus linguistics). Which approach is taken is determined by the questions that are asked rather than by the syntactic phenomenon itself. For example, in her discussion of the plural vocative you guys, Heyd (this volume) could focus on the role of the expression as a marked variant of the simple pronoun you or as an example of a general trend towards more colloquial expressions in written English (Mair and Leech 2006). Instead of following either of these options, however, she is interested in analyzing its popularity in a specific genre of computer-mediated communication as resulting from the conscious choices that speakers make to express plurality explicitly in a specific discourse situation. By contrast, Jantos’ investigation of subject-verb agreement in Jamaican English (also this volume), with its emphasis on the role of formality and register, may be termed a classic variationist study. Overall, it may be easier to approach the question of the limits of syntactic variation bottom-up, through illustrating the patterns that are treated as syntactic variants in different linguistic frameworks. In a language with rather strict word order, such as English, patterns of postverbal constituent order variation, such as the placement of verbal particles or the order of indirect and direct object, are classic cases of syntactic variation (Deh´e 2002; Gries 2002; Wasow and Arnold 2003; Mukherjee 2005). Another type of syntactic variation might be called constructional variation, as exemplified by the co-existence of the of -genitive and the s-genitive (Rosenbach 2002; Stefanowitsch 2003), or the analytic and the synthetic comparative (Mondorf 2003; Mondorf 2009), or the

Introduction

5

option to either move prepositional phrases to the clause-initial position (pied piping) or to move just the noun phrase and leave the preposition behind (preposition stranding) in wh-questions and relative clauses (Hornstein and Weinberg 1981; Hoffmann 2005). Here, the variants fulfill a requirement of structural comparability, or sameness, in that they belong to the same syntactic domain: noun phrases in the case of the genitive, adjective phrases in the case of the comparative, and prepositional phrases in the case of pied piping and preposition stranding. However, oftentimes, there are subtle, and sometimes not-so-subtle, meaning differences between the two (or more) existing variants of a wordorder pattern or a construction. For example, Goldberg (1995) discusses that the two patterns of complement sequencing after ditransitive verbs (bake Harry a cake/bake a cake for Harry) do not have the same range of meanings and should therefore not be considered variants of a common construction at all (only the V-NP-PP sequence can mean that someone baked a cake so that Harry would not have to do it). In addition to these semantic differences, the overlap between the constructional variants, i.e. contexts in which both variants are equally acceptable, may be rather small. A case in point is the English comparative: Clearly, variation (between the analytic and the synthetic form) exists at the level of the system, but not necessarily at the level of each individual adjective. Assuming that there is at least some sort of overlap between two variants, and therefore true choice, which of the co-existing, or competing, options does a speaker employ, and for which reason? In this volume, the answers to these questions generally go beyond the structure of the sentence and relate to the level of the text. The factors that contribute to the choices between variants – some might prefer the bolder verb “determine” here (compare Rohdenburg and Mondorf 2003) – include structural and situational features, among them the following: – lexical factors, such as the distinction between animate and non-animate nouns (relevant, for example, in the case of the choice between the two genitive variants, see, e.g., Rosenbach 2008), or in the form of genre-specific formulae (as discussed in Imo, this volume), – structural factors, such as the status of a prepositional phrase (PP) as a verbal complement or adjunct (relevant, for example, in the analysis of preposition stranding, see, e.g., Hornstein and Weinberg 1981), or the “heaviness” of a phrase (relevant, for example, in the discussion of postverbal PP sequences, see Wasow and Arnold 2003), – cognitive factors, such as the function of language to recreate the visual experience of objects and space (Langacker 1991; Radden and Dirven 2007), relevant, for example, in the analysis of English inversion as a “Ground-

6

–

– –

–

Heidrun Dorgeloh and Anja Wanner

before-Figure” construction (Chen 2003) or as mirroring natural perception (Kreyer 2006, also this volume), psycholinguistic factors, such as the role of accessibility in sentence production (Ferreira and Engelhardt 2006), or the need to maximize efficiency (Hawkins 2004) by minimizing the processing load of a phrase, which, as argued by Mondorf (this volume), may be most pronounced in informal genres, discourse-related factors, such as information structure (Ward and Birner 1998) and relevance (Carston 1996), other pragmatic factors, such as the need to negotiate the face-threatening potential of an utterance (Brown and Levinson 1978), as discussed in this volume by G¨unthner, social factors that relate to the formality of the text (see, for example, Jantos, this volume) or to the social background (age, gender, location) of the speaker (Mondorf 2004).

Suggested criteria ( top-down)

Examples (bottom-up)

Suggested determinants

– Semantic: equivalence in meaning – Structural: shared underlying structure – Functional: choice among competing constructions – Contextual: variation reflecting situation

– order of sentence constituents (subject/ verb, direct/ indirect object, verbal particles) – pied piping/ preposition stranding – of -genitive vs. sgenitive – analytic vs. synthetic comparative

– lexical (e.g. animacy) – structural (e.g. heaviness) – cognitive (e.g. visualisation) – psycholinguistic (e.g. processing efficiency) – textual (e.g. information structure) – pragmatic (e.g. politeness) – social (e.g. formality)

Figure 1. Dimensions of syntactic variation

Most of the factors listed in Figure 1 are touched on in this volume, with the common thread lying in the interest in genre-specific form-function correlations. However, it is not our goal to suggest that the variation phenomena discussed in this volume are determined by genre only, nor do we want to pretend that the various approaches to syntactic variation and genre represented here easily complement each other, which we do not consider a conceptual weakness, but rather a reflection of the many facets of the topic. As Gries (2003: 172) forcefully argues in his study on particle placement variation, there are usually more than

Introduction

7

a handful of variables at work and “nothing is gained by simplifying the picture too much.”

3. On genre and related concepts 3.1. Intraspeaker variation: Genre and register The concept of genre is best positioned somewhere in the intersection of sociolinguistics, discourse studies, and literary or rhetorical theory. For the sociolinguist, genre constitutes one type of “intra-speaker variation:” Just as speakers often switch into specific dialects and registers depending on the necessities of a situation, they may also switch “into and out of genres” (Schilling-Estes 2002: 375). For example, a researcher may change into the scientific register when receiving a phone call from a colleague at home, but s/he will switch into academic genres, such as a research report, an abstract preceding a research article, or a lecture, only under more specific conditions, defining the message type (Ferguson 1994: 20) rather than the context. Where registers involve “shifts in usage levels” for features associated with situations in general (Schilling-Estes 2002: 375), we use genres “to get things done” (Martin and Rose 2003: 7). As will be pointed out later in this volume, this level of verbal action implies that genres, in contrast to registers, exist at a rather “low level of generality” (Giltrow, this volume: 30), the difference between the two concepts therefore is gradient rather than absolute. Genre and register thus share the departure from functional rather than formal differences: Which of them is used “indexes properties of your present situation and social activity” (Irvine 2001: 26). The formal choices we make in relation to this social and situational context in principle involve any area of language organization, from phonology and lexicon, to morphosyntax and discourse-pragmatics (e.g. Schilling-Estes 2002: 376). But it is not always clear which kind of variation takes place under which circumstances: While in the sociolinguistic tradition, variation refers to “alternative ways of saying ‘the same’ thing” (Labov 1972: 188), “sameness” becomes “challenging” if extended beyond the phonological level (Schiffrin 1994: 288). This has led Halliday, for example, to define functional variation as showing itself primarily in differences in semantics, and only in consequence of the different things to be expressed in functional varieties to variation in lexicogrammar (Halliday 1978: 35). And although it has now certainly been acknowledged that there are also many nonsemantic differences among registers (take the use of non-standard agreement in educated Jamaican English across registers described by Jantos, this volume),

8

Heidrun Dorgeloh and Anja Wanner

intraspeaker variation still substantially covers “different ways of saying different things” (cp. Halliday 1978: 35), “most particularly in types of written texts, insofar as such types are conceived as registers (rather than, say, as genres)” (Irvine 2001: 27, our emphasis). Genre variation, therefore, is language variation beyond the limits of semantic equivalence, which is why syntax – even more than phonology or the lexicon – provides a promising area of study. Absence of the semantic equivalence criterion also means we are dealing with differences here that are “not categorical (such that one variety has a certain grammatical element or syntactic construction which another has not)” (Kortmann 2006: 603). In the first place, such non-categorical variation has been observed for certain non-standard varieties of English, i.e. for variation by user, but it also applies to variation by use. Instead of being categorical, choices motivated by and reflecting genre are “meaningful choices” in the sense of serving “the various needs of the language user” (Schulze 1998: 7). As such they are conditioned by the surrounding discourse rather than within the sentence. While below the sentence-level constructional relations hold, it is above the sentencelevel, i.e. when sentences become discourse, that “non-constructional forms of organisation take over [. . . ]. Changing the order of sentences in a text is about as meaningless as putting the end before the beginning” (Halliday 1985: xxi). The study of variation according to genre therefore has an inherent “interest [. . . ] for grammatical structure in discourse” and, with that, a natural “preoccupation with [. . . ] linguistic form-function relationships” (Sankoff 1988: 141, emphasis in the original). These, however, tend to be polyvalent and unstable (Sankoff 1988: 141; Newmeyer 2010). For example, some of the syntactic features symptomatic of the genre of personal ads in newspapers, such as ellipsis or lack of subordination, are identical with those prevalent in conversational, highly interactive speech; however, the functional explanations for their occurrence are rather different. The observed lack of syntactic elaboration can be due to a lack of planning ahead or, alternatively, to a particularly high degree of planning, in line perhaps with an intention to minimize both space and costs. In the same vein, most of the syntactic parameters treated in this volume cannot be considered “markers” of genres (nor of registers) in the sense of being unique to one genre or a set of them. For example, syntactic features that typically occur in academic texts, such as hedging by modals or nominalization (see Giltrow or Gotti, this volume), also occur – for other or the same reasons – in other kinds of writing. Similarly, heavy, i.e. long and complex, noun phrases, in the scientific register as a consequence of frequent nominalization, are overall more frequent in other kinds of formal writing (Pérez-Guerra and Martínez-Insua, this volume). In principle, therefore, (syntactic) form underdetermines genre (Stein and Giltrow 2009: 3), but that does not mean that form is not an important aspect

Introduction

9

of genre. While in rhetorical theory it has been “the disavowing of form’s priority that has accelerated genre’s conceptual revival”, for most linguists, among them the contributors to this volume, it is form, and here morphosyntactic form in particular, that constitutes “a prior condition for reasoning about genre” (Stein and Giltrow 2009: 14). 3.2. Variation in form: Genre, style, and text type The discussion of genre as a form of intraspeaker variation has shown that genre in linguistics is a conglomerate concept, uniting aspects of context, discourse function, as well as internal structure and thus showing substantial overlap with a range of related concepts (see Figure 2 below). If the focus is laid on function, studies question how a certain meaning, or social action, is encoded in language, while if the focus is laid on variation in form, we ask “how speakers mean by selecting particular forms among other possible variants in any given context” (Taavitsainen and Fitzmaurice 2007: 25). Variation in the linguistic form of texts is also studied under the heading of style (Jucker 1992; Eckert and Rickford 2001), or speech styles, although both terms are found to apply to all intraspeaker variation as well (Jucker 1992: 24–25). Style is therefore another “holistic and multilevel phenomenon” (Auer 2007: 13), but one that focuses more on the formal aspects reflecting social identities and social practice (Eckert 2000). Text types, by contrast, are “groupings of texts that are similar in their linguistic form” (Biber 1988: 170, emphasis ours); in Biber’s approach, this is entirely “irrespective of genre” (Biber 1988: 170), though others acknowledge that the “sharp distinction between an internal and an external text dimension is [. . . ] often difficult to uphold” (Pilegaard and Frandsen 1996: 3). Methodologically, the study of style and text type is associated with sets of “co-occurring variables” (Eckert and Rickford 2001: 5), and both are usually studied on the basis of large corpora. With regard to genres, it is one of the major outcomes of Biber’s work on register variation and the resulting text types (of English) that genre distinctions never prove to “adequately represent the underlying text types” (Biber 1989: 6), meaning that they do not follow the co-occurrence patterns determined by linguistic form. One explanation has been that genres are text categories primarily determined by the speakers of a language (Biber 1989: 5), and, with that, by their needs for categorization. However, the categories used in the analysis of large corpora are often not to be considered genres at all (see Giltrow, this volume), in the sense of referring to a type of communicative event (Swales 1990; Swales 2004), a kind of social action (Miller 1984; also Giltrow, this volume), or to a shared communicative purpose (Askehave and Swales 2001: 200). By its very nature, genre

10

Heidrun Dorgeloh and Anja Wanner

is therefore defined by function much more than by form. As a consequence, as some articles in this volume show, one and the same genre, e.g. the encyclopedic article, may contain a lot of formal variation, i.e. differences in text type (Virtanen, this volume), while research genres, for example, normally known for their objectivity, contain interpersonal features, such as deontic modality, which they share with otherwise less “neutral” and more overtly argumentative genres (Giltrow, this volume).

– – –

Text type linguistic form text structure varying levels of generality

– – –

Register situation function high level of generality

– – –

Genre social action patterned practice low level of generality

– – –

Style linguistic form social practice low level of generality

Figure 2. Overview of categories for modeling intraspeaker variation

There is other work that links given, external parameters to variables of linguistic form. Such work usually takes the form of linguistic profiles of what are limited areas of discourse: Examples are differences between spoken and written genres (Sanchez-Maccaro and Carter 1998), profiles of specialized genres from the academic world (Swales 2004; Flottum, Dahl, and Kinn 2006), of communicative genres in everyday interaction (for references see G¨unthner, this volume), or – highly popular in recent years – of the language and genres in computer-mediated communication (Herring 1996; Siever, Schlobinski and Runkehl 2005; Beißwenger 2007; Stein and Giltrow 2009). There is also the dimension of language variation, including variation by genre, inducing language change (Diller and G¨orlach 2001; Milroy 1992; Nevalainen and Kahlas-Tarkka 1997; Mair and Lindquist 2004), which is why a range of corpus studies relevant to the field come from the area of historical pragmatics (Jucker 1995; Fitzmaurice and Taavitsainen 2007). Some of these approaches have highlighted further problems one encounters when patterned usage is investigated via frequencies of occurrence: First, while corpora supply us “with ample sources of form; [. . . ]

Introduction

11

[they] subsume but rarely make manifest [. . . ] the context in which texts and their forms occur” (Fitzmaurice and Taavitsainen 2007: 25). And, more importantly, it is neglected that linguistic phenomena, which one might be right to interpret as conditioned by situational and functional factors, are themselves constitutive of these external conditions. It is this interrelationship of form and function that studies with a focus on syntactic variation and genre need to take into account. Despite an inconsistent usage of variationist terminology across disciplines in linguistics, we therefore think that there is room, and in fact need, for both the concepts of text types and genres when studying syntactic variation in use, and that one cuts across the other (Virtanen, this volume: 55). While research focusing on genres generally departs from external criteria, “such as intended audience, purpose, and activity” (Biber 1988: 170; Lee 2001: 38), linguistic form is equally relevant, but this is “form under pressure from function” (Giltrow, this volume: 30). By contrast, focusing on internal, i.e. structural, criteria distinguishes text types as aggregates of form, open to direct and indirect uses, and which can therefore serve entirely different functions (Virtanen, this volume). Since phenomena of syntactic variation are usually functionally motivated (Rohdenburg and Mondorf 2003; also Mondorf, this volume), both text types as “aggregates” of form (Virtanen, this volume: 55) and genres as patterned communicative practices (G¨unthner, this volume) are relevant concepts for the study of the nature of form-function relationships in grammar.

4. Studying syntactic variation and genre 4.1. Paradigms of exploration Sankoff (1988) presents three paradigms of linguistic inquiry into syntactic variation: the introspective-generative paradigm, the experimental-evaluative, and the descriptive-interpretive paradigm. In each of these, syntactic variation is studied, but with different goals, methods, and research questions. In the introspective-generative paradigm, syntactic variation is studied with regard to the options that are offered within the language system, often at the typological level. In the Principles & Parameters approach to syntactic theory, variation arises from different settings of parameters (Chomsky and Lasnik 1993) that “should govern clusters of empirical phenomena” (Biberauer 2008: 16). Many parameters, although not all, were conceived of as binary, i.e. as “Yes/No choice point(s)” (Biberauer 2008: 19). For example, with regard to the pro drop parameter, English is specified as “minus” and Italian as “plus” (Burzio 1986). What

12

Heidrun Dorgeloh and Anja Wanner

that means is that in English, pronominal subjects in finite clauses have to be expressed overtly (*Walked away), while in Italian, they need not be realized overtly. Whether or not a language is a pro drop language has consequences for language acquisition. If child grammar has an initial setting, it must be [- pro drop], because then they can reset the parameter based on positive evidence, i.e. based on naturally occurring grammatical sentences. While some linguists assume that parameters are set in an all-or-nothing fashion (Radford 1990), others point out that, if language acquisition is parameter-setting at all, it still begins locally and based on items that occur frequently in the child’s linguistic input (Tomasello 2003). If a child assumes that English is a [+ pro drop] language, he or she would need negative evidence (i.e. information that sentences without overt subjects are ungrammatical) to reset the parameter correctly, and it has been shown that negative evidence is not a relevant factor in first language acquisition (Hyams 1986). Critics of the generative approach to language acquisition argue that one has to take into consideration the input that children actually have. According to Tomasello (2003), of the 5000–7000 utterances that children hear every day, only 15% have the canonical subject-verb-object (SVO) form; most utterances are fragments. The setting of parameters can also be studied along a historical scale and variation can be framed as competition between two different grammars. For example, in Modern English the verb generally precedes its object (SVO word order), but in Old English, the SOV pattern was also available. “In the course of any syntactic change, which typically follows an S-curve, there is a transitional state in which the older and newer patterns are attested side by side at shifting frequencies” (Fischer et al. 2000: 150). When it comes to the structure of Old English, the generative method of inquiry – constructing sentences that test a hypothesis and having their grammaticality evaluated by native speakers – is not an option, and one has to turn to textual material instead. However, data from Old English may be limited with regard to the genre they represent. For example, Fischer et al. (2000) make the point that an analysis of Old English syntax that relies on data from Beowulf only (see Pintzuk and Kroch 1989) does not easily carry over to prose text material. Normally, however, the concept of genre is not brought up in the study of variation in the generative paradigm. The focus of inquiry is on “competence”, the native speaker’s tacit knowledge of the rules of language that enable him or her to produce grammatical sentences. Whether or not these sentences actually occur is irrelevant. When Chomsky talks about the native speaker, the person he has in mind is “an ideal speaker-listener, in a completely homogeneous speech-communication” (Chomsky 1965: 3). However, this does not mean that generativists do not work with data from language use. Since linguistic competence cannot be examined directly, one has to turn to

Introduction

13

data from language performance. Syntactic variation itself is approached with an eye to the question of how two clauses with the same meaning but different surface form can be derived from each other or from a common underlying source, in line with the principles of Universal Grammar. A common mechanism to relate two different surface forms to each other is move alpha – movement of a syntactic head or phrase. Analyses of patterns of syntactic variation often make use of this device (e.g., Pintzuk and Kroch 1989; Chomsky and Lasnik 1993). Increasingly, it is recognized in generative grammar that there is no such thing as free variation between two syntactic constructions and that discoursebased factors, such as the distinction between given and new information, need to be included in the discussion of syntactic variation. For example, many particle verbs allow both the V-Prt-NP and the V-NP-Prt word order (look up the information/look the information up), but this is not sufficient reason to assume that the choice between the two options is either purely stylistic or is fully determined by lexical and syntactic factors only, such as the choice of the verb and the weight or complexity of the noun phrase (NP). Deh´e (2002) argues that the choice between the two patterns is also driven by the information structure of the construction. For example, the answer to the question What happened to the cat? would more likely be Somebody locked the cat in (rather than Somebody locked in the cat), i.e. the NP previously established in the discourse would be positioned between the verb and the particle. Thus, it does not quite seem to be the case that the two surface forms are truly variants of each other. However, while generative theory is reaching into the domain of the text in such kinds of studies, the concept of genre is still largely left unexplored. In the experimental-evaluative paradigm, variation is studied with regard to cognitive processes, such as the acquisition and processing of language. Data come from a variety of sources, including transcripts of spontaneous speech (Bowerman 1974; MacWhinney 2000) and psycholinguistic experiments, such as elicitation tasks (e.g. Pinker, Lebeaux, and Frost 1987), or responding to items presented on a computer screen (e.g., Mauner and Koenig 1999), as well as from neurological imaging (e.g., Grodzinsky 2003). Unlike research that is entirely based on naturally occurring speech or writing, it is the norm in this paradigm that data are created in a controlled environment to test a hypothesis. For example, as discussed above, the active and the passive may be considered syntactic variants of the same propositional content. Since not many passives occur in the natural speech of children, it is not immediately clear whether or not the grammar of young children includes a rule of passivization. If not, it cannot be argued that they choose to use the active in situations in which the passive would also be appropriate. Therefore, linguists use a number of elicita-

14

Heidrun Dorgeloh and Anja Wanner

tion and comprehension experiments to test if the grammar of children includes a productive rule of passivization (Maratsos et al. 1985; Pinker, Lebeaux, and Frost 1987). Increasingly, insights (and methods) from psycholinguistic research are being incorporated by linguists who take a functional approach to language. A case in point is the principle of efficiency, which, following Hawkins (2004), is assumed rather than measured in many usage-based accounts of syntactic variation (see below). The impact of the need to be efficient may be even greater in spoken discourse. In an overview of studies carried out on the basis of transcribed conversations (Quaglio and Biber 2006), the authors emphasize that conversations are not just instances of language use in a different context; they also occur under greater pressure on the production side of language. “Given the pressure of real-time production, speakers tend to opt for these easier and more efficient forms” (Quaglio and Biber 2006: 702). One example would be the use of non-clausal units and “utterance launchers” (Quaglio and Biber 2006: 704) that function as attention getters. Both of these trends are found in the data from spoken German analyzed by G¨unthner (who discusses clause-like structures without finite verbs) and Imo (who discusses types of phrases that function as “attention getters” and signal a change from one genre to another) in this volume. Another concept based on language processing is that of entrenchment. The more frequently an item occurs in a speaker’s linguistic input, the more deeply it will be entrenched in the speaker’s mental grammar (Bybee and Hopper 2001). Strongly entrenched forms are accessed faster in language production and comprehension and resist change. Psycholinguistic experiments confirm this tendency at the level of the individual speaker: The more often a speaker encounters a verb in a specific construction, the less likely he or she is to use the word in a new pattern. Tomasello (2003: 179) reports a study which showed that children are less inclined to use high-frequency verbs like come in novel syntactic contexts (as in The car came him). The children showed more flexibility with low-frequency verbs, such as vanish. What this means for the questions at hand is that frequency (of type and token) is an important factor in predicting where syntactic variation and change are to be expected to occur. However, as Mondorf points out (this volume), efficiency may be at odds with the conventions of a genre. She found that a general trend towards the use of the particle instead of the reflexive in order to express telicity was not reflected in scientific and religious texts and concludes that “these highly conventionalized genres seem to be a stronghold of the reflexive” (Mondorf, this volume: 239). It is this tension between variation and conventionalization that provides the common ground for the discussion in this volume.

Introduction

15

Finally, the descriptive-interpretive approach is concerned with identifying, analyzing, and predicting patterns of syntactic variation based on data from language use. Data come from a variety of sources, but one source that clearly stands out is the use of representative linguistic corpora, often subdivided by time periods, register, and/or text type. For example, the British National Corpus is a synchronic corpus divided into various text categories according to medium (book, periodicals) and domain (imaginative, informative), which allows for the comparison of linguistic features across these categories. The goal of such corpus-based methodology is to interpret quantitatively established patterns in a functional way (Biber, Conrad, and Reppen 1998). In the Longman Grammar of Spoken and Written English, Biber et al. (1999: 41) highlight three approaches to interpreting quantitative findings qualitatively and with an emphasis on the function of linguistic expressions for the overall discourse: One can look at “(1) the work that a feature performs in discourse; (2) the processing constraints that it reflects; and (3) the situational or social distinctions that it conventionally indexes” (Biber et al. 1999: 41). All three paths are taken in the contributions to this volume, but the analyses are not always based on data from large corpora, due to their inherent limitations. Although, in principle, they all fall under the descriptive-interpretive paradigm and thus share a functional orientation, they also show the diversity that has been noted oftentimes about functional approaches (e.g., Penke and Rosenbach 2004). This concerns not only the phenomena studied, but even more so the nature of their data and the methods of investigation. Quantitative, corpus-based studies usually follow predetermined corpus categorizations that are, in practice, a mixture of registers, domains, genres and sub-genres (Lee 2001) conditioned by sheer matters of practicability (see Mondorf, this volume). However, if one is interested in the characteristics of a specific genre, a corpus that has only four broad categories, such as the one the Longman Grammar is based on (conversation, fiction, newspaper language, and academic prose), is not very helpful. Other problems arise if one turns to the language of emerging genres. Corpora typically are based on well-established categories, which are not of much use if one is interested in a register that is still very much in flux, such as computer-mediated communication today (Heyd; Puschmann, both this volume) or emerging genres in the course of a language’s history (Gotti; Obrist, Kabatek and Vincis, both this volume). Another problem for a corpus-based approach occurs if the phenomenon under scrutiny typically occurs in non-standard varieties of a language. Kortmann (2006: 603) points out that computerized corpora are now being compiled for non-standard varieties of English, “even if they cannot rival in size the megacorpora for the national standard varieties of English.” Among the variation phenomena typically found

16

Heidrun Dorgeloh and Anja Wanner

within corpora of non-standard varieties is the weakening of subject-verb agreement, which is also the subject of a pilot study in one contribution to this volume (Jantos). Finally, some aspects of form concerning genre are simply not reflected by quantities of occurrence, or the relevant constructions are not at all easy to retrieve from a corpus, as when a construction in question is not tied to any visible lexical material (such as the string -self or -selves for reflexive constructions), or because the corpus is not parsed, i.e. syntactically annotated (making it difficult, for example, to retrieve word order patterns). Issues of this kind are discussed quite explicitly in some of the corpus-based studies in this volume (Mondorf; Kreyer). One limitation of corpus-based methodology with regard to syntactic variation is that one can never be too sure that “different levels for particular features in different contexts [. . . ] genuinely reflect stylistic choices” (Schilling-Estes 2002: 394). For that reason, the study of intraspeaker variation has always tended to be also “qualitative in approach” (Schilling-Estes 2002: 377, our emphasis). It is especially the work that a feature performs in discourse, the first of the three tasks listed by Biber et al. (1999) and quoted above, that often has to be studied taking other methodological paths. In particular, this applies to detailed work on interactional data, where the format of one genre quite easily drifts into another one, this drift being marked linguistically (Imo, this volume), or where the use of a certain construction per se, not necessarily its frequency, can also become instrumental in constructing a genre (G¨unthner, this volume). Such discourse functions and the resulting relevance for genre is best unveiled by detailed analyses of syntactic constructions in context, i.e. by more text-based, qualitative work. 4.2. Overview of articles The studies included in this volume cover the full range from quantitative, corpus-based (Heyd; Jantos; Kreyer; Mondorf; Obrist, Kabatek, and Vincis; Pérez-Guerra and Martínez-Insua; Puschmann) to more qualitative, i.e. textbased, approaches to syntactic variation (Giltrow; Gotti; G¨unthner; Imo; Virtanen). Within the former group, several contributions make use of large-scale corpora (Jantos: Internation Corpus of English (ICE)-Jamaica; Kreyer: British National Corpus; Mondorf: LOB and Brown corpora as well as their Freiburg matches FLOB and Frown; Pérez-Guerra and Martínez-Insua: ARCHER corpus), while others refer to more small-scale, specialized databases, their design, as a consequence, being closer to the genre level, rather than to the register level (contributions by Heyd; Obrist, Kabatek, and Vincis; Puschmann). Mondorf, in her study of genre effects on syntactic variation, also offers a more principled consideration of the possibilities and limits of the corpus-based ap-

Introduction

17

proach, i.e. of quantified data. Other contributions follow a more qualitative approach, including the ones giving an overview and a more theoretical consideration of the issues of genre (Giltrow), communicative genre (G¨unthner), and text and discourse type (Virtanen); other qualitative studies are on genre emergence (Gotti) and merging (Imo). Notwithstanding these differences in methodology, the contributions are ordered depending on the question of whether they are more about a given (set of) genre(s) (contributions by Giltrow; Virtanen; Heyd; Puschmann; Gotti; Imo), or whether they primarily deal with syntactic, i.e. formal, phenomena (G¨unthner; Jantos; Kreyer; Mondorf; Obrist, Kabatek, and Vincis; Pérez-Guerra and Martínez-Insua). Within the group of studies on genre, the chapters by Giltrow and Virtanen are the ones most concerned with explicating theoretical positions: Giltrow explicitly emphasizes the relevance of form for defining genre, but also shows that most approaches to syntactic variation have so far dealt with the function(s) of form(s) at an all too high level of generality. She shows that although the category deontic modality, due to its general function of evoking authority, would be counter-indicated for occurrence in research genres, it nevertheless comes to be used regularly, but with more specialized functions that no longer have to do with obligation, but with shared interests of the researchers. Virtanen in her contribution reconciles form and function in a two-level model separating text types from discourse types, the latter being defined by function and thus also related to genre. Such a model can account for apparent mismatches between form and function in that direct and indirect uses of form (text types), signaled morphosyntactically (e.g., by clause-initial adverbials, tense, and referential choice) can still represent one and the same genre. As features of this kind reflect, encyclopedic articles show up in narrative, descriptive, and sometimes also in overtly expository form. Other contributions in this section focus on the interplay of genres and their form and also take into account detailed facts from the socio-pragmatic context: Gotti examines the dynamics of the emergence of a genre, the experimental essay, in 17th century English and pays particular attention not only to contentrelated and syntactic aspects (especially nominalization and premodification), but also to how these features came to be used and interpreted for the achievement of specific goals in specialized contexts. He focuses on the writing of a single author, the British scientist Robert Boyle, who is not just singled out because he made exemplary contributions towards the experimental essay as an established genre, but also because he wrote about this development. We also encounter a contemporary emergence of new genres, especially on the Internet, which is the point of interest in the contributions by Heyd and Puschmann: The study by Heyd is about email hoaxes, as part of the genre “ecology” of digital folklore. Using 2nd person plural forms as morphosyntactic variable, she shows

18

Heidrun Dorgeloh and Anja Wanner

that new genres typically trigger linguistic variation and change, such as providing a first point of entry into the written domain for previously oral forms. In a similar vein, Puschmann also finds patterns of pronominal expression that are highly characteristic of another emerging form of computer-mediated communication: (corporate) blogging. He finds that the function of self-reference of first and second-person pronouns is well-suited to the communicative contexts that hold for corporate blogs. While blogs appear to share many similarities with other, pre-digital, forms of writing, there are also other genres related to each other in format and context of occurrence: A case in point is presented in the contribution by Imo on two related radio phone-in formats; he shows that such different communicative, i.e. functionally defined, genres are often signaled as well as recognized by the interactants by way of syntactic constructions and set phrases, to the extent that such formal devices can come to signal a “hybridization” of genres. The approaches to syntactic variation range in diversity from what might still be considered one single morphosyntactic variable (Jantos; Mondorf), to sets of related constructions (G¨unthner; Kreyer) to finally more global syntactic (co-)occurrence patterns, presented as corpus-based profiles, such as for clause linkage techniques (Obrist, Kabatek, and Vincis) or noun phrase complexity (Pérez-Guerra and Martínez-Insua). In this section, Mondorf sets out with an empirical investigation of historical and present-day English corpus data on the replacement of reflexives by particles, such as it can be noted in change over time, as well as in different distributions according to medium, national standard, and genre. Results from different corpora indicate that the introduction and spread of the particle with formerly exclusively reflexive verbs affects different textual categories to different degrees. Jantos investigates non-standard subject-verb agreement in Jamaican English through an analysis of variation across three corpus categories – direct conversations, class lessons, and broadcast news – that differ in their relative level of formality and thus finds register to have a strong influence on grammatical non-agreement in Jamaican English. Both G¨unthner and Kreyer deal with a limited set of constructions: G¨unthner ties the use of Infinitkonstruktionen (‘infinite constructions’) and was-Fragen (‘what-questions’) to communicative genres of spoken German interaction; repeated use of these constructions proves instrumental in constructing these genres. Kreyer explores six constructions with fronted locative constituents in one written corpus category (fictional prose from the British National Corpus) and analyzes them as linguistic strategies that are typically used in the verbalization of spatial relations and thus for the recreation of immediate visual experience. In that sense, these constructions constitute plausible form-function correlates for the study of written narrative discourse.

Introduction

19

Finally, there are two exclusively quantitative approaches included in the volume: The paper by Pérez-Guerra and Martínez-Insua is concerned with finding syntactic reflections of textual formality, in this case by studying linguistic complexity in two genres and over time, namely letters and news from 1750 to present-day English. Using this approach towards a quantification of syntactic complexity, one can in principle place – and compare – genres on a scale of complexification. There is also a strong diachronic component in the contribution by Obrist, Kabatek, and Vincis. Here, clause linkage techniques are analyzed in a series of historical Romance corpora: juridical texts from 13th century Spain, an Old Spanish Bible corpus, a corpus containing texts of various genres in Old Spanish, and a Rheto-Romance newspaper corpus from the last two centuries. The authors find diachronic as well as synchronic variation in their data, not only in terms of different clause-linking techniques being used in different genres, but also in that the relative frequencies are symptomatic of different textual traditions, which change over time. For some genres, overt clause linkage declines over the years, but this affects certain semantic types more strongly than others and at the same time goes along with evolution of an overall more integrative style. Obrist, Kabatek, and Vincis (this volume: 251) therefore criticize studies that “consider diachrony as a linear evolution of one homogeneous object, with variation other than the diachronic one being nothing but a matter of noise in the data that can be easily compensated for by large quantities of data.” Seen from the context of this volume, it is therefore not so much the amount of data that determines whether we get a realistic view on syntactic variation, but the variety of methodologies, both synchronic and diachronic, as well as the taking into account of (con)textual traditions, i.e. genres, that provide us with a theoretical foundation for the inherent complexity of form-function relationships in syntax. References Askehave, Inger and John Swales 2001 Genre identification and communicative purpose: A problem and a possible solution. Applied Linguistics 22: 195–212. Auer, Peter (ed.) 2007 Style and Social Identities. (Language, Power, and Social Process 18.) Berlin/New York: Mouton de Gruyter. Beißwenger, Michael 2007 Sprachhandlungskoordination in der Chat-Kommunikation. (Linguistik – Impulse & Tendenzen 26.) Berlin/New York: de Gruyter. Biber, Douglas 1988 Variation Across Speech and Writing. Cambridge: Cambridge University Press.

20

Heidrun Dorgeloh and Anja Wanner

Biber, Douglas 1989 A typology of English texts. Linguistics 27: 3–43. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad and Edward Finegan 1999 Longman Grammar of Spoken and Written English. Harlow: Longman. Biber, Doublas and Susan Conrad 2009 Register, Genre, and Style. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Biberauer, Theresa 2008 Introduction. In: Theresa Biberauer (ed.), The Limits of Syntactic Variation (Linguistik Aktuell/Linguistics Today 132.), 1–72. Amsterdam: John Benjamins. Bowerman, Melissa 1974 Learning the structure of causative verbs: A study in the relationship of cognitive, semantic, and syntactic development. Papers and Reports on Child Language Development 8: 142–178. Brown, Penelope and Stephen Levinson 1978 Universals in language usage: Politeness phenomena. In: Esther Goody (ed.), Questions and Politeness: Strategies in Social Interaction, 56–311. Cambridge: Cambridge University Press. Burzio, Luigi 1986 Italian Syntax: A Government-Binding Approach. (Studies in Natural Language and Linguistic Theory.) Dordrecht: Kluwer. Bybee, Joan and Paul Hopper (eds.) 2001 Frequency and the Emergence of Linguistic Structure. (Typological Studies in Language 45.) Amsterdam/Philadelphia: John Benjamins. Carston, Robyn 1996 Syntax and pragmatics. In: Keith Brown and Jim Miller (eds.), Concise Encyclopedia of Syntactic Theories, 306–313. Oxford: Pergamon. Chafe, Wallace (ed.) 1980 The Pear Stories: Cognitive, Cultural and LinguisticAspects of Narrative Production. New York: Ablex. Chen, Ron 2003 English Inversion: A Ground-Before-Figure Construction. (Cognitive Linguistics Research 25.) Berlin/New York: Mouton de Gruyter. Chomsky, Noam 1965 Aspects of the Theory of Syntax. Cambridge, MA: MIT Press. Chomsky, Noam and Howard Lasnik 1993 The theory of principles and parameters. In: Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld and Theo Vennemann (eds.), Syntax: An International Handbook of Contemporary Research, 506–569. Berlin: Walter de Gruyter.

Introduction

21

Croft, William 2010 The origins of grammaticalization in the verbalization of experience. Linguistics 48: 1–48. Deh´e, Nicole 2002 Particle Verbs in English: Syntax, Information Structure, and Intonation. (Linguistik Aktuell/Linguistics Today 59.) Amsterdam/Philadelphia: John Benjamins. Diller, Hans-J¨urgen and Manfred G¨orlach (eds.) 2001 Towards a History of English as a History of Genres. Heidelberg: Winter. Eckert, Penelope 2000 Linguistic Variation and Social Practice. Oxford/Malden, MA: Blackwell. Eckert, Penelope and John Rickford (eds.) 2001 Style and Sociolinguistic Variation. Cambridge: Cambridge University Press. Ferguson, Charles 1994 Dialect, register, and genre: Working assumptions about conventionalization. In: Douglas Biber and Edward Finegan (eds.), Sociolinguistic Perspectives on Register, 15–30. Oxford: Oxford University Press. Ferreira, Fernanda and Paul Engelhardt 2006 Syntax and production. In: Morton Traxler and MortonAnn Gernsbacher (eds.), Handbook of Psycholinguistics, 2nd Edition, 61–91. Amsterdam: Elsevier. Fischer, Olga, Ans van Kemenade, Willem Koopman and Wim van der Wurff 2000 The Syntax of Early English. Cambridge: Cambridge University Press. Fitzmaurice, Susan and Irma Taavitsainen (eds.) 2007 Methods in Historical Pragmatics. (Topics in English Linguistics 52.) Berlin/New York: Mouton de Gruyter. Flottum, Kjersti, Trine Dahl and Torodd Kinn (eds.) 2006 Academic Voices: Across Languages and Disciplines. (Pragmatics & Beyond New Series 148.) Amsterdam/Philadelphia: John Benjamins. Goldberg, Adele 1995 Constructions: A Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. Gries, Stefan 2002 The influence of processing on syntactic variation: Particle placement in English. In: Ray Jackendoff, Nicole Deh´e, Silke Urban and Andrew McIntyre (eds.), Verb-Particle Explorations. (Interface Explorations 1.) 269–288. Berlin/New York: Mouton de Gruyter. Gries, Stefan 2003 Grammatical variation in English: A question of ‘structure vs. function.’ In: G¨unter Rohdenburg and Britta Mondorf (eds.), Determinants

22

Heidrun Dorgeloh and Anja Wanner

of Grammatical Variation in English, 155–173. Berlin/New York: Mouton de Gruyter. Grodzinsky, Yosef 2003 Imaging the grammatical brain. In: Michael A. Arbib (ed.) Handbook of Brain Theory and Neural Networks, 551–556. Cambridge, MA: MIT Press. Halliday, Michael 1978 Language as Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Halliday, Michael 1985 Functional Grammar. London: Edward Arnold. Hawkins, John 2004 Efficiency and Complexity in Grammars. Oxford: Oxford University Press. Herring, Susan (ed.) 1996 Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural Perspectives. Amsterdam/Philadelphia: John Benjamins. Hoffmann, Thomas 2005 Variable vs. categorical effects: Preposition pied piping and stranding in British English relative clauses. Journal of English Linguistics 33: 257–297. Hornstein, Norbert and Amy Weinberg 1981 Case theory and preposition stranding. Linguistic Inquiry 12: 55–91. Hyams, Nina 1986 Language Acquisition and the Theory of Parameters. (Studies in Theoretical Psycholinguistics.) Dordrecht/Boston: Reidel. Irvine, Judith 2001 ‘Style’as distinctiveness: The culture and ideology of linguistic differentiation. In: Penelope Eckert and John R. Rickford (eds.), Style and Sociolinguistic Variation, 21–43. Cambridge: Cambridge University Press. Jucker, Andreas 1992 Social Stylistics: Syntactic Variation in British Newspapers. (Topics in English Linguistics 6.) Berlin/New York: Mouton de Gruyter. Jucker, Andreas (ed.) 1995 Historical Pragmatics: Pragmatic Developments in the History of English. Amsterdam/Philadelphia: John Benjamins. Kortmann, Bernd 2006 Syntactic variation in English: A global perspective. In: Bas Aarts and April MacMahon (eds.), The Handbook of English Linguistics, 603–624. Malden, MA/Oxford: Blackwell.

Introduction Kreyer, Rolf 2006

23

Inversion in Modern Written English: Syntactic Complexity, Information Status and the CreativeWriter. (Language in Performance 32.)T¨ubingen: Gunter Narr.

Labov, William 1972 Language in the Inner City: Studies in Black English Vernacular. Philadelphia: University of Pennsylvania Press. Langacker, Ronald 1991 Foundations of Cognitive Grammar, Volume II: Descriptive Application. Stanford: Stanford University Press. Lee, David 2001 Genres, registers, text types, domains, and styles: Clarifying the concepts and navigating a path through the BNC jungle. Language Learning and Technology 5: 37–72. MacWhinney, Brian 2000 The CHILDES Project: Tools for Analyzing Talk. Mahwah, NJ: Lawrence Erlbaum. Mair, Christian and Geoffrey Leech 2006 Current changes in English syntax. In: Bas Aarts and April MacMahon (eds.) Handbook of English Linguistics, 318–342. Malden, MA/Oxford: Blackwell. Mair, Christian and Hans Lindquist (eds.) 2004 Corpus Approaches to Grammaticalization in English. Amsterdam/Philadelphia: John Benjamins. Maratsos, Michael, Dana Fox, Judith Becker and Mary Anne Chalkley 1985 Semantic restrictions on children’s passives. Cognition 19: 167–191. Martin, J.R. and David Rose 2003 Working with Discourse: Meaning Beyond the Clause. London: Continuum. Mauner, Gail and Jean-Pierre Koenig 1999 Lexical encoding of event participant information. Brain and Language 68: 178–184. Miller, Carolyn 1984 Genre as social action. Quarterly Journal of Speech 70: 151–167. Milroy, James 1992 Linguistic Variation and Change. Oxford: Blackwell. Mondorf, Britta 2003 Support for more-support. In: G¨unter Rohdenburg and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 251–304. Berlin/New York: Mouton de Gruyter. Mondorf, Britta 2004 Gender Differences in English Syntax. (Linguistische Arbeiten 491.) T¨ubingen: Max Niemeyer.

24

Heidrun Dorgeloh and Anja Wanner

Mondorf, Britta 2009 More Support for More-Support: The Role of Processing Constraints on the Choice between Synthetic and Analytic Comparative Forms. (Studies in Language Variation 4.) Amsterdam/Philadelphia: John Benjamins. Mukherjee, Joybrato 2005 English DitransitiveVerbs: Aspects of Theory, Description, and a UsageBased Model. Amsterdam: Rodopi. Nevalainen, Terttu and Leena Kahlas-Tarkka (eds.) 1997 To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen. (M´emoires de la Soci´et´e N´eophilologique de Helsinki.) Helsinki: Soci´et´e N´eophilologique. Nevalainen, Terttu, Irma Taavitsainen, P¨aivi Pahta and Minna Korhonen 2008 The Dynamics of Linguistic Variation: Corpus Evidence on English Past and Present. Amsterdam/Philadelphia: John Benjamins. Newmeyer, Frederick 2010 What conversational English tells us about the nature of grammar: A critique of Thompson’s analysis of object complements. In: Kasper Boye and Elizabeth Engberg-Pedersen (eds.), Usage and Structure: A Festschrift for Peter Harder, 3–43. Berlin/New York: Mouton de Gruyter. Penke, Martina and Annette Rosenbach 2004 What counts as evidence in linguistics? An introduction. Studies in Language 28: 480–526. Pilegaard, Morten and Finn Frandsen 1996 Text type. In: Jef Verscheuren, Jan-Ola Östman, Jan Blommaert and Chris Bulcaen (eds.), Handbook of Pragmatics, 1–12. Amsterdam/Philadelphia: John Benjamins. Pinker, Steven, David Lebeaux and Loren Frost 1987 Productivity and constraints in the acquisition of the passive. Cognition 26: 195–267. Pintzuk, Susan and Anthony Kroch 1989 The rightward movement of complements and adjuncts in the Old English of Beowulf. Language Variation and Change 1: 115–143. Quaglio, Paulo and Douglas Biber 2006 The grammar of conversation. In: Bas Aarts and April MacMahon (eds.), The Handbook of English Linguistics, 692–723. Oxford: Blackwell. Radden, G¨unter and Ren´e Dirven 2007 Cognitive English Grammar. Amsterdam/Philadelphia: John Benjamins. Radford, Andrew 1990 Syntactic Theory and the Acquisition of English Syntax. Oxford: Blackwell. Rohdenburg, G¨unter and Britta Mondorf (eds.) 2003 Determinants of Grammatical Variation in English. (Topics in English Linguistics 43.) Berlin/New York: Mouton de Gruyter.

Introduction

25

Rosenbach, Annette 2002 Genitive Variation in English: Conceptual Factors in Synchronic and Diachronic Studies. (Topics in English Linguistics 42.) Berlin/NewYork: Mouton de Gruyter. Rosenbach, Annette 2008 Animacy and grammatical variation: Findings from English genitive variation. Lingua 118: 151–171. Sanchez-Maccaro, Antonia and Ronald Carter (eds.) 1998 Linguistic Choice across Genres: Variation in Spoken and Written English (Current Issues in Linguistic Theory 158). Amsterdam/Philadelphia: John Benjamins. Sankoff, David 1988 Sociolinguistics and syntactic variation. In: Frederick Newmeyer (ed.), Linguistics: The Cambridge Survey, 140–161. Cambridge: Cambridge University Press. Schiffrin, Deborah 1994 Approaches to Discourse. Oxford: Blackwell. Schilling-Estes, Natalie 2002 Investigating stylistic variation. In: J.K. Chambers, Peter Trudgill and Natalie Schilling-Estes (eds.), The Handbook of Variation and Change, 374–401. Oxford: Blackwell. Schulze, Rainer (ed.) 1998 Making Meaningful Choices in English: On Dimensions, Perspectives, Methodology and Evidence. T¨ubingen: Gunter Narr. Siever, Torsten, Peter Schlobinski and Jens Runkehl (eds.) 2005 Websprache.net: Sprache und Kommunikation im Internet. (Linguistik – Impulse & Tendenzen 10.) Berlin/New York: Mouton de Gruyter. Stefanowitsch, Anatol 2003 Constructional semantics as a limit to grammatical alternation: The two genitives of English. In: G¨unter Rohdenburg and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 413–441. (Topics in English Linguistics 43.) Berlin/New York: Mouton de Gruyter. Stein, Dieter and Janet Giltrow 2009 Genres in the Internet: Innovation, evolution, and genre theory. In: Janet Giltrow and Dieter Stein (eds.), Genres in the Internet: Issues in the Theory of Genre, 1–25. Amsterdam/Philadelphia: John Benjamins. Swales, John 1990 GenreAnalysis: English inAcademic and Research Settings. Cambridge: Cambridge University Press. Swales, John 2004 Research Genres: Explorations and Applications. Cambridge: Cambridge University Press.

26

Heidrun Dorgeloh and Anja Wanner

Taavitsainen, Irma and Susan Fitzmaurice 2007 Historical pragmatics: What it is and how to do it. In: Susan M. Fitzmaurice and Irma Taavitsainen (eds.), Methods in Historical Pragmatics, 11–36. (Topics in English Linguistics 52.) Berlin/New York: Mouton de Gruyter. Tomasello, Michael 2003 Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, MA: Harvard University Press. Ward, Gregory and Betty Birner 1998 Information Status and Non-Canonical Word Order in English. Amsterdam/Philadelphia: John Benjamins. Wasow, Thomas and Jennifer Arnold 2003 Post-verbal constituent ordering in English. In: G¨unter Rohdenburg and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 119–154. (Topics in English Linguistics 43.) Berlin/New York: Mouton de Gruyter.

Part 1: Genre: Form, constitution, and change

Genre as difference: The sociality of linguistic variation Janet Giltrow

In the study of syntactic variation, genre has been an unstable term: fluctuating in the level of generality at which it is applied; intuiting rather than ascertaining the social situations it suggests. In contrast, rhetorical studies of genre have fixed genre at a low level of generality, in local socio-historical scenes, and claimed priority for situation over form. This chapter reviews the debates which led to genre’s rhetorical definition as “social action,” (Miller 1984) and the benefit and also the cost in disavowing form as definitive of genre. Rhetorical and variation studies of genre can seem fundamentally incompatible in their perspectives on form, yet there may be fertile meeting ground for them. Both invoke function for form, but each in ways incomplete for the study of genre: While rhetorical studies insist on the functionality of form in situation, they do not inquire into form itself; syntactic-variation studies also assume function for form, but define function at such a high level of generality as to fail to capture the social motives in genre’s domain. By revisiting function as acutely sensitive to situation – responding to local exigencies and also indicating them – rhetorical and variation studies can meet on the common ground of form, each contributing to the other’s discoveries of genre as a site of social differentiation.

1. Introduction In the language disciplines, genre has been a term both easily summoned and easily displaced. Easily summoned, it can name what people recognise as broad similarities in ways of thinking or it can name much narrower formations – predictable wordings or familiar collocations. Easily displaced, it can give way to discourse in the broader perspectives or style or register in the narrower ones. This consortium of terms recognises at once formal regularity and social ties: Although long associated with form and formality, genre also fraternises easily with discourse and register to indicate socialities of speech style, often alluding to institutional or professional settings as in, for example, “medical discourse” or “legal register.” Genre can be a term with poor traction perhaps, liable to

30

Janet Giltrow

theoretical slippage; it is also one which comes to hand to link form and scene, language and society. And it may be this coupling which tends to disable and dislocate the term. When the dynamics between form and scene are interrupted for inspection – that is, when genre is the object of inquiry – form tends to separate from scene. In different language disciplines, form then takes different roles: as mere formality, for example, or as compliance with convention, or as features’ replication and variation. In genre as an everyday speech phenomenon, rather than an object of study, feature and activity “knit,” as Bakhtin says in The Dialogic Imagination (1981: 288). But in The Problem of Speech Genres he also says that genres are hard to study (1986: 61). The problem may be that, under study, the knitting ravels, and form is undone from scene. Or that is the tendency. This chapter briefly reviews some uses and applications of the term genre, especially in its back and forth with register in linguistic approaches to language variation. Turning to other disciplines’approaches to genre, the chapter mentions recent interest in genre in literary studies but concentrates mainly on recent decades of discussion of genre in rhetorical studies and writing studies. Newrhetorical conceptions of genre insisted from the start on the fusion of form and situation, yet form still caused problems, some of which I will outline. These problems were eventually addressed by declaring for situation over form (and in one bold proposal eliminating form all together). But, as I will show, form continued to haunt the project, returning as various issues and nagging concerns. Equally, genre continues to attend studies of formal variation, although still unstable as a category. Evidently, form belongs with genre, as we might expect from the use of the term when speakers are noticing form and its social ties. The chapter concludes with a demonstration of form under pressure from function – but function at a low level of generality, taking the direct imprint of the social scene. This is the level at which we should look for genre.

2. Genre and register In the linguistic disciplines, genre has been a relatively untheorised term. When summoned to lay the groundwork or establish the territory for the study of linguistic variation, it can assume rather than examine similarity among instances. Such was Biber and Finegan’s (1988) finding in their important studies of the linguistic variation they call stance styles: In our earlier work, the texts within each genre were assumed to constitute a coherent linguistic whole, and each traditional genre category was assumed to be linguistically distinct from the others. Subsequent research of ours has indi-

Genre as difference: The sociality of linguistic variation

31

cated that this assumption is not valid; in fact, for some genre categories, greater linguistic differences exist among texts within the categories than across them. (Biber and Finegan 1988: 3)

In other words, the categories deductively assumed as genres were not demonstrated by the linguistic evidence (a hint, perhaps, that the categories were not, in fact, genres). Seeking groupings according to formal features, Biber and Finegan’s methods found them not by means of their genre categories but by means of cluster analyses, analyses which discover text types according to speech styles: “speech styles are [. . . ] sets of texts that are similar in linguistic form” (Biber and Finegan 1988: 3). Register, in the meantime, groups texts according to “the relations among participants and other characteristics of the communicative situation,” while “genre” groups texts “according to topic and purpose” (Biber and Finegan 1988: 4). Yet the means of measuring participant relations and situation characteristics can be uncertain, despite the field/tenor/mode parameters donated by Systemic Functional Linguistics (SFL), and so can the means of knowing topic and purpose as criteria unifying a set of texts. As Askehave and Swales (2001: 200) observe in their discussion of “genre identification”, “what is immediately evident to the genre analyst is not purpose but form and content.” Purpose has been taken for granted in their view, either by the very general genre sets named by analysts, such as “instruction, description, recount” (Askehave and Swales 2001: 206) in the SFL-derived series of Sydney School genre study, or by the terms derived from qualitative study, where researchers accept language-users’ answers to questions about the purpose of the genres under study. Such names may themselves be as untested as those which researchers come up with – and may indeed come from the same fund of common experience. Askehave and Swales recommend that communicative purpose be recursively revisited in genre study, for what they call a “re-purposing” (Swales 2001: 207), always in recognition of the complexity, we might add, of speakers’ motives in taking up a way of speaking. Despite continuing mention of genre in formal corpus studies of linguistic variation across text instances, and despite attempts to distinguish between genre and register, genre on the whole seems to have given way to register. By 2006, Biber suggests that genre and register are interchangeable, and opts for register to name “situationally-defined varieties described for their characteristic lexicogrammatical features” (Biber 2006: 11). Register itself is a porous term, as when Biber, Csomay and Jones (2004) refer to both “written narrative” and scientific research articles as registers, these two classifications deriving from different criteria. Such incommensurability sometimes goes unnoticed, or it may be part of an expedient instability which register inherits from genre. Biber (2006: 12)

32

Janet Giltrow

says register can be named at any level of generality, and one will find more formal similarity at low levels, less at higher levels. Yet even as register is in the ascendancy, what could be called genre is still playing a role in identifying social scenes for speech – classroom talk, or live exchanges on television shows, or guest lectures, to take examples from a recent collection (Partington, Morley, and Haarman 2004) of corpus-based studies of linguistic variation – without those identifications being examined for their criteria or definition, for their measure of purpose, let alone for their complex articulation with the social order. And described for their social purpose, genres can still be known only intuitively, as when Bruti (2004) describes the genre of the biology textbook as “a kind of specialised text, especially designed for didactic purposes, i.e. for an audience who needs to be taught the basic notions of the discipline which are generally ignored in expert-to-expert communication because they are part of the specialists’shared knowledge. As a consequence, the genre is one that presents specialist knowledge couched in a didactic style” (Bruti 2004: 125). Although still only an intuitive category, genre is nevertheless called on to establish the territory of many studies of linguistic variation, appearing and disappearing opportunistically. Current studies of linguistic variation, by increasingly sensitive methodologies, uncover the complexity of form, finding interlacing veins of formal similarity and variation. These veins can be, and often are, mined for function. There might be promise here of reattaching form to sociality and thereby reincorpo-

Terms which can displace or be displaced by genre

Terms which can render genre

Terms which fraternise with genre: those in the first column are more likely to be stipulated; those in the second are more likely to devalue the phenomenon Term which may or may not be stipulated Categories intuited by analysts (e.g., academic, guest lecture) or inherited from traditions of analysis (e.g., narrative, description; forensic, epic)

Figure 1. The terminological field of genre

- speech style - register - text type

- formulaic wordings - predictable, routine speech - form

- discourse

- available speech categories

Genre as difference: The sociality of linguistic variation

33

rating form in genre. But function is by and large named at such a high level of generality – for example, evaluation, involvement, clarification, or stance – that the promise is unfulfilled, and genre remains an abstract category rather than the site of social interaction, that is, a site which is historically situated and open to the experience of language users.

3. New-rhetorical1 concepts of genre Considered a sterile concept for the very formalism which has, in the study of linguistic variation, proved productive, genre has until relatively recently been left fallow in the adjacent fields of literary and rhetorical studies. In literary studies, there have been signs of a genre revival. As long ago as the mid 1980s, Cohen (1986) contradicted genre orthodoxies – namely, that genres were determinate classes – by introducing what might be called a socio-formal view: The naming of genres was a social process and open to the transformation of forms’ meanings and functions, according to historical contexts. More recently, an issue of the Publications of the Modern Language Association (PMLA, October 2007), literary criticism’s flagship journal, is devoted to re-thinking genre, without much advance over Cohen’s earlier claims for form, and possibly losing ground to the orthodoxies. The term genre once again appears to be available for any category of similarity, from the ancient genres of epic or lyric, to entirely unsuspected types detected by the most nuanced of literary-critical interpretation, to notice of cross-cultural transfers of motifs and figures. Despite efforts to rehabilitate genre, form appears once again as whipping boy: a policing regime or pigeonholing disdained by literary sensibility. But it also appears as a new dispensation: a “liquefaction” (Dimock 2007: 1378), a “general solvent,” a “pool” of features in which texts “swim” (Dimock 2007: 1379). In being separate from situation and the social order, the swimming pool may not be entirely different from the currents of features explored by studies of linguistic variation, but its study is methodologically much less disciplined. At this point in the literarycritical revival of genre studies, there seems to be little on offer to advance the study of linguistic variation. 1. New rhetorical here can mean, generally, new concepts of genre in rhetorical studies, and, more specifically, concepts deriving from the new rhetoric. The new rhetoric extends the notion of persuasion beyond the classical and obvious cases of language designed to manipulate audiences’ beliefs and actions to include all uses of language (see Virtanen, this volume, on argumentation). New rhetoricians hold with Kenneth Burke’s claim that all language is persuasive and his recognition of “the necessarily suasive nature of even the most unemotional scientific nomenclature” (Burke 1966: 45).

34

Janet Giltrow

Rhetorical studies, however, have more to offer. In U.S. rhetorical studies, mainly neo-classical and concerned deeply but not exclusively with the speech of American leaders, the revival of the concept of genre began in the late 1960s and developed in the 1970s and 1980s a rich discussion of issues and principles. In the 1990s and into this century the discussion spread into the study of writing in institutions, workplaces, and the public sphere – some of the areas occupied by linguistic study of what are called “specialized” texts (e.g., Gotti, this volume). In this latter phase, the discussion also spread beyond America. This chapter recommends new-rhetorical concepts of genre: These concepts stabilise the level of generality at which the term is applied and re-attach it to the social order and the socio-historical experience of language users. Genre can become, thereby, a means of re-connecting to the social order the rich information now flooding in from corpus studies of linguistic variation. While recommending new-rhetorical concepts of genre, I also point to the deficiencies of these concepts and ways those deficits can be compensated for by the study of linguistic variation. For rhetorical concepts of genre made the advances they did by disavowing form and acclaiming situation as the signal criterion for the category. This conceptualisation provides means for culling merely intuitive or unexamined or inherited applications. But it also leaves form as the problem for genre rather than the site of genre’s energies, of genre’s capacity to effect the means and motive of variation. For both linguistic and rhetorical studies, the problem for genre has been how it sorts with form. In both fields, the problem has been addressed by divesting genre of formal definition. In linguistic studies, genre works a-theoretically to rationalise corpora or gives way to avowedly formal study under the auspices of text type, as in Pérez-Guerra and Mart´ınez-Insua (this volume). In rhetorical studies, form has been dismissed from the scope of investigation. But, as I will show below, form does not go quietly: It returns to pester the rhetorical project. Equally, genre, with the speech phenomenon it names, may be waiting to be invited back as a serious player in the study of linguistic variation, not simply to oust register or to cancel text type but to find the level at which differentiation is socially meaningful. 3.1. “Genre as social action” The landmark essay in rhetorical genre theory is Carolyn Miller’s Genre as social action (1984). It synthesised 15 years of preceding discussions in the U.S., and also raised the stakes. Miller argues that genres are to be classified not by their forms but by the actions they perform, this performance being mutually recognised by groups of language users. So categories which disappoint or

Genre as difference: The sociality of linguistic variation

35

merely contain the study of linguistic variation could be interrogated for “social action”: Can instances captured under a general category be said to perform the same social action?2 From a definition of genre which classified not types of texts but types of social action, certain principles followed as Miller’s proposals were taken up and applied in many research venues: – Genre cannot be separated from the social scenes of its performance; it cannot be understood separate from situation. (So, “recount” would not be a genre, nor would “description,” for neither involves a social scene: no inferable motives for recounting or describing. What moves someone to recount or describe?) – Once genre is inseparable from situation, the set of genres must open. As forms of life – work, trade, family, government, institution – change, so too must genres. If a form of life appears, disappears, or transforms, genres must also appear, disappear, or transform. – Tied to history and locality, the set – open and contingent – drops down from a handful of timeless universals to dated, ground-level inventories of cultural occasions: those mutually recognisable moments when language users feel that something of a certain kind should be said. So, for example, narrative is not a genre but a workplace-incident report is a genre – for employees at my university, for example. This can seem highly particular, but it may be the strict locality of genre that is also its flexibility. Rather than adherence to rules, it is a linguistic experience of roles, material circumstances, and personal histories, and thus prone to revision and adjustment, always open to recognising close similarities (for example, between incident reports at different unionised workplaces) and more distant similarities (for example, to accident reports to police or to insurance bureaus). – Local and context sensitive, genre knowledge is tacit. You can’t learn a genre by being told the rules, that is, by being instructed in the form.You have to experience the situation, which itself is socio-historically, culturally embedded. This principle would challenge the pedagogical goals professed by corpus studies aiming to improve L2 instruction by making a full inventory of the formal features of target genres. But more importantly, this principle, recog-

2. Equally, one might ask if formally identical instances used in different contexts perform the same social action. The telling example from later studies of genre is the business letter. Does a text using the format and wording of a typical business letter perform the same action when composed by a student in a business-writing class as it does when composed by a manager in a corporate setting?

36

Janet Giltrow

nisable from common experience, corroborates the theoretical priority of situation in new-rhetorical genre theory. By disavowing form, which cannot account for situated experience, or even for itself – that is, it cannot account for why text features recur across instances – rhetorical genre theory opened many new perspectives and research directions: ethnographic- or qualitative-style research especially. But in its disavowal in the 1980s and the 1990s and in this century, too, form has perhaps been taken for granted and remains a troubling presence, disturbing the stages of discussion in rhetorical theory which led up to Miller’s landmark article, and returning to haunt the discussion which followed, to this day. 3.2. Trouble with form In the discussion that led up to Miller’s synthesis, form was at one level taken for granted, and treated as well understood. Possibly because form was traditionally apprehended in rhetorical tropes and schemes, “nothing more complicated than patterns of arrangement” (Gronbek 1978: 140), rhetorical appeals (logos, ethos, pathos), and the common and special topoi, form may indeed have been selfevident to rhetorical theorists in the 15 years of discussion which culminated in Miller’s article. And many genre theorists, as if to confirm this established understanding, cited Kenneth Burke’s statement on form as “the creation of an appetite in the mind of the auditor, and the adequate satisfying of that appetite” (Burke 1954: 31). But while form was apparently taken for granted at one level, in the early stages of the renewal of interest in genre, it was still the pivot of claims for genre’s significance, and as Devitt (2009) demonstrates, still centrally involved in the reasoning of the period. One important advocate of rhetorical genre theory claimed in 1973 – ten years before Miller put form to one side – that “[w]hen one knows what characteristics will inform an inaugural [presidential speech] not yet composed, one has isolated the generic membranes of the inaugural” (Jamieson 1973: 163, emphasis mine), “generic membranes” being form. Students of linguistic variation, particularly those who catalogue genre features on behalf of L2 learners, might recognise and share this goal. But not all American rhetoricians of the period were happy with predictability. Some theorists were ambivalent, one of these both shunning and entertaining predictability: Genre “should not be thought of as a predictive category necessarily,” but “[. . . ] we might be able to interpret or predict” according to the presence of certain conditions in the socio-political situation (Hart 1973: 251, 261). Others were briskly dismissive, seeing genre as a name for what is routine and trivial in speech, genre being only form, or mere formality. For one critic, a certain type of event may be “so ritu-

Genre as difference: The sociality of linguistic variation

37

alized that it is uninteresting to analyze it rhetorically” (Vatz 1973: 160). While students of linguistic variation separate genre from form, these opponents of genre criticism equate genre with form and, in addition, with mechanical compliance, a kind of automatic speaking. For another critic, it is not this aspect of language that rewards study, but the creative parts of a speech; these parts, not the predictable parts, are the interesting ones. In dismissing genre as unworthy of study, this critic is dismissing what he calls “conventionalized” messages (e.g., scientific discourse) (Wilkerson 1970: 89–91), and possibly what for students of linguistic variation has become the rich field of “specialised texts.” These are just a few samples of the period’s extensive discussion on predictability and choice, routine and surprise, these themes all attaching to the recurrence of form. And among critics who advocated genre study – those for whom form was not so oppressive a presence – predictability was a more involving problem than it might be if it were only a matter of calculating text features and extrapolating future instances. It was other than this, however, for function had entered the picture in the first stages of the genre revival. So one prominent critic said that once we knew the form/function formula for a situation, we could predict the instance, even if it occurred only once, or never (Black [1965] 1978: 137). (We might pause to notice how far this claim is from the assumptions supporting the study of linguistic variation, which, with quantitative methods, would be unlikely to discover the single instance, let alone the zero instance. But we might also notice, before we go on, that this position shares with the study of linguistic variation a comfort with prediction based on function, and, in addition, an intimation of a system of form-function resources, some of which may remain unexploited.) Another prominent critic said that a person who had never heard a eulogy would, on the occasion of the death of a friend or leader, rise to speak in eulogistic ways (Jamieson 1973: 163), so functional are the forms of eulogistic speech in the situation of a recent death. On this view, form is functional rather than conventional, that is, it derives from linguistic resources rather than from standards or customs of use.Yet this is not exactly the principle established by the formulation advanced by situational rhetoric (Bitzer 1968, restated 1980), the foundational statement from which much of new genre theory developed. Bitzer saw situations as evolving in recurring, their recurrence activating functional speech, itself recurring accordingly, and gathering audiences and constraints, the latter including rhetorical forms. Situations (recurring) → Functional Speech (recurring) [recurring Functional Speech = formal regularities (including “rhetorical forms”)]

While the form of speech in the first instance is activated by its functionality, its recurrence in subsequent instances must be owing in addition to an accumulating

38

Janet Giltrow

social experience of the form – still functional rather than conventional but also not only functional. In its first appearance, and its many subsequent applications by others, Bitzer’s original, influential statement aroused some objections: Opponents complained that the critic’s role as evaluator of aesthetic achievement and adjudicator of ethical responsibility was cancelled if speakers are simply and automatically responding to situation. Bitzer himself must have been sensitive to this kind of attack for he defended his proposals by saying there is no contradiction between predictability and “freedom and creativity” (Bitzer [1968] 1980: 34). While students of linguistic variation are not so worried about freedom and creativity, the form/function problem may not be so far away. Can matches and differences in function alone explain the replication and variation of form? Is each speaker linked directly to the system of linguistic resources, or is the link mediated by interaction with other speakers? In emphasising situation, Bitzer emphasised function. But he did not deny sociality: The recurrence of situations must be a social matter, and so must the accumulation of formal regularities – the fluent transfer of wordings amongst speakers involved in these situations. In addition, in the matter of the set of genres, rhetorical theorists and students of linguistic variation may share issues. We can see that Bitzer’s situational rhetoric opens the door to new situations and functional responses to these situations: new genres or the open set. But for most genre critics who followed Bitzer’s lead, the set was left closed, Aristotle’s classical categories (epideictic, forensic, deliberative) doing a brisk business. For classically trained rhetoricians in America, the closed set may have been politically satisfying. A timeless handful of universal types, the closed set spoke to the American rhetorical taste for keeping in sight at once the ancient instance and the contemporary moment, Socrates and Richard Nixon, both supervised by classical theory. In the study of linguistic variation, the closed set also has its attractions, although different ones. The closed set has oriented study of formal variation, but the closed set in linguistic study has not been shut tight. As Virtanen (this volume) notes, while narrative, for example, remains a common term amongst the many versions of the closed set, other terms come and go, a problem Virtanen partly addresses by distinguishing between text type (more formal) and discourse type (more functional). In American rhetorical theory, the terms of the closed set, inherited from and mandated by antiquity, were non-negotiable. Moreover, the closed set alleviated the problem of predictability. From the high level of generality of the closed set, predictability does not show as a problem: Instances are alike in such a general way that their difference rather than their similarity may be the salient experience, and only a specialist critic would be able to see the resemblance, so

Genre as difference: The sociality of linguistic variation

39

abstract are its terms. So, if we simply predict epideictic speech in the situation “death” then we can still be surprised by the instance and its inventiveness. If, however, we drop to a lower level of generality to identify not just “eulogy” but “mourners’ brief, serial recollections of the dead person at Canadian memorial services in the late 20th century” we can begin to predict the statements of the speakers. This is precisely the range of difference and likeness Biber (2006) observes: the higher and more general the register, the fewer the likenesses; the lower and more particular, the more numerous the likenesses. Attractive as the closed set was, many critics nevertheless recognised the difficulties in applying Aristotelian categories to modern discourse and found hybrids and half-types. But tinkering with the timeless categories was not enough to stop genre theory from moving towards the open set – the set opening for, for example, the social worker’s predisposition report or the tax accountant’s letter or, more recently, the homeless person’s blog, just as it opens in the study of linguistic variation for corporate blogs (Puschmann, this volume), live TV exchanges or guest lectures or biology textbooks, even while general categories continue to preside at their own level. As new genre theory transferred from the reasoning of classically trained rhetoricians to other areas of discourse studies – and eventually from America to other national discussions – the door to the set of genres opened wide. Genre theory was at the open door when Miller’s “social action” article was published. (Miller herself was studying Environmental Impact Statements as a genre, not Socrates or Nixon.) And Miller’s disavowal of formalism, her claim for the priority of situation over linguistic features, may have seemed to put an end to the skirmishes over form: its bringing into disrepute genre itself as mere formality. But form would not be dismissed so easily. Ten years after Miller’s statement, Devitt (1993) wrote about genre for a wider, but still American, audience, defending it against suspicions of being a name for merely formal compliance. On genre’s behalf, she bids for the attention of a sceptical audience of writing specialists; these specialists made up a constituency which had replaced product with process, a renovation which made form, and accordingly, in their view, genre irrelevant. Pleading for genre, Devitt emphasises choice and efficiency. Writers choose genres, choose “conventions” within genres to suit “their situation” (Devitt 1993: 579): Choosing genres, writers find themselves with a “template” (Devitt 1993: 582). Even though Devitt is as convincing as Miller in disavowing form, form returns, for what else could a “template” be? Or “convention,” or even genre? With form back on the table, genre is now recommended for efficiency: Writing would be slowed if we didn’t have genres (Devitt 1993: 576). The form has already been vetted for function, by prior speech or speakers. It is on the basis of form – and its pre-selected suitability for situation – that genre can be recommended for respectful attention.

40

Janet Giltrow

By the mid 1990s, freedom and creativity were making less mischief for genre, but the concept of genre was about to be assailed by critical theory: the feminist, post-colonial, and post-structural analyses of power. So, for example, in 2002 the editors of a volume of essays on genre (Coe, Lingard, and Teslenko 2002), like Devitt, recommend genre for convenience, but now the conveniently transferable “formal structures” embody politically significant “attitudes, motives, actions” (Coe, Lingard, and Teslenko 2002: 5), like carriers of nutrients – or infections. Coe, Lingard, and Teslenko are as eloquent as any of the genre theorists in proclaiming how genres arise and evolve interactively in situations, but their concerns in fact again focus on form, this time not for the privilege of choice but for the risk of indoctrination: how long, they ask, can writers in institutional settings comply with “formal structures” before they identify with their values (Coe, Lingard, and Teslenko 2002: 4)? While form continues to haunt the discussion, neither the editors of this volume nor Devitt before them, nor Miller, actually present any formal structures for inspection. These areas of discourse studies took form seriously not for being an object of inquiry to be measured, recorded, or even defined, but instead for its capacity to install in language users efficient templates or, more ominously, the interests of the privileged classes. Each of these phases in the rhetorical re-conception of genre has been a struggle with sameness: how to interpret formal regularities. Are formal regularities to be known for predictability and therefore the trivial, supposedly uninteresting, even regrettably non-creative aspect of utterance? Or should we think of them not in terms of predictability but in terms of expectations and appreciate formal regularities as ways language users know one another? Is formal regularity the basis for the analyst’s prediction or the language user’s expectations? Are formal regularities mechanical compliance or rational efficiency? Are they indoctrinating infestations or vectors of cultural continuity? To be sure, these questions are not ones which immediately arise for languagevariation specialists. Especially for students of language variation who aspire to L2 pedagogical effectiveness, veins of sameness across instances are a boon – a windfall for the classroom and not a troubling reiteration. Yet the circumstances of this controversy have parallels in language study. Just as rhetorical theories of genre have taken form for granted – as if we know what it is, and we know intuitively which forms bespeak which genres – students of language variation take for granted the extant genres, as if we know the social actions in which the formal aggregations participate. Of rhetorical theorists we might ask, how has it been possible to go ahead without facing form? Possibly, amongst classical rhetoricians, form was silently understood. But of later theorists we might simply ask, why not study form? One answer may be that the disavowal of form, or at least its relegation, has resulted

Genre as difference: The sociality of linguistic variation

41

in powerful analyses based in situation and sociality: action rather than simply function. And then, the situational analysis, beginning in classical rhetorical studies and travelling to other areas of discourse studies, has been so powerful as to produce research of great significance, especially in the area of professional communities (social workers, central-bank policy analysts, academic researchers, engineers, doctors and medical students), all without methodical attention to form. So vigorous and explanatory has been the social action principle that one important researcher has proposed that form may play no part at all in some genres. Form is finally chased off by Medway’s (2002) study of student-architects’ sketchbooks – a genre apparently entirely diverse in instances: in “physical format” (coil or no coil, from 8 × 11 to 6 × 4 , black, green, maroon. . . ) (Medway 2002: 128); in “internal organization”; in “‘inscriptional’ [. . . ] semiotic modes” – proportions of and relations between writing and drawing differed widely; in “spatial and graphic form” – from lots of white space to no space (Medway 2002: 129); and from no sentences to “solid prose” (Medway 2002: 129–130). Yet people recognise the sketchbooks as such. Medway estimates that definitions of genre have been in the hands of those studying highly institutionalised genres: As a result genre has been defined as reusable “solutions,” “plug-ins,” “templates,” and “ossifications” (Medway 2002: 125, 135). Medway rescues the definition of genre from these kinds of studies and proposes that “when people do roughly similar sorts of textual things in circumstances perceived as roughly similar, then we are in the presence of a social fact – and let’s call it a genre” (Medway 2002: 141), and he thereby forfeits all criteria relating to “textual regularity” (Medway 2002: 142). This is a brave proposal, and one way of getting rid of the problem of sameness, but not perhaps of making peace with form. Rhetorical genre theory is important for having directed attention to horizons of social experience, and the disavowal of form may have been necessary to opening this spectacle. But form has also haunted genre theory, seeking its day. Similarly, studies of linguistic variation, by default disavowing genre, are haunted by the interchangeability of register and genre, by the skirmishes between text type and genre, and by the eternal return of genre itself. But it is possible to arrange for genre to meet form in the light of day, without also returning to the impasse over sameness in rhetorical studies, or having to strike a bargain with convention, or having to leave genre only with the housekeeping duties it is assigned in the study of language variation. We can think of genre as the site of differences rather than sameness: the place where function differentiates form, that is, produces language variation – but only when function is latched to action situated in a social scene. There is

42

Janet Giltrow

precedent for doing this, from a theorist who, from the mid-stages of new genre theory to the present, has played an influential role. Mikhail Bakhtin was very interested in genre, seeing genres as the engines of heteroglossia, the responsiveness of language to groups’ activities and coalitions of interest. He mentions genres first amongst centrifugal forces, the ones which pull language away from the standardising centre towards the variety and variation to be found in local scenes of shared interest and activity:3 “This stratification is accomplished first of all by the specific organisms called genres. Certain features of language (lexicological, semantic, syntactic) will knit together with the intentional aim, and with the overall accentual system inherent in one or another genre [. . . ]” (Bakhtin 1981: 288). Unfortunately, Bakhtin is more likely to list the profusion of everyday genres4 and proclaim their stratifying, diversifying effects than actually to demonstrate these effects. Still, the spirit of Bakhtin’s claim about the knitting of features of language is hospitable to today’s corpus-based projects in language-variation studies. At the same time, his perspective on genre is hospitable to the situationbased, rhetorical theory of genre. In these capacities, genre works as a multiplier New-rhetorical studies FORM

Form

Function SOCIAL ACTION Purpose Context

Situation Motive

Figure 2. Emphases in syntactic-variation and new-rhetorical studies of genre

3. As usual, very sensitive to the prospect of sameness, genre theorists have read Bakhtin as attributing genre to the centripetal forces which produce the official, unitary language of the centre – the forces, that is, roughly, of standardisation. Genre theorists have taken Bakhtin’s concept of genre as centralising when in fact he says the opposite. 4. Following the quotation above, a list of genres unrolls: “oratorical, publicistic, newspaper and journalistic genres, the genres of literature (penny dreadfuls, for instance), or finally, the various genres of high literature” (Bakhtin 1981: 288–289). Similar lists continue to roll out in The Problem of Speech Genres: for example, “chronicles, contracts, texts of laws, clerical or other documents, various literary scientific, and commentarial genres, official and personal letters, rejoinders in everyday discourse [. . . ]” (Bakhtin 1986: 62). Oliver (2005) notes that Bakhtin’s lists are a “grab-bag,” and taunt attempts at taxonomy.

Genre as difference: The sociality of linguistic variation

43

of difference, situation impinging on form to perform local functions indigenous to a social scene. A brief look at the function of deontic modality in the research genres will show this differentiation. Deontic modality, as we will see below, is by functional definition counter-indicated for the research genres, but inviting attention nevertheless for its unpredicted appearance in these genres.

4. Demonstrating difference: Deontic modality in research genres To demonstrate genre as the engine of variation, we have to confront function, and the level at which it is named. According to general definitions, should, must, ought to, have to involve social authority – this is the function of the form deontic modality: – obligation and necessity (Coates 1983), – what is obligatory, permitted, or forbidden (Palmer [1979] 1990), – evaluation of the moral acceptability, desirability or necessity of a state of affairs; involving notions of “avoidance,” “permission,” “obligation” (Nuyts 2001: 25), – a prescriptive use of language, “recommendation” (Vihla 1999: 62), involved in obligations and regulations (Vihla 1999: 23); associated with “norm authorities” (Vihla 1999: 31). According to these definitions of its function, the form deontic modality is counter-indicated for the research genres, academic writing being known for the stance style “facelessness” (Biber and Finegan 1989: 103): in other words, characterised by a relative absence of “lexical and grammatical expression of attitudes, feelings, judgements, or commitments concerning the propositional content of a message.” Others describe the “neutrality” of “technical texts,” which “resonate with the history of a scientific tradition in society. They are expected to be ‘objective’ and interpersonally ‘neutral,’ realizing semantics of validity rather than morality” (Fuller 1998: 48). Accordingly, editors of a volume on modality (Gotti and Dossena 2001) predict that, in fulfilling the expectation of neutrality, writers of “specialist” texts (like the research genres) “often choose not to use the type of modality commonly employed to place somebody under an obligation – that is, deontic modality – for this would produce the opposite effect. Their mastery, instead, is shown by the adoption of a more neutral tone and the use of less subject-oriented modality such as dynamic or epistemic” (Gotti and Dossena 2001: 14). So far, function is named at a high level of generality and the purpose of the relevant genres is named only intuitively or by common sense: almost anyone

44

Janet Giltrow

could say that technical, scientific, or research genres are “neutral” sounding. According to these criteria for function and genre, deontic modality is not predicted to appear in the research genres. Yet in fact deontic modals do appear in the research genres. So the question then becomes, what factors in this scene of activity constrain their appearance, or what factors both call for deontic modality and particularise its use? In a study (Giltrow 2005) of 105 articles5 in three disciplines (Social Psychology, Urban Geography, and Forestry) I found two types of deontic modals. The first type appeared in contexts where the making of knowledge itself was represented: the methods, motives, directions of research, and researchers. I call these modals “knowledge deontics.” The second type appeared in contexts representing actions by people who are not researchers and not involved in knowledge making: people not, figuratively or literally speaking, in the laboratory but in the field. I call these modals “field deontics.” These are custom terms: that is, drawing on high-level definitions of function but stipulated at a low level of generality, and possibly inapplicable to any other genres. 4.1. Knowledge deontics Knowledge deontics tend to appear towards the beginning and end of articles in these disciplines. At the end – a position favoured more by Social Psychology than by the other disciplines – the deontic calls for more research in light of the present findings. For example, (1) Future research should examine these possible contingencies and assess the effects of these factors more systematically. Specifically, the accountability conditions [. . . ] should be more thoroughly investigated [. . . ]. (Social Psychology, quoted in Giltrow 2005: 181)

The broad functional definition of the form might analyse this occurrence as risking a Face Threatening Act (FTA), as being personal and committed rather than impersonal and neutral, but it seems to be, instead, not a literal dictate, but a way of emphasising or qualifying the present findings: a highly specialised function for this form – “knitting,” in Bakhtin’s sense, a feature with the locality of research activity. More common in my sample than end-positioned deontic modals were those positioned at the beginnings of articles, where they represent the knowledge project as dictated by the responsibilities of the discipline and 5. In each of the three disciplines six premier journals were identified through “snowball” citation analysis and through consultation with discipline specialists. From each of the six journals in each of the three disciplines, three articles were selected from 1998: the first article in the first issue, the third in the second, and so on. Articles were also selected from 1999 issues of the journals: the sixth in the first issue, the fourth in the second, and so on.

Genre as difference: The sociality of linguistic variation

45

impose obligations on others to do what, in fact, the reported research does do. For example, in (2) We must develop a more sensitised appreciation of the role of sound in making and remaking space [. . . ]. (Urban Geography, quoted in Giltrow 2005: 180)

that role of sound is the topic of the article. We might wonder, if deontics are about social obligation and high moral ground, why they are useful in the neutral and impersonal research genres. The answer may be that the general function defined for this form is, in actuality, differentiated by the push of the genre: the push of the situation’s roles, interests, circumstances, and identifications. We can analyse deontics as not imposing obligation or enforcing moral judgement but as enacting a disciplinary consensus, introducing or confirming research motives and the highly specialised orientations which distinguish the disciplines both from the non-research world and from one another. The sound/space approach may be new and about to be taken up, but most knowledge deontics prescribe actions already agreed upon. They are not really things researchers have to be told to do. So the Urban Geography sample turned up similarity in prescriptions for research: In the sample of 33 articles, seven obligate researchers to watch for the complicities of race and place. Two are nearly identical: (3) Chinese immigrants in LA vs. NYC race should be understood as socially constructed [. . . ]. (4) Male youth unemployment racial difference should be understood as socially constructed [. . . ]. (Urban Geography) (both quoted in Giltrow 2005: 181)

Others repeat similar obligations: for example, we must consider the geographically contingent nature of prejudice and discrimination, and we must unnaturalize landscapes in terms of gender and race. The broad form/function definition of deontic modality would tell us that people are being placed under obligation – imposed on – to do something they might not otherwise do. But in fact there appears to be agreement on what ought to be done. Modal-deontic expressions may be performing functions which derive but also depart from directive ones. They may consolidate research interests and emphasise the researcher’s share in these interests: a response to a situation in which activity is disciplined by a research paradigm. 4.2. Field deontics Field deontics formally impose obligations on actors who are not members of the research community, as in the following examples:

46

Janet Giltrow (5) [. . . ] cooperation must be maintained [. . . ]. (Forestry) (6) [. . . ] fire managers should start the day with [. . . ]. (Forestry) (7) [. . . ] intervention strategies should focus on old and poor neighbourhoods and minority neighbourhoods [. . . ] intervention strategies should address housing quality and maintenance [. . . ]. (Urban Geography) (8) [. . . ] selection interchange sites should take account of Arab potential; [. . . ] the areas under consideration should vary in distance [. . . ]. (Urban Geography [+ 11 more obligations]) (all quoted in Giltrow 2005: 184–185)

Strict conditions both limit and, by limiting, activate the expression of obligation. First, field obligations are a special kind of prescription, one that finds a responsibility to act on research findings rather than on some other basis, such as good will or propriety: We have found out x, so y should be done. The new knowledge ratifies the dictate, but more importantly, the dictate shows that the knowledge presented is attested by research: the only basis for prescription. Second, the actors who are obligated make up a specified population: urban planners or forest managers, for example. But whether the field, consisting of practitioners, professions, and institutions, ever hears the dictates or not, the dictates signal the quality of the research findings by indicating that they can be acted on. And the very containment of the prescription, limited to a field of designated professionals, corroborates the status of the findings. Both conditions – the restriction to action based on research-attested knowledge and the containment within a professional field – not only reconcile deontic modality with the neutrality of the research genres but also make deontic modality a testament to that neutrality. These constraints are not what we would predict from broad definitions of function applied at a high level of generality, such as academic writing. In these genres, under these circumstances, deontic modality’s function converts to the service of the neutral authority of science. The form will function differently in financial advice columns, and differently again in parliamentary speeches, or in judicial reasons. Beyond these patterns, the function is further differentiated amongst these three disciplines:At a still lower level of generality, the form functions differently in Forestry, Social Psychology, and Urban Geography, according to the social and institutional affiliations of these disciplines. Forestry, for example, projects itself onto a field densely regulated by both official policy and optimum timber yield, and its pattern of prescription is accordingly different from that in Social Psychology or Urban Geography. In turn, in Urban Geography, deontics are more likely than in the other disciplines to be found in reported speech, possibly attesting to that discipline’s articulation with political activism. But even the broader patterns show that once mobilised in a genre, and subject to the push of

Genre as difference: The sociality of linguistic variation

47

situation, local interest, and current coalitions, form is differentiated. By tracing the contingency of form and its function, we get close to the diversifications and stratifications of language which Bakhtin attributes to genre. But, to draw near this hotbed of variation, we need to track genre through the open set to the low level of generality and locality which rhetorical theories of genre, after much struggle, recommend. Genre is the name for these ground-level differentiations, where the sociality of, in this case, research – its internal affiliations and its external ones, its cultivation of the cultural phenomenon of methodology – press formal function into service.

