244 20 2MB
English Pages [506] Year 2019
Andrii Danylenko, Motoki Nomachi (Eds.) Slavic on the Language Map of Europe
Trends in Linguistics Studies and Monographs
Editors Chiara Gianollo Daniël Van Olmen
Editorial Board Walter Bisang Tine Breban Volker Gast Hans Henrich Hock Karen Lahousse Natalia Levshina Caterina Mauri Heiko Narrog Salvador Pons Niina Ning Zhang Amir Zeldes
Editor responsible for this volume Daniël Van Olmen
Volume 333
Slavic on the Language Map of Europe Historical and Areal-Typological Dimensions Edited by Andrii Danylenko, Motoki Nomachi
ISBN 978-3-11-063497-6 e-ISBN (PDF) 978-3-11-063922-3 e-ISBN (EPUB) 978-3-11-063517-1 Library of Congress Control Number: 2019941344 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2019 Walter de Gruyter GmbH, Berlin/Boston Typesetting: Integra Software Services Pvt. Ltd. Printing and binding: CPI books GmbH, Leck www.degruyter.com
Contents Contributors
VII
Andrii Danylenko Searching for a place of Slavic in Europe as a linguistic area
1
Part I: Issues in Methodology and Pre-History Björn Wiemer 1 Matrёška and areal clusters involving varieties of Slavic. On methodology and data treatment 21 Vít Boček 2 Common Slavic in the light of language contact and areal linguistics: Issues of methodology and the history of research 63 Robert Orr 3 Intertwining trees, eddies, and tentacles — some thoughts on linguistic relationships in Europe, mainly Slavic-non-Slavic 87
Part II: Slavic and Standard Average European Jadranka Gvozdanović 4 Standard Average European revisited in the light of Slavic evidence 113 Bridget Drinka 5 The perfects of Eastern “Standard Average European”: Byzantine Greek, Old Church Slavonic, and the role of roofing 145 Nataliya Levkovych, Lidia Federica Mazzitelli and Thomas Stolz 6 Slavic vis-à-vis Standard Average European: An areal-typological profiling on the morphosyntactic and phonological levels 187
VI
Contents
Paul Wexler 7 How Yiddish can recover covert Asianisms in Slavic, and Asianisms and Slavisms in German (prolegomena to a typology of Asian linguistic influences in Europe) 225
Part III: Slavic in Areal Groupings in Europe Helena Kurzová 8 Defining the Central European convergence area
261
George Thomas 9 Some morpho-syntactic features of the Slavic languages of the Danube Basin from a pan-European perspective 291 Andrey N. Sobolev 10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor 315 Andrii Danylenko 11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
347
Walter Breu 12 Morphosyntactic changes in Slavic micro-languages: The case of Molise Slavic in total language contact 385 Bernd Heine 13 On formulas of equivalence in grammaticalization: An example from Molise Slavic 433 Motoki Nomachi 14 Placing Kashubian on the language map of Europe Index of subjects Index of languages
491 495
453
Contributors Vít Boček Czech Academy of Sciences Department of Etymology Veveří 97, 60200 Brno Czech Republic [email protected] Walter Breu University of Konstanz Department of Linguistics Slavistik, Fach 179 78457 Konstanz Germany [email protected] Andrii Danylenko Pace University Dyson College of Arts and Sciences Dept. of Modern Languages and Cultures 41 Park Row, New York, NY 10038 USA [email protected] Bridget Drinka University of Texas at San Antonio Dept. of English Main Building, 1 UTSA Circle, San Antonio, TX 78249-1644 USA [email protected] Jadranka Gvozdanović Heidelberg University The Slavic Institute Schulgasse 6, 69117 Heidelberg Germany [email protected] Bernd Heine University of Cologne Institute for African Studies and Egyptology
https://doi.org/10.1515/9783110639223-201
Nonnenwerthstr. 48 50937 Köln Germany [email protected] Helena Kurzová Czech Academy of Sciences Institute of Philosophy Jilská 1, 11000, Praha 1 Czech Republic [email protected] Nataliya Levkovych Bremen University Faculty 10: Linguistics and Literary Studies Universitäts-Boulevard 13, Gebäude GW 2, 28359 Bremen Germany [email protected] Lidia Federica Mazzitelli University of Cologne, Slavic Institute Weyertal 137, 50931 Köln Germany [email protected] Motoki Nomachi Hokkaido University Slavic-Eurasian Research Center Kita 9, Nishi 7, Kita-ku, Sapporo, 060-0809 Japan [email protected] Robert Orr Independent Scholar 2044 Arrowsmith Dr. #104C Ottawa, ON, K1J 7V8 Canada [email protected] Andrej N. Sobolev Russian Academy of Sciences Institute for Linguistic Studies
VIII
Contributors
Tuchkov pereulok 9, 199053, St. Petersburg Russia [email protected] Thomas Stolz Bremen University Faculty 10: Linguistics and Literary Studies Universitäts-Boulevard 13, Gebäude GW 2, 28359 Bremen Germany [email protected] George Thomas McMaster University Dept. of Linguistics and Languages 1280 Main Street West, Hamilton, Ontario, L8S 4L8
Canada [email protected] Paul Wexler Tel-Aviv University Dept. of Linguistics Webb, 405, Ramat Aviv, 69978, Tel Aviv Israel [email protected] Björn Wiemer Johannes Gutenberg University Mainz Institute for Slavic, Turkic and CircumBaltic Studies Jakob-Welder-Weg 18, 55099 Mainz Germany [email protected]
Andrii Danylenko
Searching for a place of Slavic in Europe as a linguistic area 1 Introduction This book is a collection of chapters largely based on the papers delivered at the International Symposium, Slavic on the Language Map of Europe, which took place on August 8–11, 2013 at the Slavic-Eurasian Research Center of Hokkaido University in Sapporo (Japan). Inspired by vivid discussions during the symposium and numerous conversations with the participants in the aftermath of this event, the editors came up with an idea of preparing a comprehensive volume focusing exclusively on the areal-typological and historical sampling of Slavic as compared with other, primarily neighboring, languages used in Europe. There were several reasons behind the initial intention to bring together scholars working in areal and historical linguistics. As recent research shows, Europe can be safely conceived as a linguistic area of some kind (Bernini and Ramat 1996; van der Auwera 1998a; König and Haspelmath 1999; Heine and Kuteva 2006: 1–47; Nomachi and Heine 2011a; Drinka 2017). This unity was labeled “Standard Average European” (SAE) by Haspelmath (1998); going back to Whorf (1941), this term is used by other scholars (e.g., Garvin 1949; Déscy 1973: 29; van der Auwera 1998a; see Heine in this volume). One should also bear in mind the success of the EUROTYP project. Sponsored by the European Science Foundation, the project resulted in an unprecedented amount of typological information on the languages of Europe. Terms such as “Eurolinguistics”, “Euroversals”, “Euro-Lingua” and the like became commonplace in the studies dealing with the “promotion” of linguistic unity in Europe (Heine and Kuteva 2006: 7; also Reiter 1999), including some central European Slavic countries (Načeva-Marvanová 2007; Marvan 2008; Stachowski 2014). It comes therefore as no surprise that some scholars raise the question of whether such sociopolitical goals, or implicit ideological orientation, might have had bearing on the linguistic results obtained within the EUROTYP project (Heine and Kuteva 2006: 7, 9, fn. 5), which could affect the methodology and approach toward the study of the languages of Europe, first and foremost Slavic (Danylenko 2013). When the idea of organizing a symposium was first hatched in 2011, it was clear to the editors that the place of Slavic in the linguistic landscape of Europe had often been discussed controversially in scholarly literature. To adduce a few examples, some authors treat Slavic languages as more “marginal” https://doi.org/10.1515/9783110639223-001
2
Andrii Danylenko
European languages (e.g., Haarmann 1976: 123–127). Others argue that Slavic belongs to the core of European languages (e.g., Hock 1986: 508–509; Haspelmath 2001), while for Déscy (1973), Russian is a true SAE language like French, German, and Italian. What remains unequivocally accepted is that a number of Slavic languages have been and still are in close contact with languages of Western Europe, which is likely to influence linguistic patterning in Slavic. Moreover, the last years have brought about a rich literature viewing contact as a sine qua non of the areal diffusion of particular features, commonly accompanied by contact-induced grammaticalization, whose description involves either isogloss mapping or isopleth mapping (see van der Auwera 1998b). The volume is composed of an introduction and 14 chapters written by renowned scholars and leading specialists in various fields of linguistics from North America, Europe, Israel, and Japan. The authors deal with evidence from both “macro” and “micro” Slavic languages spoken in Europe, while introducing new empirical data and theoretical generalizations about language contact and grammaticalization in different Slavic languages and their near and distant neighbors. What is most commendable is the fact that some of the studies are premised on the results obtained from fieldwork and linguistic experiments. The book sheds light on some methodological and descriptive issues related to Slavic as viewed from the areal-typological and historical perspective (see Boček 2014, also in this volume). First, the chapters offer new insights into a number of theoretical issues, including language contact, grammaticalization, mechanisms of borrowing, the relationship between areal, genetic, and typological sampling, conservative features versus innovation, and socio-linguistic aspects of linguistic alliances conceived of both synchronically and diachronically. Second, the volume integrates new approaches towards the areal-typological profiling of Slavic as a member of several linguistic areas within Europe, including SAE (Haspelmath 2001; also Gvozdanović in this volume), the Balkan Sprachbund (Asenova [1989] 2002; also Sobolev in this volume) and Central European (CE) groupings(s) like the Danubian or Carpathian areas (Thomas 2008, also in this volume; Kurzová 1996, also in this volume), as well as the CarpathianBalkan linguistic macroarea (see Danylenko in this volume). The latter grouping is known to be primarily in the focus of an East Slavic linguistic enterprise launched long ago by some innovative studies written by Ukrainian and Russian scholars (e.g., Nimčuk 1993; Klepikova 2003). Third, some of the studies in this book concentrate on structural affinities between Slavic and other European languages that arose as a result of either grammatical replication or borrowing (see Heine and Kuteva 2005) or, to use
Searching for a place of Slavic in Europe as a linguistic area
3
the terminology of Matras and Sakel (2007), “pattern replication” or “borrowing of matter”. Special emphasis is placed in those chapters on contact-induced grammaticalization, especially in Slavic micro-languages such as Molise Slavic and Kashubian. When working on the volume, the editors focused on the relationship of the three key notions that essentially triggered the inception and subsequent realization of this project, to wit, language contact, grammaticalization, and areal grouping. Indeed, the intimate interrelation of these notions is more than obvious, a fact which is reflected in the structure of the proposed volume. Uncontroversial though the aforementioned connection might seem, some explanation of these notions is in order here.
2 Areal grouping To begin with, the discussion of areal groupings in Europe has recently displayed an obvious bias in favor of the Western European languages, a bias which appears even more pronounced in the treatment of SAE as a linguistic unity (Danylenko 2013: 135–137; also van Pottelberge 2001). Most of the Slavic languages have been literally pushed over to the backyard of political and linguistic Europe, thus being assigned a peripheral status in practically each of its linguistic areas. One may pose legitimate questions as to whether this peripheralization is a result of sociopolitical negligence or a lack of sufficient data of both genealogical and areal-typological sampling of Slavic. Leaving discussion of the former option for future research, especially in some adjacent fields of the humanities, one can agree with Heine and Kuteva (2006: 35) that, at least at the descriptive level, “we know much more about these [Romance and Germanic] languages than the languages of Eastern Europe”. I venture to claim that the situation, despite substantial progress in introducing Slavic in the context of linguistic Europe (see Wiemer 2011; Drinka 2017: 288–394), has not changed for the better. Suffice it to mention that, for instance, the non-standard varieties of Ukrainian, especially its more archaic western and northern dialects, and Belarusian, in particular its transitional Polessian dialects, are less researched in this sense. We can state that the sampling of North Slavic and some Belarusian dialects in the Circum-Baltic context (see Dahl and Koptevskaja-Tamm 2001; Seržant and Wiemer 2014), but divorced from its Ukrainian neighbor, can hardly produce an adequate areal-typological and historical picture of the languages used in the area around the Baltic Sea.
4
Andrii Danylenko
Placed in the context of areal typology and dynamic taxonomy, significant differences emerge among the European languages and within Slavic itself. East Slavic, for instance, appears to be a mere replica language which is capable of only borrowing surface structures or replicating deep structures on the model of SAE which presupposes the existence of the center and periphery of SAE. The overall situation with the propagation of SAE and its convergent features in areal linguistics today is reminiscent of the Eurocentrism of the entire enterprise of 19th-century comparative linguistics. Beginning with Wilhelm von Humboldt’s fascination with the “Sanskritic” (Indo-European) languages coming closest to the most perfect form, most of them believed that, taken together, the non-European category of languages was inferior to Indo-European. That was not linguistic relativism but linguistic absolutism; not surprisingly, Aarsleff (1988: x) called this position a kind of “incipient racism” in linguistics (see Danylenko 2016: 115–117). To use the same assessment framework, the default form is assigned today to the most advanced (core) languages of SAE (French, northern Italian dialects, German and Dutch) while the periphery Slavic languages, with rare exceptions (like Polish), characterized by a minimum of default features, appear inferior or “underdeveloped” in terms of the level of their convergence within Europe as a linguistic area. While regarding SAE as the European of Western Europe, van der Auwera (1998a: 817) suggested, and rightly so, that SAE should more appropriately be called “Standard Average Western European”. Haspelmath (2001: 1505) noted, however, that the latter term would be misleading inasmuch as the SAE Sprachbund in fact includes languages (e.g., Greek, Albanian, Latvian, and Russian) which would not really qualify as Western European languages (see Heine and Kuteva 2006: 27). Yet, admitting that the term SAWE is not perfect, van der Auwera (1998a: 824) advanced a theory of a “Charlemagne Sprachbund”, appropriating the name of the first ruler of an area where French, Italian, German and Dutch are today the core languages. Polish, as the only representative of the Slavic branch, is found very close to the center of this convergence area. The notion of the “Charlemagne Sprachbund” has been recently elaborated on by Drinka (2017: 144–168; also in this volume) as regards the diffusion of the periphrastic perfect in Europe. She argued that the existence of the Charlemagne Sprachbund is remarkably well supported by the distribution of the have and be auxiliaries in the perfects of Western Europe. As a result of the political and social ties established during the time of Charlemagne, innovative strengthening of the dyadic relationship between have and be tended to occur in the core area, with be periphrastic undergoing significant growth; in the peripheral areas, this
Searching for a place of Slavic in Europe as a linguistic area
5
growth did not occur. In sum, the borders of the Carolingian realm correspond, according to Drinka (2017: 167), closely to the boundaries of the have / be core. Leaving aside the nature of the latter correspondence for future research, one still needs to juxtapose the spread leading, in the theory of Drinka, to the replication of the periphrastic perfect in West Slavic and, via Low German in the Hanseatic league, in North Russian and some adjacent languages with other concurrent morphosyntactic and even phonological changes (Zubkova 2010: 481–494). One wonders in this respect: if neighboring languages show convergent features, does it mean that they are always reflexes of areal diffusion? Other models besides that of wave diffusion in the spirit of Johannes Schmidt show promise. Suffice it to mention here the “triangulation” approach of Wiemer, Seržant and Erker (2014: 25) according to which our understanding of structural convergence in different kinds of linguistic area has to systematically reckon with three kinds of factor, namely: whether the features encountered are (1) inherited from common ancestors, (2) typologically frequent, (3) contact-induced. At first blush, the introduction of the notion of “contact superposition zone” by Koptevskaja-Tamm and Wälchli (2001: 728) for the Circum-Baltic area could help rectify the aforementioned situation. In fact, this notion implies that there has never been a political unit that covered the whole area. Many similarities can be explained as buffer zone phenomena between two other larger areas (Western Europe or SAE and Central Eurasia) (Wiemer and Wälchli 2012: 15). Yet, this notion does not change the overall pattern of the diffusion of particular features crossing both sociopolitical and linguistic boundaries. Without accepting the historical dimension of linguistic changes as a complementary aspect of the areal-typological profiling of languages found in contact, the aforementioned notion does not look methodologically viable. Quite in the same vein, Drinka (2017: 43) has recently argued that the twodimensional image of a Sprachbund like SAE or the Circum-Baltic area only succeeds in presenting the end product of some complex changes, in synchronic fashion, but does not account for the sources of these similarities. This is why only a three-dimensional, chronologically stratified model, according to her, can adequately represent such a development, one that recognizes the essential role of contact across a Sprachbund but that also emphasizes the layered nature of that contact across time and space. The image of a dynamic Sprachbund is reformulated by Drinka (2017: 349) as a “stratified convergence zone” which allows her to account for the shared, stratified histories of the perfects and resultatives in -n/-t and in *-ŭes in the contiguous varieties of Baltic, Slavic, and Finno-Ugric, sometimes in contact with other languages, as well. Yet, synthetic though this model seems, it can still be refined by considering the nature of
6
Andrii Danylenko
driving forces behind grammatical changes in low- or high-contact languages (Mel’nikov 2003: 107, 138–141; Trudgill 2010: 300, 011), especially in Slavic (Danylenko 2015).
3 Language contact and grammaticalization If viewed from the point of view of the interrelation between language contact and grammatical change (see Heine and Kuteva 2005), Slavic appears less “pro-active” in comparison with the SAE languages. In fact, both standard and non-standard varieties of (East) Slavic function almost always as a replica system in the sense of Heine and Kuteva (Wiemer and Hansen 2012). Moreover, some cases of contact-induced grammaticalization tend to be built on both standard and non-standard data analyzed indiscriminately with an eye to proving that contact is indeed the primary driving force for language change, first and foremost in Slavic. To put it bluntly, a standard variety of one language is considered to exert influence upon a non-standard variety of another language, a premise which is likely to yield distorted results (see Kortmann 2004). Among the languages exerting influence on Slavic one finds (Low and High) German, Italian, non-Slavic Balkan languages (Albanian, Romanian, closely related Aromanian and, more arguably, Greek), Latin, and Finno-Ugric languages in the European part of Russia, most likely also Scandinavian. Yet the influence exerted by particular model languages can be different. For instance, in the Balkans, the role played by Turkic languages in grammatical changes in Slavic languages turns out to be, with rare exceptions, insignificant; by a similar token, the role that Greek has played in grammaticalization phenomena in the Balkans is considered smaller than one would assume “in view of the fact that Greek shares almost all core Balkanisms with Macedonian, Bulgarian, and Albanian” (Wiemer and Wälchli 2012: 72). Moving now northwards, there are very few clear cases of grammaticalization induced in East Slavic by Finnic languages (Grenoble 2010: 584; see Veenker 1967). As evidence shows, the idea of contact as one of the major factors triggering a grammatical change is somewhat exaggerated and warrants serious revision (see Danylenko in this volume). At least, the notion of high- and low-contact languages as suggested by Nomachi and Heine (2011b) for Slavic might appear less convincing than expected, especially when challenged by a more granular comparative-historical and dialectological approach (see Danylenko 2001, 2005, 2015).
Searching for a place of Slavic in Europe as a linguistic area
7
4 What makes Slavic so important? The book is so designed as to tackle the aforementioned issues, thus providing (1) objectively a more balanced perspective on Slavic as a member of possible areal groupings in Europe like SAE, the Central European (Danubian) area, the Carpathian-Balkan macroarea, and (2) a comprehensive overview of Sprachbunddefining features as identified in both high- and low-contact Slavic languages. In addition, the book offers (3) a series of new interpretations of well-known convergences found in the languages of Europe, including Slavic, and (4) discusses Slavic material which, save for Balkan Slavic, is usually not accounted for in the study of Europe as a linguistic area. As has been recently demonstrated in areal studies, the linguistic map of Europe has been shaped significantly by language contact of various kinds within the course of the last two millennia. But one can raise the legitimate question of why clusters of the European linguistic properties are particularly pronounced in most of the varieties of Romance and Germanic which serve purportedly as models of transfer with regard to Slavic (see Wiemer and Wälchli 2012: 32–33). Moreover, the extent of contact-induced grammaticalization is likely to differ in various contact situations, some of them belonging to wellestablished linguistic areas like the Balkan Sprachbund and others to less studied areas like the Carpathian-Balkan macroarea. In addition to these queries, the authors of this volume wish to address the following issues: – What is the relation of Slavic to SAE and other alliances in Europe? – What is the ratio of areal, genetic, and typological generalizations that researchers have used in determining areal groupings in Europe? – What are the mechanisms behind the divergent and convergent development of the contiguous (both high- and low-contact) and distant languages, including primarily Slavic? – What is the typological and areal distance between Slavic languages and other languages belonging to particular linguistic areas in Europe? – To what extent can the shared properties be products of the contact among the neighboring languages? What other societal factors, in addition to contact, are likely to determine grammatical change? Is there any correlation between societal and linguistic patterning? As has been mentioned, the pioneering character of the volume lies in its aim to answer the aforementioned questions, offering new insights into the commonly accepted notions and methods in areal studies today, in
8
Andrii Danylenko
particular about the peripheral status of Slavic within SAE. New approaches and methods with regard to ascertaining the position of Slavic in SAE and other linguistic areas as construed for Europe in synchrony and diachrony are discussed in the first part of the book. The second part deals with the typological and historical dimensions of Slavic and its participation in SAE; the chapters of this part make use of new data and well-known methods of areal linguistics, which help further our knowledge of the convergences within SAE. The third part, consisting of four chapters, focuses on the involvement of Slavic in various linguistic areas in Europe, in particular in the Central European (Danubian) area, the Carpathian area, and the Balkan Sprachbund. Finally, the last part deals with the products of contact-induced grammaticalization and grammatical replication in the domains of verbal and nominal categories drawn on the data excerpted from such Slavic microlanguages as Molise Slavic (see Breu in this volume) and Kashubian (see Nomachi in this volume). In addition to the reconstruction and the (integral) result perspective as applied in areal studies (Wiemer and Walchli 2012: 9–14), the third part contains examples of the use of the sociolinguistic typology of Trudgill (2011) and the systematic typology of Mel’nikov (2003) with regard to the arealtypological sampling of the Carpathian-Balkan macroarea. According to the sociolinguistic/systemic typology, a grammatical change can be triggered by changes in the external determinants whose list can be reduced to the following major parameters (Trudgill 2011: 13–14): (1) small versus large community size (see Haudricourt 1961), (2) dense versus loose social networks (Milroy and Milroy 1985), (3) social stability versus instability of speech communities (see Dixon 1997), (4) high versus low degree of shared background information (see Martinet 1979), (5) degree of contact versus isolation. As follows from the above, contact is not the essential trigger for language change – it is only one of a number of societal factors determining linguistic patterning. What is crucial in the sociolinguistic (systemic) explanation of grammatical changes is the fact that all these factors tend to concertedly exert their influence on linguistic communication. Thus, arguably, the low density of social networks with a small amount of shared background information and the low level of social stability of a particular community are likely to be provoked by a high level of language contact. Conversely, a high level of (adult) language contact presupposes a loosening of social networks within an ever-growing community whose speakers are gradually losing amounts of shared background information due to the exponential increase of their number.
Searching for a place of Slavic in Europe as a linguistic area
9
5 Coda Several conclusions may be drawn at this point. First, a systematic investigation of the diversity of grammaticalization phenomena observed in both low and high-contact Slavic speech communities becomes indispensable for furthering our understanding of contact and its link with grammaticalization and areal grouping. This volume represents an effort to fill in some lacunae in the exploration of the place of Slavic in the linguistic landscape of Europe. Second, empirical studies investigating the aforementioned link are likely to remain circumstantial and heuristically weak if they are not enhanced by genealogical and sociolinguistic probing. Such a systemic approach toward Slavic, together with other members of Europe as a linguistic unity, is likely to modify our understanding of Sprachbund-forming features and their clustering due to language contact of various kinds and in various periods within the course of the last two millennia. Third, I also argue that, to obtain an adequate picture of grammaticalization processes in Slavic and other languages belonging to Europe as a linguistic area, one should investigate non-standard varieties of languages in contact. Consequently, the impact of the “cultured” or, prestigious written (standard), languages on non-standard varieties should not be overestimated in areal studies, thus belonging primarily to another level of sociolinguistic sampling. Even if it was not possible for every aspect to be investigated thoroughly, the place of Slavic on the language map of Europe has been successfully addressed in this volume from various research angles, which allowed the authors to outline its areal-typological and historical contours in a more distinctive way.
6 Summary of the Volume The volume is subdivided into four thematic parts,:(1) issues in methodology and pre-history, (2) lavic and Standard Average European, (3) Slavic in areal groupings in Europe, and (4) Slavic in Contact. This structure made it possible to organize the diversity of research topics and current problems facing the students of Slavic in its areal-typological and historical dimensions in a coherent way. In the remainder, I will briefly summarize the chapters of the volume in the order of their occurrence in the respective parts. Opening Part I, Björn Wiemer, in “Matrёška and areal clusters involving varieties of Slavic: On methodology and data treatment”, assumes that any
10
Andrii Danylenko
“linguistic area is an arbitrarily defined areal subsample of the global sample” (Wiemer and Wälchli 2012: 6). The granularity of the features, according to him, must fit within the scope of areal comparison. While discussing methods commonly used to quantify areally salient features, the author explores several cases of the internal diversification of Slavic features that have evolved from distinctions inherited from Common Slavic but which have drifted in different directions, often as a result of language contact. This chapter further elaborates on the role of contact in the conventionalization of minor usage patterns and looks into considerations meant to demystify the notion of “Sapirian drift”, recently exemplified by Joseph (2013). Vít Boček, in “Common Slavic in the light of language contact and areal linguistics: Issues of methodology and the history of research”, presents a general overview of numerous and often mutually contradictory views on the role of contact influence of individual languages on the development of Common Slavic in the period of its disintegration (Boček 2014). In this chapter, the author looks into the problem of Slavic and (Early-)Romance (Proto-Romanian) language contact, while critically assessing different interpretations of Slavic-Romance similarities. He argues that in this particular case, one should speak not of unidirectional contact influence, but mutual areal convergence. In “Intertwining trees, eddies, and tentacles – some thoughts on linguistic relationships in Europe, mainly Slavic-non-Slavic”, Robert Orr offers a brief overview of the almost infinite number of ways of viewing linguistic interrelations in Europe, including the well-known SAE Germanic-Romance Sprachbund linguistic core. The modification “Standard Central European” (SCE), coined to incorporate the Slavic-Celtic periphery, is only a small improvement; within Europe, Icelandic and Faroese are located even further to the West than Celtic, albeit with a structure far closer to the SAE/SCE template. Languages do not have to influence each other across lengthy boundaries. There is also an abundance of more tenuous pathways of influence, down to and including the often unique trajectories travelled by individual words, and, on occasion, even metaphors such as “tentacles reaching around peripheries”. According to Orr, we must also consider phenomena such as various non-Indo-European ad-, sub-, and superstrata, languages playing the role of bridges, linguistic interactions on the peripheries of the Slavic world, linguistic ring continua, and the unique linguistic position of Hungarian. Part II is concerned with Slavic and its relation to SAE. Jadranka Gvozdanović, in “Standard Average European revisited in the light of Slavic evidence”, believes that Sprachbund phenomena must be reevaluated both in terms of their substantive and formal properties, a reassessment which is likely to lead to a modification of the commonly accepted notion of SAE. While
Searching for a place of Slavic in Europe as a linguistic area
11
examining the historical dimension of SAE as a true Indo-European linguistic area, the author reminds us that Celtic does historically pertain to it and the origin of SAE could be older since some of the Slavic-Germanic convergences are traced back to a period well before the Great Migrations. Only after that time were congruent phenomena greatly facilitated by the languages of Christianity, and this is where the more recent division between the eastern and western parts of SAE comes into the picture. In “The perfects of Eastern “Standard Average European”: Byzantine Greek, Old Church Slavonic, and the role of roofing”, Bridget Drinka treats SAE as a dichotomy: SAE East and SAE West. The author focuses on SAE East and the role of the Greek Orthodox Church and the Byzantine Empire as influential superstructures in the East, in parallel with the Roman Catholic Church and Roman Empire as foundational structures for SAE West. The importance of this influence is demonstrated through a close analysis of the role of the l-perfect in Old Church Slavonic and an examination of the role of Greek in this development. The perfect construction itself, consisting of an l-participle plus ‘be’ auxiliary, is of Slavic origin. However, the influence of Greek can be seen in the expansion of this construction, the late development of an incipient preterital use of the perfect in parallel with the syncretization of perfect and aorist in Greek, and the development of viewpoint aspect, facilitated by this late de-aspectualization of the perfect. Within Slavic, the east-west division was replicated in the traditions of Slavia Orthodoxa and Slavia Romana (Picchio 1980). Thus, the “roofs” of Orthodoxy and Catholicism, of Greek and Latin, created, according to Drinka, a cultural and linguistic divide not only between eastern Europe and western Europe, but also within Slavic itself. In “Slavic vis-à-vis Standard Average European: An areal-typological profiling on the morphosyntactic and phonological levels”, Nataliya Levkovych, Lidia Federica Mazzitelli, and Thomas Stolz address two topics. First, the authors check to what extent the phoneme charts of the Slavic languages correspond to the patterns which are assumed to be characteristic of the members of the SAE Sprachbund. Second, the chapter focuses on the category of possession, in particular on a small selection of phenomena which are often viewed as cases of “marginal possession”. For these phenomena too, the authors check the factors which distinguish some or all of the Slavic languages from some or all of the SAE languages and what the factors are that provide evidence for convergence. Paul Wexler, in “How Yiddish can recover covert Asianisms in Slavic, and Asianisms and Slavisms in German (prolegomena to a typology of Asian linguistic influences in Europe)”, argues that scholars ignore multiethnic and multilingual Slavic confederations within and beyond Europe in the first
12
Andrii Danylenko
millennium AD which could have contributed to the formation of SAE. According to him, the Ashkenazic Jews, descended primarily from Iranian, Slavic and Turkic converts to Judaism, are the last identifiable triadic SlavoIrano-Turkic confederation. They spoke Yiddish – originally a cryptic relexified Slavic trade language that has absorbed elements of their parallel obsolete “Iranian and Turkic Yiddishes” (recoverable also from some Afro-Eurasian trade lexicons). This chapter shows how Yiddish can identify covert Iranian and Turkic (and other Asian) lexical, phonological and syntactic elements in Slavic and German, both of which are core members of SAE. Areal groupings are the focus of Part III. Helena Kurzová, in “Defining Central European convergence area”, focuses on the definition of the Central European convergence area, whose languages – belonging to three different genetic stocks, i.e., Indo-European in its Germanic (German) and Slavic (Czech and Slovak) branches and Finno-Ugric (Hungarian) – show substantial convergence in the common structuring of semantics and syntax as well as in the typological position of the word. In the Balkan Sprachbund, the convergence is more conspicuous, however, with the Balkan correspondences being more unique. In relation to the rest of Europe, the Balkan Sprachbund is characterized by typical and peculiar features and, consequently, Balkan Slavic too shows more conspicuous differences with regard to other Slavic languages. This is not the case with CE languages that are typologically an integral part of non-Balkan Europe. The author concludes that standardization influenced by Western cultural languages in these languages, subsumed under the term SAE, has not had as strong an impact on the European languages as is often argued. The Slavic languages of the Danube Basin are investigated by George Thomas in “Some morpho-syntactic features of the Slavic languages of the Danube Basin from a pan-European perspective”. As the author shows, the languages of the Danube Basin share a number of features resulting from long-term contact. Since varieties of German, a core language of Haspelmath’s reformulated SAE, participate in that contact, the possibility that varieties of Slavic spoken in the Danube Basin also share features of SAE to a greater extent than reported hitherto seems worth exploring further. According to the author, it is important not to fall into the trap of thinking of this as “Europeanization” as if the other Slavic languages were somehow “un-European”, although the western orientation is a possible ingredient in the typological variation of the Slavic languages as a whole. Since the varieties of Slavic spoken in the Danube Basin are themselves transitional to the other Slavic languages, the investigation of the distribution of certain morphosyntactic features in Slavic as a whole may require a reappraisal of the position of Slavic within SAE.
Searching for a place of Slavic in Europe as a linguistic area
13
Andrey N. Sobolev, in “Slavic dialects in the Balkans: Unified and diverse, recipient and donor”, offers a geolinguistic (areal-typological) analysis of the Balkan Slavic languages in the Southeastern European and Eurasian context. The convergent and divergent developments in areally contiguous Bulgarian, Macedonian, Eastern Serbian, Romanian, Albanian, and Greek are analyzed in this chapter against the background of language contact and the modern Sprachbund theory, showing how the mechanisms of borrowing and calquing, language shift or language and ethnic separation interplayed in the area. Andrii Danylenko, in “Balkanisms and Carpathianisms or, Carpathian Balkanisms?”, explores areal-typological discrepancies in the treatment of the so-called Balkanisms and Carpathianisms. The author believes that the emergence of Sprachbund-forming Balkanisms and Carpathianisms is not immediately dependent on the areal diffusion of the pertinent features either via matter borrowing or pattern borrowing, or in other terms, via borrowing proper or replication. According to the author, the solution might be sought in a specific constellation of societal factors as discussed in the sociolinguistic typology of Trudgill (2011). In other words, the appearance of parallel Balkanisms and Carpathianisms within the Carpathian-Balkan linguistic macroarea might be linked via a resulting configuration of societal factors with developmental similarities of the languages in contact. Part IV consists of three chapters. In “Morphosyntactic changes in Slavic micro-languages: The case of Molise Slavic in total language contact”, Walter Breu concentrates on an overview of the morphosyntactic changes in tense, mood, case, gender, and definiteness that characterize the convergence between Molise Slavic and the contact languages involved. The author tackles the problem of “pertinacity” in language contact, by discussing factors that contradict the opinion that in language contact situations all types of changes are possible and result in a resistance to an overall drift to the Romance type. Among other phenomena, the author looks into the conservation of a case system in a “caseless environment”, the conservation of the aspect opposition of the Slavic derivational type, the conservation of the auxiliary ‘be’ for the formation of the present perfect, but also the development of a new type of past tense formed with a particle, found neither in Slavic nor in Romance, and some other pertinent morphosyntactic changes. Bernd Heine, in “On formulas of equivalence in grammaticalization: An example from Molise Slavic”, states that much of what was proposed to constitute SAE is claimed to be in some way or other due to contact-induced grammaticalization and to grammatical replication in general (Heine and Kuteva 2006). One central problem concerns the mechanism that induces people to relate the structures of two (or more) languages in contact to one
14
Andrii Danylenko
another. This problem has received some attention in recent research and a number of general notions have been proposed, such as calquing, identification, structural borrowing, loan translation, polysemy copying, pivotmatching, distributional assimilation, etc. The chapter utilizes another notion, referred to as “formula of (translational) equivalence”, being concerned with the outcome of the mechanism (Heine in press). Using an example from Molise Slavic, the author looks into the relevance of this notion for the study of language contact, especially what it means with reference to grammaticalization. Finally, in “Placing Kashubian on the language map of Europe”, Motoki Nomachi analyzes morpho-syntactic features of Kashubian in the context of SAE. To that end, the author picks up five SAE features (definite and indefinite chapters, the have-perfect, the negative concord, the non-pro-drop feature, and the comitative-instrumental syncretism) as relevant areal features borrowed from German as one of the core SAE languages with which Kashubian has had close contact for centuries. The author comes to the conclusion that, against the general belief reached in various previous studies on Kashubian, this language tends to lose these features, though speed and degree of loss vary much in each case. This fact may be explained by significant sociolinguistic changes among the Kashubian speakers who, discarding German as the dominant language, have become full Polish-Kashubian bilinguals. Acknowledgments: Motoki Nomachi and I would like to thank all the participants in the International Symposium Slavic on the Language Map of Europe, which took place on August 8–11, 2013 at the Slavic-Eurasian Research Center of Hokkaido University in Sapporo (Japan), as well as all the contributors to this volume, who have been patient in going through several stages of editing and proofreading of their chapters. The present author is particularly grateful to Bridget Drinka (University of Texas at San Antonio) for insightful comments on the draft of the introduction. The editors are also thankful to the external reviewers, who kindly shared with us their expertise and demonstrated a willingness to help us improve the quality of this book. We want to express our gratitude to Volker Gast for useful suggestions concerning the content of chapters and the overall structure of the book. Needless to say, all editorial shortcomings fall under the responsibility of the editors of this volume.
Searching for a place of Slavic in Europe as a linguistic area
15
References Aarsleff, Hans. 1988. Introduction. In: Wilhelm von Humboldt. On Language: The Diversity of Human Language-Structure and its Influence on The Mental Development of Mankind, trans. by Peter Heath, vii–lxv. Cambridge: Cambridge University Press. Asenova, Petja. [1989] 2002. Balkansko ezikoznanie. Osnovni problemi na balkanskija ezikov săjuz [Balkan Lingustics. Fundamental Problems of the Balkan Sprachbund]. Sofia: Faber. Bernini, Giuliano and Paolo Ramat (eds.). 1996. Negative Sentences in the Languages of Europe: A Typological Approach. (Empirical Approaches to Language Typology, 16.) Berlin: Mouton de Gruyter. Boček, Vít. 2014. Praslovanština a jazykový kontakt [Proto-Slavic and Language Contact]. Prague: Nakladatelství Lidové noviny. Dahl Östen and Maria Koptevskaja-Tamm (eds.). 2001. Circum-Baltic Languages: Typology and Contact, Vol. 1: Past and Present; Vol. 2: Grammar and Typology (Studies in Language and Companion Series 54, 55.) Amsterdam/Philadelphia: John Benjamins. Danylenko, Andrii. 2001. Russian čto za, Polish co za, Ukrainian ščo za, etc. “was für ein”: A case of contact-induced or parallel change? Diachronica 18 (2): 241–265. Danylenko, Andrii. 2005. Is there any possessive perfect in North Russian? WORD 56 (3): 347–379. Danylenko, Andrii. 2013. Ukrainian in the language map of Central Europe: Questions of areal-typological profiling. Journal of Language Contact 6 (1): 134–159. Danylenko, Andrii. 2015. On the mechanisms of the grammaticalization of comitative and instrumental categories in Slavic. Journal of Historical Linguistics 5 (2): 267–296. Danylenko, Andrii. 2016. Oleksandr Popov (1855–80) and the reconstruction of Indo-European noun inflection. Language and History 59 (2): 112–130. Déscy, Gyula. 1973. Die linguistische Struktur Europas. Vergangenheit, Gegenwart, Zukunft. Wiesbaden: Harrassowitz. Dixon, Robert M. W. 1997. The Rise and Fall of Languages. Cambridge: Cambridge University Press. Drinka, Bridget. 2017. Language Contact in Europe. The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Garvin, Paul L. 1949. Standard Average European and Czech. Studia Linguistica 3: 65–85. Grenoble, Lenore A. 2010. Contact and the development of the Slavic languages. In: Raymond Hickey (ed.), The Handbook of Language Contact, 581–597. Malden, MA: Wiley-Blackwell. Haarmann, Harald. 1976. Aspekte der Arealtypologie: Die Problematik der europäischen Sprachbünde. Tübingen: Gunter Narr. Haspelmath, Martin. 1998. How young is Standard Average European? Language Sciences 20 (3): 271–287. Haspelmath, Martin. 2001. The European linguistic area: Standard Average European. In: Martin Haspelmath, Ekkehard Konig, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, Vol. 2, 1492–1510. (Handbuch zur Sprach- und Kommunikationswissenschaft 20.2.) New York: Mouton de Gruyter. Haudricourt, André Georges. 1961. Richesse en phonèmes et richesse en locutions. L’Homme. Revue française d’anthropologie 1 (1): 5–10.
16
Andrii Danylenko
Heine, Bernd and Tania Kuteva. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva. 2006. The Changing Languages of Europe. Oxford/New York: Oxford University Press Hinricks, Uwe (ed.). 2010. Handbuch der Eurolinguistik. Wiesbaden: Harrassowitz. Hock, Hans Heinrich. 1986. Principles of Historical Linguistics. (Trends in Linguistics 34.) Berlin: Mouton de Gruyter. Joseph, Brian. 2013. Demystifying drift: A variationist account. In: Martine Robbeets and Hubert Cuyckens (eds), Shared Grammaticalization: With Special Focus on the Transeurasian Languages, 43–65. Amsterdam/Philadelphia, Philadelphia: John Benjamins Klepikova, Galina Petrovna. 2003. Karpatskoe jazykoznanie i “Obščekarpatskij dialektologičeskij atlas” [Carpathian linguistics and The Common Carpathian Dialect Atlas]. Studia Slavica Hungarica 48 (4): 357–374. König, Ekkehard and Martin Haspelmath. 1999. Der europäische Sprachbund. In: Norbert Reiter (ed.), Eurolinguistik: ein Schritt in die Zukunft, 111–127. Wiesbaden: Harrassowitz. Koptevskaja-Tamm, Maria and Bernhard Wälchli. 2001. The Circum-Baltic languages: an areal-typological approach. In: Östen Dahl and Maria Koptevskaja-Tamm (eds.), Circum-Baltic Languages: Typology and Contact, Vol. 1: Past and Present, 615–750. (Studies in Language and Companion Series 54.) Amsterdam/Philadelphia: John Benjamins. Kortmann, Bernd. 2004. Dialectology Meets Typology. Dialect Grammar from a CrossLinguistic Perspective. Berlin/New York: Mouton de Gruyter. Kurzová, Helena. 1996. Mitteleuropa als Sprachareal. Germanistica Pragensia 13: 57–73. (Acta Universitatis Carolinae. Philologica 5.). Martinet, André. 1979. Shunting on to ergativity or accusative. In: Frans Plank (ed.), Ergativity. Towards a Theory of Grammatical Relations, 39–43. London: Academic Press. Marvan, George J. 2008. Introducing Europe to Europeans through their Language. One Europe for Centuries – The Euro-Czech View. Prague: Faculty of Arts, Charles University. Matras, Yaron and Jeanette Sakel. 2007. Investigating the mechanisms of pattern replication in language convergence. Studies in Language 31 (1): 829–865. Mel’nikov, Gennadij Prokop’evič. 2003. Sistemnaja tipologija jazykov [The Systemic Typology of Languages]. Moscow: Nauka. Milroy, James and Lesley Milroy. 1985. Linguistic change, social network and speaker. Journal of Linguistics 21: 339–384. Načeva-Marvanová, Mira. 2007. Evropský lingvistický areál (ELA) jako společný prostor jazykový a kulturní [The European Linguistic Area (ELA) as a common linguistic and cultural space], In: Jiří Koten and Patrik Mitter (eds.), Prostor v jazyce a literatuře. Sborník z mezinárodní konference, 36–39. Ústí nad Labem: Universitas Purkyniana. Nimčuk, Vasyl’ Vasyl’ ovyč. 1993. Ukrajins’ki hovory ta balkans’kyj movnyj sojuz [Ukrainian dialects and the Balkan Sprachbund]. In: Slov”jans’ke movoznavstvo, 41–63. Kyiv: Naukova dumka. Nomachi, Motoki and Bernd Heine. 2011a. Is Europe a linguistic area? In: Motoki Nomachi (ed.), Grammaticalization in Slavic Languages from Areal and Typological Perspectives, 11–26. Sapporo: Slavic Research Center, Hokkaido University. Nomachi, Motoki and Bernd Heine. 2011b. On predicting contact-induced grammatical change. Evidence from Slavic languages. Journal of Historical Linguistics 1 (1): 48–76.
Searching for a place of Slavic in Europe as a linguistic area
17
Picchio, Riccardo. 1980. Church Slavonic. In: Alexander Schenker and Edward Stankiewicz (eds.), The Slavic Literary Languages: Formation and Development. 1–33. New Haven: Yale Concilium on International and Area Studies. Reiter, Norbert (ed.). 1999. Eurolinguistik: ein Schritt in die Zukunft. Wiesbaden: Harrassowitz. Seržant, Ilja A. and Björn Wiemer (eds.). 2014. Contemporary Approaches to Dialectology. The Area of North, North-West Russian and Belarusian Dialects. (Slavica Bergensia 12.) Bergen: Dept. of Foreign Languages, University of Bergen. Stachowski, Marek. 2014. Eurolinguistics – what it is and what it should not be. Studia Linguistica Universitatis Jagellonicae Cracoviensis 131: 383–394. Thomas, George. 2008. Exploring the parameters of a Central European Sprachbund. Canadian Slavonic Papers 50 (1–2): 123–153. Trudgill, Peter. 2010. Contact and sociolinguistic typology. In: Raymond Hickey (ed.), The Handbook of Language Contact, 299–319. Malden, MA: Wiley-Blackwell. Trudgill, Peter. 2011. Sociolinguistic Typology. Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. van der Auwera, Johan. 1998a. Conclusion. In: Johan van der Auwera (ed.), Adverbial Constructions in the Languages of Europe, 813–836. (Empirical Approaches to Language Typology. EUROTYP 20/3.) Berlin/New York: Mouton der Gruyter. van der Auwera, Johan. 1998b. Revisiting the Balkan and Meso-American linguistic area. Language Sciences 20 (3): 259–270. van Pottelberge, Jeroen. 2001. Sprachbünde: Beschreiben sie Sprachen oder Linguisten? Linguistik online 8, 1/01: 1–26. Veenker, Wolfgang. 1967. Die Frage des finnougrischen Substrats in der russischen Sprache. (Uralic and Altaic Series 82.) Bloomington: Indiana University Press. Whorf, Benjamin Lee. 1941. The relation of habitual thought and behavior to language. In: Leslie Spier et al. (eds.), Language, Culture, and Personality. Essays in Memory of Edward Sapir, 75–93. Menasha, WI: Greenwood Press. Wiemer, Björn. 2011. Grammaticalization in Slavic languages. In: Bernd Heine and Heiko Narrog (eds.), Handbook of Grammaticalization, 740–753. Oxford/ New York: Oxford University Press. Wiemer, Björn and Bernhard Wälchli. 2012. Contact-induced grammatical change: diverse phenomena, diverse perspectives. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 3–64. (Trends in Linguistics 242.) Berlin/Boston: Mouton de Gruyter. Wiemer, Björn and Hansen Björn. 2012. Assessing the range of contact-induced grammaticalization in Slavonic. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 67–156. (Trends in Linguistics 242.) Berlin/Boston: Mouton de Gruyter. Wiemer, Björn, IljaSeržant and Aksana Erker. 2014. Convergence in the Baltic-Slavic zone. In: Juliane Besters-Dilger, Cynthia Dermarkar, Stefan Pfänder, and Achim Rabus (eds.), Congruence in Contact-induced Language Change, 15–42. Berlin/Boston: Walter de Gruyter. Zubkova, Ljudmila Georgievna. 2010. Ptincip znaka v sisteme jazyka [The Principle of Sign in the System of Language]. Moscow: Jazyki slavjanskoj kul’tury.
Part I: Issues in Methodology and Pre-History
Björn Wiemer
1 Matrёška and areal clusters involving varieties of Slavic. On methodology and data treatment 1 Introduction It is certainly regrettable that descriptions of areally relevant features of European languages are often skewed towards so-called Standard Average European (SAE), which automatically yields a genealogical focus on Germanic and Romance. This tendency can in part be explained from the fact that “we know much more about these languages than the languages of Eastern Europe” (Heine and Kuteva 2006: 35). However, what happens if we move this focus into Eastern Europe? Recently, Danylenko has advocated for a Central European Area, with Ukrainian as an intermediate variety1 (Danylenko 2013: 149–150, 154): “I venture to claim that East Slavic constitutes the core of, what can be labeled, ‘Standard Average IE’ (SAIE), with Russian serving as its focus language and (Southwest) Ukrainian holding the intermediate position between SAIE and the neighboring European languages”. He conceives of SAIE as an area that stands out both typologically and genealogically. What is different in Danylenko’s approach is, first and foremost, the center of irradiation and, as a consequence, a focus placed on languages with other genealogical affiliations. Beyond these straightforward differences, there are some issues which are either implicit to Danylenko’s argument or which tend to be neglected or overlooked with regard to areal diffusion. Diffusion implies variation, and sometimes this leads to areal clusters. How do linguists establish clusters, and how do they deal with variation? In typology, areal diffusion has predominantly been assessed on a macro-level, both in terms of geographic size and of coarseness of distinctions, or features, by which languages can be compared. By contrast, contact linguistics and dialectology concentrate on diffusion and variation on lower levels, again in terms of the size of the area and the more fine-grained level of features, but also inasmuch as the investigated
1 In order to avoid clumsy circumlocutions, I use here and hereafter the term variety as an umbrella term for any sort of socially distinguished lect (be it a language, a dialect, or even a doculect), whenever I do not want to narrow down the view either to an established (standard) language or to some regionally or socially more restricted type of language. https://doi.org/10.1515/9783110639223-002
22
Björn Wiemer
varieties are usually much more closely affiliated genealogically (cf. Wiemer and Seržant 2014: 15–19). In the following I want to discuss issues which have become recurrent topics in more recent research conducted by specialists interested in linguistic variation in space and time. A more intense and effective exchange between Slavicists (and specialists of other language groups), typologists, and usage-based linguists (or variationists) should be undertaken. I will try to pinpoint some recurrent shortcomings in the areal assessment of linguistic variation, in particular, if on the basis of such an assessment, inferences concerning diachronic change are drawn. Furthermore, I will argue that there is no “Slavic type” as such and that in assuming areas including varieties of Slavic (or of any other group) one must be careful not to skew the picture by preconceived notions of typicality. Instead, features become salient only against some immediate areal background (apart from social stratification); see particularly Section 4. Saliency is, of course, a relative feature (close synonyms are ‘outstanding’, or ‘prominent’), and since it emerges only as a figure–ground phenomenon, the (areal, social, diachronic, or else) background is crucial. Before we can reach these conclusions, however, we must first consider objections against delimiting areas (with or without Slavic in their center) and show weaknesses inherent to discussions of how the areal distribution of features relates to diachronic evolution and/or contact either within Slavic varieties or with non-Slavic varieties. These issues will be laid out in Section 2, while in Section 3, I will use convenience sets of Slavic varieties and features in order to highlight some problems that arise with methods of quantifying variation or (dis)similarity between related varieties. Section 4, as mentioned above, addresses the concept of “Slavic type”. Section 5 presents case studies on different structural levels of Slavic and Baltic designed to illustrate the general points made in the preceding discussion. The chapter is capped with a summary, some conclusions and postulates (Section 6).
2 Begging questions and how to handle them In this section I want to disentangle a couple of problems which are in general known, but continue to be glossed over in many analyses of particular groups of languages showing genealogical or areal affiliation.
1 Matrёška and areal clusters involving varieties of Slavic
23
2.1 Areas are constructed: they always overlap and intersect with other areas Is there a procedure that prevents us from introducing biases into areal clines and clusters in Europe? As such, this question is too imprecise because it leads immediately to the question: Biases in relation to what? We have to specify, for instance, the geographical backdrop (see Section 4). But apart form this, the question relates both to the choice of the center(s) of diffusion (a.k.a. hotbeds) and to the choice of structural features selected by linguists to make their choices appear reasonable enough. The notion of SAE has often been criticized, and rightly so, for its being arbitrary and overly focused on preconceived features of the best known languages in Europe. Haspelmath (1998, 2001: 1505) is aware of the relative and noncomplementary character of areas, and he concedes that the features he highlights have not been selected randomly: “They were included precisely because they were known to show a distribution that supports the SAE hypothesis”. What is more important: It is perfectly possible that we will some day discover another Sprachbund, based on a different set of features, that has Russian at its core and extends all the way to western Siberia in the east and central Asia in the south, but within Europe comprises only the Slavic, Balkan, and Scandinavian languages. This area would overlap with SAE, but it would not contradict it. Thus, a language may in principle belong to different linguistic areas, and different linguistic areas may coexist “on top of” each other. (Haspelmath 2001: 1505)
Crucially, what changes in researchers’ specific interests and claims concerning linguistic areas is not the method of inquiry, but only the center of irradiation used as sort of gauge value for other languages (language varieties) both in geographic and linguistic terms; and more often than not efforts toward grouping languages into areas reveal genealogical and/or typological biases which arise from the researcher’s specific interests. This applies, for instance, to the so-called “Donau-Sprachbund” (Skalička 1968), the Central European Area (Thomas 2008: 126), or aforementioned Standard Average Indo-European (Danylenko 2013: 149–150, 154). This should make us aware of the consequence that any area posited by another linguist would just beg this question as to how to interpret and handle biases that are unavoidable in practice. Haspelmath (2001: 1492) defined linguistic areas as follows: “A linguistic area can be recognized when a number of geographically contiguous languages share structural features which cannot be due to retention from a common protolanguage and which
24
Björn Wiemer
give these languages a profile that makes them stand out among the surrounding languages”. Linguistic areas have been acknowledged as sometimes having a descriptive value, but there is no method to establish areas as distinct entities from a random sample of languages (or on the basis of random features); cf. Bickel and Nichols (2006), Bisang (2010), among others. Linguistic areas are constructed by linguists (van Pottelberge 2001; Nau 2012; Wälchli 2012; Wiemer and Wälchli 2012: 14–18), and it is only to be expected that “two neighboring languages are likely to be more similar to one another than two languages that are not neighbors” (Heine and Nomachi 2011: 4), given the usual situation that people living in some contiguous areas naturally tend to communicate with each other. Such insights sound like truisms, but quite often they do not seem to be accounted for in research, so that we should be wary of methodological shortcuts when claims about some particular area are raised. Two points should be added here. First, it is not only often empirically impossible, but even methodologically not advisable to try to tell apart the impact of genealogical affiliation from the consequences of contact (or diffusion). A good reason not to do so may be situations in which languages from a former dialect continuum, which first had dissimilated in separation from each other, got into more intense contact at a later stage and, thus, again were prone to assimilating to each other. Such scenarios have been postulated, in particular, with respect to the so-called Balto-Slavic unity, but they have also been vital with regard to dialect convergence in more recent or modern stages (see Sections 2.3 and 5.2). Second, every area is itself part of some other, larger area. This, again rather trivial, insight has created the matrёška-metaphor to be introduced in Section 4. Moreover, linguistic areas are often located at the intersection of areal clines from different geographic directions. In fact, the latter case (successive overlap) is known from dialect geography as well, and it has, from a broader perspective, been captured by the term contact superposition zone (KoptjevskajaTamm and Wälchli 2001) or stratified convergence zones (Drinka 2017).
2.2 What does “Slavic (type)” mean? The concept Slavic type, or even common Slavic heritage, proves problematic. Is there any typical representative of contemporary Slavic? How much do later stages of Slavic correspond to the structural (and thus typological) make-up of what we can reconstruct as properties of Proto- or Common Slavic? Apart from the fact that all Slavic languages have developed a distinction of perfective vs. imperfective aspect—which is based on stem derivation and whose inception can
1 Matrёška and areal clusters involving varieties of Slavic
25
be detected already in Common Slavic—contemporary Slavic languages are so diverse that it will be difficult to find a larger number of intersecting elements covering all of them. Their similarities are rather reminiscent of family resemblance, even if only categories (i.e., types) are taken into consideration and gradient criteria such as the admissible lexical input to the categories are neglected. Table 1a presents some structural properties of Slavic as they existed before the dissolution of the commonly assumed dialect continuum. These properties are not selected randomly, but as with Haspelmath’s SAE-features (see Section 2.1) they are chosen because they are convenient and salient. By contrast, Table 1b lists some features which were lacking at that early and relatively homogeneous stage of Slavic. Of course, one could choose other features instead, or add more, but since here the purpose is to demonstrate some problem which exists in principle, it is of minor importance whether features and varieties are chosen randomly or, by contrast, arbitrarily for a convenience sample. The crucial point is: What should be taken as the status quo of the alleged “Slavic type”? It would be arbitrary to assign any of the contemporary languages (or varieties thereof) as the typical representative, and the number of properties that really represent all, or at least a majority, of today’s Slavic varieties is very small (see below).
Table 1a: Structural properties of Common Slavic (by approximately 500–700 AD).
(CS+)
nominal inflection, consisting of cases,a distributed over different declensions
(CS+)
dual number in all inflected parts of speech
(CS+)
nominal vs. pronominal declension of adjectives
(CS+)
inflectional aspect opposition in the past (aorist: imperfect)
(CS+)
Supine
(CS+)
enclitic pronouns (beside other clitics)
(CS+)
syllabic tones (metatony) and phonological distinction of vowel length
a
I disregard the vocative here.
Therefore, one might want to search for such a type at some starting point, at which Slavic was spoken in a more or less homogeneous dialect continuum, say, late Common Slavic, i.e., the 5th–6th centuries AD. The following positively and negatively defined features (CS+1–7 and CS-1–7, respectively) should give us a good idea of some landmarks in the structural profile of Slavic before its diversification began:
26
Björn Wiemer
Table 1b: Absent structural properties of Common Slavic. (CS-)
articles (definite, indefinite)
(CS-)
gerunds (adverbial participles), i.e., a kind of converb
(CS-)
animacy distinctions in morphosyntax
(CS-)
modal auxiliariesa
(CS-)
analytical causative constructions (e.g., with ‘to give’ or ‘to let’) b
(CS-)
stem-derivative aspect opposition (perfective: imperfective)
(CS-)
future tensec
a
Modal auxiliaries as a category developed in historic times. The earliest documents (in Old Church Slavonic) showed only mošti and lьzě (both ‘canʼ), but neither of them was frequent, and both units occurred in lexical meanings ‘strong, mightyʼ as well (Hansen 2001: 275–279). From this we may safely infer that no core representatives of modal auxiliaries had developed before the onset of written documentation. The situation was very similar for analytical causatives (see next criterion and next note). b According to von Waldenfels (2012: 247), the oldest Slavic documents (in Old Church Slavonic from the 11th century AD) showed only a very rudimental stage of DA(VA)TI+INF as permissive constructions. Thus, although he did not exclude the possibility that this permissive construction developed during the Proto-Slavic (or Common Slavic?) period, we can argue at best for only the very initial stage of the evolution of analytical causatives. c All future grams in Slavic developed after the breakdown of Common Slavic. This is reflected in their heterogeneity both in shape and origin all over the Slavic territory. Many of them correspond to different contact situations and areal affiliations; cf. the survey in Wiemer and Hansen (2012: 104–112), as for the diversity and relatively late development of future grams, see also Andersen (2009: 128, 131–136).
There are only a handful of properties which are shared by all contemporary Slavic varieties, despite inner-Slavic variation in many details (see comments to [1(a–c)]): [1] (a) Aspect opposition: The common property is the morphological basis (stem derivation) and the fact that clear-cut grammatical oppositions between perfective and imperfective verbs (often grouped into pairs with virtually identical lexical meaning) exists; the functional range and the extent to which the opposition of perfective: imperfective has expanded and strengthened in different Slavic varieties is subject to some considerable variation. (b) Pronominal (“long form”) adjective declension: The pronominal origin of this declension (which made adjectives differ from nouns) is however obliterated in many varieties, especially in West Slavic. Compare, for instance, Polish where only remnants of the nominal inflection exist in the masculine singular
1 Matrёška and areal clusters involving varieties of Slavic
27
(see remark a to Table 2) and distinctions of vowel length have been lost, so that contractions that led to lengthening can no longer be perceived (compare NOM.SG.F forms like dobr-a ‘good’ Polish stage < dobr-ā Czech stage < dobr-a-ja Russian stage). (c) Preservation of a system of (finite and non-finite) verb forms based on two or three different stems, with a clear-cut distinction of past vs. non-past tense stems maintained as in Common Slavic: This preservation concerns only the overall architecture of the tense system with present tense stems being the constant component. Note, however, that even the stem-final morphonological changes typical of present tense inflection (due to the first palatalization of velars and the elimination of j-clusters), which are still commonplace, for instance, in East Slavic and Polish (cf. Russian mog-u ‘I can’ vs. mož-e-š’ ‘you (SG) can’ < *mog-ti (INF) as an instance of the first palatalization of velars, and rez-a-t’ (INF) ‘to cut’ vs. rež-u (PRS.1SG) < *rez-ju ‘I cut’ as an example of the elimination of j-clusters), have been replaced by conjugations in which paradigm-internal alternations are avoided, e.g. colloquial Czech muž-u, muž-eš ‘I can, you (SG) can’ instead of moh-u, můž-eš in standard Czech (spisovná čeština). Now, which of the aforementioned languages should be considered as a (more) typical representative of Slavic? Other properties can be associated at best with the majority of contemporary Slavic varieties. To the (still small) inventory of such properties belong [2] animacy distinctions: This feature is found in the domain of differential argument marking (DAM).2 Slavic looks special in a European context, because only Slavic has developed DAM on the basis of animacy; compare, e.g., Russian videt’ stoly (ACC.PL=NOM.PL) ‘to see tables’ vs. videt’ vorob’ёv (ACC.PL=GEN.PL) ‘to see sparrows’. Non-Slavic languages appear to have acquired it on this basis only under the influence of Slavic. Compare, for instance, Romanian and Ossetic (Bossong 1998: 211, 220, 231–232, 254–255). However, the animacy-based DAM shows considerable differentiation among Slavic varieties themselves. Without going into detail, it suffices to look at the differences of East Slavic vs. standard Polish and Czech: the latter have DAM not only for objects, but also for (canonical) subjects, and the affected segment on animacy scales is smaller than in East Slavic. The animacy-based DAM in western South Slavic and Sorbian has remained restricted (or has been reduced) to the accusative singular of masculine nouns, and it excludes Balkan Slavic (due to its loss of morphological case distinctions).
2 Cf. Witzlack-Makarevich and Seržant (2017) for the most recent survey and collection of papers concerning this cluster of phenomena.
28
Björn Wiemer
Other properties often associated with Slavic are either typical only for North Slavic and the northwestern part of South Slavic (first of all Slovene), like the lmarker of a general past tense, and would thus exclude Balkan Slavic. Or they are characteristic only for standard varieties, like, for instance, gerunds, which developed out of paradigmatically isolated nominally inflected forms of participles (Wiemer 2014: 1634–1639). Above that, many properties acquired in some postCommon Slavic period, fit well into larger clines comprising all of Europe or the middle part of the continent and/or are gradient (based on different criteria). Here we may mention modal auxiliaries (Hansen 2005; van der Auwera, Ammann, and Kindt 2005) and analytical causatives (von Waldenfels 2012; Levshina 2015), but also the general past marked with the l-form, which is just the Slavic manifestation of an areally pervasive perfect > past-shift (Breu 1994: 56–58; Thieroff 2000: 282–286). Moreover, here belong two different types of future known for their complementary inner-Slavic distribution: the de-volitive future (< xоtěti ‘to want’) of South Slavic (except Slovene) and the future based on an originally inchoative marker (*bǫd- ‘to be(come)’), which has developed and spread in North Slavic and Slovene. Here I am not concerned with the reasons behind or mechanism of how these gram types spread and became entrenched in contiguous Slavic-speaking areas; I also gloss over some minor cases, e.g., the fact that some Kajkavian and Čakavian dialects of Croatian have the inchoative future, while the Resian dialect of Slovene has the de-volitive future (Lukežić 2015: 347–349; Ramovš 1935: 35). To account for such pockets would certainly amount to just going down to more finegrained taxonomic levels, for which the same question of mutual overlap and layering of gram types reoccurs. After all, it is uncontroversial that the mentioned gram types are known also in overlapping or neighboring non-Slavic languages on a larger areal scale.3 In sum, I find it difficult to determine what can be regarded as “typically Slavic”, or which Slavic language (variety) might be regarded the “best representative” of Slavic. I doubt whether anybody is able to do so, unless on some non-linguistic biases or preconceptions. When we speak of Slavic (and can reasonably unite all these disparate contemporary varieties under this one rubric),
3 On the controversy concerning the inchoative future in Slavic cf. the survey by Wiemer and Hansen (2012: 104–108), for the much less controversial history of the de-volitive future in South Slavic and in the Balkans, see Heine and Kuteva (2005: 188–192), who refer to ample preceding work. The de-volitive future has been mentioned also for Polabian (Polański 2002: 815), but no more far-reaching conclusions can be made from this rudimentarily documented language. After all, attestations of *xоtěti ‘to wantʼ used as a future marker might have reflected just a stage attested also for older periods of other North Slavic languages which were subsequently suppressed by the inchoative future.
1 Matrёška and areal clusters involving varieties of Slavic
29
we can do so only on the basis of results gained from standard assumptions in genealogical classification: (i) systematic sound equivalences of assumed cognate morphemes (roots, affixes), (ii) coincidences in the sound shape and paradigmatic ordering of morphological elements, (iii) shared words (or morphemes) from the elementary stock of the lexicon. However, apart from this common stock of phonological and morphological (including paradigmatic) equivalences, we find various layers which, as it were, tear apart the former unity in form and/or function. As a result, the Slavic-speaking landscape in Europe should better be assessed in terms of a larger contact superposition zone for which Slavic varieties are but one component. Such zones are characterized by multi-sided layering of binary contacts and their partial overlay in successive steps: “(. . .) intensive micro-contacts superimposed on each other sometimes create an impression of an overall macro-contact among the languages in the area, which has not necessarily been there” (Koptjevskaja-Tamm and Wälchli 2001: 626; cf. also Koptjevskaja-Tamm 2006: 192–193). This characterization is, to some extent, reminiscent of a formulation by Shevelov (1964). He refused to explain inner-Slavic dialectal differentiation and groupings on the basis of simple, linear branching from ancestors into daughter dialects; if any metaphor was appropriate, he argued, “the most suitable would be the image of clouds in the sky on a stormy day, with their constant changes in shape, their building up, overlapping, merging, separating and their ability to vanish in an instant” (Shevelov 1964: 611–612).4
2.3 Does diffusion need to be opposed to inheritance? Campbell (2006: 2) reproached research on linguistic areas in that “too much effort has been wasted on trying to define the concept, that little progress has been made, and that it would be more productive just to investigate the facts of linguistic diffusion without the concern for defining linguistic areas”.5 Additionally, researchers should not only state the synchronic facts and circumstances (circumstantialist account), but also ask for the processes that led to the observable situation. However, it is not always possible to answer the question “What happened?” satisfactorily for the simple reason that data are lacking (and none can be supplied post factum); cf. Nau (2012: 465). Diffusion usually implies contact, inasmuch as diffusion encompasses contiguous or overlapping speech communities. However, a caveat is in place
4 I owe this information to Andrii Danylenko. 5 Cf. Bisang (2004: 28, 2006a: 96, 2006b: 76) for a similar stance.
30
Björn Wiemer
here: the circumstantialist observation that in a given area certain features which are rare (or absent) in immediately surrounding (or neighboring) areas occur in not too closely related languages, does not in all cases allow for the conclusion that diffusion is the result of contact.6 Presumably, this caveat proves even more well-placed in case the area of convergence becomes larger. The mere fact of co-occurrence may be the result of spontaneous polygenesis (i.e., a parallel, but independent development). Moreover, spontaneous polygenesis is the more likely, the more structurally similar the languages in question are, and it is, in turn, sister languages of a common ancestor which are particularly likely to be structurally similar (van Pottelberge 2001). Finally, sister languages often occur in geographical vicinity. Thus, for smaller areas in which most varieties are genealogically closely affiliated it becomes particularly difficult to tell apart effects of contact from spontaneous polygenesis (see Section 5.2 for a comparison of cases). In turn, if change spreads from different hotbeds in a larger area, their spread zones can begin to overlap, and they may even merge, so that later macro-effects turn out as coalesced smaller, originally separate spread zones. Ideally, all such scenarios should be checked before claims about the role of contact are made.7 After all, if alternative hypotheses can be dismissed the claim for contact as the driving force in areal clustering becomes particularly strong. These caveats do not only make us aware of problems in the analysis of specific instances, they also show that one should not in principle put the close association of diffusion and contact into contrast with genealogical affiliation: these different reasons for convergence need not contradict, but may strengthen one another. From a typological perspective, a very similar point was made by Dahl (2001: 1457): “(. . .) putting diffusion against inheritance leads us astray. If diffusion is defined as a spread of an innovation from one location to another, this is part and parcel of virtually every process of linguistic change.” In this connection, one should not focus on spectacular cases alone (cf. also Wälchli 2012).
6 Remarkably, the term Sprachbund was invented by Trubetzkoy (1930) in a strictly circumstantialist manner, i.e., it was not designed to answer the question of “What happened?” (van Pottelberge 2001). 7 Extreme positions—blindly assuming contact as the main explanans for convergence vs. ruling out contact by all means—are never helpful. Ironically, considerations of structural closeness have prompted many linguists (like, for instance, Jakobson) to assume that a language is “ready” for borrowings (i.e., diffusion by contact); for discussion, see Thomason and Kaufman (1988: 17–18). Again, it is easy to be caught in a vicious circle: What is the explanans and what is the explanandum?
1 Matrёška and areal clusters involving varieties of Slavic
31
3 Problems with taxonomies and unidimensional measures Let us now address some rather basic questions of how diachronic diversification within genealogically related languages may be classified and quantified. Breu (1994) elaborated on a schema according to which the diachronic development of categorial oppositions can be ordered. It distinguishes between conservative and innovative behavior, and each of them is further subdivided: conservative behavior can manifest itself as preservation of inherited features or as the non-introduction of distinctions (categories) otherwise existent in surrounding or sister languages; likewise innovative behavior can surface as loss (or decrease) of an inherited feature or as the introduction (i.e., addition) of a new feature. See Figure 1 below.
behavior compared with the initial feature (category)
conservative
preservation
non-introduction
innovative
loss
addition
Figure 1: Taxonomy of diachronic behavior (Breu 1994: 59).
On the basis of Breu’s schema we can evaluate the lists in Tables 1a–1b that were presented in 2.2. If we examine some modern Slavic varieties and leave aside various intermediate stages and details (see the footnotes to Table 2), we get Table 2. I have deliberately chosen standard varieties, since these are normally referred to in the typological literature, and have added Northwest Russian (the Russian dialects of the Pskov-Novgorod region) which has often been discussed for its marked differences from standard Russian. The most obvious, though superficial, conclusion from this table is that, all in all, innovations prevail over conservative properties by a ratio of more than 3 : 1. Moreover, the higher the value of the coefficients, the more conservative the language is. Thus, prima facie, among the five Slavic varieties for the 14 features compared in Table 2 the following ranking (from most to least conservative) applies: Serbian/Croatian > Bulgarian > (standard) Russian / Polish (on an equal footing) > Northwest Russian.
32
Björn Wiemer
Table 2: Conservative and innovative features of some Slavic varieties as compared to Common Slavic. Old features (present in Common Slavic) Feature
Standard Polish
Standard Russian
Northwest Russian
Standard Bulgarian
Standard Serbian/ Croatian
nominal inflection
Kept
Kept
Kept
Lost
(Kept)
dual
Lost
Lost
Lost
Lost
Lost
nominal vs. pronominal adjective declension
Losta
(Kept)
Lost
Kept
Kept
aorist: imperfect
Lost
Lost
Lost
Kept
(Kept: Loss)
supine
Lost
Lost
Lost
Lost
Lost
enclitic pronouns
(Lost)
Lost
Lost
Kept
Kept
syllable tones
Lost
Lost
Lost
Lost
(Kept)b
Σ: Kept / Lost
.: . = .
.: . = .
: = . : = . .: . =
New features (lacking in Common Slavic) articles (definite/ indefinite)
NotAdd
NotAdd
NotAddc
(Added, only definite)
NotAdd
gerunds
Added
Added
NotAdd
Added
Added
animacy distinctions
Added
Added
Added
Added
Added
modal auxiliaries
Added
Added
Addedd
Added
Added
e
analytical causatives
Added
Added
NotAdd
Added
Added
perfective: imperfective aspect
Added
Added
Added
Added
Added
future tense
Added
Added
Added
Added
Added
Σ: NotAdd / Add
: = . : = .
: = . .: . = .
: = .
mean of coefficients:
.
.
.
.
.
1 Matrёška and areal clusters involving varieties of Slavic
33
Values in brackets were counted ½ (and ½ was added to the opposite value). Legend and general figures: conservative
innovative
Lost: Added: Coefficient of conservative vs. innovative features: ..
Kept: NotAdd:
feature preserved feature not introduced
feature lost feature introduced
A handful of remnants (like pełen ‘full (of)’, godzien ‘worthy’, wart ‘worthy’) can be left aside. They are discernible only in the masculine singular and are occasionally used also in object position. b Syllable tones are still alive, although their distribution has changed (in comparison to Common Slavic). c Neither in East Slavic dialects nor in colloquial standard Russian can the postposed -to (inflected or not) be regarded as an article. It should rather be characterized as a topic marker (Wiemer and Hansen 2012: 114–116). d Actually, the extent to which modal auxiliaries have become entrenched in these dialects has remained uninvestigated. Whether, for instance, usage of nado ‘(one) must’ as in nado trava (NOM) kosit’ (INF) ‘one has to mow the grass’ has been conditioned by standard Russian is an open question. e The same caveat as for the preceding feature holds here. At least, in dialect materials no such feature appears to occur. a
While these observations may represent fairly accurately the communis opinio concerning the innovative and conservative nature of Slavic varieties, it must be admitted that they have serious drawbacks. First of all, the comparison is based on particular phenomena that have somehow risen to salience in the literature on change in Slavic, without clearly articulated parameters. Second, the central tendency of both coefficients (old vs. new features) can result from diametrically opposed values for the single coefficients (compare, for instance, Northwest Russian vs. Bulgarian). Third, the feature list is not sensitive to the type of gram expressing a given category. This becomes most obvious with the future tense. In North Slavic the appearance of the future rests on the reinterpretation of the perfective present stem, whereas the followers of *bǫd- have been entrenched as future marker only with imperfective verbs (apart from regionally much more restricted markers, which I gloss over here). This complementary distribution is conditioned by the establishment of the opposition of perfective : imperfective stems and, therefore, does not represent an independent process among the features under consideration. Contrary to this, in South Slavic the underlying feature is connected with an additional morpheme (developed from Common Slavic xоtěti ‘to want’) and in principle independent of the new aspect opposition. Slovene, in turn, is intermediate between North and
34
Björn Wiemer
South Slavic, insofar as its distinct future gram derives from *bǫd- as in North Slavic, but it is not restricted to imperfective verbs and thus behaves like xоtětibased future grams in the rest of South Slavic. Fourth, many features exist on a gradient, in several respects. Trivially, they should be assessed from the point of view of frequency (type/token), and, as a rule, new categories (or grams) arise and gain territory only via stepwise expansion in the lexicon. To give several examples of gradient trends, in Slavic the imperfect first deteriorates and then disappears before the aorist does8; the inventory of modal auxiliaries is constituted item by item; the level of grammaticalization of analytical causatives depends on their range of functions; animacy distinctions develop to different degrees depending on syntactic position and grammatical number (especially in the accusative singular masculine, but also in the accusative plural and even in the nominative plural); the decrease of paradigm size in nominal inflection usually starts in the plural (compare the DativeInstrumental-Locative-syncretism in Serbian and Croatian), and so forth. Fifth, the chosen features are of different relevance for central parts of the language system. Relevance can be operationalized as degree of persistency and frequency. For instance, the supine, the dual, the gerunds or even the analytical causatives are less relevant than derivational aspect, the case system, or articles. Sixth, across Slavic varieties many features may be nondistinctive either in general (like the rise of stem-derivational aspect) or with respect to the small sample of varieties in question. For instance, in the given sample the dual and supine are not discriminative, because all compared Slavic languages have lost these forms; but if Slovene or Sorbian were included, the picture would become slightly more interesting. From this another question arises: whether the chosen distinctive feature is really crucial enough to give the respective variety a profile clearly differentiated from the rest. This issue has been noted from a macro-areal point of view by Dahl (2001: 1457), according to whom “the finding that two languages share a certain feature is not interesting as long as we have not found a sufficient number of languages that do not share the feature in question. In other words, the study of non-trivial similarities between languages presupposes an understanding of linguistic diversity.” This holds true not only for genealogically less-related or non-related groups of languages, but also—on a lower taxonomic level—among closely related language varieties, e.g., varieties deriving from a former dialect continuum.
8 An exception is Molise Slavic, and this exception can readily be explained by Italian influence (Breu 1998: 349, 2005: 41).
1 Matrёška and areal clusters involving varieties of Slavic
35
Feature lists as presented in Table 2 have been used for isopleth maps (van der Auwera 1998), too. As useful as they have proved for coarse areal comparisons, they can be criticized for some aspects (part of which are the same as the aforementioned ones) such as9: (i) Features cannot be weighted or hierarchized. (ii) Features must be binary, they cannot be scalar. (iii) No account of lexicon–grammar interaction is given (in particular, of range of admissible lexical input). (iv) Similarity is the net effect of a certain number of pre-conceived features (considered for some reason or other to be outstanding), but not necessarily of the same features. (v) No information is provided about the origin and direction of diffusion, or on (relative or absolute) chronology (i.e., “only” a circumstantialist account is possible). (vi) The area under consideration must be homogeneous with respect to the given feature(s). In particular: (vii) (vi.a) Features may turn out irrelevant for one or more varieties, or they may apply only partially (e.g., the definite article in Greek is not postposed, in contrast to other languages of the so-called Balkan Sprachbund). (viii)(vi.b) “Deviant” sub-varieties cannot be accounted for. (ix) (vi.c) The languages (varieties) must have remained in stable locations (in other words: no noticeable migrations of speaker populations must have occurred). Part of these shortcomings are inherent to the method itself; other limitations arise more generally from the nature of the data. For instance, the relation between dialectal (i.e., diatopic) and diastratic variation (in Coseriu’s (1988) sense) can hardly be accounted for. In fact, this relation poses one of the greatest problems for any empirically reliable assessment of the motifs of convergence phenomena among (non-standard) varieties of different language groups within larger areal clines. Traditional wave models cannot account for diastratic diffusion either, and they, too, rely greatly on the assumption of varietyinternal homogeneity. Of course, these caveats against commonly applied methods of areal typology (or dialect geography) should make us aware of their inherent restrictions and implicit assumptions. If we begin thinking seriously about how different parts of the Slavic-speaking world might be integrated into areal accounts or,
9 Some of these drawbacks were discussed in van Pottelberge (2001) and Laakso (2002: 241–242).
36
Björn Wiemer
more broadly, into accounts of patterns in intra- and cross-linguistic variation, one should further elaborate on the respective methods. Once we are forced to conclude that (a) there is no “most typical” representative of Slavic and that (b) linguistic areas cannot be established on a strictly random and nonpreconceived basis (see Section 2.1), the question of how Slavic varieties are integrated into the linguistic landscape of Europe seems to boil down to asking for areal clines, or clusters, to which various Slavic varieties, or rather, certain of their properties, belong. In the remainder of this chapter, I will leave aside other nagging problems and turn to one more consideration (Section 4), before offering several case studies (Section 5).
4 Matrёška: the figure-ground problem Typologists have often, and rightly so, been criticized for a neglect of diastratic diversification. In particular, so-called SAE-features (or Europeanisms) have been considered on the basis of standard varieties. However, it is standard varieties which often prove to be the odd ones out in comparison to the areal diffusion of structural properties of vernacular (non-standard) varieties. Of course, this holds true not only for Romance and Germanic (De Vogelaer and Seiler 2012: 14; Bucheli Berger, Glaser, and Seiler 2012: 96–97), but also for Slavic (and Baltic). Here are just three examples. (a) The standard Polish [± virile]-distinction, being just one variant among several Slavic animacy distinctions (see Section 2.2), stands out not only in comparison to its Slavic neighbors, but even on the background of variation within Polish (Wiemer 2004a: 507–511). (b) The existence of uninflected nouns in standard Russian (mostly in the neuter, e.g., kino ‘cinema’, pal’to ‘coat’, kakao ‘cacao’, kupe ‘(train) compartment’, but also animates like kenguru ‘kangaroo’ or kakadu ‘cockatoo’) proves unusual on the background of non-standard varieties, both in the diatopic and the diastratic dimension (the so-called gorodskoe prostorečie ‘urban mixed non-standard’), where uninflected nouns (as well as suppletion and other irregularities) tend to be avoided (either by inflecting them, as in v kin-e (LOC) ‘in the cinema’, or by adapting them to more convenient declension classes and changing the gender, e.g., kakav-a (F) for kakao, indeclinable in standard Russian). (c) Standard Lithuanian distinguishes between two evidential constructions based on different participles and with virtually complementary functional
1 Matrёška and areal clusters involving varieties of Slavic
37
distribution (Wiemer 2006a). The co-occurrence of both constructions is, however, a rather artificial result of codification: in terms of dialect geography, these two constructions derive from opposite regions (Holvoet 2007: 92–96); see Section 5.1. What holds true for a critical diastratic assessment of alleged outstanding features applies a fortiori to a more or less purely diatopical perspective, in particular if we want to speak of convergence zones or linguistic areas. Recall Haspelmath’s definition of linguistic area quoted in Section 2.1. Regardless of how large or small they are, zones of convergence become visible only against a background of larger geographical units (Wiemer 2004a). Concomitantly, one needs to define the size of the varieties that are being compared (see Section 2.3). Therefore, it is evident that, if one wants to apply variationist tools to areas (of whatever size), any outstanding feature can be established only on the background of the immediate surrounding of a given area; the occurrence of a typologically (i.e., on a world-wide scale) rare feature in non-related and areally contiguous languages may per se be a good indicator of contact being the medium of diffusion. But we may even come across a typologically frequent phenomenon in some small area (with genealogically not too closely affiliated languages) which turns out as noteworthy if this small area is assessed in relation to its immediate surrounding. Compare, for instance, future grams in European languages. Future grams for which the sources are movement verbs (first of all, ‘to go’ and ‘to come’) are reasonably widespread in the languages of the world (Bybee, Perkins, and Pagliuca 1994: 251–253; Majsak 2005: 349–356, 390–392, 399). In Europe, this kind of future gram is well represented in Romance and, therefore, in Europe’s (south)western part. Contrary to these, we do not find such future grams in Slavic at all; for Germanic, only the English going tofuture, the de-venitive source (‘to come’) in Scandinavian and in Swiss German – located at the other end of the Germanic landscape – can be found. Swiss German can be suspected to be in “micro-areal conspiracy” with Romansh, thus a Romance variety (for data, see Dahl 2000: 320–321). De-andative and de-venitive grams seem to have arisen independently of one another. Polygenesis comes as no surprise in view of the frequency with which de-andative and de-venitive futures are attested worldwide. However, each of these regions is close to, or even encircled by, larger areas in which other future grams are salient, or which do not have any firmly established future gram at all. By comparison, according to Bybee, Perkins, and Pagliuca (1994: 252–253), the verb ‘to want’ seems to be less well represented as
38
Björn Wiemer
a source expression of future grams, but it occurs in two clearly separated (and thus distinct) areas of Europe, in the North Sea area (most prominently English, but also Scandinavian; Dahl 2000: 322) and in the Balkans. These areas are, again, surrounded by larger areas with different future grams (or none at all).10 Their salience in areal terms should be more or less comparable to the salience of future grams based on movement verbs, but, at first blush at least, this applies only if we restrict our survey to Europe (including whatever one might want to include into Europe’s fuzzy edges). Whether it applies on a global background can only be judged if the worldwide distribution is checked (including smaller pockets, such as in the aforementioned Romansh-Swiss German de-venitive case), which does not seem to have been done. Moreover, the application of variationist approaches seems to imply that the smaller the area, the more fine-grained the criteria (features) may and should be, and the more differentiated with regard to the lexical input. In addition, token frequency should become more pronounced as a factor of change and be evaluated in combination with type frequency. This combination is crucial for an assessment of the productivity of patterns. Typefrequency has predominated in large-scale areal typology studies in recent years, while token-frequency can be assessed only on the basis of sufficiently representative and reliable corpora. Let us sum up. Regardless of how any particular case may be decided upon (after thorough empirical investigation), the fact that a particular feature emerges as salient for some small area A1 is due to its lack (or much less noticeable manifestation) in the immediate areal surrounding A2, and this larger area encircling A1 may itself be salient on a yet larger areal background A3, and so forth (at least theoretically). In addition to clines (and their multi-sided intersections resulting in contact superposition zones), we get an idealized model of concentric circles, or of something that resembles the well-known Russian nesting dolls called matrёška. Accordingly, I will refer to this constellation as the matrёška-model. A check of cases should reveal whether this idealized situation happens to exist in reality, or whether the model of contact superposition zones is not only the dominant, but the only existing one.
10 Note that the areal picture can quickly change if we classify future grams (or any other gram type) not on the basis of their source expressions, but on the basis of how much the morphological exponents of future meaning have coalesced (cliticized > agglutinated > fused) with the stems. For instance, the North Sea de-volitive future is much less bound than the respective morphemes in the Balkans.
1 Matrёška and areal clusters involving varieties of Slavic
39
5 Case studies The following case studies are meant to give more detailed illustrations of some of the points made above. The case studies deliberately rest on a circumstantialist standpoint. I am not primarily interested here in the diachronic background, but rather in stating what one can see from a “synchronic snapshot”, thus showing how areas of different size may overlap and to what extent smaller areas might be nested into larger ones. A too superficial glance may create the illusion that an area is homogeneous in terms of some structural property, and that even on a synchronic (i.e., circumstantialist) level it can become quite obvious that such shared properties have grown from different hotbeds, i.e., independent lines of development, which just happened to meet and accrete somewhere in that larger area. However, I do not intend to solve this question for any particular case, but I do want to suggest that a more nuanced approach is called for. The examples chosen are not restricted to cognate units (in related languages), but it is instructive to learn how cognate material distributes itself differently over adjacent (sub)areas, while non-cognate material (from rather unrelated languages) often underlies calquing in a much more consistent way across genealogical boundaries. Leaning on Matras and Sakel (e.g., Sakel 2007), I will use the term PAT-borrowing for patterns reshaped by some contact variety on the model of another variety and the term MAT-borrowing for cases in which also the phonological shape (“material”) has been taken over.
5.1 Resultatives and related constructions Resultatives are forms (or constructions) that focus on a current state caused by a telic event. Such are, for instance, West Slavic constructions with an inflected HAVE-verb and participles with the n/t-suffix showing object-agreement, e.g., Polish Mamy spraw-ę (ACC) załatwio-n-ą (ACC) ‘We have the matter settled’. European languages abound with participles whose primary or at least original function is/was to denote such states, concomitantly they derank (or eliminate) agentive arguments by focusing on the most affected participant of the situation (Haspelmath 1994). Cross-linguistically, resultatives often supply the input to various grammaticalization paths yielding different, though related, constructions on the clause level. In the eastern part of the Circum-Baltic Area (CBA), where Slavic, Baltic and Finnic have come into contact, constructions based on resultatives or anterior participles are distributed unevenly, but they cluster into subgroups within that rather small area. This suggests that the Slavic-Baltic-Finnic contact region lies in an overlap of larger clines, or at least
40
Björn Wiemer
of different regions each of which stretches beyond this particular contact region. For instance, Southwest Ukrainian and written varieties of Ukrainian spoken in the western part of Ukraine have demonstrated similarity to standard Polish,11 thus appearing intermediate between other East Slavic varieties, Polish and/or Baltic (cf. Wiemer and Giger 2005: 61–66; Danylenko 2006: 276–279; Wiemer, forthcoming). Grammaticalization paths which are relevant for the contact region of Slavic, Baltic and Finnic are presented in Figure 2.12 Arrows with solid lines show developmental steps which, for the given source, are without an alternative, arrows with broken lines show alternative paths. Of course, this picture is a simplification, but exactly for this reason it serves our purposes well in discerning both area-internal differentiation and how entire (rather small) areas can be nested in larger ones.
(a) perfect (anterior) > general past
resultative
voice-neutral domain (b) inferential --------------------------------------------------------------------------(c) (foregrounding) passive (d) backgrounding passive (i.e., with canonical subject) (=impersonal) voice-relevant domain
Figure 2: Resultative participles as input to clausal categories in the eastern CBA.
The developmental steps (a-c) are well-founded for cognitive reasons, thus it is not surprising that they are found occurring time and again in different languages of the world. Both object-oriented resultatives and foregrounding passives code the most patient-like argument in a syntactically privileged way (namely, as nominative subject). For this reason they are often conflated in different languages (e.g., in most Slavic languages, Lithuanian, French). (a) The step toward a perfect amounts to losing the restriction to state-changing events.
11 This situation was common in the 19th century in the Galician and Bukovyna varieties, and even earlier: Middle Ukrainian and Middle Belarusian demonstrated the same pattern with the Polish one. Deviations from Polish occurred when Ukrainian became more influenced by Russian in the Russian Empire and especially during the Soviet period (Andrii Danylenko, p.c.). 12 The facts used in the rest of this case study are based mainly on Seržant (2012), Wiemer (2006a, 2006b, 2007a, forthcoming), the relevant chapters in Wiemer and Giger (2005), also on Wiemer and Hansen (2012: 86–95), Wiemer, Seržant, and Erker (2014: 30–36) and Nau (2011), although the argument to be derived from them differs in some respects. See these works also for further references.
1 Matrёška and areal clusters involving varieties of Slavic
41
(b) The evolution into an inferential can be motivated either directly from a resultative gram or via a perfect gram; in any case this change consists in a tip-effect between communicative back- and foreground: the inferential asserts (or: focuses on) the event which the perfect or resultative implies. For instance, the Lithuanian inferential coded by a non-agreeing participle with the t-suffix states that something happened prior to the actual speech act and implies that the result or some kind of relevance for the present moment holds in (1), whereas a perfect (and even more so a resultative) focuses on a state that holds at (or is relevant for) the moment of speech and allows one to infer that this state was caused by some event (2–3): Lithuanian (1)
Čia kiški-o tupė-ta. here hare-GEN sit-ANT ‘Obviously, a hare sat here.’
(2)
A hare has been here.
(3) The dog has eaten the steak (that’s why it is looking so happy). These changes do not per se influence voice, although the Lithuanian inferential with the non-agreeing t-participle is more or less restricted to one-place (and thus intransitive) verbs and demotes its argument syntactically by coding it in the genitive. No other language in this region has this sort of construction, and at present Lithuanian is the only language in this region which has two different evidential constructions at its disposal.13 The other peculiarity is that the demoted argument is in the genitive; this is the same case as for the demoted argument in the Lithuanian foregrounding passive or, rather marginally, in constructions with predicative possession. Other languages of the region have either a dative (e.g., Latvian, Latgalian), adessive phrases (e.g., the adessive case in the Finnic languages, the PP u ‘at’ + genitive in varieties of Russian), a translative PP (e.g., Polish przez ‘through’ + accusative), or the bare instrumental (standard Russian), which is of particularly ancient descent (at least for Slavic). I will return to oblique agent-phrases below. Furthermore, it is important to realize that perfects have evolved out of resultative or anterior participles only in the northern half of the examined region, i.e., in Baltic, Finnic and the Russian dialects of the Pskov-Novgorod region
13 The other construction, with past active participles (which in the Baltic languages agrees with the nominative subject in number and gender), is common in Baltic and the Finnic languages Estonian and Livonian, but it is rather used with a reportive function.
42
Björn Wiemer
(Northwest Russian), while a perfect is absent in the southern half, i.e., in Polish.14 In fact, since in Northern Slavic the Common Slavic perfect, based on the l-participle and an ESSE-auxiliary verb, declined and became the marker of the general past (see Section 2.4), the West Slavic way of forming new resultatives (> perfects) has been constructions of the HAVE+n/t-participle type (see above) which closely correspond to the western European type known from Germanic and Romance. As for Northwest Russian, its perfects based on n/t-participles with an (often omitted) ESSE-auxiliary as in (4) were claimed by Seržant (2012) to have been a hotbed of diffusion which affected neighboring Baltic and Finnic dialects. His argument is mainly based on the coding of the demoted agentlike argument and its acquisition of subject properties. Regardless of how justified this claim is, anteriority participles as sources of a perfect are attested fairly well among the autochthonous Finno-Ugric languages occupying the region (north)east from Northwest Russian. Remarkably, in many of them perfect paradigms have been reinterpreted as evidentials (contrary to Russian dialects).15 Compare example (4) from Northwest Russian with the Estonian example in (5), and the Udmurt examples in (6a–b). Note that the Estonian example lends itself to divergent interpretations (typical of many resultatives and perfects): Northwest Russian (4) U syn-a ženё-no-s’. at son-GEN married-ANT-RM ‘Our son is married.’ (lit. ‘at our son (is) married’) Estonian (5)
Mul on auto varasta-tud. 1SG.ADES be.PRS.3SG car.NOM steal-ANT ‘My car is stolen.’, or: ‘I have my car stolen.’, or: ‘I have stolen a car.’ (Example thanks to courtesy of Christoph Unger)
Udmurt (6a) mon č́i̮lkak vunet-iśkem. 1SG completely forget-PRF.1SG ‘I have completely forgotten.’ (Serebrennikov 1960: 119)
14 The perfect status of the -no/-to-construction in Ukrainian has been a subject of considerable debates; see Wiemer and Giger (2005: 62–64) for a survey. 15 Cf. Serebrennikov (1960: 118), Csúcs (1998: 289) and other contributions in Abondolo (1998).
1 Matrёška and areal clusters involving varieties of Slavic
43
(6b) Vetl-em. go-PRF.3SG ‘She has gone (obviously).’ (Csúcs 1998: 289) Therefore, the micro-areal diffusion which probably spread from a tiny region into the borderlands of East Slavic, Baltic and Finnic may, on closer inspection, well turn out to be but a cluster segment from a larger area in the northern part of Russia (possibly stretching over both sides of the Ural mountains). The question is whether this cluster is really a uniform cline or rather a loose agglomeration of tiny areas which all comprise offsprings of polygenetic (thus, independent) origin. Let us now come to the lower part of Figure 2, i.e., to voice-related operations. The canonical passive (with a subject coding the most patient-like argument in the nominative) often evolves from resultatives, because the natural link between both is the focus on the most affected participant (and the demotion of some agentive counterpart, i.e., the causer of a state-changing event). In many languages (such as Russian, Lithuanian or French) object-oriented resultatives and passives prove to be difficult or even impossible to differentiate, so that only the context or situation of speech can disambiguate them. Many passives have evolved from object-oriented resultatives (Wiemer 2004b: 286–288). Passives can, in turn, be the basis from which impersonals evolve, i.e., constructions which syntactically demote the most agent-like participant, but do not promote another argument to a syntactically privileged position (subject).16 The demoted argument may be the only one, so that we get some kind of subjectless passive, as in German Hier wird getanzt ‘There is (some) dancing here’. In other languages, as in Polish, impersonals based on participles can be derived without almost any lexical restrictions from any verb provided its most agent-like argument is a human being; for instance Polish (7) Na t-ej sal-i nie palo-no nigdy. on this-LOC room-LOC NEG smoke[.IPFV]-ANT never ‘In this room nobody ever smoked.’ (lit. ‘. . . (it) was never smoked’) (8)
Wczoraj zaaresztowa-no złodziej-a. yesterday arrest[.PFV]-ANT thief-ACC ‘Yesterday they arrested a/the thief.’ or ‘Yesterday a/the thief was arrested.’
16 The diachronic relation between passives and impersonals varies depending on diverse conditions (Wiemer 2011: 545–546).
44
Björn Wiemer
Another source of impersonals may be perfects (anteriority participles), which, as already mentioned, do not themselves imply any voice orientation. As for the Polish -no/-to-construction the question which one of the two possible construction types—passives proper or simply anteriority participles—gave the original input (or whether both had a share in this development) has remained a question open to debate (cf. Wiemer, forthcoming). In any case, an array of systematic differences between the Polish -no/-to-construction and the socalled new possessive perfect as represented in Northwest Russian make us strongly inclined to assume that these East Slavic dialects and Polish form different hotbeds for the spread of constructions based on non-agreeing participles in the (south)eastern CBA. First of all, (i) the Polish -no/-to-construction does not at all function like a perfect, but, even more importantly, (ii) there are no problems in determining the syntactic alignment of arguments (since no other changes than the syntactic demotion of the most agent-like argument take place), (iii) the demoted argument cannot be expressed (any more), but the PPs and the instrumental which could be used in earlier stages (see below) did not belong to the dative-adessive domain of demoted agents which are typical of the northern region grouped around Northwest Russian. In turn, what speaks in favour of both Northwest-Russian and Polish having been independent centers of irradiation is the fact that among all varieties in the southeastern CBA the non-agreeing predicative participles in Polish and Northwest-Russian prove to be least restricted in terms of lexical input (although for different reasons). Neither aspect nor actional classes of verbs induce any restrictions, either, whereas all other varieties (including standard Latvian and standard Russian, which do not belong properly to the Baltic-Slavic-Finnic contact region) show quite severe restrictions as for both lexical input and aspect or actional classes. Remarkably, Lithuanian is situated between these two assumed hotbeds not only in geographic terms, but also formally: its non-agreeing participles (taparticiples, see above) stand out within the entire contact region examined here in terms of function (inferential) and argument structure (genitive marking for the demoted argument). The history of the Lithuanian non-agreeing taparticiples differs entirely from the diachronic background of the other participles (regardless of whether they are cognate or not), as this is the case for the history of the genetivus auctoris (Holvoet 1995; Wiemer 2004b: 310–313). In a sense, Lithuanian is set off from the other two independent spread zones, each of which has developed its own way of marking oblique actors. Explicit marking of the most agent-like argument is blocked in the Polish -no/ -to-construction, but this is a fairly recent development; Polish up to the late 18th century allowed this argument to be marked with przez ‘through’ + accusative
1 Matrёška and areal clusters involving varieties of Slavic
45
or od ‘from’ + genitive. A similar remark is justified with respect to the restriction to past reference, which was established approximately in the same period (Wiemer, forthcoming, with further references). Neither of these restrictions occurs with the modern Ukrainian equivalent, which can mark the most agent-like argument with the bare instrumental and combines with an ESSE-auxiliary in all tenses and the subjunctive. However, the chronological relation to agent markers in Polish has yet to be settled. On the other hand, the Ukrainian -no/-to-construction has not extended the admissible lexical input or loosened actional and aspect restrictions to the same extent as modern Polish has; it shows some properties of a perfect and tends to be conflated with the passive proper (Wiemer and Giger 2005: 61–66).
5.2 What happened: accretion of different spread zones, or disintegration of a unitary spread zone? Danylenko (2013) dwelt on the spirantization of /g/ (*g > γ) in a sort of “central European belt”. He argued that this process took place in East Slavic independently from dialects farther to the west (Bohemian, Moravian, etc.). His argument relied mainly on absolute chronology (before and after the respective disappearance of the jers): In fact, the presence of an uninterrupted belt γ / h from the Bavarian frontier to the Oka river which cuts through the entire Slavic territory, with g-languages both north (North Russian, Polish, Lower Sorbian) and south (Slovene, except some westernmost dialects, Serbian and Croatian, save some littoral dialects, Macedonian and Bulgarian) can hardly be a result of convergence, involving German, Hungarian, Czech and Slovak within the C[entral] E[uropean] Sprachbund. (Danylenko 2013: 141–142)
However, convergence appears much less unlikely if we assume that an effect over a larger area (“from Bavaria to the Oka”) might have been created by the accretion of diffusion from different hotbeds. Has this alternative hypothesis been considered? Can it be investigated? Exactly the opposite constellation may obtain when there are two or more (very) small areas A and B simultaneously belonging to the same language group (e.g., Slavic) and to a larger area in which this group dominates, and both A and B show the same (or a very similar) feature which is absent in the other varieties belonging to their group, in particular it is absent in the region between A and B. Now the question would arise whether A and B have the same origin, stemming from a larger area C which has shrunk, so that A and B are just the last relics of that former area C, or whether they represent two separate points of innovation.
46
Björn Wiemer
A case in point might be the Slavic preposition za + ACC with the initial meaning ‘instead, in place of’ (e.g., Polish Janek wypłukał brudne naczynia za mnie ‘Janek washed the dirty dishes instead of me’) which marks the basis of comparison in West Ukrainian (Carpathian region) and in the Slavic dialects in close contact with Lithuanian and Latgalian.17 The two spots are not too far away from each other in terms of physical space, but we may wonder whether (i) they have developed independently from each other or whether (ii) their parallel occurrence reflects earlier diffusion from one of these spots to the other, or (iii) from another hotbed between them that has vanished in the meantime. Hypothesis (ii) seems not improbable if one relies on the data in the AUM (1988): its map 261 shows that the diffusion zone of za + ACC as comparison marker stretches quite consistently from the northeastern edges of the Carpathian mountains to the north into Polissja, i.e., the border region with Belarusian, while it appears to be only scarcely attested in the Carpathian mountains themselves. In fact, za + ACC is attested in the more easterly parts of the Ukraine as well (Danylenko 2006: 280–297) and was found in various Ukrainian regions by Hujenko (1930: 57–58). At present, this leaves us with the whole issue to be settled anew. Another case in point may be the parallel occurrence of DAT-INS-syncretisms in the plural of nouns (and adjectives). We encounter them in some Northwest Russian dialects (e.g., s vёdr-am ‘with buckets’, sva-im glaz-am videl ‘(I) saw (it) with my (own) eyes’; cf. Bromlej and Bulatova 1972: 99; Bukrinskaja, Karmakova, and Ter-Avanesova 2008: 123) as well as in Serbian and Croatian, where it further involves the locative (with the ending {ima}). However, this phenomenon can hardly have arisen from a common hotbed, for the following reasons: (i) there is too much space and too many dialects (without any traits of this syncretism) in between; (ii) the endings which originated in some earlier variation, are different ({am} vs. {ima}); (iii) the DAT/INS/LOC-syncretism in Serbian and Croatian fits well into a South Slavic cline of the loss of case (Slovene having preserved all cases, Balkan Slavic having lost almost all of them), while the DAT/INS-syncretism in the Russian dialects has parallels in Baltic.18 In this case, it seems more justified to assume polygenesis here than in the case of za+ACC as marker of the basis of comparison.
17 As for the Baltic-Slavic contact region cf. Wiemer (2004a: 505–506), Nau (2011: 72–73). In Wiemer (2004a) this preposition was still (as in many other sources) erroneously identified with the meaning ‘behind’ (which would govern another case). 18 Compare, for instance, colloquial Lithuanian Krintantį vaiką sugriebiau ab-iem rank-om (is) ‘I caught the falling child with both hands’. The last syllable is often omitted so that the difference in the endings of the dative plural and the instrumental plural is nivellated. Again, it is a separate question whether this coincidence of dative-instrumental-syncretism rests on
1 Matrёška and areal clusters involving varieties of Slavic
47
5.3 Reflexive-reciprocal polysemy The third case study concerns contemporary Russian (and probably all East Slavic) and how it differs from the West and South Slavic languages in its limited range of a reciprocal meaning of verbs marked with the reflexive marker (RM). If we distinguish between the “heavy” and “light” markers (Kemmer 1993) and between canonical and natural reciprocals (Nedjalkov 2007a), we find that the Russian light marker -sja/-s’ (allomorphs of a postfix) indicates reciprocity only with natural reciprocals, such as obnimat’sja ‘to embrace each other’, celovat’sja ‘to kiss each other’, vstretit’sja ‘to meet each other’. Canonical reciprocals need a complex marker of the each other-type (drug druga (ACC/GEN) / drug-u (DAT), and the like, with the second component inflected for case); compare, for instance, bojat’sja drug drug-a (GEN) ‘to fear each other’, zavidovat’ drug drug-u (DAT) ‘to envy each other’, soglašat’sja drug s drug-om (INS) ‘to agree with each other’. This complex marker is a recent development, while in earlier East Slavic the light RM was used for a larger class of reciprocals (Zarickij 1961: 33–34). All West Slavic languages still reflect this earlier stage; moreover, they have an extensive range of verbs (with different types of valency structures) whose RM-derivatives allow for variable interpretations in the domain of reciprocity and reflexivity proper. For instance, in Polish we notice regular polysemy (or vagueness) of reciprocal vs. reflexive meaning not only with the light RM (the enclitic się), but even with the heavy marker (siebie (ACC/GEN), sobie (DAT), etc.); see examples (10) and (11). “Regular” means that many verbs which are not inherently oriented toward a reciprocal situation can be used as nuclei of clauses which together with the RM and a nominative subject in the plural code either canonical reciprocals or reflexives (with possible intermediate interpretations). Interlocutors derive the appropriate interpretation from the given linguistic and/or situational context. Compare, for instance, Polish (9) Koledz-y karci-l-i się za zł-ą colleagues-NOM reproach.[IPFV]-PST.3PL RM for bad-ACC.SG.F prac-ę. work[.F]-ACC.SG ‘The colleagues reproached (i) one another, (ii) themselves for bad work.’
a parallel, but independent development. Lithuanian dialects and Russian in the Pskov region are not directly connected. Whether there was a connection in previous times, and whether nonetheless this kind of syncretism did not develop independently, remains to be studied.
48
Björn Wiemer
(10) Małżonk-owie kupi-l-i sobie prezent-y. spouses-NOM.PL buy[.PFV]-PST.3PL REFL.DAT present-ACC.PL ‘The spouses bought presents (i) for each other, (ii) for themselves.’ (11) Przyjaciel-e opowiada-l-i o sobie. friends-NOM.PL tell[.IPFV]-PST.3PL about REFL.LOC ‘The friends told stories about (i) each other, (ii) themselves.’ This process is still productive (Wiemer 1999, 2007b), and analogous examples could be supplied for any of the other West Slavic languages. The regular reciprocal-reflexive polysemy positions itself into a larger cline stretching westwards, as it comprises German and French (and possibly the rest of continental West Germanic and the large part of Romance).19 This polysemy applies, for instance, to the following Polish and German equivalents when the RM is added (and the subject is in the plural): Polish (imperfective verbs) German (12) pchać ⇒ pchać się głaskać ⇒ głaskać się czesać ⇒ czesać się pobudzać ⇒ pobudzać się dotykać ⇒ dotykać się chwalić ⇒ chwalić się bronić ⇒ bronić się
stoßen ⇒ sich stoßen streicheln ⇒ sich streicheln kämmen ⇒ sich kämmen anspornen ⇒ sich anspornen berühren ⇒ sich berühren loben ⇒ sich loben verteidigen ⇒ sich verteidigen
‘to push’ ‘to caress’ ‘to comb’ ‘to encourage’ ‘to touch’ ‘to praise’ ‘to defend’
We thus get a west-east belt in the middle of the European continent. English is not included, because of the radical decrease of the RM as a marker of recessive diatheses there (in terms of Geniušienė 1987). Natural reciprocals are not marked with the RM, either. Compare, for instance, They met at the railway station; The two samples matched; The roads crossed; They kissed; They argued / quarrelled / agreed). But canonical reciprocals are consistently marked with a complex device (each other, one another) very much resembling the one we know from contemporary standard Russian (see above) and from other East Slavic (e.g., Ukranian ljubyty odyn
19 The area of the reciprocal-reflexive polysemy probably spreads over the larger part of continental Europe. It is attested also in Bulgarian; see the language-specific contributions in volume 2 of Nedjalkov (2007). How consistently this area has evolved should be investigated more thoroughly.
1 Matrёška and areal clusters involving varieties of Slavic
49
(NOM) odnoho (ACC) ‘to love one another’) as well as from Baltic varieties. Thus, English and (probably) most of East Slavic standard and non-standard varieties are set off from a “middle-European continental belt” in that they use morphosyntactic means to consistently distinguish canonical reciprocals from both natural reciprocals and true reflexives. English and East Slavic are beyond the west and the east end of this area, respectively, but their coincidence pertains only to the consistent special treatment of canonical reciprocals, while the diachronic reasons evidently differ. Now it may seem that we have clear areal cuts (just like ideal isoglosses) in a west-east direction across the middle of Europe. On this background, however, Lithuanian displays an intermediary stage. Like modern East Slavic and English, it sets off canonical reciprocals from reflexives by an obligatory complex marker with the same internal structure as in East Slavic (vien-as kit-ą, vien-as kit-am, etc. ‘one the other’—the first part, in the nominative, is inflected for gender and number as well); compare canonical reciprocals as in (13a–c) with natural reciprocals like bučiuotis ‘to kiss each other’, matytis ‘to see each other, to meet’, ribotis ‘to border on each other’: Lithuanian (13a) Draug-ai kritikav-o vien-as kit-ą. friend-NOM.PL criticize-PST.3 one-NOM.SG other-ACC.SG ‘The friends criticized one another.’ (13b) Sutuoktini-ai pirk-o vien-as kit-am dovan-as. spouse-NOM.PL buy-PST.3 one-NOM.SG other-DAT.SG present-ACC.PL ‘The spouses bought each other presents.’ (13c) Jų reikal-ai vien-i nuo kit-ų priklaus-o. 3PL.GEN matter-NOM.PL one-NOM.PL from other-GEN.PL depend-PRS.3 ‘Their matters depend on each other.’ Moreover, like Russian, Lithuanian no longer realizes a regular derivation of verbs with the light RM (-si-/-s) which would allow for either a reciprocal or a reflexive interpretation (as this is the case in West Slavic, German and French; see above). This concerns also benefactive functions of the RM, which in Lithuanian is equally widespread as, for instance, in German and Italian or applies also to the Polish heavy RM sobie. Thus, according to Geniušienė (2007: 641), a sentence like Jie nu-si-pirko knygų can have only a reflexive interpretation, not a reciprocal one (‘They bought themselves / *each other (some)
50
Björn Wiemer
books’), as opposed to the formal equivalent in German Sie kauften (für) sich Bücher ‘They bought books (i) for each other / (ii) for themselves’ or in Polish; see (10). In general, contemporary Lithuanian (like Russian) marks true reflexivity distinctly from canonical reciprocals. Concomitantly, the light RM hardly ever yields a proper reflexive meaning,20 this meaning requires the heavy RM (inflected for case: save (ACC), savęs (GEN), sau (DAT), and so on), comparable to Russian sebja (ACC/GEN), sebe (DAT), and so on: Lithuanian Jis kritikuoja save (*kritikuoja-si), Russian On kritikuet sebja ‘He criticizes himself’ (≠ kritikuet-sja, which can only be understood as a passive: ‘is (being) criticized’). Therefore, “reflexive derivation of reciprocals is not an active process in Lithuanian in the sense that no new reflexive reciprocals are formed” (Geniušienė 2007: 663), and the number of RM-verbs with a reciprocal-reflexive polysemy has practically been reduced to zero. Nonetheless, in terms of type frequency, natural reciprocals with the light RM are still more prominent than in Russian. Unfortunately, at the moment I am unable to provide precise figures showing the number of natural reciprocals with the light RM in Lithuanian in comparison to Russian. But in Lithuanian natural reciprocals are derived even from intransitive verbs as, for instance, from talkinėtis ‘to support each other’ (< talkinėti kam (DAT) ‘to help somebody’), šnibždėtis ‘to whisper into each other’s ear’ (< šnibždėti kam (DAT) ‘to whisper into somebody’s ear’) with the dative, or pyktis ‘to be cross with each other’ (< pykti ant ko (GEN) ‘to be angry at somebody’), pasižvilgčioti ‘to lurk at each other’ (< pažvilgčioti į ką (ACC) ‘to lurk, look at somebody’) with a PP. Apart from that, the semantic relation to the corresponding non-RM-verb is always transparent, i.e., Lithuanian has virtually no reciproca tantum, in contrast to Russian (Geniušienė 2007: 644–646, 671). Compare the following equivalences (# indicates that the verb with RM exists, but in other than reciprocal meaning): (14) Lithuanian sveikinti ⇒ ‘to welcome, greet’
Russian *zdorovat’
sveikintis ‘to exchange greetings’ pravardžiuoti ⇒ prasivardžiuoti —— ‘to call by nickname’ ‘to call each other by nicknames’
⇒
zdorovat’sja ‘to exchange greetings’
20 An exception would be pri-si-versti ‘to force oneself’ (priversti ‘to force’), which is practically synonymous to priversti save ‘idem’. There is also a limited (and now closed) class of grooming verbs (e.g., rengti-s ‘to dress oneself’), which could better be treated as autocausatives.
1 Matrёška and areal clusters involving varieties of Slavic
riboti ‘to border’
⇒
prašyti ⇒ ‘to ask, request’ siekti ⇒ ‘to reach for’
ribotis ‘to border (on each other)’ prašytis ‘to ask each other’ siektis ‘to reach, touch each other’
51
graničit’ ⇒ *graničit’sja ‘to border’ prosit’ ⇒ #prosit’sja (kuda-libo) ‘to ask, request’ ‘to ask, apply for’ (autocausative) dostigat’ ⇒ #dostigat’sja ‘to reach, achieve’ (only passive)
All these details about RM-marked valency changes amount to an important peculiarity which makes Lithuanian, once again, stand out in areal terms: even in the grey zone between canonical and natural reciprocals, one hardly finds items showing an overlap with the reflexive function, although the derivational relation to the transitive bases has remained transparent and there is no conceptual reason why a reflexive interpretation should not be possible. We observe an almost complementary distribution between reciprocal and reflexive function of RM-verbs (Geniušienė 2007: 652). Consider, for instance, a Lithuanian pair of quasi-synonyms in complementary distribution: Lithuanian (15a) J-ie pa-si-ragin-o skubė-ti. PFX-RM-encourage-PST.3 hurry-INF 3-NOM.PL.M (only reciprocal interpretation) ‘They encouraged each other / *themselves to hurry.’ (15b) Pa-si-drąsin-ome ei-ti į rūs-į. PFX-RM-encourage-PST.1PL go-INF in(to) cellar-ACC (only reflexive interpretation) ‘We encouraged ourselves / *each other to go into the cellar.’ This pair can be compared with its translational equivalents in Polish (16) and German (17), which show reciprocal-reflexive polysemy (‘∨’ indicates inclusive disjunction): Polish (16a) Oni pobudzali się, by iść szybciej. reciprocal ∨ reflexive ‘They encouraged each other ∨ themselves to go faster.’
52
Björn Wiemer
(16b) Zachęcaliśmy się do zejścia w piwnicę. reciprocal ∨ reflexive ‘We encouraged each other ∨ ourselves to go into the cellar.’ German (17a) Sie drängten sich zur Eile. reciprocal ∨ reflexive ‘They urged each other ∨ themselves to hurry up.’ (17b) Wir machten uns Mut, in den Keller zu gehen. reciprocal ∨ reflexive ‘We encouraged each other ∨ ourselves to go into the cellar.’ Compare further with Russian, where reciprocal and reflexive meaning need to be marked explicitly and distinctly (‘/’ indicates different interpretations): (18a) Oni toropili drug druga / (každyj) sebja (samogo). ‘They urged themselves / each other to hurry up.’ reciprocal / reflexive (18b) My podtalkivali drug druga / (každyj) sebja (samogo), čtoby spustit’sja v podval. ‘We pushed ourselves / each other to go down into the cellar.’ reciprocal / reflexive Other Lithuanian pairs of RM-verbs with a complementary distribution of reciprocal and reflexive meaning are: (19a) Jie glamonėja-si. only reciprocal (vs. glamonėja save ‘carress themselves’) ‘They are caressing each other.’ (19b) Jie lepina-si. only reflexive (vs. lepina vienas kitą ‘pamper each other’) ‘They are coddling/pampering themselves.’ (20a) Žiūrovai stumdo-si. only reciprocal ‘The spectators are shoving (and justling) one another.’ (20b) Jie stumia-si į priekį. only reflexive (or rather autocausative) ‘They nose forward.‘ (21a) Jie pliekia-si (dvare). only reciprocal ‘They are lashing each other (in the yard).’ (21b) Jie peria-si (pirtyje). mostly reflexive ‘They are birching themselves / ?each other (in the sauna).’
1 Matrёška and areal clusters involving varieties of Slavic
53
This almost complementary distribution of reflexive vs. reciprocal RM-verbs in the lexicon seems to be quite unique in terms of the closer areal surroundings of Lithuanian, and it is tempting to explain this finding as the synchronic result of two intersecting tendencies: The first one amounts to a shrinkage of lexical RM-marked reciprocals, a trend which can be observed at least in East Slavic, but the number of RM-marked lexical reciprocals decreases also to the north, namely in Latvian (Lithuanian’s closest relative) and in Finnish (as well as in the Scandinavian languages); cf. Geniušienė (2007: 671). The other tendency consists in the preservation of reciprocal-reflexive polysemy in the area (south) west from Lithuanian (stretching from Polish at least to French). These two tendencies constitute spread zones which intersect in Lithuanian. They lead to an areal superposition (i.e., overlap) of two opposed tendencies and let Lithuanian look as if it forms the western periphery of a zone in which an innovative pattern (i.e., different treatment of canonical reciprocals against reflexive meaning) has been spreading or, inversely, the eastern periphery of a zone which, in this respect, shows conservative behaviour. This peculiar areal status of Lithuanian becomes visible only if we look at the concentric surroundings of Lithuanian in both directions, to the (north)east and to the (south)west (for each, so to say, in a matrёška-fashion, or like a Venn diagram). This has not been done systematically here, because data for many languages in either direction are lacking. On this background, we might furthermore realize that the areally outstanding synchronic situation of Lithuanian most probably reflects a lexically restricted residuum of RM-marked reciprocals which are otherwise replaced by a more distinct device of the each other-type. Russian seems to be fairly distinct from the rest of Slavic, which, rather, display reciprocal-reflexive polysemy. The extent to which the Russian situation holds true for the rest of East Slavic remains to be studied.
6 Summary, conclusions and postulates In this chapter I discussed some of the problems which linguists face when trying to establish areal clines, hotbeds of diffusion, and language types based on areal and genealogical criteria. These considerations started from the question, what changes if, instead of some “mainstream area” (or feature cluster) like SAE, we postulate an alternative area, choose different criteria, and shift hotbeds of diffusion because criteria have changed. Does this change the methodology? And what makes an area (or feature cluster) A1 more appropriate than some other area A2? Among other things, one sometimes gets the impression
54
Björn Wiemer
that spread zones and their hotbeds in Europe, especially if they are established in relation to features associated with SAE-type languages, are motivated by other than linguistic reasons. This impression might have given rise to the following complaint: “(. . .) there seems to be a tradition of identifying the eastern border of European civilization with that of the author’s home country (or, in a self-castigating way peculiar to many small nations, with its western border). This ideologically loaded use of the term European is all too easily extended to languages as well (. . .)” (Laakso 2002: 235). Moreover, I tried to point out some problems in determining the relation between a merely synchronic account of observations (i.e., the circumstantialist approach) and the diachronic processes which have led to convergence, or divergence, in some particular area (“What happened?”). I hope this study contributes to a serious examination of the basis of (areal) linguistic objectification of assumptions concerning the position of Slavic languages in Europe. This position should, of course, be unbiased by any (implicit or explicit) political ideology. Apart from this, one must free oneself from the untenable (often tacit) implication that the word Slavic refers to a monolithic or easily definable notion. The same applies, a fortiori, to the notion European. Linguistic varieties called Slavic are extremely diversified among one another, certainly no less than Germanic or Romance languages (or Semitic, or FinnoUgric, etc.), so that hardly can any one of them be regarded as the “best representative” of Slavic. As far as I know, the extent to which the degree of diversity within Slavic, or the degree of differences against some background of, say, “typical SAE languages” (or features), may be quantified (and thus substantiated), has never been stated in a methodologically impeccable manner. An impeccable way of establishing such distances requires, among other things, that one does not impute the result of one’s differences, areal clusterings, or any typicality judgments on preconceived criteria. Otherwise one will get what one implied from the start, i.e., areal patterns will turn out as “the product of the criteria used to identify them” (Heine and Nomachi 2011: 4). This applies even from a merely circumstantialist viewpoint: the structure of areal diffusion cannot be adequately captured if one does not look at it from the angle of various criteria, some of which I have mentioned in this chapter. And, of course, if we have at our disposal sufficiently reliable data (or their faithful reconstructions) from earlier stages, we ought to take advantage of them in order to first establish what really happened in each particular case. This should be done, however, with an eye to more widely known tendencies in order not to arrive at judgments about isolated phenomena that can hardly be generalized.
1 Matrёška and areal clusters involving varieties of Slavic
55
As trivial as these conclusions essentially are, it should have become clear that the investigation of convergence (or clusters, clines, and so forth) requires the application of quantifying methods. Since methods of quantification in general have only been touched upon here, this remark may be taken as an outlook for future research. Quantifying methods have been implied, but rarely implemented, by referring to minor use patterns, which are the locus of change. It has also been admitted that change consists in an increase of frequency, functional extension, and other parameters that are gradual and for which we need to know critical thresholds beyond which quantity turns into quality (of change, or growth of (dis)similarity). Concomitantly, some sort of variation is always present and some immediate surrounding is always necessary in order to notice salient patterns or convergence zones. Quantification is therefore necessary in order to make these patterns palpable and to segregate them from everyday variable usage as it occurs in any speech community countless times without any discernible consequences for change. If this is agreed upon, the question “What happened?” involves some other queries which should be answered even before we deal with this big question. To these implied queries belong: What should we count? Types or tokens? And at what level of granularity?
Abbreviations ACC ADES ANT CBA DAM DAT GEN INF INS IPFV LOC NEG NOM PFX PFV PL PP PRF
accusative adessive indeclinable anteriority participle Circum-Baltic area differential argument marking dative genitive infinitive instrumental imperfective locative negation nominative prefix perfective plural prepositional phrase perfect
56
PRS PST REFL RM SAE SAIE SG
2 3
Björn Wiemer
present past non-enclitic reflexive pronoun reflexive marker (clitic or agglutinated) Standard Average European Standard Average Indo-European singular second person third person
Acknowledgments: First of all, I want to thank Andrii Danylenko and Motoki Nomachi for their helpful remarks on a first draft of this chapter. I am also obliged to Bridget Drinka for valuable comments and a thorough proofreading of two versions of this paper. Furthermore, I want to thank Birutė Spraunienė, Jolanta Šinkūnienė, Vaiva Žeimantienė and Aistė Žemaitytė for their native speaker judgments on Lithuanian examples, Christoph Unger for the discussion of Estonian data as well as Svetlana Edygarova and Rogier Blokland for their consultation on Udmurt examples. Marfa Tolstaja and Andrii Danylenko supplied me with valuable information concerning Ukrainian dialects. Of course, the entire responsibility for mistakes or misunderpretations rests on me.
References Abondolo, Daniel (ed.). 1998. The Uralic Languages. London/New York: Routledge. Andersen, Henning. 2009. On the origin of the Slavic aspects: Questions of chronology. In: Vit Bubenik, John Hewson, and Sarah Rose (eds.), Grammatical Change in Indo-European languages (Papers Presented at the Workshop on Indo-European linguistics at the XVIIIth International Conference on Historical Linguistics, Montreal, 2007), 123–140. Amsterdam/ Philadelphia: John Benjamins. AUM – Atlas ukrajins’koji movy [Atlas of the Ukrainian Language], Vol. 2. Kyiv: Naukova dumka, 1988. Bickel, Balthasar, and Johanna Nichols. 2006. Oceania, the Pacific Rim, and the theory of linguistic areas. In: Zhenya Antić et al. (eds.), Proceedings of the 32nd Annual Meeting of the Berkeley Linguistics Society: General Session and Parasession on Theoretical Approaches to Argument Structure. Berkley, CA: Berkley Linguistic Circle. Bisang, Walter. 2004. Dialectology and typology – An integrative perspective. In: Bernd Kortmann (ed.), Dialectology Meets Typology (Dialect Grammar from a Cross-Linguistic Perspective), 11–45. Berlin/New York: Mouton de Gruyter. Bisang, Walter. 2006a. Contact-induced convergence: Typology and areality. In: Keith Brown (ed.), Encyclopedia of Language and Linguistics, Vol. 3, 88–101. Oxford: Elsevier.
1 Matrёška and areal clusters involving varieties of Slavic
57
Bisang, Walter. 2006b. Linguistic areas, language contact and typology: Some implications from the case of Ethiopia as a linguistic area. In: Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas. Convergence in Historical and Typological Perspective, 75–98. Houndmills/New York: Palgrave Macmillan. Bisang, Walter. 2010. Areal language typology. In: Peter Auer and Jürgen E. Schmidt (eds.), Language and Space. An International Handbook of Linguistic Variation, Vol. 1: Theories and Methods, 419–440. Berlin/New York: Mouton de Gruyter. Bossong, Georg. 1998. Le marquage différentiel de l’objet dans les langues d’Europe. In: Jack Feuillet (ed.), Actance et valence dans les langues de l’Europe, 193–258. Berlin/New York: Mouton de Gruyter. Breu, Walter. 1994. Der Faktor Sprachkontakt in einer dynamischen Typologie des Slavischen. In: Hans-Robert Mehlig (ed.), Slavistische Linguistik 1993, 41–64. Munich: Otto Sagner. Breu, Walter. 1998. Romanisches Adstrat im Moliseslavischen. Die Welt der Slaven 43: 339–354. Breu, Walter. 2005. Verbalaspekt und Sprachkontakt. Ein Vergleich der Systeme zweier slavischer Minderheitensprachen (SWR/MSL). In: Sebastian Kempgen (ed.), Slavistische Linguistik 2003, 37–95. Munich: Otto Sagner. Bromlej, Sof’ja Vladimirovna and Lidija Nikolaevna Bulatova. 1972. Očerki morfologii russkix govorov [Studies on the Morphology of Russian Dialects]. Moscow: Nauka. Berger, Claudia Bucheli, Elvira Glaser, and Guido Seiler. 2012. Is a syntactic dialectology possible? Contributions from Swiss German. In: Andrea Ender, Adrian Leemann, and Bernhard Wälchli (eds.), Methods in Contemporary Linguistics, 93–119. Berlin/New York: John Benjamins. Bukrinskaja, Irina Anatol’evna, Ol’ga Evgen’evna Karmakova, and Aleksandra Valer’evna TerAvanesova. 2008. Govory belorussko-russkogo pogranič’ja [Dialects of the BelarusianRussian border region]. In: Ljudmila Ėduardovna Kalnyn’ (ed.), Issledovanija po slavjanskoj dialektologii 13: Slavjanskie dialekty v situacii jazykovogo kontakta (v prošlom i nastojaščem), 118–179. Moscow: Izdatel’stvo RAN. Bybee, Joan, Revere Perkins, and William Pagliuca. 1994. The Evolution of Grammar (Tense, Aspect, and Modality in The languages of the World). Chicago/London: The Chicago University Press. Campbell, Lyle. 2006. Areal linguistics: A closer scrutiny. In: Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas. Convergence in Historical and Typological Perspective, 1–31. Houndsmill/New York: Palgrave Macmillan. Csúcs, Sándor. 1998. Udmurt. In: Daniel Abondolo (ed.), The Uralic languages. London/ New York: Routledge, 276–304. Coseriu, Eugenio. 1988. Die Begriffe “Dialekt”, “Niveau” und “Sprachstil” und der eigentliche Sinn der Dialektologie. In: Jörn Albrecht, Jens Lüdtke, and Harald Thun (eds.), Energeia und Ergon (Sprachliche Variation – Sprachgeschichte – Sprachtypologie). Studia in honorem Eugenio Coseriu, 15–43. Tübingen: Narr. Dahl, Östen. 2000. The grammar of future time reference in European languages. In: Östen Dahl (ed.), Tense and Aspect in the Languages of Europe, 309–328. Berlin/New York: Mouton de Gruyter. Dahl, Östen. 2001. Principles of areal typology. In: Martin Haspelmath, Ekkehard König, Wolfgang Oesterreicher, and and Wolfgang Raible (eds.), Language Typology and Language Universals. An International Handbook, Vol. 2, 1456–1470. Berlin/New York: Mouton de Gruyter. Danylenko, Andrii. 2006. Slavica et Islamica. Ukrainian in Context. Munich: Otto Sagner.
58
Björn Wiemer
Danylenko, Andrii. 2013. Ukrainian in the language map of Central Europe: Questions of areal-typological profiling, Journal of Language Contact 6 (1): 134–159. De Vogelaer, Gunther and Guido Seiler. 2012. The Dialect laboratory: Introductory remarks. In: Gunther De Vogelaer and Guido Seiler (eds.), The Dialect Laboratory: Dialects as a Testing Ground for Theories of Language Change, 1–31. Amsterdam/Philadelphia: John Benjamins. Drinka, Bridget. 2017. Language Contact in Europe: The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Geniušienė, Emma. 1987. The Typology of Reflexives. Berlin/New York: Mouton de Gruyter. Geniušienė, Emma. 2007. Reciprocal and reflexive constructions in Lithuanian (with references to Latvian). In: Vladimir P. Nedjalkov (ed.), Reciprocal Constructions, Vol. 2, 633–672. Amsterdam/Philadelphia: John Benjamins. Hansen, Björn. 2001. Das slavische Modalauxiliar. Semantik und Grammatikalisierung im Russischen, Polnischen, Serbischen/Kroatischen und Altkirchenslavischen. Munich: Otto Sagner. Hansen, Björn. 2005. How to measure areal convergence: a case study of contact-induced grammaticalization in the German-Hungarian-Slavonic contact area. In: Björn Hansen and Petr Karlík (eds.), Modality in Slavonic Languages. New Perspectives, 219–237. Munich: Otto Sagner. Haspelmath, Martin. 1998. How young is Standard Average European? Language Sciences 20 (3): 271–287. Haspelmath, Martin. 2001. The European linguistic area: Standard Average European. In: Martin Haspelmath, Ekkehard König, Wolfgang Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Universals. An International Handbook, Vol. 2, 1492–1510. Berlin/New York: Walter de Gruyter. Haspelmath, Martin. 2004. Passive participles across languages. In: Barbara Fox and Paul J. Hopper (eds.), Voice. Form and Function, 151–177. Amsterdam/Philadelphia: John Benjamins. Heine, Bernd and Tania Kuteva. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva. 2006. The Changing Languages of Europe. Oxford: Oxford University Press. Heine, Bernd and Motoki Nomachi. 2011. Is Europe a linguistic area? In: Motoki Nomachi (ed.), Grammaticalization in Slavic Languages. From Areal and Typological Perspectives, 1–26. Sapporo: Slavic Research Center, Hokkaido University. Holvoet, Axel. 1995. On the evolution of passive in Lithuanian and Latvian. In: Wojciech Smoczyński (ed.), Analecta Indoeuropaea Cracoviensia. I. Safarewicz memoriae dedicata, 173–182. Cracow: Universitas. Holvoet, Axel. 2007. Mood and modality in Baltic. Cracow: Wydawnictwo Uniwersytetu Jagiellońskiego. Hujenko, S. H. 1930. Pryjmennyk za v ukrajins’kyx dijalektax [The preposition za in Ukrainian dialects]. In: Olena Kurylo (ed.), Zbirnyk sekciji hramatyky ukrajins’koji movy, Vol. 1, 41–71. Kyiv: Naukovo-doslidnyj instytut movoznavstva VUAN. Kemmer, Suzanne. 1993. The Middle Voice. Amsterdam/Philadelphia: John Benjamins. Koptjevskaja-Tamm, Maria. 2006. The circle that won’t come full: two potential isoglosses in the Circum-Baltic area. In: Yaron Matras, April McMahon, and Nigel Vincent (eds.),
1 Matrёška and areal clusters involving varieties of Slavic
59
Linguistic Areas. Convergence in Historical and Typological Perspective, 182–226. Houndsmill/New York: Palgrave Macmillan. Koptjevskaja-Tamm, Maria and Bernhard Wälchli. 2001. The Circum-Baltic languages. An areal-typological approach. In: Östen Dahl and Maria Koptjevskaja-Tamm (eds.), CircumBaltic Languages, Vol. 2: Grammar and Typology, 615–761. Amsterdam/Philadelphia: John Benjamins. Laakso, Johanna. 2002. Finno-Ugrians vs. European: problems of European areal typology from a finno-ugrist’s point of view. In: Rogier Blokland and Cornelius Hasselblatt (eds.), Finno-Ugrians and Indo-Europeans: Linguistic and Literary Contacts (Proceedings of the Symposium at the University of Groningen, November 22–24,2001), 234–245. Maastricht: Shaker. Levshina, Natalia. 2015. European analytic causatives as a comparative concept. Evidence from a parallel corpus of film subtitles. Folia Linguistica 49 (2): 487–520. Lewy, Ernst. 1942. Der Bau der europäischen Sprachen. (Proceedings of the Royal Irish Academy. 48.) Dublin: Hodges and Figgis. Lukežić, Iva. 2015. Zajednička povijest hrvatskih narječja [The Common History of Croatian Dialects], Vol. 2: Morfologija [Morphology]. Zagreb: Hrvatska sveučilišna naklada. Majsak, Timur A. 2005. Tipologija grammatikalizacii konstrukcij s glagolami dviženija i glagolami pozicii [The Typology of the Grammaticalization of Constructions with Verbs of Motion and Verbs of Position]. Moscow: Jazyki slavjanskoj kul’tury. Nau, Nicole. 2011. A Short Grammar of Latgalian. Munich: LINCOM Europa. Nau, Nicole. 2012. Modality in an areal context: The case of a Latgalian dialect. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 465–508. (Trends in Linguistics 242.) Berlin/ New York: Mouton de Gruyter. Nedjalkov, Vladimir Petrovič. 2007a. Overview of the research (Definitions of terms, framework, and related issues). In: Vladimir Petrovič Nedjalkov (ed.), Reciprocal constructions, Vol. 1. Amsterdam/Philadelphia: John Benjamins, 3–114. Nedjalkov, Vladimir Petrovič (ed.) 2007b. Reciprocal Constructions, Vol. 1–2. Amsterdam/ Philadelphia: John Benjamins. Polański, Kazimierz. 2002. Polabian. In: Bernard Comrie and Greville G. Corbett (eds.), The Slavonic Languages, 795–824. London/New York: Routledge. Ramovš, Fran. 1935. Historična gramatika slovenskega jezika [The Historical Grammar of the Slovenian Language], Vol. 7: Dialekti [Dialects]. Ljubljana: Učiteljska tiskarna. Sakel, Jeanette. 2007. Types of loan: Matter and pattern. In: Yaron Matras and Jeanette Sakel (eds.), Grammatical Borrowing in Cross-Linguistic Perspective, 15–29. Berlin/New York: Mouton de Gruyter. Serebrennikov, Boris Aleksandrovič. 1960. Kategorii vremeni i vida v finno-ugorskix jazykax permskoj i volžskoj grupp [The Categories of Tense and Aspect in the Permian and Volgaic Groups of the Finno-Ugric Languages]. Moscow: Izdatel’stvo AN SSSR. Seržant, Ilja. 2012. The so-called possessive perfect in North Russian and the Circum-Baltic area. A diachronic and areal account. Lingua 122: 356–385. Shevelov, George Y. 1964. A Prehistory of Slavic. The Historical Phonology of Common Slavic. Heidelberg: Carl Winter. Skalička, Vladimír. 1968. Zum Problem des Donausprachbundes. Ural-altaische Jahrbücher 40: 3–9.
60
Björn Wiemer
Thieroff, Rolf. 2000. On the areal distribution of tense-aspect categories in Europe. In: Östen Dahl (ed.), Tense and Aspect in the Languages of Europe, 265–305. Berlin/New York: Mouton de Gruyter. Thomas, George. 2008. Exploring the parameters of a Central European Sprachbund. Canadian Slavonic Papers 50 (1–2): 123–153. Thomason, Sarah Grey and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley/Los Angeles, CA: University of California Press. Trubetzkoy, Nikolaj Sergeevič. 1930. Proposition 16. Actes du premier Congrès international de linguistes: à La Haye, du 10–15 avril 1928, 17–18. Leiden: A.W. Sijthoff. van der Auwera, Johan. 1998. Revisiting the Balkan and Meso-American linguistic areas. Language Sciences 20: 259–270. van der Auwera, Johan, Andreas Ammann, and Saskia Kindt. 2005. Modal polyfunctionality and Standard Average European. In: Ales Klinge and Henrik Hoeg Müller (eds.), Modality. Studies in Form and Function, 295–322. London: Equinox. van Pottelberge, Jeroen. 2001. Sprachbünde: Beschreiben sie Sprachen oder Linguisten? Linguistik online 8, 1/01: 1–26. Von Waldenfels, Ruprecht. 2012. The Grammaticalization of ‘give’ + Infinitive. A Comparative Study of Russian, Polish and Czech. Berlin/New York: Mouton de Gruyter. Wälchli, Bernhard. 2012. Grammaticalization clines in space: Zooming in on synchronic traces of diffusion processes. In: Björn Wiemer, Wälchli Bernhard, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 233–272. (Trends in Linguistics 242.) Berlin/New York: Mouton de Gruyter. Wiemer, Björn. 1999. The light and the heavy form of the Polish reflexive pronoun and their role in diathesis. In: Katharina Böttger, Markus Giger, and Björn Wiemer (eds.), Beiträge der Europäischen Slavistischen Linguistik (POLYSLAV) 2, 300–313. Munich: Otto Sagner. Wiemer, Björn. 2004a. Population linguistics on a micro-scale. Lessons to be learnt from Baltic and Slavic dialects in contact. In: Bernd Kortmann, Bernd (ed.), Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective, 497–526. Berlin/ New York: Mouton de Gruyter. Wiemer, Björn. 2004b. The evolution of passives as grammatical constructions in Northern Slavic and Baltic languages. In: Walter Bisang, Nikolaus P. Himmelmann, and Björn Wiemer (eds.), What Makes Grammaticalization? A Look from its Fringes and its Components, 271–331. Berlin/New York: Mouton de Gruyter. Wiemer, Björn. 2006a. Grammatical evidentiality in Lithuanian (a typological assessment). Baltistica 41 (1): 33–49. Wiemer, Björn. 2006b. Relations between Actor-demoting devices in Lithuanian. In: Werner Abraham and Larissa Leisiö (eds.), Passivization and Typology: Form and Function, 274–309. Amsterdam/Philadelphia: John Benjamins. Wiemer, Björn. 2007a. Kosvennaja zasvidetel’stvovannost’ v litovskom jazyke [Indirect evidentiality in Lithuanian]. In: Xrakovskij, Viktor S. (ed.), Ėvidencial’nost’ v jazykax Evropy i Azii. Sbornik statej pamjati Natalii Andreevny Kozincevoj [Evidentiality in the Languages of Europe and Asia. Collection of Articles to the Memory of Natalia Andreevna Kozinceva], 197–240. Sankt-Peterburg: Nauka. Wiemer, Björn. 2007b. Reciprocal and reflexive constructions in Polish. In: Nedjalkov, Vladimir P. (ed.), Typology of Reciprocal Constructions, Vol. 2, 514–559. Amsterdam/ Philadelphia: John Benjamins.
1 Matrёška and areal clusters involving varieties of Slavic
61
Wiemer, Björn. 2011. The grammaticalization of passives. In: Bernd Heine and Heiko Narrog (eds.), Handbook of Grammaticalization, 535–546. Oxford: Oxford University Press, Wiemer, Björn. 2014. Umbau des Partizipialsystems. In: Tilman Berger, Karl Gutschmidt, Sebastian Kempgen, and Peter Kosta (eds.), Slavische Sprachen. Ein internationales Handbuch zu ihrer Struktur, ihrer Geschichte und ihrer Erforschung, Vol. 2 (Reihe HSK 32.2.), 1625–1652. Berlin/Boston: De Gruyter Mouton. Wiemer, Björn. (forthcoming). On the rise, establishment and continued development of subject impersonals in Polish, East Slavic and Baltic. In: Seppo Kittilä and Leonid Kulikov (eds.), Diachronic Typology of Voice and Valency-Changing Categories. Amsterdam/ Philadelphia: John Benjamins. Wiemer, Björn and Markus Giger. 2005. Resultativa in den nordslavischen und baltischen Sprachen. Bestandsaufnahme unter arealen und grammatikalisierungstheoretischen Gesichtspunkten. Munich/Newcastle: LINCOM Europa. Wiemer, Björn and Björn Hansen. 2012. Assessing the range of contact-induced grammaticalization in Slavonic. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 67–155. (Trends in Linguistics 242.) Berlin/New York: Mouton de Gruyter. Wiemer, Björn, Ilja Seržant and Aksana Erker. 2014. Convergence in the Baltic-Slavic contact zone. Triangulation approach. In: Juliane Besters-Dilger, Cynthia Dermarkar, Stefan Pfänder, and Achim Rabus (eds.), Family Effects in Language Contact. Modeling Congruence as a Factor in Contact Induced Change, 15–42. Berlin/New York: Mouton de Gruyter. Wiemer, Björn and Ilja A. Seržant. 2014. East Slavic dialectology: achievements and perspectives of areal linguistics. In: Ilja A. Seržant and Björn Wiemer (eds.), Contemporary Approaches to Dialectology. The Area of North, North-West Russian and Belarusian Dialects. (Slavica Bergensia 12.), 11–80. Bergen: University of Bergen. Wiemer, Björn and Bernhard Wälchli. 2012. Contact-induced grammatical change: the diversity of grammatical transfer phenomena and the diversity of different perspectives on areas. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 3–64. (Trends in Linguistics 242.) Berlin/New York: Mouton de Gruyter. Witzlack-Makarevich, Alena and Ilja A. Seržant. 2017. Differential argument marking: Patterns of variation. In: Ilja A. Seržant and Alena Witzlack-Makarevich (eds.), The Diachronic Typology of Differential Argument Marking. Berlin: Language Science Press. Zarickij, Nikolaj S. 1961. Formy i funkcii vozvratnyx glagolov (na materiale drevnerusskogo jazyka) [Forms and Functions of the Reflexive Verbs (on the Basis of the Material from Old Russian]. Kyiv: Izdatel’stvo Kievskogo universiteta.
Vít Boček
2 Common Slavic in the light of language contact and areal linguistics: Issues of methodology and the history of research 1 Introduction Linguistic disciplines are not like cells in a prison, where one would be completely isolated from others. Nor are they discrete rooms of a quiet hotel building, where it would be possible to spend hours and hours without seeing or even noticing other guests. They rather resemble different departments in a company, or houses in a small village where everyone knows their neighbors and can or even needs to cooperate with at least some of them. Of course, the number and degree of contacts vary depending on various factors and purposes. An area with very multifarious contacts is undoubtedly areal linguistics. Its relations to other linguistic disciplines were best shown by Pilarský (2001: 15), who introduced the following scheme:
linguistic typology in general areal linguistics A areal typology B
C
inter-systemic areal linguistics
descriptive linguistics
contrastive linguistics
language contact studies
diachronic linguistics
In this chapter I will focus on the bottom right corner of the diagram, i.e., on the relations between areal linguistics (with areal typology within it), language contact studies, and diachronic (historical-comparative) linguistics. An obvious overlap in research interests in these disciplines is language contact in general. https://doi.org/10.1515/9783110639223-003
64
Vít Boček
As far as the particular overlap between areal linguistics and contact linguistics is concerned, the main question obviously is whether or not long and intensive language contact is to be seen as the conditio sine qua non for admitting areal convergence, i.e., as the ultimate cause of areal convergence. (If not, then also independent, parallel developments in neighboring languages should be included in areal(-typological) research.) It is sometimes stated that no satisfactory general theory of language contact has yet been developed that could serve as a basis for explaining the convergence phenomena ascertained by areal typology (cf. Pilarský 2001: 19). Yet, I find this opinion ungrounded and too skeptical. In fact, there are very good models at our disposal, above all the already classic language contact model by Thomason and Kaufman (1988; Thomason 2001) and the rather alternative model by van Coetsem (1988, 2000; cf. Boček 2014: 111–200 for a thorough discussion and comparison of the two models). The overlap between contact linguistics and diachronic linguistics is also very close. Thinking about the possible role of language contact in the development of languages has become an intrinsic part of diachronic linguistics, as can be seen, for instance, from the simple fact that standard textbooks and guides to historical linguistics (and even to etymology as its sister discipline) commonly contain quite long chapters on language contact (see Bynon 2004: 171–261; Hock and Joseph 2009: 241–278, 347–424; Durkin 2009: 132–178). Despite this, however, many issues are still to be resolved or at least discussed. Clearly, the principal question here is how to demonstrate contact-induced language change in the distant past or prehistory, i.e., in reconstructed languages. In what follows, I discuss the aforementioned overlaps in the context of research into the development of Slavic protolanguage. I first present some general issues characteristic of research into changes in Common Slavic that are assumed to be contact-induced and claim that for a correct evaluation, and hence the acceptance or the rejection of extant contact hypotheses, it is necessary to uncover the deep methodological principles that have been applied by individual scholars (Section 2). To achieve this, I then offer and present the concept of paradigms of historical-comparative linguistics, according to which three different approaches to the significance of language contact in the development of languages can be distinguished: conventional, revisionist, and revolutionary (Section 3). In Section 4, the concept is applied to research into the contact between Common Slavic and Early Romance dialects: individual works on this topic are classified as belonging to one of the paradigms. Since some changes are assumed to be contact-induced both in Slavic from Romance and in Romance from Slavic, which means that convergence processes were possibly at work, the concept is eventually incorporated into the framework of areal linguistics (Section 5). In general, the main objective of the paper is to achieve
2 Common Slavic in the light of language contact and areal linguistics
65
a deeper understanding of the relations between the three aforementioned disciplines: diachronic linguistics, language contact studies, and areal linguistics. The suitability of research into Slavic for such an endeavor is highlighted.
2 Common Slavic and issues of language contact This section sets the stage for the planned investigation. Three general issues characteristic of research into contact-induced changes in Common Slavic are discussed: the high diversity of contact hypotheses, the varying degree of their elaboration, and the significance of the relationship of explanations by contact to the system-internal description of the changes under consideration. There exist very different approaches to the delimitation and periodization of Common Slavic and, consequently, also to the question of its designation (for a discussion, see Weiher 1967). In the present chapter, I will leave this topic aside and, following Brackney (2007), I will use the term “Common Slavic” for the whole time span of the commonly assumed existence of the Slavic protolanguage, i.e., roughly for the last two millennia BC and the first millennium AD. It is certain that Common Slavic in its last stage, that is, in the second half of the first millennium AD, was exposed to many external influences from other languages, because its speakers, the historical Slavs expanding in the given period to new areas, were in several and multifarious contacts with other peoples. There are indications that contact influences existed also in the preceding stages, from which there is no direct historical or linguistic evidence (Arumaa 1964: 17–46; Shevelov 1964: 613–624; Boček 2014: 203–373). The question is which of the contacts left traces in the Slavic linguistic system, i.e., which of them resulted in the rise of contact-induced language phenomena in Slavic. A large number of hypotheses of this type have already been proposed, with varying degrees of plausibility. Interestingly, quite often one and the same structural change in Slavic has been explained by different scholars as contact-induced by different languages. An illuminating example from the development of the Slavic sound system is the monophthongization of diphthongs, which has been interpreted as a result of contact influence from Iranian (Zaliznjak 1963; Pisani 1974), Uralic (Falkenhahn 1968), Altaic (Shevelov 1964; Galton 1997a; Brackney 2007), Celtic (Gvozdanović 2009), and Romance (Bonfante 1966; Enrietti 1982). Similarly, in the area of grammar, a typical case is the rise of the category of verbal aspect in Slavic, which has been explained as contact-induced by Germanic (Vaillant 1966: 463; Schmitt-Brandt
66
Vít Boček
1971), Altaic (Menges 1958, 1983; Schelesniker 1991a, 1991b; Galton 1997b), and Uralic (Polák 1964, 1970; Bednarčuk [Bednarczuk] 1997). Of course, there are considerable differences in the degree of elaboration of contact hypotheses among individual scholars working on individual Slavic changes. As was aptly put by Vermeer (2003), when reading scholarly papers, it is necessary to be aware of the difference between two types of statements: (a) those that are explicitly related to the relevant facts, i.e., those that are properly argued or substantiated, and (b) those that are just preliminary claims or even “shots in the dark”. In further research, statements of the latter type may be converted into statements of the former type or, on the contrary, may turn out to be wrong. Undoubtedly, many propositions of the contact influence of various language groups on Slavic (for an overview, see Boček 2014: 374–378) are statements of the aforementioned latter type. This can certainly be perceived as their disadvantage, while at the same time it can be a challenge for scholars to rethink these preliminary ideas. Another striking fact is that most Slavists do not consider a broader context to the chosen topic. When considering a contact-induced change in Slavic they often use comparative data from only one other group of languages and do not even acquaint themselves with alternative contact explanations. Thus, specialists in Altaic languages see the source of many Slavic changes in Altaic (e.g., Menges), Celticists prefer Celtic (e.g., Gvozdanović), experts in Romance studies readily admit the influence of Romance on Slavic (e.g., Enrietti), and so forth. What is essential, of course, is the relationship of explanations by contact to the non-contact-related description of a given change, i.e., above all, to its internal conditions and dating. One can mention here, for instance, three different hypotheses about external influence with respect to the Slavic change g > γ (> h). This change was supposed to be contact-induced by Iranian (Abaev 1965: 41–52), Altaic (Novák 1939–1940), or Celtic (Gvozdanović 2009: 101–104). However, all these hypotheses are conceivable only when the very early, i.e., late Common Slavic, dating of the change is assumed, which is advocated above all by Jakobson (1925–1926), Trubetzkoj (1988: 277–289) and Isačenko (1936), but rejected by most scholars, who, instead, consider the change to be a later, independent development in individual Slavic languages (cf. Komárek 2006: 312–322). A truism should be mentioned in this context: before we can answer the question of which of the possible contact influences is the most likely, we should first examine whether a system-external explanation is actually needed or whether a system-internal explanation is sufficient. To give an example from Slavic, let me mention different explanations of the retraction of s into š after r, u, k, and i. There are strong advocates of an Iranian impulse behind this Slavic change (Pisani 1935, 1974; Shevelov 1964; Schmitt-Brandt 1971;
2 Common Slavic in the light of language contact and areal linguistics
67
Galton 1997a; Erhart 2006) as well as strong detractors of this theory (Filin 1962; Zaliznjak 1963; Sedov 1978; Trubačev 1983). Admittedly, the abovementioned issues can be satisfactorily resolved primarily by a careful examination of language data. Nevertheless, the purpose of the present chapter is to show that in searching for the most probable answer to the question of possible external influences on Slavic, one needs to remember something else. The point is that an understanding of the methodological principles that were applied by individual scholars dealing with the development of Slavic can be helpful and desirable. According to Lunt (1997: 9–12), when trying to understand a linguistic problem, a student can either look first at the primary data, or, in contrast, search for bibliographical sources with an eye to finding how other scholars tackled a particular problem. There are certainly some advantages as well as disadvantages to both approaches. Concerning the disadvantages, it is clear that collecting data afresh can be tiresome and tedious and the final interpretation of them can turn out to be unoriginal (thus, the whole procedure is pointless, like reinventing the wheel); at the same time, concentrating solely on the extant literature may convince the investigator that the obtained solution is correct, even if it is not (so, eventually, he only helps to perpetuate a myth). In my opinion, the latter disadvantage can be eliminated just by constantly taking into account the methodological principles which individual scholars followed. Such “accounting” is essentially a historiographic tool for assessing linguistic enquiries. An approach which tries to do this consistently is my concept of the paradigms of historical-comparative linguistics to be presented in Section 3.
3 Paradigms of historical-comparative linguistics In this section, I first outline the roots of the concept of paradigms of historicalcomparative linguistics, then present my recent elaboration on it, and finally give some general examples of its applicability in order to have this theoretical framework ready for a more specific application to be given in the subsequent section. The concept was inspired by a paper by Janhunen (2001) written as part of very sharp polemics on the nature of basic principles to be followed in Uralic comparative studies (for an overview of the polemics, see Boček 2012). Janhunen distinguished four different paradigms in Uralic comparative studies. According to him, the first, called the conventional paradigm, is based on four assumptions: (1) the Uralic languages are related in the context of a language family; (2) the internal relations within the Uralic family are the result of diachronic divergence,
68
Vít Boček
which can be described in terms of a family tree; (3) the diachronic divergence presupposes a more or less uniform prehistorical protolanguage; and (4) the protolanguage, although only fragmentarily known, must have been a natural language with a limited geographical distribution, i.e., a homeland (Janhunen 2001: 30–31). The second paradigm is called revisionist by Janhunen, and it consists of assumptions which consider the Uralic homeland to be wider (diffuse), or limited, but situated elsewhere (i.e., more to the west) than traditionally thought. A logical consequence of these approaches is the rejection of the binary model of the Uralic family tree diagram and its replacement by other models, above all the “bush” model and the “comb” or “rake” model of the Uralic languages, with more parallel branches at one and the same level. The third, revolutionary paradigm grows from the revisionist one and further radicalizes its premises, which eventually leads to the rebuttal of all four abovementioned conventional principles. Encouraged by the rejection of the family tree, the “revolutionaries” claim that the whole concept of Uralic genetic affinity is wrong. Instead, they assume that the similarities shared by individual Uralic languages are the result of various areal influences among originally unrelated languages. The comparative method as the main tool of historical-comparative linguistics is repudiated. Finally, the “counterrevolutionary” paradigm is a reaction to the revolutionary one, showing the implausibility and unacceptability of the latter. Essentially, the “counterrevolutionary” paradigm is constituted by works of the present-day conventionalists, in which revolutionary claims are explicitly discussed and criticized. In my recent work (Boček 2014), I elaborate on Janhunen’s line of reasoning. I show that the paradigms can be generalized and, subsequently, applied to the historical-comparative research tradition of essentially any language or group of languages. Thus, the concept of paradigms can be used to evaluate the situation in Indo-European studies in general as well as in a chosen branch of this discipline, e.g., in Slavic historical-comparative studies. In addition, I propose also certain methodological improvements to the original concept. Janhunen used the notion “paradigm” in a Kuhnian sense, i.e., as a shared model, something that is commonly adhered to or something which the scientific community – at a given period – agrees with (Kuhn 1962; for an early discussion of Kuhn’s paradigm with respect to linguistics, see Percival 1976). Thus, in Janhunen’s theory, the paradigm means a specific period of the development of a chosen scientific field (Uralic studies) in which specific methodological principles are advocated more or less uniformly. The conventionalist paradigm is represented above all by the founding fathers of Uralic comparative studies from the 19th century, while the revisionist paradigm is constituted by scholars who question the very foundations of the field without completely denying them, as exemplified in many works extant from
2 Common Slavic in the light of language contact and areal linguistics
69
the 20th century. Yet the revolutionary paradigm, aimed at destroying all preceding achievements, has been created by a group of present day Uralists at the threshold of the new millennium. By contrast, I claim that the paradigm is an abstract set of inter-connected principles that are, in fact, not advocated all together, but only partially (by a group of scholars, a number of studies, or even only parts of studies) and variably in different periods. At the same time, paradigms represent, in some respects, a chain of actions and reactions to previously expressed ideas on a given topic. A paradigm as a bundle of related ideas is essentially a construct which is not manifest in specific texts in extenso (meaning that all the principles from a particular paradigm would be sustained in a specific text). On the contrary, one and the same text can represent (and often does represent) a combination of ideas borrowed from more – usually two – paradigms. As a rule, a “dominant” of the text can be identified, i.e., a paradigm which is manifested in the examined text to the greatest extent. Beside this, a subordinate, often much less visible tendency to take into consideration some thoughts from another paradigm is detectable. For example, a predominantly conventional approach can be accompanied by a tendency to revisionism, or a predominantly revisionist text can reveal a revolutionary undertone. In order to achieve a more accurate delimitation of the aforementioned paradigms, I propose to view the role of the contact-induced development of languages as a decisive factor in this case. Accordingly, the paradigms can be reframed as follows: (1) the conventional paradigm states that in the development of languages the main role is played by language divergence, whereas no or only a minor role is played by language convergence; (2) the revisionist paradigm is characterized by the balanced roles of divergence and convergence; (3) according to the revolutionary paradigm, the propelling force in language development is language convergence; (4) the counter-revolutionary approach reverses the effect of a previous revolution by returning to conventionalism or revisionism. Overall, the conventional, the revisionist, and the revolutionary opinions constitute a kind of tripartite system. Based on systematic variables, other tripartite systems are also possible such as the degree of trust in the comparative method (‘absolute’ × ‘reserved’ × ‘no’), the nature of the explanation of language change (i.e., change arises ‘exclusively or at least predominantly from internal factors’ × ‘from a balanced interplay of internal and external factors’ × ‘exclusively or predominantly from external factors’), the way of viewing the notion of a homeland (‘a geographically limited area of the presence of a protolanguage’ × ‘the broader area more connected with other homelands’ × ‘total rejection of the very notion of homeland’), or the way of construing the notion of the genetic affiliation of languages (‘every language has only one
70
Vít Boček
parent language’ × ‘some kinds of languages, namely pidgins and creoles, cannot be classified genetically’ × ‘a language can have several protolanguages’), and so forth. Individual texts generated by historical-comparative linguists can be incorporated into the system of established parameters of the paradigms. Thus, texts advocating only the traditional Stammbaum model, assuming a gradual disintegration of protolanguages into daughter languages, typically belong to the conventional paradigm. The revisionist paradigm is mainly represented by texts in which the tree model is relativized and complemented by considering the role of convergence of the Sprachbund type in language development. Additionally, there may also arise immediate, essentially conventional, reactions against revisionism. These reactions can be called counterrevisionist. Texts belonging to the revolutionary paradigm tend to make the revisionist point of view more extreme, thus leading to a complete rejection of the comparative method as such and to the assumption of language mixing as the propelling force in language development. In addition, the rejection of previous approaches can be accompanied by some amateurish reasoning. The corresponding texts might be better placed and evaluated beyond the basic tripartite system of scientific paradigms since such amateurish reasoning is not a paradigm, but a deviation thereof. Yet another notion should be mentioned in this context, namely ideology. In fact, any paper regardless of its belonging to a specific paradigm can be more or less influenced by a certain ideological framework. However, works from the revolutionary paradigm tend to be overloaded by ideology. Finally, revolutionary works often trigger counter-revolutionary responses in the form of either defending and rehabilitating basic conventional positions or proposing a return to a more moderate revisionist standpoint, whence one can distinguish between two slightly different types of counter-revolutionary reactions. The whole system of different paradigms and reactions can be shown in the following scheme: ideology amateurism conventionalism
revisionism counter-revisionism
revolutionism counter-revolutionism I
counter-revolutionism II traditionalism
avantgardism
2 Common Slavic in the light of language contact and areal linguistics
71
The core of the diagram is formed by the conventional, the revisionist, and the revolutionary paradigms, which are situated on the axis between traditionalism and avantgardism as two opposite poles of the continuum of the openness of a given paradigm to innovations. Revisionism and revolutionism can be called “primary” reactions, since they represent the most usual and direct form of rethinking the principles of historical-comparative linguistics as far as language contact is concerned. Counter-revolutionism is no longer considered one of the basic paradigms, since its set of maintained principles is the same as that of one of the basic paradigms, i.e., either the revisionist or the conventional one. There are essentially three possible “secondary” reactions to basic paradigms, namely counter-revolutionism I, counter-revolutionism II, as well as the counter-revisionism. With this conceptual and theoretical framework established, it is now possible to objectively assess specific works of historical-comparative linguistics. Subjective criticism is at work only when the field of linguistic historiography is disregarded and the evaluator takes on the role of historical linguist who has his own attitude to particular works and their ideas, i.e., he accepts, further develops, or – on the contrary – refutes them. I am ready to advance some typical examples of texts with a clear dominant, as far as paradigms are concerned. In the field of Indo-European studies, the conventional paradigm is represented mainly by the works of the founders of Indo-European studies, of the Schleicherian generation, as well as of the Neogrammarians. One of the representative cases of revisionism is Erhart’s (1976) concept, according to which, in the development of individual Indo-European language groups, divergent (differentiating) and convergent (integrating) forces had been equally important. Classic instances of revolutionism are the works of Uhlenbeck (1934), Trubetzkoy (1939), and Pisani (1940), who argued that the Indo-European languages had been formed primarily by convergence. A revolutionary approach with an amateurish flavor is, in my opinion, Dixon’s (1997) punctuated equilibrium model (Boček 2014: 248–263). The best example of the revolutionism strongly influenced by (Marxist) ideology is the “new science of language” (Marrism) developed in the 1920s by the Soviet linguist Nikolaj Marr. A similar scale of examples of works from individual paradigms can be constructed also in the field of Slavic historical-comparative studies, starting with conventional works in which language contact is hardly or not at all taken into account (cf. Mareš 1999, 2001), continuing with revisionist works accepting the balanced roles of divergence and convergence (cf. Martynov 1982, 1983), and finishing with “revolutionaries” assigning the decisive role in the development of Slavic to external influences (cf. Brackney 2007, applying Dixon’s model to Slavic, or
72
Vít Boček
Pritsak 1983, with his even more problematic theory, assuming the rise of Slavic as a pidgin). To return to the topic of putative external sources of individual Slavic changes, the concept of paradigms of historical-comparative linguistics can shed additional light on some of the explanations. For instance, in the case of the retraction of s into š after r, u, k, and i, the advocates of Iranian influence belong to different paradigms, because their understanding of the problem and their explanation of the change is very different. Shevelov’s (1964) study is predominantly a conventional work, albeit with a certain revisionist undertone. The author understands language contact rather as shared, simultaneous innovations taking place in the case of both Slavic and Iranian retraction. This corresponds to a general problem of most of the conventional works, which fail to distinguish between the primary (genetic) and secondary (areal) isoglosses. As has been mentioned, Erhart’s (1976) work is revisionist at its core. Slavic in his concept is seen, due to many mainly morphological characteristics, as originally belonging to the western part of the IndoEuropean dialect continuum. Therefore, many phonological features that link Slavic with the East Indo-European part must be explained as secondary, i.e., induced by contact with some of the eastern group of Indo-European dialects, most likely from Iranian. Finally, Pisani’s (1935, 1974) writings belong to the revolutionary paradigm. According to Pisani, there was no Proto-IndoEuropean and its daughter protolanguages, but only a dialectal continuum of Indo-European, and it is only our a posteriori observation that the change s > š took place in the territories where Iranian and Slavic languages eventually emerged. As follows from the above, the range of applicability of the concept of paradigms of historical-comparative linguistics may vary considerably from whole linguistic schools or trends to random scholarly texts on individual language changes. I believe that the best approach could be a middle way in which a particular, not broadly defined topic is chosen as a starting point and the corresponding texts are evaluated with the help of the aforementioned concept. A typical topic could be the history of research into contacts between two languages or groups of languages. In this case, whole scales of paradigms and reactions could be established, from conventional works, on the one hand, up to counter-revolutionary reactions, on the other. In section 4, I will demonstrate this with the help of a case study dealing with Slavic-Romance language contact. Similar case studies could involve contact between Slavic and other languages. However, there is a specific reason for focusing on Slavic-Romance language contact, which differs from the contact between Slavic and other language groups such as Iranian, Altaic,
2 Common Slavic in the light of language contact and areal linguistics
73
Uralic, Celtic, or Germanic. In all these cases, Slavic is seen only as the recipient language, while in the case of Slavic-Romance contact both vectors of influence are considered possible. As a result, a student can cover both Slavic influence on Romance languages and Romance influence on Slavic languages as well as certain connections and relations between them. Thus, the issue of Slavic-Romance language contact will bring us closer to questions of areal linguistics.
4 A case study: early Slavic-Romance language contact In this section, the concept of paradigms of historical-comparative linguistics will be applied to research into early Slavic-Romance language contact. First, two centers of contact will be distinguished and their characteristics given. Then, the scales of paradigms and reactions will be established with respect to particular problems facing research on both centers. Two centers of (Common) Slavic-(Early) Romance contact are distinguished, not only geographically, but also in terms of (1) the chronology of a particular contact, (2) the transferred linguistic material, and (3) the vector of the putative contact influence. The first center was located further to the west, comprising areas of the former Roman provinces of Noricum, Pannonia, and Dalmatia. It can be roughly described as the western part of the South Slavic area, since it involved areas where those Common Slavic dialects were spoken, from which the Croatian and Slovene languages eventually emerged. They had been in contact with Early Romance dialects, which later gave rise to the Dalmatian language (once spoken at the Dalmatian Coast, now extinct), the Friulian and the Ladin languages (both spoken in specific regions of Northeast Italy), and several northeastern Italian dialects. The contact started at the turn of the 7th century AD and continued up to the second millennium. In this area, the contact progressed in one direction and at one level of the system: the vector of influence went from Romance to Common Slavic and the only linguistic level affected was the lexical level. Tens or even hundreds of words (if toponyms are included) were borrowed, primarily from Christian terminology (e.g., Common Slavic *križь ‘cross’ < *kr’ū(d)žĭ < Romance *krū́(d)že < Latin crux ‘cross’) and material culture (e.g., Common Slavic *ocělь ‘steel’ < *ăt’ālĭ < Romance *ăt’ā́le < Latin aciarium ‘steel’) (cf. Rocchi 1990; Holzer 2007; Boček 2010).
74
Vít Boček
The second center of Slavic-Romance contact was situated more to the east (and south) and involved areas of present-day Romania, Bulgaria, Macedonia, and Serbia, i.e., roughly the former Roman provinces of Dacia, Thracia, and Moesia. Here, contact took place between Common Slavic dialects which ultimately gave rise to the Bulgarian and Serbian languages, on the one hand, and Romance dialects from which the Romanian language sprang, on the other hand. Research into Slavic-Romance contact in the east has constantly faced manifold problems. The first problem concerns the dating of the beginning of this contact. It is commonly accepted that the contact began in the 6th century AD, when Slavs, expanding from their putative homeland to the south, spread to the lower Danube basin area. This view must be interpreted as conventional, since it is premised on the most traditional localization of the Slavic primary homeland to the northeast of the Carpathian Mountains. However, one can cite also other hypotheses. According to Bonfante (1966), the Romanized peoples of Dacia were in direct contact with Slavs in their homeland, which would mean that their contacts had begun already in the 2nd century AD. Nevertheless, Bonfante maintained that the actual linguistic influence of Romance upon Common Slavic commenced around 600 AD. Thus, his view can be considered revisionist. Recently, Del Gaudio (2013, 2014) advanced a new theory about the initial contacts between Common Slavic and Early Romance. He posited the presence of Roman culture also in some areas to the north of Dacia – to be specific, in the southwestern and southeastern regions of contemporary Ukraine, and also in the Black Sea areas. In Del Gaudio’s opinion, several words were borrowed from Latin/Romance into Common Slavic already before its expansion, to be specific, in the 2nd century AD, some possibly through Greek mediation. This hypothesis should be interpreted as revolutionary since, when questioning the very concept of Common Slavic, the author almost completely rejected the real value of the notion of a protolanguage as a scientific tool for describing genetic relations. There are also other warnings. The author is only vaguely familiar with scholarly literature on the topic, while treating historical facts rather selectively. It is not surprising that Caldarelli’s (2014) discussion of Del Gaudio’s theory seems to be a typical counter-revolutionary response from the conventional standpoint (counter-revolutionism II). The second problem related to Slavic-Romance contact lies in the language material and the vector of its transfer, as illustrated, for instance, in Del Gaudio’s theory about early lexical borrowings into Common Slavic. If some of the main stock of Early Romance loanwords in Common Slavic were not borrowed in the western, but in the eastern area (e.g., Balkan Romance *čĕrĕ́s’a ‘cherry, Prunus avium L.’ > Common Slavic *čĕrĕša > *čerša, cf. Boček 2010:
2 Common Slavic in the light of language contact and areal linguistics
75
61–65), the issue of lexical influences from Romance to Common Slavic seems to be exhausted and more or less satisfactorily resolved. More complicated, however, is the “opposite” problem of Common Slavic lexical influences upon Balkan Romance or, better, Proto-Romanian (cf. Mihăilă 1960; Rosetti 1968: 308–319). It is still not sufficiently clear when the process of the borrowing of Slavic words into (Proto-)Romanian began. The majority of scholars assume that about a dozen Romanian words purportedly displaying some early Slavic sound features (e.g., nasalized vowels) had already been borrowed in the second half of the first millennium (cf. Leschber 2007, 2012). However, this view has been repeatedly questioned, most recently by Paliga (2010, 2012), who tries to prove that the oldest Slavic borrowings in Romanian can hardly be dated earlier than the 12th century. Overall, the topic can be subsumed under research into the Balkan linguistic area, since borrowings from Slavic into Romanian belong to the group of lexical Balkanisms (cf. Solta 1980: 85–101; Kahl, Metzeltin, and Schaller 2012). In addition to the lexical influence, one should mention structural features which are presumed to have been transferred in the eastern area. In this case, two research traditions seem to compete. First, there are numerous studies dealing with the role of Slavic structural features in the rise and further development of Romanian (e.g., Seidel 1957; Petrovici 1956, 1957, 1958; Rosetti 1968: 302–308; Trummer 1983; Petrucci 1999; Klimkowski 2014). Second, there are only a very limited number of works investigating the possible influence of Balkan Romance (Proto-Romanian) on Slavic (cf. Bonfante 1966; Enrietti 1982, 1987, 1992–1993, 1998, 2001, 2009; Bednarczuk 2010). Thus, Slavic is viewed more often than not as the source language in this contact situation. With the help of the concept of paradigms of historical-comparative linguistics, papers of both aforementioned traditions can be interpreted. I will start with the vector spreading from Slavic to Proto-Romanian and mention typical instances. An example of a revolutionary work with a strong ideological load is Seidel (1957), who ascribed a vast number of Romanian syntactic features to Slavic influence. However, as recently shown by Klimkowski (2011: 336–337) in his counter-revolutionary (I) reaction, Seidel’s analysis is problematic from both the chronological and geographical points of view, since features from different periods are compared and linguistic data from Russian are mainly taken into account. The latter aspect is probably a result of the Sovietization of Romanian science in the 1950s. Likewise, Petrovici’s work can be interpreted as revolutionary. In a series of studies extant from the 1950s, this author argued that Romanian is “a Romance language with Slavic pronunciation” (Petrovici 1957: 43). According to Petrovici, both vowel and consonant subsystems of Romanian
76
Vít Boček
were heavily influenced by Slavic. This scholar rearranged the traditional Romanian phonemic inventory using the distinctive features of palatalization and labialization. He posited the existence of four series of consonants: neutral, palatalized, labialized, and palato-labialized (e.g., there were p, p’, p°, and p’°). Thus, the number of consonants increased considerably (from 20 in the traditional system to 72) and, consequently, the number of vowels and semivowels decreased. This development was the result, as Petrovici claimed, of contact with Slavic with a full-fledged correlation of palatalization. As a counterrevolutionary reaction (counter-revolutionism I), I can mention Trummer (1983), who modified Petroviciʼs conclusion by stating that the Romanian language should rather be viewed as “a Romance language with only limitedly applied Slavic pronunciation” (Trummer 1983: 141), since Romanian was only partly and temporarily exposed to Slavic influence. To give another example of a counter-revolutionary (I) reaction, one can mention Petrucci (1999: 41–49), who completely rejected Petroviciʼs theory, referring to its improbability from the typological point of view and, more importantly, from the perspective of language contact theory in general. Curiously enough, as Petrucci (1999: 45) pointed out, Petrovici’s claim for extensive Slavic influence on the phonemic system of Romanian continues to be rather popular among some historical linguists. For instance, Hock (1991: 673) and Bynon (2004: 243) accept this hypothesis without any reservation; Petrovici’s unusual interpretation of the Romanian phonological system is supported also by Stadnik (2002: 96–98). An obvious example of a revisionist work is Petrucci’s (1999) monograph. With the help of a modern language contact model developed by Thomason and Kaufman (1988), the author discusses all the phonological features of Romanian (including stress, palatalizations, pre-ioticization, the change sk > št before front vowels, the phonemes /z/ and /x/, etc.) as well as its morphological features (like the neuter gender, the vocative, the construction of the numerals ‘11ʹ through ‘19ʹ, the shortening of the infinitive) that have often been viewed as contact-induced by Slavic. Most Romanian linguists clearly belong to the conventional paradigm since they typically admit only the restricted influence of Slavic upon Romanian, or even no influence at all (cf. Vasiliu 1968; Coteanu and Rosetti 1969; Djuvara 2007). A similar evaluation can be made in the case of the contact influence direction from Proto-Romanian to Slavic. Both Bonfante and Enrietti belong to the tradition of Italian Neo-linguistics, a trend which arose in the 1920s in opposition to Neogrammarians and represented a radicalization of the wave theory and areal geography (cf. Bartoli and Bertoni 1925). Since Neo-linguistic works assign the decisive role in language development to external forces, they must
2 Common Slavic in the light of language contact and areal linguistics
77
be interpreted as revolutionary (for a counter-revolutionary reaction, see Hall 1946). However, in their studies of Romance-Slavic language relations, both Bonfante and Enrietti present a more moderate view and their treatment of the famous neo-linguistic areal norms is undogmatic. Thus, their papers can be interpreted as results of the late withdrawal of their authors from revolutionary positions to revisionism. Still, both authors assume Proto-Romanian to have had quite an intense contact influence on Slavic. According to Bonfante (1966), the following Slavic changes were induced by Proto-Romanian: the tendency to change the subsystem of vowels from the three-level to a four-level arrangement; the monophthongization of diphthongs; the open syllable tendency; changes in palatalization; the rise of analytic verbal patterns; the rise of the passive reflexive. The same Slavic changes, although with many differences in the detail, have been proposed as contact-induced by Proto-Romanian by Enrietti (1982, 1987, 1992–1993, 1998, 2001, 2009) and Bednarczuk (2010). Again, to cover the whole scale of reactions, one should mention the counterrevisionist reaction of Caldarelli (2012) to Enrietti; Caldarelli completely rejects the possibility of external influences on the rise of a Slavic open syllable structure. Finally, Shevelov (1964: 621) can be mentioned among several conventional works completely denying any influence of Proto-Romanian on Slavic. What has never been discussed in this respect is the fact that there is a certain overlap between phonological changes supposed to be induced in the direction from Slavic to Romance and changes considered as induced in the opposite direction from Romance to Slavic. To be specific, it concerns several palatalization processes and the process during which the ProtoRomanian and Slavic systems of vowels became similar, each displaying a tendency to a four-level structure. Different authors have described this tendency somewhat differently, but they largely agree that the system was not quite usual from the typological point of view, which means that the probability of contact influence increases. Bonfante (1966) presented the corresponding systems as follows: Proto-Romanian
o u
a ă ȋ
Slavic a
ę (> ie) e i
ĕ o ŭ u y
e i ĭ
Now, the following question arises: which of the perspectives is more plausible than the other? Was the influence directed from Slavic to Proto-Romanian or vice versa? In fact, two more explanations are possible.