5. Partnerships for genre and variation Where the research concern is indeed strictly with forms, their changes over time, or their distribution over populations, genre is a dispensable concept. When the concern is not so strictly formal, when it broaches function and text type, genre is a useful partner in the study of linguistic variation. To be a useful partner, genre must locate at a lower level of generality than may have been traditionally assumed, at a level of social activity which already attracts the attention of students of linguistic variation: on-line medical advice, the live TV exchange, the guest lecture, or the corporate blog, for example. In turn, at these levels, study of linguistic variation becomes an indispensable partner to the study of genre, coaxing out the actualities of situation. To establish this partnership, terms such as function and purpose may need to be re-tooled to contribute to the more complex and realistic notion of social action. For example, interpersonal pronouns (discussed by Puschmann, this volume) can be defined for their function as indexing communicative roles. But Puschmann’s study of the genre of the corporate blog (a low-level classification, lower even than blog) shows that in practice the form is pushed by situation to more complex performance. This performance is in turn contained in its range, for some combinations hint at other genres, such as advertisement. Puschmann’s study suggests how analysis of a form and its variation, inter- and intra-generically, can be interrogated for its functional reach, in this case designing the positions of corporate self-representation. Other forms may be equally eligible for interrogation of situated function, such as those forms contributing to syntactic complexity (Pérez-Guerra and Mart´ınez-Insua, this volume), or those associated with narrative or description (Virtanen, this volume). Inquiries into the local action of forms may expand and enrich the notion of function. So, too, may historical portraits of genres at their emergence also elaborate our sense of what counts as “purpose.” Gotti’s account (this volume) can tell us that

48

Janet Giltrow

there is opportunity in heeding Askehave and Swales’ (2001) cautions against complacency around communicative purpose. His outline of the social scene from which the experimental article emerged shows that, while it is sensible to describe the purpose of this emergent genre as being to disseminate information widely and promptly, it is also not enough. The forms called on to execute this purpose also ramify in function: They are swept into a recursive series of motives, an interlocking sociality which attends the forms in, as Gotti (this volume) says, quoting Valle (1999: 111), “concentric” circles of reception. At the inner circle of reception, the forms consolidate and promote a community of practice, functioning also in the outer circles as measures of character which articulate with the stratifications of the social order. These effects could be captured neither by the agent-effacement function of the passive, for example, nor by the explicit purpose of disseminating information. Yet neither could the linguistic dynamism of this social scene be captured without methodologically precise attention to form. Possibly the high levels of classification, the levels deserted by genre, are most productively occupied by inquiry which, by means of relatively abstract categories, establishes the forms which are worth pursuing. In their broad or narrow distributions, their development or decline, these forms will be the ones with a guarantee of meaning, searchable and potential at the ground-level of genre.

References Askehave, Inger and John M. Swales 2001 Genre identification and communicative purpose: a problem and a possible solution. Applied Linguistics 22: 195–212. Bakhtin, Mikhail M. 1981 The Dialogic Imagination, trans. Caryl Emerson and Michael Holquist. (University of Texas Press Slavic Series 1.) Austin: University of Texas Press. Bakhtin, Mikhail M. 1986 The problem of speech genres. In: Caryl Emerson and Michael Holquist (eds.), Speech Genres and Other Late Essays, 60–102.Austin: University of Texas Press. Biber, Douglas, Eniko Csomay and James K. Jones 2004 Vocabulary-based discourse units in university registers. In: Alan Partington, John Morley and Louann Haarman (eds.), Corpora and Discourse, 23–40 (Linguistic Insights 9.) Bern: Peter Lang. Biber, Douglas and Edward Finegan 1988 Adverbial stance styles in English. Discourse Processes 11: 1–34.

Genre as difference: The sociality of linguistic variation

49

Biber, Douglas and Edward Finegan 1989 Styles of stance in English: Lexical and grammatical marking of evidentiality and affect. Text 9: 93–124. Biber, Douglas 2006 University Language: A Corpus-based Study of Spoken and Written Register. (Studies in Corpus Linguistics 23.) Amsterdam/Philadelphia: John Benjamins. Bitzer, Lloyd F. [1968] 1980 Functional communication: a situational perspective. In: E. E. White (ed.), Rhetoric in Transition: Studies in the Nature and Uses of Rhetoric, 21–38. University Park, PA: Pennsylvania State University Press. Black, Edwin [1965] 1978 Rhetorical Criticism: A Study in Method. Madison: University of Wisconsin Press. Bruti, Silvia 2004 Paraphrase types in the Pavia biology corpus: Some appositional constructions. In:Alan Partington, John Morley and Louann Haarman (eds.), Corpora and Discourse, 125–138. Bern: Peter Lang. Burke, Kenneth 1954 Counter-statement. Los Altos, California: Hermes. Burke, Kenneth 1966 Terministic Screens. In: Language as Symbolic Action, 44–62. Berkeley: University of California Press. Coates, Jennifer 1983 The Semantics of the Modal Auxiliary. London: Croom Helm. Coe, Richard, Lorelei Lingard and Tatiana Teslenko (eds.) 2002 The Rhetoric and Ideology of Genre. (Research and Teaching in Rhetoric and Composition.) Cresskill, NJ: Hampton. Cohen, Ralph 1986 History and genre. New Literary History 17: 203–219. Devitt, Amy J. 2009 Re-fusing form in genre study. In: Janet Giltrow and Dieter Stein (eds.), Genres in the Internet: Issues in the Theory of Genre, 27–48. (Pragmatics & Beyond New Series 188.) Amsterdam/Philadelphia: John Benjamins. Devitt, Amy J. 1993 Generalizing about genre: New conceptions of an old concept. College Composition and Communication 44: 573–586. Dimock, Wai Chee 2007 Genres as fields of knowledge. Periodical of the Modern Languages Association 122: 1377–1388. Fuller, Gillian 1998 Cultivating science: Negotiating discourse in the popular tests of Stephen Jay Gould. In: J. R. Martin and Robert Veel (eds.), Reading Science:

50

Janet Giltrow Critical and Functional Perspectives on Discourses of Science, 35–62. London/New York: Routledge.

Giltrow, Janet 2005 Modern conscience: Modalities of obligation in research genres. Text 25: 171–200. Gotti, Maurizio and Marina Dossena 2001 Introduction. In: Maurizio Gotti and Marina Dossena (eds.), Modality in Specialized Texts, 9–20. (Linguistic Insights 1.) Bern: Peter Lang. Gotti, Maurizio this volume A new genre for a specialized community: The rise of the experimental essay. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 85–110. (Topics in English Linguistics 70.) Berlin/NewYork: Mouton de Gruyter. Gronbeck, Bruce 1978 Celluloid rhetoric: on genres of documentary. In: Karlyn K. Campbell and Kathleen H. Jamieson (eds.), Form and Genre, 139–161. Falls Church, VA: The Speech Communication Association. Hart, Roderick P. 1971 The rhetoric of the true believer. Speech Monographs 38: 249–261. Jamieson, Kathleen M. Hall 1973 Genre constraints and the rhetorical situation. Philosophy and Rhetoric 6: 162–170. Medway, Peter 2002 Fuzzy genres and community identities: The case of architecture students’sketchbooks. In: Richard Coe, Lorelei Lingard and Tatiana Teslenko (eds.), The Rhetoric and Ideology of Genre, 123–154. (Research and Teaching in Rhetoric and Composition.) Cresskill, NJ: Hampton Press. Miller, Carolyn R. 1984 Genre as social action. Quarterly Journal of Speech 70: 151–167. Nuyts, Jan. 2001 Epistemic Modality, Language, and Conceptualization: A CognitivePragmatic Perspective. (Human Cognitive Processing 5.) Amsterdam/ Philadelphia: John Benjamins. Oliver, Rob 2005 Interdisciplinary Bakhtin: Revisiting the notions of ‘speech genre’ and ‘voice’. Paper presented at Conference on College Composition and Communication, March 16–19, 2005. San Francisco, USA. Palmer, Frank Robert [1979] 1990 Modality and the English Modals. (Longman Linguistics Library 23.) London/New York: Longman. Partington, Alan, John Morley and Louann Haarman (eds.) 2004 Corpora and Discourse. (Linguistic Insights 9.) Bern: Peter Lang.

Genre as difference: The sociality of linguistic variation

51

P´erez-Guerra, Javier and Ana E. Mart´ınez-Insua this volume Do some genres or text types become more complex than others? In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 111–140. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Puschmann, Cornelius this volume Thank you for thinking we could: Use and function of interpersonal pronouns in corporate web logs. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 167–191. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Valle, Ellen 1999 A Collective Intelligence: The Life Sciences in the Royal Society as a Scientific Discourse Community, 1665–1965. (Anglicana Turkuensia 17.) Turku: University of Turku. Vatz, Richard E. 1973 The myth of the rhetorical situation. Philosophy and Rhetoric 3: 154– 161. Vihla, Minna 1999 Medical Writing: Modality in Focus. (Language and Computers 29.) Amsterdam/Atlanta, GA: Rodopi. Virtanen, Tuija this volume Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 53–84. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Wilkerson, Kenneth Errol 1970 On evaluating theories of rhetoric. Philosophy and Rhetoric 3: 82–96.

Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre Tuija Virtanen

The concern of this chapter is with two different kinds of categorizations of text and discourse: text/discourse types and genres. In the study of text/discourse types, the emphasis has traditionally been on the first member of the reflexive text-context pair of notions. Investigations of genres, again, have tended to shift the focus increasingly onto its second member. It is shown that there is added value in taking into account both genre dynamics and text/discourse types, each in their own right, in studies of variation across texts and discourses. A two-level model separating text types from discourse types is presented, to account for apparent mismatches between the two. Direct and indirect uses of text types are analyzed in texts representing one and the same genre. The syntactic signals under attention include clause-initial adverbials, tense, and referential choices in subject position. Further, genres as social action are approached from the perspectives of emergence and change that affect their linguistic and contextual characteristics. Both text/discourse types and genres invite analyses that focus on the nexus of the linguistic, social, and cognitive dimensions of discourse; yet they disclose very different aspects of the forms and functions that are emergent in and across texts and contexts.

1. Introduction Linguistic variation contributes to the emergence of categories of text and discourse, ranging from text/discourse types and styles to genres and registers. Based on intertextuality, such categories help interlocutors – speakers and hearers, writers and readers, signers, onlookers, lurkers and other potential participants – to (co-)construct, mediate, maintain, or alter contexts and cultures through discourse practices. Variation across texts and discourses has been discussed in many different fields of language study and the focus has accordingly shifted between text and context. The linguistic signals of the various catego-

54

Tuija Virtanen

rizations of texts and discourses include a wide variety of syntactic, lexical, and phonological phenomena. Visuals, too, can be indicative of such categorizations, as shown, for instance, by differences in the shape of various groupings of texts. The concern of the present volume is with syntactic variation across texts and discourses, in particular as it manifests in emergent genres. The constitution and change of prototypical categories at the text-context interface has been investigated diachronically in historical linguistics, and recently also in environments such as the internet, which invites linguists to witness ongoing language change. This chapter deals with two different kinds of categorizations: text/discourse types and genres. In the study of text/discourse types, the emphasis has traditionally been on the first member of the text-context pair of notions; investigations of genres, again, have tended to shift the focus increasingly onto the second member of this reflexive pair. While text types are thus usually explored using text-internal criteria, genres are characterized in terms of text-external criteria, in terms of the two-way traffic between texts and their situational and sociocultural contexts. Text/discourse types are grounded in cognitively based, goal-oriented, formfunction relationships, which evolve through recursive and reiterative strategies and intertextual practices across contexts, thus developing into heuristics that facilitate discourse production and interpretation (see, e.g., Adam 2005; de Beaugrande and Dressler 1981; Kinneavy 1980; Longacre 1983; Pilegaard and Frandsen 1996; Smith 2003; Virtanen 1992a, 1992b; Werlich 1976). The prototypical text type is the narrative. The number of identified types is usually limited to a small set. As will be argued below, for analytical purposes it is expedient to distinguish two matching levels of types, one for form and the other for function. Genres are generally approached from the point of view of social action, as classes of communicative events having shared communicative goals and cognitive schemata (see, e.g., Bakhtin 1986; Bhatia 1993; Briggs and Bauman 1992; Giltrow, this volume; G¨unthner, this volume; Halliday and Hasan 1985; Hanks 1987; Hymes 1972; Miller 1984; Swales 1990). Examples of what people, and analysts of different orientations, recognise as genres seem to vary maximally: Non-fiction genres might, for instance, include news stories, editorials, a range of different kinds of workplace documents and reports, legal contracts, sermons, lectures of various kinds, reality shows, weather forecasts, and so forth, but also everyday speech genres such as service encounters, gossip, or complaints and reproaches. Central to the present understanding of genre is its symbiosis with linguistic and cultural communities of various kinds, ranging from speech communities for public genres, to (professional) discourse communities and communities of practice for semi-public and private genres.

Variation across texts

55

Text/discourse types and genres understandably tend to feature in different studies. Taken together, however, the study of both kinds of categorizations allows linguists to explore different dimensions of the interplay between cognitive and social concerns and constructs, as manifest in discourse through aggregates of concrete linguistic elements. An integrated view will, however, have to take into account the fact that text types cut across genres in interesting ways. This may go unnoticed if we opt for an appropriation of text/discourse types into models of genre. Other concepts used for linguistic variation across texts and discourses include registers and styles, which will not be further elaborated on in the present chapter. Registers are related to variables of the communication situation, at various levels of abstraction, usually as a notion to be kept apart from genre but sometimes used interchangeably with it. The term may refer to broad fields of language use, such as the language of religion or law. (For discussions of register, see, e.g., Biber 2006; Crystal and Davy 1969; Dittmar 1995; Eggins 1994; Halliday 1978). Style, again, primarily refers to the characteristics of particular texts. Style always involves an element of choice. An expedient use of the term in the present context would be to view styles as the outcome of variation within text types and genres, or more generally, as the outcome of the omnipresent variability of all language use (see, e.g., the discussion in Enkvist 1985b; cf. also Virtanen 2008). Linguists sometimes reduce style to a simplified generic distinction between relatively formal or informal language, or they use it as a convenient label for residual variation. Yet, there is often a real need for an additional concept to come to grips with intracategorical variation; the notion of style might be a good candidate for systematizing such observations. In what follows, the discussion starts with text and discourse types in section 2. Section 3 focuses on linearization as a cue to text type. Chains of textstrategic signals appearing as the starting points of new textual units are examined in different types of text that represent one and the same conventionalized genre, the encyclopedic article. The syntactic devices under study include adverbial placement, tense, and referential choice manifest in the subject. The second part of the chapter deals with genre dynamics (section 4), a concern shared by the majority of the contributions to the present volume. We will see that there are good reasons for maintaining these two different kinds of categorizations as separate in our attempts to understand syntactic (and other kinds of linguistic) variation across text and discourses (section 5).

56

Tuija Virtanen

2. Text and discourse types Repertoires of text and discourse types found in the relevant literature tend to include narrative and a small number of non-narrative types. After the opening discussion of the narrative and argumentative types of text in 2.1, 2.2 presents a two-level typological model that separates form from function. 2.1. The narrative and argumentative types of text There is a basic distinction between narrative and non-narrative texts, first, in terms of whether or not the dominant text strategy is one of contingent temporal succession (often including some element of causality), and second, in terms of whether the text is built around (a group of) participants. Linguistic exponents of the narrative type of text have been widely discussed in the literature, and at times discourse has been explored through narrative alone without any particular note being made of such an orientation. Narrative is, indeed, a basic type of text: It can be used to (co-)construct discourses of many different kinds as it can serve a maximal range of discourse functions, as different as, for instance, entertaining, instructing, describing, explaining, and judging various experienced or imagined phenomena. Narrative is a very strong type of text in the sense that few linguistic signals are needed for a text to be interpreted as such. But this also makes narrative into a very conspicuous type of text: It tends to have a high degree of figureness in relation to the co-text – and the same applies to its situational and socio-cultural context. People tend to recognise a narrative as distinct from a non-narrative text. The basic character of narrative among various repertoires of text types would thus rather seem to leave a very backgrounded status to the non-narrative. Yet, at the other end of the cline from temporally sequential, participant-oriented texts to those that manifest neither contingent temporal succession nor participant orientation, we find the equally important text type of overt argumentation. Here, however, the opposite is true. Overtly argumentative texts do not constitute a basic type of text: the range of discourse functions that they can serve is limited to judging, to expression of beliefs and attitudes in the construction of an argument. In contrast to narrative, the overtly argumentative type of text will need a great deal of linguistic signalling, one of a very different kind from narrative. It is therefore, in this explicit form, a relatively weak contribution to the communicative goal of persuading others to share one’s viewpoint or change their own – it might indeed backfire, in a way an explicit form of narrative would not. Argumentation is much more effective if left implicit, and it therefore readily appears in the guise of another type of text (see 2.2, below). The

Variation across texts

57

overtly argumentative type of text is essentially structured according to various thought patterns constituting an implicit dialogue with the audience. The ratio of syntactic and lexical signals pertaining to text type is likely to differ between texts categorized as narrative or argumentative, as is the use of metatextual signals. At this point, a note on persuasion is in order. All language use is in a sense persuasive. It is therefore expedient to make a distinction between the omnipresent persuasion in human communication (including efforts to silence it) and persuasion in a particular type of text, argumentation, which is explicitly structured to make an overt argument. Similarly, it is important to keep apart persuasion and the argumentative discourse type of functional character. We can speak of such a discourse type in the instances where people are engaged in communicative activities that primarily involve judging, rather than, say, instructing others, entertaining them, or explaining to them the essence of a particular phenomenon, irrespective of whether such discourse functions are made explicit in the text. While not the only ones possible, all of these discourse functions are inherently persuasive in the sense that people are interested in making others see things from their point of view, or at least understand what the viewpoint is, and hence regard this point of view as a possible way of approaching what they are talking about. This is, indeed, the common ground needed for the fundamental pragmatic activities of (co-)construction, negotiation, and adaptation in which interlocutors engage through discourse, in interconnected but always unique communicative situations. 2.2. A two-level typology: Text types and discourse types To come to grips with the apparent mismatches between text types and discourse types we need a two-level typological model (Virtanen 1992a, 1992b). Since discourse is generally used as an overarching notion (with varying scope) and text constitutes a central part of it, I have chosen to label the two parallel levels accordingly. Discourse types are closely connected with discourse functions; text types, again, clearly relate to the form that these categories will take through aggregates of linguistic exponents of the particular text strategies that are associated with them. Such a theoretical distinction has the advantage of approaching the use into which text types are put, and through which they evolve, in terms of whether or not there is a direct link between the two levels. In what follows, I will speak of direct (primary) and indirect (secondary) uses of text types. In direct use, the corresponding text type and discourse type agree; in indirect use, there is a mismatch between the two.

58

Tuija Virtanen

Hence, the narrative type of text can realize any discourse type with ease while the overtly argumentative type of text can only function as a realization of the corresponding argumentative type of discourse. Inversely, the narrative discourse type matches with some form of the narrative text type exclusively while the argumentative type of discourse can be realized through practically any type of text. Narrative is thus a strong text type that can be labeled basic in terms of the ease that it manifests in being able to agree with a maximal range of different discourse types. At the functional level it is the very strong argumentative discourse type that can be granted the status of the basic type of discourse on similar but reversed grounds: the argumentative type of discourse can take the form of any type of text. Dorgeloh (1997: 143–144) relates this difference to metafunctional motivations at the discourse type level: The narrative type of discourse relies primarily on the ideational metafunction of language while the interpersonal metafunction is the predominant one in the argumentative type of discourse (for metafunctions of language, see Halliday 1978). Text types are exhibited in texts through aggregates of linguistic exponents. We can speak of text strategies to refer to the micro and macro level ways of constructing coherent texts (see Enkvist 1985a, 1987; Virtanen 1992a, 2004). In their idealized, prototypical form, individual text types may be related to particular text strategies. But investigations of systematic linguistic variation across texts will disclose complex patterns of co-occurrences and combinations of different text strategies. Text strategies are the focus in section 3, below, where the discussion of their syntactic signals includes clause-initial adverbials, tense, and referential choice between full noun phrase, pronominal, and zero subjects. In contrast to the functional level of discourse types, text types are thus relatively closely associated with the form of a text. But texts are seldom unitype; text types usually appear in embedded and hybridized forms, resulting in multitype texts which may, but need not, show dominances of particular types. Yet, the existence of a relatively loose set of shared prototypical categories of texts facilitates discourse processing by serving as cognitive heuristics for interlocutors. Such abstract categories provide people with templates of various kinds that can be recycled in forms that are appropriate in particular situational and socio-cultural contexts. Matching texts with shared prototypes allows people to detect and interpret deviations from them and set out to (co-)construct approximations to them. Renewal comes from combinations of convention and innovation. Text types are thus dynamic categories that evolve and are maintained or altered through the use of language. At a very high level of abstraction, at least some of them are, to an extent, shared across speech communities, discourse communities, and communities of practice. But the latter two may, for instance, signal group identity through preference for, or repulsion

Variation across texts

59

of, particular types of text, as used in the genres that they own, cherish, and develop. The potential of particular text and discourse types to be used indirectly was discussed above in terms of the two endpoints on a scale reaching from narrative to argumentation. Typologists work with small sets of categories. The repertoires of types vary so that some, for instance, include description or instruction while others do not. Expository texts, however, appear in many typologies in one form or another. What is shared is the presence of narrative. In corpus-based studies of linguistic variation (see, e.g., Biber 1988) the distinction between narrative and non-narrative is conspicuously present, in contrast to overtly argumentative texts, which can be expected to enjoy less popularity as the argumentative discourse type is, for persuasive ends, typically realized indirectly. The instructive or procedural type of text manifests (inherent) temporal text structuring and thus comes close to the narrative end of the continuum. Near the argumentative end, again, there is some form of expository type of text. Expository and argumentative texts are organized in accordance with patterns of thought processes such as additions, elaborations, contrasts, concessions, and so forth – each according to its own prototypical combination. Both instructive and expository texts realize fewer types of discourse than the narrative type of text, but unlike overt argumentation, they are clearly capable of functioning in the service of discourse types other than the corresponding one. The interplay between discourse types and text types constitutes a dynamic interface, an area where renewal can take place. Kroon (2002) studies Latin ghost story letters in terms of the two levels. Herring, van Reenen, and Schøsler (2001) discuss the use of such a model in historical linguistics. Esser (2006) posits decisions concerning the fit between the two levels of types as a major receiver-oriented stylistic choice resulting in variance of presentation in speech and writing. In the process of being used indirectly, text types are assigned characteristics that serve to index the relevant discourse type and discourse function. Hence, examining the changing patterns of use of the sentence-initial and in written discourse, Dorgeloh (2004: 1778) found its functional potential over time to be connected to the “de-narrativization” of academic genres. Dorgeloh (2005) shows how early scientific writing increasingly moves away from indirect (secondary) narrative towards an impersonalized argument, changes that affect the linguistic expression of agentivity. Gotti’s discussion, in this volume, of the conscious development of the early experimental essay as different from contemporary scholarly genres includes the change in discourse type from argumentative to what is identified as “informative.” There is a “conspicuous interval” (Gotti, this volume: 102) in such texts between the report of the experimental findings and the author’s reflections on them,

60

Tuija Virtanen

the two parts thus manifesting different types of text in direct and indirect use. The status of the descriptive type of text would at first sight seem to be very different from the other types of text singled out in the literature. It is very versatile in adopting various guises in accordance with a dominant type of text in multitype texts, and examples of independent texts that would be descriptive to the exclusion of other types are difficult to find. In this light, the discourse function of merely describing observable or imaginative phenomena appears to be of minor interest to people.Yet, description is a common metadiscursive label in speech and writing for text and talk that deals with characteristics of people, their activities, objects, pictures, places, scenery, and so forth. And descriptive texts constitute a distinct category as they share linguistic characteristics that are unlike those of the other prototypical types of text. Still, object descriptions of, for instance, apartments or properties for sale and technical product descriptions in commercial contexts are embedded in, or indeed, manifestations of, the argumentative type of discourse. In scholarship, again, there is a striving for explanation and argumentation, rather than mere description, even though informative descriptions are an essential part of it all. Indirect use of the descriptive type of text in the service of the instructive type of discourse is prevalent in travel guides, motivated by textual reasons (see, e.g., Virtanen 1992a). In such contexts, entire texts can often be classified as descriptive but their function is clearly instructive or procedural. Ummelen (1997) investigates the effects of instructive and descriptive types of text in the service of the instructive type of discourse in software manuals. Merlini Barbaresi (2004) examines embedded descriptions of environments in various genres of fiction and non-fiction. Linguists working exclusively with narrative are concerned with the internal variation of the text type across contexts: From this perspective, descriptions, and perhaps any other non-narrative types of text, are indeed an inherent and integral part of narrative discourse (see, e.g., Shiro 2003). Linguists working with expository or argumentative texts similarly see descriptions as subordinate to these dominant types; here, too, the descriptive text type predominantly appears in indirect use and readily adopts linguistic characteristics from the main types of text (see, e.g., Lindeberg 2004 on scholarly argument). Further, discussions of text type benefit from making a distinction between evocative texts, such as narrative, and operational texts, such as instructions of various kinds (see Enkvist 1985a). While argumentation can be likened to evocative texts and exposition to operational texts, descriptive texts are again different as they are able to serve both evocative and operational ends. Text types cut across genres in ways that motivate separating the two. Instances of one and the same text type, such as the narrative, instructive, or

Variation across texts

61

descriptive, can be found in many different genres. In the next section we will see how one and the same genre, in this case the encyclopedic article, has been given various textual forms that do not all conform to the expository type of text, which might be expected to be the default in such contexts. The concern here is with linearization and the focus is on the choice of linguistic elements to be placed at the beginning of clauses and sentences, a slot that carries a great deal of textual potential.

3. Direct and indirect uses of text types: Evidence from beginnings Keeping the genre dimension relatively constant, the examples below illustrate direct (primary) and indirect (secondary) uses of text types in encyclopedic articles from one and the same source (Myers and Copplestone 1981). Text types show attraction to particular text strategies and the focus here is accordingly on text-strategic variation, in terms of the participant- or topic-oriented text strategies, the temporal and locative text strategies, the listing and taxonomic strategies, the contrastive and concessive strategies, and many more. Text strategies consist of writers’ and interlocutors’ recursive decisions, conscious or not, concerning the (co-)construction of text and talk in particular situational and socio-cultural contexts, in view of a communicative goal. They are manifest in thematically or topically uniform text-structuring orientations, exposed through choices of concrete linguistic elements that serve to create coherence and indicate text segmentation (Enkvist 1985a, 1987; Virtanen 1992a, 1992b, 2004). The focus of this section is on linearization, in particular, on the linguistic material placed at the beginning of clauses and sentences. In textual terms, the prominent positions are beginnings and endings – but linguistic material is placed in these positions for very different reasons and the two types of placement have very different effects. We will see that the elements placed at the outset of clauses and sentences may contribute to the construction of a particular text strategy and hence also serve to signal text segmentation. Their wide textual scope allows them to mark boundaries in the text and introduce new textual units, of various sizes and at various levels of text-structural hierarchy. In such texts there is typically some degree of text-strategic silence when one and the same textual unit continues, as linguistic signals of the activated text-strategy are then positioned non-initially. Another way to distinguish between new and continuing textual units is to use initial elements of smaller size as compared to the ones functioning as text-strategic markers. In terms of text-structural iconicity, local boundaries tend to be indicated by fewer markers, markers of smaller size and/or markers conveying less new information than those signalling ma-

62

Tuija Virtanen

jor boundaries in the same text. Both major and minor boundaries can thus be signalled linguistically, and absences of such indices are related to absences of textual boundaries. (For a detailed discussion of text strategies of various kinds, see Virtanen 1992a, 1992c). Sections 3.1. and 3.2. deal with text strategies in five different articles of the encyclopedia; the first is analysed in its entirety. The syntactic signals discussed include clause-initial adverbials, tense, and referential choices in subject position. 3.1. In-depth analysis of a text Let us start with a scrutiny of an entire text that exhibits several shifts from one text type to another. The overall discourse type is expository (indicated in capital letters to the right), and the text types in its service include the expository, descriptive, and narrative (in small capitals, followed by the bidirectional arrow of the interface). (1) NAR ↔ EXP Leonardo da Vinci (1452–1519). By 1476, Leonardo was living in the house of his master Verrochio and Ø had already painted the portrait which anticipates his famous Mona Lisa by 25 years, the Ginevra di Benci, now in Washington. In 1481 he was commissioned to paint the Adoration of the Magi. The unfinished picture in the Uffizi EXP ↔ EXP reveals the extent of his originality. The central group is a tightly composed pyramid of figures from which the landscape recedes into distant atmospheric depth. It prefigures his later studies of the Virgin and Child with St Anne with its densely unified composition. The complicated poses could only be achieved after many preliminary DES ↔ EXP studies of drapery and anatomy. Leonardo was a compulsive draughtsman whose thousands of drawings, together with notebooks, record his artistic experiments and scientific explorations. NAR ↔ EXP In 1483, Leonardo offered his services to Lodovico Sforza, Duke of Milan, as painter, sculptor, military engineer and architect. He remained at the Milanese court until it fell to the French in 1499. EXP ↔ EXP For a projected equestrian monument to Francesco Sforza he made a huge clay model of a riderless horse, and in Santa Maria delle Grazie, Milan, he painted in oil his famous Last Supper (about 1497) a work of psychological subtlety unprecedented at that date. His painting of the Madonna of the Rocks, of which two versions survive in London and Paris, shows a growing mastery of sfumato, of atmospheric blending and blurring of light and shade by which he achieved in oil painting a new and subtle range of tonal values. In 1517, Leonardo removed NAR ↔ EXP to France to the Chˆateau of Cloux, where he continued until his death

Variation across texts

63

to investigate the nature of the world around him respected both for his intellectual integrity and for his artistic genius. (from: Myers and Copplestone 1981)

The text consists of three narrative passages. These are structured along the lines of a temporal and a participant-oriented text strategy. The absolute indications of time serve to shift the text from one textual unit to another as they indicate points of time in the temporally sequential flow of an artistic lifetime that forms the basis of this text strategy: The starting point, by 1476 is followed by in 1481, in 1483, and in 1517. In texts realizing the narrative discourse type, such signals would be expected to be more varied, including text-related and situationally anchored elements. The absolute adverbials in use here convey information that may be inferred to be part of the period of time given in parentheses in the heading.The indirect use of narrative is thus signalled through the expected chain of temporal markers whose absolute form and information status as inferrable or unused/new (Prince 1981) serve as cues to the expository discourse type, which is the default in this genre. The first narrative passage introduced by by 1476 contains other signals of temporality in its second clause: the flashback marked by the past perfect followed by the adverbial already, and the two foreshadowings, one marked by the lexeme anticipates in the non-past tense and the adverbial by 25 years, the other marked by the adverbial now, referring to the time of writing and possibly that of reading. Yet it is the sentence-initial signal in 1481 that conveys the textstrategically important information that the reader is here shifted to a new textual unit within the narrative. The narrative is resumed at the outset of the second paragraph and for the third time, in the last sentence of the text, which could be a paragraph of its own had it been expanded to cover more information. Each time we find the simple past referring to (a sequence of) events (was commissioned; offered; made, painted; removed). The sentence-final adverbial clause, until it fell to the French in 1499, functions as a signal of closure of the second narrative unit. The adverbial until his death in the final narrative passage is placed in the text-structurally non-conspicuous middle zone of the clause; end-weight is here given to elements that are central to the genre, fulfilling expectations raised by the expository discourse type. The verbs in these two clauses (remained; continued to investigate) refer to an action or activity that has duration until the end-point indicated by the adverbial. The article is tied together by a participant-oriented text strategy, formed through a chain of nominal, pronominal, and zero references to Leonardo da Vinci. Even though they are much more frequent in the text than indications of the temporal text strategy, these references still contribute to the signalling

64

Tuija Virtanen

of textual boundaries. The full noun phrase, Leonardo, is used at the outset of four textual units, three of which are narrative while one is descriptive (starting Leonardo was a compulsive draughtsman [. . . ]). The three major textual boundaries manifesting both a lexically weighty signal of the temporal text strategy and the full noun phrase constitute the starting points of the narrative passages. Within these passages, pronominal and zero references are used to indicate that readers should keep the main participant active in their minds and focus on what is said about him, his work, and life. There are no intervening references to other people or artistic works between the last sentence of the first paragraph and the beginning of the second paragraph, in the way that there is interference through references to Wolgemut in the main text-strategic chain referring to D¨urer in the middle of example (2), below. Yet, the shift from a descriptive passage to a new narrative one is here signalled using a full noun phrase reference to the continuous participant Leonardo. Full noun phrase references can be expected irrespectively of text-strategic needs when there is a risk of interference from other participants or a lack of adjacency between two mentions. Sentence-initial signals of the temporal text strategy are therefore the first indication of narrative, while the frequent references to (a) continuous (group of) participants can only occasionally contribute to the signalling of textual boundaries through the form in which they appear. The text also manifests expository passages, i.e. direct or primary use of the text type. These are signalled at their outset by definite full noun phrase references to the artistic work under attention in the particular textual units: for example, the unfinished picture in the Uffizi, rather than the pronominal it referring to the Adoration of the Magi introduced at the end of the preceding sentence and textual unit. This crowded noun phrase is shorthand for a great deal of information, serving the condensed style of the encyclopedia, and it allows the writer to expand the expository passage focusing on the originality of the picture. In this textual unit centred around artistic work there are several shifts between abstraction levels. Hence, the second sentence starts with another definite full noun phrase moving the reader spatially to a particular part of the picture and then zooming out from there (the central group [. . . ] from which the landscape recedes into [. . . ]; for a detailed discussion of the locative text strategy and the processes of zooming in and out through text, see Virtanen 1992a; see also the discussion in Kreyer, this volume.) The locative text strategy is, however, not adopted in this passage; instead the expository text type continues with its topic-oriented references to the picture, whereby a pronominal it is enough. The final definite full noun phrase in this expository passage, the complicated poses, refers to a characteristic of another picture introduced at the end of the preceding sentence, which is presented as a continuation of previous work. The shift in

Variation across texts

65

the abstraction level from the individual paintings and their merits to a general look-back at the repeated activities of the artist (after many preliminary studies of drapery and anatomy) may make the reader anticipate a textual boundary already before the outset of the paragraph-final descriptive passage, signalled by a full noun phrase reference to Leonardo. The second expository passage is embedded in the narrative of Leonardo in Milan. The shift is indicated through the sentence-initial adverbial, for a projected equestrian monument to F.S., a conspicuous signal after which no more than a pronominal reference to the continuous participant will be needed (compare also the presence of the temporal closure at the end of the narrative passage, discussed above). The default reading is that the temporal sequentiality of the text continues even in this expository passage in terms of a time window into the period introduced in the first two sentences of the paragraph. The equestrian monument would thus have been made before the Last Supper: This interpretation, even though not the only one possible and not explicitly stated in the text, rises from the use of and between two simple past-tense verbs in a text that has previously activated expectations of a temporal text strategy and that displays a continuous participant. The clause-initial locative adverbial moves the readership mentally to another place in Milan and then instructs them to focus on the famous painting – a text strategy typically found in travel guides, where people need to be instructed to move from one spot to another, stop in the right places, and then take a look at particular sights (for a detailed discussion, see Enkvist 1985a; Virtanen 1992a, 2004). The third sentence of this expository passage starts with a full noun phrase reference to another painting from the same period of time, anchored to the artist through the genitive determiner his, whose pronominal form contributes to signalling continuation of the given textual unit. It is only in the full noun phrase reference to the continuous participant at the outset of the last narrative passage that we again find a text-strategic signal of a new textual unit, reinforcing the temporal marker at this major boundary. What this relatively close scrutiny of a well-formulated, condensed encyclopedic article tells us is that the expository type of discourse is here realized through narrative. The major boundaries of the text are the ones exhibiting more than one marker of the various text strategies used to structure the text: the beginnings of the three narrative passages. The absolute form of the temporal markers, however, points to an indirect use of this text type. Two expository passages, in direct use, are embedded in the narrative. The first of these is followed by a descriptive passage, which also realizes the expository discourse type. New textual units are linguistically signalled in ways that contribute to continuity and coherence. Such beginnings also incorporate markers of text segmentation and the hierarchy of text structure. While the outset of a new unit is a conspic-

66

Tuija Virtanen

uous position in the text, the end of such a unit may also include cues that help readers anticipate a textual boundary. Such signals may be part of the semantic dimensions functioning in the service of text structuring; yet, it is at the outset of a new textual unit that we can expect to find explicit signals of the current text strategies, signals that differ from what is characteristic of the rest of the text. This is the motivation for the initial placement of the various adverbials in the above text. 3.2. Other texts In what follows, another four extracts are presented from the same encyclopedia (Myers and Copplestone 1981). The first three realize the expository type of discourse, default in this genre. Example (2) is similar to (1) as concerns participant orientation; however, signals of a temporal text strategy are practically missing, apart from an occasional and connecting two temporally sequential events, and the sequence of the adverbials first and later. This text fragment can be characterized as a mixture of the descriptive, narrative, and expository types of text. Example (3), again, illustrates the expository type of text in its primary/direct use, conforming to topic continuity – a topic-oriented text strategy – and organized around a taxonomy of the various parts and aspects of the discourse topic. Example (4) is expository and descriptive in nature: The expository text type here embeds a short descriptive passage, structured according to a locative text strategy, which mentally moves the reader from one place to another in order to better understand the exposition of the phenomena under attention. The text helps us zoom in from the independent cities of the Etruscans, outside to their cemeteries, and finally to the tombs of their aristocratic families. We are then redirected to one location after another, and at each stop the text is zooming in to the details of the particular tombs in a particular locative order. The last text fragment, example (5), is fundamentally different from the others: It is argumentative in character. In this example from the last chapter of the book, explicitly presenting the authors’ views, the argumentative discourse type and the corresponding text type agree. Argumentative text strategies function in a very different way from those presented above. More than syntax, the cues to the thought patterns exhibited in the text rely on discourse phenomena such as various metatextual comments, attitudinal connectors, and intricate lexical ties. All of these convey evaluative meanings. In the sample the majority of the lexical items indicate stance and affect (e.g. death; announce, countless times, always premature; point out; bewildering; great priority; appeal; regard, and many more). In accordance with the conventions of learned exposition, metatextual comments may appear in the passive as in having said that it should

Variation across texts

67

be pointed out that; but they can also be encapsulated anaphora such as the announcement. The connectors of the sample include however, and some occurrences of ands, in the first instance indicating temporal succession and contrast, and in the second instance, the third and final point of a list of increased weight, giving evidence for the claim made earlier in this long sentence. Grammatical devices include the use of negation for evaluative purposes, the semi-auxiliary seem to, hedging the evaluative statement in accordance with the conventions of learned exposition, the auxiliary should, denoting the obligation of pointing out, the use of vague quantifiers for hedging purposes – many, much, one of the most, little, and indeed the layers of syntactic embedding in what is pointed out. (2) DES / EXP / NAR ↔ EXP Albrecht Durer ¨ of Nuremberg was a painter and graphic artist by profession and he was endowed with intellectual powers that put him in line with the great Italian masters. He was profoundly concerned with the theory of art and Ø wrote treatises on various subjects including proportion and artistic theory. He was the son of a goldsmith and Ø first trained in his father’s shop but later Ø transferred to painting. He was apprenticed to Michael Wolgemut, a competent but uninspired painter who ran the largest workshop in Nuremberg. Wolgemut had close links with the publishers including [. . . ]. Despite the limitations of a rigid workshop training, Durer ¨ was not content to confine himself to his own artisan class. He set about widening his education by studying mathematics, Latin and the Humanities, thus broadening his horizons intellectually and artistically. (3) EXP ↔ EXP Romanesque architecture is remarkable for its variations between one region and another. It was dome-vaulted in the south-west of France, but Ø retained timber ceilings in many Anglo-Norman churches and also in Italy. The materials employed were generally local. Romanesque mural-painting was inspired by Byzantine models. Enough remains to point to the existence of at least two schools of painting. One is Greek in origin and Ø is found in Burgundy and the Germanic empire. It uses several layers of grounding and colour, fixed with wax. It is thick and brilliant. The other prevails in the Loire valley and Ø consists of distemper, with the colours diluted in a mixture of water and liquid glue applied on the dry plaster; the colours are light and mat. (4) EXP / DES ↔ EXP The Etruscans are the first native Italian civilisation in the area north and west of the River Tiber during the 8th century BC. Their art was influenced by [. . . ]. The Etruscans lived in independent cities. Outside these, large cemeteries contained the rock-cut tombs of aristocratic families. Tombs were regarded as the houses of the dead, and Ø imitated those of the living. At Caere, their interiors were carved with doorways, windows and rafters, features of timber construction copied in stone, and Ø had several connecting rooms. At Tarquinia and elsewhere, the ceilings and walls were painted.

68

Tuija Virtanen

(5) ARG ↔ ARG [. . . ] The death of painting and sculpture has been announced countless times, and the announcement has always been premature. Having said that, however, it should be pointed out that for many people one of the most bewildering aspects of much artistic activity today is that it seems to place no great priority on visual appeal (as opposed to communication by visual means), has little regard to the traditional yardstick of manual dexterity and craftsmanship, and in many cases is not concerned to leave any permanent record. (all texts from: Myers and Copplestone 1981)

Evaluation is omnipresent in texts of various types, in one form or another (including efforts to silence it). Yet, recent studies suggest that the argumentative discourse type is, in fact, the default in much of art criticism and history (see, e.g., the discussion in Merlini Barbaresi 2004; Tucker 2003). In contrast, the texts under investigation here come from an encyclopedia, thus activating expectations of an expository type of discourse. And this type of discourse, as we have seen, may adopt several different text-typical guises. Ignoring, for the moment, the presence of implicit or explicit evaluation in all discourse, it is, however, obvious that the borderline between argumentation and exposition is especially fuzzy. This is facilitated by the fact that both rely on text strategies that are fundamentally different from the text-structuring continuities manifest in narrative, instructive/procedural, and descriptive texts. As pointed out above, Dorgeloh (1997: 143–144) draws attention to the fact that the expository and argumentative types of discourse are predominantly interpersonal while the ideational metafunction is primary in the narrative and descriptive discourse types. Hence, writers of expository and argumentative texts are said to be intruding into the minds of their readership through explaining and evaluating, in a way that is very different from the observer-based viewing of experienced, or imagined, events or states conveyed through narrative (dynamic viewing) and description (static viewing). In this light, instructive (procedural) discourse is special. The interpersonal metafunction is predominant like in expository and argumentative discourse; yet, the prototypical text strategy in instructive texts is action-oriented, inherently temporal, and hence rather like the text-strategic continuities in narrative and descriptive texts. The above examples come from one and the same source, an encyclopedia of art. While the encyclopedic article constitutes an established genre, there are great differences between encyclopedias and various sub-genres can be singled out. It has been repeatedly pointed out above that genres are fundamentally different from text/discourse types. Let us therefore devote the next section to genre, examining it in its own right before turning to a discussion of both kinds of categorizations in section 5.

Variation across texts

69

4. Genre dynamics While investigations of text and discourse types are primarily concerned with the first member of the reflexive text-context pair of notions, genre studies set out to explore its second member. Genres are characterized in contextual terms, relating them to situational and socio-cultural criteria, and in many approaches, to a view of discourse as social action. They are regarded as prototypical categories, which serve to link the world to the words and the words to the world. The context to be focused on in genre studies varies from concrete, situated investigations of genres in the making, to increasingly abstract socio-cultural contexts where the focus is on speech and discourse communities of various kinds. Genres also provide a way of investigating the workings of discourse in the formation and dynamics of various communities of practice. There is, however, a fundamental difference between approaches that view genres as relatively concrete manifestations of situated social action (see, e.g., the discussion of workplace incident reports in Giltrow, this volume, and complaints and reproaches in impromptu speech in G¨unthner, this volume) and others where genre is understood as a highly abstract concept mediating between texts and contexts (see, e.g., Bazerman 1988; Eggins 1994; see also the discussion in Bhatia 2005). The different approaches to genres are related to very different theoretical and pedagogical roots and goals (see, e.g., Bakhtin 1986; Bhatia 1993; Briggs and Bauman 1992; Giltrow, this volume; Halliday and Hasan 1985; Hanks 1987; Hymes 1972; Maingueneau 2002; Miller 1984; Paltridge 1995; Swales 1990, 2004; for a recent survey of different approaches, see Bauman 2006; for literary genres, see, e.g., Fowler 1982). Linguistic studies of concrete, text-externally definable categories sometimes prefer the term text type to genre, in line with the German tradition of Textsorte (see, e.g., Kohnen 2001; also the discussion in Dorgeloh 1997: 140–141; Taavitsainen 2001). All in all, however, text types as understood in the present chapter and genres in a wide range of interpretations are constructs of a wholly different nature, and section 5, below, elaborates further on their differences and interrelationships. As genre repertoires are, in principle, open-ended, discourse phenomena of very different kinds are given this epithet. Established genre labels are usually considered evidence of their existence and they are therefore examined as potential sources of insight into the particular community and its communicative needs and goals (Swales 1990). In genre studies, it is thus especially important to respect what Lyons (1963: 5) called “material adequacy.” Speech and discourse communities and various communities of practice are said to own particular genres as it is through these constructs that they are able to function as such for people in various constellations. Genres can be public, semi-public,

70

Tuija Virtanen

or private. But in other ways, too, they are not all equally accessible to people: the genres of the various discourse communities generally demand some expertise, often a very high level of specialization. While not directly involved with genre analysis, Sarangi (2007: 579) draws attention to a central problem in investigations of professional discourse other than our own, in terms of the “analyst’s paradox” pertaining to the interpretation of the insights of members of particular discourse communities. Studying genre as social action demands reflection on the extent and nature of alignment between discourse analysts’ and various participants’ perspectives. To understand the form and function of a particular genre in relation to a given discourse community it is often helpful to investigate the genre through comparisons with adjacent genres (cf. Todorov 1976). But we also need to understand how genres help construct interdiscursive chains through which they are mediated (Fairclough 1992). Methodological dimensions include: (i) choices between data-driven and theory-driven approaches; and (ii) whether to focus on linguistic and/or situational characteristics, or rather explore user/interlocutor/participant awareness and/or understanding of particular genres. The outcome is likely to be very different if we, for instance, study (a set of) predetermined genres as manifested in a particular body of data, set out to investigate (clusters of) linguistic elements in a large corpus, or perform a situated analysis of interlocutors doing genre face-to-face or online. In each case it is instructive to carefully consider the authenticity of the data (cf. Gill 2008). These are all central issues in the study of other kinds of text categorizations as well (see also the discussion in Virtanen 2008 of the use of corpora in the study of discourse). 4.1. Emergent genres Genres tend to emerge from the familiar: from already established genres, through conscious efforts to develop something that is unlike them, or through processes of intergenre formation. But they may also evolve afresh to fill gaps that exist in particular socio-cultural or professional genre spaces. Such gaps may be the result of social or technological changes in the mediation of discourse, new communicative needs, goals and possibilities, and the formation of new discourse communities and/or communities of practice. Apart from highly conventionalized genres, such as the tax form and other products of document design (which are changed over time through a conscious, goal-oriented effort), texts realizing particular genres exhibit a great deal of variation. This is so because of the dynamism and flexibility needed for them to be of use to members of communities of various kinds who engage in situated social action through discourse. But variability is also a potential source of generic change.

Variation across texts

71

Halmari and Virtanen (2005: 241) devised an intergenre model based on the two processes of minimizing versus maximizing intertextual and interdiscursive gaps. The dynamic and dialectic nature of genres as well as changes in genreexternal and genre-internal pressures may result in a maximization process of the intertextual gaps within established genres (i.e., gaps between individual texts realizing and constructing the particular genre). This process may, over time, lead to the emergence of intergenres as a particular genre may thus acquire characteristics of another genre to a degree that will motivate speaking of a minimization of interdiscursive gaps (i.e., gaps between genres). For instance, infomercials can in this light be viewed as an intergenre, having their roots in advertisements and product information. Limiting their discussion to persuasive genres, Halmari and Virtanen (2005: 239) argue that “a lot of the ‘categorical ambiguity’ [. . . ] encountered in the attempts to typologize various texts into their generic ‘types’ can be explained by looking at diachronic changes caused by (1) the need for new, fresh forms of persuasion within one genre and (2) the need to keep persuasion implicit. The first factor is often embedded in the second.” But not all genres evolve through intergenre processes. In a many-faceted study, Gotti (this volume) examines the conscious development of a new scholarly genre, the experimental essay, within a specialized discourse community. The chapter sheds light on the extent and limits of metapragmatic awareness manifested by members of the scientific community in relation to the new genre. Understanding current genres and their dynamics is greatly facilitated by insights gained from their historical developments as well as those of other, adjacent genres. Diachronically oriented studies often approach generic change in terms of particular linguistic phenomena, which may be related to text and discourse types and/or changes in genre-internal or genre-external pressures. Hence, Wanner (2009) is concerned with the rise and fall of the passive in academic writing. Dorgeloh and Wanner (2009) approach genre conventions in scientific discourse from the two perspectives of syntactic productivity and formulaic language. Kabatek, Obrist, and Vincis (this volume) combine three studies of clause linkage in various genres in different Romance languages in order to investigate the potential of clause linkage to signal systematic text-strategic differences across text types and genres, as well as changes in them. A finding relative to the historical roots of current genres comes from Rheto-Romance newspapers: Commentaries and lead articles included in them are still closer to older journalistic language than news texts. P´erez-Guerra and Mart´ınez-Insua (this volume) focus on changes in linguistic complexity in news texts and letters in a decontextualizing study that assumes generic comparability over time. Drawing a distinction

72

Tuija Virtanen

between syntactic and lexical complexity, they found little variation in syntactic complexity over time and between the two genres, but a higher degree of lexical complexity, increasing over time, in news as compared to letters. Syntactic and lexical complexity are distinct from textual complexity, which would be another topic worthy of diachronic enquiry in terms of both text types and genres (for synchronic approaches to textual complexity in relation to text type, see, e.g., the discussions in Virtanen 1992a; Merlini Barbaresi 2004). Historical linguists are usually acutely aware of the effects that the limited availability of data has on the kinds of conclusions that they will be able to draw, and it is indeed a field where linguists have engaged in discussions of the burning issues of the relative authenticity and comparability of the data (see, e.g., Kyt¨o 2001; W˚arvik 1990). Herring, van Reenen, and Schøsler (2001: 9–10) advocate inductive approaches from available evidence to investigate text types and genres in older languages as these are likely to differ considerably from their possible present-day equivalents. Access to the Internet has, in recent years, produced a growing number of new textual and/or contextual categories, many with roots in one or several established offline genres that have been adapted to the demands, constraints, and possibilities of electronic discourse. Such categories may gradually develop into full-fledged genres owned by virtual discourse communities or virtual communities of practice. Members of virtual communities show a high degree of metapragmatic awareness concerning the existence and development potential of the categories they are using. Irrespective of whether these can be as yet called genres, such electronic categories of text and/or context are the outcome of users’ efforts, conscious or not, technological affordances, and the rapid recontextualization processes taking place in computer-mediated communication. Of interest to studies of genre formation is also the coining and subsequent development of names for new computer-mediated modes and other categories of the verbal and the visual. People are communicating with one another through computers and mobile phones, adapting to the technology and profiting from new platforms and programs for a variety of communicative goals. Electronic genres and related categories occasionally leak (back) offline, thus affecting a range of established genres in a particular genre space. It is crucial that linguists set out to examine emergent online genres and other categorizations of computer-mediated texts and contexts, in view of the processes of genre dynamism and the impact that such processes have on people’s lives, at work and leisure. (For discussions of the emergence of genres in computer-mediated communication, see, e.g., Herring 2007; Herring, Stein, and Virtanen, forthcoming; Heyd, this volume; the contributions to Kwa´snik and Crowston 2005; Miller and Shepherd 2004; Myers 2009.)

Variation across texts

73

While emergent genres are always in some sense new, we may, however, well ask ourselves whether they can ever be totally detached from the familiar, rather than being, for instance, the outcome of unused possibilities or efforts to deviate from some set of existing genre conventions in order to create the opposite. Traces of their former identities and contexts may, however, fade or vanish over time. Comparing the development of electronic news stories as separate from corresponding news in print, Byrman and Virtanen (2001) came to think of the invention of the car engine: Daimler’s 1886 car engine was built into a coach body and hence incorporated in a familiar, established genre – rather than creating a brand-new genre for this new invention. 4.2. Genres on the move Giltrow (this volume) points out that genres are always on the move. And recycling them affects their linguistic and contextual characteristics. This happens, for instance, when established genre conventions are adapted to new environments, or appropriated, and sometimes relabelled, for new purposes. Borrowing terms from translation studies (Chesterman 1997: 108), we can, for instance, speak of people setting out to “foreignize” or “domesticate” genres. Genres can be foreignized from spontaneous everyday speech (their primary source; cf. Bakhtin 1986), for instance, to serve a given discourse community. This is true of many scholarly genres, recounts of various kinds (police, insurance reports), a number of marketing genres (e.g., sales talk in object descriptions), and many others. Yet, it is important to recall that there are also many (institutional) genres that are inherently written, bearing no trace of having once been foreignized from spontaneous spoken interaction, for instance, genres such as tax forms, book-jacket blurbs, or annual reports. In contrast to genres that have been foreignized from face-to-face impromptu speech for various socio-cultural purposes, genres originating in professional discourse communities can be domesticated, to serve a wider number of people in non-professional contexts. Examples include counselling and therapy sessions, which have been spread through popular media into everyday “psychobabble,” adopting linguistic characteristics of the corresponding professional genres. But genres can also be appropriated to new environments: witness, for instance the use of commercial genres in the service of education (edutainment). Bhatia (2005) discusses the recent large-scale “colonization” by promotional discourse of a wide range of academic, professional and institutional genres (cf. also Fairclough 1992). The genre of the public apology (discussed, for instance, in Lakoff 2000) has been foreignized from its everyday use and pragmatic functions, but it is still reminiscent enough of the everyday apology to succeed in

74

Tuija Virtanen

forwarding given political purposes. At the same time, it has been appropriated to new contexts, serving a context-constructing function through institutionalization. Genres can also be relabelled to fit new uses. It is not difficult to think of, for instance, sensational revelations presented as news in tabloids that have been very quickly turned into analytical columns in quality papers – because such sensational revelations are not appropriate material for respectable papers. Thus there is potential for insight not only in genre labels as such but also in interdiscursive relabelling processes, or indeed the lack of them in appropriated genres. Genres are an important source of information for our understanding of the workings of discourse in society. In the words of Greg Myers (2000: 188): “We both use genres and are channelled by them; that is the tension that must guide any of our analyses, whether of corporate annual reports, handbooks, e-mail, newspapers, or leaflets. Change comes, not from the inside or outside, but in that tension.”

5. Genres and text types: Same or different? Genres manifest linguistic differences that need to be explored in terms of context-sensitive frameworks oriented towards the performativity and/or collaborative sense-making of social action. In a situated study, text types would seem to be abstractions that are made use of in concrete instances of doing genre. Adopting an abstract view of genre, however, would appear to result in very general categories encompassing the choice of text type. If, for the purposes of a particular study, it is necessary to construe genre as having effect simultaneously on more than one abstraction level, the resulting model of genres would seem to cut across the two-level model of text and discourse types without overlap. There is thus no need for Occam’s Razor. Appropriating parts of text-typological enquiry into the field of genre studies is quite possible, although this avenue fails to do justice to the role of text and discourse types in language use. The insight into the dynamism of the two levels of text types and discourse types provided by context-sensitive scrutiny of texts is very different from insight gained from observing interlocutors do genre in particular communicative situations or exploring the communicative purposes of speech or writing as experienced by participants in a discourse community or community of practice. Investigating texts and discourses in terms of text/discourse types and genres (in its various senses) opens different avenues to the kinds of communicative goals, text strategies, and social action in which people are (re-)engaged through

Variation across texts

75

discourse. The fact that genres can be talked about using shared labels but text type differences are recognized in experiments (and perhaps they influence people’s choices of what to read and how to write) suggests that they constitute two very different phenomena, even though both have to do with similarities and differences between texts and discourses (and they are not the only ones: recall registers and styles, too). Assimilating text type selection and the degree of prototypicality to the generic structure potential of a particular genre in the modelling of variation across texts and discourses runs the risk of silencing the tension between text type as form and discourse type as function, which is an important source of insight into their interplay and the evolution of text types. The tension between form and function in relation to genre is of a different nature, discussed by Giltrow (this volume). While both fiction and non-fiction may manifest similar repertoires of text types, in fiction the match between text type and discourse type is determined by the reader. Students of literature work with the supertype of (literary) narrative. When subcategorizations are needed, it may therefore be expedient to rather focus on genre differences within narrative, for instance taking into account categories such as myth, thriller, fairy tale, romance, science fiction, and so forth. Fanfiction is an emergent computer-mediated genre; others will no doubt follow suit as a result of the various kinds of interactive, collaborative storytelling taking place online. Narrative is a central notion in the study of literature and story has acquired a specialized meaning in narratology, even though it can appear in compound genre labels such as detective story, or short story, as opposed to novel. But there are also many non-fiction genres that may exhibit characteristics of narrative text, such as reports, chronicles, biographies, or news stories. Some of these are called stories, while others are not, perhaps in order to distinguish them from the storytelling genres. Everyday storytelling, too, can be approached through its genres, such as gossip, jokes, accounts, confessions, reminiscences, and many more (see, e.g., Ochs and Capps 2001). Such genres realize the narrative type of text/discourse but they are very different from one another: witness a joke and a parable, both narrative. Everyday storytelling takes place face-to-face and through phones and computers; new storytelling genres may thus also emerge in online contexts such as blogs, various discussion fora, egocasts and other modes of computermediated communication. Even though narrative is sometimes used in the sense of a genre in studies of impromptu speech, it is best reserved to discussions of narrative and non-narrative types of text/discourse, as separate from genres (see also Giltrow, this volume). Narrative is not a genre on a par with, say, sermon, report, diary, weather forecast, horoscope, or book proposal. Nor is it a genre such as an anecdote, parable, or face-to-face reproach. While a story, account

76

Tuija Virtanen

or recount is a genre, narrative is a type of text or discourse. A narrative genre would thus refer to a genre that realizes the narrative type of text/discourse – a genre which carries another label as interlocutors tend not to speak of narratives. Linguists also use the term narrative as an umbrella term for storytelling genres in planned and unplanned speech. Narratives are found in myriad genres of fiction and non-fiction, of speech and writing. But non-narrative types of text also appear in a number of different genres: witness the use of instructive text, for instance, in recipes, guidebooks of various kinds, or computing pages on the internet; the use of expository text in encyclopedias, textbooks, and many other scholarly genres, or indeed in a number of organisational genres such as various plans; the use of overt argumentation in reviews, judicial opinions, or expert statements concerning candidates who have applied for a particular academic position. We have seen in section 2.2., above, that descriptive texts adapt to the dominant type of text; hence, descriptions in fictional and everyday storytelling genres are closely woven into the narrative and as such very different from spoken or written object descriptions in art criticism or in marketing, technical product descriptions, or descriptions of scenery, sights, and the recommended route in travel guides. Keeping apart text/discourse types and genres in linguistic analysis helps us disclose very different aspects of the text-context interface. People are capable of making good use of both kinds of categorizations, simultaneously but in different ways and for different purposes. Both genres and text types are prototypical notions that allow for variability. Text types have to be flexible for people to be able to use them as heuristics for discourse processing in particular communicative situations. And genres have to be flexible to be able to function as social action. Linguists need separate tools for the analysis of text/discourse types and genres to do justice to this complex field, and an integrated model will have to take both into account in their own right.

6. Concluding remarks One and the same text type can be put to use in very different genres, and one and the same genre easily manifests texts that can be related to very different types. Particular genres, owned by communities of various kinds, may grow fond of particular types of text and/or resist others. But all these practices leave traces, in text types and in genres. Concrete linguistic analyses of these two different kinds of linguistic constructs contribute to our understanding of the emergence and change of each phenomenon, as well as their interplay in particular contexts

Variation across texts

77

and cultures. Text types and genres invite investigations focusing on the nexus of the linguistic, social, and cognitive dimensions of discourse. People easily pay attention to lexis and pronunciation, but many find syntactic structures difficult to grasp. Similarly, people pay attention to content and affect, and make sense of contexts and cultures through genres, as in the very concrete sense demonstrated by shared genre labels. But perhaps many are much less aware of their or other people’s use of text strategies related to text types. Linguists, however, need to be aware of both kinds of phenomena in order to study the forms and functions that are emergent in and across texts and contexts. For linguists of any orientation there is added value in taking into account both genre dynamics and text/discourse types. As the present volume shows, theoretical and methodological variation leading to results that are not necessarily concomitant provides us with new vistas for the study of (the syntactic intricacies of) genre dynamics and the interface between text and discourse types.

References Adam, Jean-Michel 2005 Les Textes: Types et Prototypes. 2nd Edition. Paris: Armand Colin. Bakhtin, Mikhail M. 1986 Speech Genres and Other Late Essays. (University of Texas Press Slavic Series 8.) Austin: University of Texas Press. Bauman, Richard 2006 Speech genres in cultural practice. In: Keith Brown (ed.), Encyclopedia of Language & Linguistics, 745–758. 2nd Edition. Amsterdam: Elsevier. Bazerman, Charles 1988 Shaping Written Knowledge: The Genre and Activity of the Experimental Article in Science. (Rhetoric of the Human Sciences.) Madison: The University of Wisconsin Press. de Beaugrande, Robert and Wolfgang Dressler 1981 Introduction toText Linguistics. (Longman Linguistics Library 26.) London: Longman. Bhatia, Vijay K. 1993 Analysing Genre: Language Use in Professional Settings. (Applied Linguistics and Language Study.) London: Longman. Bhatia, Vijay K. 2005 Generic patterns in promotional discourse. In: Helena Halmari and Tuija Virtanen (eds.), Persuasion Across Genres: A Linguistic Approach, 213– 225. (Pragmatics and Beyond, New Series 130.) Amsterdam/Philadelphia: John Benjamins.

78

Tuija Virtanen

Biber, Douglas 1988 Variation Across Speech and Writing. Cambridge: Cambridge University Press. Biber, Douglas 2006 University Language: A Corpus-based Study of Spoken and Written Registers. (Studies in Corpus Linguistics 23.) Amsterdam/Philadelphia: John Benjamins. Briggs, Charles L. and Richard Bauman 1992 Genre, intertextuality, and social power. Journal of Linguistic Anthropology 2: 131–172. Byrman, Gunilla and Tuija Virtanen 2001 Det lokala globalt – om nyheter p˚a webben [The local globally – on news on the web]. Humanetten 9. Available at http://vxu.se/hum/publ/ humanetten/nummer9/art0118.html. Chesterman, Andrew 1997 Memes of Translation: The Spread of Ideas in Translation Theory. (Benjamins Translation Library 22.) Amsterdam/Philadelphia: John Benjamins. Crystal, David and Derek Davy 1969 Investigating English Style. (English Language Series.) London & Harlow: Longmans. Dittmar, Norbert ¨ 1995 Register. In: Jef Verschueren, Jan-Ola Ostman, Jan Blommaert and Chris Bulcaen (eds.), Handbook of Pragmatics.Amsterdam/Philadelphia: John Benjamins. Dorgeloh, Heidrun 1997 Inversion in Modern English: Form and Function. (Studies in Discourse and Grammar 6.) Amsterdam/Philadelphia: John Benjamins. Dorgeloh, Heidrun 2004 Conjunction in sentence and discourse: Sentence-initial and and discourse structure. Journal of Pragmatics 36: 1761–1779. Dorgeloh, Heidrun 2005 Patterns of agentivity and narrativity in early science discourse. In: Janne Skaffari, Matti Peikola, Ruth Carroll, Risto Hiltunen and Brita W˚arvik (eds.), Opening Windows on Texts and Discourses of the Past, 83–94. (Pragmatics & Beyond 134.) Amsterdam/Philadelphia: John Benjamins. Dorgeloh, Heidrun and Anja Wanner 2009 Formulaic argumentation in scientific discourse. In: Roberta Corrigan, Edith A. Moravcsik, Hamid Ouali and Kathleen M. Wheatley (eds.), Formulaic Language: Volume 2. Acquisition, Loss, Psychological Reality, and Functional Explanations, 523–544. (Typological Studies in Language 82.) Amsterdam/Philadelphia: John Benjamins.

Variation across texts

79

Eggins, Suzanne 1994 An Introduction to Systemic Functional Linguistics. London: Pinter. Enkvist, Nils Erik 1985a A parametric view of word order. In: Emil S¨ozer (ed.), Text Connexity, Text Coherence: Aspects, Methods, Results, 320–336. (Papiere zur Textlinguistik 49.) Hamburg: Helmut Buske. Enkvist, Nils Erik 1985b Text and discourse linguistics, rhetoric, and stylistics. In: Teun A. van Dijk (ed.), Discourse and Literature: New Approaches to the Analysis of Literary Genres, 11–38. (Critical Theory 3.) Amsterdam/Philadelphia: John Benjamins. Enkvist, Nils Erik 1987 A note towards the definition of text strategy. Zeitschrift f¨ur Phonetik, Sprachwissenschaft und Kommunikationsforschung 40: 19–27. Esser, J¨urgen 2006 Presentation in Language: Rethinking Speech and Writing. T¨ubingen: Gunter Narr. Fairclough, Norman 1992 Discourse and Social Change. Cambridge: Polity Press. Fern´andez Toledo, Piedad 2005 Genre analysis and reading of English as a foreign language: Genre schemata beyond text typologies. Journal of Pragmatics 37: 1059–1079. Fowler, Alistair 1982 Kinds of Literature. Oxford: Oxford University Press. Gill, Martin ¨ 2008 Authenticity. In: Jef Verschueren and Jan-Ola Ostman (eds.), Handbook of Pragmatics Online. Amsterdam/Philadelphia: John Benjamins. http://www.benjamins.com/online/hop. Giltrow, Janet this volume Genre as difference: The sociality of linguistic variation. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 29– 51. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Gotti, Maurizio this volume A new genre for a specialized community: The rise of the experimental essay. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 85–110. Berlin/New York: Mouton de Gruyter. G¨unthner, Susanne this volume Grammatical constructions and communicative genres. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 195–217. (Topics in English Linguistics 70.) Berlin/NewYork: Mouton de Gruyter.

80

Tuija Virtanen

Halliday, M.A.K. 1978 Language as a Social Semiotic: The Social Interpretation of Language and Meaning. London: Edward Arnold. Halliday, M.A.K. and Ruqaiya Hasan 1985 Language, Context, and Text: Aspects of Language in a Social-Semiotic Perspective. (Specialised Curriculum: Language and Learning.) Victoria: Deakin University Press. Halmari, Helena and Tuija Virtanen 2005 Towards understanding modern persuasion. In: Helena Halmari andTuija Virtanen (eds.), Persuasion across Genres: A Linguistic Approach, 229– 244. (Pragmatics & Beyond, New Series 130.) Amsterdam/Philadelphia: John Benjamins. Hanks, William 1987 Discourse genres in a theory of practice. American Ethnologist 14: 668– 692. Herring, Susan C. 2007 A faceted classification scheme for computer-mediated discourse. Language@Internet 4: Article 1. Retrieved 12 May, 2008, from http://www. languageatinternet.de/articles/2007/761. Herring, Susan C., Pieter van Reenen and Lene Schøsler 2000 On textual parameters and older languages. In: Susan C. Herring, Pieter van Reenen and Lene Schøsler (eds.), Textual Parameters in Older Languages, 1–31. (Current Issues in Linguistic Theory 195.) Amsterdam/ Philadelphia: John Benjamins. Herring, Susan C., Dieter Stein and Tuija Virtanen (eds.) forthcoming Handbook of the Pragmatics of Computer-Mediated Communication. Berlin/New York: Mouton de Gruyter. Heyd, Theresa this volume I know you guys hate forwards: Address pronouns in digital folklore. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 333–358. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Hymes, Dell 1972 Models of the interaction of language and social life. In: John J. Gumperz and Dell Hymes (eds.), Directions in Sociolinguistics: The Ethnography of Communication, 35–71. New York: Holt, Rinehart & Winston. Kabatek, Johannes, Philipp Obrist and Valentina Vincis this volume Clause linkage techniques as a symptom of discourse traditions: Methodological issues and evidence from Romance languages. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 247– 275. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter.

Variation across texts

81

Kinneavy, James L. [1971] 1980 A Theory of Discourse. New York: W.W. Norton. Kohnen, Thomas 2001 On defining text types within historical linguistics: The case of petitions/statutes. European Journal of English Studies 5: 197–203. Kreyer, Rolf this volume Syntactic constructions as a means of spatial representation in fictional prose. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 277–303. Berlin/New York: Mouton de Gruyter. Kroon, Caroline 2002 How to write a ghost story? A linguistic view on narrative modes in Pliny Ep. 7.27. In: Donna Shalev and Lea Sawicki (eds.), Donum Grammaticum. Studies in Latin and Celtic Linguistics in Honour of Hannah Ros´en, 189–200. Leuven: Peeters. Kwa´snik, Barbara H. and Kevin Crowston (eds.) 2005 Genres of Digital Documents. Special issue of Information Technology & People 18. Kyt¨o, Merja 2000 Robert Keayne’s Notebooks: A verbatim record of spoken English in Early Boston? In: Susan C. Herring, Pieter van Reenan and Lene Schløsler (eds.), Textual Parameters in Older Languages, 273–308. (Current Issues in Linguistic Theory 195.) Amsterdam/Philadelphia: John Benjamins. Lakoff, Robin Tolmach 2000 The Language War. Berkeley: University of California Press. Lindeberg, Ann-Charlotte 2004 Promotion and Politeness: Conflicting Scholarly Rhetoric in Three Dis˚ bo: A ˚ bo Akademi University Press. ciplines. A Longacre, Robert E. 1983 The Grammar of Discourse. (Topics in Language and Linguistics.) New York: Plenum Press. Lyons, John 1963 Structural Semantics. (Publications of the Philological Society 20.) Oxford: Publications of the Philological Society. Maingueneau, Dominique 2002 Analysis of an academic genre. Discourse Studies 4: 319–342. Merlini Barbaresi, Lavinia 2004 A model for defining complexity in descriptive text type. Folia Linguistica 38: 355–381. Miller, Carolyn R. 1984 Genre as social action. Quarterly Journal of Speech 70: 151–67.

82

Tuija Virtanen

Miller, Carolyn R. and Dawn Shepherd 2004 Blogging as social action: A genre analysis of the weblog. In: Laura Gurak, Smiljana Antonijevic, Laurie Johnson, Clancy Ratliff and Jessica Reyman (eds.), Into the Blogosphere: Rhetoric, Community, and Culture of Weblogs. Retrieved 25 May, 2008, from http://blog.lib.umn.edu/blogosphere/ blogging as social action a genre analysis of the weblog.html. Myers, Bernard L. and Trewin Copplestone (eds.) 1981 The Macmillan Encyclopedia of Art. London: Papermac. Myers, Greg 2000 Powerpoints: Technology, lectures, and changing genres. In: Anna Trosborg (ed.), Analysing Professional Genres, 177–192. (Pragmatics & Beyond, New Series 74.) Amsterdam/Philadelphia: John Benjamins. Myers, Greg 2009 The Discourse of Blogs and Wikis. London: Continuum. Ochs, Elinor and Lisa Capps 2001 Living Narrative: Creating Lives in Everyday Storytelling. Cambridge, MA: Harvard University Press. Paltridge, Brian 1995 Working with genre: A pragmatic perspective. Journal of Pragmatics 24: 393–406. P´erez-Guerra, Javier and Ana E. Mart´ınez-Insua this volume Do some genres or text types become more complex than others? In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 111–140. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Pilegaard, Morten and Finn Frandsen ¨ 1996 Text type. In: Jef Verschueren, Jan-Ola Ostman, Jan Blommaert and Chris Bulcaen (eds.), Handbook of Pragmatics. Amsterdam/Philadelphia: John Benjamins. Prince, Ellen F. 1981 Toward a taxonomy of given–new information. In: Peter Cole (ed.), Radical Pragmatics, 223–225. New York: Academic Press. Sarangi, Srikant 2007 The anatomy of interpretation: Coming to terms with the analyst’s paradox in professional discourse studies. Text & Talk 27: 567–584. Shiro, Martha 2003 Genre and evaluation in narrative development. Journal of Child Language 30: 165–195. Smith, Carlota S. 2003 Modes of Discourse: The Local Structure of Texts. (Cambridge Studies in Linguistics 103.) Cambridge: Cambridge University Press.

Variation across texts

83

Swales, John M. 1990 Genre Analysis. (Cambridge Applied Linguistics.) Cambridge: Cambridge University Press. Swales, John M. 2004 Research Genres: Explorations and Applications. (Cambridge Applied Linguistics.) Cambridge: Cambridge University Press. Taavitsainen, Irma 2001 Changing conventions of writing: The dynamics of genres, text types, and text traditions. European Journal of English Studies 5: 139–150. Todorov, Tzvetan 1976 The origin of genres. New Literary History 8: 159–70. Tucker, Paul 2003 Evaluation in the art-historical research article. Journal of English for Academic Purposes 2: 291–312. Ummelen, Nicole 1997 Procedural and Declarative Information in Software Manuals: Effects on Information Use, Performance and Knowledge. (Utrecht Studies in Language and Communication 7.) Amsterdam: Rodopi. Virtanen, Tuija 1992a Discourse Functions of Adverbial Placement in English: Clause-Initial Adverbials of Time and Place in Narratives and Procedural Place De˚ bo: A ˚ bo Akademi University Press. scriptions. A Virtanen, Tuija 1992b Issues of text typology: Narrative – a ‘basic’ type of text? Text 12: 293– 310. Virtanen, Tuija 1992c Given and new information in adverbials: Clause-initial adverbials of time and place. Journal of Pragmatics 17: 99–115. Virtanen, Tuija 2004 Point of departure: Cognitive aspects of sentence-initial adverbials. In: Tuija Virtanen (ed.), Approaches to Cognition through Text and Discourse, 79–97. (Trends in Linguistics. Studies and Monographs 147.) Berlin/New York: Mouton de Gruyter. Virtanen, Tuija 2009 Corpora and discourse analysis. In:Anke L¨udeling and Merja Kyt¨o (eds.), Corpus Linguistics: An International Handbook, Volume 2, 1043–1070. Berlin/New York: Mouton de Gruyter. Wanner, Anja 2009 Deconstructing the English Passive. (Topics in English Linguistics 41.) Berlin/New York: Mouton de Gruyter. W˚arvik, Brita 1990 On the history of grounding markers in English narrative: Style or typology? In: Henning Andersen and Konrad Koerner (eds.), Historical

84

Tuija Virtanen Linguistics 1987: Papers from the 8th International Conference on Historical Linguistics, 531–542. (Current Issues in Linguistic Theory 66.) Amsterdam/Philadelphia: John Benjamins.

Werlich, Egon 1976 AText Grammar of English. (Uni Taschenb¨ucher 597.) Heidelberg: Quelle & Meyer.

A new genre for a specialized community: The rise of the experimental essay Maurizio Gotti

The paper examines the formal and dynamic characteristics of the birth of the genre of the experimental essay. Particular attention is paid to a comprehensive understanding not only of its content and syntactic aspects (especially the processes of nominalization and premodification) but also of how this genre was constructed, interpreted and used in the achievement of specific goals in specialized contexts. In this respect, the paper investigates the specific needs and requirements of the community of practice that prompted its appearance, as well as the links established by this new genre between the discursive practices of the members of this community and their research activities. The comments and examples provided to support each point are mainly drawn from Robert Boyle’s texts. This British scientist has been selected as a paradigmatic figure not only because his influence on the formation of this new genre has been widely acknowledged by contemporary and subsequent scientists and critics, but also because of the specific metatextual indications provided in his Proemial Essay [. . . ] with Some Considerations Touching Experimental Essays in General (1661) and the many practical exemplifications of this genre written by him.

1. Introduction The Late Middle English and the Early Modern English periods witnessed great changes in major genres used both in general and specialized contexts. In the field of narrative texts, for example, there was a great evolution from the short anecdotes found in jestbooks and collections of exempla to the longer and more coherent narrative works leading to the rise of the novel. Also, the fragmentary listing of unrelated events found in chronicles evolved into a more coherent account and description of events, opening the way to the development of newspaper articles. In the administrative field, early petitions acquired new stylistic

86

Maurizio Gotti

features and functions, contributing to the rise of the genre of statutes.1 Although many of these new genres were the result of changes in the nature of existing ones, in several cases their typical structures and communicative purposes became markedly different and thus led to radical generic innovations. One of these cases is represented by the experimental essay. The aim of this paper is to examine the formal and dynamic characteristics of the birth of the genre of the experimental essay. Particular attention will be paid to a comprehensive understanding of the interactions between the new genre and its context, focusing not only on its content and syntactic aspects (especially the processes of nominalization and pre-modification), but also on how this genre was constructed, interpreted, and used in the achievement of specific goals in specialized contexts. This is in line with the evolution of genre analysis, as the emphasis on text and context in the latest approaches in genre analysis have been almost reversed (cf. Swales 1998; Bhatia 2004), with context attracting more serious attention – though without undermining the importance of linguistic form – in the description of specialized genres. In this respect, the paper will investigate the general context in which the genre of the experimental essay originated, the specific needs and requirements of the community of practice (Lave and Wenger 1991) that prompted its appearance, and the links established by this new genre between the discursive practices of the members of this community and their research practices. The examples provided in support of each point will mainly be drawn from Robert Boyle’s texts. This British scientist has been selected as a paradigmatic figure both because of the specific metatextual indications provided in his Proemial Essay [. . . ] with Some Considerations Touching Experimental Essays in General (1661) and the many practical exemplifications of this genre written by him, such as his New Pneumatical Experiments about Respiration (1665–1678).2 Indeed, the influence of Boyle on the formation of this new genre has been widely acknowledged by both his contemporary and later scientists and critics. Boas Hall asserts: “[Boyle] not only described experiments, he taught the world how to write up experiments so that they could readily be repeated” (Boas Hall 1965: 43). Paradis not only recognises Boyle’s merit in creating this new genre but also acknowledges his influence on succeeding specialists, such as Newton: “Boyle was both a conscious inventor of the form he called the

1. For more details about these changes see Kohnen (2001); for a presentation of more general changes in the history of English see Diller and G¨orlach (2001); as regards the Early Modern English period see Moessner (2001). 2. These essays by Boyle, together with any references to this author, can be found in The Works (Boyle [1772] 1965).

A new genre for a specialized community: The rise of the experimental essay

87

‘experimental essay’ and its first extensive exploiter. We can hardly doubt that Newton took many of his literary cues from Boyle, whose work he had studied and mastered” (Paradis 1987: 85).

2. The context 2.1. The research practices The development of the experimental essay in the Early Modern English period was consequential to the great epistemological and methodological innovations that took place in seventeenth-century England (Vickers 1987; Hunter 1989; Jardine 1999; Shapiro 2000). Indeed, these innovations determined the need for corresponding changes both in regard to the methods of communicating information about new scientific discoveries and the most suitable means of expression chosen to describe and discuss the new phenomena then being observed and analysed. The realisation that the English language was inadequate for the needs of expression by men of science led to its gradual amelioration, both from a quantitative and a qualitative point of view. British scientists made great efforts to increase the number of specialized terms and to improve the exactness of their meanings. This new perspective determined a growth not only in movements whose aim was a reform of the existing verbal code in order to limit its inaccuracy and increase its power of expression, but also in movements concerned with the creation of novel expository genres able to guarantee the prompt and widespread diffusion of information relating to the new developments in the various specialized fields. Indeed, the needs of seventeenth-century natural philosophers could no longer be satisfied by the traditional essay, as this mainly followed the principles and employed the techniques of a prevalently literary type. The innovative characteristics of this new genre derived from the great importance attributed to the experimental process in the research programmes of Early Modern English men of science, scientists who – elaborating on Francis Bacon’s intuitions – shared the principle that the progress of knowledge could not be based on the servile observance of traditional theory, but should rely on the observation of natural phenomena and accurate experimental activity. To carry out their communicative task, scientists needed a new expository form. Already Bacon, by means of the aphorism, had tried out a writing form different from the traditional ones to convey personal observations related to short and specific items; this form, however, while useful for brief comments and reflections, was not deemed appropriate for the description of experiments.

88

Maurizio Gotti

The other main forms available to the scientist were the essay, the dialogue, and the treatise. The first, however, had been used also for literary or philosophical purposes, and therefore retained some rhetorical features unsuitable to scientific purposes. The latter two were more suitable to writings on lengthy issues and, when the purpose was mainly argumentative, would be used when a new theory was to be set out or when old ones were to be debated. The writing of treatises or dialogues was instead deemed inappropriate for tentative suggestions or the early formulation of hypotheses. This point of view was effectively expressed by Boyle in simile form: For as those, that apply themselves to procreation too young, and before they have attained to their full vigour and strength, do generally both hinder their own growth, and become the parents but of weak and short-lived children; so they, that too early, and before their judgment and experience be fully ripe, addict themselves to write books, do commonly both hinder their own proficiency in knowledge, and write but immature, and therefore seldom lasting treatises. (Boyle [1772] 1965: I, 299–300)

There was a need, instead, for a shorter form, one which would offer the scientist the opportunity to briefly report on experiments carried out, procedures followed, results obtained, and any relevant personal comments. This genre would allow for immediacy of communication, and would protect the writer from any accusation of incomplete theoretical exposition, as its purpose would be mainly descriptive rather than argumentative. The lack of such a genre would have detrimental effects on the development of the sciences, as it either compelled the experimenter to write unnecessarily long and theoretically still immature books, or it discouraged him from revealing his experiences and tentative opinions, as is confirmed by Boyle’s comments: It has long seemed to me none of the least impediments of the real advancement of true natural philosophy, that men have been so forward to write systems of it, and have thought themselves obliged either to be altogether silent, or not to write less than an entire body of physiology: for, from hence seem to have ensued not a few inconveniences. (Boyle [1772] 1965: I, 300)

The means of communication identified as appropriate for specialized purposes was the experimental essay, a form meant to enable the researcher to report his experiences with immediacy and precision. Such a genre, however, would differ from the traditional essay, as the traditional essay followed foreign models such as Montaigne’s, models structured on principles inspired either by literary or argumentative criteria.

A new genre for a specialized community: The rise of the experimental essay

89

2.2. The community of practice This was the opinion of the large group of scientists who in the seventeenth century promoted the formation of the Royal Society. In their preliminary meeting on 28 November 1660 the founders of the Royal Society determined that the purpose of their gathering was “a designe of founding a Colledge for the promoting of Physico-Mathematicall Experimentall Learning” (quoted in Boas Hall 1991: 9), and chose the phrase Nullius in verba (‘On the words of no one’) as their motto. The composite nature of this early scientific community is well described by Valle: It consisted of an inner “esoteric” core of committed Baconian experimental philosophers, above all Boyle, Hooke and Desaguliers, whose aim was the creation of new knowledge. [. . . ] This community, however, was defined not in terms of object of study, but according to epistemological and rhetorical norms and practices: how they went about constructing scientific facts. Around this inner circle was a larger concentric zone of the exoteric community, the bulk of the Society: men who were capable of taking an informed interest in what was happening, and of contributing to it on a more minor level. Surrounding this again was a sizeable penumbra of the genteel London public, who went to the weekly meetings of the Society for entertainment, and because it was the fashion, but who had no clear understanding of what they were hearing or witnessing; men, in fact, very like Pepys. (Valle 1999: 111)

These new researchers were convinced that many natural philosophers in the past had been anxious to provide explanations and theories before having enough evidence on which to base them, and therefore emphasized the need for an experimental approach, so as to better facilitate the collection of abundant data from which correct generalizations could be derived. Apart from this emphasis on experimental activity, another important aspect of the new scientific approach consisted in the need for both the procedures and the results of these experiments to be made known to the entire learned world. The publicity given to the work of the members of the Royal Society would further distinguish them from the group of alchemists, a group that considered secrecy to be one of the main characteristics of its research method. The publication of experiments would also have a socializing function, as this exchange of information could promote new professional relationships and strengthen existing links, thus favouring the formation of a new scientific community. Specialized matters were attracting wider interest, especially among aristocratic and cultured people, and this select group of people, who found their proper identification with the newly founded Royal Society, were gradually separating themselves from the less learned group of non-scientific practitioners.

90

Maurizio Gotti

Some of the features characterizing the members of the new scientific community can be found in the following abstract from a letter sent by Gascoines to Oldenburg in which the qualities of Isaac Newton and Francis Line – although presented as adversaries from a theoretical point of view – are extolled: “Therefore in this let us suppose them equal; that they were both great scholars in their kind; great lovers of truth and haters of contest for itself; that both trusted to nothing but to their eyes and experience, nor delivered any thing but what they thought they had truly found” (Gascoines to Oldenburg 15 December 1675; in Rigaud 1965: I, 223). Specialized publications were also used for a proselytizing purpose, as the members of this select group often took advantage of the chance, through their writing, to inform others of the new principles they shared and to gain others’ consensus. Moreover, there was a need to socialize the discoveries made and the new ideas developed, thanks to a collaborative spirit that inspired seventeenthcentury scientists, in contrast to the individualism that characterized philosophers in the Renaissance period. The development of the sciences was now seen as a result of public discussion and knowledge sharing, in the conviction that “In Assemblies, the Wits of most men are sharper, their apprehensions readier, their thoughts fuller, than in their closets” (Sprat 1667: 98).

3. Main purposes of the experimental essay 3.1. Sharing knowledge The main function of the experimental essays was informative, as they were principally written to circulate information about research being carried out in Britain and in other parts of the world. Experiments were usually reported very fully, and even the slightest detail was described, as this would enable the reader to carry out an accurate evaluation of the contents of the essay. Indeed, a feature of experimental essays that immediately strikes the reader is the richness of detail to be found in them. Descriptions are vivid, with an abundance of data that provides a precise and immediate representation of the experience reported. The details are very specific in regard not only to the procedures followed, but also to the equipment used and the times at which certain events occurred. Here is an example: About 11 of the clock in the forenoon we put a frog into a small receiver, containing about 15½ ounces troy weight of water, out of which we had tolerably well drawn the air (so that when we turned the cock under water, it sucked in about 13¼ ounces of water); the frog continued in it (the receiver all the while under water)

A new genre for a specialized community: The rise of the experimental essay

91

lively enough until about 5 of the clock in the afternoon, when it expired. (Boyle [1772] 1965, 3: 358)

The precision of the narration is also visible in the use of very specific terminology appropriate to specialized writings: “The jaws remained mightily opened, and somewhat distorted; the epiglottis with the rimula laryngis, which remained gaping was protruded almost to the farther end of the nether-chap” (Boyle [1772] 1965, 3: 357); and “I had a mind to observe, whether, [. . . ] there would not [. . . ] appear some sudden swelling, greater or less, of the body of the animal, by the spring and expansion of some air (or a¨erial matter) included in the thorax or the abdomen” (Boyle [1772] 1965, 3: 360). The provision of maximal information confirms the main pragmatic goal of the experimental essay, which is to give as many details as possible in the clearest possible way. 3.2. Widening the community The abundance of detail and the precision of the narration of the experimental events may be attributed to the writer’s willingness to provide his readers with as many opportunities as possible to understand his report clearly, not only in order to enable them to come to appropriate conclusions, but also to repeat the same experiment in their own laboratories, and thus prove his results to be more reliable. Indeed, the minuteness of detail and the accuracy of the narration are meant to make repeatability easier and thus encourage the growth of empirical practice in the community of scientists. Boyle confirms the proselytizing aim of experimental essays in his preface to the Proemial Essay, where he encourages his model reader, Pyrophilus, to carry out experimental activity: “I was also hopeful, that the easiness of divers things inviting you to make trial of them, and keeping you from being disappointed in your expectations, the success of your first attempts would incourage you to make trial also of more nice and difficult experiments” (Boyle [1772] 1965, 1: 306). This emphasis on the promotional purpose of the experimenter’s writings betrays the strong feeling of belonging to the new scientific community that animated the members of the Royal Society. This zealous spirit helps to explain why so much detail is provided in Early Modern English experimental essays. Although it led some of them into a sort of prolixity and drudgery, their purpose was to contribute to the advancement of science and not to gain any personal advantage: By the way of writing, to which I have condemned myself, I can hope for little better among the more daring and less considerate sort of men, should you shew them these papers, than to pass for a drudge of greater industry than reason, and fit for little more, than to collect experiments for more rational and philosophical heads to explicate and make use of. But I am content, provided experimental learning be

92

Maurizio Gotti really promoted, to contribute even in the least plausible way to the advancement of it; and had rather not only be an under-builder, but even dig in the quarries for materials towards so useful a structure, as a solid body of natural philosophy, than not do something towards the erection of it. (Boyle [1772] 1965, 1: 307)

3.3. Obtaining official recognition A further reason that justifies the experimenter’s recourse to this detailed narrative technique is his need to acquire official recognition of his results. Indeed, the detailed and accurate description of his personal scientific experience was considered to be one of the requisites for transforming a personal account into an official protocol to be submitted to the broad community of men of science. The careful and objective narration of one’s experiments could provide the materials for proper scrutiny and reliable judgement, and thus permit the transformation of personal results into facts widely accepted by the scientific world. Having obtained the consensus of a wider public in this way, experimental data could become matters of fact and part of scientists’ shared culture. Moreover, it was easier to secure acceptance of experimental results by the scientific community if a more official character was conferred on the experiment itself. This is why the researcher often invited persons of a certain standing to be present while he was carrying out his activities. The presence of these reliable witnesses, explicitly noted in the experimental report, provided official substantiation of the procedures and results narrated. For example, to confirm his experiments concerning respiration, Boyle adds the following statement: For confirmation of which, I have this to alledge, that, having in the presence of some virtuosi provided for the nonce a very small receiver, wherein yet a mouse could live some time, if the air were left in it, we were able to evacuate it at one suck, and by that advantage we were enabled, to the wonder of the beholders, to kill the animal in less than half a minute. (Boyle [1772] 1965, 3: 370)

The confirmation provided by the presence of esteemed and reliable witnesses proved particularly useful in cases in which a novel or expensive apparatus was used to carry out the experiment. Although not able to repeat the experience in his own laboratory, the reader – provided with a detailed account of the events and reassured by the presence of reliable spectators – was nevertheless able to scrutinize the contents of the report, and – by means of this process of “virtual witnessing” (Shapin 1984: 490) – he could make a correct evaluation of them. Robert Boyle himself chose to publish several of his experimental reports in the Philosophical Transactions of the Royal Society of London in order to obtain official recognition of them, and was encouraged to do so by Henry Oldenburg, the editor of this journal. Indeed, as the establishment of priority in discoveries

A new genre for a specialized community: The rise of the experimental essay

93

often led to acrimonious disputes, Oldenburg was fully aware of the importance of obtaining recognition, and this is confirmed by the following extract from a letter to Robert Boyle in which he showed him his willingness to carry out this registering function by publishing his experimental reports: I acknowledge, yt yt yealousy, about the first Authors of Experiments, wch you speak off, is not groundlesse: And therefore offer myselfe, to register all those, you or any person shall please to communicate, as new, wth yt fidelity, wch both of ye honor of my relation to the R. Society (wch is highly concerned in such Experiments) and my owne inclination doe strongly oblige me to. (Oldenburg to Boyle 29 August 1665; in Oldenburg 1965–77: II, 486)

3.4. Establishing stylistic criteria Apart from outlining clear principles of an epistemological nature, the experimental essay also had another purpose, one linked to the stylistic issues connected with the writing of this new genre. Seventeenth-century natural philosophers were not simply interested in widening the community of scientists among the nobler minds of the period, but they also clearly perceived that the causes of differentiation of this group from that of practitioners was not merely methodological and conceptual, but also linguistic and stylistic. Therefore, a gentleman was expected to structure his discourse in an appropriate manner, not only to guarantee a more successful perlocutionary result for his own argumentative text, but also because in that way he could facilitate his interlocutors’ interpretative task. As successful argumentative activity implied the need for people to judge the validity of the various issues, it was very important that the language used in discussions should be clear and readily comprehensible. One of the writers most commonly praised for his style was Boyle. For example, in the following letter, Huygens praises the style adopted by Boyle in his writing activity: “Having yet but perused Mr Boyle’s book, I could not take notice of all the fine things which it contains; but where I have read, I see appear much spirit and modesty, with that ordinary retinue, which keeps him from speaking definitively, as most of your present philosophers do” (Huygens to Moray 14 July 1662; in Rigaud 1965: I, 93). Boyle was often presented as a true gentleman, who showed equilibrium and modesty not only when he made his own personal suggestions, but also when he took various opinions into consideration. Huygens’appreciation of Boyle’s style, however, did not prevent the author of the letter from expressing his criticism, albeit in a civil way and in a modest manner: “but there remains one difficulty with [me] by this experiment, in which I do not see that either Linus’s hypothesis or Mr Boyle’s is satisfactory, that is, why the siphon sticks to the finger, so that

94

Maurizio Gotti

one must use some little force to draw it off ” (Huygens to Moray 14 July 1662; in Rigaud 1965: I, 93). As can be seen, in the expression of criticism the tone was usually humble and polite, in line with the features of a civil style. An important reason for the use of this civil style is the fact that the members of this community knew each other either directly or indirectly. This higher degree of cooperation and esteem was reflected in the frequent use of positive adjectives referring to personal qualities such as celebrated, expert, great, industrious, ingenious, learned, and worthy. The rules of politeness, however, did not prevent people from expressing themselves freely. Indeed, errors or omissions were pointed out clearly, and motivations and explanations for challenged views were requested from the other party. However, even in the most polemic controversies, the tone remained civil and criticism was expressed in an objective way, avoiding a direct attack on the opponent. Indeed, the civility corresponding to a real gentleman’s discourse implied the adoption of a fair attitude towards his interlocutors and respect for the people whose opinions he was arguing against. This behaviour was meant to follow the rules typical of polite conversation among gentlemen, according to which participants speak in an appropriate way, paying compliments even to the people with whom they are disagreeing. This respectful and cooperative attitude is also one of the features that Sprat points out as being typical of the behaviour of the members of the Royal Society: “They could not be much exasperated one against another in their disagreements, because they acknowledge, that there may be several methods of nature, in producing the same thing, and all equally good” (Sprat [1667] 1959: 92). What distinguished a gentleman’s behaviour, therefore, was his respect for the person whose views he was criticising and his limiting his objections to the points he saw as incorrect without any unfair recourse to excessive aggressiveness. According to this criterion, ad hominem argument was considered unacceptable, as criticism should be directed towards the issue under debate rather than the opponent(s). On this subject Boyle writes: “For I love to speak of persons with civility, though of things with freedom” (Boyle [1772] 1965, 1: 312). The civility of a scientist’s behaviour could also be seen from the way in which, even when objecting to a certain methodological supposition, he was grateful to the person who expressed it. Indeed, in the course of the discussion the contestants kept repeating that the purpose of their objections was to clarify their own positions and not simply to quarrel. Their tone was commonly very polite, as can be seen in the use of the “stance marker” (Hyland 1999: 99) I pray in the following quotation: Touching the Solutions given by Mr Newton to the scruples by me propos’d about his Theory of Colors, there were matter to answer them, and to form new difficulties; but seeing that he maintains his opinion with so much concern, I list

A new genre for a specialized community: The rise of the experimental essay

95

not to dispute. But what means it, I pray, that he saith; Though I should shew him, that the White could be produced if only two Un-compounded colors, yet I could conclude nothing from that. (Huygens 1673: PT 3 VIII, 97, 6112)

This objective way of analyzing other scientists’ conclusions and attacking wrong or inconsistent ideas allowed the argumentative paper to be accepted more readily and highlighted its persuasive strength, thus increasing its perlocutionary effect, which consisted in convincing the reader of the validity of the writer’s thesis. The advantages of this approach were clearly perceived by Boyle: And as for the (very much too common) practice of many, who write, as if they thought railing at a man’s person, or wrangling about his words, necessary to the confutation of his opinions; besides that I think such a quarrelsome and injurious way of writing does very much misbecome both a philosopher and a Christian, methinks it is as unwise, as it is provoking. For if I civilly endeavor to reason a man out of his opinions, I make myself but one work to do, namely, to convince his understanding; but, if in a bitter or exasperating way I oppose his errors, I increase the difficulties I would surmount, and have as well his affections against me as his judgment: and it is very uneasy to make a proselyte of him, that is not only a dissenter from us, but an enemy to us. (Boyle [1772] 1965, 1: 312)

4. Main features of the experimental essay 4.1. Brevity As already noted above, one important characteristic of the experimental essay is its shorter length compared with that of a treatise or a dialogue. This conciseness enables the researcher to report even limited experiences without compelling him to describe them at book length. Indeed, several of the seventeenth-century experimental essays consist of reports of single experiments and resemble the rewording, in continuous prose, of the notes that the researcher made in his diaries. The principles of conciseness and of economy of discourse are often pointed out by specialists. Sentences should be as concise as possible with no space given to unnecessary details. Here, for example, is the advice Bacon gives as regards the style of a scientific text: Never cite an author except in a matter of doubtful credit: never introduce a controversy unless in a matter of great moment. And for all that concerns ornaments 3. All quotations from the Philosophical Transactions, which is referred to as PT within the in-text citations in this chapter, are taken from an original copy in the British Library in London.

96

Maurizio Gotti of speech, similitudes, treasury of eloquence, and such like emptinesses, let it be utterly dismissed. Also let all those things which are admitted be themselves set down briefly and concisely, so that they may be nothing less than words. (Bacon [1620] 1968, 4: 254)

These stylistic indications were accepted by subsequent men of science and codified by the Royal Society. In fact, Article IV of Chapter V of the Statutes of the Royal Society (1728) reads: “In all Reports of Experiments to be brought into the Society, the Matter of Fact shall be barely stated, without any Prefaces, Apologies, or Rhetorical Flourishes, and entered so into the Register-Book, by order of the Society.” This preference for a clear, simple style was generally accepted by seventeenth-century scientists. Indeed, from Sprat’s description of the Royal Society’s activity we can see that the way its members wrote was in keeping with the principles set out by Bacon: They have therefore been most rigorous in putting in execution, the only Remedy, that can be found for this extravagance: and that has been, a constant Resolution, to reject all the amplifications, digressions, and swellings of style: to return back to the primitive purity, and shortness, when men deliver’d so many things, almost in an equal number of words. They have exacted from all their members, a close, naked, natural way of speaking; positive expressions; clear senses; a native easiness; bringing all things as near the Mathematical plainness, as they can: and preferring the language of Artizans, Countrymen, and Merchants, before that, of Wits, or Scholars. (Sprat [1667] 1959: 113, emphasis in the original)

However, the criterion of conciseness is subordinate to the higher principle of clarity of exposition, which is considered the most important in terms of perlocutionary value. Indeed, the principle of the avoidance of verbosity so often maintained by Sprat and the other members of the Royal Society is also neglected whenever the author runs the risk of being too brief and therefore unclear. This hierarchy of criteria is confirmed by the following passage, in which Boyle apologizes for the impression of verbosity that some parts of his essays might give: I have knowingly and purposely transgressed the laws of oratory in one particular, namely, in making sometimes my periods or parentheses over-long: for when I could not within the compass of a regular period comprise what I thought requisite to be delivered at once, I chose rather to neglect the precepts of rhetoricians, than the mention of those things, which I thought pertinent to my subject, and useful to you, my reader. (Boyle [1772] 1965, 1: 305)

A new genre for a specialized community: The rise of the experimental essay

97

4.2. Lack of assertiveness A second feature of this genre is that there is no need for the author to arrive at definite conclusions or to systematise the results obtained; the data are to be reported as they are observed, without the writer being required to accompany them with hypotheses or comments. This allows the researcher to report all the details of his experimental activity, even those that he might not be able to explain, thus reducing his theoretical responsibilities and the risk of being criticized, as he, “having for the most part the liberty to leave off when he pleases, is not obliged to take upon him to teach others what himself does not understand” (Boyle [1772] 1965, 1: 303). These experimental accounts, however, provide the identification of useful superstructures on which other scientists might be able to erect appropriate theories. The same principle enables writers also to report experiments that have been unsuccessful, as the analysis of these experiences might help the reader to avoid making the same mistakes as those reported or could enable him to draw interesting conclusions. In promoting the reporting of even unsuccessful experiments seventeenth-century scientists were also following Bacon’s teachings: “No one should be disheartened or confounded if the experiments which he tries do not answer his expectation. For although a successful experiment be more agreeable, yet an unsuccessful one is oftentimes more instructive” (quoted in Hacking 1983: 247). A feature that clearly stands out in reading Early Modern English experimental essays is the frequent use of the narrating technique compared with the very limited space given over to the author’s reflections. Also, the introduction to the experimental essay is usually very short, and briefly outlines the purpose of a series of connected experiments. This confirms the non-argumentative, and mainly informative, purpose of the genre. 4.3. Perspicuity As regards the way experimental essays should be written, the prevailing opinion is that authors should adopt a philosophical rather than a rhetorical style – that is, a style that does not coincide with the traditional way of writing, one typical of literary and philosophical works. Underlining a need widely felt by the contemporary world of science, Boyle stresses the fact that the expressions used in experimental writings should be “rather clear and significant, than curiously adorned” (Boyle [1772] 1965, 1: 304). This choice of a more referential language is due to the very function of the paper, which is to provide information in as clear a way as possible: “And certainly in these discourses, where our design is only to inform readers, not to delight or persuade them, perspicuity ought to be esteemed at least one of the best qualifications of a style” (Boyle

98

Maurizio Gotti

[1772] 1965, 1: 304). This is the reason why we find a strong condemnation of metaphors in several scientific texts of this period, as metaphors were usually seen as deceitful devices. Here, for example, is Hobbes’s opinion: “In Demonstration, in Councell, and all rigourous search of Truth, Judgement does all; except sometimes the understanding have need to be opened by some apt similitude; and then there is so much use of Fancy. But for Metaphors, they are in this case utterly excluded. For seeing they openly professe deceipt; to admit them into Councell, or Reasoning, were manifest folly” (Leviathan, quoted in Jones 1951: 81). The scientific community is convinced that the use of unnecessary rhetorical devices in specialized literature can hinder the comprehensibility of the text; with the aid of the following comparison, Boyle depicts the negative consequences of such devices on the informative value of the paper itself: And to affect needless rhetorical ornaments in setting down an experiment, or explicating something abstruse in nature, were little less improper, than it were (for him that designs not to look directly upon the sun itself) to paint the eyeglasses of a telescope, whose clearness is their commendation, and in which even the most delightful colours cannot so much please the eye, as they would hinder the sight. (Boyle [1772] 1965, 1: 304)

As can be seen from this passage, rhetorical devices are not to be banned from scientific texts completely, but they should be avoided when unnecessary, or when they do not improve the illustrative value of the paper but, on the contrary, only make the text more confusing and more difficult to understand. On the other hand, Boyle is also careful to note the opposite risk, that is, the adoption of a dull mode of expression: “For though a philosopher need not be sollicitous, that his style should delight its reader with his floridness, yet I think he may very well be allowed to take a care, that it disgust not his reader by its flatness, especially when he does not so much deliver experiments or explicate them, as make reflections or discourses on them” (Boyle [1772] 1965, 1: 304–305). And, to reinforce his point, he makes use of a very effective simile: “Thus (to resume our former comparison) though it were foolish to colour or enamel upon the glasses of telescopes, yet to gild or otherwise embellish the tubes of them, may render them more acceptable to the users, without at all lessening the clearness of the object to be looked at through them” (Boyle [1772] 1965, 1: 305). Boyle also condemns the habit that certain scientists have of using cryptic language so as not to make their discoveries comprehensible to their readers. He maintains, on the contrary, that all works should be written in clear language so that all readers can decode the contents and thus improve their knowledge of the subject. This universal sharing of single discoveries is considered essential for the formation of a specialist community and for the progress of scientific

A new genre for a specialized community: The rise of the experimental essay

99

thought. This principle justifies the harsh criticism of all those who wilfully use obscure language and who, therefore, do not deserve to be admitted into the select society of men of science. On several occasions Boyle mentions a close relationship between obscure language and obscure thought. As the clearest example of this criticism he cites “alchemists,” whom he attacks in The Sceptical Chymist for “their obscure, ambiguous, and almost aenigmatical way of expressing what they pretend to teach, that they have no mind to be understood at all, but by the sons of art” (Boyle [1772] 1965, 1: 460). Moreover, “they deliver their hypotheses as darkly as their processes; and it is almost as impossible for any sober man to find their meaning, as it is for them to find their elixir” (Boyle [1772] 1965, 1: 470–1). In the same work, Boyle suggests another close relationship, establishing an equation between the author’s clarity of language and his honesty of behaviour. As an example of this questionable behaviour he once again quotes alchemists, whom he accuses of using obscure language to mask the inaccuracies and unreliability of their principles: “[They] write thus darkly, not because they think their notions too precious to be explained, but because they fear, that if they were explained, men would discern, that they are far from being precious” (Boyle [1772] 1965, 1: 521). Boyle’s conclusion is that “when they pretend to teach the general principles of natural philosophers, this equivocal way of writing is not to be endured” (Boyle [1772] 1965, 1: 521). The solution proposed is that “judicious men, skilled in chymical affairs, shall once agree to write clearly and plainly of them, and thereby keep men from being stunned, as it were, or imposed upon by dark or empty words” (Boyle [1772] 1965, 1: 522). This need for clarity also applies to the use of specialized terminology. One accusation frequently made against scientists by non-specialists is the obscurity of the terms used, many of them being either new or adapted from foreign languages. A very balanced position on the subject is adopted by the majority of the members of the Royal Society, who condemn the unjustified use of foreign loans or strange terms and therefore try to avoid them when other English words with the same meaning are available. But they are also aware that recourse to specific terminology is at times unavoidable, as certain specialized concepts require new names that are not available in English. This is the reason why the use of ambiguous terminology was considered unacceptable, as it was perceived as a serious obstacle to correct argumentation that could render communication among scientists impossible. This terminological issue was central to scientific procedures, as the obscure use of language on the writers’ part would thus not only prevent their being understood, but also prevent their being fully accepted into the scientific community. Indeed, criticism of ambiguous language was commonly conveyed in very strong terms,

100

Maurizio Gotti

and was often expressed in the metaphorical equivalence between obscurity and falseness. This is the reason why writers so often appeared keen to deal with matters concerning terminology and expression, and it also explains the authors’ wishes to use clear terms and to specify their meanings, particularly when the terms were invented anew. It is, however, true that when redefining the properties of already studied compounds or facts, a scientist would frequently continue to employ existing vocabulary in spite of the fact that the use of an old term to refer to a new concept would create a mismatch between the existing signans and its new signatum. Boyle himself followed this practice. Indeed, when reporting his pneumatic discoveries in New Experiments he specified the concept of a totally void space by making use of the existing term vacuum, which was commonly employed by the English-speaking community to refer to the amount of air present in an empty container. However, he tried to avoid problems of ambiguity by specifying the context of reference to which his terms should be attributed. For example, in using the word vacuum, which many of his contemporaries continued to understand in terms of its traditional metaphysical meaning, Boyle was careful to specify the physical significance he conferred on it: “By which I here declare once for all, that I understand not a space, wherein there is no body at all, but such as is either altogether, or almost totally devoid of air” (Boyle [1772] 1965, 1: 10). 4.4. Objectivity Another quality that, it was said, should be reflected in the style adopted by the experimental writer was his honesty.This quality requires the researcher to report events faithfully and sincerely, and to express his opinions and conclusions with the degree of positiveness corresponding to the certainty of the facts described. Confirmation of this attitude can be found in the following quotation from Boyle: Perhaps you will wonder, Pyrophilus, that in almost every one of the following essays I should speak so doubtingly, and use so often, perhaps, it seems, it is not improbable, and such other expressions, as argue a diffidence of the truth of the opinions I incline to, and that I should be so shy of laying down principles, and sometimes of so much as venturing at explications. But I must freely confess to you, Pyrophilus, that having met with many things, of which I could give myself no one probable cause, and some things, of which several causes may be assigned so differing, as not to agree in any thing, unless in their being all of them probable enough; I have often found such difficulties in searching into the cause and manner of things, and I am so sensible of my own disability to surmount those difficulties, that I dare speak confidently and positively of very few things, except of matters of fact. (Boyle [1772] 1965, 1: 307)

A new genre for a specialized community: The rise of the experimental essay

101

This careful attitude prompts the author to use the various modal expressions that English offers to suit the different degrees of certainty of the facts reported and to adopt verbs like to seem and to appear to report with caution the actions as he perceives them, as can be seen in the following quotation: “His body seemed to be perpendicular to the horizon. [. . . ] We perceived him to lie stark dead with his belly upwards. [. . . ] The other frog [. . . ] seemed to be distressed. [. . . ] She appeared to be very much disquieted” (Boyle [1772] 1965, 3: 359). The same prudence is shown by hedging expressions, usually placed in parentheses: “The frog was perfectly alive, and continued to appear so (if I am not mistaken) near an hour” (Boyle [1772] 1965, 3: 359); “She seemed at first (which yet I am not too confident of, upon a single trial) to continue well somewhat longer than a hen in her condition would have done” (Boyle [1772] 1965, 3: 355); and “Such an inflation (though not great) we thought we observed; but until farther trial, I dare not acquiesce in it” (Boyle [1772] 1965, 3: 360). Hedging expressions were often used also to soften the tone of the divergence of opinions: When Mr. Line has tryed this, I could wish, he would proceed a little further to try that which I call’d the Experimentum Crucis, seeing (if I mis-remember not) he denies that as well as the other. For when he has tryed them (which by his denying them, I know he has not done yet as they should be tryed) I presume he will rest satisfied. (Newton 1676: PT X, 121, 502)

In taking this cautious attitude, the experimenter not only shows his professional correctness in that he reports actions as he perceives them with the appropriate degree of certainty, but he also offers a picture of himself as a reliable and faithful witness to the events that he is reporting. Moreover, in order to make his narration even more reliable, he carefully inserts the testimony of his collaborators: “I could not, I say, discern the difference to amount to above, if so much as an hair’s breadth; and the chief operator in the experiment professed, that, for his part, he could not perceive any difference at all” (Boyle [1772] 1965, 3: 362–3). Although the narration mainly reports the events in an objective way, the essay also includes mention of the experimenter’s reactions, but these are usually expressed in brief hints, often relegated to parenthetical phrases: “Whilst the air was drawing out, the lesser frog skipped up and down very lively, and, somewhat to our wonder, clambered up several times to the sides of the receiver” (Boyle [1772] 1965, 3: 358). “Evidentiality” (Chafe 1986: 271) was usually attained by means of observation and perception, two processes that were deemed basic and preliminary to induction (Taavitsainen 2001). The actions regarding observation and perception were usually expressed by verbs having a first-person pronoun subject:

102

Maurizio Gotti

I found, notwithstanding all my care to preserve the Vessels, when I was freeing them, as heedfully as I could, from the supposed Parenchyma, that in every breach, I made, either with my fingers or otherwise, all my endeavours were destructive to my purpose; and if, upon examination of those bits, much of which is called Parenchyma, I met in them more Vessels, than I had preserved in the parts whence they came: And though the Portion were never so small, yet my bare eye could make this discovery; much more could I, when assisted by a Microscope, perceive, I had destroyed more Vessels, than preserved, in despite of the exact care, I was capable to use. (King 1666: PT I, 18, 316)

Another important principle followed in the writing of experimental essays was that of reproducing in the writing the same distinction correctly adopted by the scientist between the setting out of the facts observed and his considerations of them. Indeed, Boyle suggests leaving ”a conspicuous interval” (Boyle [1772] 1965, 1: 2) on the page between the two textual parts (the report of experimental findings and the reflections on them) so as to show the methodological procedure adopted while also underlining the rhetorical and pragmatic differences between those two parts.

5. Syntactic features of experimental essays The seventeenth century showed interesting developments in the syntactic features of scientific English. However, unlike the lexical field, changes took place in a non-explicit way, as scientists were usually unaware of the syntactic modifications that they were introducing into the language by means of their writings. The new perspective that characterized their research method conflicted with the constraints of the language and forced them to adapt its rules to their expressive needs. As Halliday’s (1988) analysis of extracts from Newton’s treatise on Opticks (written in 1675–87 and published in 1704) shows, certain syntactic changes took place in that period. An analysis of the following two theorems formulated by Newton may help in the identification of the main features of such changes: 1. The Excesses of the Sines of Refraction of several sorts of Rays above their common Sine of Incidence when the Refractions are made out of divers denser Mediums immediately into one and the same rarer Medium, suppose of Air, are to one another in a given Proportion. 2. The proportion of the Sine of Incidence to the Sine of Refraction of one and the same sort of Rays out of one Medium into another, is composed of the Proportion of the Sine of Incidence to the Sine of Refraction out of the first Medium into any third Medium, and of the Proportion of the Sine of Incidence to the Sine of

A new genre for a specialized community: The rise of the experimental essay

103

Refraction out of that third Medium into the second Medium. (Newton [1704]: Experiment 8)

As we can see from this passage, sentences are quite long, with very lengthy noun phrases preceding and following the verb. Moreover, the verb (exemplified in this case by are and is composed of ) starts assuming the copular function usual in modern scientific English, where the verb merely links the very long nominal phrases coming before and after it. Sentences start becoming structurally simple, with few or no subordinate clauses, complying with the modern preference for coordination rather than subordination in sentence structure. If we examine another sentence from the text, other interesting details may be noted: “Now those Colours argue a diverging and separation of the heterogeneous Rays from one another by means of their unequal Refractions.” In this quotation there are a large number of words referring to processes (such as diverging, separation, and refraction). This indicates a distinct preference for the use of nouns deriving from verbs. The increase in the use of nominalization was part of a gradual tendency towards a loss of importance of the verb, compensated by a growth of importance of the noun. This change was mainly for textual reasons, as the process of nominalization enabled the scientist to include more information within a single sentence and guaranteed a better flow of discourse. Examples of the use of deverbal nouns are also frequently found in experimental essays: “We [. . . ] observed that divers violent convulsions [. . . ]” (Boyle [1772] 1965, 3: 360); and “As soon as the Cooks hand was thrust into the water, it began to have a glimmering;” (Beale 1666: PT I, 13, 227). In seventeenth-century experimental essays we can also see the emergence of a prevalence of pre-modifying constructions over post-modifying ones; the former make up a prevalent feature of modern specialized discourse (see Gotti 2008: 73) as they allow for the construction of more complex sentences. Moessner’s (2006: 67) analysis of 50 of Boyle’s experimental essays gives the following results: 62.73% pre-modifiers versus 37.26% post-modifiers. The use of pre-modification offers advantages in terms of greater textual conciseness, as does the use of non-finite forms, which were often employed to avoid further coordinated or subordinated clauses. The following examples show how non-finite forms were frequently used in experimental essays: These glasses being conveyed into a fit receiver, and the air being leisurely pumped out, and somewhat slowly readmitted, the numerous bubbles that had ascended during the operation, constituted at the top an a¨erial aggregate, mounting to 8 /10 , wanting about an hundredth part of an inch. Presently after the tube (by and by to be described) was filled again with the same water, and inverted; and the water being drawn down to the surface of the vesseled water, and the air let in again,

104

Maurizio Gotti

the water was impelled up to the very top, within a tenth and half a tenth of an inch. (Boyle [1772] 1965, 3: 361, emphasis added) The water, till it was stirr’d, gave no light, but was thick and dark, as we saw by day-light, and by candle-light. [. . . ] but being gently stirr’d by the hand moving round [. . . ] it did so shine [. . . ]. (Beale 1666: PT I, 13, 227, emphasis added)

This frequent use of the processes of nominalization and pre-modification in experimental essays is in line with Biber and Finegan’s (1997) observation that specialist written registers developed from being less to more informative between the seventeenth and the twentienth centuries. Another syntactic feature of experimental essays – which is meant to guarantee maximum comprehension – is the adoption of simple verb forms and sentence constructions. Indeed, the voice commonly used in experimental essays is active, often putting the researcher in a thematic position and in the grammatical function of the subject, so as to emphasize his active role in the experimental activity: We put a full grown duck [. . . ] into a receiver [. . . ]; then pump[ed] out the air, [. . . ] she appeared much discomposed; [. . . ] from which we presently rescued her by letting in the air upon her; [. . . ] we soon after included the same bird in the same receiver, and having by a special way cemented it on very close, we suffered her to stay thus shut up with the air for five times as long as formerly (by our guess, helped by a watch) without perceiving her to be discomposed. (Boyle [1772] 1965, 3: 355–356)

The active presence of the scientist is emphasized not only when he is performing an operation, but also when the behaviour of the animal or object used in the experiment is being described. In many cases, indeed, the behaviour of the animal is framed by, and even secondary to, the scientist’s experience of perceiving the behaviour, in order to underline the fact that what is being narrated is not happening spontaneously but only as a result of the experimenter’s actions. This framing continues to make the experimenter’s role central in all the parts of the report: “I perceived some little motions, which made me conclude him alive” (Boyle [1772] 1965, 3: 360). Although the active voice is quantitatively prevalent in the narration, there is also the appropriate use of the passive form. This is usually employed for specific cases, such as the underlining of unexpected results, as if to diminish the scientist’s responsibility in those cases. Here are some examples: “This duck being reduced, in our receiver, to a gasping condition [. . . ]” (Boyle [1772] 1965, 3: 356); and “His jaws, which were formerly shut, gaped exceeding wide, as if they had been stretched open by some external violence” (Boyle [1772] 1965, 3: 358). A confirmation of the close connection between the use of the passive and the attribution of the responsibility for an

A new genre for a specialized community: The rise of the experimental essay

105

unpleasant consequence to external forces, and not to the experimenter himself, can be seen in the following sentence, in which Boyle prefers to use the passive of to kill rather than the active of the verb to die: “She continued, by our estimate, above two hours and half in the exhausted receiver without giving clear proof of her being killed” (Boyle [1772] 1965, 3: 357). Another case in which the passive is frequently used is in reporting how certain procedures (usually involving the utilization of equipment) have been carried out.The shift from the active to the passive voice in these cases underlines the passing from the active role of narration of events to the description of procedures that are becoming standardized in the experimenter’s repertoire or in that of scientists in general: A chemical pipe sealed at one end, and 36 inches, or somewhat less, in length, was filled with water, and inverted into a glass vessel, not two inches in diameter, but 1/4 of an inch, or little more in depth. [. . . ] Presently after the tube (by and by to be described) was filled again with the same water, and inverted; and. . . the water was impelled up to the very top, within a tenth and half a tenth of an inch. (Boyle [1772] 1965, 3: 361)

The confirmation that the actions described here are not novel, but rather a standardized procedure comes from the text itself, as the author clearly refers to them as “matter of fact” (Boyle [1772] 1965, 3: 362).4 Indeed, when he comes to underlining the originality of the experimental events he is describing in the following paragraphs, the writer resorts to the use of the active form: “We provided a clear round glass [. . . ]. We conveyed the glass [. . . ]” (Boyle [1772] 1965, 3: 362). Experimental accounts often started from the observation of a natural phenomenon that had aroused the curiosity and intervention of the researcher. It is interesting to note that these introductory sections usually contain passive verbs, a form that correlates well with the researcher’s rather passive attitude in this phase, as his role is limited to the observation of events taking place without his direct intervention: “[. . . ] fresh Mackrels were boyl’d in Water, with salt and sweet herbs; [. . . ] the Mackrels were left in the Water for pickle. [. . . ] more fresh Mackrels were boyl’d in like Water;” (Beale 1666: PT I, 13, 226). When instead the experimental part is reported, verbs are usually in the active form with the frequent use of I /we subjects: [. . . ] we repeated the same Trial, and found the same effects. [. . . ] I took a piece that shin’d most, and fitted it as well as I could devise in the night, both to my 4. More examples of passive forms used in Boyle’s papers to report standardized procedures can be found in Dear (1985).

106

Maurizio Gotti

great Microscope, and afterwards to my little one; but I could discern no light by any of these Glasses; [. . . ] I examin’d, in my great Microscope. [. . . ] We numbred them, and agreed in the number, order and place. (Beale 1666: PT I, 13, 227)

6. Conclusion The analysis carried out in this paper has shown the main functions performed by the experimental essay, consequential to the great epistemological and methodological innovations which were taking place in seventeenth-century England. As has been seen, in the early phases of the formation of a specialized community, the experimental essay played a fundamental role in the dissemination of scientific communication. However, the communicative role of this writing activity was not limited merely to the informative aspect. The experimental essay also fulfilled other important goals linked to socialization purposes, favouring the creation of a spirit of solidarity among the members of a new social group sharing innovative intellectual interests and professional practices. The essays published by this new specialized discourse community also served a stylistic purpose, as the metatextual considerations expressed in them helped to create new discursive practices. The analysis has also demonstrated that although this new genre has evident links with its traditional existing model, the essay in Montaigne’s style, it cannot be considered the result of “secondary generic productivity” (Fishelov 1999: 56), as the new genre has not merely been conceived as an alteration of the existing genre, but as a clear distinction and innovation compared to it, endowed not only with new pragmatic functions and stylistic features but also with a specific name to denote it. Indeed, Boyle first used the term, “experimental essay,” in his seminal work Proemial Essay [. . . ] with Some Considerations Touching Experimental Essays in General (1661), which also contains fundamental metatextual indications on the functions and features of this new genre. The dynamic process that has led to the origin of experimental essays can thus be awarded the status of primary generic productivity,5 as it represents a clear case of “constitutive intertextuality” or “interdiscursivity” (Fairclough 1992: 117), deriving from a conscious design to create a novel construct by exploiting available generic resources within a new form that is more suited to the new stylistic and pragmatic conventions needed. The publication of experimental essays greatly contributed to the advancement of specialized disciplines as it favoured the establishment of widely shared 5. See also Moessner (2006) for a similar conclusion.

A new genre for a specialized community: The rise of the experimental essay

107

experience and knowledge within the scientific community, providing it with a common empirical and methodological approach. Moreover, the study of the texts taken into consideration here has pointed out a remarkable degree of consistency in the various functions of the experimental essay in this period, as all the various aspects observed – ethical, methodological, and stylistic – greatly contributed to the creation of the new patterns of scientific communication required by the specialized community that had formed in England in the seventeenth century. In this way, the experimental essays provided a fundamental contribution to the evolution of the community formed by the members of the Royal Society, who thus were able to characterise themselves not merely as a sociolinguistic group but as a “sociorhetorical” one (Swales 1990: 24), endowed with its own generic and stylistic practices. The experimental essay was widely accepted by the scientific community of the seventeenth century. The various advantages it offered caused it to become widespread, and its popularity became even vaster with the increase in circulation of scientific journals, which mainly consisted – apart from letters and reviews – of brief reports of experimental activity. The Early Modern English model of experimental essays greatly influenced the subsequent realizations of this genre (see Bazerman 1988). To appreciate its importance, we should consider the fact that, in the following centuries, while some forms of specialized writings – such as the dialogue – almost disappeared, the experimental essay survived and became an essential part of specialized literature. The rapid diffusion of scientific journals made the experimental essay an established genre commonly used by men of science. In the course of time the experimental essay has certainly evolved and has found a more cohesive structure and more welldefined stylistic norms. Important changes have taken place, such as a shift from an author-centred to an object-centred rhetoric, expressed linguistically in moreor-less involved and verbal terms (see Atkinson 1999). From the syntactic point of view, the centuries following the seventeenth century have also developed and consolidated the features identified in experimental essays, with an increase in the processes of nominalization and pre-modification of specialized discourse. However, many of its early features have remained and testify to the seminal role played by the reflections and writing practices of the discourse community constituted by the members of the Royal Society in the widespread adoption of this important specialized genre.

108

Maurizio Gotti

References Atkinson, Dwight 1999 Scientific Discourse in Sociohistorical Context: The Philosophical Transactions of the Royal Society of London, 1675–1875. (Rhetoric, Knowledge, and Society.) Mahwah, NJ: Lawrence Erlbaum. Bacon, Francis [1620] 1968 Novum Organum. In: James S. Spedding, Robert Leslie Ellis and Douglas Denon Heath (eds.), The Works of Francis Bacon. 14 Vols. London: Longmans; rpt. New York. Bazerman, Charles 1988 Shaping Written Knowledge. (Rhetoric of Human Sciences.) Madison: The University of Wisconsin Press. Bhatia, Vijay K. 2004 Worlds of Written Discourse: A Genre-based View. (Advances in Applied Linguistics.) London: Continuum. Biber, Douglas and Finegan, Edward 1997 Diachronic relations among speech-based and written registers in English. In: Terttu Nevalainen and Leena Kahlas-Tarkka (eds.), To Explain the Present. Studies in the Changing English Language in Honour of Matti Rissanen, 253–75. (M´emoires de la Soci´et´e N´eophilologique de Helsinki 52.) Helsinki: Soci´et´e N´eophilologique. Boyle, Robert [1772] 1965 The Works, Ed. Thomas Birch, 6 Vols. London: J. & F. Rivington, rpt. Hildesheim: Georg Olms. Boas Hall, Marie 1965 Robert Boyle on Natural Philosophy. Bloomington: Indiana University Press. Boas Hall, Marie 1991 Promoting Experimental Learning: Experiment and the Royal Society 1660–1727. Cambridge: Cambridge University Press. Chafe, Wallace 1986 Evidentiality in English conversation and academic writing. In: Wallace Chafe and Johanna Nichols (eds.), Evidentiality: The Linguistic Coding of Epistemology, 261–312. Norwood: Ablex. Dear, Peter 1985 Totius in verba: Rhetoric and authority in the Early Royal Society. Isis 76: 145–161. Diller, Hans-J¨urgen and Manfred G¨orlach (eds.) 2001 Towards a History of English as a History of Genres. (Anglistische Forschungen 298.) Heidelberg: Winter. Fairclough, Norman 1992 Discourse and Social Change. Cambridge: Polity Press.

A new genre for a specialized community: The rise of the experimental essay

109

Fishelov, David 1999 The birth of a genre. European Journal of English Studies 3: 51–63. Gotti, Maurizio 2008 Investigating Specialized Discourse. Bern: Peter Lang. Hacking, Ian 1983 Representing and Intervening. Cambridge: Cambridge University Press. Halliday, Michael A.K. 1988 On the language of physical science. In: Mohsen Ghadessy (ed.), Registers of Written English, 162–177. (Open Linguistics Series.) London: Pinter. Hunter, Michael 1989 Establishing the New Science:The Experience of the Early Royal Society. Woodbridge: Boydell Press. Hyland, Ken 1999 Disciplinary discourses: Writer stance in research articles. In: Christopher N. Candlin and Ken Hyland (eds.), Writing: Texts, Processes and Practices, 99–121. (Applied Linguistics and Language Study Series.) London: Longman. Jardine, Lisa 1999 Ingenious Pursuits: Building the Scientific Revolution. London: Little/ Brown. Jones, Richard Foster 1951 The Seventeenth Century. Stanford: Stanford University Press. Kohnen, Thomas 2001 On defining text types within historical linguistics: The case of petitions/statutes. In: Lilo Moessner (ed.), Early Modern English Text Types, 197–203. Monographic issue of European Journal of English Studies 5. Lave, Jean and Wenger, Etienne 1991 Situated Learning: Legitimate Peripheral Participation. (Learning in Doing: Social, Cognitive & Computational Perspectives Series.) Cambridge: Cambridge University Press. Moessner, Lilo (ed.) 2001 Early Modern English Text Types. Monographic issue of European Journal of English Studies 5. Moessner, Lilo 2006 The birth of the experimental essay. In: Vijay Bhatia and Maurizio Gotti (eds.), Explorations in Specialized Genres, 59–77. (Linguistic Insights. Studies in Language and Communication 35.) Bern: Peter Lang. Newton, Isaac 1704 Opticks: Or a Treatise of the Reflections, Refractions, Inflections and Colours of Light. London.

110

Maurizio Gotti

Oldenburg, Henry 1965–77 The Correspondence of Henry Oldenburg. Ed. A. Rupert Hall and Marie Boas Hall, 11 Vols. Madison: University of Wisconsin Press. Paradis, James 1987 Montaigne, Boyle, and the essay of experience. In: George Levine (ed.), One Culture: Essays in Science and Literature, 59–91. (Science & Literature.) Madison: The University of Wisconsin Press. Rigaud, Stephen J. 1965 Correspondence of Scientific Men of the Seventeenth Century. 2 Vols. Hildesheim: Georg Olms Verlag. Shapin, Steven 1984 Pump and circumstance: Robert Boyle’s literary technology. Social Studies of Science 14: 481–520. Shapiro, Barbara 2000 A Culture of Fact: England, 1550–1720. Ithaca: Cornell University Press. Sprat, Thomas [1667] 1959 A History of the Royal Society of London for the Improving of Natural Science. London: Martyn, rpt. London: Routledge & Kegan Paul. Swales, John 1990 GenreAnalysis (The CambridgeApplied Linguistics Series.) Cambridge: Cambridge University Press. Swales, John M. 1998 Other Floors, Other Voices: A Textography of a Small University Building. (Rhetoric, Knowledge, and Society.) Mahvah, NJ: Erlbaum. Taavitsainen, Irma 2001 Evidentiality and scientific thought styles: English medical writing in Late Middle English and Early Modern English. In: Maurizio Gotti and Marina Dossena (eds.), Modality in Specialized Texts, 21–52. (Linguistic Insights. Studies in Language and Communication 1.) Bern: Peter Lang. Valle, Ellen 1999 A Collective Intelligence: The Life Sciences in the Royal Society as a Scientific Discourse Community, 1665–1965. (Anglicana Turkuensia 17.) Turku: University of Turku. Vickers, Brian (ed.) 1987 English Science: Bacon to Newton. (Cambridge English Prose Texts.) Cambridge: Cambridge University Press.

Do some genres or text types become more complex than others?1 Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

This chapter reports on an exploration of the structural and syntactic (processing) complexity of some text types or genres in the recent history of English. It concentrates on the complexity of constituents functioning as (non-pronominal) nominal subjects, objects and adverbials in declarative sentences. Data are drawn from a corpus of texts from 1750 to present-day English, namely the British component of ARCHER (A Representative Corpus of Historical English Registers). It has been claimed that unmarked subject and object positions are of particular importance in the determining of complexity, hence of processing. In an attempt to connect linguistic complexity and textual formality, we focus here on two text types, letters and news, which have been subjected to the same experiment. The notion of complexity is understood in the present study to be relational and relative and as such is based on a series of metrics which, on the one hand, measure the degree of complexity of the constituents (size/length, syntactic density, syntactic depth) and, on the other hand, allow us to place genres on a scale of complexification, for the purposes of comparison among genres or text types.

1. Introduction This chapter deals with the connection between linguistic complexity and the evolution of text types, or genres, across time. A few words are thus in order as regards the notions of text type and linguistic complexity to be used in this study. First, while in multifactorial multidimensional studies the labels register, genre, and text type have often been used interchangeably (see, for example, 1. This research was been funded by the Spanish Ministry of Education and Innovation (grant number FFI2009-11274/FILO) and Xunta de Galicia (grant number INCITE08PXIB204016PR), which is hereby gratefully acknowledged. This is part of a broader project on the degree of variation experienced by the English language in its recent history as far as the syntactic complexity of clausal constituents is concerned.

112

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

Biber 1994: 51–53, for the definition of these terms within a multidimensional approach), in this study we will use the term text type when we refer to categories of discourse established on a structural linguistic basis, following other contributions to this volume (see, for example, Virtanen, this volume). Since our approach relies exclusively on the way in which overt linguistic features can be understood as indicators of linguistic variation across time and texts, in what follows we shall avoid conceptual labels such as genre or register (see, for example, Giltrow’s, Mondorf ’s and Virtanen’s papers, this volume, which take the social scene/action or the functional/communicative motivation as the main triggers of textual – or, better, genre – variation). Providing that in what follows we shall undertake the study of textual variants from a strictly linguistic perspective, text type will be preferred to its alternatives mentioned above. Second, whereas the concept of linguistic complexity has been approached from different perspectives (cognitive, informative, syntactic, and lexical; see section 3), both cross- and intra-linguistically, in this chapter we have favoured an intra-linguistic, bidimensional, and multifactorial approach. This is: (i) intralinguistic because the main research goal is to characterize one language in its recent history; (ii) bidimensional because it integrates structural constituency and syntactic processing; and (iii) multifactorial because it is not defined by means of a single variable. The principal assumption underlying this study is that certain (linguistic) aspects within a given language2 can exhibit different degrees of (linguistic) complexity across time, which, in consequence, paves the way for a diachronic characterization of linguistic phenomena in terms of linguistic complexity. In fact, such an intra-linguistic perspective can lead to the study of the linguistic complexity of different clausal constituents, different text types, and different periods of the history of the language under investigation.3 In section 2 we describe the corpus from which the data have been retrieved and justify the rationale of focusing specifically on subject complexity in two text types. In section 3 we describe linguistic complexity both from a theoretical perspective and an empirical perspective. In section 4 we explore the metrics

2. From a cross-linguistic point of view, McWhorter (2001: 127) claims that “it is a truism in linguistics in general that all languages are equally complex” and concludes by saying that one cannot compare languages but only specific linguistic phenomena according to their linguistic complexity. 3. Whereas this chapter is devoted to the connection between diachrony and complexity, P´erezGuerra (2007) concentrates on (synchronic) textual variation in present-day English and, more specifically, on the measurement of the structural and syntactic complexity of the subjects in declarative clauses found in both written-to-be-read and spoken texts.

Do some genres or text types become more complex than others?

113

which will be used as indicators of linguistic complexity, and in section 5 we discuss the results of the application of the metrics to the corpus material. Finally, in section 6 we summarize the findings and suggest lines for further research.

2. The corpus Data for this study were taken from the British component of ARCHER (A Representative Corpus of Historical English Registers). In trying to focus on written-to-be-read (and written-to-be-spoken) texts and informal ([possibly] speech-based) textual material, two text types were analysed, news and letters. The texts analysed belong to the second halves of the eighteenth, nineteenth, and twentieth centuries (1750–1799, 1850–1899, 1950–1990), and the number of words analysed amounts to 109,911, distributed as shown in Table 1. Table 1. Word totals Text-type/period

1750–1799

1850–1899

1950–1990

Total

News Letters

26,147 12,906

23,252 11,215

24,245 12,146

73,644 36,267

Total

39,053

34,467

36,391

109,911

The analysis focuses on unmarked nominal arguments and constituents. A database was created with all the preverbal nominal subjects (external arguments), postverbal nominal objects (internal arguments), and nominal adverbials (nonsubcategorized constituents) of the declarative sentences in the texts under investigation. A large number of non-nominal and pronominal arguments and constituents, however, were disregarded in our study. These included: (i) postverbal subjects and preverbal objects; (ii) pronominal subjects, objects, and adverbials with modification (as in [1]); (iii) prepositional phrases (as in [2]); (iv) coordinative constructions, or conjunction phrases (as in [3]); and (v) clausal constituents (as in [4]): (1)

They all have difficult cases (1968WINNICOTT.X9)

(2)

Between 15,000 and 17,000 Vietnamese will land on Malaysian shores each month for the next six months. (1979STM2.N9)

(3)

Mr. Knox and his Son were at the Execution (1762PUB1.N3)

(4)

Whether Tito, who attended in 1960 when Khrushchev held the stage, will come is still uncertain. (1967STM2.N9)

114

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

As regards the text types chosen here, we believe that news and letters are two textual variants (or text types) within a given period of the history of the language. News writing is here taken as representative of formal language that is written to be published. The text type of letters has been chosen because of its idiosyncrasy, between formal and informal language, between written and oral language, and between private and public language, thus permitting further comparison with other text types that are taken as prototypically representative of these dimensions. With respect to the opposition of formal versus informal language, letters permit the inclusion of informal speech since they do not normally require any kind of literary expertise. As Hopper (1997: 3) puts it, letters are examples of vernacular written narrative, where “‘vernacular’ is meant to suggest that it is the kind of narrative that does not usually find its way into print, and is not held up as a model of prose to be imitated.” As far as the opposition between binomial written versus oral language is concerned, letters occupy an intermediate position between prototypically written text types and other speech-based linguistic discourses. In fact, in a corpus of fifteenth- and seventeenth-century letters, Markus (2001) explores characteristics such as lack of agreement, dangling constructions with unclear referents, omissions and marked word order, all of which are typical of oral language. Biber and Finegan (1997: 75) investigate the linguistic evolution of letters, as well as of other genres, from 1650 to 1990 and corroborate the existence of a “general historical trend towards more ‘involved production’ [. . . ] and more ‘situated reference”’, that is, a drift toward orality. In this respect, Biber (2001: 212), dealing with the functional interpretation of the application of his multidimensional analysis to eighteenth-century letters, claims that “written registers were generally more sharply distinguished from spoken registers in the eighteenth century than they are at present.” Finally, as regards the public versus private dichotomy, Markus (2001: 182) claims that “[t]he language of letters is and, as can be assumed, was as close to the linguistic usage of individuals as can be, perhaps with the exception of diaries.” Since text types are here understood as codifications of linguistic features (Taavitsainen 2001: 141), one can characterize a given text type by exploring its linguistic characteristics. Following, for example, Crain and Shankweiler’s (1988) processing deficit hypothesis, we contend that complexity is influenced by linguistic circumstances and is not inherent to a given clause type. That said, the present study investigates complexity in such text variants by quantifying several linguistic factors, factors which, according to the relevant literature, are determinants of complexity. As noted in section 1, the working hypothesis is that text types can be graded in terms of complexity.

Do some genres or text types become more complex than others?

115

The two text types selected in this study, news and letters, will be analysed for unmarked nominal subjects, objects, and adverbials in unmarked or kernel clauses. We have focused on the subjects partly because their structure has been claimed to have consequences for the syntactic (and cognitive) processing of a sentence (see, for example, Ferreira [1991], who investigates sentential complexity based exclusively on the syntactic complexity of subjects). As an illustration of this, Davison and Lutz (1985: 60) maintain that “the high load of processing would occur in subject position of the target sentence. [. . . ] The syntactic element in this position is what matches or fails to match with information in the context sentence.” Furthermore, in connection with his syntactic prediction locality theory, Gibson (1998: 27) emphasizes the relevance of subjects (or external arguments) to the determination of the processing cost of a sentence; in his words, “modifying the subject should cause an increase in the memory cost for predicting the matrix verb, whereas modifying the object should not cause such an increment.” However, we have also paid attention to the description of unmarked objects and adverbials in an attempt to pinpoint the syntactic and distributional differences between the subjects and other core nominal constituents in the clause.4 More specifically, the null hypotheses here are: (i) that the degrees of linguistic complexity exhibited by the subjects (external arguments), the objects (internal arguments), and the adverbials (non-subcategorized constituents) are alike (syntactic null hypothesis); and (ii) that the degrees of linguistic complexity exhibited by the subjects (preverbal and sentence-initial), the objects (postverbal) and the adverbials (either preverbal or postverbal5 ) are alike (distributional null hypothesis). A final remark is in order concerning the categorial nature of the constituents under analysis. To rule out that the syntactic structure of the segments fulfilling the functions of subject, object, and adverbial interferes with the statistical description of such functional elements, we have concentrated exclusively on nominal categories, that is, noun-phrase based subjects, objects, and adverbials. Whereas Tables 2.1 and 2.2 give the global raw figures and the proportions of pronominal and non-pronominal subjects, objects, and adverbials investigated, in Figure 1 we classify the subjects, objects, and adverbials categorially and offer the relative proportions (nf = normalised frequency per 1,000 words) of the data given in Tables 3.1, 3.2, and 3.3, respectively.

4. In an attempt to establish differences between adverbials and other constituents in terms of processing complexity, Shapiro et al. (1992: 450) maintain that adverbials imply an extra processing cost since they have to be thematically accommodated in the sentence. 5. We leave for further research the detailed analysis of the linguistic complexity of the adverbials according to the position that they occupy in the clause.

116

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

Table 2.1. Distribution of pronominal subjects and objects6 text type \ period News

Letters

1750–1799

1850–1899

1950–1990

Total

subjects

579 nf = 22.15

468 nf = 20.13

534 nf = 22.1

1581

objects

110 nf = 4.2

120 nf = 5.16

77 nf = 3.18

307

subjects

778 nf = 60.29

600 nf = 53.5

712 nf = 58.62

2,090

objects

267 nf = 20.7

208 nf = 18.55

166 nf = 13.67

641

1,734

1,396

1,489

Total

4,619

Table 2.2. Distribution of non-pronominal subjects, objects, and adverbials text type \ period News

Letters

Total

1750–1799

1850–1899

1950–1990

Total

subjects

830 nf = 31.75

968 nf = 41.63

1,078 nf = 44.47

2,876

objects

301 nf = 11.51

649 nf = 27.92

735 nf = 30.32

1,685

adverbials

127 nf = 4.85

66 nf = 2.84

87 nf = 3.58

280

subjects

231 nf = 17.90

192 nf = 17.12

190 nf = 15.65

613

objects

375 nf = 29.10

228 nf = 20.33

265 nf = 21.82

868

adverbials

52 nf = 4.02

40 nf = 3.6

38 nf = 3.13

130

1,916

2,143

2,393

6. The data for pronominal adverbials (such as there, etc.) are statistically insignificant.

6,452

Do some genres or text types become more complex than others?

117

Figure 1. Distribution of subjects, objects, and adverbials Table 3.1. Pronominal and non-pronominal subjects (percentages per text type and period) 1750–1799 non-pron pron

1850–1899 non-pron pron

1950–1990 non-pron pron

Totals

News

830 58.91%

579 41.09%

968 67.41%

468 32.59%

1,078 66.87%

534 33.13%

4,457

Letters

231 22.89%

778 77.11%

192 24.24%

600 75.76%

190 21.06%

712 78.94%

2,703

Totals

1,061

1,357

1,160

1,068

1,268

1,246

7,160

text type

Table 3.2. Pronominal and non-pronominal objects (percentages per text type and period) 1750–1799 non-pron pron

1850–1899 non-pron pron

1950–1990 non-pron pron

Totals

News

301 73.24%

110 26.76%

649 84.40%

120 15.60%

735 90.52%

77 9.48%

1,992

Letters

375 58.41%

267 41.59%

228 52.29%

208 47.71%

265 61.48%

166 38.52%

1,509

Totals

676

377

877

328

1,000

243

3,501

text type

118

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

Table 3.3. Adverbials (percentages per text type and period; pronominality issues considered irrelevant) text type \ period

1750–1799

1850–1899

1950–1990

Totals

News Letters

127 52

66 40

87 38

280 130

Totals

179

106

125

410

Some clarifying remarks are needed here with respect to the categories in which the subjects have been divided in Figure 1 and Tables 3.1–3.3. First, the nonpron(ominal) category only includes (multi- and one-word) noun phrases whose heads are not pronouns. As already mentioned, clauses, coordinative structures, and prepositional phrases functioning as subjects, objects, or adverbials have not been addressed in this chapter. Second, the pron(ominal) column embraces overt (personal, demonstrative, and indefinite) pronominal subjects, objects, and adverbials. For reasons of coherency, we have not paid attention to expletive forms (it, there), null categories in contexts of ellipsis or anaphora (pro), or wh-proforms whose (unmarked) distribution in the sentence is conditioned by word-order requirements. Figure 1 reveals that there are no significant statistical differences in the periods under investigation as far as the relative distribution of subjects, objects, and adverbials is concerned in the periods examined in this chapter. As shown in Tables 3.1–3.3, the subjects in the letters are commonly pronominal and the frequency of pronominal objects is higher than 35 percent throughout the three periods. By contrast, non-pronominal subjects and objects are predominant in the news, where the frequency of pronominal constituents (especially in the case of objects) decreases as we approach contemporary English. The previous findings accord, on the one hand, with the subjective style of the letters, which demands a large number of personal pronouns fulfilling the argument functions of subject and object. On the other hand, the increase of non-pronominal objects can be explained by the fact that the objects which have been considered in this study are postverbal and thus conditioned by principles such as end-weight. These results, taken alone, lead to the conclusion that the subjects in the letters are less elaborate than those in the news and, in consequence, the strategies of subject complexification are more productive in the news text type. However, in what follows we will seek to test this conclusion by examining linguistic complexity in exclusively non-pronominal subjects, objects, and adverbials in order

Do some genres or text types become more complex than others?

119

that such an analysis can either corroborate or refute the conclusion resulting from the observation of pronominal and non-pronominal subjects.

3. Linguistic complexity Briefly, linguistic complexity is here interpreted as a theoretical concept related to both syntactic expansion and phrase-marker configuration, in which the size (Wasow 1997: 94), the weight (Arnold et al. 2000: 35) and the “depth” (Ferreira 1991: 226; Sampson 2001: 47) of the constituents under examination play a significant role. Since this study is based not on sentence structure but exclusively on the syntactic organisation of the subjects, objects, and adverbials, the analysis of syntactic relations such as embedding and subordination will be restricted to the scope of such constituents. However, we do not intend here to address issues such as the degree of complexity relativity among constituents, that is, the parsing differences depending on positioning, since our methodology addresses only to the structure of the grammatical subjects, objects, and adverbials occurring in the corpus. The theoretical assumptions on which this study is based are as follows. First, we contend that the process of decoding and interpreting a message is unambiguous on the hearer/reader’s part. In other words, we assume that the human parser is able to interpret correctly the linguistic input and thus discard any sort of processing ambiguity. As Carlson (2002: 8) points out, this assumption fits a “garden-path” approach to language rather than a constraint-satisfaction theory approach that subsumes the existence of parallel processings of the same utterance to which an optimality model must be applied. When we analyse a database subject, object, or adverbial, we stick to only one syntactic interpretation of the constituent under study, one which will be regarded as the optimal interpretation. Second, complexity is understood as a relational notion; put differently, something is always less/more complex than something else, and in consequence, one can compare, for example, the behaviour of a given textual variant in a certain period of the history of a language in terms of complexity. That stated, the hypothesis is that texts can be diachronically graded in terms of complexity. In this respect, we have selected here two representative text types of the written and the speech-based dimensions, which will be analysed in the same way. Third, complexity is a relative concept. As Frazier (1988: 204) puts it, “there is no general unit of complexity (defined either in terms of time or number of computational operations) which would permit us to predict in ‘absolute’ terms

120

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

the complexity of a sentence.” Our account of complexity, which will be based on not one but several metrics, will inevitably be partial. Fourth, syntactic or structural complexity is not associated with linguistic richness or McWhorter’s (2001) “ornamentation.”7 McWhorter’s concept of ornamentation is applicable to variation across languages – either unrelated dialects or dialects derived from the same source – and has not been applied to the study of variation of a given language or dialect in a short time period. The notion of complexity as functional richness is often coupled with linguistic explicitness or transparency. In this vein, Rohdenburg (1996: 151) claims that “[i]n the case of more or less explicit grammatical options the more explicit one(s) will tend to be favored in cognitively more complex environments.”8 In the same vein, Szmrecsanyi’s (2009) and Szmrecsanyi and Kortmann’s (forthcoming) metrics for analyticity, syntheticity and grammaticity measure redundancy in a language, the assumption being that increased frequencies of grammatical markers imply higher redundancy and hence lower output economy and increased complexity. From this perspective, grammatical systems with many overt functional options will be more complex since they comprise more ornamentation and are more transparent. This is because the different functional meanings are overtly materialized in those languages. Fifth, syntactic or structural complexity is not linked to conceptual complexity (animacy, etc.) and/or informative complexity (givenness versus newness). Conceptual complexity is strongly tied to the referentiality of the entities conveyed by the referring categories (mainly nouns). Conceptual (as well as syntactic) complexity is covered by, for example, Gibson’s (1998, 2000) dependency locality theory, according to which the measure of the integration cost of a constituent involves counting, in broad outline, the number of new discourse referents that have to be processed so that the constituent under examination can be integrated in the syntactic structure, see Warren and Gibson (2002: 86–87) for 7. In McWhorter’s (2001: 135) words, “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.” Such a concept of ornamentation, in line with Hawkins’ (2004: 38) Minimize Forms principle, is also based on the reduction not only of linguistic forms but also of the functional properties associated to them. Yet it does not necessarily coincide with Dahl (2004: 43), who contends, for instance, that more vocabulary and less phonetic reduction, that is, more phonetic weight, implies more complexity, independently of whether such lexical and phonetic richness is functional in the language. 8. To give an example of complexity as linguistic explicitness, Rohdenburg (1996) maintains that (i) is more explicit and thus more complex than (ii) since the infinitival particle to in (i) marks write the paper as an embedded nonfinite clause: (i) I help him to write the paper. (ii) I help him write the paper.

Do some genres or text types become more complex than others?

121

a treatment of this issue. Informative and/or semantic complexity will not play a role in this study either. Since the concern of this chapter is the investigation of complexity in strictly unmarked subjects, objects, and adverbials, rather than in the whole sentence, we cannot assess the influence that informative factors such as givenness and/or newness (see Arnold et al. 2000: section 1) or semantic determinants such as animacy (see Rosenbach 2005), among others, exert on complexity. Finally, complexity is regarded as a synonym of processing difficulty or cost, either for the speaker or for the hearer. Whereas we agree with McWhorter (2001: 134) when he says that “all languages [in all periods] are acquired with ease by native learners” and, in consequence, historical periods of a given language cannot be graded on a scale of processing difficulty, we contend that, within a language, certain syntactic phenomena in certain textual contexts are particularly demanding (Dahl 2004: 40) in terms of processing cost and, with no doubt, more demanding than others. Put another way, linguistic phenomena can be graded in terms of processing cost and, consequently, can be used to characterize a textual variety or text type of the language under consideration. Moreover, if such an analysis is carried out by paying attention to different periods of the history of the language, then we will be shedding light on the diachronic linguistic characterisation of the text types under investigation. A further remark is in order here regarding the connection between processing and complexity. The relationship between processing cost and syntactic apparatus is not one-to-one, since a sentence which is apparently complex from the point of view of its syntactic organisation is not necessarily difficult to process (see Gayraud and Martinie 2005). The measurement of the processing cost of, for example, a subject will not be based on derivational mechanisms that explain the emergence of a given structure. On the one hand, this justifies the fact that the basic formal representation used in this chapter does not utilize empty or covert nodes and simply combines the different constituents in a syntactically hierarchical structure based on the notion of dependency. As a consequence of this, the related complexity metrics (metrics 5 and 6; see section 4) will not encompass processes of syntactic derivation or, in Dahl’s (2004: 44) terminology, “steps,” and will have no validity in proving whether a given (Externalized) manifestation of a language is due to a more primitive/elaborated version of Internalized Language or Universal Grammar. On the other hand, since we have avoided null terminal nodes in this (basic) analysis, we have not made any distinction between dependents that are subcategorized by the heads and dependents that are adjoined to some sort of intermediate projection (see, in this respect, Shapiro et al. 1992, or Gayraud and Martinie 2005, who claim that such a distinction has consequences for the processing cost of an utterance).

122

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

That there is a need for connectivity in syntax is supported by, for example, Hawkins’ (2006: 208) minimize domains and combination principles, or his (2006: 211) notion of phrasal combination domain, which would disfavour the analysis of (5) as (5’) and would give support to phrasal combinations such as those in (5”): (5)

the students from Siegen

(5’)

the students from Siegen (5”)

the students from Siegen

the students from Siegen

In this section we have favoured a concept of complexity based on an objective analysis of the surface structure of core clausal components, something which permits the scaling of the constituents in terms of syntactic processing cost. It must be emphasized that we have avoided any sort of connection between complexity and systemic functionality, semantic content or referentiality in the external world, and cognitive processing or derivational intricacies.

4. The metrics Providing that our concept of linguistic complexity is relative (see section 3), and also that this exploration relies on the diachronic scaling of subjects, objects and adverbials in different text types according to their syntactic and/or structural complexity, this investigation requires a set of objective quantitative metrics that can evaluate objectively the complexity of the constituents, independently of uses and users. In what follows we describe the metrics that will be used in the analysis of the data in section 5.

Do some genres or text types become more complex than others?

123

4.1. Metrics 1 and 2: Size or length of constituent Size, or length, has commonly been associated with complexity in the literature. For example, Wasow (1997: 81) maintains that grammatical weight implies “size of complexity”, Yaruss (1999: 330) claims that “attempts to separate length and complexity are somewhat artificial”, and Hawkins (2004: 9) argues that “[c]omplexity increases with the number of linguistic forms.” Metrics 1 and 2 measure the size or length of the constituent by counting the number of words of the subject, object, or adverbial (metric 1) and the number of words up to (and including) what we call the marker of the last immediate constituent of the subject, object, or adverbial (metric 2). Whereas counting all the words of a constituent (metric 1) provides us with its objective length, the computation of the number of words from the first word to the syntactic marker of the last constituent of the subject, object, or adverbial (metric 2) offers the empirical value of the length of the portion of subject, object, or adverbial which the hearer/reader has to process in order to grasp the syntactic structure of the whole subject, object, or adverbial. In consequence, it reflects the amount of the construction that has to be memorized so that the parsing of the whole constituent can be realised correctly. Metric 2 relativizes the effect of metric 1 since the former focuses on the length of the minimal segment in the construction that is capable of providing information about the overall syntactic structure independently of the length of the final constituent (see Wasow [1997: 102] on the felicitousness of relativising weight). The portion of lexical structure up to what we are calling the marker of the last immediate constituent is similar to Hawkins’ (2004: 32) “Constituent Recognition Domain,” which is defined as follows: “[t]he [Constituent Recognition Domain] for a phrasal mother node M consists of the smallest set of terminal and non-terminal nodes that must be parsed in order to recognize M and all [Immediate Constituents] of M.” Our concept of a syntactic marker (a preposition in a prepositional phrase, a determiner or a possessive specifier in a noun phrase, a participle in a participial phrase, a coordinating conjunction in a coordinative structure, a subordinator [or complementizer] in a subordinate clause [or CP], a wh-proform in a relative clause, etc.) borrows Prideaux and Baker’s (1986: 32) concept of “bracketing,” according to which “[t]he language user assumes that when a new unit for processing is encountered or initiated, it will be marked as such.” Examples (6) to (15) illustrate the most frequent types of markers (in boldface) in our corpus examples (see appendix for a detailed list of markers):

124

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

(6)

The ceremony took place at nine o’clock (1866PAL1.N6) [determiner as the marker of the noun/determiner phrase]

(7)

Your Ladyship dares me to stop in my new work! (1751Richardson.X3) [possessive specifier as the marker of the noun/determiner phrase]

(8)

Nato’s first mission was now complete (1989TIM1.N9) [possessive marker ’s as the marker of the noun/determiner phrase]

(9)

Pleasure-seekers are notoriously the most aggrieved and howling inhabitants of the universe, (1869Eliot.X6) [noun as the marker of a determiner-less noun phrase]

(10)

Demands [. . . ] without which I can no longer answer the Occasions of my Family (1751Smollett.X3) [preposition as the marker of the postmodifying prepositional phrase]

(11)

The humility which you laud in a character such as that of Macready has always to me a certain falseness about it – (1876Trollope.X6) [relativizer as the marker of the wh-relative clause]

(12)

[France has made such proposals to our Court as cannot be accepted, and] at last added that they could not desert the Dutch (1785GEN2.N3) [complementizer as the marker of the that-clause]

(13)

Helen & Bill, by the way, send their fondest regards to you both. (1950Thomas.X9) [coordinating conjunction as the marker of the coordinative construction]

(14)

The declaration of neutrality demanded by the Minister of France, might have been considered as superfluous (1793STA1.N3) [ed-participle as the marker of the participial clause]

(15)

The apotheosis of Scobie – culminating for me in the shower of rockets from H.M.’s Navy – is sublimity. (1960Aldington.X9) [ing-form as the marker of the nexusless nonfinite clause]

This notion of marker is heavily tied to the concept of “incrementality,” whose main assumption, in Pickering et al.’s (2000: 5) words, is that “the language processing system must very rapidly construct a syntactic analysis for a sentence fragment, assign it a semantic interpretation.” The identification of marker relies not only on theoretical claims (for example, on Chomsky’s notion of syntactic head, see Kimball’s [1973] new nodes principle, as reported by Frazier [1979: 43], or Hawkins’ [2006: 209] dependency) but also on statistical information (Corley and Crocker 2000: 137).

Do some genres or text types become more complex than others?

125

After calculating the number of words up to the marker, metric 2 expresses the ratio of such a figure to the number of words of the subject, object, or adverbial, already measured by metric 1. As generally accepted in the literature, the higher the ratio in metric 2, the more (syntactically) complex the constituent under discussion. In Hawkins’ (2004: 9) words, “efficiency is increased by selecting and arranging linguistic forms so as to provide the earliest possible access to as much of the ultimate syntactic and semantic representation as possible.” 4.2. Metrics 3 and 4: Syntactic density Inspired by Hawkins’ (1994) IC-to-non-IC ratio,9 metrics 3 and 4 have been designed in order to calculate syntactic density (in Smith’s [1988: 272] terminology). Metric 3 measures the number of basic immediate constituents and metric 4 gives the ratio of words per immediate constituent. A large number of immediate constituents (metric 3) implies a higher degree of complexity. Longer immediate constituents (metric 4) signify a higher load of structural complexity in the constituent under analysis. 4.3. Metrics 5 and 6: Syntactic depth The syntactic depth of the subjects, objects, and adverbials has been measured by way of the number of abstract (non-terminal) nodes in a simple syntactic structure of the constituent (in metric 5) and the ratio of non-terminal-to-terminal nodes (our metric 6). As suggested in Frazier (1985: 156), based on Miller and Chomsky (1963), “the complexity of a [constituent] may be determined by simply dividing the number of nonterminals [= intermediate nodes] in the sentence by the number of terminals [= minimal nodes].” The figure in metric 6 gives an idea of the amount of syntactic apparatus associated with the words of a constituent and thus implies a higher degree of syntactic complexity. As a consequence of the assumption that few non-terminal nodes implies less complexity, metric 5 will also be seen as indicating syntactic complexity. To give an example, the phrases (16) and (17), with their corresponding syntactic analyses in (16’) and (17’), differ in the degree of linguistic complexity: (16)

the spy with binoculars from Italy (‘the spy is from Italy’)

(17)

the spy with binoculars from Italy (‘the binoculars were made in Italy’)

9. Hawkins’ IC-to-non-IC ratio links the ratio in (our) metric 3 not to the total number of words, that is, the total number of terminal nodes, but to the total number of non-terminal plus terminal nodes up to the marker.

126

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

(16’)

the spy with binoculars from Italy (17’)

the spy with binoculars from Italy (16’) contains three non-terminal levels, represented by dotted lines. By contrast, (17’) includes four non-terminal levels. In the light of well-known principles such as minimal attachment (Frazier 1979, favoured also by Clifton et al. [1991: 266]), late closure (Frazier 1979) or recency (Gibson et al. 1996), the analysis in (16’) is less complex than the one in (17’). Metrics 5 and 6 receive further support from the relevant literature. Among others, Beaman (1984: 45) declares that “[i]t has generally been accepted that syntactic complexity in language is related to the number, type, and depth of embedding in a text”; Sampson (2001: 47) maintains that the depth of a word or constituent is the “total number of those nonterminal nodes in the word’s lineage which have at least one younger sister”; Warren and Gibson (2002: 79– 80) claim that “nested10 [. . . ] syntactic structures are more difficult to process than non-nested structures. Increasing the number of nestings makes a sentence unprocessable;” and finally, Dahl (2004: 44) says that “‘structural complexity’ [is] a general term for complexity measures that pertain to the structure of expressions, at some level of description.”

10. Gibson’s nesting does not coincide with our notion of embedding. For Gibson (2000: 96), the definition of nesting is as follows: “[a] syntactic category A is said to be nested within another category B if B contains A, a constituent to the left of A, and a constituent to the right of A.” The defining factor of embedding in this chapter is simply inclusion, subcategorisation is not relevant to our notion of embedding.

Do some genres or text types become more complex than others?

127

4.4. Metrics 7 and 8: Syntactic efficiency Hawkins (1994) measures the degree of syntactic efficiency by way of his IC[Immediate Constituent]-to-word ratio (similar to our metric 7) and his online IC-to-word ratio (comparable to our metric 8). The former consists of the division of the number of immediate constituents by the number of words up to the marker of the last immediate constituent. By contrast, Hawkins’ on-line IC-to-word ratio is the aggregate of the division of the number of the immediate constituents by the number of words that belong to it up to the marker, as illustrated in (18): (18)

The application of the Bank of Victoria to the Supreme Court for its approval of the amended scheme of reconstruction which was adopted by the colonial shareholders and depositors at meetings held on the 26th inst. was heard to-day. (1893MAN2.N6)

[The application]

[of the Bank of Victoria]

[to the Supreme Court]

[for its approval of ...]

Immediate Constituent 1

Immediate Constituent 2

Immediate Constituent 3

Immediate Constituent 4

2 words

5 words => 7 words up to here

4 words => 11 words up to here

1 word up to and including the marker => 12 words up to the marker

1/2 = 50%

2/7 = 28.57%

3/11 = 27.27%

4/12 = 33.33%

aggregate 34.79%

As Hawkins (2004: 32–33) states, “[t]he higher these ratios, the more minimal is the Constituent Recognition Domain’; “[t]he human parser prefers linear orders that minimize [Constituent Recognition Domains] (by maximizing their IC-tonon-IC [or IC-to-word] ratios).” So that all the metrics can lead positively to the measurement of complexity of the subjects, we have opted for changing the order of the terms of the division of Hawkins’ ratios, resulting in word-to-IC (our metric 7) and on-line word-to-IC (metric 8). Such ratios will thus describe lack of (or negative) syntactic efficiency. A few words are in order here to clarify the effect of metrics 7 and 8. Metric 7 gives information about the structure that has to be processed in order to retrieve the overall syntactic design of the constituents (that is, words up to the

128

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

marker) per immediate constituent. To give some examples, in a phrase with the design in (19), metric 7 will tell us that by computing 2.66 words in each immediate constituent (8 words, 3 immediate constituents) we can grasp the syntactic structure of the phrasal element. Similarly, in the design in (20), the value for metric 7 will also be 2.66 since both the number of words up to the marker and the number of immediate constituents are identical to the figures in the phrase in (19). In consequence, the degrees of complexity attributed to the structures in (19) and (20) by metric 7 are the same. (19)

[3 words]constituent1 [4 words]constituent2 [1 word up to the marker]constituent3

(20)

[5 words]constituent1 [2 words]constituent2 [1 word up to the marker]constituent3

Metric 8 also calculates the amount of structure that has to be processed per immediate constituent but discriminates against leftmost heavy constituents. Compare, for example, the designs in (19) and (20). As already noted, both ideal phrasal segments contain the same number of words up to the marker (8 words). However, they differ in the way in which such words are distributed in the three immediate constituents: the first constituent comprises three words in (19) and five words in (20), the second immediate constituent contains four words in (19) and two in (20), and the marker is the first word in the third constituents in both (19) and (20). In metric 8 (or Hawkins’ reversed on-line IC-to-word ratio) heaviness plays a significant role in the definitive computation and, since the first immediate constituent in (20) is heavier than the first constituent in (19), the value for the design in (20) will be larger. In fact, the figures for (19) and (20) are 3.03 and 3.7, respectively.

5. Analysis of the data In what follows, we will discuss the results of each of the most significant metrics and will connect the findings with the characterisation of the two text types under investigation in terms of linguistic complexity. 5.1. Size Figure 2 shows that there are no significant diachronic differences within each of the text types in late Modern English and present-day English as far as the values for the first metric are concerned. As regards the differences among functional constituents, one can draw the following conclusions. First, the subjects are shorter than the objects in the news and the letters. Second, the subjects in the

Do some genres or text types become more complex than others?

129

Table 4. Metrics type text

period

1750–1799

news

1850–1899

1950–1990

1750–1799

letters

1850–1899

1950–1990

function subject object adverbial subject object adverbial subject object adverbial subject object adverbial subject object adverbial subject object adverbial

1 4.39 6.4 2.46 3.58 5.41 3.67 4.75 6.24 2.12 3.32 4.93 2.42 4.37 6.19 1.83 3.22 4.52 2.18

2 2.27 2.7 1.7 1.8 2.19 1.84 2.33 2.35 1.63 1.76 2.19 1.5 3.47 2.45 1.32 1.82 2.05 1.34

3 1.49 1.54 1.25 1.29 1.49 1.3 1.46 1.53 1.27 1.27 1.46 1.15 1.37 1.5 1.11 1.31 1.41 1.18

metrics 4 5 2.68 3.39 3.51 5.4 1.96 1.46 2.58 2.58 3.26 4.41 2.62 2.67 2.91 3.75 3.58 5.24 1.55 1.12 2.46 2.32 3.03 3.93 1.96 1.42 2.86 3.37 3.6 5.19 1.5 0.83 2.27 2.22 2.89 3.52 1.92 1.18

6 0.62 0.66 0.47 0.57 0.66 0.59 0.64 0.66 0.3 0.57 0.63 0.39 0.6 0.66 0.21 0.51 0.61 0.48

7 1.38 1.48 1.33 1.31 1.36 1.39 1.44 1.43 1.23 1.32 1.35 1.27 1.9 1.52 1.17 1.28 1.33 1.11

8 1.5 1.57 1.38 1.37 1.48 1.46 1.55 1.58 1.29 1.38 1.47 1.21 1.56 1.68 1.18 1.35 1.44 1.13

Figure 2. Metric 1 (number of words)

news are slightly longer than the subjects in the letters. Notice in this respect that we are not considering pronominal subjects; in fact, the size of the subjects in the news practically coincides with the size of the objects in the letters in the three periods. Third, the objects in the news are clearly longer than the objects

130

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

in the letters. These observations accord with O’Donnell’s (1974) conclusion that the average length of syntactic units in written language is greater than in spoken language. Fourth, the nominal adverbials are shorter than the nominal subjects and objects, despite the fact that most of them are (either absolutely or relatively) clause-final. Despite the differences evinced in the light of metric 1, Figure 3 displays close values for our metric 2, that is, for the lexical material that has to be processed in order to grasp the syntactic structure of the constituents. The figures are extremely similar among functional constituents, text types, and periods. In view of these data, a likely hypothesis is that syntactic complexity is not associated either with the syntactic functions of subject, object, and adverbial, or with the type of text (news and letters), at least in nominal constituents in the periods under research. Lexical complexity, by contrast, would make the difference. Whereas syntactic complexity is a qualitative notion that embraces complexity issues related to the syntactic organisation of a given constituent (among others, the amount of structure that has to be processed before the recognition of the syntactic design of a constituent, that is, before our marker), lexical complexity is a mere quantitative concept that reports the structural or distributional dimension of a constituent by computing the size of its structure. To give an example, the size of the post-marker material will have lexical and not syntactic consequences since such material is not relevant to the determination of the overall syntactic structure of the constituent. We shall come back again to the hypothesis that syntactic and textual variation does not condition syntactic complexity when we discuss the results of the ensuing metrics.

Figure 3. Metric 2 (number of words until and including the marker)

Do some genres or text types become more complex than others?

131

5.2. Density

Figure 4. Metric 3 (number of immediate constituents)

As was the case with metric 2, the results for metric 3 are practically identical in the three variables to which we are paying attention, that is, text types, syntactic functions, and periods. Such a surprisingly similar number of immediate constituents, which is effectively a metric of syntactic (and not lexical) design, favours the hypothesis that syntactic functions and text-type typology do not play a role in the determination of the degree of syntactic complexity, at least in the recent history of the English language. The diachronic variation evinced by metric 4 in Figure 5 is not statistically significant. However, several remarks seem in order here so that the results in Figure 5 can be couched within a general theory of linguistic complexity. First, metric 3 suggested that there were no differences between functions and text

Figure 5. Metric 4 (words per immediate constituent)

132

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

types as far as syntactic complexity was concerned, since the syntactic structure of the examples under investigation, that is, the number of their immediate constituents, was comparable. Second, metric 2 showed that the amount of lexical material previous to the marker was similar in the two text types and in the three functional clausal constituents. Third, metric 4 now tells us that the average lengths of the immediate constituents are similar. Such findings can only be in keeping with the fact that there are differences in the total number of words of the constituents, evinced by metric 1, if we hypothesize that the nominal clausal elements under investigation contain a very small number (one maybe) of postmarker immediate constituents, and that it is in these immediate constituents where the differences of size suggested by metric 1 can be found. In other words, a nominal clausal component is more complex than another one because of the lexical complexity of the immediate constituent(s) occurring after the marker, something which does not imply more syntactic complexity on the part of the subject, object, or adverbial. 5.3. Depth In section 4 we reported that two metrics illustrate the concept of syntactic depth: metric 5, which measures the number of non-terminal nodes, and metric 6, which computes the ratio of non-terminal to terminal nodes. As regards metric 5, we must recognise that, since in the vast majority of the examples the number of non-terminal (intermediate) nodes coincides with the number of words minus 1, and the exceptions to this tendency do not modify the resulting figure at all (see Table 4), this metric does not shed any light on complexity variation. In consequence, it will be discarded in the analysis of the data here. On the other hand, metric 6 offers the ratio of non-terminal nodes (in metric 5) to words (in metric 1). Since we have already seen a strong mathematical connection between metrics 1 and 5, the computation in metric 6 is, on a theoretical basis, also futile and thus will be disregarded here. 5.4. Lack of efficiency The diachronic tendency affecting the late Modern English and present-day English periods is already corroborated by Figure 6, given that there are no significant statistical differences in the values for metric 7 across time. Besides, the results for the two text types and the three functional constituents in the three periods are practically identical. From a theoretical perspective, given that statistical variation in metric 7 would imply a change in syntactic complexity, the fact that the amount of lexical structure that has to be processed per syntactic node up to the marker is identical in the subjects, objects, and adverbials in

Do some genres or text types become more complex than others?