78
Vít Boček
First, both the influence of Slavic on Proto-Romanian and the influence of Proto-Romanian on Slavic might have been at work at the same time, only in different areas. It is generally agreed that to the north of the Danube, Slavs had gradually been assimilated by Romance peoples, whereas, to the contrary, to the south of the Danube, Romance peoples had been assimilated by Slavs (cf. Sala 2005: 21; Klimkowski 2014). It is legitimate to assume that in the north the evolving arising Proto-Romanian system had been influenced by the vanishing Slavic language before its extinction, whereas in the south, the Slavic system might have been influenced by the disappearing Balkan Romance language (Boček 2014: 356–357). However, this assumption does not explain much since it does not answer the question of how and where the alleged contactinduced language features first emerged. The second possibility is to admit that the linguistic situation in the given areas was so complex that it would be perhaps problematic to distinguish opposite vectors of contact influence. Given that in Early Medieval Europe, multilingualism was not an exception but the rule (cf. Wolfram 1997), it is possible to maintain that the linguistic convergence of Slavic and Romance was of a fuzzy character. Many people were probably symmetrical bilinguals with a roughly equal proficiency in the two languages. Moreover, recent research into the questions of early medieval ethnicity has shown that the group identity was perhaps less stable than traditionally thought (see Pohl 1988). I believe that the situation in the lower Danube basin might have been similar when different types of identity, manifesting themselves differently according to the specific communicative situation, were practiced. To determine a vector of contact in such a situation is highly problematic. Moreover, it is also very difficult to distinguish between external influence and parallel developments.
5 Areal-typological outlook With the last solution described in Section 4, we are again faced with the problem already suggested in the introduction: it is not quite clear how to understand or conceive the very notions of linguistic area and areal convergence (cf. Thomason 2001: 99–128; Campbell 2006). At any rate, it should be borne in mind that some of the aforementioned Slavic-Romance convergences were found not only in Proto-Romanian and Slavic, but also in other languages of the area. Thus, it would be perhaps reasonable to speak about a Sprachbund in a similar way as in Novák (1984). According to him, the development of Slavic in the second half of the first millennium was part of a complex convergence in
2 Common Slavic in the light of language contact and areal linguistics
79
which several other languages were taking part, including Altaic, Vulgar Latin / Early-Romance, and Greek koine. It remains out of the scope of the present paper to consider whether the interpretation here suggested can be subsumed under the assumptions about the very early beginnings of the Balkan linguistic area (Joseph 2010). From the general point of view, the last solution presented in Section 4 seems to be similar to those explanations which have been, at least over the last few decades, advanced for various linguistic areas in Europe. Thus, within research into the Balkan linguistic area, Lindstedt (2000, 2014), assuming the situation of stable multilingualism with stable prestige relations among the Balkan languages, speaks about contact-induced changes by “mutual reinforcement”. A similar tendency, brought about by multilingualism, is discussed in extensive literature on the Central European linguistic area (e.g., Pilarský 2001; Thomas 2008; also Kurzová in this volume). Interestingly enough, the same approach is detected in recent research into Standard Average European. In the latter case, it appears problematic to ascertain the center of the diffusion of common European features (cf. Haspelmath 1998). According to our system of historical linguistics paradigms, interpretations assuming no source language and no target language within a certain linguistic area belong quite obviously to the revisionist paradigm. This does not mean that every work presuming the existence of a linguistic area is revisionist. Some of such theories, especially those which describe the rise of whole language groups or branches due to convergence (Trubetzkoy, Uhlenbeck, Pisani), should be viewed as revolutionary. Nevertheless, standard works on commonly recognized linguistic areas are revisionist at their core inasmuch as they do not deny genetic affiliations of individual languages belonging to a particular area. In other words, it is maintained in such works that every language belonging to a certain linguistic area retains its membership in a language family. One should agree that, when viewed diachronically, research on language contact is faced with variegated problems. The scarcity or even lack of empirical data is only the most visible one. A hidden, yet very significant problem is that authors who treat some language changes as contact- or non-contact-induced adhere to different theoretical frameworks and employ different methodological principles. The foregoing survey shows that, as far as language contact is concerned, three different standpoints can be distinguished within historicalcomparative linguistics, these being conventional, revisionist, and revolutionary. I also demonstrated the extent to which the uncovering of reasons behind admitting or rejecting a hypothesis about external influence can prove profitable for future research not only in Indo-European studies in general, but also in the field of Slavic comparative linguistics.
80
Vít Boček
The achieved results can be summarized as follows: First, the concept of paradigms in historical-comparative linguistics is useful not only from the perspective of language contact in general but also from the perspective of areal linguistics and, more narrowly, areal typology. Consequently, three traditional disciplines – historical-comparative linguistics, language contact studies, and areal linguistics – are linked by a concept which helps us to better understand their interrelations. Second, works focusing on areal similarities among languages and assuming the existence of linguistic areas belong primarily to the revisionist paradigm, which accepts that linguistic divergence and linguistic convergence play balanced roles in the development of languages. Third, the moderate revisionist approach to areal-typological phenomena lies in the reluctance of its advocates to determine a source language and a target language in a given situation; such a situation is revealed, for instance, in the case of early Slavic-Romance contacts involving, in particular, some convergent phonological features. Acknowledgments: This paper was written with the support of a grant provided by the Czech Science Foundation, No. 13-17435S. I would like to thank the editors of the volume for valuable comments on earlier versions of this paper. I alone am responsible for any shortcomings.
References Abaev, Vasilij I. 1965. Skifo-evropejskie izoglossy. Na styke vostoka i zapada [Scytho-European Isoglosses. On the Border of East and West]. Moscow: Nauka. Arumaa, Peeter. 1964. Urslavische Grammatik. Einführung in das vergleichende Studium der slavischen Sprachen, Vol. 1: Einleitung. Lautlehre. Heidelberg: Carl Winter. Bartoli, Matteo and Giulio Bertoni. 1925. Breviario di neolinguistica. Modena: Società Tipografica Modenese. Bednarčuk, Lešek [Leszek Bednarczuk]. 1997. Konvergencii balto-slavjanskix i finno-ugorskix jazykov v strukturnom i areal’nom aspekte [Convergences of Baltic-Slavic and Finno-Ugric languages in the structural and areal aspects]. In: T. M. Sudnik and Jevgenij A. Xelimskij (eds.), Balto-slavjanskie issledovanija 1988–1996. Sbornik naučnyx trudov, 91–108. Moscow: Indrik. Bednarczuk, Leszek. 2010. Parallèles phonétiques romano-slaves. Neophilologica 22: 11–22. Boček, Vít. 2010. Studie k nejstarším romanismům ve slovanských jazycích [Studies on the Early-Romance Borrowings in Slavic]. Prague: Nakladatelství Lidové noviny. Boček, Vít. 2012. A note on prehistoric language contact or what Slavic studies can learn from Uralic studies. In: Ilona Janyšková and Helena Karlíková (eds.), Theory and Empiricism in
2 Common Slavic in the light of language contact and areal linguistics
81
Slavonic Diachronic Linguistics, 365–376. (Studia etymologica Brunensia 15.) Prague: Nakladatelství Lidové noviny. Boček, Vít. 2014. Praslovanština a jazykový kontakt [Proto-Slavic and Language Contact]. Prague: Nakladatelství Lidové noviny. Bonfante, Giuliano. 1966. Influences du protoroumain sur le proto-slave? Acta Philologica. Societas Academica Dacoromana 5: 53–69. Brackney, Noel C. 2007. The Origins of Slavonic: Language Contact and Language Change. Munich: Lincom Europa. Bynon, Theodora. 2004. Historical Linguistics. Cambridge: Cambridge University Press. Caldarelli, Raffaele. 2012. Open syllable once again. Endogenous or exogenous? Studi Slavistici 9: 157–166. Caldarelli, Raffaele. 2014. On Latin-Protoslavic language contacts. Some remarks on a recent paper by Salvatore Del Gaudio. Studi Slavistici 11: 171–181. Campbell, Lyle. 2006. Areal linguistics: a closer scrutiny. In: Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas: Convergence in Historical and Typological Perspective, 1–31. Basingstoke: Palgrave Macmillan. Coetsem, Frans van. 1988. Loan Phonology and the Two Transfer Types in Language Contact. Dordrecht: Foris Publications. Coetsem, Frans van. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Carl Winter. Coteanu, Ion and Alexandru Rosetti (eds.). 1969. Istoria limbii române [History of the Romanian Language], Vol. 2. Bucharest: Editura Academiei Republicii Populare Române. Del Gaudio, Salvatore. 2013. Latin-Protoslavic language contacts and their reflexes in early Old Church Slavonic texts. In: Marcello Garzaniti, Alberto Alberti, Monica Perotto, and Bianca Sulpasso (eds.), Contributi italiani al XV Congresso Internazionale degli Slavisti (Minsk, 20–27 agosto 2013), 49–75. Firenze: Firenze University Press. Del Gaudio, Salvatore. 2014. Sposoby proniknovenija latinizmov v staroslavjanskij jazyk [The ways of penetration of Latinisms into Old Church Slavonic]. Palaeobulgarica 38 (1): 62–76. Dixon, Robert M. W. 1997. The Rise and Fall of Languages. Cambridge: Cambridge University Press. Djuvara, Neagu. 2007. O scurtă istorie a românilor povestită celor tineri [A Short History of Romanians Narrated for the Youth]. Bucharest: Humanitas. Durkin, Philip. 2009. The Oxford Guide to Etymology. Oxford: Oxford University Press. Enrietti, Mario. 1982. Considerazioni sul costituirsi dell’unita linguistica slava. La legge della sillaba aperta. Atti del Sodalizio Glottologico Milanese 23: 60–98. Enrietti, Mario. 1987. L’apertura e la richiusura delle vocali in protoslavo. Europa Orientalis 6: 7–24. Enrietti, Mario. 1992–1993. Die zweite slavische Palatalisierung im Lichte der Sprachinterferenz. Ricerche Slavistiche 39/40: 7–27. (Contributi italiani all’XI Congresso Internazionale degli Slavisti, Bratislava, 30 agosto – 8 settembre 1993.) Enrietti, Mario. 1998. Osservazioni sull’origine e la scomparsa degli jer. In: Jerzy Rusek and Wiesław Boryś (eds.), Prasłowiańszczyzna i jej rozpad, 117–124. Warsaw: Wydawnictwo Energeia. Enrietti, Mario. 2001. L’origine e il diffondersi di št’ < šč e žd’ < ždž in slavo. Europa Orientalis 20 (1): 7–15.
82
Vít Boček
Enrietti, Mario. 2009. “Andere Zeiten, andere Lautgesetze”. Replica a Raffaele Caldarelli. Studi Slavistici 6: 309–312. Erhart, Adolf. 1976. Geneze indoevropských jazyků – diferenciace či integrace? [The genesis of Indo-European languages – a differentiation or an integration?] Listy Filologické 99 (4): 193–205. Erhart, Adolf. 2006. Ausgewählte Abhandlungen zur indogermanischen vergleichenden Sprachwissenschaft. Edited by Bohumil Vykypěl. Hamburg: Verlag Dr. Kovač. Falkenhahn, Viktor. 1968. Die sprachgeographische Lage der balto-slawischen Sprachen als Ursache für ihre Besonderheiten. Zeitschrift für Slawistik 13: 226–235. Filin, Fedot Petrovič. 1962. Obrazovanie jazyka vostočnyx slavjan [The Rise of the Language of the East Slavs]. Moscow/Leningrad: Izdatel’stvo Akademii nauk SSSR. Galton, Herbert. 1997a. Der Einfluß des Altaischen auf die Entstehung des Slavischen. Wiesbaden: Harrassowitz Verlag. Galton, Herbert. 1997b. Gedanken über den Ursprung des slavischen Verbalaspekts. Wiener Slavistisches Jahrbuch 43: 67–76. Gvozdanović, Jadranka. 2009. Celtic and Slavic and the Great Migrations. Reconstructing Linguistic Prehistory. Heidelberg: Carl Winter. Hall, Robert A. Jr. 1946. Bartoli’s ‘neolinguistica’. Language 22 (4): 273–283. Haspelmath, Martin. 1998. How young is Standard Average European? Language Sciences 20 (3): 271–287. Hock, Hans Henrich. 1991. Principles of Historical Linguistics. Berlin/New York: Mouton de Gruyter. Hock, Hans Henrich and Brian D. Joseph (eds.). 2009. Language History, Language Change, and Language Relationship. An Introduction to Historical and Comparative Linguistics. Berlin/New York: Mouton de Gruyter. Holzer, Georg. 2007. Historische Grammatik des Kroatischen. Einleitung und Lautgeschichte der Standardsprache. Frankfurt am Main: Peter Lang. Isačenko, Aleksander Vasil’evič. 1936. Zur Frage der “Zentralslavischen” Lautveränderungen. Der Wandel g > γ im Slovenischen. Sborník Matice Slovenskej 14: 56–63. Jakobson, Roman Osipovič. 1925–1926. Review of: Trávníček, František: Příspěvky k nauce o českém přízvuku. Brno 1924. Slavia 4: 805–816. Janhunen, Juha. 2001. On the paradigms of Uralic comparative studies. Finnisch-Ugrische Forschungen 56: 29–41. Joseph, Brian D. 2010. Language contact in the Balkans. In: Raymond Hickey (ed.), The Handbook of Language Contact, 618–633. Malden, MA: Wiley-Blackwell. Kahl, Thede, Michael Metzeltin, and Helmut Schaller (eds.). 2012. Balkanismen heute – Balkanisms Today – Balkanizmy segodnja. Vienna: LIT Verlag. Klimkowski, Tomasz. 2011. Romanitatea balcanică și contactele ei cu slavii [Balkan Romanians and their contacts with Slavs]. Annales Universitatis Apulensis. Series Philologica 12 (1): 325–338. Klimkowski, Tomasz. 2014. Preiotarea în limba română [Preiotacization in the Romanian language]. Studia Romanica Posnaniensia 51 (2): 41–57. Komárek, Miroslav. 2006. Studie z diachronní lingvistiky [Studies from Diachronic Linguistics]. Olomouc: Univerzita Palackého v Olomouci. Kuhn, Thomas. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press.
2 Common Slavic in the light of language contact and areal linguistics
83
Leschber, Corinna. 2007. Archaik, Reliktinseln und Kontinuität: Zu den arealen Faktoren in der historischen Sprachwissenschaft. Romanistik in Geschichte und Gegenwart 13 (2): 151–180. Leschber, Corinna. 2012. Kriterien zur Analyse von Slavismen im Rumänischen. Philologica Jassyensia 8: 167–179. Lindstedt, Jouko. 2000. Linguistic balkanization: contact-induced change by mutual reinforcement. In: Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in Contact, 231–246. (Studies in Slavic and General Linguistics 28.) Amsterdam/Atlanta,GA: Rodopi. Lindstedt, Jouko. 2014. Balkan Slavic and Balkan Romance: From congruence to convergence. In: Juliane Besters-Dilger, Cynthia Dermarkar, Stefan Pfänder, and Achim Rabus (eds.), Congruence in Contact-Induced Language Change: Language Families, Typological Resemblance, and Perceived Similarity, 168–183. (Linguae & litterae 27.) Berlin/Boston: Mouton de Gruyter. Lunt, Horace Gray. 1997. Common Slavic, Proto-Slavic, Pan-Slavic: what are we talking about? I. About phonology. International Journal of Slavic Linguistics and Poetics 41 (1): 7–67. Mareš, František Václav. 1999. Diachronische Phonologie des Ur- und Frühslavischen. Frankfurt am Main: Peter Lang. Mareš, František Václav. 2001. Diachronische Morphologie des Ur- und Frühslavischen. Frankfurt am Main: Peter Lang. Martynov, Viktor Vladimirovič. 1982. Stanovlenie praslavjanskogo jazyka po dannym slavjanoinojazyčnyx kontaktov [The Formation of Common Slavic on the Basis of Its Contacts with Other Languages]. Minsk: Nauka i texnika. Martynov, Viktor Vladimirovič. 1983. Jazyk v prostranstve i vremeni. K probleme glottogeneza slavjan [Language in Space and Time: On the Problem of the Glottogenesis of Slavs]. Moscow: Nauka. Menges, Karl. 1958. Influences altaïques en slave. Bulletin de la Classe des Lettres et des Sciences Morales et Politiques. 5e Série 44: 518–541. Menges, Karl. 1983. Slawisch-Altaische Kontakte. Die Slawischen Sprachen 4: 37–61. Mihăilă, Gheorghe. 1960. Împrumuturi vechi sud-slave în limba romînă. Studiu lexicosemantic [The Early South Slavic Borrowings in Romanian Language. A Lexico-Semantic Study]. Bucharest: Editura Academiei Republicii Populare Romîne. Novák, Ľudovít. 1939–1940. Neznáme nemecké vplyvy na západoslovanský, východoslovanský a maďarský konsonantizmus. Germanoslavistický príspevok k porovnávacej jazykovede stredoeuropskej (o zmenách γ (x) > h a r’ > ř) [Unknown German influences on West Slavic, East Slavic, and Hungarian consonant systems. A Germano-Slavic contribution to the comparative Central European linguistics (on the changes γ (x) > h and r’ > ř)]. Linguistica Slovaca 1/2: 106–117. Novák, Ľudovít. 1984. Vznik Slovanov a ich jazyka. (Základy etnogenézy Slovanov.) [The emergence of Slavs and their language. (The foundations of the ethnogenesis of Slavs)]. Slavica Slovaca 19 (3): 209–232. Paliga, Sorin. 2010. When could be dated ‘the earliest Slavic borrowings in Romanian’? Romanoslavica 46: 101–116. Paliga, Sorin. 2012. The trichotomical character of Proto-Slavic and the long-debated issue of the oldest Slavic borrowings in Romanian. In: Ilona Janyšková and Helena Karlíková (eds.), Theory and Empiricism in Slavonic Diachronic Linguistics, 347–363. (Studia etymologica Brunensia 15.) Prague: Nakladatelství Lidové noviny.
84
Vít Boček
Percival, W. Keith. 1976. The applicability of Kuhn’s paradigms to the history of linguistics. Language 52 (2): 285–294. Petrovici, Emil. 1956. Zum slavischen Einfluß auf das rumänische Laut- und Phonemsystem. In: Hans Holm Bielfeldt (ed.), Vorträge auf der Berliner Slawistentagung (11.–13. November 1954), 109–129. Berlin: Akademie-Verlag. Petrovici, Emil. 1957. Kann das Phonemsystem einer Sprache durch fremden Einfluß umgestaltet werden? Zum slavischen Einfluß auf das rumänische Lautsystem. ’S-Gravenhage: Mouton & Co. Petrovici, Emil. 1958. Javlenija singarmonizma v istoričeskoj fonetike rumynskogo jazyka – sledstvie slavjano-rumynskoj jazykovoj interferencii [The effects of synharmonism in the historical phonetics the Romanian language: a result of Slavic-Romanian linguistic interference]. Romanoslavica 2: 5–37. Petrucci, Peter R. 1999. Slavic Features in the History of Rumanian. Munich: LINCOM Europa. Pilarský, Jiří. 2001. Donausprachbund. Das arealistische Profil einer Sprachlandschaft. Inaugural dissertation, Department of Germanic Studies, University of Debrecen. Pisani, Vittore. 1935. Slavo e iranico. In: Bruno Migliorini and Vittore Pisani (eds.), Atti del III Congresso Internazionale dei Linguisti. Roma, 19–26 settembre 1933-XI, 371–379. Firenze: Monnier. Pisani, Vittore. 1940. Geolinguistica e indoeuropeo. Roma: Giovanni Baroli. Pisani, Vittore. 1974. Indogermanisch und Europa. Munich: Fink. Pohl, Walter. 1988. Die Awaren. Ein Steppenvolk in Mitteleuropa 567–822 n. Chr. Munich: Beck. Polák, Václav. 1964. Les éléments finno-ougriens en slave. Orbis 13: 568–588. Polák, Václav. 1970. Considérations sur l’origine de l’aspect verbal en slave. Orbis 19: 187–201. Pritsak, Omeljan. 1983. The Slavs and the Avars. In: Gli Slavi occidentali e meridionali nell’alto medioevo, Spoleto, 15–21 aprile 1982, Vol. 1, 353–435. (Settimane di studio del Centro italiano di studi sull’alto Medioevo 30.) Spoleto: Centro italiano di studi sull’alto Medioevo. Rocchi, Luciano. 1990. Latinismi e romanismi antichi nelle lingue slave meridionali. Udine: Campanotto. Rosetti, Αlexandru. 1968. Istoria limbii române de la origini pînă în secolul al XVII-lea [A History of the Romanian Language from Its Origins to the 17th Century]. Bucharest: Editura pentru Literatură. Sala, Marius. 2005. From Latin to Romanian: The Historical Development of Romanian in a Comparative Romance Context. [n.p.]: The University of Mississippi. Schelesniker, Herbert. 1991a. Slavisch und Indogermanisch. Der Weg des Slavischen zur sprachlichen Eigenständigkeit. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Schelesniker, Herbert. 1991b. Das slavische Verbalsystem und seine sprachhistorischen Grundlagen. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Schmitt-Brandt, Robert. 1971. Die Herausbildung der slavischen Sprachgemeinschaft. In: Robert Schmitt-Brandt (ed.), Donum Indogermanicum. Festgabe für Anton Scherer zum 70. Geburtstag, 224–243. Heidelberg: Carl Winter. Sedov, Valentin Vasil’evič. 1978. Slavjane i irancy v drevnosti [Slavs and Iranians in the distant past]. In: Istorija, kul’tura, etnografija i fol’klor slavjanskix narodov. VIII Meždunarodnyj s”ezd slavistov Zagreb – Ljubljana, sentjabr’ 1978 g. Doklady sovetskoj delegacii, 227–240. Moscow: Nauka.
2 Common Slavic in the light of language contact and areal linguistics
85
Seidel, Eugen. 1958. Elemente sintactice slave în limba romînă [Slavic Syntactic Elements in the Romanian Language]. Bucharest: Editura Academiei Republicii Populare Romîne. Shevelov, George Y. 1964. A Prehistory of Slavic. The Historical Phonology of Common Slavic. Heidelberg: Carl Winter. Solta, Georg Renatus. 1980. Einführung in die Balkanlinguistik mit besonderer Berücksichtigung des Substrats und des Balkanlateinischen. Darmstadt: Wissenschaftliche Buchgesellschaft. Stadnik, Elena. 2002. Die Palatalisierung in den Sprachen Europas und Asiens. Eine arealtypologische Untersuchung. Tübingen: Gunter Narr Verlag. Stanciu, Ioan. 2005. Die frühen Slawen in der rumänischen archäologischen Forschung. Kurze kritische Untersuchung. In: Piotr Kaczanowski and Michał Parczewski (eds.), Archeologia o początkach Słowian. Materiały z konferencji, Kraków, 19–21 listopada 2001, 567–582. Cracow: Księgarnia Akademicka. Stanciu, Ioan. 2013. The problem of the earliest Slavs in Intra-Carpathian Romania (Transylvania and the north-west vicinity). Slovenská Archeológia 61 (2): 323–370. Thomas, George. 2008. Exploring the parameters of a Central European Sprachbund. Canadian Slavonic Papers 50 (1/2): 123–153. Thomason, Sarah Grey. 2001. Language Contact: An Introduction. Edinburgh: Edinburgh University Press. Thomason, Sarah Grey and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley/Los Angeles, CA: University of California Press. Trubačev, Oleg Nikolaevič. 1983. Jazykoznanie i etnogenez slavjan. Drevnie slavjane po dannym etimologii i onomastiki [Linguistics and the ethnogenesis of Slavs. Ancient Slavs in the light of etymology and onomastics]. In: Samuil Borisovič Bernštejn, Viktor Ivanovič Borkovskij, Nikita Il’ič Tolstoj, and Oleg Nikolaevič Trubačev (eds.), Slavjanskoe jazykoznanie. IX Meždunarodnyj s”ezd slavistov Kiev, sentjabr’ 1983 g. Doklady sovetskoj delegacii, 231–270. Moscow: Nauka. Trubetzkoy, Nikolai Sergeevič [Nikolaj Sergeevič Trubetzkoj]. 1939. Gedanken über das Indogermanenproblem. Acta Linguistica. Revue Internationale de Linguistigue Structurale 1: 81–89. Trubetzkoj, Nikolaj Sergeevič. 1988. Opera Slavica Minora Linguistica. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Trummer, Manfred. 1983. Slawisches in der Entwicklung des rumänischen Lautsystems. Die Slawischen Sprachen 4: 133–142. Uhlenbeck, Christianus Cornelius. 1934. Oer-Indogermaansch en Oer-Indogermanen. Mededeelingen der Koninklijke Akademie van Wetenschappen, Afdeeling Letterkunde A 77 (4): 125–148. Vaillant, André. 1966. Grammaire comparée des langues slaves, Vol. 3: Le verbe. Paris: Éditions Klincksieck. Vasiliu, Emanuel. 1968. Fonologia istorică a dialectelor dacoromâne [A Historical Phonology of the Daco-Romanian Dialects]. Bucharest: Editura Academiei Republicii Socialiste România. Vermeer, Willem. 2003. Comedy of errors or inexorable advance? Exploring the dysfunctionality of the debate about the progressive palatalization of Slavic. In: Jos Schaeken, Peter Houtzagers, and Janneke Kalsbeek (eds.), Dutch Contributions to the Thirteenth International Congress of Slavists, Ljubljana: Linguistics, 397–452. (Studies in Slavic and General Linguistics 30.) Amsterdam/New York: Rodopi.
86
Vít Boček
Weiher, Eckhard. 1967. Urslavisch-Gemeinslavisch-Dialekte des Gemeinslavischen (?). Anzeiger für slavische Philologie 2: 82–100. Wolfram, Herwig. 1997. The ethno-political entities in the region of the Upper and Middle Danube in the 6th–9th centuries A.D. In: Przemysław Urbańczyk (ed.), Origins of Central Europe, 45–57. Warsaw: Stowarzyszenie Naukowe Archeologów Polskich. Zaliznjak, Andrej Anatol’evič. 1963. O xaraktere jazykovogo kontakta meždu slavjanskimi i skifo-sarmatskimi plemenami [On the nature of linguistic contact between Slavic and Scytho-Sarmatian tribes]. Slavjanskoe jazykoznanie. Kratkie soobščenija Instituta slavjanovedenija 38: 3–22.
Robert Orr
3 Intertwining trees, eddies, and tentacles — some thoughts on linguistic relationships in Europe, mainly Slavicnon-Slavic 1 Introduction: Multi-level linguistic relations in Europe This chapter will focus on various apparently neglected aspects of the complex interplay of language interrelations in Europe, already covered in a vast and still growing literature, mainly consisting in a detailed “molecular-level” look at certain items. Its approach will involve a good deal of cross-referencing, as several of the lines of thought to be followed are intertwined, and do not admit easy disentangling. To take just one example, leaving aside for the moment the further complicating issue of Balto-Slavic relations,1 any research into diachronic/synchronic linguistic typologies, contacts, parallels, eddies,2 sub- and superstrata in Europe3 citing Slavic data has recently been showing a bewildering amount of complexity, with possible leads involving Romance, Germanic, Albanian, Greek, Iranian,4 Uralic, Turkic, and Celtic, in addition to intra-Slavic relations.5 Celtic6 is the only
1 For recent work, see, e.g., Boček (2014: 370–374) for an overall view. 2 Following Orr (2003a: 272), an eddy is “any current that goes against the general trend of an overall linguistic change, or is at the periphery of such a change.” 3 Up to now much of the pertinent research and discussion of the broader themes covered in this paper has revolved around the issues of sub- and superstrata, and these will be touched on throughout this article. That terminology will continue to be used here, even if the actual concepts need a lot of work, since most of the material to be cited involves discussion of the work of other scholars rather than adding fresh comments. 4 At least up to the time of Christ, and probably later, Iranian extended far into Europe, and far further northwards in Asia than its current range, with numerous points of contact with various European and Uralic languages; one relic from that period that might engender some surprise for the modern reader is the presence of a larger number of Iranian elements in Polish than in other Slavic languages, see Trubačev (1965). 5 Cf. most recently Boček (2014), and the citations therein (see also his table of contents, which lists the following potential linguistic contacts: Slavic-Iranian, Slavic-Thracian, SlavicCeltic, Slavic-Germanic, Slavic-Albanian, Slavic-Romance, and Slavic-Uralic). 6 In this context it might be noted that an adequate knowledge of Celtic is often lacking among linguists working on pan-European issues, which often gives rise to treatments of https://doi.org/10.1515/9783110639223-004
88
Robert Orr
one of these groups which has had no direct contact with Slavic for at least a millennium, although relationships continue to be postulated.7 For Slavic-Celtic relations, such proposals may be grouped into two: suggestions of prehistoric contacts, most recently Gvozdanović in a series of works (e.g., Gvozdanović 2008, 2009), possibly before the two groups were separated8 by the respective expansions of Germanic and Romance, and suggestions of typological parallels, with both groups tending to occupy Europe’s Eastern and Western peripheries respectively (see Orr 1992). For an earlier period (up to 1000 BC), one might reconstruct two more language families in Europe north of the Alps: Atlantic (Hamito-Semitic; mainly Europe’s Western littoral as far north as South-Western Norway), and Old European (surviving in Basque, albeit currently covering a much truncated area), which would be descended from very different protolanguages than Proto-Indo-European and Uralic. Most relationships involving Proto-IndoEuropean and Uralic, on the one hand, and Atlantic and Old European, on the other, and individual descendants of all those groups, however, would have been submerged by subsequent migrations and invasions. Most recently, Vennemann (1998, 1999) has provided the launching pad for a comprehensive view of Atlantic-Old-European-Indo-European linguistic relations; much of his research, however, has centered on the reconstruction of a Semitic superstratum, imposed by Phoenicians and mainly in languages close to Europe’s North-West coasts, which he uses to explain several problematic etymologies,
Celtic material and individual forms leaving something to be desired. Apart from the all too frequent misinterpretation of forms, one example involves the discussion of prefixation as illustrated by Décsy (2000: 367–369) who states that “ . . . the Celtic languages know of no praefixia verbalia”. Actually, Common Celtic inherited an extensive set of prefixes from Indo-European, and most historical Celtic grammars, either of various stages of individual languages (e.g., Calder 1923: 261–88; Thurneysen 1980: 495–524) or of Celtic as a whole (Lewis and Pedersen 1961: 69–70; 245–251; 252–267, 328–403), devote major sections to them. This is not immediately apparent from the modern Celtic languages without extensive study. Historical Celtic studies are made more complicated by the massive lack of early documentation (although recently the situation has improved somewhat); it is almost as if French were attested, but hardly any Latin, and scholars had to work out the affinities of French thus handicapped. Even the names of major Celtic languages can give rise to confusion; see, e.g., Gvozdanović (2008: 163, 2009: 35, 252–253); Kurkina (2012: 146) cites the same form, troed, as uel. (Welsh) and kimr. (Cymric) on the same page. 7 It might also be noted that, as Comrie (1981: 142) points out, Celtic was the only living IndoEuropean branch not represented on the territory of the former USSR. 8 We can only speculate about the very existence of other, earlier, or maybe even later expansions which split Celtic from Slavic; possibly even Holzer’s (1989: 14) “Temematisch”.
3 Intertwining trees, eddies, and tentacles
89
mainly Germanic,9 while at the same time suggesting that the large number of Semitic-Celtic syntactic matches might also be due to the reconstructed superstratum. These Semitic-Celtic syntactic matches, however, appear to be typological parallels rather than the result of Semitic influence on Celtic: the basic VerbSubject-Object word order, with all its implications, plus the lack of a verb have,10 possessive phrases where only one noun may be marked as overtly definite, e.g., Welsh llyfr y dyn ((the) book the man) ‘the man’s book’; llyfr fy mab ((the) book my son) ‘my son’s book’,11 cf. the Semitic idaafa construction (construct state), e.g., Arabic kitāb-u (book-Nominative) al-rajul-i (the man-Genitive) ‘the man’s book’,12 where the presence of one definite article renders the whole phrase definite (Brockelmann 1985: 163–169); plus the widespread use of prepositional pronouns common to both groups. The languages of Europe are often viewed as being grouped round a Germanic-Romance linguistic core, “Standard Average European” (SAE),13 often described as a Sprachbund, which offers a useful and conventional starting point. All studies of SAE should include discussion of the division of the languages of Europe into H-languages (languages making extensive use of a verb have, and the other syntactic features entailed; see Isačenko 1974; more recently Grković-Major 2010) and B-languages (languages making restricted use of such a verb, or lacking it altogether, and using various constructions with a verb which might be loosely glossed as ‘be’).14 Isačenko (1974: 44) suggested
9 Vennemann’s proposals mainly involve putative borrowings from Atlantic and Old European into Germanic, and many of them are good as far as they go, but continuing the discussion might yield some more interesting results. Many of the forms appear to admit of alternative explanations with, for instance, perfectly good Slavic cognates and Indo-European pedigrees. 10 Actually, it is the presence of a transitive verb have that is quite rare cross-linguistically: the widespread distribution of have in Europe has distorted perceptions, see, for instance, Orr (2002) and the literature cited therein; for a recent interpretation, see Drinka (2017). 11 This construction is very common in place names all over Celtic countries, see note 20 below for a further example. It should be noted that in contrast to the Arabic and Goidelic constructions, the Welsh construction has no cases; the juxtaposition noun-article-noun is sufficient. 12 No definite article can be reconstructed for Proto-Semitic, which supports the claim that these similarities are typological, and not due to contact. 13 For more on the term “Standard Average European”, and detailed discussion, see also Dahl 1990. The concept goes back at least to Whorf ([1941] 1956: 138). 14 This topic is further complicated by the distinction between what might be described as existential and copular be, a distinction not treated by Isačenko. In any case, have and be are both fairly vague concepts; the main difference may be seen as transitive (have) and
90
Robert Orr
that within Europe, all the Romance and Germanic languages are H-languages, and all the Uralic languages are B-languages, with the boundary between the two types actually crossing Slavic and Baltic territory, including some degree of transition zones from B- to H-languages, thus approximating closely to SAE. A conventional view of the range of SAE, which fairly closely corresponds to the area covered by Isačenko’s H-languages, is given by Haspelmath (2001: 53): “The core members of this Sprachbund are the Continental West-Germanic and Gallo-Romance languages (especially Dutch, German, French), but it also comprises the other Romance and Germanic languages and the Balkan languages [unmentioned by Isačenko – RAO], as well as the Slavic languages (particularly West Slavic)”. Elsewhere (Orr 1992), I suggest that Celtic is crossed by an H-B line similar to that proposed by Isačenko, with numerous transition zones similar to those described for Slavic and Baltic above, with the Goidelic group, possibly apart from later Manx, providing almost canonical examples of the B-pattern, and Brythonic, especially Breton, tending towards the H-pattern, and go on to suggest syntactic parallels between Slavic and Celtic as an area of possible fruitful research, shedding new light on several problems in Slavic linguistics (see also Clancy 2010). In the context of the modification to Isačenko (1974) suggested in Orr (1992), therefore, SAE might better be described as “Standard Central European (SCE)”,15 and this usage will be followed in this paper as far as possible, although SAE may still have to be used in certain contexts. The proposed SCE framework is based on the distinction between H- and B-languages. In this context Celtic data are an important, albeit often ignored,
intransitive (be). This may also be illustrated by Putonghua yŏu 有 ‘there is/has’ (negative méiyŏu 没有) used to convey both ideas, almost as if, for instance, French were to use il y a ‘there is/there are’ both as intransitive/existential be and transitive have: Putonghua (1) Wŏ yŏu yīběn shū 我有一本書 ‘I have a book’ [transitive] (2) Wŏ hái méiyŏu qùguo Běijīng 我還沒有去過北京 ‘I have not been to Beijing yet’ [intransitive] (3) Mén koŭ yŏu rén 門口有人 ‘There is someone at the door’ [intransitive] French (1) Moi il-y-a un-volume livre ‘Me there-is one-volume book’ (2) Moi encore il-n’y-a-pas aller-passer Beijing ‘Me yet there-is-no go-pass Beijing’ (3) Porte bouche il-y-a homme ‘Door mouth there-is man’ Something similar actually seems to have developed in Brythonic, see Orr (1992: 252–253) and the literature cited therein, and below. 15 See Kurzová (1996, and also in this volume), for discussion of a proposed “Central European” Sprachbund narrower than SCE.
3 Intertwining trees, eddies, and tentacles
91
part of the approach to delimiting SCE’s Western boundary, with Slavic being an important component of its Eastern, and therefore some space will be allotted to the citation thereof, and the comparison with Slavic. Briefly, within both Slavic and Celtic, the evolution of have itself is also complicated, offering some common partial matches. All Slavic languages use, to varying degrees, forms of a verb have which might be reconstructed as *jьměti (see Meillet 1924: 203), for instance, Serbo-Croatian imam knjigu ‘I have a book’, often coexisting with a possessive construction using the preposition u ‘at’ to express have, most widespread in Russian, albeit present in every other Slavic language, with the apparent exception of Slovenian (Orr 1992: 247–251). These constructions appear to have supplanted an earlier pattern, attested in Old Church Slavonic, using be plus a dative (see Grković-Major 2010: 40). The developments involving have seems to have taken a partially similar course in Celtic to that outlined for Slavic. Both Old Irish and Brythonic had constructions using be as SCE uses have plus bound pronominal forms with dative force, e.g., Old Irish táth-um (is-to-me) ‘I have’; báth-um (was-to-me) ‘I had’; Breton ur velo c’hlas am eus (a bike blue to-me is) ‘I have a blue bike’, later replaced in Irish with a construction using the preposition ag, with Brythonic developments being a little more complicated (Orr 1992, 2002).16 There is some degree of variation in the developments,17 but it should be noted that Russian and Irish show a striking parallel: constructions close to the compound H-perfects in SCE have their agents marked by u and ag respectively, with these agents appearing to show some subject properties (Timberlake 1976: 554–562; Orr 1989: 9–11, 1992: 252–254). In contrast to Slavic, however, no verb have seems to have arisen in Goidelic. B-languages use a variety of nominalized forms to express modality, in contrast to the modal verbs used in H-languages, as may be illustrated by juxtaposing some examples from Russian (B) and Czech (H). While Czech makes much use of verbal constructions with close parallels in Western European languages such as, for instance, nemáš kouřit ‘you should not smoke’, Russian tends to express modality using either adjectival constructions, e. g., ja dolžen ‘I must’ or constructions with the logical subject in the dative accompanied by indeclinable forms, which are often referred to as predicatives or members of a special class called the “category of state” (Russian kategorija sostojanija), e.g., mne (I.DAT) nado (necessary-Predicative) ‘I need’. Furthermore, in some Slavic H and transitional 16 Welsh uses a construction with the prepositions gan or gyda ‘with’, e.g., mae llyfr gen i/ gyda fi ‘is book with me/with me ‘I have a book’. 17 For Bulgarian, Mirčev (1971: 83) documents an eddy, where the use of u ‘at’ to denote possession expanded and contracted, see also Orr (2003a: 274–275).
92
Robert Orr
H-B languages, such nominals have acquired verbal endings, e. g., formally Czech nelze and Russian nel’zja are identical, but nelze can take a past tense ending in certain registers, e.g., nelzelo, where *nel’zjalo would be ungrammatical in Russian. Sometimes verbs and predicatives are borrowed from H-languages into B-languages, and change their status from verbs to predicatives or vice versa, e. g., Manx laik (predicative) from English like; laikal (verb < laik < like). In contrast to Czech musit, ‘must’, ultimately from German muss (müssen; see Isačenko 1974: 75), Hungarian has muszáj < muss, adapted to the B-pattern, similarly Upper Sorbian dyrbjeć ‘have to’ < Old High German durfan, Breton fellout < French falloir (Orr 1992: 261–264).18 The term “Central” in SCE, however, is called into some question when one notes that within Europe Icelandic and Faroese are located even further to the West than Celtic, and yet their structure is far closer to the SAE/SCE template. One might, therefore, suggest that the spread of Scandinavian to the Faroes and Iceland from Norway postdated the earliest stages of the formation of SAE/ SCE. Lockwood (1977a) notes several idioms and collocations in Faroese which clearly arose under Gaelic influence, as well as some phonological/phonetic points of contact, for instance, preaspiration in Scottish Gaelic, Icelandic and Faroese. These can be traced back to the Viking period. And yet Gaelic nonphonological/phonetic influence on Icelandic and Faroese does not seem to have gone beyond the phraseological; Greene’s (1979) distinction between grammaticalized verbal paradigms and “free syntactic constructions” is useful here. Faroese has an idiom ótti er á honum ‘fear is on him’/ótti kom á hann ‘fear came on him’, clearly modelled on (Scottish) Gaelic tha an t-eagal air ‘is the fear on-him’/ thàinig an t-eagal air ‘came the fear on-him’, absent from Lockwood (1977a), and coexisting with a verb óttist ‘to fear’. Some uses of the Faroese preposition hjá actually recall constructions with Russian u and Irish ag primarily ‘by’, ‘with’, and French chez; it can also be used in some genitive constructions, e.g., hesturin hjá mær horse-the at me – ‘my horse’. From a glance at Lockwood (1977b: 144–152; Thrainsson et al. 2004: 149–151, 304–310), however, it may be noted that Faroese has a fully functioning verb hava ‘to have’, plus a verb eiga ‘to own’, as well as the full range of
18 One excellent example of a B-H transition in Brythonic is provided by Welsh piau/Cornish pewo/Breton biou, originally an interrogative (‘to-whom-is?’), which subsequently developed into a relative verbal form (‘to-whom-is’). In Modern Welsh it is normally classed as a defective verb (the only form remaining is piau (present), and it behaves like a transitive verb; Middle Breton further developed an infinitive piaout ‘to own’, see Orr (1992: 255–256) and the literature cited therein.
3 Intertwining trees, eddies, and tentacles
93
modal verbs typical of SCE. In any case, currently contact between Insular Celtic and Insular Scandinavian is probably between very low and non-existent, thus pretty well eliminating the possibility of ongoing contacts,19 in contrast to situations elsewhere in Europe. One might tentatively suggest that by the time of large-scale contacts between Celtic and Scandinavian the H-B patterns had already firmly taken shape, and in the case of those two groups raw geography, changing trade patterns, and finally language death, made it far easier for such contacts to be broken and that such contacts did not last long enough or run deep enough to exert the kind of influence we see further south and east in Europe (see also below).20 Another approach to the comprehensive classification and overview of the languages of Europe is taken by Décsy (2000). He suggests the division of Europe into eight linguistic zones, all of which cut across both genetic and typological boundaries. Some of his criteria are a little surprising: he also cites “Standard Average European” (Décsy 2000: 54–77, 493), but his definition thereof, however, is different from the commonly accepted meaning of the term. He takes SAE (SCE) to mean the five European languages with the most speakers—English, French, German, Italian, and Russian, even though the lastnamed is well outside the area normally covered by SCE. Most of the features he enumerates, whether cultural or linguistic, apply to all European languages,
19 Such “contacts” as exist nowadays are mostly manifested through the medium of English. I recall a conversation in Gweedore, Donegal, where I was told by a local, who had worked on a boat where all the rest of the crew were from Eriskay, that Scottish Gaelic was just “Irish with a Norwegian accent”. On being told this anecdote a senior professor pooh-poohed it, asking how much contact there could be between Norway and the zone covered by the West of Ireland and the Scottish Islands. I replied by citing, rather feebly, frequent visits by Norwegian scholars such as Alf Sommerfelt and Magne Oftedal. Later on I realized that the “Norwegian accent” cited in the conversation probably referred to imitations heard through the medium of English stereotypes. 20 The period 1263–1266 (the Battle of Largs and the Treaty of Perth) might be suggested as a terminus circa quem for the breaking of Scandinavian-Celtic contacts, a possible textbook example of Entwistle and Morison’s “stretching of elastic until breaking point” (see also fn. 30 below). Drummond (1991: 140) suggests that the actual linguistic boundary between Gaelic and Scandinavian may have been fixed on the island of Canna, whose high-point is called Carn a’Ghaill ‘cairn of the foreigner [Scandinavians in this case – RAO]’. There is also Beinn Hògh, on the nearby island of Coll, which may well be a “River Avon”-type name – “hill hill” from Beinn Haugr, where Gaelic beinn ‘hill’ is prefixed to Scandinavian haugr, admittedly ‘burial mound’ in Old Icelandic, but Faroese heyggjur, Norwegian haug both mean ‘hill’, and it is possible that in Western Isles Scandinavian, which died out sometime after 1263–1266, haugr may also have had that meaning. Currently, of course, English has spread to cover almost all the areas where Scandinavian-Celtic contacts may have taken place.
94
Robert Orr
except Russian in two instances. Elsewhere he does concede that other languages also have features in common with SCE, for example, Danish (Décsy 2000: 81). Some of his zones are self-explanatory, for instance, the Viking Zone includes Scandinavian, Lapp, and various other Finno-Ugric languages, plus the extant Celtic languages (see also Orr (1994/95) for a view of the interrelations in Décsy’s Viking Zone from a slightly different angle), and his Rokytno Zone, named after Rokytno (Ukrainian Rokytne) in Ukraine includes Polish, Lithuanian, Belarusian, Ukrainian, Kashubian,21 thus closely corresponding to Isačenko’s eastern H-B transitional zone.22
2 Family trees, shifting centers and peripheries, intersections, lost contacts, substratum theories, “linguistic rings”, the case of Hungarian, tentacles reaching around peripheries This part of the paper will include brief discussion of the complexities of family trees, shifting centers and peripheries, intersections, lost contacts, substratum theories, “linguistic rings”, the case of Hungarian, tentacles reaching around peripheries, and their interrelations.
2.1 Family trees Nevertheless, linguistic interrelations in Europe are much more complicated than one might suppose by citing the useful, albeit crude, demarcation represented by
21 Décsy’s other zones, the Littoral Zone, the Peipus Zone, the Danube Zone, the Balkan Zone, and the Kama Zone, are also grouped according to his own idiosyncratic sets of criteria, which are occasionally a little unclear (see Orr 2009). 22 Oddly enough for such an extensive treatment, Décsy (2000: 364–365, 386–387) offers far too brief a discussion of have (in the broadest sense, subsuming both possession and auxiliation) and the related distribution of be and have, omitting a great deal of the relevant literature, especially Isačenko’s seminal article. Another major omission in his framework comes right at the beginning, when he suggests that Proto-Indo-European and Uralic are the “original European protolanguages”, omitting any mention of Basque or Hamito-Semitic, although hinting at the presence of other languages (Décsy 2000: 23, 391–392). In this context the omission of Basque is astonishing, although elsewhere Décsy (2000: 455) does acknowledge that Basque does not belong to either Indo-European or Uralic.
3 Intertwining trees, eddies, and tentacles
95
SAE/SCE, East/West, or have/be. There are several other ways of viewing European linguistic typology, often going to the “molecular level”. Languages do not have to influence each other across lengthy boundaries. There is also an abundance of more tenuous pathways of influence, down to and including the often unique trajectories travelled by individual words.23 The family tree has been used for illustrating linguistic relationships for a very long time, although linguists have long been questioning its utility. One problem is that family trees illustrate historical, and not modern/typological realities. Recently, there have been several attempts to deny the very validity of the family tree. Dixon (1997: 28–53), in particular, has offered an alternative framework involving the concepts of diffusion and punctuated equilibrium, based mainly on his work on Australian languages, and suggests that the family tree cannot be applied to many language families as well as it can to, for instance, Indo-European, Uralic, Semitic (see Orr 1999: 140, 2005; McWhorter 2001: 122–123). According to Dixon’s adaptation, languages and language families also exist for long periods in states of rough equilibrium, “punctuated” by disturbances, after which new states of equilibrium gradually emerge, reminiscent of the oft-discussed Balkan Sprachbund, and he suggests that reconstructed proto-languages are probably the results of periods of linguistic equilibrium. He goes on to suggest (1997: 54–56) that during periods of rapid language change new categories tend to appear suddenly, for example, the development of postposition-based systems into case systems; the shift of ergative systems to accusative systems, but that such categories tend to be lost gradually. From a close reading of Dixon’s relevant passages, however, he actually seems to arguing that family tree diagrams should be far more complex than normally presented, rather than ditching them altogether. Therefore, one modification that might be considered is a much more intricate model, analogous to bushes which can sprout branches which later can fuse again, or can fuse with branches from other bushes, or may even be cut and forced together, for example, fungal mycelia bushes, which might extend the analogy further.24 Dawkins (1988: 259–260) draws attention to the fact that animal species never merge or converge, or jump from one branch of a family tree to another
23 Cf. the dictum attributed to Gilliéron (actually traceable back to Jakob Grimm; see Orr 1993: 303, editor’s note) that “every word has its own history.” 24 One should also cite Shevelov’s (1964: 607) likening of language development to “clouds in the sky on a stormy day”; Entwistle and Morison’s (1964: 23) “the separation of the IndoEuropean peoples to be like the stretching of elastic until breaking point”; Wolpoff and Caspari’s (1997: 283) “river splitting up into channels, recombining, and then splitting up again”.
96
Robert Orr
(let alone from one tree to another!), whereas different languages often do, citing the examples of Icelandic as opposed to Swedish, Norwegian, and Danish, and Modern English, and suggesting that Icelandic had split off from the Continental Scandinavian languages. Dawkins’ selected example illustrates the differences even more clearly than he puts it. The earliest division in Scandinavian is actually between West (Norwegian, Faroese, Icelandic) and East (Swedish, Danish); the similarities between Swedish and Danish, on the one hand, and Norwegian, on the other, are due to later parallel evolution and typological convergence (large-scale loss of morphology, etc.), thus illustrating the essentially historical nature of family trees. The division into Continental and Insular Scandinavian is later (see Haugen 1976: 222), or indeed, to put it another way, Continental Scandinavian is the result of later convergence (see Alter 1999: 102). Gaps in attested material often lead to incomplete family tree diagrams, even in well-attested languages and language families such as SCE. One example is provided by Boček (2014: 17), who reproduces from Radoslav Katičić a complex, albeit incomplete, family tree diagram for English, omitting any mention of influence from East Indian languages, Slavic, and, surprisingly, Celtic. On the other hand, areas such as isolated islands or mountain valleys offer the best environments for classical family trees to develop. Europe as a whole, with its mixture of plains and mountain ranges crisscrossed by rivers and passes, may well not be an ideal example of such an area, although individual localities may well be. Rouse (1992: 39) offers a brief description of how linguists construct family trees, and suggests an example which is conducive to being illustrated by family tree diagrams, and also providing support for Entwistle and Morison’s “stretching of elastic until breaking point”: . . . the situation among the ancestors of the Tainos is favourable to the establishment of family trees. The ancestral speakers expanded linearly, first along the major rivers of northeastern South America and then along the chain of islands that constitutes the West Indies. As a result, each successive speech community became more and more isolated from those that preceded it, with less and less opportunity for strong interaction with them.
Meanwhile, Marcantonio (2002) mounts a comprehensive attack on the concept of the family tree as applied to Uralic, usually considered one of the most solid family-tree-based reconstructions, and cited as such by Dixon (1997: 28). She (Marcantonio 2002: 277) concludes by drawing a useful parallel with Gould’s (1991) treatment of the Cambrian Explosion, and his shoehorn metaphor, to illustrate how Samoyed came to be classified as a Uralic language, proposing that:
3 Intertwining trees, eddies, and tentacles
97
Early researchers were seduced by the Darwinian theory, and they shoe-horned what little evidence that they had into a Darwinian model of language development. . . . . At the end of the 19th Century, the evidence [for classifying Samoyed – RO] was highly equivocal, with the correlations between Samoyed and other Altaic languages being approximately equivalent to the correlations with the Ugric and Finnic languages. However, in order to fit the Darwinian model, researchers were obliged to jump in one direction or the other. They had to make a simple and clear choice of one originating parent language. Accordingly, Samoyed was classified as a Uralic language on nothing more than the ‘on-balance’ opinion of one author (Donner). It is as though the scientific paradigm of the time would accept nothing less.
Family trees may be viewed from at least two angles: either starting from the roots and tracing the history via the branches and twigs, or starting from an individual twig and tracing it back to the roots. Most linguistic histories start off from the roots, with, perhaps, Baudouin de Courtenay (1984: 24), Strang (1970), constituting exceptions. For an application of the second approach from biological evolution, see Dawkins (2005), which is loosely modeled after Chaucer’s Canterbury Tales, where the eponymous tales themselves, some originals, and others not, are contained inside a frame tale narrated by a group of pilgrims on the way from Southwark to Canterbury Cathedral to visit the shrine of Saint Thomas Becket. Dawkins’ animal pilgrims embark on a path to find their common ancestor, many contributing to the overall tale of the evolution of life. Following this model, Dawkins describes the tracing of life itself, right up to its very source. Starting with modern humans, Dawkins conducts us on a “pilgrimage” back to the hypothesized origin of life, backwards along the branches of a sort of reverse family tree to the roots, encountering on the way larger and larger groups of related life-forms. How would a comprehensive Family Tree covering European languages look? Obviously it would include the fungal mycelia bush analogy, with extensive intertwinings. To take the documented history of Russian, one would have to include, for instance, Belarusian, Ukrainian,25 and a multiplicity of Baltic, Uralic, and Turkic languages, later followed by Bulgarian, Polish, German, French, etc., recently overwhelmed by English, presenting a great deal of difficulty in such an exercise (see Dixon 1997: 45–46). To apply a modification of Dawkins’ approach (twig to roots) to the history of Russian, from the point of view of direct genetic ancestry, it would first emerge from East Slavic dialect combinations and recombinations, as Slavs and their speech spread eastwards and northeastwards, rendezvousing with 25 “Russian”, “Belarusian”, and “Ukrainian” are used here to denote later formations.
98
Robert Orr
many extinct Baltic, Uralic, Iranian, and Turkic26 languages and dialects, some of which have not even left behind their glossonyms (see Nuorluoto 2006, and the review thereof in Orr 2012). Russian would then rendezvous with Byzantine Greek via Old East Slavic, followed by Western European languages, whether directly or via (Czech) Polish/Ukrainian/Belarusian, culminating in the currently overwhelming influence of English. Such contacts sometimes involve breaks of varying length, followed by resumptions. Even for languages such as English, geographically far more isolated than Russian before around 1600, one would have to include intertwinings with at least Scots, West Indian Creoles, Forth and Bargy Hiberno-English, etc. To adapt Dawkins’ approach, from the point of view of direct genetic ancestry English would in theory first rendezvous with the other Ingvæonic languages, for example, Frisian. However, at the same time English would have had close encounters with French, which left a major mark, and then Scandinavian, followed by a rendezvous at the Common West Germanic and Common Germanic nodes. As with Russian, there is at least one example of contact being lost and then resumed: English and Scandinavian, see Townend (2002: 41). But even the citation of such intertwinings does not begin to encompass the potential complexity of language contact and family trees. It is a truism that all language contact goes on inside an individual speaker’s head. To gain a complete picture, therefore, one might have to draw any tree diagram to accommodate myriad contacts involving numerous influential individual speakers. The reach and extent of attested language change has often obscured this fundamental point. On occasion it just takes one speaker to have a role in language change: Peter the Great might be said to be solely responsible for Russian-Dutch contacts, in the form of loanwords from Dutch into Russian (Kiparsky 1975: 111–121). Another issue is raised by the status of literary canons: great works which provide bodies of quotes which become part of the idiomatic stock of the respective language. If such canons are, in turn, under heavy foreign influence, this may have some effect on the language. Dr. Johnson’s well-known assessment of John Milton comes to mind: “He was desirous to use English words with a foreign [Latin] idiom”, which had a palpable effect on the structure of any of his phrases. A more startling example is provided by Morzinski (1994), who shows that Joseph Conrad, usually considered a master of English style, was
26 The actual forces involved in the “Mongol” invasion of Kyivan Rus’ would have been mostly Turkic-speaking.
3 Intertwining trees, eddies, and tentacles
99
actually strongly influenced by his native Polish; Morzinski (1994: 43–67, 69–92) devotes whole chapters to areas such as aspect and Conrad’s use of voice. The issue is further complicated by a discernible reduction in Polish influence over Conrad’s literary career (Morzinski 1994: 118). How, then, are such contacts as described for Milton, Conrad, or Peter the Great to be captured in any linguistic family tree, and what would that do for any sort of completeness? Perhaps part of the problem is the family tree model itself: tree diagrams tend to be oversimplified. To paraphrase Anttila (2000: 107): “Family tree diagrams are there, and we have to live with them.”
2.2 Shifting Centres and Peripheries Cyxun (2012, 2013) has recently synthesized suggestions that in any sort of center-periphery framework changes often take place on various peripheries of the territory covered by a given language family, as opposed to the commonly cited contrast of an innovating center and conservative peripheries. He points out that within Slavic alone there are a wide variety of linguistic phenomena suggesting that innovations can spread in any direction, and are complicated by the emergence of several centers within the Slavic linguistic area, e.g., Belarusian, Slovak, and Ukrainian, with Macedonian, Lower Sorbian, and Slovenian emerging as peripheries. The split into North and South Slavic would have come later.
2.3 Intersections Focusing mainly on Belarusian, Marvan (2013: 175) suggests that the main axis of linguistic influence in Europe is Greek > Latin. He identifies four more minor axes with the cardinal points of the compass, and concludes with the suggestion that Belarusian is at the junction of all those minor axes crossing Europe, all of which exert mutual influences, covering areas such as morphology, phraseology, and so forth (Marvan 2013: 177). Marvan’s framework might be adapted to cover other European languages.
2.4 Lost contacts One area worth scrutinizing is where former links have been almost obliterated by geography, migration, intrusion, language death, and the like. Two
100
Robert Orr
examples of lost contacts have already been discussed above: Celtic/Slavic and Celtic/Scandinavian. A further example involves Slavic/Old English as explored by Martynov (1996: 87, 1998: 5–6), part of whose extensive research into the comparatively neglected area of Slavic borrowings into Germanic,27 includes Slavic loanwords unique to Old English and absent from the rest of Germanic. Any such Slavic forms would have been borrowed into the immediate pre-Old English stage of Germanic when the Angles, Saxons, and Jutes were still on the Continent, or when contacts were still maintained.28 Some of Martynov’s more convincing examples include: Common Slavic *lěsъ /Old English lǣs ‘forest’ < Indo-European *loiḱ-; cf. Latin lūcus29 (Martynov 1996: 86–87, 1998: 14–15); Common Slavic *světьlъ < Old English sweotol ‘clear’ (Martynov 1996: 87, 1998: 23–24; see also Orr 2016), for which an etymology derived from CS *světьlъ looks almost too good to be true. Another example, not cited by Martynov, is English warlock < Germanic *wārlog < Iranian (ScythoSarmatian) *wrkolak, which followed a trajectory Iranian > Slavic > Germanic (Nichols 1987).
2.5 Substratum theories Many scholars view Russian as being built on a Finno-Ugric substratum; it has often been proposed that the increased use of B-constructions in Russian and Latvian is an innovation, partially due to Finno-Ugric influence, although see also Clancy (2010: xvi, 126, 128, 247) for further discussion. The nominalized modal forms cited above for B-languages, of course, constitute one set of constructions where advocates of substratum theories have already drawn attention; see Veenker (1967: 97–99, 109–117) for Slavic, especially Russian; Schmidt (1990: 188–189) for Celtic. Various non-Indo-European ad-, sub-, and superstrata have been proposed, for both Slavic and Celtic, which has generated a vast literature, for Slavic studies, specifically Russian, chiefly involving the postulation of a Finno-Ugric, or sometimes Altaic substratum, which may
27 It might be noted that many of Tolkien’s superficially Germanic forms have Slavic echoes, see Orr (2016) and the literature cited therein. 28 Cf. Lockwood’s (1961: 12–14) suggestion that dun ‘duck’ is a borrowing from Old English into Old Low German, an example of Entwistle and Morison’s (1964: 23) “stretching of elastic until breaking point”, also Orr (2003b). 29 Although Martynov cites Latin lūcus as a cognate of lǣs, Old Latin loucos renders that suggestion unlikely.
3 Intertwining trees, eddies, and tentacles
101
have also affected Latvian. The topic of pre- or non-Indo-European substrata in Celtic also seems to be an ongoing discussion, with a literature similar in scope (see Orr 1992). Finnic and Scandinavian contacts, according to many scholars, provided East Slavic with a substratum and a superstratum respectively (see Veenker 1967; Strumiński 1996). Before the contact was broken Scandinavian provided a bridge connecting Northern Russia and the Scottish Highlands, from one periphery of Europe to another, e.g., Old Scandinavian Ragnhild(r), borrowed into Gaelic as Raonaid (perceived as a Gaelic form of “Rachel”) and into East Slavic as Rogъnědь (Orr 1994, 2011). A further partial parallel is provided by the forms Rogъvolodъ (< Рёгнвалд - Ragnvaldr) and Raguilъ/Raguilovičъ, the latter two taken from the Book of Enoch rather than Scandinavia, but all seen as related forms. They recall the name Campbell in the history of the Scottish Highlands: despite its obvious derivation from cam ‘crooked’30 plus beul ‘mouth’; although attempts were made to derive it from Latin De Campo Bello.31 Two peripheries of the Slavic world also support the idea of strong external linguistic influence, with both examples to be cited involving place names showing local irregularities in the iconic developments of Common Slavic *-VRC- sequences, explicable by contact with neighboring non-Slavic languages. In the context of influence on Russian, Holzer (2006: 135) cites the modern place name Narva, referring to both a city and a river (the latter name being primary), a non-pleophonic form maintained under Finnic influence, contrasted with the attested Old Russian form Norova, which it has ultimately superseded. Narva is currently just on the Estonian side of the border, despite having a Russian population of over 90 percent (see Shevelov 1964: 415).32 This recalls another city at the other end of the Slavic world, Varna in Bulgaria, which, again, does not show the Slavic development, in this case metathesis, of *-VRC-sequences either. Shevelov
30 From Common Celtic *kambos > Latin cambiare > French changer; also > Scottish Gaelic caman/camanachd > Faroese Grímur Kamban. 31 In Scotland names from the classical languages are occasionally taken as equivalents to Gaelic names on the basis of a perceived similarity, e.g., Aeneas for Aonghus (Angus), Hector for Eachann. 32 This example could support Vermeer’s (1986) more general theory of the presence of large numbers of Finnic speakers acting as a brake on certain Slavic sound changes, although Holzer omits any reference to Vermeer.
102
Robert Orr
(1964: 406) explains Varna by suggesting that the presence of a large Greek population hindered metathesis until the process had run its course.
2.6 “Linguistic rings” Another form that linguistic influence can take is “linguistic rings”, a concept borrowed from Dawkins’ (2005: 308–320) “ring species”, defined as cases of unbroken gene flow continua, taking the geographical shape of rings, with gradual differentiation as one moves along the ring, in either direction. At one point in such rings, however, there is an abrupt break, the site of a collision between two originally closely related forms, separated for a prolonged period, and divergent as a result, and then from which prehistoric migrations may be traced back to their approximate origins. Dawkins cites two examples, one comparatively local, involving salamanders in the mountains around California’s Central Valley, and one global, involving seagulls all the way round the Northern Hemisphere. The same phenomenon appears to exist in language, both at the dialectal and lexical levels. In such linguistic rings abrupt breaks are easier to find, due to common phenomena such as war, mass migration, and the like, than in biological rings, although the massive human intervention in fragmenting the areas of species evolution currently under way may change that situation. Two examples of such ring continua, albeit broken, may be traced within Slavic, with their respective abrupt breaks on the border between East Slavic and Polish, and the Vidin-Osogov line in the Balkans, both of which can be traced back to early migrations among the Slavs (Ivić 1956: 12–18; Sławski 1962). In the case of Vidin-Osogov (separating Serbian and Bulgarian) we are probably dealing with two different waves of Slavic migration to the Balkans, which started off from different points further north and then later collided further south. This can be clearly seen if we note that Serbian merges gradually with Croatian, the Kajkavian dialects of which in turn merge with Slovenian. Dialectal data from North-West Slovenian and Central Slovak point to a situation where Slovenian would in turn have merged with central Slovak dialects if the Hungarians had not sliced through that part of the Slavic dialect continuum. Further east, Northeast Bulgarian shows similar evidence of a severed link with Southwest Ukrainian. The history is not as clear here, as this time Romanian sits athwart the severed link, cf.:
3 Intertwining trees, eddies, and tentacles
Bulgar ian
manian) (Ro
n) aria ng
Sor b
Ea st
)
ch/Slovak Czen
(Jatvingi an
vic Sla
ian
(a b r up tb
k) rea
Polish
103
(H u
ian/Croatian/Serb ven ian Slo
It might be noted that the break between East Slavic and Polish depicted above is rather the result of their apparent original separation by Jatvingian, an extinct Baltic language, whose speakers shifted to East Slavic in the East and Polish in the West. As Jatvingian was squeezed, leading to East and West Slavic (re)establishing contact, their mutual influence and interplay (re)commenced, similar to the situation along the Vidin-Osogov line farther south, albeit against the background of a bundle of isoglosses (Orr 2008: 194-195).33 The same phenomenon appears to exist in the paths taken by individual loanwords along paths involving more than one language. These paths can become fairly complex, with the ultimate source often at several removes, and sometimes the borrowing comes full circle, forming what may be dubbed lexical rings, e.g., English garda ‘Irish policeman’ < Irish garda < French garder/garde < Frankish ward (English ward); Hiberno-English gossoon ‘boy’ < Irish garsún < French garçon (gars) < West Germanic/Medieval Latin *warkjone < Frankish *wrakjo ‘mercenary soldier’; cf. Old English wræcca ‘wretch’.34 All these “rings” would have
33 Holzer (2006: 137–138) suggests that there were at least two separate expansions of Slavic, the first one covering the area between the Gulf of Finland and the Lower Danube, followed by a later expansion to the South and West, with the First Palatalization taking place between these expansions. Such processes would have been very conducive to the development of some minor “rings”. 34 A glance at the semantics of the wider “etymological nest” represented by gossoon/ garsún/garçon (gars)/*warkjone/*wrakjo/wræcca offers a nice illustration of the often mind-boggling complexity involved in distinguishing loanwords from native forms, the disentangling of which is essential in the type of study being pursued here. On the Germanic level the above “nest” includes forms such as, e.g., Gothic wrikan ‘persecute’,
104
Robert Orr
taken time to emerge. In both cases the original donor language had ceased to exist as such35 by the time the given loanword had come full circle, long after *ward and *warkjone had been borrowed into “French” as the first link in the chain. Holzer (2006: 128, fns. 129 and 131) lists a set of East Slavic place names clearly from Finnic and Baltic originals which illustrate the complexities of the chronological processes involved, pointing out that in the light of the possible number of lost Baltic and Finnic languages/dialects it cannot be stated with certainty whether, for example, the First Palatalization was shared with parts of Baltic and Finnic or not, in contrast to Shevelov’s (1964: 250, 1971: 309) dismissal of such possibilities, going on to point out that in some cases Slavic and Baltic palatalizations even affect individual Finnic and Baltic languages.36
2.7 The case of Hungarian Ostler (2005: 261–262, 309–314) points to two areas in Europe where the details of the process of language replacement, on whatever level, remain a mystery: the Slavicization of the Balkans and the Germanicization of England. Another linguistic mystery is provided by the current geographical location of Hungarian (honfoglalás ‘conquest’), far from its closest relatives close to the Urals, which has always been slightly anomalous in the linguistic history of Europe. In the course of a treatment of migrations Nichols (2011: 187–188) offers a brief discussion of the origins of Hungarian, mainly in the form of questions, suggesting that at least one ten-man all-Hungarian-speaking squadron under
and on the Indo-European level it includes Latin urgeo ‘chase’, and a host of Slavic and Baltic nominal and verbal forms: Ukrainian voroh, ‘enemy’, Slovenian vrag ‘devil’; Old Church Slavonic vrěšti ‘throw away; expel’, Russian izverg ‘monster’; Lithuanian, vargti ‘suffer’; Old Prussian wargan ‘evil’. Also clearly related to the above “nest” is the Old Germanic form *wargaz, popularized by J.R.R. Tolkien (Old Icelandic vargr ‘outlaw, criminal, wolf’), very probably an original borrowing from Slavic into Germanic (Orr 2003b, 2016: 115) and the literature cited therein. 35 “Frankish” is customarily used as a surrogate to refer to the speakers of the West Germanic dialect who ultimately gave their name to France. Their dialect would certainly have been mutually intelligible with the West Germanic dialects which ultimately gave rise to English at the time the borrowings would have taken place, possibly even part of the same dialect continuum. 36 This example can also illustrate Holzer’s (2006: 129) reminder that from the point of view of historical records South and West Slavic have preserved relevant material far better than North Slavic.
3 Intertwining trees, eddies, and tentacles
105
a charismatic leader in a mostly Turkic-speaking37 army might have acted as a nucleus for the later linguistic expansion of Hungarian into a slightly larger area covering what is now Hungary and parts of Romania, Slovakia, and the former Yugoslavia.38
2.8 Tentacles reaching around peripheries At first sight, the use of “tentacle” as a linguistic metaphor might seem a little odd. Nevertheless, there are some linguistic and cultural phenomena for which “tentacle” appears to be the best metaphor.39 Some elements of the Scandinavian contacts with both Celtic and Slavic as discussed above might be viewed in this context. One very interesting example, albeit non-linguistic, is provided by the possibility of a Sarmatian origin for the legend of King Arthur’s sword; see Cochran and Harpending (2010: 147–148) who trace its presence in British40 folklore to it being told by about 5,000 Sarmatians brought into Britain by Marcus Aurelius.
3 Conclusion Dixon (1997: 364–365) suggests that the linguistic situation in Europe at some stage during the Neolithic period might have resembled that currently found in New Guinea or the Caucasus. In several well-argued papers, Vennemann (1998: 1–4) proposes an interesting counterargument: the large-scale settlement of Europe after the end of the last Ice Age would have been comparatively recent (ca. 10,000 BCE, possibly more recent than the peopling of much of the Americas). It might also be noted that this settlement would have been fairly gradual, following the retreat of the glaciers, and might even have matched similar expansions by Eskimo and Athabaskan. What this entails is that, however complex the interplay of various languages may have been throughout European history, such complexity may not have remained unchanged over
37 There may also have been Mongolian and Iranian speakers, see also fn. 26, certainly far more than Uralic speakers. 38 Ostler (2005: 309 fn.) devotes only one line to the problem of the Hungarian honfoglalás ‘conquest’. 39 The presence of tentacles in modern culture may be attributed to Howard Phillips Lovecroft. 40 In this case English, taken over and adapted from earlier Brythonic material.
106
Robert Orr
time, and future research may well confirm that post-Ice Age Europe was substantially less linguistically diverse, typologically, than many later periods, even after the spread of Indo-European and Uralic. In turn, this may ease the task of studying prehistoric language contact in Europe: this article cannot claim to be more than one indicator out of many in this direction.
References Alter, Stephen G. 1999. Darwinism and the Linguistic Image. Baltimore/London: Johns Hopkins University Press Anttila, Raimo. 2000. Greek and Indo-European Etymology in Action: Proto-Indo-European *aǵ-. Amsterdam/Philadelphia: John Benjamins. Baudouin de Courtenay, Jan Niecisław. 1984. O jȩzyku polskim. Warsaw: Państwowe Wydawnictwo Naukowe. Boček, Vít. 2014. Praslovanština a jazykový kontakt [Proto-Slavic and Language Contact]. Prague: Nakladatelství Lidové noviny. Brockelmann, Carl. 1985. Arabische Grammatik, 22nd ed. Leipzig: Verlag Enzyklopädie. Calder, George. 1923. A Gaelic Grammar. Glasgow: Gairm. Clancy, Steven J. 2010. The Chain of Being and Having in Slavic. (Studies in Language Companion Series 122.) Amsterdam/Philadelphia: John Benjamins. Cochran, Gregory and Henry Harpending. 2010. The 10,000-Year Explosion: How Civilization Accelerated Human Evolution. New York: Basic Books. Comrie, Bernard. 1981. Languages of the Soviet Union. Cambridge: Cambridge University Press. Cyxun, Gennadij Afanas’evič [Henadz’ Apanasavič Cyxun]. 2012. Ekskljuzivnyja mižslavjanskija izaleksy i praslavjanskaja areal’naja struktura [Exclusive inter-Slavic isolexes and the Common Slavic areal structure], In: Metka Furlan and Alenka Šivic-Dular (eds.), Praslovanska dialektizacija v luči etimoloških raziskav, 61–66. Ljubljana: SAZU. Cyxun, Henadz’ Apanasavič. 2013. Aspekty slavjanskaj arėal’naj linhvistyki. Daklad na XV Mižnarodnym z’ezdze slavistaŭ [Aspects of Slavic Areal Linguistics. A Report Delivered at the 15th International Congress of Slavists]. Minsk: Prava i ėkanomika. Dahl, Östen. 1990. Standard Average European as an exotic language. In: Johannes Bechert, Giuliano Bernini, and Claude Buridant (eds.), Toward a Typology of European Languages, 3–8. The Hague: Mouton de Gruyter. Dawkins, Richard. 1988. The Blind Watchmaker. Harmondsworth: Penguin. Dawkins, Richard. 2005. The Ancestor’s Tale: A Pilgrimage to the Dawn of Life. London: Phoenix Paperbacks. Décsy, Gyula. 2000. The Linguistic Identity of Europe, Parts 1–2. Bloomington, IN: Eurolingua. Dixon, Robert M. W. 1997. The Rise and Fall of Languages. Cambridge: Cambridge University Press. Drinka, Bridget. 2017. Language Contact in Europe: The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Drummond, Peter. 1991. Scottish Hill and Mountain Names. Glenrothes: Scottish Mountaineering Trust.
3 Intertwining trees, eddies, and tentacles
107
Entwistle, William J. and Walter A. Morison. 1964. Russian and the Slavonic Languages. London: Faber & Faber Gould, Stephen J. 1991. Wonderful Life: The Burgess Shale and the Nature of History. Harmondsworth: Penguin. Grković-Major, Jasmina. 2010. The development of predicative possession in Slavic languages. In: Motoki Nomachi (ed.), The Grammar of Possessivity in South Slavic Languages: Synchronic and Diachronic Perspectives (Slavic Eurasian Studies 24.), 35–54. Sapporo: Slavic Research Center, Hokkaido University. Greene, David. 1979. Perfects and perfectives in modern Irish, Eriu 30: 122–141. Gvozdanović, Jadranka. 2008. Centum elements in Slavic revisited. In: Sebastian Kempgen et al. (eds.), Deutsche Beiträge zum 14. Internationalen Slavistenkongress, Ohrid 2008 (Die Welt der Slaven. Sammelbände. Sborniki 32.), 159–169. Munich: Otto Sagner. Gvozdanović, Jadranka. 2009. Celtic and Slavic and the Great Migrations. Heidelberg: Carl Winter. Haspelmath, Martin. 2001. Non-canonical marking of core arguments in European languages. In: Alexandra Y. Aikhenvald, Robert M.W. Dixon, and Masayuki Onishi (eds.), Non-canonical Marking of Subjects and Objects (Typological Studies in Language 362.), 53–83. Amsterdam/Philadelphia: John Benjamins. Haugen, Einar. 1976. The Scandinavian Languages. Cambridge, MA: Harvard University Press. Holzer, Georg. 1989. Entlehnungen aus einer bisher unbekannten indogermanischen Sprache im Urslavischen und Urbaltischen. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Holzer, Georg. 2006. Methodologische Überlegungen zur Auswertung der slavisch-baltischen und slavisch-finnischen Lehnbeziehungen für die slavische Siedlungs- und Lautgeschichte. In: Juhani Nuorluoto (ed.), The Slavicization of the Russian North: Mechanisms and Chronology. 128–139. Helsinki: Helsinki University Press. Isačenko, Aleksander. 1974. On ‘Have’ and ‘Be’ languages: A typological sketch. In: Michael S. Flier (ed.), Slavic Forum, 43–77. The Hague: Mouton. Ivić, Pavle. 1956. Dialektologija srpskohrvackog jezika [The Dialectology of the Serbo-Croatian Language]. Novi Sad: Matica Srpska. Kiparsky, Valentin. 1975. Russische Historische Grammatik, Vol. 3: Entwicklung des Wortschatzes. Heidelberg: Carl Winter. Kurkina, Ljubov’. 2012. K probleme interpretacii leksičeskix izogloss [On the problem of the interpretation of lexical isoglosses]. In: Metka Furlan and Alenka Šivic-Dular (eds.), Praslovanska dialektizacja v luči etimoloških raziskav, 135–150. Ljubljana: SAZU. Kurzová, Helena. 1996. Mitteleuropa als Sprachareal. Germanistica Pragensia 13: 57–73 (Acta Universitatis Carolinae. Philologica 5.). Lewis, Henry and Holger Pedersen. 1961. A Concise Comparative Celtic Grammar. Göttingen: Vandenhoeck & Ruprecht. Lockwood, William Burley. 1961. The Faroese Bird Names. Copenhagen: Munksgaard. Lockwood, William Burley. 1977a. Some Traces of Gaelic in Faroese”. Fróðskaparrit 25: 9–25. Lockwood, William Burley. 1977b. An Introduction to Modern Faroese. Tórshavn: Føroya Skúlabókagrunnur. Marcantonio, Angela. 2002. The Uralic Language Family: Facts, Myths, and Statistics. Oxford: Blackwell. Martynaŭ, Viktar Uladzimiravič. 1998. Praradzima slavjan. Linhvistyčnaja veryfikacyja. [The Homeland of the Slavs: a Linguistic Verification] Minsk: Belaruski kamitėt slavistaŭ.
108
Robert Orr
Martynov, Viktor Vladimirovič [Viktar Uladzimiravič Martynaŭ]. 1996. Slavjanizmy v drevneanglijskom [Slavic Forms in Old English]. In: Adam Jevgen’evič Suprun [Adam Iaŭhenavič Suprun] and Xel’mut Jaxnov [Helmut Jachnow] (eds.), Slavjano-germanskie jazykovye paralleli. Slawisch-germanische Sprachparallelen. Minsk: Belorusskij gosudarstvennyj universitet. Marvan, Jiří. 2013. Belarusian within the Slavic-Baltic Euro-Zone: morpho-lexical processes. Slavia 82: 175–188. McWhorter, John. 2001. The Power of Babel. London: Heinemann. Mirčev, Kirill. 1971. Predlog u v possessivnoj funkcii v istorii bolgarskogo jazyka [The preposition u in the possessive function in the history of Bulgarian]. Issledovanija po slavjanskomu jazykoznaniju. 74–84. Moscow: Nauka. Morzinski, Mary. 1994. Linguistic Influence of Polish on Joseph Conrad’s Style. Boulder, CO: East European Monographs. Nichols, Johanna. 1987. Russian vurdalak ‘werewolf’ and its cognates. In: Michael S. Flier and Simon Karlinsky (eds.), Languages, Literature, Linguistics. In Honor of Francis J. Whitfield, 165–177. Berkeley: Berkeley Slavic Specialties. Nichols, Johanna. 2011. Forerunners to globalization: the Eurasian steppe and its periphery. In: Cornelius Hasselblatt, Peter Houtzagers, and Remco van Pareren (eds.), Language Contact in Times of Globalization, 177–195. Amsterdam: Rodopi. Nuorluoto, Juhani (ed.). 2006. The Slavicization of the Russian North: Mechanisms and Chronology, 128–139. Helsinki: Helsinki University Press. Orr, Robert A. 1989. A Russo-Goidelic Syntactic Parallel: U nego svoja izba postavlena/Tá sé déanta agam, General Linguistics 29: 1–21. Orr, Robert A. 1992. Slavo-Celtica. Canadian Slavonic Papers 34: 245–268. Orr, Robert A. 1993. More on borrowings: comments on Manaster Ramer. Diachronica 10: 301–306. Orr, Robert A. 1997. Review of: Janda, Laura A. Back from the Brink. LINCOM Europa. Journal of Slavic Linguistics 5: 150–163. Orr, Robert A. 1994/95. A Slavic-Baltic-Finnic-Hebridean Sprachbund? SLAVONICA: A New Series of Scottish Slavonic Review 22: 53–64. Orr, Robert A. 1999. Evolutionary biology and historical linguistics. Diachronica 16: 123–157. Orr, Robert A. 2002. Demythologising Celtic – Celtic as non-exotic compared with other linguistic systems. Journal of Celtic Language Learning 7: 5–44. Orr, Robert A. 2003a. Eddies in language and biology. In: Douglas Coleman, William J. Sullivan, and Arle Lommel (eds.), LACUS Forum XXIX: Linguistics and the Real World, 271–281. Houston: LACUS. Orr, Robert A. 2003b. Murk – a neglected Slavic loanword in Germanic? Canadian Slavonic Papers 16: 47–60. Orr, Robert A. 2005. Family trees in historical linguistics and evolutionary biology. In: Adam Makkai, William J. Sullivan, and Arle Lommel (eds.), LACUS Forum XXXI: Interconnections, 277–288. Houston, TX: LACUS. Orr, Robert A. 2008. Dawkins’ the ancestor’s tale: A linguist’s view. In: Patricia Sutcliffe, Lois Stanford, and Arle Lommel (eds.), LACUS Forum XXXIV: Speech and Beyond, 185–198. Houston, TX: LACUS. Orr, Robert A. 2009. Review of: Décsy, Gyula. The Linguistic Identity of Europe. Word 55: 456–464.
3 Intertwining trees, eddies, and tentacles
109
Orr, Robert A. 2012. Review of: Nuorluoto, Juhani (ed.), The Slavicization of the Russian North: Mechanisms and Chronology (Slavica Helsingiensia 27.). Helsinki: Helsinki University Press. 2006, Journal of Slavic Linguistics 20: 121–143. Orr, Robert A. 2012. An irregular Indo-European sound change reflected in Slavic? In: Andrii Danylenko and Serhii Vakulenko (eds.), Studien zu Sprache, Literatur, und Kultur bei den Slaven: Gedenkschrift für George Y. Shevelov aus Anlaß seines 100 Geburtstages und 10 Todestages: 7–17. Munich: Otto Sagner. Orr, Robert A. 2016. Slavic morphology in Tolkien? Sméagol and Ungol. In: Bohumil Vykypěl and Vít Boček (eds.), Perspectives of Slavonic Etymology: 113–120. Prague: Nakladatelství Lidové noviny. Ostler, Nicholas. 2005. Empires of the Word: A Language History of the World. New York: Harper Collins. Meillet, Antoine. 1924. Le slave commun. Paris: Édouard Champion. Rouse, Irvine. 1992. The Taínos: Rise and Decline of the People who Greeted Columbus. NewHaven/London: Yale University Press. Schmidt, Karl Heinz. 1990. The postulated pre-Indo-European substrates in insular Celtic and Tocharian. In: Markey L. Thomas and John A. C. Greppin (eds.), When Worlds Collide? The Indo-Europeans and pre-Indo-Europeans, 179–202. Ann Arbor: Karoma. Shevelov, George Y. 1964. A Prehistory of Slavic. The Historical Phonology of Common Slavic. Heidelberg: Carl Winter. Shevelov, George Y. 1971. Issues and non-Issues. Five years after A Prehistory of Slavic. In: George Y. Shevelov, Teasers and Appeasers: Essays and Studies on Themes of Slavic Philology (Forum Slavicum 32.), 297–325. Munich: Wilhelm Fink. Shevelov George Y. 1979. A Historical Phonology of the Ukrainian Language. Heidelberg: Carl Winter. Sławski, Franciszek. 1962. Zarys dialektologii południowosłowiańskiej z wyborem tekstów gwarowych [An Outline of the South Slavic Dialectology with a Selection of Dialect Texts]. Warsaw: Państwowe Wydawnictwo Naukowe Strang, Barbara. 1970. A History of English. London: Methuen. Strumiński, Bohdan. 1996. Linguistic Interrelations in Early Rus. Northmen, Finns, and East Slavs (Ninth to Eleventh Centuries) (Colana di filologia e letterature slave 2.). Rome: La Fenice Edizioni; Edmonton/Toronto: Canadian Institute of Ukrainian Studies Press. Thrainsson, Hoskuldur, Hjalmar Petersen, Jógvan í Lon Jacobsen, and Zakaris Svabo Hansen 2004 Faroese: An Overview and Reference Grammar. Torshavn: Føroya Fróðskaparfelag. Thurneysen, Rudolf. 1980. A Grammar of Old Irish. Dublin: Dublin Institute for Advanced Studies. Timberlake, Alan. 1976. Subject properties in the North Russian passive. In: Charles Li (ed.), Subject and Topic, 545–570. New York: Academic Press. Townend, Matthew. 2002. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English (Studies in the Early Middle Ages 6.) Turnhout: Brepols. Trubačev, Oleg Nikolajevič. 1965. Iz slavjano-iranskix leksičeskix otnošenij [From the SlavicIranian lexical relationships]. Etimologija: 3–81. Veenker, Wolfgang. 1967. Die Frage des finnougrischen Substrats in der russischen Sprache, Bloomington, IN: Indiana University Press. Vennemann, Theo. 1998. Andromeda and the Apples of the Hesperides. In: Karlene Jones-Bley, Angela Della Volpe, Miriam Robbins Dexter, and Martin Huld (eds.), Proceedings of the
110
Robert Orr
Ninth Annual UCLA Indo-European Conference, 1–68. Washington, D.C.: Institute for the Study of Man. Vennemann, Theo. 1999. Testing the West: Hesperia, Euskal, Herria, Europe, Abendland and supporting etymologies. In: Neile A. Kirk and Paul Sidwell (eds.), From Neanderthal to Easter Island, 85–102. Melbourne: Association for the History of Language. Vermeer, Willem. 1986. The rise of the North Russian dialect of Common Slavic. In: Dutch Studies in Russian Linguistics (Studies in Slavic and General Linguistics 8), 503–515. Amsterdam: Rodopi. Whorf, Benjamin Lee. 1941. The relation of habitual thought and behavior to language. In: Leslie Spier (ed.), Language, Culture, and Personality, Essays in Memory of Edward Sapir, 75–93. Menasha, WI: Sapir Memorial Fund. Whorf, Benjamin Lee. 1956. The relation of habitual thought and behavior to language. In: John B. Carroll (ed.), Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf, 134–159. Cambridge, MA: MIT Press. Wolpoff, Milford and Caspari, Rachel. 1997. Race and Human Evolution. New York: Simon & Schuster.
Part II: Slavic and Standard Average European
Jadranka Gvozdanović
4 Standard Average European revisited in the light of Slavic evidence 1 Introduction “Standard Average European” (henceforth, SAE) refers to the central and western European language-contact area, with room for some discussion about marginal languages and how far it extends to the East and the West. Generally, non-Indo-European languages such as Basque do not belong to it, and adherence of some other, partly marginal, language groups has been disputed (Whorf 1941, 1956: 138; Haspelmath 1998, 2001). According to Haspelmath (2001: 1492) the criteria for demonstrating that a structural feature is a Europeanism are: “(i) that the great majority of core European languages possesses it; (ii) that the geographically adjacent languages lack it (i.e., Celtic in the West, Turkic, eastern Uralic, AbkhazAdyghean and Nakh-Daghestanian in the East, and perhaps Afro-Asiatic in the south); (iii) that the eastern Indo-European languages lack it (Armenian, Iranian, Indic); and (iv) that this feature is not found in the majority of the world’s languages”.
Concerning the timeline for SAE, Haspelmath proposes membership assignments such as Celtic being outside SAE but Italic inside it, with the consequence that SAE must have arisen later than the presumed Italo-Celtic prehistoric unity. In light of SAE features not being common to all IndoEuropean languages, SAE must, in his view, have arisen after the dissolution of the Proto-Indo-European unity during the great migrations (Haspelmath 2001: 1507). According to this approach, SAE is a result of congruent developments of neighboring languages, all of which have common Indo-European roots. SAE has been assumed by Haspelmath (2001) and others to be based on the following main features: (a) presence of definite and indefinite articles, (b) relative clauses with a relative pronoun, (c) ‘have’-perfects, (d) nominative experiencers, (e) participial passives, (f) so-called “anti-causative prominence” (i.e., an intransitive counterpart to a transitive verb, which expresses an event affecting its subject, while giving no semantic or syntactic indication of the cause of the event), (g) dative external possessors, (h) negative pronouns without verbal negation, (i) particles in comparative constructions, (j) subject-person affixes as strict agreement markers, (k) relative-based equative constructions, and (l) https://doi.org/10.1515/9783110639223-005
114
Jadranka Gvozdanović
intensifier-reflexive differentiation. Based on their number of shared SAE properties, languages can be classified either as core or periphery members. The surface-oriented formulation of SAE membership criteria has two consequences: (1) restricted ability to identify structurally similar properties with different morphosyntactic manifestations (such as definiteness expressed by an article vs. adjectival definiteness, as is characteristic of historic Slavic), and (2) restricted ability to group the relevant features and thereby recognize typological patterns. The main areas of SAE in need of further development are: (1) the proper procedure for establishing the essential features and the borders of SAE, (2) the principled evaluation of these features, and (3) the historical IndoEuropean dimension and possible paths of development. Haspelmath and most other researchers of SAE (see, however, Heine and Kuteva 2006) assumed that the common properties of SAE arose between Late Antiquity and the Early Middle Ages, particularly since the great migrations of the first millennium AD, which brought the main bulk of the European branch of Indo-Europeans together in a growing cultural continuum. But do these properties truly have their origin in this period? Heine and Kuteva (2006) offered a more detailed and diversified picture of the core phenomena considered to pertain to SAE by regarding them not in terms of product but of process, characterized by intermediate stages that yield the observable patterns of grammatical variation. Their study shows that congruence phenomena among the European languages developed in processes that occurred in phases and lasted over longer periods of time (in several instances until the modern times). The present chapter looks into the core phenomena of SAE in the Slavic languages showing that their development was slow, depended on the existing structural properties, and remained incomplete in several instances. Grammatical change involves a critical leap that is qualitative in nature, but what we find in several instances in Slavic are adaptive changes that do not really modify the categories. This chapter discusses the data and proposes a discovery procedure for discerning the status of ongoing changes. This chapter discusses critically the properties of SAE in Slavic and their historical context. Section 2.1 gives a critical survey of processes assumed to have yielded articles in Slavic, showing that most of them did not yield a grammatical change. Section 2.2 discusses relative clauses with relative pronouns and shows that this phenomenon is older than SAE. Section 2.3 analyzes the development of so-called possessive perfect in the Slavic area and proposes a subdivision of stages at variance with Heine and Kuteva (2006). Section 2.4 surveys the remaining SAE features. Conclusions about the emergence of SAE and the applied methodology are presented in Section 3.
4 Standard Average European revisited in the light of Slavic evidence
115
2 The defining grammatical properties of SAE in the light of Slavic evidence 2.1 The article The article is a relatively specific category: more languages in the world lack it than have it (see Dryer 2013). Of the languages that have the article, most have the definite article, whereas both definite and indefinite articles, or only the indefinite article occur much less frequently. Dryer (2013) views the definite article as a morpheme which accompanies nouns and which codes definiteness and specificity. However, not only do languages such as German show that specificity is not coded by means of the definite article but the indefinite article instead (see von Heusinger 2012), there are also many examples showing that the definite article can accompany different parts of noun phrases (particularly attributes). In addition, the definition should be further elaborated as follows. Information that is actually or virtually present and accessible in both the speaker’s and the addressee’s frame of reference may be termed definite; information that is present in the speaker’s frame of reference but only accessible (without being previously present) to the hearer may be termed specific; and information in neither of these frames of reference is indefinite and unspecific. Definiteness is as a rule expressed by means of a morpheme (the definite article, a suffix, etc.), but specificity is a broader category with a variety of expression means, including word order. Absence of specificity implies absence of definiteness; most languages do not mark indefiniteness or lack of specificity. Definiteness denotes that the referent has been deictically situated in the actual situation; therefore, markers of definiteness usually develop from demonstrative pronouns. Specificity denotes that the referent has been deictically situated independently of the actual situation (in language which distinguish definiteness and specificity, a specific referent can be deictically situated outside of the actual situation). Indefiniteness, on the other hand, denotes that a referent has been individuated, but not (yet) deictically situated; lack of specificity lacks any information about the referent(s). Demonstratives conceptualize accessibility of referents in the deictic setting of (actual, preceding or sometimes following) discourse. Accessibility can be situational or knowledge-based. By denoting accessibility, demonstratives essentially resemble definite articles, but they always refer to the deictic setting, whereas the definite article can be used also independently of the deictic setting, for knowledge-based accessibility. In other words, accessibility can be
116
Jadranka Gvozdanović
situational or knowledge-based; specificity equals (the speaker’s) knowledgebased accessibility, and definiteness equals both. Proto-Indo-European did not have articles, but this does not mean that it did not have definiteness and/or specificity. There was presumably an incipient category of combination with a demonstrative/relative pronoun that provided the basis for article development in Central Europe in the Middle Ages, but Greek, for example, had this development earlier due to its inhabitation history (and possible early language contacts).1 For the expression of nominal determination in Slavic, various strategies of combination with demonstrative/relative pronouns could be observed in different periods: first adjectival definiteness comparable to Germanic, and in recent centuries increased use of determiners in contact with the surrounding article languages. As a Balkan areal phenomenon, Bulgarian and Macedonian developed their demonstratives to increasingly acquire article functionality, particularly for the “neutral” t-form, which occupies an intermediate position between (deictic-center oriented) proximal and distal denotations. As illustrated by the following examples excerpted from Mayer (1988: 102, 104), Bulgarian has made the decisive leap and become an article language indeed. (1)
Bulgarian a. Otivam na učilište. ≠ Otivam na učilišteto. go.PRS.1SG on/to school.DEF go.PRS.1SG on/to school ‘I am going to school.’ ≠ ‘I am going to the school (building)’. b. Zemjata e planeta. earth.DEF be.PRS.3SG planet ‘Earth is a planet.’ c. Zemjata e planetata, na kojato živeem nie. earth.DEF is planet.DEF on which.DEF live.PRS.1PL we ‘The Earth is the planet on which we live.’
These examples show that accessible entities are marked by the definite article in Bulgarian as in the second part of (1a) and in (1c), whereas newly introduced entities as in (1b) and abstract entities as in the first part of (1a) lack any article.
1 The existence of a definite article is well known for Classical Greek, but not so for Classical Latin. The definite article emerges in Germanic, Romance and Insular Celtic in the course of the Middle Ages (in Old High German and in Old Irish before the end of the first millennium AD, in Old French somewhat later; see Irslinger 2015), and the indefinite article emerges later or not at all.
4 Standard Average European revisited in the light of Slavic evidence
117
Macedonian differs from Bulgarian by using proximal demonstrative endings (-ov/-va/-vo) and distal endings (-on/-na/-no) in addition to neutral endings (-ot/-ta/-to), still in their original deictic meanings, although there is incipient article-like use of the latter, see the following examples (Topolinjska 2009: 179, 187) (2)
Macedonian a. Srcevo me boli koga go gledam. heart.PROX.DEM me ache.PRS.3SG when he.ACC look.PRS.1SG ‘This (my) heart is aching as I look at him.’ b. Narodecon ne misli taka. people.DIS.DEM not think.PRS.3SG thus ‘Those people do not think like that.’ c. Pette novodojdeni studenti ni pristapija. five.DEF newly.arrived students we.DAT come.PRS.3PL ‘The five newly-arrived students come to us.’ d. Go vidov čovekot. he.ACC see.AOR.1SG man.ACC.SG.M.DEF ‘I saw (him) the man.’
Examples (2a) and (2b) show that the original distinction between demonstrative proximity and distance is still preserved in Macedonian. Examples (2c) and (2d) illustrate the development of the deictically neutral t-demonstrative to acquire the functionality of a definiteness marker in contexts in which other markers—initial word order in (2c) and object doubling in (2d)—already point to specificity. We can conclude that the t-demonstrative of Macedonian is an incipient article. The other Slavic languages which developed increased uses of demonstrative pronouns did so in contact situations as well. In central Europe, the main source language was German and its influence particularly strong on Czech. In Czech, the demonstrative pronoun ten/ta/to is used to disambiguate anaphoric, cataphoric or deictic reference; it is also used in correlative constructions, with superlatives and in nominalizations (Berger 1993: 359, 179). (3) Czech a. Podivej se vidíš toho pána v wonder.IMPER REFL see.PRS.2SG that.ACC.SG.M man.ACC.SG.M in cylindru? cylinder.LOC.SG.M ‘Look here, do you see this man with (a) cylinder?’
118
Jadranka Gvozdanović
b. Prosím vás, kde je tady ředitel. ask.PRS.1SG you.ACC where is here director.NOM.SG.M (*ten ředitel)? (*that director) ‘Please, where here is (the) director (*this director)?’ d. Lidé musí je hlídat, a People.NOM.PL.M must.PRS.3PL they.ACC.PL watch.INF and tím hlídáním se to this.INSTR.SG.N watching.INSTR.SG.N REFL this zdraží. become.expensive.PRS.3SG ‘People must watch them, and due to this watching it becomes more expensive.’ Example (3a) illustrates the demonstrative use of the demonstrative pronoun, which is not possible, if the denoted entity is deictically absent as in (3b); example (3c) illustrates the use of the demonstrative for anaphoric co-reference, tím hlídaním ‘(due to) this watching’. Czech has not reached the tipping point of transition from demonstrative to definite article because demonstrative uses are still anaphorically (or possibly cataphorically) or deictically bound. Polish follows the same pattern and has also not reached this tipping point; demonstrative uses remain deictically or anaphorically (or cataphorically) bound, as shown by the following examples from Topolińska (1981: 48). (4) Polish a. Wiesz, nareszcie wyburzono stary blow.up.N.SG.PP old.ACC.SG.M know.PRS.2SG finally dworzec. Ten budynek/*budynek railway station.ACC.SG.M this building.NOM.SG.M/*building strasznie oszpecał centrum miasta. terribly disfigure.PST.M.SG center.ACC.SG.N town.GEN.SG.N ‘You know, finally, the old railway station was blown up. That building used to disfigure the city center terribly.’ b. Do tramwaju wsiadła jakaś onto tram.GEN.SG.M get.on.PST.SG.F certain.NOM.SG.F
4 Standard Average European revisited in the light of Slavic evidence
119
młoda para studencka. (*Ta) young.NOM.SG.F couple.NOM.SG.F student.REL.NOM.SG.F. (*That) dzwiewczyna rozejrzała się, przeszła od razu girl.NOM.SG.F look.around.PST.SG.F REFL go.over.PST.SG.F immediately na przednią platformę, i (*ten) chłopiec on front.ACC.SG.F platform.ACC.SG.F and (*that) guy.NOM.SG.M zatrzymał się przy konduktorze. stop.PST.SG.M REFL by conductor.LOC.SG.M ‘A young student couple got on a tram. The young woman looked around and went immediately to the front platform, and the young man stopped by the conductor.’ As is shown in (4a), anaphoric demonstratives are required whenever ambiguity must be avoided. Whenever there is no ambiguity, as in (4b), anaphoric demonstratives are not used. Again, paralleling Czech, the Polish demonstrative ten/ta/to is still a demonstrative, not a definite article. Slavic minority languages in long-lasting contact with Germanic or Romance article languages have developed slightly different patterns. This group is made up particularly of Upper and Lower Sorbian, in contact with German, and Resian Slovene and Molise Slavic, in contact with Italian. Standard Upper Sorbian uses the demonstrative tón for deictic and anaphoric reference, with nominalized adjectives and superlatives, and in contrastive instances illustrated by (5). In all these instances, the demonstrative denotes situation-specific accessibility, as illustrated by the following example (Faßke 1980: 568). (5) Standard Upper Sorbian Wječor porjedźeše wón piwo tym Evening.ACC.SG.M offer.PRT.3SG he.NOM beer.ACC.SG.N that.DAT.PL.M buram a tym chěžkarjam nutřka w peasant.DAT.PL.M and that.DAT.PL.M local.DAT.PL.M inside in korčmarskej stwě a tym čeladnym room.REL.LOC.SG.F inn.LOC.SG.F and that.DAT.PL.M servant.DAT.PL.M wonka w wobchodźe. outside in store.LOC.SG.M ‘In the evening, he offered beer to the peasants and the locals inside the inn and to the servants outside in the store.’
120
Jadranka Gvozdanović
The demonstrative tón in contemporary Colloquial Upper Sorbian has been described by Scholze (2007) as an article regularly used in pragmatic functions but not in semantic functions. In terms of the analysis proposed here, this description points to essentially the same uses as in Czech and Polish in which use remains within the realm of a demonstrative (see Scholze 2007: 146, 157, 139): (6) Colloquial Upper Sorbian a. Mó smó rjanu stu měli. we.NOM AUX.PRS.1PL nice.ACC.SG.F room.ACC.SG.F have.APP.PL ‘We had a nice room.’ b. Kedźwi na to psyka! watch.IMPR on that.ACC.SG.N dog.ACC.SG.N ‘Pay attention to the dog (i.e., who will be there)!’ c. Also, jow jo jen muž, tón so here AUX.PRS.3SG a man.NOM.SG.M, that.NOM.SG.M šibejži chětrě. come.PRS.3SG hastily ‘So, here is a man, he is coming hastily.’ Examples (6a) and (6b) illustrate the fact that in Colloquial Upper Sorbian the demonstrative pronoun is not used as much as an article, for instance, in German; contrastively used noun phrases denoting a new referent in a situation occur without the demonstrative pronoun as in (6a). Another article language within the SAE area that exerted strong influence on neighboring Slavic is Italian. This can be observed in Resian Slovene, a regional minority language in the Alpine region probably since the great migrations, and in Molise Croatian, a minority that moved to southern Italy half a millennium ago. Resian Slovene uses the demonstrative pronoun deictically, anaphorically, cataphorically, and in nominalizations in an eighteenth century translated Catechism, as illustrated by (7a). Contemporary spoken Resian offers a slightly different picture, illustrated by (7b), (7c), and (8) excerpted from Steenwijk (1992: 80, 126, 189). (7) Resian Slovene a. Te din na that.ACC.SG.M day.ACC.SG.M on je wstal AUX.PRS.3SG stand.up.APP.M.SG
sveti Petak on Pentecost.ACC.SG.M he.NOM od from
4 Standard Average European revisited in the light of Slavic evidence
121
tih martvih. that/the.GEN.PL.M dead.GEN.PL.M ‘On that day on Pentecost he was resurrected from the dead.’ b. Somo̤ narédili din vlí̤ki kɔ́p AUX.PRF.3PL make.APP.PL one/a big.ACC.SG.M cup.ACC.SG.M kafɛ́ toga bílaga. coffee.GEN.SG.N the.GEN.SG.N white.GEN.SG.N ‘We made a big cup of coffee (out) of the white one.’ c. Dín zéc si zdélal, an ma one/a rabbit.NOM.SG.M REFL.DAT do.APP.SG.M as have.PRS.3SG sé̤dan te̤ mládi. seven that.ACC.PL young.GEN.PL ‘A/One rabbit has brought forth (lit. has done so as to have) (these) seven young ones.’ (8) Resian Slovene Alóre trí̤ kópije so paršlɔ́. Jsé trí̤ So three couples.NOM.PL AUX.PRS.3PL come.APP.PL these three kópije ni so ostali wsɔ́ jzdé̤ couples.NOM.PL not AUX.PRF.3PL remain.APP.PL all.NOM.PL here w Réziji. Alóre so bíli wsé AUX.PRF.3PL be.APP.PL all.NOM.PL in Resia.LOC.SG.F so te trí̤ kópije, né̤, ni so mé̤li that.NOM.PL three couples.NOM.PL no not AUX.PRF.3PL have.APP.PL otrokɛ́, to bílu múš nu children.ACC.PL that be.APP.PL man.NOM.SG.M and źanɐ́, niso mé̤li otrokɛ́. wife.NOM.SG.F not.PRT.AUX.PRS.3PL have.APP.PL children.ACC.PL ‘So, three couples arrived. These three couples did not all stay here in Resia. So there they all were, the three couples, no, they had no children, (it was) just husband and wife, they had no children.’ There are differences between the eighteenth century translated Catechism in (7) and the vernacular stories collected by Steenwijk (1992). This Catechism was more recently analyzed by Benacchio (2014), who distinguished three functions of the emerging definite article in Resia (implying three stages of its functional development): (1) the anaphoric function, (2) referent denotation by general information, and (3) generalization. In addition, the emerging indefinite article as dín zéc ‘a/one rabbit’ in (7c) was used for specific, nonspecific and exemplary functions. Benacchio assumed a straightforward line of development between
122
Jadranka Gvozdanović
Resian of the eighteenth century and the contemporary language. However, we should keep in mind that the Catechism as a religious text was prone to following the Italian tradition of religious texts, possibly different from the spoken language.2 Indeed, Runić (2013) showed that modern spoken Resian uses the definite article only preceding an attribute (in most instances an adjective), not preceding a bare noun.3 This is compatible with Steenwijk’s examples of the spoken language. This marking of definiteness attached to the adjective is a remarkable continuation of the historical Slovene (and more generally Slavic) marking of definiteness on the adjective. In this sense, Resian exhibits only an adaptive change in the realm of definiteness after almost one and a half millennium of Romance influence.4 The other Slavic minority island under Italian influence, Molise Croatian (called Molise Slavic in Breu’s publications, see also in this volume) has, according to Breu (2008), developed an indefinite article (from ‘one’) in the absence of a definite article, thus showing that presence of an indefinite article need not imply presence of a definite article.
Molise Croatian Italian
indefinite | indefinite
0 / definite
\ 0
Figure 1: Molise Croatian and Italian articles (Breu 2008: 83).
Colloquial Upper Sorbian German
indefinite | indefinite
tón 0 | / | definite 0
Figure 2: Colloquial Upper Sorbian and German articles (Breu 2008: 85).
2 The first Slovene protestant text, Cerkovna ordninga ‘Church Rules’ by Primož Trubar (1564), shows a strong influence of German sources in the increased use of the demonstrative pronoun (except in prepositional phrases), which almost fully matches the German sources, but Trubar’s commentaries in the margins show in part a different usage (for example, on page 17 the main text repeatedly has ta Greh ‘the sin‘, but the commentary in the parallel line has only Greh ‘sin’). 3 The Catechism did have a few exceptions before bare nouns, but this may have been due to the influence of the article distribution in the Italian source text. 4 The emerging indefinite article from ‘one’ described by Benacchio (2014) does not follow the same restriction to pre-attributive positions, but the functionality of ‘one’ is basically the same as in Croatian, Bosnian and Serbian, where it is used for specificity, i.e., ‘a (certain)’. What we can observe in Resian are more frequent uses, but no change of functionality. In this sense, the categorical leap has not yet occurred.
4 Standard Average European revisited in the light of Slavic evidence
123
Breu (2012) explicitly states that Colloquial Upper Sorbian and Molise Croatian have developed full-fledged indefinite articles in referential, nonreferential and generic functions. He also states that Colloquial Upper Sorbian has developed a grammaticalized definite article, whereas Molise Croatian has not done it because the contacting language, Italian, does not exhibit ambiguity between the demonstrative and the article. Indeed, this may be one factor but certainly not the only one. The additional factor is that Molise Croatian preserved relative freedom of word order, and placement before the finite verb is a means for expressing specificity. Molise Croatian is thus an interesting case because it has preserved word order variation expressing specificity (and definiteness) and apparently added an indefinite article as an overt expression of lack of definiteness. This extended use of ‘one’ developed as an adaptive change (see Andersen 1973) under preservation of the essential features of the original system. This contrasts with the increased usage of the demonstrative taking on the functions of the definite article in Colloquial Upper Sorbian, under strong German influence. The explanation for this disparity can be found in the difference between the two conditioning systems. Italian has relatively free word order, which has had a preserving influence on the originally relatively free word order of Slavic, capable of expressing specificity in positions preceding the finite verb. This capacity of Slavic was strongly restrained under German influence in Sorbian, which developed a predominantly verb-second word order in its spoken Catholic variety (Scholze 2007: 330). This fact was beneficial to the development of articles, but not a sufficient condition for them. It can be generally observed that the indefinite article develops in Slavic bilingual situations. In addition to Resian Slovene, Molise Croatian and Colloquial Upper Sorbian, Balkan Slavic also shows incipient uses of ‘one’ as an indefinite article, but Friedman (2015) convincingly showed that the uses remain restricted to specific-referential contexts (thereby in fact remaining within the realm of the original Slavic uses of ‘one’). Friedman (2015: 130) therefore proposed to use the term indefinite marker instead of indefinite article for a bearer of the grammatical meaning of indefiniteness that is not fully grammaticalized as an indefinite article (i.e., semantically bleached and subject to rules of obligatory occurrence). For example, Macedonian does not use any article in non-referential contexts, e.g., Vikni lekar! ‘Call a doctor!’ in which English or German requires an article. In view of these findings by Friedman we should refer back to the Colloquial Upper Sorbian example (6a) mó smó rjanu stu měli ‘we had a nice room’, in which there is also an article lacking where the article languages would expect it. Also Molise Croatian should be inspected for similar examples in order to establish whether we are dealing with indefinite markers or indefinite articles.
124
Jadranka Gvozdanović
In the north-east area of Slavic, in contact with Finnic and related languages, dialectal Russian has developed postpositive -to, functionality comparable to the increased demonstrative use in other Slavic languages, yet with an important difference: -to can mark any autosemantic constituent, including substantival, adjectival, verbal, adverbial, numeral, and pronominal elements. The marking by this postpositive particle is strictly local (in contradistinction to demonstrative words in the other languages, which mark entire syntagms, and the postpositive marking of Bulgarian and Macedonian, which also has the entire syntagm as its domain). As an example, compare particle stacking in North Russian in (9) taken from Meščerskij (1972: 250) and in accordance with the data in Kuz’mina and Nemčenko (1962). (9) North Russian Nynče-to na bol’šoj-to doroge now-DEM on big.LOC.SG.M-DEM road.LOC.SG.M ručej-to u bajni-to u Pavla-to creek.NOM.SG.M-DEM at bath.GEN.SG.F-DEM at Paul.GEN-DEM doma ne tekë sej god. home.GEN.SG.M not flow.PRS.3SG this.ACC.SG.M year.ACC.SG.M ‘Now on the big road the creek near the bath at Pavel’s home does not flow this year.’ (10) Standard Russian a. No, no razmaxalas’. Uberi ruki-to, hey hey strike.out.PST.F.SG put.away.IMPER.2SG hands.ACC.PL.F-DEM bešenaja! mad.woman.NOM.SG.F ‘Hey, you merely got into striking out. Put your hands away, mad woman!’ b. Xorošo peli u Vozdvižen’ja. Noč’ nicely sing.PST.PL at Exaltation.GEN.SG.N night.NOM.SG.F tixaja. Sjuda donosilo vozduxom. quiet.NOM.SG.F here bring.PST.SG.N air.INSTR.SG.M Peli-to xorošo. Da mne mat’ sing.PST.PL-DEM nicely so I.DAT mother.NOM.SG.F moja, ploxo. my.NOM.SG.F badly ‘They were singing nicely at the Exaltation of the Cross. The night was calm. The sounds were brought here by air. So they were singing nicely. But I, my dear mother, felt badly.’
4 Standard Average European revisited in the light of Slavic evidence
125
In dialectal Russian, as illustrated in (9), there is a -to marking attached to ‘now’, ‘big’, ‘river’, the name ‘Pavel’, i.e., all the elements that are accessible and immediately relevant to the deictic setting of the speech situation; -to is similar to the West Slavic demonstrative ten/tón, except that its usage is not restricted to noun phrases.5 In standard Russian (illustrated by example (10), taken from Pasternak’s novel Doktor Živago), -to is a topicality marker. This example shows that -to is by no means restricted to the north and north-west of Russia, although the higher frequency of -to in the northern areas can be attributed to contacts with Finnic (e.g., Veps) and Permic (e.g., Komi) languages, which, alongside Finnish, have similar particles (see Zajceva 1981; Gvozdanović 1997: 1077). (11) a. Finnic (Veps) Lapš6-se iunoz. child-DEM fall.asleep.PST b. Russian Rebenok-to usnul. child-DEM fall.asleep.PST.M.SG ‘The child fell asleep.’ Leinonen (1998) analyzed northern Russian dialectal texts (in which this postpositive particle appears in declined forms as well) and came to the conclusion that the use of the northern Russian -to is indeed similar to the use of the Finnish demonstrative (and anaphoric) pronoun se; both are used as markers of what Leinonen (1998: 75) calls “prominence”, defined as importance within the discourse and accessibility of the referent to the addressee at a certain point in discourse. According to Leinonen’s (1998: 78–83) analysis, the postpositive particle can mark (a) accessibility of the referent, (b) topic, (c) switching in reported dialogues, (d) contrast, (e) emphasis, (f) right dislocation, (g) setting of time and place, and (h) inalienable possession. This may be illustrated by (12),
5 Historically, postposed instances of the demonstrative were attested in early texts, as shown by the Uspenskij Sbornik (Anthology of the Dormition Cathedral in Moscow): obrete ostrovŭ sredě morja . . . se že i donyně ostrovŭ tŭ zovomŭ est’ boljarovŭ ‘he found an island in the middle of a sea . . . until now this island has been called Boljarov’ (Knjaževskaja et al. 1971: 54: 8). 6 In contrast to Zajceva (1981), Veps dictionaries contain laps’ (cf. http://vepsnoid.blogspot. de/2010/10/veps-english-dictionary-vepsa.html, retrieved on July 30, 2013).
126
Jadranka Gvozdanović
from Leinonen (1998: 81), in which two instances of a narratively accessible road are marked by -to, followed by a third instance of a narratively inaccessible road, in which -to is lacking. (12) North Russian Pora-to ot dorogi-to (1) ni xodim/ da vit’ Time-DEM off road.GEN.SG.F-DEM not walk.PRS.1PL/ so well pamjat’-to była xoroša/ dak memory.NOM.SG.F-DEM be.PST.F.SG good.NOM.SG.F/ so-as primetili/ v kakoj storone doroga-to (2)/ note.PST.PL/ on which.LOC.SG.F side.LOC.SG.F road.NOM.SG.F-DEM/ da vot kak i privykli// da kak so actually also and get.used.PST.PL// so that esli tak i /zabudem/ v kakoj if so and /forget.PERF.NONPST.1PL/ on which.LOC.SG.F storone doroga (3). side.LOC.SG.F road.NOM.SG.F ‘At that time we didn’t walk away from the road (1) / because our memory was good / so we noted which side the road (2) is on / and so we got used to it // and then we forget /on which side there is a road (3).’ It is essential to note that the non-accessible road appears in a new intonation unit, marking a new information unit (in line with Chafe 1994), setting a new frame of reference in which the formerly accessible road became inaccessible. In this sense, accessibility holds within information units and (transitively) pairs of information units. In order to understand how the postpositive particle -to may have developed, it is important to pay attention to both external (contact-induced) processes and the available language-internal means. Concerning language-internal means, we should mention that Old East Slavic had a conjunction to, used between conditional (or temporal) subordinate and main clauses, or, more generally, at the border between prothesis and apodosis. Originally used at the beginning of the apodosis as in example (13), it developed into the final part of the prosthesis in contexts in which the prothesis and apodosis functioned as Topic and Focus. The following examples are from Uspenskij sbornik (Anthology of the Dormition Cathedral in Moscow) (Knjaževskaja et al. 1971: 162b: 23–25 and 39b: 12–14) from the twelfth century (Knjaževskaja et al. 1971).
4 Standard Average European revisited in the light of Slavic evidence
127
(13) Old East Slavic a. Reče že kŭ nima epifanŭ to say.AOR.3SG PRT to they.DUAL.DAT Epifan then věrueta li čadě. believe.PRS.2DUAL Q child.DUAL.NOM/VOC ‘And Epiphanius said to them both: do you believe then, children?’ b. Lěpo bo namŭ jestĭ narekŭšemŭ sę good for we.DAT.PL be.PRS.3SG sworn.DAT.PL REFL čĭrnĭcemŭ to po vĭsja dni kajati sę monk.DAT.PL then for all.ACC.PL day.ACC.PL regret.INF REFL grěxŭ svoixŭ. sin.GEN.PL our.GEN.PL ‘For it is good for us, sworn monks, then to regret our sins forever.’ The analysis of article-like expressions in Slavic points to essential similarity among the Slavic languages, which are in the process of expanding the functional domain of deictic and anaphoric accessibility to accessibility based on general knowledge and correspondingly enhancing the employment of the tdemonstrative. However, only Bulgarian has clearly undergone the “quantum leap” of creating the new category of “definite article”. Macedonian is less advanced on this development path and still deictically rooted, whereas the other Slavic languages in contact with socio-culturally dominant article languages developed far-reaching adaptive mechanisms without reaching the critical point of a grammatical change. We can observe such developments in the dialects, but critical reflection about language norms usually had the effect of keeping such phenomena outside of the standard languages. Significantly, the Bulgarian adoption of the definite article occurred as part of a complex process of adopting so-called Balkanisms (see e.g., Friedman 1983; Hinrichs 1990) including loss of case morphology, restriction of wordorder freedom and object doubling, which together changed the typological profile of the language. The space for changes is determined by sets of categories that together constitute a language type and the tipping point for the definite article in Slavic is apparently reached when the case system decays and word order becomes heavily constrained. For the indefinite article, the tipping point seems to be reached when – in addition to referential uses – also nonreferential uses obtain the indefinite marker. In Slavic, the articles emerge in contact with article languages. The early Bible texts show increased uses of the demonstrative t-pronoun (as illustrated above), but not to the extent equivalent to an article language (therefore called “articloid” by Mendoza 2004). The development beyond this level, yielding
128
Jadranka Gvozdanović
a grammaticalized article, occurs only after extended periods of bilingualism with an article language and depending on other system-inherent properties (such as adjectival definiteness or specificity expressed by means of word order). Indefiniteness markers are (in line with the definition by Friedman 2015) even more restricted in their uses. The development of the articles is in all the instances in Slavic a relatively new phenomenon, to be dated since the late Middle Ages.
2.2 Relative clauses Slavic prototypically builds relative clauses with relative pronouns. Historically these were *jĭ-type demonstrative pronouns (also referred to as the *yo-type) and *kŭ-type interrogative/relative pronouns. This distinction was lost in many Slavic languages centuries ago (see, e.g., Rappaport 2000: 3). Zubatý (1918) claimed, concerning jenž and který, that the former is definite and the latter non-specific and general (e.g., dám ti knihu již ~ kterou chceš ‘I will give you the book which you want ~ any book which you want’), but this distinction was not confirmed as still valid by his contemporaries. This contrast was simplified or eliminated in the other Slavic languages as well (for a detailed survey of East Slavic, see Danylenko 2014), but the conceptual distinction between these two types of relative reference essentially survives into modern Slavic languages. As shown by Fried (2010), contemporary Czech uses the bare (indeclinable) relativizer co for determinative relative clauses and co with an (accusative) resumptive pronoun for non-determinative (explicative) relative clauses. The strategy of using an indeclinable relativizer with resumption in non-determinative (i.e., non-restrictive) relative clauses is a phenomenon attested in Slavic from Slovene in the West to Ukrainian, Belarusian (and to a limited extent Russian) in the East (see Danylenko 2014 concerning East Slavic; Browne 1986 concerning Croatian and Serbian). In most languages, the indeclinable form is ‘what’ (e.g., ščo in Ukrainian, co in Czech, što in Croatian and Serbian), but in Slovene it is ki ‘that’, resembling the Gallo-Romance type (French que, northern Italian che).7 In order to understand system-internal reasons for the development of indeclinable relativizer strategies, it is instructive to take a look at Freising Fragments, written around the year 1000 in Pannonia, but presumably a copy of a two
7 The term ‘where’ is used as an indeclinable relativizer in the languages of the Balkan Sprachbund (e.g., deto ‘where’ in Bulgarian) and also in East Slavic (Southwest and East Ukrainian and some other dialects), discussed, for example, in Murelli (2011) and Danylenko (2014).
4 Standard Average European revisited in the light of Slavic evidence
129
centuries older late Common Slavic (Proto-Slovene) text. The only relativizing strategy in the Freising Fragments was based on declined demonstrative *jĭ followed by the focusing particle že. There are 18 instances of this demonstrative/relative pronoun meaning ‘which, who, that, what’ and additional 10 instances of this demonstrative/relative pronoun meaning ‘because, for, when, that’. The former instances became replaced by ki + resumptive pronoun (‘which, who, that’) or kar ‘what’ in Slovene, as illustrated by (14), whereas the latter instances were in due course replaced by specified relativizers ker ‘because, for’, ko ‘when’, kot ‘as’, da ‘that’ (Bernik et al. 1993). (14) Freising Fragments (Proto-Slovene) Taje dela načnem delati jaže that.ACC.PL deed.ACC.PL shall.PERF.NONPST.1PL do.INF which.ACC.PL oni delaše (2, 43/4). they.NOM do.IMPERF.3PL Contemporary Slovene translation Ista dela začnemo delati, ki so same.ACC.PL deed.ACC.PL shall.PERF.NONPAST.1PL do.INF which AUX.PL jih oni delali. they.ACC.PL they.NOM do.APP.PL ‘We (shall) perform the same deeds that they performed.’ Relative clauses with relative pronouns were thus historically well attested in Slavic. The same can be shown, for instance, for Celtic, where the relative *io, attested in Hispano-Celtic, is assumed to have had a Proto-Celtic origin (see Eska and Evans 1993: 35). Old Irish Milan Glosses of the ninth century probably indicate the beginning of the development towards a unified indeclinable relativizer (in view of the variation discussed by McKone 1980).8 Relative clauses with relative pronouns consequently span Europe from east to west and have attestations older than SAE. Their origin can be placed into the prehistoric period.9
8 The Old Irish relative *io had a Proto-Celtic origin; it existed already in Hispano-Celtic (Eska and Evans 1993: 35). 9 Equative constructions based on the relative marker are wide-spread across Europe and may be considered related to relative clauses, e.g., Russian tak (že) . . . kak . . . ; Czech tak . . . jako ‘such as’. Exceptions such as those in Celtic (cf. Irish chomh . . . le) or Scandinavian (lika . . . som) are due to later developments but typologically not fundamentally different. Their origin is also presumably older than SAE.
130
Jadranka Gvozdanović
2.3 Possession and habeo-perfect Early stages of Indo-European languages were characterized by a change towards increased transitivity (see Klimov 1974; Gamkrelidze and Ivanov 1984; Bauer 1997). In the realm of possession, this can be illustrated by the Latin shift from the mihi est construction to the transitive habeo-construction. Celtic languages in the west and Slavic languages in the east of Europe preserved essentially the spatial expression of possession (of the type ‘at me is’ in the sense of ‘I have’),10 but the core of the European languages shifted towards a transitive expression of possession, followed by a corresponding shift towards a habeoperfect of transitive verbs (see Drinka in this volume, also Drinka 2017). As shown by Bauer (1997: 288–290), Latin had two types of mihi est constructions, the possessive and the verbal one, illustrated by the following examples (with the gerundive equaling a participle); example (15) is excerpted from the Early Roman playwright Titus Maccius Plautus’s play Casina (line 519).11 (15) Possessive mihi est construction in Early Latin cui sit uxor who.DAT be.SUBJ.3SG wife.NOM ‘to whom may be the wife’, i.e., ‘to take her as his wife’ (16) Verbal mihi est construction in Early Latin Liber legendus mihi est. book.NOM read.GERV.NOM me.DAT be.PRS.3.SG ‘I have the/a book to read.’ It is from the latter type of construction, in which the (participial) gerundive patterned with the noun expressing the semantic goal, that habeo-perfect developed through an intermediate stage of participle agreeing with the object before it became reanalyzed as part of the perfect tense. Parallel developments could be observed in other Romance and in Germanic languages, which also developed a transitive ‘have’ verb. In Old Church Slavonic there were three main strategies for expressing possession (cf. Danylenko 2002; also McAnallen 2011): (1) the verb im-ě-ti ‘to have’ (most widely spread; this atelic verb was related to the telic verb *im-tei > (j)ę-ti 10 Danylenko (2002) showed that East Slavic combines these ‘be’-type properties with (newer) expressions pertaining to the ‘have’-type. 11 Titus Maccius Plautus (254 – 184 BC), see. http://www.thelatinlibrary.com/plautus/cas. shtml
4 Standard Average European revisited in the light of Slavic evidence
131
‘to take, catch’), (2) the mihi est type of construction, known from Latin and rendered in Old Church Slavonic by the dative of the possessor, the verb ‘to be’ and the nominative of the possessed, and (3) locative expression (u ‘at’ + the genitive of possessor). Verbal mihi est constructions also existed, expressed by the infinitive (and a goal-object with transitive verbs). The development of the perfect tense crucially involved the active participle in Slavic. Old Church Slavonic had a variety of active participle + verb and ‘be’ verb constructions in which the active participle agreed with the agent in the nominative. As these constructions are in many instances attested in scriptural translations from Greek, we cannot rule out Greek influence (Večerka 1961: 72). However, such constructions were also attested in translations of other Greek syntactic patterns (e.g., the genitive absolute) as illustrated by (17a) from Večerka (1961: 72), and in Slavic texts between the eleventh and the sixteenth centuries without any Greek model (cf. Večerka 1961: 74) (17) Slavic ‘be’ + active participles a. Old Church Slavonic12 Svętaja že běsta pojąšta i xvalęšta holy.NOM PRT be.IMP.DUAL singing.NOM.DUAL and praising.NOM.DUAL boga. .god.ACC.SG.M ‘The holy ones were singing and praising God.’ b. Old Czech13 Těch, ješto jsú svět milujíc. these.LOC.PL who/that be.PRS.3PL light.ACC.SG.M loving ‘These who are loving the light.’ c. Middle Russian14 bě bo prosę u b(og)a. be.AOR.3SG because begging.NOM.SG at god.GEN.SG.M ‘because he was begging God.’
12 Codex Suprasliensis (183: 26–27), see http://titus.uni-frankfurt.de/texte/etc/slav/aksl/ suprasl/supra.htm 13 Štítného naučení křesťanské (Christian Doctrine by Štítný, 1376) (cited after Večerka 1961: 74) 14 Žitie Nifonta (Nifont’s Vita, from the 16th century) (cited after Večerka 1961: 74)
132
Jadranka Gvozdanović
d.
Old Polish15 Za iste iest bog sødzøcz ye for same.ACC.PL be.PRS.3SG god.NOM.SG.M judging.NOM they.ACC.PL na zemi. on earth.LOC.SG.F ‘For the same God is judging them on earth.’
e.
Old Croatian16 Ki li bi Poličanin svadiv se who Q be.COND.3SG inhabitant.of.Poljica.NOM.SG.M quarrel.APP REFL ili skarav se tere ubio Poličanina or offend.APP REFL then killed.APP inhabitant.of.Poljica.ACC.SG.M a ga Polica ne plaćaju. but he.ACC.SG Poljica not pay.PRS.3.PL ‘If an inhabitant of Poljica would kill an(other) inhabitant of Poljica after quarreling or offending (each other), then Poljica do not pay for him.’
The last example is highly indicative. It originates from the statutory of Poljica that emerged as an indigenous collection of tribal legal regulations in the fourteenth century. Here we have two active past participles with the conditional ‘to be’ verb and an -l (> -o) past participle with a resultative meaning. These examples show that—at least by the end of the fourteenth century—the use of active participles with the verb ‘to be’ was fully entrenched in the system, and only -l participle had the resultative meaning characteristic of the perfect tense. The perfect formation in Slavic may have been reinforced by translations from Greek, but it firmly rested on indigenous means of expression which were expanded. The formation of ‘habeo’ (so-called possessive) perfect in Slavic was a different matter. Both in the Balkan area and in Slavic languages under strong German influence, a syntactic reinterpretation occurred, cf. Kashubian miec + (passive past participle + accusative complement) -> (miec + PPP) + accusative complement (Nomachi 2012: 132). Recipient passive construction also emerged on a restricted scale and with abrupt grammaticalization, pointing to borrowing.
15 Psałterz floriański (Psalterium Florianum, 14th century), 57, 11; cited after Večerka (1961: 74) 16 Poljički statut (Statutary of Poljice, ca 1400), 38; cited after Večerka (1961:74).
4 Standard Average European revisited in the light of Slavic evidence
133
In line with Nomachi’s (2012) findings about Kashubian, the decisive point is reached when the passive past participle ceases to agree with the object.17 Otherwise, ‘to have’ + PPPs congruent with the object noun are encountered across Slavic, for which no reanalysis can be claimed. Incongruent past passive participles without ‘to have’ encountered in northern and eastern Slavic are a different matter, for they can be interpreted in reference to the entire clausal predication, not only the object.18 In (18), excerpted from Wiemer (2012: 32), the first (indeclinable) past passive participle is incongruent, it refers to the entire situation of tucking together the hands with a strap, in contrast to the final, congruent past passive participle, which determines the object ‘legs’. (18) Ukrainian Zakručeno ruky syryceju, nohy tighten.PPP.N.SG hand.NOM/ACC.PL strap.INSTR.SG leg.NOM/ACC.PL v kolodky zabyti. in wooden.shackle.ACC.PL.F lock.PPP.PL ‘The hands tightened with a strap, the feet (are) locked in blocks.’ Polish exhibits a further development, by which a past passive participle can be formed from intransitive verbs in what has been described as an active construction with an implied animate agent (cf. Siewierska 1988: 274). (19) Polish Kupiono mi jeszcze tornister. schoolbag.ACC.SG.M bought.PPP.SG.N I.DAT also ‘There was also a schoolbag bought for me.’ This is an existential possessive construction (cf. the Latin construction mihi est), in which the resultative state is ascribed to the situation as a whole. Whenever this happens, PPP-agreement with the goal argument is lost in a language. Varieties at the borders of Slavic (North Russian, Thracian Bulgarian, southwestern Macedonian) have undergone two mutually connected changes: (a) loss of congruence between objects and participles (formally past passive participles, but without passive meaning when attached to intransitive verbs)
17 It was Nitsch (1954: 275) who attested non-agreeing past passive participles in Kashubian, e.g., òni miëlë tégò króla pòbité/ pòbity (*pòbitégò) ‘they had this king killed/ they killed this king’, which can be viewed as a clear signal of grammaticalization. 18 See also the discussion of the northern Russian constructions in Orr (1989).
134
Jadranka Gvozdanović
and (b) loss of restriction to transitive verbs, as illustrated by (20a) from Vasilev (1968: 220), (20b) from Vasilev (1968: 217) and (20c) from Friedman (1976: 99).19 (20) Peripheral Slavic varieties: a. North Russian20 U nego v gorod uexano. at he.GEN in town.ACC.SG.M go.off.PPP.N.SG ‘He has gone to town.’ b. Southern Thracian Bulgarian Tugava ošte nemax xodeno na kasabata. then yet not.have.IMP.1SG gone.PPP.N.SG onto town.DEM ‘Then I had not yet gone to the town.’ c. Macedonian south and west of the Vardar River Nožov me ima isečeno. knife.PROX.DEM I.ACC have.PRS.3SG cut.PPP.N.SG ‘This (here) knife has cut me.’ All these instances lack participial congruence. Moreover, the past passive participle is no longer restricted to transitive verbs, and its original passive meaning is thereby lost. In northern Russian, this may be related to Scandinavian influence (Breu 1996: 31), in Southern Thracian Bulgarian to Greek influence, and in southwestern Macedonian to Romance influence. It is important to note that these neighboring languages have conflated passive and active participles. In Macedonian, ‘have’ + PPP perfects are marked for resultativity, whereas the original perfect and pluperfect are only marked temporally (i.e., for taxis) as shown by the following contrasting Macedonian examples (Friedman 2015: 18). (21) Macedonian a. Taa mi go pokaža, no jas vekje already she me.DAT he.ACC show.AOR.3SG but I go bev videl. he.ACC be.AOR.3SG see.APP
19 Danylenko (2005) pointed to the resultative meaning of such constructions in North Russian. 20 Also discussed in Timberlake (1974).
4 Standard Average European revisited in the light of Slavic evidence
b.
135
Taa mi go pokaža, no jas vekje already she I.DAT he.ACC show.AOR.3SG but I go imav videno. he.ACC have.AOR.3SG see.PPP ‘She pointed him out to me, but I had already seen him.’
In the development of so-called possessive perfect, consisting of ‘have’, an object and a past passive participle, several stages can, in my view, be distinguished: (1) a stage at which the past passive participle determines the object, (2) a stage at which the past passive participle determines the entire predicate, and (3) a stage at which the past passive participle refers to the entire predication.21 The critical transition point occurs between stages 1 and 2. It is there that ‘have’ loses the possessive meaning and becomes a relational auxiliary. The spread of such constructions in the European languages may have started earlier than tacitly assumed by Heine and Kuteva (2006), i.e., already in the Early Middle Ages. Drinka (2011, 2017) pointed to the role of Biblical texts in the spread of periphrastic perfects of the type discussed above. New Testament Greek provided only an impetus carried on by Latin and Gothic. Wulfila’s Gothic Bible from the fourth century AD provides evidence for close copying of the Greek objective-complement construction (see Drinka 2011: 59), equaling stage 1 in my classification of the Slavic developments, and also the beginning of a ‘have’ perfect (with both ‘have’ and main-verb participles) as a translation of the Greek synthetic perfect participle providing the basis for later more grammaticalized ‘habeo’ perfect attested in the sixth century AD (see Drinka 2011: 60), equaling my stage 2. The critical transition between stages 1 and 2 in Gothic apparently started in Wulfila’s Bible of the fourth century. More than a millennium later, Slavic peripheral varieties followed the same pattern of development.22
21 Heine and Kuteva (2006: 148) distinguish the following stages in the rise of possessive perfects: (0) the meaning of possession, (1) resultative, (2) the perfect meaning, (3) past meaning, (4) loss of aorist or preterite function, (5) the possessive perfect is generalized as a past time marker. The stages discussed by me all occur within what Heine and Kuteva call “resultative”, in my view glossing over important changes within this phase. 22 Participial passives were closely connected with this development, but their origin was clearly older by virtue of the Indo-European reconstruction of past passive participle desinences.
136
Jadranka Gvozdanović
2.4 Other SAE features Even a brief survey of the other core features of SAE in the Slavic area points to their presence on a non-incidental scale. Participial passives are attested throughout the Slavic languages. Anti-causative prominence has different forms of expression in Slavic, most widespread is a variant of reflexive formation.23 Expressions for experiencer are interestingly varied across Slavic. Physical experiences combine with an accusative goal-experiencer (e.g., ‘I have a headache’ translates into Croatian as boli me.ACC glava ‘(my) head is aching (me)’), a possessor-experiencer (e.g., Russian u menja.GEN bolit golova ‘head is aching at me’), or a dative experiencer (e.g., Ukrainian jomu.DAT bolyt’ holova ‘to him/his head is aching’). Perceptual experiencers occur prototypically in the dative or nominative case, and emotional experiencers cover the full range and depend on the degree of intentionality. Of the remaining properties, it is the basic and traditional pattern in Slavic that negative pronouns do not occur without verb negation. However, in contact with Standard Italian and the surrounding southern dialects of Italian, Molise Croatian developed an optional omission of the negative verb particle if a negative pronoun or adverb precedes the verb (cf. Breu in this volume). On the positive side, Slavic does have particles in comparative constructions, subject-person affixes are strict agreement markers, relative-based equative constructions occur in Slavic as well, as does intensifier-reflexive differentiation. Intensification is expressed by prosodic means, particles and, for instance, absolute superlatives. Among the specific case features classified as SAE, Haspelmath (2001: 1498) discusses also the dative external possessor (in contradistinction to locative external possession, characteristic of Northern and Northeastern Europe, and NP-internal possession, characteristic, for example, of English). Yet a closer look into Slavic shows that the matter is more complicated because the dative possessor is more widespread and, moreover, it is used in the first place for intimate relations of possession, as in the following Croatian example (which has parallels in Russian and other Slavic languages). In view of such uses, the label “dative of external possessor” does not seem fully appropriate.24
23 Anti-causative formation is viewed as an intransitive counterpart to a transitive verb, which expresses an event affecting its subject, while giving no semantic or syntactic indication of the cause of the event. 24 The comitative-instrumental polysemy is of newer origin.
4 Standard Average European revisited in the light of Slavic evidence
137
(22) Croatian To mi je sin. this I.DAT be.PRS.3SG son.NOM ‘This is my son.’ *To mi je uzvanik/poštar. this I.DAT be.PRS.3SG invitee.NOM/postman.NOM *‘This is my invitee/postman.’ Nomachi (2016) established a typological cline by which West Slavic and Slovene, which show more habere-type features, exhibit very limited dative external possession, in contrast to Southeast Slavic, that preserves more esse-type features and also allows for more uses of dative external possession.25 Croatian seems to occupy an intermediate position in this continuum. Slavic exhibits widespread strategies of relative formation from interrogatives. Particularly manner interrogatives are also involved in the formation of equatives, as shown by most Slavic and Romance languages, the Balkan languages, Germanic (excluding Scandinavian), as well as Hungarian, Romani and Georgian (Heine and Kuteva 2006: 240). By virtue of such distribution, this formation has low indicativeness for SAE. Following Ultan (1978), Haspelmath also mentions verb-fronting as a SAE strategy for interrogation, in contrast to the use of interrogative particles in Basque, Irish, Scottish Gaelic, Albanian, Hungarian, Lithuanian, Russian, Finnish, Turkish, and, according to Haspelmath (2001: 1501), Nakh-Daghestanian. What remains in the end are Romance and Germanic languages, and not all of them (see French est-ce que, which in fact developed into a question particle). For Slavic, verb fronting is an interrogation strategy only if accompanied by a question particle (i.e., li) or an interrogative intonation.26 On the other hand, interrogative intonation by itself suffices for interrogation (in so-called yes-no questions). Interrogative particles are focused, but do not require interrogative intonation. Slavic thus has a multitude of expression means for interrogation, and verb fronting is a subordinate strategy due to primarily pragmatically governed word-order rules. It is in this perspective that the languages listed above should be reconsidered. Strict agreement markers are also a relatively widespread phenomenon in Slavic and beyond. Finally, distinction of emphatic (so-called intensifier) and
25 Concerning possession in Old Russian, Eckhoff (2006) established a major division between relational and non-relational nouns. 26 It is suitable to speak about verb-fronting only if there is a predominant position for the verb in a language (such as the verb-second position in Germanic).
138
Jadranka Gvozdanović
non-emphatic reflexive pronouns is probably an old phenomenon in IndoEuropean and is found in various early attestations, including Celtic and Slavic.27 Slavic distinguishes between ‘self’ (e.g., the Russian genitive-accusative form sebja ‘self’) and the emphasizing reflexive pronoun (e.g., sam ‘self’ in Russian, see Klenin 1980). By having such formations, Slavic resembles the core languages of SAE (see German sich ‘self’ and selbst (emphatic), Italian si stesso (emphatic), and so on).28
3 Conclusions about the origins of SAE How did SAE arise? Haspelmath’s (2001: 1507) pointed out that the distribution of the SAE features29 in part crosscuts the old Indo-European dialectal boundaries, so these properties could not have been brought to Europe from the IndoEuropean homelands. This leads us to the question of the SAE origins. First of all, it is important to state that some of the presumed SAE features have a more general distribution in Indo-European and attestations earlier than the presumed origin of SAE between late Antiquity and the Early Middle Ages. This holds for relative clauses with relative pronouns (Section 2.2) and intensifier-reflexive differentiation (discussed in Section 2.4). Dative possessor constructions are also more widely spread (they were also attested across Celtic, but were lost due to attrition of inflectional morphology, see Irslinger 2015: 75). The intensifier-reflexive distinction (Section 2.4) can on the basis of distribution be assumed to be older than SAE. Relative-based equative constructions (mentioned in Section 2.2) have also been reconstructed as prehistoric.30 In addition, participial passives developed from participles whose formation also antedated the formation of SAE.31
27 For example, Old Irish had an emphasizing pronoun side (sede), based on ede, following the independent pronoun si ‘she’ (Lewis and Pedersen 1989: 201). ‘Self’, on the other hand, was expressed by a form of ‘be’ (f-, f-ad), and this persists into Modern Irish for all persons as féin. 28 Western South Slavic preserves emphatic pronouns (not only reflexive) as accented (e.g., sebe ‘self’) and non-emphatic pronouns as a clitic (i.e., se). 29 These are definite and indefinite articles, relative clauses with relative pronouns, the ‘habeo’-perfect, participial passive, dative possessors, negative pronouns without verbal negation, relative-based equative constructions, subject-person affixes as strict agreement markers, and intensifier/reflexive differentiation. 30 E.g., for Celtic based on *isetero > -ithir in Old Irish, cf. McKone (1994: 125). 31 For Celtic, the origin of participial passives was assumed by Irslinger (2015) to coincide with the formation of SAE.
4 Standard Average European revisited in the light of Slavic evidence
139
Adjectival definiteness arose in Slavic during the former part of the first millennium AD, possibly under Germanic influence, and articloid definiteness (partly also indefiniteness) has been developing in contact situations since the Late Middle Ages. In both processes, demonstrative markers provided the sources for markers of definiteness. Compared to the core SAE languages, Slavic had an attenuated development and partial preservation of adjectival definiteness, in addition to specificity expressed by means of word order. The other changes dated after late Antiquity also affected Slavic in a specific way. Slavic preserved the distinction between active and passive past participles, enabling their distinct employment in passive and perfect formations; as a consequence, ‘have’-perfects developed across Slavic only as secondary phenomena. Participation of Slavic in SAE thus did not discontinue the older structural traits (such as the distinction of active and passive past participles), but adapted and complemented then (e.g., Slavic preserved esse-perfect and introduced habeo-perfect as resultative, as illustrated by Macedonian in Section 2.3). The remaining properties discussed here were shown to be either older (probably dialectal Indo-European, for instance, uniting Celtic and Italic)32 or (in most instances) younger than the presumed origin of SAE. It is important to maintain that these properties could arise due to the similar Indo-European basis, but their formation also depended on other concurrent processes such as inflectional simplification or word order constraints. Finally, one should bear in mind that in addition to regional effects of vernacular contacts, Bible translations had overarching effects and gave an important impetus to the formation of SAE. This may later have been reinforced by Latin as the language of governance and learning in much of Europe, and by Vulgar Latin. This survey has also stressed the importance of recognizing the sudden effect of grammatical change on the level of the system by showing that for demonstrative pronouns, grammaticalization to articles hinges on bleaching of deictic semantics enabling nonreferential uses (Section 2.1). For the perfect tense (Section 2.3), grammaticalization hinges on bleaching of the participial semantics and the corresponding loss of syntactic congruence with the object. The presented Slavic data on change in progress thus provide an insight into the formation processes of SAE.
32 Concerning Celtic, see Irslinger (2015: 75).
140
Jadranka Gvozdanović
Abbreviations ACC ADJ AOR APP AUX COND DAT F GEN GERV IMP IMPERF INDECL INF INSTR LOC M N NOM NONPST PERF PL PP PPP PRS PRT PST Q REL SAE VOC
accusative adjective aorist active past participle auxiliary conditional dative feminine genitive gerundive imperfect imperfective indeclinable infinitive instrumental locative masculine neuter nominative nonpast tense perfective plural past participle past passive participle present tense particle past tense question particle relative (adjective) Standard Average European vocative
4 Standard Average European revisited in the light of Slavic evidence
141
References Andersen, Henning. 1973. Abductive and deductive change, Language 49 (4): 765–793. Bauer, Brigitte. 1997. Nominal syntax in Italic: a diachronic perspective. In: Jadranka Gvozdanović (ed.), Language Change and Functional Explanations, 273–301. Berlin: Mouton de Gruyter. Benacchio, Rosanna. 2014. Grammatikalizacija v situacijax jazykovogo kontakta: razvitie artiklja v rez’janskom dialekte [Grammaticalisation in situations of language contact: development of the article in the Resian dialect], In: Motoki Nomachi, Andrii Danylenko, and Predrag Piper (eds.), Grammaticalization and Lexicalization in the Slavic Languages, 205–217. Munich: Otto Sagner. Bernik, France et al. 1993. Brižinski spomeniki: znanstvenokritična izdaja [Freising Manuscripts: A Scholarly-Critical Edition]. Ljubljana: Znanstvenoraziskovalni center SAZU. Berger, Tilman. 1993. Das System der tschechischen Demonstrativpronomina – Textgrammatische und stilspezifische Gebrauchsbedingungen (unpublished Habilitationsschrift). Munich: Ludwig Maximilian University. Breu, Walter. 1996. Überlegungen zu einer Klassifikation des grammatischen Wandels im Sprachkontakt (am Beispiel slavischer Kontaktfälle). Sprachtypologie und Universalienforschung (STUF) 49: 21–38. Breu, Walter. 2008. Razvitie system artiklej v slavjanskix mikrojazykax v absoljutnom jazykovom kontakte [Development of the article system in Slavic micro-languages in absolute language contact]. In: Sebastian Kempgen et al. (eds.), Deutsche Beiträge zum 15. Internationalen Slavistenkongress Ohrid 2008, 75–88. Munich: Otto Sagner. Breu, Walter. 2012. The grammaticalization of an indefinite article in Slavic micro-languages. In: Björn Wiemer, Bernhard Wälchli and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 275–322 (Trends in Linguistics 242.). Berlin/ Boston: Mouton de Gruyter. Browne, Wayles. 1986. Relative Clauses of Serbo-Croatian in Comparison with English. Zagreb: University of Zagreb. Chafe, Wallace. 1994. Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: University of Chicago Press. Comrie, Bernard and Tania Kuteva. 2013. Relativization strategies. In: Matthew Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chap ter/s8, Accessed on September 1, 2017.) Danylenko, Andrii. 2002. The East Slavic HAVE: Revising a developmental scenario. In: Karlene Jones-Bley et al. (eds.), Proceedings of the Thirteenth UCLA Indo-European Conference, 105–127. Washington, DC: Institute for the Study of Men. Danylenko, Andrii. 2005. Is there any possessive perfect in North Russian? Word 56 (3): 347–379. Danylenko, Andrii. 2014. On the relativization strategies in East Slavic. In: Motoki Nomachi, Andrii Danylenko, and Predrag Piper (eds.), Grammaticalization and Lexicalization in the Slavic Languages, 183–204. Munich: Otto Sagner. Drinka, Bridget. 2011. The sacral stamp of Greek: periphrastic constructions in New Testament translations of Latin, Gothic and Old Church Slavonic. In: Eiric Welo (ed.), Indo-European Syntax and Pragmatics: Contrastive Approaches, 41–74. (Oslo Studies in Language 3.3.) Oslo: University of Oslo.
142
Jadranka Gvozdanović
Drinka, Bridget. 2017. Language Contact in Europe. The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Dryer, Matthew. 2013. Definite articles. In: Matthew Dryer and Martin Haspelmath (eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. Eckhoff, Hanne Martine. 2006. Old Russian Possessive Constructions: A Construction Grammar Account. Oslo: University of Oslo. Eska, Joseph F. and David Ellis Evans. 1993. Continental Celtic. In: Martin Ball (ed.), The Celtic Languages, 26–63. London: Routledge. Faßke, Helmut. 1980. Grammatik der obersorbischen Schriftsprache der Gegenwart. Morphologie. Bautzen: Domowina. Fried, Miriam. 2010. Accusative resumptive pronoun in the Czech relative clauses with absolute relativizer co. Korpus–gramatika–axiologie 1 (1): 16–29. Friedman, Victor A. 1976. Dialectal synchrony and diachronic syntax: The Macedonian perfect. In: Sanford B. Steever, Carol A. Walker, and Salikoko S. Mufwene (eds.), Chicago Linguistic Society, Papers from the Parasession on Diachronic Syntax, 96–104. Chicago: Chicago Linguistic Society. Friedman, Victor. 1983. Grammatical categories and a comparative Balkan grammar. In: Ziele und Wege der Balkanlinguistik, 81–98. (Balkanologische Veröffentlichungen 8.) Berlin: Osteuropa-Institut an der Freien Universität Berlin. Friedman, Victor. 2015. ‘One’ as an indefinite marker in Balkan and non-Balkan Slavic. In: Victor Friedman, Macedonian Studies 2, 129–164. Skopje: Macedonian Academy of Sciences and Arts. Gvozdanović, Jadranka. 1993. Language change: complex in its sources and directions, yet simple in systematicity. In: Jaap van Marle (ed.), Historical Linguistics 1991, Papers from the Tenth International Conference on Historical Linguistics, Amsterdam, August 12–16, 1991, 135–152. Amsterdam: John Benjamins. Gvozdanović, Jadranka. 1997. How languages living apart together may innovate their systems (as illustrated by -to in Russian). In: Raymond Hickey and Stanisław Puppel (eds.), Language History and Linguistic Modelling. A Festschrift for Jacek Fisiak on his 60th Birthday, 1069–1081. Berlin/New York: Mouton de Gruyter. Haspelmath, Martin. 1998. How young is Standard Average European? Language Sciences 20 (3): 271–287. Haspelmath, Martin. 2001. The European linguistic area: Standard Average European. In: Martin Haspelmath, Ekkehard König, Wolfgang Österreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, 1492–1510. Berlin/New York: Mouton de Gruyter. Haspelmath, Martin, Matthew S. Dryer, David Gil, and Bernard Comrie (eds.). 2005. The World Atlas of Language Structures. Oxford: Oxford University Press. [Revised version as 2008 World Atlas of Language Structures Online. Munich: Max Planck Digital Library, Available online at http://wals.info/index Heine, Bernd and Tania Kuteva. 2006. The Changing Languages of Europe. Oxford: Oxford University Press. Hinrichs, Uwe. 1990. Das Slavische und die sogenannten Balkanismen. Zeitschrift für Balkanologie 26 (1): 43–62. Irslinger, Britta. 2015. Standard Average European and the Western fringe: a reconsideration. Historische Sprachforschung 126: 33–88.
4 Standard Average European revisited in the light of Slavic evidence
143
Klenin, Emily. 1980. Sentential and discourse prominence: the case of the emphatic pronoun sam. Russian Linguistics 4: 269–280. Klimov, Georgij Andreevič. 1974. On the character of languages of active typology. Linguistics 131: 11–25. Knjaževskaja, Ol’ga A., Vladimir Georgievič Dem’janov, Maja Valentinovna Ljapon, and Sergej Ivanovič Kotkov (eds.). 1971. Uspenskij sbornik XII-XIII vv. [Anthology of the Dormition Cathedral, 12th–13th Century]. Moscow: Nauka. Kuz’mina, Irina Borisovna and Elena Vasil’evna Nemčenko. 1962. K voprosu o postpositivnych časticach v russkix govorax [Towards the question of post-positive particles in Russian dialects] In: Varvara Georgievna Orlova (ed.), Materialy i issledovanija po russkoj dialektologii. Novaja serija, Vol. 3, 5–32. Moscow: Nauka. Leinonen, Marja. 1998. The postpositive particle -to in Northern Russian dialects compared with Permic languages. Studia Slavica finlandensia 15: 74–90. Lewis, Henry and Holger Pedersen. 1989. A Concise Comparative Celtic Grammar. Göttingen: Vandenhoeck und Ruprecht. Mayer, Gerald L. 1988. The Definite Article in Contemporary Standard Bulgarian. Berlin: FU Osteuropa-Institut. McAnallen, Julia. 2011. Predicative possession in Old Church Slavic Bible translations. In: Eirik Welo (ed.), Indo-European Syntax and Pragmatics: Contrastive Approaches, 155–172. (Oslo Studies in Language 3.3.) Oslo: University of Oslo. McKone, Kim. 1994. An tSean-Ghaeilge agus a réamhstair [Old Irish and its Prehistory]. In: Kim McKone, Damian McManus, Cathal Ó Háinle, Nicolas Williams, and Liam Breatnach, (eds.), Stair na Gaeilge. In ómos do Pádraig Ó Fiannachta, 61–219. Maynooth: Maynooth University, Department of Old Irish. Mendoza, Imke. 2004. Nominaldetermination im Polnischen. Die primären Ausdrucksmittel. LMU München: Habilitationsschrift. Meščerskij, Nikita Aleksandrovič (ed.). 1972. Russkaja dialektologija [Russian Dialectology]. Moscow: Nauka. Murelli, Adriano. 2011. Relative Constructions in European Non-standard Varieties. Berlin/ Boston: Mouton de Gruyter. Nitsch, Kazimierz. 1954. Wybór pism polonistycznych [Selected Polonistic Writings], Vol. 3. Wrócław: Ossolineum. Nomachi, Motoki. 2006. Polska konstrukcja rezultatywno-posesywna mam to zrobione a kaszubskie jô móm to zrobioné [The Polish resultative-possessive construction mam to zrobione and Kashubian jó móm to zrobioné]. Język polski 86 (3): 173–183. Nomachi, Motoki. 2012. The recipient-passive construction and its grammaticalization in Kashubian. In: Andrii Danylenko and Serhii Vakulenko (eds.), Studien zu Sprache, Literatur und Kultur bei den Slaven; Gedenkschrift für George Y. Shevelov aus Anlass seines 100. Geburtstages und 10. Todestages, 109–135. Munich: Otto Sagner. Nomachi, Motoki. 2016. Dative of external possession in Croatian: From an areal typological perspective. Jezikoslovlje 17 (1/2): 453–474. Orr, Robert. 1989. A Russo-Goidelic syntactic parallel: U nego svoja izba postavlena/Tá sé déanta agam. General Linguistics 29 (1): 1–21. Rappaport, Gilbert Charles. 2000. The Czech possessive reflexive pronoun jehož and its paradigm: synchronic morphosyntax and developmental pathways of the Slavic relative clause. Slavonic and East European Journal 44 (1): 1–18.
144
Jadranka Gvozdanović
Runić, Marija. 2013. The definite article in the Slovene dialect of Resia. Quaderni di lavoro Atlante Sintattico d’Italia 16: 91–108. Scholze, Lenka. 2007. Das grammatische System der obersorbischen Umgangssprache unter besonderer Berücksichtigung des Sprachkontakts. Konstantz: Universität Konstanz. Siewierska, Anna. 1988. The passive in Slavic. In: Masayoshi Shibatani (ed.), Passive and Voice, 243–289. (Typological Studies in Language 16.) Amsterdam: John Benjamins. Steenwijk, Han. 1992. The Slovene Dialect of Resia: San Giorgio. Amsterdam: Rodopi. Timberlake, Alan. 1974. The nominative object in North Russian. In: Richard Brecht and Catherine Chvany (eds.), Slavic Transformational Syntax, 219–243. Ann Arbor: University of Michigan. Topolińska, Zuzanna. 1981. Remarks on the Slavic Noun Phrase. Wrócław: Ossolineum. Topolinjska [Topolińska], Zuzanna. 2009. Definiteness. In: Sebastian Kempgen et al. (eds.), Die slavischen Sprachen. Ein internationales Handbuch zu ihrer Struktur und Geschichte, 176–188. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 32.2.) Berlin/ Munich/Boston: Mouton de Gruyter. Trubar, Primož. 1564. Cerkovna ordninga [The Church Ruling]. Tübingen: Ulrich Mohart. Ultan, Russel. 1978. Some general characteristics of interrogative systems, In: Joseph Greenberg (ed.), Universals of Human Language, Vol. 4, 211–248. Stanford: Stanford University Press. Vasilev, Christo. 1968. Der romanische Perfekttyp im Slavischen. In: Erwin Koschmieder and Maximilian Braun (eds.), Slavistische Studien zum VI Internationalen Slavistenkongress in Prag, 215–230. Munich: Trofenik. Večerka, Radoslav. 1961. Syntax aktivních participií v staroslověnštině [The Syntax of Active Participles in Old Slavic]. Prague: Státní pedagogické nakladatelství. von Heusinger, Klaus. 2012. Referentialität, Spezifizität, und Diskursprominenz im Sprachvergleich. In: Lutz Gunkel and Gisela Zifonun (eds.), Deutsch im Sprachvergleich: Grammatische Kontraste und Konvergenzen, 417–455. Berlin: de Gruyter. Whorf, Benjamin Lee. 1941. The relation of habitual thought and behavior to language. In: Leslie Spier et al. (eds.), Language, Culture, and Personality. Essays in Memory of Edward Sapir, 75–93. Menasha, WI: Greenwood Press. Wiemer, Björn and Markus Giger. 2005. Resultativa in den nordslavischen und baltischen Sprachen. (LINCOM Studies in Language Typology 10.) Munich: LINCOM Europa. Wiemer, Björn. 2012. Zum wechselndem Status nichtkongruierender Prädikationstypen im nördlichen Slavischen und Litauischen. In: Andrii Danylenko and Serhii Vakulenko (eds.), Studien zu Sprache, Literatur und Kultur bei den Slaven; Gedenkschrift für George Y. Shevelov aus Anlass seines 100. Geburtstages und 10. Todestages, 31–57. Munich: Otto Sagner. Zajceva, Marija I. 1981. Grammatika vepsskogo jazyka [The Grammar of the Veps Language]. Leningrad: Akademija nauk SSSR. Zubatý, Josef. 1918. Jenž, který, kdo, co [That, which, who, what], Naše řeč 2: 37–44.
Bridget Drinka
5 The perfects of Eastern “Standard Average European”: Byzantine Greek, Old Church Slavonic, and the role of roofing 1 Europe as a linguistic area Recent inquiries into the possibility of a European Sprachbund have yielded promising results. In a series of volumes produced in connection with the EUROTYP Project, scholars have provided evidence for an areal distribution in Europe of converbs (Haspelmath and König 1995), negatives (Bernini and Ramat 1996), adverbial constructions (van der Auwera 1998a), tense and aspect (Dahl 2000), and other structures. Haspelmath (1998, 2001) uses Whorf’s (1956) term “Standard Average European” (SAE) to characterize this Sprachbund, and assembles, as evidence, nine crucial features which are shared by a number of European languages but which are not well-attested in the languages of the world (Haspelmath 2001: 1503–1505): 1. 2. 3. 4. 5.
Definite and indefinite articles Post-nominal relative clauses with inflected relative pronouns The HAVE perfect Participial passives Dative external possessors (Grm Die Mutter wäscht dem Kind die Haare ‘The mother washes the child his hair’) 6. Negative pronouns and lack of verbal negation 7. Relative-based equative constructions (Cz tak Z jako X ‘as Z as X’) 8. Subject person affixes as agreement markers (non pro-drop) 9. Intensifier-reflexive differentiation (Grm sich vs. selbst)
Using these features, Haspelmath constructs a cluster map, based on the concept of “isopleths” or quantified isoglosses (van der Auwera 1998a), to display how many of these features each language has (Map 1). As can be easily seen, French and German have all nine features, and are considered to be the “nuclear” members of SAE – a finding congruent with van der Auwera’s (1998b) designation of the nucleus as the “Charlemagne Sprachbund”. Several other languages surrounding the nucleus also participate fully in these https://doi.org/10.1515/9783110639223-006
146
Bridget Drinka
Nnts
Fin
Ice
Est
Swd
Nor 6
Ir
Lit
Wel 1
2
Grm Fr
Rus
Pol
Dut
Eng 7
Brt
Kom
Ltv
Cz
9
Udm
5 Hng
Tat
Ukr
Sln
Bsq It
8 Srd
Spn Prt Mlt 2
SCr Lzg
7 Rom Alb Blg Grk
Grg Trk
l Arm
Map 1: Cluster map of SAE Features (Haspelmath 2001: 1505).
trends: Spanish, Portuguese, Italian, Sardinian, Albanian, Dutch, all possess 8 key features, while Greek, Romanian, and English have 7, making them participants, but less centrally so. Notably, most Baltic and Slavic languages are even less central, with 5 features, and other European languages are classed by Haspelmath as non-members of the Sprachbund: Basque, Breton, and Maltese with 2 features, Welsh, Georgian, and Armenian with 1, and Irish and Turkish with none.
2 Standard Average European east and west and the role of roofing Alongside Haspelmath’s contributions to the identification of the distributional patterns of morphosyntactic features in Europe can be placed the innovative work of Kortmann (1998a, 1998b), whose thoroughgoing analysis of adverbial subordinators reveals a significant fact: among the languages of Europe, an east-west split exists in subordinator use, with eastern languages relying more heavily on temporal subordinators (e.g., when, while, before) and western languages using many more causal, conditional, or concessive (CCC) subordinators (e.g., because, if, although) (Kortmann 1998b: 530–535). When Kortmann compared the ratios1 of CCC: TIME subordinators in Classical Greek and Latin to
1 These ratios are calculated by dividing the percentage of CCC readings of the subordinator inventories of each language by the percentage of temporal readings.
5 The perfects of Eastern “Standard Average European”
147
those of the modern European languages, he was surprised to discover that they followed the same patterns of east and west, respectively: Latin showed a considerably higher ratio of CCCs than Classical Greek. Kortmann proposes that the “drastically” divergent profiles of these two ancient languages, one resembling the languages of the east, the other the west, suggests that Classical Greek and Latin served as models for languages in their own spheres of influence. As languages of learning, religion, law, and literacy, Greek and Latin acted as “guiding” or “roof” languages (Kortmann 1998a: 221), providing a classical template for later writers to follow. Striking evidence for this claim emerges when Kortmann calculates the average percentage of CCC and TIME subordinators for 6 western SAE languages, viz., French, Spanish, Italian, German, English, and Dutch, and finds almost identical percentages to those found in Latin. Similar correspondences exist in the east: Romanian has a much lower ratio than its Romance congeners, and, indeed, has a profile almost identical to that of Classical Greek. Gothic, which is attested almost exclusively in 4th-century translations of the Greek New Testament, does not conform to the profile of other Germanic languages, but has a much lower CCC: TIME ratio, approximating that of Greek. What these facts suggest is that Romanian and Gothic, positioned as they are in the Balkans, conform more closely to the pattern of Greek than to that of their geographically remote relatives. The role of religious affiliation and cultural hegemony is clearly in evidence here. As Kortmann notes (1998b: 535), Romanian speakers have always been Greek Orthodox, and Greek was spoken by the upper classes in Romania until the early 19th century. The “sacral stamp” of Greek is likewise pre-eminent in Gothic (Drinka 2011). The east / west split, charted out by Kortmann in Map 2, thus appears to be intimately tied to the division between Greek Orthodoxy and Roman Catholicism. These insights will emerge as essential as we turn our attention in the following section to a similar division among the periphrastic perfects in eastern and western Europe.2
2 Gaeta and Luraghi (2002: 136) chart the distribution of gender and case in the languages of Europe, and produce a map which bears striking resemblance to Maps 2 and 3 here: languages in the eastern half tend to retain case distinctions and a three-gender system, while those in the west are losing case marking and tend to have at most two genders. The authors do not provide an explanation for this distribution, other than the shared membership of many of the western languages in the Mediterranean linguistic area, but this remarkable parallelism may hint at motivations for this distribution which are similar to those proposed here.
148
Bridget Drinka
Ice Fin ScGl Ltv Mnx
Ir
Lith Dan
Wls
Rus
Eng Dut
Pol Yidd
Brt
Grm Fr Hng
Bsq
Goth
Lat
Rom
SCr Prt
Rmni
It Spn
Blg
Ctl
Grg Trk
Arm
Mcd Alb ClGr Grk Mlt Map 2: Distribution of CCC: TIME ratios in the languages of Europe (after Kortmann 1998b: 534).
Bold: I languages with a ratio of 1.5 and higher Italic: II languages with a ratio higher than 1.0 but lower than 1.5 Normal: III languages with a ratio of 1.0 or lower Solid line demarcates the linguistic core area of Europe Dotted line indicates the east-west division of the linguistic core areas
3 The perfect, east and west As illustrated in Map 3, the periphrastic perfects of Europe illustrate an east / west split similar to that described by Kortmann: while the Western European languages form the perfect with HAVE / BE + PPP (past passive participle), the Eastern European languages form it with BE + PAP (past active participle), or with a secondary adoption of HAVE perfects beside older BE forms in languages in contact with western BE / HAVE languages (see Drinka 2003, 2013, and Heine and Kuteva 2006: 140–182 for more details).
5 The perfects of Eastern “Standard Average European”
149
Ice Far Swd
Fin (Est) Dan (Ltv) IR Slc Kash (Lith) (Rus) WLS Eng Fris (Blr) Dut (LSrb) (Pol) Brt St Grm (USrb) Yidd StFr (Cz) (Ukr) SGrm (Slva) Rmns Frin Hng Lad [Glc] Bsq Occ NIt Sln SCr Rom [Prt] {Cast} {Ctl} StIt Gag Srd Alb Mcd (Blg) Trk Sic Cal Grk Nor
SCGL
Arm
Map 3: HAVE/BE perfect auxiliation (Drinka 2017: 3).
Bold Underline Italics [Brackets] {Curly brackets} CAPS (Parentheses) Strikethrough Double Strikethrough
+ HAVE BE only HAVE only ter / tener used as aux haver used as aux; main vb. > tener BE (+ ‘after’) + verbal noun (Ir, Wls, ScGl) historically BE, with some examples of HAVE (esp. WSlav, Circum-Baltic) BE auxiliary lost (Rus, Blr, Ukr, Hung) No perf (Gag, Turk) or PRET greatly preferred (Sic, Cal, Prt, Grk) BE
As explored in detail in Drinka (2017), the evidence for areal spread is extensive. In western Europe, the present-day distribution resulted from at least three layers of diffusion, with each layer superimposing itself upon the next: an early spread of HAVE perfects across the Romance and Germanic territory, through inheritance from or influence of Latin; a later expansion of the BE perfects alongside the HAVE perfects in the territory under Carolingian rule, probably assisted by scribal expansion of the use of deponents; and, finally, a shift of anterior meaning of the perfect to preterital meaning, found especially in the more circumscribed area of northern French, southern German, and northern
150
Bridget Drinka
Italian, initiated in the vernacular language of Paris. The role of Latin as a “roof” language, both as a classical foundation and as the language of prestige and learning in the Middle Ages, is well reflected in the perfects of western Europe. In the east, the “roofing” effects of Greek were likewise essential to the development of the temporal-aspectual system of Slavic, and it is to this complex development that we now turn.
4 Historical relations of Byzantium and Church Slavonic We begin with a brief examination of the historical and sociolinguistic factors that influenced this development. After the division of the Roman Empire into east and west at the end of the 4th century AD, Greek came to be the language of culture, learning, and official transactions among educated speakers across four-fifths of the Roman Empire (Horrocks 1997: 150). By the 6th century, it had become the language of prestige and power in the eastern Roman Empire; its influence continued to grow as Byzantine holdings increased, stretching in the 9th and 10th centuries across the eastern Mediterranean and as far east as the Caucasus. Byzantine influence was also greatly expanded by the work of missionaries, who were sent north to the Balkans and Central Europe, bringing Orthodox Christianity to the Slavic peoples. As a result, the culture of the Slavs was thoroughly infused with elements of Greek literature, art, and language. The intimate relationship between Church Slavonic and Byzantine Greek dates back to the earliest manifestations of Old Church Slavonic when, in the 9th century, missionaries Constantine (Cyril) and Methodius brought their translations of Greek religious texts to the Slavs of Greater Moravia. The brothers, who had grown up in bilingual Thessalonica, were native Greek speakers but were clearly in complete command of the Slavic variety of their city. Invited to Greater Moravia by Rostislav, the Grand Prince of Moravia, to counterbalance the German hierarchy that existed there, the brothers inaugurated Old Church Slavonic as a conduit of Byzantine and Orthodox culture and religion to the Slavic peoples. It was therefore not a coincidence that these missionary efforts took place in Greater Moravia, at the border of the eastern and western church territories. When Methodius died in 885, his followers were expelled from Moravia, and these disciples carried Christianity to other Slavic peoples, especially, initially, to the Bulgarians and Macedonians (Le Feuvre 2009: 7).
5 The perfects of Eastern “Standard Average European”
151
The earliest surviving Old Church Slavonic works date to the 10th and 11th centuries and exist only in later copies of these manuscripts, but their many archaic features, along with the apparent co-intelligibility between the Macedonian missionaries and their central European flock (Molnár 1985: 24), have led some to characterize Old Church Slavonic as a fairly accurate representative of Common Slavic – a designation which presents serious difficulties. For example, while Old Church Slavonic does possess many archaic features, it also displays innovative trends and variability across texts. More importantly, Old Church Slavonic itself is not so much a language as a corpus of texts, and thus does not represent a spoken language at a given point in space or time. What comes down to us are not exact renditions of earlier texts, but copies of copies of those texts, some of which have undergone extensive alteration, such as the removal of Moravianisms, the calcification of the liturgical language, and the reshaping of the texts according to Greek models (Shevelov 1988: 600, 613–615).3 The temporal or geographical provenance of a particular archaism or innovation is therefore not easily identified. While valuable insights can be gained about Common Slavic through the analysis of Old Church Slavonic forms, we will not go so far as to equate it with Common Slavic. The conservative nature of Church Slavonic throughout its history stems especially from the reverence for the written language of the old Slavonic texts, or crъkovnьnyę kъnigy ‘church books’, as revealed truth, in what might be regarded as an extreme example of the power of the “sacral stamp of Greek”. At various points in time, scholars and scribes sought to remove all deviations due to human intervention from the divine language of revelation represented in the Slavonic writings, and the essential role of Greek in this process of the “correcting of the books” is evident throughout the history of Slavonic (Lunt 1977: 441). Examples of this reliance on a Greek template can be found throughout the history of the Byzantine Empire. For example, after the fall of the First Bulgarian Empire in 1018, the Byzantine administration introduced hellenized features to the Church Slavonic of the Bulgaro-Macedonian area, and these innovations spread through missionary efforts to the newly-converted Kyivan (Kievan) Rus’ territories. Later, at the
3 MacRobert (1986: 162–164) demonstrates the subtlety of the relationship between native Slavic structures in Old Church Slavonic and those influenced by Greek, distinguishing such blatant foreign borrowings as the unparallel “compromise” forms of the Suprasliensis from “naturalized” forms like the neuter relative pronoun eže used as a “makeshift copy” of the Greek definite article.
152
Bridget Drinka
time of the Second Bulgarian Empire in the 14th century, a number of normative policies based on Greek grammatical patterns and stylistic principles were established in the Church Slavonic of Serbia and Bulgaria, as Church Slavonic became what Mathiesen (1984: 60) terms an “icon” of Orthodoxy, elevated but artificial (Shevelov 1988: 615). As spiritual leadership shifted from South Slavic to East Slavic, a number of southern innovations were introduced into the written language, including orthographic and syntactic features influenced by Greek. The written varieties of Rus(s)ian provide evidence of strong, persistent influence from Church Slavonic, and hence Greek, throughout history (MacRobert 1986: 150). A final, remarkable example of the use of Greek as a model for Slavonic is to be found on the border between Slavia Orthodoxa and Slavia Romana (Picchio 1980: 22) in the late 16th and early 17th centuries: as a reaction to the claims of some Catholic clergymen and scholars that only Latin, not Church Slavonic, was a suitable language for liturgical purposes, Orthodox scholars of the Ruthenian lands, that is, Belarus’ and Ukraine, composed grammars to demonstrate the validity of Slavonic, relying heavily on the grammatical patterns of Greek, and at times even regarding Slavonic and Greek as sharing one grammar (see Frick 1985). The classical model of Greek was seen as capable of imbuing Church Slavonic with dignity and connection to humanistic values, allowing it to rival or even surpass Latin (Picchio 1980: 26–30). For Church Slavonic, then, the prestigious model of Greek remained close at hand throughout the centuries, providing examples of “correct” usage and syntactic and stylistic patterns to be imitated.
5 The sacral stamp of Greek and the Old Church Slavonic verb system The influence of the “sacral stamp” of Greek, witnessed likewise in Latin and Gothic, is even more decidedly in evidence in Old Church Slavonic, where the matching of word to word and form to form is often exact (Hannick 1972: 424; Tzitzilis 1999: 605). As demonstrated by linguistic and statistical analyses (Section 10), the verbal structures of Old Church Slavonic reflect those of Greek with remarkable precision: Greek aorists are almost always translated as Old Church Slavonic aorists, for example, and Greek imperfects are likewise usually translated as Old Church Slavonic imperfects.
5 The perfects of Eastern “Standard Average European”
153
5.1 The influence of Greek on Old Church Slavonic participial constructions One of the clearest examples of the effect of Greek influence on the Old Church Slavonic verb system is to be found in the participial constructions, including periphrastic verb forms. The identification of Old Church Slavonic past active participles with Greek aorist and perfect participles greatly increased their productivity, and extended their semantic range (Růžička 1963: 365–370). Růžička’s impressive array of examples of such influence illustrates not only how Old Church Slavonic translators patterned their participial usage on that of their Greek models, but also how precisely the construction of the replica reflects that of the original, as witnessed even by the use of the že particle in imitation of the Greek particle dè as in the following example: (1)
Luke 23: 55 Grk Katakolouthḗsasai dè hai follow.after.AOR.ACT.PTCP.F.NOM.PL PTCL DEF.ART.F.NOM.PL gunaîkes, haítines ē̃san woman.F.NOM.PL who.F.NOM.PL be.IMPERF.ACT.3PL sunelēluthuîai come.with.PERF.ACT.PTCP.F.NOM.PL ek tē̃s galilaías autō̃i. out.of DEF.ART.F.GEN.SG Galilee.F.GEN.SG him.DAT.SG OCS vъ slědъ že šьdъšę in tracks PTCL follow.PAP.F.NOM.PL ženy, jęže běaxǫ woman.F.NOM.PL who.F.NOM.PL be.IMPERF.ACT.3PL sъ nimь prišьly otъ galileję. with him.INST.SG come.PAP.PL from Galilee.F.GEN.SG ‘Following after, the women who had come with him from Galilee.’ (Růžička 1963: 17)
The Old Church Slavonic translator uses a participle construction to describe the following women, as Greek does, and copies exactly the Greek pluperfect construction in BE + perfect active participle (ē̃san sunelḗluthuîai) by using the imperfect form of BE + past active participle (PAP) (běaxǫ prišьly). It is clear that both participial formations illustrated here conform precisely to the Greek original.