133

Figure 6. Metric 7 (word-to-IC ratio)

the database indicates likeness of syntactic complexity. The conclusion, already hypothesized in the previous subsection, is as follows: The degrees of syntactic complexity, that is in the material previous to the marker, are comparable in the three functional constituents. The conclusions drawn for metric 7 in Figure 6 are also applicable to the picture displayed in Figure 7 for metric 8, namely, coincidence in the data for periods, text types, and functional constituents as far as the on-line word-to-IC ratio, which analyses syntactic complexity in terms of regularity previous to the marker (see section 4 for the description of this metric). Once more, the conclusion must be that syntactic complexity is statistically equivalent in the subjects, objects, and adverbials of the news and letters investigated.

Figure 7. Metric 8 (on-line word-to-IC ratio)

134

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

6. Concluding remarks and further research This chapter has reported on the exploration of lexical and syntactic complexity in two text types – news and letters – in the recent history of English. It has focused on the degree of linguistic complexity exhibited by constituents functioning as (unmarked preverbal) subjects, (unmarked postverbal) objects, and adverbials occurring in declarative sentences in a corpus of texts from 1750 to present-day English, taken from the British component of ARCHER. Whereas in the literature it has been claimed that subjects (external arguments) and objects (internal arguments) are crucial as far as the determination of complexity and processing are concerned, we have also paid attention to adverbials (nonsubcategorized constituents) in an attempt to determine whether complexity is conditioned by syntax. This chapter has concentrated on a tiny aspect of linguistic complexity in the English language, namely, complexity in the syntactic design of the subjects, objects, and adverbials, and, in consequence, cannot lead to major claims on complexity variation in the language. The main conclusions reached in this chapter are as follows. First, the proportion of pronominal subjects in the informal texts (letters) is greater than the rate in the formal written material (news). This fact, according to Gibson and Thomas (1999: 232), is interpreted as a sign of low complexity in the former texts. By contrast, non-pronominal subjects and objects are predominant in the news and their frequency increases across time. Second, the statistical variation of non-pronominal subjects, objects, and adverbials in the periods under scrutiny here is not significant. Third, the adverbials and the subjects are shorter than the objects, and the average length of all the syntactic units in the news is greater than in the letters. On the one hand, this indicates that the text type of news, as representative of written formal language, shows a higher degree of lexical complexity. On the other hand, the objects are lexically more complex than the subjects and the adverbials, which implies that endweight has been conditioning exclusively the postverbal argument positions in the clause, at least from 1750 to the present. Fourth, the degree of syntactic complexity, that is, processing complexity, has been shown to be much alike in the two text types and in the three types of constituents in the periods examined. This study, then, has demonstrated that the syntactic design of the nominal constituents realising functions such as subject, object and adverbial has not undergone significant variation in the last three centuries since the amount of linguistic structure which the reader/hearer has to process in order to determine the overall syntactic category of the constituents has been practically identical across time. By contrast, there are clear differences among text types as far as the degree of lexical complexity is concerned, that is, as regards the

Do some genres or text types become more complex than others?

135

length of specifically the post-head constituents (postmodifiers and complements). We have left for further research both the issue of the analysis of marked subjects and objects, and also that of a fine-grained analysis taking into account the positions occupied by the adverbials in the clause.

Appendix: Taxonomy of markers In phrases: – noun phrases: · with no postmodification/complementation: – consisting of only a noun: the noun – introduced by a determiner: the determiner (DP-hypothesis) – introduced by a possessive phrase: the possessive clitic ’s, which counts as a word · with postmodification/complementation: the marker of the postmodifier/complement – prepositional phrases: the preposition – adjective phrases: the adjective – adverb phrases: the adverb In clauses: – clauses introduced by a non-wh complementizer: complementizers (that, while, whether, etc.) – (relative and interrogative/exclamative) wh-clauses: · with overt wh-form: the wh-form · with covert wh-form: the empty wh-form In the wh-clause in the girl [I know], an empty position previous to the underlined segment is taken as the marker of the relative clause because: (i) of analogy with the girl who(m) I know – complexity differences between the girl I know and the girl who(m) I know would find no justification, (ii) the speaker/reader processes I know as a subordinate clause depending on (the) girl once s/he has accepted the existence of an empty wh-position which warrants cohesion with the preceding constituent. – finite clauses not introduced by a complementizer: the inflected verbal form (the operator if the verbal group is complex and/or phrasal, or the verb in simple verbal groups) – nonfinite clauses not introduced by a complementizer: · with to-infinitive: the infinitival marker to · with an -ing form: the -ing verbal form · with an -ed form: the -ed verbal form In coordinative structures: the coordinating nexus

136

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

References Arnold, Jennifer E., Thomas Wasow, Anthony Losongco and Ryan Ginstrom 2000 Heaviness vs. newness: The effects of structural complexity and discourse status on constituent ordering. Language 76: 28–55. Beaman, Karen 1984 Coordination and subordination revisited: Syntactic complexity in spoken and written narrative discourse. In: Deborah Tannen (ed.), Coherence in Spoken and Written Discourse, 45–80. (Advances in Discourse Processes 12.) Norwood, NJ: Ablex. Biber, Douglas 1994 An analytical framework for register studies. In: Douglas Biber and Edward Finegan (eds.), Sociolinguistic Perspectives on Register, 31–56. (Oxford Studies in Sociolinguistics 3.) New York: Oxford University Press. Biber, Douglas 2001 Dimensions of variation among eighteenth-century speech-based and written registers. In: Susan Conrad and Douglas Biber (eds.), Variation in English: Multi-Dimensional Studies, 200–214. Harlow: Longman. Biber, Douglas and Edward Finegan 1997 Diachronic relations among speech-based and written registers in English. In: Terttu Nevalainen and Leena Kahlas-Tarkka (eds.), To Explain the Present. Studies in the Changing English Language in Honour of Matti Rissanen, 253–276. Helsinki: Soci´et´e N´eophilologique. Carlson, Katy 2002 Parallelism and Prosody in the Processing of Ellipsis Sentences. (Outstanding Dissertations in Linguistics 72.) New York: Routledge. Clifton, Charles Jr., Shari Speer and Steven P. Abney 1991 Parsing arguments: Phrase structure and argument structure as determinants of initial parsing decisions. Journal of Memory and Language 30: 251–271. Corley, Steffan and Matthew W. Crocker 2000 The modular statistical hypothesis: exploring lexical category ambiguity. In: Matthew W. Crocker, Martin Pickering and Charles Clifton Jr. (eds.), Architectures and Mechanisms for Language Processing, 135– 160. Cambridge: Cambridge University Press. Crain, Stephen and Donald Shankweiler 1988 Syntactic complexity and reading acquisition. In: Alice Davison and Georgia M. Green (eds.), Linguistic Complexity and Text Comprehension: Readability Issues Reconsidered, 167–192. (Psychology of Reading and Reading Insruction.) Hillsdale, NJ: Lawrence Erlbaum.

Do some genres or text types become more complex than others? ¨ Dahl, Osten 2004

137

The Growth and Maintenance of Linguistic Complexity. (Studies in Language Composition Series 71.) Amsterdam/Philadelphia: John Benjamins. Davison, Alice and Richard Lutz 1985 Measuring syntactic complexity relative to discourse context. In: David R. Dowty, Lauri Karttunen and Arnold M. Zwicky (eds.), Natural Language Parsing. Psychological, Computational, and Theoretical Perspectives, 26–66. (Studies in Natural Language Processing 23.) Cambridge: Cambridge University Press. Ferreira, Fernanda 1991 Effects of length and syntactic complexity on initiation times for prepared utterances. Journal of Memory and Language 30: 210–233. Frazier, Lyn 1979 On Comprehending Sentences: Syntactic Parsing Strategies. Bloomington, IN: Indiana University Linguistics Club. Frazier, Lyn 1985 Syntactic complexity. In: David R. Dowty, Lauri Karttunen and Arnold M. Zwicky (eds.), Natural Language Parsing. Psychological, Computational, and Theoretical Perspectives, 129–189. (Studies in Natural Language Processing 23.) Cambridge: Cambridge University Press. Frazier, Lyn 1988 The study of linguistic complexity. In: Alice Davison and Georgia M. Green (eds.), Linguistic Complexity and Text Comprehension. Readability Issues Reconsidered, 193–221. (Psychology of Reading and Reading Instruction.) Hillsdale, NJ: Lawrence Erlbaum. Gayraud, Fr´ed´erique and Bruno Martinie 2005 Does structural complexity (defined as sentential subordination) necessarily involve processing difficulty (measured in term of time of planning)? Paper delivered at the Symposium ‘Approaches to Complexity in Language,’ University of Helsinki, 24–26 August 2005. Gibson, Edward 1998 Linguistic complexity: Locality of syntactic dependencies. Cognition 68: 1–76. Gibson, Edward 2000 The Dependency Locality Theory: A distance-based theory of linguistic complexity. In: Alec Marantz, Yasush Miyashita and Wayne O’Neil (eds.), Image, Language, Brain. Papers from the First Mind Articulation Symposium, 95–126. Cambridge, MA: MIT Press. Gibson, Edward, Neal J. Pearlmutter, Enriqueta Canseco-Gonzalez and Gregory Hickok 1996 Recency preference in the human sentence processing mechanism. Cognition 59: 23–59.

138

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

Gibson, Edward and James Thomas 1999 Memory limitations and structural forgetting: The perception of complex ungrammatical sentences as grammatical. Language and Cognitive Processes 14: 225–248. Giltrow, Janet this volume Genre as difference: The sociality of linguistic variation. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 29– 51. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Hawkins, John A. 1994 A Performance Theory of Order and Constituency. (Cambridge Studies in Linguistics 73.) Cambridge: Cambridge University Press. Hawkins, John A. 2004 Efficiency and Complexity in Grammars. Oxford: Oxford University Press. Hawkins, John A. 2006 Gradedness as relative efficiency in the processing of syntax and semantics. In: Gisbert Fanselow, Caroline F´ery, Ralf Vogel and Matthias Schlesewsky (eds.), Gradience in Grammar. Generative Perspectives, 207–226. Oxford: Oxford University Press. Hopper, Paul J. 1997 Dispersed verbal predicates in vernacular written narrative. In: Akio Kamio (ed.), Directions in Functional Linguistics, 1–18. (Studies in Language Companion Series 36.) Amsterdam/Philadelphia: John Benjamins. Markus, Manfred 2001 The development of prose in early Modern English in view of the gender question: Using grammatical idiosyncracies of 15th and 17th century letters. European Journal of English Studies 5: 181–196. McWhorter, John H. 2001 The world’s simplest grammars are creole grammars. LinguisticTypology 5: 125–166. Miller, George A. and Noam Chomsky 1963 Finitary models of language users. In: Duncan Luce, R. Robert R. Bush and Eugene Galanter (eds.), Handbook of Mathematical Psychology, Vol. 2, 419–492. New York: Wiley. Mondorf, Britta this volume Genre effects in the replacement of reflexives by particles. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 219– 245. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter.

Do some genres or text types become more complex than others?

139

O’Donnell, Roy 1974 Syntactic differences between speech and writing. American Speech 49: 102–110. P´erez-Guerra, Javier 2007 Am I more complex when I speak or when I write? A corpus-based study on linguistic complexity in spoken and written Present-day English. In: Mari Carmen Campoy and Mar´ıa Jos´e Luz´on (eds.), Spoken Corpora in Applied Linguistics, 127–146. (Linguistic Insights: Studies in Language and Communication 51.) Frankfurt/Berlin/Bern: Peter Lang. Pickering, Martin J., Charles Clifton Jr. and Matthew W. Crocker 2000 Architectures and mechanisms in sentence comprehension. In: Matthew W. Crocker, Martin Pickering and Charles Clifton Jr. (eds.),Architectures and Mechanisms for Language Processing, 1–28. Cambridge: Cambridge University Press. Prideaux, Gary D. and William J. Baker 1986 Strategies and Structures: The Processing of Relative Clauses. (Current Issues in Linguistic Theory 46.) Amsterdam/Philadelphia: John Benjamins. Rohdenburg, G¨unter 1996 Cognitive complexity and increased grammatical explicitness in English. Cognitive Linguistics 1: 149–182. Rosenbach, Anette 2005 Animacy versus weight as determinants of grammatical variation in English. Language 81: 613–644. Sampson, Geoffrey 2001 Empirical Linguistics. (Open Linguistics.) London: Continuum. Shapiro, Lewis P., Patrick McNamara, Edgar Zurif, Susan Lanzoni and Laird Cermak 1992 Processing complexity and sentence memory: Evidence from amnesia. Brain and Language 42: 431–453. Smith, Carlota S. 1988 Factors of linguistic complexity and performance. In: Alice Davison and Georgia M. Green (eds.), Linguistic Complexity and Text Comprehension. Readability Issues Reconsidered, 247–279. (Psychology of Reading and Reading Instruction.) Hillsdale, NJ: Lawrence Erlbaum. Szmrecsanyi, Benedikt 2009 Typological parameters of intralingual variability: Grammatical analyticity versus syntheticity in varieties of English. Language Variation and Change 21: 319–353. Szmrecsanyi, Benedikt and Bernd Kortmann 2009 Between simplification and complexification: Non-standard varieties of English around the world. In: Geoffrey Sampson, David Gil and Peter Trudgill (eds.), Language Complexity as an Evolving Variable, 65–79.

140

Javier P´erez-Guerra and Ana E. Mart´ınez-Insua

(Studies in the Evolution of Language 14.) Oxford: Oxford University Press. Szmrecsanyi, Benedikt and Bernd Kortmann (forthcoming) Typological profiling: Learner Englishes versus indigenized L2 varieties of English. Advance online preprint http://www.benszm.net/omnibuslit/ Szmrecsanyi Kortmann profiling webversion.pdf (last accessed on March 30, 2010). Taavitsainen, Irma 2001 Changing conventions of writing: The dynamics of genres, text types, and text traditions. The European Journal of English Studies 5: 139–150. Virtanen, Tuija this volume Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 53–84. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Warren, Tessa and Edward Gibson 2002 The influence of referential processing on sentence complexity. Cognition 85: 79–112. Wasow, Thomas 1997 Remarks on grammatical weight. Language Variation and Change 9: 81–105. Yaruss, J. Scott 1999 Utterance length, syntactic complexity, and childhood stuttering. Journal of Speech, Language, and Hearing Research 42: 329–344.

Mein Problem ist/mein Thema ist (‘My problem is/my topic is’): How syntactic patterns and genres interact Wolfgang Imo

Two closely related radio phone-in formats will be analyzed in terms of their status as communicative genres. The first genre is that of a psychological radio counselling programme, the second one is a talk or chat programme. Every now and then, there are instances of hybrid or merged genres where a listener calls at the talk programme with the intention of receiving serious psychological counselling. The question is, then, how the interactants – the caller and the host – manage the “reciprocal adjustment of perspectives” (Luckmann 1992: 229) that is necessary for the formal and functional change which is needed to transform something akin to small talk into a bona fide counselling session. The solution proposed here is that the interactants make use of certain syntactic and sequential constructions – e.g. the fixed phrase “my problem is” for counselling formats, as opposed to “my topic is” for talk formats – to perform the hybridization of genres. In the end, a re-analysis of genres in terms of an integrational approach of interactional Construction Grammar is proposed, which might help unify the description of the complex interrelated processes that are part of the hybridization of genres.

1. Radio phone-in formats: psychological counselling versus small talk The aim of this study is to compare two radio phone-in formats, formats which are closely related and sometimes merge to form hybrid genres. The first format is exemplified by the German counselling programmes Ratgeber Lebensfragen (‘advice on general questions of life’) and Von Mensch zu Mensch (‘from person to person’) and the second format by the German talk programme Domian (named after its host J¨urgen Domian). One of these formats involves the psychological counselling of callers by professional psychologists or trained priests. The other format is centred on a popular host, who usually only wants to chat with his callers (the self-description of the format accordingly is that of talk ra-

142

Wolfgang Imo

dio), but who sometimes also offers lay psychological help if need be. Usually, both of these formats can easily be kept apart, as they show many differences in terms of how the hosts talk, of the sequential structure of the calls, of the lexical, syntactic, and prosodic material that is used, and of the topics that are discussed. But while the serious radio counselling formats never develop into mere small talk, the radio talk format sometimes evolves into a serious counselling session and, accordingly, the structural and formal features of this genre are suddenly produced by both the caller and the host. The question is: How do the interactants manage to adjust their expectations about what is going to happen so that their perspectives converge on a common communicative aim, namely that of a serious psychological counselling session? How is it possible to start with a format that favours small talk and sensationalism and end with one that favours serious advice and sensitivity about the caller’s problems? The solution is that the interactants use certain fixed phrases – for example my problem is or, alternatively, my topic is – as well as a range of syntactic patterns that act as contextualization cues (Gumperz 1982) that signal which type of communication is about to take place. These signals are comparable to Puschmann’s (this volume: 188) “deictic bridges,” which can also be used for contextualization cues. For example, by changing the phrase my topic is (which is typical of the radio talk format) to my problem is (which is typical of the psychological counselling format), callers can offer a re-framing of the genre or even the activation of a new genre to the host, who can then decide to go along with the caller by using the word problem himself. So the interaction between a large and complex pattern – a genre – and a smaller syntactic construction is vital for managing the “reciprocal adjustment of perspectives” (Luckmann 1992: 229) of caller and host. I will show in detail what the typical formal features of the two genres are and how hybrid genres can be activated by my problem is, a phrase that bridges the gap between local (phrasal and syntactical) and global (genre-related) structures. Before doing so, a short introduction to the theory of communicative genres will be given. In the end – and as a perspective for further research – I will look at the possible advantages of merging the concept of communicative genres with the theory of Construction Grammar. Both approaches share a lot of common ground in terms of their aims, ideas, and methods. A merger of both approaches may facilitate the analysis of the complex relations of small syntactic structures (constructions) to larger ones (genres, text types).

How syntactic patterns and genres interact

143

2. The concept of communicative genres The concept of communicative genres was developed by Luckmann (1986, 1988, 1992) and in the following decades it has been extended by G¨unthner (1995, 2006), G¨unthner and Knoblauch (1994, 1995), and Bergmann and Luckmann (1995) to take into account the special properties of genres that rely heavily on interactive and co-productive features, such as gossiping or reproaching (see also G¨unthner, this volume). According to Luckmann (1992: 227), communicative genres “offer solutions to specifically communicative problems.” These problems are typically dialogical ones, in other words, they have to do with “the interlocking perspectives in the ‘dialogical’ production and reception of communicative action” and provide for the “reciprocal adjustment of perspectives” of the interactants. Both the producer and the recipient of a set of utterances need to agree on what agenda they are both following. In order to relieve interactants of having to discuss these agendas all the time on a meta-communicative level – a time-consuming and cumbersome procedure – genres have become part of the social stock of knowledge of a society. Genres consist of more or less fixed patterns of speech that include uniform lexical material and syntactic structures, recurrent patterns of sequential and turn-taking organization, and a high degree of predictability within certain situations or institutions. This co-occurrence of features simplifies communication in a twofold way: First, once a genre is activated, it triggers routinized expectations in the recipients. One can pretty accurately project the course an activated genre will take. Second, certain situations – a job interview, for example – are associated with one or more genres that the interactants can use and adhere to. By using the genre job interview (Birkner and Kern 2000) both the applicant and the personnel officer have lexical, syntactic, and other types of structures at hand, “relieving them [. . . ] of having to co-ordinate every communicative action anew” (G¨unthner and Knoblauch 1995: 6). Actions, situations and genres are closely linked together: “If one may say that the individual ‘chooses’ a genre according to his communicative project and the perceived requirements of the social situation, one may say correspondingly that, once a genre is ‘chosen,’ it is the model which ‘chooses’ the parts in executing the communicative project” (Luckmann 1992: 226). If a special interactive setting or situation demands that a communicative job has to be done again and again, linguistic routines slowly congeal into fixed patterns that can then become blueprints for future interactions. Once a genre is generated, the communicative processes within that genre are no longer spontaneous but follow prepatterned routines: “By prepatterned, we refer to the observation that the occurrence of a certain feature of communicative actions makes the occurrence of another feature expectable or predictable” (G¨unthner and Knoblauch 1995: 8). This, of

144

Wolfgang Imo

course, does not mean that genres force their rules onto every communication within a certain setting. The degree of fixedness of a genre itself is variable and can reach from a fairly rigid and inflexible pattern (e.g., a wedding ceremony) to a loose preference for certain communicative actions (e.g., a reproach). Typically, genres are complex patterns with a high degree of fixedness. In order to analyze these complex patterns, three levels are taken as the starting points of analyses: “The level of internal structure, the situative level and the level of external structure” (G¨unthner and Knoblauch 1995: 8). The level of internal structure includes the material basis a genre consists of: words, phrases, idioms, formulae, syntactic structures, prosodic and nonverbal features, semantic content, and stylistic devices. The situative level includes those aspects that are related to the interactive setting of communication: turn-taking, adjacency pairs, pre-sequences, sequential ordering, co-production of utterances, patterns of storytelling, and many other structures which have been described and analyzed in the wake of conversation analysis (Schegloff, Jefferson, and Sacks 1977; Sacks and Schegloff 1979; Sacks 1992). Furthermore, the participation framework consisting of the production format as well as the participation status is part of the analysis of the situative level. Production format refers to the relation of the speaker to his or her utterance and to the figures portrayed (for example, is the speaker the original producer of the utterance or does he merely report someone else’s utterance?). Participation status refers to the “portrayed relation between the communicating participants and their utterances” (G¨unthner and Knoblauch 1995: 14). The last aspect of the situative level is that of the social occasion which forms the context of a given communicative interaction. The external structure (“outer structure” according to Bergmann and Luckmann 1995: 291) comprises those aspects of a genre that link it to the larger social structure: milieus, ethnic and cultural groups, gender-related situations, and so on. Thus, the category of genre is “located between the linguistic, code-related, and the institutional, social-structure-related determination of communicative processes” (Bergmann and Luckmann 1995: 189). One of the most important features of genres is that they have to be viewed as interactively constructed, dialogical products that are intertwined with the actual interactive, situational, and social contexts within which they occur (G¨unthner 1995: 210). Because of this focus on the on-line construction and cooperative talking-into-being (Boden 1994) of genres, recipient design and recipient reactions as well as meta-communicative negotiations of the right form of a genre are important aspects. Furthermore, the reflexive structure of genres has to be taken into account: Not only do certain contexts provide the basis for the selection of certain genres, the choice of a genre itself can create the context for this genre anew.

How syntactic patterns and genres interact

145

One last aspect refers to the relation of genres to each other. Genres do not exist as isolated patterns, but form the center of the communicative budget of a society. The communicative budget includes all forms of communication, those that have become sedimented into fixed patterns (or discourse traditions, according to Kabatek, Obrist and Vincis, this volume) as well as those that happen on an ad hoc basis and do not orient to genres (“the whole communicative dimension of social life” [Bergmann and Luckmann 1995: 301]). Due to the fact that genres are solutions to important and recurrent social problems, they provide the heart of the communicative budget. By describing the genres of a society, we can describe the vital communicative situations and processes of that society. Genres can often be grouped into “families” (Bergmann and Luckmann 1995: 300) that share certain features (for example, they occur in the same situation, they have similar sequential structures, they rely to some degree on the same lexical or syntactic material). Sometimes the borders between related genres may be clear, but often hybrid forms of genres emerge (Heyd and Giltrow, this volume). How such hybrids emerge and how interactants signal and ratify the change of one genre into another by syntactic means will be the aim of this study.

3. The data As examples of the format psychological counselling programme I will use ten transcribed phone-in calls at two broadcasting stations in southern Germany. Both offer advice by trained psychologists or priests. One is called Ratgeber Lebensfragen (‘advice on general questions of life’), broadcast by the station Bayern 2, the other Von Mensch zu Mensch (‘from person to person’) by the station SWF 1. The topics are usually open – any kind of private problem can be the reason for a call. Sometimes, though, special topics are designated by the radio hosts, for example the topic of “obsessions.” In that case, only callers who suffer from an obsession (obsessive washing, for example) are put through to the psychologist or priest who is counselling that day. The samples for the format radio talk programme are taken from the programme Domian, run by the popular host of the same name and broadcast “bimedially” (Schramm 1999: 23) both on TV (on WDR) and radio (Eins live).1 Twenty-six calls have been recorded and 1. Domian is not a real TV programme, but rather an instance of “filmed radio” (Schramm 1999: 23), because there is just a single, fixed camera showing Domian sitting in the radio broadcasting studio. The emphasis on radio can both be seen in the reaction of the callers, who usually say they “listen to” Domian and in the self-description of the programme as “Radiotalk” on the Eins Live homepage, see also Krause (2006). Unlike Ratgeber Lebensfragen and Von Mensch

146

Wolfgang Imo

transcribed. Domian also features open sessions as well as sessions in which the host J¨urgen Domian designates a special topic. Topics can be problem-oriented (e.g., “children can be cruel”), but more often entertaining topics (e.g., “sex at uncommon places”) are preferred. The audio files of the broadcasts are part of Susanne G¨unthner’s database of spoken German and have been transcribed by me according to the standards of the transcription system GAT (Gespr¨achanalytisches Transkriptionssystem, see Selting et al. 1998; see also G¨unthner, this volume). In the following paragraphs I will first briefly sketch out some of the features of the outer structures of both formats and will then present the situational levels and the inner structures of both formats. Next, some examples of Domian sessions that start out as tokens of the genre radio talk programme but are interactively changed into tokens of the genre psychological counselling programme will be analyzed. In the final section, I will discuss how the theory of communicative genres can be extended by using the terminology of Construction Grammar.

4. Two genres: Psychological counselling programmes and radio talk programmes 4.1. The outer structure of the psychological counselling formats The outer structure of the counselling programmes is framed by three main factors: the callers, the audience, and the producers of the programme (Willmann 1996: 204). The producers include the actual consultant, the radio host, the telephone service of the broadcasting station, and the station’s technical personnel. The consultants (either psychologists or priests) vary and usually have fixed days on which they counsel. Thus, callers of the programme who have listened to it for some time have the opportunity to listen to those consultants they like best. The radio hosts only appear at the beginning and the end of – but never during – a counselling session. They greet and say goodbye to the callers and sometimes ask the experts to elaborate on a problem or clear up aspects of a topic after a call that might be unclear to the audience. They are also responsible for keeping up the speed of the sessions by reminding the callers to be concise zu Mensch, Domian is a format or genre very much geared to the personality of the host himself. This, of course, is to be expected because talk formats rely heavily on the ability of the host to draw an audience by personal style and charisma. Counselling formats, on the other hand, do not have to rely on particular individuals as they offer a service that can be delivered by any qualified counsellor.

How syntactic patterns and genres interact

147

and by stressing how many callers are waiting in line or how much time has already elapsed. This results in a marked temporal limitation of psychological counselling programmes (Willmann 1996: 205), which is often referred to by the callers. 4.2. The outer structure of Domian On the surface, the outer structure of the radio talk programme Domian is very similar to that of the counselling programmes. Here, too, the triad of callers, audience, and producers sets the external frame. A closer look, though, reveals some striking differences. The persons on air consist only of the host and the callers. The whole format is centred on J¨urgen Domian, who has become a popular figure in the German mass media. While the broadcasting station also employs psychologists, they never go on air but work behind the scenes and only talk to some of the callers. Due to the fact that Domian is both host and main conversational partner (not so much a consultant), he is the most prominent part of the programme. Domian often gives advice, but as he has no psychological training (he studied German, politics, and philosophy [Schramm 1999: 24]), he can only give common-sense advice – a fact he often stresses in the midst of giving advice, sometimes with the ulterior motives of admonishing, ridiculing, or attacking the caller. The time frame is not controlled in as rigid a way as it is in the counselling sessions discussed above. Domian himself decides whether he wants to continue talking or finish a conversation. 4.3. The situational level of the psychological counselling formats The situational level comprises all kinds of phenomena that are related to the interactional setting of a communicative genre. However, I will only look at those aspects that involve syntactic and formal patterns which are correlated to the structure of the two genres. The first of these aspects is the greeting sequence. The following excerpt shows the typical structure of such a greeting: Example 1 01

HO

02 03

CA

counselling format: introduction .hh ich will UMgehend den (.) n¨ achsten Anrufer wieder einen h¨ orer begr¨ ußen; ‘.hh I want to greet the (.) next caller again a male one immediately;’ guten Abend? ‘good evening?’ guten Abend. ‘good evening.’

148

Wolfgang Imo

04

CO

05

CA

06

07 08

CO

09

CA

10

guten Abend. ‘good evening.’ (0.5) .hhh ja:; ‘(0.5) .hhh yes:;’ Also hh das problEm Aller steht a ¨h auch ¨ ah so ¨ ah vor MIR jetzt irgendwo; ‘well hh everybody’s problem now also eh presents itself to me somehow;’ wie wIe und wo Anfangen am besten; ‘how how and where best to begin;’ mhm, ‘mhm,’ ¨ ahm (.) es is eine etwas AUßergew¨ ohnliche proble- probleMAtik, ‘ehm (.) it is a somewhat uncommon problem,’ die vielleicht nicht so ganz a ¨h: H¨ AUfig in ihrer sendung erscheint, ‘which maybe doesn’t occur so very often in your programme,’

The example is typical insofar as it contains all the routinized formal and lexical elements commonly found in this type of phone-in call. First, the radio host often uses adverbs and adverbial phrases referring to the limited amount of time. In this case the host uses the adverb UMgehend (‘immediately’) in line 1. These adverbs are used to “talk into being” the seriality of the counselling format and force the callers to focus on their topics. Here a first close interaction between a syntactic format – sentences containing adverbs referring to (a lack of) time – and genre – seriality of counselling sessions – is established. The next step in the opening sequence is the actual greeting, where the formal, distanced, and polite greeting guten Abend (‘good evening’) is used in all cases. From the first utterances on, the participation status is thus signaled as one of politeness and formality, which is ratified by the greetings of the caller (CA) and the consultant (CO). Usually, the callers start talking about their troubles without being prompted by the host or consultant. Only rarely does the consultant say something like Tell me your problem. In example (1), the caller opens the conversation himself with an utterance including a word that is typical of the psychological counselling format: problem. A phrase containing this word occurs regularly in all counselling sessions and provide a very strong link between a syntactic pattern (my problem is) and the type of genre (counselling session). The ending of a conversation, too, offers many clues as to what has been going on during the conversation, and, again, formal phrases and patterns link local to global structures. The following example is typical for the closing sequence of a psychological counselling format:

How syntactic patterns and genres interact Example 2

counselling format: closing sequence

86

dAnn (.) hat die TANte, (.) ‘then (.) the aunt, (.)’ KAUM mehr m¨ oglichkeiten sie zu verlEtzen. ‘will have almost no chance to hurt you.’ (1.0) ah JA, ‘oh yes,’ DANke, ‘thank you,’ s war n GUter rat. (.) ‘that was good advice. (.)’ ja. ‘yes.’ [wir DANken ihnen.] ‘we thank you’. [ich w¨ unsche ihnen] n sch¨ onen Abend und‘I wish you a nice evening and-’ einen sch¨ onen Abend eine gute nAcht. ‘a nice evening and good night.’ ja gut NACHT. ‘yes good night.’ GUT nacht. ‘good night.’ auf WIEderh¨ oren. ‘good bye.’

CO

87 88 89

CA

90 91 92

CO

93

HO

94

CA

95

HO

96

CA

97

CO

98

HO

149

The first part of the closing sequence deals with the final ratification of the advice the consultant has given the caller. The caller first thanks the consultant and then explicitly calls her piece of advice a good one (line 90–91). Again, this phrase of ratification actively refers to the type of communicative genre to which the interactants are relating. By means of the locally produced syntactic pattern containing a positive evaluation of advice the callers “talk themselves into being” as clients of – and recipients of advice from – a professional consultant. Furthermore, the polite final closing sequences create a serious and professional situation close to that of visiting a paid professional consultant in person. An aspect of the production format that is very prominent in the counselling programmes is the consultants’ technique of reformulating the utterances of the caller: Example 3

counselling format: reformulation

43

.hh also ich verstehs SO. ‘.hh well that’s how I understand it.’ [sie] SAgen einen satz? ‘you say a sentence?’ [mhm] ja: .h‘mhm yes .h-’

CO

44 45

CA

150

Wolfgang Imo

46

CO

47

CA

48 49

CO

und ¨ ah dann MERken sie dass sie jetzt einen GANZ ¨hnlichen FAST den gleichen NOCHemal sagn. a ‘and ehm then you realize that you say a very similar one almost the same sentence again.’ ja::, ‘yes,’ .hh ¨ ah ¨ ah diREKT den gleichen; ‘.hh ehm ehm exactly the same one;’ diREKT den gleichen. ‘exactly the same one.’

By reformulating the utterances, the consultants appropriate the callers’ views and signal that they are working to get a close understanding of the problem. These reformulations are part and parcel of counselling programmes (Willmann 1996: 207), but not of the talk programme Domian. They rely on a set of fixed, formal syntactic patterns that usually start with the discourse marker also, which signals a conclusion, then go on with a phrase of try-marking (also ich verstehs SO, ‘that’s how I understand it’), and then deliver the reformulation in portions that are often produced in the format of questions. These formal patterns again act as contextualization cues referring to the genre of counselling session. 4.4. The situational level of Domian The following opening sequence immediately reveals the striking differences between programmes such as Von Mensch zu Mensch and Domian: Example 4

Domian: introduction

01

KINder k¨ onnen grausam sein; ‘children can be cruel;’ CARSten dreißig jahre ZU diesem thema guten MOrgen; ‘Carsten thirty years about this topic good morning;’ hallo DOmian [GR¨ USS dich ]; ‘hello Domian hi;’ [hallo CARSten]; ‘hello Carsten;’ ja du hast zu DEM thema angerufen waRUM; ‘well you called to talk about that topic why;’ .h JA weil ich einfach eh KLEINw¨ uchsig bin, ‘.h well simply because I am small,’

DO

02 03

CA

04

DO

05 06

CA

There are no adverbs or adverbial phrases referring to the limited amount of time. Instead, Domian (DO) either repeats the special topic, in this case KINder k¨onnen grausam sein, (‘children can be cruel’), or stresses that there is no fixed topic that night. Domian usually greets his callers with a formal Guten Morgen (‘good

How syntactic patterns and genres interact

151

morning’), but most callers either immediately switch to a more informal variety ¨ dich (‘hi’, line 3). This informal of greeting, such as hallo (‘hello’) and GRUSS greeting is always ratified by Domian, who prefers an informal participation frame (hallo CARSten, ‘hello Carsten’, line 4). The atmosphere of informality, though, is not initialized by the callers but by the fact that Domian always uses first names and the informal second person pronoun du (‘you’) instead of the polite and distanced form Sie (‘you’). So, the informal and easy mode of interaction is produced and ratified by both parties. The choice of pronouns hints at the fact that the genre is not one of serious and potentially face-threatening counselling but one of talk. After the opening sequence, some of the callers start talking on their own. More often, though, Domian asks about the caller’s reason for calling. He usually uses phrases containing the word Thema (‘topic’) for this purpose, thereby setting the frame for the conversation as one of telling stories instead of detailing problems. The endings, too, show the same signs of informality as the beginnings: Example 5

Domian: closing sequence

185

DO

186

HO

187

DO

188

HO

189

DO

190

HO

191

DO

192

HO

193

DO

194

HO

h GUT aber im moMENT scheints ja doch alles ganz ganz oK zu laufen, ‘.hh well but at the moment everything seems to be quite o.k.,’ ja BIN ich auch FROH [dr¨ uber]; ‘yes I’m happy about that too;’ [ja:, ] ‘yes,’ aber es war eben wirklich SEHR sehr grausam die jahre; ‘but it has been really very cruel all those years;’ GLAUB ich DAS glaub ich wirklich; ‘I believe I really believe that;’ ja; ‘yes;’ .h dann w¨ unsch ich EUCH DIR der faMIlie dem jungen Alles [gute von HERzen; ] ‘.h then I wish you all you the family the boy all the best from the bottom of my heart;’ [DANke (.) ;] ‘thank you ’ [tsch¨ uss HILtrud;] ‘bye Hiltrud;’ [oK jo tsch¨ uss; ] ‘o.k. yes bye;’

This example is taken from the ending of a conversation about a caller’s son, who suffers from Tourette’s Syndrome. In spite of this topic, which seems to invite counselling, Domian and the caller just talk about the case history of the caller’s son, details of the disease, and the problems a mother has to cope with

152

Wolfgang Imo

when a child suffers from Tourette’s Syndrome. Counselling is neither asked for nor given, and the caller herself stresses that everything is under control at the moment. So Domian just repeats the caller’s view (line 185) and gives her his best wishes for her and her family (line 191). Both Domian and the caller construct the situation as one of equal partners in a conversation, not as one of consultant and client. The typical ratification sequences shown in 4.3. can sometimes occur when Domian gives real advice, but usually they are missing. The final closing is very informal. Domian uses the casual phrase tsch¨uss (‘bye’) together with the caller’s first name and the caller responds accordingly. Again, the participation status, one of informality, is produced interactively. While patterns of reformulation form a standard technique in Von Mensch zu Mensch and Ratgeber Lebensfragen, they rarely occur in Domian. Instead, two other interactional patterns are prevalent. One is the strategy used by Domian to elicit as many sensational details of a story as possible. While the questions about details in the counselling programmes only serve to obtain vital information in relation to the actual problem, the questions from Domian are meant to elicit responses that detail the more sensational aspects of the caller’s topic. As there are no special syntactic formats that are applied for that purpose, I will not discuss this strategy in detail here. A second, much more fixed pattern involves Domian reproaching his callers. Example 6

Domian: reproaches 1

81

CA

82

DO

ja also wir treffen uns ja eigentlich NUR daf¨ ur; ‘yes well actually we meet only for it;’ ich bin entSETZT aber wIrklich; (3.0) ‘I’m aghast, really;’

((...)) 87 DO

eh (.) STIRBST du nich vor schlechtem gewissen deine deinem mann gegen¨ uber? ‘ehm (.) are you not dying because of your bad conscience regarding your husband?’

Domian regularly uses typical small constructions expressing disgust or reproaches, such as ich bin entSETZT (‘I’m aghast’), to give his personal view of the affair. Such structures that sensationalize the problems of the callers, as well as reproach them for their actions, are never found in psychological counselling formats.2 Therefore, their existence itself already marks the participation format as an informal one and signals the audience that the genre is one of talk. 2. Once Domian even gets berated himself by a caller who thinks Domian should not reproach callers, a criticism to which he responds by saying that he does not think his job should be to accept everything he hears. Klemm (1996: 257) describes such patterns as strategies of “confrotainment” and “emotainment.”

How syntactic patterns and genres interact

153

4.5. The internal structure of the psychological counselling formats The main differences between the psychological counselling and the talk programmes in terms of their internal structures lie in the words that occur, the prosody that is used, and the topics that are talked about on the programmes. As may be expected, the experts in Von Mensch zu Mensch and Ratgeber Lebensfragen use vocabulary that is – albeit on a popular level – part of an expert register. Examples of this vocabulary include eigenes ich (‘internal ego’), Eustress (‘positive stress’), mit sich ins Reine kommen (‘making peace with oneself ’), and Trauerweg (‘path of mourning’). Furthermore, the counsellors usually avoid emphatic and stressed utterances in favour of a quiet and calming way of speaking; often advice is given in a try-marked (rising pitch) prosodic format. The topics all centre on the typical problems one might expect to form the basis of psychological counsellings that take place during one-on-one telephone counsellings (such as the Samaritan telephone service) or at a psychologist’s office. One pattern that occurs in all of the calls analyzed here involves phrases containing the word problem. Most of the time this word is uttered as part of a phrase by the caller (mein Problem ist, ‘my problem is’; Ich habe ein Problem, ‘I have a problem’); in some cases the counsellor asks the caller to tell him/her what his/her problem is. The utterance containing problem thus has the status of a key phrase in the psychological counselling programmes, as it helps with “talking into being” the situation and roles of advice-seeker and counsellor and with setting the mode for the interaction as one of giving/receiving advice. At the end of a counselling session, the word Rat (‘advice’) has a similar function: It serves as a cue as to what the interaction was about. Therefore, an explicit reference to the preceding sequence as one in which good advice has been given can often be found at the end of a call (ah ja DANke, es war n GUter rat,‘oh yes, thank you, that was good advice’). With the help of phrases containing the words problem and advice, the interactants locally activate and highlight the genre they are using as the background structure of their conversation. 4.6. The internal structure of Domian The internal structure of Domian differs strongly from that of the other two formats. Sometimes, expert vocabulary occurs, but usually only after a caller has first used such a word or phrase. Domian himself uses everyday expressions such as Schuld (‘guilt’), Angst (‘fear’), and Scham (‘shame’) and sometimes even strong colloquial expressions, such as Scheiße (‘shit’) or even abusive words that are used to dramatize the situation for the audience. On a prosodic

154

Wolfgang Imo

level, Domian prefers a very expressive and emphatic manner of speaking; he often sighs, utters interjections, breathes in and out heavily and uses a reproaching or accusing tone. Though serious psychological problems can sometimes also be found in Domian, they only form a small subset of possible topics. The main aim of Domian is to attract callers who have something sensational and interesting to tell. Therefore, topics revolving around sexuality play a large role in the show, as they can easily be exploited as sensational issues. In the same way that the psychological counselling formats contain phrases with the word Problem (‘problem’) as contextualization cues which help to frame the situation of counselling, phrases with the word Thema (‘topic’) are used as key phrases in Domian’s programme. During the introduction to his programme and between the calls the word Thema is often used. Domian does not ask his callers to tell him their problems, instead he invites them to talk to him about the topics that interest them. Every time Domian refers to talking about topics as the aim of his programme, he sets the tone for the callers. By using the neutral word Thema (‘topic’), he avoids constructing a counselling situation. As the general setting of the programme is not one of a counselling situation, one also does not usually find phrases containing thanks for the advice given. In a typical closing sequence, the callers thank Domian for listening to them and for offering a good programme, but not for his advice.

5. Small forms interacting with larger ones: The “talking into being” of genres So far, I have been describing the prototypical features of the two formats. While the psychological counselling programmes follow a fixed pattern and can easily be described as being part of a sedimented genre, the talk programme is much less fixed because of the range of possible topics and the informal mode of interaction, both of which allow for a variety of other small genres (such as greetings to friends, praise for the host, and so on) to occur within the programme. Yet, the typical syntactic and prosodic features, the sequential structure, and the similarity of the sensational topics provide for at least a genre-like structure: Callers to Domian expect a chat (or talk) and not a counselling session. Sometimes, though, the two patterns of counselling programme and talk programme mix to generate hybrid structures. It is only the pattern for the talk programme, though, that is open to be transformed into a hybrid. The pattern of Von Mensch zu Mensch and Ratgeber Lebensfragen is so fixed that it is not open to such a change.

How syntactic patterns and genres interact

155

The question is: How do the interactants (Domian and the caller) manage the necessary “communicative adjustment of perspectives” (Luckmann 1992: 219) to settle on a new pattern? The solution lies in the use of “contextualization cues” (Gumperz 1982: 131) which signal the contextual presuppositions on which the interactants are relying. Due to the striking persistence of phrases such as Mein Problem ist (‘my problem is’) or Ich hab da ein Problem (‘I’ve got a problem’) in psychological counselling programmes, these phrases have the function of contextualization cues: They can be used by the callers to signal to Domian that it is not just a “topic” that forms the basis of the conversation but rather a “problem,” which demands a different genre geared not to chatting but to advice giving. The following example illustrates a case where a call at Domian’s programme starts out on the tracks of the radio talk pattern but is then transformed into a counselling pattern: Example 7

Domian: Christin’s call 1

01

chrisTIN sIEbenundzwanzig is jetzt dran. ‘it’s Christin’s twenty-seven turn now.’ guten MORgen; ‘good morning;’ mOrgen DOmian; ‘morning Domian;’ chrisTIN; ‘Christin;’ dein THEma. ‘your topic.’ hhhhhhhh also MEIN thema ich arbeite als (.) prostituIErte, ‘hhhhhhhh well my topic I work as (.) a prostitute,’ ja, ‘yes,’ .h und f¨ uhl mich ziemlich beSCHISSen auf deutsch gesagt; ‘.h and plainly speaking I feel quite fucked up;’ wie lange ARbeitest du schon als eh prostituIErte? ‘how long have you been working as a prostitute?’ seit einem JAHR jetzt. (.) ‘for a year right now. (.)’ .h das probLEM is ich habe (.) f¨ Unftausend euro schulden; ‘.h the problem is I have (.) debts of five thousand euros’;

DO

02 03

CA

04

DO

05 06

CA

07

DO

08

CA

09

DO

10

CA

11

The caller phones Domian during one of the “free” topic nights. Domian sets the frame of the talk by mentioning the word Thema (‘topic’, line 5: dein THEma/‘your topic’). The caller first ratifies the agenda as one of talking about (interesting) topics by taking up the word (also MEIN thema ich arbeite als (.) prostituIERte/‘well my topic I’m work as (.) a prostitute’). Domian then goes on

156

Wolfgang Imo

to ask for details and background information (line 10) and the caller suddenly uses the key phrase of counselling formats: das probLEM is (‘the problem is’) in line 12. After that, the conversation begins to show striking similarities to conversations found in dedicated counselling programmes: Even though Domian still sticks to a questioning routine whose aim is to elicit as many interesting details as possible, the conversation continues to centre on the problems of the caller (her debts to her pimp and her wish to stop working as a prostitute). Domian’s main focus is on helping her find solutions for her problems. Accordingly, we do not find any reproaches of the caller’s behaviour: She is taken very seriously, and Domian’s pattern of interaction resembles that of a counsellor. He even uses the counsellors’ formal strategy of reformulation – including the discourse marker also (‘so’), a phrase of try-marking (‘the way I understand it’), and the reformulation is offered in a question format – to signal to the caller that he is striving to gain a solid understanding of her situation: Example 8 153

DO

154

CA

155

DO

Domian: Christin’s call 2 .h also versteh ich es SO dass dass die HAUPT angst eigentlich DArin besteht dass jemand dir geWALT antut? ‘.hh so the way I understand it is that your main fear actually is that someone will hurt you?’ ja; ‘yes;’ und dass die SCHULdengeschichte da hast du AUCH angst vor aber eigentlich ZWEITrangig is; ‘and that your debts you are afraid of them but not as much;’

The conversation culminates in a sequence where Domian does his best to give advice, and the best advice he can give as a lay person is to refer her to the programme’s trained psychologist and to tell her to go to a credit counselling organization: Example 9

Domian: Christin’s call 3

175

.hh ich M¨ OCHte dir gerne etwas ANbieten; .‘hh I want to offer you something;’ ich m¨ ochte dir ANbieten dass du mit meiner psychoLOgin sprichst; ‘I want to offer you to talk to my psychologist;’

‘’ .h und eh wir werden dich mal verSORgen dann auch GLEICH mit eh mit entSPREchenden .h ANlauftelefonnummern ANlaufaDREssen, ‘.h and ehm we will provide you immediately with ehm with the right telephone numbers contact addresses,’

D

176

177

CA

178

DO

((...))

How syntactic patterns and genres interact 185

DO

186

CA

157

so und dar¨ uber hiNAUS auf JEden fall auch mal n termin bei der SCHULDnerberatung machen; ‘that’s that and above that you should definitely contact a credit counsellor;’ .h ja; ‘.h yes;’

The caller is not reproached for her situation but given the best help she can get from Domian. The whole participation status of this call has been transformed: The absence of reproaches, extreme formulations, and sensationalizing prosody, combined with the presence of a pronounced task-oriented attitude and the use of formal patterns of reformulations have transformed the standard pattern of talk into that of counselling. The importance of contextualization cues at the beginning of a call to its reinterpretation as a token of the genre counselling programme can be shown by an example where a caller never uses a phrase that includes the word problem throughout the whole conversation. Therefore, Domian quite obviously assumes the caller just wants to talk about her sexual affair with her father-in-law, rather than looking for advice on how to end the situation. The longer the talk goes on, the more Domian reproaches the caller and the less seriously she is taken by him. The following example shows one of his reproaches: Example 10

Domian: reproaches 2

121

ERSte frage zu dieser thematik wieSO: WENN du schon so etwas machst; ‘first question about this topic why if you have to do something like that;’ is nicht ordentlich verH¨ Utet worden; ‘ you contracept properly;’

DO

122

The strong reproach is signaled here by a range of formal syntactic means – the modal particle schon in line 121 (the word cannot be translated directly, a paraphrase would be something like ‘If you really had to . . . ’) and the question format (line 122), which is frequently used for reproaches in German (G¨unthner, this volume) – as well as strong accents and staccato speaking (see G¨unthner 2000 for a detailed analysis of the syntactic structures of reproaches in German). When the caller answers that she and her lover just did not think of contraception when they had sex, Domian implicitly calls her juvenile and irresponsible and questions her mental maturity: Example 11

Domian: reproaches 3

126

ja aber da MACHT man sich du bist doch auch ne‘yes but really you also are a-’

DO

158

Wolfgang Imo

127

du: bist doch auch ne erwachsene FRAU; ‘you also are a grown-up woman;’ da macht man sich doch geDANken (.) u ¨ber dieses thema doch nun WIRKlich; ‘one’s bound to think (.) about this topic really;’

128

Domian’s tone in this conversation is strikingly different from that illustrated in (7)–(9), where the hybridization of genres is managed successfully. In Christin’s call, the caller is never reprimanded for choosing an easy way of earning money, in spite of the fact that she claimed that this was her initial incentive for becoming a prostitute. This conversation at hand, however, terminates abruptly when the problem of differing perspectives is explicitly referred to by the caller after another reproach by Domian: Example 12 225

DO

226

CA

227

DO

228 229

230

Domian: reproaches 4 das is ja so SCH¨ Abig alles; .h eh‘everything is so cheap; .h eh-’ nur ich hab jetzt eigentlich nIch Angerufen dass ich vOrw¨ urfe krieg sondern eher n RAT von dir; ‘only I didn’t call you just to be reproached but rather to get an advice;’ jA der RAT, ‘yes the advice’, weißt du wie der RAT aussieht? (.) ‘do you know what the advice looks like? (.)’ der rAt sieht AUS eh leg die kArten aufn tisch und TRAge‘the advice looks like that put your cards on the table and bear-’ TRAge die verantwortung f¨ ur All das was du Angerichtet hast, ‘bear responsibility for everything you did wrong,’

At this stage it is too late to “talk into being” the genre of psychological counselling programme. Domian refuses to switch into a mode of giving serious, objective advice but continues to tell the caller that he is disgusted with her behaviour. Shortly after he ends the conversation: Example 13 251

252

DO

Domian: reproaches 5 jetzt hab ICH dir MEIne meine klare MEInung gesagt die dir jetzt vielleicht n bisschen bisschen KRASS erscheint, ‘now maybe I clearlytold you my opinion which may seem to you to be a bit harsh,’ .h eh du WEIßT ich hab im hIntergrund Immer einen psychologen; ‘.h ehm you know that I always have a psychologist at hand in the background;’

How syntactic patterns and genres interact 253

CA

254

DO

255

256

257

CA

258

DO

259

CA

260

DO

261

CA

262

DO

159

‘

DER hat dieses gespr¨ ach jetzt auch MITgeh¨ ort; he has also listened to this conversation; .h eh ich w¨ ur w¨ urde GERN dass du dich mit ihm AUCH mal unterh¨ altst; ‘.h ehm I woul would like to have you talk to him too;’ dann kann ER .h eh DIR seine meinung AUCH sagen zu dem fall; ‘then he can .h ehm give you a piece of his mind about your case too;’

‘

’ ok? ‘o.k.?’

‘

’ das heißt DU legst jetzt auf und mein psychoLOge der PEter ruft dich gleich an; ‘so you’ll now put the receiver down and my psychologist Peter will call you immediately;’ alles klar Ok; ‘all right o.k.;’ auf WIEderh¨ oren; ‘goodbye;’

Domian himself decides that he has had enough of this conversation. His apology for being ’n bisschen krass (‘maybe a bit harsh’, line 251) is only halfhearted, as can be seen in line 256, where he says – in an emphatic way – that the psychologist, too, will give the caller a piece of his mind, too. With this biased expression, Domian distances himself from any serious counselling. The conversation ends without a thanking sequence. Domian tells the caller to put the phone down (line 260), and they only exchange goodbyes. This misfire of a conversation, which was framed by Domian as talk and intended by the caller as a counselling session, illustrates the importance of formal constructions that operate as contextualization cues signalling what kind of pattern or genre the caller wants to activate at as early a stage of the conversation as possible.

6. Genres, constructions, and construction discourse It has been shown through the analysis of the two genres radio talk and radio counselling as well as their hybrid forms that syntactic constructions serve a twofold function: On the one hand, a phrase can signal that a certain genre

160

Wolfgang Imo

is about to be activated. On the other hand, that phrase may be an integral component of a genre (as part of its inner structure). The problem of the theory of communicative genres is that the relations between small constructions (for example, sentences or sentence patterns such as questions, phrases, or even words) and a whole genre have to involve a change in the frame of analysis: One has to switch between morphology, syntax, lexis, and genre as separate analytic domains. This complicates the illustration of the interrelatedness between small and large units. A theory that would be able to bridge the gap between these units by treating them all as equal instances could offer an analytic frame that can be used for all – or at least many – aspects of a genre. One solution might be supplied by the theory of Construction Grammar, especially a newly emerging branch which is empirically oriented and has been extended to the description of the syntactic structures of spoken language (Auer 2006; Deppermann 2006; G¨unthner 2006; G¨unthner and Imo 2006, 2007a, 2007b; Thompson 2002; Hopper 2004). One of the basic tenets of Construction Grammar is to view words, phrases, idioms, and sentences as basically the same thing: more or less complex signs which consist of a coupling of phonological material and a meaning or function and which have to be described holistically. Not only do the traditional two parts of a sign – meaning and form – have to be taken into account if one wants to describe a construction, but also prosody, sequential position, context, the preferred situation in which the sign occurs, and so on. Furthermore, constructions are not treated as isolated categories but as members of a complex network. Some constructions can have close connections, some may be further apart, and others may even merge (Imo 2007b). While the origin of Construction Grammar and its further development (Fillmore 1988; Fillmore, Kay, and O’Connor 1988; Goldberg 1996; Croft 2002; Langacker 1987) has been in the field of syntax, there have been attempts to ¨ extend the theory to units larger than sentences. Ostman (2005: 121) suggests the label “construction discourse” to describe complex patterns such as texts or discourses, and G¨unthner (2006) highlights the affinity between constructions and communicative genres. The advantage of speaking of constructions instead of words, syntactic patterns, and genres is that constructions can be crossreferenced: A phrase such as mein Problem ist (‘my problem is’) can carry the information that it occurs within counselling sessions, and a genre such as radio counselling programme can include the information that it contains phrases such as mein Problem ist/‘my problem is.’ Similarly, reformulations in question formats and the use of discourse markers can provide vital information about the genre those syntactic structures are part of. For the sake of clarity, I will pick out only the phrase mein Problem ist as an illustration of how the interaction

How syntactic patterns and genres interact

161

between syntactic structures and genres could be visualized. The entries under the genre radio counselling programme, as well as under the construction mein Problem ist (‘my problem is’), have been drastically reduced to keep the schema simple. In Figure 1, both constructions are interlinked because one occurs within the other: construction discourse: radio counselling programme sequential structure fixed pattern of greeting − telling of topic − reformulation − acceptance of reformulation − advice − acceptance of advice − exchange of thanks and goodbyes prosody calming way of speaking ... lexis/syntax no extreme formulations contains phrases such as mein Problem ist expert vocabulary reformulations in question format ... function giving advice for people with problems ... (…)

construction: syntax semantics function context (…)

mein Problem ist/my problem is fixed phrase ... referring to something one has troubles with ... signalling that one is going to talk about something one has troubles with and that one wants advice/help ... occurs in radio counselling programmes ...

Figure 1. Interaction between syntactic structures and genres

As is illustrated by the connecting lines, the entries in both constructions now refer to each other, or, in other words, both constructions contain each other. This explains why it is possible to “talk into being” a genre with the use of a simple phrasal construction. Words, phrases, and sentences are not the reduced units consisting only of semantic and phonetic material, as is commonly held, but contain a lot of additional information, such as the preferred contexts in which these words, phrases, and sentences occur. When such a construction as mein Problem ist is used within a token of a construction discourse that does not contain this construction as its typical feature (as in the case of radio talk programmes), the connection the smaller construction has with the larger construction discourse it is usually embedded in (i.e. radio counselling programme) establishes a link to this construction discourse. Thus, the possibility for switching over to other construction discourses or for blending both discourses into a new hybrid construction discourse can be explained and illustrated as in Figure 2:

162

Wolfgang Imo

construction: radio talk programme

construction: radio counselling protramme

seq. structure prosody lexis/syntax function

seq. structure prosody lexis function

fixed pattern emphatic mein Thema ist provide entertainment

construction:

mein Problem ist

syntax semantics function context

(described above)

(described above)

occurs in radio couns. programme

Figure 2. Linking of construction discourses/genres

When a caller utters the phrase mein Problem ist in an actual token of the construction radio talk programme, the connection of mein Problem ist to the construction radio counselling programme is provided via the contextual entry of the phrase. So, indeed, “linguistic knowledge includes far more than just a set of abstract rules and a lexicon; it also includes knowledge of sedimented constructions and their usage in particular genres” (G¨unthner, this volume: 210; see also Dorgeloh and Wanner, this volume). The same, of course, holds for the other syntactic patterns that are linked regularly to only one special genre of the two genres analyzed here, such as the complex syntactic patterns of reformulations (in the case of the counselling format) or reproaches (in the case of the talk format). The basic idea of Construction Grammar is to provide constructions with contextual entries and to conceptualize them as forming a network. This idea helps to illustrate how hybrids of genres can emerge and why small constructions have such an important role in creating these hybrids.

How syntactic patterns and genres interact

163

References Auer, Peter 2006

Auer, Peter 1995

¨ Construction Grammar meets conversation. Einige Uberlegungen am Beispiel von so-Konstruktionen. In: Susanne G¨unthner and Wolfgang Imo (eds.), Konstruktionen in der Interaktion, 291–314. (Linguistik – Impulse & Tendenzen 20.) Berlin: de Gruyter.

Reconstructive genres of everyday communication. In: Uta Quasthoff (ed.), Aspects of Oral Communication, 289–304. (Research in Text Theory 21.) Berlin: de Gruyter. Birkner, Karin and Friederike Kern 2000 Impression management in East and West German job interviews. In: Helen Spencer-Oatey (ed.), Culturally Speaking: Managing Relations in Talk across Cultures, 253–271. London: Cassell Academic. Boden, Deirdre 1994 The Business of Talk: Organization in Action. Cambridge: Cambridge University Press. Croft, William 2002 Radical Construction Grammar. Oxford: Oxford University Press. Deppermann, Arnulf 2006 Construction Grammar – eine Grammatik f¨ur die Interaktion? In: Arnulf Deppermann, Reinhard Fiehler and Thomas Spranz-Fogasy (eds.), Grammatik und Interaktion, 43–66. Radolfzell: Verlag f¨ur Gespr¨achsforschung. Dorgeloh, Heidrun and Anja Wanner this volume Introduction. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 1–26. (Topics in English Linguistics 70.) Berlin/ New York: Mouton de Gruyter. Fillmore, Charles J. 1988 The mechanisms of Construction Grammar. Proceedings of the Annual Meeting of Berkeley Linguistics Society 14: 35–55. Fillmore, Charles, Paul Kay and Catherine O’Connor 1988 Regularity and idiomaticity in grammatical constructions: The case of let alone. Language 64: 501–538. Goldberg, Adele 1996 Construction grammar. In: Keith E. Brown and Jim E. Miller (eds.), Concise Encyclopedia of Syntactic Theories, 68–70. New York: Pergamon. Gumperz, John J. 1982 Discourse Strategies. (Studies in Interactional Sociolinguistics 1.) Cambridge: Cambridge University Press. G¨unthner, Susanne 1995 Gattungen in der sozialen Praxis. Deutsche Sprache 3: 193–218.

164

Wolfgang Imo

G¨unthner, Susanne 2000 Vorwurfsaktivit¨aten in der Alltagsinteraktion. Grammatische, prosodische, rhetorisch-stilistische und interaktiveVerfahren bei der Konstitution kommunikativer Muster und Gattungen. (Reihe Germanistische Linguistik 221) T¨ubingen: Max Niemeyer. G¨unthner, Susanne 2006 Von Konstruktionen zu kommunikativen Gattungen: Die Relevanz sedimentierter Muster f¨ur die Ausf¨uhrung kommunikativer Aufgaben. Deutsche Sprache 34: 173–190. G¨unthner, Susanne this volume Grammatical constructions and communicative genres. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 195–217. (Topics in English Linguistics 70.) Berlin/NewYork: Mouton de Gruyter. G¨unthner, Susanne and Wolfgang Imo (eds.) 2006 Konstruktionen in der Interaktion. (Linguistik – Impulse & Tendenzen 20.) Berlin: de Gruyter. G¨unthner, Susanne and Hubert Knoblauch 1994 “Forms are the food of faith”. Gattungen als Muster kommunikativen Handelns. K¨olner Zeitschrift f¨ur Soziologie und Sozialpsychologie 4: 693–723. G¨unthner, Susanne and Hubert Knoblauch 1995 Culturally patterned speaking practices – the analysis of communicative genres. Pragmatics 5: 1–32. Heyd, Theresa this volume I know you guys hate forwards: Address pronouns in digital folklore. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 333–358. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Hopper, Paul 2004 The openness of grammatical constructions. Proceedings of the 40th Annual Meeting of the Chicago Linguistic Society, 239–256. Chicago: Chicago Linguistic Society. Imo, Wolfgang 2007a Der Zwang zur Kategorienbildung: Probleme der Anwendung der Construction Grammar bei der Analyse gesprochener Sprache. Gespr¨achsforschung – Online Zeitschrift zur verbalen Interaktion 8: 22–45. Imo, Wolfgang 2007b Construction Grammar und Gesprochene-Sprache-Forschung: Konstruktionen mit zehn matrixsatzf¨ahigen Verben im gesprochenen Deutsch. T¨ubingen: Max Niemeyer. Kabatek, Johannes, Philipp Obrist and Valentina Vincis this volume Clause-linkage techniques as a symptom of discourse traditions: Methodological issues and evidence from Romance languages. In: Hei-

How syntactic patterns and genres interact

165

drun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 247–275. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Klemm, Michael 1996 Streiten ‘wie im wahren Leben’? “Der heiße Stuhl” und “Einspruch!“ im Kontext der Personalisierung und Emotionalisierung des Fernsehprogramms. In: Bernd Ulrich Biere and Rudolf Hoberg (eds.), M¨undlichkeit und Schriftlichkeit im Fernsehen, 135–163. (Studien zur Deutschen Sprache 5) T¨ubingen: Gunter Narr. Krause, Daniel 2006 Beratung, Therapie oder doch bloß ‘Show’? Motivationen und Gratifikationen von Domian-Anrufern. Ph.D. Dissertation, Department of Sociology, University of M¨unster. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar. Stanford: Stanford University Press. Luckmann, Thomas 1986 Grundformen der gesellschaftlichen Vermittlung des Wissens: Kommunikative Gattungen. K¨olner Zeitschrift f¨ur Soziologie und Sozialpsychologie 27: 191–211. Luckmann, Thomas 1988 Kommunikative Gattungen im kommunikativen ‘Haushalt’einer Gesellschaft. In: Gisela Smolka-Koerdt, Peter M. Spangenberg and Dagmar Tillmann-Bartylla (eds.), Der Ursprung von Literatur, 179–288. (Materialit¨at der Zeichen.) M¨unchen: Wilhelm Fink. Luckmann, Thomas 1992 On the communicative adjustment of perspectives, dialogue and communicative Genres. In: Astri Heen Wold (ed.), The Dialogical Alternative, 219–234. Oslo: Scandinavian University Press. Mondorf, Britta this volume Genre effects in the replacement of reflexives by particles. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 219– 245. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. ¨ Ostman, Jan Ola ¨ 2005 Construction discourse: A proglegomenon. In: Jan Ola Ostman and Mirjam Fried (eds.), Construction Grammars: Cognitive Grounding and Theoretical Extensions, 121–144. (Constructional Approaches to Language 3.) Amsterdam/Philadelphia: John Benjamins. Puschmann, Cornelius this volume Thank you for thinking we could: Use and function of interpersonal pronouns in corporate web logs. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 167–191 (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter.

166

Wolfgang Imo

Sacks, Harvey 1992 Lectures on Conversation. Oxford: Blackwell. Sacks, Harvey and Emanuel A. Schegloff 1979 Two preferences in the organization of reference to persons in conversation and their interaction. In: George Psathas (ed.), Everyday Language: Studies in Ethnomethodology, 15–21. New York: Irvington. Schegloff, Emanuel A., Gail Jefferson and Harvey Sacks 1977 The Preference for self-correction in the organization of repair in conversation. Language 53: 361–382. Schramm, J¨urgen 1999 Eins Live, Domian: Die Leute rufen lieber uns an als die Telefonseelsorge. In: Monika Maaßen, Thomas Groll and Hermann Timmerbrink (eds.), Mensch versteht sich nicht von selbst, 23–33. (Kommunikationso¨ kologie 5.) M¨unster: LIT Verlag. Selting, Margret, Peter Auer, Birgit Barden, J¨org Bergmann, Elizabeth Couper-Kuhlen, Susanne G¨unthner, Christoph Meier, Uta Quasthoff, Peter Schlobinski and Susanne Uhmann 1998 Gespr¨achsanalytisches Transkriptionssystem. Linguistische Berichte 173: 91–122. Thompson, Sandra A. 2002 Constructions and conversation. Unpublished working paper, University of Konstanz. Willmann, Thomas 1996 Privates in der o¨ ffentlichen Kommunikation. In: Hubert A. Knoblauch (ed.), Kommunikative Lebenswelten, 201–214. Konstanz: Universit¨atsverlag Konstanz.

Thank you for thinking we could: Use and function of interpersonal pronouns in corporate web logs1 Cornelius Puschmann

This paper examines the function and use of first person and second person pronouns (inter-personal pronouns or IPPs) in Internet blogs maintained by several major U.S. companies for marketing and public relations purposes and evaluates their significance as a genre marker. After briefly assessing prior research on blogs in general and corporate blogs in particular, the linguistic properties of pronouns and pronominal distributions across different written genres are described. The paper analyzes usage data from a corpus of corporate web logs and develops an explanation of the function of IPPs in specific communicative contexts, especially that of corporate blogs. The evaluation of the results is followed by a concluding discussion of the factors that make pronominal expression highly characteristic of this emerging form of computer-mediated communication (CMC).

1. Introduction Despite the popularity of digital forms of communication and their academic exploration, linguistic studies of emerging registers and genres are not quite as prevalent as one might assume. Blogging – a term and a technology that have only been with us for a short period – is a much-discussed phenomenon in media studies and social sciences (see Schmidt 2007 for a detailed overview), yet linguists have been slightly hesitant to investigate this new kind of written expression.2 At first glance, blogs appear to share many similarities with pre-digital forms of writing, such as diaries and editorials. It is perhaps for this 1. I am indebted to Theresa Heyd, two anonymous reviewers and especially Anja Wanner for extensive feedback on both formal and conceptual aspects of earlier versions of this paper. 2. There are certain exceptions: See, for example, Herring and Paolillo (2006) for a discussion of blogs and issues of gender and genre analysis, a topic that has sparked interest in computational linguistics recently (Koppel, Argamon and Shimoni 2003; Schler, Koppel, Argamon and Pennebaker 2006). Also see Nowson, Oberlander and Gill (2005) for issues of style and author personality in blogs. Nilsson (2003) is one of the few attempts to assess blogs linguisti-

168

Cornelius Puschmann

reason that digital genres that differ more markedly from conventional genres of written language and can be described as conceptually more oral than blogs have received more attention in writings about computer-mediated communication (CMC). Following the assumption that the systematic study of a new form of CMC can contribute to an improved overall understanding of language use, this paper examines a single linguistic phenomenon in a single emerging genre: the use of inter-personal pronouns (IPPs)3 in corporate blogs. Personal pronouns as the realization of person deixis serve as a natural bridge between language and environment, making them an ideal object of study for investigations into register. This paper approaches the object of study from two principal directions: After sketching the development of blogs (section 2) and corporate blogs (section 3) and providing a description of pronouns and a survey of pronominal usage across existing genres (section 4), I then turn to an analysis of interpersonal pronouns in corporate blogs (section 5) and make concluding observations based on my findings (section 6). This research is based on a corpus of corporate web log texts (roughly 5 million word tokens), assembled between August 2006 and December 2007 in the context of my doctoral thesis project.4

2. Situational and functional properties of blogs Web logs (frequently clipped to “blogs”) are a novel form of online publishing that has gained significant popularity in recent years (for an overview of blog history, see Blood 2000, 2006). While blogs can be used in a multitude of ways, to many users they serve functions similar to those of a personal diary (McNeill 2005). The following example is a post from what can be characterized as a typical journal-style personal blog. Text (1) “My Birthday,” 19 February 2007 Well, it was my first birthday away from the family. It was interesting . . . teasin! it wasnt as bad as i thought and i practiced for my mission. I woke up and texted sis cally. However, her assessment uses a relatively small corpus and examines a limited number of features. 3. I adopt the term from Wales (1996), who uses it to denote pronouns of the first and second person in English. 4. See Puschmann (2010) for a detailed description of the data. The corpus assembled for my PhD project (henceforth abbreviated as CBC/Corporati) consists of 25,476 posts (5,356,486 word tokens) from 137 corporate blogs maintained by large and medium-sized U.S. companies. It was assembled between April 2, 2006, and December 15, 2007.

Interpersonal pronouns in corporate web logs

169

because she was born on my b-day, for those of you who forogt. Teasin sis! Umm i talked to mom and dad and the rest of the family throughout the day. Thanks for the calls everyone. Uhh . . . i went to breakfast (the good ol Galley) with Lizzie and Kenzie and afterwards ran home to find a package sent from Mom and the famoly in cali. I went to a basketball game and we won. Then i went out to dinner with a bunch of friends and ate hecka food. Afterwards we came back to Lizzie’s and opened presents from Lizzie and Max. Max gave me a journal his mom gave him . . . that was special and i needed one so im thankful for it. Lizzie gave me the Spanish Book of Mormon. I was pretty excited cause now im gonna be fluent in Espanol! She also gave me an awesome CTR ring because I have been wearing Max’s old one for the last year. Umm it also spins so it is way tight. Then I came back to my dorm and Max and I opened the rest of my presents. The Abillas gave me some nice notes and hecka food which is gettin me fat. haha thanks! O and thanks Ang for the Head and Shoulder bottle. You’ll be gettin it back sometime haha! And then I opened the fam in Cali’s package. I got a sweet shirt that had my Mission on the back. Kaley gave me the hymn book in Spanish, its pocket sized. And then a CD and hecka more food. Grandma also sent a letter. Shes awesome! And that was my b-day! This might be boring sorry!

It should be stressed that the narrative style in this example is only one form of blogging common in personal blogs, although narrative is, as Virtanen (this volume) points out, the most prototypical text type. Blogs can also be used to filter and collect hyperlinked references to other sources on the web (i.e. filter blogs, see Herring et al. 2004: 3), to discuss issues of public interest (politics, news) and to engage others in debate. Personal blogs are distinguished from uses in settings such as news reporting, education, and business by the fact that the individual blogger may freely assign a topical focus with each entry. By contrast, when blogging is mandated in an organizational or professional context, it can be argued that some sort of conscious strategic goal is always associated with the activity. One of relatively few linguistic descriptions of blogs comes from David Crystal (2006: 15): [The blog] takes the form of a personalized web page where the owner can post messages at intervals. Many blogs are personal diaries, ranging in length from brief notes to extended essays; many are on topics of general interest or concern, such as a hobby or political issue. Some blogs are monologues; some have shared authorship, some are interactive.

The description shows that several paradigmatically different definitions can be given to explain what a blog is: a technical one (as a publishing technology), one of formal convention (as a genre, personal diaries) and one focused on its

170

Cornelius Puschmann

users.5 Crystal makes quite a sweeping generalization in his description of the production and resulting stylistic aspects of blogs: From a linguistic point of view, they have one thing in common: the written language is unmediated. The page you are reading now is a remove away from the text that originally appeared on my computer, because it has been regularized by the publisher’s copy-editor and checked for house-style consistency by a proofreader. That is the fare of virtually all published material [. . . ] but in blogging there is no such intermediary. The language of blogs displays the process of writing in its naked, unedited form. For precedents, we have to go back to the Middle Ages, before standard English evolved. (Crystal 2006: 15)

It is noteworthy that blogging is on the onset compared with other forms of publishing. By contrast, chats and instant messaging are not regarded as publishing, most likely due to their synchronous and ephemeral character, but from a purely technical point of view this makes little difference. Chats are produced in real time, not necessarily archived by default and not always available to a general public. Blogs, however, are generally more asynchronous than synchronous (but consider microblogging formats such as Twitter), they exist in a persistent form on the Web and can be found via search engines. The special characteristic of blogs that Crystal highlights is thus that they have the potential to act as publishing media (comparable to books, magazines, newspapers etc.) but, unlike books, they are (according to him) produced in an “unmediated” way, without any editorial oversight or stylistic polishing. While this may apply to a prototypical personal blog, as illustrated in text (1), it is problematic to generalize such attributes to the entire genre, which is probably why other scholars favor safer (i.e. more technical) definitions. The way in which standard software tools for blogging function and the dynamic of blog conversations – exchanges between bloggers who quote and link to each other – both work to structure and constrain blogs in regards to their visual presentation and the language they use. While they are topically highly diverse, the fact that blog entries are dated and attributed to the blogger, that they are presented in reverse chronological order, and that they are stored under an unchanging web address (the permalink) for archival purposes is not a conscious choice of the blogger but is preconditioned by the software he or she

5. Many definitions highlight the technical aspects of blogs since only they can be regarded as fully stable. For example, Herring et al. (2004: 101) call blogs “frequently modified web pages in which dated entries are listed in reverse chronological sequence.” Other definitions place stronger emphasis on their use: Schmidt (2007) describes blogs as “a tool for information, identity and relationship management.”

Interpersonal pronouns in corporate web logs

171

uses.This fact is relevant because the schematic presentation of text together with contextually salient meta-data enables the use of context-dependent language. While theoretically a blogger can avoid the use of explicit self-reference, such a style creates a discrepancy, because he or she is per default credited by name for each entry. Since the writer of a blog post is also the publisher, it is generally assumed that the person to whom an entry is credited must also be the I of the post if any self-reference occurs. Conversely, the blogger is in a sense encouraged to use self-reference by the contextual factors: As the “owner” of the blog, he or she is not forced to discuss any specific topic and may write about whatever he or she likes; only the blogger has access to the publishing software, and every single post automatically identifies the blogger as the author. The cognitive analogy with (spoken) discourse is drawn because the blog is effectively an extension of the self in textual terms. As Crystal points out, a blog may be monologic (if there is no addressee) or dialogic (if there is). A similar description comes from Nilsson (2003: 31): “It could be argued that blogs combine both the monologue and the dialogue in a space-bound, electronic environment. They are simultaneously self-reflective thoughts presented publicly, and continuous conversations.” Lyons (1977) suggested that face-to-face interaction can be regarded as the generic template onto which virtually all kinds of mediated interpersonal communication (such as telephone conversations or letters, and, by extension, Internet chats) are conceptually mapped. The contextual information that the blogging frame provides – dated, attributed and archived entries, “about” pages that describe the blogger, hyperlinks to other pages and the use of deictic expressions – make it possible to conceptualize a blog entry as a sort of uninterrupted, planned, organized and persistent conversation. These surroundings evoke the concept of the deictic center first proposed by Karl B¨uhler ([1934] 1990). The deictic center encodes an origin (or origo) of space, time, and person as the point of orientation of all deictic expressions. In spoken discourse the deictic center can clearly be located by all discourse participants, but this is not generally the case in written genres where the reader must construct it from the text and from contextual clues, such as his or her knowledge of the conventions of the genre. While the deictic center may shift in third-person literary narratives – spatially, temporally and in terms of actors (see Rapaport et al. 1989) –, blogs differ both from pre-digital and other digital forms of publishing in that they always provide a fixed set of contextual clues to the reader and in that the origo is always with the writer. The meta-data provided with a blog entry follow B¨uhler’s origos of space, time, and person. Only the category space is omitted, but since the blog itself is frequently conceptualized

172

Cornelius Puschmann

as a virtual place this is not problematic.6 The deictic center in blogs can be regarded as stable in terms of who the blogger is and when a particular entry was posted, but as shifting with regard to the addressee. Apart from blog deixis, two other aspects support the conversation metaphor cognitively: Blogs permit feedback via comments and trackbacks and they are perhaps the sole form of mass publishing media that can be maintained by an individual.

3. The corporate blog as a domain-specific genre The corporate blog as a domain-specific type of blog can be considered a fairly recent innovation, although in recent years a steadily growing number of major international companies has taken up the practice. Reasons for the interest in blogging and other forms of social media are numerous: Blogs can be used internally for knowledge management, to improve communications and simplify project management, and externally for purposes such as recruiting, marketing, and public relations. A significant selling point of blogs for corporations, however, is specifically their function as persistent conversations between individuals. The paradigm of the social web is problematic for organizations in the sense that they are not equivalent to (human) social actors and are perceived accordingly. Blogs allow companies to personalize their communications – an important measure in a medial world that is becoming increasingly individualized. While the aims of independent private bloggers are highly diverse and may be reasonably unspecific and self-focused (recognition, friendship, fame, recording one’s thoughts), the goals of corporate bloggers can be assumed to be precisely defined. The goal of public relations is generally to positively influence how an organization is perceived by the public and it can be assumed that this holds true for company blogs which are created for that purpose as well. In order to achieve such a positive effect, however, it is crucial that company blogs are perceived first and foremost as real blogs and only secondly as material published by a corporate entity. Applying the concepts of communicative purpose and cognitive schemata (Virtanen, this volume), it can be argued that corporate blogs formally follow the schema of personal blogs, but are more constrained and strategically aligned with regard to their purpose. Since they follow the conventions of personal blogs 6. In terms of person, the deictic center is virtually always the blogger, the temporal deictic center is generally coding time, and the spatial deictic center is either the place of coding (the location of the blogger at the time of writing) or the blog itself, conceived as a virtual place. The (older) genre of the personal homepage seems to differ in this respect (see de Saint-Georges 1998). Spatial metaphors are generally prevalent when conceptualizing the Internet, making blogs virtual personal spaces where people can think aloud.

Interpersonal pronouns in corporate web logs

173

closely and since blogging technology allows the use of self-centered deictic expressions, corporate blogs do not differ significantly from other kinds of blogs in their principally high frequency of interpersonal pronouns, a characteristic that markedly places them apart from other forms of business discourse, where IPP usage is generally low. Table 1 shows the 20 most frequent tokens in a corporate blogging corpus of 5 million words (CBC/Corporati) vs. the 20 most frequent tokens in the written section of the British National Corpus.7 Table 1. Most frequent tokens in CBC/Corporati and the written section of the BNC Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

TokenCBC/Corporati the to and of a in I is for It you on with this that we are be at was

Part-of-speech tag DT TO CC IN DT IN PP VBZ IN PP PP IN IN DT IN PP VBP VB IN VBD

TokenBNC(written) the of and a in to is to was it for with he be on I that by at you

Part-of-speech tag AT0 PRF CJC AT0 PRP TO0 VBZ PRP VBD PNP PRP PRP PNP VBI PRP PNP CJP PRP PRP PNP

The pronouns of the first and second person (labeled with the part-of-speech tag PP in CBC/Corporati and PNP in the BNC and referred to in this paper as 1PP and 2PP) rank clearly higher in CBC/Corporati, although their proximal distance to one another is the same (I and you are four positions apart). This is a notable contrast, especially when considering the overall similarities of the two lists, one of which (the BNC) is representative of a wide range of genres, while the other one is based on a sample from a new and highly specialized subgenre 7. The list of the most frequent words in the BNC (all sections) reproduced in Table 1 has been adopted from Kilgarriff (1998). Also see Kilgarriff (1998) for a description of the BNC tagset.

174

Cornelius Puschmann

of computer-mediated communication. The interpersonal pronouns of the first and second person significantly outrank those of the third person in blogs, while the situation is reversed in the BNC, where he outranks I and you. Notably, nonpersonal it occupies the exact same slot in both lists. The overall impression of pronominal egocentricity in blogs is confirmed when comparing specifically a list of the 20 most frequent pronouns in CBC/Corporati and the BNC (Table 2). Table 2. Most frequent pronouns in CBC/Corporati and the BNC Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Pronoun-CBC/Corporati I it you we they he me them us she him myself her one itself themselves yourself himself ourselves herself

Pronoun-BNC(all sections) it I you he they she we who them him me her one us something nothing himself anything itself themselves

A close comparison reveals several interesting details: – The use of interpersonal pronouns dominates in CBC/Corporati, whereas third-person reference is more prevalent in the BNC. (This is not affected by the relatively high frequency of I and you in the BNC – full noun phrases are still much more frequent than pronouns in the BNC.) – The rank sequence of the IPPs I, you, we, me, us, myself, yourself, ourselves is identical in both corpora, with all instances ranking equally or higher in CBC/Corporati.

Interpersonal pronouns in corporate web logs

175

– The significant role of interpersonal pronouns in CBC/Corporati is the only notable difference between the two corpora – otherwise they are relatively similar. The virtual context of blogging thus seems to mirror that of a real-world conversation to the extent that both are egocentric and assume a temporal and spatial orientation that is transparent and accessible to all participants. As outlined, a blog can be regarded as a technological frame in which the blogger linguistically creates a self-focused discourse situation that allows him or her to address a variety of communicative partners. This mediated environment shares some similarities with a face-to-face conversation, the planned and structured elements of written language not withstanding. Importantly, and in significant contrast to spoken language, bloggers can voice their thoughts, opinions and experiences in an uninterrupted manner, without having to cede the floor to a communicative partner, and they can plan and organize their language before publishing a post. What blogs share linguistically with spoken discourse is an ego-centered orientation that uses the spatial and temporal context to define a deictic center. The here and there of the blog are hyperlinked, the now and then are specified by time stamps associated with blog entries, and the I and you of the blog discourse are blogger and reader, although both self-reference and reference to others can be used with different meanings. While blogs exist that are topically introspective and linguistically monologic (they have 1PP, but no 2PP), this kind of writing is relatively rare in corporate blogs – after all, a company seeks to present itself effectively to an outside readership, making audience-oriented writing an important goal.