154
Bridget Drinka
It will be instructive to examine how Latin renders these participial constructions, since its divergence from Greek and Old Church Slavonic points to key differences between eastern and western methods of forming these periphrastic constructions. The Latin translation is presented in (2): (2) Lat subsecutae autem mulieres, woman.F.NOM.PL. follow.after.PST.DEPON.PTCP.F.NOM.PL. PTCL quae cum eo venerant de Galilaea. who.F.NOM.PL with him come.PLUPERF.ACT.3PL from Galilee Consider first the introductory participle. While Greek and Old Church Slavonic, as illustrated in (1), have past active participles at their disposal, Latin has only the perfect deponent participle, identical in form to the perfect passive participle, to connote both pastness and activeness at once. As a deponent verb, then, subsecutae accurately replicates the aorist active participle of Greek. However, if a transitive verb had been used, such a precise replication could not have been made, since the perfect participles of transitive verbs in Latin are passive. The second participial construction, the pluperfect, differs more noticeably from the Greek model: Latin translates the Greek BE pluperfect (ē̃san sunelēluthuîai) as a synthetic pluperfect (venerant). As illustrated here, Greek periphrastic pluperfects tend to appear as other structures in Latin, such as synthetic pluperfects. Again, with no past active participle, Latin cannot easily replicate this structure, as Old Church Slavonic can; a semantically similar but structurally dissimilar form is chosen instead. The limited range of participles in Latin in comparison with those of Greek and Old Church Slavonic turns out to have major consequences: both the extension of HAVE as a perfect auxiliary for transitive verbs and of BE for unaccusative verbs can be seen as the result of an extension of those auxiliaries to address the lack of a productive past active participle. Both auxiliaries engendered subjectorientation in their participles: the HAVE auxiliary, once grammaticalized, allowed for the eventual interpretation of the past passive participle (PPP) as active, and the BE auxiliary, connected with expanded use of the deponent, created periphrastic deponents, allowing the PPP to be interpreted, like synthetic deponents, as passive in form but active in meaning. Greek and Old Church Slavonic, with their numerous participles, did not need to stretch their auxiliaries in this way. This fact is one of the fundamental reasons for the distribution of HAVE / BE perfects in the west, and BE perfects in the east.
5 The perfects of Eastern “Standard Average European”
155
5.2 The influence of Greek on Old Church Slavonic periphrastic constructions Turning to the periphrastic structures of Old Church Slavonic, we begin by reviewing the status of the synthetic perfect and the various periphrastic formations in Koine and Byzantine Greek, collected in Table 1.
Table 1: Periphrastic perfects and aorists in Koine and Byzantine Greek. SYNTHETIC PERFECT
KOINE & NEW Synthetic active TESTAMENT indicative perfects GREEK are losing ground at end of Koine; synthetic middle perfects less so. Synthetic pluperfects in decline. Aorists and perfects are synchretizing.
BYZANTINE GREEK
Synthetic perfect is lost in th c. AD
BE
+ AORIST
BE
+ PERFECT
HAVE
+ AORIST/
PARTICIPLE
PARTICIPLE
PERFECT PARTICIPLE
st century AD: new use as pluperfect, as synthetic pluperfect declined. Aorist and perfect participles are synchretizing, like synthetic finite correlates.
BE
+ active or passive perfect participle > frequent, often replacing synthetic forms. Perfect > narrative tense in NT. Periphrastic perfect of subjunctive, optative, imperative, infinitive go out of use.
HAVE + aorist participle are no longer used. HAVE + object + participle (especially mediopassive) as in Ancient Grk, but more frequent. Also HAVE + adjunct of time + participle appears.
Aorists and perfects merge.
BE
+ passive participle ( v / #); (2) the cluster often contains isoglosses of lexicalized unique phonetic phenomena whose distribution, following linguogeographical logic, usually diverges from the main separating lines that demarcate the action of phonetic laws; (3) the isoglosses are concentrated in the cluster so compactly as to form a virtually single line and almost never produce transitional types. The number of isoglosses in the cluster indicates that this is the most important line of language differentiation in the South Slavic linguistic territory, a point consistently disputed by Bulgarian linguists. The age of the oldest of these isoglosses indicates the pre-Balkan character of the opposition between the western and eastern parts of the South Slavic language area and provides evidence in favor of that opposition that does not, as has been claimed, “cover, and in so doing, erase the more ancient dialectal structure of the South Slavic languages” (Kurkina 1992: 19–20), but, on the contrary, at least for the dialect of East Serbia and West Bulgaria, should be taken in direct relation to that ancient dialectal structure. It would be wrong to assume, however, that the modern localization of the genetic cluster of isoglosses in western Bulgaria coincides with the location of the line separating the South Slavic east and west since the time of the Slavs’ first settlement in the region, as Popović (1960), for one, has suggested. The present line emerged, rather, during the modern era, not before the 17th century, due to migration of population from the West. It is symptomatic that the most ancient level of Slavic toponyms in the region points toward an initially different origin of the *tj > št / č, *dj > žd / ǯ, *ǫ > ǝ / u isoglosses, further to the West for some parts of the territory (Loma 1994). Before the migration to the East started, these dialects had occupied the eastern geographic periphery of the West South Slavic territory, without any direct contact with the East South Slavic area. In general, viewed genetically within the framework of areal linguistics, Slavic in the Balkans cannot be divided into any entities directly corresponding to any Slavic ethnic labels. While in theory the diasystem of any Balkan language is understood as a sum and a function of all of its territorial units, in practice one can describe any Balkan language as a sum and a function of the smallest sufficient number of purposefully chosen representative dialects, or category members that represent the whole category better than any others.
324
Andrey N. Sobolev
Balkan linguists would be more accurate if they referred not to languages “as a whole”, but rather to their representative dialects.15
3 Slavic in the Balkans viewed arealtypologically 3.1 The areal data Speakers of Balkan Slavic were immersed in a zone of active processes of convergence that led to the Balkan Sprachbund (convergent group).16 A unique system of similarities and differences among the Balkan languages and dialects (Miklosich 1861; Sandfeld 1930; Asenova [1989] 2002; Hinrichs 1999b), examined according to the methods of areal linguistics (Cyxun 1981) and presented in the Minor Dialect Atlas of Balkan Languages (MABL), reveals: a. heterogeneous areal and systemic distribution of each intra-Balkan linguistic phenomenon, as represented by individual isoglosses; b. irradiation centers of phenomena, elucidating their origin and the chronology of their emergence; c. stages of development of phenomena as demonstrated by their variants and frequency; d. linguistic microregions which enjoy a certain independence in the Balkans irrespective of any borders; e. general territorial division of the Balkan Sprachbund linguistic landscape irrespective of any borders, as represented by clusters of isoglosses.
15 A representative dialect of a particular language is a dialectal unit, purposefully selected by implementing a linguistic procedure – a unit which belongs to the core of a large dialectal area of a single language and, consequently, exhibits all the characteristic developmental tendencies associated with this area. Our list of representative Balkan dialects includes the Croatian neo-štokavian younger i-dialect, Serbian Zeta-Lovćen dialect, Serbian Timok dialect, West Macedonian Ohrid dialect, Southwest Bulgarian Pirin dialect, Bulgarian Rhodopi dialect, Bulgarian Moesia dialect, Albanian Gheg Dibra dialect, Albanian Tosk Skrapar dialect, North Greek Westmacedonian dialect, South Greek Peloponnissos dialect, South Arumanian Pindos Non-Farsherot dialect. 16 The structural innovation involving the transition from an inflective system to nominal analytism became the most important structural change among the Balkan Slavic languages. Bulgarian linguists have relied on this fact as justification for classifying this generally South Slavic type as Bulgarian (Kočev 2001); for another point of view, see Stojkov (1963).
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
325
A large portion of the grammatical and lexical phenomena mapped thus far in MABL is concentrated in the east and southeast area, which reflects mostly Greek – Bulgarian, and subsequently Greek – Bulgarian – Macedonian, linguistic links; the South-East area stands in opposition to the West area (see the map below). A range of phenomena affords the possibility of seeing the South-East area as the central area, that is, as the source of a large number of Balkan innovations that are absent within the relatively more archaic West area. In some cases, this area is evidently connected to the geographic location of Modern Greek as the main source language and as the most important mediating language for a large number of Balkan phenomena under consideration It is especially interesting that the major direction of area borders – northnortheast – south-southwest – repeats the direction of major genetic isoglosses in the South Slavic dialectal region. For instance, in the field of morphosyntax and syntax, the western group of Balkan dialects, which includes East Serbian, East Montenegrin, Gheg Albanian, Tosk Albanian, and, in some instances, Pindean South Aromanian and even some Greek dialects, is characterized in general by the following features (examples are excerpted from Sobolev 2003): (1) An accusative ending of feminine substantives is present in the function of direct object, in the case of Albanian cumulating case and definiteness: 1)
East Serbian ok'al-a sǝm žen-'u-tu call-PAP.F.SG be.PRS.1SG woman-F.SG.ACC-DEF.F.SG.ACC ‘I called that woman.’
2)
Gheg Albanian vel'i-u e sh'a-u gru-n Veli-DEF Her rant-AOR.3SG wife.F.SG-DEF.ACC ‘Veli ranted against his wife.’
There is no such ending in West Macedonian, Southwest, Rhodopian, and Northeast Bulgarian, North Greek, or Pindean South Aromanian. (2) The negative future form is built with a (not inflected) auxiliary ascending to ‘to want’17 (at times along with ‘to have’): 3)
Aromanian nu vaĭ ad'aru kǝl'iva not want.FUT Build the shepherdhouse ‘You will not build the shepherd shack.’
17 In some idioms, the East Serbian preserves the conjugated verb ču, češ and so on ‘to want’ as an auxiliary when forming the absolute future tense.
326
Andrey N. Sobolev
4)
East Serbian n'eče da=vr'evi-mo18 v'iše want.FUT.NEG PART.CONJUNCTIVE=talk-PRS.1PL more ‘We will talk no more.’
5)
East Serbian n'ema da=d'oĭd-u ĭ'utre have.FUT.NEG PART.CONJUNCTIVE=come-PRS.3PL tomorrow ‘They will not come tomorow.’
On the contrary, the forms with ‘to have’ are exclusively used in Southwest Bulgarian and Northeast Bulgarian, and preferred in Rhodopian Bulgarian: 6)
Rhodopian Bulgarian n'emʌ gọ v'ȇruv-ʌm have.FUT.NEG him trust-PRS.1SG ‘I will not trust him.’
7)
Rhodopian Bulgarian p'ǝrvʌtʌ nǝd'ȇl’ȇ zʌ 'avgus first week of august nǝ=št-'ȇš v'ig’ȇ k'ȏštʌ bǝs p'ečȇnọ NEG=want-PRS.2SG see.INF house without grilled
‘In the first week of August you will not see a household without a grilled lamb.’ (3) The preference for omitting the definite article (or the possibility of omitting it) in locative prepositional constructions has been observed:
8)
Aromanian fič'or-li iš'ǝrǝ tu k'al-i child-PL.DEF run on street-INDF ‘The children ran into the street.’
9)
East Serbian det'e se izĭur'ilo na put-# on street-INDF child-SG run ‘A child ran into the street.’
On the contrary, the (definitive) article is obligatory in phrases referring to the same state of affairs in West Macedonian, Southwest, Rhodopian and Northeast Bulgarian, North, and South Greek:
18 The conjunctive particle da is optional in the dialect.
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
327
10) Rhodopian Bulgarian vǝl’k’-'ȏn č'ȇkʌ-šǝ pọd ĭ'elʌ-nʌ wolf.M-DEF.DIST wait-IMPRF.3SG under spruce.F-DEF.DIST ‘The wolf waited under the spruce.’ (4) Conjunctions are exclusively used (or at least preferred against the prepositions) in comparative constructions with adjectives in comparative degree: 11)
se tëi Gheg Albanian ma=i=gjaet COMP=ART=high than.CONJ you ‘higher than you’
12)
East Montenegrin nije ĺepše viđela, no ńu be.NEG.AUX.PRS.3SG beautiful.COMP see than.CONJ her ‘She never saw anything more beautiful than that female creature.’
13)
East Serbian p'etar p'o=golem n'ego ĭ'ovan Petar PART.COMP=big than.CONJ Jovan ‘Petar is bigger/older than Jovan.’
(5) The short form of personal pronouns as an indirect or direct object precedes negation, auxiliary verbs, and participles in the (plu)perfect tense in the middle of a sentence with negation: 14) East Serbian t'i mu ne=s'i d'a-l p'are you.NOM to.him NEG=be-PRS.2SG give-PTCP.M.SG money ‘You didn’t give (the) money to him.’ Macedonian and Bulgarian dialects exhibit a high level of variety here, but none of the options is identical to the Western Balkan pattern (MABL 2003: Map 43). Compare: 15)
Rhodopian Bulgarian ĭ'ȇ gọ sǝm ne=r’'ukʌ-lʌ I.NOM him be.PRS.2SG NEG=call-PTCP.F.SG ‘I didn't call him.’
328
Andrey N. Sobolev
16)
Rhodopian Bulgarian ĭ'ȇ nǝ=s'ȏm gọ r’'ukʌ-lʌ I.NOM NEG=be.PRS.1SG him call-PTCP.F.SG ‘I didn't call him.’
17)
Rhodopian Bulgarian nǝ=m'ȏ sʌ ml'ogu tr'ȏsǝ-lǝ NEG=I.ACC be.PRS.3PL much search-PTCP.PL ‘They didn’t search a lot for me.’
(6) General questions can be formed by using nothing more than interrogative intonation, with or without negation: 18)
Tosk Albanian n çorrov'od qe? in Çorrovoda be.AOR.3SG ‘Were you in (the town of) Çorrovoda?’
19)
East Serbian 'oǯ da=d'oĭdeš? want.PRS.2SG PART.CONJ=come.PRS.2SG ‘Will you come?’
20)
West Macedonian ne
znaĭš? know.PRS.3SG ‘Don’t you know?’ NEG
21)
Gheg Albanian mo e shet at l'opën? NEG it sell this cow ‘Don’t you sell this cow?’
(7) Demonstrative particles used as predicatives allow for nominative instead of genitive or nominal accusative case: 22)
Gheg Albanian ja
Sefed'in-i Sefedin-DEF ‘Here is Sefedin.’ DEM
23)
East Serbian et'e čov'ek DEM man.M.SG.NOM.INDF ‘Here is the man.’
In the field of lexis, as shown in MABL, indicative instances of preservation in the Western area include Slavic archaisms (which may be both archaisms in
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
329
West Balkan Slavic and loans from Slavic to Albanian, Aromanian, and Greek), and Palaeo-Balkan, Latin, and even some specific Turkish lexical items. In the East, Greek loans and Slavic lexical archaisms are indicative. a. Slavic lexical items There are many Slavic lexical items that are very common in the West but not at all (or with significant formal, areal or semantic limitations) distributed in the East. Some random examples are as follows: (1) The reflex of Proto-Slavic *osojь ‘the shady side of a mountain’ is present in East Serbian osʹoĭ, Dalmatian Croatian osoĭ, East Montenegrin osoĭe, West Macedonian sʹoĭnica, Tosk Albanian soʹishtë, and Rhodopian Bulgarian usʹoĭkʌ. By contrast, we find the same semanteme expressed in Southwest Bulgarian zʹavrʌt, Northeast Bulgarian s’ʹȇnčẹstʌ stǝrnʹa, Gheg Albanian jʹeze, North Greek anʹilĭu, and Pindean South Aromanian tu k’eʹari. (2) The reflex of Proto-Slavic *loky ‘a puddle’ is present in East Serbian lʹokvina, Dalmatian Croatian lokva, East Montenegrin lokva, and West Macedonian łʹokma, in opposition to Southwest Bulgarian g’oł and g’ʹol’ište, Rhodopian Bulgarian bʌrʹug’i PL., Northeast Bulgarian g’oł, Gheg Albanian hurdh, Tosk Albanian pellgʹoskë, North Greek bʹara, and Pindean South Aromanian bʹarǝ and groʹapǝ. (3) The reflex of Proto-Slavic *lěnivъjь ‘lazy’ is present in East Serbian leńʹiv ADJ.M., Dalmatian Croatian linčina and lin ADJ.M., and East Montenegrin liĭen ADJ.M., opposing West Macedonian mʹǝrʒa and mʹǝrʒlif ADJ.M., as well as the following Turkisms: Rhodopian Bulgarian kʌłpʌzʹan’in, Northeast Bulgarian kʌłpʌzʹan’in (and mǝrz’ẹl’ʹif ADJ.M.), Southwest Bulgarian dembʹel’in, Gheg Albanian (njerʹi) demʹel ADJ.M., Tosk Albanian dembʹel, North Greek timbʹelis, and Pindean South Aromanian timbʹelu. (4) The reflex of Proto-Slavic *nevěsta with the meaning ‘daughter-in-law (son’s wife)’ is present in East Serbian nevʹesta, Dalmatian Croatian nevista, East Montenegrin nevĺesta and mlada, West Macedonian sna: and, rarely, nʹe:sta, Rhodopian Bulgarian nǝvʹȇstʌ, and Pindean South Aromanian nveʹastǝ, as opposed to Southwest Bulgarian snʹa:, Northeast Bulgarian snʌ(x)ʹa, Gheg Albanian nʹuse gjʹalit and e re:, Tosk Albanian nʹuse e djʹalit, and North Greek nif’. b. Palaeo-Balkan, Latin and Turkish lexical items There are also cases where Palaeo-Balkan, Latin and Turkish lexemes, uncommon in eastern Balkan dialects, are observed precisely in East Serbian, Dalmatian Croatian, East Montenegrin, Gheg and Tosk Albanian, and Pindean
330
Andrey N. Sobolev
South Aromanian. One such case involves reflexes of the Latinism *vessica: East Serbian bešʹika ‘a bladder’, Gheg Albanian mshegz ‘a blotch’, and Tosk Albanian fshʹikës ‘a blotch’. c. Greek lexical items Areal data enables us to characterize the West Balkan area through negative features as well, such as the lack (or significant formal, areal or semantic restrictions) of certain Balkan Graecisms, which are predominantly spread across Bulgarian and Greek language territory, for instance: (1) αρέσω / arʹesō ‘to appeal’ is absent in Dalmatian Croatian, East Montenegrin, East Serbian, West Macedonian, and Gheg and Tosk Albanian; (2) δάσκαλος / dʹaskalos ‘a teacher’ is characterized for example by informants in East Serbia as “Bulgarian”, compared to the authentic učʹiteĺ (the lexeme *dʹaskalos is absent in Dalmatian Croatian, East Montenegrin, and Gheg and Tosk Albanian); (3) ζευγάρι / zeugʹari ‘a pair’ is not observed in East Serbian, Dalmatian Croatian, East Montenegrin, or Gheg and Tosk Albanian; (4) κάθε / kʹathe ‘each’ is absent in East Serbian, as well as Dalmatian Croatian, East Montenegrin, and Gheg and Tosk Albanian. Some Greek and Balkan lexemes have been borrowed into the East Serbian dialect without any morphological adaptation that makes use of its own means of derivation and inflection. Greek αργάτης / argʹatis ‘a hired worker’ follows this pattern in East Serbian argat (same in East Montenegrin, West Macedonian, Gheg Albanian, Tosk Albanian, and Pindean South Aromanian), while in Bulgarian dialects the lexeme is formed with a singularity suffix (ʌrg’atʹin; compare North Greek arγ’ac < αργάτης / argʹatis).
4 Slavic in the Balkans as a donor Linguists should discern between inherited and acquired elements in a language. Among acquired elements, linguists should further distinguish those borrowed by the recipient language from those acquired through semantic shift or structural copying (that is, through a substitution of its own language material or function for another language’s material or function). Both of these mechanisms were at work in the contact situations typical for the Balkans, where South Slavic languages took form by borrowing something foreign and influencing other languages by giving away something of their own in turn. Lexical items loaned from Slavic into Balkan languages like Greek or
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
331
Albanian provide good illustrations of this relationship (see Budziszewska 1991; Ylli 1997; Sobolev 2012). Our own data do not include sufficiently reliable evidence for direct Slavic influence, in the form of lexical borrowing, on terms related to the spiritual culture of neighboring peoples; their area of distribution (with extremely rare exceptions) is usually limited to the geographic area of Macedonia and immediately adjacent regions, that is, to an area where there is no doubt as to the existence of a Slavic population today or in the recent past. Relevant data include a reflex of *kolęda with the meaning of ‘Eve’ in some of the Greek and Aromanian dialects of Aegean Macedonia and Epirus; the lexeme ta sʹurva ‘evening before New Year, New Year’s Eve’ in North Greek; and a reflex of Macedonian kʹǝrčma in Albanian and Greek dialects, meaning ‘celebration of the end of reaping and threshing’ (MABL 2005a). By contrast, Slavic languages have undoubtedly, although somewhat unexpectedly, served as Balkan-wide donor languages in the semantic field of animal husbandry (MABL 2009), a fact that has, unfortunately, been completely overlooked by Balkanists (Asenova [1989] 2002: 43–47, 2003). An Aromanian mediation ought to be presupposed in the diffusion of a certain portion of Slavisms. The phonetic characteristics of lexemes derived from Slavic languages in Balkan recipient languages point to both ancient (*strǫga ‘milking stanchion’, *sěno ‘hay’, *koryto ‘trough’) and recent influences. The thematic spectrum of animal husbandry Slavisms is very wide. It includes denotations of wool by its quantity, appearance, and quality (e.g., *rudъ ‘type of wool’); physical features of animals, relevant for husbandry (*mьrša ‘lean, meager’, *slabъ ‘weak, ill’); gelding, breeding, and genitals of animals (*skopiti ‘to castrate’, *pьrčiti (sę) ‘to be in rut’); gender and age denotations (*šelegъ ‘two years old ram’, *pьrčь ‘billy-goat’); color denotations irrelevant for husbandry (*bělъ ‘white’, *šarьgavъ ‘colored’); animal rearing, nutrition, and their corresponding equipment (*gordjь ‘fence’, *kopana ‘trough’); socio-economic organization of a collective of herders (*čelьnikъ ‘owner of a large amount of animals’); buildings, parts and complexes of buildings (*lěsa ‘fence plank’, *stanъ ‘animal farm’; and economically relevant body parts (*kostь ‘bone’). All of this material unambiguously demonstrates a symbiotic co-existence among Slavic, Aromanian, Albanian, and Greek herder populations during the whole period of Slavic presence on the Balkan Peninsula.19 The distribution of
19 In a contrasting case, the phonetic features of genetically Albanian animal husbandry terms (*bardza ‘with white-colored body parts’, kuq ‘red-haired’, *laura ‘with white- or redcolored body parts’, ngjesh ‘with white- and red-colored body parts’, shkurt ‘short’, and *verk ‘nape’) indicate their fairly recent borrowing into Greek dialects (more rarely into Aromanian and Slavic). It is possible to link them to the direction of mass migrations prominent in the
332
Andrey N. Sobolev
Slavisms by number and area in various languages and dialects indicates that contacts were most intensive on the territory of Macedonia and South Albania. The area of West South Slavic (Serbian) influence on Albanian dialects in some cases clearly contrasts with the area of East South Slavic influence, while the innovative Central Balkan Slavic area contrasts with archaic areas. Greek, Albanian, and Aromanian Slavisms that are no longer present in contemporary Slavic languages and dialects in their corresponding meanings (e.g., *rastoka ‘a corridor for leading to the milking stanchion’) are particularly interesting. Finally, the series of inter-Slavic isoglosses like *kopana ~ *koryto ‘trough’, and the like distinctly continue into the present-day non-Slavic language territory, which indicates both the unity of the Slavic linguistic landscape in the past and the absence of barriers between segments of the polyphonic Balkan landscape as a whole (MABL 2009). In sum, Slavic donations to the languages of the Balkans happened to varying degrees, ranging from inconsequential (as in the lexis of traditional spiritual culture) to clearly dominant (as in animal husbandry terminology). But there is no evidence of a significant pan-Balkan influence in the fields of morphology or syntax.20
5 Balkan Slavic as a recipient The list of linguistic phenomena called “Balkanisms” that make Balkan Slavic part of the Balkan convergent group has been constantly under investigation since Kopitar and often generally qualified as “iconical”, “analytic”, “polysemantic” and “redundant” (Hetzer 1995; Hinrichs 1999a; Sobolev 2004a). The standard list in Hinrichs (1999a: 432, 434) includes the “non-labialized midvowel, reduced case system, locative numerals 11–19, postpositive article, vocative, analytical comparative, adnominal dative, object doubling, periphrastic
ethnic history of the Balkans as well as to the labor migrations of Albanians to the South, a resurgent movement at the end of 20th century. The concentration of Albanianisms in the area of color denotations, as well as other descriptions for external features of animals (irrelevant from the point of view of production, but important for the identification of individual animals in the herd during grazing) is most probably explained by the role of Albanians as hired herders. It might be supposed that our data point towards Albanian and Greek symbiosis, i.e., a difference in the distribution of social and economic roles between the Greek population and Albanian migrants. 20 “The Slavs have been a passive element in the process of syntactical and morphological evolution . . . There is just a single language category, where all Balkan languages would present Slavic elements” (Seliščev 1925: 51).
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
333
future tense with volo ‘to want’, substitution of infinitive, evidential mood, hypertrophic verbal system, some lexical items” (see Haarmann 1976: 77–96; Hetzer 1995: 26–27). This section concentrates on new data that lead, together with the well-known facts, to some new interpretations. Language data enable us to reconstruct the diverse roles played by ethnic groups in linguistic, cultural, social, and economic processes. In contrast to terms associated with certain economic or household practices, borrowings of terms associated with kinship and affinity should directly indicate some connection between different ethnic groups. Following this notion, the group of Aromanians in Pindus we have studied was obviously historically connected to the Slavs (see the Aromanian Slavisms lʹalā ‘uncle’, nveʹastā ‘daughter-in-law’, tʹetā ‘aunt’, tʹatā ‘father’), while the Muslim Slavs in Albania and the Albanians themselves were evidently more likely21 to be connected with Ottoman Turks (cf. Albanian turkisms baeb ~ babʹa ‘father’, daexh ~ dʹajo ‘uncle on mother’s sideʹ, and others in Sobolev 2004b; Sobolev and Novik 2013). Much attention has been given to the issue of the non-Slavic lexical fund in South Slavic languages (Hinrichs 1999b). It is useful to look here into lexical calques, that is, borrowing of internal forms of designation. Under the influence of Balkan languages, calquing kinship terms introduced certain lexical neutralizations into South Slavic languages, such as ‘generation’ ~ ‘belt’, ‘become a wife’ ~ ‘become a husband’, ‘man’ ~ ‘husband’, ‘fiancé’ ~ ‘son-in-law’, ‘male cousin’ and ‘nephew’ ~ ‘grandson’. The same Balkan influence causes the Slavic model ‘older’ ~ ‘younger’ (son, daughter, brother, sister, grandson, etc.) to be replaced by the scheme ‘big’ ~ ‘little’ (MABL 2005a). A marvelous example of a Greek (and perhaps Aromanian) calque can be seen in the West Macedonian lexemes prʹitatko and paratʹatko ‘stepfather’, prʹisin ‘stepson’, prʹikerka ‘stepdaughter’ (from Greek pʹarapatʹeras ‘stepfather’, Aromanian paratʹatā, parafičʹor). Calquing of pre-Christian Balkan calendar terms22 is probably evident in the name given to the Grandma March Day (the 1st or 14th of March). While the day’s designation in the Greek – Macedonian – Bulgarian area is motivated by mart ‘March’, a lexical item originating from a Latin source through Greek mediation (Greek prutumartjʹa, Macedonian bʹabʌ mʹartʌ, Bulgarian pʹȏrvʌ mʹartʌ), its name
21 See, however, Ylli (1997: 315) on Slavic kinship terms in Albanian: çejadë ‘family member’, dedo ‘grandfather’, pastërk ‘adopted son’, and teto ‘aunt’. 22 Calquing from Slavic to Balkan languages is not recorded in the sphere of traditional folk culture terms. Designations characteristic of some South Greek dialects that use lexemes with the inner form ‘beard’ to denote last ears in the field and part of the field intentionally left unreaped are supposedly Slavic in origin: brʹada ‘beard; last ears in the field’ > to mʹuśi (tu xorafx’ʹu) ‘beard (of the field)’.
334
Andrey N. Sobolev
in the West Balkan area is a lexeme inherent to each language that bears an inner form and motivation ‘summer’ (that is, for West South Slavic languages, it is possible to talk about a calque from Albanian): Albanian dʹita (e) vʹerës ‘day of summer’ > East Montenegrin proĺetńak, West Macedonian lʹetnik. To denote days of the Orthodox calendar, Slavic languages in the Balkans use an Orthodox Saint’s proper name (usually borrowed from Greek or through Greek mediation), calque a Greek syntagm, ‘saint X’, form them with a possessive construction, ‘X’s day’, or form a suffixal -ica-derivative (see East Macedonian svetʹa pʹetkʌ ~ pʹetkọvden ~ petkʹovica ‘Saint Friday’).23 In several cases, the ‘(Saint) X’ model is more persistent in western and/or central parts of the Balkans (Montenegro, East Serbia, West Macedonia) and usually has direct parallels in Greek, Albanian, and Aromanian, while it is opposed in East Macedonia and Bulgaria by a Slavic possessive model ‘X’s Day’ (compare Serbian svʹeti ĭʹovan ~ Macedonian ivʹan(ọv)den, Macedonian svʹeti tanʹasia ~ Macedonian ʌtʌnʹasọvden, Serbian sveti dźordźija ~ Macedonian g’ʹurg’evden, Macedonian svʹeti mʹitriĭa ~ Macedonian mʹitrovden). In the Central Balkan area, genetically diverse languages and cultures are so intensely blended that calendar terms are sometimes absolutely pan-Balkan and identical in form and ritual function (compare Macedonian svʹeti tanʹasia ‘(Saint) Athanasios’, Albanian shejtanʹas, Greek aithanʹas, Aromanian ʹaγ’u θanʹasi; Macedonian svʹeti mʹitriĭa ‘(Saint) Dimitrios’, Albanian shënmʹitër, Greek and Aromanian aĭdimʹitr) (Sobolev 2006b).
23 Muslim influence on terms in Balkan spiritual cultures differs from Christian influence in being accomplished through direct borrowing from the main Muslim Balkan donor language, that is Turkish (which may, in turn, serve as a mediator of lexis derived from Arabic, Iranic, or other languages). For instance, the traditional calendar of Rhodopian Muslim Slavs includes Ramazan baĭryam, Kurban baĭryam or Koch baĭryam, Kadır gedzhe or Kadır Avshımı, Nevruz, Edrelez, Kasım, Redzhep, Shabat, Ramazan months, Yuch aĭlar or Khairnite dene fast, two days before Baĭrams – Golyamo and Malko Arfe, the fortieth day after Kurban Baĭryam – Ashchura. Many of these elements are also, for example, observed among Muslim Slavs in Bosnia and Albanian Muslims in the Skrapar and Dibër regions (Popov 1994: 114–116; Ylli and Sobolev 2002: 391–416; 2003: 420–427). These data may be compared with the rather ordinary recoding of proper nouns from Greek to Bulgarian (Θόδωρας > Božidar, Πέτρος > Kamen), from Aromanian to Macedonian (Floria > Cvetanka) (Sedakova 2007: 108), from Aromanian to Albanian (Buna > Mira), which seems to be characteristic primarily of (Orthodox) Christians. Krasimira Koleva from Shumen University communicated personally that Bulgarian Muslims prefer direct borrowing to re-coding of Turkish and (mostly) Arabic names, although they are perfectly aware of their internal form (Khasan ‘beautiful’, Aĭshe ‘prosperous’, Dzhemile ‘beauty’, Aĭsun ‘light’, etc.). To be fair, we should emphasize that re-coding occurs here as well, as in Sevda and Lyubov, with the inner form ‘love’.
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
335
Calquing a Greek inner form also produced such terms of traditional spiritual culture as Macedonian čʹisti ponʹedełnik < Greek i katharʹa deftʹera lit. ‘clean Monday’; Macedonian velʹikʌtʌ nedʹel’ʌ < Greek megʹalē evdomʹada lit. ‘big week’; Macedonian vʹeli-čʹetvǝrtok < Greek megʹalē pʹemtē lit. ‘big Thursday’; Macedonian velʹipetok < Greek i megʹalē paraskeuʹē lit. ‘big Fridayʹ; Bulgarian peddesʹetnicʌ < Greek ts penʹēnda ts paskaliʹas lit. ‘fifty of Easterʹ; Macedonian svetʹi duh < Greek ‘agio pnʹeuma lit. ‘Saint Spirit’, i.e., ‘Holy Spirit’. In cases where the Greek language utilizes, or has utilized, two or more denomination models (due to the differences among local Greek traditions and the chronology of their emergence), their reflexes have been fixed in various parts of the South Slavic area, generating corresponding isoglosses. Examples can be found in the cases of Slavic *vъskrьsъ ‘resurrection’ in the Serbian area vs. *velikъ dьnь ‘big Day’ in the East Serbian, Macedonian, and Bulgarian areas to denote Easter; Macedonian četʹiries mʹučenici < Greek sarʹanta mʹartyres ‘forty Martyrs’ vs. Bulgarian sv’ẹt’ʹi čẹt’ʹir’is’ʹe < Greek οι ʹagioi sarʹantēs ‘Saint forty’, i.e., ‘Holy forty’; Macedonian velʹivdenska nʹedela ‘Easter week’ vs. Bulgarian svʹetłʌ nedʹel’ʌ lit. ‘bright week’ < Greek deutʹera tēs lamprʹēs, trʹitē tēs lamprʹēs ‘Monday, Tuesday of Lighting.ʹ Traces of Romance influence have formed extremely important isoglosses that also serve to demarcate Balkan linguistic and cultural space. The Greek model for denoting the last week (and the last day) before the Lenten fast (Greek ē tyrinʹē evdomʹada ‘cheese shrovetide’, ‘cheese week’ > Macedonian sʹirnʌ nedʹel’ʌ, Bulgarian s’ʹirn’i zʹagọv’ẹzn’i) is opposed to the Romance model ‘white shrovetide’, ‘white week’ (Aromanian siptumʹāna ‘albā > Macedonian bʹeła nʹedela, Serbian bijeli pokladi). Each has its distinct distribution area. The receptivity of Slavic in grammar in addition to the already mentioned well-known phenomena can be demonstrated by pointing to the reconstruction of the functional-semantic field of Slavic instrumental under the influence of the pan-Balkan (originally Greek) system, with its expanded patientivity, whence the so-called “accusativus graecus”. In Balkan Slavic dialects that experience close contact with Greek, this process affects the semantic and functional core of the whole category (Sobolev 2006a): 25) Rhodopian Bulgarian z'ȏ dʌ mǝ m'ax-ʌ rʌk’'i-tǝ take.AOR.3SG PART.CONJ to.me wave-PRS.3SG hand.PL.-DEF ‘He/she started to wave his/her hands at me.’ (Sobolev 2001)
336
Andrey N. Sobolev
The maps in (MABL 2005b) align with the results of dialectological research (see, for example, Topolinjska 1995–1997) to demonstrate the evident nonhomogeneous geographic distribution of accusative constructions among Slavic dialects from East to West, a fact that fully represents their remoteness from the center of the most intensive interactions among Balkan languages in Southern Albania, Northern Greece, and Macedonia, along with corresponding differences in their degree of balkanization. The reconstruction of the Slavic instrumental sphere in the overall Balkan context complies with the rules for the hierarchy of case meanings that affect parts of the core as well as the periphery of the corresponding functionalsemantic field (Sobolev 2009) and leads to the division of previously undivided meanings and to the creation of new structural oppositions, previously unknown to the Slavic linguistic system, under the simultaneous influence of the Greek model’s patientivity (accusativity)24 field and the Balkan Romance model’s agentivity (ablativity) field: 26)
West Macedonian se='iskap-if od ml'eko-to REFL=spill-AOR.3SG of milk.N.SG-DEF ‘I spilled milk all over myself.’
In general, we find almost no evidence of restrictions that would keep any of the (Balkan) Slavic languages from borrowing and adopting foreign (in our case, Balkan) substance and functions.
6 Susceptibility of typically Slavic categories to borrowing According to Rusakov (2004), regularities emerge from the examination of contact-induced change in Balkan Slavic languages from the point of view of the degree of resistance or penetrability associated with various areas of their linguistic structure. In the area of grammar, this examination confirms assumptions about which factors facilitate foreign influence. The more a phenomenon complies with the following parameters, the more likely it is to generate interference (Harris and Campbell 1995: 122–136):
24 It seems very convenient to name these fields with terms denoting central semantic roles (agent, patient, instrument).
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
337
a. the formal expression of meaning is clear and simple (auxiliary word > agglutinative affix > flexion); b. the meaning is clear and monosemantic; c. the phenomenon appears in discourse-generating and emotive spheres of language; d. there is a structural affinity in a particular fragment of the grammatical systems of two languages. When we consider contact-induced grammatical change in Balkan Slavic, it becomes obvious, for example, that the noun system (with declension loss, rise of the definiteness and object doubling) is more penetrable than the verbal system. In the latter, moreover, the modality sphere is more penetrable than the tense-aspect sphere. The presence of aspect as a basic category in all Slavic languages is therefore especially interesting. Indeed, this category has not disappeared from any Slavic language, evidently because the Slavic type of aspect is marked by extreme formal complexity and lack of transparency in its expression and by tightly intertwined grammatical and lexical characteristics in the semantics of concrete lexemes, and so resists interference from languages that do not have this category. When it comes to aspect, Slavic has been neither a donor nor a recipient in the Balkans. There is a striking parallel in the field of lexis, more precisely among kinship terms. Let us consider, for example, a semantic microfield composed of denotations for brothers and sisters of parents, together with their spouses. In this area, Slavic languages are neither donors nor recipients. In the South Slavic settlements that have been studied, the Slavic lexical archaisms *stryjь ‘uncle on father’s side (father’s brother)’, *stryjьna ‘aunt on father’s side (father’s sister)’, *ujь ‘uncle on mother’s side (mother’s brother)’ and *ujьna ‘aunt on mother’s side (mother’s sister)’ are well preserved. This microfield is not unique, as we see from a similar state of affairs with the lexemes *bratučędъ ‘cousin’, *děverь ‘husband’s brother’, *šurъ ‘wife’s brother’, *zъly ‘husband’s sister’, *jętry ‘wife’s sister’, among others (MABL 2006). Not only have these denotations been preserved by South Slavs, they have not been borrowed by non-Slavic Balkan languages with less differentiated kinship systems that display a large number of structural neutralizations. Based on these examples from grammar and lexicon, we can assume that at least some prototypical structures result in strong borrowability restrictions. It is possible to compile an ostensive list of such categories of South Slavic languages, which can be labeled as “antibalkanisms” (Sobolev 2011; for Albanian, see Rusakov 2013). Among these categories, I can name the following:
338
Andrey N. Sobolev
(1) phonological enclitics; stress shift on proclitics (as in Bulgarian bʹez=žena ‘without a wife’); tonic stress, inherent to a Slavic lexical item (Gvozdanović 2009: 101, 111), but paradigmatic for inflection; phonological relevance of vocalic quantity; phonological relevance of vocalic nasality; syllabic sonants, constituting a subclass at the border between vowels and consonants; the sonant status of /v/; (2) the category of peripheral case forms as opposed to structural cases (for instrumental, see Miloradović 2003; MABL 2005; Sobolev 2009a); the category of animateness and personness; the category of collectivity, with contradictory plural semantics and singular form; the category of dualis; opposing “short” and “long” forms of adjectives with unclear intrasystemic functions; the class of possessive adjectives and pronouns (see Ivanov 1989: 21–28); (3) absence of any categorial marker for definiteness on any member of the nominal group, that is, of an explicit marker for individualizing, generic, specific, or indefinite meaning (see MABL 2005b: Maps 93–126):25 27)
Macedonian Toj dojde vo (edna) bela košulja. he come.AOR in (one) white shirt ‘He came wearing a/the white shirt.’
28)
Macedonian Toj dojde vo bela-ta košulja. he come.AOR in white-DEF shirt ‘He came wearing a/the white shirt.’
(4) the category of verbal aspect with the admittedly vague general meaning of terminativity, expressed by a root morpheme or a suffix (the Serbian imperfective infinitive pada-ti vs. the perfective infinitive pas-ti ‘to fall down’; Bulgarian čʹist’-a clean[IPFV]-PRS.1SG ‘I clean’ vs. plat’-ʹa pay.[PFV]-PRS.1SG ‘I pay’); synthetic forms of future tense, expressed through present indicative of perfective verbs (izbere in the example to follow refers to future): 29)
Bulgarian Toj šte si kupi knigata, kojato si izber-e. which REFL choose[PFV]-PRS.3SG he will REFL buy book ‘He will buy a book of his choice.’
25 This new category, expressed through an article morpheme with individualizing meaning, emerged historically on adjectives in attributive syntagma (Sobolev 2009a, 2009b; see Mladenova 2007).
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
339
(5) forms of infinitive (and supine) like Rhodopi Bulgarian Nǝ štʹȇš vʹig’ȇ ‘You will not see’ (see Miklosich 1861; Cyxun 1981). Possible reasons why these phenomena rank low on the scale of contactinduced influence might include their inherence in lexical items, relatively unclear intersystemic distribution, function, semantics and others.
7 Conclusion The genetic and areal-typological analysis of Balkan Slavic languages reveals divergent and convergent developments, which can be interpreted against the background of comparative-historical theory, geolinguistic theory, language contact and Sprachbund theory, including dialectology of convergent linguistic groups. General mechanisms of genetic splitting and typological merging, borrowing and calquing, language shift, and language and ethnic separation interplayed to make this an extremely peculiar area of Europe and Eurasia. We can conclude that South Slavic, on the one hand, was delivered to the Balkan peninsula disintegrated in at least two genetically opposed subgroups (so-called West South Slavic and East South Slavic), while on the other hand, it became part of the Balkan linguistic landscape irrespective of its primary genetic subdivision. Due to multilingualism the Balkan linguistic landscape can be viewed as uninterrupted continuum of closely and distantly related dialects (languages) shaped of isoglosses that run irrespective of “language borders”. Fitting into the landscape, the Slavic languages followed its constant geolinguistic subdivision in West and East. Among the results of integration in the history of Slavic in the Balkans neither do we observe hybridization, nor creolization. Together with the Balkan Sprachbund features, (Balkan) Slavic seems to be able to integrate almost any foreign phenomenon, while the degree and depth of integration often depend on the contact setting. Yet we observe evident internal linguistic constant characteristics of Slavic languages like aspect that do not depend on any historical, cultural, and/or contact settings and do not propagate beyond Slavic itself, behaving like barriers to language integration. These constant characteristics of Slavic languages perhaps make it easier for non-Slavic speakers to acquire Slavic as a whole communication code and switch completely, over time, to (Balkan) Slavic, than to integrate prototypical Slavic structural phenomena into another language system.
340
Andrey N. Sobolev
West versus East Balkan dialects
Καστέλλι
Legend South Slavic dialects Albanian dialects Greek dialects Aromanian dialect The West Balkan dialects are marked black. Map 1: The Balkan Peninsula
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
341
References Asenova, Petja. [1989] 2002. Balkansko ezikoznanie. Osnovni problemi na balkanskija ezikov săjuz [Balkan Lingustics. Fundamental Problems of the Balkan Sprachbund]. Sofia: Faber. Asenova, Petja (ed.). 2003. Aktualni problemi na balkanskoto ezikoznanie. Aspekti na izsledvaneto na obščobalkanskata leksika. Dokladi ot meždunarodna naučna konferencija (Sofia, 30.09 – 01. 10.2002) [Current Problems in Balkan Linguistics. Aspects of Research on the Shared Balkan Lexicon. Papers from an International Scholarly Conference]. Sofia: Faber. Belić, Aleksandar. 1905. Dijalekti istočne i južne Srbije [Dialects of Eastern and Southern Serbia]. Belgrade: Srpska akademija nauka. Birnbaum, Henrik. 1965. Balkanslavisch und Südslavisch. Zur Reichweite der Balkanismen im südslavischen Sprachraum. Zeitschrift für Balkanologie 3: 12–63. Birnbaum, Henrik. 1975. Common Slavic. Progress and Problems in its Reconstruction. Cambridge, MA: Slavica Publishers. Bratanić, Branimir. 1951. Uz problem doseljenja južnih Slovena. Nekoliko etnografsko-leksičkih činjenica [On the problem of South Slav settlement. Some ethnographic and lexical facts]. In: Zbornik radova, Vol. 1, 221–250. Zagreb: Filozofski fakultet, Sveučilište u Zagrebu. Budziszewska, Wanda. 1991. Zapożyczenia słowiańskie w dialektach nowogreckich [Slavic Borrowings in Modern Greek Dialects]. Warsaw: Instytut Slawistyki PAN. Campbell, Lyle. 2006. Areal linguistics: a closer scrutiny. In: Yaron Matras et al. (eds), Linguistic Areas: Convergence in Historical and Typological Perspective, 1–31. New York: Palgrave Macmillan. Croft, William. 2000. Explaining Language Change. Harlow: Longman. Cyxun, Gennadij Afanas’evič. 1981. Tipologičeskie problemy balkanoslavjanskogo areala [Problems of the Typology in the Balkan Slavic Area]. Minsk: Nauka i texnika. Gvozdanović, Jadranka. 2009. Melodischer Akzent (Flexion). In: Karl Gutschmidt et al. (eds.), The Slavic Languages. An International Handbook of their Structure, their History and their Investigation, Vol. 1, 100–114. (Handbooks of Linguistics and Communication Science 32.1) Berlin: Mouton de Gruyter. Haarmann, Harald. 1976. Aspekte der Arealtypologie. Die Problematik der europäischen Sprachbünde. Tübingen: Narr. Harris, Alice Carmichael and Lyle Campbell. 1995. Historical Syntax in Cross-Linguistic Perspective. Cambridge: Cambridge University Press. Hetzer, Armin. 1995. Nominalisierung und verbale Einbettung in Varietäten des Albanischen. Wiesbaden: Harrassowitz. Hinrichs, Uwe. 1999a. Die sogenannten ‘Balkanismen’ als Problem der SüdosteuropaLinguistik und der allgemeinen Sprachwissenschaft. In: Uwe Hinrichs (ed.), Handbuch der Südosteuropa-Linguistik, 429–461. Wiesbaden: Harrassowitz. Hinrichs, Uwe (ed.). 1999b. Handbuch der Südosteuropa-Linguistik. Wiesbaden: Harrassowitz. Ivanov, Vjačeslav Vsevolodovič. 1989. Sinxronnaja i diaxroničeskaja tipologija posessivnosti [Synchronic and Diachronic Typology of Possessivity]. In: Vyačeslav Vsevolodovič Ivanov (ed.), Kategorija posessivnosti v slavjanskix i balkanskix jazykax, 5–43. Moscow: Nauka. Ivić, Pavle. [1958] 1991. Značaj lingvističke geografije za uporedno i istorijsko proučavanje južnoslovenskih jezika i njihovih odnosa prema ostalim slovenskim jezicima [Significance of linguistic geography for the comparative and historical study of South
342
Andrey N. Sobolev
Slavic languages and their relationship with other Slavic languages]. In: Pavle Ivić, Izabrani ogledi. O slovenskim jezicima i dijalektima, 167–199. Niš: Prosveta. Ivić, Pavle and Dalibor Brozović. 1988. Jezik, srpskohrvatski / hrvatskosrpski, hrvatski ili srpski [Language, Serbo-Croatian / Croato-Serbian, Croatian or, Serbian]. Zagreb: Jugoslavenski leksikografski zavod “Miroslav Krleža”. Joseph, Brian. 2010. Language contact in the Balkans. In: Raymond Hickey (ed.), The Handbook of Language Contact, 618–633. Malden, MA: Wiley-Blackwell. Kočev, Ivan (ed.). 2001. Bălgarski dialekten atlas. Obobštavašt tom. 1–3. Fonetika. Aktsentologija. Leksika [Atlas of the Bulgarian Dialects. Collected Volumes 1–3. Phonetics, Accentology, Lexicon]. Sofia: Bălgarska akademija na naukite. Kurkina, Ljubov’ Viktorovna. 1992. Dialektnaja struktura praslavjanskogo jazyka po dannym južnoslavjanskoj leksiki [The Dialect Structure of Proto-Slavic on the Basis of the South Slavic Lexicon]. Ljubljana: Slovenska akademija znanosti in umetnosti. Loma, Aleksandar. 1994. Jezička prošlost jugoistočne Srbije u svetlu toponomastike [The linguistic past of Southeastern Serbia in the light of toponomastics]. In: Pavle Ivić (ed.), Govori prizrensko-timočke oblasti i susednih dijalekata. Zbornik radova sa naučnog skupa (Niška Banja, juna 1992), 107–136. Niš: Srpska akademija nauka i umetnosti. MABL. 2003. – Andrej Nikolaevič Sobolev (ed.), Malyj dialektologičeskij atlas balkanskix jazykov. Probnyj vypusk [A Minor Dialect Atlas of the Balkan Languages. A Trial Issue]. Munich: Otto Sagner. MABL. 2005a. – Andrej Nikolaevič Sobolev (ed.), Malyj dialektologičeskij atlas balkanskix jazykov. Serija leksičeskaja, Vol. 1: Leksika duxovnoj kul’tury [A Minor Dialect Atlas of the Balkan Languages. The Lexicon Series. Vol. 1. The Vocabulary of Spiritual Culture]. Munich: Otto Sagner. MABL. 2005b. – Andrej Nikolaevič Sobolev (ed.), Malyj dialektologičeskij atlas balkanskix jazykov. Serija grammatičeskaja, Vol. 1: Kategorii imeni suščestvitel’nogo [A Minor Dialect Atlas of the Balkan Languages. The Grammar Series. Vol. 1: Noun Categories]. Munich: Otto Sagner. MABL. 2006. – Andrej Nikolaevič Sobolev (ed.), Malyj dialektologičeskij atlas balkanskix jazykov. Serija leksičeskaja, Vol. 2: Čelovek. Sem’ja [A Minor Dialect Atlas of the Balkan Languages. The Lexicon Series. Vol. 2: Human Being. Family]. Munich: Otto Sagner. MABL. 2009. – Andrej Nikolaevič Sobolev (ed.), Malyj dialektologičeskij atlas balkanskix jazykov. Serija leksicheskaja, Vol. 3: Životnovodstvo [A Minor Dialect Atlas of the Balkan Languages. The Lexicon Series. Vol. 3: Animal Husbandry]. Sankt-Peterburg: Nauka / Munich: Otto Sagner. MABL. 2013. – Andrej Nikolaevič Sobolev (ed.), Malyj dialektologičeskij atlas balkanskix jazykov. Serija leksicheskaja, Vol. 7: Polevodstvo. Ogorodničestvo [A Minor Dialect Atlas of the Balkan Languages. The Lexicon Series. Vol. 7: Field Crop Production. Horticultura. Sankt-Peterburg: Nauka / Munich: Otto Sagner. Miklosich, Franz. 1861. Die slavischen Elemente im Rumänischen. Vienna: Österreichische Akademie der Wissenschaften. Miloradović, Sofija. 2003. Upotreba padežnih oblika u govoru Paraćinskog Pomoravlja: balkanistički i etnomigratsioni aspekt [The Use of Case Forms in the Paraćin Pomoravlje Dialect]. Belgrade: Srpska akademija nauka i umetnosti. Mladenova, Olga. 2007. Definiteness in Bulgarian. Modelling the Processes of Language Change. Berlin/New York: Mouton de Gruyter.
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
343
Popov, Račko. 1994. Kalendarni praznitsi i običai [Calendrical holidays and customs]. In: Račko Popov and Slavka Grebenarova (eds.), Rodopi. Tradicionalna narodna kultura i socialnonormativna kultura, 82–117. Sofia: Bălgarska akademija na naukite. Popović, Ivan. 1960. Geschichte der serbokroatischen Sprache. Wiesbaden: Otto Harrassowitz. Rusakov, Aleksandr Jur’evič. 2004. Interferencija i pereključenie kodov [Interference and Code Switching]. Sankt-Peterburg: Leningradskij gosudarstvennyj universitet. Rusakov, Aleksandr Jur’evič. 2013. Nekotorye izoglossy na albanskoj dialektnoj karte (k voprosu o vozniknovenii i rasprostranenii balkanizmov albanskogo jazyka [Some isoglosses on the Albanian dialect map (On the question of the emergence and diffusion of Balkanisms in Albanian]. In: Vjačeslav Vsevolodovič Ivanov (ed.), Issledovanija po tipologii slavjanskix, baltijskix i balkanskix jazykov, 113–174. Sankt-Peterburg: Aletejja. Rusakov, Aleksandr Jur’evič and Andrey Nikolaevič Sobolev. 2008. Substancial’no-funkcional’naja teorija balkanskogo jazykovogo sojuza i slavjanskie jazyki [The Substantial and Functional Theory of the Balkan Sprachbund and the Slavic Languages]. Sankt-Peterburg: Nauka. Sandfeld, Kristian. 1930. Linguistique balkanique. Problèmes et résultats. Paris: Édouard Champion. Schuster-Šewc, Heinz. 2014. Der Zerfall der slavischen Spracheinheit (des Urslavischen) und die genetische Gliederung der slavischen Sprachen. In: Karl Gutschmidt et al. (eds.), The Slavic Languages. An International Handbook of their Structure, their History and their Investigation, Vol. 2, 1153–1165. (Handbooks of Linguistics and Communication Science 32.2.) Berlin: Mouton de Gruyter. Seliščev, Afanasij Matveevič. 1925. Des traits linguistiques communs aux langues balkaniques: Un balkanisme ancien en bulgare. Revue des études slaves 5 (1/2): 38–57. Sobolev, Andrej Nikolaevič. 1998. Sprachatlas Ostserbiens und Westbulgariens., Vol. 1: Problemstellung, Materialien und Kommentare, Kartenanalyse, Vol. 2: Karten, Vol. 3: Texte. Marburg: Biblion. Sobolev, Andrej Nikolaevič. 2001. Bolgarskij širokolykskij govor [The Bulgarian Široka Lăka Dialect]. Marburg: Biblion. Sobolev, Andrej Nikolaevič. 2003. Južnoslavjanskie jazyki v balkanskom areale [South Slavic in the Balkan Area]. Marburg: Institut für Slavische Philologie. Sobolev, Andrej Nikolaevič. 2004a. On the areal distribution of syntactical properties in the languages of the Balkans. In: Olga Mišeska-Tomić (ed.), Balkan Syntax and Semantics, 59–100. (Linguistik Aktuell/Linguistics Today 67.) Amsterdam/Philadelphia: John Benjamins. Sobolev, Andrej Nikolaevič. 2004b. Opyty issledovanija tjurkizmov v balkanskix dialektax [On research of Turkisms in the Balkan dialects]. Zeitschrift für Balkanologie 40 (1): 61–69, (2): 206–229. Sobolev, Andrej Nikolaevič. 2006a. Južnoslavjanskij instrumental i ego balkanskie ekvivalenty [The South Slavic instrumental and its Balkan equivalents]. Makedonska akademija na naukite i umetnostite. Prilozi 31 (2): 59–71. Sobolev, Andrej Nikolaevič. 2006b. Ob obščebalkanskoj leksike duxovnoj kul’tury: nepodvižnye prazdniki narodnogo kalendarja [On the Common Balkan lexicon of spiritual culture: Immovable holidays of the national calendar]. Probleme de filologie slavǎ 14: 165–173. Sobolev, Andrej Nikolaevič. 2009a. From synthetic to analytic case: variation in South Slavic dialects. In: Andrej Malchukov and Andrew Spencer (eds.), The Oxford Handbook of Case, 716–729. Oxford: Oxford University Press.
344
Andrey N. Sobolev
Sobolev, Andrej Nikolaevič. 2009b. O kategorii opredelennosti v bolgarskom jazyke (po povodu knigi Olgi Mladenovoj Definiteness in Bulgarian. Modelling the Processes of Language Change. Berlin/New York, 2007) [On the category of definiteness in Bulgarian (on the occasion of Olga Mladenova’s Definiteness in Bulgarian. Modelling the Processes of Language Change, Berlin/New York, 2007)]. Zeitschrift für Balkanologie 45 (2): 228–252. Sobolev, Andrej Nikolaevič. 2011. Antibalkanizmy [Antibalkanisms]. Južnoslovenski filolog 67: 185–195. Sobolev, Andrej Nikolaevič. 2012. Slavische Lehnwörter in albanischen Dialekten. In: Bardhyl Demiraj (ed.), Aktuelle Fragestellungen und Zukunftsperspektiven der Albanologie: Akten der 4. Deutsch-Albanischen Kulturwissenschaftlichen Tagung “50 Jahre Albanologie an der Ludwig-Maximilians-Universität Munich” (23. – 25. Juni 2011, Gut Schönwag bei Wessobrunn), 215–234. Munich: Harrassowitz. Sobolev, Andrej Nikolaevič. 2013. Osnovy lingvokul’turnoj antropogeografii Balkanskogo poluostrova [The Foundation of Linguo-Cultural Anthropogeography of the Balkan Peninsula], Vol. 1. Sankt-Peterburg: Nauka / Munich: Otto Sagner. Sobolev, Andrej Nikolaevič. 2014. Theoriebildung in der Dialektologie: historischvergleichende Beschreibung. In: Karl Gutschmidt, Sebastian Kempgen, Tilman Berger, and Peter Kosta (eds.), The Slavic Languages. An International Handbook of their Structure, their History and their Investigation, Vol. 2, 2067–2074. (Handbooks of Linguistics and Communication Science 32.2) Berlin: Mouton de Gruyter. Sobolev, Andrej Nikolaevič. 2015. Mrkovići (i Gorana): jazyki i dialekty Černogorskogo Primor’ja v kontekste novejšix balkanističeskix issledovanij [Mrkovići (and Gorana): languages and dialects of the Montenegrin coast in the context of recent Balkan linguistics research]. In: Bardhyl Demiraj (ed.), Akten der 5. Deutsch-Albanischen Kulturwissenschaftlichen Tagung. (Pogradec, Albanien. 6. – 7. Juni 2014.). Munich: Harrassowitz. Sobolev, Andrej Nikolaevič. 2016. Sravnitel’no-istoričeskoe i areal’no-tipologičeskoe izučenie balkanskix dialektov: aktual’nye voprosy teorii [Comparative-historical and arealtypological research of the Balkan dialects: key issues in theory]. Linguistique Balkanique 55: 53–74. Sobolev, Andrej Nikolaevič and Aleksandr Aleksandrovič Novik (eds.). 2013. Golo Bordo / Gollobordë, Albaniya. Materialy balkanskoi ekspedicii RAN i SPbGU 2008–2010 gg. [Golo Bordo / Gollobordë, Albania. Materials of the Balkan Expedition of the Russian Academy of Sciences and Sankt-Peterburg State University, 2008–2010]. Sankt-Peterburg: Nauka / Munich: Otto Sagner. Stojkov, Stojko. 1963. Osnovnoto dialektno delenie na bălgarski ezik [The foundation of the division of the Bulgarian dialects]. In: Slavjanska filologija, Vol. 3, 105–119. Sofia: Bălgarska akademija na naukite. Šuvalov, Petr Valer’evič. 1998. Proniknovenie slavjan na Balkany [The invasion of Slavs in the Balkans]. In: Agnija Vasil’evna Desnitskaja and Nikita Il’ič Tolstoj (eds.), Osnovy balkanskogo jazykoznanija. Jazyki balkanskogo regiona, Part 2: Slavjanskie jazyki, 5–28. Sankt-Peterburg: Nauka. Toxtas’ev, Sergej Remirovič. 1998. Drevnejšie svidetel’stva slavjanskogo jazyka na Balkanax [The oldest evidence of the Slavic language in the Balkans]. In: Agnija Vasil’evna Desnitskaja and Nikita Il’ič Tolstoj (eds.), Osnovy balkanskogo jazykoznanija. Jazyki balkanskogo regiona 2. Slavjanskie jazyki, 29–57. Sankt-Peterburg: Nauka.
10 Slavic dialects in the Balkans: Unified and diverse, recipient and donor
345
Tolstoj, Nikita Il’ič. 1997. Iz geografii slavjanskix slov [From the geography of Slavic words]. In: Nikita Il’ič Tolstoj, Izbrannye trudy, Vol. 1, 122–260. Moscow: Jazyki russkoj kul’tury. Tolstoj, Nikita Il’ič. 1999. O sootnošenii central’nogo i marginal’nogo arealov v sovremennoj Slavii [On the interrelation of the central and peripheral areas in contemporary Slavia]. In: Nikita Il’ič Tolstoj, Izbrannye trudy, Vol. 3, 10–29. Moscow: Jazyki russkoj kul’tury. Topolinjska, Zuzana. 1995–1997. Makedonskite diјalekti vo Egeјska Makedoniјa. Sintaksa. Vols. 1, 2. [Macedonian Dialects in the Aegean Part of Macedonia. Syntax.] Skopјe: Makedonska akademija na naukite i umetnostite. Trubetzkoy, Nikolaj Sergeevič. 1928. Proposition 16. In: Actes du premier Congrès international de linguists: à La Haye, 10–15 avril 1928, 17–18. Leiden: A.W. Sijthoff. Videnov, Mixail and Bojan Bajčev. 1999. Velikotărnovskijat ezik [The Language of Veliko Tărnovo]. Veliko Tărnovo: Abagar. Ylli, Xhelal. 1997. Das slavische Lehngut im Albanischen, Part 1: Lehnwörter. Munich: Otto Sagner. Ylli, Xhelal and Andrej Nikolaevič Sobolev. 2002. Albanskii toskskii govor sela Lešnja [The Tosk Albanian Dialect of the Village of Leshnja]. Marburg: Biblion. Ylli, Xhelal and Andrej Nikolaevič Sobolev. 2003. Albanskii gegskii govor sela Muhurr [The Gheg Albanian Dialect of the Village of Muhurr]. Munich: Biblion.
Andrii Danylenko
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms? 1 Introduction Areal relationship among the languages of Europe has been extensively discussed in recent literature (e.g., Heine and Kuteva 2006). Yet the process of areal diffusion, leading to contact-induced grammaticalization, has not yet received definitive explication. It is, at least, generally acknowledged that linguistic areas, either recognized or rather constructed by linguists, are notoriously messy and “what we understand about linguistic areas is depressingly meager” (Thomason 2001: 99; Campbell 2006: 1). Even the “oldest” notion of the Balkan Sprachbund is no exception, although it differs from the traditional, ethno-cultural interpretation of the Balkans. As a canonical construct in contact linguistics, the Balkan Sprachbund reveals extensive convergences on various features in a group of distantly related languages found in this area. Called since Seliščev (1925) “Balkanisms”, these features have attracted considerable attention in the rather large literature on Balkan linguistics (Schaller 1975, 96–10; Hinrichs 1989/1991; Kahl, Metzeltin, and Schaller 2012; Feuillet 1986, 2012). Paradoxically, some of the Balkanisms occasionally occur not only in the languages and dialects spoken in the Balkans but also in other closely or distantly related languages found outside Southeastern Europe (Hetzer 2010: 463; Joseph 2010: 623). However, among the “non-Balkan” languages, one can hardly find (Southwest) Ukrainian, although since the publication of The Carpathian Dialect Atlas (CDA) in 1967, Southwest Ukrainian has been viewed in the Carpathian linguistic enterprise as a core language in the “CarpathianBalkan macrozone” (Klepikova 2004: 304; Gricenko [Hrycenko] 2008: 30). In fact, Carpathian linguistics, has been largely concerned with the “CarpathianBalkan convergences and direct correspondences” (Plotnikova 2008: 24), which are seen as marginal in Balkan linguistics. Depending in part on just how the Carpathian area and the Balkans are defined areal-typologically by the Note: Research on this work was partly financed by the Japan Society for the Promotion of Science and Dyson College of Pace University (New York) in June–August, 2013. I would like to thank Motoki Nomachi (Hokkaido University), Ronelle Alexander (University of California, Berkeley), Vladimir Žobov (Sofia University), George Thomas (McMaster University) for their valuable comments in preparing this chapter. https://doi.org/10.1515/9783110639223-012
348
Andrii Danylenko
representatives of Carpathian or Balkan linguistics, some convergences may be called either Balkanisms or Carpathianisms The first attempt at reconciling the Balkan and Carpathian perspectives was undertaken as early as 1967 by the Soviet scholar Samuil Bernštejn (1967) and his followers. In one of his programmatic articles, this scholar (Bernštejn 1972: 13; Nimčuk 1988: 305–308) maintained that the existence of shared features in the Balkans and the Carpathian areas provided enough evidence for postulating the existence of a Carpathian (northern periphery) type of the Balkan Sprachbund, a thesis that has never been seriously considered in Balkan linguistics. One can tentatively explain this lacuna by the fact that the Carpathian features are primarily identifiable in the vocabulary and partly in phonology (Popova 2005), while Balkanisms, with rare exceptions, are commonly found in the morphosyntax and, to a lesser extent, in phonology (Priestly 2010: 281; Asenova 2012: 333). At first blush, it may be connected with the current conceptualization of linguistic areas which are implicitly associated with the lack of the so-called shared core vocabulary (Dahl 2001: 1457), whence a focal point on what was called by Matras and Sakel (2007) pattern rather than matter borrowing. Yet it is not clear, in this respect, why the Carpathian area is taken in scholarly literature to be mostly an ethno-cultural conglomerate, while the Balkan Sprachbund is perceived as a linguistic league due to a closed number of convergences on phonological and morphosyntactic features (Balkanisms). All in all, the aforementioned areal-typological and structural discrepancies in the interpretation of both Balkanisms and Carpathianisms warrant critical revision. In this chapter, I intend to explore the relations between the Balkanisms and Carpathianisms as both synchronic typological characteristics and historical convergences. First, I will offer a critical survey of some of the “classic” Balkanisms (Friedman 2008: 131) and their most popular common classifications. They have been under scrutiny since the times of Jernej Kopitar (1829) and Franz Miklosich (1861). This is why I will flesh out only some most significant morphosyntactic features which are commonly treated as Sprachbund-forming (Section 2). In Sections 3–3.3.2, I will hone in on the areal-typological profiling of Carpathianisms and Balkanisms as they are postulated for Southwest Ukrainian. All this will help me to ascertain with more precision the distributive parallelism of the aforementioned properties in the Carpathian-Balkan macroarea as propounded in Carpathian linguistics. Finally, in Section 4, I will advance an alternative – sociolinguistic-typological – explanation of the emergence of convergent features in the Balkan Sprachbund, on the one hand, and the Carpathian area, on the
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
349
other.1 The main question, which I intend to address in this paper, is whether there are grounds for speaking about a Carpathian-Balkan macroarea as viewed from the Carpathian and Balkan perspective? In order to tackle this question, I argue that the emergence of Sprachbundforming convergences is not immediately dependent on the areal diffusion of the pertinent features either via matter borrowing or pattern borrowing, or in other terms, via borrowing proper or replication (Heine and Kuteva 2005: 2–6). Moreover, neither matter nor pattern borrowing can explain why and how the corresponding change could have involved particular structures or elements in the entire Carpathian-Balkan macroarea. I also believe that a particular contact-induced change is not to be directly sought in the multiple causation brought about by a multilingual contact situation (Lindstedt 2000: 231; Joseph 2010: 625; Alexander 2012: 39). The latter theory, profitable as it may look, furnishes only a partial explanation to the aforementioned query. It remains unclear what exactly in a multilingual contact situation makes members of a particular speech community to resort to a matter or pattern borrowing from a language used within another community. As a working hypothesis, I venture to claim that the palliative solution might be sought not in a substratum or adstratum explanation or pidiginization effects (Joseph 1983, 2010: 624–625) but rather in a specific constellation of societal factors and their valuables as discussed in terms of the sociolinguistic typology of Trudgill (1997: 349–350) or the systemic typology of Mel’nikov (2003). Specific variables of these factors are likely to account for interrelated and concomitant changes at different levels of the language system (Plank 1998: 224). Tentatively, if particular language systems acquire identical valuables for the corresponding societal factors shaping the respective grammaticalization pathways, one is likely to observe parallel changes in the respective (multiple) languages in contact. In other words, the appearance of parallel Balkanisms and Carpathianisms within the Carpathian-Balkan linguistic macroarea might be linked to the development of similarities in the languages in contact due to a particular configuration of the pertinent societal factors (see Section 4).
1 I use the notions of “Sprachbund” and “area” on a par, irrespective of the differences in the conceptualization of these terms by different scholars (Wiemer and Wälchli 2012: 14). Both “Sprachbund” and “(macro)area” are treated here equally, thus referring to linguistic rather than ethno-linguistic groupings.
350
Andrii Danylenko
2 Balkanisms The Balkan Sprachbund is the oldest and most famous linguistic area that has been in the scholarly focus over the longest period of time. To use Heine and Kuteva’s terminology, its “uncontroversial” members (Albanian, Aromanian, Bulgarian, Modern Greek, Macedonian, Balkan Romani, Romanian, the southernmost and eastern dialects of Serbian, in particular the Prizren-Timok dialect group) (Stanišić 1985/1986) demonstrate a number of convergent properties (Balkanisms).2 One should add here Turkish – not a “full” structural participant but crucial nevertheless, according to Joseph (2010: 619). Suffice it to mention that there are many Turkish loanwords in the languages of the Balkans as a result of the Ottoman occupation of much of the area in the 14th – early 20th century. Contrary to Sandfeld (1930: 213), the role Greek has played in the linguistic Balkanization is smaller than one would expect from this Balkan language. Other neighboring languages may also be considered members of this Sprachbund, for instance Judezmo (also known as Ladino or Judeo-Spanish; maybe only at the phonological and lexical levels) (Joseph 2010: 619–620). The epicenter of Balkanisms, according to Lindstedt (2000: 234), seems to be somewhere south of the lakes Ohrid and Prespa where the Greek, Albanian, Macedonian, Aromanian, and Romani languages meet. Based on both areal factors and indices of Balkanization, Macedonian, Bulgarian, Albanian and Romanian (Daco-Romanian and Megleno-Romanian) are viewed, with rare exceptions, as central to the Balkan linguistic enterprise. Moreover, Balkan Slavic (Bulgarian, Macedonian, and the southernmost and eastern dialects of Serbian, including the Prizren-Timok and Torlakian dialect groups) appears to be “most Balkan”, that is, having most of the Balkanisms (Schaller 1975: 100; Lindstedt 2000: 234; Hetzer 2010: 457).3 The number of Balkanisms does not commonly exceed twenty in the literature on the Balkan languages (see Hinrichs 1999: 432–434; Joseph 2010: 621–623). The number of most significant features shared by various of these languages looks to depend on the delimitation of a linguistic group proper in the same way a language type or a language group is determined within a traditional philological approach (Wiemer and Wälchli 2012: 16). The rigid classification becomes, however, untenable inasmuch as sporadic occurrences of some of those features are observed in some Germanic and Baltic languages
2 Rather than Romanian one should speak here more accurately about Daco-Romanian and Megleno-Romanian. 3 Having constructed an isopleth map based on ten isoglosses reflecting “the most significant Balkanisms”, van der Auwera (1998: 262–263) contended that Bulgarian does happen to be included in all isoglosses but Macedonian and Tosk Albanian are only half-in for some of them.
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
351
(Schaller 2012: 10) as well as Slavic and Hungarian (Hinrichs 1989/1991). For instance, detecting some “traditional Balkanisms” (analytic properties) in English (and elsewhere in Western and Central Europe), Aronson (2007: 11–12) stated that they serve, though imperfectly, to delimit a geographical area, rather than a typological area. In fact, there is a circularity in the definition of the Balkan languages: they are Balkan because they are spoken in the Balkans.4 Some scholars advanced several solutions to the aforementioned conundrum. The simplest solution was elaborated on by Schaller (2012: 10–12, 1975: 101–103) who proposed to distinguish three groups of the Balkanisms: (1) “Primary Balkanisms” which are not paralleled in the languages outside the Balkans, e.g., the demise of infinitive and its replacement by a finite complement clause, and the formation of a future tense based on a reduced, often invariant, form of the verb ‘to want’. (2) Balkanisms whose parallels in other languages testify to their genetic (Indo-European) relatedness, e.g., analytic comparative adjective formations which are found, among other “non-Baltic” languages, in Russian, Lithuanian, and Latvian, and the formation of the “teen numerals” (between 11 and 19) which occurs, in particular, in Slavic and Latvian. (3) Balkanisms that are coincidently paralleled outside the Balkans and thus qualify to be typological correspondences. One of such properties is the use of an enclitic (postposed) definite article, typically occurring after the first word in the noun phrase, in the Balkan and Scandinavian languages. The proposed classification provides areal-typological parameters for the delimitation of the Balkan Sprachbund. However, it fails to adequately explain why Balkanisms are attested not only within but also outside the Balkan Sprachbund whose delimitation, in its turn, depends on the identification of such features. The indices of Balkanization calculated by Lindstedt (2000: 234) can hardly help in this case. First, the author employed a preset number of 4 Having noted circularity in the definition of Balkanisms, Sobolev (2011) offered, instead, a concept of “anti-Balkanism”. According to him, anti-Balkanisms are features like the periphrastic conditional, the predicative instrumental case, and the genitive of negation that are “hard to borrow or calque” (Sobolev 2011: 187, 188, 190–191). Showing their ambivalence toward analyticity and synthetism, economy and redundancy, and some other double parameters, such anti-Balkanisms belong to the Slavic, even Indo-European language type (Sobolev 2011: 191). Heuristic at its core, Sobolev’s suggestion is highly speculative since the inherent features of this type are determined rather intuitively. In short, this theory brings us back to the traditional problem of the borrowability of language features as investigated in Wiemer, Wälchli, and Hansen (2012).
352
Andrii Danylenko
twelve shared grammatical innovations within the Balkans. Second, Lindstedt premised his indices on the descriptive statistical approach that offers only a partial solution to the problem of the delimitation of the Balkan Sprachbund and other linguistic leagues in Europe (see Danylenko 2013). In order to break this type of “circularity”, Heine and Kuteva (2005: 182–183) advanced the notion of “grammaticalization area” conceived as a group of geographically contiguous languages that have undergone the same contactinduced grammaticalization, thus ruling out such factors as genetic relationship, drift, and chance. According to these authors, of the ten-plus properties commonly associated with the Balkan Sprachbund, the majority is suggestive of contact-induced grammaticalization; accordingly, with respect to these properties, one can speak of distinct grammaticalization areas. For the Balkan Sprachbund, Heine and Kuteva (2005: 188) accepted the following Balkanisms where (vi) is an additional feature to be found in all the “uncontroversial” Balkan languages: i. the formation of a periphrastic de-volitive future; ii. the pleonastic use of weak object pronominal forms in combination with full noun phrases; iii. the use of a suppressive marker for numerals between ‘11ʹ and ‘19ʹ, i.e., ‘1 on 10ʹ, ‘2 on 10ʹ and so forth; iv. the infinitive loss and its replacement with a finite complement clause; v. replacement of synthetic adjectival comparative forms with analytic ones; vi. the use of the quantifying interrogative ‘how much?’ in expressions for a degree marker (‘inasmuch as, insofar as’). According to Heine and Kuteva (2005: 182), the contribution that the study of grammaticalization can make to defining areal relationship is a modest one. First of all, it is concerned only with a limited spectrum of linguistic phenomena that do not encompass phonological and lexical ones. Leaving aside the “questionable significance” of grammaticalization areas for Balkan linguistics as discussed by Wiemer and Wälchli (2012: 17–18), there are some other shortcomings in the theory of Heine and Kuteva. First, the list of Balkanisms appears to be predetermined as these features are largely amenable to the ethno-linguistic account. Second, while avoiding the multitude of hypotheses that have been voiced concerning the diffusion of features in the Balkan Sprachbund, Heine and Kuteva’s approach fails, nevertheless, to offer other than areal criteria for determining what languages can be included in the grammaticalization area. In other words, it is still unclear at what areal point and under what “contact conditions” a particular grammatical replication would start. Even if contact-induced grammaticalization is a factor
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
353
that is relevant for understanding areal relationship (Heine and Kuteva 2005: 188), its relevancy looks largely subordinate to the idea of a mere areal diffusion process.
3 Balkanisms or Carpathianisms? The idea of areal diffusion has always remained pivotal in Balkan and Carpathian linguistics. In particular, it has been held responsible for the spread of Balkanisms in Slavic and other languages of Europe (Hinrichs 1989/1991). I will first outline the place of Southwest Ukrainian in the alleged diffusion processes as postulated in contemporary Balkan linguistics (Section 3.1), followed by a discussion of Carpathianisms (Section 3.2) and Carpathian Balkanisms as distinguished in Carpathian linguistics (Section 3.3).
3.1 Balkanisms in Southwest Ukrainian Before discussing Ukrainian in its relation to the Balkan Sprachbund, it is worth mentioning the Serbian-Croatian-Bosnian dialectal continuum in the context of the areal diffusion of Balkanisms. In view of its vicinity to the Balkan Sprachbund, this dialectal continuum is allegedly found to be at a crossroads of several linguistic groupings, one of which is the Carpathian area with Southwest Ukraine at its core (Hamp 1989: 44) and a Central European linguistic area (Thomas 2008; see Danylenko 2013: 137–140). With respect to Balkanisms and how they tended to diffuse over the adjacent territories, Thomas (2010) advanced a rather “reconciliatory theory” premised on the analysis of a set of five morphosyntactic features: (1) periphrastic future, (2) perfect as a sole preterit, (3) pluperfect, (4) prepositive definite article, and (5) prepositive indefinite article. Leaving aside the indefinite article – which is a special issue (Friedman 2003) – Thomas (2008: 140–141) tried to demonstrate that Balkan Sprachbund innovations did not reach Kajkavian, which shares many features with Slovene dialects, and Central European features stood well short of Torlak (the Prizren-Timok dialect group) spoken in the extreme southeast corner of the Serbian-Croatian-Bosnian continuum. On the other hand, Kajkavian and Torlak constitute bridgeheads into Serbian and Croatian territories for the Central European and Balkan features, respectively. Some – but not all – of these bridgeheads were built out into Štokavian, the true bridge between the two Sprachbünde (Thomas 2010: 382). According to the author, the large
354
Andrii Danylenko
movements of population caused by the Ottoman invasions and the opportunities opened by Habsburg victories brought about a mixing of dialects and the spread of Balkanisms. Undoubtedly, this not only fostered this spread but also favored resistance to Central Europeanisms (Thomas 2010: 384). Attractive as it may look, this scenario cannot explain how some of the Balkanisms managed to reach the Carpathian basin as part of the Central European Sprachbund, including its peripheral southwestern Ukrainian, Slovak, and southern Polish dialects (Plotnikova 2008: 10; Nimčuk 1993), where the corresponding features are called Carpathianisms (Rot 1967: 267; Bernštejn 1972: 10). With rare exceptions going back to the 1930s, Southwest Ukrainian has remained conspicuously irrelevant for the study of the Balkan Sprachbund.5 Pogorelov (1939) singled out, for instance, the following Bulgarian features in Transcarpathian Ukrainian: (1) possessive use of dative pronouns; (2) the use of the absolute (indeclinable) relativizer ščo ‘what’; (3) analytic comparative formations with the marker maj ‘more’ from *naj which also can occur with verbs (see Sections 3.1 and 3.3.1); (4) the use of prefix po- with a comparative diminutive meaning, e.g., SWU poteplyj ‘less warmer, lukewarm’; (5) reduplication of the comitative preposition sъ > sъs > zoz ‘with’ (see Pan’kevyč 1955: 211; Danylenko 2015: 279–283). Pan’kevyč (1955) agreed that some phonological and grammatical features did reach, primarily via Romanian, Southwest Ukrainian. However, unlike lexical borrowings from what he called the “Balkan circle”, most of the phonological and grammatical “Bulgarianisms” supported similar features in the local dialects like the medial l and enclitic forms of the dative like my ‘to me’, ty ‘to you’ (Pan’kevyč 1955: 238) or reduplication of the comitative preposition in zoz (Vasmer 1940/1941: 50; Danylenko 2015: 279–283).
5 In fact, the relevance of this material was taken for granted in both Slavic and Germanlanguage scholarship long before this date; for literature, see Kobyljans’kyj (1971: 420–422). Besides Pan’kevyč, Pogorelov’s theory was heavily criticized by Vasmer (1940/1941) who easily identified most of the so-called “Bulgarianisms” as Common Slavic. Subsequently, a few Ukrainian scholars showed their zeal in expanding possible connections between South Slavic and Southwest Ukrainian. Thus, Onyškevyč (1971) attempted at localizing features shared by Southwest Ukrainian and Bulgarian as far as the Bojkian group of dialects spoken in the central and western part of the Carpathian region; for discussion, see Nimčuk (1988: 297–301).
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
355
One finds no Ukrainian parallels in the latest book dealing with Balkanisms (Kahl 2012). Neither were they referred to in a programmatic article about the diffusion of Balkanisms in Slavic published by Hinrichs (1989/1991), although Jagić (1898) and especially Vasmer (1940/1941) were cognizant of the significance of the Ukrainian material. Instead of Southwest Ukrainian, Hinrichs (1989/1991: 52) made use of some Russian features which, to his mind, may be viewed as true Balkanisms: (1) the use of prepositions in analytic constructions of the type R prepodavatel’ po (on) jazykoznaniju (DAT.SG) ‘professor of linguistics’; (2) the use of a postpositive definite article, e.g., R dial. dom-ot ‘the house’; (3) the formation of the “teen numerals” like R odinnadcat’ ‘eleven’ as ‘one-onten’; (4) the pleonastic use of weak object pronominal forms together with full noun phrase, e.g., R kuda eё (it.ACC.F) postavit’ mašinu (car.ACC.F) ‘Where to park it, the car?’; (5) the possessive use of dative pronouns, e.g., R put’ emu (he.DAT) ležit ‘his trip is [. . .]’; (6) the formation of a future tense with the auxiliary ‘to be’; (7) analytic comparatives, e.g., R bolee (more) otkrytyj (honest.M.SG) ‘more honest’; (8) reduplication of words, e.g., bolee-menee ‘more or less’. The choice of Balkanisms in Hinrichs’ list looks controversial. On the one hand, the author believes in the irrelevance of the aforementioned Balkan features since they are likely to arise in practically any substandard Slavic language. Yet, on the other hand, he relies primarily on the analytic features as found in Bulgarian, Romanian, Albanian, and Greek (Hinrichs 1989/1991: 44–45). Consequently, one can take areal features to be genetic-typological since Balkanisms are likely to be potentially reproduced in all (substandard) Slavic languages. The latest enquiry into the diffusion of Balkan features into Southwest Ukrainian was made by Priestly (2010), who traced the use of Rm maj ‘more’ in its function as the first element in analytic comparative formations in the Ukrainian dialects spoken in Romania, e.g., Rm mai bun and U dial. maj dobryj ‘better’, a replacement for the synthetic form, now U lipšyj ‘better’. In particular, the author noted that the word maj in the analytic constructions occurs to the north of the international boundary, in Hucul dialects as far as Kolomyja. In some villages where the Slavic superlative marker naj co-exists with maj only, which supports a supposition that the prior existence of the superlative marker naj assisted in the spread of maj in Transcarpathian dialects (Priestly 2010: 272–273); the latter fact might explain the use of maj with superlative
356
Andrii Danylenko
forms as in maj bilšyj ‘best’, literally ‘more best’ in most of the Transcarpathian and Bukovyna dialects (AUM 2, maps 223, 224). Priestly’s study offers a glimpse into a far more serious problem pertaining to the alleged diffusion of Balkanisms into the Carpathian basin. Under the disguise of traditional terminology, the question arises as to what exactly one deals with in such cases. Is this a Balkanism, Carpathianism or a Carpathian Balkanism, if the corresponding feature was borrowed from one of the Balkan languages into Southwest Ukrainian, Slovak and southern Polish dialects? Priestly believes that maj is obviously a mere borrowing, but given its wide distribution it appears to be part of “some wider phenomenon” (Priestly 2010: 280), that is, the interrelations between the Balkan Sprachbund and the Carpathian linguistic area, in general (Onyškevyč 1971: 442). Without considering the use of maj in the analytic comparative/superlative formations as a Balkanism, Priestly (2010: 280–281) deems it necessary to stress that the literature on the Balkan Sprachbund is lengthy, while that on the Carpathian (and/or the Central European) one is less lengthy but still considerable. What is remarkable, he adds, is that the former exhibits a preponderance of syntactic features, whereas such features tend to often be subsidiary to phonological and lexical features in the latter (Priestly 2010: 281, see Gricenko [Hrycenko] 2008; Plotnikova 2008). In sum, there are linguistic grounds for admitting that Southwest Ukrainian, demonstrating some features shared with the Balkan languages, has “the right” to belong in the scope of Balkan linguistics.
3.2 Carpathianisms Similarities between the languages of the Carpathian area, encompassing both Slavic (Ukrainian, Polish, Slovak, and Moravian) and non-Slavic (Hungarian, some Romanian, and Romani) dialects and languages (Gricenko [Hrycenko] 2008: 32, Rot 1973: 248), and the Balkans have been in the focus of Carpathian linguistics since the late 1960s (Rot 1967, 1973; Bernštejn 1972; Kobyljanskij [Kobyljans’kyj] 1973; Klepikova 2004) and even earlier (see Kobyljans’kyj 1971: 430–432; Nimčuk 1988). Disregarded in the West, Carpathian linguistics is premised on two major tenets. First, both Carpathian and Balkan languages constitute a kind of “Carpathian-Balkan macrocontinuum” (macroarea) (Gricenko [Hrycenko] 2008: 27). Second, as a result of contacts between the dialects belonging to this continuum, one encounters numerous – primarily, lexical and grammatical – convergent features commonly called “Carpathianisms”. In view of the convergent features spread over the Carpathian-Balkan linguistic macroarea, Bernštejn (1972: 13) distinguished a Carpathian type of the
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
357
Balkan Sprachbund. Whence, purportedly, the existence of Carpathian-Balkan convergences in the phonology, morphosyntax, and the vocabulary (Rot 1967: 267). However, as is evidenced in The Common Carpathian Dialect Atlas (CCDA), the main emphasis in Carpathian linguistics has been placed on the study of lexical and derivational similarities within the Carpathian-Balkan linguistic macroarea. In fact, grammatical and phonological phenomena have largely remained beyond the scope of Carpathian linguistics. The latter fact is likely to explain why the Carpathian areal studies have not kindled thus far an interest of the Balkanists. Today, Carpathian linguistics is largely concerned with an “areal study of the cultural-linguistic tradition in the Carpathian-Balkan continuum” with an eye of ascertaining compact and juxtaposed areas in this zone (Plotnikova 2008: 13). Pursuing a complimentary goal in comparison with Balkan linguistics, this scholarly approach is particularly important for reconstructing the ethnic origins of Slavs and other peoples living in the Balkans. This is why the use of ethno-linguistic methods is aimed at the historical study of lexical borrowings as well as ways of the diffusion of both lexical and cultural phenomena within the Carpathian-Balkan linguistic macroarea (Klepikova 2003: 358; Plotnikova 2008: 13). Thus, the study of genetic and areal interrelations between Southwest Ukrainian and South Slavic has resulted in the formation of a separate linguistic discipline, Carpathian or Carpathian-Balkan linguistics (Gricenko [Hrycenko] 2008: 46–47). This discipline is designed to explain under what conditions the ethnic contacts took place within this linguistic area, as well as how these contacts influenced the structure and development of the dialects under consideration. Overall, Gricenko’s [Hrycenko] view is well harmonized with the fact that the Slavic ethnic element has always been present in the Carpathian-Balkan macroarea for more than 1,500 years, while interacting with neighboring peoples and their languages (Klepikova 2004: 305). This is why the analysis of cultural linguistic interrelations should be based primarily on the study of lexical isoglosses and their etymological interpretation (Gricenko [Hrycenko] 2008: 47; see Kobyljanskij [Kobyljans’kyj] 1973). As a result, despite declarations about the presence of Carpathianisms at all levels of the language system (Rot 1967: 267; Bernštejn 1976: 8; Popova 2005), morphosyntactic phenomena tend to remain beyond the scope of this discipline. The aforementioned methodological limitation becomes obvious in a dichotomy advanced by Bernštejn at the dawn of Carpathian linguistics. The author suggested to distinguish between Carpathianisms, on the one hand, and Balkanisms, on the other (Bernštejn 1972: 10). The first group of convergences contains “common particularities” that emerged in the Carpathian region under the influence of substratum. The second group is characterized by those features
358
Andrii Danylenko
(primarily from Bulgarian, Albanian, and Turkish) which appeared in this area as a result of a later colonization of the Carpathian basin by the newcomers from the Balkans (Kobyljans’kyj 1971: 433). Remarkably, Carpathianisms, according to Bernštejn (1972: 10), could be found at various levels of the language system, while Balkanisms reflected relatively recent lexical transfers. Moreover, such morphosyntactic Carpathianisms as the possessive use of dative (enclitic) pronouns or analytic comparative adjective/adverb formations and so forth were largely identical with the corresponding Balkanisms (Bernštejn 1976: 8). That was a declaration of the existence of hybrid “Carpathian Balkanisms” which would unite the two adjacent areas into one Sprachbund (Kobyljanskij [Kobyljans’kyj] 1973: 29). In hindsight, Bernštejn’s theory reveals some shortcomings. First, most of the Carpathian convergences are traced to the Common Slavic period – this might explain why they are attested in North Ukrainian and Belarusian likewise (Vasmer 1940/1941). Second, if Balkanisms are primarily lexical and derivational properties, it is highly problematic to speak about a certain Carpathian (northern or peripheral) type of the Balkan Sprachbund since the notion of linguistic area presupposes sharing structural features. If shared vocabulary by itself were enough to establish a linguistic area, then the entire world would be one huge linguistic area (Thomason 2000: 312). This is why using vocabulary as a sole criterion for determining the Carpathian area and its Sprachbund-forming features would trivialize the very concept of this area and its shared traits, whether they are identified as Balkanisms or Carpathianinisms, e.g., Ab vatra ‘fire’ which made its way into West and East Slavic, P watra, SWU vatra, R dial. vatruga (Desnickaja 1976: 20–21). Overall, with an emphasis on the ethno-linguistic aspects of the alleged Carpathian-Balkan linguistic area, the distinction between Carpathianisms and Balkanisms gets blurred, especially if the center of possible diffusion of the shared features is localized in the Carpathian basin (Klepikova 2004: 304). Since Slavic speakers appeared in the Balkans later, coming into contacts with the local Albanian and Greek population, the presence of Carpathianisms, now treated almost exclusively as non-grammatical convergences in Carpathian linguistics, seems to be at variance with the “local” layer of convergent features.
3.3 Carpathian Balkanisms? Unlike Bernštejn, Rot suggested to distinguish Carpathianisms as structural features shared by the languages spoken in the Carpathian basin only. According to him (Rot 1967: 248–249, 1973: 15–16), Bulgarian, Hungarian, Romanian and
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
359
Ukrainian came into multilateral contacts which propelled convergent processes in Southwest Ukrainian, Hungarian dialects of Transylvania and Transcarpathia, northern Romanian dialects, as well as northwestern and even Banat Bulgarian dialects. All in all, Rot (1973: 22–23) was the first to describe the linguistic interrelations and multilingual contacts in the Carpathian linguistic area, where Southwest Ukrainian played one of the most prominent roles. In addition to “regular” phonological and morphosyntactic Carpathianisms, these contacts triggered purportedly the diffusion of Ukrainianisms in bilingual situations like Ukrainian-Hungarian, although Rot (1967: 266) provided only lexical Ukrainianisms of the type H baraboj from U barabol’/barabolja ‘potato’. Among the morphological Carpathianisms, Rot (1973: 47) mentioned preterits of the type vjux ‘I lead’, pljux ‘I swam’ (see Danylenko 2012a: 16, 2012b) and the vocative desinence -o which allegedly spread from Ukrainian into Romanian and well beyond, e.g., U nevist-o (VOC) and Rm nevast-o (VOC) ‘woman’.6 A pioneering contribution to the study of Southwest Ukrainian in its relation to the Balkan Sprachbund was made by Nimčuk (1991, 1993) who tried to systematize all the known convergent features in Southwest Ukrainian and Balkan Slavic (see Kobyljans’kyj 1971; Desnickaja 1987: 10). In total, Nimčuk (1993: 50–59) mentioned 27 phonological, derivational, and morphosyntactic Balkanisms which are paralleled, according to him, in (Southwest) Ukrainian as a core member of the Carpathian linguistic area (Hamp 1989: 44; Gricenko [Hrycenko] 2008: 30). Leaving aside some minor phonological features, lexical transfers and some minor morphosyntactic borrowings, I will refer only to those phonological features (a, b, c) together with primary (d, e, f, g, h, i) and one secondary morphosyntactic (k) features in Nimčuk which are found in major Balkan languages and dialects (Joseph 2010: 621–623), see Table 1.
3.3.1 Phonological Carpathian Balkanisms The phonological Carpathian Balkanisms (a), (b), and (c) seem to resist both areal and typological interpretation. All of them tend to be discussed in genetic terms, thus representing convergent realizations of Common Slavic tendencies as is the case of Feature (a), see the rise of ў from CS ū in East Slavic, Polish, Sorbian, and
6 In this case one deals with a matter borrowing of the whole lexeme with a subsequent extraction of the vocative desinence -o as encountered in other forms like Rm draguț-o ‘my dear’.
360
Andrii Danylenko
Table 1: Common Ukrainian-Balkan features (after Nimčuk 1991, 1993). Feature
Ukrainian Dialects
Balkan Languages
a.
A mid-to-high central vowel (ў)
Transcarpathian, Lemkian dialects
Albanian, Romanian, Bulgarian, some dialects of Macedonian and Bosnian-Croatian-Serbian, some Romani dialects, and Turkish
b.
The change d’, t’ > g’, k’
Hucul, Bukovyna, some Bojkian and Dniester dialects
Bulgarian and Macedonian dialects
c.
The existence of medial l’
Transcarpathian, southeastern and some northern Ukrainian dialects
Romanian, Hungarian
d.
The use of dative enclitic pronouns
Transcarpathian, Lemkian, Bojkian, and some southeastern Ukrainian dialects
South Slavic and Greek
e.
Analytic comparative adjective/adverb formations
Transcarpathian, Lemkian, Bojkian and Hucul dialects
Bulgarian, Macedonian, Southeast Serbian (Prizren-Timok dialect group), Romanian, Greek, Albanian, Romani, and Turkish
f.
Location model ‘one on ten’ in the derivation of numerals from ‘ʹ to ‘’
Transcarpathian, some Dniester and western Volhynian dialects
South Slavic, Albanian, Aromanian, Megleno- and DacoRomanian
g.
The formation of a devolitive future tense
Transcarpathian and some southeastern Ukrainian dialects
Greek, Tosk Albanian, Romanian, South Slavic, including Southeast Serbian (Prizren-Timok dialect group) Romani
h.
The use of relativizer ščo ‘what’
Across Ukrainian dialects
Romanian, South Slavic, Albanian, Greek
i.
The merger of goal and location in de (< kъde)
Across Ukrainian dialects
Romanian, South Slavic, including Southeast Serbian (Prizren-Timok dialect group), Greek
k.
The use of the preposition za with verba dicendi and verba sentiendi
Across Ukrainian dialects
Bosnian-Croatian-Serbian
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
361
Table 1 (continued ) Feature l.
Ukrainian Dialects
The use of paratactic Across Ukrainian constructions with i and ta dialects
m. Impersonal constructions (with the direct object) denoting natural phenomena
Across Ukrainian dialects
Balkan Languages South Slavic, Albanian, Romanian, Greek South Slavic, Albanian, Greek
some Bulgarian dialects (Šumen, Tixomir near Kărdžali, Catholic settlers near Svištov and Plovdiv, and Bešenov in Romania) (Shevelov 1979: 82–83). Feature (b) is irrelevant for our discussion since it is found in the dialects that do not border with Romanian as one of the major players in the Balkan Sprachbund and a potential transmitter of the feature in question (see AUM 2, map 12). The change d’, t’ > g’, k’ is attested in Balkan Slavic, in particular in East Bulgarian, including the Kotel-Elena-Drjanovo dialect (Stojkov 2006: 111–112; Kobyljans’kyj 1960: 258). A connection between that Bulgarian innovation and the 17th-century southwestern Ukrainian change is possible during or soon after the Wallachian colonization in the 13th – 16th centuries. However, it is not possible to positively prove such a link in terms of either relative or absolute chronology (Shevelov 1979: 690, 539). Deserving of attention is the Ukrainian medial l which is attested both in East Ukrainian (Poltava and Starobil’s’k dialects) and West Ukrainian (Bukovyna and Pokuttja) (Žylko 1966: 64), but especially in those southwestern Ukrainian dialects which border with Hungarian and Romanian (Pan’kevyč 1938: 55) but not with Polish whose phonological system is free from this sound. In view of this well-established sound and its connection with the loss of palatalization across Ukrainian dialects (Shevelov 1979: 745; Kobyljans’kyj 1960: 258), there are no grounds for considering medial l either as a Central European or Balkan Sprachbund-forming feature (Danylenko 2013: 143–144; see Thomas 2008: 133). Altogether, the corresponding phonological features in the Carpathian and Balkan areas are hard to consider as contact-induced even if one postulates here a case of matter borrowing when loan words tend to introduce new phones into a target language (Thomason and Kaufman 1988: 124–125). Tentatively, this type of convergence looks distinctive at the micro-level rather than macrolevel. One can thus speak more accurately of Carpathian phonology rather than the Carpathian-Balkan one (Friedman 2008: 143).
362
Andrii Danylenko
The above conclusion holds true for two more phonological Carpathian Balkanisms which were not mentioned by Nimčuk and other specialists in Carpathian linguistics. The first one is devoicing of word-final obstruents. According to Joseph (2010: 622), this feature is found in Bulgarian, Macedonian, Romanian, Modern Greek, some Romani dialects, South Montenegrin and Torlak Serbian, and Turkish. The same feature is typical of West and especially Southwest Ukrainian which show complete neutralization of phonemic voicing before all obstruents and word boundary, e.g., U did# [t] ‘grandfather’. It should be borne in mind, however, that this neutralization is a result of the historical transformation of the phonetic protensity system into the phonetic voicing system in the Ukrainian-speaking territories (Danylenko 2006: 189–190). In other words, devoicing of word-final stop in the adjacent (Carpathian and Balkan) linguistic areas might be a corollary to different processes. First, in the case of Southwest Ukrainian and Balkan Slavic, one can posit independent parallel (convergent) realizations of genetically similar tendencies. Second, the aforementioned convergent features appear to be results of internal processes taking place in genetically distant and typologically different systems of Romanian, Modern Greek, Romani, and Turkish. The second feature is the development of *’a and *’ȩ into e in certain phonetic environments in Southwest Ukrainian. Thus, ’a underwent an umlaut into ’e after palatalized consonants, while ȩ changed into e after palatalized and some other obstruents (AUM, 2, maps 42–47). Since the ’a umlaut sprang first in the Moldova Ukraine, it is tempting to postulate a Romanian influence in this case, e.g., moldoveán ‘Moldovian’ and moldovéni (PL), a phenomenon largely shared with Bulgarian: mljako ‘milk’ next to mlečen ‘milky’. But the conditions are far from identical. In Romanian the ’a umlaut took place after palatalized consonants or a syllable with a front vowel, and this is the situation in most Bulgarian dialects (Shevelov 1979: 546). The Rodopian dialects where the presence of the preceding palatalized consonant suffices for the fronting of a, as in Ukrainian, is an exception, e.g., jejce ‘egg’, žeba ‘frog’, also šĕpka ‘hat’, poleni ‘logs’ in those Aegean Macedonian dialects, e.g., in Sjarsko and the northeastern part of Valovištko, where *a and *ę coalesced in one front vowel (Popova 2005: 287–291; BDA, maps F25–F32, F46–F53). As Shevelov (1979: 546) argued, in the Ukrainian dialects the character of the following consonant or next syllable, as well as the stress place, plays no role. In the Romanian and Bulgarian alternations consonants are palatalized before a but not before e, while in Southwest Ukrainian the palatalization which preceded a is also retained before e (provided the given consonant still exists in both palatalized and non-palatalized variants). Thus, contacts with Romanian
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
363
and Bulgarian, with their alternations ’a: e, at best, could have propelled the Ukrainian dialectal change ’a > ’e, doubtful as this assumption is, but the change itself was uniquely local, operating in accordance with its own tendencies. Rejecting an areal-typological interpretation of these phonological convergences, Popova (2005: 289) posited a common tendency in the treatment of both *a and *ȩ in Southwest Ukrainian and some Bulgarian (Rodopian) and Macedonian dialects; engendered in late Common Slavic, this tendency could have run to completion in individual Slavic dialects. Thus, it is tempting to accept the existence of phonological convergences (Carpathian Balkanisms) that could have developed in the Carpathian-Balkan macroarea as a result of different developmental pathways of common tendencies shaped in late Common Slavic.
3.3.2 Primary Carpathian Balkanisms 3.3.2.1 Dative enclitic pronouns Feature (d) is identified by Nimčuk (1993: 51) as the dative possessive. In fact, the use of enclitic pronouns in constructions like brat mu (he.DAT) ‘(to, for) his brother’, commonly attested in Transcarpathian, Lemkian, and Bojkian dialects and even some southeastern dialects (AUM 2, maps 273–275), is not possessive. One has to distinguish between the ethic (1a) and the possessive use (1b) of (enclitic) pronouns in the dative case (see Pogorelov 1939: 58). (1)
SWU a. zaslabla my žinka. get.sick.PST.F.SG I.CLT.DAT wife.NOM ‘My wife got sick.’ b. ja i syn mi. I.NOM and son.NOM I.CLT.DAT ‘My son and I.’
Together with other sentential clitics, these enclitics follow closely the ranking rule, especially in some central Transcarpathian dialects (Tolstaja 2012: 195; Danylenko 2012a), which is the corollary of their Common Slavic origin (Vaillant 1977: 88–89; Zaliznjak 2008: 35–36). Surprisingly, they are found in abundance in the language of administration in the Middle Ukrainian period (Nimčuk 1993: 51). Thus, the feature in question has its own, Common Slavic provenance, although later its primary use could have been maintained in the Carpathian region longer than in the rest of the Ukrainianspeaking territories. According to Tolstaja (2012: 201–202), the sentential
364
Andrii Danylenko
clitics in Southwest Ukrainian are closer to the Slovak type with a slightly different order of clitics though, rather than the Serbian-Croatian-Bosnian system. Thus being of Common Slavic origin, the enclitics in Southwest Ukrainian seem to demonstrate a more conservative sentential type as compared with the Balkan enclitics, in particular the emergence of their possessive use in South Slavic, Romanian and Greek as illustrated in (2) (Feuillet 2012: 174): (2) a. Bu kăstata mu ‘his house’ b. Ma (dial.) glasot mu ‘his voice’ c. Rm păru-mi ‘my hair’
3.3.2.2 Analytic comparative formations Feature (e) is viewed as a “Balkanism of Romanian origin” (Priestly 2010) or a “post-Byzantine Balkanism” (Friedman 2000: 100). However, the situation with the comparative marker maj in Southwest Ukrainian is far less obvious (Onyškevyč 1971: 442, see Section 3.1). First, from the late 16th century onward this marker has also been used sometimes with superlative forms of the type maj bil’šyj, literally ‘more bigger’, rather than maj velykўj ‘bigger’, a pattern typical of Romanian. Second, the marker maj is employed not only with verbs (3a) but also with adverbs and nouns, see (3b) and (3c) respectively (Nimčuk 1963: 24–25; Rot 1973: 49). (3)
SWU a. maj ljublju more love.PRS.1SG ‘I love more.’ b. maj rano more early ‘earlier’ c. maj rôzumak more clever.person.NOM.SG.M ‘A more clever person.’
Tentatively, the use of maj-forms alongside synthetic forms (as well as with verbs and nouns) is likely to testify to a different level of the grammaticalization of this marker, especially if the latter was borrowed or replicated on the Romanian maj as shown in (4).
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
(4)
365
Rm Fiică-mea cea mai mare Daughter-my.F DEF.F more big.F ‘My older daughter.’
However, the latter scenario is hard to prove since the marker maj can be paired with the local (eastern Slavic) dialectal marker naj which merged in the stressed position with maj (Bevzenko et al. 1978: 214–215; Nimčuk 1993: 52) as a result of matter borrowing (see Matras and Sakel 2007).7 What is worth mentioning in this respect is a similar pattern of comparatives with preposed pó- and superlatives with náj- in Balkan Slavic including the PrizrenTimok dialect group (Ivić 1985: 110–113; Stanišić 1985/1986: 257; Feuillet 2012: 210–211): Bu dobăr ‘good’ po-dobăr ‘better’ naj-dobăr ‘best’ b. Gk pio ‘more’ o pio ‘most’ c. Ab ё ‘more’ mё (+ article) ‘most’
(5) a.
Based on the above areal-typological and diachronic arguments, one can concur with Nimčuk (1963: 25) that the analytic comparative formations emerged in Bulgarian and Southwest Ukrainian independently (Rot 1973: 49–50) and are likely to represent different stages of their grammaticalization in the two systems. Thus, the independent developmental cline of the marker maj ~ naj in Ukrainian may be traced in such analytic comparatives as NU (Western Polissian) najbol’š ‘most of all’ (Hromyk 2006: 39) and SEU najupered ‘first of all’ (Nimčuk 1963: 26).
7 The marker maj in this function also occurs to the north of the international boundary: a map in Rieger (1996: 226) shows such forms in Hucul dialects as far north as Kolomyja; it also shows some villages where the Slavic superlative marker naj co-exists with maj, and a zone with naj only. Priestly (2010: 203) supports a supposition that the prior existence of the superlative marker naj assisted in the spread of Rm mai. I believe, however, that this might be a case of the borrowing of matter rather than pattern replication.
366
Andrii Danylenko
3.3.2.3 Derivation of numerals from ‘11ʹ to ‘19ʹ Feature (f) casts doubt on the Balkanization of the corresponding pattern. Among other languages (Hetzer 2010: 463, Heine, Kuteva 2005: 194), the location model of the “teen numbers” is found in all Slavic languages, especially in Transcarpathian Ukrainian. The advanced phonetic erosion of the second part of the location model [‘X on (top of) 10ʹ] inherited from Common Slavic is revealing in Transcarpathian and other eastern Slavic dialects (Bevzenko et al. 1978: 218), e.g., -nac’cat’, -nac’c’at’, -nac’c’it’ ‘on (top of) 10ʹ (Nimčuk 1993: 53). There are therefore no compelling grounds for treating this derivational pattern as a unique Balkanism or Carpathian Balkanism. 3.3.2.4 The de-volitive future The de-volitive future using an invariant particle derived from a verb meaning ‘to want’ or ‘to will’ was one of the Balkanisms to be identified as such, see Feature (g) (Seliščev 1925: 48; Sandfeld 1930: 180; Friedman 2008: 132). However, one wonders whether a parallel form in Southwest Ukrainian can be viewed as a result of borrowing from the Balkans, in particular from Romanian. In the latter language decategorialization is less advanced since the future marker voi is still an inflected auxiliary (Graur 1966: 269–270) while, for instance, in Modern Greek (tha + subjunctive), Bulgarian (šte and other dialectal varieties like ša, še, ši and others), Macedonian (ḱe and other dialectal varieties like ḱe, ža, za), and Albanian (do + subjunctive) decategorialization and erosion have given rise to an uninflected tense form (Koneski 1982: 201–201; Stanišić 1985/1986: 261–262; Heine and Kuteva 2005, 192). (6) a.
Rm Voi scrie. FUT.AUX.1SG write.INF b. Ma Ḱe pišam. FUT.AUX write.NONPAST.1SG ‘I shall write.’
Remarkably, de-volitive periphrastic constructions with various semantic nuances are liberally attested not only in the Ukrainian-speaking territories adjacent to Romanian. They are found more often than not, although with another meaning, in Southeast Ukrainian likewise (Nimčuk 1993: 54): (7) a.
SWU Xôtiv’-ym umerty z bol’y. wish-PST.AUX.M.SG die.INF from pain.GEN.M ‘I almost died from pain.’
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
367
b. SEU Vona xotila prymirjaty, ščo jij krašče she wish.PST.F.SG try.on.INF what she.DAT better bude do lycja. be.FUT.3SG to face.GEN ‘She was going to try on what will suit her best.’ The areal distribution, as well as the history of periphrastic constructions with ‘to wish, to will’ used for future events in Slavic (Křížková 1960), prompt us to search for non-areal mechanisms of the grammaticalization of the de-volitive periphrastic constructions in Ukrainian (Danylenko 2011: 165–177). Different stages of grammaticalization in Southwest (7.a) and Southeast Ukrainian (7.b) seem to corroborate an independent development in the aforementioned dialect groups. 3.3.2.5 The use of an uninflected relativizer Feature (h) is the occurrence of relative clauses with the uninflected (morphologically-invariant) relativizer U ščo (Be što, R čto) with or without the resumptive pronoun in parallel use with the inflected relative pronoun of the type U jakyj/kotryj ‘which, that’ (Danylenko 2014). This usage tentatively signals a strengthening of the uninflected relative clause linkage in the Carpathian linguistic area in contrast to some other eastern Slavic dialects, e.g., Russian, whose relative clause combining is based primarily on the inflected relative clause linkage. In fact, this correlation can be applied to the rest of the Indo-European languages, conceived in terms of Standard Average European. Not surprisingly, from the point of view of relative clause combining, Fiorentino (2007) divided Europe into two parts. On the one hand, Continental West Germanic languages (Dutch, German) still use an inflected Indo-European relative pronoun, while Romance languages, Greek, and English, on the other, adopted a mixed system with an invariant marker, see Fr and S que, I che (Murelli 2011: 324–329, 261; Danylenko 2018a: 366–371).8
8 Indeed, Romance and Germanic have demonstrated an analytic tendency, based on the uninflected relative clause linkage, to simplify at the cost of inflecting technique. Suffice it to mention here those German dialects, e.g., Oberrotweil, Balse, East Pomeranian and others, which use the uninflected wo ‘where’ without resumption (Fleischer 2004: 224–227). Of interest also is another, even closer German parallel, i.e., the use of uninflected was ‘what’ in combination with a resumptive pronoun, one of the most wide spread innovations in the north-west (in the North Saxon dialect of Husby in Schleswig) and, predominantly, in the east, namely, in East Pomeranian, the Upper Saxon dialect of north-western Bohemia, North Bavaria, and the Lubica linguistic island within the Slovak language area. Fleischer (2004:
368
Andrii Danylenko
I believe that in Ukrainian as well as other (Indo-European) languages the use of relative clauses with the uninflected relativizer speaks of a strengthening of the secondary analytic features in the inflecting language system characterized by primary synthetic features (Haarmann 2004: 82–83). Thus, unlike the inflected relative pronoun strategy premised on hypotactic subordination, the use of the relativizer in Ukrainian and Balkan Slavic, using absolute relativizers, tends to demonstrate a similar level of analyticity in paratactic constructions (Krapova 2009). To use Hinrichs’s words (1989/1991: 55), analytic features appear simpler in semantic content, more redundant in expression, more economic in substance, more friendly in communication, more economic in morphosyntactic expression, whence a predilection for parataxis (Trudgil 2011: 62, 146; Danylenko 2018b: 86–93; Danylenko 2018c). It follows from the above that the use of constructions with uninflected relativizers is subordinate to changes in the nature of communication in a multilingual, “colloquial” context, whence the use of the uninflected relativizer predominantly in dialects and colloquial speech where interpersonal understanding should be fast, efficient, and “friendly” (Joseph 2010: 624–625). It is not surprising that the Ukrainian locative-specialized relative element de ‘where’ (< hde < gde < kъde), which is used cross-dialectally, reminds of a similar relativization strategy in the Balkan languages, e.g., deto (also det, detu as used in the Bulgarian dialects of Western and Eastern Thrace) ‘where’ (< kădeto) in Bulgarian (Bojadžiev 1991: 118, 125, 132, 140), gdje ‘where’ in Serbian-Croatian-Bosnian, de ‘where’ (< unde) in Romanian in conjunction with resumption (Murelli 2011: 184–191): (8) a.
SEU Ta baba pomerla, de vony z neji that old.lady.NOM.F.SG die.PST.F.SG UREL they at her.GEN nasmixalys’. sneer.PST.PL ‘That old lady, at which they had sneered, died.’ (Žylko 1966: 126) b. Bu Ženata, deto govorix s neja. woman.DEF UREL talk.PST.1SG with her.ACC ‘The woman with whom I talked.’ (Krapova 2009: 1241)
The wh-adverbial marker of the type BU deto which has been used as an invariable relativizer since the 17th century, can be compared to the uninflected
235) hypothesized that, under the Slavic influence, the eastern occurrences of this relative clause type might have been contact-induced which, however, needs theoretical and empirical verification.
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
369
relative marker ščo in Ukrainian and its parallel relativizers in Balkan Slavic in that it does not encode the syntactic role of the head noun in the matrix clause. The Ukrainian relativizer ščo requires resumption in all the oblique cases while the Bulgarian deto behaves like što with respect to the resumptive pronoun in Serbian-Croatian-Bosnian. Thus, the pronoun (in its enclitic form) is obligatory with animates and more often than not with inanimates (Browne 1986: 32–35, 87; Kordić 1995: 160), although resumption in the oblique cases tends to become today obsolete (Thomas and Osipov 2012: 514). On toj čolovik, ščo z nym he that.NOM.SG.M man.NOM.SG.M UREL with whom.INS ja rozmovljav. I talk.PRT.SG.M ‘That is the man with whom I was talking.’ b. SCB Lòpata što se njóme žȉto vȅjē. spade UREL REFL it.INS rye winnow.PRS.SG.F ‘The spade with which rye is being winnowed.’
(9) a.
U
Vis-à-vis the typological parallelism between U de and Bu deto one can hardly concur with Gołąb and Friedman (1972: 45), according to whom, the function of deto as a general hypotactic conjunction meaning ‘where’ is a calque from the Greek. However, other Slavic languages and dialects have grammatized similar wh-adverbial markers. For instance, one should mention the wh-adverbial gde in the transitional Ukrainian-Russian dialects of the Kursk and Voronež regions (Akimova 1964: 142, Danylenko 2014: 196) and in the insular Russian dialects spoken in the territory of modern Latvia (Ruke-Dravinja [Ruke-Dravina] 1964); the marker de is encountered in some resettled Ukrainian dialects spoken in Southern Bessarabia between the estuaries of the Dniester and the Danube rivers (Kolesnykov 2015: 374). 3.3.2.6 The “merger” of goal and location in de Feature (i) was identified by Seliščev (1925: 47) as a Balkanism. The merger of goal and location in de (< kъde) is commonplace in the history and synchrony of East Slavic, as well as some West and South Slavic languages (Nimčuk 1993: 56; Pavliuc and Robciuc 2003: 77–78). Although in the neighboring Romanian language unde means both goal and location, this fact is not relevant in the context of this discussion. A similar syncretism is typical of other languages of Europe and can be traced to the prehistorical stage of Indo-European, when the accusative case served as a general oblique case before turning into the
370
Andrii Danylenko
accusative object. In other words, the primary independence of the accusative presupposed its indifference with regard to the semantics of direction and location, whence the residual use of Fr ici ‘here’, Gr poi ‘where’, L kur ‘where’ and many others combining the meaning of direction and location (Popov 1881: 17–20, 111–117; Danylenko 2016).
3.3.3 Secondary Carpathian Balkanisms Almost half of the Carpathian Balkanisms listed in Nimčuk (1993) are the socalled secondary features. A cursory look at them, including Features (k), (l), and (m) (see Section 3.3), is enough to argue that the so-called secondary Carpathian Balkanisms have either a particularly wide or extremely limited extent of diffusion as compared with the so-called primary Carpathian Balkanisms. They can be therefore left aside for future discussion in areal-typological or simply areal terms, although some of them demonstrate an independent (language-internal) grammaticalization of some Indo-European features. Among such Indo-European features, Nimčuk (1993: 55, 57–58) mentioned unwittingly Features (l) and (m), that is, the use of paratactic constructions with the prepositions i and ta as shown in (10a) and (10b) and the use of constructions denoting natural phenomena and taking the direct object as illustrated in (11a) and (11b): (10) a. SWU Hrebu s’ino, ta sonce uže rake.PRS.1SG hay.ACC.SG.N and sun.NOM.SG.N already zajšlo. go.down.PST.3SG.N ‘I rake hay, although the sun has already gone down.’ (Nimčuk 1993: 55) b. BCS Što činiš te se ne ženiš? what.ACC do.PRS.2SG and REFL not marry.PRS.2SG ‘What do you do and [why] you do not marry?’ (Sandfeld 1930: 197–198) (11) a. SWU Zvyčyrilo nja ў poli. get.dark.PST.3SG.N me.CLT.ACC in field.LOC.SG.N ‘Dusk fell on me in the field.’ (Nimčuk 1993: 58) b. Bu Go stămni večer-ta. him.CLT.ACC get.dark.PST.3SG.N evening.DEF ‘Dusk fell on him in the evening.’ (Sandfeld 1930: 210–211)
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
371
c. Bu (Soflijsko) Toj gu e strax. he.NOM he.CLT.ACC is fear ‘He is afraid.’ The impersonal construction with a clitic and a transitive use of the intransitive as demonstrated in (11a) and (11b) is of Indo-European origin (Popov 1881: 170–171, 223–224; Bauer 2000: 97–109). Yet it has a limited distribution both in Transcarpathia and in the Bulgarian dialects spoken, in particular, in Eastern and Western Thrace, now in Turkey and Greece; there, it is attested in a wider set of impersonal environments, as a rule, with the pleonastic use of weak (accusative or dative) object pronominal forms together with full pronominal forms of the nominative object as in (11c) (Bojadžiev 1991: 101, 118, 125, 132, 140, 198). The use of paratactic constructions in (10a) and (10b) can be explained through the lens of Givón’s (2009) theory about the rise of syntactic complexity rather than areal diffusion. This scholar argued that the developmental cline in the genesis of syntactic complexity, in diachrony, ontogeny, and no doubt in evolution, is primarily compositional (synthesis), following the general trend from simple clause via clause chains (parataxis) to complex (embedded) clause (syntax). Each of the stages is dependent on a historically prevalent type of communication within a particular speech community. The development (retention) of clause chains (parataxis) can be sustained due to the analytic tendency in the inflecting language system used in a speech community with an infringed homogeneity and changes in the social network due to extensive contacts of its speakers (see Trudgill 2011: 62; Danylenko 2018b and 2018c). In view of the above, I argue that the convergence of paratactic means in Southwest Ukrainian and Balkan languages could be propelled by similar changes in the abovementioned societal factors.
4 A sociolinguistic outlook The foregoing survey of the so-called “Carpathian Balkanisms” showed that most of the phonological and morphosyntactic convergences in the languages belonging to the Carpathian area, on the one hand, and the Balkan Sprachbund, on the other, are hard to ascertain as areal features only. Apart from Feature (f), which can be viewed as a contact-induced grammaticalization with Romanian serving as a model language, and Feature (e) combining elements of independent development with matter borrowing (see fn. 7), the rest of the shared features go back to the Common Slavic period or the history of East Slavic; they are results of the
372
Andrii Danylenko
language-internal (parallel) grammaticalization. This is why the Carpathian Balkanisms must be analyzed from a threefold – areal, genealogical, and typological – perspective. It should be noted at this point that, although faced with some conspicuous similarities primarily in the morphosyntax, the Balkanists are not ready to perceive the two areas under consideration as two parts of one Carpathian-Balkan linguistic macroarea. By contrast, the proponents of the existence of such a macroarea base their arguments on ethno-linguistic and cultural finds which are primarily reflected in some lexical convergences as attested in CDA and CCDA. Thus, to disregard the arguments of the representatives of Carpathian linguistics would be rather problematic, especially if one takes into consideration features shared by the languages spoken in the Carpathian basin and the Balkans as discussed by Nimčuk, Gricenko [Hrycenko], and other linguists. One wonders if the discrepancy between the two approaches can be somehow reconciled. Tentatively, in view of the prehistoric migration of Slavs to the Balkans and the Turkish presence in this area from the 14th century onward, one can distinguish between an early Slavicization of the Balkans and a later Balkanization of the Carpathian area correspondingly (Klepikova 2004: 304). The two consecutive waves of assimilation could have brought about the distinction of “Balkan Carpathianisms” (Balkanisms) for the Balkans and “Carpathian Balkanisms” (Carpathianisms) for the Carpathian area. Technically speaking, the “Balkan Carpathianisms” can be viewed as lexical and derivational features spread in the 1st millennium in the languages of the Carpathian basin and, subsequently, the Balkans. In their turn, the “Carpathian Balkanisms” tend to be conceived as phonological and morphosyntactic features developed relatively late in the Carpathian-Balkan linguistic macroarea (Asenova 1989: 80). Leaving aside the core vocabulary and derivational patterns that, arguably, have a minimum impact on the delimitation of the Carpathian-Balkan macroarea, two principle questions arise at this stage of argumentation. First, one should determine if the shared morphosyntactic features in the Carpathian-Balkan macroarea are the results of language-internal (parallel) or contact-induced grammaticalization. Second, both language-internal and contact-induced grammaticalization should be screened for possible propelling forces of the corresponding changes. I venture to claim that in the two cases the immediate reason behind these changes has one and the same nature that can be stated in terms of the sociolinguistic (systemic) typology. As a starting point in our theorization, I agree with Lindstedt (2000) who posited the existence of stable multilingualism in the Balkans from the 6th century onward which could have led to structural convergences between the respective languages. This is why for the Balkans, the author argued, there is no single source
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
373
language for this or that grammatical Balkanism. The source language simply did not exist in the traditional sense (Lindstedt 2000: 238, 241). A similar sociolinguistic situation can be reconstructed for the Carpathian area where, according to Rot (1973: 22), multilingual interference could influence the development of similar features across the dialects spoken in the Carpathian Mountains. To use Lindstedt’s logic, some features in the two adjacent linguistic areas could be results of either borrowing of matter or pattern replication due to mutual contact. Lindstedt (2000: 241) also noted that the multilingual contact situation would favor explicit syntactic marking in the Balkan languages: structural conflicts between the languages, according to him, are solved analytically, by syntactic means, because cross-language identification between analytic structures is easier than between inflectional categories (see Hinrichs 2004: 148; Joseph 2010: 625). Consequently, multilingual speakers across the Carpathian-Balkan macroarea could favor even such analytic structures that are not common in any of the contacting languages (see Thomason, Kaufman 1988: 96). However, if the origins of most of the (Carpathian) Balkanisms are to be sought in multilingual contact, the argument about borrowing of matter and pattern replication loses its validity in the light of the existence of multiple model and replica languages. To assume borrowing and replication with ensuing grammaticalization as ultimate results of multilingual contact, one still needs to adequately justify why, of all possible forms and structures available in the Balkan and Carpathian languages, particular features are borrowed or replicated and others are not. Otherwise, it is impossible to explicate the strengthening of analyticity in the Carpathian and Balkan languages found in multilingual contact situations through, for instance, a replication process. As a possible resolution to this conundrum, one should restore the intermediate step – between multilingual contact and replication – which is missing in the modern views on grammatical replication and borrowability in language contact (Wiemer, Wälchli, and Hansen 2012). If accounted for, this step is likely to explain why similar grammaticalization processes were taking place in both of the linguistic areas, thereby triggering convergent changes commonly called Balkanisms or Carpathianisms. To determine the intermediate step, one needs to resort to the postulates of the sociolinguistic typology as elaborated in Trudgill (2011). In this respect, one has, in particular, to concede that a particular configuration of societal factors (“external determinant”) preconditions the shaping of an “internal determinant” defined as a principal feature optimizing the whole system of a particular language system (Mel’nikov 2003: 57; Danylenko 2018b). Depending on changes in the external and ultimately internal determinants, an inflecting language can either accrue complexity through gaining new synthetic features, or simplify through the strengthening of its analytic features.
374
Andrii Danylenko
The external determinant is characterized through a number of factors, of which (1) small versus large community size, (2) dense versus loose social networks, (3) a large versus small amount of shared background information, and (4) social stability versus instability may be only some of the most important (Trudgill 2010: 300–301, 314; Mel’nikov 2003: 114; Danylenko 2018b: 83–88). Arguably, over a long-term period of social upheavals both Carpathian and, to a larger extent, Balkan language communities have been characterized by relatively small sizes, dense social networks, and large amounts of commonly shared information. In other words, long-term, intense multiple contacts between small communities in the two linguistic areas under consideration have involved all members of the respective communities, including adult speakers accommodating to imperfect skills of an interlocutor (Joseph 2010: 624). Such a post-critical threshold learning tended to involve selection or creation of structures that are “comfortable” (acceptable) to all speakers, a process which would result in simplification. In the case of Balkan Slavic and, to a lesser degree, of Carpathian Slavic as inflecting systems one can speak about an ever increasing analyticity conducive to the regularization of irregularities, an increase in lexical and morphological transparency, and the loss of redundancy (Trudgill 2011: 62–63). Among such changes, one can name, for instance, (a) deflexion (reduction in overt case-marking), which is particularly pronounced in Balkan Slavic, (b) an increase of prepositional usage, which is observed in both Balkan and Carpathian Slavic, (c) reduction/loss of conjugations and declensions, and (d) an increase in periphrastic verb forms as found in the two linguistic areas (see Danylenko 2011, 2015). Thus, convergences of the external and, ultimately, internal determinants of the corresponding Balkan and Carpathian languages rather than other motivations (Wiemer and Wälchli 2012: 19) appear to constitute driving forces of the grammaticalization of similar features in most of the idioms of the CarpathianBalkan macroarea.9 In some cases, however, multiple or parallel creation of separate features can be explained with the help of borrowing or replication (both 9 The postulated convergence of external determinants of the respective communities can be linked to the fact that, in the Balkans and in the Carpathian area, both rough and mountainous regions, such communities coexisted co-territorially in multilingual villages, towns, and cities and needed to communicate with one another on a variety of levels. According to Joseph (2010: 625), members of such communities in the Balkans were necessarily familiar with one another’s language to some degree. Very similar multilingual communities existed in the Carpathian contact zone, delineated in the north by the ring of the Carpathian Mountains extending from the Bavarian Forest through the Sudeten, the Tatras to the Transylvanian Alps; in the south by the Julian Alps and the Dinaric Range; and in the west by the Tauern and Dolomites of the Alpine Range (Thomas 2008: 126).
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
375
grammatical and lexical). Yet in no way can the latter processes be treated as the principle mechanisms of the grammaticalization of shared features since their formation is not dependent on the directionality inherent of borrowing or replication. In a sense, we are dealing here with multiple and parallel grammaticalization driven by changes in the internal determinant, which is ultimately contingent on a particular constellation of external (societal) factors.
5 Conclusions In sum, one can speak about the existence of the Carpathian-Balkan macroarea defined from the standpoint of both Carpathian and Balkan linguistics. Due to historical differences in the variables of external (societal) factors, the level of convergence tends to look different within the Balkan and Carpathian areas correspondingly, whence presumably weaker links between these two linguistic areas within the postulated macroarea. To ascertain this with certainty, one needs to bear in mind that the respective convergences can hardly be explicated through the areal studies perspective involving either borrowing or replication in the form of contact-induced grammaticalization. The adequate explanation should be based on areal, genetic, and typological (sociolinguistic or systemic) tenets. As it follows from this threefold approach, the alleged spread of shared features in the two areas was not immediately propelled by contact or even multilingual contacts but rather by a historical coalescence of the societal factors and the attendant variables giving rise to similar (analytic) changes in the internal determinants. The study of the Carpathian and Balkan languages, whose development has been determined by a similar constellation of societal factors and their variables, should be harmonized in what regards the history and areal typology of convergent features. Thus, lexical and derivational Carpathianisms as studied in Carpathian linguistics may serve as an “extralinguistic” illustration of the multilingual contacts during the migrations of Slavs to the Balkans that, through historical changes of the societal factors, reshaped the internal determinant of languages in contact. Finally, the emergence of grammatical Balkanisms that have long been in the focus of Balkan linguistics can be viewed as results of the analytic simplification of the corresponding inflecting systems due to the respective modifications of the internal determinant. These observations lead to, at least, two questions. First, how can we reconcile different levels of convergence in the two linguistic areas constituting the Carpathian-Balkan macroarea? Second, under what societal variables, were the internal
376
Andrii Danylenko
determinants changing and coalescing through different developmental stages in the Carpathian and Balkan languages? Hopefully, future research will furnish plausible answers to these questions.
Abbreviations AAR
Ab ACC AUX BCS
Be Bu CLT CM CS DAT DEF
dial F Fr FUT GEN
Gr H I INF INS L LOC M Ma NOM NONPST NU P PL PRS PST R REFL Rm S SEU
adverbial absolute relativizer Albanian accusative auxiliary Bosnian-Croatian-Serbian Belarusian Bulgarian clitic comparative marker Common Slavic dative definite article dialectal feminine French future genitive Greek Hungarian Italian infinitive instrumental Lithuanian locative masculine Macedonian nominative nonpast tense North Ukrainian Polish plural present tense past Russian reflexive Romanian Spanish Southeast Ukrainian
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
SG SWU UREL VOC
377
singular Southwest Ukrainian uninflected relative particle vocative
References Akimova, Galina Nikolaevna. 1964. Tendencii v razvitii otnositel’nogo podčinenija v sovremennyx vostočnoslavjanskix jazykax [Tendencies in the development of relative subordination in modern East Slavic languages]. Izvestija Akademii nauk SSSR. Serija literatury i jazyka 23 (2): 138–144. Alexander, Ronelle. 2012. Convergence and causation in Balkan Slavic. In: Thede Kahl, Michael Metzeltin, and Helmut Schaller (eds.), Balkanismen heute – Balkanisms Today – Балканизмы сегодня, 39–45. Vienna: LiT Verlag. Aronson, Howard I. 2007. The Balkan linguistic league, “Orientalism” and linguistic typology. In: Kenneth Naylor Memorial Lecture Series 4, 1–77. Ann Arbor/New York: Beech Stave Press. Asenova, Petja. 1989. Balkansko ezikoznanie: osnovni problemi na balkanskija ezikov săjuz [Balkan Linguistics: Pain Problems of the Balkan Linguistic League]. Sofia: Izdatelstvo Nauka i Izkustvo. Asenova, Petja. 2012. O statuse obščix osobennostej dvux balkanskix jazykov – balkanizmy oni ili net? [On the status of shared peculiarities in two Balkan languages – are they Balkanisms or not?] In: Thede Kahl, Michael Metzeltin, and Helmut Schaller (eds.), Balkanismen heute – Balkanisms Today – Балканизмы сегодня, 333–343. Vienna: LiT Verlag. AUM – Matvijas, Ivan Hryhorovyč et al. 1984–2001. Atlas ukrajins’koji movy [Atlas of the Ukrainian Language], Vols. 1–3. Kyiv: Naukova dumka. Bauer, Brigitte. 2000. Archaic Syntax in Indo-European. The Spread of Transitivity in Latin and French. Berlin/New York: Mouton de Gruyter. BDA – Stojkov, Stojko and Samuil Borisovič Bernštejn. 1964. Bălgarski dialekten atlas [Bulgarian Dialect Atlas], Vol. 1: Jugo-iztočna Bălgarija [Southeastern Bulgaria]. Sofia: Bălgarska Akademija na naukite. Bernštejn, Samuil Borisovič. 1972. Problemy karpatskogo jazykoznanija [Problems of Carpathian linguistics]. In: Galina Petrovna Klepikova (ed.), Karpatskaja dialektologija i onomastika, 3–15. Moscow: Nauka. Bernštejn, Samuil Borisovič. 1976. Lingvističeskie aspekty karpatistiki [Linguistic aspects of Carpathian studies]. In: Obščekarpatskij dialektologičeskij atlas, 5–15. Chişinău: Štiinca. Bevzenko, Stepan Pylypovyč et al. 1978. Istorija ukrajins’koji movy. Morfolohija [A History of the Ukrainian Language. Morphology]. Kyiv: Naukova dumka. Bojadžiev, Todor. 1991. Bălgarskite govori v zapadna (Belomorska) i iztočna (Odrinska) Trakija [Bulgarian Dialects in Western (Belomorska) and Eastern (Odrin) Thrace]. Sofia: Sv. Kliment Oxridski. Browne, Wayles. 1986. Relative Clauses in Serbo-Croatian in Comparison with English. Zagreb: Institute of Linguistics.
378
Andrii Danylenko
Campbell, Lyle. 2006. Areal linguistics: a closer scrutiny. In: Yaron Matras, April McMahon, and Nigel Vincent (eds.), Linguistic Areas: Convergence in Historical and Typological Perspective, 1–31. Basingstoke: Palgrave Macmillan. CCDA – Udler, Rubin Jakovlevič. et al. Obščekarpatskij dialektologičeskij atlas [Common Carpathian Dialect Atlas], Vols. 1–7. Chişinău /Moscow, 1989–2003. CDA – Bernštejn, Samuil Borisovič et al. 1967. Karpatskij dialektologičeskij atlas [Carpathian Dialect Atlas]. Moscow: AN SSSR. Dahl, Östen. 2001. Principles of areal typology. In: Martin Haspelmath, Ekkehard Konig, Wulf Oesterreicher, and Wolfganag Raible (eds.), Language Typology and Language Universals. An International Handbook, Vol. 2, 1456–1470. New York: Walter de Gruyter. Danylenko, Andrii. 2006. Slavica et Islamica: Ukrainian in Context. Munich: Otto Sagner. Danylenko, Andrii. 2011. Is there any inflectional future in East Slavic? A case of Ukrainian against Romance reopened. In: Motoki Nomachi (ed.), Grammaticalization in Slavic Languages, 147–177. Sapporo: Slavic Research Center, Hokkaido University. Danylenko, Andrii. 2012a. Auxiliary clitics in Southwest Ukrainian: Questions of chronology, areal distribution, and grammaticalization. Journal of Slavic Linguistics 20 (1): 3–34. Danylenko, Andrii. 2012b. Lota vўvjuh˝, a pjat’ varўšuv’ pohubyl’: A case of Ukrainian auxiliary clitics with the element x. In: Andrii Danylenko and Serhii Vakulenko (eds.), Studien zu Sprache, Literatur und Kultur bei den Slaven. Gedenkschrift für George Y. Shevelov, 74–94. (Sammelbände – Sborniki 42.) Munich: Otto Sagner. Danylenko, Andrii. 2013. Ukrainian in the language map of Central Europe: Questions of areal-typological profiling. Journal of Language Contact 6 (1): 134–159. Danylenko, Andrii. 2014. On the relativization strategies in East Slavic. In: Motoki Nomachi, Andrii Danylenko, and Predrag Piper (eds.), Grammaticalization and Lexicalization in Slavic languages, 183–204. Munich: Otto Sagner. Danylenko, Andrii. 2015. On the mechanisms of the grammaticalization of comitative and instrumental categories in Slavic. Journal of Historical Linguistics 5 (2): 267–296. Danylenko, Andrii. 2016. Oleksandr Popov (1855–80) and the reconstruction of Indo-European noun Inflection. Language and History 59 (2): 112–130. Danylenko, Andrii. 2018a. A tale of two pathways: on the development of relative clause chaining in East Slavonic. In: Jasmina Grković-Major, Björn Hansen, and Barbara Sonnenhauser (eds.), Diachronic Slavonic Syntax, 361–386. Berlin/Boston: De Gruyter Mouton. Danylenko, Andrii. 2018b. The correlation of linguistic patterning and societal structures in systemic typology. Studia Linguistica Universitatis Iagellonicae Cracoviensis 135 (1): 81–96. Danylenko, Andrii. 2018c. Complexity development and simplification in systemic typology: types of changes in the internal determinant. Studia Linguistica Universitatis Iagellonicae Cracoviensis 135 (2): 113–129. Desnickaja, Agnija Vasil’evna. 1976. K interpretacii balkanizmov v karpatskoj leksike [On the interpretation of Balkanisms in Carpathian lexicon]. In: Obščekarpatskij dialektologičeskij atlas, 15–27. Chişinău: Štiinca. Desnickaja, Agnija Vasil’evna. 1987. Albanskaja literatura i albanskij jazyk [Albanian Literature and Albanian Language]. Leningrad: Nauka. Feuillet, Jack. 1986. La linguistique balkanique. (Cahiers balkaniques 10.) Paris: Publication Langues’O.
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
379
Feuillet, Jack. 2012. Linguistique comparée des langues balkaniques. (Collection de grammaires de l’Institut d’études slaves 14.) Paris: Institut d’études slaves. Fiorentino, Giuliana. 2007. European relative clauses and the uniqueness of the relative pronoun type. Rivista di linguistica 19 (2): 263–291. Fleischer, Jürg. 2004. A typology of relative clauses in German dialects. In: Bernd Kortmann (ed.), Dialectology Meets Typology, 211–243. Berlin/New York: Mouton de Gruyter. Friedman, Victor A. 2000. Romani in the Balkan linguistic league. In: Ch. Tzitzilis and Kh. Symeonides (eds), Balkanlinguistik: Synchronie und Diachronie, 95–105. Thessaloniki: University of Thessaloniki. Friedman, Viktor A. 2003. ‘One’ as an indefinite marker in Balkan and Non-Balkan Slavic. In: Alan Timberlake and Michael S. Flier (eds.), American Contributions to the 13th International Congress of Slavists, 93–112. Bloomington, IN: Slavica. Friedman, Viktor A. 2008. Balkan Slavic dialectology and Balkan linguistics: periphery as center. In: Christina Y. Bethin (ed.), American Contributions to the 14th International Congress of Slavists. Ohrid, September 2008, Vol. 1: Linguistics, 131–148. Bloomington, IN: Slavica. Gołąb, Zbigniew and Victor A. Friedman 1972. The relative clause in Slavic. In: Paul M. Peranteau, Judith Levi, and Gloria Phares (eds.), The Chicago Witch Hunt. Papers from the Relative Clause Festival, April 13, 1972, 30–46. Chicago: Chicago Linguistic Society. Graur, Alexandru. 1966. Gramatica limbii române. Bucharest: Editura Academiei Republicii Socialiste România. Gricenko, Pavel Efimovič [Pavlo Juxymovyč Hrycenko]. 2008. Carpato-balcanica v svete “Obščekarpatskogo dialektologičeskogo atlasa” [Carpat[h]o-Balcanica in the light of The Common Carpathian Dialect Atlas] In: Anna A. Plotnikova (ed.), Karpato-balkanskij dialektnyj landšaft: jazyk i kul’tura. Pamjati Galiny Petrovny Klepikovoj, 26–57. Moscow: Institut Slavjanovedenija RAN. Haarmann, Harald. 2004. Prozesse sprachlichen Strukturwandels im Spannungsfeld von Kontaktlinguistik und Sprachtypologie. In: Uwe Hinrichs (ed.), Die europäischen Sprachen auf dem Wege zum analytischen Sprachtyp, 67–85. Wiesbaden: Harrassowitz. Hamp, Eric P. 1989. Yugoslavia – a crossroads of Sprachbünde. Zeitschrift für Balkanologie 25 (1): 44–47. Heine, Bernd and Tania Kuteva. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva. 2006. The Changing Languages of Europe. Oxford/New York: Oxford University Press. Hetzer, Armin. 2010. Das südosteuopäische Areal. In: Uwe Hinrichs (ed.), Handbuch der Eurolinguistik, 457–473. Wiesbaden: Harrasowitz. Hinrichs, Uwe. 1989/1991. Das Slavische und die sogenannten Balkanismen. Für eine pragmatische Balkanlinguistik. Zeitschrift für Balkanologie 26 (1): 43–62. Hinrichs, Uwe. 1999. Die sogenannten ‘Balkanismen’ als Problem der Südosteuropa-Linguistik und der algemeinen Sprachwissenschaft. In: Uwe Hinrichs (ed.), Handbuch der Südosteuropa-Linguistik, 429–461. Wiesbaden: Harrasowitz. Hinrichs, Uwe. 2004. Orale Kultur, Mehrsprachigkeit, radikaler Analytismus: Zur Erklärung von Sprachstrukturen auf dem Balkan und im kreolischen Raum. Ein Beitrag zur Entmystifizierung der Balkanlinguistik. Zeitschrift für Balkanologie 40 (2): 141–174.
380
Andrii Danylenko
Hromyk, Jurij. 2006. Stupenjuvannja pryslivnykiv u zaxidnopolis’kyx i sumižnyx hovirkax [Degrees of comparison of adverbs in Western Polissian and adjacent dialects]. Rozprawy slawistyczne 20: 27–42. Jagić, Vatroslav. 1898. Verwandtschaftsverhältnisse innerhalb der slavischen Sprachen. Archiv für slavische Philologie 20: 13–53. Joseph, Brian D. 1983. The Synchrony and Diachrony of the Balkan Infinitive: A Study in Areal, General, and Historical Linguistics. Cambridge: Cambridge University Press. Joseph, Brian D. 2010. Language contact in the Balkans. In: Raymond Hickey (ed.), The Handbook of Language Contact, 618–633. Malden, MA: Wiley-Blackwell. Ivić, Pavle. 1985. Dijalektologija srpskohrvatskog jezika: uvod u štokavsko narečje [The Dialectology of the Croatian Language: An Introduction into the Štokavian Dialect]. Novi Sad: Matica srpska. Kahl, Thede, Michael Metzeltin, and Helmut Schaller (eds.). 2012. Balkanismen heute – Balkanisms Today – Balkanizmy segodnja. Vienna: LiT Verlag. Klepikova, Galina Petrovna. 2003. Karpatskoe jazykoznanie i “Obščekarpatskij dialektologičeskij atlas” [Carpathian linguistics and The Common Carpathian Dialect Atlas]. Studia Slavica Hungarica 48 (4): 357–374. Klepikova, Galina Petrovna. 2004. Novye raboty o slavjanskom vlijanii na jazyki karpatoslavjanskogo areala [New works on the Slavic influence on the languages of the Carpathian Slavic area]. In: Issledovanija po slavjanskoj dialektologii 9, 304–315. Moscow: Institut slavjanovedenija RAN. Kobyljans’kyj, Bronyslav Volodymyrovyč. 1960. Dialekt i literaturna mova [Dialect and literary language]. Kyiv: Radjans’ka škola. Kobyljans’kyj, Bronyslav Volodymyrovyč. 1971. Leksyčni paraleli v hovorax ukrajins’koji i pivdennoslov"jans’kyx mov [Lexical parallels in Ukrainian dialects and South Slavic]. Praci XII Respublikans’koji dialektolohičnoji narady, 429–440. Kyiv: Naukova dumka. Kobyljanskij, Bronislav Vladimirovič [Bronyslav Volodymyrovyč Kobyljans’kyj]. 1973. Možno li govorit’ o svoeobraznom balkano-karpatskom sojuze jazykov [Can one speak about a kind of Balkan-Carpathian linguistic league]? In: Simpozium po problemam karpatskogo jazykoznanija (24–26 aprelja 1973 g.). Tezisy dokladov i soobščenij, 28–29. Moscow: Nauka. Kolesnykov, Andrij Oleksandrovyč. 2015. Morfolohija ukrajins’kyx pivdennobesarabs’kyx hovirok: heneza i dynamika [The Morphology of the Ukrainian Southern Bessarabian Dialects: Genesis and Dynamics]. Izmajil: SMIL. Koneski, Blaže. 1982. Istorija na makedonskiot jazik [A History of the Macedonian Language]. Skopje: Kulura. Kopitar, Jernej. 1829. Albanische, walachische und bulgarische Sprache. Jahrbücher der Literatur 46: 59–106. Kordić, Snjěžana. 1995. Relativna rečenica [The Relative Sentence]. Zagreb: Hrvatsko filološko društvo, Matica hrvatska. Krapova, Iliana. 2009. Bulgarian relative and factive clauses with an invariant complementizer. Lingua 120 (5): 1240–1272. Křížková, Helena. 1960. Vývoj opisného futura v jazycích slovanských, zvláštĕ v ruštinĕ [The Development of Periphrastic Future in Slavic Languages, Especially in Russian]. Prague: Státní pedagogické nakladatelství. Lindstedt, Jouko. 2000. Linguistic Balkanization: contact-induced change by mutual reinforcement. In: Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
381
Contact, 231–246. (Studies in Slavic and General Linguistics 28.) Amsterdam/Atlanta, GA: Rodopi. Matras, Yaron and Jeanette Sakel. 2007. Investigating the mechanisms of pattern replication in language convergence. Studies in Language 31 (1): 829–865. Mel’nikov, Gennadij Prokop’evič. 2003. Sistemnaja tipologija jazykov: Princypy, metody, modeli [A Systemic Typology of Languages: Principles, Methods, Models]. Moscow: Nauka. Miklosich, Franz. 1861. Die slavischen Elemente im Rumunischen. Denkschriften der kaiserlichen Akademie der Wissenschaften, philosophisch-historische Klasse 12: 1–70. Murelli, Adriano. 2011. Relative Constructions in European Non-Standard Varieties. Berlin/ Boston: Mouton De Gruyter. Nimčuk, Vasyl’ Vasyl’ovyč. 1963. Stupenjuvannja u verxnjonadboržavs’kyx hovirkax [Degrees of comparison in the Upper Bârzava river dialects]. Pytannia hramatyky i leksykolohii ukrajins’koji movy, 17–29. Kyiv: Akademija Nauk Ukrajins’koji RSR. Nimčuk, Vasilij Vasil’evič [Vasyl’ Vasyl’ovyč Nimčuk]. 1988. Karpatоukrainsko-ju žnoslavjanskie jazykovye paralleli i toždestva [Carpathian Ukrainian and South Slavic linguistic parallels and equivalencies]. In: Obščeslavjanskij lingvističeskij atlas. Materialy i issledovanija, 294–313. Moscow: Nauka. Nimčuk, Vasyl’ Vasyl’ovyč. 1991. Balkans’kyj movnyj sojuz ta ukrajins’ki dialekty [The Balkan Sprachbund and Ukrainian dialects]. In: Rozvytok duxovnoji kul’tury slov”jans’kyx narodiv, 4–30. Kyiv: Naukova dumka. Nimčuk, Vasyl’ Vasyl’ovyč. 1993. Ukrajins’ki hovory ta balkans’kyj movnyj sojuz [Ukrainian dialects and the Balkan Sprachbund]. In: Slov”jans’ke movoznavstvo, 41–63. Kyiv: Naukova dumka. Onyškevyč, Myxajlo Josypovyč. 1971. Bojkivs’ko-pivdennoslov”jans’ki movni paraleli [BojkianSouth Slavic language parallels]. In: Praci XII Respublikans’koji dialektolohičnoji narady, 440–445. Kyiv: Naukova dumka. Pan’kevyč, Ivan. 1938. Ukrajins’ki hovory Pidkarpats’koji Rusy i sumežnyx oblastej, [The Ukrainian Dialects of Subcarpathian Rus’ and Adjacent Regions] Part 1: Zvučnja i morfolohija [Phonetics and Morphology]. (Knihovna Sboru pro výzkum Slovenska a Podkarpatské Rusi při Slovanském Ústavu v Praze 9.) Praha: Orbis. Pan’kevyč, Ivan. 1955. Ukrajins’ko-bolhars’ki movni zv”jazky v Semyhorodi [The UkrainianBulgarian linguistic relations in Semyhorod]. Slavia 24: 211–240. Pavliuc, Nicolae and Ion Robciuc. 2003. Ukrainian Dialects in Romania. Dialectal Texts. Edmonton: Canadian Institute of Ukrainian Studies. Peco, Asim. 1991. Pregled srpskohrvatskih dijalekata [An Overview of the Serbo-Croatian dialects]. Belgrade: Naučna knjiga. Plank, Frans. 1998. The co-variation of phonology with morphology and syntax: A hopeful history. Linguistic Typology 2: 195–230. Plotnikova, Anna Arkad’evna. 2008. Karpato-balkanskij dialektnyj landšaft – novye pespektivy: jazyk i kul’tura vo vzaimodejstvii [A Carpathian-Balkan dialectal landscape–new perspectives: language and culture in cooperation]. In: Anna Arkad’evna Plotnikova (ed.), Karpato-balkanskij dialektnyj landšaft: jazyk i kul’tura. Pamjati Galiny Petrovny Klepikovoj, 10–25. Moscow: Institut slavjanovedenija RAN. Pogorelov, Valerij. 1939. Bolgarizmy v karpatorusskix govorax [Bulgarianisms in the CarpathoRusyn dialects]. In: Valerij Pogorelov. Karpatorusskie ètjudy 5, 57–62 (Spisy Filozofickej fakulty Slovenskej univerzity v Bratislave 29.) Bratislava.
382
Andrii Danylenko
Popov, Aleksandr Vasil’evič. 1881. Sintaksičeskie isslědovanija I [Syntactic Studies I]. Voronež: V. I. Isaev. Popova, Tat’jana Vital’evna. 2005. K voprosu o sxoždenijax meždu ukrainskimi i bolgarskimi dialektami v oblasti fonetiki [On the question of convergences between the Ukrainian and Bulgarian dialects in the field of phonetics]. In: Issledovanija po slavjanskoj dialektologii 6, 279–291. Moscow: Institut slavjanovedenija RAN. Potebnja, Aleksandr Afanas’evič. 1968. Iz zapisok po russkoj grammatike [From the Notes on Russian Grammar], Vol. 3. Moscow: Prosveščenie. Priestly, Tom. 2010. On the diffusion of Romanism maj in Ukrainian dialects. Balkanistica 23: 267–284. Rieger, Janusz (ed.). 1996. A Lexical Atlas of the Hutsul Dialects. Warsaw: Semper. Rot, A. M. 1967. Osobennosti vzaimodejstvija jazykov i dialektov Karpatskogo bassejna i vengersko-ukrainskaja jazykovaja interferencija [Particularities of the interrelations of the languages and dialects of the Carpathian Basin and the Hungarian-Ukrainian interference]. Studia Slavica Hungarica 13: 247–268. Rot, A. M. 1973. Osobennosti vzaimodejstvija jazykov i dialektov karpatskogo areala [Particularities of the Interrelations of the Languages and Dialects of the Carpathian Basin]. Užhorod: Užgorodskij gosudarstvennyj universitet. Ruke-Dravinja [Ruke-Dravina], Velta. 1964. Dial. gde [=kotoryj]. In: Ihor Vahros and Martti Kahla (eds.), Lingua Viget. Commentationes Slavicae in Honorem V. Kiparsky, 115–119. Helsinki: Suomalaisen Kirjallisuuden Kirjapaino Oy. Sandfeld, Kristian. 1930. Linguistique balkanique. Paris: Klincksieck. Schaller, Helmut W. 1975. Die Balkansprachen. Eine Einführung in die Balkanologie. Heidelberg: Carl Winter. Schaller, Helmut W. 2012. Balkanismen in Vergangenheit und Gegenwart – Theorie und Wirklichkeit. In: Thede Kahl, Michael Metzeltin, and Helmut Schaller (eds.), Balkanismen heute – Balkanisms Today – Balkanizmy segodnja, 5–12. Vienna: LiT Verlag. Seliščev, Afanasij. 1925. Des traits linguistiques communs aux langues balkaniques: un balkanisme ancient en bulgare. Revue des études slaves 5 (1/2): 38–57. Shevelov, George Y. 1979. A Historical Phonology of the Ukrainian Language. (Historical Phonology of the Slavic languages 4.) Heidelberg: Carl Winter. Sobolev, Andrej. 2011. Antibalkanizmy [Anti-Balkanisms]. Južnoslovenski filolog 67: 185–195. Stanišić, Vanja. 1985/1986. Balkanizmi u srpskohrvatskom jeziku [Balkanisms in the Serbo-Croatian language]. Balkanika 16/17: 245–265. Stojkov, Stojko. 2006. Bǎlgarska dialektologija [Bulgarian Dialectology]. Sofia: Prof. Marin Drinov. Thomas, George. 2008. Exploring the parameters of a Central European Sprachbund. Canadian Slavonic Papers 50 (1/2): 123–153. Thomas, George. 2010. Serbo-Croatian as a bridge between the Balkan and Central European Sprachbünde. Balkanistika 23: 371–388. Thomas, Paul-Louis and Vladimir Osipov. 2012. Grammaire du bosniaque, croate, monténégrin, serbe. (Collection de grammaires de l’Institut d’études slaves 8.) Paris: Institut d’études slaves. Thomason, Sarah G. 2000. Linguistic areas and language history. In: Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in Contact, 311–327. (Studies in Slavic and General Linguistics 28.) Amsterdam/Atlanta, GA: Rodopi.
11 Balkanisms and Carpathianisms or, Carpathian Balkanisms?
383
Thomason, Sarah G. 2001. Language Contact: An Introduction. Edinburgh/Washington, DC: Georgetown University Press. Thomason, Sarah G. and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkley/Los Angeles, CA: University of California Press. Tolstaja, Marfa Nikitična. 2012. Karpatoukrainskie ènklitiki v južnoslavjanskoj perspective [Carpathian-Ukrainian enclitics in the South Slavic perspective]. In: Karpato-balkanskij dialektnyj landšaft. Jazyk i kul’tura 2, 190–210. Moscow: Institut Slavjanovedenija RAN. Trudgill, Peter. 1997. Typology and sociolinguistics: linguistic structure, social structure and explanatory comparative dialectology. Folia Linguistica 31 (3/4): 349–360. Trudgill, Peter. 2010. Contact and Sociolinguistic Typology. In: Raymond Hickey (ed.), The Handbook of Language Contact, 299–319. Malden, MA: Wiley-Blackwell Trudgill, Peter. 2011. Sociolinguistic Typology. Social Determinants of Linguistic Complexity. Oxford: Oxford University Press. Vaillant, Andre. 1977. Grammaire comparée des langues slaves, Vol. 5: La syntaxe. Paris: Klincksieck. van der Auwera, Johan. 1998. Revisiting the Balkan and Meso-American linguistic areas. Language Sciences 20 (3): 259–270. Vasmer, Max. 1940/1941. Gibt es bulgarische Einflüsse in den ukrainischen Karpatenmundarten? Zeitschrift für slavische Philologie 17: 48–51. Wiemer, Björn and Bernhard Wälchli. 2012. Contact-induced grammatical change: Diverse phenomena, diverse perspectives. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 3–64. (Trends in Linguistics 242.) Berlin/Boston: Mouton De Gruyter. Wiemer, Björn, Bernhard Wälchli, and Björn Hansen (eds.). 2012. Grammatical Replication and Borrowability in Language Contact. (Trends in Linguistics 242.) Berlin/Boston: Mouton de Gruyter. Zaliznjak, Andrej Anatol’evič. 2008. Drevnerusskie ènklitiki [Old Russian Enclitics]. Moscow: Jazyki slavjanskix kul’tur. Žylko, Fedot Troxymovyč. 1966. Narysy z dialektolohiji ukrajins’koji movy [Essays on the Dialectology of the Ukrainian Language]. Kyiv: Radjans’ka škola.
Walter Breu
12 Morphosyntactic changes in Slavic micro-languages: The case of Molise Slavic in total language contact 1 Introduction Several Slavic micro-languages are spoken in non-Slavic countries in the western and southern periphery of the Slavic language area. Examples of such microlanguages would be the well-established literary languages Lower and Upper Sorbian (Germany), but also include lesser-known varieties like Burgenland Croatian (Austria, Slovakia, Hungary), Carinthian Slovene (Austria), Slovenebased Resian (Italy) and Molise Slavic (Italy). All of them have been in situations of language contact for centuries and have now entered a stage of “total” (or “absolute”) language contact. This means that all the speakers of these varieties are fully bilingual (and have been so for a long time), as they use at least one variety of the dominant language of the country in question in addition to their own minority language. As a result of the historical and present influence of the dominant languages and varieties many contact-induced changes have manifested themselves in these Slavic micro-languages on all linguistic levels. Such cases of long-term collective bilingualism drift towards a kind of compound bilingualism, characterized by the formula “one form – two substance systems”, as Sasse (1992: 61) puts it. It is in such situations that the concept of a common diasystem or diagrammar of the two (or more) languages in question becomes evident. The concept of “diagrammar” is a theoretical model based on the hypothesis that multilingual speakers do not strictly separate the grammars of their two or more languages but combine them in the most economical way possible. One could argue that the grammars of the individual languages in contact are synchronically derived from such a diagrammar, in terms of a common deep structure, by means of language-specific rules. The fewer the rules, the more economical the management of the languages is. Therefore, the reduction of the rules by means of a more and more comprehensive diagrammar is the overall purpose of language change in total contact situations (see Breu 2011a: 440). We will show in the following that there are two main procedures leading to such a favorable diagrammar, the “adaptation of the semantic structure” of the replica language to that of the dominant model (semantic calque) and the “loan translation” (formal calquing) of periphrastic elements from the dominant language system. https://doi.org/10.1515/9783110639223-013
386
Walter Breu
Although contact-induced developments in phonology/phonetics and in the lexicon are more obvious to speakers and linguists at first glance, changes in grammatical categories have a deeper effect on the structure and the typological characteristics of the replica languages. This is why this chapter will be concerned with the morphosyntactic changes that characterize convergences between the contact languages involved. We will focus largely on one special case, i.e., Molise Slavic in Southern Italy, while other Slavic varieties in non-Slavic speaking countries will only be considered as points of comparison for the developments found in Molise Slavic.1 As syntactic questions of word order also deeply affect language structure, we will at least marginally touch on the contactinduced changes in this sphere, too, especially with regard to the order of clitics. Morphosyntactic changes in Molise Slavic and the other micro-languages do not involve changes in the “external” morphology, in the sense that no inflectional endings have been borrowed (= no matter borrowing), but rather in that they concern the internal structure of the categories (= pattern borrowing).2 In Molise Slavic we find changes in the oppositions and functions of the grammemes of all grammatical categories of the verb and the noun, making this micro-language, in many respects, distance itself from what could be called the common Slavic basis and bringing it closer to typically Romance structures. As for the developments themselves, even some Slavic “diachronic constants” of language change have been cancelled out by Romance diachronic constants. By “diachronic constants” we mean evolutionary tendencies (loss, preservation, restructuring) in a language family (phylum) that ideally are observed by all its members with respect to a certain linguistic phenomenon. Examples are the Slavic languages’ strong tendencies to lose synthetic tenses, to preserve their case and aspect systems, to not develop articles, and especially to follow a certain pathway in a given change. An example for the latter type of constants certainly is the loss of the imperfect before the loss of the aorist, whenever a Slavic language reduces its system of past tenses (see Breu 2006: 71–73). In the present chapter we claim that all violations of the diachronic constants of Slavic in the micro-languages in question are due to contact with languages belonging to families that have different diachronic constants. On the other hand,
1 For a direct comparison of most of the aforementioned minority languages with respect to verbal aspect and to their article systems, see Breu (2011a). 2 For the terms “matter loan” and “pattern loan”, corresponding to the traditional concepts of Materialentlehnung and Strukturentlehnung, see Sakel (2007).
12 Morphosyntactic changes in Slavic micro-languages
387
the cases of resistance against the dominance of foreign diachronic constants should not be neglected, especially in the case of total language contact, which displays a general tendency towards maximal isomorphism.3 After presenting a wide range of contact-induced developments, we will include a short discussion about the effect of these changes on the position of Molise Slavic in the SAE (Standard Average European) typology in our conclusion at the end of this paper.
2 The actual and the historical situation of Molise Slavic Molise Slavic is still spoken in the coastal hinterland of the Province of Campobasso, about 35 kilometres away from the Adriatic Sea, in the Southern Italian Region of Molise. Nowadays, this Slavic-speaking area, having been larger in the past, is restricted to the territory of three bordering municipalities with the villages of Acquaviva Collecroce, Montemitro, and San Felice del Molise in their centers. There are only about one thousand persons left who actively use Molise Slavic or at least are able to understand it, out of an overall number of less than two thousand inhabitants of these villages.4 Language knowledge and behavior differs from one village to another, with the smallest village, Montemitro, being most conservative with respect to both the influence of language contact and language usage. In San Felice only very few older people still use the language, while Acquaviva, which historically was considered the cultural center of the Molise Slavs, is situated in the middle between these two extremes, at least as far as everyday usage of Molise Slavic is concerned,
3 See Sasse’s (1992: 61) claim with respect to totally bilingual communities: “The ideal goal over the long term in such a situation is a total isomorphism of the two languages”. On the use of the term “isomorphism” in contact linguistics see Heine (2013). When applied specifically to parallel structures in grammar we also use the term “isogrammatism”, following largely Gołąb (1956; 1959). For the concept of “diachronic constants”, see Guxman (1981) who, however, applied it only to language-internal rather than contact-induced developments. 4 The last 2013 official census resulted in a total number of 1784 inhabitants, after 2081 in 2001 and 4883 in 1951, a reduction that is mainly due to emigration into major centers of Italy but also to other European countries and overseas, especially to West Australia. The actual numbers for the individual villages are the following: Acquaviva and San Felice 674 each, Montemitro 436 (Source: Italian National Institute of Statistics ISTAT).
388
Walter Breu
and shows the effects of language contact most clearly. The grammars of the three villages show considerable dialectal variation, for example through three different forms for the feminine dative singular: -u (Acquaviva), -i (San Felice) and -Ø (Montemitro).5 If not stated differently, our examples come from the dialect of Acquaviva. From a genetic point of view, Molise Slavic belongs to the ŠtokavianIkavian dialect group in the Central South Slavic (Bosnian-Croatian-Serbian) language continuum, contrary to the (I)jekavian and Ekavian Bosnian-Croatian -Serbian standards, e.g., mblika ‘milk’ ≠ mlijeko, mleko and lit ‘summer’ ≠ ljeto, leto, respectively. This and other linguistic characteristics show that, in addition to determining the moment in time in which the separation of the main body of Štokavian speakers occurred, the ancestors of Molise Slavs probably immigrated to Italy from the Neretva valley in Herzegovina (part of the state of Bosnia and Herzegovina today) in the 16th century (Rešetar 1911: 50–139). After the immigration, their South Slavic variety came under the influence of the Molisian dialect of Italian first. After Italy’s unification in 1861, standard Italian acquired its role as an additional donor language. With respect to loanwords we find two types, of which one reflects source forms of the nearby Molisian dialect, while more recent borrowings are nearer to standard Italian.6 As for grammar, most contact-induced changes in Molise Slavic seem to go back to the first centuries after the Slavic settlement in Molise, as they can best be explained from dialectal models. Italian (together with its southern varieties) has always been the only umbrella language for Molise Slavic, while Slavic
5 Diachronically the initial form is -i (coming from the ja-stems), which has been preserved in the San Felice dialect. The ending -u in Acquaviva is introduced analogically either from the masculine dative or, more probably, from the feminine accusative, while -Ø in Montemitro is the result of the phonological rule *i > Ø in word-final position: žen-i, žen-u, žen-Ø ‘wife-DAT.SG.F’. 6 There is, however, also some kind of analogical integration of newer borrowings on the basis of existing loans. For example Standard Italian verbs of the -ere classes (both accented and unaccented) are integrated into the -it-class instead of passing to the phonetically nearer class with infinitives in -et. This integration type ultimately derives from the fact that in Southern Italian dialects the classes in -ere and -ire had collapsed, resulting in an integration of older loans from dialectal Molisian (irrespective of their historical endings in Latin) into only the infinitive class in -i(t), see Molise Slavic lejit (with dialectal phonetics and morphology) unlike Italian leggere ‘to read’. So only this classes in -i(t) had become productive for borrowings in older times (alongside the class in -a(t) which is not of importance here). As a consequence new borrowings from the Italian classes in -ere become part of this class, too, instead of the -et-class, as we would expect from a purely synchronic point of view; e.g., kuredžit (← Italian correggere ‘to correct’). A more detailed analysis of the analogical integration of verbs (proportionally adapted to an older stratum) in situations involving different source languages in Europe is found in Breu (1991: 40–50).
12 Morphosyntactic changes in Slavic micro-languages
389
standard varieties, for example standard Croatian, have never played any substantial role in everyday life in the Molise Slavic villages.7
3 Molise Slavic and the Slavic diachronic constants of grammatical change 3.1 The categories of the Molise Slavic Verb 3.1.1 The category of TENSE Besides the cumulative aspectual-temporal grammemes imperfect and perfect, which will be dealt with in the section on verbal aspect, there are the following tense grammemes in Molise Slavic: the present, two modally differentiated future tenses, two modally differentiated future tenses in the past and the past perfect. All of them – with the sole exception of the present tense – are formed analytically with the help of the clitic forms of the three modal auxiliaries bit ‘to be’, jimat ‘to have’, and tit ‘to want’. They are combined with the infinitive in the future tenses and with the l-participle and a specialized particle in the past perfect. It is this particle that differentiates the past perfect from the (present) perfect, incorporating the l-participle, too. The present tense is of Slavic origin, and so is the volitive future, formed with the help of the clitic present of tit ‘to want, will’. But contrary to the Balkan varieties of Bosnian-Croatian-Serbian, in Molise Slavic the tit-future is restricted to situations marked for “possibility, probability”, while its opposite, the jimat-future of “necessity, obligation” is used for all other cases, including those situations in which plans or fixed time schedules exist. The modern opposition of two modal futures has been induced by language contact, with the jimat-future being a calque from the future tense in the dialects of Southern Italy, especially also of Molise, formed by the auxiliary avé, meaning both ‘to have’ and ‘must’ (Giammarco 1968: 282). The de-obligative meaning of the dialectal future goes back, of course, to its partial meaning ‘must’, which is confirmed by the use of dovere ‘must’ as a future auxiliary in the modern colloquial varieties of Italian in Molise.
7 For an extended overview of the grammatical structures and the lexical characteristics of Molise Slavic together with a large text corpus (commented and fully glossed) of all three Molise Slavic dialects, see Breu (2017).
390
Walter Breu
In Molise Slavic, jimat has also become the overall modal verb for expressing obligation outside the future sphere, replacing in this respect other candidates like the corresponding forms morati, treba in Bosnian-Croatian-Serbian. It seems reasonable to claim that after the Slavs’ immigration to Italy the lexical meaning of South Slavic (j)imat ‘to have’ was initially expanded to include ‘must’ by following the Italian dialectal polysemic model of avé, whose logical consequence was also the development of a future of necessity. This new deobligative future contrasted with the native tit-future from the very beginning, which, as a consequence, led to the latter being kept within the bounds of a future of probability. This procedure of copying an existing polysemy from the model language to the replica language, leading to the isomorphism in question, which we call “adaptation of the semantic structure”, may be symbolized by the two-stage scheme in Figure 1 with respect to the lexical change in jimat ‘to have’.8 The historically initial stage before the change, showing an asymmetry between the polysemy in the dominant language L2 and different expressions for the two concepts (meanings) in the minority language L1, figures here as IS (= initial stage). RS symbolizes the resulting situation after the polisemization in the minority language, being changed in this way to an L1ʹ. The Bosnian-CroatianSerbian form treba ‘must’ is inserted here as a “dummy”, as we are not informed about the form actually used in earlier Molise Slavic:
IS:
L2
avé
RS:
L2
avé
have must
jimat (treba)
L1
have must
jimat
L1’
Figure 1: Polisemization of Molise Slavic jimat ‘to have’.
The subsequent or parallel grammatical development of a new necessitative (de-obligative) future in Molise Slavic, ultimately leading to the new modal opposition within the future formations, is presented in Figure 2 (see Breu 2011b: 156–157).9 Here, again, contact-induced polisemization was the main cause of language change:
8 Further examples from Molise Slavic showing an adaptation of its semantic structure to a polysemic model in the sphere of the lexicon are furnished in Breu (2003: 354–358). 9 As stated by Rešetar (1911: 226) the formation of a “habeo-future”, as he calls it, has nothing to do with similar formations in Old Church Slavonic, which had already disappeared in the
12 Morphosyntactic changes in Slavic micro-languages
must IS:
L2
avé
FUT
RS:
L2
avé
must FUT
jimat tit jimat ≠ tit
391
L1 L1’
Figure 2: Development of the modal opposition in the future by polisemization.
On the whole, the jimat-future is a case of the contact-induced development of a new grammatical opposition (of grammemes) in a contact situation. Examples of the two futures in modern Molise Slavic are presented in (1) and (2)10: (1)
Si riva znat jenu stvaru do tihi if arrive.PFV.3SG know.INF one.ACC.SG.F thing.ACC.SG.F of those.GEN Ndžik ča nasa ubit. PR.NOM.SG AUX.FUT.3SG us.ACC kill.INF ‘If he comes to know anything of it, Ndžik will (probably) kill us.’
(2)
Ono ka je prola, je veča lipo that.NOM.SG.N REL AUX.PRF.3SG go.PFV.PTCP.SG.N is more nice.NOM.SG.N do no ka ma dokj. than that.NOM.SG.N REL AUX.FUT.NEC.3SG come.INF ‘What has passed is better than what will (necessarily) come.
The contact-induced developments of the future have not yet reached their final stage, as there is a certain tendency towards dropping the original Slavic tit-future in Acquaviva. At least some younger speakers seem to replace it more
oldest Serbo-Croatian texts. The Molise Slavic future of obligation must also be regarded separately from the Ukrainian future of the type pysaty-mu ‘write=FUT.1SG’, which contrary to the Romance future formations goes back to the amalgamation of the infinitive with enclitic forms of the present of *jęti/jaty ‘to take’ as an auxiliary, with an inceptive connotation (Danylenko 2011). This means that the Ukrainian future of this type differs not only from the Italian synthetic future in standard Italian scriverò (< scribere habeo) ‘I shall write’ but also from the periphrastic avé-type in Southern Italian dialects. 10 Molise Slavic examples in this chapter are exclusively first-hand examples, coming from recordings done during field research (Breu and Piccoli 2011, 2012; Breu 2017) or, to a lesser extent, from the modern Molise Slavic literature in the Acquaviva dialect, especially from Nicola Gliosca’s novels, for instance Sep aš Mena ‘Sep and Mena’ (2010). They are cited, as a rule, in abridged form or with omissions.
392
Walter Breu
or less completely with the jimat-future, with the result of losing the modal differentiation again and arriving at a 1:1 correspondence of only one future in both languages in contact (see Marra 2005). This means that the contactinduced modal opposition only existed at an intermediate stage of the longterm development. Nevertheless it has been stable for more than one hundred years as is evidenced from the texts in Rešetar (1911), and continues to be functional for most language users, especially if we do not consider “semi-speakers”. Interestingly, Rešetar (1911: 225–226) mentions both futures, but without any hint at their functional differentiation.11 The case of the past perfect seems still more complicated; see below Table 1. Traditionally, the past perfect, lost or very much restricted in usage in most modern Slavic languages, with the exception of Bulgarian and Macedonian, was formed with the help of the imperfect of bit ‘to be’ and the lparticiple of the main verb, which was parallel to the construction of the perfect derived with the help of the present of bit and the l-participle. In spite of the conservation and productivity of both the ancient imperfect and the analytic perfect in Molise Slavic and in spite of the Italian model being formed in a similar way as the traditional Slavic one (i.e., like a present perfect but with the imperfect of the auxiliary instead of the present), the past perfect of Molise Slavic shows a form of its own. Most surprising is the development of the pastperfect particle bi (Acquaviva) or ba (Montemitro) or both (San Felice), being inserted into the perfect formation. Its etymology is not quite clear, but we may assume that it derives from the l-participle of bit, which in its full form Table 1: Past perfect of dat ‘to give’ in the Molise Slavic dialect of Acquaviva. SINGULAR
. . .
masculine
feminine/neuter
zbi da ~ sa bi da zbi da ~ si bi da je bi da
zbi dala ~ sa bi dala zbi dala ~ si bi dala je bi dala
PLURAL
zbima dal (< *sma bi dal) zbita dal (< *sta bi dal) zbi dal ~ su bi dal
11 It is worth noting that, unlike the dialect of Acquaviva, in the more conservative dialect of Montemitro, the tit-future is used far more frequently than the (j)imat-construction (Marra 2005: 154–155), still restricted, as it seems, to certain types of obligation with only secondary reference to the concept of future. Nevertheless direct opposition has been noted even with aged speakers, for example: Ono ke ma do, ja verjem sfe ke če do grubo ‘What will (necessarily) come, I always believe that it will (probably) come bad’ (Breu and Piccoli 2012: 76).