4. Interpersonal pronouns in English Before conducting a more detailed analysis of personal pronoun use in corporate blogs, I will present an overview of their properties in English as such and across different genres. Most linguists classify pronouns as a subtype of noun (e.g., Huddleston and Pullum 2002). Pronouns are able to act as the head of a noun phrase in the major noun phrase positions and they are distinguished by their inability to take determiners. English personal pronouns retain several levels of grammatical information, such as case, number and gender, not all of which are marked morphologically on full nouns. The tendency to reserve specific pronouns for different types of discourse participants (self, addressee, others) and to encode information relating to categories such as gender and social status in them exists

176

Cornelius Puschmann

in many languages and points to the cognitive salience of these factors. In language acquisition, gaining an understanding of pronouns as contextual pointers that may shift their reference with every turn of a conversation comes relatively late in the development of children (around age three, according to Wales 1996). Pronoun systems show significant variation across languages and dialects, e.g. by marking inclusive and exclusive pronominality, dual and trial number, social deixis and other categories that English does not explicitly indicate. The pronoun paradigm below reflects only standard Modern English and omits non-standard varieties, many of which morphologically encode more grammatical information than the standard permits (e.g. American English y’all marking plural on the second person; see also Heyd, this volume). Table 3. Personal pronouns in standard Modern English Singular Subject First I Second You Masculine He Third Feminine She Neuter It

Plural Object Reflexive Subject Object Reflexive me myself we us ourselves you yourself you you yourselves him himself they them themselves her herself it itself

While the term pronoun suggests that they stand in for nouns, this needs qualification. Most pronouns can be used either deictically or endophorically, and in the latter case the pronoun effectively replaces the entire noun phrase, a form of pragmatic economy that “enhances communicative efficiency by avoiding needless repetition” (Langacker 1987: 490). When used deictically, however, a pronoun does not replace anything, rather, it is usually the default referring expression. Thus while Cornelius Puschmann and I would semantically be coreferential when used by me, both the full nominal and the third person pronoun would be highly marked if I were to use them to refer to myself in actual discourse. This is observed by Wales, who remarks: The 1PP and 2PP are characteristically used in the situational context, and refer normally to human beings in a ‘dialogue’, the speaker (‘I’) and addressee (‘you’): properly, ‘inter-personal’ pronouns. It is harder here to see their function linguistically as substituting for a noun, since the speaker and addressee can be referred to by an infinite number of nouns, depending on their known or perceived attributes: woman, mother, teacher, harridan, wit, etc. [. . . ] the choice of terms depends on the speaker’s perspective, and is more varied than applies to the 1PP. (Wales 1996: 3)

Interpersonal pronouns in corporate web logs

177

While use of 3PP can thus be understood largely via economic principles of linguistically structuring information that applies in more or less any situation of language use, 1PP and 2PP are tied intrinsically to the roles filled by speakers in concrete discourse situations (in other words, in situations where a deictic center exists). In such settings, their use is the default and use of full noun phrases is generally marked. Exceptions are restricted to specific contexts, such as particular speech acts, as in I, Cornelius Puschmann, hereby solemnly swear, when swearing an oath. The simple example functions both as a first-person speech act and as a third-person, context-detached declarative statement. The full noun phrase is inserted to neutralize context as a required source of information – to extract the deictic center, in B¨uhler’s terms – and through this it is made perfectly explicit who I is beyond the situation in which the oath was sworn. At the same time, 1PP is not entirely omitted in favor of the full noun phrase because this would prevent the oath from functioning as a speech act. Essentially, when speakers refer to themselves in the third person they disembody themselves from their discourse role, just as referring to a discourse partner via 3PP or full noun phrase ejects him or her from the discourse, a linguistic behavior that is regarded as antisocial in most contexts. The basic distinction between self, discourse partner and third persons is thus clearly reflected in the pronominal paradigm and applied even before the mechanisms of linguistic economy (see Lyons 1977, who describes deixis as the more basic pronominal reference than anaphora). From a cognitive point of view, it is noteworthy that 1PP and 2PP are perfectly mono-referential in spoken discourse: Both Cornelius Puschmann and he can hypothetically refer to more than one individual, but in a face-to-face conversation I and you unambiguously point to the speaker and the addressee(s) in the instance they are used. Because the spatial and temporal footing is much less solid in a computer-mediated discourse environment, the deictic anchoring of IPPs (especially 1PP-pl and 2PP) is significantly less rigid than in face-to-face conversation – a fact that is consciously exploited by many bloggers.8 Person as a grammatical category is generally interpreted as being structured as a strict hierarchy: If the speaker himself or herself is included in those referred to, use of 1PP-pl is the next plausible option following 1PP-sing. After 1PP-sing as the default IPP, the pronominal circle of reference extends outwards, from the speaker to the speaker plus others, to an addressee (minus the speaker)

8. Note that there is a gradual decline of precision when it comes to how clearly the referring IPP is linked to its referent. While 1PP-sing clearly identifies the blogger, 1PP-pl can be coreferential (see section 5). Use of 2PP can refer to discourse partners but also to indefinite referents (see also section 5).

178

Cornelius Puschmann

plus others, to third parties. Pragmatically, 1PP-pl is frequently an attractive alternative to 1PP-sing, as it creates a basic dichotomy between the speaker and all those with him or her (physically, or sometimes only rhetorically) on one side and all others persons on the other. The following example demonstrates the versatility of 1PP-pl in this respect: Text (2) Cenk Uygur, The Huffington Post, “The Republicans Lost Iraq” (excerpt), 2 December 2006 http://www.huffingtonpost.com/cenk-uygur/therepublicans-lost-iraq b 35432.html (last accessed on March 30, 2010) As Fareed Zakaria pointed out in his last column in Newsweek, there were two consecutive days in Baghdad recently where we protected the Mahdi Army one day and fought them the next.We are alternately protecting and fighting the Shiites and the Sunnis. We are clearly caught in the middle of a chaotic civil war. It is impossible to “win” in this situation.

Journalist Cenk Uygur’s blog is part of the liberal news and blog aggregator site The Huffington Post (www.huffingtonpost.com) and the text in (2) is one of many contributions critical of the Iraq war. In the excerpt, Uygur uses 1PP-pl in the function of a national we – it denotes all citizens of the United States collectively, which may or may not include the reader. This is one of many examples where 1PP-pl is extended beyond a physical situation while also being non-anaphoric (the reader must infer to whom we refers) and it exemplifies how a skillful writer can seek to create solidarity with his or her readership through the choice of the appropriate pronouns. As a later example, text (3), will show, this usage potential is especially relevant in texts where the intended referent appears to shift from one instance of 1PP-pl usage to the next. A specific aspect of mediated uses of 1PP-pl is brought to the surface in (2): The distinction between inclusive and exclusive uses of the pronoun is often difficult to recover (i.e. the blogger’s intended reference may not always be clear) and whatever his or her intention, the meaning can shift in either direction at the time of reading, depending on who the reader is. While anaphoric uses will exclude the reader unless he or she has been explicitly named, virtually all purely deictic uses of 1PP-pl potentially include the reader and at the same time depend on the reader’s readiness to consider himself or herself as included. A similar dynamic can be observed with regard to 2PP. While most uses of 2PP exclude the speaker, its scope can cover a range of referents as well, as text (3), taken from a pregnancy blog (“The Baby Blawg”), illustrates:

Interpersonal pronouns in corporate web logs

179

Text (3) “Natural ‘Birth Control’,” 21 August 2006 Doctors need more training in natural birth control. Well, duh. Didn’t need a study to tell me that one. Point 1: Most people refer to “natural family planning” or “natural family spacing” (or even “fertility awareness”) rather than “natural birth control”. That is not a term that you hear often. Point 2: I think it’s hillarious when people say “you can get pregnant while breastfeeding, you know”. Well yes, but if your baby is under six months old and you are exculsively breastfeeding (and not sleep-training) you are about as likely to get pregnant as if you are on the pill. But you don’t hear people warning each other “you can get pregnant on the pill, you know”.

The sample, taken from a pregnancy blog, exemplifies use of 2PP to refer alternately to the blogger, the reader, and to a generic third person. Potentially, the blogger is referring purely to herself (the blogger self-identifies as a woman in her profile), but use of 2PP actively involves the reader without risking the kind of logical confusion that might arise if 1PP-pl were used. As English does not mark 2PP for gender, number or social distance (i.e. T-V distinction9 ), it achieves maximal scope in blogs, where, like 1PP-pl, it can potentially address almost any reader. As outlined, pronouns of the third person can for several reasons be regarded as distinct from IPPs because of the status of the former as true noun replacements versus the status of the latter as distinct realizations of discourse roles. In instances of language use where no temporal and spatial orientation is indicated, the use of such discourse roles can be confusing and ambiguous, which explains the low overall frequency of IPPs in written language. It is only on a second level that ways in which the issue of a certain kind of content conditions the frequency of pronouns comes into play: Utterances that involve the speaker and addressee inevitably contain more pronouns, whereas statements about third parties contain fewer. Since there is no definite discourse partner and since we often find what amounts to self-omission in many kinds of written texts, everyone referenced in such writings is essentially a third person. Finally, deliberate rhetorical distancing plays an important role in genres that seek to embody objectivity, that is, the ejection from any single context of interpersonal interaction, such as academic writing, legal texts, news reporting and instruction manuals.10 It should 9. Many languages encode a distinction of social intimacy or distance between discourse partners in 2PP (e.g. Tu – Vous in French; Du – Sie in German). For an analysis of social deixis and the T-V distinction in 2PP use, see Brown and Gilman (2003). 10. Television news are exemplary in this regard: While there is arguably a technologically mediated discourse situation, frequently discourse roles are not assigned, and there is no use of interpersonal pronouns. On a second level (and in more recent usage) 2PP is used to address

180

Cornelius Puschmann

be noted that the discourse role of outsider that 3PP inevitably assigns is purely rhetorical in blogs. Since both we and you are flexible in their reference when used in a blog (where there is no physical discourse situation), they can always be interpreted to include the reader, unless the reader is explicitly named and designated as a third person. This is not possible in face-to-face conversation, where the situation shapes the deictic configuration to a large extent. Across genres, explicit self-reference of the author both marks involvement and suggests a degree of self-presentation on his part. In an early study, Chafe and Danielewitcz (1987: 23) observed that: If people have a natural inclination to talk about themselves, it is apparently in letters that they have the best opportunity to do so, being freed of whatever inhibitions might be imposed by the immediate presence of an interlocutor. The use of first person pronouns is thus not necessarily a feature which differentiates spoken from written language, but rather a feature which the absence of a direct audience may even foster when the circumstances are right. At the same time, as we can see from the figure for academic papers, writing can create a context in which maximum suppression of one’s own identity is possible.

Three factors help explain the relative lack of IPPs in academic writing: (a) the absence of an immediate discourse situation, (b) the non-involvement of the speaker in what is verbalized, and (c) the conscious motivation to suppress the self. The high frequency of self-reference in letters that Chafe and Danielewitcz (1987: 23) notice can similarly be explained via the combination of speaker involvement, the interpretation of the text as an uninterrupted chance to express oneself, and a lack of need to suppress self-centered language. In essentially the same way as in letters, the pronominally reflected self-suppression in academic writing stands in marked contrast to the self-centered mode of blogs (as discussed in detail in section 5).11 This has important consequences specifically for corporate blogs, which seek to personalize the discourse between an institution and individuals. Whereas academic discourse seeks to distance the author from the text and thereby to objectify the content, corporate blogs aim to achieve the opposite. Similar results in terms of underrepresentation of personal pronouns in academic writing and, to a lesser extent, news texts, are visible in Johansson’s analysis (1985) of the Lancaster-Oslo/Bergen Corpus (LOB). However, the picthe audience and in some contexts self-reference is also possible. News formats illustrate the duality that shapes pronominal distribution: At the same time, news formats seek to appear objective and the subject matter is markedly detached from the speaker as a person. 11. See Miller and Shepherd (2004), who point to early blog researcher Rebecca Blood’s (2000) characterization of blogging as “an outbreak of self-expression,” which Blood relates to the reverse chronological ordering of posts that is conventional in blogs.

Interpersonal pronouns in corporate web logs

181

ture is slightly more nuanced: While Chafe and Danielewitcz examined selfinvolvement only, the analysis of LOB took reference to other actors into account. Johansson (1985: 31) notes: With all these forms, except we, the fictional texts have a much higher frequency than the other category groups. [. . . ] The personal pronouns do not all behave in the same way, however. The frequency differences are especially large with I and you, he and she. This is a reflection of the proportion of dialogue as well as of subject-matter; fictional texts contain a great deal of dialogue (hence, the high frequency of I and you) and typically deal with persons and their experiences (hence, the higher frequency of he and she).

While Johansson attributes the high 1PP frequency to the prevalence of dialogue, his description of fiction containing much language related to “persons and their experiences” is significant, albeit with qualification. Blogs, for example, are centrally concerned with persons and their experiences, yet these are presented from the blogger’s point of view (1PP) to a discourse partner (2PP), with the third person that describes others used much less frequently. It is significant that a presentation from this subjectified perspective is possible in blogs, whereas it is not possible in many other genres, where the relevant contextual information is unavailable.

5. Use of interpersonal pronouns in corporate blogs As I have shown, the blog is a form of publishing that echoes certain situational parameters of spoken discourse by providing meta-information on the speaker and the time of coding, thereby constructing a deictic center that can be retrieved by the blog reader. Furthermore, blogs can be characterized as planned, structured and uninterrupted speech in the sense that they are maintained by individuals who are not restricted to a single topic and not subject to the constraints of on-line speech production. Based on these circumstances, the following hypotheses can be formulated: – Bloggers are likely to use egocentric deictic reference because this allows for an easier conceptualization of blogging in terms of spoken discourse, which is a familiar form of communication to both the blogger and the readers. – Bloggers are likely to feature prominently in their own discourse, either explicitly (via 1PP-sing) or more implicitly (via 1PP-pl, 2PP, or other strategies). – Bloggers are likely to conceptualize blog entries on a continuum between monologue and dialogue, because the fluid referentiality of 2PP allows them to frequently shift focus, in extreme cases from one clause to the next.

182

Cornelius Puschmann

Corporate blogs, specifically those that place emphasis on public relations, seek to establish a tri-part relationship between the writer of the blog, its readership, and the company as a whole. Therefore, the basic assumptions that hold true for blogs in general must be extended to reflect this dynamic: – Corporate bloggers can be assumed to address readers more explicitly than the more introverted and diarist-like personal bloggers, due to their purposeful and goal-oriented nature. – Corporate bloggers are likely to consciously exploit the inclusive rhetorical potential of 1PP-pl in order to set off the company as a unitary whole and contrast it with third parties. – Corporate bloggers similarly are able to use to rhetorical potential of 2PP to address both a generic readership and very specific focus groups (such as clients, investors, or competitors). Text (4), in which Jonathan Schwartz, CEO of Sun Microsystems, addresses both employees and external stakeholders, illustrates the full potential of corporate blogs in this regard: Text (4) Jonathan Schwartz, Jonathan’s Blog, “We Think We Can,” 30 July 2007 http://blogs.sun.com/jonathan/entry/we think we can (last accessed on March 29, 2010) Paraphrasing Henry Ford, “You think you can, or you think you can’t – either way, you’re right.” That quote struck me as the perfect summary of our fiscal year 2007 performance. We did what we said we’d do a year ago. As you may have seen, we’ve announced our fourth quarter and full fiscal year results (our fiscal year ends with the school year, in June) . We exceeded the commitments made a year ago, to restore Sun to 4% operating profitability in Q4, and did so by delivering our single best operational quarter since 2001. On an annual basis, we improved Sun’s profitability by over a billion dollars. A billion. We grew revenue, expanded gross margins, streamlined our operating expenses – and closed the year with an 8% operating profit in Q4, more than double what some thought to be an aggressive target a year ago. We did this while driving significant product transitions, going after new markets and product areas, and best of all, while aggressively moving the whole company to open source software (leading me to hope we can officially put to rest the question, “how will you make money?”) .And we’re not done – not by any stretch of the imagination. We have more streamlining to do, more commitments to meet, more customers to serve and developers to attract. But it’s evident we’ve got the right foundation for growing Sun – with real innovation the market values, as shown by Q4’s 47% gross margins, the highest on record in five years. I’ll be with a variety of external audiences most of this week – and I’ll summarize their questions and comments in a few days.

Interpersonal pronouns in corporate web logs

183

In the interim. . . to our customers, partners, and most of all, our amazing global employee base – thank you for thinking we could. You were right. Keep thinking that way. You ain’t seen nothin’ yet.

As in (2), Schwartz uses 1PP-pl with rhetorical extension to refer to all employees of the company, including himself. But use of 1PP-pl in this function does not prevent him from using 1PP-sing to present himself as distinct from everyone at Sun, nor to use the full noun Sun as denoting something other than we. Instead, an intricate interplay of discourse roles is developed: – Use of 1PP-pl, used as subject with a range of dynamic verbs of action and movement and semantically as agent, denotes the employees of Sun Microsystems (or, in a more restrictive interpretation, the management team). – 1PP-sing is used in contexts where agency of the institutional we would not result in a semantically well-formed proposition. E.g., Jonathan Schwartz will summarize what external audiences have to say, since this is something that the collective company is not literally capable of doing. – 2PP can denote the single blog reader (and all readers collectively), any single employee of Sun (and all employees collectively), and customers and partners in those contexts where Schwartz is the speaker, while it links back to him and the company in cases where he is paraphrasing the speech of others (how will you make money?). – Sun, the company name, is used only three times and always in object position (restore Sun / grow Sun) or as part of a noun phrase in object position (Sun’s profitability), while 1PP-pl dominates in subject position. – The indefinite pronoun some is used once to explicitly refer to a third party, as is the expression external audiences to describe people who are not employees of the company. The rhetorical significance of institutional we becomes clear when one examines its verbal collocates, for example say, do, announce, grow, improve, and exceed, all of which are used transitively. The company’s annual earnings have not been announced by an indefinite collective of employees, but by the management team, and what we said we’d do a year ago refers to a controversial plan to refocus Sun formulated by Schwartz. But it is obviously of strategic value for the CEO to downplay his role and instead to present the company’s success as the result of a team effort (using we instead of I ). At the same time, the use of we is preferable to any other conceivable reference (you, Sun, the company, etc.) because (a) it includes Schwartz himself, (b) it asserts a discourse situation that is made plausible via the blog, and (c) it allows to position the collective we apart from some outsiders who have been skeptical about the restructuring plans.

184

Cornelius Puschmann

On the basis of the described contexts and functions in which IPPs appear, it is possible to develop a taxonomy of post types and evaluate them against the background of the strategic goals they can potentially serve. It it furthermore possible to associate quantitative usage date from CBC/Corporati with these post types to highlight which uses are most common in which blogs. In terms the assignment of discourse roles and consequent designation of IPPs, there are several possible alternative post types: – Nobody: The blogger avoids self-reference and does not address an audience. – Only you: The blogger avoids self-reference and addresses an audience (2PP). – Only me: The blogger refers to himself or herself (1PP-sing) and does not explicitly address an audience. – Only us: The blogger refers to the company, the blog team etc. (1PP-pl) and does not explicitly address an audience. – You and I : The blogger refers to himself or herself (1PP-sing) and addresses an audience of 1-to-X people (2PP). – You and us: The blogger refers to the company, the blog team etc. (1PP-pl) and addresses an audience of 1-to-X people (2PP). Table 4 illustrates these different strategies with regard to their use of IPPs. Table 4. Strategies of IPP use in corporate blogs Nobody 2PP 1PP-sing 1PP-pl

Only you

Only me

Only us

X X X

You and I

You and us

X X

X (X) X

In the following, individual examples for each strategy will be provided. 5.1. Nobody strategy Text (5) Greg Sterling, SearchEngineLand.com, “Microsoft’s Johnson: ‘Once Yahoo And Microsoft Agree On A Transaction’,” 22 February, 2008 http://searchengineland.com/080222-163028.php (last accessed on March 29, 2010) This afternoon, Kevin Johnson, president of Microsoft Platforms & Services Division, issued an email to Microsoft employees (that was released via its PR firm). The letter has a tone of confidence that assumes a deal with Yahoo will ultimately take place. It reiterates what Microsoft believes to be the benefits of that deal and looks beyond regulatory approval to integration issues. It downplays Yahoo

Interpersonal pronouns in corporate web logs

185

resistance, potential culture conflicts between the companies and potential layoffs with an upbeat message about the future. Below is the full text of the email: [quoted text follows]

Omission of overt self-reference is possible in scenarios where the writer is not himself a part of the events or actions described. In Nobody-type entries, 2PP reference is missing as well, so that what is verbalized is not told to anyone in particular. The most natural topic for such a post is an event that, in terms of personal involvement, has a certain distance from both blogger and reader. 5.2. Only you strategy Text (6) Anon, Thomson Holidays Blog, “Top Places to Party,” 30 November 2008 http://thomsonholidays.blogs.com/my weblog/2007/10/top-places-to-p.html (last accessed on March 29, 2010) If you’re feeling a little down-in-the-dumps lately, the chances are it is our good old English weather that is to blame. So if you fancy hotting things up a bit, and you’re after a bit of liveliness into the bargain, then check out our Top Places to Party around Europe. From Ibiza to Malta, Bulgaria to Valencia, our guide gives you the best night-spots for that much-deserved break. Go on, you know you want to. . . Top Places to Party

Blog entries such as (6) are encountered primarily in those rare cases where no individual blogger is credited as the author of a post. Because the contextual information about the author is not provided, blog deixis cannot be relied on, and therefore self-reference must be largely omitted (but note the use of the plural possessive determiner, our). The lack of a discernible speaker in conjunction with the use of the imperative (go on) makes the function of the blog as advertising apparent. Very few contexts are imaginable in which the speaker is so significantly backgrounded in favor of the addressees, paired with suggestions regarding what they should do. Such peculiarities place these types in significant distance from mainstream blogs and provoke accusations of being “fake blogs” (or “flogs,” see Puschmann 2010: 122). 5.3. Only me strategy Text (7) Marc Monseau, JNJ BTW, “Through the Eyes of Another. . . ,” 30 July 2007 http://jnjbtw.com/?p=92 (last accessed on March 29, 2010) For doctors and family members caring for someone who is unwell, it’s often very difficult to understand and emphasize with what they are going through. Even

186

Cornelius Puschmann

more so if the disorder is pyschological. I’ve often heard about organizations trying to help folks see the world through the eyes of patients often turning to new technology and other tools that create a virtual reality. Recently, Sri Ramaswami, a friend of mine in PR at Johnson & Johnson’s Janssen company, told me about a new virtual reality tool that they had developed to help people understand what it’s like to walk in the shoes of a schizophrenic. Some coverage of this initiative here, here and here.

While there is clearly an implied reader addressed in posts such as (7), the reference is not made explicit. The audience is neither directly involved in what is communicated, nor is it invited to comment, surveyed or ordered to do anything. At the same time it is not possible for the blogger to present the exact same information without referring to himself and his (indirect) involvement via his friend. 5.4. Only us strategy Text (8) Lionel Menchaca, Direct2Dell, “Ratings & Reviews in More Languages,” 15 October 2007 http://direct2dell.com/one2one/archive/2007/10/15/31771.aspx (last accessed on March 29, 2010) Late last week, we rolled out ratings and reviews functionality in France, Germany and Spain for Consumer and Business systems. Since we began offering the ability to share ratings and reviews on October 4, we’ve now expanded the functionality to the following countries: Canada, France, Germany, Ireland, Spain, United Kingdom, United States. We plan to offer in more languages moving forward.

Entries of this type are especially characteristic for corporate, institutional or team blogs due to the fact that the corporate we can be plausibly used without a prior specification of who it refers to in such blogs. Blog entries such as (8) are a part of the organizational website and the actions described have not been conducted by a single individual (see also [4] and [5]). 5.5. You and I strategy Text (9) Thomas Mahon, English Cut, ”greetings from u.s.a.,” 16 November 2007 http://www.englishcut.com/archives/000232.html (last accessed March 29, 2010) I arrived in America safely, and am having a lovely time. As I said in my last post, I am only visiting Atlanta and New York. I’m in Atlanta today and tomorrow, and New York Monday, Tuesday and until Wednesday noon. As always, if you wish

Interpersonal pronouns in corporate web logs

187

to meet up, please feel free to get in touch, I look forward to seeing Everybody again. Thank you.

These types of posts exemplify a configuration where both discourse roles (blogger and reader) are fully applied. The request of the blogger to contact him makes 2PP plausible, although one could also imagine the use of other pronouns (anyone, someone), since this is what the blogger does when he expresses that he looks forward to seeing Everybody. 5.6. You and us strategy Text (10) The Official Palm Blog, “Happy America Recycles Day!” 15 November 2007 http://blog.palm.com/palm/2007/11/happy-america-r.html (last accessed March 29, 2010) In the spirit of America Recycles Day, we just wanted to reiterate that Palm fully supports recycling. As mentioned in a previous post, Palm’s recycling program is free and takes cell phones and handhelds in any condition, from any brand and also any old accessories that go with it. If you want to participate, just download a pre-paid mailing label here, or pick one up at a Palm retail store. (Don’t forget to erase all your personal information before sending it in). Here are the links to Palm’s recycle site and the National Recycling Coalition. And of course there’s no reason not to recycle everyday! -Palm Recycle Team

As the most likely type of post in terms of how discourse roles are assigned, this strategy fully establishes the noted tri-part relationship between blogger, reader and company via IPPs. As in (4) and (5), the blogger uses both we and the company name (Palm), the 1PP-pl standing in for Palm Recycle Team.

6. Conclusion This paper has presented an analysis of the use of inter-personal pronouns (IPPs) in corporate blogs. It has done so assuming that the corporate blog is an emerging domain-specific genre of computer-mediated communication. The goal was to depict the implicit conceptualization that bloggers and blog readers alike have of the specific communicative situation that blogs create and that are reflected linguistically in IPP use. It has been shown (a) that blog technology provides important meta-data that bloggers can rely on to encode deictic information in their writing through the use of pronouns (via a deictic center) and that blog readers can use to reconstruct said context, and (b) that it is the virtual nature of

188

Cornelius Puschmann

the discourse situation that makes rhetorical extensions of the pronouns possible, both in ways anticipated by the blogger and in unforeseen ways. The specific communicative goals associated with corporate blogging favor personalized and subjectified expressions over a more detached style (as can be found, for example, in academic writing). For this reason, strategies that assign more discourse roles are overall preferred to those that assign fewer ones. Another reason for why the latter type is used rarely is that it ignores contextual information that implies a discourse situation that blog readers have become to expect. The specific example of pronouns – deictic bridges between text and context – serves to highlight the significance of genre all the way down to the foundations of grammar in real-world discourse, because genre, understood broadly as situational context paired with convention, is extremely salient in computer-mediated communication. The openness and dynamicity of the communicative situation found in blogs result in a continuum between interpersonal communication and expository writing and in an interplay between different communicative actors (blogger, addressee, company). It is specifically this interplay which is at once a challenge and an opportunity for corporate bloggers, who must appear as competent discourse partners in a constantly shifting communicative environment.

References Blood, Rebecca 2000 Weblogs: A history and perspective. [Web log message]. Retrieved from Rebecca’s Pocket, http://www.rebeccablood.net/essays/weblog history.html. Boyd, Dana 2006 A blogger’s blog: Exploring the definition of a medium. Reconstruction 6: Article 2. Retrieved from http://reconstruction.eserver.org/064/boyd.shtml. Brown, Roger, and Albert Gilman 2003 The pronouns of power and solidarity. In Christina Bratt Paulston and G. Richard Tucker (eds.), Sociolinguistics: The Essential Readings, 156– 176. Oxford: Blackwell. B¨uhler, Karl [1934] 1990 Theory of Language: The Represenational Function of Language. Amsterdam/Philadelphia: John Benjamins. Chafe, Wallace L. and Jane Danielewicz 1987 Properties of Spoken and Written Language. Berkeley, CA: National Center for the Study of Writing.

Interpersonal pronouns in corporate web logs

189

Crystal, David 2006 Language and the Internet. 2nd Edition. Cambridge: Cambridge University Press. Efimova, Lilia, and Jonathan Grudin 2007 Crossing boundaries: A case study of employee blogging. In: Proceedings of the 40th Hawaii International Conference on System Sciences (HICSS-40), 86–105. Los Alamitos, CA: IEEE Press. Herring, Susan C. and Jonathan C. Paolillo 2006 Gender and genre variation in weblogs. Journal of Sociolinguistics 10: 439–459. Herring, Susan C., Lois A. Scheidt, Sabrina Bonus and Elijah Wright 2004 Bridging the gap: A genre analysis of weblogs. In: Proceedings of the 37th Hawaii International Conference on System Sciences (HICSS-37), 101–111. Los Alamitos, CA: IEEE Press. Heyd, Theresa 2008 Email Hoaxes: Form, Function, Genre Ecology. (Pragmatics & Beyond New Series 174.) Amsterdam/Philadelphia: John Benjamins. Heyd, Theresa this volume I know you guys hate forwards: Address pronouns in digital folklore. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 333–358. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Huddleston, Rodney D. and Geoffrey K. Pullum 2002 The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press. Johansson, Stig 1985 Word frequency and text type: Some observations based on the LOB corpus of British English texts. Computers and the Humanities 19: 23– 36. Kilgarriff, Adam 1998 BNC database and word frequency lists. [Website]. Retrieved from http://www.kilgarriff.co.uk/bnc-readme.html. Koppel, Moshe, Shlomo Argamon and Anat R. Shimoni 2003 Automatically categorizing written texts by author gender. Literary and Linguistic Computing 17: 401–412. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar. Stanford, CA: Stanford University Press. Lyons, John 1977 Semantics. Cambridge: Cambridge University Press.

190

Cornelius Puschmann

McNeill, Laurie 2005 Genre under construction: The diary on the Internet. Language@ Internet 2: Article 1. Retrieved from http://www.languageatinternet.de/articles/2005/120. Miller, Carolyn R. and Dawn Shepherd 2004 Blogging as social action: A genre analysis of the weblog. In Laura J. Gurak, Smiljana Antonijevic, Laurie A. Johnson, Laurie A. Johnson, Clancy Ratliff, and Jessica Reyman (eds.), Into the Blogosphere: Rhetoric, Community, and Culture ofWeblogs [Website]. http://blog.lib.umn.edu/blogosphere/blogging as social action a genre analysis of the weblog.html. Nilsson, Stephanie 2003 The function of language to facilitate and maintain social networks in research weblogs. Ph. D. Dissertation, Ume˚a Universitet, Sweden. Nowson, Scott, Jon Oberlander and Alastair Gill 2005 Weblogs, genres and individual differences. In: Proceedings of the 27th Annual Conference of the Cognitive Science Society, 1666–1671. Hillsdale, NJ: Lawrence Erlbaum. Puschmann, Cornelius 2010 The Corporate Blog as an Emerging Genre of Computer-Mediated Communication: Features, Constraints, Discourse Situation. (G¨ottinger Schriften zur Internetforschung.) G¨ottingen: Universit¨atsverlag G¨ottingen. Rapaport, William J., David A. Zubin, Gail A. Bruder, Judith F. Duchan, Michael J. Almeida, Joyce H. Daniels, Mary Galbraith, Janyce M. Wiebe and Albert H. Yuhan 1989 Deictic centers and the cognitive structure of narrative comprehension. Manuscript, SUNY Buffalo Department of Computer Science, Buffalo, NY. Saint-Georges, Ingrid de 1998 Click here if you want to know who I am. Deixis in personal homepages. In: Proceedings of the 31st Hawaii International Conference on System Sciences (HICSS-31), 68–77. Los Alamitos, CA: IEEE Press. Schler, Jonathan, Moshe Koppel, Shlomo Argamon and James W. Pennebaker 2006 Effects of age and gender on blogging. In: Proceedings of the AAAI Spring Symposium Computational Approaches to Analyzing Weblogs: 191–197. Menlo Park, CA. Schmidt, Jan 2007 Blogging practices: An analytical framework. Journal of ComputerMediated Communication 12, Article 13. Retrieved from http://jcmc. indiana.edu/vol12/issue4/schmidt.html. Virtanen, Tuija this volume Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre. In: Heidrun Dorgeloh and Anja

Interpersonal pronouns in corporate web logs

191

Wanner (eds.), Syntactic Variation and Genre, 53–84. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Wales, Katie 1996

Personal Pronouns in Present-Day English. Cambridge: Cambridge University Press.

Part 2: Syntactic variation based on genre

Grammatical constructions and communicative genres1 Susanne G¨unthner

As empirical studies of everyday interactions show, participants in formal as well as informal situations gravitate towards sedimented patterns on various levels. They do this in forms that range from those of grammatical constructions (Fillmore, Kay and O’Connor 1988) to communicative genres (Bakhtin [1979] 1986; Luckmann 1986; G¨unthner and Knoblauch 1995). In this paper, I argue that specific grammatical constructions (e.g., initial positioning of the finite verb in German declarative sentences) are closely tied to particular communicative genres (such as jokes). However, not only are specific grammatical constructions preferred within particular genres (jokes, chats, complaints, exemplary stories, and lamentos, for example), but their repeated use is instrumental in constructing genres. On the basis of empirical analyses of Infinitkonstruktionen (‘infinite constructions’) and was-Fragen (‘what-questions’) I discuss the relationship between communicative genres and syntactic patterns in spoken German interaction. In addition, questions concerning routinization and sedimentation of linguistic forms in social actions are addressed.

1. Introduction As empirical studies of everyday interactions show, participants in formal as well as informal situations orient towards sedimented patterns on various levels. They do this in forms that range from those of grammatical constructions to communicative genres. The concept of communicative genres, as developed within Sociology of Knowledge (Luckmann 1986, 1988; Bergmann 1987; Bergmann and Luckmann 1999a, 1999b; G¨unthner and Knoblauch 1994, 1995, 1996, 2007) and 1. Thanks to Lisa Roebuck and Katie Gilbert for checking for the proper use of English in this chapter. I would also like to thank Heidrun Dorgeloh, Wolfgang Imo, Katharina K¨onig, Anja Wanner, and the participants of the DGfS workshop “Syntactic Variation and Emerging Genres” (March 2007, University of Siegen) for their helpful comments on an earlier version of this chapter.

196

Susanne G¨unthner

Anthropological Linguistics (G¨unthner 1995, 2000, 2007; Hanks 1987; Kotthoff 1998), ties back to the work of Bakhtin (1986) and Voloshinov ([1929] 1986), as well as to the Ethnography of Communication (Gumperz and Hymes 1972). According to Bakhtin, speaking occurs in speech genres that guide the interaction and that are determined by social structures: Speech genres organize our speech in almost the same way as grammatical (syntactical) forms do. We learn to cast our speech in generic forms and, when hearing others’ speech, we guess its genres from the very first words; we predict a certain length . . . and a certain compositional structure; we foresee the end; that is, from the very beginning we have a sense of the speech whole, which is only later differentiated during the speech process. (Bakhtin 1986: 78–79)

Recent studies within Interactional Linguistics also show that grammatical constructions, as well as, for example, prosodic designs and lexical constraints, are often closely connected with particular communicative genres. In German for example, a language which has verb-second positioning in non-dependent declarative sentences, initial positioning of the finite verb (within declaratives) can be used to contextualize the beginning of a particular genre. When recipients hear utterances such as Kommt Fritzchen nach Hause (‘Comes little Fritz home’) or Treffen sich zwei alte Freunde. Fragt der eine. . . (‘Meet two old friends. Asks the first one. . . ’), they expect a joke. In everyday German narratives, speakers also make frequent use of initial positioning of the finite verb in combination with the narrative present tense, for example, GEH an die t¨ur, schAU RUM (‘go to the door, look around’) (G¨unthner 2000, 2006a). Or, as Schlobinski (2001) shows, so-called inflective constructions (*malheftigumarm*, ‘justvehementlyhug’) are characteristic features of chat-room conversations. Thus, specific grammatical constructions are closely tied to particular communicative genres. However, not only are specific grammatical constructions preferred within particular genres (jokes, chats, complaints, exemplary stories, and lamentos, for example), but their repeated use is instrumental in creating new genres. Thus the relationship between communicative genres and grammatical constructions is reflexive. On the basis of empirical analyses of Infinitkonstruktionen (‘infinite constructions’) and was-Fragen (‘what-questions’) I shall illustrate the relationship between communicative genres and syntactic patterning in spoken German interaction.

Grammatical constructions and communicative genres

197

2. The concept of communicative genres Over the last 20 years, studies within the Sociology of Knowledge2 as well as within Anthropological Linguistics3 have repeatedly addressed the issue of communicative genres, or discourse genres. These studies provide a theoretical conceptualization linking the notion of genre to the theoretical model of Social Constructivism (Berger and Luckmann 1966). As communicative genres represent a central communicative means in the construction of social reality, they play a major role within any approach to communicative practice (Hanks 1987; G¨unthner 2000). In everyday interaction, participants do not continuously invent new ways of speaking; instead, they orient to culturally sedimented patterns, patterns which are part of their communicative knowledge. These patterns, which are called communicative genres (Luckmann 1986, 1992), represent historically and culturally specific, prepatterned, and complex solutions to recurrent communicative problems (Luckmann 1986, 1992). On the one hand, they guide interactants’ expectations about what is to be said (and done) and in what way in the particular context. On the other hand, they are the sediments of socially relevant communicative processes (Luckmann 1986; G¨unthner and Knoblauch 1994, 1995). If, for instance, a speaker begins an utterance with Did you hear the one about. . . , a certain genre (a joke) is contextualized and specific expectations arise on the recipient’s part. If we hear Once upon a time there lived a king. . . , the speaker also constructs expectations concerning a particular genre (a fairytale). Thus, the knowledge that communicative processes with specific functions occurring in certain social situations take on recurrent forms not only guides the communicative actions but also their interpretations. Genres facilitate the coordination of communicative actions by guiding interactants’ expectations as to a course of communicative action. Therefore, the function of communicative genres may be seen as one which takes the burden of having to coordinate every communicative action anew away from the actors (Luckmann 1986; G¨unthner and Knoblauch 1994, 1995). By way of routinization, these interactive tasks become non-problematic and speakers may concentrate on other tasks. Communicative genres can be treated as historically and culturally specific conventions and frames according to which speakers compose talk or texts and recipients interpret them (Hanks 1987; G¨unthner 2000, 2007; G¨unthner and 2. See Luckmann (1986, 1992), G¨unthner and Knoblauch (1994, 1995, 1996), Bergmann and Luckmann (1995, 1999a, 1999b), and Knoblauch (1995, 2001). 3. See Hanks (1987), Briggs and Bauman (1992), G¨unthner (1993, 2000, 2007, 2008), and Kotthoff (1994, 1998). See also Imo (this volume).

198

Susanne G¨unthner

Luckmann 2001).4 In choosing a particular genre, a speaker makes use of culturally segmented solutions to communicative problems. As historical and cultural products, communicative genres are open to change and cultural variation (G¨unthner 1993, 2001; G¨unthner and Luckmann 2001). Therefore, if we take communicative genres as socially constructed solutions that organize, routinize, and standardize individuals’ dealings with particular communicative situations (Luckmann 1986, 1988), it follows that different cultural groups may construct different solutions for one and the same communicative problem (G¨unthner 2007). Thus, the repertoire of communicative genres varies from one cultural group to the next (G¨unthner 1995, 2007; G¨unthner and Luckmann 2001). An essential element of genre-related knowledge is an understanding of the appropriate use of genres, or in other words when to use what genre. Members of a cultural community usually have a knowledge of those genres that are necessary to use in their life-world; for example they know how to tell a joke, they recognize when someone else is telling a joke, and they also know in which situations it is appropriate to tell what kinds of jokes (G¨unthner and Knoblauch 1995; G¨unthner and Luckmann 2001). There are situations that require the use of a particular genre (when someone dies in Caucasian Georgia, for example, lamentos are expected, see Kotthoff 1999, 2002). In other situations, a speaker may have more than one genre from which s/he may choose. If a speaker intends to criticize the misbehaviour of a co-participant, for example, he might reproach, tease, or make fun of that co-participant. The particular choice depends on various aspects such as the social context, the specific communicative situation, the relationship between the participants, their communicative habitus, the degree of the misbehaviour, and so on. In everyday interactions, we frequently meet blendings and hybridizations of various generic models; a professor might tell a joke within her lecture and thus integrate a particular genre (a joke) into a more complex one (a lecture), or she might move into a sermonic tone of voice within her lecture and thus transform the lecture genre. Communicative genres are composed of various elements on three different levels (G¨unthner and Knoblauch 1994, 1995, 1996): (i) The level of internal features: Features located on the internal level include of the use of particular lexicosemantic elements, phonological and prosodic devices, grammatical constructions, phrases, registers, formulas, rhetorical figures and tropes, stylistic devices,

4. Thus, the concept of communicative genre has always been tied to a theory of social action (Berger and Luckmann 1966; Luckmann 1986, 1992): Communicative genres are treated as ways of performing communicative actions in social contexts. See Giltrow (this volume) and Virtanen (this volume) for the concept of genre as social action.

Grammatical constructions and communicative genres

199

specific linguistic varieties, repair strategies, prescriptions for topics and topical areas, and so on. (ii) The level of interactional features: This level consists of those elements that are part of the ongoing interaction, in other words, the interactive organization of conversations including patterns of turn taking, adjacency pairs, pre- and post-sequences, preference organizations, strategies for longer stretches of conversation, and the participation framework. (iii) The level of external features: The external level reflects the relationship between the use of genres and particular communicative milieus, communicative situations, the selection of types of participants (according to gender, age, status, etc.), and the institutional distribution of genres.

Communicative genres are typically associated with specific milieus or communities of practices, which are often marked by regional, class, ethnic, age, occupational, and gender characteristics. Examples of such milieus or communities of practices could be families and their conversations at the dinner table, personnel offices of corporations conducting job interviews, or neighborhood gossip. Communicative genres are also linked to social institutions. Legal, political, pedagogical, and military institutions use particular genres. Universities, for example, are marked by strong preferences for certain genres, among them lectures, discussions, faculty meetings, seminars, office hours, and papers. Generic forms of communicative practices, thus, include a combination of expected features that are located on all three analytical levels and form (more or less) complex communicative patterns. The concept of communicative genre links the internal structure of the text and the dynamic of dialogue to the systems of the social and cultural order (Luckmann 1988). The analysis of genres allows not only for the description and explanation of communicative practices in detail, but also provides access to interpretative processes by establishing an essential analytic link between speaking activities in the ongoing interaction, the social context, and the cultural ideologies (for example, the expectations and communicative norms) of a particular group. Thus, genre analysis goes far beyond the task of classifying discursive activities. It mediates between situationally produced texts and larger sociocultural contexts. However, one should emphasize that communicative genres are not to be considered as static products to be described by their structural features. Rather, genres are constructed within communicative actions; they are emergent products of locally managed interactions. This dynamic aspect of genres can best be expressed by the notion of performance as it is used in Linguistic Anthro-

200

Susanne G¨unthner

pology (see also Imo, Giltrow, and Virtanen, all in this volume) on a dynamic conception of genres.) In using a particular genre, speakers link their utterances “to generalized or abstracted models of discourse production and reception” (Briggs and Bauman 1992: 147). Through this kind of linkage, which Briggs and Bauman in applying Bakhtin’s concept of intertextuality describe as an “intertextual relationship,” texts are rendered “ordered, unified, and bounded on the one hand, and fragmented, heterogeneous, and open-ended on the other” hand (1992: 147). By reproducing a genre, a prior discourse becomes recontextualized and the speaker creates historical and social links; in other words, the speaker recontextualizes these connections to fit the current interactional setting. In this recontextualization speakers may follow the canonized model, deviate from it, or create hybrid forms – the possibilities for variations are manifold. Each recontextualization, however, has implications for the genre, as the situational use of a genre is not only oriented towards the generic model but also modifies it at the same time. On the basis of empirical studies of infinite constructions and what-constructions, I shall illustrate the relationship between communicative genres and grammatical constructions and argue that sedimented patterns located on various levels (syntactic as well as textual) form essential components of communicative practice.

3. Infinite Constructions Within everyday narratives, speakers frequently make use of grammatical constructions that do not follow the rules of German standard grammar, yet the constructions represent conventionalized constructions participants use to fulfill communicative tasks. These peripheral patterns (Fries 1987) form recurrent, conventionalized, and even grammaticized resources with formal and functional characteristics used by speakers as solutions to particular communicative problems. One of these patterns is that of the so-called infinite constructions, constructions which lack a finite verb, as illustrated in the following data:5 (1)

I: (.) soFORT anKALte I (.) immediately stopped-PAST PARTICIPLE ‘I immediately stopped.’

5. Infinite constructions belong to what I call the family of “Dichte Konstruktionen” (‘dense constructions’), see G¨unthner (2006a, 2006b).

Grammatical constructions and communicative genres

(2)

201

Ich (.) raus ausm wagen I (.) out of-the car ‘I [jumped] out of the car.’

Infinite constructions exhibit the following features:6 – the construction forms a turn construction unit (TCU) of its own; – the construction starts with a deictic or anaphoric element (usually a firstperson pronoun) which refers to the subject of the action (in general to the protagonist); – no finite verb is realized (sometimes even the infinite part of the verb, the past participle, is omitted); – and the construction shows a specific prosodic design. In the following transcript, Klaus reconstructs a panic attack, which struck him while driving his car: Transcript (1) PANIK-ATTACKEN: KLAUS (‘PANIC-ATTACKS: KLAUS’) 145 Klaus: 146 147 148 149 150 151

da: han ich=s gmErktthen I noticed itwie=s kOmmt, how it started, I: (.) soFORT anKALte; 7 I (.) immediately stopped; s=AUto parkt, parked the car, ond mir GSA:, and said to myself,



The utterance I: (.) soFORT anKALte (‘I (.) immediately stopped’) (line 147) starts with the agent of the event by means of the pronoun I (‘I’). 8 After a short break, “(.)”, the narrative event is introduced by ways of the past participle

6. Sandig (2000) describes similar constructions as “Emphase-Satzmuster” (‘emphatic sentence patterns’); Redder (2003, 2006) examines a specific type of infinite construction: the “autonome Partizipialkonstruktionen” (‘autonomous participle constructions’). 7. The verb forms anKALte, parkt and GSA (line 149) are past participle forms in German; the perfective auxiliaries habe (‘have’) or bin (‘am’) have been omitted. 8. I is the Swabian dialect form of ich (‘I’).

202

Susanne G¨unthner

soFORT anKALte (‘immediately stopped’). By introducing the agent (I ) first, then referring to his/her action (soFORT anKALte; ‘immediately stopped’), and at the same time leaving the position of the finite auxiliary verb (habe [‘have’]) empty, the utterance, which forms a turn construction unit, is condensed. The finite auxiliary verb is not the only element that is omitted, other semantically redundant material is left out too. Consequently, the recipients are only confronted with the who and what of the narrative event. Thus, infinite constructions, such as I: (.) soFORT anKALte (‘I (.) immediately stopped;’), form bi-partite constructions, consisting in the introduction of the agent and the mentioning of the new narrative event; both parts are adjacently placed within one intonation unit. Even though the finite verb, and thus the inherent grammatical linkage, is missing, the two parts are semantically tied. Furthermore, the construction reveals a specific prosodic design with a marked rhythm. It shows dense accentuation (Uhmann 1996); in these short turn construction units, nearly every accentuable syllable is accented. This combination of marked rhythm and dense accentuation contributes to the contextualization of prosodic emphasis (Selting 1994; Sandig 2000; Schwitalla 2003). In the following transcript, Loni reconstructs how someone tried to steal her purse while she was sleeping on a train: Transcript (2) SCHLAFWAGEN¨ UBERFALL (‘RAID IN A SLEEPING CAR’) 71 72 73 74 75 76 77 78 79 80

Loni:

und dann (...) so eh merk=ich=so, and then (...) so eh I=notice, so ein ein rAscheln, a kind of rustling sound, [so ¨ A]hnlich; [or something] like that; Rosi: [hm] Loni: irgendwIe war da was; somehow there was something there; da seh ich so=n schatten ¨ Uber MIR, then I see a kind of shadow over me, merk ne HAND an meinem kopfkissen; notice a hand at my pillow;

brutAl. Echt. brutal. really. da sind die andern (auf)gewacht. then the others woke up.

Grammatical constructions and communicative genres

203

The construction in line 78, (“ I (.) totally screamed;>”), deviates again from standard German grammar. We are dealing with the juxtaposition of two components: the pronoun ICH (‘I’) and the past participle voll geschrIEn (‘totally screamed’). The finite auxiliary verb (habe; ‘have’), which serves to link the two parts, is omitted. Nonetheless, although the grammatical cohesion between the two components is greatly reduced, the pattern forms a routinized, even grammaticalized construction used in everyday narratives. Once again, the speaker (Loni) first introduces the agent and protagonist of the action ICH (‘I’), then after a micro pause, she presents the protagonist’s action, and thus the narrative-advancing information (Hopper 1979). The infinite construction at hand also reveals a specific prosodic design: accentuated syllables and a marked rhythm set the construction off from the preceding utterances. Formally, the construction shows parallels with reported speech, where the narrator first introduces the character and then begins her/his communicative action (e.g. I: you are crazy!; he: what’s going on?). The bi-partite structure, in combination with the specific prosodic design (short intonation units, several accentuated syllables, a micro pause between the two parts, the marked rhythm), contributes to a scenic performance of the narrated event. Past events are not only reconstructed, but in using syntactic patterns such as infinite constructions, the narrated events are mise-en-sc`ene: The speaker not only reports her or his past actions but re-stages them and contextualizes a highly emotional stance towards the reconstructed happenings. The use of dense constructions, such as infinite constructions, is closely related to what Hopper and Thompson refer to as “narrative foregrounding”, and thus, to material which supplies essential points of the narrative and “comprises the backbone or skeleton of the text, forming its basic structure” (1980: 280–281). In everyday German narratives, the production of an infinite construction often leads to lists of further syntactically dense constructions, as can be seen in the following segment. Tina reconstructs a panic attack she suffered:

Transcript(3) PANIK-ATTACKEN: TINA II (‘PANIC ATTACKS: TINA II’) 95 96

97

Tina:

der FAHRstuhl war mir SO egAl gewesen; I didn’t care about the elevator; ; ; (.) dann; (.) then;

204

Susanne G¨unthner

98

they did- it was not really open yet,>











holt FEUerWEHR, call emergency ,

99 100 101 102 103 104 105 106 107

In line 99 Tina reconstructs how she rushed into her apartment after suffering a panic attack: (‘’). Again, the infinite construction begins with the first-person pronoun ich (‘I’), which introduces the agent. Then, after a short pause the prepositional phrase mit der HAND rein (‘with the hand in’), which functions as a sort of predicate substitute, is used to reconstruct the protagonist’s action. Subsequent to this infinite construction, further syntactically dense constructions (G¨unthner 2006a, 2006b) line up, reconstructing the “narrative-advancing information” (Hopper 1979: 249) in condensed form: 101 102 103 104

Tina:

These syntactically dense patterns resemble infinite constructions in as much as they consist of participle constructions or prepositional phrases without any

Grammatical constructions and communicative genres

205

(finite) verbs, forming independent turn construction units. However, in contrast to infinite constructions, the agent is no longer explicitly mentioned; the subject is omitted. Instead, the agent introduced in the infinite construction (line 99) remains interactively and cognitively active throughout the whole sequence. A stringing together of such dense constructions, which leaves out semantically redundant material and is organized in short intonation units, functions to contextualize a dynamic, even hectic rendition of a number of events happening in rapid succession. A string of such dense constructions is generally triggered by an infinite construction. To sum up the characteristics of infinite constructions: By the omission of contextually given information and by the leaving of the seemingly obligatory topological positions of German syntax empty (that is the position of the finite verb), the presented components – the agent as well as narrative-advancing activities – are foregrounded in infinite constructions. The bi-partite organization (subject plus participle or prepositional element) is characteristic of infinite constructions, as illustrated in Table 1: Table 1. Bi-partite organization of infinite constructions introducing the agent9

introducing the narrative-advancing event

I: (.) ‘I’

soFORT anKALte ; ‘immediately stopped’

ICH (.) ‘I’

voll geschrIEn ; ‘totally screamed’

ich (-) ‘I’

mit der HAND rein ; ‘with the hand in’

Even though infinite constructions are syntactic patterns that deviate from the rules of German standard grammar due to the fact that an obligatory syntactic position remains empty (as the finite verb is omitted), they represent a conventionalized, even grammaticized, interactionally fully functional construction, one which interactants use as a resource for particular communicative tasks in specific genres (such as everyday narratives). With the marked deletion of interactively given information and the juxtaposition of the agent and the particular action, the narrator contextualizes immediateness, expressivity, and a number of events happening in rapid succession. Furthermore, the specific prosodic gestalt

9. The observation that in our examples the agents are always introduced by first-person pronouns is based on the fact that in these everyday narratives speakers reconstruct their own (mostly affectively loaded) experiences.

206

Susanne G¨unthner

of the construction (the dense accentuation and the marked rhythm) contributes to the signaling of emphasis. As Goffman ([1974] 1986: 506) points out, narrators of everyday narratives do not only reconstruct past events, they also stage these past happenings as “little shows” (Goffman 1986: 506); they present them as something for the recipients “to re-experience, to dwell on, to savor” (Goffman 1986: 506). In German, infinite constructions form grammatical and rhetoric-stylistic devices that narrators use to stage past events. By foregrounding the agent and her/his action and thus the central components of the plot, the seemingly grammatically deviant infinite construction becomes a grammaticized way of creating narrative density, of constructing a “lively” narrative form (Sandig 2000: 299–300), and thus, of dramatizing the story-telling.

4. Was-Konstruktionen (‘what-constructions’) Another syntactic gestalt frequently used in particular everyday genres, specifically in reproaches and complaints, is that of constructions with the interrogative pronoun was (‘what’). Reproaches are communicative activities that speakers use to evaluate their co-participants’ behaviour negatively and to construct it as “deviant” (G¨unthner 2000: 83). They can be treated as a form of “remedial interchange” (Goffman 1967: 57–60): In uttering a reproach, the speaker focuses on the recipient’s misdeed and demands a remedial reply, an apology or an account. As remedial interchanges are face-threatening activities and thus endanger the “ritual order” (Goffman 1967: 42), the initiation of a remedial interchange inevitably opens up the possibility that instead of remedial work the co-participant may challenge the initiator’s legitimacy and a quarrel may ensue. A study of reproaches in informal German interactions reveals that speakers – in reproaching their co-participants – orient to sedimented forms that have specific prosodic, lexico-semantic, and syntactic rhetoric as well as specific interactive characteristics (G¨unthner 2000). Reproaches are “secondary genres” (Bakhtin 1986: 61), sedimented on all three analytical levels (the internal, the interactional, and the external level). In German, speakers frequently make use of question formats when reproaching their co-participants. For example: Wieso ¨ ↓↑LASCH du st¨andig dein geschirr rUmstehen.; (‘why do you always leave your dirty dishes lying around’); warum ↑↓TU::N sie=s dann nicht.; (‘why don’t you do it then’); and Kannst du nich endlich damit AUFh¨orn? (‘can’t you finally stop doing that?’). Besides question formats with warum (‘why’), wieso (‘why’), as well as yes/no questions with modal verbs, for example, k¨onnen, (‘can’) or m¨ussen

Grammatical constructions and communicative genres

207

(‘must’), speakers also make use of the interrogative pronoun was (‘what’): ¨ was ↑↓FAHRSCH du au JE.DES. wochenende nach schtU:gert. (‘what are you doing driving to Stuttgart every weekend’). In the following transcript segment, taken from a family conversation, the mother (M), the father (V), and the children (Ernst and Ulfa) are talking about habitual drinkers. In lines 10–11 M reproaches her husband (V) because of his drinking habits: Transcript (4) ALKI: OBERSCHWABEN (‘ALCOHOLICS: HIGH SWABIA’) 1M: 2Ulfa:

3M: 4V: 5Ulfa: 6Ernst: 7M:

8 9 10 11 12 13Ernst: 14V: 15 16

da trinkt ma- s¨ auft ma doch nimmer f¨ ur der DUr[scht.] then on- one doesn’t drink because of thi[rst.] [Also] [well] ga[nz] to[tally] [] [] [.] [.] =

(-) (-) und was and what des isch ja gwOhnheit. that is pure habit. da KANN ma doch koi DURSCHT meh han. you surely cannot be thirsty anymore. ( [ )] [(ja=no=i) ich kann dir] sagen warum [(yeah=no=I) I can] tell you why (0.5) wenn ich fort bin when I go out

208 17 18 19 20 21

Susanne G¨unthner da weiß ich wie viel ich trinken kann; then I know exactly how much I can drink; (-) wenn noch autofahre; (-) when I am the designated driver; 2.0) un wenn da bloß zwei bier getrunken habe, an- when I have then drunk only two beers, und drei eh fff is irgendwie en en (-) limitierte SAche. and three eh fff is somehow a a (-) limited amount.

In line 10 M uses a was-question (‘what-question’) to confront V with his seemingly deviant behaviour. The prosodic configuration of the was-question (und was ‘and what ’; line 10) contextualizes an affectively loaded, reproachful voice with a falling terminal pitch, a verum focus, a lengthening and glide on the verb ↑MA:CHsch (‘make’) and an increase in volume (G¨unthner 1996). M’s explanation, des isch ja gwOhnheit. da KANN ma doch koi DURSCHT meh han. (‘this is pure habit. you surely cannot be thirsty anymore.’; lines 11–12) signals her interpretation of V’s behaviour. The possible reason, Durst haben (‘to be thirsty’), is rejected as implausible; instead M insinuates that V drinks out of “pure habit.” German reference grammars refer to the function of the interrogative was as a pronoun begging some unknown piece of information that the asker presumably doesn’t know (in the sense of Was k¨onnte ihr denn gefallen? [‘What is it she might like?’]). However in the above sequence, the interrogative pronoun was asks for the reason behind a particular action or behaviour. V’s reaction in line 14 orients towards the structural patterning of the was-questions: V provides an explanation: ich kann dir sagen warum. . . (‘I can tell you why. . . ’). This causal use of the interrogative pronoun was (‘what’) differs from the syntactic form of a question asking for some unknown fact insofar as in the causal use all argument positions of the verb are filled. When we ask “what do you open?” the second verbal argument (the object) is missing and we expect the recipient to provide it in her/his answer. In the case of und was (‘and what ’), however, all syntactic constituents are available; the object is part of the construction. Thus, this was-construction formally and functionally varies from asking a question for some unknown fact (Günthner 2000). Many German reference grammars do not mention this use of the interrogative was at all. In his Deutsche Grammatik, Erben (1972: 231), however, refers to “uses of was,” asking for reasons (in the sense of why) and purposes (in the sense of what for) and calls them “colloquial uses” of the interrogative pro-

Grammatical constructions and communicative genres

209

noun was. The Duden Grammar (2005: 314) – the principal German reference grammar – also mentions “colloquial uses of was in the sense of why.” Both grammars portray causal uses of was as if they were colloquial variants of asking a why-question – neither referring to functional differences nor to contextual restrictions. The specific genres (reproaches, complaints, indignations10 ) in which was is used in a causal sense in any way are not mentioned. Thus, the question arises: Why do speakers frequently use was-constructions in the minor genre of reproaches? Or, to put this question in another way: Why do we interpret questions asking for the reason for a particular action as a reproach? First of all, by posing such a question, the speaker questions the plausibility of the presented action or behaviour of her/his co-participant. Questioning the action or behaviour of a co-participant, however, is a delicate, face-threatening act, as the consensus concerning the appropriateness of the action or behaviour is at issue, and the speaker demands an explanation. Sometimes, speakers explicitly refer to the absence of a plausible explanation, or they explicitly evaluate the action at issue as inadequate and deviant: was ↑MA:CHSCH=DU= NOMAL=EIN BIER=UFF. des isch ja Gewohnheit. da KANN ma doch koi DURSCHT meh han. (‘and what this is pure habit. you surely cannot be thirsty anymore.’). The combination of the questioner presenting a seemingly improper action with the absence of a plausible explanation for the action or the presence of a morally problematic reason for the action seems to constitute the architecture of reproaches in was-formats. As with reproaches, here not only a general action or type of behaviour is judged according to someone’s moral norms, but the behaviour of the coparticipant is negatively evaluated, and this activity bears a face-threatening potential and can provoke counter-reproaches, teasing, and may even lead to quarrelling (G¨unthner 2000). Goffman refers to the interactive risk of “priming moves,” a term he uses for the initiating steps of a remedial interchange, and remarks that, due to this risk, “priming moves” tend to be uttered “in various disguises” (Goffman 1967: 156). Instead of an “outright challenge, we are likely to find devices such as a ‘set-up question’: should the ‘asker’receive the expected answer, he will be in a clear position to challenge or to negatively sanction the respondent, but at the same time he leaves a little room open for an unanticipated answer that might adequately account for the apparent infraction” (Goffman 1967: 156). Consequently, question formats function as such “set-up questions” or conversational “pre-sequences” (Levinson 1983: 345–348) and are used to test out whether there might be any plausible reason that led, or even forced 10. Speakers also use was-questions to mark their indignation about someone’s behaviour, e.g. was fragt der dich auch so penetrant! (‘what [meaning: why] does he ask you in such a pushy way!’).

210

Susanne G¨unthner

the addressee to carry out the seemingly inappropriate act. Thus, reproaches (and complaints) in question formats can be treated as “off record” strategies (Brown and Levinson 1978: 216). “Off record” strategies are – as Brown and Levinson (1978: 216) show – important devices that speakers use when they want to perform a face-threatening act but avoid the responsibility for doing so. On the surface, in using a question format, speakers ask for an explanation concerning the specific behaviour and thus open up an opportunity for their co-participant to initiate remedial work.

5. Conclusion Empirical analyses of infinite constructions and was-constructions in everyday German reveal the interrelation between uses of communicative genres and syntactic constructions. As we have observed, certain syntactic gestalts – due to their form-function relations – are predestined for usage in particular communicative genres (see also Imo, this volume). In the case of infinite constructions, their fragmentary, condensed character, in combination with their prosodic design, contextualizes a sudden reflex action of the protagonist. Thus, this construction functions as an interactive resource that speakers use in reconstructions of dynamic and even dramatic events in everyday narratives. With was-constructions, in asking the co-participant for reasons for her/his behaviour, speakers question the plausibility of this behaviour and demand a remedial response. Due to the face-threatening character of reproaches, question formats turn out to be ideal “off record” (Brown and Levinson 1978: 216) strategies used in this highly sensitive minor or secondary genre. As the analyses reveal, linguistic knowledge includes far more than just a set of abstract rules and a lexicon; it also includes knowledge of sedimented constructions and their usage in particular genres.

References Bakhtin, Mikhail M. [1976] 1986 The problem of speech genres. In: Caryl Emerson and Michael Holquist (eds.), Speech Genres and Other Essays, 60–102. (University of Texas Press Slavic Series 8.) Austin: University of Texas Press. Berger, Peter L. and Thomas Luckmann 1966 The Social Construction of Reality. New York: Doubleday.

Grammatical constructions and communicative genres

211

Bergmann, J¨org 1987 Klatsch. Zur Sozialform der diskreten Indiskretion. Berlin/New York: de Gruyter. Bergmann, J¨org and Thomas Luckmann 1995 Reconstructive genres of everyday communication. In: Uta M. Quasthoff (ed.), Aspects of Oral Communication, 289–304. (Research in Text Theory 21.) Berlin/New York: de Gruyter. Bergmann, J¨org and Thomas Luckmann (eds.) 1999a Kommunikative Konstruktion von Moral. Band 1: Struktur und Dynamik der Formen moralischer Kommunikation. Opladen: Westdeutscher Verlag. Bergmann, J¨org and Thomas Luckmann (eds.) 1999b Kommunikative Konstruktion von Moral. Band 2: Von der Moral zu den Moralen. Opladen: Westdeutscher Verlag. Briggs, Charles L. and Richard Bauman 1992 Genre, intertextuality and social power. In: Journal of Linguistic Anthropology 2: 131–172. Brown, Penelope and Stephen Levinson 1978 Universals in language usage: Politeness phenomena. In: Esther N. Goody (ed.), Questions and Politeness: Strategies in Social Interaction, 56–311. (Cambridge Papers in Social Anthropology 8.) Cambridge: Cambridge University Press. Duden 2005 Duden Grammatik der deutschen Gegenwartssprache. Mannheim/Leipzig/Wien/Z¨urich: Dudenverlag. Erben, Johannes 1972 Deutsche Grammatik. M¨unchen: Max Hueber. Fries, Norbert 1987 Zu einer Randgrammatik des Deutschen. In: J¨org Meibauer (ed.), Satzmodus zwischen Grammatik und Pragmatik, 75–95. (Linguistische Arbeiten 180.) T¨ubingen: Max Niemeyer. Giltrow, Janet this volume Genre as difference: The sociality of linguistic variation. In: Heidrun Dorgeloh and Anja Wanner (eds.): Syntactic Variation and Genre, 29– 51. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Goffman, Erving 1967 Interaction Ritual: Essays on Face-to-Face Behavior. New York: Doubleday Anchor. Goffman, Erving [1974] 1986 Frame Analysis: An Essay on the Organization of Experience. NewYork: Harper & Row.

212

Susanne G¨unthner

Gumperz, John J. and Dell Hymes (eds.) 1972 Directions in Sociolinguistics:The Ethnography of Communication. New York/Chicago: Holt, Rinehart and Winston. G¨unthner, Susanne 1993 Diskursstrategien in der interkulturellen Kommunikation. Analysen deutsch-chinesischer Gespr¨ache. (Linguistische Arbeiten 286.) T¨ubingen: Max Niemeyer. G¨unthner, Susanne 1995 Gattungen in der sozialen Praxis. Deutsche Sprache 3: 193–217. G¨unthner, Susanne 1996 The prosodic contextualization of moral work: An analysis of reproaches in why formats. In: Elizabeth Couper-Kuhlen and Margret Selting (eds.), Prosody in Conversation: Interactional Studies, 271–302. (Studies in Interactional Sociolinguistics 12.) Cambridge: Cambridge University Press. G¨unthner, Susanne 2000 Vorwurfsaktivit¨aten in der Alltagsinteraktion. Grammatische, prosodische, rhetorisch-stilistische und interaktiveVerfahren bei der Konstitution kommunikativer Muster und Gattungen. (Reihe Germanistische Linguistik 221.) T¨ubingen: Max Niemeyer. G¨unthner, Susanne 2006a Grammatische Analysen der kommunikativen Praxis – “Dichte Konstruktionen” in der Interaktion. In: Arnulf Deppermann, Reinhard Fiehler and Thomas Spranz-Fogasy (eds.), Grammatik und Interaktion – Untersuchungen zum Zusammenhang von grammatischen Strukturen und Gespr¨achsprozessen, 95–122. Radolfzell: Verlag f¨ur Gespr¨achsforschung. G¨unthner, Susanne 2006b Von Konstruktionen zu kommunikativen Gattungen: Die Relevanz sedimentierter Muster f¨ur die Ausf¨uhrung kommunikativer Aufgaben. Deutsche Sprache 34: 173–190. G¨unthner, Susanne 2007 Intercultural communication and the relevance of cultural specific repertoires of communicative genres. In: Helga Kotthoff and Helen SpencerOatey (eds.), Handbook of Intercultural Communication, 127–152. (Handbooks of Applied Linguistics 7.) Berlin/New York: de Gruyter. G¨unthner, Susanne 2008 Interactional sociolinguistics. In: Gerd Antos and Eija Ventola (eds.), Handbook of Interpersonal Communication, 53–76. (Handbooks of Applied Linguistics 2.) Berlin/New York: Mouton de Gruyter.

Grammatical constructions and communicative genres

213

G¨unthner, Susanne and Hubert Knoblauch 1994 Forms are the food of faith. Gattungen als Muster kommunikativen Handelns. K¨olner Zeitschrift f¨ur Soziologie und Sozialpsychologie 4: 693– 723. G¨unthner, Susanne and Hubert Knoblauch 1995 Culturally patterned speaking practices – The analysis of communicative genres. Pragmatics 5: 1–32. G¨unthner, Susanne and Hubert Knoblauch 1996 Die Analyse kommunikativer Gattungen in Alltagsinteraktionen. In: Susanne Michaelis and Doris Tophinke (eds.), Texte – Konstitution, Verarbeitung, Typik, 35–57. M¨unchen: Lincom. G¨unthner, Susanne and Thomas Luckmann 2001 Asymmetries of knowledge in intercultural communication: The relevance of cultural repertoires of communicative genres. In:Aldo Di Luzio, Susanne G¨unthner and Franca Orletti (eds.), Culture in Communication: Analyses of Intercultural Situations, 55–86. (Pragmatics & Beyond 81.) Amsterdam/Philadelphia: John Benjamins. Günthner, Susanne and Hubert Knoblauch 2007 Wissenschaftliche Diskursgattungen – PowerPoint et al. In: Peter Auer and Harald Baßler (eds.), Reden und Schreiben in der Wissenschaft, 53– 65. Frankfurt a.M.: Campus Hanks, William F. 1987 Discourse Genres in a theory of practice. American Ethnologist 14: 668– 692. Hopper, Paul J. 1979 Aspect and foregrounding in discourse. In: Talmy Giv´on, Discourse and Syntax, 213–241. (Syntax and Semantics 12.) New York: Academic Press. Hopper, Paul J. and Sandra A. Thompson 1980 Transitivity in grammar and discourse. Language 56: 251–299. Imo, Wolfgang this volume Mein Problem ist/mein Thema ist (‘My problem is/my topic is’): How syntactic patterns and genres interact. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 141–166. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Knoblauch, Hubert 1995 Kommunikationskultur: Die kommunikative Konstruktion kultureller Kontexte. (Materiale Soziologie 5.) Berlin/New York: de Gruyter. Knoblauch, Hubert 2001 Communication, contexts and culture. A communicative constructivist approach to intercultural communication. In: Aldo Di Luzio, Susanne G¨unthner and Franca Orletti, Culture in Communication: Analyses of

214

Susanne G¨unthner Intercultural Situations, 3–34. (Pragmatics & Beyond New Series 81.) Amsterdam/Philadelphia: John Benjamins.

Kotthoff, Helga 1994 Verbal duelling in Caucasian Georgia. In: Uta M. Quasthoff (ed.),Aspects of Orality, 112–137. (Research in Text Theory 21.) Berlin/New York: de Gruyter. Kotthoff, Helga 1998 Spaß Verstehen. Zur Pragmatik von konversationellem Humor. (Reihe Germanistische Linguistik 196.) T¨ubingen: Max Niemeyer. Kotthoff, Helga 1999 Affect and meta-affect in Georgian mourning rituals. In: J¨urgen Schl¨ager and Gesa Stedman (eds.), Representations of Emotion, 149–172. T¨ubingen: Gunter Narr. Kotthoff, Helga 2002 Gender and poeticity in Georgian mourning rituals. In: Bettina Baron and Helga Kotthoff (eds.), Gender in Interaction: Perspectives of Femininity and Masculinity in Ethnography and Discourse, 283–329. (Pragmatics & Beyond New Series 93.) Amsterdam/Philadelphia: John Benjamins. Levinson, Stephen C. 1983 Pragmatics. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Luckmann, Thomas 1986 Grundformen der gesellschaftlichen Vermittlung des Wissens: Kommunikative Gattungen. K¨olner Zeitschrift f¨ur Soziologie und Sozialpsychologie, Sonderheft 27: 191–211. Luckmann, Thomas 1988 Kommunikative Gattungen im kommunikativen ‘Haushalt’einer Gesellschaft. In: Gisela Smolka-Koerdt, Peter Spangenberg and Dagmar Tillmann-Bartylla (eds.), Der Ursprung der Literatur, 279–288. M¨unchen: Wilhelm Fink. Luckmann, Thomas 1992 Theorie des sozialen Handelns. (Sammlung G¨oschen 2108.) Berlin/New York: de Gruyter. Redder, Angelika 2003 Partizipiale Ketten. In: Ludger Hoffmann (ed.), Funktionale Syntax: Die pragmatische Perspektive, 155–188. Berlin/New York: de Gruyter. Redder, Angelika ¨ 2006 Nicht-sententiale Außerungsformen zur Realisierung konstellativen Schilderns. In: Arnulf Deppermann, Reinhard Fiehler and Thomas Spranz-Fogasy, Grammatik und Interaktion – Untersuchungen zum Zusammenhang von grammatischen Strukturen und Gespr¨achsprozessen, 123–146. Radolfzell: Verlag f¨ur Gespr¨achsforschung.

Grammatical constructions and communicative genres

215

Sandig, Barbara 2000 Zu einer Gespr¨achs-Grammatik: Prototypische elliptische Strukturen und ihre Funktionen in m¨undlichem Erz¨ahlen. Zeitschrift f¨ur Germanistische Linguistik 4: 291–318. Schlobinski, Peter 2001 *knuddel – zuerueckknuddel – dichganzdollknuddel*. Inflektive und Inflektivkonstruktionen im Deutschen. Zeitschrift f¨ur Germanistische Linguistik 29: 192–218. Schwitalla, Johannes 2003 Gesprochenes Deutsch. Eine Einf¨uhrung. (Grundlagen der Germanistik 33.) Berlin: Erich Schmidt. Selting, Margret 1994 Emphatic speech style – with special focus on the prosodic signalling of heightened emotive involvement in conversation. Journal of Pragmatics 22: 375–408. Uhmann, Susanne 1996 On rhythm in everyday German conversation: Beat clashes in assessment utterances. In: Elizabeth Couper-Kuhlen and Margaret Selting (eds.), Prosody in Conversation: Interactional Studies, 303–365. (Studies in Interactional Sociolinguistics 12.) Cambridge: Cambridge University Press. Virtanen, Tuija this volume Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 53–84. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Voloshinov, Valentin N. [1929] 1986 Marxism and the Philosophy of Language. Cambridge, MA: Harvard University Press.

Appendix: Transcription conventions (based on GAT) Sequential structure [] overlap [] = latching of new turns or single units Pauses (.) (-), (–), (—)

micro-pause short, middle or long pauses of ca. 0.25 – 0.75 seconds, up to ca. 1 second

216

Susanne G¨unthner

(2.0) (2.85)

estimated pause of more than ca. 1 second measured pause (measured to hundredths of a second)

Other segmental conventions and=uh slurring within units :, ::, ::: lengthening, according to duration uh, ah, etc. hesitation signals, so-called “filled pauses” ’ glottal stop Laughter so(h)o haha hoho heehee ((laughing))

laughter particles during speech syllabic laughing description of laughter

Final pitch movements ? high rise , rise to mid level pitch ; fall to mid . low fall Other conventions ((cough))

( ) (such) al(s)o (such/which) ((. . . )) ->

paralinguistic and non-linguistic actions and events accompanying paralinguistic and non-linguistic actions and events over a stretch of speech interpretive comments over a stretch of speech unintelligible passage, according to its duration presumed wording presumed sound or syllable possible alternatives omission of text specific line in the transcript which is referred to in the text

Accents ACcent Accent !AC!cent

primary or main accent secondary accent extra strong accent

Pitch step-up/step-down ↓ pitch step-down ↑ pitch step-up

Grammatical constructions and communicative genres

217

Change of pitch register

low pitch register

high pitch register Intra-linear notation of pitch movement within an accent ‘SO fall ‘SO rise ? SO level ˆSO rise-fall ‘’SO fall-rise ↑‘ small pitch step-up to the peak of the accented syllable ↓’ small pitch step-down to the bottom of the accented syllable ↑‘SO or ↓’SO conspicuously high or low pitch step-up or down to the peak or the bottom of the accented syllable Volume and tempo changes forte, loud

fortissimo, very loud

piano, soft

pianissimo, very soft

allegro, fast

lento, slow

crescendo, becoming louder

diminuendo, becoming softer

accelerando, becoming faster

rallentando, becoming slower Breathing in and out .h, .hh, .hhh in breathe, according to duration h, hh, hhh out breathe, according to duration

Genre effects in the replacement of reflexives by particles Britta Mondorf

Some changes in the English verb system appear to affect different genres to different degrees. These changes concern, for instance, the use of reflexives and particles with English verbs. While reflexive self has been observed to be on the decline with a range of verbs, alternative strategies measuring out the verbal action have come to the fore, such as the use of particles, see (1) and (2), or way-constructions, see (3): (1) to straighten (oneself) versus to straighten up (2) to brace (oneself) versus to brace up (3) to work oneself to the top versus to work one’s way to the top First analyses indicate that the introduction and spread of the particle with formerly exclusively reflexive verbs differ according to genre. An empirical investigation of historical and present-day English corpus data traces both the trajectories of change in the use of reflexives and particles as well as their distribution according to medium, national standard, and genre. The present article relates the observed distributional differences to genre differences and discusses how far the syntactic variation between reflexives and particles reflects and contributes to the emergence of genre-specific, form-function correlations.

1. Introduction Syntactic variation is a pervasive and highly revealing phenomenon that any theory aiming to explain language needs to be able to account for. The challenge of explaining syntactic variation concerns at least (a) the multifarious factors that govern syntactic variation, (b) their interrelatedness manifested in synergetic or antagonistic tendencies within an intricately intertwined network that continually shapes and designs language, and (c) their relevance for language production and processing.

220

Britta Mondorf

Traditional grammatical theory had, for a long time, relegated language variation to the realm of performance, thereby declaring it out of bounds for linguistic theory building. But even within functionally minded sociolinguistic circles, research in the area of syntactic variation, as opposed to phonological or lexical variation has occasionally been impaired by erroneous assumptions concerning an alleged uniformity of syntax (see Hudson 1980: 44, 47), which was regarded as “the marker of cohesion in society.” By contrast, functional approaches, whose main concern is to explain language form on the basis of its function, have always stressed that the true extent of grammatical variation is still to be acknowledged and that syntactic variation poses an instructive challenge for linguistic theory building. The occupation with grammatical variation phenomena has recently seen the publication of two volumes explicitly dedicated to this research area (Rohdenburg and Mondorf 2003; Rohdenburg and Schl¨uter 2009). They provide ample evidence for the view that grammatical variation is much more pervasive than formerly expected. The multifarious internal and external factors that shape and design grammar can by no means be relegated to performance if a linguistics theory is to reach descriptive adequacy and predictive power. The sheer number and diversity of factors that affect the choice of grammatical variants can help us to sharpen our eye for the detection of underlying cognitive principles and generalizations that operate across allegedly autonomous language modules (see Gries 2003; Mondorf 2009 b). Modular models assuming a delimitation of, for example, syntax and morphology, are notoriously ill suited to explain grammatical variation, since different variants can, for instance, system-internally trade syntax against morphology (see Rosenbach 2002; Vosberg 2006; Mondorf 2009 b). Similarly, approaches that reject phonological influences on grammar by assuming that in top-down processing we first determine grammatical structure and then arrange pronunciation are doomed to failure (see Schl¨uter 2005). The present volume gives centre stage to a factor that has often been neglected even in variation research, i.e. the effect of genre. Notable exceptions are recent studies on syntactic variation according to text type or genre, which have been promoted by the availability of present-day English and historical corpora, see, e.g., Elsness’ (2009) analysis of the decline of the present perfect and Lyne’s (2006) analysis of pronouns preceding verbal gerunds, and a collection of articles in Kyt¨o, Ryd´en and Smitterberg (2006). The neglect of genre effects has never been axiomatic but rather has resulted from difficulties in operationalizing genre effects in studies on syntactic variation. One of the reasons why syntactic variation does not easily lend itself to research in terms of genre is, for instance, that the vast amounts of data required for analyzing syntactic variation need to be broken down into smaller lots according to genre, which occasionally

Genre effects in the replacement of reflexives by particles

221

leaves the resultant lots too small to allow for meaningful comparison. What is more, the study of syntactic variation according to genre requires computerized databases that easily lend themselves to grouping texts by genres, such as LOB (Lancaster/Oslo-Bergen Corpus) and Brown and their Freiburg counterparts FLOB and Frown, the British National Corpus, the American National Corpus, the Helsinki Corpus, ARCHER (A Representative Corpus of Historical English Registers) or CONCE (Corpus of Nineteenth-Century English). However, the classification by text type or genre in these corpora is by no means consistent, which is partly due to the fact that genre is not established through structural linguistic markers/categories, but is determined through language use and social structures (see Dorgeloh and Wanner, this volume). These parameters prove far more difficult to quantify than formal characteristics. What is more, we often do not know how fine-grained our classification of different genres ought to be, how to account for inconsistencies in the delimitation of genres, and how to distinguish between genre and text type. As regards the crucial relation between processing and the function or purpose of communication, it has often been convincingly argued that “in natural contexts, people seek to establish reference with respect to their behavioral goals during the earliest moments of linguistic processing. Moreover, referentially relevant nonlinguistic information immediately affects the manner in which the linguistic input is initially structured” (Tanenhaus et al. 1995: 1636). Research findings indicate that syntactic variation is generally functionally motivated, i.e. driven by demands exerted by processing efficiency (see Rohdenburg and Mondorf 2003). Similarly, genre analyses are likely to show that genre-specific variation is likewise functionally motivated and shaped by the communicative purpose characterizing a certain communicative event. Any theory that seeks to hold explanatory potential for syntactic variation thus needs to pinpoint the underlying functional motivations triggering such variation. Instead of simply stipulating constraints in competence-based frameworks, we need to detect the functional motivations underlying these constraints and their relevance ought to be assessed on the basis of usage-based criteria. To give an example, the Stay constraint as it applies to – relatively immobile – complements is functionally motivated by requirements of adjacency, processing efficiency, and ultimately by our cognitive design. Similarly, the replacement of, for example, reflexives by particles in certain genres might be motivated by pragmatic and cognitive requirements relating to characteristics of these genres. What compounds matters is that the strengths of determinants are subject to diachronic and synchronic variation. While the reflexive is less popular in some genres investigated in the present article, in earlier stages of English it was the majority or even the exclusively available form. A genre-specific diachronic comparison of the use of this

222

Britta Mondorf

variant thus needs to monitor its diachronic development across genres in order to prise apart historical effects and genre effects and in order to account for the possibility that certain genres facilitate the use of new variants while others tend to retain older forms. What is more, it might not only be the case that some genres introduce new forms at a delayed rate, but that they make differential use of coexisting variants by establishing a certain division of labour. Such patterns call for functionally motivated accounts that model a reranking of constraints over time, much in the spirit of Haspelmath’s (1999) diachronic adaptation approach. Such adaptation might be particularly needed as new genres emerge (see Gotti, this volume). As regards genre-related syntactic variation, the present section is a first step in taking up the challenge of tackling some of the above-mentioned issues by first assessing the relevance of genre effects on a highly restricted set of syntactic variation phenomena. From there, we need to progress towards establishing possible ways of quantification of genre effects (see P´erez-Guerra and Mart´ınez-Insua, this volume; Kabatek, Obrist and Vincis, this volume), their interrelatedness with other determinants of syntactic variation, their theoretical status and the formulation of what, at this stage, are still highly tentative explanations. 1.1. The decline of the reflexive Several linguists have observed that reflexive self is becoming obsolete in a wide range of environments (see Kirchner 1951: 158; Jespersen 1909–49 [1961]: 325–331; Strang 1970: 153; Peitsara 1997: 321; K¨onig and Siemund 2000: 55; Rohdenburg 2009; Mondorf 2006, forthc.). Empirical support for this claim has recently been adduced by Rohdenburg (2009), who witnesses a decline of the reflexive for some 120 verbs, and by Mondorf (forthc.), who observes the reflexive to be supplanted by the more recent way-construction. Both authors concur that the replacement of the reflexive not only affects its overall frequency of use with certain verbs but also its range of applications. For instance, while the reflexive covered both concrete and abstract meanings in past centuries, its distribution is increasingly becoming restricted to abstract uses. The more recent competitor, the way-construction, starts out in concrete domains and, by a process of grammaticalization, conquers new territories, i.e. abstract domains that were formerly almost exclusively covered by reflexives (see Israel 1996: 227; Mondorf forthc.). These observations lead to the question of why the reflexive should be on the decline in these environments and how the English language system expresses those functions formerly fulfilled by the reflexive.

Genre effects in the replacement of reflexives by particles

223

1.2. Competitors of the reflexive Recent research shows that there are several functionally partly overlapping strategies used by the language system to express meanings that were formerly (almost) exclusively associated with the reflexive. Examples (5)–(9) illustrate five observed alternatives exploited for the verb to raise, summarised in (4). (4)

reflexive ⎧ Ø ⎪ ⎪ ⎪ ⎪ ⎨ way-construction particle ⎪ ⎪ reflexive + particle ⎪ ⎪ ⎩ ...

⎫ ⎪ ⎪ ⎪ ⎪ ⎬ ⎪ ⎪ ⎪ ⎪ ⎭

(5)

Reflexive: The thick eyebrows had raised themselves, while a mischievous smile did battle with the grief in his eyes. [BNC wridom1]

(6)

Zero: Gary watched the manager’s eyes raise Ø in disbelief and savoured that wonderful moment. [Daily Mail 1994]

(7)

Way-construction: Buchan’s compassionate novel has an integral wisdom that raises its way above your run-of-the-mill home-counties marriage saga. [Daily Mail 1995]

(8)

Particle: Just before we left, I raised up to straighten my coat and sneaked a look at the McLaren girl. [BNC wridom1]

(9)

Reflexive + particle: The meal finished, Stan Carver, placing his hands on each side of his plate, slowly raised himself up from the table and, standing still for a moment, said . . . , Thank God for a good dinner. [BNC wridom1]

At least four of these alternatives can also be illustrated with the verb to empty. (9) (10)

Reflexive: My head seemed to have emptied itself of words. [BNC wridom1] Zero: [. . . ] where the Volga empties Ø into the Caspian sea. [Guardian 1991]

224

Britta Mondorf

(11)

Particle: [. . . ] gulping sobs, went through to a noise like a drain emptying out and finished up with a sort of throaty . . . . [BNC wridom1]

(12)

Reflexive + particle: Her heart emptied itself out, [. . . ]. [BNC wridom1]

The first strategy (illustrated in examples [6] and [10]) simply drops the reflexive and thus renders a formerly reflexive – and in the sense of Hopper and Thomson (1980), semi-transitive – use of the verb less transitive.1 Cases in point are empty (into), trouble, qualify (for), oversleep, keep (from/to), etc., for which the reflexive has increasingly been ousted by a zero option during the past centuries (see Jespersen [1909–49] 1961: 325–331; Visser 1963: 420–439; Peitsara 1997: 348–349; Rohdenburg 2009). Another strategy discussed in Mondorf (forthc.) replaces the reflexive by way-constructions as illustrated in example (7) and (13): (13)

She worked herself to the top. She worked her way to the top.

Yet another strategy combines reflexives and particles. It is exemplified by (9), (12), and (14): (14)

He looked me all over, as if he had been going to buy me; and then straightening himself up with a grunt, he said, “‘He’s the right sort for you, Jerry; [. . . ].”. [NCF: Sewell, Anna. Black Beauty. 1877]

These reflexive plus particle combinations are, however, rare, as we will see in section 2.2. The historical development of such verbs appears to move either in the direction of exclusively using the particle or exclusively taking the reflexive. And finally, the strategy that forms the focus of the present study on genre effects in the replacement of reflexives is illustrated in (8), (11) and (15). In these cases a particle is used in a function formerly performed by a reflexive. (15)

a. b.

Brace yourself for the impact. [. . . ]. he said, in tones of mock-comfort, Brace up, Merrill. [BNC wridom1]

This article is structured as follows: Section 2 introduces a diachronic corpus study on the transition from reflexive to particle verbs covering the sixteenth to 1. This does not, of course, necessarily extend to other fully transitive uses and senses of the respective verb, such as to empty a bucket, etc.