12 Morphosyntactic changes in Slavic micro-languages
393
nowadays is bija (masculine), bila (feminine; also neuter in Acquaviva and San Felice), bilo (neuter in Montemitro) and bil in the plural (with the variant bili in San Felice). If this etymology is true, then the past perfect of Molise Slavic must be derived from a double perfect, functioning as a secondary pluperfect, with two l-participles (analytic present perfect of the auxiliary bit(i) + past participle of the main verb). A double perfect of this type was used in many Slavic languages and still exists in Bosnian-Croatian-Serbian, e.g., bili smo dali ‘(we) had given’. As it seems, this type of double perfect had already fully replaced the traditional past perfect of the type bjesmo dali (imperfect of biti + l-participle) in those parts of Herzegovina which the ancestors of today’s Molise Slavs came from. Otherwise the Italian model would certainly have cancelled out the double perfect, by supporting the use of the ancient type with the imperfect of the auxiliary. Nevertheless, the question remains as to why the Italian model was not copied, which would have resulted in a renewal of the ancient form. It is worth noting, however, that Molise Slavic, though not copying the Italian construction directly, approximated to it by reducing the analytic form of the auxiliary in the past perfect of the double-perfect type to only one inflected form. As a matter of fact, this was the very result of the transformation of the l-participle of bit, originally inflected for gender and number in the l-participle, into an uninflected particle. Molise Slavic even further followed this pathway in reducing the auxiliary complex of its past perfect to a mono-inflectional auxiliary as found in Italian. The most surprising result of this way of adapting was a partial incorporation of the particle into the auxiliary wherever it was possible, i.e., in the bisyllabic forms of the first and second person plural. In the other cases the auxiliary forms were reduced and amalgamated with the particle. Only with the third person singular such an amalgamation was not possible because of the syllable structure of Molise Slavic, disallowing a combination featuring jb- word-initially. In the end, these different reconstructions resulted in a new periphrasis which still does not correspond directly to the Italian auxiliary construction in the imperfect but nevertheless consists of only one inflected auxiliary and an l-participle and nothing else, which could be claimed to be a maximum adaptation without fully copying the model of the dominant language. While in the Acquaviva dialect the insertion of the particle into the auxiliary (a case of infixation, rare in Indo-European) is obligatory, it is optional in Montemitro with older non-amalgamated forms still in use. The paradigm in Table 1 shows the maximally adapted forms of the past perfect in the dialect of Acquaviva together with their more analytic historical source, still in use in the 1st and 2nd person singular as well as in the 3rd
394
Walter Breu
person plural12. The sentences (3) and (4) are examples of the Molise Slavic past perfect: (3)
Je
poša napri jene butege, go.PFV.PTCP.SG.M in.front one.GEN.SG.F shop.GEN.SG.F di je bi sta jenu votu where AUX.PRF.3SG PTL.PQP be.PTCP.SG.M one.ACC.SG.F time.ACC.SG.F s tatom. with father.INS.SG.M ‘He passed in front of a shop, where he had once been with his father.’ AUX.PRF.3SG
(4)
Mi ni z-bi-ma vidil maj vuca fina nonda. we not AUX.PRF-PTL.PQP-1PL see.PTCP.PL never wolf.ACC.PL until then ‘We had never seen wolves before.’
The Molise Slavic past perfect confirms that contact influence in analytic constructions does not necessarily lead to formal calquing by means of loan translation, but that other ways of parallelism may be selected. On the other hand, the conservation and productivity of the past perfect as such, in spite of the conservation of the aspectual system allowing for an alternative expression of taxis relations outside the relative tense, finds its basis in the high productivity of the Italian past perfect as a relative tense (taxis). This influence of the contact language is indirectly confirmed by a relatively restricted usage of the past perfect or even its loss in most Slavic languages outside the contact areas. This means that the past perfect is a good example for the conservation of a historical form due to language contact in contrast to a global tendency towards its loss in the Slavic language family as a whole, constituting one of the diachronic constants of Slavic. The Italian model of using relative tenses for the expression of taxis relations is the reason behind the formation of “futures in the past”, too. These two relative tenses, which had to be introduced into Molise Slavic as a consequence of taking over the Italian principle of the agreement (concordance) of tenses (see Breu 2011b),13 are formed by the imperfect of the auxiliaries tit and jimat in combination with the infinitive. Just like in the case of the two simple futures,
12 By omitting the particle bi in the individual forms, this paradigm also serves as an example for the analytic perfect. 13 An example showing the imperfect biša ‘WAS.3SG’ as the past tense form of the present je ‘is’ agreeing with the perfect is: Za hi držat kalm je njimi reka ka ova biša sama prija gvera ‘In order to keep them calm, he told them that this was only the first war’.
12 Morphosyntactic changes in Slavic micro-languages
395
this does not correspond to the standard Italian future in the past, homonymous with the conditional. Once again the Molise Slavic tenses reflect the Southern Italian construction, in this case the imperfect of the auxiliary avé with the infinitive. Initially, it was this construction which was copied. The creation of a “future of probability in the past” is a language-internal reaction to the asymmetry with respect to the simple futures, going back to this development of the “future of necessity in the past”. This subsequent development shows that even in situations of total language contact internal processes of analogy continue to be at work. Note that in the more conservative dialect of Montemitro only the first step of forming a “necessitative future in the past” formed with (j)imat ‘to have, must’ has been taken, while there is no volitive “future of probability in the past”. A past tense of the tit-clitic does not even exist, unlike čahu ‘will.pst.1SG’, etc. in the Acquaviva variety.14 Whenever obligation is absent, Montemitro simply does not apply temporal concord at all, by using the imperfect instead of a future. Examples for both futures in the past in the Acquaviva dialect are provided under (5)15: Botin čaša bit veča velki aš veča prey.NOM.SG.M AUX.FUT.PST.3SG be.INF more big.NOM.SG.M and more boghat. rich.NOM.SG.M ‘[In the future] the prey would (probably) be bigger and richer.’ b. Napri čuminere čekaša ka Ndžeta before chimney.GEN.SG.F wait.IPRF.3SG COMP PR.NOM.SG.F maša mu ponit džurnal. AUX.FUT.PST.NEC.3SG him bring.INF journal.ACC.SG.M ‘In front of the chimney, he waited that Ndžeta would (necessarily, as planned) bring him the journal.’
(5) a.
Unlike Italian, Molise Slavic has not formed a future perfect (futurum exactum) of the Romance type and it has even dropped – if it had existed at all in the original émigré dialect – the future perfect of the Slavic type formed with the help of the future of the auxiliary bit ‘to be’ and the l-participle. This is probably
14 See for example its absence in the paradigm of tit in the grammar of Sammartino (2004: 98–100), which, contrary to what it says in the title, only refers to the dialect of Montemitro. 15 Full paradigms of the two futures in the past can be found in Breu (2011b: 158).
396
Walter Breu
due to the fact that this tense is rarely used in colloquial discourse in Italian. Actually, Molise Slavic matches colloquial Italian in introducing the perfect instead of the missing future perfect as in (6): (6)
Čaš
ga pitat dòp, AUX.FUT.NEC.2SG him.ACC ask.INF then kada sma čil funera. when AUX.PRF.1PL make.PTCP.PL funeral.ACC.SG.M ‘You will ask him later, when we have made the funeral.’
3.1.2 The category of ASPECT Molise Slavic has a double aspect system, consisting of the typically Slavic “derivational type”, the so-called opposition of perfectivity, expressed in all verb forms, and an “inflectional type”, restricted to the past, traditionally based on the opposition of imperfect vs. aorist, but transformed in Molise Slavic into a formally more complex opposition of (inflected) imperfect vs. (analytic or, periphrastic) perfect. First of all, one should keep in mind another Slavic diachronic constant, which was mentioned in the introduction, saying that if only one component of the old inflecting opposition of imperfect vs. aorist is lost, it is always the imperfect. Even in those Slavic varieties outside the contact areas, where both inflecting forms were replaced by the perfect, the imperfect was always to disappear first (Breu 2011a: 442–444, 2011b: 162–169, Broj [Breu] 2006: 71–74). In Molise Slavic, however, we find the opposite development with the imperfect still being highly productive, while the aorist has been replaced completely by the perfect. On the other hand, a look at the history of the Romance languages shows that here, whenever one of the two components of the inflectional aspect category was lost, it was always the aorist. Thus, we could call this type of reduction a “Romance diachronic constant”. The neighboring Italian dialects of Molise Slavic underwent this type of reduction (which originally came down from Northern Italy along the Adriatic coast) and influenced the Slavic minority language in such a way that – instead of following the Slavic diachronic constant of losing the imperfect first – it copied the aforementioned Romance reduction type. In other words, language contact made Molise Slavic to replace its “genetic” diachronic constant with that borrowed from the foreign model language. In this respect, we could say that Molise Slavic stopped behaving like a Slavic language. That this development is indeed a result of the language contact with a Romance variety is confirmed by the Resian micro-language in
12 Morphosyntactic changes in Slavic micro-languages
397
Northern Italy, showing the same preference for losing the aorist before the imperfect. In this case the model for the non-Slavic development came from the Romance dialects in the region of Friuli – Venezia Giulia. Although nowadays – contrary to Molise Slavic – the perfect in Resian (just like in Slovene or in Russian) is the only productive past tense, we know that this is a relatively recent development patterned on a historical stage corresponding to that in Molise. As a matter of fact, there are still many formal relics of the imperfect, while the aorist forms have completely disappeared (Breu 2011a: 445; Breu 2011b: 175–177; Broj [Breu], Pila, and Šol’ce [Scholze] 2017: 71–72). Apart from this succession of mergers violating a diachronic constant of Slavic, the coalescence of the aorist with the perfect in itself might be viewed as a case of adaptation of the semantic structure by copying the polysemy in question from the neighbouring Italian dialects, but only if the necessary polysemy of the perfect (for aoristic and perfect functions) already existed at the time of the Slavic immigration in Lower Molise. Alternatively, this development could have happened simultaneously in both the Slavic and the Romance varieties of Molise in compliance with the Romance diachronic constant. An independent merger in Molise Slavic is excluded, precisely because in this case the imperfect would have been lost instead of the aorist, according to the diachronic constant of Slavic in question. With respect to the other aspect category in Molise Slavic, expressed by pairs of perfectives and imperfectives, i.e., by “grammatical derivation” with the help of suffixes, prefixes and suppletion, the opposite holds true. All the morphological means needed to form aspectual pairs continue to exist (Breu 2006: 74–75). We find, for example, prefixation in sijat/posijat (ipf./pf.) ‘to sow’, krest/ukrest (ipf./pf.) ‘to steal’, suffixation in kupit/kupivat (pf./ipf.) ‘to buy’, umbrit/umirat (pf./ipf.) ‘to die’, and suppletion in vrč/mečat (pf./ipf.) ‘to put’ or hot/pokj (ipf./pf.) ‘to go’. This “derivational” aspectual opposition encompasses all verb forms, including the imperfect and perfect with the “contradictory” combinations of a perfective imperfect and an imperfective perfect having functions of their own. The sentences under (7) are examples of all four possible combinations, with an imperfective imperfect in (7a), a perfective perfect in (7b), a perfective imperfect in (7c) and an imperfective perfect in (7d).16
16 The main functions of the imperfective imperfect and the perfective perfect correspond to those of the imperfective and the perfective l-past respectively, for example in Russian or Croatian. This means that the first combination expresses, for instance, processual, habitual, and static situations, as in (7a), while the perfective perfect fits into the realm of single total (completed) states of affaires; see (7b). Yet the perfective imperfect as in (7c) expresses, among
398
Walter Breu
(7) a.
One kumbatahu sendza elmeta. they fight.IPFV.IPRF.3PL without helmet.GEN.SG.M ‘They fought / were fighting without a helmet.’ b. Ja sa pa nazjamu, sa I AUX.PRF.1SG fall.PFV.PTCP.M on.the.earth AUX.PRF.3SG zvenija. faint.PFV.PTCP.3PL ‘I fell to the ground, I fainted.’ c. Večaru, kada ostanahma rabi, jope nasa in.the.evening when stop.PFV.IPRF.1PL work again us vamjahu. take.PFV.IPRF.3PL ‘In the evening, when we stopped working, they used to take us again.’ d. Sa rabija na gošta mbača jene AUX.PRF.1SG work.IPFV.PTCP one year at ART.INDF.GEN.SG.F magine. engine.GEN.SG.F ‘I worked for one year on an engine.’
Even in loan verbs derivational aspect is fully productive, though restricted to suffixation as the only morphological means of forming aspectual pairs. The integration rule is very strict: all telic verbs are integrated as perfectives, forming an imperfective partner with the help of the suffix -iva-, e.g., calare ‘to sink’ → kalat (pf.) => kalivat (ipf.), promettere ‘to promise’ → prmitit (pf.) => prmičivat (ipf.), partire ‘to depart’ → partit (pf.) => parčivat (ipf.); for more details, see Broj [Breu] (2006: 82–85), Breu (2017: 65–66). The reason for the stability of the opposition of perfectivity comes from the very absence of such a category in the dominant varieties. Again, we find the same situation in Resian, with the full conservation of
other things, habitually repeated sequences of events, while the imperfective perfect in (7d) is used for the temporal delimitation of atelic actions and states (see Broj [Breu] 2006, 2014). Out of all these forms only the perfective imperfect may be replaced with another form (the imperfective imperfect) without any major change in its meaning, with only the marked reference to the subsequence of situations being lost. But it is impossible to replace the imperfective perfect as in (7d) with the imperfective imperfect, which – in accordance with the Italian imperfect – does not allow for the expression of a situation delimited in time. It is equally impossible to replace the imperfective imperfect in (7a) with the imperfective perfect since the Molise Slavic imperfective perfect – just like the Italian perfect – never refers to unlimited processual, habitual or static situations, contrary to the imperfective l-past in other Slavic languages, e.g., Bosnian-Croatian-Serbian or Russian.
12 Morphosyntactic changes in Slavic micro-languages
399
aspectual pairs and the high productivity of the suffixation type found in borrowings (Benacchio 2009: 189–190). In contact situations of Slavic minorities with German, the situation is different. In colloquial Upper Sorbian, for instance, most telic German verbs are integrated biaspectually. Here the suffix -owa-, historically related to Molise Slavic -iva-, serves as an integration suffix without any aspectual properties. In the relatively rare cases of aspectually paired borrowings they are characterized by prefixation, probably due to the telicizing function of German prefixes (a characteristic absent from Italian as the contact language of Molise Slavic). German does not have grammatical aspect, neither of the inflectional nor of the derivational type. This is why the opposition of perfectivity has been changed with respect to the inherited Slavic vocabulary, too. Even though pairs of verbs of the Slavic derivational type continue to exist not only in standard Upper Sorbian but also in its colloquial variety, colloquial Upper Sorbian transformed their usage in such a way that a new category was born, the grammatical opposition of telicity, strongly influenced again by the characteristics of prefixation in German (Scholze 2008: 230–255). Unlike lexical telicity, the distribution of the partners in these pairs follow grammatical rules, which often make the use of “perfective” (= telic) or “imperfective” (= atelic) verbs compulsory. This means that the German influence on Slavic aspect in Sorbian resulted in something entirely new with components from both sides. Burgenland Croatian has not (yet?) developed such an opposition. But it is similar to colloquial Upper Sorbian in that it only uses prefixes for the secondary formation of aspectual pairs in borrowings, again as a consequence of the role of the prefixes in German. In spite of the full conservation of the derivational aspect in Molise Slavic, we find clear evidence of its subordination to the inflectional category, whose components are supported by their Italian equivalents. First of all, we see that the usual development of the imperfective perfect to a grammeme expressing “process”, “iteration” and “habituality”, as in Russian and Croatian, was completely blocked, presumably because these functions are incompatible with the perfect in Italian, where only the imperfect expresses these functions. Another piece of evidence comes from the treatment of the so-called inceptively static verbs, in which, for example in Russian, the perfective verb (irrespective of the type of derivation of the aspectual pair) expresses the beginning of the state denoted by the imperfective verb, e.g., in videt’/uvidet’ ‘to see’ (derived by prefixation) and ponjat’/ponimat’ ‘to understand’ (derived by suffixation, with a subsequent root alternation). In Molise Slavic such verbs are normally biaspectual on the derivational level, while the opposition between the state and its beginning is expressed through the inflectional aspect
400
Walter Breu
(imperfect vs. perfect), just like in Italian. The following Molise Slavic examples come into consideration: vidahu ‘I could see’: sa vidija ‘I saw, caught sight’, kapahu ‘I understood, was in the state of understanding’: sa kapija ‘I understood, comprehended’, znadahu ‘I knew’: sa znaja ‘I learnt, came to know’, and even jimahu ‘I had’: sa jima ‘I got’, just like in Italian, vedevo: ho visto, capivo: ho capito, sapevo: ho saputo and avevo: ho avuto, respectively. Apart from the changes in the derivational and inflectional aspect oppositions we also find contact-induced developments in the field of periphrastic constructions. Here the syntagmatic principle of formal calquing is at work, to wit, the loan translation of single parts of the construction. In Italian there is a progressive, formed by the auxiliary stare ‘to be, to stay’ and the gerund, while southern Italian dialects prefer other constructions. In the latter case, the most widespread type is the use of the grammaticalized particle mo ‘now’ in combination with the present in order to express ongoing actions. Molise Slavic has copied this model, according to the syntagmatic principle of language contact by translating mo with its own particle sa ‘now’, whence, e.g., sa gre ‘s/he is coming’. A periphrastic model was also copied in the case of the “imminentive”, expressing actions at the point of being realized or in their very initial phase, e.g., stoji za partit exactly translating the Italian periphrasis sta per partire ‘is about to leave’, literally ‘is for leave’. The same holds true for the corresponding forms of the imperfect, in this case stojaša za partit, calquing Italian stava per partire ‘was about to leave’. Just like in Italian all these periphrases individualize (grammatize) specific functions of the highly polyfunctional imperfect.
3.1.3 The category of MOOD Within the category of mood the most important changes have taken place with respect to irreality. Like all other Slavic languages, Molise Slavic has an analytic conditional, expressing, in particular, both potential and counterfactual situations.17 It is used either independently as in (8a) or in both parts of the hypothetical period as in (8b)18:
17 There is some terminological confusion in the sphere of unreal moods. In this chapter, we adopted a terminology in which “irreality” is the opposite to “reality” and serves as a cover term for unreal situations that could potentially still be fulfilled and counterfactual situations, where this is not the case. 18 Example (8a) contains a still accomplishable (potential) condition, while (8b), referring to a situation in the past, is counterfactual. Without the wider context, which is omitted here,
12 Morphosyntactic changes in Slavic micro-languages
401
(8) a.
Kisa benja si bima si vazal who.knows perhaps if AUX.COND.1PL REFL.DAT take.PFV.PTCP.SG.M oni dite? that.ACC.SG.M child.ACC.SG.M? ‘How is it, if we possibly would adopt that child?’ (= What about adopting that child?) b. Ona si bi bila drita bi jiskodila she if AUX.COND be.PTCP.SG.F right AUX.COND pass.PFV.PTCP.SG.F vrata! door.ACC.PL.M ‘If she had been intelligent, she would have passed the door!’
Although Italian differentiates between counterfactual and potential conditions,19 this model could not be transferred to Molise Slavic, as it would have required the contact-induced disambiguation of a polysemy in L1. But as it seems, even in situations of total language contact, new oppositions cannot develop on their own. They are only possible as a by-product of copying a polysemy in other fields. This was the case with the above-mentioned development of the de-obligative future as a secondary result of copying the polysemic model of avé which included also its function as a future auxiliary besides its two lexical meanings of ‘to have’ and ‘must’.20 In the given case the impossibility of introducing a new differentiation becomes even more surprising as Italian has a specialized form for the counterfactual that, indeed, has been copied by means of the adaptation of the semantic structure. We are referring to the imperfect, which in colloquial Italian, besides its aspectual-temporal indicative function of expressing an imperfective past,
both could be easily understood the other way round. Sentence (8a) is from a text from the conservative dialect of Montemitro, where the bi-conditional is still used frequently, while in Acquaviva it tends to lose its hypothetic functions, which preferably are expressed by the imperfect; see below. 19 In the case of potential conditions standard Italian applies the imperfect subjunctive in the protasis (e.g., fossi ‘I were’) and the simple conditional (sarei ‘I would be’) in the apodosis, whereas counterfactual situations are expressed by the past perfect subjunctive (fossi stato ‘(if) I had been’) and the analytic conditional (sarei stato ‘I would have been’). See also the standard Italian translations of (8a) ‘Chissà, forse, se ci prendessimo quel bambino’ and (8b) ‘Se fosse stata intelligente, avrebbe varcato la porta!’. 20 For the discussion of the difficulties in differentiating polysemic entities of the lexicon in contact situations, see Breu (2003: 358–360). Obviously it is very difficult for the bilingual speakers to accept the necessary monosemization in their L1, so that even borrowing new terms is not a general way for removing cases of polysemy.
402
IS:
Walter Breu
L2
IPRF SUBJ/COND
RS:
L2
IPRF SUBJ/COND
indicative
IPRF
counterfactual potential
bi-COND
L1
IPRF
L1’
indicative counterfactual potential
bi-COND
Figure 3: Transfer of the counterfactual meaning of the imperfect by polisemization.
also appears in counterfactual situations. This polysemy is regularly conferred to Molise Slavic by expanding the functions of the imperfect from the indicative into the field of irreality, too. In Figure 3 this adaptation process is again symbolized in a two-stage scheme, which encompasses not only the imperfect and its functions, but also the existing polysemy of the bi-conditional21. As can be seen from this scheme, the imperfect not only follows the Italian model in becoming polysemic for the imperfective past and counterfactuality, but contrary to Italian this new means of expressing irreality has become just as polysemic as the bi-conditional, in expressing both counterfactual and potential situations; see examples (9a) and (9b) respectively22:
21 The auxiliary of the Molise Slavic bi-conditional is inflected only in the first and second person of the plural (bima, bita), otherwise the uninflected form bi is used. Though the biconditional is formed just like the past perfect with the l-participle and in spite of the partial homonymy of the particle bi of the past perfect and the auxiliary of the bi-conditional, the complete analytic forms are always distinct, as only the past perfect additionally has the perfect auxiliary, e.g., ona bi dala ‘she would have given’ ≠ ona je bi dala ‘she had given’. 22 The aspect used in hypothetical constructions mainly depends on the feature of telicity: for atelic situations imperfective verbs are used, while telic situations have a preference for the perfective aspect. Alleged exceptions of the type čakivahu in (9a) result from the distributive or habitual iteration of telic states of affairs. Thus we could render a situation like (9a) more explicitly by “would have destroyed us all one after the other”. But the imperfective imperfects in (9b) could, indeed, indicate that in the case of potential irreality, referring to the future, the imperfective aspect is preferred also in the case of single telic situations. So possibly the Slavic derivative category of aspect could secondarily allow for a formal differentiation between potential and counterfactual irreality, but, of course only with telic verbs. Special research will have to be done, in order to find an adequate answer to this open question. In any case sentence (9a) demonstrates that the perfective imperfect also has functions outside the abovementioned field of habitual sequences of situations (see Broj [Breu] 2014: 345–347).
12 Morphosyntactic changes in Slavic micro-languages
403
(9) a.
One dojahu s jenme kararmatam those.NOM come.PFV.IPRF.3PL with ART.INDF.INS.SG.M tank.INS.SG.M a nasa čakivahu tuna! and us.ACC destroy.IPFV.IPRF.3PL all ‘Those would have come with a tank and would have destroyed us all!’ b. Si sa skinjivahu kude kundrada “Pietrafitta”, if REFL descend.IPFV.IPRF.3PL towards location.NOM.SG.F PR.NOM.SG.F rivivahu prije di fundica. arrive.IPFV.IPRF first where fountain.NOM.SG.F ‘If they descended at the location of “Pietrafitta”, they would arrive at the fountain first.’
Even if the bi-conditional were to disappear as a superfluous synonym in the field of irreality – and there is a certain tendency of this occurring – the imperfect will continue expressing its polysemy in the realm of irreality.23 This case actually shows that there are, indeed, differences in the grammatical structure of two languages in contact that cannot be levelled out, due to an existing polysemy in L1. In other words, an adaptation of the semantic structure of two languages in contact only happens in the case of a polysemy in L2, while a polysemy in L1 prevents it. On the contrary, the polysemy in L1 leads to corresponding mistakes in Italian in this area: By transferring the function of potentiality to the Italian imperfect, this form becomes as polyfunctional as the genuine Slavic conditional (Breu 2011b: 175), meeting in its own way the requirement of maximal isomorphism as the prerequisite for an ideal diagrammar.
3.2 The categories of the Molise Slavic noun 3.2.1 The category of CASE In principle, Molise Slavic has preserved the Slavic category of CASE, except for the old vocative and the locative. While the vocative disappeared elsewhere in
23 This does not necessarily mean that the bi-conditional will disappear completely, as it also has functions outside the field of irreality, for example in rendering the Italian conditional in its function of politeness, e.g., vorremmo = bima til ‘we would like’ (conditional of volere = tit ‘to want’), for example: Mi bima sa til fermat odeka vičeras ‘We would like to settle here this evening’. In addition, it also functions as a modally neutral alternative to the two aforementioned modal futures in the past, in this case by copying one of the functions of the Italian past conditional, e.g. ona bi dola, Ital. sarebbe venuta ‘she would have come [in the future]’.
404
Walter Breu
Slavic, for instance in Russian or Slovene, the loss of the locative needs an explanation. We cannot simply say that losing this case is a reaction to Italian not having it. This should be true then for the instrumental, too, or even for the entire case system. The reason for the full coincidence of the locative with the accusative, found in the dialects of Acquaviva and San Felice, lies deeper. It comes from the Italian polysemy for the indication of “place” and the “motion towards a place”, for example, with in Italia meaning both ‘in Italy’ and ‘to Italy’. Traditionally, the regular way of expressing indications of place in modern Slavic is by means of the locative, while the indication of motion to a place is expressed by the accusative as, for instance, in Russian v Italii ‘in Italy.LOC’ vs. v Italiju ‘to Italy.ACC’. The Italian model has led to the formal merger of this opposition in Molise Slavic,24 according to the principle of the adaptation of the semantic structure in the replica language; see Figure 4.25 Because of the wider field of functions the accusative had in comparison with the locative, the forms of the accusative survived, and those of the locative disappeared, with the exception of some fossilized adverbial constructions.
IS:
L2
ACC
RS:
L2
ACC
state motion
LOC ACC
L1
state motion
ACC
L1’
Figure 4: Loss of the locative in Molise Slavic.
The examples in (10) demonstrate this contact-induced ambiguity of the accusative in Molise Slavic, with a movement towards a target in (10a) and the indication of a state (place) in (10b):
24 The Molise Slavic development is, of course, different from the general loss of the category of case in the Balkan languages, as in this micro-language only the opposition between place and motion is lost, whereas the case system as such continues to be used. 25 For the sake of simplicity the polysemic case in this scheme is called “accusative” in caseless Italian, too. But for bilingual speakers even some psychological reality seems to be involved in such an equation as Molise Slavs express isolated nouns in the accusative, when translating them from Italian to Molise Slavic. They say for example ženu (ACC.SG) instead of the nominative form žena, when asked for a translation of the (isolated) word (la) donna ‘(the) woman’.
12 Morphosyntactic changes in Slavic micro-languages
405
(10) a. Nosahu divojku u crikvu. carry.IPFV.IPRF.3PL girl.ACC.SG.F in church.ACC.SG.F ‘They used to accompany the girl (= the bride) to the church.’ b. Z-bi-ma prisegl u crikvu, però. AUX.PRF-PTL.PQP-1PL marry.PTCP.PL in church.ACC.SG.F however ‘We had married in the church, however.’ In the dialect of Montemitro the situation is more complicated. In fact, we find the same merger of the locative with the accusative with an exception of the feminine nouns, which today belong exclusively to the historical a-declension in all the Molise Slavic dialects.26 It is the traditional homonymy of the locative with the dative in this declension that has blocked its formal coincidence with the accusative in the Montemitro dialect. This is true only with respect to the form itself, however, as the adaptation of the semantic structure, i.e., the formal transfer of the polysemy of “place” and “motion to a place” of the dominant model language to the replica one has occurred in this case as well. This means that the “task” of merging the expression of the two functions in question was fulfilled also in this special case, with the homonymy of dative and locative preserved in this declension, by transferring the function “motion to a place” in the opposite direction, namely, from the accusative to the dative (locative). As a matter of fact, the pressure coming from the dominant model was so strong that even the resulting complication within the case system as a whole was adopted, to wit, the joint expression of “place” and “motion” by the accusative in one declension and by the dative in the other. This situation is further complicated by the fact that the functions of “place” and “motion” of the dative are restricted to non-attributed feminines, while an attribute obligatorily entails its regular replacement by the accusative. We might speculate that in the future Montemitro’s non-attributed feminine nouns will follow the rest in also using the accusative throughout. But this is only a hypothesis about possible processes of internal restructuring not directly connected with language contact. Things could develop quite unexpectedly, as was the case in Acquaviva. As a matter of fact, what looks like an overall formal dominance of the accusative here is only a synchronic surface phenomenon. Again it is the (feminine) a-declension to disturb this pattern, at least from a historical point of view, by bringing about a merger of the ending of the dative with that of the accusative. In this way, both problems have been resolved.
26 For the transformation of the old system of declensions intersecting with gender towards a purely gender-based inflectional system of the nouns in Molise Slavic, see Section 3.2.2.
406
Walter Breu
The system acquired the joint expression of “place” and “motion” in one form and the different behavior of the functions of the cases dependent on the declension/gender class. However, all this was achieved to the detriment of a linguistic constant of Slavic, violated by such a merger of dative and accusative in part of the nominal paradigm. The dialect of San Felice, on the other hand, does not show any of these complications. Here the accusative expresses “place” and “motion” in both declensions/genders with its original endings, i.e., with dative and accusative remaining heteronym throughout the nominal system.27 As for the vocative, it has merged with the nominative completely. Given the productivity of the vocative in the modern Štokavian varieties as used in the countries of Ex-Yugoslavia, this coalescence can also be attributed to the influence of the corresponding polysemy of functions of the nominative in Italian.28 Apart from the loss of the locative and the vocative, there is at least one more development which makes the Molise Slavic case system interesting in terms of language contact, namely the new roles of prepositions. Contrary to the procedure of the adaptation of the semantic structure, affecting the paradigmatic structure of the system of the replica language, we are dealing in this case with the syntagmatic pressure coming from the model language, resulting, again, in a type of loan translation. Unlike the Balkan Slavic languages Bulgarian and Macedonian, which due to their language contact of the substrate type29 lost entirely their nominal inflection by replacing it with prepositional phrases, Molise Slavic did not lose its case endings. Nevertheless the Italian prepositional model, actually, exerted its
27 For more details about the complex interplay of dative, accusative, and locative in developing the modern case systems of the three dialects, see Breu (2008). 28 A new type of the vocative, however, is emerging, based on a procedure copied from colloquial Southern Italian (see Rohlfs 1966: 448–449), namely, the omission of the final syllable or consonant in proper names, for example Duna’! (vs. Dunat NOM.SG) = Italian Dona’! (vs. Donato) or Li’! (vs. Lina NOM.SG). This development is, of course, independent of a similar phenomenon found in proper names of the a-declension (both genders) of colloquial Russian, i.e., the omission of the final vowel (see Comtet 2003): Nataš! (vs. Nataša NOM.SG.F), Volod’! (vs. Volodja NOM.SG.M). 29 We are speaking here about the supposed language change from Balkan Latin to Balkan Slavic (Bulgarian/Macedonian) in the early Middle Ages. For the two opposite types of contactinduced language change, which we call the adstrate/superstrate type (L2 is active), found for example in Molise, and the substrate type (L1 is active), found in Balkan Slavic, see Breu (2011a: 430–431, 2011b: 151–152).
12 Morphosyntactic changes in Slavic micro-languages
407
influence on Molise Slavic by adding prepositions to the nouns with preserved case endings. This was clearly the case in the instrumental, resulting in an obligatory preposition s ‘with’ for this case when expressing the “means” with which an action is being carried out, e.g., s nožam (knife.INS.SG.M) ‘with (the) knife’, corresponding exactly to Italian con il coltello. Actually, this development was doubly motivated, as it also copes with the pressure coming from a semantic asymmetry within the realm of the instrumental functions of the two languages in contact. As a matter of fact, already before the addition of the preposition s to the “instrumental of means” the corresponding prepositional phrase existed in Molise Slavic. But it only had a comitative function, roughly meaning “together with the knife”, which the corresponding Italian phrase ambiguously also expressed. So, the syntagmatic change in question was automatically accompanied by an adaptation of the semantic structures. In other words, Molise Slavic adapted to the polysemic model of the Italian with-phrase, in our example con il coltello ‘by means of the knife / together with the knife’, by expanding the meaning of its own prepositional phrase.30 Actually, the adaptation of the semantic structure would only have led to a merger of the expression of both functions be it with or without a preposition in both cases, but the principle of the syntagmatic adaptation to the structure of the model language led to the dominance of the prepositional phrase with the pure instrumental becoming obsolete.31 In the genitive we also find a preposition, but contrary to the instrumental it is optional. The preposition used here is do, which in all other Slavic languages means ‘until, towards’, while in Molise Slavic it corresponds to the Italian di ‘of, from’. The usual explanation for this semantic change comes from the phonetic similarity of dialectal də with Slavic do, leading to an amalgamation of form and
30 See Nomachi and Heine (2011) for a development parallel to Molise Slavic in minority varieties influenced by German. Danylenko (2015) proposes an alternative explanation of comitative-instrumental polysemy, based on the fact that it is also found in Slavic varieties outside the areas in which contact has a strong influence. We do not exclude a general tendency towards analytic constructions nor do we deny an overall possibility of the comitativeinstrumental distinction being lost, but we claim that language contact is one of the criteria favoring the manifestation of such a tendency, especially when a clear analytic model exists in the dominant contact language. 31 There are some remnants of the instrumental used without a preposition in adverbial function as in putam (road.INS.SG.M) ‘on the road’, referring to a motion taking place on the road, in Italian per la strada.
408
Walter Breu
meaning with the inherited preposition *od ‘of, from’. Be that as it may, the Molise Slavic expansion of the genitive to an analytic construction of the type hiža do jene žene ‘the house of a woman’ corresponds syntagmatically to the Italian la casa di una donna, but, just like in the instrumental case, the genitive ending is preserved. The optionality of the preposition do in the genitive could be an intermediate stage before reaching the final stage of an obligatory preposition that we find in the instrumental. In addition to the nominative and the accusative, lacking a preposition in the model language as well, only the dative remains without a preposition, and not even an optional one. In Italian the case functions, expressed by the dative in Molise Slavic, always require a preposition accompanying the nouns in question, normally a ‘to’. As of now, it is hard to explain convincingly why the dative fails to follow the model of the genitive and the instrumental, though, for instance, na ‘on’, used in Balkan Slavic, would have been a candidate in Molise Slavic, too. In a way, the very preposition a in Italian could give us a clue, as it is absent in the local Romance dialects in many cases in which the standard has it. The most striking fact is again the indication of “place” and “motion to a place”, in this case with respect to place names. There are, indeed, place names which combine with a fossilized preposition, normally well-known villages or towns of the neighbourhood like na Palatu ‘in/to Palata’, u Napulu ‘in/ to Naples’. But the only really productive means is Ø (no preposition), for instance, grem/stojim Kambavaš ‘I go to Campobasso/I stay at Campobasso’, corresponding to the Italian vado/sto a Campobasso and even grem/stojim Mundimitar ‘I go to Montemitro/I stay at Montemitro’ as a variant of grem/stojim na Mundimitar. As is demonstrated by the feminine nouns, having distinct forms for the nominative and accusative, the place name itself appears in the accusative: grem/stojim Čelendzu (nom. Čelendza ‘Celenza sul Trigno’). All in all, Italian a corresponds to Ø in many such cases, so it might not have been conceived as a model meant to expand the Molise Slavic dative through a preposition either. There is also a syntactic reason, giving the dative a special position among the three cases in question, namely, object doubling, which could also have influenced its special treatment with respect to prepositions; see Section 4.3 below. The data from the Molise Slavic case system show therefore that the cases with a prepositional model in the dominant language do not copy this model in the same way. Apart from the full preservation of the case endings, there is a scale which ranges from complete copying (instrumental) to partial copying (genitive) to no copying at all (dative).
12 Morphosyntactic changes in Slavic micro-languages
409
3.2.2 The categories of GENDER and DECLENSION Molise Slavic has restructured its system of declensions and genders.32 As far as declensional classes are concerned, the most radical development is the complete loss of the old feminine i-declension (zero-ending in the nominative singular). Contrary to a possible language-internal gender-based change of simply passing over to the productive feminine a-declension, language contact was decisive for the behavior of individual nouns in Molise Slavic. Whenever a feminine i-noun had a masculine equivalent in the dominant contact language, it changed its gender to masculine and entered the masculine declension with the zero ending in the nominative-accusative singular. In the opposite case, it remained feminine and passed over to the a-declension, for example with a new nominative in -a, instead of its original zero-ending. Examples are kost ‘bone’ and peč ‘oven’, having become masculine because of Italian osso M, forno M, contrary to *noč > noča ‘night’ and *stvar > stvara ‘thing’, keeping their feminine gender due to the Italian notte F, cosa F, but having changed their endings (see Breu 2013: 99–103).33 As for the grammatical category of gender itself (defined in agreement with attributes, predicates and pronouns), the neuter was completely lost in nouns.34 The model for this was, of course, the Italian two-gender system. Again neutral nouns did not simply pass over to the masculine gender, as one might presume, given their identity of endings in most cases. Nor did they simply assume the gender of their Italian equivalent, as one might have expected, given the behavior of the nouns of the former i-declension. The whole picture, however, is not so simple (Broj [Breu] 2013: 94–97). While in the Montemitro dialect all former neuters, indeed, became masculine,
32 For a detailed presentation of the individual processes involved in creating a pure twogender declension system, which we will sum up only briefly in the remainder, see Broj [Breu] (2013). 33 A tendency towards the reduction of nouns inflecting according to the old i-paradigm also exists in other Slavic languages, for example in Ukrainian or in Bosnian-Croatian-Serbian. On the other hand, we find the substitution of individual endings of the i-declension with those of the a-declension, e.g., in the genitive singular of Upper Sorbian. But to the best of our knowledge, there is no other Slavic language or variety which shows a complete transition of former feminine i-nouns into the o- or the a-declension triggered exclusively by the gender of their correspondents in the contact language. 34 In other spheres the neuter was preserved, especially in substantivized adjectives and adverbs, but also in impersonal sentences. The preservation of the third gender in these cases is also based on language contact, in particular on the model offered by Italian varieties in Central and Southern Italy (Broj [Breu] 2013: 105–111).
410
Walter Breu
in the Acquaviva variant there are several exceptions which could only partially be interpreted as an analogy to the feminine gender of their Italian equivalents, e.g., *sedlo N > sedla F ‘saddle’ (Italian sella F, Montemitro sedlo M), *zvono N > zvona F ‘bell’ (Italian campana F, Montemitro zvono M). There are clear counterexamples, contradicting any such rule,35 e.g., nebo N > neba ‘sky F’ (≠ Italian cielo M, Montemitro nebo M) or gn(j)izdo N > njizda ‘nest F’ (≠ Italian nido M, Montemitro gnjizdo M). This means that, notwithstanding the contactinduced character of the loss of neutral nouns as such, the subsequent distribution we find today is not a direct consequence of language contact. The other question is why the dialects of Montemitro and Acquaviva behave differently. Here we actually can offer a contact-based explanation. As the aforementioned examples show, we find different endings for the former neuters in the nominative in these two variants. While Montemitro has kept the original ending o, we find a in Acquaviva. This change, however, is not due to morphology but is the result of a contact-induced phonological change of e, o to a in unstressed syllables in Acquaviva and San Felice, the so-called “Molise Slavic Akan’e” (Broj [Breu] 2013: 83). The resulting form -a of the nominative singular ending coincided with the original ending of the feminine nouns, which opened a way (but not a requirement) to a transition from the original declension to the feminine a-declension. The reasons for actually making use of this possibility, missing in the Montemitro variant, have to be explored individually for each noun in question. Apart from the loss of the neuter and the i-declension, there is still another change which eventually turned the Molise Slavic declension into a fully two-gender-based system, the passing over of the masculine nouns that traditionally belonged to the a-declension into the remaining only masculine declension. Morphologically they normally keep their a in the nominative singular, just as most ex-neuters do, but their inflection follows the other masculine nouns, for example, tata ‘father’ (NOM, GEN, ACC), tatu (DAT), tatam (INS), and so on. Italian borrowings of the masculine gender ending in -a normally even replace their nominative form with the unmarked masculine zero-ending, for example, statista ‘statesman M’ → statišt.
35 As a matter of fact, some authors have followed Muljačić (1973: 35) in positing a development based on the gender of the Italian equivalents of the ex-neuters. But, unfortunately, his examples are wrong. For instance, the former neuter meso N > mesa M ‘meat’ (dialects of Acquaviva and San Felice, meso M in Montemitro), contrary to what Muljačić said, became masculine in all three dialects, in spite of the feminine gender of its Italian equivalent carne.
12 Morphosyntactic changes in Slavic micro-languages
411
Unlike the Bosnian-Croatian-Serbian standards and also Italian, Molise Slavic has completely lost its gender distinction in the plural. Yet this effect is secondary to the merger of the nominative, the only form distinguishing masculine and feminine gender in the attributes, with the accusative, a coalescence that follows a diachronic constant of Slavic, realized, for example, also in Russian. This diachronic constant was not blocked but rather spurred by the caseless structure of the Romance contact varieties, showing a model with identical subject and object forms. In addition, this model seems to be an important factor for the rise of a tendency in Molise Slavic to lose the distinction of animacy, based on the differentiation of animate masculine nouns from inanimates in the accusative singular in opposition to the nominative.36 Like in the case of the neuter, the loss of the animate gender represents an instance of the reduction of gender distinctions in accordance to Italian expressing only two of them.
3.2.3 The category of NUMBER With respect to number Molise Slavic is conservative. Just like in Italian, singular and plural continue to exist, and just like in the standard Bosnian-CroatianSerbian varieties or in Russian, the former dual has transformed into a paucal. This paucal was not affected by the Italian two-number system, probably because it is not a free number grammeme, but is governed by the numbers 2–4. It is worth noting, however, that the number category also seems to be especially conservative in contact situations outside Molise Slavic, as is shown, for example, by the preservation of the (free) dual in Slovene and standard Upper and Lower Sorbian, three languages strongly influenced by German without such a number grammeme. Only in very recent times, colloquial Upper Sorbian has reduced its independent dual to a governed form of the numerals meaning ‘two’ and ‘both’ (Scholze 2008: 135–139).
3.2.4 The category of DEFINITENESS Slavic languages outside the contact areas do not have a grammaticalized opposition of definiteness expressed by articles. The absence of an article system
36 By now, special animate accusatives like popa (ACC.SG) ‘preast’, formally coinciding with the genitive, have become optional and are normally replaced by the forms that overlap with the nominative, whence pop.
412
Walter Breu
IS:
L2
UNO
RS:
L2
UNO
one indef. article
JENA
one indef. article
JENA
L1 L1’
Figure 5: Creation of an indefinite article by copying the polysemic model of UNO.
might even be claimed to be a diachronic constant of Slavic.37 Molise Slavic is different in this respect, as it shows an obligatory indefinite article in all the relevant positions in which Italian also uses it. The obvious model for creating an indefinite article comes from the Italian polysemy of the form UNO for the number ‘one’ and the indefinite article, which was copied in Molise Slavic (adaptation of the semantic structure),38 see figure 5. This development took place not only in contradiction to the abovementioned diachronic constant of Slavic, but it also violated a typological universal, according to which an indefinite article emerges only after the creation of a definite article, still missing in Molise Slavic. Interestingly enough, this situation is also due to the mechanism of the adaptation of the semantic structure, in this case in terms of the absence of a polysemic model in Italian. In particular, the Italian article IL does not have a demonstrative function, thus banning even a language-internal creation of a definite article in Molise Slavic from its demonstratives, found in many languages. Such a development would have infringed upon the parallelism between the replica and the model language, by introducing a polysemy in Molise Slavic which is not found in Italian. Nevertheless, the Molise Slavic article system, based only on an indefinite article, became fully functional by reinterpreting the absence of the indefinite article as a “definite zero-article”. This was possible due to the fact that besides pragmatic definiteness (anaphoric, cataphoric and so on), the Italian definite article also expresses innate definiteness (semantic definiteness of
37 Though it is true that Slavic languages outside the areas of strong language contact may have articloids in the sense of demonstratives and the numeral ‘one’ expressing some article functions, it is also true that such articloids have by no means acquired the status of fully grammaticalized articles. The (postponed) definite article in Balkan Slavic is, of course, one of the cases of contact-induced developments, and so is the Macedonian indefinite “article”, still missing some characteristic functions (Breu 2012: 280–294). 38 UNO and JENA symbolize the whole paradigm of their forms, including the short forms na and nu. For the whole range of functions of the Molise Slavic indefinite article compared with Colloquial Upper Sorbian and Macedonian see Breu (2012), presenting also the full paradigm of JENA (Breu 2012: 279). See also Heine (this volume) for their interpretation of the development of the Molise Slavic indefinite article.
12 Morphosyntactic changes in Slavic micro-languages
DEM
IS:
L2
IL
deictic DEM definite article
413
L1
Figure 6: Missing polysemic model in Italian for the creation of a definite article in Molise Slavic.
unique referents) and generic definiteness; see figure 6. Therefore, it did not leave any special function to be claimed by the absence of an article which otherwise would have impaired the supposed model of exactly two articles represented in Molise Slavic by the opposition of na (indefinite article): Ø (definite zero-article). Here is just one example showing the interplay of these two “articles”, with the indefinite article na introducing the rheme zvizda in (11a) and Ø anaphorically referring to it as the theme in (11b): (11) a. Je
ju pala na zvizda her fall.PFV.PTCP.3SG ART.INDF.NOM.SG.F star.NOM.SG.F do zlata na čelu. of gold.GEN.SG.M on forehead.ACC.SG.F ‘A golden star fell on her forehead.’ b. E ova divojka je pola doma, and this.NOM.F girl.NOM.SG.F AUX.PRF.3SG go.PFV.PTCP.SG.F home a Ø zvizda sfitlaša na čelu. and DEF star.NOM.SG.F glow.IPRF.3SG on forehead.ACC.SG.F ‘And this girl went home, and the star glowed on her forehead.’ AUX.PRF.3SG
Just like Molise Slavic, colloquial Upper Sorbian developed an article system, too, in this case by adapting to the semantic structure of the German indefinite article EIN, polysemous with respect to the number ‘one’ and the indefinite article. As a consequence, we find in this variety an obligatory indefinite article JEN, homonymous with the number ‘one’. But in this variety of Upper Sorbian, one also encounters a definite article modelled on the German one whose form DER, unlike Italian IL, has demonstrative functions in addition to that of a definite article. This model was copied by making the existing Sorbian demonstrative TÓN a polysemous form, expressing also the function of an obligatory pragmatic definite article. It did not, however, include the level of semantic definiteness which is still expressed by means of the absence of articles, or, in other words, there is a semantic definite zero-article Ø in opposition with the pragmatic one (Breu 2012: 304–306). We could claim this to be an intermediate stage, but we should keep in mind that in German, contrary to Italian, the absence of articles also has functions of its own, for example, with regard to mass
414
Walter Breu
nouns and partly so in the case of genericity, which could have helped to grammatize Ø as an article of its own. In addition, a system with two definite articles is in no way alien to historical and actual varieties of German, e.g., Bavarian dea (M) ‘the (pragmatically definite)’: da (M) ‘the (semantically definite)’.39
3.3 The category of COMPARISON of adjectives and adverbs Like all Slavic languages, Molise Slavic has a category of comparison for adjectives and adverbs. In this section we will concentrate on the grammemes “comparative” and “superlative”, traditionally expressed by suffixes in Slavic. In general, the synthetic type remains unchanged in most Slavic languages, while others, especially Russian and even more so Balkan Slavic show a clear development towards analyticity. In language contact, a development from synthetic to analytic constructions is defined by the syntagmatic principle. Therefore it is no wonder that the Italian model of expressing the comparative by means of a comparator and the absolute (positive) form of the adjective or adverb has been copied by Molise Slavic, to wit, with the help of the adverb veča ‘more’, corresponding to Italian più, while the Bosnian-Croatian-Serbian standards continue to use the suffix type almost exclusively (see Breu 2009). In contact with German, which contrary to Italian expresses comparison synthetically, the Sorbian varieties40 and also Burgenland Croatian have preserved their suffixed type as well.
39 At first glance, the fact that we find a formal definite article in Resian seems to contradict our line of argumentation with respect to Italian-Slavic language contact (Benacchio 1996). But it should be borne in mind that in the Resia valley, besides Italian and Friulian, German historically also played an important role and could therefore be held “responsible” for the differing developments in Resian and Molise Slavic (Breu 2011a: 441). 40 In colloquial Upper Sorbian there is a stronger tendency than in the standard variety towards an analytic expression of comparison by means of the comparator bóle, the comparative of the adverb jara ‘very’, which corresponds to the Russian comparator bolee, the comparative of očen’ ‘very’, and differs from jace (standard wjace), the comparative of wyle (standard wjele) ‘much’, just like in Russian bolee from the adjectival bol’še ‘more’ (see Scholze 2008: 88–90). Further investigation of this conspicuous correspondence between Russian and Upper Sorbian might be useful. As long as we do not have a contact-related explanation for the Russian analytic comparative, we could perhaps claim that language contact is not the only reason for changing the morphosyntactic expression of the category of comparison in Slavic. However, the Balkan Slavic languages Bulgarian and Macedonian with their fully analytic type of comparison, using the particle po as a comparator, e.g., Bulgarian po-dobăr ‘CMP-good = better’, have to play a role in any discussion about the influence of language contact on the overall development of Slavic comparatives (and superlatives).
12 Morphosyntactic changes in Slavic micro-languages
415
In accordance with Italian, Molise Slavic has also preserved some suppletives. Thus, we find bolji ‘better’ (adjective) and bolje (adverb), corresponding to Italian migliore and meglio respectively, and in the same way gori ‘worse’ (adjective) and gore (adverb), corresponding to Italian peggiore and peggio, and also manji, manje ‘less’, the negative counterparts of veča ‘more’, which just like Italian più does not distinguish between an adjectival and an adverbial form. There are some more instances of isogrammatism with respect to comparison, especially between Molise Slavic and its closest contact variety, the Romance dialects of Basso Molise. Among other things, at this point we could nominate the normal way of expressing the superlative function by means of the formal comparative, again with the definite Ø-article corresponding to the Romance article, for instance, veča velki ‘(the) biggest’ corresponding to Italian il più grande, alongside a more explicit way by means of the synthetic superlative naveča ‘most’, functionally corresponding to Italian il più, plus the positive form, e.g., naveča velki ‘the biggest one’. Just like in the Romance dialects of the neighbourhood, analytic variants or pleonastic extensions are possible for the suppletives, i.e., bolji = veča dobri = veča bolji ‘better’, including a pleonastic variant even in the case of veča ‘more’ = veča čuda (see Italian dialectal più molto, literally ‘more much’). The most striking correspondence, induced by language contact, is certainly the complete replacement of adjectival comparatives (and superlatives) by adverbial ones in the suppletives, for example bolje ‘better’ (and its analytic and pleonastic variants), which now has expanded its original adverbial character to including adjectival use like Italian meglio, replacing the adjective migliore in colloquial speech.41 Like Italian, Molise Slavic has two types of syntactic linkers, the preposition do (with the genitive), corresponding to Italian di ‘of’, and the conjunction ka, corresponding to (or rather borrowed from) Italian che ‘that’. Both can be used with compared nouns, but do refers to subjects, ka to objects and compared properties, a rule perhaps even more consistently observed than in Italian. For the superlative both languages exclusively use the corresponding conjunctions ka, che.
41 See Breu (2008) for more details, including a contact-based explanation for the preservation of the separate adjectival form manji ‘less’ in spite of its absence in Italian, the lack of a synthetic superlative *namanje and a scenario explaining the sequence of adaptations leading to the almost complete isogrammatism between the Molise Slavic and the Italian comparison.
416
Walter Breu
4 Some syntactic characteristics of Molise Slavic 4.1 The position of the attribute Molise Slavic has in many respects adapted to Italian word order, following the syntagmatic principle of contact-induced change. This is especially evident in the case of the position of Molise Slavic adjectival attributes as compared with other Slavic languages, including the Bosnian-Croatian-Serbian standards as its closest relatives. In Molise Slavic, just like in Italian and contrary to Croatian,42 “differentiating” attributes are placed after the noun, for example, adjectives meaning colours like crni ‘black’, crnjeli ‘red’, nationalities like tedesk ‘German’, taljan ‘Italian’ or participles like kuhani ‘cooked’, dizgracijani ‘miserable’. On the other hand, descriptive adjectives, in accordance with Italian usage, should be placed before the noun, e.g., lipi ‘beautiful’. On the whole, postpositive attributes are by far more frequent than prepositive ones. The different positions of the attributes lead in some cases to oppositions of the type Je na brižna žena ‘She is a poor woman’ (pitiable, because something awful has happened to her) ≠ Je na žena brižna. ‘She is a poor woman’ (she has no money) as in Italian: È una povera donna ≠ È una donna povera; similarly, na dobri ljud ‘a good man’ (good character) ≠ na ljud dobri ‘a talented man’. The contact-induced dominant rule of postposing the attribute brought about in some cases “hypercorrect” usage, thus resulting in a preference for the position after the noun where even Italian would place the adjective before it. For instance, in the case of na hiža nova ‘a new house’ only postposition is possible whereas in Italian one encounters a difference between una casa nuova ‘a newly built house’ and una nuova casa ‘a different house’.43 The same is true of na hiža stara ‘an old house’. Other adjectives displaying a clear tendency to postposition are velki ‘big’, mali ‘small’ unlike their Italian equivalents grande and piccolo, where the unmarked position of the attribute is before the noun and postposition gives them a special meaning. Of course, such a discrepancy
42 In general, in Slavic the position of attributes before the noun is basic, while placing them after it, is rare and “symptomatic”. Some instances of unmarked postposition occurring in Polish, however, are probably due to the influence of Latin. For some examples of postpositive attributes in the standard languages, see Uhlířová (2014: 2228–2230). In the areas of contact, postposition may be found more often. This is especially true in cases of contact with Romance languages as, for instance, in Resian. 43 In other combinations, in which ‘new’ cannot be understood as ‘newly built’, both positions are possible, e.g., na žena nova = na nova žena ‘a new wife’, with no difference in meaning.
12 Morphosyntactic changes in Slavic micro-languages
417
can lead to errors in L2 when Molise Slavs speak Italian, like in the abovementioned case of the contact-induced hypothetic imperfect, expanding its functional range from counterfactual to potential irreality. In other words, semantic rules regulating a syntactic alternation in the model language disallow direct calquing of its word order in the recipient language. They correspond to the lexical or grammatical polysemies in L1 impeding the induction of an opposition existing in L2 into L1. In both cases an isogrammatism reducing the differentiating rules of their two languages in the diagrammar of the bilinguals can emerge only from transferring L1 characteristics into L2, which is equivalent to mistakes in the dominant language. This is a clear contrast to the adaptation of the semantic structure of L1 to L2, based on a polysemy in the model language L2, and to direct calquing by loan translation from the model language in cases without specific alternation rules.44
4.2 Position of the clitics Traditionally Slavic clitics follow Wackernagel’s law, i.e., they are placed in the second position after the first word with full accent.45 If a sentence has more than one clitic, they follow each other in a fixed order with the whole cluster of clitics in Wackernagel’s second position. In Italian (and its southern varieties), clitics are verb-adjacent. Without going into greater detail, we can say that they precede the verb in declarative and interrogative clauses, while they follow it in the imperative. The ranking of clitics in Molise Slavic is a rather complicated question that warrants further research. Generally speaking, Molise Slavic has fully adapted to the Italian positioning of clitics, without remnants of the former Wackernagel rule. This happened, of course, in accordance with the syntagmatic principle of contact-induced language change. Thus we find declarative sentences like (12a) and questions like (12b), with clitics preceding the verb, 44 There is a rather similar case of copied word order in total language contact which ultimately leads to complications, namely the effect of the German sentence-final verb position (V-end) on Upper Sorbian. While it was easy to copy the German model, it was impossible for Sorbian to restrict this new word order only to subordinate clauses as German does, while displaying “verb second” position (V2) in the main clause, as in Sorbian subordinate and main clauses traditionally do not behave differently (Breu and Scholze 2006: 46–55). As a result, Upper Sorbian now exhibits V-end as a general rule in its syntax, leading to compensatory mistakes in the German word order of untrained Sorbian bilinguals. 45 For a comparison of the realizations of Wackernagel’s law in Slavic standard languages, see Uhlířová (2014: 2226–2228).
418
Walter Breu
and imperatives like (12c) with clitics postpositioned as in Italian. The clitics are marked with the equals sign pointing towards the verb as their host. For the sake of simplicity the following sentences all contain a loan construction based on the Italian idiomatic phrase andarsene ‘to leave’46: (12) a. Su= bi= si= ga= pol tuna. AUX.PRF.3PL= PTL.PQP= REFL.DAT= 3.SG.N= go.PFV.PTCP.PL all.NOM ‘They all had left.’ b. Si= ga= gre? REFL.DAT= 3.SG.N= go.IPFV.PRS.3SG ‘Is s/he leaving?’ c. Po =si =ga!47 go.PFV.IMP.2SG =REFL.DAT =3.SG.N ‘Leave!’ Unlike variation in the Italian model, in which clitics either follow the infinitive or precede the modal verb, as shown in (13), the Molise Slavic infinitive accepts only proclitics. While blocking postposition could be a special restriction on the Molise Slavic infinitive, the impossibility of putting the clitics before the modal verb seems to be linked to the fact that the modal verb mam in (13), functioning also as the auxiliary of the de-obligative future, is part of the cluster of clitics. (13) Mam= si= ga= po. must.PRS.1SG= REFL.DAT= GEN.SG.N= go.PFV (Italian Me=ne=devo andare ~ Devo andar=me=ne.)48 ‘I must leave’ or ‘I will (necessarily) leave.’
46 In the infinitive (written in one word according to the Italian orthography) this phrase consists of the elements andar-se-ne (‘go’ + reflexive pronoun + partitive particle), giving si ga pokj in Molise Slavic with a different (but nevertheless verb-adjacent) word order. In the present tense we have se ne va ‘s/he is leaving’, giving si ga gre in Molise Slavic. The clitic ga in this phrase is originally the genitive singular neuter of the anaphoric pronoun, but it is used in this construction as an equivalent of the Italian partitive particle ne. This example shows that even if there is no direct equivalent to a syntactically separable form of the model language, the replica language is able to replace it by a functionally similar one. 47 The Molise Slavic imperative corresponds exactly to Italian vattene (va=te=ne)! ‘Leave!’. 48 According to a general rule of the Italian orthography, prepositive clitics are written separately, while postpositive ones are attached to the verb: me ne devo andare ~ devo andarmene. This rule is also responsible for the opposition between indicative forms like se ne va, corresponding to the question in (12b), and the imperative vattene, corresponding to (12c).
12 Morphosyntactic changes in Slavic micro-languages
419
The same restriction allies to the future of probability, formed with the help of the auxiliary tit ‘to want, will’, in which only ču= si= ga= po ‘I will (probably) leave’ would be possible. However, variation is, indeed, possible with non-clitic modal verbs like moč ‘can, may’, i.e., si ga moraš po leč ~ moraš si ga po leč ‘you may go to bed’, corresponding to the Italian variation te ne puoi andare ~ puoi andartene. This means that, while the use of clitics after the infinitive corresponding to the Italian andartene is excluded here, too, they optionally precede the modal. But here we have to take into consideration once again the question of the specific model language in question. As a matter of fact, Southern Italian, the only model for Molise Slavic at the beginning, prefers pre-modal clitics to the detriment of the post-infinitival position which is commonplace in standard Italian. Especially in older times, but to a certain extent also in modern dialects, even the same variation of clitics before and after the modal verbs as in Molise Slavic existed (Rohlfs 1968: 173–174). So it could well be the case that, apart from the special rule determining the clustering of clitics, Molise Slavic copied an older local model and refrained from additionally copying the later model of the literary standard. A more detailed analysis of the restrictions and variations in the clustering of clitics in Molise Slavic, as well as their exact model in the local dialects, must be left for further research. But from what we know now, we could argue that even the combination with the infinitive does not contradict the rule that Molise Slavic clitics, if used in a declarative sentence, are always oriented towards the verb and no longer used in Wackernagel position. So, while the orientation of the cluster of clitics towards the verb has, actually, been copied, the internal ranking of clitics is not necessarily identical in the two languages in contact; see (14) with the perfect auxiliary je at the beginning of the cluster in Molise Slavic unlike the corresponding auxiliary è at the end of the cluster in Italian. (14) Je=
si=
ga=
AUX.PRF.3SG= REFL.DAT= GEN.SG.N=
poša. go.PFV.PTCP.SG.M
‘He left.’(≠ Italian Se n’è andato.) This means that the slot for clitics as such has changed its position in accordance with the Italian model, but inside the cluster clitics still entertain their own ranking, not susceptible to language contact. A reason for this may be that, due to their high degree of grammaticalization, clitics are similar to inflectional endings, and therefore behave similarly in certain respects like bound morphemes, being less influenced by language contact than other syntagmatic
420
Walter Breu
elements. In other words, the internal ranking of clitics is comparable to that of suffixes. Another case of this type is the position of the negative clitic which in Molise Slavic always occurs at the end of the cluster, while the corresponding negative adverb non in Italian precedes it: si ga ne gre ≠ Italian non se ne va ‘s/ he does not leave’. Apart from these peculiarities, the internal structure of the clitics cluster in all the above-cited examples exactly matches the Italian clitic ordering: si-ga = Italian se-ne. There is no doubt that the syntactic change from Wackernagel’s second position to a verb-adjacent position, found in Molise Slavic, also corresponds to the overriding of a Slavic syntactic constant by the corresponding Romance one. We find a similar change in Balkan Slavic, probably also due to the Romance (or Vulgar Latin) influence, especially on Macedonian.49
4.3 Double negation It is a diachronic constant of Slavic that negative pronouns or adverbs are doubled by the negative particle, normally more or less immediately proclitic to the verb. Italian has also a double (pleonastic) negation which, however, has some restrictions, the most important being that negative pronouns or adverbs preceding the verb are not doubled. Such an asymmetry between L2 and L1 is, of course, another area where the syntagmatic principle of contact-induced change could apply, in this case not simply in the sense of the adaptation of word order but with respect to the number of elements concerned. This means that in trying to rectify this asymmetry a special type of formal calquing could be claimed to be relevant that again translates element by element, but with one of the elements transferred as a zero-element (Ø). This is, indeed, what happened. The Italian model requires the omission of the negative particle if a negative pronoun like nessuno ‘nobody’, niente ‘nothing’, or an adverb like mai ‘never’ precedes the verb. This model is copied in Molise Slavic by replacing the traditional negative particle ne, ni- with Ø in such cases, i.e., whenever the corresponding negative pronouns and
49 Cimmerling (2013: 56–57) considers Molise Slavic to be the only Slavic language fully corresponding to a “V-system”, in which chains of clitics are localized in the “verb group” without a fixed place in the sentence, surpassing in this respect even Macedonian. His hypothesis is based on only three single-clitic translations from Italian to Molise Slavic (1 clitic pronoun + finite present), but it is corroborated by the first-hand examples in the present paper with real chains of clitics. However, as evidenced from the examples in Benacchio (2009: 185), the Slovene-based Resian dialect in Friuli is also a candidate for a full “V-system”.
12 Morphosyntactic changes in Slavic micro-languages
421
adverbs nikor ‘nobody’, nišča ‘nothing’, maj ‘never’ and so on precede the verb. Here are two such examples, in which the corresponding Italian translations are added: (15) a. A nikor Ø je vasa pita si and nobody.NOM Ø AUX.PRF.3SG you.PL.ACC ask.PTCP.SG.M if tijahta stat s tedeski? want.IPRF.2PL stay.INF with German.INS.PL (= Italian E nessuno Ø vi ha domandati se volevate stare con i tedeschi?) ‘And nobody asked you, if you wanted to stay with the Germans?’ b. Je bi dža dečidija e nišča AUX.PRF.3SG PTL.PQP already decide.PFV.PTCP.SG.M and nothing.NOM e nikor mu Ø morahu čit kanjat and nobody.NOM him Ø can.IPRF.3PL make.INF change.INF moždane. mind.ACC.PL (= Italian Aveva già deciso e niente e nessuno Ø potevano fargli cambiare idea.) ‘He had already decided and nothing and nobody could make him change his mind.’ c. Maj mangu jena čita merikan Ø never not.even one.NOM.M city.NOM.SG.M American.NOM.SG.M Ø je bi mu para AUX.PRF.3SG PTL.PQP him.DAT seem.PFV.PTCP.SG.M naka lip. so beautiful.NOM.SG.M (= Italian Mai nemmeno una città americana Ø gli era parsa così bella.) ‘Never not even one American city had seemed to him so beautiful.’ Nevertheless, double negation of the Slavic type, showing the negative particle before the finite verb (auxiliary), is still possible and even seems to be preferred. This means that we have a case of variation in which the Italian model is retained only partially as an optional rule; consider sentence (16) containing double negation.50
50 The pleonastic negative particle appears here in the form ni-, combining with some auxiliaries, instead of ne with full verbs. Transformed into the (imperfective) present, the second part of (16) would be: . . .nikor ne rispunjiva.
422
Walter Breu
(16) Su AUX.PRF.3PL
mitraljival, nikor ni-je machine.gun.IPFV.PTCP.PL nobody.NOM not-AUX.PRF.3SG
rispunija. answer.PFV.PTCP.SG.M (≠ Italian Hanno mitragliato, nessuno Ø ha risposto.) ‘They machine-gunned (for a while),51 nobody answered.’ When the negative pronoun is positioned after the verb, like nikor ‘nobody’ in (17a), then, of course, nothing happens, as the traditional Slavic and the Italian constructions coincide, i.e. we regularly find the pleonastic negation before the verb. The same is true for the negative adverb maj in (17b)52: (17) a. Nonda ni-maša nišča nikor! there not-have.IPRF.3SG nothing.ACC nobody.NOM (= Italian Lì non aveva niente nessuno!) ‘Nobody did not have anything there.’ b. Ni-sa ukrela maj nišča nikromu. not-AUX.PRF.1SG steal.PFV.PTCP.3SG.F never nothing.ACC nobody.DAT (= Italian Non ho rubato mai niente a nessuno.). ‘I have never stolen anything from anybody.’ The incomplete adaptation to the Italian model could be again explained by the fact that the contact-induced new rule is not followed consistently inasmuch as it led to a complication of the once simple rule of redoubling. Just like in the case of the contact-induced partial postposition of the attributes, mistakes in the Italian of Molise Slavs, here in the form of undesired redoubling, serve as a kind of remedy for, nonetheless, creating a syntactic parallelism between L1 and L2. We should, however, also consider the hypothetic possibility of local or dialectal Italian usage of the negation being different from the standard, in that it could presumably allow optional doubling with preverbal negative pronouns. But all the examples attested up to now in the neighboring village of Palata show omission of the negative particle like in the standard variety, e.g.,
51 One should remember that the Molise Slavic imperfective perfect never expresses an unlimited process or state in the past (see fn 16 above), but only temporally delimited ones, explained here with “for a while”. An ongoing process would require the imperfective imperfect, i.e., mitraljivahu in this case. 52 Another example of this type, but with only one negative pronoun, is (18b) in Section 4.4 below.
12 Morphosyntactic changes in Slavic micro-languages
423
[niʃʃunə Ø ɛ mmɛʎʎə də issə] (= Italian nessuno Ø è meglio di questo) ‘Nobody is better than this one’.53 The same holds true for the relevant examples recorded in Apulia by Poletto and Rasom (2006: 81–85), which indirectly confirm Rohlf’s (1968: 215) observation that in the northern dialects doubling in the case of a preverbal negative pronoun is optional, while he does not mention any southern dialects in this respect. In any case the omission of the negative particle in the cases in question is not a recent phenomenon. Rešetar (1911) cited it in several sentences with verbpreceding nikor ‘nobody’, for instance (in Rešetar’s own transcription): Nikor d’onihi mičiciji, ka on je bi mitiva, Ø zovaše ńega’ (=Italian Nessuno di quegli amici che aveva invitato Ø lo chiamava) ‘None of those friends he had invited called him’. As a matter of fact, the omission was just as optional as today; consider the following example with preserved pleonastic negation, contrary to Italian: tȗna kačatúr su-potégnili ȍvu kóz – nȉkor nìje ju-kȍlij! (≠ Italian tutti i cacciatori spararono a questa capra – nessuno Ø la colpì!) ‘All hunters shot at this goat – nobody hit it’. From all this follows that both possible models as found in standard and Southern Italian, were only partially copied by Molise Slavic, with mistakes occurring in Italian as the secondary language of the Molise Slavic bilinguals.
4.4 Object doubling The last instance of syntagmatic copying we will briefly deal with is the doubling of nouns and pronouns in object position by means of clitic short pronouns, which, in some respect, seems to mirror what has happened in the case of negation in the opposite way. Once again we do not have to deal with a simple reordering of words but with a different number of elements in the two languages in contact. In the case of object doubling contrary to negation it was Molise Slavic that had fewer elements than the dominant model language. The adaptation of L1 to L2 is therefore tantamount to the addition of elements in the replica language. As is well-known, (pleonastic) object doubling is a feature of Bulgarian and Macedonian and also of other languages of the Balkan Sprachbund, following special rules triggered by features like object type (direct, indirect), position of the object and definiteness. The Bosnian-Croatian-Serbian standard languages do not share this property. Molise Slavic, on the other hand, also has object
53 My thanks go to Domenica Catino for this and other examples from Palata.
424
Walter Breu
doubling, but it clearly differs from the Balkan Slavic rules. This is, of course, due to its specific contact situation. As a matter of fact, object doubling is a property of Southern Italian and – with different rules – of standard Italian. While standard Italian, in principle, only doubles objects that precede the verb, Southern Italian allows for doubling also in other cases (Rohlfs 1968: 169). Both models are copied by Molise Slavic, revealing a result that lies in between both rules: obligatory doubling in the case of objects preceding the verb and optional doubling in other cases. Such a situation inevitably leads to mistakes in the Italian of Molise Slavs, too, because just like Southern Italians they transfer their wider usage to the Italian standard. Let us have a look at some examples showing object doubling as a contactinduced feature of Molise Slavic. As we can see, the clitic pronoun hi in (18a) pleonastically doubles the verb-preceding noun Greka ‘the Greeks’, while ga in (18b) doubles the verb-preceding, fully accented demonstrative pronoun toga, just like in standard Italian li and l’ (=lo) respectively.54 (18) a. Nondeka mi Greka ni-sma hi vidil there we Greek.ACC.PL not-AUX.PRF.1PL ACC.PL see.PTCP.PL penjend. at.all (= Italian Lì noi i Greci non li abbiamo visti per niente.) ‘There, we did not see the Greeks at all.’ b. Toga ni-je ga potegnija DEM.MID.ACC.SG.M not-AUX.PRF.3SG ACC.SG.M shoot.PFV.PTCP.SG.M nikor, nobody ta je sa potegnija sam! DEM.MID.NOM.SG.M AUX.PRF.3SG REFL shoot.PFV.PTCP.SG.M alone (= Italian Costui non l’ha sparato nessuno, costui si è sparato da solo!) ‘Nobody has shot that one, he has shot himself alone!’ On the other hand, we also found pleonastic objects in Molise Slavic in positions, in which standard Italian, unlike the local Romance dialects, does not allow for doubling as, for instance, ga doubling the postverbal proper noun Matija in (19a) and the postverbal fully accented personal pronoun njega in (19b). The wrong pleonastic forms that Molise Slavs at least optionally use when speaking Italian are given in brackets in the Italian translations:
54 For similar constructions in Resian, see Benacchio (2009: 186).
12 Morphosyntactic changes in Slavic micro-languages
425
(19) a. Ga
vidahu Matija: jimaša ACC.SG.M see.IPRF.1SG PR.ACC.SG.M have.IPRF.3SG goča do pota velke. drop.ACC.PL of sweat.GEN.SG.M big.ACC.PL (≠ Italian [Lo] vedevo Mattia: aveva gocce di sudore grosse.) ‘I saw Mattia: he had big beads of sweat.’ b. Pa su kalal zdolu, then AUX.PRF.3PL descend.PFV.PTCP.PL down su ga vazal njega. AUX.PRF.3PL ACC.SG.M take.PFV.PTCP.PL him.ACC.SG.M (≠ Italian Poi sono scesi, [lo] hanno preso lui.) ‘Then they went down, they took him.’
In addition, pleonastic object clitics also refer to whole sentences as in (20a) and relative clauses as in (20b). Such cases once again emerge as mistakes in the Italian of the Molise Slavs, while they would be possible, as it seems, in Southern dialects. (20) a. Ga
ne kapaša not understand.IPRF.3SG ka maša mu da patrusinulu. COMP must.IPRF.3SG him give.INF parsley.ACC.SG.F (≠ Italian Non [lo] capiva che gli doveva dare il prezzemolo.) ‘She did not understand that she should give him the parsley.’ b. Je bi vidija priju votu AUX.PRF.3SG PTL.PQP see.PTCP.SG.M first.ACC.SG.F time.ACC.SG.F jenga ART.INDF.ACC.SG.M nimaldža ka ni-je bi ga not-AUX.PRF.3SG PTL.PQP ACC.SG.M animal.ACC.SG.M REL vidija maj. never see.PTCP.SG.M (≠ Italian Aveva visto per la prima volta un animale che non [lo] aveva mai visto.) ‘He had seen an animal for the first time that he had never seen (before).’ ACC.SG.N
426
Walter Breu
5 Conclusion This chapter was devoted to several morphological and syntactic characteristics of Molise Slavic, a South Slavic micro-language still spoken in three small villages in Southern Italy in the hinterland of the Adriatic coast. In some cases other Slavic minority languages served as points of comparison, especially Resian in North-Eastern Italy and colloquial Upper Sorbian in East Germany. The ancestors of today’s Molise Slavs immigrated to Molise roughly 500 years ago. Since then Molise Slavic has been in a situation of total language contact with the dominant Romance varieties of its new homeland. Most differences between Molise Slavic and its nearest cognates of the Bosnian-Croatian-Serbian language continuum were induced by language contact. But in spite of the farreaching changes, total isomorphism with the dominant Romance varieties has not developed, especially where no corresponding categories existed in standard Italian and in the local Molisian dialects. One can cite here, for example, the category of aspect of the Slavic derivative type; the very absence of a model for the reconstruction of Molise Slavic might be claimed responsible for preserving such categories unchanged. With respect to morphology or, rather morphosyntax, two main areas were examined, the verbal complex, comprising a lot of changes in the categories of Tense, Aspect and Mood, on the one hand, and the grammar of the noun, regarding the categories of Case, Gender/Declension, Number, and Definiteness, on the other. As for the adjective and the adverb, the morphosyntactic characteristics of the category of comparison were shown to have developed completely from the Slavic synthetic type to the Romance analytic one. In the syntax, changes in the position of the attribute (from pre-position to postposition) and of the clitics (from Wackernagel’s position to verb-adjacent) were explained by language contact, and so were the reduction of double negation and the introduction of object doubling. In theoretical terms, the far-reaching isomorphism between the languages in contact was the result of a tendency to reduce the rules necessary for the derivation of language-specific forms from the common diagrammar as conceived by the bilinguals. In most cases the isogrammatism in question was reached with the help of the adaptation of the replica to the dominant language. Two major procedures served to fulfil this task, a semantic and a syntagmatic one. The most evident changes with respect to the grammatical categories of Molise Slavic were triggered by the “adaptation of its semantic structure” to that of the dominant varieties by means of copying polysemies existing in L2 to L1. This procedure led to such grammatical developments in Molise Slavic as the development of modal differentiation in the future tense, the rise of a hypothetical
12 Morphosyntactic changes in Slavic micro-languages
427
imperfect, the merger of the locative with the accusative case, the loss of the neuter gender, and the formation of an indefinite article. It blocked, however, the development of a definite article because of a missing polysemy in L2, so that an article opposition could emerge only language-internally through the opposition between the contact-induced indefinite article and a definite zeroarticle. The syntagmatic procedure contributing to grammatical isomorphism in Molise Slavic was loan translation (in the broadest sense). In addition to word-for-word copies of the dominant models of its contact languages leading to the development of periphrastic constructions like a progressive and an imminentive, Molise Slavic also followed the syntagmatic principle as applied to the word order of existing elements. In this respect, it adapted the position of its attributes and clitics to that of their equivalents in the model languages, even omitting or adding some syntactic elements. All this lead to the formation of new rules for double negation and object doubling respectively. In general, the main bulwark against contact-induced changes should have been constituted by the diachronic constants of Slavic, a theoretical concept describing the overall parallelism of language change and persistence within the Slavic language group. But the contact with dominant varieties from another group with its own diachronic constants largely resulted in overriding the constants of their own genetic group (phylum). As a consequence, Molise Slavic followed in many respects typically Romance changes, for example, in developing an indefinite article or in losing the aorist instead of the imperfect. Unlike polysemies in L2, being copied to L1, an existing polysemy in L1 impeded the rise of isogrammatism due to a change in L1. For example, it was not possible to disambiguate the potential and the counterfactual functions, neither in the traditional conditional nor in the newly acquired hypothetical imperfect, in spite of distinct forms for these functions in the model language. In this case, isogrammatism could be achieved only by transferring the problematic polysemies from Molise Slavic to Italian, which brought about mistakes in L2 in terms of an imperfect with the (incorrect) function of potentiality in Italian. Other mistakes of the Molise Slavic bilinguals in Italian are due to the incomplete transfer of complex syntactic rules requiring additional information, in order to ascertain if they would apply or not. Such cases show a tendency towards hypercorrect usage of the postposition of attributes and the optionality of pleonastic double negation with pre-verbal negative pronouns. Bernd Heine asks in the conclusion of his chapter in this volume whether, as a consequence of the substantial transformations in its grammatical and
428
Walter Breu
semantic structure due to massive language contact, Molise Slavic has “moved up the scale of types from peripheral to core or from core to the nucleus SAE” (Standard Average European). He argues that there is no information on most of the twelve features of Haspelmath (2001) and claims “on impressionistic grounds” that, in spite of all the transformations, Molise Slavic has retained an overall Slavic typological profile. This is not the place to give a final answer to this question which has to be addressed in a separate study. But apart from the fact that Haspelmath (2001: 1492) already includes the Romance and the Balto-Slavic languages in the “core European languages” forming the SAE area, Molise Slavic has certainly moved up the scale further in the direction of the nucleus of this area as a result of its far-reaching adaptation to Italian. In fact, Italian itself does not belong to the nucleus of SAE constituted by German and French which are characterized by nine convergent features, according to Haspelmath (2001: 150). This language has only eight of them, thus predetermining the maximum result to be reached by Molise Slavic as a result of language contact as compared with only five features shared by the genealogically close Bosnian-Croatian-Serbian and other standard Slavic languages. As evidenced from the foregoing analysis, Molise Slavic developed an article system (SAE feature 1) and adapted the lack of verbal negation in the case of negative pronouns to the Italian rules, at least optionally (SAE feature 8). The only different feature remaining, the ‘have’-perfect (SAE feature 3), has not been dealt with in the present chapter, and, indeed, the Slavic ‘be’construction with the l-participle still remains the central way of deriving the perfect. But if Czech purportedly belongs to the Slavic languages which have a have-perfect (Haspelmath 2001: 1495), in spite of its many restrictions in terms of a resultative perfect and its low level of grammaticalization (Giger 2003), then we could at least marginally attribute this feature to Molise Slavic as well. Indeed, this language knows have-perfects with the resultative meaning of the type Jimaš (have.PRS.2SG) tuna redžištrana? ‘Have you recorded all?’, derived with the help of the have-auxiliary and the past passive particle. Consequently, in terms of its grammatical structure, Molise Slavic moved indeed away from its nearest cognates and came much closer to the nucleus of SAE. It remains open to discussion, however, if the features used in the SAE classification really fit for deciding on a Slavic or Romance “typological profile”.
12 Morphosyntactic changes in Slavic micro-languages
Abbreviations and glosses 1,2,3 = ACC ART AUX COMP COND DAT DEF DEM
F FUT GEN IMP INDF INS IPF., IPFV IPRF IS LOC L1 L2 M MID N NEC NOM PF., PFV PL PQP PR PRF PROB PRS PTCP PTL REFL REL RS SAE SG SUBJ
1st, 2nd, 3rd person clitic (boundary) accusative article auxiliary complementizer conditional dative definite demonstrative feminine future genitive imperative indefinite instrumental imperfective imperfect initial stage locative primary language secondary language masculine middle neuter necessitative nominative perfective plural past perfect proper noun perfect probability present participle participle reflexive relative resulting stage Standard Average European singular subjunctive
429
430
Walter Breu
References Benacchio, Rosanna. 2009. Il contatto slavo-romanzo nel croato del Molise e nei dialetti sloveni del Friuli. In: Lenka Scholze and Björn Wiemer (eds.), Von Zuständen, Dynamik und Veränderung bei Pygmäen und Giganten, 177–191. Bochum: Brockmeyer. Benacchio, Rosanna. 1996. L’articolo nel dialetto resiano: sulla questione della determinatezza nelle lingue slave. In: Rosanna Benacchio, Francesca Fici, and Lucyna Gebert (eds.), Determinatezza e indeterminatezza nelle lingue slave, 43–59. Florence: Unipress. Breu, Walter. 1991. Abweichungen vom phonetischen Prinzip bei der Integration von Lehnwörtern. In: Klaus Hartenstein and Helmut Jachnow (eds.), Slavistische Linguistik 1990, 36–70. Munich: Otto Sagner. Breu, Walter. 2003. Bilingualism and linguistic interference in the Slavic-Romance contact area of Molise (Southern Italy). In: Regine Eckardt, Klaus von Heusinger, and Christoph Schwarze (eds.), Words in Time. Diachronic Semantics from Different Points of View, 351–373. Berlin/New York: Mouton de Gruyter. Breu, Walter. 2008. Der slavische Lokativ im Sprachkontakt. Ein Beitrag zur Binnendifferenzierung des Moliseslavischen. In: Peter Kosta and Daniel Weiss (eds.), Slavistische Linguistik 2006/2007, 59–102. Munich: Otto Sagner. Breu, Walter. 2009. La comparazione nello slavomolisano. Un risultato tipico del contatto linguistico assoluto. In: Paola Desideri et al. (eds.), Alloglossie e comunità alloglotte nell’Italia contemporanea, 9–36. Rome: Bulzoni. Breu, Walter. 2011a. Language contact of minority languages in Central and Southern Europe: A comparative approach. In: Bernd Kortmann and Johan van der Auwera (eds.), The Languages and Linguistics of Europe. A Comprehensive Guide, 429–451. Berlin/ New York: Mouton de Gruyter. Breu, Walter. 2011b. Il verbo slavomolisano in confronto con altre lingue minoritarie: mutamento contatto-dipendente, resistenza e sviluppo autonomo. In: Walter Breu (ed.), L’influsso dell’italiano sul sistema del verbo delle lingue minoritarie, 149–184. Bochum: Brockmeyer. Breu, Walter. 2012. The grammaticalization of an indefinite article in Slavic micro-languages. In: Björn Wiemer, Björn Hansen, and Bernhard Wälchli (eds.), Grammatical Replication and Borrowability in Language Contact, 275–322. Berlin/Boston: Mouton de Gruyter. Breu, Walter. 2014. Funkcii nastojaščego i imperfekta soveršennogo vida i perfekta nesoveršennogo vida v molizsko-slavjanskom mikrojazyke. [The functions of the perfective present and imperfect, and of the imperfect perfect in the Molise-Slavic microlanguage]. Scando-Slavica 60 (2): 321–350. Breu, Walter. 2017. Slavische Mikrosprachen im absoluten Sprachkontakt. Band I. Moliseslavische Texte aus Acquaviva Collecroce, Montemitro und San Felice del Molise (Slavistische Beiträge, 505.) Wiesbaden: Otto Harrassowitz. Breu, Walter and Giovanni Piccoli. 2011. Südslavisch unter romanischem Dach. Die Moliseslaven in Geschichte und Gegenwart im Spiegel ihrer Sprache, Part 1: Texte gesprochener Sprache aus Acquaviva Collecroce. Munich/Berlin: Otto Sagner. Breu, Walter and Giovanni Piccoli. 2012. Südslavisch unter romanischem Dach. Die Moliseslaven in Geschichte und Gegenwart im Spiegel ihrer Sprache, Part 2: Texte
12 Morphosyntactic changes in Slavic micro-languages
431
gesprochener Sprache aus Montemitro und San Felice del Molise. Munich/Berlin/ Washington, DC: Otto Sagner. Breu, Walter and Lenka Scholze. 2006. Sprachkontakt und Syntax. Zur Position des Verbs im modernen Obersorbischen. In: Tilman Berger, Jochen Raecke, and Tilman Reuther (eds.), Slavistische Linguistik 2004/2005, 41–88. Munich: Otto Sagner. Broj, Val’ter [Walter Breu]. 2006. Flektivnyj i derivacionnyj glagol’nyj vid v molizskoslavjanskom jazyke [Inflectional and derivational verbal aspect in Molise Slavic]. Voprosy jazykoznanija 3: 70–87. Broj, Val’ter [Walter Breu]. 2013. Jazykovoj kontakt kak pričina perestrojki kategorij roda i sklonenija v molizsko-slavjanskom jazyke [Language contact as a reason for the reorganization of the categories of gender and declension in Molise Slavic]. In: Pёtr Mixajlovič Arkad’ev and Vjačeslav Vsevolodivič Ivanov (eds.), Tipologija slavjanskix, baltijskix i balkanskix jazykov (preimuščestvenno v svete jazykovyx kontaktov), 81–112. Sankt-Peterburg: Aletejja. Broj, Val’ter [Walter Breu], Malinka Pila, and Lenka Šol’ce [Scholze]. 2017. Vidovye pristavki v jazykovom kontakte (na materiale rez’janskogo, molizsko-slavjanskogo i verxnelužickogo mikrojazykov). In: Rosanna Benacchio, Alessio Muro, and Svetlana Slavkova (eds.), The role of prefixes in the formation of aspectuality: issues of grammaticalization, 59–84. Florence: Unipress. Cimmerling, Anton Vladimirovič. 2013. Tipologija porjadka slov v slavjanskix jazykax [A typology of word order in the Slavic languages]. In: Pёtr Mixajlovič Arkad’ev and Vjačeslav Vsevolodivič Ivanov (eds.), Tipologija slavjanskix, baltijskix i balkanskix jazykov (preimuščestvenno v svete jazykovyx kontaktov), 7–80. Sankt-Peterburg: Aletejja. Comtet, Roger. 2003. Peut-on parler d’un “néo-vocatif” en russe contemporain? In: Sebastian Kempgen, Ulrich Schweier, and Tilman Berger (eds.), Rusistika. Slavistika. Lingvistika, 83–90. Munich: Otto Sagner. Danylenko, Andrii. 2011. Is there any inflectional future in East Slavic? A case of Ukrainian against Romance reopened. In: Motoki Nomachi (ed.), Grammaticalization in Slavic Languages: From Areal and Typological Perspectives, 147–177. Sapporo: Slavic Research Center, Hokkaido University. Danylenko, Andrii. 2015. On the Mechanisms of the grammaticalization of comitative and instrumental categories in Slavic. Journal of Historical Linguistics 2: 276–296. Giammarco, Ernesto. 1968. Dizionario Abruzzese e Molisano, Vol. 1: A–E. Rome: Edizioni dell’Ateneo. Giger, Markus. 2003. Resultativa im modernen Tschechischen. Unter Berücksichtigung der Sprachgeschichte und der übrigen slavischen Sprachen. Bern: Peter Lang. Gołąb, Zbigniew. 1956. The concept of isogrammatism. Biuletyn Polskiego Towarzystwa Językoznawczego 15: 1–12. Gołąb, Zbigniew. 1959. Some Arumanian-Macedonian isogrammatisms and the social background of their development. Word 15 (3): 415–435. Guxman, Mirra Moiseevna. 1981. Istoričeskaja tipologija i problema diaxroničeskix konstant [Historical Typology and the Problem of Diachronic Constants]. Moscow: Nauka. Haspelmath, Martin. 2001. The European linguistic area: Standard Average European. In: Martin Haspelmath et al. (eds.), Language Typology and Language Universals/ Sprachtypologie und sprachliche Universalien/La typologie des langues et les universaux linguistiques. (HSK 20.2.), 1492–1510. Berlin/New York: Mouton de Gruyter.