Genre effects in the replacement of reflexives by particles

225

twentieth centuries. The analysis of a selection of 27 verbs documents marked changes in the verb system, most of which have occurred after the eighteenth century. Three synchronic corpus studies are presented in section 3. While section 3.1 is concerned with the considerable effects of the medium on the use of reflexives and particles, section 3.2 analyses British-American differences. The effects of genre and text type are described in section 3.3, which provides empirical support for the claim that the spread of the particle affects different genres to different degrees. The particle is shown to be particularly prominent in fiction and humorous texts, a finding that is related to similar tendencies observed for other competitors of reflexive self and the emergence of genre-specific formfunction correlations.

2. Diachronic corpus study on the competition between reflexives and particles Having shown that the assumed competition between reflexives and particles exists, the central claim that we are dealing with a spread of the particle at the expense of the reflexive still requires empirical support, which is provided in the following sections. 2.1. Methodology Four time periods have been analyzed in order to trace the historical development of particles with verbs that were formerly almost exclusively reflexive. The choice of the data is mainly motivated by the fact that the historical corpora investigated share the characteristic of containing British English prose data and thus form a comparatively – though not entirely – homogeneous set. This allows us to compare the use of the constructions under investigation across several centuries. The corpora used for the diachronic part of this study are listed in Table 1. They amount to roughly 77 million words. As a first step, three subsections of the British National Corpus were investigated in order to find a selection of verbs that display variation between reflexive and particle constructions in present-day English. The three subsections of the 2. The Eighteenth-Century Fiction corpus and the Nineteenth-Century Fiction corpus have been split into subcorpora in order to provide time periods that start or finish at the turn of a century. Since the emergent subcorpora (ECF 1, ECF 2, NCF 1, NCF 2) were ordered by authors’ birth dates rather than by publication dates, birth dates have been chosen for locating the historical texts in time. This inconsistency in dating the texts is negligible, since all historical texts are dated by birth dates while only the present-day English data are based on publication dates.

226

Britta Mondorf

Table 1. Overview of British English historical corpora Corpus Early English Prose Fiction (EEPF) Eighteenth Century Fiction (ECF 1)

Period *1460–1700 *1660–1699

Eighteenth Century Fiction (ECF 2) Nineteenth Century Fiction (NCF 1)

*1700–1752 *1728–1799



Time Spans *1460–1700



Mio Words 10 5

*1700–1800

5 11

Nineteenth Century Fiction (NCF 2)

*1800–1869

*1800–1900

26

British National Corpus (wridom1)

p 1960–1993

p 1960–1993

20

Total

77

* birth dates, p publication dates2

British National Corpus (spokcont, spokdem, and wridom1) served as the basis for a pilot corpus in which those verbs that served as both particle verbs and reflexive verbs were selected. In order to work with manageable numbers, the range of particles was additionally reduced to out and up. A concordance software, WordSmith, was then used to find all instances of the searchstrings *sel* (whereby the asterisk stands for any number of characters, thus allowing for themselves, myself, etc., and for a wide range of spelling variants still current in the time periods investigated) immediately followed by out or up.3 The resulting hits were then manually edited so as to include only those verbs that meet the following criteria: – The verb had to be able to occur without the reflexive (e.g. She curled herself up on the sofa would have been included because curl can occur without the reflexive; e.g., in She curled up on the sofa). – The verb had to have roughly the same meaning in reflexive and particle constructions. This latter criterion eliminated verbs like haul, since the particle construction meaning is not equivalent to the reflexive construction, as illustrated in (17). (16)

But as I know how the land lies, d’ye see, and the current of my inclination sets me off, I shall haul up close to the wind, and mayhap we shall clear Cape Margery [. . . ]. [NCF: Smollett, T. G. Sir Launcelot Greaves. 1762]

3. This means that examples such as the following have been eliminated from the tally: “. . . repulse but still unawares – could cause her to coil up in herself in shame, they’ve come here. . . . ” [BNC wridom1].

Genre effects in the replacement of reflexives by particles

227

Conversely, it led to the inclusion of hire, since its meaning in the following constructions was considered equivalent. (17)

. . . , when he came at Huntington, he hired himselfe to an Inkeeper where he played the vnder Oastler for the space of a yeare to his great grief and discontent, [. . . ]. [EEPF, anonymous]

(18)

[. . . ], who placed his daughter in a common brothell, to hire out the vse of her body for monie. [EEPF: Heywood, Thomas. Sir Richard Whittington. 1656]

Instances of the passive had to be excluded from the tally because passives have no reflexive variant. Thus, cases such as example (19) have been discarded from the analysis. (19)

They were curled up to fit into the pod, [. . . ]. [F-LOB] *They were curled themselves to fit into the pod, [. . . ].

This procedure rendered a list of 27 verbs that formed the basis for the diachronic study: boost, bow, brace, burn, coil, curl, dress, ease, empty, fit, freshen, gear, heave, hire, jerk, launch, lock, open, prop, psyche, raise, rouse, smarten, straighten, stretch, work, and wrap. 2.2. Results of the Diachronic Study The underlying hypothesis that the reflexive is on the decline in certain constructions in English can be based on a range of observations (see, e.g., Kirchner 1951: 158; Strang, 1970: 153; Siemund 2003: 488). Empirical support is drawn from a comparison of reflexive uses in competition with way-constructions. A diachronic study of ten verbs introduced by Mondorf (forthc.) reveals that the historically earlier reflexive is increasingly being ousted by the way-construction in resultative uses. The alternation between the two functionally overlapping constructions is illustrated in the following examples from the seventeenth and the twentieth centuries, respectively: (20)

[. . . ].; yet I am not unknown to the African part of the Macrocosme, where my single Sword hath eaten its way through thousands, and hath afterwards drank it self into a Surfeit, with the blood of those Helldyed Infidels. [Richard Head and Francis Kirkman. The English Rogue. 1668]

228

Britta Mondorf

(21)

Worked himself into a frenzy and gave himself indigestion. [BNC wridom1] .

(22)

[. . . ] he workedhis way down the steep bank toward the stream [Frown, from Mondorf forthc.]

For an empirical study of the hypothesis that the way-construction is supplanting reflexive self in resultative uses, 1,146 occurrences of ten verbs were investigated in four historically stratified corpora. The vertical axis provides the percentages for each variant, while the horizontal axis displays the time line.

Figure 1. Diachronic development of the competition between way-constructions versus reflexives (N = 1146) (based on Mondorf forthc.)

Each curve is additionally labelled with the absolute figures in order to indicate if the sample is large enough to permit the deduction of meaningful claims. The analysis reveals that the reflexive is the majority form in the earliest time period investigated. It is used in 48 cases with all the ten verbs investigated, while the way-construction is only used 27 times, which amounts to roughly 36%. The crossover around the eighteenth century indicates that, by then, the way-construction had become the majority form with these verbs. With 95 cases of reflexive self as opposed to 358 instances of way in present-day English, the reflexive today accounts for a mere 20%.4 The historically earlier reflexive has largely been supplanted by the way-construction in the course of the past five centuries. 4. The decline in resultative uses of the reflexive has also been documented for several of the individual verbs investigated in Mondorf (forthc.).

Genre effects in the replacement of reflexives by particles

229

In Mondorf (forthc.), these quantitative changes are related to qualitative changes in the use of the reflexive. It turns out that reflexive verbs are particularly threatened with concrete meanings, while they can – to some degree – stand their ground with abstract meanings. This distribution lends itself to an explanation in terms of functional approaches and grammaticalization theory (see, e.g., Israel 1996: 227; Hopper and Traugott 1993: 2; Mondorf forthc.). Innovations tend to be initiated in local contexts and, by metaphorical extension, expand to new semantic domains. This is in line with the finding that the way-construction was first implemented with concrete meanings and only gradually spread to less prototypical, abstract contexts. If the distribution of the reflexive is thus increasingly restricted by competing constructions, we might now ask whether this phenomenon can also be observed for particles, which are our main concern in the present article. Is the decline of the reflexive to the benefit of the way-construction part of a more general pattern? Figure 2 maps the trajectory of change for the distribution of 27 verbs that can occur with the reflexive (e.g., She braced herself for the impact), or the particle (e.g., She braced up for the impact), or a combination of both (e.g., She braced herself up for the impact), throughout the four time periods investigated. A total of 3,619 occurrences that fulfilled the selection criteria described in section 2.1 were analyzed. The vertical axis provides the percentages for each of the three variants. The horizontal axis indicates the four time periods investigated. Again, in order to provide information on the sample size, each graph has also been labelled with the absolute number of occurrences in each time period.

Figure 2. Diachronic development of the competition between reflexives versus particles (N = 3619)

230

Britta Mondorf

The solid line representing the use of the reflexive with the 27 verbs investigated shows a dramatic slump from about 80% to roughly 30% after the nineteenth century. Conversely, the percentages for particle verbs increase from about 10% before the nineteenth century to 60% in the present-day English data. The particle construction is clearly supplanting the reflexive in the group of 27 verbs that form the basis for the present study. The combination of reflexives plus particles has remained fairly stable across the five centuries observed, hovering around a mere 10%. Similar to the competition between reflexive self and way-constructions, we observe that the reflexive is also rapidly declining in comparison to the particle. The crossover marking the time when the particle becomes the majority form occurs in the twentieth century. This indicates that the replacement process seems to be a fairly recent development, even more recent for particles than for way-constructions (though the onset of these developments varies for individual verbs, see Mondorf forthc.).

3. Synchronic corpus studies on the competition between reflexives and particles In order to find out whether certain factors are particularly likely to promote the spread of the particle at the expense of the reflexive, three determinants are investigated: the effect of the medium (spoken or written English), the effect of national variety (British or American English), and the effect of the genre or text type (news writing, religious texts, fiction, etc.). 3.1. The effect of the medium The present section hypothesizes that the spread of the particle will generally be more advanced in spoken than in written texts. The following subsections derive the hypothesis, introduce methodological issues, and present the findings ascertained in a synchronic corpus study regarding the effect of the medium on the choice between reflexives and particles. 3.1.1. Deriving the hypothesis According to Rohdenburg (2009), the well over 120 reflexive verbs analysed in his study share the characteristic of being largely incompatible with a semantically near-equivalent zero variant. Being almost exclusively reflexive, they belong to a group of verbs that has been described as relatively rare, formal, and more closely associated with written than spoken registers (see Rohdenburg

Genre effects in the replacement of reflexives by particles

231

2009; Christophersen and Sandved 1969: 122; Biber et al. 1999: 345). If this observation applies to reflexive verbs in general, we can thus hypothesize that reflexive verbs should be more closely associated with written discourse than with spoken discourse. 3.1.2. Methodology The selection procedure follows the one described for the diachronic study in section 2.1. The database consists of three domains of the British National Corpus: the spoken context-governed domain, the spoken demographic domain, and the written domain 1. The so-called spoken context-governed domain comprises mostly premeditated speech recorded during business meetings, public talks, etc., while the spoken-demographic domain has been named for the demographically sampled selection of speakers who were asked to record their speech and that of their interlocutors by means of portable tape recorders. Finally, written domain 1 contains imaginative prose, such as novels, etc., and has been chosen to ensure comparability to the historical databases. In order to find out if the medium affects the choice between reflexive and particle verbs, the British National Corpus has been searched for all verbs that occur with the reflexive in combination with the immediately following particles up or out. From the resultant list of verbs, only those that have been judged by the investigator to be also acceptable without the reflexive (i.e., with only the particle) have been elicited for further study. This led to the inclusion of cases such as to curl (oneself) up. The verb to cheer has been excluded from the tally because its high frequency in collocations such as cheer up is likely to skew results in favour of the particle. The 38 verbs that fulfilled the selection requirements are: boost, bow, brace, bundle, burn, coil, curl, doll, dress, ease, empty, fit, freshen, gear, heave, hire, hitch, hoist, hole, hurl, inch, jerk, juice, launch, lever, lock, open, prop, psyche, raise, rear, rouse, smarten, straighten, stretch, tart, work, and wrap. The occurrence of these verbs with reflexives and/or particles has then been analyzed and manually edited. 3.1.3. Results Figure 3 presents the distribution of reflexives and particles according to medium. We observe a clearly stratified pattern. As expected, the reflexive is rarely chosen in unpremeditated speech (spoken demographic) and partly premeditated speech (context-governed). By contrast, it occurs far more often in written discourse. Its use in the written data is roughly ten times higher than in spoken texts.The situation for the particle is quite similar, though the effect of the medium is considerably weaker. The particle is used

232

Britta Mondorf

Figure 3. Ratio of reflexives and particles in spoken and written English (N = 2177)

roughly five times more often in writing than in spoken discourse. Combinations of a reflexive plus particle are extremely rare in spoken texts but considerably more typical of written discourse. The hypothesis stating that the reflexive is more closely associated with written texts than with spoken texts is confirmed. Similarly, particle verbs are more often found in the written data investigated here than in the spoken material. When comparing the reflexive and the particle in relation to the medium we find that the overall effect of the medium is more pronounced for the reflexive. Thus, we can conclude that the reflexive is even more sensitive to variation according to the medium than the particle. 3.2. The effect of national variety As the diachronic study in section 2 shows, the increase of the particle at the expense of the reflexive gains momentum in the twentieth century. In light of a wide range of divergent developments in British versus American English since late Modern English (see, e.g., Rohdenburg and Schl¨uter 2009), we might reasonably expect differences in the implementation of these grammatical changes in the two national varieties. 3.2.1. Deriving the hypothesis Traditional generalizations concerning British-American differences have often been found to be too simplistic to explain the wide range of contrasts and their trajectories of change (see Rohdenburg and Schl¨uter 2009). According to the traditional “colonial lag” explanation of British-American differences (Marckwardt 1958: 80), one would expect the ousting of the reflexive by the particle to be

Genre effects in the replacement of reflexives by particles

233

more advanced in British English than in American English. However, the “colonial lag” hypothesis has failed to explain a wide variety of British-American contrasts and the literature abounds in counterexamples (see Mondorf 2009a; Elsness 2009; Kjellmer 2009; Schl¨uter 2009; G¨orlach 1987: 54). For this reason, we turn to another generalization concerning British-American differences that has been phrased in terms of colloquialization (see, e.g., Mencken 1973: 94–96; Biber 1987: 108–113; Mair 1998: 153–154; and Tottie 2002: 176). This approach assumes that different societal structures promote the use of forms that are generally associated with spoken, more colloquial or informal registers in American English rather than in British English. In line with the colloquialization hypothesis concerning British-American differences, we should expect the reflexive to be more widespread in British English, as the particle can be considered more colloquial and hence more closely associated with American English. This expectation would be in line with a finding reported in Rohdenburg 2009. He finds American English to be more advanced in the replacement of reflexive self by the Ø-variant than British English. If the reflexive is more strongly associated with formal styles, this observation lends itself to an explanation in terms of colloquialization. 3.2.2. Methodology The basis for the second synchronic study are the four matching corpora of British and American English, described in Table 2. The retrieval procedure again follows that described for the first corpus study in section 2.1. Although these corpora are relatively small in comparison to the British National Corpus, they have the additional advantage of being matched for text type or genre. Table 2. Four matching corpora of British and American English Corpus

Period

Mio Words

BrE

LOB FLOB

1960s 1990s

1 1

AmE

Brown Frown

1960s 1990s

1 1

3.2.3. Results Figure 4 displays the distribution of the reflexive, the particle, and the combination of reflexives and particles in British and American English.

234

Britta Mondorf

Figure 4. Reflexives and particles in British and American English (N = 168)

The reflexive is used more than twice as often in the British standard than in the American standard. It is thus far less typical of American English. Similarly, the particle is found more often in British English than in American English. However, the difference between both national varieties is less pronounced for the particle than for the reflexive. The figures for the combination of reflexives plus particles are identical but instances of the combination are rare. Returning to the hypothesis formulated in section 3.2.1, we find that the data do not provide any indication of a “colonial lag.” More revealing, however, is an explanation that appears in terms of colloquialization. The colloquialization approach holds some appeal for the analysis of the British-American differences in syntax observed here, since it can be credited with relating style, medium, and national variety to each other. After all, as Figure 4 in section 3.2.3 indicates, the more advanced decrease of the reflexive in spoken English is matched by the more advanced decrease of the reflexive in American English. If spoken language is generally more closely associated with colloquial styles than with written language, this finding points towards an explanation of the low ratio of reflexives in American English in terms of colloquialization. In other words, the British-American differences are positively correlated with differences in terms of formal versus colloquial styles. Such parallels have also been observed for other constructions (see Mondorf 2009a), in the use of which American English has been observed to prefer variants that were typical of colloquial styles. There is, however, one crucial aspect that is not covered by the colloquialization explanation of British-American contrasts, namely the role of process-

Genre effects in the replacement of reflexives by particles

235

ing complexity associated with reflexive structures. According to Rohdenburg (2009), there is some indication that reflexive structures exert a high processing effort on their users. This is shown in a series of studies involving the tendency to avoid the reflexive in sentences that display a high degree of syntactic complexity. In a series of preliminary studies, Rohdenburg (2009) observes that increased grammatical complexity discourages the use of the reflexive as opposed to its zero variant. For instance, it has been suggested that the presence of argument complexity increases cognitive complexity (see Rohdenburg 1996, 2003; Mondorf 2003, 2009b). Consequently, we should expect a lower ratio of reflexives in examples such as (23) than in those illustrated by (24):5 (23)

Presence of a complement Now the home of her eldest son, William, Cliveden hit the headlines in 1963 as the place where Christine Keeler disported Ø with Jack Profumo. [www.time.com/time/magazine]

(24)

Absence of a complement While Patton disported himself outside the system, Eisenhower worked from within. [www.americanheritage.com]

Empirical evidence shows that the reflexive, with potentially complement-taking verbs such as disport, is used less often in American than in British English (see Rohdenburg 2009). Additionally, there appears to be more of a tendency to avoid certain cognitively complex structures in informal styles than in formal styles. Thus, informal tabloids are found to compensate more strongly for the syntactic and phonological complexity effects than are formal broadsheets (see Mondorf 2009a). Such compensatory strategies, as well as avoidance strategies, have also been observed for American as opposed to British English. Avoidance of cognitively complex structures thus might be a characteristic that American English shares with more colloquial styles. We have only just begun to analyze the reflexive in terms of cognitive complexity and the explanation for the distribution of the reflexive suggested here is, at this stage, still highly tentative. In view of parallel avoidance strategies in colloquial styles as opposed to formal styles and in American English as opposed to British English, this correlation might, however, be a promising avenue for further research. The observation that factors such as medium and national

5. For a discussion of the observation that the Complexity Principle (see Rohdenburg 1996) does not apply here, see Rohdenburg (2009).

236

Britta Mondorf

variety affect the choice between reflexive and particle verbs leads us to another, possibly related, factor, the effect of genre. 3.3. The effect of genre or text type There is considerable overlap between what some researchers have dubbed as text type and genre. Analyses of syntactic variation according to genre face several methodological problems. First, those databases that lend themselves to an analysis in terms of genre are often too small for analyzing syntactic variants. Second, the classifications of genres prevailing in the corpora available to us are often not entirely satisfactory (see Lee 2001). Consider the following definition provided by Bhatia (1993: 13, drawing on Swales 1990) according to which a non-fictional genre is a [. . . ] recognizable communicative event characterized by a set of communicative purpose(s) identified and mutually understood by the members of the professional or academic community in which it regularly occurs. Most often it is highly structured and conventionalized with constraints on allowable contributions in terms of their intent, positioning, form and functional value. These constraints, however, are often exploited by the expert members of the discourse community to achieve private intentions within the framework of socially recognized purpose(s).

One would be hard put to decide whether the British National Corpus domain arts, which does not differentiate between academic and non-academic texts, is to be counted as containing texts of the same or of different genres. The stated aims of genre analysis are to characterize typical textual features of genrespecific texts in order to find pedagogically utilizable form-function correlations and to explain this characterization with regard to socio-cultural and cognitive constraints operative in the relevant area of specialization (see Swales 1990 and Bhatia 1993). This leads us to another caveat for those studying genre effects on syntax: the problem of prising apart genre effects from those associated with style or the medium, on the one hand, and the problem of isolating socio-cultural and cognitive constraints operative in certain genres, on the other hand.Other weaknesses of the British National Corpus classification concern the problem that imaginative texts comprise novels, dramas, poems, and several texts labelled as lectures, which even contain classroom discussions. Keeping these problems in mind, we can nevertheless start culling as much information as possible from the sources available.

Genre effects in the replacement of reflexives by particles

237

3.3.1. Deriving the hypothesis In the previous sections we have witnessed that the reflexive is on the decline in certain uses and that its competitor, the particle, has continually been gaining ground. Since we have already seen that the reflexive is particularly threatened in spoken texts, we can reasonably ask whether this replacement process affects different genres or text types to different degrees. The hypothesis that can be derived from the observation that spoken texts are generally more closely linked to less formal styles can be phrased as follows: Given the relation between genre or text type and medium, we can expect marked differences in the use of the reflexive in different genres. Traditionally written and formal genres will tend to be more closely associated with the reflexive than will spoken and more colloquial texts. 3.3.2. The way-construction according to genre or text type One of the closest competitors of the reflexive, the way-construction, can be observed to display marked and systematic variation according to text type. Figure 5 provides the ratios of the way-construction in the four matching corpora LOB, F-LOB, Brown, and Frown, differentiated by nine genres or text types. The segments of the pie charts are labelled according to text type. They reflect the ratio of the way-construction per million words for the respective text type. Additionally, the numbers labelling each segment provide the actual number of occurrences of the way-construction. Thus, in the LOB corpus, 30 instances of the way-construction have been found for fiction texts, 1 in humorous texts, 13 in journalistic press texts, etc. The reason why the 3 occurrences of the wayconstruction yield a larger segment in the pie chart for the skills, trades, and hobbies texts than for the science texts is that the former category contains fewer words than does the latter. Although the figures are scarce in some categories, the overall tendency appears to be that the way-construction is used particularly often in fiction and humorous texts. It is least frequent in scientific and religious texts. The latter text types tend to be highly conventionalized and hence might repel more recent, less established, and more colloquial structures. However, the observable differences are, at the present stage of the investigation, merely tentative results that need to be confirmed in larger and more consistently differentiated databases. We can now turn to an analysis of the relation between the reflexive and the particle according to text type.

238

Britta Mondorf

Miscellaneous: 1

LOB

Balance: 3

Miscellaneous: 0

Balance: 3

FLOB Fiction: 39

Belles Lettres/ Blogr/ Essays: 16

Fiction: 30

Belles Lettres/ Blogr/ Essays: 7 Popular Lore: 4

Popular Lore: 4 Humour: 1 Skills/ Trades/ Hobbies: 3

Skills/ Trades/ Hobbies: 4

Religion: 0

Press: 13 Religion: 1

Miscellaneous: 1

Press: 10

Balance: 4

Brown

Balance: 3

Humour: 2

Miscellaneous: 1

Frown Fiction: 27

Fiction: 30 Belles Lettres/ Blogr/ Essays: 8

Belles Lettres/ Blogr/ Essays: 7 Popular Lore: 4

Popular Lore: 4

Skills/ Trades/ Hobbies: 4

Skills/ Trades/ Hobbies: 3 Humour: 2 Religion: 2 Press: 6

Religion: 0

Press: 3 Humour: 4

Figure 5. Way-constructions according to text type in four matching British and American corpora (N = 412)

3.3.3. Results Figure 6 displays the distribution of the reflexive and the particle in different text types, or genres. The vertical axis plots the ratio of each variant per million words (pmw), while the horizontal axis provides the eight different text types investigated. Each column is also labelled with the actual number of occurrences. For six text types, or genres, the use of the particle exceeds that of the reflexives. This is in line with the diachronic development described in section 2, a development in which the particle has been found to have largely ousted the reflexive in present-day English in those functions investigated in the present study. The only exceptions to the preferential use of particles over reflexives are first, religious texts, and second, scientific texts, in which the reflexive is on a par with its competitor. Though the instances for these two text types are extremely small (merely one case for each variant in religious texts and merely two cases in scientific texts), the phenomenon parallels the finding observed for another competitor of the reflexive, the way-construction. Just as with particles, way-

Genre effects in the replacement of reflexives by particles

239

Figure 6. Ratio of reflexives and particles by text type in the LOB, F-LOB, Brown and Frown corpora (N = 156)

constructions had more difficulties in gaining a foothold in these two genres. These highly conventionalized genres seem to be a stronghold of the reflexive. A second parallel between the competition of the reflexive with way-constructions, on the one hand, and with particles, on the other hand, is found in fiction and humorous texts. In these genres, the particle and the way-constructions seem to be particularly frequent, which might be some indication that the replacement of the reflexive is comparatively advanced in fiction and humorous texts. These text types are also less rigorously conventionalized, which might help to explain why they can become strongholds of the more recent constructions and hence of the domains in which the historically earlier reflexive is losing ground.

4. Conclusion The present article has provided empirical support for the hypothesis that the reflexive is increasingly being ousted by the particle in certain resultative constructions. This development holds irrespective of national variety for both British and American English, though American English is more advanced in supplanting the reflexive by the particle, a finding that has been related to the American English tendency towards colloquialization. The reflexive appears to be more

240

Britta Mondorf

successful in standing its ground in the written medium than in spoken texts. This result can be attributed partly to the more colloquial and less formal nature of spoken discourse as opposed to written discourse. Reflexive self appears to be associated primarily with formal rather than informal texts, whereas the particle prevails in informal domains. As regards genre differences in the syntactic constructions investigated, we find that the particle is particularly successful in fiction and humorous texts, a finding that is paralleled by another competitor of the reflexive, the way-construction. Conversely, both competitors of the reflexive appear to be dispreferred in religious and scientific texts. Given their highly conventionalized character, these text types seem to be the ones in which the reflexive can, to some extent, stand its ground. More research is needed to prise apart the effects of genre and to quantify its effect on the choice between syntactic variants as opposed to other possibly related factors. The present article is meant to have contributed towards this aim.

References Primary sources ARCHER (A Representative Corpus of Historical English Registers). Described in Biber, Douglas, Edward Finegan and Dwight Atkinson, 1994, “ARCHER and its challenges: Compiling and exploring a Representative Corpus of Historical English Registers.” In: Udo Fries, Peter Schneider and Gunnel Tottie (eds.), Creating and Using English Language Corpora. Papers from the Fourteenth International Conference on English Language Research on Computerized Corpora, Z¨urich 1993, 1–13. Amsterdam: Rodopi. American National Corpus (ANC) 2006 Second release. Linguistic Data Consortium. British National Corpus (BNC) 1995 BNC Consortium/Oxford University Computing Services. Brown ICAME 1961 Collection of English Language Corpora. Bergen: Norwegian Computing Centre for the Humanities. Early English Prose Fiction (EEPF) 1997 Chadwyck-Healey, Cambridge. Eighteenth-Century Fiction (ECF) 1996 Chadwyck-Healey, Cambridge.

Genre effects in the replacement of reflexives by particles F-LOB 1990s

241

Match of the Lancaster-Oslo-Bergen Corpus. University of Freiburg. ICAME Collection of English Language Corpora, 2nd edition, 1999.

Frown 1990s

Match of the Brown Corpus. University of Freiburg. ICAME Collection of English Language Corpora, 2nd edition, 1999. The Daily Mail and the Mail on Sunday on CD-ROM 1993–1997 Chadwyck-Healey, Cambridge. Secondary sources Bhatia, Vijay 1993 Analysing Genre: Language Use in Professional Settings. (Applied Linguistics and Language Study.) London: Longman. Biber, Douglas 1987 A textual comparison of British and American writing. American Speech 62: 99–119. Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad and Edward Finegan 1999 Longman Grammar of Spoken and Written English. Harlow: Longman. Christophersen, Paul and Arthur O. Sandved 1969 An Advanced English Grammar. London: Macmillan. Dorgeloh, Heidrun and Anja Wanner this volume Introduction. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 1–26. (Topics in English Linguistics 70.) Berlin/ New York: Mouton de Gruyter. Elsness, Johan 2009 The present perfect and the preterite. In: G¨unter Rohdenburg and Julia Schl¨uter (eds.), One Language, Two Grammars? Grammatical Differences between British and American English, 228–245. (Studies in English Language.) Cambridge: Cambridge University Press. G¨orlach, Manfred 1987 Colonial lag? The alleged conservative character of American English and other ‘colonial’ varieties. English World-Wide 8: 41–60. Gotti, Maurizio this volume A new genre for a specialized community: The rise of the experimental essay. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 85–110. (Topics in English Linguistics 70.) Berlin/NewYork: Mouton de Gruyter. Gries, Stefan Th. 2003 Multifactorial Analysis in Corpus Linguistics: A Study of Particle Placement. London: Continuum Press. Haspelmath, Martin 1999 Optimality and diachronic adaptation. Zeitschrift f¨ur Sprachwissenschaft 18: 180–205.

242

Britta Mondorf

Hopper, Paul J. and Sandra A. Thompson 1980 Transitivity in grammar and discourse. Language 56: 251–299. Hopper, Paul J. and Elizabeth C. Traugott 1993 Grammaticalization. Cambridge: Cambridge University Press. Hudson, Richard A. 1980 Sociolinguistics. (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Israel, Michael 1996 The way constructions grow. In: Adele E. Goldberg, (ed.), Conceptual Structure, Discourse and Language, 217–230. Stanford, CA: CSLI Publications. Jespersen, Otto [1909–49] 1961 A Modern English Grammar. London: Allen & Unwin. Kabatek, Johannes, Philipp Obrist and Valentina Vincis this volume Clause linking techniques as a symptom of discourse traditions: Methodological issues and evidence from Romance languages. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 247– 275. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Kirchner, Gustav 1951 A special case of the object of result. English Studies 32: 153–159. Kjellmer, G¨oran 2009 The revived subjunctive. In: Rohdenburg, G¨unter and Julia Schl¨uter (eds.), One Language – Two Grammars: Grammatical Differences between British and American English, 246–256. (Studies in English Language.) Cambridge: Cambridge University Press. K¨onig, Ekkehard and Peter Siemund 2000 The development of complex reflexives and intensifiers in English. Diachronica 17: 39–84. Kyt¨o, Merja; Mats Ryd´en and Erik Smitterberg (eds.) 2006 Nineteenth-Century English: Stability and Change. (Studies in English Language 16.) Cambridge: Cambridge University Press. Lee, David 2001 Genres, registers, text types, domains, and styles: Clarifying the concepts and navigating a path through the BNC jungle. Language Learning and Technology 5: 37–72. Lyne, Susanna 2006 The form of the pronoun preceding the verbal gerund: Possessive or objective? ICAME Journal 30: 37–54. Mair, Christian 1998 Corpora and the major varieties of English: Issues and results. In: Hans Lindquist, Staffan Klintborg, Magnus Levin and Maria Estling (eds.),

Genre effects in the replacement of reflexives by particles

243

The Major Varieties of English: Papers from MAVEN 97, 139–157. V¨axj¨o: Acta Wexionensia. Marckwardt, Albert H. 1958 American English. New York: Oxford University Press. Mencken, Henry L. 1973 The American Language: An Inquiry into the Development of English in the United States. New York: Knopf. Mondorf, Britta 2003 Support for more-support. In: G¨unter Rohdenburg and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 251–304. (Topics in English Linguistics 43.) Berlin: Mouton de Gruyter. Mondorf, Britta 2006 Competing constructions. Paper Presented at the Third Meeting of the German Construction Grammar Network ‘Language Variation and Change in Construction Grammar,’ D¨usseldorf, Germany, 31 March– 2 April 2006. Mondorf, Britta 2009a Synthetic and analytic comparatives. In: G¨unter Rohdenburg and Julia Schl¨uter (eds.), One Language – Two Grammars? Grammatical Differences between British and American English, 86–107. (Studies in English Language). Cambridge: Cambridge University Press. Mondorf, Britta 2009b More Support for More-Support: The Role of Processing Constraints on the Choice between Synthetic and Analytic Comparative Forms. (Studies in Language Variation 4.) Amsterdam/Philadelphia: John Benjamins. Mondorf, Britta forthc. Variation and change in English resultatives. Language Variation and Change. Peitsara, Kirsti 1997 The development of reflexive strategies in English. In Matti Rissanen, Merja Kyt¨o and Kirsi Heikkonen (eds.), Grammaticalization at Work: Studies of Long-term Developments in English, 277–370. (Topics in English Linguistics 24.) Berlin/New York: Mouton de Gruyter. P´erez-Guerra, Javier and Mart´ınez-Insua, Ana this volume Do some genres or text types become more complex than others? In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 111–140. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Rohdenburg, G¨unter 1996 Cognitive complexity and increased grammatical explicitness in English. Cognitive Linguistics 7: 149–82.

244

Britta Mondorf

Rohdenburg, G¨unter 2003 Cognitive complexity and horror aequi as factors determining the use of interrogative clause linkers in English. In: Rohdenburg, G¨unter and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 205–249. (Topics in English Linguistics 43.) Berlin/New York: Mouton de Gruyter. Rohdenburg, G¨unter 2009 Reflexive structures. In: Rohdenburg, G¨unter and Julia Schl¨uter (eds.), One Language – Two Grammars? Grammatical Differences between British and American English, 166–181. (Studies in English Language.) Cambridge: Cambridge University Press. Rohdenburg, G¨unter and Britta Mondorf (eds.) 2003 Determinants of Grammatical Variation in English. (Topics in English Linguistics 43.) Berlin/New York: Mouton de Gruyter. Rohdenburg, G¨unter and Julia Schl¨uter (eds.) 2009 One Language – Two Grammars: Grammatical Differences between British and American English. (Studies in English Language.) Cambridge: Cambridge University Press. Rosenbach, Anette 2002 Genitive Variation in English: Conceptual Factors in Synchronic and Diachronic Studies. (Topics in English Linguistics 42.) Berlin/NewYork: Mouton de Gruyter. Schl¨uter, Julia 2005 Rhythmic Grammar: The Influence of Rhythm on Grammatical Variation and Change in English. (Topics in English Linguistics 46.) Berlin/New York: Mouton de Gruyter. Schl¨uter, Julia 2009 Phonology and grammar. In: G¨unter Rohdenburg and Julia Schl¨uter (eds.), One Language – Two Grammars? Grammatical Differences between British and American English, 108–129. (Studies in English Language.) Cambridge: Cambridge University Press. Siemund, Peter 2003 Varieties of English from a cross-linguistic perspective: Intensifiers and reflexives. In: G¨unter Rohdenburg and Britta Mondorf (eds.), Determinants of Grammatical Variation in English, 479–506. (Topics in English Linguistics 43.) Berlin/New York: Mouton de Gruyter. Strang, Barbara M.H. 1970 A History of English. London: Methuen. Swales, John M. 1990 Genre Analysis. (Cambridge Applied Linguistics.) Cambridge: Cambridge University Press.

Genre effects in the replacement of reflexives by particles

245

Tanenhaus, Michael K., Michael J. Spivey-Knowlton, Kathleen M. Eberhard and Julie C. Sedivy 1995 Integration of visual and linguistic information in spoken language comprehension. Science 268: 1632–1634. Tottie, Gunnel 2002 An Introduction to American English. (The Language Library.) Malden, MA: Blackwell. Visser, Fredericus Theodorus 1963 An Historical Syntax of the English Language, Vol. 1. Leiden: Brill. Vosberg, Uwe 2006 Die Große Komplementverschiebung. Außersemantische Einfl¨usse auf die Entwicklung satzwertiger Erg¨anzungen im Neuenglischen. (Language in Performance 34.) T¨ubingen: Gunter Narr.

Clause linkage techniques as a symptom of discourse traditions: Methodological issues and evidence from Romance languages Johannes Kabatek, Philipp Obrist, and Valentina Vincis

Throughout the last two decades, corpus linguistic methods have continuously been used to account either for the general historical evolution of particular linguistic phenomena or for their variation across genres, but rarely for both at the same time. In the present contribution, we will first introduce the notion of discourse traditions, a theoretical concept introduced by scholars in the field of Romance linguistics for the description of textual traditions. The main postulate associated with this concept is that textual traditions must be regarded independently from the diachronic evolution of language as a system; nonetheless, they may have a strong influence on it. In the second step, clause linkage techniques, commonly known to be highly symptomatic for the identification of textual traditions, are analysed. A quantitative approach will be used on a series of historical Romance corpora: juridical texts from 13th century Spain, an Old Spanish Bible corpus, a corpus containing texts of various Old Spanish genres and, finally, a Rheto-Romance newspaper corpus from the last two centuries. Results show the importance of genre differentiation in corpora in two different dimensions: vertically, distinguishing texts traditions with globally different characteristics; and, horizontally, identifying differences within the single texts.

1. Introduction In the last years, the idea that languages are synchronically homogeneous has been criticised in several studies that have claimed that there is not only linguistic variation due to different varieties of a language but also due to different textual traditions. Not all linguistic elements appear in the same way in all texts, and there are elements that tend to be attached to certain textual genres but not to others. This is of course true for lexical items, but it seems to be valid as well, at least to some degree, for grammatical elements. A well-known example from the

248

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

Romance languages is the French pass´e simple, a verbal tense almost entirely restricted to a series of written text types in France. From a diachronic view, this means that innovations can emerge in some genres without affecting others at first, but also that a loss of grammatical forms does not have to affect all genres at the same time (see Kabatek 2005a; Koch 2008; Pons 2008). Another example of this phenomenon can be seen in the Spanish subjunctive future form amare (‘love’), which was still in general use in Old Spanish, but had almost completely disappeared by the beginning of the eighteenth century; however, it has been preserved until the present, almost solely due to its use in juridical texts. In the last years, several scholars in Romance linguistics have introduced and developed the concept of discourse traditions, by taking the textual variation in a given language into account (Koch 1997; Oesterreicher 1997; Wilhelm 2001; Aschenberg 2003; Kabatek 2005b; Kabatek 2008). Discourse traditions are textual traditions that are related not only to certain elements of content, but also to the lexical and grammatical items predominantly used in that tradition. The notion of discourse traditions was introduced as a purely theoretical concept with convincing arguments, intuitively applied to a series of textual traditions, but without a properly elaborated methodology. Several examples, generally based on detailed philological analysis, have shown that textual traditions could strongly determine the choice of lexical and grammatical elements in a text. However, the problem of detailed philological analyses is that they cannot be applied to larger text corpora. In order to differentiate and to compare discourse traditions we need systematic criteria like those applied by scholars such as Douglas Biber and others for the identification of text genres both in synchronic and diachronic corpora (see Biber 1993, 2004; Biber, Conrad, and Reppen 1998; Biber, Davies, Jones, and Tracy-Ventura 2006; Parodi 2007, among others). Operating in a fashion that is similar to Biber’s method, but with a series of crucial differences, the present paper will present some results of genredifferentiating corpus analyses that are not based – as is Biber’s work – on a large number of different elements analysed statistically in order to show which of the elements can be regarded as relevant for genre differentiation. Instead, our method is limited to the analyses of a reduced number of elements considered to be particularly symptomatic of genre differentiation. The basic hypothesis is that elements that fulfill the task of linking clauses can strongly indicate certain particularities of texts and textual traditions. Evidence of this phenomenon becomes particularly clear when one pays attention not only to the elements themselves, but also to their relative frequency in the texts. The following pages will present some of the results of a research project on the relationship between clause linkage techniques and discourse traditions, presenting empirical

Clause linkage and discourse traditions

249

data from diachronic corpora of two Romance languages along with some general implications for theory and methods. The section following this introduction is dedicated to the presentation of the framework of discourse traditions theory and in particular to analyses of clause linkage (or “junction,” in Raible’s terminology, see Raible 1992: 28) as applied in the T¨ubingen project B 14/SFB 441.1 Section 3 offers the first results of textual differentiation in Old Spanish texts, in which texts that belong to the same genre (legal texts) show strong internal differences due to three different historical backgrounds, representing thus three different discourse traditions within the same genre. In section 4, a larger Old Spanish corpus will be presented and the question of the relationship between diachrony and textual tradition will be discussed. Section 5 is dedicated to the internal differences in a single text. In this section, different versions of Old Spanish Bibles will be compared in order to show, on the one hand, that clause-linking elements can be attributed to textual structure, and on the other hand, that different models (in this case, Latin and Hebrew origins) can lead to rather different results in each respective Bible version. Finally, section 6 will show how clause linkage techniques have evolved in Rheto-Romance newspapers from the nineteenth century to the present, and how a rather stable textual tradition changes over time from a more subordinated to a more nominalised style. The general objective will be to show: 1) the relevance of genre distinction and the notion of discourse traditions in relation to synchronic and diachronic studies; and 2) the possibilities for and limitations of the analysis of clause linkage patterns for textual differentiation.

2. On discourse traditions and the qualitative-quantitative method The theoretical framework for the empirical and methodological issues discussed in this paper is discourse tradition theory. The notion of discourse traditions was first presented by Peter Koch (1997) and Wulf Oesterreicher (1997) and has earlier roots in Eugenio Coseriu’s functionalism (Coseriu 1988: 161–181) and in Schlieben-Lange’s (1983) theory of historical pragmatics. During the last decade, a series of papers and monographs have further elaborated the basic notion of discourse tradition in its theoretical and methodological aspects and have 1. The project B 14 Discourse traditions of Romance languages and multidimensional analysis of diachronic corpora, directed by Johannes Kabatek and with collaboration of Philipp Obrist and Valentina Vincis, is part of the linguistic collaborative research centre SFB 441 Linguistic Data Structures: On the Relation between Data and Theory in Linguistics at the University of T¨ubingen. The research centre is funded by the German Research Council (DFG). See http://www.sfb441.uni-tuebingen.de/index-engl.html.

250

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

demonstrated in several empirical contributions its useful application, above all, in the field of historical linguistics. However, since the basic assumption of the framework is that every utterance might be determined not only by grammar and lexicon but also by textual tradition, discourse tradition theory is by no means limited to diachronic issues but fundamental for synchronic linguistics as well.2 The basic assumption of our approach is very simple and obvious. Producing utterances is not simply an activity that results from a certain grammar of a language, but that is also (and simultaneously) shaped by concrete utterances that have already been produced: utterances that are part of the tradition of a community, or discourse traditions. What seems so simple and evident has a number of important consequences, of which we will mention only a few: – The tradition of the texts does not only lead to the addition of elements (such as genre-specific markers, textual identifiers, or textual formulae etc.), but also determines the selection of elements in texts, and not only lexical items but also grammatical elements. – Synchronic variation in texts is not only a matter of differences between dialects or varieties, but also of differences between textual traditions. – Even if the origin of a textual tradition might often be described by means of concrete pragmatic needs, discourse traditions are more than only situationspecific coincidences: It might be true that there are speech acts that are strongly determined by situational constellations (asking for a light for a cigarette, for example), but the concrete linguistic form (Do you have a light?) of resolving the social task is very frequently taken over by formerly pronounced texts and is not spontaneously invented. – Discourse tradition is a general notion for tradition in language, and it includes all linguistic results of tradition, from simple formulae up to complex literary traditions. – Synchronic and diachronic variation must be differentiated according to different discourse traditions. In diachrony, by looking more closely at discourse traditions, it seems possible to shed light on abstract views that regard linguistic change as putative invisible-hand phenomena such as the s-curve or other apparently regular evolutions (Kroch 2001). – Discourse tradition is not only a relevant notion for the repetition of texts, text types or genres, formulae, and so on, but for all linguistic phenom2. The insistence on text tradition is not new at all; see e.g. Hockett’s claim that tradition is a linguistic universal: “3.5 Every human language has tradition [. . . ]. If we design and build a collection of machines that communicate among themselves with a language, this property will be lacking.” (Hockett 1966: 12).

Clause linkage and discourse traditions

251

ena related to traditions of utterances, including a significant absence of textual repetition in pragmatically analogous situations, social stratification and group phenomena correlating with textual traditions, politeness, textual interferences, and so on. From its first presentation onwards, the notion of discourse tradition has appeared in theoretical as well as in empirical studies. Koch and Oesterreicher presented it as a fundamental contribution to linguistic theory, and the ensuing elaborations of the concept (see, among others, Wilhelm 2001; Aschenberg 2003; Kabatek 2005a; L´opez Serena 2007; Pons 2007; S´aez Rivera 2007) have elaborated and added a series of aspects. At the same time, several empirical studies have identified discourse traditions in different Romance languages (Wilhelm 1996; Stoll 1997; Oesterreicher, Stoll, and Wesch 1998; Wilhelm 2001; Schmidt-Riese 2003; Kabatek 2005b; Ciapuscio, Jungbluth, Kaiser, and Lopes 2006; Guzm´an River´on 2007; Castilho 2009) and others have shown the importance of this notion for historical syntax (Sim˜oes 2007; Kabatek 2008). The elements that have been identified as relevant for discourse traditions are, on the one hand, peripheral elements such as explicit notions of a certain textual tradition, formula or the overall textual form, and on the other hand, certain lexical or grammatical items. The project we are currently presenting has its starting point in a critical view on monolithic diachronic studies that consider diachrony as a linear evolution of one homogeneous object, with variation other than the diachronic one being nothing but a matter of noise in the data that can be compensated for by large quantities of data. At the same time, we point out that the detailed philological analyses of individual texts, perfect as they might be, cannot serve as a basis for comparison with other texts: To be able to compare and to distinguish by means of a tertium comparationis, abstraction is necessary at least to some degree. The first step towards this abstraction consists of the selection of possibly relevant elements from textual traditions and their comparison in different texts. In many cases, a discourse tradition can be identified simply by certain keywords contained in a text. In other cases, however, it is not only the mere appearance of a word but also the frequency of its appearance that helps us to identify the discourse tradition. The twenty most frequent words of a text may give strong evidence of its discourse tradition. Our claim is that the same holds for certain grammatical elements and that their relative frequency in texts should correlate with different discourse traditions. Changing the point of view from reception (and observation) to production, this means that speakers will make their choices of grammatical variants according to the textual tradition they wish to represent. The relative frequency and the distribution of elements along a text will be

252

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

considered as grammatical indicators of discourse traditions, accompanied by all the other indicators (lexical, formulaic, and so on). According to the findings on corpora and genres applied to different languages by Biber and others, the continuum between orality and literacy is the most salient factor that determines the choice of elements in different texts. Koch and Oesterreicher stated in their seminal work on the continuum between what they call “language of immediacy” and “language of distance” (Koch and Oesterreicher 1985: 21; see also Koch and Oesterreicher 1990, 1994) that with regards to clause-linking, there is a strong correlation between this continuum and a scale between aggregation and integration. Whereas at the pole of aggregation (correlated with “language of immediacy”), we can find simple juxtaposition of clauses, at the pole of integration (correlated with “language of distance”), we will find clause integration by techniques of condensation, like e.g. in nominalising constructions. In 1992, Wolfgang Raible, in the tradition of the Cologne school of typology, offered a complex theory of what he calls Junktion (‘clause linkage’, see Raible [1992: 28]; also Raible 2001). He suggests the following stages between aggregation and integration (Raible 2001: 595), some of which are illustrated by English examples:3 I II

simple juxtaposition junction by phrasal adverbs

III IV

coordinating conjunctions subordinating conjunctions

V VI VII

infinite verbs (gerunds or infinitives) prepositional groups nominalisations

Joan remains at home. She is ill. Joan is ill. This is why she remains at home. Since Joan is ill, she remains at home. Being ill, Joan remains at home. On account of her illness, Joan remained at home.

Raible combines this continuum between aggregation and integration with a second axis of semantic categories. This second axis, based on a cognitive and logical scheme of increasing relational complexity, arranges the semantic relationship between clauses on an ordered scale: Condition is considered to be less complex than cause; “counter-cause” (the concessive relationship) is consid-

3. The original examples of the whole classification were French. In Raible (2001), the English examples quoted here are given for some of the levels. Not all levels must have a representation in all languages; Raible considers his continuum to be universal, with concrete forms differing from one language to another.

Clause linkage and discourse traditions

253

Figure 1. Two dimensions of “junction,” according to Raible (1992)

ered to be more complex than cause, and so on (Raible 1992: 131; see Figure 1). This cognitive scheme (Kortmann 1997, 2001) is related by Raible to language acquisition, language change, and the difference between orality and literacy. In an onomasiological perspective, the junction scheme is a field of two universal continua; in a semasiological perspective, every language offers a series of elements (junctors) and techniques of clause linkage that can be arranged in accordance with the two dimensions. The junction scheme can be applied to a number of linguistic fields: Typologically, language comparison can be based on the junction techniques languages offer; in a diachronic perspective, evolutions of junction techniques and junctors can be observed. The complete inventory of junctors in a language can be described as a characteristic pattern. In synchrony, language patterns and textual patterns can be distinguished, the former consisting of the complete clause-linking grammar of a language, the latter consisting of the clause-linking elements found in a text. These textual junction patterns will be considered, in our own studies, as relevant for the textual structure (see section 4), and, as we will see in the next section, for different discourse traditions.

3. The first step: Discourse traditions in Old Spanish texts In a detailed linguistic analysis of Old Occitan and Old Spanish juridical texts (Kabatek 2005b), we found striking textual and linguistic differences between several Spanish legal texts from the thirteenth and fourteenth centuries. Even if all of these texts could be categorised as, or even had been categorised as, laws, there are still fundamental differences that could not be explained in terms of diachronic evolution: Even within a very clearly defined genre, there were

254

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

important subcategories. Although many historical studies suppose that a simplistic genre distinction between general text types like juridical prose, literary prose, poetry, etc. is enough to give a representative view on textual variation in a language, the existence of subcategories seemed to suggest itself. The explanation we offered for the variation was a historical one: In thirteenth-century Castile, three different law systems co-existed, the first one being based on the so-called faza˜nas (‘descriptions of cases’) with implicit laws, the second one, the fueros, offering general norms or more abstract laws, and the third one deriving from the Renaissance of Roman law from the twelfth century onwards (for the “Renaissance of the 12th Century,” see Haskins 1927) and based on a complex juridical system with elaborate institutional and meta-juridic instructions. The texts, and accordingly, the linguistic elements that represent the three traditions, are dissimilar: Whereas the case laws are based on simple aggregative case descriptions with short, unconnected simple sentences, the abstract Spanish local laws tend to be lists of subordinated, conditional constructions. The texts based on Roman law, finally, are elaborated and complex texts with more integrative structures and categories that are more semantically complex than those found in the faza˜nas and the fueros. Applying the Raible scheme to the clause-linking elements in our three juridic traditions, we found strong differences in the junction patterns they presented. The differences were even clearer when measuring the relative frequency of these elements, since even in the least complex tradition of case law, some isolated highly integrative exceptions (probably due to Latin influence) appeared. The idea of measuring the relative frequency was adopted from Biber (1993). As in Biber’s studies, we calculated the relative frequency per 1000 words. The following two schemes offer some general results of different junction patterns in three Old Spanish juridic traditions (see Kabatek 2005a, 2005b).

Figure 2. Relative number (per 1,000 words) of clause-linking elements in three Old Spanish legal texts ordered by semantic relationships between clauses. The numbers 1–9 represent degrees of increasing semantic complexity (1=addition to 9=concession).

Clause linkage and discourse traditions

255

Figure 3. Relative number (per 1,000 words) of clause-linking elements in three Old Spanish legal texts ordered by degrees of integration (syntactic density) between clauses. The numbers from II–VII represent degrees of increasing integration (II=adverbs; III=coordination; IV=subordination; V=infinite verbs; VI=PrepP with infinite verbs; VII= nominalisations).

The result of this study was that the junction patterns differed clearly in the three texts and that these differences were not purely idiosyncratic ones: The three texts can be considered representative of the three legal traditions we identified. The comparison of the junction patterns of our three texts to those of other Old Spanish legal texts confirmed that the patterns could be regarded as representative features of the three traditions, since similar patterns were found in other texts belonging to each of them. This means that there not only existed three legal traditions in thirteenth-century Spain, but that they corresponded to three discourse traditions with respective linguistic characteristics which we were able to identify. The questions arising from this first study were the following: 1) Could the observations made in the case of the legal texts be generalised? 2) To what degree is junction-pattern differentiation a helpful instrument for the identification of discourse traditions? 3) Can we distinguish diachronic variation from textual variation? These and other questions will be presented in the following sections.

4. Clause linkage, diachrony, and discourse traditions in Old Spanish In order to see how clearly junction patterns can be used to identify different textual traditions, we expanded our perspective on Old Spanish by enlarging the corpus in two directions: other genres were included in the analysis on a synchronic axis (the thirteenth century); and, texts from the early thirteenth until the late fifteenth century were added on a diachronic axis. As to the methodology, a computational tool, the program TraDisc, was elaborated in order to make larger portions of texts accessible. TraDisc is a semi-automatic annotating instrument

256

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

that allows for the creation of textual junction patterns. This program must be trained by introducing the junction techniques of a language. After that, the ambiguous cases must be checked manually in order to extract a textual junction pattern.4 The set of Old Spanish texts we chose for a first enlargement of the corpus is as follows: – The Cantar de Mio Cid (CMC), a medieval lay on a hero of the Reconquista, supposedly written down at the beginning of the thirteenth century; – the Milagros de nuestra Se˜nora (‘Miracles of Our Lady’, Mil) by Gonzalo de Berceo, a devotional text in verse from the middle of the thirteenth century (between 1246 and 1254), attached to the literary genre of the Mester de Clerec´ıa (‘Ministry of Clergy’); – the Lapidario (‘Book of Stones’, Lap), from 1250, and the Libro de los judizios de las estrellas (‘Book of Judgments of Astrology’, Est), from 1254; these are two of the scientific prose works attributed to Alfonso X, both mainly translated from earlier Arabic works, and partially updated; – the Conde Lucanor (‘Count Lucanor’, CL), a prose work with didactic and moralistic purposes, written by don Juan Manuel between 1325 and 1335; – the Libro de Buen Amor (‘Book of Good Love’, LBA) by Juan Ruiz, the Archpriest of Hita, contemporary to the CL (from between 1330 and 1343); another example of the Mester de Clerec´ıa, mainly in verse; – the Celestina (Cel) from 1499, a novel that is seen as marking the end of medieval literature and the beginning of the literary Renaissance in Spain.5 We will now turn to an analysis of the data. Again, the junctors have been calculated per 1000 words (with a portion of 10,000 to 30,000 tokens per text being analysed) and they have been ordered according to a modified scheme based on the two Raible dimensions. The results are shown in Table 1.

4. Program and help files are available at: http://www.sfb441.uni-tuebingen.de/b14/tools.html. 5. The extent to which these texts are representative may be questionable. As the project advances, we are hoping to include more and more texts; however, two problems with such inclusion are the limited number of text series in different traditions in this early period of the Spanish language and the lack of free access to reliable electronic full texts, access which is necessary for annotation. Leaving corpus design issues aside, some first observations seem to be nonetheless possible.

Clause linkage and discourse traditions

257

Table 1. Junction analyses in a series of Old Spanish texts CMC

Mil

Lap

Est

CL

LBA

Cel

∼ 1200

∼ 1250

∼ 1250

1254

∼ 1330

∼ 1330

1499

epic hagiographic scientific scientific

didactic hagiographic fictional

Coordination II–III

22.9 54.5%

18.3 46.6%

70.6 65.4%

60.2 70.2%

63.8 61.5%

24.3 39.3%

26.3 38.7%

Subordination IV

11.4 27.1%

13.0 33.1%

26.3 24.4%

24.8 28.9%

26.0 25.1%

17.3 28.0%

19.3 28.4%

Gerunds V

3.4 8.1%

2.1 5.3%

5.5 5.1%

0.5 0.6%

3.9 3.8%

3.0 4.9%

9.7 14.3%

PrepP VI–VII

4.3 10.2%

5.9 15.0%

5.6 5.2%

0.2 0.2%

10.0 9.6%

17.2 17.8%

12.7 18.7%

Total II–VII

42.0

39.3

108.0

85.7

103.7

61.8

68.0

IV–VII

19.1

21.0

37.4

25.5

39.9

37.5

41.7

At first glance, the data seem not to allow for very clear conclusions. However, a closer look allows for the following very general observations about the results: – There is a diachronic shift to higher scores of more integrative junction (subordination, infinite constructions, and nominalisations). While three out of four texts from the thirteenth century range show a ratio of less than 26 per 1000 higher density clause linkage elements, the three later texts show a ratio of 37 and above. – Variation within genres is much higher than originally expected: Mil and LBA may be partly considered to pertain more or less to a similar genre (Mester de Clerec´ıa); and Lap and Est are usually taken as examples of scientific prose. Both pairs show significant differences, and while the former pair is separated by a span of 80 years, Lap and Est are coetaneous. – Finally, there seems to be no direct relation of concurrence between the more aggregative and more integrative junctors in Old Spanish. The comparison of Lap and Est rather suggests that clause linkage by different means works fairly independently, since both texts show high frequencies of aggregative junctors, but considerable divergence of hypotactic ones. The first finding, a growth in the more integrative sector, is the one that fits best with the traditional view on the emergence of written Romance languages and with the further elaboration, or Ausbau, of these languages – the idea that these languages are written traditions that were and are in the process of becoming

258

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

more and more diversified and elaborate. But these data – and a closer look at the texts – also suggest that an overall comparison of clause-linking patterns in texts only makes sense if we are dealing with rather homogeneous ones, as was the case with the legal texts presented above in section 3. From the second half of the thirteenth century onwards, more heterogeneous texts appear, combining and elaborating former traditions. Thus, the application of our analysis to more heterogeneous texts makes it necessary to differentiate the progression of clauselinking elements. In some of our texts, considerable differences could be found between separable parts of the texts. This is why easily identifiable sub-texts were regarded as more or less representative of independent discourse traditions. An interesting example is the tradition of prologues of Old Spanish texts: Because of the rhetorical necessity of explaining the intentio and the utilitas of the text in the foreword, the clause-linking elements found frequently in this context express, on the semantic axis, causality and finality (Estrada Garc´ıa 2007). In some texts, like the fourteenth-century El Conde Lucanor, a collection of exempla, the rather stereotypical scheme of introduction, exemplum, and moral conclusion, implies a high degree of diversity in the corresponding clauselinking elements. This is why Kamilla Tarasiewicz (2006) proposed, as a complementary element to the global junction pattern analysis, depicting the junction progression in a junctogram, a horizontal graph that follows the tokens of a text to show the position of the different clause linkers within the text. The junctogram, which has been integrated as an export function into the program TraDisc, allows us to see where the linking structure of the text changes considerably; it is possible, for example, to distinguish more narrative from more argumentative passages within texts with the help of this tool. The question that arises from the junctogram visualisation is whether there is a methodological possibility of distinguishing clause-linking elements that derive from the argumentation structure of a text from those serving pragmatic needs (thus somewhat independent from other factors), or if we can show that other factors can play a role. In short: Does textual difference derive only from the different content in the texts or can the same content be expressed in different ways according to different textual traditions or styles? This question will be discussed in the following section.

5. Clause linkage and text structure: The case of Old Spanish Bibles A fortunate circumstance made it possible that in 2007 the T¨ubingen project on clause linkage and discourse traditions could establish research collaboration with the Spanish Bible Corpus project coordinated by Andr´es Enrique-Arias at

Clause linkage and discourse traditions

259

the University of the Balearic Islands (Spain).6 By merging the two approaches, it was possible to apply junction analysis to large parallel corpora of different Old Spanish Bible versions. When comparing different versions of the same text, very similar junction patterns and junction progressions should be expected. In fact, as can be shown in Figure 4, the clause-linking elements are almost identical in four different Old Spanish Bible translations of an example extracted from the book of Numbers (22: 21–35):

Figure 4. Junctogram Numbers (22: 21–35) in four different Old Spanish Bibles. The numbers under the x-axis represent the verses of the chapter under analysis; every clause-linking element in each of the four versions appears once in the upper half of the scheme (semantic complexity) and once in the lower half of the scheme (degree of integration).

What we can see in this example is a rather homogeneous textual progression in all versions: Whereas in verses 21–28 and 31, narrative passages present, above all, additive elements of coordination (III–1) and other elements appear only sporadically, in the two passages marked by the two black rectangles – a passage corresponding to the dialogue between Balaam and the donkey, in contrast to the surrounding narrative context – more integrative structures (degree IV, 3, 5, 6, 7, 8) can be found in all versions.

6. This joint project is funded by the DAAD/MEC program PPP/Acciones Integradas.

260

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

However, there are also passages where not all versions proceed in the same way. In the following example, one of the versions offers a different solution than the rest:

Figure 5. Junctogram Numbers (22: 27–31) in four different Old Spanish Bibles

Here we can see in the passages highlighted in dark grey that one of the versions, E8, shows slightly different solutions in several cases (in 12 cases in this example). In the light-grey passages, all the versions show the same solution or, as in 4 cases, one of the other versions has a divergent solution. Even a “blind” analysis of the different Bible versions based only on objective visualizations of junctors shows that version E8 seems to be different from the rest, and not only in the example presented here, but almost systematically throughout the text. What we can see here is confirmed by the historical information on the exceptionality of E8: As opposed to the other versions that were translated from Hebrew, E8 is a thirteenth-century version translated from the Latin Vulgata. The result is perfectly well-formed Old Spanish in both cases, but the way of saying things is slightly different according to the models behind the texts. The junctogram analysis of nine Old Spanish Bible versions also allowed for a global comparison of the similarity between them.7 The objective abstract projection of junction patterns in this case confirms the results of detailed philological analysis: the more frequent that similar junction patterns between the different versions are, the more these versions have to do with each other historically. E3 and Aj, which are almost 100% identical, are two manuscripts of 7. For a comparison of similarity between Old Spanish Bibles see also Enrique-Arias (2006).

Clause linkage and discourse traditions

261

the same version; GE1 and Faz show disparate values since they are not only translations but also elaborations of the original text, with a number of additions and comments in the case of GE1 and with considerable reductions in Faz. The similarity between GE1 and E8 might be explained by the fact that both texts have the Latin Vulgata as model for the translation.

Figure 6. Similarity between different Old Spanish Bibles according to clause linkage techniques (overall score)

The comparison of several versions of the same text (different versions of the Old Spanish Bible) has shown some interesting results of clause linkage analysis when completed with the more detailed junctogram: The junctogram makes it possible to identify junction patterns that are based on the textual structure of content and it can show deviations from the rule when several versions of one text are compared. In the next section, a different stability, that of a textual genre, will be taken as a starting point for a diachronic study of junction patterns.

6. Clause linkage in Rheto-Romance newspapers Throughout the previous sections, we have changed our focus from a general overview on junctors in the Old Spanish period to a more precise direction: variation and change within a determined passage of the Bible re-elaborated in different traditions. Notwithstanding the fact that we know much about the literary origins of the Old Spanish texts or the Bible translations, it remains diffi-

262

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

cult to reconstruct the exact stylistic effect that the different junction techniques might have created. For this reason, the last section of this paper will be dedicated to a more modern genre. We will, nonetheless, evaluate it in its historical context: the evolution of journalistic style in Rheto-Romance, and more precisely, Romontsch Sursilvan, one of its regional variants spoken in the Upper Rhine valley in the Swiss canton of Grisons. While this variant has a literary history of less than 400 years, journalistic activity has only taken place throughout the last 170 years. This offers an excellent opportunity to observe an emerging genre in an emerging written language within an environment that is much closer to the present than medieval stages of language.8 In a brief study, we have analysed three small partial corpora of approximately 10,000 tokens each, focusing on clause linkage structure. They are organised in chronological order: – From January 1839 we used the first two issues of the Grischun Romontsch (GrR), the valley’s first newspaper, founded in 1836, and the Amitg della Religiun e della Patria (ARP), founded in 1838. – From January 1906, we used excerpts from the first two issues of both the Gasetta romontscha (GaR), a conservative newspaper issued for the first time in the 1850s, and from the Grischun (Gr), its progressive counterpart. – Since there has been a press monopoly in the upper Rhine Valley since World War I, the last diachronic partial corpus we selected is only taken from a single newspaper: La Quotidiana (LQ). The results, displayed in a similar fashion as those for Old Spanish, are displayed in Table 2. The most general observations concerning this corpus are: – In contrast to the Old Spanish corpus, the overall score of junctors shows a decrease rather than an increase through time. – The frequency of overt linkage of finite clauses, both coordinated and subordinated, is clearly diminishing, while the prepositional phrases have remained stable or are even increasing slightly. Comparing the relative weight of junction techniques, we can see that prepositions are becoming more important. This happens at the expense of coordination, but not only that; unlike in Old Spanish, subordination also diminishes with coordination. – Variation within chronological pairs is, again, considerable. It is interesting to observe that the newer of two newspapers usually anticipates the evolution

8. For a more detailed vision of the historical context and for further illustration of the argument outlined below, please refer to Obrist (2009).

263

Clause linkage and discourse traditions Table 2. Junction analyses for Rheto-Romance newspapers (syntax) 1839

1906

2006

GrR

ARP

Total

GaR

Gr

Total

LQ

Coordination II–III

27.3 57.0%

18.2 61.7%

22.9 58.6%

19.0 52.3%

10.4 47.2%

14.4 50.3%

13.6 41.3%

Subordination IV

10.0 20.9%

4.5 15.3%

7.4 18.9%

6.7 18.4%

1.9 8.6%

4.1 14.3%

5.1 15.5%

Gerunds V

0.6 1.3%

1.0 3.4%

0.8 2.0%

2.1 5.8%

1.5 6.8%

1.8 6.3%

0.4 1.2%

PrepP VI–VII

10.0 20.9%

5.8 19.7%

8.0 20.5%

8.5 23.4%

8.2 37.3%

8.3 29.0%

13.8 41.9%

Total

47.9

29.5

39.1

36.3

22.0

28.6

32.9

to come: ARP was not founded until 1839, and Gr in 1905, whereas their counterparts GrR and GaR were already established; however, we cannot claim statistical significance in this respect. We will seek insight into the statistical evolution using illustrative passages of the underlying texts. The opening article from the international news section of the ARP, for example, begins as follows: (1) Frontscha. A Paris fuva la generala creta che cun la redunonza dils deputai vignessan las gassas bubrontadas cun saung, ni silmains rumpessi ora ina ravolta. Denton ein las combras aviartas senza disturbar strusch ina miur. Bein zvar che sez il retg haveva tema de malemperneivlas novadats. Pertgei tutt Paris fuva sin quella chischun emplenius cun postas militaras et il retg circumdaus cun extraordinarias excortas. Finalmegn arrivaus cun il tgau saun el saal della redunonza, ha el teniu tier ils deputai in plaid de quest contegn: [. . . ] Aber per miserabla memoria ha Philipp gnianc tratg a strada las strepitusas demonstratiuns e tiarratriembels sin ils confins della Svizzera, sonder lura sin in solemna vivat dils deputai turnaus cun piall sauna en siu casti. ‘In Paris, the general opinion was that the meeting of the Parliament would cause the streets to flood with blood, or at least, that a revolt would break out. However, the houses opened without even startling a mouse. It is true that even the king was prepared for bad news. For this reason, all of Paris was full of military checkpoints on this occasion, and the king [himself] surrounded by an unusual amount of guards. Finally, upon arriving in an uninjured state, he addressed the deputies with the following words: [. . . ] But because of his bad memory, Philipp did not even bring up the loud demonstrations and unrest on the Swiss

264

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

boarder; instead, he [then] returned safely to his castle with a solemn ‘vivat’ to the deputies.’

The content of the article is, actually, a simple list of facts. The arguments can be summarised as follows: Paris had been fearing a general uprising on the occasion of the parliament meeting. Security measures were taken. The king held a speech without referring to the agitation on the Swiss border; he returned to his palace. However, if we were given only the series of junctors: denton (‘all the same’). . . bein zvar che (‘it is true that’). . . pertgei (‘for this reason’). . . aber (‘but’). . . per (‘because of ’). . . gnianc/ sonder (‘not even’/‘instead’), our first impression would probably be that this is an editorial article. Indeed, we have what Biber (1993: 11) calls an overt expression of persuasion: a high frequency of conjunctions expressing causality and similar relations (concessivity is in fact based on causality). The following article from the first issue of Grischun in 1906 also deals with a situation of crisis. However, in comparison to the former text, its style seems somewhat more modern, if still not quite contemporary to us: (2) Russia. La situaziun ei cheu disparada; eun paucs plaids: Revoluziun e confusiun sin tut la lingia. Ad Odessa han ils consuls dellas differentas naziuns perfin decidiu de tener prompts ils bastiments, per transportar a casa lur compatriots pil cass de basegns. Il consul svizzer ha domondau il cussegl federal, sch’el vegli surprender las spesas de transport de quels Svizzers, che seigin senza miezs. Il cussegl federal ha respondiu en senn affermativ, resalvont il regress vid ils cantuns u las vischnauncas. ‘The situation here has exploded: in a few words, revolution and confusion everywhere. In Odessa, the consuls of several countries have finally decided to prepare ships in order to repatriate their citizens in case of necessity. The Swiss consul has asked the government if it is willing to pay for the transport of these Swiss, [alleging] that they are without means. The government has answered affirmatively, reserving [for itself] [the right of] recourse against the cantons and municipalities.’

The differences between (1) and (2) are obvious: The newer article (2) shows less overt linkage and when it uses junctors, the junctors are in the form of noun and infinitive phrases more often than the junctors are in the older article (1): per transporar a casa ‘in order to bring home,’ pil cass de basegns ‘in case of necessity,’ resalvont il regress ‘reserving the regress.’ Furthermore, the underlying semantic relations are not made explicit: In the latter example, the possiblity of holding the repatriates responsible for the expenses they have caused of course establishes a relation of condition or contrast with the assistance they are given by the government, but there is no but or although.

265

Clause linkage and discourse traditions

Does overt clause linkage drop between 1839 and 1906 because of a more implicit structure of argumentation? A re-classification of the data seems to confirm this intuition. We have split up the coordinating and subordinating junctors into two semantic groups: those usually considered as descriptive and narrative (e ‘and,’ lura ‘then,’ etc.), and those that show expository or argumentative functions (sche ‘if,’ perquei che ‘because,’ aber ‘but,’ etc.). The results show a more differentiated image: The high overall junctor score for 1839 is indeed based on argumentative functions, including condition, cause, and concession. Table 3. Junction analyses for Rheto-Romance newspapers (semantics) 1839

1906

2006

GrR

ARP

Total

GaR

Gr

Total

LQ

Addition & temporal

18.4 38.4%

10.3 34.9%

14.5 37.1%

19.4 53.4%

8.9 40.5%

13.7 47.9%

12.8 38.9%

of which et

10.0

6.5

8.4

10.2

4.5

7.2

5.7

Argumentative functions

29.5 61.6%

19.2 65.1%

24.6 62.9%

16.9 46.6%

13.1 59.5%

14.9 52.1%

20.1 61.1%

Total

47.9

29.5

39.1

36.3

22.0

28.6

32.9

A major deviance from the general evolution is the value of additional and temporal junctors in the Gasetta Romontscha. This is mainly due to the junctor e (‘and’), the frequency of which is twice the value as of the coetaneous Grischun paper. It is of course legitimate to inquire about the reasons for this important variation. We must insist, however, that this is not a mere issue of corpus design: E-junctors are well distributed throughout the text, and it is impossible to find a passage with an excessive presence of enumerations. If the reason for the variation is idiosyncrasy, the idiosyncrasy must pertain to the style of the newspaper as a whole, and not to that of a particular article. Consider the following example, taken from the local news section: (3) Als vischins dallas tracziuns de Puschlav vuleva la malvegnida prescripziun de stuer pinar la lena sort da cumminonza buc el tgau e han els perquei concludiu de reparter dad els anora la lena marcada ed ein era, encuter el scammond della suprastonza comunala, sepatroni de quella. ‘To the residents of the Puschlav villages, the unwelcome prescription of dividing timber among the community did not make any sense, and they therefore decided to distribute the timber of their area by themselves, and they also, against the prohibition of the council, took possession of it.’

266

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

The information here is given as a sequence: malcontentedness with the rules set by the authorities, disobedient autonomous repartition, and finally usurpation. There is topic continuity in the sense that the thematic position of all three sentences is occupied by the Puschlav residents. In two of the three sentences, they also make up the grammatical subject. There is a clear structure in this passage; however, the information is presented in a style that does not seem to be as close to an objective, journalistic reporting of the facts it could be. Its incremental character (conjunctions followed by adverbs) makes it somehow rhetorically marked. Consider the local news from 2006, here, in form of an obituary: (4) Gest il mument ch’il temps da pensiun era arrivaus, ha sia sanadat entschiet a far quitaus. Ina malsogna maligna s’annunzia e s’avonza cun ried. Liungs tractaments medicinals, diversas dimoras en spitals e la buna tgira a casa han bein purtau levgera, mo buc il migliurament ton desiderau. ‘At the very moment of reaching the age of retirement, his health began to decline. A malign illness announced itself and proceeded quickly. Long medical treatments, long stays at hospitals, and good care at home brought relief indeed, but not the recovery so hoped for.’

This passage is similar to the precedent in its length (three sentences) as well as in the organisation of its content: Someone has reached retirement age, he is diagnosed with cancer, he is given every possible treatment, and he still dies. The way this information is presented is different, however: The uniting semantic feature of the three sentences (and the whole article), namely the patient, is neither featured in thematic function nor is it the grammatical subject. Instead, each sentence begins with a different agent: first the general health condition, then the illness, and finally the treatment. Coherence between the sentences is mainly lexical; there are hardly any overt signals of cohesion, except perhaps for the verb tense. It is important to observe that (4) can easily be paraphrased in such a way that topic continuity be respected: Just in the moment of reaching the age of retirement, his health started to decline (and) he was diagnosed with a fast growing cancer. As observed by Kreyer (this volume), a topic shift as in (4) appears to be an indicator in and of itself. It would indeed be interesting to correlate the presence of junctors – especially coordinators – with phenomena such as inversion and dislocation. For the time being, the findings of this section may be summarised as follows. The main goal of newspaper texts is transmitting news, and this has been so ever since newspapers have existed. Still, clause linkage shows considerable diachronic and synchronic variation. First, overt clause linkage tends to decline over the years. Second, this decline affects conditionals and causals more

Clause linkage and discourse traditions

267

strongly than it affects other semantic areas of clause linkage. Third, there is a diachronic shift from more aggregate to more integrate types of linkage. Many of the phenomena described in this section have been said to be functional necessities of journalistic writings (see L¨uger 1983; Kirstein 1997; Carreira 1999). For example, Eggers (1973: 46) replaces a noun cluster taken from a newspaper article by a series of verb phrases that reach twice the length of the initial structure. He cites this as evidence that noun phrases fit more easily into journalistic language because they allow for more concise formulations (see also Puschmann, this volume). However, it is important to distinguish between textual function and traditionality of discourse. The argument of conciseness does not necessarily apply to examples like (4), where pil cass de basegns (‘in case of necessity’) could easily be rendered alternatively with sche quei fuss necessari (‘if this is necessary’). Moreover, apart from a very modest contraction of the text, there is no need to avoid causal markers and series of e-junctors. From a synchronic point of view, it can be argued that newspaper language, as we describe it, obeys conventionalised patterns or discourse traditions. From a diachronic point of view, we can retrace these patterns to their origins and describe their evolution. An important terminus in this respect is genre differentiation, as described by Wilhelm (2001: 472), a fundamental mechanism in the creation of new discourse traditions. Many of the nineteenth-century newspaper articles we analysed resemble essays or sermons. Wilhelm’s research on the origins of journalism in Italy shows that these are precisely the genres which served as models for emerging journalism. If it is true that today’s commentaries and lead articles still show some affinity to older journalistic language, as we have stated with respect to example (1), we might even postulate that commentaries follow a more archaic discourse tradition than ordinary news texts.9

7. Conclusions The question we have discussed in this paper is to what extent textual traditions, in the sense of discourse tradition theory, have to do with patterns of clauselinking elements. The initial hypothesis, that clause linkage might be a strong indicator of different textual traditions, was supported by the traditional view on textual elaboration and by our previous studies, where a strong correlation between the two was found in some Old Spanish texts. We therefore argued that 9. An interesting formal mark in this respect is that the renowned German newspaper Frankfurter Allgemeine Zeitung still prints the titles of its commentaries in Gothic letters, while normal serifs are used for all other headings.

268

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

not only different clause-linking techniques, but also their relative frequencies, are symptomatic of different textual traditions, and this is so not only in the context of genres but also in the context of sub-tradition(s) within a single genre. In the initial design of our study, we aimed to create global descriptions of the clause linkage patterns found in each text and we applied this process to three corpora: a diachronic corpus of Old Spanish texts including different genres, an Old Spanish parallel corpus of Bible texts, and a Rheto-Romance diachronic newspaper corpus from the nineteenth and twentieth centuries. It is obvious that texts are complex entities and that any reductive projection of text properties, even a bi-dimensional and elaborated one such as the one that was proposed in our study, necessarily requires that strong limitations be imposed upon it. But instead of giving up on looking for criteria that can be formalised and limiting ourselves to purely philological descriptions of texts as incomparable individuals, we believe that formalisations are necessary, and this is perhaps even more the case for purposes other than textual typology alone. Textual typology, in the sense of a self-sufficient discipline, can work with complex criteria to define the individual properties of a text, but if we think, for example, that corpus linguistics is interested in genre differentiation in order to adequately describe grammatical variation in a language or the evolution of grammar in diachrony, we need reductive criteria that also allow for a comparison of larger amounts of texts. Nevertheless, and even if clause linkage patterns are strong indicators of certain textual traditions, other features that might be even more indicative of the differentiation of certain text traditions should also be taken into account. On the other hand, in some cases it seems clear that textual traditions can be identified via clause linkage analysis. In the case of the Bible corpus, it was even possible by objective measure of clause-linking elements to confirm the philological reconstruction of similarity and difference between different versions, and the junctogram analysis allowed, in some of the Bible examples, for the distinguishing of factors of textual progression from individual deviances between the different versions. Obviously, clause linkage analysis also allows for the evolution of clause linkage techniques themselves when applied to a diachronic corpus that is limited to one genre. In this way, the evolution of the Rheto-Romance newspapers showed the Ausbau (‘elaboration’) of this genre from a more aggregative to a more integrative style in the last two centuries. What our results perhaps more clearly show are the possible achievements and limitations of the study of junction patterns for historical linguistics. For more general observations on the significance of clause linkage patterns for genre characterisation (or, in a more general way, for the identification of discourse traditions), we will need to study a larger number of texts from different

Clause linkage and discourse traditions

269

languages, not only from Romance languages and not only in a diachronic perspective, and we will have to compare the clause linkage patterns with other linguistic features. It seems that the tendency in linguistics to simplify or even to avoid the question of genre differentiation has to do with the enormous complexity of textual traditions. Corpus linguistics deals with an ever-increasing mass of data and sometimes this simple mass allows for important observations. However, the analysis of a large data mass also creates new questions, and since the mass only allows for a first look at the complex and differentiated situation or evolution behind the data, we absolutely need criteria for the objective analysis of the discourse traditions of which our corpora are composed. This differentiation will not only lead to an atomistic view that does not allow for more general states or evolutions to be seen, but it will contribute to an adequate conception of what a language really is and how some phenomena are text specific or discourse tradition specific, while others affect the language as a whole. This is surely the most important contribution the notion of discourse traditions can make to linguistic theory.

Primary Sources Bibles: Aj = Alb = E3 = E4 = E7 = E8 = E19 = Faz = GE 1 =

Biblioteca de Ajuda 52-xii-1 [manuscript] Biblia de Alba, Madrid, Palacio de Liria [manuscript] Escorial I.i.3 [manuscript] Escorial I.i.4 [manuscript] Escorial I.i.7 [manuscript] Escorial I.i.8 [manuscript] Escorial I.ii.19 [manuscript] Fazienda de Ultramar, Salamanca, Universidad 1997 [manuscript] Madrid, Biblioteca Nacional, 816 [manuscript]

Rheto-Romance newspapers: GrR = Il Grischun romontsch. Cuera: Stamparia S. Benedict, 1836–1839 [excerpt: 1–8 Jan 1839] ARP = Amitg della Religiun e della Patria. Surrhein/ Somvitg, 1838–1839 [excerpt: 7–21 Jan 1839] GaR = Gasetta romontscha. Must´er: Condrau, 1857–1996 [excerpt 4–11 Jan 1906] Gr = Il Grischun. Cuera, 1905–1916 [excerpt: 5–12 Jan 1906] LQ = La Quotidiana. Cuira: Gasser, 1997- [excerpt: 12 Jan 2006]

270

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

References Aschenberg, Heidi 2003 Diskurstraditionen – Orientierungen und Fragestellungen. In: Heidi Aschenberg and Raymund Wilhelm (eds.), Romanische Sprachgeschichte und Diskurstraditionen, 1–18. (T¨ubinger Beitr¨age zur Linguistik 464.) T¨ubingen: Max Niemeyer. Barra, Mario 2008 Tradici´on discursiva, creaci´on y difusi´on de innovaciones sint´acticas: la cohesi´on de los argumentos nominales a partir del siglo XIII. In: Johannes Kabatek (ed.), Sintaxis hist´orica del espa˜nol y cambio ling¨u´ıstico: Nuevas perspectivas desde las Tradiciones Discursivas, 127–150. Frankfurt/Madrid: Vervuert/ Iberoamericana. Biber, Douglas 1993 Using register-diversified corpora for general language studies. Computational Linguistics 29: 219–241. Biber, Douglas 2004 Historical patterns for the grammatical marking of stance: A crossregister comparison. Journal of Historical Pragmatics 5: 107–135. Biber, Douglas, Susan Conrad and Randi Reppen 1998 Corpus Linguistics: Investigating Language Structure and Use. (Cambridge Approaches to Linguistics.) Cambridge: Cambridge University Press. Biber, Douglas, Mark Davies, James K. Jones and Nicole Tracy-Ventura 2006 Spoken and written register variation in Spanish: A multi-dimensional analysis. Corpora 1: 7–38. Carreira, Maria Helena Araujo 1999 Faits et effets linguistiques dans la presse actuelle (Espagne, France, Italie, Portugal). (Travaux et Documents 4.) Paris: Universit´e Paris 8. Castilho, Alessandra 2009 Transforma¸ca˜ o de gˆeneros discursivos em uma perspectiva diacrˆonica: o exemplo da not´ıcia [Transformation of discourse genres in a diacronic perspective: The example of the not´ıcia]. In: Ataliba de Castilho (ed.), Hist´oria do Portuguˆes Paulista [History of S˜ao Paolo Portuguese], 637– 663. (S´erie Estudos, volume 1.) Campinas: IEL/Unicamp/FAPESP. Ciapuscio, Guiomar, Konstanze Jungbluth, Dorothee Kaiser and C´elia Lopes (eds.) 2006 Sincron´ıa y diacron´ıa de tradiciones discursivas en Latinoam´erica. (Bibliotheca Ibero-Americana 107.) Frankfurt/Madrid:Vervuert/Iberoamericana. Coseriu, Eugenio 1988 Sprachkompetenz. Grundz¨uge derTheorie des Sprechens. (Uni-Taschenb¨ucher 1481.) T¨ubingen: Francke. Eggers, Hans 1973 Deutsche Sprache im 20. Jahrhundert. (Serie Piper 61.) M¨unchen: Piper.

Clause linkage and discourse traditions

271

Enrique-Arias, Andr´es 2006 Sobre el parentesco entre la Biblia deAlba y la Biblia de la RealAcademia de la Historia Ms. 87. Romance Philology 59: 241–263. Estrada Garc´ıa, Rosa Mar´ıa 2007 La tradici´on de los pr´ologos en el espa˜nol medieval. M.A. Thesis, Universit¨at T¨ubingen. Fern´andez Alcaide, Marta and Araceli L´opez Serena (eds.) 2007 Cuatrocientos a˜nos de la lengua del Quijote. (Actas: Universidad de Sevilla 64.) Sevilla: Universidad de Sevilla. Frank, Barbara, Thomas Haye and Doris Tophinke (eds.) 1997 Gattungen mittelalterlicher Schriftlichkeit. (Script Oralia 99.) T¨ubingen: Gunter Narr. Guzm´an River´on, Marta 2007 Tradiciones discursivas e historia de la lengua espa˜nola en Am´erica. In: Fern´andez Alcaide and L´opez Serena (eds.), Cuatrocientos a˜nos de la lengua del Quijote, 79–88. (Actas: Universidad de Sevilla 64.) Sevilla: Universidad de Sevilla. Haskins, Charles Homer 1927 The Renaissance of the Twelfth Century. Cambridge, MA/London: Harvard University Press. Hockett, Charles [1963] 1966 The problem of universals in language. In: Joseph Greenberg (ed.), Universals of Language, 1–29. Cambridge, MA/London:The MIT Press. Jacob, Daniel and Johannes Kabatek (eds.) 2001 Lengua medieval y tradiciones discursivas en la Pen´ınsula Ib´erica. Descripci´on gramatical – pragm´atica hist´orica – metodolog´ıa. Frankfurt: Vervuert. Kabatek, Johannes 2001 ¿C´omo investigar las tradiciones discursivas medievales? El ejemplo de los textos jur´ıdicos castellanos. In: Daniel Jacob and Johannes Kabatek (eds.), Lengua medieval y tradiciones discursivas en la Pen´ınsula Ib´erica. Descripci´on gramatical – pragm´atica hist´orica – metodolog´ıa, 97–132. Frankfurt: Vervuert. Kabatek, Johannes 2005a Tradiciones discursivas y cambio ling¨u´ıstico. Lexis 29: 151–177. Kabatek, Johannes 2005b Die Bolognesische Renaissance und der Ausbau romanischer Sprachen. Juristische Diskurstraditionen und Sprachentwicklung in S¨udfrankreich und Spanien im 12. und 13. Jahrhundert. (Beihefte zur Zeitschrift f¨ur romanische Philologie 321.) T¨ubingen: Max Niemeyer.

272

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

Kabatek, Johannes (ed.) 2008 Sintaxis hist´orica del espa˜nol y cambio ling¨u´ıstico: Nuevas perspectivas desde las Tradiciones Discursivas. Frankfurt/Madrid: Vervuert/ Iberoamericana. Kirstein, Corinna Manuela 1997 Textlinguistische Analyse informationsbetonter Textsorten der spanischen Zeitung El Pa´ıs. (Europ¨aische Hochschulschriften 186.) Frankfurt/Berlin/Bern: Peter Lang. Koch, Peter and Wulf Oesterreicher 1985 Sprache der N¨ahe – Sprache der Distanz. M¨undlichkeit und Schriftlichkeit im Spannungsfeld von Sprachtheorie und Sprachgeschichte. Romanistisches Jahrbuch 26: 15–43. Koch, Peter and Wulf Oesterreicher 1990 Gesprochene Sprache in der Romania. Franz¨osisch, Italienisch, Spanisch. (Romanistische Arbeitshefte 21.) T¨ubingen: Max Niemeyer. Koch, Peter and Wulf Oesterreicher 1994 Schriftlichkeit und Sprache. In: Hartmut G¨unther and Otto Ludwig (eds.), Schrift und Schriftlichkeit. Ein interdisziplin¨ares Handbuch internationaler Forschung, 587–604. (Handb¨ucher der Sprach- und Kommunikationswissenschaft 20.) Berlin/New York: de Gruyter. Koch, Peter 1997 Diskurstraditionen. Zu ihrem sprachtheoretischen Status und ihrer Dynamik. In: Barbara Frank, Thomas Haye and Doris Tophinke (eds.), Gattungen mittelalterlicher Schriftlichkeit, 43–79. (Script Oralia 99.) T¨ubingen: Gunter Narr. Koch, Peter 2008 Tradiciones discursivas y cambio ling¨u´ıstico: el ejemplo del tratamiento vuestra merced en espa˜nol. In: Kabatek (ed.), Sintaxis hist´orica del espa˜nol y cambio ling¨u´ıstico: Nuevas perspectivas desde las Tradiciones Discursivas, 53–88. Frankfurt/ Madrid: Vervuert/ Iberoamericana. Kortmann, Bernd 1997 Adverbial Subordination. A Typology and History of Adverbial Subordinators Based on European Languages. (Empirical Approaches to Language Typology 18.) Berlin/New York: Mouton de Gruyter. Kortmann, Bernd 2001 Adverbial conjunctions. In: Martin Haspelmath, Ekkehard K¨onig, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals/ Sprachtypologie und sprachliche Universalien/ La typologie des langues et les universaux linguistiques – An International Handbook/ Ein internationales Handbuch/ Manuel international. (Handb¨ucher zur Sprach- und Kommunikationswissenschaft 20.), 842– 854. Berlin/New York: de Gruyter.

Clause linkage and discourse traditions

273

Kreyer, Rolf this volume Syntactic constructions as a means of spatial representation in fictional prose. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 277–303. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Kroch, Anthony 2001 Syntactic change. In: Mark Baltin and Chris Collins (eds.), The Handbook of Contemporary Syntactic Theory, 699–730. (Blackwell Handbook of Linguistics.) Malden, MA: Blackwell. L´opez Serena, Araceli 2007 Las tradiciones discursivas en la historiograf´ıa ling¨u´ıstica y en la historia de la lengua espa˜nola. In: Marta Fern´andez Alcaide and Araceli L´opez Serena (eds.), Cuatrocientos a˜nos de la lengua del Quijote, 49–54. (Actas: Universidad de Sevilla 64.) Sevilla: Universidad de Sevilla. L¨uger, Heinz-Helmut 1983 Pressesprache. (Germanistische Arbeitshefte 28.) T¨ubingen: Max Niemeyer. Obrist, Philipp 2009 Lungatg quotidian e lungatg dalla quotidiana – Der syntaktische Ausbau des Bündnerromanischen aus diskursiver Sicht. In: Cristina Albizu et al. (eds.), Alltag – Quotidien – Quotidinne – Quotidiano, 197–216. Aachen: Shaker. Oesterreicher, Wulf 1997 Zur Fundierung von Diskurstraditionen. In: Barbara Frank, Thomas Haye and Doris Tophinke (eds.), Gattungen mittelalterlicher Schriftlichkeit, 19–41. (Script Oralia 99.) T¨ubingen: Gunter Narr. Oesterreicher, Wulf, Eva Stoll and Andreas Wesch (eds.) 1998 Competencia escrita, tradiciones discursivas y variedades ling¨u´ısticas. T¨ubingen: Gunter Narr. Parodi, Giovanni (ed.) 2007 Working with Spanish Corpora. (Corpus and Discourse.) London: Continuum. Pons Rodr´ıguez, Lola 2007 Ret´orica y tradiciones discursivas. In: Fern´andez Alcaide and L´opez Serena (eds.), Cuatrocientos a˜nos de la lengua del Quijote, 67–78. (Actas: Universidad de Sevilla 64.) Sevilla: Universidad de Sevilla. Pons Rodr´ıguez, Lola 2008 El peso de la tradici´on discursiva en un proceso de textualizaci´on: un ejemplo en la Edad Media castellana. In: Johannes Kabatek (ed.), Sintaxis hist´orica del espa˜nol y cambio ling¨u´ıstico: Nuevas perspectivas desde las Tradiciones Discursivas, 197–224. Frankfurt/Madrid: Vervuert/Iberoamericana.

274

Johannes Kabatek, Philipp Obrist, and Valentina Vincis

Puschmann, Cornelius this volume Thank you for thinking we could: Use and function of interpersonal pronouns in corporate web logs. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 167–191. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Raible, Wolfgang 1992 Junktion. Eine Dimension der Sprache und ihre Realisierungsformen zwischen Aggregation und Integration. (Sitzungsberichte der Heidelberger Akademie der Wissenschaften, Berichtheft 2.) Heidelberg: Winter. Raible, Wolfgang 2001 Linking clauses. In: Martin Haspelmath, Ekkehard K¨onig, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals/ Sprachtypologie und sprachliche Universalien/ La typologie des langues et les universaux linguistiques – An International Handbook/ Ein internationales Handbuch/ Manuel international, 590–617. Berlin/New York: de Gruyter. S´aez Rivera, Daniel 2007 Tradiciones discursivas, historiograf´ıa ling¨u´ıstica e historia de la lengua. In: Fern´andez Alcaide and L´opez Serena (eds.), Cuatrocientos a˜nos de la lengua del Quijote, 89–112. Sevilla: Universidad de Sevilla. Schlieben-Lange, Brigitte 1983 Traditionen des Sprechens. Elemente einer pragmatischen Sprachgeschichtsschreibung. Stuttgart: Kohlhammer. Sim˜oes, Jos´e da Silva 2007 Sintaticiza¸ca˜ o, discursiviza¸ca˜ o, e semanticiza¸ca˜ o das ora¸co˜ es de ger´undio no portuguˆes brasileiro [Syntaxization, discoursization, and semantization of gerund phrases in Brazilian Portuguese]. PhD. Thesis, Universidade de S˜ao Paulo. Schmidt-Riese, Roland 2003 Relatando M´exico. Cinco textos del per´ıodo fundacional de la colonia en Tierra Firme. (Textos y documentos espa˜noles y americanos 3.) Madrid: Iberoamericana. Stoll, Eva 1997 Konquistadoren als Historiographen. Diskurstraditionelle und textpragmatische Aspekte in Texten von Francisco de Jerez, Diego de Trujillo, Pedro Pizarro und Alonso Borreg´an. (Scriptoralia 91.) T¨ubingen: Gunter Narr. Tarasiewicz, Kamilla 2006 “Junktographie”. Eine Methode zur Darstellung vonTextkoh¨asion. Staatsexamen Thesis, Universit¨at T¨ubingen.

Clause linkage and discourse traditions

275

Vincis, Valentina 2008 T´ecnicas de junci´on y tradiciones discursivas medievales: el testimonio de los textos. In: Concepci´on Company (ed.), Actas del VII Congreso Internacional de Historia de la Lengua Espa˜nola, 2201–2222. Madrid: Arco. Wilhelm, Raymund 1996 Italienische Flugschriften des Cinquecento (1500–1550). (Beihefte zur Zeitschrift f¨ur romanische Philologie 279.) T¨ubingen: Max Niemeyer. Wilhelm, Raymund 2001 Diskurstraditionen. In: Martin Haspelmath, Ekkehard K¨onig, Wulf Oesterreicher and Wolfgang Raible (eds.), Language Typology and Language Universals/ Sprachtypologie und sprachliche Universalien/ La typologie des langues et les universaux linguistiques – An International Handbook/ Ein internationales Handbuch/ Manuel international, 467– 477. (Handb¨ucher zur Sprach- und Kommunikationswissenschaft 20.) Berlin/New York: de Gruyter.

Syntactic constructions as a means of spatial representation in fictional prose1 Rolf Kreyer

The present paper explores how six syntactic constructions with fronted locative constituents are used for the representation of space in fictional prose. More specifically, on the basis of 188 tokens drawn from the British National Corpus the paper describes how these constructions are exploited to create an effect that variously has been described as imaginary guided tour (Drubig 1988), camera movement (Dorgeloh 1997) or immediate-observer effect (Kreyer 2007). Although all of the six construction types are equally suited to create this particular effect, the data show large differences in the distribution of the individual types. The paper discusses various factors that might influence the choice of construction (e.g. information status and textual functions) and explores to what extent the individual constructions contribute to our understanding and our perception of the genre fictional prose.

1. Introduction Over the last two decades, the study of full-verb inversion (e.g. in the garden was a man) in narrative discourse has brought to light the potential of this construction to help readers immerse themselves into a narrative. Researchers make use of concepts like “imaginary guided tour” (Drubig 1988: 87), “eyewitness perspective,” “camera movement” (Dorgeloh 1995: 228 and 1997: 110), or “immediate-observer effect” (Kreyer 2006a, 2006b, 2007) to describe this phenomenon. This particular use of full-verb inversions is motivated by the fact that the reader is usually confronted with a situation of displacement: Narrative texts report events that are distant in time and place; the reader does not have immediate access to the described events. As a consequence, “the speaker’s discourse must be designed to make the addressee construct a vivid representation of a perceptually inaccessible reference situation” (Drubig 1988: 88). Such a 1. I would like to thank J¨urgen Esser, Harold Fish, Sebastian Patt, Sharmila Vaz, and the editors of this volume for comments on earlier versions of this paper.

278

Rolf Kreyer

“vivid representation,” according to Kreyer (2006a and 2006b), can be achieved if the text (and the constructions therein) mirrors or simulates natural perception. “Natural perception” refers to the fact that perception is usually smooth: If, for instance, a person enters a room, he or she will find a desk at the end of the room. Looking at the desk, the person will become aware of a folder lying on that desk. A closer look at the folder might reveal that something is written on this folder, and so on. That is, natural perception moves from an object A to an object B, where object B is usually perceived in relation to object A. The potential of inverted constructions to mirror natural perception lies in the fact that these constructions provide a prime means to represent, iconically, the shift of focus from object A to object B through the order of constituents: (A part of) the preposed constituent makes reference to object A, while the postposed subject makes reference to B. As has been hinted at in Kreyer (2006a: 129–131), however, other constructions, such as clauses with preposed locative adverbials or cases of there-insertion (i.e. existential or presentational there-constructions with preposed locative expression), have a similar potential since they, too, allow for an initial locative phrase and then the introduction of a new object. This suggests, that “inversions, although different with regard to syntactic properties, seem to be equivalent to clauses with preposed locative adverbial and thereinsertion as regards their function in narrative discourse, namely to simulate natural perception in order to enable the reader to immerse into the narrative” (Kreyer 2006a: 130–131). The present study explores the use of full-verb inversions and other constructions with regard to the immediate-observer effect in prose fiction. To that end, an exhaustive corpus analysis is employed. This is the only way to get an idea of the extent of use of particular constructions within a particular genre, and thus, such an analysis can provide information on the association of individual constructions with the genre prose fiction. The paper falls into two major parts. The first part analyses the immediate-observer effect in a number of different constructions. After a discussion of experimental findings on the verbalization of spatial configurations, the paper identifies strategies of representing space through language and, in a corpus-based analysis, explores the ways in which six construction types are exploited to achieve the immediate-observer effect. The second part is rather exploratory and tries to identify factors that may influence the choice of one construction over others. The study will conclude with some suggestions for further research along the lines sketched out in the second part of the paper.

Syntactic constructions as a means of spatial representation in fictional prose

279

2. Syntactic constructions and the representation of space in language 2.1. Theoretical background Relevant to the present discussion are a number of experimental studies on the verbalization of spatial relations (e.g. Linde 1974; Linde and Labov 1975; Ullmer-Erich 1982; Ehrich and Koster 1983; see Kreyer 2006a for a detailed discussion).These studies suggest that, to a large extent, the description of spatial relations can be described by two strategies, namely what Tversky (1996) terms “route” and “survey descriptions.” She writes: A route description takes ‘you,’ the reader, on a mental tour; it uses a changing view from within the environment, and locates landmarks with respect to you in terms of ‘your’ front, back, left, and right. A survey description, in contrast, takes a static view from above the environment and locates landmarks with respect to each other in terms of north, south, east, and west. (Tversky 1996: 479)

The two types of descriptions are illustrated in examples (1) and (2), respectively. Note that elements of both types are usually mixed in descriptions, but in most cases one of the two strategies predominates, as in the examples below: (1)

As you open the door, you are in a small five-by-five room which is a small closet. When you get past there, you’re in what we call the foyer [. . . ]. If you keep walking in that same direction, you’re confronted by two rooms in front of you [. . . ]. And on the right side, straight ahead of you again, is a dining room which is not too big [. . . ]. And on your left you would find the master bedroom [. . . ]. (Linde and Labov 1975: 929–30, emphasis added)

(2)

Well, er as you enter the door, immediately to the right of it is the desk against the wall, connected to it is the bed and then comes the corner going up to the window, and there between the window and the wall is this bookshelf [. . . ]. (Ullmer-Ehrich 1982: 233, emphasis added)

From the above it becomes clear that both strategies correlate with different construction types. While in both cases, reference to the location is made clause initially, the object to be located is encoded in the direct object of the clause with route descriptions (the master bedroom), whereas survey descriptions often make use of inverted constructions in which the postposed subject denotes the entity to be located (the desk against the wall). Leaving aside for the moment other constructions, the distinction above hints at one important aspect with regard to the verbalization of spatial relations in narrative discourse. While in the first case a perceiving entity is mentioned explicitly in the clause, the perceiver

280

Rolf Kreyer

remains implicit in the second case. Drawing on Langacker (1990), Dorgeloh (2006: 105) argues that this leads to differences in the way the described spatial configurations are construed. More specifically, the explicit mention of a perceiving entity will lead to a more objective construal of the discourse universe. In contrast, “inverted constructions are more subjective, because the reader’s focus of attention is directed only to the (conceptual) scene expressed via language”; events are presented “in a way similar to immediate experience” (Dorgeloh 1997: 165). The reader, through the text, can assume an “eyewitness perspective” (Dorgeloh 1995: 228). It follows that inverted constructions should be a prime means to create the immediate-observer effect, whereas constructions that make a perceiving entity explicit are less suited in this respect. This point will be discussed in more detail below. In an attempt to explain why inverted constructions are such a suitable means for representing spatial relations, Kreyer (2006a) distinguishes two kinds of survey descriptions, depending on the point of reference that is taken. One possibility is to take the observer of a scene as the point of reference. This observer is usually fixed and the entities perceived are located relative to the observer, i.e. the observer serves as point of reference. The effect can be likened to a series of photographs where the individual pictures show different and nonoverlapping parts of the same scene: this way, it is impossible to establish the location of the entities relative to each other. In the following I well refer to this strategy as photo series.2

Figure 1. The observer as point of reference (photo series)

(3)

In front of him was a bungalow. It seemed expensive. Behind him grew a palm tree. It looked healthy but did not carry any coconuts. Also behind him stood a bicycle. It was an ugly red colour and looked rather old-fashioned.

2. Note that this strategy could also be understood as a special instance of a route description, since objects are located with regard to an observer. However, I think the main difference between route and survey descriptions is that between a “mental tour” and a “static view.” From this perspective, the photo series can be understood as an instantiation of survey descriptions.

Syntactic constructions as a means of spatial representation in fictional prose

281

Another (more frequently realized) possibility is to locate entities of the discourse universe in relation to one another, i.e. we are dealing with shifting points of reference. If the location of entities is described in this way, a path between two objects is established which the reader (as observer) can follow. The effect, here, is that of a camera moving, hence (following Dorgeloh 1995: 228, 1997: 110) I will denote this strategy with the term camera movement:

Figure 2. Shifting points of reference (camera movement)

(4)

He looked at the bungalow. It seemed expensive. On the right of the building was a palm tree. It looked healthy but did not carry any coconuts. Below the tree stood a bicycle. It was an ugly red colour and looked rather old-fashioned.

With the second case, inverted constructions are a useful means to mirror the shift of the observer’s gaze from an object A, which has been under scrutiny up to this point, to an object B, which is going to be looked at.3 Kreyer (2006a) finds that the order of constituents in inverted constructions is in line with studies on perception in that the preposed constituent usually makes reference to those objects that would be perceived first, due to their higher relative visual prominence. Capitalizing on studies by Talmy (1983), Ehrich and Koster (1983), and Wenz (1996), Kreyer (2006a) identifies four possible relations that mirror the higher visual prominence of the object referred to in the preposed constituent compared to the lower visual prominence of the object denoted by the postposed subject: First, the former is larger than the latter; second, it is more permanently located; third, the preposed constituent makes reference to an object that is closer to the perceiver than the object denoted in the postposed subject; and, fourth, in those cases where one object is contained in another, the entity represented by 3. This is an extremely common strategy used by tourist guides: “The text receiver is systematically guided from one spot to another in a particular order. Once on a specific spot, s/he is shown the relevant sight and then provided with the pertinent information about the sight” (Virtanen 1992: 191).

282

Rolf Kreyer

the postposed subject is usually contained in the object that is referred to by the preposed constituent. Consider the examples below: (5)

Ahead was a short marble staircase, leading to what appeared to be a lecture-room on the next floor. On the left of the stair was a heavy mahogany door. (ANL: 2607–8)4

(6)

On the next shelf were twelve pieces of matching sculpture, each about twenty centimetres high. (JY2: 606)

(7)

Beyond the desk is a plain maroon carpet. At the edge of the carpet are a filing cabinet and a hatstand and two upright chairs. (J17: 9–10)

(8)

All attention was fixed on a car shaped like a bullet. In it sat a tiny beetle wearing goggles. (FNW: 1704–5)

On the whole, the successful simulation of natural perception through inverted constructions is due to the fact that the order of constituents in inverted constructions follows the sequences given in Table 1 below. Table 1. Relations holding between entities denoted in the preposed and by the postposed constituent in inverted constructions (see Kreyer 2006a: 125) entity denoted in the preposed constituent more permanently located larger closer containing

entity denoted by postposed subject more movable small further away5 is contained

These findings can also be applied to other constructions. Any construction that enables the writer to make a clause-initial reference to a location and to introduce an item to be located towards the end of the clause could be exploited to mirror natural perception, if the respective constituents followed the order specified in Table 1.6 In particular, it is possible to distinguish three broad classes, namely: (I) constructions with preposed locative phrases that involve inversion of subject and main verb; (II) constructions that, apart from the preposed locative, keep the canonical order of constituents; and (III) constructions with a reversed 4. All examples are taken from the British National Corpus (BNC). The three-letter code in front of the colon refers to the BNC-document; the number following the colon identifies the sentences within the document. 5. Note that this configuration is always given with the observer as a fixed point of reference, since nothing can be closer to the observer than the observer him- or herself. 6. The left-hand column would then list properties of the location, whereas the right-hand column would describe the entity to be located.

Syntactic constructions as a means of spatial representation in fictional prose

283

order of arguments and verbal ellipsis, which cannot be assigned to one of the other two classes. On the whole, the six constructions shown in Table 2 can be distinguished: Table 2. Potential constructions for the simulation of natural perception Class

Construction

I

LVS7 LtVS tLVS

II

LSV LSVX8

III

LS

Example Next to the till was a box of chocolate Christmas tree decorations. (HTL: 602) On the right hand bank, there was first the well used tow path. (B3J: 1704) There, in the middle of the white-tiled floor, was a beetle. (FRF: 1813) On the right, a corridor disappeared into the depths of the building. (ANL: 2609) From under her cloak, Maisie took the scroll of paper that he had given her back in August. (HR8: 1002) To his right, a nervous, embarrassed Katherine. (FNT: 1704)

As can be seen from the table above, class I consist of full-verb inversions and two kinds of what is often called there-insertion, namely one with there preceding and one with there following the locative phrase. The second class encompasses tokens where either the location of the subject of the clause is specified or where the entity located is referred to by some other postverbal constituent – usually the object. Class III contains one element only, namely verbless constructions that have an initial locative phrase, followed immediately by a noun phrase. The aim of the present paper is to analyse the use of these six constructions types with regard to their exploitation in the simulation of natural perception. The simulation of natural perception can be regarded as a communicative problem in the sense of Luckmann (1986; see also the discussion by G¨unthner, this volume). Accordingly, the constructions discussed above can be understood as being functionally motivated, i.e. geared towards solving this particular communicative problem. Following Bitzer’s ([1968] 1980) concept of situational rhetoric (see Giltrow, this volume), we can assume that these constructions 7. L refers to locative, V refers to verb, and S refers to subject. Other symbols used in the following are t for there and X for any postverbal constituent. 8. Note that the codes in brackets do not necessarily represent the full structure of the clause. For instance, in the sentence Above it, an old woman in a white headscarf raises her hands [. . . ] (HGN: 3201), the structure of the clause is ASVO, but the entity to be located is the subject of the clause, and the token will therefore be grouped under LSV.

284

Rolf Kreyer

contribute to our conception of the genre narrative: The recurring situational demand of simulating natural perception leads to the recurring activation of functional speech, i.e. the recurring activation of the six constructions under scrutiny. These particular constructions can thus be assumed to be conventionally associated with narrative texts. The present study will explore to what extent this statement holds. The next section describes the material at hand and the way the data to be analyzed was obtained. 2.2. The data The search for constructions of the kind exemplified in Table 2 above is not unproblematic. In obtaining a suitable database, linguists usually turn to one of the two following options: One method is to gather tokens they come across during their everyday reading; this is obviously far from ideal, since it is not clear to what extent their reading preferences may lead to a representative corpus of a particular genre, say, for example, fictional prose. At best, we may hope that a researcher works his or her way through fifty narratives. As a remedy, the researcher may turn to a corpus that contains a representative amount of narrative texts (the BNC, for instance, contains 423 prose-fiction samples) and scan (portions of) these texts for the constructions that he or she seeks. While this approach may solve the problem of representativeness, another problem remains: the unreliability of the human reader. Constructions that are most conspicuous are more likely to gain the attention of the reader and thus become part of the database. Also, the human reader may become primed to particular patterns and thus miss relevant tokens that do not match this pattern. As a consequence, a database gathered in this way is far from objective and it is difficult to find out how representative the collected material actually is. Instead, the data at hand are collected in a different, i.e. a semi-automatic, way, as will be described below. With this option, the data are drawn from the second edition (World Edition) of the British National Corpus (BNC), that is, they are based on a representative sample of the English language. The BNC contains a total of 100 million words from the 1960s onwards (Burnard 1995: 11), 10% of which are taken from spoken and 90% from written texts (Aston and Burnard 1997: 28–29). Within the written component of the BNC, Lee (2001) distinguishes 46 different genres, e.g. fictional prose, school essays, academic writing in specific disciplines such as medicine or social sciences, newspaper articles, and so on. Genre, in Lee’s understanding, denotes “groups [of texts that] are all more or less conventionally recognized as text categories, and are associated with typical configurations of power, ideology and social purposes, which are dynamic/negotiated aspects of situated language use” (Lee 2001: 47). Genres are thus understood to be

Syntactic constructions as a means of spatial representation in fictional prose

285

recognisable by text-external criteria, as opposed to Biber’s (1989) concept of text type, defined by text-internal linguistic criteria. Relevant to the present study are the criteria fiction and the distinction between drama, poetry and prose, leading to three fictional genres. The genre underlying the present study is what Lee has classified as prose fiction, a component of the BNC consisting of 423 text samples totalling 16.7 million words. This amount, of course, is too much to be analyzed as a whole. As a consequence, a sub-corpus was created which included all sentences of prose fiction that ended in a number ending in either 01 or 02, e.g. 101, 401, or 3902. This way, the complete prose-fiction section of the BNC is represented, i.e. all 423 prose-fiction texts, and each document is represented faithfully, i.e. the database is not skewed towards the beginnings, middles, or ends of the documents. On the whole, the sub-corpus contains circa 26,000 sentences and circa 330,000 words. The search for relevant constructions in the corpus is not easy, since the BNC is not parsed (i.e. syntactically annotated).9 In view of these problems the present study applies a method that has already been used successfully in identifying full-verb inversions in the BNC (Kreyer 2006b). Underlying this approach is the fact that although the BNC is not parsed, because it is tagged, and thus provides information about the word class of each word, it is still useful for the automatic retrieval of constructions of the kind exemplified above; consider example (9), which is the tagged version of the example of the LSV construction in Table 2 above: (9)

On the right, a corridor disappeared into the depths of the building. (ANL: 2609)

Many more instances of preposed locative adverbials follow the same pattern of part-of-speech tags: (10)

Beyond the door, Mother Bombie’s feet in soft felt slippers could be heard slapping angrily down the stairs. (AEB: 805)

(11)

In the office, the phone rang. (JY0: 5403)

(12)

On the ground, the mist was less noticeable but it was a dull day and cold for May. (J54: 9)

9. Parsed material is very rare. The British Component of the International Corpus of English, for instance, contains only one million words. The constructions at issue, however, are rather rare and the resulting database would be too small.

286

Rolf Kreyer

All of these examples share a particular part-of-speech sequence: an initial preposition followed by an article, a singular common noun and a comma. In terms of tags, this sequence would be represented by “PRP AT0 NN1 PUN.” In principle, it would thus be possible to identify candidates for preposed locative phrases by letting a computer search for this particular tag sequence. Obviously, a number of relevant tokens would be missed that way, since instances like prepositional phrases containing noun phrases with pre- or postmodification do not fit the pattern: (13)

In the midnight stillness, Rozanov pointed out that we were three men and a woman as in Pushkin’s story about improvisation, Egyptian Nights. (AE0: 804)

(14)

At the foot of her bed, the cat Hastings rose, turned around and around, pushing his paw gently into the eiderdown, and then curling into the soft silk hollow. (AD1: 1104)

(15)

In the Leningrad of 1942, besieged by the Nazis, explosions everywhere, he could imagine himself shrinking from birth. (AE0: 3504)

The answer to this problem is either to try to predict all possible formal realizations of preposed locative phrases and let a computer search for the relevant tag sequences, or to let a computer search for sequences that are not fully specified but leave room for some variation. As has been argued in Kreyer (2006b: 105–110) it is reasonable to follow the latter alternative. The data underlying the present study were extracted from the corpus with the help of over 400 tag sequences of the kind exemplified above. These sequences aimed at identifying preposed locative phrases with initial prepositions, past participle or present participle forms, all of which could optionally be preceded by adverbs. In addition, preposed adverb phrases were included if they allowed for a locational interpretation. Consider the following examples: (16)

To the right a small, bobbing motor-boat was moored to a black post. (JY2: 702)

(17)

Attached to the rope was a four-pronged metal grapnel. (HJD: 901)

(18)

Delicately placed on a slight ridge, left-centre and midway between gatehouse and river, the surviving columns of the forum balanced, lifting the eyes to the exactly right focal point in a sky of scintillating, tearful blue feathered with airy clouds. (H8L: 323)

(19)

Turning off the motorway at the service station he entered a different world. (AC2: 2)

Syntactic constructions as a means of spatial representation in fictional prose

287

(20)

Still leaning against her side of his desk he watched her, waiting for something. (FPF: 74)

(21)

Ahead, a narrow sliver of light showed where the chapel door had not been properly closed. (HH1: 701)

Of course, not all tokens identified this way are relevant for the present discussion. In many cases, for instance, the preposed constituent merely functions as a circumstantial adverbial of space. (22)

On the bus he reflected that his interview with Mrs Wilson had gone well. (ASN: 1102)

(23)

On the drive back to St Petersburg the girl once again dropped off to sleep. (CML: 2102)

It is true that the preposed locative adverbials above also provide information as to the location of entities referred to in the rest of the clause. (22) tells the reader that he is on the bus and (23) tells the reader that the girl is on her drive back to St Petersburg; however, the locations are not focal in these cases and they do not simulate natural perception in the same way as the examples below: (24)

From under her cloak, Maisie took the scroll of paper that he had given her back in August. (HR8: 1002; = example (13))

(25)

Across the gardens the river was wide, flat and moonlit [. . . ]. (G03: 201)

Here, the sentences represent iconically how a particular entity comes into view, and locating this entity seems to be in the focus of attention. Accordingly, cases like (22) and (23) above have been excluded from the database while tokens similar to (24) and (25) were included. On the whole, we are left with a total of 188 instances. Of these, the three most frequent kinds are LSV and LSVX, the non-inverted, non-elliptical clause where the preposed locative phrase specifies the location of the subject and another postverbal constituent, respectively, (35.6% and 25.5%) and LVS, i.e. full-verb inversions (29.8%). Next are there-insertions of the type LtVS, with 13 tokens (6.9%). The verbal-ellipsis construction (LS) is only attested four times (2.1%). There-insertions with there preceding the locative phrase have not been found. The findings on the formal realizations of the observer effect are summarized in Table 3 and Figure 3 below. As Figure 3 shows, narrative texts seem to show a clear predominance for certain constructions in the representation of spatial relations. This is partly

288

Rolf Kreyer

Table 3. Absolute and relative frequencies of construction types

Freq. (abs.) Freq. (rel.)

LVS

LtVS

56 29.8%

13 6.9%

Construction type tLVS LSV 0 0.0%

67 35.6%

LSVX

LS

Total

48 25.5%

4 2.1%

188 99.9%

Figure 3. Relative frequencies of construction types (n=188)

surprising, since all of the constructions could, in principle, be seen as possible solutions to the communicative problem of simulating natural perception and thus could be expected to be associated with the genre of prose fiction to a similar extent. The rest of the paper explores this problem by looking at the five attested constructions in more detail. Our first concern will be with the way in which these constructions are used to represent space through language. 2.3. The verbalization of spatial relations In the discussion above, attention was drawn to two major ways of representing spatial relations, namely route and survey descriptions. With regard to the constructions predominantly used, the first is characterized by an explicitly mentioned observer (encoded in the subject of the clause) who perceives a particular object (usually encoded in the direct object). It is reasonable to assume that this way of presentation will destroy the illusion of immediacy: The reader cannot

Syntactic constructions as a means of spatial representation in fictional prose

289

truly obtain an eyewitness perspective if he or she is made aware of the person who is actually observing the scene. In contrast, inverted constructions, which are often found in survey descriptions, do not put the observer onstage, which makes it easier for the reader to assume this role. As a consequence, the description of spatial relations should make more frequent use of constructions that do not encode an observer, if they aim at an illusion of immediate perception. This is exactly what is shown in the present data. Of the 188 tokens 48 were of the LSVX type. However, in only 29 of these is the subject actually the observer (i.e. 15.4% of 188 instances). In the remaining 19 cases, subject and observer are not identical. Compare the examples below: (26)

Opening his desk drawer, he drew out the current Operation Cuckoo folder, and in a calm, businesslike voice, got straight down to it. (CDA: 802)

(27)

At the end of the note, she had written: Don’t be afraid. (FP7: 3402)

It follows that in 84.6% of all cases, the immediate-observer effect is not threatened through the explicit mention of an observer. Spatial relations are construed subjectively rather than objectively (in the sense of Langacker 1990), and thus presented “in a way similar to immediate experience” (Dorgeloh 1997: 165). Another potential threat to the illusion of immediate perception, in my view, is the use of what above I have called the “photo series”, i.e. that way of description where the observer serves as a fixed point of reference. In contrast to the “camera movement” technique, where the point of reference shifts from one object to another, the photo series makes explicit mention of an observer in the preposed locative. One would therefore assume that of the two strategies, the camera movement is more frequently realized. Also, one would assume that to a large extent the instances of camera movement follow the predictions of a natural order of perception shown in Table 1 above. This order of perception was based on the relative visual prominence of the objects at which the observer is looking. That is, of two objects, the one that is usually perceived first is the one that is closer to the observer, larger, and more permanently located; in addition, if one object contains the other, it is the containing object that is perceived first. Applying these facts to inverted constructions, one would predict that the entity denoted in the preposed constituent is closer to the observer, larger, etc. than the entity denoted by the postposed subject. This prediction was borne out by the data in Kreyer (2006a): Of the 86 instances of inversions that qualified as candidates, 69 (90.8 %) followed one of the four relations listed above. The following analysis will show if similar findings hold for the other four constructions.

290

Rolf Kreyer

Of the 188 cases underlying the present analysis, 12 instances cannot be assigned to one of the two ways of description: In some cases, there is no point of reference given in the preposed constituent and cannot be inferred from the cotext, in other cases it is not clear if the point of reference is the observer or another object. With the remaining 176 cases, though, we find that the photo series is far less frequently used than camera movement. The latter is used more than seven times as often as the former, i.e. 155 tokens as opposed to 21 tokens. This predominance of camera movement holds for all five types involved, as Figure 4 below makes clear.

Figure 4. Photo series and camera movement across construction types (n=176)

The four constructions on the left-hand side of the diagram show a clear preference for the camera-movement strategy. Only the LS type shows equal frequencies for each of the two strategies. However, it would be premature to conclude that the type LS does not follow the general preference for the camera movement – a larger database is needed to determine whether the construction type LS is indeed an exception. The next question concerns the degree to which each of the five constructions follows the rules regarding the order of perception shown in Table 1. Obviously, all instances of the photo series comply automatically since nothing is closer to the observer than the observer him- or herself. The 21 instances of this strategy thus are all instantiations of the sequence closer – further away. More interesting are the 155 tokens that instantiate camera movement, i.e. shifting points of reference. Of these, the vast majority, i.e. 122 tokens (79%),

Syntactic constructions as a means of spatial representation in fictional prose

291

comply with the rules of natural order of perception stated in Table 1. And again, as with the distribution of photo series and camera movement, all constructions seem to be similar with regard to their proportion of compliant cases, with percentages hovering around 80% (cf. Figure 5). Note that the number of LtVS constructions in the data at hand is rather low, namely 12 tokens; a larger number may also show instances that do not comply with the general rule. Similarly, the number of occurrences of the type LS (2) is, of course, too small to allow any claims with regard to this particular construction.

Figure 5. Compliance with natural order of perception across construction types (n=155)

Of the remaining 33 tokens, 8 tokens describe the location of a sound, whereas with another 7 tokens either the location or the item to be located is of a rather abstract kind; consider the examples below: (28)

From the direction of the grandstands came the sound of cheers and applause. (HTJ: 3301)

(29)

Outside was darkness, and as usual the dark came in with him on hair and clothes. (GUM: 401)

292

Rolf Kreyer

In cases like these, it is impossible to make statements regarding relative visual prominence. In the remaining 18 cases, the context does not provide enough information as to which of the two relevant entities is the visually more prominent, as in the following example: (30)

Beneath this mirror stood Miranda’s surprisingly feminine dressing table, skirted in spotted white muslin frills. (FPB: 601)

To sum up the findings on the verbalization of spatial relations, it was found that out of a total of 176 instances, 21 followed the photo series strategy, that is, the location is established with regard to an observer and entities are located with regard to this fixed point of reference. The majority of cases, i.e. 155 tokens, showed camera movement, that is, in these constructions the location of one entity is specified with regard to the location of another entity. It was further found that what holds for the data in general also holds for each construction type individually: Camera movement is the preferred means of encoding spatial relations. Another part of the analysis concerned the extent to which the 155 instances of camera movement mirror a natural order of perception. In this respect, the data showed that with the vast majority of tokens (121 out of 155, or 78%), the locative phrase made reference to the more visually prominent item, while the entity located usually was the less visually prominent one. Again, all of the constructions seem to be similar with regard to the proportion of instantiations that follow this pattern as opposed to those that do not. It is important to note, though, that of the 33 instances that could not be evaluated with regard to relative visual prominence, no token was found that ran counter to the predictions made in Table 1 above.

3. Influences on the choice of construction Above it was found that all of the construction types analyzed in this study are similar with regard to their potential for mirroring natural perception, and thus they help readers to immerse themselves into the narrative. That is, all constructions are equally suited to create the immediate-observer effect, at least as far as the natural order of perception is concerned. The question that arises, of course, concerns other factors that might explain the choice of one construction over the other. This aspect will be explored in the following sections.

Syntactic constructions as a means of spatial representation in fictional prose

293

3.1. The influence of information status Information status has always figured prominently in studies concerning word order phenomena. It is reasonable to assume that the choice of construction depends, to some extent, on the givenness or newness of the constituents of the clause. The present section will look into this aspect in more detail. Concepts like givenness and retrievability are usually of a graded kind and, accordingly, publications in this field of study distinguish three, four, or even more distinct information states (see, for instance, Prince 1981; Birner and Ward 1998; Lambrecht 1994; Chafe 1994). However, a look at real language data shows that it is often extremely difficult to assign information states to constituents in authentic texts. An illustrative example is given in Kreyer (2006b: 142–146). He compares his results on the distribution of information states in inverted constructions with those of Birner (1996) and finds large discrepancies. While in Birner’s material, almost three quarters of the preposed constituents either refer to items that have been mentioned in the previous discourse or are inferable from such items; in Kreyer’s data only 19.8% of all cases are of this kind. This sizeable difference is due to the rather fuzzy concept of inferable. One definition that is often used (also by Birner) is given by Prince (1981: 236), who writes that an item A is inferable from an item B if it is linked to item B by “logical – or, more commonly, plausible – reasoning.” What “plausible reasoning” refers to exactly remains elusive and various authors apply the notion differently. In addition, the concept of information status is extremely difficult to apply to locations. Consider the examples below: (31)

Directly beneath my grandmother’s room was the cellar, approached through a white-painted door in the sitting-room. On the wall of that room was a patch where the barometer had hung – so familiar a face that I had hardly realized it was there. (AC7: 1702-3)

(32)

They were mercenaries, wearing the black and white linen bands of the Free Company, but there were two or three others in full Hearthware armour who seemed to be in charge. Around their waists were black and white sashes, and swords were naked in their hands. (GWF: 3300-1)

The noun phrase the wall of that room could be considered inferable, since it is common knowledge that rooms have walls. However, does that also mean that the location designated by the prepositional phrase on the wall of that room is inferable? This question is difficult to decide. Similarly, we all know that humans have waists and as a consequence could interpret their waists as representing

294

Rolf Kreyer

inferable information. But again, does that mean that the location designated by the prepositional phrase around their waists is related to the previous text by logical or plausible reasoning? In addition to problems in judging the information status of constituents in authentic language samples, there is also evidence that too fine-grained distinctions are irrelevant. In Kreyer (2006b), attention is drawn to the central role of creating a textual world on the basis of a linearized text. This process is facilitated by the reader or hearer being enabled to integrate new information into the already existing discourse model (see also the discussion of the principle of Immediate Textual Integration in Kreyer 2006b: 196–202). With regard to full-verb inversions, Kreyer concludes that “the exact information status of a preposed constituent is not relevant for the integration of the postposed constituent [usually presenting new information]. The relevant variable is whether the preposed item provides a link to the previous discourse or not” (Kreyer 2006b: 149). The central distinction that has to be made, then, is that between linked to the previous discourse and not linked to the previous discourse. The exact nature of the link (i.e. the exact information state) is immaterial as far as the creation of the textual world is concerned: Different kinds of links will lead to different mental processes, but the outcome of these processes, i.e. the integration of new information, is the same. The same accounts for the representation of spatial relations. Here, the focus is on the creation of a visual scene and the location of objects within this scene. This location does not need to be known to the reader beforehand, it is sufficient if the reader can establish a new location in relation to some object within the scene that has already been established. Consider the example below: (33)

a. b. c. d. e.

He looked at the palm tree. Under the palm tree was a bicycle. Under the palm tree there was also a wooden box. Sitting on the leaves were many coloured birds. In front of the house was a car.

The objects in (33b) to (33d) can all be located, irrespective of the exact information status of the preposed constituent: In (33b) the preposed constituent represents a new location that is related to the given palm tree, in (33c) we are dealing with a given location, since it has already been mentioned in the previous sentence, and in (33d) the location is new, but related to an element that is associated with the previously mentioned palm tree, namely the leaves. In all of these cases, it is possible to create the visual scene. This, of course, is not possible with (33e), since the clause does not provide any link to an object

Syntactic constructions as a means of spatial representation in fictional prose

295

that has already been located before. Accordingly, it seems reasonable to draw a distinction between elements that are in some way or other linked to the previous discourse and those that are not; further, more fine-grained, categories seem unnecessary. In the following, we explore the linkedness of location and located entities in the data at hand, focussing on the existence or non-existence of a link to the previous discourse. A look at the data shows that the preposed constituent has a link to the previous discourse in 79.3% of all 188 cases, whereas only 18.6% are not linked in any way (in 2.1% there is not enough context preceding the individual tokens). This also holds across construction types, as we can see in Figure 6.

Figure 6. Linkedness of the locative element across constructions (n=188)

The locative phrases are rarely unlinked across all constructions. In addition, all constructions seem to behave fairly similarly with regard to the linkedness of the initial locative expression: There are no large deviations among the distributions of linked and unlinked locatives among the individual constructions. One point that deserves mention, though, is the comparatively high proportion of unlinked locatives with the construction type LSVX. Here, 27% of all tokens start with a constituent that does not provide a link to the previous discourse. This is most

296

Rolf Kreyer

likely due to the fact that in addition to the location and the item to be located, constructions of this type provide a third slot, the subject, which often is an anaphoric pronoun and thus can fulfil the linking function, as in the examples below: (34)

On the other side of the road at the front of the house a builders’ skip was full to the brim. Ashley walked across. She had looked everywhere else and, although it seemed a long shot, she might as well look in there. Hanging over the edge, she peered down. Buried beneath a fretwork of ancient planks, polythene sacks, and what looked like the treadle off a sewing machine, she saw a scruffy hessian-wrapped parcel. Eureka! (JY9: 1697–1702)

(35)

And she might not like Guy Sterne very much, but he was curiously easy to talk to. In Castries they strolled through the busy streets [. . . ]. In the duty-free shopping centre she noticed a couple of English girls buying jewellery and leather handbags, who ogled Guy with open interest. (JY3: 1700-2)

We can thus conclude that with the vast majority of all cases, the initial locative serves to establish a link to the previous discourse. With constructions of the LSVX type, the preposed locative seems to be less constrained since often the subject of the clause serves as an anaphoric link; as a consequence, the author is freer with regard to the presentation of irretrievable information in the preposed constituent. As regards the entity to be located, interesting observations can be made, as Figure 7 shows. First of all, one would expect the proportion of unlinked constituents among the located entities to be higher if the constructions under scrutiny were mainly used to uphold a progression of linked (given) before unlinked (new) in the clause. In this context, the fact that 43.6 % of the entities to be located are related to entities referred to in the previous discourse is surprising. The constructions analyzed here do not seem to perform an information-packaging function. Also, the diagram shows a slight amount of variation among construction types as regards the information status of the item located. Two classes can be discerned: With the types LSV and LtVS, the majority of the subjects (the items being located) are in some way linked to the previous discourse. In contrast, with LSVX and LVS, there is a slight predominance of items that are completely new to the discourse. A possible explanation for this difference could lie in different textual functions of the individual structures – we will explore this point and related issues in what follows.

Syntactic constructions as a means of spatial representation in fictional prose

297

Figure 7. Linkedness of the entity located across constructions (n=188)

4. A glance at textual functions So far, our analysis has shown that all of the construction types are equally suited to simulate natural perception, since all of the tokens (regardless of the construction type) follow the natural order of perception as summarized in Figure 4 above. In addition, the two strategies in verbalizing spatial relations are used to similar extents by all constructions. That is, there does not seem to be any difference in the use of the constructions regarding the simulation of natural perception. With regard to linkedness, slight differences between the individual constructions could be found. While the preposed locative in all constructions usually has a link to the previous discourse, with more than a quarter of all instances of the LSVX type the preposed constituent is unlinked. This was explained by the fact that the subject in these cases usually is anaphorically related to the previous discourse. Another finding concerns the generally small proportion of new items among the entities to be located and dissimilarities among constructions with regard to the linkedness of these entities. These somewhat mixed results pose the question as to the function of the individual constructions. As far as mirroring natural perception is concerned, all constructions are essentially similar. Some minor differences can be found with regard to linkedness but it does not seem to be the case that the five construction types primarily serve an information packaging function. Maybe the individual construction types serve more global textual functions, such as text structuring

298

Rolf Kreyer

or topic managing (see Kreyer 2006b for a detailed discussion of this point with regard to inversions). A good starting point in this respect is to look at those parts of the text that follow the constructions at issue. In particular, does the located entity recur in those parts of the text that follow the construction? Note that this does not yet make any claims as to the specific function of the located entity. It may merely serve as a new point of reference for the location of another object (and thus essentially fulfil the immediate-observer effect function) or it may function as the topic for a number of following sentences (and thus serve text-structuring purposes). Consider example (36) and (37), respectively: (36)

Beyond that lay the Germans’ barbed wire, and beyond that still the Germans, waiting for them in their trenches. (K8T: 902)

(37)

Lying on the ground near the car Pete spotted something round and shiny. It was the driving wheel of the car. Someone had ripped it out and thrown it away. “Can I have it, Mum?” Pete asked and Mum grinned and said “Well, I suppose so but I’ll give it a good wash when we get home.” Pete ran ahead and opened the big glass doors as he usually did and Mum lifted him up to press the button for the lift. He couldn’t wait to play with his car wheel. (HSA: 202-7)

Leaving these functional differences aside for the moment, the data show that in 129 out of 188 cases, i.e. 68.6%, the located entity recurs in the text that follows the construction under analysis. Similar proportions also show across the individual constructions. Figure 8 makes clear that, similarly to the variables analyzed so far, differences between individual constructions are of a gradual kind. The proportion of located items that do not recur in the following discourse generally is rather low for all construction types, ranging from 20.8% for LSVX to 37.5% for the type LVS. An exception might be the LtVS construction: Here almost half of all located entities are not part of the following discourse, but further cases would have to be analyzed to rule out random effects, since a sample of 13 tokens is too small to make any definite claims. One striking observation, however, concerns the two types of LSVX and LVS; recall that both constructions were highly similar with regard to the information status of the located entity. With the majority of tokens (i.e. 58.3% and 62.5%, respectively), the located entity was new to the discourse, and it was suggested that these two constructions in particular are exploited to introduce new entities into the discourse world. Figure 8, however, suggests that this is a function more frequently fulfilled by constructions of the LSVX type. This dis-

Syntactic constructions as a means of spatial representation in fictional prose

299

Figure 8. Occurrence of located entity in the following discourse (n=188)

crepancy hints at interesting interactions between information status and textual function in the constructions under scrutiny. So far, however, the main insight from this first analysis is that in most of the cases the located entity recurs in the following discourse and that, by and large, there are no big differences between the individual constructions. Further analyses would have to take a closer look at the more specific functions that the located entity fulfils in the previous discourse. Then it might, for instance, be shown that with some constructions the located entity is primarily used as a point of reference for the location of another entity, while with other constructions the located entity is also introduced as the new discourse topic for a stretch of text.

5. Conclusions and outlook The present paper has explored the use of six construction types (LVS, LtVS, tLVS, LSV, LSVX, and LS) in the verbalization of spatial relations in narrative discourse. In a corpus of approximately 26,000 sentences, 188 instances of syntactic constructions in immediate-observer effect function have been identified. It was shown that there are significant differences with regard to the use of individual constructions. While some were extremely rare or did not even occur in the corpus, others accounted for one third of the data. From this we can conclude that only some of these constructions contribute to our conception

300

Rolf Kreyer

of the genre prose fiction. It also shows that there are different constraints on the use of the individual options. What these constraints are, however, so far remains elusive. On the other hand, with regard to the linguistic representation of spatial relations, all constructions behave surprisingly similarly, and, more importantly, seem to be equally useful for solving the communicative problem of simulating natural perception and thus creating “a vivid representation of a perceptually inaccessible reference situation” (Drubig 1988: 88); all constructions are “experientially iconic” (Enkvist 1981). Slight differences have been shown among the construction types concerning the distribution of information states within the clauses and concerning the recurrence of the located entity in the following discourse. These findings suggest a more detailed analysis of the interaction of linkedness and textual functions, such as topic-managing or text segmentation (see Virtanen 2004), fulfilled by the constructions at issue. Related to textual functions is the question as to the use of those constructions that construe the discourse world in an objective way, i.e. through the explicit mention of an observer in the clause. As has been argued above, such uses run counter to the immediate-observer effect; it becomes more difficult for the reader to assume an “eyewitness perspective” (Dorgeloh 1995: 228). What is the function of such instances and how does this relate to the immediate-observer effect? In this context, the role of the verb needs to be taken into consideration, too. It might be that the writer’s primary focus is on using a particular verb (e.g. see, found, discover, etc.), which then automatically leads to an objective construal of the discourse world. Finally, further research will have to look into the effects that are created by clusters of syntactic constructions in an immediate-observer effect function. A useful starting point in this respect may be Virtanen’s (1992: 85) concept of “textstrategic continuity,” defined as “a thematic or topical uniform text-structuring orientation chosen to attain, in view of the communicative goal, a maximally profitable text organization, for the benefit of the text receiver” (see Hasselgard 2010: chapter 4 for a recent discussion). Clusters of syntactic constructions in an immediate-observer effect function can be understood as enhancing textstrategic continuity. The question that arises is whether individual constructions have specific functions in this respect. Do some constructions coincide with the beginning of text-strategically continuous portions of a text, while others signal the ending of such portions or mark important turning points? In summary, while the present paper has been able to provide some insight into the exploitation of different syntactic constructions for the verbalization of spatial relations, future research will be needed to explore further the role of information status and the function of these syntactic constructions in the organization of texts.

Syntactic constructions as a means of spatial representation in fictional prose

301

References Aston, Guy and Lou Burnard 1997 The BNC Handbook: Exploring the British National Corpus with SARA. (Edinburgh Textbooks in Empirical Linguistics.) Edinburgh: Edinburgh University Press. Biber, Douglas 1989 A typology of English texts. Linguistics 27: 3–43. Birner, Betty J. 1996 The Discourse Function of Inversion in English. (Outstanding Dissertations in Linguistics.) New York/London: Garland. Birner, Betty J. and Gregory L. Ward 1998 Information Status and Noncanonical Word Order in English. (Studies in Language Companion Series 40.) Amsterdam/Philadelphia: John Benjamins. Bitzer, Lloyd F. [1968] 1980 Functional communication a situational perspective. In: Eugene E. White (ed.), Rhetoric in Transition Studies in the Nature and Uses of Rhetoric, 21–38. University Park, PA: Pennsylvania State University Press. Burnard, Lou (ed.) 1995 Users Reference Guide British National Corpus Version 1.0. Oxford: Oxford University Computing Services. Chafe, Wallace L. 1994 Discourse, Consciousness,and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: Chicago University Press. Dorgeloh, Heidrun 1995 Viewpoint and the organisation of informative discourse: On the discourse function of full inversion in English. In: Brita Warvik, SannaKaisa Tanskanen and Risto Hiltunen (eds.), Organization in Discourse. Proceedings from the Turku Conference, 223–230. (Anglicana Turkuensia 14.) Turku: University of Turku. Dorgeloh, Heidrun 1997 Inversion in English: Form and Function. (Studies in Discourse and Grammar 6.) Amsterdam/Philadelphia: John Benjamins. Dorgeloh, Heidrun 2006 Inversion in descriptive and narrative discourse: A text-typological account following functional principles. Cahiers de Recherche 9: 101–114. Drubig, Hans B. 1988 On the discourse function of subject verb inversion. In: Josef Klegraf and Dietrich Nehls (eds.), Essays on the English Language and Applied Linguistics on the Occasion of Gerhard Nickel’s 60th Birthday, 83–95. (Studies in Descriptive Linguistics 18.) Heidelberg: Groos.

302

Rolf Kreyer

Ehrich, Veronika and Charlotte Koster 1983 Discourse organization and sentence form: The structure of room descriptions in Dutch. Discourse Processes 6: 169–195. Enkvist, Nils Erik 1981 Experiential iconism in text strategy. Text 1: 97–111. Hasselgard, Hilde 2010 AdjunctAdverbials in English. (Studies in English 39.) Cambridge: Cambridge University Press. Kreyer, Rolf 2006a ‘Observer effect,’ ‘eyewitness perspective,’ and ‘imaginary guided tour’. What is so natural about the way inversions represent spatial relations?. Cahiers de Recherche 9: 115–132. Kreyer, Rolf 2006b Inversion in Modern Written English: Syntactic Complexity, Information Status and the Creative Writer. (Language in Performance 32.) T¨ubingen: Gunter Narr. Kreyer, Rolf 2007 Inversion in modern written English: Syntactic complexity, information status and the creative writer. In: Roberta Facchinetti (ed.), Corpus Linguistics Twenty-five Years on: Selected Papers of the Twenty-fifth International Conference on English Language Research on Computerised Corpora, 187–203. (Language and Computers 62.) Amsterdam: Rodopi. Giltrow, Janet this volume Genre as difference: The sociality of linguistic variation. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 29– 51. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. G¨unthner, Susanne this volume Grammatical constructions and communicative genres. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 195–217. (Topics in English Linguistics 70.) Berlin/NewYork: Mouton de Gruyter. Lambrecht, Knud 1994 Information Structure and Sentence Form: Topic, Focus, and the Mental Representations of Discourse Referents. (Cambridge Studies in Linguistics.) Cambridge: Cambridge University Press. Langacker, Ronald W. 1990 Subjectification. Cognitive Linguistics 1: 5–38. Lee, David Y.W. 2001 Genres, registers, text types, domains, and styles: Clarifying the concepts and navigating a path through the BNC jungle. Language Learning and Technology 5: 37–72.

Syntactic constructions as a means of spatial representation in fictional prose

303

Linde, Charlotte 1974 The linguistic encoding of spatial information. Ph.D. Dissertation, Columbia University. Linde, Charlotte and William Labov 1975 Spatial networks as a site for the study of language and thought. Language 51: 924–939. Luckmann, Thomas 1986 Grundformen der gesellschaftlichen Vermittlung des Wissens: Kommunikative Gattungen. K¨olner Zeitschrift f¨ur Soziologie und Sozialpsychologie, Sonderheft 27: 191–211. Prince, Ellen F. 1981 Toward a taxonomy of given-new information. In: Peter Cole (ed.), Radical Pragmatics, 223–255. New York: Academic Press. Talmy, Leonard 1983 How language structures space. In: Herbert L. Pick and Linda P.Acredolo (eds.), Spatial Orientation Theory, Research and Application, 225–282. New York: Plenum Press. Tversky, Barbara 1996 Spatial perspective in descriptions. In: Paul Bloom, Mary A. Peterson, Lynn Nadel and Merrill F. Garrett (eds.), Language and Space, 463– 492. (Language, Speech, and Communication.) Cambridge, M.A.: MIT Press. Ullmer-Ehrich, Veronika 1982 The Structure of living space descriptions. In: Robert J. Jarvella and Wolfgang Klein (eds.), Speech, Place, and Action, 219–249. Chichester: John Wiley and Sons. Virtanen, Tuija 1992 Discourse Functions of Adverbial Placement in English: Clause-Initial Adverbials of Time and Space in Narratives and Procedural Place Descriptions. Abo: Abo Akademis Foerlag. Virtanen, Tuija 2004 Point of departure. Cognitive aspects of sentence-initial adverbials. In: Tuija Virtanen (ed.), Approaches to Cognition through Text and Discourse, 79–97. (Trends in Linguistics. Studies and Monographs 147.) Berlin/New York: Mouton de Gruyter. Wenz, Karin 1996 Iconicity in verbal description of space. In: Martin P¨utz and Ren´e Dirven (eds.),The Construal of Space in Language andThought, 269–286. (Cognitive Linguistics Research 8.) Berlin/New York: Mouton de Gruyter.

Agreement in educated Jamaican English: A corpus-based study of spoken usage in ICE-Jamaica1 Susanne Jantos This paper investigates non-standard subject-verb agreement in Jamaican English through the analysis of variation across three text types which differ in their relative level of formality. A sample of 40,000 words from each of the text categories – direct conversations, class lessons, and broadcast news – taken from the spoken section of ICE-Jamaica, serves as the data for the study. The data reveal that there is a substantial amount of variation between the text categories. However, the correlation is not one of unidirectional decrease of frequencies with increasing formality; instead the pattern found is rather complex. In fact, the data provide evidence of four trends. First, overall rates of grammatical non-agreement decrease with increasing formality. Second, register correlates differently with various types of non-agreement. Third, there is no evidence of a general tendency to simplify inflection on verbs, and fourth, creole influence is strongest in the most informal text category and literally absent in higher formality texts. Register can thus be described as having a strong influence on grammatical non-agreement in Jamaican English.

1. Introduction The standard view on subject-verb agreement in English is that verbs normally agree in person and number2 with their respective subjects. Agreement is defined as “the relationship between two grammatical units such that one of them displays a particular feature (e.g. plurality) that accords with a displayed (or semantically implicit) feature in the other” (Quirk et al. 1985: 755). To give an 1. I wish to thank Christian Mair, Dagmar Deuber, and Ingrid Rosenfelder for their input and reactions on an earlier version of this paper and Katie Gilbert for her helpful comments and corrections. 2. Hudson (1999) rejects this view and argues that verbs only have agreement features in present tense and that they agree with their subject merely in number but not in person. Only the exceptional verb be agrees in number and person with its subject and shows agreement features in past and present tense alike.

306

Susanne Jantos

example of agreement, in (1) both subject and verb are plural, while in (2) the subject is singular, but the verb lacks singular marking. Example (2) therefore represents a case of non-agreement. (1)

Uhm they teach you the science behind electronics (S1A-046)3

(2)

A student represent all the committees on the boards (S1A-059)

However, non-standard subject-verb agreement has been characterised as a widespread phenomenon in varieties of English. The underlying factors determining whether subject and verb agree in number may be different for different varieties of English. These factors include, but are not limited to: 1) general tendencies toward simplification and regularization found in native varieties (see Trudgill 1990 and Hudson 1999 for dialects of British English); 2) partial loss of inflectional morphology due to language contact (see Schneider 2007) and the levelling of irregularities, which has been attributed to “L2-acquisition difficulty” (Szmrecsanyi and Kortmann 2009: 69); 3) the influence from substrate languages (Platt, Weber, and Ho 1984); and 4) the focus on function rather than form in English as a lingua franca (ELF) communication. However, apart from the type of variety, there also may be social and situational conditions, such as age or gender, or the level of formality and characteristics of the register, which may favour or disfavour standard agreement (see Crawford 2005 for American English). The latter two will be the focus of the present analysis of subject-verb agreement in three spoken registers of educated Jamaican English, all of which differ in their relative degrees of formality. Although “specific grammatical constructions are closely tied to particular communicative genres” (G¨unthner, this volume: 196), the analysis in this paper will be based on differences in register rather than genre, because “the study of syntactic variation according to genre requires computerized databases that easily lend themselves to grouping texts by genres” (Mondorf, this volume: 221), and the ICE-corpora unfortunately do not allow such grouping by genres. Instead, the text categories from ICE-Jamaica that are used here are already classified as text types according to formal criteria such as level of formality, public versus private, monologue versus dialogue, and so on. Each such text type in ICE-Jamaica is heterogeneous in that it consists of a wide range of spoken genres, which, however, are not separately analysed (cf. section 2.1). The paper is structured as follows. Section two presents a short overview of the language situation in Jamaica and previous research on subject-verb agree3. The code in brackets refers to the corpus text in ICE-Jamaica (that is, the Jamaican component of the International Corpus of English) from which the example was taken; in this case text 046 of the spoken category S1A, direct conversations.

Agreement in spoken educated Jamaican English

307

ment in native and non-native varieties of English. This is followed by a description of the data and methodology used for this study, including a brief discussion of the registers investigated, as well as an examination of subject-verb agreement with lexical verbs with regard to variation along formality and register lines. The type of non-agreement is the second main feature to be considered, with notional considerations, the proximity principle, and lack of inflection as prominent factors. After that, I turn to the exceptional verb be. Here I analyse not only the effects of register and formality but also the types of non-agreement and factors such as subject number and tense, and these factors’ respective potentials for influencing or determining subject-verb agreement with be. Concluding remarks follow a discussion of the results.

2. Theoretical background Jamaican English is a so-called New English (Platt, Weber, and Ho 1984) with a particularly interesting language development. The variety developed under British dominance during colonial times, but in the post-colonial era it has been subject to growing influence from American English due to America’s closer geographical proximity. Apart from these competing influences, the language situation in Jamaica is further complicated by the presence of a substrate language, Jamaican Creole (JC or Patois). Jamaican English and Jamaican Creole are not to be seen as two strictly separate linguistic systems but rather as two varieties located at opposite ends of a creole continuum. This continuum has gradual transitions, ranging from an acrolect at one pole, which is the variety closest to International Standard English, to a basilect at the other pole, with a number of mesolectal varieties in between (see Sand 1999: 50). In the past few decades, most research on the creole continuum characterising the Jamaican language situation has concentrated on the basilectal, i.e. the creole end of the continuum (cf. Patrick 2004), while only a few studies have focused on its acrolectal end, the underlying assumption being that the latter variety was identical to standard British English, the language traditionally regarded as the norm in Jamaica due to its association with the British colonial administration. However, a number of sociolinguists studying the acrolect have shown a tendency for Jamaican English to move away from the postulated British norm (see Sand 1999: 13–14) and concluded that the emergence of a new standard, Jamaican Standard English, must be taking place (cf. Mair 2002). Since it is in spoken usage where language change usually takes place first, this paper focuses on the spoken section of the ICE-Jamaica.

308

Susanne Jantos

Mair (2002: 36) postulates that “[p]atois is clearly the dominant shaping influence on spoken English in Jamaica,” not only as regards phonological features, but also in terms of syntax and lexis. He does not, however, go into detail about “the complicated interaction of the creole and standard English norms on the level of textual macro-structure” (Mair 2002: 37). In fact, there have been few comparative studies dealing with variation across registers in Jamaican English. Sand (1999), for instance, focussed on the mode of production in her analysis of radio and press texts, while Mair (1992, 2002) analysed one particular written register at a time, namely newspaper usage (1992) and student essays (2002). Systematic investigations of register characteristics or effects of situational factors, such as formality, are still rare. For this reason, the present paper will systematically analyse three spoken registers which differ in their relative levels of formality (for an investigation of spoken genres with regard to grammatical constructions see G¨unthner, this volume). In her analysis of 40 direct conversations, Deuber (2009) found that nonagreement of regular verbs with their subjects is relatively frequent (around 20%). She did not find evidence that the proximity principle or the semantic notion of the noun were strong explanatory factors in her data. Rather, “simple unmarking of third person singular verbs is clearly the predominant phenomenon” (Deuber 2009: 22) due to influence from the creole substrate, where third-person singular s-marking on verbs does not exist (cf. also Patrick 1999). Sand (1999) found evidence for creole influence in the use of quantifiers such as much, which in Jamaican English often occurs with nouns that function as count nouns in International Standard English, and she attributes this to the fact that much can be used irrespective of count-noun versus mass-noun distinctions in Jamaican Creole (see Bailey 1966: 30). However, Sand (1999) suggests that reclassifications of count and mass nouns are not restricted to the quantifier, but that there is a general, though slight, tendency in Jamaican English to extend notional concord, and this assumption is supported by Deuber (2009) in that there are frequent examples of everybody or phrases with every in her data, a word and phrases which can be interpreted as notionally plural, as they are followed by plural verbs. Proximity agreement seems to be an extremely infrequent phenomenon in private conversations (see Deuber 2009), but was found rather frequently in Sand’s (1999) spoken and written media texts. Sand did not find any influence of the mode of production on proximity agreement, as the percentage was similar in Jampress (written) and Jamradio (spoken), with 10.1% and 10.8% respectively. This result is in stark contrast with Quirk et al.’s observation that “Proximity concord occurs mainly in unplanned discourse. In writing it will be corrected to grammatical concord if noticed” (Quirk et al. 1985: 757). Deuber’s

Agreement in spoken educated Jamaican English

309

and Sand’s results lead to the conclusion that formality might exert a strong influence on this type of agreement. Mair relativizes this hypothesis, since in his investigation of newspapers “uncorrected instances of the proximity principle in action are the most frequent grammatical peculiarity observed” (Mair 1992: 82), but his analyses of student essays, which have a lower average degree of formality, yield similar amounts of proximity agreement (Mair 2002: 52). The possibility of formality being a strong factor influencing proximity agreement will be addressed later. With regard to the verb be, studies so far have focussed on native varieties of English, such as New Zealand English (Hay and Schreier 2004), American English (Crawford 2005), British English (Tagliamonte 1998), and Canadian English (Tagliamonte and Smith 2000), and very rarely on New Englishes. In the native varieties mentioned above, a number of linguistic features have been identified which influence agreement with be. These are, among others, tense, subject type, polarity, subject number, and contractedness. Numerous studies have consistently found the present tense paradigm to be less prone to agreement regularization than the past tense of be (cf. Hay and Schreier 2004). In contrast to this, non-agreement with be seems to be literally nonexistent in the past tense and rather infrequent in the present tense in Deuber’s 40 direct conversations in Jamaican English. She found eight cases of the latter, of which one each can be explained by proximity and notional concord (Deuber 2009), and attributes these instances of present tense be nonagreement to the fact that mesolectal creole has invariant is as a copular form for all persons (Deuber 2009). The subject type is important insofar as existential there + be constructions (ETBs) have been found to show higher rates of non-agreement in past tense contexts with third-person plural subjects than non-existentials, and personal pronoun subjects have the lowest rates (see Hay and Schreier 2004: 211). Some scholars found an effect of polarity, in that was is frequently extended to positive contexts, while weren’t is often found in negative contexts, irrespectively of subject number and person (Britain 2002). Schilling-Estes and Wolfram (1994: 280) therefore assumed a “remorphologization of the was and were allomorphs of past be along positive/negative, rather than person-number, lines.” However, in Hay and Schreier’s (2004) data, there was no evidence of polarity playing any role in determining agreement with be. Meecham and Foley (1994) found subject number and contractedness to be particularly important factors determining (non-)agreement with be in Canadian English and showed that the rate of plural subjects with singular verbs is much higher than that of singular subjects with plural verbs, 72% versus 1%, respectively. The authors identify the contracted singular form of be, i.e. there’s,

310

Susanne Jantos

as the most significant indicator of non-agreement, and Crawford (2005) even labels this the default form for there + be constructions. The social factors investigated so far include gender, age, social class (Hay and Schreier 2004), and the level of education (Cheshire 1999). Cheshire assumes a close interrelation of the social and situational variables educational level and formality in influencing agreement, and expects “agreement in existential there constructions to occur variably: more frequently in formal speech styles than in informal styles, and more frequently in the speech of educated speakers” (Cheshire 1999: 138). In other words, non-agreement would be expected in informal language usage by speakers of lower educational levels. These hypotheses were not confirmed by Crawford (2005) in his investigation of conversations, lectures, textbooks, fiction texts, and chat samples. He found existential there + be constructions to be more frequent in academic lectures than in private conversations and explains this, on the one hand, by referring to the higher cognitive load in lectures and, on the other hand, by the different discourse functions of existential there constructions in the text types concerned (cf. Mondorf, this volume). Since, apart from Crawford’s work, “no study to date has actually compared the extent to which variation is found in a number of different situations of language use” (Crawford 2005: 39), the present paper will focus on register variation along the lines of Biber et al. (1999) and on differences in the level of formality between the registers in order to investigate to what extent these factors influence subject-verb agreement in Jamaican English.

3. Data and methods of analysis 3.1. Data The data used for this study are taken from the Jamaican component of the International Corpus of English, which comprises 300 spoken and 200 written texts of around 2000 words each. The ICE-Jamaica corpus does not allow investigations into the influence of social factors, such as age or educational level, on subject-verb agreement, because age information is not given for all speakers and all speakers included in the corpus have at least a secondary school education. For this reason, social factors will not be taken into account here. For the present study, 20 texts each were taken from the text categories of direct conversations, class lessons, and broadcast news of the spoken section of ICE-Jamaica to represent three macro-registers differing in their respective levels of formality. Direct conversations are the most informal text type used, class lessons are considered to be of medium formality, and broadcast news

Agreement in spoken educated Jamaican English

311

represent the highest formality level. However, these formality levels are not to be taken as absolute. They are relative in comparison to the other text types investigated. While class lessons are relatively high in formality when compared to direct conversations, their degree of formality is lower than that of broadcast news. The registers chosen differ not only in terms of formality but also with regard to other factors such as the size of the audience, the participant constellation, i.e. monologue or dialogue, and the relation of the speakers and audience toward one another, as well as the sphere of communication in which the interactions take place, in private or public. However, most of the registers are not homogeneous and comprise various sub-registers, and the texts within each register may differ considerably in the above-mentioned aspects and include a wide range of different speech genres. The text category of direct conversations consists of private interactions, which usually involve two or more participants who know each other well and are roughly equal in terms of social status and the proportion of utterances. Normally there is no large audience present and topics often focus on private matters (see G¨unthner, this volume, on everyday narratives). On the other hand, there is a lot of intertextual variation in this category, as there are also conversations of a more interview-like nature, where the participants are not acquainted or where one participant might be dominating the conversation by choosing the topics. The utterances in this category are produced on the spot, as there is no time to plan grammatical structure or choose words well ahead of time (cf. Biber et al. 1999: 8). For this reason, direct conversations are characterised by high numbers of hesitations, repetitions, or repair sequences. Participants frequently interrupt one another or speak simultaneously and sentences4 are often left incomplete and tend to have few embedded sub-clauses or complex syntactic structures. The class of broadcast news typically consists of monologues in the public sphere where one speaker communicates something to a large unknown audience which is not present in the immediate communicative surrounding. However, these monologues are sometimes interrupted by reports by other speakers, whether they be correspondents or interview partners, but only in very rare cases do these interruptions lead to a more interactive communicative style. Since broadcast news are pre-formulated and written for an audience, there are few signs of online production, such as repetitions, hesitations, or false starts, 4. Because text units often remain incomplete in spoken conversation, it is questionable as to whether or not they can actually be called sentences. For instance, there are numerous sentences with zero subjects or missing auxiliary verbs in the conversations of ICE-Jamaica. However, the term “sentence” here denotes text units in spoken language, which are typically signalled by means of intonation and often preceded/followed by a pause.

312

Susanne Jantos

and the grammatical information tends to be encoded more synthetically than in most other text types (see Szmrecsanyi and Kortmann 2009). The text category of class lessons comprises texts that range somewhere in between the interactive direct conversations and the news monologues. Some of the class lessons belong to the sub-register of lectures and can be described as monologues, in which the teacher is the only active participant and the students assume the more passive role of listeners. Apart from occasional questions or interruptive comments, they do not contribute verbally to the communication process. On the other hand, there are also a number of class lessons where the teacher and students together discuss a certain topic very actively. The students here contribute considerably to the utterances, so these examples can be described as highly interactive sequences, sequences in which the proportions of utterances are almost equally distributed between the teacher and the students. In terms of production characteristics, too, class lessons range between direct conversations and broadcast news. In the sub-register of lectures, where the speaker is not in danger of losing the floor if s/he needs time for thinking and formulating utterances, there are comparably few signs of online production. On the other hand, the highly interactive and often fast-paced discussion sequences are characterized by frequent overlaps or interruptions, repetitions, false starts, and repair sequences, since here, speakers cannot plan ahead but have to compete to take or keep the floor. The following analysis will concentrate on agreement patterns in the macro-registers. 3.2. Methods of analysis For the analysis of lexical verbs, I manually counted all verbs with third-person singular or plural subjects in present tense except modal auxiliaries and be. I excluded from my counts all sentences with null subjects, as in (3), or deleted copular or auxiliary verbs, as in (4), as well as subjunctive forms, questions, and tag questions. (3)

Seems like she’s funny (S1A-056)

(4)

And a lot of times because they’re sitting up there they not see what’s happening on the ground in terms of the students (S1A-059)

Unclear cases do not appear in any of the analyses either. The four most frequent reasons for exclusion are: because the subject is unclear, as in (5); because the tense is unclear, as when there is neither marking for past tense nor number, as in (6); because the sentence is not clearly interpretable or shows clusters of creole features, as in (7); and because there is non-agreement within the subject noun

Agreement in spoken educated Jamaican English

313

phrase as when the head of the subject noun phrase is not marked for number but the plural is indicated by other means, such as quantifying expressions, as in (8): (5)

This in an effort to improve cooperation and communication in areas which affect the whole of Ireland (S2B-007)

(6)

Something drop (S1A-050)

(7)

. . . they start cuss and how they spent ten thousand dollars on campaigning and them never win is for me the why me win (S1A-060)

(8)

Oh man alright so uhm that’s uh why I oh yeah yeah get in trouble and all sort of thing happen (S1A-050)

Contributions by extra-corpus speakers, quotations, and creole passages, but also instances of unclear speech, self-corrections or repetitions, were excluded from the analyses. The few cases where the word following the verb begins with a sibilant had to be disregarded as well because the pronunciation of these sentences made it impossible to identify whether or not the verb was in fact marked for number with third person singular –s. For the analysis of be I looked at present and past tense verbs following subjects in all persons and numbers. Non-agreement was almost nonexistent in all persons except third. There was one single instance of is with a second-person singular subject. This is surprising, since, as Deuber (2009) notes, is is used in the creole as an invariant copular form with all persons, hence a considerable number of such instances were expected, especially in direct conversations. Since non-agreement occurred almost exclusively with third-person subjects, I will restrict my analysis to these cases. Excluded from my counts were the same cases as above. 4. Results and discussion 4.1. Verbs other than be Table 1 shows the distribution of the 2055 verbs analysed and the rates of (non-)agreement and across the three text types (direct conversations, class lessons, and broadcast news), which are here represented by the low, medium, and high formality levels, respectively. In the category of direct conversations, there are 713 verbs (34.7 % of all lexical verbs), in the class lessons, there are 638 verbs (31.0%), and in the broadcast news, there are 704 verbs (34.3%). The lexical verbs are thus distributed fairly evenly across the three categories.

314

Susanne Jantos

Table 1. Distribution of verbs in each register Grammatical Agree Tokens Formality

Non-agree Percent

Tokens

Total

Percent

Tokens

Low

58

82.5

12

17.5

71

Medium

58

91.8

5

8.2

63

High

67

96.4

2

3.6

70

Total

185

90.2

20

9.8

205

The overall rate in all 60 spoken texts is 9.8% for non-agreement5 and 90.2% for agreement. That means that non-agreement is not a highly frequent phenomenon. However, looking separately at each level of formality, some systematic differences between the categories come to the fore in that non-agreement rates decrease steadily with increasing levels of formality, namely from 17.5% in low formality, to 8.2% in medium formality, and then to 3.6% in high formality. Table 2 shows the percentages of grammatical agreement versus non-agreement by subject number for the three registers. Here agreement rates are displayed separately for verbs with singular as opposed to plural subjects and a very clear agreement pattern emerges. Table 2. Grammatical agreement versus non-agreement by subject number Grammatical Agreement Agree Tokens Formality

Low

Medium

High

Subject Number

Subject Number

Subject Number

Non-agree Percent

Tokens

Percent

Singular

223

65.4%

118

Plural

365

98.9%

4

1.1%

Total

588

82.8%

122

17.2%

Singular

354

87.6%

50

12.4%

Plural

232

99.1%

2

.9%

Total

586

91.8%

52

8.2%

Singular

460

96.2%

18

3.8%

Plural

219

96.9%

7

3.1%

Total

679

96.4%

25

3.6%

34.6%

5. The point of departure for the categorization of the verbs and their inflection is grammatical agreement. Agreement marking that does not correspond to the grammatical number of the noun but is allocated according to the semantic notion of the subject (notional agreement) is marked as non-agreement, a marking, however, which is not to be taken as an evaluation of the correctness of the form. It merely expresses that there is no grammatical agreement. The exceptional noun police is not included in the counts and will be discussed separately below.

Agreement in spoken educated Jamaican English

315

In all levels of formality, non-agreement rates are much higher when the subject is singular, as in (9), than when the subject is plural, as in (10). (9)

And we find it help you know when the spouses are there (S1A-054)

(10)

. . . and you have people out there who wants to see these things. . . (S1A051)

In direct conversations more than one-third of all verbs with singular subjects show a lack of –s marking (34.6%). This rate decreases, with increasing formality, to 12.4% in class lessons and then to 3.8% in broadcast news. On the other hand, the rate of non-agreement with plural subjects, which is around 1% in low and medium formality texts, increases in the most formal category, where 3.1% of the verbs following plural-subject noun phrases show hypercorrect third-person singular –s marking. This distribution suggests that the lack of –s marking is stigmatized, as it decreases with increasing formality, while the third-person singular inflection is a prestigious form, which increases with increasing formality. In all text types, non-agreement is much more likely to be characterized by a lack of marking rather than by hypercorrect marking. In order to find out whether non-agreement is caused by the same factors in all text types we have to look at what might account for hypercorrect marking of verbs with plural subjects and what might account for the frequent lack of marking with singular subjects. In the broadcast news data, one example of hypercorrect marking may be attributed to the proximity principle, according to which the verb sometimes does not agree with the syntactic head but with a nearer noun in a complex noun phrase: (11)

The moves to replace Mr. Sangster follows three successive losses to Mr Buchanan in general elections (S2B-009)

In (11), the syntactic head is the moves, which is plural, but the verb has thirdperson singular –s marking due to the more proximal singular noun Mr. Sangster. None of the hypercorrect markings in the other categories can be characterized as proximity agreement. Three of the verb forms in the broadcast news may have been marked for singular due to notional agreement, as in (12), whereas only one instance in the class lessons category and none in direct conversations category can be so explained. (12)

The ten includes three unidentified men two security guards and a gunman (S2B-011)

316

Susanne Jantos

In (12), the head of the subject noun phrase is interpreted as “the total of ten” and therefore causes singular agreement. The other instances of hypercorrect marking cannot be attributed to such alternative explanations and can only be explained due to third person –s being the more prestigious form. This prestige of –s marking is also reflected in the fact that the noun police, which takes a plural verb in an overwhelming majority of cases in Jamaican English, is occasionally followed by a verb marked for singular in the class lessons and broadcast news (13). (13)

The Red Hills Police says it has received no formal report. . . (S2B002)

As concerns the lack of marking on verbs following singular subjects, the strongest factors seem to be strikingly different in each of the categories, as is shown in Figure 1, where notion means “notional agreement”, environment means “creole context”, and don’t simpl. (that is don’t-simplification) means the simplified, invariant use of the contracted don’t as the negative form with all persons and numbers in the present tense (see Trudgill 1990 for a discussion of don’t-simplification in British dialects).

Figure 1. Types of non-agreement in each text category

For direct conversations, of the 118 verbs that do not grammatically agree in number with their subjects, only 1 case could be described as proximity agreement, see (14):

Agreement in spoken educated Jamaican English

(14)

317

There’s one oh one of my lecturers come 6 from I think not Linstead Sligoville (S1A-050)

Here, it is not the syntactic head of the NP one, but the closer noun lecturers that functions as the controller of agreement. This ties in well with Deuber (2009), who finds that proximity agreement is extremely rare in direct conversations. Her data yield only one token in a sample of about 80,000 words. In class lessons and broadcast news, two lexical verbs each are not marked for thirdperson singular due to proximity agreement. Therefore, one cannot speak of a striking difference between the registers and formality levels with regard to the frequency of proximity agreement with lexical verbs. Notional agreement is somewhat more frequent in the most informal text category; 15 instances, or 14.4% of the cases of non-agreement in conversations, can be thus explained. Of these, 10 occur in phrases that have an indefinite pronoun as the subject, see (15), while the other 5 occur in phrases in which the head noun may be given a collective reading, see (16). (15)

Everybody have to go to the chapel (S1A-056)

(16)

Otherwise the I’ll say that administration have a lot of respect for the Student Guild (S1A-059)

Interestingly, there is variation in the marking of verbs following such collective nouns, in one case even within a single utterance, see (17). (17)

Hollywood puts two and two together and just give you what you want (S1A-051)

In class lessons, only 2 instances (4%) of grammatical non-agreement can be attributed to notional considerations, one where the verb is preceded by an indefinite pronoun and one where the head can be understood collectively. In broadcast news, 10 instances of notional agreement can be found (55.6% of non-agreement), all of which involve nouns that can be understood as denoting either a group or several individual members of that group (18): (18)

Despite the uncertainty of the tour taking place the West Indies Cricket selection committee have chosen a sixteen-man squad. . . (S2B-017)

6. If only the orthographic transcription is considered, example (14) could be a zero relative clause, but from the audio recording it becomes clear that oh is stressed and the following clause is introduced as new information. Therefore, this utterance is probably not a case of zero relative.

318

Susanne Jantos

The frequency pattern of notional agreement seems rather surprising; the numbers are highest in direct conversations but there are hardly any instances in class lessons, and the numbers increase considerably again in broadcast news. However, when the type of notional agreement is taken into account, it becomes clear that register characteristics are highly influential here. Conversations are the register in which most of the phrases with indefinite pronouns, such as everybody and somebody, occur (86 tokens). These constructions are rather infrequent in class lessons. Of the 57 tokens of *body, most occur in questions, as in (19), and thus they were excluded from analysis. (19)

Anybody want to help me with it (S1B-011)

Indefinite pronouns are absent from the 20 broadcast news texts. Therefore, the distribution of indefinite pronouns correlates negatively with formality levels. In contrast to this, the number of collective nouns rises with the level of formality. A search of Hundt’s (2006) 35 collective nouns and an additional 7 nouns (administration, exec(utive),7 faculty, Hollywood, office, Seacole and UWI ) revealed that these are most frequent in broadcast news (518 instances) and almost equally frequent in class lessons and direct conversations (196 and 201 cases, respectively). Therefore, the rise in notional agreement in news texts is not really surprising anymore. As the above discussion indicates, the difference in the rate of notional agreement between conversations and class lessons is not due to different distributions of collective nouns but instead to the high number of indefinite pronouns in the former category and their infrequency in the latter category. However, not only do the numbers of occurrences differ across the registers, but so too do the types of collective nouns. While direct conversations are particularly rich in collective nouns such as family, university, and UWI, the most frequent collective nouns in class lessons are (in order of frequency) group, class, and government. In broadcast news, ministry, government, and company far outnumber all other collective nouns. These frequencies reflect the characteristic distribution of topics in the text categories discussed and thus provide evidence for the strong influence of formality and register characteristics on the choice of vocabulary and syntactic structures such as subject-verb agreement patterns. The analysis of different types of notional agreement given above shows that Deuber’s assumption of a tendency in Jamaican English to extend notional agreement can be seen only in low formality conversational texts, where indefi7. In the direct conversations, executive was occasionally abbreviated to exec, so a WordSmith search in the three text categories was carried out for exec*. UWI is an acronym that stands for the University of the West Indies.

Agreement in spoken educated Jamaican English

319

nite pronouns are most frequent, but not in broadcast news, where the number of collective nouns is highest. In the former category, indefinite pronouns, such as everybody, which are grammatically singular in International Standard English, are notionally interpreted as plural, thus notional agreement is extended. In the latter register, on the other hand, there are few such reclassifications of nouns. The collective nouns in this category, government, for example, can be interpreted even in International Standard English as entities or groups of persons, hence singular, or as the individual members of the category, hence plural. The same is true for the extension of the use of the quantifier much to contexts in which it determines (plural) count nouns, as in (20): (20)

I don’t have that much loads (S1A-046)

This phenomenon is more frequent in conversations (56%) than in class lessons (33%), and it is completely absent from broadcast news. While the number of occurrences is too low in this case to make reliable assumptions about reclassifications of nouns along the count versus non-count distinction, these numbers do tentatively support Sand’s (1999: 134–135) results. A few initial observations will help in interpreting the results as we turn from notional considerations to further types of non-agreement. Since casual speech is furthest removed from the standard variety and closest to Jamaican Creole, a wide range of upper and even some lower mesolectal forms, for instance the absence of s-marking, can be expected particularly in those environments in which other creole features occur, such as the use of the third-person pronoun him in subject position (see Patrick 1999: 429 and Deuber 2009). By contrast, in highly formal language, such as that commonly used in broadcast news, very few or no creole features are expected, as highly formal language is located closest to the standard on the creole continuum. The data support the above assumption: There are 16 instances of absent –s marking (13.6% of non-agreement) in mesolectal environments in direct conversations, as in (21), while there are no such instances in either class lessons or broadcast news. (21)

So him just still go at that school until now (S1A-061)

In all text categories except broadcast news, the simple unmarking of verbs for third-person singular is the most frequent type of non-agreement, accounting for 60% of all non-agreement in direct conversations and 76% in class lessons. In the most formal category, unmarking is the second most influential factor, accounting for 33.3% of non-agreement, after notional considerations, which are responsible for 55.6% of non-agreement. Although the percentages suggest that unmarking is more frequent in class lessons than in direct conversations,

320

Susanne Jantos

the number of tokens shows that this is not the case: The number declines from 71 instances in texts that are least formal to 39 in texts of medium formality and finally to 6 tokens in texts with the highest formality. Lastly, we will consider what is above called don’t-simplification. Figure 2 shows the distribution of do in positive versus negative and in full versus contracted forms for the three categories; here do means that the form is positive, as in (22). Cases in which the sentence is negated but the full form is used are denoted (not) do, for example (23), and all contracted negative forms appear as not do contracted in Figure 2, of which (24) is an example:

Figure 2. Agreement versus non-agreement with do

(22)

. . . yes that’s what she do8 language and communication I think (S1A050)

(23)

Miss isn’t it that inflection do not change the meaning (S1B-015)

(24)

Cos this guy don’t look like a true geek. . . (S1A-063)

Don’t-simplification with third-person singular subjects occurs 15 times in direct conversations, 7 times in class lessons, and not at all in broadcast news. Thus this feature decreases sharply across registers with increasing formality levels in total 8. Because of the low number of tokens, no distinction was made between cases where do was used as a full verb as opposed to an auxiliary verb. The examples given in (22) through (24) are counted as instances of non-agreement. The verb in each case is not marked for third-person singular and thus does not agree with its subject. The corresponding cases of grammatical agreement would have been she does, inflection does not change, and this guy doesn’t look, respectively.

Agreement in spoken educated Jamaican English

321

numbers, but not in percentages of non-agreement in the respective category (12.7%, 14%, and 0%, respectively). Although the non-agreement contraction don’t is more frequent in direct conversations than is the agreement form doesn’t, it is not a general simplification feature in Jamaican English, not even in informal unmonitored speech. All speakers also use the standard form, which does make number distinctions – doesn’t for third-person singular and don’t for third-person plural subjects. The partial loss of subject-verb agreement that Trudgill (1990) found for some British dialects cannot be postulated for Jamaican English. There is merely a tendency to extend the use of don’t to third-person singular contexts. A phenomenon that qualifies as a plausible explanation for the high frequency of contracted don’t with third-person singular subjects in low formality texts is the presence of the invariant duon in mesolectal Jamaican Creole (cf. Deuber 2009), an invariant negator used with all persons and numbers. In class lessons, where the influence of the creole sharply decreases, agreeing contracted verb forms are far more frequent than are contracted verb forms that do not agree with the subject, and in broadcast news contracted verb forms are totally absent. In all text categories in which the full positive do and negative (not) do occur, agreement is more frequent than is non-agreement. In the following section, subject-verb agreement with be will be examined across the three text categories. Again, the dependence of agreement on features such as subject number or tense, and also the type of non-agreement, will be the focus of attention. 4.2. The verb be Table 3 shows the distribution of the verb be with third-person singular versus plural subjects for the three text categories and the corresponding levels of formality: Table 3. Distribution of be with singular versus plural subjects across text categories

Subject sg Subject pl Total

Conversation

Class Lesson

News

1209 310 1519

1333 334 1667

783 364 1147

The majority of all sentences have singular subjects, only 20.4% of the instances of be in the category of lowest formality have plural subjects, 20.0% in medium formality and 31.7% in high formality texts. In comparison to the many hundreds of instances of grammatical agreement, non-agreement with be is extremely rare in the data, regardless of the subject number or the text category and register.

322

Susanne Jantos

While the rates are negligible, being consistently below 1%, the pattern can nevertheless reveal interesting facts about subject-verb agreement in educated Jamaican English. Non-agreement is marginally higher with plural subjects in all categories, and it increases slightly with increasing formality. The same is true when be is analysed with regard to tense. In direct conversations and class lessons, there is a marginal difference in the frequency of non-agreement, which is slightly higher in present tense (0.5% each in both categories) than in past tense (0.3% in direct conversations versus 0.4 % in class lessons). In broadcast news, this difference is not evident (0.8% in both tenses). An examination of the types of non-agreement that occur may prove insightful if compared to the types that occur with verbs other than be. If the types of non-agreement are similarly distributed across the categories when the verb be is used rather than a lexical verb, this distribution would point to general tendencies of non-agreement that do not depend on the verb type but on other factors, such as formality and register. Figure 3 gives an overview of the types of non-agreement in each register.

Figure 3. Types of non-agreement with the verb be in each text category

A pattern emerges similar to that visible for other types of verbs, namely that notional agreement is highest in texts of high formality. However, with direct conversations and class lessons the pattern is reversed. With be, notional agreement is more frequent in class lessons than it is in direct conversations, whereas this pattern was reversed with other verbs. While the rate of notional agreement is around 10% in direct conversations, 40% of all non-agreement in class lessons and more than 60% in broadcast news can be attributed to notional considerations, which means that in these cases it is not the grammatical number of the subject that determines agreement but instead the noun is given a collective reading. Proximity agreement with be is absent from low formality texts but

Agreement in spoken educated Jamaican English

323

accounts for about 7% of all non-agreement in medium formality texts and for about 22% in high formality texts, as in (25). (25)

Preparations for the ceremony is now in high gear (S2B-020)

Here, the verb agrees with the more proximal singular noun ceremony, although the syntactic head of the subject noun phrase is the plural noun preparations. My data do not support the first part of Quirk’s observation that “[p]roximity concord occurs mainly in unplanned discourse. In writing it will be corrected to grammatical concord if noticed” (Quirk et al. 1985: 757). The analysis instead yields results similar to those of Sand and Deuber, who show that proximity agreement is extremely infrequent in informal, spoken direct conversations (Deuber 2009) and more frequent in news texts (Sand 1999). This is true for both be and other types of verbs. As only spoken data were analysed, it is not possible to make assertions about the validity of the second part of Quirk’s assumption. While the amount of proximity agreement correlates positively with the level of formality, this correlation is not to be taken as evidence of a direct dependence of proximity agreement on register. The correlation might be caused by other factors, such as the number of complex noun phrases, which tends to be higher in more formal texts due to the fact that these texts usually have a higher lexical and informational density than do informal conversations. Verb number9 refers to instances of non-agreement where the verb does not agree with the subject noun in number, but no alternative explanations, such as proximity of another noun or notional number, can be found. In the overwhelming majority of cases, the subject number is plural and the verb used is singular, as in (26), but occasionally a plural verb is used with a singular subject (27). (26)

The features such as ka and na which is uhm. . . (S1B-011)

(27)

Protocol are a treaty or a convention (S1B-008)

This type of grammatical non-agreement decreases from about 45% in direct conversations to 33% in class lessons and to 6% in broadcast news. One of the creole features described by Patrick (2004) explains some of the variation, namely the invariant is, which functions as the copular form for all persons in mesolectal Jamaican Creole and thus is extended to plural contexts in ways that 9. The category “other” consists of verbs occurring in phrases that display syntactic structures like coordinated subjects, for example, which may influence agreement.

324

Susanne Jantos

are not grammatical in International Standard English. The decrease might then be due to the lower influence of Jamaican Creole in texts of higher formality. However, an analysis of a larger data set is needed in order to gain reliable insights into agreement with be. The last phenomenon to be mentioned here is non-agreement in existential there + be constructions, which accounts for about 18% of non-agreement with be in direct conversations, 20% in class lessons, and 5.6% in broadcast news. The numbers are very low when compared to all occurrences of the verb be, but the analysis of only existential constructions reveals some interesting patterns. First, as Figure 4 shows, the non-agreement rate is consistently lower than 10% in all three text categories, but it decreases steadily with increasing formality.

Figure 4. Agreement in existential there + be constructions

While it seems that formality is the crucial influencing factor here, there need not be a causal relation between both variables. The figure merely indicates a slight negative correlation between non-agreement and the level of formality. Alternative explanations for this pattern might be found by considering characteristics of the text categories that are here associated with a particular (relative) level of formality, factors such as the higher syntactic complexity and lexical density of utterances in class lessons and broadcast news as compared to that in direct conversations or the higher cognitive load in lectures, which Crawford (2005) found to be highly influential on agreement in existential constructions. The overall frequency of existential there constructions is highest in class lessons (111 tokens), as predicted by Crawford (2005), and lowest in broadcast news (33 tokens), with direct conversations ranging in between the two (77 tokens). Following Crawford (2005), the cognitive load can be assumed to be highest in

Agreement in spoken educated Jamaican English

325

lectures and thus might cause the high number of there + be constructions in the category of class lessons. This is because, on the one hand, lectures are not pre-formulated in the way news texts are, but, on the other hand, they are also not subject to the competitive strategies of turn keeping and turn taking as are unplanned casual conversations. The necessity of managing discourse in this way may cause existential there + be constructions to be used rather frequently for summarizing topics or introducing new issues. As the results show, however, non-agreement in existential there + be constructions does not seem to be affected by such differences in function but instead follows the overall pattern of non-agreement in educated Jamaican English: The rate of non-agreement decreases steadily with increasing formality of the text, from direct conversations to class lessons to broadcast news.

Figure 5. Agreement in existential there + be constructions with contractions

Figure 5 presents non-agreement rates in existential there constructions depending on whether or not the construction is contracted, as in (28), or not contracted, as in (29). (28)

I know that there’s some negatives to that. . . (S1B-018)

(29)

You’re not allowed to put a lot of things in the fridge because there is so many persons living on hall (S1A-046

There is indeed an effect of contraction on non-agreement in educated Jamaican English in that non-agreement is more frequent in contracted there + be constructions (6.5%) than in non-contracted ones (0.9%). My data thus support Meecham and Foley’s (1994) findings that contractedness influences subjectverb agreement in existential there + be constructions and also provide evidence for Crawford’s assumption that there + be constructions occur more frequently

326

Susanne Jantos

in lectures than they do in direct conversations. However, non-agreement rates were not particularly frequent in the analysed class lessons, possibly owing to the fact that this text category is comprised not only of lectures but also of more interactive class lessons, such as discussions or tutorial groups. Again, a systematic analysis of more homogeneous text types would be valuable and might yield important insights into the highly complex pattern of subject-verb agreement in existential there + be constructions. In sum, the verb be behaves rather differently from regular verbs. First, rates of non-agreement are so low as to be insignificant and second, non-agreement is more frequent with plural subjects than it is with singular subjects. However, the analysis of the types of non-agreement yields a pattern similar to that found for other verbs, namely that notional and proximity agreement increases with increasing formality, while non-agreement of verb number and in existential there + be constructions decreases. Existential there + be constructions have been found to be more frequent in class lessons than in other text types, but the effects of register on non-agreement in there + be constructions across the categories were not as strong as those described by Crawford (2005), a difference which may be due to the heterogeneity of my categories.

5. Conclusion In my analysis of subject-verb agreement in three text categories of ICE-Jamaica, a substantial amount of variation has emerged with regard to phenomena that have frequently been identified as characteristic features of Jamaican English. Subject number is a factor which exerts a strong influence on (non-)agreement with regular verbs as well as with the verb be, regardless of register. As described above, the type of verb, whether it is the exceptional verb be or a regular verb, also plays a role in determining agreement. However, the more interesting and complex findings appear in a systematic analysis of the prevailing types of non-agreement in the categories under consideration. In my data, evidence was found in support of Deuber’s (2009) assumption that a lack of inflection on verbs following third-person singular subjects is a frequent phenomenon in informal Jamaican English. Moreover, a slight tendency to extend notional concord (see Sand 1999) appeared in my data. However, these features did not prove to be evenly distributed in all registers; there was no unidirectional tendency for all types of agreement to decrease with increasing formality. In fact, four trends appeared in my data which contribute to the rather complex pattern of subjectverb agreement in Jamaican English: First, overall rates of grammatical nonagreement decrease with increasing formality; grammatical non-agreement is

Agreement in spoken educated Jamaican English

327

most frequent in direct conversations and least frequent in broadcast news. This finding, together with the occasional hypercorrectly marked verbs in the higher formality levels, indicates that leaving verbs unmarked is stigmatized, while the use of s-marking on third-person singular verbs indicates speaker prestige. Second, register correlates differently with the various types of grammatical non-agreement. Proximity agreement, for instance, increases with increasing formality and thus definitely cannot be characterized as a feature of unplanned discourse (cf. Quirk et al. 1985). The same seems to be true for notional agreement, but here the pattern is not straightforward. While the tendency to extend notional agreement to indefinite pronouns such as somebody or everybody was found in direct conversations (cf. Deuber 2009), it was rather infrequent in class lessons and completely absent from broadcast news. That the rate of notional agreement is nevertheless higher in news texts can be explained by the higher number of collective nouns in this category. Here, register characteristics are extremely influential in that they account for an opposing trend, particularly regarding the choice of vocabulary. While the high frequency of indefinite pronouns in direct conversations and their absence in news texts can be explained by the prevalence of interpersonal topics in conversations and the lack of such topics in news texts, the high number of collective nouns in news also can be attributed to common news topics. In broadcast news, topics such as politics, where the nouns government and ministry are particularly well represented, or economics, where recurrent reference is made to company, are much more frequent than in conversations, where collective nouns, such as family, do occur, but much more rarely. Third, a simplification process of the type Trudgill (1990) found in dialects of England or a general trend in varieties of English towards erosion of inflectional morphology do not qualify as a plausible explanation for agreement variation in my data, as these features do not consistently occur in all categories. What I call don’t-simplification occurs mostly in direct conversations and occasionally, but not very frequently, in class lessons. However, such instances are not found in broadcast news. Thus, one cannot speak of a general trend in Jamaican English to simplify inflection. The more likely explanation for the extension of the contracted form don’t to third-person singular subjects is an influence of the creole form duon, which can be used with all persons and numbers. Creole influence is the most compelling explanation for most of the types of non-agreement in the category of direct conversations, namely the extension of notional agreement to indirect pronouns, the use of the quantifier much with plural count nouns, the use of is and don’t with plural subjects, and the simple unmarking of verbs.

328

Susanne Jantos

However, and this is the fourth trend revealed by my data, creole influence decreases strongly with increasing formality. In direct conversations, lack of inflection occurs particularly frequently in environments where other creole features can be found, such as the use of him or them as subject pronouns. Both the number of such creole environments and the instances of unmarked verbs decrease in class lessons and decrease even more in broadcast news, where creole environments are no longer found and unmarked verbs are rather scarce. Thus, it appears that Mair’s (2002: 36) conclusion that “[p]atois is clearly the dominant shaping influence on spoken English in Jamaica” can be fully supported in the category of direct conversations, but is less supported in the category of class lessons, and does not seem to hold in the case of broadcast news, since in this last category notional agreement is far more influential than are creole features. The analysis indicates that conversations and class lessons resemble one another in many respects, especially when they are compared to broadcast news. This is true for the prevailing types of non-agreement and for occurrences of different kinds of notional agreement, but also for the rates of don’t-simplification and proximity agreement. The contrast between conversations and class lessons is much less pronounced than that between class lessons and broadcast news. The heterogeneity of the categories might account for this similarity, as the text type direct conversations comprises not only casual conversations but also more formal interviews, and class lessons include not only rather formal lectures, but also more interactive and informal tutorial groups. The texts within the category broadcast news, however, are more homogeneous and consistently on a very formal level. Moreover, broadcast news data consist of scripted texts, while all other texts are produced in time. Thus this text type contrasts sharply with the other two, while the differences between the other two text types are not so marked. In conclusion, an analysis of register variation proves to be a very valuable tool for investigating subject-verb agreement in Jamaican English, both because it enables linguists to gain interesting and more accurate insights into this rather complex phenomenon and because of its dependence on situational factors, which the analysis of a whole corpus or small sections thereof may overlook. What remains to be investigated is register variation in corresponding written sections of ICE-Jamaica. From such an investigation one will be able to draw conclusions as to what differences are caused by register characteristics or formality and how much variation can be attributed to the mode of production.

Agreement in spoken educated Jamaican English

329

References Bailey, Beryl L. 1966 Jamaican Creole Syntax: A Transformational Approach. Cambridge: Cambridge University Press. Biber, Douglas, Stig Johannson, Geoffrey Leech, Susan Conrad and Edward Finegan 1999 Longman Grammar of Spoken and Written English. London: Longman. Britain, David 2002 Diffusion, levelling, simplification and reallocation in past tense be in the English Fens. Journal of Sociolinguistics 6: 16–43. Cheshire, Jenny 1999 Spoken Standard English. In: Tony Bex and Richard Watts (eds.), Standard English: The Widening Debate, 129–148. London: Routledge. Crawford, William 2005 Verb agreement and disagreement: A corpus investigation of concord variation in existential there + be constructions. Journal of English Linguistics 33: 35–61. Deuber, Dagmar 2009 The English we speaking: Morphological and syntactic variation in educated Jamaican speech. Journal of Pidgin and Creole Languages 24: 1–52. G¨unthner, Susanne this volume Grammatical constructions and communicative genres. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 195–217. (Topics in English Linguistics 70.) Berlin/NewYork: Mouton de Gruyter. Hay, Jennifer and Daniel Schreier 2004 Reversing the trajectory of language change: Subject-verb agreement with be in New Zealand English. Language Variation and Change 16: 209–235. Hudson, Richard 1999 Subject-verb agreement in English. English Language and Linguistics 3: 173–207. Hundt, Marianne 2006 The committee has/have decided . . . : On concord patterns with collective nouns in inner- and outer-circle varieties of English. Journal of English Linguistics 34: 206–232. Mair, Christian 1992 Problems in the compilation of a corpus of Standard Caribbean English. A pilot study. In: Gerhard Leitner (ed.), New Directions in English Language Corpora, 75–96. (Topics in English Linguistics 9.) Berlin/New York: Mouton de Gruyter. Mair, Christian 2002 Creolisms in an emerging standard: Written English in Jamaica. English World-Wide: A Journal of Varieties of English 23: 31–58.

330

Susanne Jantos

Meecham, Marjory and Michele Foley 1994 On resolving disagreement: Linguistic theory and variation – There’s bridges. Language Variation and Change 6: 63–85. Mondorf, Britta this volume Genre effects in the replacement of reflexives by particles. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 219– 245. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Patrick, Peter 1999 Urban Jamaican Creole: Variation in the Mesolect. (Varieties of English around the World 17.) Amsterdam/Philadelphia: John Benjamins. Patrick, Peter 2004 Jamaican Creole: Morphology and syntax. In: Bernd Kortmann, Edgar W. Schneider, Clive Upton, Rajend Mesthrie and Kate Burridge (eds.), A Handbook of Varieties of English. Vol. II: Morphology and Syntax, 407–438. Berlin/New York: Mouton de Gruyter. Platt, John, Heidi Weber and Mian Lian Ho 1984 The New Englishes. London: Routledge and Kegan Paul. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik 1985 A Comprehensive Grammar of the English Language. London: Longman. Sand, Andrea 1999 Linguistic Variation in Jamaica: A Corpus-Based Study of Radio and Newspaper Usage. (Language in Performance 29.) T¨ubingen: Gunter Narr. Schilling-Estes, Natalie and Walt Wolfram 1994 Convergent explanation and alternative regularization patterns: Were/ weren’t leveling in a vernacular English variety. Language Variation and Change 6: 273–302. Schneider, Edgar 2007 The dynamics of New Englishes: From identity construction to dialect birth. Language 79: 233–281. Szmrecsanyi, Benedikt and Bernd Kortmann 2009 Between simplification and complexification: Non-standard varieties of English around the world. In: Geoffrey Sampson, David Gil and Peter Trudgill (eds.), Language Complexity as an Evolving Variable, 64–79. (Studies in the Evolution of Language 13.) Oxford: Oxford University Press. Tagliamonte, Sali 1998 Was/were variation across the generations: View from the city of York. Language Variation and Change 10: 153–191.

Agreement in spoken educated Jamaican English

331

Tagliamonte, Sali and Jennifer Smith 2000 Analogical leveling in Saman´a English: The case of was and were. Journal of English Linguistics 27: 8–26. Trudgill, Peter 1990 The Dialects of England. Oxford: Blackwell.

I know you guys hate forwards: Address pronouns in digital folklore Theresa Heyd

While the Internet is a vibrant site of genre emergence and change, it remains disputed whether this macro-change plays a role in actual linguistic micro-change. This paper argues for a more subtle role of computer-mediated communication (CMC) genres: Due to their intermediate spoken/written status, they can act as catalysts for ongoing processes of linguistic change and provide a first “point of entry” into the written domain for previously oral forms. This hypothesis is explored here within the genre ecology of digital folklore. The study is based on a corpus of 150 email hoaxes, deceptive messages that spread in digital social networks and are a form of 1-to-n communication fostering open plurality. The morphosyntactic variable under investigation is the emergence of 2nd person plural forms such as you guys that strive to fill the gap in the English pronoun paradigm. The occurrence of such forms in the email hoax corpus is analyzed: what is their function within the genre? How may the technicality of the medium ultimately support the process of linguistic change? As an outlook, comparative data from digital and traditional genres are discussed.

1. Introduction 1.1. Genre theory in computer-mediated communication: An overview The notion of genre – however one may wish to define it – has strongly influenced studies in computer-mediated communication (CMC) for a number of years. Basically, two more or less complementary strands of digital genre research can be made out: On the one hand, the more academic approaches, in particular from a linguistic perspective, have grappled with the ontological status of digital genre, with its conditions of emergence and potential family resemblances with pre-digital genres (for examples, see Gains 1999; Zitzen and Stein 2004; Barron 2006). In parallel – and largely untouched by this academic discussion – so-called document genres have become a central tool for text classification in applied fields such as data mining, library science, and other

334

Theresa Heyd

applied technologies (see Crowston and Williams 2000; Kwasnik and Crowston 2005 for an overview). The differences in these approaches are striking: While the issue of precise and principled categorization is an end unto itself in genre theory, the more applied fields tend to rely on purely pragmatic, best-practice strategies in categorizing and classifying texts. This co-existence of research strands may serve as an initial explanation as to why the notion of genre – once a rather static concept of literary theory – has undergone such a remarkable renaissance. If the approach presented here is grounded in the linguistic CMC tradition, the potential of the genre concept in applied settings can be seen as a powerful driving force behind digital genre theory in general. In a very general way, genre is considered here as a form of textual categorization: By viewing a piece of text as pertaining to a genre, we assign it a discursive category. From a linguistic perspective, the proximity to prototype theory and other cognitive approaches to semantics is self-evident. Making sense of the world by segmenting and grouping its entities into categories is a fundamental cognitive given in humans, and there is no reason why this strategy should not apply to the discourse universe. Linguistic categorization has been explored in depth in recent decades (see Aarts 2006 for an overview); while most approaches to genre do not explicitly make this connection, it still seems justified to describe genre theory as the prototype theory of discourse linguistics. It fits in well with this semantic approach that many researchers in recent years have pointed out the horizontal interrelations between neighboring phenomena: genres are not isolated items but form continua with weak and strong ties, hierarchies, and the possibility of overlap. Thus early papers by Devitt (1991) and Bazerman (1994) have described genre sets or systems where one genre entails the production of another; Orlikowski andYates (1994) have coined the term genre repertoires for the shared genre continuum of discourse communities (see also Virtanen, this volume). In CMC studies, the notion of a genre ecology has become popular; based on a study by Erickson (2000), Kwasnik and Crowston (2005) suggest that we “extend [Erickson’s] apt metaphor because it captures succinctly how, like any organism in an ecological community, genres have effects on each other and depend on each other for their effectiveness” (Kwasnik and Crowston 2005: 81–82). This approach is taken up in the case study presented here: By describing digital folklore with its subgenres, predigital antecedents and genre neighbors, it outlines a genre ecology for, as it were, one particular ecological niche within a discourse continuum. Of course, this very general description leaves open many of the recurring questions regarding genre theory. For one, it does not define the ontological status of genres: While many linguistic approaches (for example, Herring 2007)

I know you guys hate forwards: Address pronouns in digital folklore

335

treat category labels as items that can be systematically defined and ascribed (so that the term “classification” might be more fitting), other approaches – and particularly the bottom-up folksonomies of Web 2.0 appliances – treat genres as cognitive givens that are primarily defined by their usage through speakers. More importantly, it leaves open how genres are to be defined – through formal or functional features, based on intra- or extratextual issues, on a micro- or macrolevel. The contributions by Virtanen and Giltrow in this volume give a detailed account of this ongoing debate, from the traditional understanding of microlinguistically defined text types to highly situation-based models such as New Rhetorical genre theory. The approach presented here is not grounded in one specific school, but attempts to draw from both form and function driven genre models. Finally, the recent interest in digital genres has highlighted a further factor in the genesis and constitution of genres, namely that of (socio-)technical determination: While textual categories do not emerge through technological innovation alone (Herring 2004), the influence of hardware and software on the shape of computer-mediated communication genres can hardly be denied. The data presented here show evidence of such dynamics in email, a wellestablished socio-technical mode of CMC. Studies on more recent phenomena (e.g. Puschmann on blogs, this volume) suggest that this tendency continues to hold as online communication has moved into its Web 2.0 phase. 1.2. Scope of this study The case study presented here tries to give an account of these interwoven problems. It takes the perspective of one specific CMC genre candidate, namely digital folklore; within this textual framework, it analyzes the occurrence of one specific morpho-syntacic variable that appears to be particularly relevant for digital folklore as a genre, namely forms of plural address.1 Pronominal usage as a linguistic feature has successfully been analyzed in other explorations of CMC and its implications for linguistic innovation and change (see, e.g., Yates 1996 for a classic example; Puschmann, this volume for a more recent approach). Section 2 offers an initial description and analysis of digital folklore, its emergence, characteristics and potential pre-digital predecessors. It also presents a two-level model of digital genre (see also Heyd 2008) that can accommodate for various descriptive levels. In short, the genre ecology of digital folklore is outlined. 1. A very early version of this study has appeared in Heyd (2008), chapter 3. The author is grateful to the editors of this volume, and to all the participants of the workshop on genre and syntactic variation at the DGfS 2007 meeting in Siegen, for their valuable comments and insights.

336

Theresa Heyd

Section 3 discusses the relevance of morpho-syntactic variation in CMC genres by introducing pronominal and vocative plurality as a morpho-syntactic issue that is relevant in digital folklore. It first gives a brief description of discursive strategies for marking plurality in English and highlights pragmatic and syntactic factors involved, such as the communicative setting or the gap in the English pronoun paradigm. It then presents quantitative and qualitative evidence of plural reference and addressing in digital folklore and discusses possible functional implications: In particular, it analyzes how plurality is exploited for sociopragmatic effects, and to what the degree the technicality of the medium encourages the occurrence of such forms. Finally, the concluding discussion in section 4 offers some speculations on how syntactic variation and language change may be intertwined with digital genres and genre change.

2. Introducing the data 2.1. Digital Folklore: A CMC genre candidate Ethnographic studies of folklore have traditionally had a strong bias towards data that are: a) spoken; and that b) pertain to indigenous, native, or at least rural communities – traditional songs, tales or rhymes are typical material. It was not until the second half of the twentieth century that the perspective was gradually widened to include much more contemporary forms of folklore, such as written (or at least paper-based) data that occur in urban, (post-)industrial societies. This widening of scope is largely due to the studies by Dundes and Pagter (for example, 1992) on officelore and other urban folklore that is transmitted in workplace environments via photocopying and similar technical means. Two functional criteria for urban folklore have been emphasized in this context, namely the multiple existence of items (engendered by some means of message reproduction), and variation in the phenotype of those items (as a result of minor deviations). Dundes and Pagter’s credo that “folklore in the modern world includes urban folklore from the paperwork empire” (1992: xxii) has been impressively underlined in large collections of officelore materials. In sum, this focus shift is reminiscent of sociolinguistics, where Labov’s studies in the 1970s engendered a transformation from largely rural dialectology studies to urban sociolinguistics. Within this progressive framework, it is not surprising that folklore has found its way into the digital dimension. Indeed, the sociotechnical givens of the World Wide Web and, more generally, the Internet, can be said to foster the development of folkloristic material: From a sociological vantage point, it connects a huge

I know you guys hate forwards: Address pronouns in digital folklore

337

number of participants in a vastly extended social network; from a technological vantage point, it offers ample (and user-friendly) means of reduplicating and transmitting material. The existence of digital folklore has been acknowledged in earlier studies. Fernback (2003) still treats “legends on the net” as an indication of “cyberspace as a site for oral culture” (Fernback 2003: 39) and thus adheres to a largely traditional view of folklore. By contrast, Kibby (2005) sees email as a quite natural, and even technically enhanced, medial environment for folklore, although she still emphasizes the intermediate status of “secondary orality” that is often ascribed to email (Kibby 2005: 771). However, these initial studies are based on a rather slim data layer. I have proposed an extensive description of digital folklore as a genre of CMC (Heyd 2008) based on a corpus of 147 texts; the main features of digital folklore are outlined below. In principle, folkloristic material may be found in many if not most sociotechnical modes of the Internet – there is no inherent reason why folklore should not be displayed in blogs, discussion groups, and so on. However, the term is here limited to discourse that is transmitted via email. This is due to the technical givens of modern email programs. It has long become standard that email messages can be easily stored, forwarded and sent to multiple recipients simultaneously. As noted above, email thus provides a perfect medial environment for folkloristic discourse as it ensures the multiple existence of items. Whereas traditional folklore is passed on via word of mouth, and officelore, in the sense discussed by Dundes and Pagter, relies on reproductive means such as the photocopier or hand drawing/typing, the technicality of email programs is ideally suited to store, manage, and disseminate folkloristic material. Digital folklore is therefore defined here as textual material that is passed along in digital social networks via the forward function of email programs. A second central aspect concerns the contents of digital folklore. In principle, there is a broad range of subject matters that occur in digital folklore – indeed, digital folklore can be subcategorized into various subgenres on this basis, some of which are briefly introduced below. However, a common denominator applies to all forms of digital folklore: In order to be successful – that is, to warrant a sufficient level of dissemination – they must contain material that is in some way non-canonical, spectacular, or even sensationalist. In the terms of narrative theory, such discourse must display a degree of tellability. Where the contents of digital folklore are text-based, this is often true in a very literal sense, as such discourse is often presented in a more or less narrative format. Where the material is not exclusively text-based (audio or video attachments, PowerPoint slideshows, pictures, and other graphic material), the notion of tellability can still be applied in a wider sense; such material must supply the recipient with stimulating, thought-provoking, or generally interesting contents.

338

Theresa Heyd

A third characteristic of digital folklore is again in line with the criteria evoked by Dundes and Pagter (1992): Different instantiations of one digital folklore item typically exhibit a degree of variation. This usually concerns minor differences that accumulate in the course of the forwarding process; variation thus is not a factor inherent in digital folklore from the very beginning, but a quality that necessarily emerges during the life cycle of a folkloristic item. In text-based digital folklore, quantitative as well as qualitative changes may be observed: Text may be added or deleted, but users will also interact with the makeup of the material itself. Changes occur on all linguistic levels, from minor typographic/orthographic corrections to lexicogrammatical changes and even paraphrases of whole paragraphs or texts. While some forms of variation are technically induced – email programs automatically insert material such as signatures and transmission information – the bulk of variation is clearly motivated by users’ wishes to engage with the material. Although the notion of variation is well known from earlier forms of folklore, it is all the more remarkable that it should occur in digital folklore, as the technical means of the medium enable a transmission of data without any manual copying or reduplication. In sum, variation between items appears to be a very central criterion of digital folklore. So far, the genre of digital folklore has been described by functional criteria that concern its sociotechnical makeup. As mentioned above, digital folklore can be further subcategorized according to more text-internal and specific aspects, such as topic and structure. Such a content-based differentiation of the supergenre goes hand in hand with bundles of other formal features, such as discourse structure or recurrent lexicogrammatical features. As of yet, no comprehensive study exists of digital folklore as a genre spectrum. The following enumeration therefore does not claim to be exhaustive. However, as a first approximation, three major types of text can be made out, namely humorous or jocular, emotional or inspirational, and political texts. A large amount of digital folklore is of an overtly jocular nature. For non-textbased items, this may include visual humor such as photoshopped pictures or more traditional cartoons with punch lines. In the text-based realm, a relatively established format is the joke list, where conventional jokes are collected, often with a certain thematic orientation – the classic case probably being the list of lawyer jokes. Themes other than that of traditional jokes can form the scope of such jocular “forwardables” (Kibby 2005); cases in point are collections of famous quotes (be they from literature or from TV shows), lists of pick-up lines, strange facts, “yo mama” jokes (the scatological sounding rituals already described in Labov 1972), and many others. In general, the list format appears to be highly amenable to digital folklore as a structural entity. Enumerations exist for virtually any conceivable subject

I know you guys hate forwards: Address pronouns in digital folklore

339

or purpose. There appears to be something deeply alluring, indeed cognitively salient, about this discourse format. One list type that has become particularly conventionalized, and is therefore likely to be a subgenre, is theYou know you are [from] x if y list. These humoristic enumerations play on perceptions of social, ethnic, or cultural groups; usually, their purpose is identity creating rather than defamatory. The variations for x are myriad: Items range from You know you’re a New Orleans native if [. . . ] to You know you’re using ICQ2 too much when [. . . ]. It can be speculated that it is the relatively rigid discourse format that makes for its attractiveness; even more importantly, it can easily accommodate for additions. A short excerpt from such a digital folklore message is given in text (1).3 Text (1) From: Rachel Keyte Date: Sun, 11 Feb 2001 12:36:25 +1100 > >Hey destinions, > I remember someone posting something like this a few years ago, but I > thought I’d share this latest piece of humour with you all again: > >You know you are a C´ eline Dion fan when: > >You know there are more than 5 official releases available >You have all the albums on CD for playing, on LP for collection, on >MC for a walkman >Your standard phrases are: "that’s the way it is", "there is only one >road I am walking", "think twice", or "where does my hart beat now" >You marry someone who is 40 years older than you are [...] (Source: http://www.cm.nu/∼shane/lists/destin/2001-02/0056.html¸ accessed September 14, 2009)

A further, highly specific type of list is the survey: Surveys are enumerations of questions that typically elicit personal information of a more or less sensitive, emotional, or humoristic nature. The recipient fills in his or her answers and forwards the questionnaire to others, who in turn complete the list. In addition to forwarding, such surveys are particularly popular for secondary utilization on personal homepages, blogs and so on.

2. ICQ is an online instant messaging service. 3. Examples have not been edited except for minimal formatting; in particular, typographic or grammatical infelicities have been left unchanged.

340

Theresa Heyd

Apart from humoristic contents, inspirational texts are an important category of digital folklore. Inspirational digital folklore can include anything from prayers, poems, to narratives, and so on. A relatively discrete subtype is the prayer chain, where spiritual support is requested for a person in need. Another distinctive subtype is the good-luck chain, where the communicant is invited to make a wish, or is promised good fortune if he or she forwards the email to a certain number of recipients. Emotional and pro-social contents in forwardables appear to be an effective trigger for the dissemination of a message; it is no wonder, then, that this orientation provides ample material for digital folklore. A prayer chain message is displayed in text (2). Text (2)

-----~Original Message~----Forward>>>> At 10am this morning I received a prayer request from church. Cathy Mitchell, who attends Spring Hills, called in to our secretary. Her husband,Tony, is an Air Force Commander in Afghanistan. She received an urgent email from him this morning. It said, "We Need Christians to pray, pray, pray." Please pray for God’s protection of your troops and HIS wisdom for their commanders. Pass this on to as Many as you think will respond. Lord, hold our troops in your loving hands. Protect them as they protect us. Bless them and their families for the selfless acts They perform for us in our time of need. I ask this in the name of Jesus, our Lord an Savior. When you receive this, please stop for a moment and say a prayer for our ground, air and navy personnel in every area of the middle East. There is nothing attached.... This can be very powerful.... Just send this to all the people in your address book. Do not stop this prayer chain, please.... Of all the gifts you Could give to anyone in the US Military, be it Army, Navy, Marines or National Guard, Prayer is the very best one.... Amen! (Source: http://www.freelists.org/archives/mhs.51/10-2003/msg00000.html, accessed September 14, 2009)

I know you guys hate forwards: Address pronouns in digital folklore

341

Finally, many items of digital folklore have a political orientation or make a comment on social, economic, or ecological issues. One distinctive form of political digital folklore is the email petition, where communicants are asked to sign their names under a petition text before forwarding the message. Email petitions are almost always fictitious in that they are not part of a principled larger political effort or campaign, and therefore remain without consequence. While the originators of email petitions usually mean well, serious political organizations or initiatives do not use this tool; even if the information contained is accurate, such forwarded texts easily become out-of-date or get changed through textual interaction. Nevertheless, this format continues to be a highly popular form of digital folklore – it offers communicants the opportunity for that effortless altruism that has aptly been termed “armchair activism” by John Ratliff (2007). A less structured but popular subtype in this spectrum is the political rant, where political claims and opinions are uttered in an often populist or radical manner. Some of these rants have become highly popularized items of the Internet; a case in point is the I am a bad American manifesto, which, incidentally, is organized over large parts into a list format; an excerpt is given in text (3). Such politically oriented digital folklore has arguably become even more important in the post-9/11 era. Text (3) I usually am against the propagation of these "please forward" emails. However, given the recent discussion topics, I thought it was relevant and entertaining to both sides of the issues. rudy --- Michael Ringer wrote: > Reply-to: [email protected] > > Bad American > > I Am Your Worst Nightmare. I am a BAD American. > I like big cars, big tits, and big cigars. > I believe the money I make belongs to me and my family, > not some mid-level > governmental functionary with a bad comb-over who wants > to give it away > to crack addicts squirting out babies. I’m not in touch > with my feelings > and I like it that way, dammit. > I believe no one ever died because of something Ozzy

342

Theresa Heyd

> Osbourne, Ice-T or > Marilyn Manson sang. > I think owning a gun doesn’t make you a killer. > I believe it’s called the Boy Scouts for a reason. > I don’t think being a minority makes you noble or > victimized. > I don’t celebrate Kwanzaa. > I believe that if you are selling me a Big Mac, you’d > better do it in > English. > I don’t use the excuse "it’s for the children" as a > shield for unpopular > opinions or actions. > I think Oprah is a big fat pig. > I think fireworks should be legal on the 4th of July. > I think that being a student doesn’t give you any more > enlightenment than > working at Blockbuster. In fact, if your parents are > footing the bill to put > your pansy ass through 4--7 years of senior high school, > you haven’t > begun to be enlightened. [...] (Source: http://www.xent.com/FoRK-archive/2001.05/0498.html, accessed September 14, 2009)

2.2. The email hoax corpus The study that forms the basis for the observations proposed here was oriented toward a further subcategory of digital folklore, namely email hoaxes. Email hoaxes are one of the most well-known instantiations of digital folklore, as they have been present in mailboxes all over the world since the mid-nineties; the fact that the genre label “email hoax” is widely recognized by computer users is a strong indicator for its status as a subgenre. Very briefly, email hoaxes can be characterized as deceptive messages that follow the pattern of digital folklore; in other words, they contain a proposition that is false or at least problematic. In this sense, they differ from many inspirational or jocular digital folklore texts that have no inherent truth value. This quality has, without a doubt, contributed to the prominence of email hoaxes in the digital folklore genre spectrum, as the violation of the Cooperative Principle that goes along with deceptive behavior is felt very strongly by communicants. Email hoaxes can be subdivided into at least three topic fields: fake virus warnings, giveaway hoaxes (where free

I know you guys hate forwards: Address pronouns in digital folklore

343

goods are promised), and charity hoaxes (where people in need are invented). An example of a giveaway hoax is given below. Text (4) I could use some new underwear. Maybe you could to? My name is Victoria Johnson, founder of Victoria’s Secret. In an attempt to get our name out to more people in the rural communities where we are not currently located we are offering a $50 gift certificate to anyone who forwards this email to 9 of their friends. Just send this email to them and you will recieve a an email back with a confirmation number to claim your gift certificate. Sincerely Victoria Johnson Founder of Victoria’s Secret Hey guys, DONT DELETE THIS EMAIL It really works, I tried it and got my Gift certificate confirmation number in 3 minutes. (Email hoax corpus, see below)

The case study presented here is based on a corpus of 147 email hoaxes (for more information see Heyd 2008); while more investigations of other forms of digital folklore are clearly needed, it is felt that the results presented in section 3 can be extended to the supergenre as a whole. Figure 1 gives a summarizing overview of digital folklore as a genre ecology. It highlights the two-level structure of digital folklore as a supergenre with numerous subgenres at the formal/content-based level. As has been pointed out, this subcategorization is an open set; the spectrum shown here is not exhaustive. The model also acknowledges the factor of genre change, or “intergenre formation” (Virtanen, this volume: 70), as it indicates urban folklore as the historical antecedent of digital folklore.

3. Morphosyntactic variation in digital folklore: The case of plural address 3.1. Forms of plural address in English: An overview The description of digital folklore up to this point has been in line with many approaches to CMC genres: Thus it is often felt that digital genres are most aptly described through their social and technical givens (see, for example, Askehave and Nielsen 2005). This is in contrast with many more traditional

344

Theresa Heyd

Figure 1. Digital folklore as a genre ecology

approaches to text types in the pre-digital sphere, where linguistic patterns are the central yardstick for textual categorization; well-established examples include the stylistic analysis by Crystal and Davy (1969) or Biber’s notion of text types (in, for example, Biber 1988). However, this difference in focus does not imply that linguistic criteria are irrelevant or inapplicable to digital genres. To the contrary, particular sociotechnical settings may give rise to particular lexical or (morpho-)syntactic patterns. One very striking and well-established example is the use of acronyms and abbreviations in text messaging (see, for example, Crystal 2001), where the linguistic form is clearly motivated by technical constraints. Other cases in point that have been studied include the grassroots English used by non-native Nigerian speakers in emails (Blommaert and Omoniyi 2006) and – from a pragmatic vantage point – the specific turn-taking patterns of chat (Zitzen and Stein 2004). The following section presents a morphosyntactic variable that is highly prevalent in digital folklore, namely open plurality in noun phrases, particularly when they are used as address forms, this concerns the use of constructions such as all of you or you guys. Open plurality has strong pragmatic motivations that can be brought in line with the sociotechnical shape of digital folklore outlined above. After a brief sketch of this pragmatic constitution an overview describes the morphosyntactic instantiations of open plurality; these are of particular linguistic interest as second-person plural pronouns make up a domain of ongoing language change in present-day English. Finally, data from the email hoax corpus are presented and discussed.

I know you guys hate forwards: Address pronouns in digital folklore

345

It is acknowledged here that open plurality is neither specific nor necessary for digital folklore as a genre. However, certain linguistic patterns can and do coincide with forms of digital genres. The more general question of syntactic variation in CMC, and of how language change and genre change may be intertwined, is taken up again in the concluding discussion. 3.1.1. Pragmatic background Lyons (1977) famously describes the default setting toward which human discourse is geared: “The canonical situation of utterance involves one-one, or onemany signalling in the phonic medium along the vocal-auditory channel, with all the participants present in the same actual situation able to see one another and to perceive the non-vocal paralinguistic features of their utterances, and each assuming the role of sender and receiver in turn” (Lyons 1977: 637–638). Of course, there is a broad range of genres that diverge from this basic pattern: There is unidirectional discourse and long-distance transmission, not to mention the realm of written discourse. Possibly the most fundamental source of variation is the one that Lyons implicitly mentions, namely the aspect of person deixis. Discourse is often directed at more than one; this fundamental given can be encoded through the deictic anchoring of the utterance with the help of distal person deictics. Indeed, there exists a wealth of genres that typically – or even necessarily – occur in a 1-to-n speech situation. While some of these are very traditional and fundamental, such as storytelling, sermons, and so on, the proportion of 1-to-n genres has greatly increased with the advent of modern mass media, which by definition address a multitude of recipients. Among the more tangible genres of this kind are TV and radio shows (typically with a host or similar speaker figure), newspaper articles and editorials, scientific papers, and novels (with a narrator persona). Quite obviously, with the advent of the Internet, a further layer of genres grounded in the mass medial – and thus prone to a 1-to-n design – has begun to evolve. For a more detailed outline of deictic anchoring and its pertinence to emergent digital genres, see also Puschmann (this volume). 3.1.2. Encoding strategies From a structural perspective, there are several morphosyntactic strategies of encoding plurality. They arise from differing pragmatic needs, which is hardly surprising: pronominal usage, due to its deictic and interpersonal nature, is often intertwined with pragmatic considerations. Essentially, the following three strategies are of relevance here:

346

Theresa Heyd

1) Subdue or leave ambiguous the deictic anchoring. It is a viable strategy to avoid pronominal reference to both first- and second-person roles in discourse as far as possible. First- and second-person pronouns can be replaced by impersonal constructions; alternatively, utterances can be phrased so as not to contain person-deictic reference at all. Such avoidance strategies are well-known in situations where politeness is at stake, in particular where a speaker is not sure whether to use t- or v-forms in addressing someone. In the case of 1-to-1 versus 1-to-n addressing, this form of ambiguity is likely to occur where the speaker wishes to leave unspecified the intended audience of his or her utterance. 2) Simulate a 1-to-1 situation. In another common strategy, speakers may use singular constructions despite a manifestly plural audience. Well-known instances of this are address forms such as Dear reader in novels, editorials, and so on. The pragmatic effect is quite clear: The use of singular forms can be used to simulate closeness or intimacy, states which are usually absent from massmedial communication. While 1-to-1 simulation is arguably easier where there is no physical co-presence of the addressees (for example, in the written medium or in TV or radio broadcasts), it may also occur in face-to-face situations, if the audience is sufficiently large; thus sermons or speeches are situations in which it is conceivable that 1-to-1 simulation may be used by the speaker. 3) Openly display the 1-to-n situation. This can be considered the default strategy for encoding plural addresses. There are several morphosyntactic options for displaying open plurality. Considering that alternative encoding strategies exist, it can be hypothesized that open plurality is a way of pragmatically foregrounding the group nature of the utterance and even of fostering a sense of group dynamics. The main focus here is on open plurality. This is only in part due to methodological issues: For morphosyntactic reasons, cases of plurality are much easier to identify and analyze within a text corpus. More importantly, however, open plurality is a strategy that pertains to digital genres and their sociotechnical givens. It is a frequently voiced hypothesis that open multiplicity is more typical of oral discourse, whereas written discourse tends to subdue the plural orientation. Thus Maynor (2000: 416) argues that “second-person contexts [. . . ] are much more likely to arise in conversation.” Given the arguably intermediate status of most CMC genres between the traditional domains of spoken and written discourse, it is of great interest to examine how the 1-to-n situation is encoded in digital genres – specifically, the question of whether these genres are geared towards the open plurality standard of oral genres, in the way that many forms of CMC incorporate aspects of spoken discourse, is considered here. Plural marking occurs in the noun phrases of an utterance. Structurally, there are two loci in a sentence where plural-marked noun phrases are found: as

I know you guys hate forwards: Address pronouns in digital folklore

347

an argument of a predicate within the sentence structure (intra-sententially, in subject or object position), or as elements that are not directly incorporated into the sentence structure (vocatives, often at the beginning or the end of utterances). Vocatives are never syntactically necessary elements in a sentence; they are pure, pragmatically charged, add-ons. As Zwicky (1974: 796) remarks, vocatives are “almost never neutral” but can “express attitude, politeness, formality, status, intimacy, or a role relationship.” Axelson (2007: 101) gives a list of additional effects that includes in-group status and pseudo-intimacy. Vocatives normally occur at the periphery of sentences as in examples (1) and (2): (1)

Hey guys, DONT DELETE THIS EMAIL

(2)

Come on you guys [. . . ] (Email hoax corpus)

However, they can occasionally occur within a sentence construction as in example (3) (note also the typical clustering of vocative items): (3)

but seriously guys, you guys i i have to get some help because i just, i like to burn things i’m a pyro. (MICASE Corpus)

Examples (1) and (2) also show that vocatives often (but not necessarily) pertain to imperatives; indeed, vocatives have frequently been analyzed in terms of resembling imperative subjects (see, for example, Leech 1999 for an overview). By contrast, plural marking can also occur in noun phrases that are structurally required in a sentence, that is, noun phrases that are in subject or object position. By consequence, they are not as pragmatically charged as vocatives. Here are two examples of plural marking in subject (example 4) and object (example 5) positions: (4)

Most of you who tried it will probably say NO.

(5)

I am asking you all, begging you to please, forward this email on to anyone and everyone you know, PLEASE [. . . ] (Email hoax corpus)

Examples (1) to (5) already give an impression of the types of items that can act as displays of open plurality. While virtually all address formulas contain the second-person pronoun you, additional elements are needed for the plural marking of a noun phrase. The loss of the thou/ye distinction is a well-known development in the history of the English language (see, for example, Wales 1996): After the sixteenth century, the element you encodes both plural and singular of the second-person pronoun and both the t- and v-forms (see Brown and Gilman 1960), as well as other functions (for example, the non-deictic, impersonal you). Due to this historical merger, addresses with you have a cer-

348

Theresa Heyd

tain degree of in-built ambiguity between plural and singular, as Wales (2004) has remarked. As a consequence, speakers have had to find alternative ways of marking plural addresses in conjunction with you. One strategy that is not limited to the second-person pronoun but is also found with the third-person plural, they/them, is the addition of an indefinite pronoun that can express plurality such as all, most, and everyone. This yields premodified constructions of plurality, such as all/most/every one of you, as well as postmodified constructions (essentially you all). In addition, there are some cases where the plural-marking occurs beyond the boundaries of the noun phrase; examples are given in (6) (premodified) and (7) (postmodified): (6)

But for those who don’t repost it, I will still pray for you [. . . ]

(7)

I am sure you are all well aware of the free offerings made from Gap and the free gift certificates offered from Abercrombie and Fitch [. . . ] (Email hoax corpus)

In light of the gap in the English pronoun paradigm, it is not surprising that alternative forms have been evolving that would allow a more elegant form of open plurality than the ones described so far. Indeed, the emergence of secondperson plural pronoun forms is one of the more active sites of ongoing language change in current English (see Kortmann 2006). Among the best-known of these forms are items that are highly marked as oral and regional, such as youse, yinz, and y’all, the last of which is still stigmatized as a Southernism (although increasingly less so, see Tillery, Wikle and Bailey 2000). The other notable forms in use are constructions of the type you + nounpl , such as you lot, you folks, you people or you guys. In particular, you guys is increasingly becoming a standard in North American varieties of English. Finally, an exception is to be made for vocative constructions where you can be deleted – thus vocatives allow for bare plural nouns as address formulas such as friends, people, guys, and so on. In sum, these are the most prevalent strategies for encoding open plurality. Their occurrence in digital folklore is summarized below. 3.1.3. Strategies of plural address in the email hoax corpus The email hoax corpus contains 147 individual texts, in other words, it contains tangibly different email hoaxes. Due to the viral nature of the transmission of existing email hoaxes and the evolution of new ones, there is a degree of textual overlap in some items (where, for example, new email hoaxes are created by virtue of copying and pasting sections of existing items). The corpus contains a total of 32,496 words at a 4,234 word vocabulary. For a first approximation of

I know you guys hate forwards: Address pronouns in digital folklore

349

how the 1-to-n nature of the genre is handled in these texts, the message openings can be analyzed. The greeting formulas used in openings are, by definition, vocatives, since they are constituted by noun phrases that are not syntactically incorporated into the ensuing sentence. Greeting formulas also lend themselves to analysis since they are structurally distinct entities within an email message. Of the 147 items in the corpus, only 44 (30%) contain a distinct opening. This relatively low quota is not unusual for email as a sociotechnical mode: In Gains’ study of text features of email, 37% of academic emails and 92% of commercial emails contained no opening formula (Gains 1999: 85, 91). The opening formulas that occur in the corpus are listed in Table 1; they are subdivided according to their plurality strategy. Table 1. Opening formulas in the email hoax corpus Plural Dear friends Dear all Hi everyone A big hello to you guys Attention friends Hello everyone Hello Disney fans Attention E-Mail Beta Test Participants Hello readers Dear AOL and IM users Hello all Hi all Happy Chanukah everyone

5 3 1 1 1 1 1

singular Dear customer Dear reader Dear friend Dear Hotmail user

1 2 1 1

non-specific Hi Hey Grace and peace Hello

13 1 1 5

1 1 1 1 1 1 19

5

20

As can be seen, neutral openings where the reference to the addressees is subdued are in the majority, particularly due to the frequent use of a simple and informal Hi as greeting. However, in the direct comparison between 1-to-1 simulation and open plurality, the tendency is very clear with 19 to 5 items. From a qualitative perspective, a striking diversity can be noted in the plural-marked items, whereas the singular items all adhere to the canonical greeting pattern dear + nounsg . By contrast, the plural items include constructions with dear, greeting words such as hi/hello, as well as more unusual addresses such as attention or the seasonal happy Chanukah. For plural marking, both indefinite pronouns (all/everyone)

350

Theresa Heyd

and plural nouns (friends/participants, etc.) are used; there is one occurrence of the quasi-pronoun you guys. In sum, there appears to be a preference for open plurality over 1-to-1 simulation in the openings of email hoaxes; in addition, a high degree of variation in linguistic choice was found. From this initial observation, the analysis can be extended to the actual text body of the messages. The corpus contains a total of 1,182 instances of secondperson pronouns, with 807 tokens of you and a remainder of 375 related tokens (such as yours/yourself, etc.). Analyzing them for their plurality strategy can only be done manually since there is no single morphosyntactic marker for it, as outlined above. Unsurprisingly, the large majority of these items is not explicitly marked for either singular or plural but uses the generic “all-purpose you” (Maynor 2000: 416), which can be employed for both singular and plural addresses (as well as other uses, such as impersonal you in the sense of a thirdperson pronoun or in collocations such as you know). With five instances, the number of explicitly singular-marked items is extremely low; two examples are given in (8) and (9): (8)

Please, if you are a kind person, send this on.

(9)

When you forward this e-mail to friends, Microsoft can and will track it ( If you are a Microsoft Windows user) For a two weeks time period. (Email hoax corpus)

By contrast, the ratio of openly plural-marked items in the corpus is considerably higher, at a total of 40 instances. These items are listed in the tables below. Table 2 shows the use of pre- and postmodified forms of you; Table 3 summarizes those constructions that can arguably be treated as quasi-pronouns. In Table 2, there are a majority of premodified, versus postmodified, constructions, with a ratio of 15 to 9 occurrences. All postmodified structures are tokens of you all, with the exception of one item where the plural-marking transcends the boundary of the noun phrase (you are all). Similarly, most of the premodified items follow the pattern of those/most/all/some of you; three of them contain arguably singular constructions, constructions which, however, imply a multiple audience, such as each one/each and every one of you. One item constitutes an exception in that the plural-marking occurs in the preceding relative clause. Finally, the last example (to all you Champagne lovers) is a vocative that is both pre- and postmodified. The items in Table 3 are of particular interest as they reflect a trend in the current development of English: They fall into the class of quasi-pronouns, or pronominal surrogates. With 12 tokens in the corpus, there is a clear tendency for the use of (you) guys. As outlined in the overview on encoding strategies,

I know you guys hate forwards: Address pronouns in digital folklore

351

Table 2. Pre-and postmodified forms of you in the email hoax corpus premodified

postmodified

is: DID IT WORK? Most of you who tried it will probably on this list. For those of you who send this along, I to pipe up, I’ll blame all of you!"" And with that she Who send this along, I thank you so much, but for those who my secratary will send all of you my screenname. I go/ some joke./ And for those of you who care please send this , Discovery Zone. . . Some of you might not be parents, but that each and every one of you please forward this on Gary "Nick" Hogman, Some of you receiving know me, some do Gary "Nick" Hogman, Some of you receiving know me, some do and may God bless each one of you who cares enough to pray I urge each and every one of you to pass this on to as many you, Tim PS: For those of you who dont take 5 minutes to is reposted. For those of you who repost, I thank you so [email protected]/To all you Champagne Lovers. . . . Send His message on. ""I am asking you all, begging you to You will, but I needed to let you all know. I love you all to let you all know. I love you all and I felt I needed to THEN FORWARD/I am asking you all, begging you to please THEN FORWARD/I am asking you all, begging you to please Robert Crensman. I am sure you are all well aware of the HEN FORWARD/I am asking you all, begging you to please was a hoax but I sent it to you all and ten(10) days later For our survival!! God bless you all!/Sincerely,/Debbie

15

9

the bare guys can be used in vocatives, often with a preceding item (such as hey or sorry). You guys occurs in the data both as a vocative (for example, Ok you guys) and intrasentential in subject and object positions (for example, I know you guys hate forwards and A big hello to you guys). By contrast, other typical constructions of the pattern you + nounpl are virtually absent from the corpus. There is just one case of you folks that occurs in an embedded narrative (and thus does not strictly constitute an address to the receivers of the message). It is not so surprising that other collocations of this pattern do not occur: Some of them are distinctly regional (for example, you lot for British varieties); most other forms, such as you people, have a clear negative or distancing connotation. Finally, there are three examples of ya’ll (alternative spellings such as yall or y’all do not occur). This may surprise since ya’ll is still being perceived as a highly marked – and stigmatized – Southernism. It cannot be determined whether its occurrence in these data is a sign of the ongoing “nationalization” of ya’ll (Tillery, Wikle and Bailey 2000), or whether it is being used as a marker

352

Theresa Heyd

Table 3. Quasi-pronouns in the email hoax corpus you guys me saying this. . . )/I know you guys hate forwards, but I minutes."/"A big hello to you guys, forward this to at least (1).jpg)"/"All of you guys who dont even try to send 95 and 98./ I know you guys hate forwards. But I a chain letter./ Ok you guys. . . . This isn’t a chain least 5 or 6./ Come on you guys. . . . And if you’re too to answer any questions you guys might have./Intel and AOL guys the Internet."/"Subj: sorry guys about this virus!/I us at: www.applebees.com/Hey guys,/DONT DELETE THIS EMAIL/ Founder of Cracker Barrel/Hey guys,/DONT DELETE THIS EMAIL/ of Victoria’s Secret/Hey guys, DONT DELETE THIS EMAIL Christine Schmidt/ GUYS. . . . PLEASE, PLEASE TAKE you folks I just said, ""Okay, you folks got my $250, and now I’m ya’ll just your time. I know that ya’ll will/ impress me !!!! / impress me !!!! I love ya’ll!!!!!/ Another that I am / I/ love/ ya’ll!!/ Love,/

7

5

1 3

of irony or humor here. Finally, it should be noted that the cases of you all in Table 2 might be interpreted as instances of ya’ll; since this is still an item of predominantly spoken use, there is no canonical lexical form yet. Nevertheless, it can be regarded as unlikely that these postmodified cases are intended as cases of ya’ll: In addition to being a highly marked stereotype, the origin of ya’ll is disputed and has been traced to constructions other than you all, such as the Hiberno-English ye aw (Bernstein 2003). It is difficult to compare the rate of open plurality in digital folklore against other genres, since the manual counting of plural-marked items is not viable for the size of the large reference corpora. However, initial comparative studies on the frequency of (you) guys in varying genres (data not shown) suggest that the occurrence of (you) guys in digital folklore indeed takes up an intermediate position between the high frequencies of explicitly oral genres (for example, TV sitcoms or the MICASE corpus) and the low frequencies in explicitly literate genres (for example, the Time Magazine corpus). In this sense, digital folklore as a digital genre can be seen as a bridging form that allows for the use of orality markers in a written environment. However, digital folklore can be expected to take an intermediate position between the more explicitly spoken and written genres. It should also be noted that the items presented here are based on a corpus of core messages. Due to the repeated forwarding process, digital folklore items tend to accumulate a cluster of framework messages at the periphery of the original core message

I know you guys hate forwards: Address pronouns in digital folklore

353

that typically contain greetings, explanations, apologies, and other forms of face management. Initial studies (Heyd 2008) have shown that the ratio of open plurality distinctly increases if these peripheral structures of digital folklore are taken into account. It should have become clear that there are solid functional motivations for the use of open plurality in digital folklore. As was outlined in section 2, digital folklore is, by definition, a 1-to-n genre that exploits the technicality of its CMC environment, namely the possibility of storing, forwarding and multipleaddressing messages. The use of plural address forms can be considered an effect in syntactic variation that is closely linked to these genre-based parameters. At the same time, psychosocial factors are clearly involved in the interaction with digital folklore. Items of digital folklore provide their forwarders with social capital since they are narrative ready-mades and thus endowed with a certain prestige (see Heyd 2008); at the same time, it is frequently reported that email users are – at least superficially – annoyed with receiving digital folklore (see, for example, the survey data in Kibby 2005: 786–788). The effect of solidarity and inclusion that can be created through the use of open plurality can therefore be used to mitigate potential averse reactions to the circulation of such messages. This can be seen particularly well in the co-text of some cases of (you) guys cited above, such as Hey guys, DONT DELETE THIS EMAIL, or I know you guys hate forwards. But I started this a month ago because I was very short on cash. In sum, open plurality is a morphosyntactic factor whose use is prompted and fostered through its specific genre environment.

4. Conclusion: Genre change – language change? Where syntactic variation is put into relation with genre variability and genre change, it is tempting to extend the analysis to the next step: How may genre change be involved with, or actually prompt, language change? Such a connection is particularly often claimed, or at least implied, with regard to CMC genres. These claims come both from the CMC community where, perhaps understandably, both researchers and users are overwhelmed by the sheer breadth and speed of innovation in digital discourse, and even more frequently, they come from prescriptivist concerns of laypeople about the proper use of English and the fear of linguistic decay (who thus carry on the complaint tradition into the digital age). The prediction, or fear, that linguistic aspects of CMC usage, such as lowered orthography standards and abbreviations or lexicalized oral forms, are spreading out into non-digital forms, has yet to be succinctly proven.

354

Theresa Heyd

The study presented here certainly provides no fuel for such bold claims. However, more modest assertions can certainly be put forward based on case studies as the one presented here. For one, digital genres provide an easy entry point into the written domain for items that are new, oral, or non-standard. Due to their intermediate status, digital genres such as digital folklore can therefore ease the integration of new items into the lexicon, as the transition from hitherto purely oral forms into written discourse is a major step in grammaticalization. In the case of plurality in digital folklore, this mechanism can be witnessed with regards to quasi-pronouns, like ya’ll, which were firmly restricted to the oral discourse universe until recently. The same point has been made for interjections such as awwa or meh (Yagoda 2007) or, in German, the rise of inflective constructions such as g¨ahn or knuddel (see, for example, Schlobinski 2001), which, as Yagoda (2007: 2) puts it, have been given “a major shot in the arm” through the advent of hybrid genres in the digital environment. In this sense, genre change can be at least a mitigating factor in language change. Finally, if it remains disputable whether digital genre variation can act as a motor to language change, the study presented here has shown that CMC is a great locus for examining linguistic innovation. From a methodological perspective, CMC genres provide easily available datasets; in this sense, digital genres are a fascinating object of study not only for research into CMC, but for studies in syntactic variation and genre quite generally.

References Aarts, Bas 2006

Conceptions of categorization in the history of linguistics. Language Sciences 28: 361–385. Askehave, Inger and Anne Ellerup Nielsen 2005 Digital genres: A challenge to traditional genre theory. Information Technology & People 18: 120–141. Axelson, Elizabeth 2007 Vocatives: A double-edged strategy in intercultural discourse among graduate students. Pragmatics 17: 95–122. Barron, Anne 2006 Understanding Spam: A macro-textual analysis. Journal of Pragmatics 38: 880–904. Bazerman, Charles 1994 Systems of genre and the enactment of social intentions. In: Aviva Friedman and Peter Medway (eds.), Genre and the New Rhetoric, 79–99.

I know you guys hate forwards: Address pronouns in digital folklore

355

(Critical Perspectives on Literacy and Education.) London: Taylor & Francis. Bernstein, Cynthia 2003 Grammatical features of Southern speech: yall, might could, and fixin to. In: Stephen Nagle and Sarah Danders (eds.), English in the Southern United States, 106–118. (Studies in English Language.) Cambridge: Cambridge University Press. Biber, Douglas 1988 Variation across Speech and Writing. Cambridge: Cambridge University Press. Blommaert, Jan and Tope Omoniyi 2006 Email fraud: Language, technology and the indexicals of globalisation. Social Semiotics 16: 573–605. Brown, Roger and Albert Gilman 1960 The pronouns of power and solidarity. In: Thomas Sebeok (ed.), Style in Language, 253–276. (The MIT Paperback Series 59.) Cambridge, MA: MIT Press. Crowston, Kevin and Marie Williams 2000 Reproduced and emergent genres of communication on the World Wide Web. The Information Society 16: 201–215. Crystal, David and Derek Davy 1969 Investigating English Style. (Indiana University Studies in the History and Theory of Linguistics.) Bloomington, IN: Indiana University Press. Crystal, David 2001 Language and the Internet. Cambridge: Cambridge University Press. Devitt, Amy 1991 Intertextuality in tax accounting: generic, referential, and functional. In: Charles Bazerman and James G. Paradis (eds.), Textual Dynamics of the Professions: Historical and Contemporary Studies of Writing in Professional Communities, 336–357. (Rhetoric of the Human Sciences.) Madison: University of Wisconsin Press. Dundes, Alan and Carl R. Pagter 1992 Work Hard and You Shall be Rewarded: Urban Folklore from the Paperwork Empire. 2nd Edition. Detroit: Wayne State University Press. Erickson, Thomas 2000 Making sense of computer-mediated communication (CMC): Conversations as genres, CMC systems as genre ecologies. In: Proceedings of the 33rd Hawaii International Conference on System Sciences (HICSS-33), 3011. Los Alamitos, CA: IEEE Press. Fernback, Jan 2003 Legends on the net: An examination of computer-mediated communication as a locus of oral culture. New Media and Society 5: 29–45.

356

Theresa Heyd

Gains, Jonathan 1999 Electronic Mail – a new style of communication or just a new medium? An investigation into the text features of e-mail. English for Specific Purposes 18: 81–101. Giltrow, Janet this volume Genre as difference: The sociality of linguistic variation. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 29– 51. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Herring, Susan 2007 A faceted classification scheme for computer-mediated discourse. Language@Internet 4: Article 1. Retrieved from http://www.languageatinternet.de/articles/2007/761/index html. Herring, Susan 2004 Slouching toward the ordinary: current trends in computer-mediated communication. New Media & Society 6: 26–36. Heyd, Theresa 2008 Email Hoaxes: Form, Function, Genre Ecology. (Pragmatics & Beyond New Series 174.) Amsterdam/Philadelphia: John Benjamins. Kibby, Marjorie 2005 Email forwardables: Folklore in the age of the internet. New Media & Society 7: 770–790. Kortmann, Bernd 2006 Syntactic variation in English: A global perspective. In: Bas Aarts and April McMahon (eds.), The Handbook of English Linguistics, 603–624. (Blackwell Handbooks in Linguistics.) Oxford: Blackwell. Kwasnik, Barbara H. and Kevin Crowston 2005 Introduction to the special issue: Genres of digital documents. Information Technology and People 18: 76–88. Labov, William 1972 Language in the Inner City. Philadelphia: University of Pennsylvania Press. Leech, Geoffrey 1999 The distribution and function of vocatives in American and British English conversation. In: Hilde Hasselg˚ard and Signe Oksefjell (eds.), Out of Corpora: Studies in Honour of Stig Johansson, 107–118. (Language and Computers 26.) Amsterdam: Rodopi. Lyons, John 1977 Semantics. Cambridge: Cambridge University Press. Maynor, Natalie 2000 Battle of the pronouns: y’all versus you-guys. American Speech 75: 416– 418.

I know you guys hate forwards: Address pronouns in digital folklore

357

Orlikowski, Wanda and JoAnne Yates 1994 Genre repertoire: Examining the structuring of communicative practices in organizations. Administrative Science Quarterly 39: 541–574. Puschmann, Cornelius this volume Thank you for thinking we could: Use and function of interpersonal pronouns in corporate web logs. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 167–191. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Ratliff, John 2007. Armchair activism. [Website]. Retrieved from http://www.breakthechain. org/armchair.html (accessed March 22, 2010). Schlobinski, Peter 2001 *knuddel zurueckknuddel dich ganzdollknuddel*. Inflektive und Inflektivkonstruktionen im Deutschen. Zeitschrift f¨ur germanistische Linguistik 29: 192–218. Tillery, Jan, Tom Wikle and Guy Bailey 2000 The nationalization of a Southernism. Journal of English Linguistics 28: 280–294. Virtanen, Tuija this volume Variation across texts and discourses: Theoretical and methodological perspectives on text type and genre. In: Heidrun Dorgeloh and Anja Wanner (eds.), Syntactic Variation and Genre, 53–84. (Topics in English Linguistics 70.) Berlin/New York: Mouton de Gruyter. Wales, Katie 2004 Second person pronouns in contemporary English: The end of a story or just the beginning? Franco-British Studies 33: 172–185. Wales, Katie 1996 Personal Pronouns in Present-Day English. (Studies in English Language.) Cambridge: Cambridge University Press. Yagoda, Ben 2007 Pardon the interjection: The Internet and the rise of awwa, meh, feh, and heh. Slate Magazine [Website]. Retrieved from www.slate.com/id/2159929/ (accessed March 22, 2010). Yates, Simeon 1996 Oral and written linguistics aspects of computer conferencing: A corpus based study. In: Susan Herring (ed.), Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural Perspectives, 29–46. (Pragmatics & Beyond 39.) Amsterdam/Philadelphia: John Benjamins. Zitzen, Michaela and Dieter Stein 2004 Chat and conversation: A case of transmedial stability? Linguistics 42: 983–1021.

358

Theresa Heyd

Zwicky, Arnold 1974 Hey, Whatsyourname! In: M.W. LaGaly, R.A. Fox and A. Bruck (eds.), Papers from the Tenth Regional Meeting of the Chicago Linguistic Society, 787–801. Chicago: Chicago Linguistic Society.

List of contributors

Heidrun Dorgeloh ([email protected]) Heinrich-Heine-Universit¨at D¨usseldorf Anglistik 3 Universitätsstraße 1 D-40225 D¨usseldorf Germany

Theresa Heyd ([email protected]) University of Pennsylvania Department of Linguistics/LDC 3600 Market Street #810 Philadelphia, PA 19104 USA

Janet Giltrow ([email protected]) Department of English University of Britisch Columbia 397-1873 East Mall (Buchanan Tower) Vancouver, BC V6T 1Z1 Canada

Wolfgang Imo (wolfi[email protected]) Westf¨alische Wilhelms-Universit¨at M¨unster Germanistisches Institut Hindenburgplatz 34 D-48143 M¨unster Germany

Maurizio Gotti ([email protected]) Universit`a di Bergamo Piazza Rosate 2 I-24129 Bergamo Italy

Johannes Kabatek ([email protected]) Romanisches Seminar Eberhard Karls Universit¨at T¨ubingen Wilhelmsstr. 50 D-72074 T¨ubingen Germany

Susanne G¨unthner ([email protected]) Germanistisches Institut/ Sprachwissenschaft Westf¨alische Wilhelms-Universit¨at M¨unster D-48143 M¨unster Germany

Rolf Kreyer ([email protected]) Institut f¨ur Anglistik und Amerikanistik Philipps-Universit¨at Marburg Wilhelm-R¨opke Str. 6 D-35032 Marburg Germany

360

List of contributors

Susanne Jantos ([email protected]) Reichenhainer M¨uhlberg 92 D-09125 Chemnitz Germany Ana E. Mart´ınez-Insua ([email protected]) University of Vigo English Department, FFT Campus Universitario E-36310 Vigo Spain Britta Mondorf ([email protected]) Department of English and Linguistics Johannes Gutenberg-Universit¨at Jakob-Welder-Weg 18 D-55128 Mainz Germany Philipp Obrist ([email protected]) Romanisches Seminar Eberhard Karls Universit¨at T¨ubingen Wilhelmsstr. 50 D-72074 T¨ubingen Germany Javier P´erez-Guerra ([email protected]) University of Vigo English Department, FFT Campus Universitario E-36310 Vigo Spain

Cornelius Puschmann ([email protected]) Heinrich-Heine-Universit¨at D¨usseldorf Anglistik 3 Universit¨atsstraße 1 D-40225 D¨usseldorf Germany Valentina Vincis ([email protected]) Romanisches Seminar Eberhard Karls Universit¨at T¨ubingen Wilhelmsstr. 50 D-72074 T¨ubingen Germany Tuija Virtanen (tuija.virtanen@abo.fi) Åbo Akademi University Department of English Language and Literature Fabriksgatan 2 ˚ bo FL-20500 A Finland Anja Wanner ([email protected]) University of Wisconsin-Madison Department of English 600 North Park Street Madison, WI 53706 USA

Index

9/11 341 academic writing 43, 46, 71, 179, 180, 188, 284 acrolect 307 active voice 4, 13, 104–105 address form 179, 181–187, 343–348 adverbial 113, 115–118, 148, 150 adverbial placement 17, 53, 55, 58, 62, 65– 66, 278, 285, 287 aggregation 252–253 agreement 4, 18, 305–306, 315–317 animacy 6, 120–121 aphorism 87 ARCHER 16, 111, 113, 134, 221, 240 author-centered rhetoric 107 be 307, 309–310, 312–314, 321–326 biblical text 245, 247, 249, 258–261, 268 blog 18, 39, 47, 75, 167–175, 179–182, 184–187 British National Corpus 15, 173–174, 221, 223–236, 277, 282, 284–285 British vs. American English 232–239 canonical situation (of utterance) 345 chat 142, 170–171, 196, 310, 344 class lesson 18, 305, 310–313, 315–329 clause linkage 18–19, 71, 247–249, 252–253, 255, 257–259, 261–262, 264, 266–269 closed set 38–39 cognitive (approach/ linguistics) 4, 112, 177, 334 cognitive (dimension of variation) 5–6, 13, 54–55, 77, 122–123, 176, 205, 220, 235, 252–253 coherence 61, 65, 266 collective noun 317–319, 327, 329 colloquialization 233–234, 239 communicative genre 17–18, 141–143, 146– 147, 149, 160–161, 195–201, 210 communicative problem 198, 283, 288, 300

communicative purpose 9, 19, 31, 48, 74, 86, 172, 221, 236 community of practice 48, 74, 85–86, 89 complexity – conceptual complexity 120 – lexical complexity 72, 130, 132, 134, – syntactic complexity 19, 72, 111, 115, 120, 125–126, 130–134, 235, 324 computer-mediated communication 4, 10, 18, 72, 80, 167–168, 174, 187–188, 333–337, 343, 345–346, 353–355 construction discourse 160–162 construction grammar 21, 141, 146, 160, 163 contextualization 72, 200, 202 contextualization cue 142, 150, 154–155, 157, 159 contraction 317, 321 convention 14, 30, 37–39, 41, 58, 66–67, 71, 73, 106, 169, 171, 188, 197, 323 conversation 8, 15, 94, 147–148, 151–153, 155–160, 171–172, 175–177, 180, 199, 207, 311, 346 conversation analysis 144, 146, 196 Cooperative Principle 342 coordination 103, 252, 255, 257, 259, 262– 263 copula 103, 309, 312–313, 323 corporate blog 18, 39, 47, 172–175, 180– 182, 187–188 corpus linguistics 4, 10, 15–16, 18–19, 31– 32, 34–35, 42, 59, 113–114, 167–168, 225–226, 230–233, 247–249, 255–256, 268–269, 278, 342–344, 348–352 creole continuum 307, 319 creole influence 305, 308, 327–328 deictic center 171–172, 175, 177, 181, 187 dense constructions 200, 203–205 density 111, 125, 131, 206, 255, 257, 323– 324 deontic modality 10, 17, 43–46

362

Index

description 31–32, 35, 47, 59–62, 64–66, 68, 73, 76, 86–88, 90, 105, 254, 265, 279–280, 288–290 descriptive text type 17, 60, 66, 76 deverbal noun 103 diachronic variation 131, 250, 254 dialect 7, 120, 176, 201, 250, 306, 316, 321, 327 dialogue 57, 88, 95, 107, 171, 176, 181, 259, 306, 311 diary 75, 95, 114, 168–169 digital folklore 13, 333–345, 348, 352–354 digital genre 168, 335–336, 344–346, 354 direct object 4, 6, 111, 113, 115–118, 129– 130, 132–134, 208, 279, 283 discourse tradition 145, 247–253, 255, 258, 267–269 discourse type 1, 17, 38, 53–59, 62–63, 65–66, 68–69, 71, 74–77 displacement 277 don’t-simplification 316, 320–321, 327–328 economy of discourse 95 efficiency 6, 14, 39, 125, 127, 176, 221 email 74, 184–185, 335, 337–338, 340–341, 343–344, 347–349, 352 email hoax 17, 333, 342–344, 347–352, 356 encyclopedic article 10, 17, 55, 61, 65, 68 essay 88, 106, 238, 267, 284, 308–309 evidentiality 101 existential construction 278, 310, 324–326 experimental report 97, 105, 107 experimental essay 17, 59, 71, 85–91, 93, 95, 97, 99, 102–104, 106–107 expository text 59, 64, 66 eyewitness perspective 280, 289, 300, 302 Face Threatening Act 44, 206, 209–210 face-to-face-communication 70, 73, 75, 171, 175, 177, 180, 346 feature 5, 7, 15–17, 30–35, 39, 42, 44, 88, 94–97, 112, 114, 142, 180, 305, 307– 309, 312, 319–321, 323, 335 fiction 15, 18, 60, 75–76, 181, 225, 237, 235, 277–278, 284–285, 288, 300, 310 field deontics 44–45 first-person pronoun 101, 204 form – complexity of form 32

– form and function 2–3, 6, 11, 17, 19, 37, 54, 70, 75, 77, 210, 219, 236, 335 – formal regularities 29, 37–38, 40 – formal variation 10, 30, 32, 38 – relevance of form 1–3, 8–11, 16–18, 29–30, 33–43 formality 4, 6, 18–19, 29–30, 36, 39, 111, 148, 305–311, 313–315, 317–318, 321– 328, 347 function – function of language 5, 53, 58, 68 – function of genre 9–10, 29, 37–38 – function of text/discourse 9, 31, 56–60, 68, 107, 258, 266–267, 277, 297, 300 – functional linguistics 4, 6–7, 11, 14–15, 31 generality 7, 10, 17, 29–30, 32–34, 38–39, 43–44, 46–47 generative grammar 4, 11–13 genre – generic productivity 106 – genre as social action 1, 34, 54, 70, 198 – genre as form 9 – genre feature 10, 36–37, 86, 90, 106– 107, 142–146, 154, 196, 198–199, 232, 236, 255, 268–269, 326–328, 337–338, 345 – genre, theory of 1, 40, 42, 47 – genre, emergence and change 1–2, 14, 17, 35, 47, 53, 70–72, 74, 85–87, 145, 198, 333, 335–336, 338, 343, 354 – genre differentiation 34, 43, 47, 248–249, 268–269, 338 – genre ecology 333, 335–336, 343–344 – genre hybridization (also: merging) 17– 18, 141, 158 grammaticalization 21, 222, 229, 354 hedging 8, 67, 101 hypercorrection 315–316, 327 humorous writing 225, 237–240, 338–340 ICE-Jamaica 16, 305–307, 310–311, 326, 328 iconicity 61, 278, 287, 300 immediate-observer effect 277–278, 280, 285, 292, 298–300 impersonal construction 350 indefinite pronoun 318–319, 327, 349

Index infinite construction 18, 195–196, 200–206, 210 inflection 305–307, 315, 326–328 information status 63, 277, 293–294, 296, 298–300 – given and new information 4, 13, 61, 205, 294, 317 information structure 6, 13, 177 integration 252, 254–255, 257–259 interaction 10, 18, 33, 38, 73, 142, 144, 148, 151, 153, 156, 161, 171, 179, 196– 197, 199, 341 interactional linguistics 196 interdiscursivity 106 intergenre model 70–71, 343 interjection 154, 354 Internet (also: World Wide Web) 17, 54, 72, 76, 167, 171–172, 333, 336–337, 341, 345 intraspeaker variation 4, 7–10, 16 inversion 5, 266, 277–278, 282–283, 285, 287, 289, 294, 298 joke 75, 195–198, 338 journalistic text 71, 237, 262, 266–267 junction 249, 252–261, 263, 265, 268 language change 10, 54, 253, 307, 336, 344–345, 348, 353–354 language processing 6, 13–14, 111, 115, 119, 121, 124, 134, 220–221 legal text 19, 179, 247–249, 253–255, 258 letter 19, 42, 59, 62, 71–72, 107, 111, 113– 116, 118, 128–130, 133–134, 171, 180 lexicon 7–8, 162, 210, 250, 354 literary theory 33, 75, 334 loan word 99 locative preposing 3, 18, 65, 277–278, 282, 285–287, 295–297 marker (of genre) 4, 8, 61, 167, 348 marker (of syntactic phrase) 123–128 medium 15, 18, 219, 225, 230–232, 234– 237, 240, 333, 336, 338, 345–346 metaphor 98, 172 MICASE 347, 352 modal 8, 44–46, 101, 206–207, 312 narrative 17–18, 31–32, 35, 38, 54–60, 62– 65, 68, 75–76, 85, 97, 114, 169, 196,

363

201–206, 258–259, 265, 277–279, 284, 287–288, 337, 351 new-rhetoric(al) 33–34, 36, 42 news writing 71–75, 113–118, 128–130, 134, 169–170, , 180, 230, 249, 261–269, 284, 325, 327 (broadcast) news 18, 305, 310–313 nominalization 8, 17, 85–86, 103–104, 107 non-agreement 18, 305–310, 312–317, 319– 328 objectivity 100, 179 officelore 336–337 oral vs. written language 126, 168, 170, 175, 180, 234 particle 4, 13, 18, 219, 221, 225–226, 229– 234, 238–239 passive voice 13, 48, 104–105, 227 perception 6, 101, 278–279, 281–283, 287– 292, 300 person deixis 168, 177, 345 perspicuity 97 persuasion 33, 57, 71, 264 phone-in show 18, 141, 145, 148 plurality 4, 305, 333, 344–353 polarity 309 politeness 6, 94, 148, 251, 346–347 post-modification 103 predictability 36–38, 40, 143 pre-modification 86, 103–104, 107 pronoun 4, 18, 47, 101, 151, 167, 173–176, 181–183, 187–188, 201, 206–208, 296, 309, 317–318, 327–328, 333, 348–350; see also: first-person pronoun and secondperson pronoun prose fiction 12, 18, 41, 225, 277–278, 284– 285, 288, 300 prototype theory 334 proximity 307–309, 315–317, 322–323, 326– 328 qualitative (methodology) 15–17, 31, 36, 87, 249, 336, 349 quantitative (methodology) 15–16, 19, 37, 87, 122, 130, 249, 336 question 150, 156, 161, 195–196, 206, 208– 210, 312, 316, 339

364

Index

radio counseling 141–142, 161–162 radio talk 142, 145–147, 155, 160, 162 referential choice 17, 53, 55, 58, 62 reflexive 18, 219, 222–223, 228–229, 232, 234, 239 register 2, 7, 10, 15–16, 18, 29–32, 34, 39, 41, 53, 55, 75, 168, 305–308, 314, 318– 320, 322, 326 reproach 75, 144, 152, 157–158, 198, 206, 209 research practice 86–87 rhetoric – object-centred rhetoric 107 – rhetorical device 98 – rhetorical studies/ theories 1, 9, 29–30, 33–40, 42, 107, 183 routinization 195, 197 Royal Society 89–92, 94, 96, 99, 107 scientific writing 14, 31, 44, 59, 87, 95, 98, 102–103, 106–107, 237–238, 240, 345 second-person pronouns 18, 151, 167–168, 174, 176, 313, 344, 346–348 sedimentation 145, 154, 195, 197, 200, 206, 210 simplification 306, 316, 320–321, 327–328 situational rhetoric 37–38, 383 social action 1, 9–10, 29, 34–36, 39–42, 53–54, 69–70, 76, 195, 198 Social Constructivism 197–200 social media 172, 333, 337 sociolinguistics 1, 7, 25, 220, 307, 336 solidarity 106, 178, 353 Southernism 348, 351 space (representation of) 3, 18, 175, 277, 279–281, 288–292, 300 spoken language 1, 10, 14, 114, 130, 146, 160, 171, 175, 181, 195, 230, 234, 305, 307–311, 323, 336, 352 stance 30, 33, 43, 66, 94, 109, 203 storytelling 75–76, 144, 200, 206, 345 style 1–2, 9–10, 29–32, 53, 55, 75, 93–98, 169–171, 234, 258, 262, 264–266, 310– 311

subgenre 15, 68, 173, 334, 339, 342–343 subject 6, 12, 53, 55, 62, 104, 111–118, 123, 127–129, 134, 183, 205, 278, 282– 283, 296, 305, 309, 312, 321, 322, 328, 347 subjectivity 118, 280, 289 subordination 8, 103, 119, 255, 257, 262– 263 substrate 306–308 synchronic variation 19, 221, 225, 230, 247, 249–250, 266 syntactic feature 8, 102–104, 162 Systemic Functional Linguistics 31, 122 tense 53, 62–63, 65, 196, 248, 266, 307, 309, 312–313, 321–322 terminology 11, 87, 91, 99–100 text segmentation 61, 65, 300 text strategy 56–58, 61–66, 68, 74, 300 text type 2, 9–11, 17, 31–32, 38, 53–61, 69, 74–77, 111–112, 114, 220–221, 230, 236, 254, 285, 305–306, 335, 344 thematic position 104, 266 topic continuity 66, 266, 300 topic management 298, 300 transitivity 5, 183, 219, 222–225 treatise 88, 95 urban folklore 336, 343 verbless construction 283 verbosity 96 vocative 4, 347–351 way-construction 219, 222–224, 227–230, 237–240 web 2.0 335 weight 13, 63–64, 67, 118–120, 123, 134, 262 what-construction 200, 206–210 word order 4–5, 12–13, 118, 293 written language 10, 18, 73, 113–114, 130, 134, 168, 170, 175, 179–180, 234, 262