432
Walter Breu
Heine, Bernd. 2013. On Isomorphism and Formulas of Equivalence in Language Contact. In: Guangshun Cao et al. (eds.), Breaking Down the Barriers: Interdisciplinary Studies in Chinese Linguistics and Beyond (Festschrift in Honor of Professor Alain Peyraube), 755–784. Taipei, Taiwan: Institute of Linguistics. Marra, Antonietta. 2005. Mutamenti e persistenze nelle forme di futuro dello slavo molisano. In: Walter Breu (ed.), L’influsso dell’italiano sulla grammatica delle lingue minoritarie. Problemi di morfologia e di sintassi, 141–166. Bochum: Brockmeyer. Muljačić, Žarko. 1973. Su alcuni effetti del bilinguismo nella parlata dei croati molisani. In: Manlio Cortelazzo (ed.), Bilinguismo e diglossia in Italia, 29–37. Pisa: Pacini. Nomachi, Motoki and Bernd Heine. 2011. On predicting contact-induced grammatical change. Evidence from Slavic languages. Journal of Historical Linguistics 1 (1): 48–76. Poletto, Cecilia and Sabrina Rasom. 2006. I quantificatori nei dialetti pugliesi. In: Federico Damonte and Jacopo Garzonio (eds.), Studi sui dialetti della Puglia, 77–98. (Quaderni di lavoro dell’Atlante Sintattico d’Italia 7.) Padua: Unipress. Rešetar, Milan. 1911. Die Serbokroatischen Kolonien Süditaliens. Vienna: Kaiserliche Akademie der Wissenschaften. Rohlfs, Gerhard. 1966. Grammatica storica della lingua italiana e dei suoi dialetti. Fonetica. Torino: Einaudi. Rohlfs, Gerhard. 1968. Grammatica storica della lingua italiana e dei suoi dialetti. Morfologia. Torino: Einaudi. Sammartino, Antonio. 2004. Grammatica della lingua croatomolisana. Montemitro/Zagreb: Profil. Sakel, Jeanette. 2007. Types of loan: matter and pattern. In: Yaron Matras and Jeanette Sakel (eds.), Grammatical borrowing in cross-linguistic perspective, 15–30, Berlin: Mouton de Gruyter. Sasse, Hans-Jürgen. 1992. Language decay and contact-induced change: similarities and differences. In: Matthias Brenzinger (ed.), Language death. Factual and theoretical explorations with special reference to East Africa, 59–80. Berlin/New York: Mouton de Gruyter. Scholze, Lenka. 2008. Das grammatische System der obersorbischen Umgangssprache im Sprachkontakt. Bautzen: Domowina. Uhlířová, Ludmila. 2014. Wortstellungstypologie. In: Sebastian Kempgen et al. (eds.), Die slavischen Sprachen. Ein internationales Handbuch zu ihrer Struktur und Geschichte, 2221–2232. (Handbücher zur Sprach- und Kommunikationswissenschaft, Bd. 32.2.) Berlin/Munich/Boston: Mouton de Gruyter.
Bernd Heine
13 On formulas of equivalence in grammaticalization: An example from Molise Slavic 1 Introduction Research that has been carried out in the course of the last decades suggests that there is some justification to consider Europe to be an areal unit of some kind (see especially Bechert et al. 1990; Bernini and Ramat 1996; Dahl 1990, 2000; Kuteva 1994, 1998; Haspelmath 1998, 2001; van der Auwera 1998; König and Haspelmath 1999). Based on a cross-linguistic typological perspective, Haspelmath (2001, see also 1998) proposes to define Europe as a Sprachbund, which he calls Standard Average European (SAE), a term going back to Whorf (1941, see also 1956: 138), also used by other scholars (e.g., Garvin 1949; Déscy 1973: 29; van der Auwera 1998).1 The place that Slavic languages occupy within the landscape of European areal typology has been discussed controversially. Some authors treat Slavic languages as more “marginal” European languages (e.g., Haarmann 1976: 123–127). Others again argue that they belong to the core of European languages (e.g., Hock 1986: 508–509; Haspelmath 2001), and for Déscy (1973), Russian is a “Standard Average European” language like French, German, and Italian. What is obvious is that a number of Slavic languages have been, and still are in close contact with languages of Western Europe, and there is by now a rich literature that describes the nature and the result of contact. More recent work suggests that grammaticalization plays an important role in contact-induced grammatical change (e.g., Heine and Kuteva 2006). Wiemer et al. (2012) provide an impressive list of contact-induced grammaticalizations and some of these changes appear to be in the direction of an SAE format. In this paper attention is drawn to an alternative force that appears to play a role in contact-induced grammatical change. It is argued that bilingual speakers in close contact with another language aim at establishing formulas of translational equivalence between the discourse structures of the languages concerned. To this end, a case study is presented involving a grammatical
1 Note that Whorf did not use “Standard Average European” with reference to linguistic typology, nor did he define it in terms of linguistic features. https://doi.org/10.1515/9783110639223-014
434
Bernd Heine
category that has been proposed by Haspelmath (2001) to be criterial for classifying languages as belonging or not belonging to SAE, namely articles as grammaticalized markers of definiteness and indefiniteness. Discussion is further restricted to indefinite articles and to one specific situation of language contact, namely that between Italian and Molise Slavic (in short: Molisean).
2 From numeral ‘1’ to indefinite article in Molisean 2.1 Unidirectionality A number of evolutionary regularities have been proposed in studies of grammaticalization to account for the fact that grammatical development is essentially unidirectional (see Börjars and Vincent 2010 for discussion). For the evolution of indefinite articles, Givón (1981: 75) proposes the semantic scale in (1) (Heine 1997: 75) to account for the conceptual continuum of grammaticalization from numeral ‘one’ to indefinite article, arguing for a cognitive implicational scale of grammaticalization: Quantification implies existence/reference, which again implies genericity/connotation. (1)
From numeral to indefinite article (Givón 1981: 75) Quantification > referentiality/denotation > genericity/connotation
The grammaticalization of indefinite articles has been shown to proceed similarly through a series of contexts and stages. Thus, it is the stages listed in (2) that mark the gradual pragmatic and semantic evolution of many indefinite articles derived from a numeral for ‘one’ (Heine 1997: 70): (2)
Stages in the semantic evolution of numeral-derived indefinite articles 1. Numeral: An item serves as a nominal modifier denoting the numerical value ‘one’ (numeral). 2. Presentative marker: The item introduces a new participant presumed to be known to the speaker but unknown to the hearer and this participant is then taken up as definite in subsequent discourse. 3. Specific indefinite marker: The item presents a participant presumed to be known to the speaker but unknown to the hearer, irrespective of whether or not the participant is expected to come up as a major discourse participant.
13 On formulas of equivalence in grammaticalization
435
4. Non-specific indefinite marker: The item presents a participant whose referential identity neither the hearer nor the speaker knows. 5. Generalized indefinite article: The item can be expected to occur in all contexts and on all types of nouns except for a few contexts involving, for instance, definiteness marking, proper nouns, predicative clauses, and so on. Grammaticalization theory would predict that if a language has reached a given stage then it likely has also passed through all preceding stages. This evolutionary scenario can also be mapped onto synchrony and be understood as a synchronic implicational scale. On the basis of this scale it is possible to formulate the generalizations in (3). These generalizations cannot be reversed, that is, presence of, for instance, Stage 2 in a given language by no means entails that that language also has a Stage 3 article. (3) Probabilistic generalization on numeral-derived indefinite markers If a language has a marker of Stage X then this marker can also be used for all preceding stages. These generalizations are probabilistic rather than absolute, and they must be taken with care. First, they rest on impressionistic observations in a number of languages in various parts of the world, that is, they are not based on a representative typological sample. And second, regularities in grammatical evolution have been shown to be overall unidirectional, but they are not without exceptions. Language history is complex and there may be other factors interfering with the evolution in a particular case, including alternative patterns of grammaticalization.
2.2 Molisean There are a number of languages that have been argued to have undergone substantial contact with other languages, and a number of these languages located at the western and southwestern periphery of the Slavic world. In Nomachi and Heine (2011), a group of Slavic languages is singled out which they refer to as Slavic “high-contact” languages. These are languages that can be shown to have had a history of long and intense interaction with German or Italian (or both), involving massive second language acquisition and bilingualism. “Highcontact” languages include Upper Sorbian, Lower Sorbian, Kashubian, extinct Slovincian and Polabian, Slovene, as well as four varieties of Croatian, namely Burgenland Croatian, Kajkavian, Čakavian, and Molisean.
436
Bernd Heine
Molise Slavic (Molisean) is the language of a community of speakers from the Hercegovinian Neretva Valley; Breu (2012: 134) assigns it to the ŠtokavianIkavian dialect group of the South Slavic languages. Molisean speakers emigrated around 1500 because of the Turkish invasion on the Balkans, settling in areas of southeastern Italy that were sparsely inhabited due to earthquakes and epidemics. Today, Molisean is spoken only in three villages, Acquaviva Montemitro, and San Felice of Molise Region in the Campobasso Province. After contact both with the local Italian varieties and with Standard Italian over a period of half a millennium, their language has been massively influenced by this Romance language (for a survey, see Breu 1998, also 1999, 2003a, 2003b, 2003c, 2004, 2012). Together with Upper Sorbian and Lower Sorbian of eastern Germany, Molisean forms one of what Breu (2003a) refers to as Slavic “micro-languages”. There are a number of differences between the two micro-languages. First, unlike Upper Sorbian, Molisean does not dispose of a standard form, and second, the model language is German in the case of Upper Sorbian but non-standard varieties of Italian spoken in the Molise region of the Campobasso Province, since about 150 years also Standard Italian. The two model languages are with reference to the subject matter of the following discussion structurally alike, and contact between model and replica languages has in both cases a long history. Both German and Italian have indefinite articles of Stage 5. Slavic languages are well-known for their absence of indefinite articles, with the possible exception of Macedonian (Weiss 2004). But as Breu (2003a) shows convincingly in his analysis of Slavic micro-languages2 (see also Heine and Kuteva 2006: 97–139), Macedonian is not unique in this respect: Molisean is another “exception”, having replicated an indefinite article from the surrounding dominant language Italian. Consider the examples below, where sentences from Molisean (M) are given, followed by an Italian (I) and an English translation (there are no interlinear glosses in Breu’s publications, and the markers in question are printed in bold, where Ø = lack of article, and * Ø = the article must not be omitted). Like the model language Italian, Molisean has a largely grammaticalized Stage 5 indefinite article (na, nu with feminine singular objects; henceforth in
2 We are not able to do justice to the fine-grained analysis presented by Breu (2003a, 2012); the reader is referred to these works for many more details. The data to be discussed below are taken overwhelmingly from Breu (Breu 1998, 1999, 2003a, 2003b, 2003c, 2004, 2012); some of them were not available to us when we were working on Heine and Kuteva (2006).
13 On formulas of equivalence in grammaticalization
437
short: na) that can be traced back to the numeral jen- ‘one’. Molisean ‘one’ disposes of a paradigm of morphophonological forms, distinguishing between long and short forms, for example, jena vs. na (nominative masculine singular). As the following examples show, their use is essentially obligatory. Example (4a) illustrates the lexical use as a numeral for ‘one’ (Stage 1), (4b) the presentative use (Stage 2), characteristic of openings in tales, (4c) the specific indefinite (Stage 3), and (4d) the non-specific indefinite (Stage 4). With abstract and generic referents as well, Molisean shows roughly the same degree of grammaticalization as the Italian indefinite article does, see (4e). (4) Molisean (Breu 2003a: 37, 2012: 278–299; M = Molisean, I = Italian)3 a. Stage1 M. Mitaj još nu divojku! *Ø I Invita ancora una ragazza! *Ø ‘Invite one more girl!’ b. Stage 2 M. Biša nu votu na žena stara. *Ø I C’era una volta una donna anziana. *Ø ‘Once upon a time there was an old woman.’ c. Stage 3 M. Sfe skup je uliza na ljud tusti. *Ø I D’improvviso entrò un uomo grasso. *Ø ‘Suddenly a fat man came in.’ d. Stage 4 M. Si ta ćuje na polidzjot, ta meće pržuna. *Ø I Se ti sente un poliziotto, ti mettein prigione. *Ø ‘If a policeman hears you, he’ll lock you up.’ e. Stage 5 M. Ti jesi jindženuv kana na dita. *Ø I Tu sei ingenuo come un bambino. *Ø ‘You are naive like a child.’ While Molisean has acquired essentially the full range of grammatical functions associated with an indefinite article, there are a few contexts that have not been affected by grammaticalization. A paradigm example is provided by “pure mass nouns”, where no indefinite article may be used (Breu 2012: 285).4 Finally, there remain a few uses that are not entirely in accordance with grammaticalization. As we will see below, this does not apply only to Molisean but in the same way also to Italian.
3 Breu gives no examples of the numeral Stage 1. 4 No example is provided by Breu (2003a, 2012).
438
Bernd Heine
2.3 On semantic relationship in grammaticalization The data on the indefinite article in Molisean suggest that speakers aiming at establishing translational equivalence have a knowledge about whether different meanings are or are not sufficiently “related” to one another, given an appropriate context, to be expressed by one and the same linguistic form. The way “related meaning” is to be defined in grammaticalization raises an empirical problem that has not been solved entirely. But at least its effects can be reconstructed on the basis of a linguistic analysis of the data available. For example, Stages 2 and 3 can be said to share the meaning feature “presumed to be known to the speaker but unknown to the hearer” but not the feature “need not be taken up in subsequent discourse”. On this analysis, there is one feature that the two stages have in common, but also one that distinguishes them. Stage 4, by contrast, differs from Stage 3 in one feature (referential identity unknown to the speaker), but not in the other (referential identity unknown to the hearer), and it shares neither of these features with Stage 2. Thus, the relative degree of semantic relationship is higher between adjacent stages along a scale of grammaticalization than between nonadjacent stages. “Semantic relationship” in the sense of the term used here as a term of contact linguistics can be likened to what in work on semantic maps is captured by means of relative distance on a map (see, e.g., Narrog and van der Auwera 2012) and can in fact be captured in terms of “semantic map connectivity” (Wiemer et al. 2012: 28). I hypothesize that relative degree of “semantic relationship” is a constraint on contact-induced grammaticalization, that is, grammaticalization in the replica language is likely when close relationship is involved but unlikely when remote semantic relationship is involved. And I further hypothesize tentatively that the same applies to “polysemy copying”, namely that it requires the two meanings associated with one linguistic form in the model language to be semantically close in order to be copied analogically in the replica language. The hypotheses just presented are based on a small sample of data and therefore need to be taken with care, subject to further research. But if confirmed by a large corpus of data, they would have a number of implications. They would be of help in particular for determining what a possible process of contact-induced grammaticalization is and what is not.
439
13 On formulas of equivalence in grammaticalization
2.4 A scale of grammaticalization A survey of a few Slavic languages suggests that these languages behave overall in accordance with the probabilistic prediction in (3) above. Table 1 shows the following in particular. Table 1: Degree of the grammaticalization from numeral ‘one’ to indefinite article in selected Slavic languages (Sources: Breu 2003a, 2012; Weiss 2004; Heine and Kuteva 2006: 97–139; Heine 2012).5 Stage
Function
Upper Molise Sorbian Slavic
Mace- Czech, Serbian, donian Bulgarian Croatian, Polish, Russian
Ukrainian, Belarusian
Numeral ‘one’ Presentative Specific indefinite Non-specific indefinite Generalized
+ + + + +
+ + (+)
+
+ + + + +
+ + (+)
+ (+)
(+)
First, it is precisely those Slavic languages that have had the most intense contact with languages having Stage 5 indefinite articles that also created corresponding articles. At one end there are Colloquial Upper Sorbian in eastern Germany with a history of nearly a millennium of contact with German, and Molisean, which has been in contact with Italian for roughly 500 years; at the other end there are the Eastern Slavic languages Ukrainian and Belarusian, both languages with the least amount of contact with article languages. And second, Table 1 also shows that contact-induced grammaticalization proceeds in one direction from one stage to the next, where a new stage is built on the stage immediately preceding it. Synchronically, this fact can once again be described in the form of the implicational scale proposed in (3). In addition to Colloquial Upper Sorbian and Molisean, there might be another Slavic variety having a largely or fully grammaticalized indefinite article, namely the Slovene dialect of San Georgio, whose indefinite article includes
5 Note that we are restricted here to non-standard, colloquial, varieties of the languages concerned. As Breu (2003a) has shown for Upper Sorbian, an entirely different picture would arise if Standard Upper Sorbian were chosen.
440
Bernd Heine
not only specific and non-specific but also predicative (Stage-5) uses (see Wiemer et al. 2012: 119). But there is also one problematic case, namely Macedonian. In this language, the item eden ‘one’ is widely used as an indefinite marker of specific (Stage 3) and generic reference (Stage 5) but apparently not of non-specific reference (Stage 4) (Weiss 2004). Furthermore, the generic use of this item is exceptional and in need of further research.
3 Formulas of equivalence (FOEs) We do not know how the indefinite article in Molisean evolved diachronically; to our knowledge, there are no historical records that would enable us to reconstruct the evolution. What we do know, however, is its present-day uses, meticulously described by Breu (2003a, 2012). With reference to the subject of this paper, this knowledge includes the following: (5) Observations on the uses of the indefinite article na in Molisean (based on Breu 2003a, 2012) a. The article is to a large extent a replica of the Italian indefinite article un-. b. Like the Italian indefinite article un- it exhibits the entire range of salient grammaticalizations as its Italian equivalent. c. There appears to be a mechanism that enables speakers of Molisean to establish regular correspondences between structures of the two languages. These correspondences take the form of formulas of equivalence (FOEs).
3.1 What are formulas of equivalence? Formulas of equivalences (FOEs) are recurrent equations that speakers are assumed to make in situations of language contact between linguistic structures of different languages (or dialects) that are conceived and treated the same (Keesing 1991; Heine 2013). They are hypothesized to represent discontinuities of language use that are stored and retrieved whole in the brain. Equivalent categories are those that are regularly used in translation work as corresponding to one another between the two languages concerned (Heine and Kuteva 2005: 219–234; Heine 2013). FOEs relate in some way or other to a number of alternative concepts that have been proposed in works on language contact. They concern translational equivalence, that is, the bilingual behavior of speakers (or writers) and hearers
13 On formulas of equivalence in grammaticalization
441
(or readers) in the translation of concepts or contents from one language into another. Translation(al) equivalence has widely been discussed in translation theory (see, e.g., Catford 1965; Uwajeh 2007). Important for our purposes is that the information sent from the source language to the target language be considered by speakers and hearers as ‘equal in value’. Whether, or to what extent, FOEs have their existence purely in reconstruction work carried out by the analyst or are in fact part of the cognitive and pragmatic knowledge and behavior of speakers in language contact is an empirical question that cannot be resolved without much further research. FOEs can manifest themselves at any level of change, be that the level of innovation, propagation, or conventionalization (Croft 2000). And they may surface in a wide range of linguistic structures, extending from phonetic features to larger stretches of discourse. Our interest here is restricted to FOEs that involve grammatical replication, that is, the transfer of structure and/or meaning between languages in contact (Heine and Kuteva 2005). Equivalence, variously referred to as “connection”, “correspondence”, “isogrammatism”, “mutual isomorphism”, or “similarity”, is a central notion in some works of contact linguistics. A survey of a number of documented cases of intense language contact suggests that looking for ways of establishing formulas of equivalence provides one of the motivations of people aiming at communicating successfully in bilingual situations. As the description by Breu (2003a, 2012) suggests, correspondences in the use of indefinite articles between Molisean and Italian can be described in terms of a cluster of FOEs. Evidence for the presence of FOEs comes from the following observations: a. Equivalence between the languages concerned is such that any explanation other than via FOEs would be clearly less plausible. b. Equivalence can be at variance with generalizations on grammaticalization. Observation (a) can be illustrated with the formula of Table 2, where a verb of the replica language Molisean having one basic meaning acquires a second meaning by replicating a polysemy pattern of the model language Italian. The example presents one sub-type of FOEs, namely polysemy copying (Heine and Kuteva 2005; Heine 2012). Considering the semantic difference between the two meanings involved, it would not seem very plausible to argue that the polysemy found in Italian and Modern Molisean is due to chance. Rather, a semantically more plausible hypothesis is one according to which Molisean speakers established a lexical equivalence pattern between the two languages. Another notion is proposed by Gast and van der Auwera (2012), which they call interlingual identification. This term is taken from Weinreich ([1964] 1953: 7–8),
442
Bernd Heine
Table 2: A lexical formula of equivalence (based on Breu 2003b).
Verb Meaning
*Pre-contact replica language Earlier Molisean
Model language Italian
Replica language Modern Molisean
nosit ‘carry’
portare (a) ‘carry’, (b) ‘drive a car’
nosit (a) ‘carry’, (b) ‘drive a car’
to describe contact-induced changes in Mayan and Mixe-Zoquean languages of Mexico. Interlingual identification concerns either “linguistic signs” (including constructions) or “linguistic categories” (Gast and van der Auwera 2012: 382). Whether, or to what extent, “categories” are in fact involved in the behavior of bilinguals aiming at establishing FOEs is an issue that is in need of more research (see Gast and van der Auwera 2012: 29). In any case, interlingual identification can be viewed as a prerequisite for FOEs to evolve. FOEs also relate to what is called structural isomorphism (Heine and Kuteva 2005). Structural isomorphism6 (henceforth, isomorphism) rests on the linguist’s theoretical constructs of categories. “Equivalent categories” then means that there is a category Mx in the model language and a category Rx in the replica language which are in some sense taken to be structurally the same. What exactly “structurally the same” stands for in a given case is contingent upon the analyst’s descriptive framework – that is, on how categories are defined in that framework.
3.2 Formula of equivalence may be at variance with grammaticalization In general, FOEs between Molisean and Italian are in accordance with grammaticalization: The various stages of grammaticalization can be interpreted as each representing an FOE of obligatory language use (see Heine 2013). But there are also examples where there is a difference. Consider the following examples. In (6a), Molisean speakers employ the expected article since Stage 5
6 Note that the term “isomorphism” has received a range of applications in linguistics; it is used especially in language-internal analysis for one-to-one mappings between linguistic form and meaning. The way the term is used in contact linguistics in general and in the present paper in particular differs from the former applications in that it is strictly comparative in nature, relating to comparisons between languages (or dialects) in contact.
13 On formulas of equivalence in grammaticalization
443
requires an article. But there is another fairly productive formula of equivalence whereby the Italian definite article corresponds to zero (Ø) expression in Molisean (Breu 2003a, 2012: 290, 314). Accordingly, when Italian uses the indefinite article as in (6a), Molisean has its Stage 5 indefinite article na, but when Italian has the definite article lo, there is a zero article in Molisean. (6) Molisean - Italian (Breu 2003a: 37, 2012: 290) a. M. Na Zlav ne goriva laž. I Uno Slavo non dice bugi. ‘A (real) Molisean Slav does not lie.’ b. M Ø Zlav ne goriva laž. I Lo Slavo non dice bugie. ‘The Molisean Slav does not lie.’ (= ‘Molisean Slavs don’t lie’) Breu (2012: 289–290) notes that (6a) is a “prescriptive” and (6b) a “descriptive” predication; nevertheless, both are suggestive of a Stage 5 (generalized article) function and, hence, would require the Molisean indefinite article na. The question then is: Why does (6b) contradict grammaticalization? I will return to this issue in Section 3.3. There are also other examples where semantic classification based on principles of grammaticalization is seemingly ignored, that is, where Molisean speakers replicate the Italian model, even though their generalized Stage 5 article would have been expected. Thus, in the following example, Molisean has again a zero article where Italian uses the definite article la: (7)
M. Ø Tigra je na nimaldža. I La tigre è un animale. ‘The tiger is an animal.’
This situation is in need of explanation. Breu (2012: 290) observes that in Italian “the definite article is compulsory in generic definitions” like (7). Molisean has acquired an indefinite article but no definite article. Whenever there is a definite article in Italian, Molisean speakers tend to use a “zero article”. Accordingly, a contrast such as the one in (8a) is rendered in Molisean “predominantly” by a distinction use vs. non-use of an indefinite article. And the formula appears to also apply to negative clauses, as (8b) shows. (8) a.
M na žena Ø žena I una donna la donna ‘a woman’ ‘the woman’
(Breu 2012: 302)
444
Bernd Heine
b. M Nimam Ø maginu. I Non ho la macchina. ‘I don’t have a car.’
(Breu 2012: 298)
Thus, translational equivalence overrules grammaticalization and functional classification in that Molisean speakers tend to draw on the formula in (9): (9) The definite-article formula of equivalence (Breu 2012: 290, 302): “Use a zero article in Molisean when there would be a definite article in Italian”. In addition to the formula in (9) there are a number of other FOEs, and in most of them, both the model and the replica languages are identical with regard to article use. In the following example (10) one would have expected an indefinite article of Stage 3 since the participant ‘day’ is known to the speaker but presumably unknown to the hearer. But since there is no article in Italian there is also none in Molisean in this context. (10) M Sa rabija Ø po dana. I Ho lavorato Ø mezza giornata. (?una) ‘I worked half a day.’
(Breu 2012: 297)
Thus, there appears to be a lexically restricted FOE of the following kind: (11) The quantifying (‘half a’) formula of equivalence (Breu 2012: 297): “Use a zero article when the head noun is qualified by ‘half’”. Another FOE is lexically or contextually determined, based on semantic distinctions shared by the model and replica languages, the result being what Breu (2012: 288) calls “fluctuations in the usage of ONE”: In classifying copula predications as illustrated in (12), both Italian and Molisean use the expected Stage 5 indefinite article with nouns such as ‘student’ but no article with nouns such as ‘professor’ or ‘policeman’: (12) M Ona je na študentesa / Ø profesoresa. I Lei è una studentessa / Ø professoressa. ‘She is a student / a professor.’ (Breu 2003a: 42, 2012: 291) Based on the description volunteered by Breu (2012: 291), the following FOE of preferred usage can be proposed:
13 On formulas of equivalence in grammaticalization
445
(13) A predicate noun formula of equivalence (Breu 2012: 297): “In a classifying copula predication, use the predicate noun without indefinite article when the noun denotes a career denomination but with article otherwise”.
3.3 Formulas of equivalence and grammaticalization For good reasons, Breu (2012) concludes that Molisean has the same kind of “full-fledged” indefinite article as Italian. But the reason for discussing specific Molisean-Italian FOEs was that there nevertheless are also differences, and that FOEs are in some way or other at variance with what would be the expected outcome of grammaticalization. For example, we saw that in some contexts where one would expect a Stage 3 or Stage 5 indefinite article in Molisean, there is an FOE that requires a zero article corresponding to the definite article in Italian. Comparative work on language contact suggests that grammaticalization tends to be more advanced in the model than in the replica language (Heine and Kuteva 2005; Wiemer et al. 2012), and Heine and Nomachi (2013: 89–90) propose the following generalization as a diagnostic of directionality in contact-induced replication: If two languages have undergone the same process of grammaticalization as a result of language contact but one of them exhibits a high and the other a low degree of grammaticalization then the former is more likely to have provided a model of replication than the other way round.
There is, however, one FOE in the Molisean-Italian contact situation that can be taken to be a counterexample to this generalization, namely the one in (14), exemplified in (15). (14) The generic noun listing formula of equivalence (Breu 2012: 283): When there is an alternative listing (‘either – or’) of two generic nouns, use the indefinite article in Molisean while the article is not obligatory in Italian. (15) M Što mam ti kupit: na kučič o na mačič? *Ø /*Ø I Che ti devo comprare: un cane o un gatto? or Ø /or Ø ‘What should I buy you: a dog or a cat?’
446
Bernd Heine
According to this FOE, Molisean uses an obligatory, that is, a fully grammaticalized article whereas Italian has a less grammaticalized situation, where the article has not yet been generalized, that is, is optional. To conclude, this formula is contrary to what generalizations on contactinduced grammaticalization would predict (Heine and Nomachi 2013): Rather than being less grammaticalized, the replica language presents the more grammaticalized situation.
3.4 Discussion The observations made in the preceding section raise the following questions: How do formulas of equivalence relate to grammaticalization? Are they independent of one another, or are they related to one another? While this issue needs much more research it would seem that, in spite of what we noted in Section 3.3, FOE are overall in accordance with principles of grammaticalization, as the listing in Table 3 suggests. First, an overview of the formulas shows that each of the five stages of semantic evolution distinguished in (2) can be described as a formula of equivalence of its own. Second, all these formulas taken together could be interpreted as a more abstract formula that summarizes the regularities that speakers of Molisean have developed in the use of the indefinite article. And third, none of the ten formulas listed really contradicts the morphological pathway of grammaticalization from non-article to full-fledged article, involving the three main stages of Table 4.
Table 3: The main formulas of equivalence between Molisean and Italian in the use of the indefinite article (based on Breu 2003a, 2012). Stage
Main function
Molisean
Italian
Presentative Specific indefinite Non-specific indefinite Generalized Alternative listing of generic nouns Classifying copula predication Generic definitions “Pure mass nouns” Quantifying ‘half a’
+ + + + + +/– – – –
+ + + + +/– +/– Definite – –
13 On formulas of equivalence in grammaticalization
447
Table 4: Final stages in the morphological evolution of the indefinite article in Molisean. Morphological stage Form
Context
I II III
“Pure mass nouns”, quantifying ‘half a’ Classifying copula predications presentative, specific indefinite, non-specific indefinite, generalized nouns, alternative listings of generic nouns
No article Optional article Obligatory article
4 Conclusions Formulas of equivalence are interlingual generalizations, that is, they apply in the same way to the model and the replica language. With reference to indefinite articles, they predict not only whether an article is or is not used, but also whether an expression is acceptable both with and without article and in the latter case also whether or not there is a semantic contrast associated with this distinction. A central question that cannot be answered without much further empirical work is what cognitive status FOEs have for speakers in language contact situations. Most probably, this status differs depending on whether or not speakers are regularly exposed simultaneously to both the model and the replica language, as appears to be the case for many speakers of Molisean. But whether, or to what extent, FOEs play a role in the behavior of other contact situations involving Slavic languages is entirely unclear. Languages in contact can be classified on the basis of the relative number of FOEs they share. In the ideal case, where all discourse options available to speakers of both the model and the replica languages are determined by FOEs there would be complete intertranslatability. Cases that have been discussed under the rubric of “metatypy” (Ross 1996, 1997, 2001), such as the Indian village Kupwar (Gumperz and Wilson 1971), Northwestern New Britain (Thurston 1982, 1987), or the situation on Karkar Island in Papua New Guinea (Ross 1996). These come close to presenting ideal cases, but for none of them is there in fact sufficient evidence to suggest that they really are, and much the same applies to Molisean and any other Slavic language. The present paper was meant to show that at least one Slavic language, namely Molisean, has made an important move towards a SAE format, undergoing substantial transformations in its grammatical and semantic structure due to massive language contact with both Standard Italian and its regional varieties (Breu 1998, 1999, 2003a, 2003b, 2003c, 2004, 2012). With regard to the
448
Bernd Heine
structure of its indefinite article, Molisean has become a paradigm member of SAE. Molisean speakers have carried their numeral for ‘one’ through all stages of grammaticalization, developing a Stage 5 indefinite article largely, though not entirely equivalent to the Italian model. But what about the overall position of Molisean within the classification proposed by Haspelmath (see Section 1)? Has Molisean moved up the scale of types from peripheral to core, or from core to nucleus SAE? In spite of the rich documentation that Breu has presented on Molisean, there is no information on most of the twelve features of Haspelmath (2001). On impressionistic grounds it would seem that in spite of all the transformations, Molisean has retained an overall Slavic typological profile.
Abbreviations FOE = formula of equivalence I = Italian M = Molisean SAE = Standard Average European Ø = zero expression
References Bechert, Johannes, Giuliano Bernini, and Claude Buridant (eds.). 1990. Toward a typology of European languages. (Empirical Approaches to Language Typology 8.) Berlin: Mouton de Gruyter. Bernini, Giuliano and Paolo Ramat. 1996. Negative Sentences in the Languages of Europe: A Typological Approach. (Empirical Approaches to Language Typology 16.) Berlin/ New York: Mouton de Gruyter. Börjars, Kersti and Nigel Vincent. 2010. Grammaticalization and directionality. In: Heiko Narrog and Bernd Heine (eds.), The Oxford Handbook of Grammaticalization, 162–176. Oxford: Oxford University Press. Breu, Walter. 1990a. Sprache und Sprachverhalten in den slavischen Dörfern des Molise (Süditalien). In: Walter Breu (ed.), Slavistische Linguistik 1989, 35–65 Munich: Otto Sagner. Breu, Walter. 1992. Das italokroatische Verbsystem zwischen slavischem Erbe und kontaktbedingter Entwicklung. In: Tilmann Reuther (ed.), Slavistische Linguistik 1991, 93–122. Munich: Otto Sagner. Breu, Walter. 1994. Der Faktor Sprachkontakt in einer dynamischen Typologie des Slavischen. In: Hans Robert Mehlig (ed.), Slavistische Linguistik 1993, 41–64. Munich: Otto Sagner. Breu, Walter. 1996. Überlegungen zu einer Klassifizierung des grammatischen Wandels im Sprachkontakt (am Beispiel slavischer Kontaktfälle). Sprachtypologie und Universalienforschung 49 (1): 21–38.
13 On formulas of equivalence in grammaticalization
449
Breu, Walter. 1998. Romanisches Adstrat im Moliseslavischen. Die Welt der Slaven 43: 339–354. Breu, Walter. 1999. Die Komparation im Moliseslavischen. In: René Métrich, Albert Hudlett, and Heinz-Helmut Lüger (eds.), Des racines et des ailes: théories, modèles, expériences en linguistique et en didactique, 37–63. Nancy: Association des Nouveaux cahiers d’allemand. Breu, Walter. 2003a. Der indefinite Artikel in slavischen Mikrosprachen: Grammatikalisierung im totalen Sprachkontakt. In: Holger Kuße (ed.), Slavistische Linguistik 2001, 27–68. Munich: Otto Sagner. Breu, Walter. 2003b. Bilingualism and linguistic interference in the Slavic-Romance contact area of Molise (Southern Italy). In: Regine Eckardt, Klaus von Heusinger, and Christoph Schwarze (eds.), Words in Time. Diachronic Semantics from Different Points of View, 351–373. Berlin/New York: Mouton de Gruyter. Breu, Walter. 2003c. Impersonales Neutrum im Moliseslavischen. In: Sebastian Kempgen, Ulrich Schweier, and Tilman Berger (eds.), Rusistika, Slavistika, Lingvistika: Festschrift für Werner Lehfeldt zum 60. Geburtstag, 57–71. Munich: Otto Sagner. Breu, Walter. 2004. Der definite Artikel in der obersorbischen Umgangssprache. In: Marion Krause and Christian Sappok (eds.), Slavistische Linguistik 2002, 9–57. Munich: Otto Sagner. Breu, Walter. 2012. The grammaticalization of an indefinite article in Slavic micro-languages. In: Björn Wiemer, Berhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 275–322. (Trends in Linguistics Studies and Monograph 242.) Berlin/New York: Mouton de Gruyter. Catford, John C. 1965. A Linguistic Theory of Translation: An Essay in Applied Linguistics. Oxford: Oxford University Press. Croft, William. 2000. Explaining Language Change: An Evolutionary Approach. Harlow: Longman. Dahl, Östen. 1990. Standard Average European as an exotic language. In: Johannes Bechert, Giuliano Bernini, and Claude Buridant (eds.), Toward a Typology of European Languages, 3–8. (Empirical Approaches to Language Typology 8.) Berlin/New York: Mouton de Gruyter. Dahl, Östen. 2000. The grammar of future time reference in European languages. In: Östen Dahl (ed.), Tense and Aspect in the Languages of Europe, 309–328. (Empirical Approaches to Language Typology, Eurotyp 20-6.) Berlin/New York: Mouton de Gruyter. Déscy, Gyula. 1973. Die linguistische Struktur Europas: Vergangenheit, Gegenwart, Zukunft. Wiesbaden: Harrassowitz. Garvin, Paul L. 1949. Standard Average European and Czech. Studia Linguistica 3: 65–85. Gast, Volker and Johan van der Auwera. 2012. What is ‘contact-induced grammaticalization’? Evidence from Mayan and Mixe-Zoquean languages. In: Björn Wiemer, Berhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact 381–426. (Trends in Linguistics Studies and Monograph 242.) Berlin/New York: Mouton de Gruyter. Givón, Talmy. 1981. On the development of the numeral ‘one’ as an indefinite marker. Folia Linguistica Historica 2 (1): 35–53. Gumperz, John J. and Robert Wilson. 1971. Convergence and creolization: A case from the Indo-Aryan/Dravidian border in India. In: Dell H. Hymes (ed.), Pidginization and Creolization of Languages: Proceedings of a Conference Held at the University of the West Indies, Mona, Jamaica, April 1968, 151–167. Cambridge: Cambridge University Press.
450
Bernd Heine
Haarmann, Harald. 1976. Aspekte der Arealtypologie. Die Problematik der europäischen Sprachbünde. (Tübinger Beiträge zur Linguistik 72.) Tübingen: Gunter Narr. Haspelmath, Martin. 1998. How young is Standard Average European?, Language Sciences 20 (3): 271–287. Haspelmath, Martin. 2001. The European linguistic area: Standard Average European. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, 1492–1510. Berlin/New York: Walter de Gruyter. Heine, Bernd. 1997. Cognitive Foundations of Grammar. Oxford/New York: Oxford University Press. Heine, Bernd. 2002. On the role of context in grammaticalization. In: Ilse Wischerand and Gabriele Diewald (eds.), New Reflections on Grammaticalization, 83–101. (Typological Studies in Language 49.) Amsterdam/Philadelphia: John Benjamins. Heine, Bernd. 2012. On polysemy copying and grammaticalization in language contact. In: Claudine Chamoreau and Isabelle Léglise (eds.), Dynamics of Contact-Induced Change, 125–166. (Language Contact and Biligualism 2.) Berlin: Mouton de Gruyter. Heine, Bernd. 2013. On isomorphism and formulas of equivalence in language contact. In: Cao Guangshun, Hilary Chappell, Redouane Djamouri, and Thekla Wiebusch (eds.), Breaking down the Barriers: Interdisciplinary Studies in Chinese Linguistics and Beyond, 755–784. Taiwan: Institute of Linguistics, Academia Sinica. Heine, Bernd and Tania Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva. 2003. Contact-induced grammaticalization, Studies in Language 27: 529–572. Heine, Bernd and Tania Kuteva. 2005. Language Contact and Grammatical Change. Cambridge: Cambridge University Press. Heine, Bernd and Tania Kuteva. 2006. The Changing Languages of Europe. Oxford: Oxford University Press. Heine, Bernd and Motoki Nomachi. 2013. Contact-induced replication: Some diagnostics. In: Martine Robbeets, and Hubert Cuyckens (eds.), Shared Grammaticalization (With Special Focus on the Transeurasian Languages), 67–100. (Studies in Language Comapnion Series 132.) Amsterdam/Philadelphia: John Benjamins. Hock, Hans Heinrich. 1986. Principles of Historical Linguistics. Berlin: Mouton de Gruyter. Keesing, Roger M. 1991. Substrates, calquing and grammaticalization in Melanesian Pidgin. In: Elizabeth C. Traugott and Bernd Heine (eds.), Approaches to Grammaticalization, vol. 1: 315–342. (Typologiacal Studies in Language 19.) Amsterdam/Philadelphia: John Benjamins. König, Ekkehard and Martin Haspelmath. 1999. Der europäische Sprachbund. In: Norbert Reiter (ed.), Eurolinguistik: ein Schritt in die Zukunft, 111–127. Wiesbaden: Harrassowitz. Kuteva, Tania. 1994. Iconicity and auxiliation. Journal of Pragmatics 22: 71–81. Kuteva, Tania. 1998. Large linguistic areas in grammaticalization: auxiliation in Europe. Language Sciences 20 (3): 289–311. Narrog, Heiko and Johan van der Auwera. 2012. Grammaticalization and semantic maps. In: Heiko Narrog and Bernd Heine (eds.), The Oxford Handbook of Grammaticalization, 318–327. Oxford: Oxford University Press. Nomachi, Motoki and Bernd Heine. 2011. On predicting contact-induced grammatical change: Evidence from Slavic languages. Journal of Historical Linguistics 1 (1): 48–76.
13 On formulas of equivalence in grammaticalization
451
Ross, Malcolm D. 1996. Contact-induced change and the comparative method: Cases from Papua New Guinea. In: Mark Durie and Malcolm D. Ross (eds.), The Comparative Method Reviewed: Regularity and Irregularity in Language Change, 180–217. New York: Oxford University Press. Ross, Malcolm D. 1997. Social networks and kinds of speech-community event. In: Roger Blench and Matthew Spriggs (eds.), Archaeology and Language I: Theoretical and Methodological Orientations, 209–261. London/New York: Routledge. Ross, Malcolm D. 2001. Contact-induced change in Oceanic languages in North-West Melanesia. In: Alexandra Y. Aikhenvald and Robert M. W. Dixon (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics, 134–166. Oxford: Oxford University Press. Thurston, William R. 1982. A Comparative Study in Anêm and Lusi. (Pacific Linguistics B-83.) Canberra: The Australian National University. Thurston, William R. 1987. Processes of Change in the Languages of North-Western New Britain. (Pacific Linguistics B-99.) Canberra: The Australian National University. Uwajeh, M. K. C. 2007. Translation Equivalence: An Essay in Theoretical Linguistics. Munich: LINCOM Europa. van der Auwera, Johan (ed.). 1998. Adverbial Constructions in the Languages of Europe. (Empirical Approaches to Language Typology, EUROTYP, 20-3.) Berlin: Mouton de Gruyter. Weinreich, Uriel. 1964. Languages in Contact. Findings and Problems. London/The Hague/ Paris: Mouton. Weiss, Daniel. 2004. The rise of an indefinite article: The case of Macedonian eden. In: Walter Bisang, Nikolaus P. Himmelmann, and Björn Wiemer, What Makes Grammaticalization? A Look from its Fringes and its Components, 139–165. (Trends in Linguistics Studies and Monograph 158) Berlin/New York: Mouton de Gruyer. Whorf, Benjamin Lee. 1941. The relation of habitual thought and behavior to language. In: Leslie Spier (ed.), Language, Culture, and Personality, Essays in Memory of Edward Sapir, 75–93. Menasha, WI: Sapir Memorial Fund. Whorf, Benjamin Lee. 1956. Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf. Cambridge, MA: MIT Press. Wiemer, Björn, Berhard Wälchli, and Björn Hansen (eds.). 2012. Grammatical Replication and Borrowability in Language Contact. (Trends in Linguistics Studies and Monograph 242.) Berlin/New York: Mouton de Gruyter.
Motoki Nomachi
14 Placing Kashubian on the language map of Europe 1 Introduction Kashubian, including the extinct Slovincian dialects, is one of the West Slavic languages spoken in the Pomeranian Voivodeship of Poland. It has not been a subject of particular analysis in the areal typology, with only a few exceptions in which this language served merely as a source of some episodic examples (see Décsy 2000: 149 for the “Rokytno zone”; Heine and Kuteva 2006: 157 for the European periphery) or even had no example provided (Dahl and Koptjevskaja-Tamm 2001 for the Circum-Baltic area). The reasons for this neglect are multiple. First, with regard to Kashubian itself, for a long time this language was dealt with as a dialect of Polish.1 In addition, although some significant achievements in this realm have been made by Polish linguists such as Stieber and Popowska-Taborska (1964–1978), Polish dialectology did not pay much attention to morphosyntax, whose elaboration remains incomplete in Kashubian (Popowska-Taborska 1980: 36). Second, particularly in the earlier stage of areal typology, the main objects of analysis were standard languages or well-established dialects/language varieties or generalized language types, which can be seen, for instance, in Haspelmath’s (1998, 2001) description of Standard Average European (henceforth, SAE). Indeed, the micro-orientation on specific non-standardized languages is one of the important directions after the EUROTYP work (van der Auwera 2011: 301). The material, used even in those studies that have mentioned Kashubian independently of Polish, is often problematic. The most elaborated work on Kashubian that covers syntax is Lorentz’s Gramatyka pomorska (Pomeranian Grammar, 1927–1937), and his publications, including earlier works such as Slovinzische Grammatik (1903), are often sources for areal typological studies from Jakobson’s Eurasian Sprachbund (1931) onward. However, Lorentz conducted his research in the late 19th – early 20th centuries, when the sociolinguistic situation among Kashubs was totally different from now, as they lived
1 The linguistic status of Kashubian has been disputable. Some scholars tended to treat it a distinct dialect of Polish, while others were of opinion that it is a separate language. Since 2005, the Kashubian language has been officially recognized as a separate regional language in Poland. https://doi.org/10.1515/9783110639223-015
454
Motoki Nomachi
in a polity where German was the dominant official language. All Kashubs were bilingual (Kashubian and German), trilingual (Kashubian, German and Polish) or even quadrilingual (Kashubian, (High) German, Low German, and Polish).2 Thus, one should be careful in using the materials of Lorentz and his contemporaries, as they do not represent the current situation, which drastically changed at the end of World War II. German is no longer dominant and there is little German multilingualism among Kashubs, while Polish has instead become the locally dominant language that all Kashubs master, and even speak better than Kashubian in many cases. Therefore, for an areal typology that deals with Kashubian, one should consider linguistic changes caused by the aforementioned, relatively rapid sociolinguistic changes – a situation not often seen with “bigger” languages such as the major SAE languages, which have had rather stable linguistic structures. Taking all the above aspects into consideration, in this chapter I analyze the dynamics of Kashubian morphosyntactic features based on fine-grained dialect data in order to place this language areal-typologically in the context of SAE. In Section 2, I discuss the relevant features according to which Kashubian should be analyzed; in Sections 3 to 7, I analyze five morphosyntactic areal features shared with SAE in general and how they have changed in the history of Kashubian. The examples used in this article are taken from various sources, including some from 19 native speakers from all three main dialects (north, central, and south).
2 Areal features of SAE and Slavic According to Haspelmath (2001: 1492–1510), there are twelve major SAE features: a. definite and indefinite articles, b. relative clauses with relative pronouns, c. the ‘have’-perfect, d. nominative experiencers, e. the participial passive, f. anticausative prominence, g. dative external possessors, h. negative pronouns and lack of verbal negation, i. particles in comparative constructions, j. relative-based equative constructions, k. subject person affixes as strict agreement markers, and l. intensifier-reflexive differentiation. Haspelmath also identifies some additional likely SAE features: m. verbal fronting in polar interrogatives, n. comparative marking of adjectives, o. “A-and-B” conjunction, p. comitative–instrumental syncretism, and q. a suppletive second ordinal.3
2 For the dynamics of multilingual situations among Kashubs, see Treder (2005: 9–24). 3 Haspelmath also provides some features that are probably non-existent in SAE.
14 Placing Kashubian on the language map of Europe
455
Considering nine of the aforementioned features (a, b, c, e, f, h, j, k, l), Haspelmath finds that German and French get the highest score (9/9) among major European languages, while most Slavic languages are somewhat peripheral, scoring six points (Czech)4 or five points (Polish, Russian, Ukrainian, Bulgarian, Serbo-Croatian, Slovene).5 Kashubian shares with other Slavic languages, including Polish, the aforementioned five features (b, e, f, j, l). Leaving them aside, I concentrate on the remaining four features of Haspelmath’s nine (a, c, h, k) because most Slavic languages do not have them and they are features that separate most Slavic languages from SAE. Also, having been in close contact for centuries with German, which does possess these four features, Kashubian might have incorporated them into its grammatical system. In addition to these major criteria, among Haspelmath’s likely SAE features, feature (p) above seems to be particularly relevant not only to the search for the properties of SAE (Heine and Kuteva 2006: 183–203; Stolz 2009: 399) but also to a typology of Slavic languages, especially in the context of language contact, as has been shown by Nomachi and Heine (2011). Also, while a “check-list” type survey of the existence/non-existence of a given category in a particular language is helpful at earlier stages for profiling the language and identifying linguistic areas, one has to keep in mind that the evolution of grammatical categories is often gradual and slow, entailing a transitional situation where two or more layers of a certain category can be found at the same time. Therefore, to assess a given grammatical phenomenon properly, one should pay attention, for instance, to its degree and direction of grammaticalization, or, as Heine and Kuteva (2006) have discussed properly, its diachrony and micro-areal distribution.
4 Haspelmath includes Czech as a language that features a have-perfect. There indeed is a construction with the verb mít ‘to have’ and past passive participle that is reminiscent of the have-perfect in West European language, although it can hardly be described as a true perfect. First, it has a meaning of resultative. Second, morphologically, it cannot be derived from intransitives. Finally, semantically, not all transitive verbs can derive this resultative construction; for details, see Giger (2003). If Czech were included among the languages that possess a have-perfect, then, at least, Polish, Serbo-Croatian, Slovene, and Bulgarian would also have to be included, because they have the same resultative construction; for details on this matter in Polish and South Slavic, see Nomachi (2006) and Nomachi (2015) respectively. 5 Haspelmath does not provide any data from Lower and Upper Sorbian, Slovak, Belarusian, and Macedonian.
456
Motoki Nomachi
3 Definite and indefinite articles: feature (a) 3.1 The definite article Common Slavic did not have articles in general (Wackernagel 2009: 598), yet some but not all Slavic languages later developed a fairly highly grammaticalized definite article, including Resian Slovene (Benacchio 2014; also Gvozdanović in this volume), Molise Slavic (Breu 2012, also Breu in this volume), and a colloquial variety of Upper Sorbian (Scholze 2012). Similar phenomena can also be found in several additional varieties, although not highly grammaticalized and therefore best not referred to with the term “definite article”: (colloquial) Czech (Krámský 1972; Berger 1993; Trovesi 2004), Slovene (Trovesi 2004), and Burgenland Croatian (Neweklowsky 1978). There are two common points to be mentioned here. First, all these varieties have experienced heavy language contact with non-Slavic languages that possess a definite article, namely, either Romance or Germanic; and second, in all instances, the source of the “definite article” is a demonstrative pronoun that has been (to a greater or lesser degree) grammaticalized, as is also the case in the historical grammaticalization in the putative donor languages. In Kashubian, according to Lorentz (1958: 41, 1962: 21), the demonstrative pronouns ten/ta/to, nen/na/no ‘that’ can be used as definite articles,6 and this property came into being under the influence of German. However, Lorentz does not give any definition of, or criteria for, a definite article nor does he provide any concrete analysis of his material (see also Nau 1995: 114), and neither do Boryś and Popowska-Taborska (1999: 319). On the other hand, Cybulski and Wosiak-Śliwa (2001: 187) point out that in Kashubian a demonstrative pronoun is overtly used for marking definiteness, without analyzing what kind of functions it has. Thus, one may wonder if the usages in question indeed entail a definite article or just a definite-article-like usage of a demonstrative pronoun. To this end, both formal and semantic aspects should be considered. Regarding the formal aspect, one important observation Lorentz made in this respect is that, in the Slovincian dialects, the demonstrative pronouns ten/ta/to and nen/na/no have two declension patterns, depending on the existence of phonological stress (Lorentz 1903: 276–277). This is thus a case of phonetic reduction, which is one of the universal parameters for grammaticalization
6 There is no difference in the use of ten/ta/to and nen/na/no. For the sake of simplicity, I will use the writing system of the contemporary Kashubian language except for Slovincian dialects.
14 Placing Kashubian on the language map of Europe
457
(Heine 2005: 579). Indeed, the non-stressed form is used only when the usage fulfills some functions of a definite article (Lorentz 1903: 222). Compare the nonstressed form tĕn in place of the stressed form tìĕn in tĕn χlùŏp ‘the guy,’ and tå in place of the stressed full form tìĕvå/tåŭ in dűŏ tå Lépskå jìĕzoră ‘to the Lipsk lake’ to the stressed form tìen in tìen přäšed ‘he (< that) arrived’, see example (1) (Lorentz 1905: 1). (1)
Tò bél jädĕn χlùŏp, a that be.APP.SG.M one.NOM.SG.M man.NOM.SG.M and tĕn χlùŏp są nãzẹvoŭl Gřänáŭtă. that.NOM.SG.M man.NOM.SG.M REFL call.APP.SG.M Grzenauta Tìen přäšed dűŏ tå Lépskå jìĕzoră. that.NOM.SG.M come.APP.SG.M onto that.GEN.SG.M Lipsk Lake ‘There was a man. The man’s name was Grzenauta. He came to Lipsk Lake.’
Considering the fact that the demonstrative pronoun seems to have a short form in use as a result of erosion (Heine 2005: 579), it might indeed be on the way to grammaticalization in Kashubian, as is the case in colloquial Upper Sorbian (Scholze 2012: 325). However, according to Piotrowski (1981: 44), one cannot find concrete rules for using a demonstrative pronoun as a definite article. It is important to note that this two-declension patterning is characteristic only for Slovincian dialects; other Kashubian dialects seemingly do not have a short form, which makes it difficult to learn the evolutionary path of the Kashubian definite marker. In the earliest dialectal recordings by Hilferding ([1862] 1965: 416) in the mid–19th century, one can find examples in which even for the most basic anaphoric meaning a demonstrative pronoun was not used as a definite marker. This means that the use of the demonstrative pronoun in the position was not obligatory, as in (2). Indeed, in Hilferding’s texts, demonstrative pronouns are used only as anaphoric forms, which is not the case in Lorentz’s texts (see below). (2)
Bél mlénorz, ten mial trze be.APP.SG.M miller.NOM.SG.M that.NOM.SG.M have.APP.SG.M three corki, a wszetkie trze bele moerą. daughter.NOM.PL.F and all.NOM.PL three be.APP.PL ghost.INS.SG.F Ø Mlénorz nie widzal nic. miller.NOM.SG.M not see.APP.SG.M nothing ‘There was a miller. He had three daughters and all three were ghosts. The miller did not see anything. . .’
458
Motoki Nomachi
According to Knoll (2012: 47–48), based on his analysis of Lorentz’s texts recorded in the late 19th–20th centuries, one can find a wide range of usages of demonstrative pronouns that are close to the functions of a definite article – from the most concrete anaphoric function exemplified by (3), which is the lowest stage of grammaticalization of a definite article, to the most abstract, generic function, as in (7), showing the highest degree of grammaticalization (see Heine 2012: 132 for the functional classification of definite articles).7 One can also find cataphoric examples, like (4), associative-anaphoric like (5), and abstract-situational, as in (6).8 (3)
Rôz bëlë jedny lëdze, tï mielë once be.APP.PL one.NOM.PL people.NOM.PL that.NOM.PL have.APP.PL dwòje dzecy, to jedno so zwa Hans a two child.NOM.PL that one.NOM.SG.N REFL call.APP.SG.N Hans and to drëdżé Gréta. A tej ti lëdze that other.NOM.SG.N Greta and then that.NOM.PL people.NOM.PL ni mielë nic do jedzeni. not have.APP.PL nothing onto eating.GEN.SG.N ‘Once there were certain people. They had two children. One’s name was Hans and the other’s name was Greta. And then, the people did not have anything to eat.’
(4)
Òn mieszkôł w ti karczmie, co terô he.NOM live.APP.SG.M in that.LOC.SG.F inn.LOC.SG.F which.NOM now nôleżi Panu Lëtwinu. belong.PRS.3SG nobleman.DAT.SG.M Litwin.DAT ‘He lived in the inn that now belongs to Mr. Litwin.’
(5)
Òn szedł he.NOM go.APP.SG.M dostôł całi get.APP.SG.M whole
kòżdi every tidzéń week
dzéń na rëbë, day onto fish.ACC.PL nic. Ta nothing that.NOM.SG.F
ale òn but he.NOM biéda poverty.NOM.SG.F
7 For the German equivalent of (7) it is noteworthy that in this case the definite article can be omitted: leben wie (die) Herren ‘to live like noblemen’. 8 For details, see Hawkins (1978) and Heine and Kuteva (2006). According to Heine (2012: 132), grammaticalization of definite articles goes through the following six stages: 1. Demonstrative, 2. Anaphoric-definite, 3. Contextually definite, 4. Semantically definite, 5. Indefinite specific, 6. Article loss. German has reached Stage 4, while Colloquial Upper Sorbian is at Stage 3.
14 Placing Kashubian on the language map of Europe
459
bëła wiôlgô. be.APP.SG.F big.NOM.SG.F ‘He went fishing every day, but he did not get anything for a week. The poverty was enormous.’ (6)
Te òna czëła, że ti lëdze then she.NOM hear.APP.SG.F that that.NOM.PL people.NOM.PL gôdelë, żebë ten król say.APP.PL SUBJ that.NOM.SG.M king.NOM.SG.M zarô przëjachôł. immediately come.APP.SG.M ‘Then she heard that the people said that the King would come immediately.’
(7)
Òna chce dobri dwór miec she.NOM want.PRS.3SG good.ACC.SG.M palace.ACC.SG.M have.INF a mieszkac jak ti pónowie. and live.INF as that.NOM.PL nobleman.NOM.PL ‘She wants to have a nice palace and to live like noblemen.’
According to Himmelmann (2001: 833), cross-linguistically, two usages can be the key for identifying definite articles: the associative anaphoric use as in (5) and the larger situation use as in (6), which is “the first mention of entities that are considered to be unique, and hence generally identifiable, in a given speech community.” One can recall here Knoll’s (2012: 48) conclusion that the Kashubian demonstrative pronouns are highly grammaticalized, but that this process was not finished at the time of Lorentz’s observations. Knoll’s observation is also supported by the fact that the use of demonstrative pronouns is not always consistent.9
9 However, one has to admit that the abstract situational and the generic meanings are rather rare and far less grammaticalized than others. For the abstract situational meaning, a word such as słuńce ‘sun’ may serve as an example. Although it denotes a unique item, this word does not take a definite marker. Here is an example from Lorentz (1924: 813): Słuńce zachòdzë w òdleglim lase ‘The sun sets in a distant forest’. With regard to the generic usage, according to the material in Tetzner (1899: 243), a Slovincian saying such as Japke pąnje nie dalik wod jablone ‘The apple does not fall far from the tree’ does not take a definite marker for japke ‘apple,’ which conveys a generic meaning. Compare its German equivalent: Der Apfel fällt nicht weit vom Stamm.
460
Motoki Nomachi
How do these usages look now in Kashubian, which is not under German influence anymore? According to my consultants, the equivalents of (3)–(7) respectively would be (8)–(12) today: (8)
Bëlë so jedny lëdze, co mielë be.APP.PL REFL one.NOM.PL people.NOM.PL that.NOM.PL have.APP.PL dwòje dzecy. Jedno sã zwało Hans, a Hans and two child.NOM.PL one.NOM.SG.N REFL call.APP.SG.N drëdżé Gréta. Ti lëdze ni other.NOM.SG.N Greta that.NOM.PL people.NOM.PL not mielë nic do jedzeniô. have.APP.PL nothing onto eating.GEN.SG.N ‘Once there were certain people. They had two children. One’s name was Hans and the other’s name was Greta. And then, the people did not have anything to eat.’
(9)
Òn mieszkôł w ti/Ø karczmie, co terô he.NOM live.APP.SG.M in that.LOC.SG.F inn.LOC.SG.F which.NOM now słëchô panu Lëtwinowi. belong.PRS.3SG nobleman.DAT.SG.M Litwin.DAT ‘He lived in the inn that now belongs to Mr. Litwin.’
(10) Òn szedł kòżdi dzéń na rëbë, ale bez he.NOM go.APP.SG.M every day onto fish.ACC.PL but through całi tidzéń nic nie dostôł. Ta/Ø Biéda whole week nothing not get.APP.SG.M that.NOM.SG.F poverty.NOM.SG.F bëła wiôlgô. be.APP.SG.F big.NOM.SG.F ‘He went fishing every day, but he did not get anything for a week. The poverty was enormous.’ (11) Tej òna czëła, że ti lëdze then she.NOM hear.APP.SG.F that that.NOM.PL people.NOM.PL gôdelë, żebë Ø król zarô przëjachôł. king.NOM.SG.M immediately come.APP.SG.M say.APP.PL SUBJ ‘Then she heard that the people said that the King would come immediately.’
14 Placing Kashubian on the language map of Europe
461
(12) Òna chce miec bëlny dwór she.NOM want.PRS.3SG have.INF nice.ACC.SG.M palace.ACC.SG.M i mieszkac jak Ø pónowie. nobleman.NOM.PL and live.INF as ‘She wants to have a nice palace and to live like noblemen.’ It is suggestive that the contemporary use of demonstrative pronouns as definite markers in Kashubian has become very close to that of Polish, as exemplified by (13)–(17): (13) Byli sobie pewni ludzie, którzy be.APP.PL REFL certain.NOM.PL people.NOM.PL that.NOM.PL mieli dwoje dzieci. Jedno miało na child.NOM.PL one.NOM.SG.N have.APP.SG.N onto have.APP.PL two imię Hans, a drugie Greta. Owi/Ci name.ACC.SG.N Hans and other.NOM.SG.N Greta that.NOM.PL ludzie nie mieli nic do jedzenia. people.NOM.PL not have.APP.PL nothing onto eating.GEN.SG.N ‘Once there were certain people. They had two children. One’s name was Hans and the other’s name was Greta. And then, the people did not have anything to eat.’ (14) On mieszkał w Ø/tej karczmie, która he.NOM live.APP.SG.M in Ø/that.LOC.SG.F inn.LOC.SG.F which.NOM teraz należy do pana Litwina. now belong.PRS.3SG onto nobleman.GEN.SG.M Litwin.GEN ‘He lived in the inn that now belongs to Mr. Litwin.’ (15) Każdego dnia szedł na ryby, ale every.GEN.SG.M day.GEN.SG.M go.APP.SG.M onto fish.ACC.PL but przez cały tydzień nic nie złowił. through whole week nothing not catch.APP.SG.M Ø Bieda to była wielka. poverty.NOM.SG.F COP be.APP.SG.F big.NOM.SG.F ‘He went fishing every day, but he did not get anything for a week. The poverty was enormous.’
462
Motoki Nomachi
(16) Wtedy usłyszała, że ci ludzie mówili, then hear.APP.SG.F that that.NOM.PL people.NOM.PL say.APP.PL żeby Ø król zaraz przyjechał. SUBJ king.NOM.SG.M immediately come.APP.SG.M ‘Then she heard that the people said that the King would come immediately.’ (17) Ona chce mieć piękny dwór she.NOM want.PRS.3SG have.INF nice.ACC.SG.M palace.ACC.SG.M i mieszkać jak Ø panowie. nobleman.NOM.PL and live.INF as ‘She wants to have a nice palace and to live like noblemen.’ However, according to my consultants, the associative-anaphoric use of a demonstrative pronoun is sometimes possible as in (18),10 which is not the case with Polish as in (19)11: (18) Òna dzysô jidze na ten/jaczis she.NOM today go.PRS.3SG onto that/certain.ACC.SG.M film, bò òna znaje film.ACC.SG.M because she.NOM know.PRS.3SG tegò/Ø reżisera. that.ACC.SG.M director.ACC.SG.M ‘Today she is going to the/a film, because she knows the director.’ (19) Ona dzisiaj she.NOM today zna Ø know.PRS.3SG
pójdzie na film, bo go.NPST.3SG onto film.ACC.SG.M because reżysera. director.ACC.SG.M
Here it is important to note that in (18), according to my consultants, the demonstrative ten ‘that’ or the indefinite marker jaczis ‘certain’ should
10 The same account is valid for (10) with regard to the presence and absence of the demonstrative pronoun ta ‘that’. 11 The original sentence is in colloquial Upper Sorbian (see Scholze 2012: 327); it is adapted here. According to Tomasz Wicherkiewicz (personal communication in 2018), in a variety of colloquial Polish spoken in the Poznań region, the demonstrative pronoun ten can be used in the same context as in colloquial Upper Sorbian, but its usage is never obligatory.
14 Placing Kashubian on the language map of Europe
463
accompany the word film ‘film’ for actualizing it for the associative-anaphoric use of a demonstrative pronoun tegò in the second part of the sentence. To sum up, at this stage, Kashubian seems to have a tendency to degrammaticalize the definite marker by replacing the German model with the Polish one, but with one exception: the associative-anaphoric usage of the demonstrative pronoun. To the best of my understanding, demonstrative pronouns cannot be used for the abstract-situational and generic meanings in modern Kashubian.
3.2 The indefinite article Lorentz (1958: 41) also argues that Kashubian possesses an indefinite article with the numeral jeden/jedna/jedno ‘one.’ The same or similar phenomena are found in the aforementioned Slavic languages that have had close contact with Romance or German (Section 3.1). Again, Lorentz did not explain situations in which it can be used as an indefinite article. Typologically, according to Givón (1981: 52), it is most probable both theoretically and empirically that on its way to becoming an indefinite article, a grammaticalized item passes through an intermediate stage, namely that of a numerically neutral indefinite determiner (see Schroeder 2006: 556). Crosslinguistically, Heine (1997: 71–77, 2012: 134) posits the following evolutionary model in which Stage 1 is the oldest and Stage 5 is the most recent: Stage 1: The numeral – lack of an indefinite article Stage 2: The presentative marker – introduction of a new participant unknown to the hearer (or introduction of a topic-to-be) Stage 3: The specific marker – introduction of any participant known to the speaker, but presumed to be unknown to the hearer (or introduction of a specific, but non-topical participant) Stage 4: The non-specific marker – introduction of a participant whose referential identity neither the hearer nor the speaker knows or cares to know Stage 5: The generalized article – this can be expected to occur in all types of nouns (which are not definite) The difference between Stages 2 and 3 here lies in that in Stage 2 the new participant will become definite in subsequent discourse, while Stage 3 does not always have to show such an implication. According to Heine (1997: 71), this model can be interpreted in two ways, that is, either in terms of the diachronic
464
Motoki Nomachi
evolution of an indefinite marker, or with a synchronic implication. If a language has an indefinite marker of stage X, then this implies that that indefinite marker is used also for stages preceding X. As with a definite marker, it seems difficult to establish the grammaticalization pattern for an indefinite marker. In the earliest dialect texts recorded by Hilferding ([1862] 1965: 403), there are examples without the Stage 2 use of jeden ‘one,’ as in (20): (20) Tą bél rôz Ø bùr, a mial landlord.NOM.SG.M and have.APP.SG.M there be.APP.SG.M once trze córki. three daughter.ACC.PL ‘Once upon a time, there was a landlord and had three daughters.’ In contrast, in the texts recorded by Lorentz (1924: 393), the frequency of jeden is in general higher, as illustrated in (21): (21) W jedny wsë béł jeden in one.LOC.SG.F village.LOC.SG.F be.APP.SG.M one.NOM.SG.M bògati pón a jeden ùbòdżi rich.NOM.SG.M nobleman.NOM.SG.M and one.NOM.SG.M poor.NOM.SG.M széwc. Ten pón miôł cobbler.NOM.SG.M that.NOM.SG.M nobleman.NOM.SG.M have.APP.SG.M jedną córkã a ten széwc one.ACC.SG.F daughter.ACC.SG.F and that.NOM.SG.M cobbler.NOM.SG.M jednégò sëna. one.ACC.SG.M son.ACC.SG.M ‘In a certain village, there was a rich nobleman and a poor cobbler. The nobleman had a daughter and the cobbler had a son.’ In this story, wies ‘village’ in the phrase w jedny wsë ‘in a village’ is not mentioned later, which means that it remains indefinite. This indefinite marker can serve therefore as a case of Stage 3. The other instances of the numeral jeden in collocation with nouns (bògati pón ‘rich man,’ ùbòdżi széwc ‘poor cobbler,’ córka ‘daughter,’ and syn ‘son’) can be regarded as examples of Stage 2, as they become later definite entities in the story. Although the frequency of the numeral jeden was high in Lorentz’s data, as in (21), Lorentz’s material shows that Kashubian had hardly reached Stage 4 in Heine’s scale, nor does it seem to have taken a shortcut to Stage 5 either with
14 Placing Kashubian on the language map of Europe
465
some exceptions.12 On the contrary, the use of the numeral jeden was not always consistent even in Stages 2 and 3 (see Knoll 2012: 50). Kashubian retains today almost the same level of grammaticalization of the numeral jeden as an indefinite marker, because none of my six consultants from various regions of Kashubia permits any example of Stage 4 as in (22) or Stage 5 as in (23): (22) Zawòłôj *jednégò/Ø dochtora! doctor.ACC.SG.M call.IMPER.2SG *one/Ø ‘Call a doctor!’ (23) *Jeden/Ø słoń je *jednym/Ø zwiérzãcém. animal.INS.SG.M *One/Ø elephant be.PRS.3SG *one/Ø ‘The elephant is an animal.’ In this context, it is worth comparing a text recorded by Lorentz (1924: 392) in Tłuczewo in Central Kashubia, in (24), with a contemporary version edited by inhabitants of the same village in 2014, in (25); the villagers do not speak German at all. (24) Lorentz’s original text recorded in Tłuczewo in the early 20th century: Rôz szedł jeden lesny w las ë widzôł jednã gapã a chcôł ją strzelac. Ta gapa rzekła: „Nie ùstrzél mie, jô jem córka królewskô a jem przeklãtô, ale të mie możesz zbawic.” Ten lesny pitôł sã: „Co jô móm robic, żebë ce zbawic?” Ta gapa òdpòwiedza: „Biôj tą drogą dali, tej të przińdzesz do jedny chałëpë. Tam sedzy jedna stôrô baba. Òna cë dô jesc i pic, ale nie wez nick, bò czej të bãdzesz jôdł a pił, tej të mie nie bãdzesz mógł zbawic. Biôj przez tã chałëpã, tej przińdzesz w jeden ògród. . . (25) The 2014 version of Lorentz’s text, edited by two current residents of Tłuczewo:
12 In a text recorded in Pażęce (Central Kashubia) by Lorentz (1924: 587), one finds jeden in a predicate as follows: To je jedna krowa, a nié jedna kòza ‘This is a cow, not a goat’. This example may mean that Kashubian jeden reached Stage IV. However, examples of this usage of jeden are very rare in Lorentz’s text, and so Lorentz’s example could be rather accidental. To add a diachronic dimension, it should be noted that during my field research conducted in 2018, all three of my consultants in Pażęce asserted that today no one there spoke as in Lorentz’s text, and for the construction to be acceptable, it should not contain jeden, just as in Polish, i.e., To je Ø krowa, a nié Ø kòza.
466
Motoki Nomachi
Rôz szedł jeden lesny w las ë widzôł Ø gapã i chcôł do ni strzelac. Ta gapa rzekła: „Nie ùstrzél mie, jô jem córka królewskô, a jem przeklãtô, ale të mie możesz zbawic.” Ten lesny pitôł sã: „Co jô móm robic, żebë ce zbawic?” Gapa òdpòwiedza: „Biôj tą drogą dali, tej të przińdzesz do Ø chałëpë. Tam sedzy Ø storô baba. Òna cë dô jesc i pic, ale nie wez nick, bò czej të bãdzesz jôdł ë pił, tej të mie nie bãdzesz mógł zbawic. Biôj przez tã chałëpã, tej przińdzesz w Ø ògród. . . ‘Once upon a time a forester went into the forest. There he saw a crow and wanted to shoot it. The crow said: “Don’t shoot me! I am a princess and I am cursed, but you can save me.” The forester asked: “What should I do to save you?” The crow answered: “Go farther this way, then you will come to an old hut. There is an old woman sitting there. She will offer you something to eat and drink, but don’t take anything, because if you eat and drink, then you will not be able to save me. Go through the old hut, then you will come to a garden . . . ’ A comparison of one text, edited by two people from a single village, may not give a definitive answer regarding the grammaticalization of the numeral jeden, even though both consultants provided identical linguistic judgments. Yet it does suggest that the frequency of the numeral jeden is reduced greatly – in the example, by four-fifths – although all five examples equally have a Stage 2 meaning as presentative markers of new participants in the narrative (lesny ‘forester,’ gapa ‘crow,’ stôrô baba ‘old woman,’ chałëpa ‘hut’ and ògród ‘garden’). The first use of jeden in the text, jeden lesny, is rather idiomatic (Sychta 1968: 92), and thus it might not be omitted even today.13 The reason for the reduced frequency in other cases may be explained by the fact that in Lorentz’s version jeden always appears in the rhematic position providing new information. Thus, as long as one does not mean to emphasize the indefiniteness, word order can work to express indefiniteness without using jeden. Thus, in the above examples, the overt use of jeden could even be redundant. The higher frequency of a given grammatical form, including its obligatoriness as in the Stage 2 usage evident in Lorentz’s example (24), can be assumed to mean its regularization, eventually leading to a higher degree of grammaticalization of
13 According to Dieter Stern (personal communication in 2018), the participant introduced first into a story usually ranks higher in topic prominence than all other participants introduced subsequently. The numeral ‘one’ serves to introduce the most prominent topic, the ‘hero’ of the narrative, around whom everyone and everything else is grouped and through whose eyes the listener perceives the story. Thus, this could be a functional difference.
14 Placing Kashubian on the language map of Europe
467
a given phenomenon. If that assumption is true, then one could say that in present-day Kashubian the degree of jeden’s grammaticalization as an indefinite marker has been regressing since the time of Lorentz’s observation. One could also still say that the indefinite marker jeden remains at Stage 2, because it would not be wrong to insert it, according to my consultants.14 Overall, Kashubian has never possessed an indefinite article per se, as a fully grammaticalized item, even under strong German influence, though there might have been some tendency to grammaticalization at that point. Today, this grammaticalization not only does not continue but even progresses in the opposite direction. In other words, degrammaticalization of indefinite markers is under way.
4 The have-perfect: feature (c) As mentioned by Lorentz (1958: 41), Kashubian incorporated have- and beperfects in its verbal system. This is particularly interesting because the other West Slavic languages, including Sorbian, and some South Slavic languages, e.g., Slovene and Burgenland Croatian, that had or still have close contact with German, have not grammaticalized the construction as much as Kashubian (except for the extinct Polabian, see Lehr-Spławiński 1929: 235).15 According to Lorentz (1962: 1057), have- and be-perfects either convey the present perfect or confirm facts and events taken from the past.16 To express these notions, both use the auxiliaries, miec ‘to have’ or bëc ‘to be,’ and past participles in -n-, -t-, and -l-.17 Although Drinka (2017: 383) states that “the -l-
14 According to these consultants, one could insert jeden as in Lorentz’s text, but it would be redundant from a semantic viewpoint. This would presumably mean that the perception of the function of jeden has reversed to Stage 1. 15 According to Pelka (2015: 151), the Silesian ethnolect has a resultative construction with the verb ‘to have’. Indeed, Tambor (2006: 170) provides the following example (to be sure, without any appropriate interpretation): mjała to wyskoczone / wylezióne (w kolanie) ‘She had jumped / crawled out (on her knee)’. Considering the fact that these PPPs are derived from the intransitive verbs (wyskoczyć ‘jump out’, wyleźć ‘crawl out’) as well, this construction may represent a higher degree of grammaticalization than the Polish have-perfect, and, naturally, it has been explained as a grammatical calque of the German haben-periphrastic. However, to the best of my knowledge, the degree of the grammaticalization of the Silesian have-perfect has yet to be determined. 16 It is important to note that the be-perfect in modern Kashubian does not mean the (actional) perfect per se, it has the resultative meaning (Nomachi 2016). 17 Décsy (2000: 150) describes these periphrastic forms as follows: “The compound (past tense) is not formed with a verb of being (as in the other Slavic languages) but with the verb of
468
Motoki Nomachi
participle can also be used with the HAVE auxiliary”, one should not confuse the -l- active past participle used in the simple past tense such as jô rzekł ‘I said’ with the past participle in -l- in periphrastic constructions calqued from German such as jô móm rzekłé ‘I have said’. The former must be a short form used only as part of the past tense, while the latter is a long form which can also be used as an adjective (Nomachi 2016: 272). Lorentz (1962: 981) claims that the have-perfect is used with intransitives other than verbs of motion, transitive and reflexive verbs, while the be-perfect is used with intransitive verbs of motion, closely following the distribution of the auxiliaries haben ‘to have’ and sein ‘to be’ in German as seen in (26) and (27) respectively: (26) Jô móm tã białkã widzałé.18 I.NOM have.PRS.1SG that.ACC.SG.F woman.ACC.SG.F see.PP.SG.N ‘I have seen the woman.’ (27) Ta białka je precz jidzonô. that.NOM.SG.F woman.NOM.SG.F be.PRS.3SG away go.PP.SG.F ‘The woman is gone away.’ Both patterns were very productive and already fairly highly grammaticalized in the mid-19th century, judging from the texts recorded by Hilferding. However, in religious texts dated from the 16th century up to the 18th century and translated from German and/or Polish into Slovincian, there is not a single example of either form19; German haben periphrases are always translated into the regular past tense form, which consists of the auxiliary bëc ‘to be’ and the short -l- active past participle. One can assume that these periphrases may have appeared appear in Kashubian in the early 19th century. It is impossible, however, to trace the concrete steps in the process of
having (as in English and German),” providing the example ja mam widzeł. Obviously, this is not just a past tense, as will be discussed in this section. His example is also wrong, as the short -l-participle cannot be a part of the have-perfect. 18 In present day Kashubian, the form widzałé ‘seen’ is rather rare. Instead, the form widzóné is more often used. 19 All German haben-periphrastic forms are translated into preterit forms that consist of the auxiliary ‘to be’ and the -l-participle. Compare the German original and its translation in Slovincian in the 18th century: Du hast ausgestanden and Znosiel jes ‘You (have) endured’ (Hinze 1967: 12).
14 Placing Kashubian on the language map of Europe
469
grammaticalization of the have-perfect in Kashubian, if any grammaticalization indeed existed, because of the lack of sufficient evidence to follow the changes in semantic and formal respects.
4.1 Semantic features of the have-perfect Lorentz’s observation on the productivity of the have-perfect implies that the Kashubian have-perfect is far more highly grammaticalized than in other major West and South Slavic languages (except for Macedonian and some Bulgarian dialects20) that have a similar structure but defined as a resultative (henceforth, have-resultatives). The resultative expresses a state caused by a preceding change, while the perfect means an action that results in some consequences relevant to the sequential temporal plane (Maslov 1988: 64; Wiemer and Giger 2005: 1–3). Cross-linguistically, the resultative is characterized as a lower stage of grammaticalization of the perfect (Heine and Kuteva 2006: 148; Bybee 2015: 142. Lorentz and other scholars such as Breza and Treder (1981) and Stone (2002) did not expand on the semantic function of the Kashubian have-perfect (Tommola 2000: 470). Drinka (2017: 381–383) discusses this phenomenon in more detail. In fact, the Kashubian have-perfect covers the wide range of types of perfect mentioned by Comrie (1976: 56), again, which is not the case with the have-resultative in other Slavic languages. Consider (28) as an example of the perfect of result, whereas (29) is an example of the experiential perfect, and (30) is an example of the perfect of the recent past. (28) Jô móm kònie ùrzaszoné. I.NOM have.PRS.1SG horse.ACC.PL terrify.PP.SG.N ‘I have terrified horses [and the horse is terrified as a result of the action].’ (29) A tëlé razy jô cë móm gôdóné, and many times I.NOM you.DAT.SG have.PRS.1SG talk.PP.SG.N że të ùpadniesz. that you.NOM.SG tumble.down.NPST.2SG ‘I have told you so many times that you will tumble down.’ 20 An exception is Macedonian, in which the have-perfect has been grammaticalized by language contact with other Balkan languages that possess it, including Aromanian (Friedman 2014; Gołąb 1984; Koneski 1966), while the have-perfect in the Thracian Bulgarian dialect (cf. Mladenov and Kodov 1935: 104) may be a result of intensive contact with Greek.
470
Motoki Nomachi
(30) Jô móm żdóné na niegò trzë gòdzënë. I.NOM have.PRS.1SG wait.PP.SG.N on him.ACC three hour.NOM.PL ‘I have been waiting for him for three hours [and still I am waiting].’ (31) Stanisłôw Janke mô przedolmaczoné “Pana Tadeusza” Stanislaw Janke have.PRS.3SG translate.PP.SG.N Pan Tadeusz.ACC.SG.M na kaszëbsczi. onto kashubian.ACC.SG.M ‘Stanislaw Janke has translated Pan Tadeusz into Kashubian.’ In this context, it is interesting to note that example (31) is perfectly acceptable to my consultants, while not all Kashubs accept example (32): (32) Aleksander Majkòwsczi mô napisóné Aleksander Majkowski have.PRS.3SG write.PP.SG.N “Żëcé i przigòdë Remùsa.” The Life and Adventures of Remus (literally) ‘Aleksander Majkowski has written The Life and Adventures of Remus.’ Although both (31) and (32) present concrete results of actions, the difference is that in (31) the author is alive and the book was published relatively recently (2010), while in (32) the author is dead and the work was published long ago (1938). Another striking difference in semantics between the have-resultative and the have-perfect is that the former is derived in most cases from perfective verbs only, whereas the have-perfect does not know such a restriction, as can be seen in (29) and (30). This is also one of the points that proves that the Kashubian have-perfect is highly grammaticalized. It is furthermore important to note that some of the examples of the Kashubian have-perfect demonstrate that it has also spread into the domain of the past tense, as evidenced in (33) with loni ‘last year,’ just like in German (Nomachi 2008). In this context, it is worth noticing that there are consultants, particularly from Southern Kashubia, who prefer not to use adverbs that indicate the past with the have-perfect as in (34), although there are indeed examples of such usage, illustrated in (33). Those speakers prefer to replace 10 minut temù nazôd ‘10 minutes ago’ with a prepositional phrase òd 10 minut ‘for 10 minutes’ as a result of interpretation of the construction as a have-resultative.
14 Placing Kashubian on the language map of Europe
471
(33) Jô gò móm loni zabité. I.NOM he.ACC have.PRS.1SG last year kill.PP.SG.N ‘I killed him last year.’ (34) Jô ju móm wësłóné tegò e-maila I.NOM already have.PRS.1SG send.PP.SG.N that.ACC.SG.M e-mail.ACC.SG.M 10 minut temù nazôd / òd 10 minut. / from ten minute.GEN.PL ten minute.GEN.PL that.DAT.SG.N ago ‘I sent this e-mail ten minutes ago/Ten minutes have passed since I sent this e-mail.’ Yet one has also to recognize that there are restrictions on forming the haveperfect. For instance, the have-perfect cannot be derived either from modals as in (35) or from atelic verbs as in (36), which is not the case in German. These restrictions seem to be determined by both grammatical and lexical factors.21 (35) *Jô móm I.NOM have.PRS.1SG Ich habe I.NOM have.PRS.1SG ‘I had to work.’
mùszoné pracowac. must.PP.SG.N work.INF arbeiten müssen. work.INF must.PP.SG.N
(36) *Jô móm I.NOM have.PRS.1SG Ich habe I.NOM have.PRS.1SG ‘I hated the man.’
tegò that.ACC.SG.M den that.ACC.SG.M
knôpa man.ACC.SG.M Mann man.ACC.SG.M
nienawidzóné. hate.PP.SG.N gehasst. hate.PP.SG.N
4.2 Morphosyntactic features of the have-perfect According to morphosyntactic criteria outlined by Heine and Kuteva (2006: 144), the have-resultative in most West and South Slavic languages typically shows the following features:
21 According to Dieter Stern (personal communication in 2018), the restriction may have to do with the fact that the German examples are semantically very clearly past and not perfect tense, which is not the case with Kashubian whose have-perfect is a true perfect, but not a simple equivalent of the past tense.
472
Motoki Nomachi
i. Only transitive verbs are allowed as main verbs. ii. The PPP-verb still has the structure of a modifier of the patient, agreeing with the patient noun phrase in case, number and/or gender (if there are such morphological categories.) iii. Nevertheless, the possessive verb tends to be interpreted as an auxiliary and the PPP-verb as the new main verb. With regard to Kashubian, the grammaticalization of the have-perfect has reached the next stage which has the following four features (Heine and Kuteva 2006: 145): i. Instead of being transitive, the main verb may be intransitive. ii. A possessive interpretation is now ruled out. iii. Agreement in number and gender between the main verb and the object gradually disappears, that is, the PPP-verb tends to be presented in one invariable form. iv. There is no more ambiguity, that is, there is only one agent, which can no longer be interpreted as a possessor. The Kashubian have-perfect indeed shows all these features, which evidences the higher degree of grammaticalization of this perfect in morphosyntactic terms. Features (i) and (ii) can be exemplified, for instance, by (30). With regard to feature (iii), according to Lorentz (1962: 980), the indeclinable neuter singular form of the past participle without agreement in gender, number, or case is the most typical pattern in this construction; this has also been illustrated by all examples of the have-perfect above. However, one can also find a construction with grammatical agreement as in (37), taken from a work of Alojzy Bùdzysz (1874–1934), a contemporary of Lorentz.22 (37) Miôł palëcã zabëtą, a pòszedł have.APP.SG.M stick.ACC.SG.F forget.ACC.SG.F and go.APP.SG.M za nią w jizbã. for it.ACC.SG.F into house.ACC.SG.F ‘He had forgotten a stick and went to his house for it.’
22 According to Rudnicki (1913: 129), in the Slovincian dialects there are also cases in which the participle agrees with the gender of the grammatical subject. However, Rudnicki provides only masculine forms as examples, and so one cannot judge if agreement is really made between the grammatical subject and the participle.
14 Placing Kashubian on the language map of Europe
473
The structure in (37) is formally identical with the resultative, but semantically it indeed is the perfect. According to Lorentz (1962: 981), the pattern with grammatical agreement as in (37) is rather rare; nevertheless, it can be still found today. In this context, it is also important to note that Kashubian, alongside the highly grammaticalized have-perfect, also keeps the have-resultative. For instance, theoretically, example (38), taken from a work by the contemporary writer Wanda Lew-Kiedrowska (1948–), has two readings: (38) Mia ju prawie robòtã skùńczoną, czej work.ACC.SG.F done.ACC.SG.F when have.APP.SG.F already just w dwiérzach stanął Jónk. in door.LOC.PL stand.APP.SG.M Janek.NOM ‘She had just finished her work, when Janek stopped at the door.’ (perfect) ‘She had her work just finished, when Janek stopped at the door.’ (resultative) According to my consultants, a grammatically agreeing form such as (38) is more natural to interpret as a resultative. However, one can also find a kind of mismatching as follows, which is not rare in contemporary Kashubian. Example (39), from Leszek Szulc (1962–), takes a variant of the formally developed have-perfect without grammatical agreement between the participles (wcësnioné ‘pulled’ and òpińtóné ‘wrapped’) and the accusative objects (mùcã ‘hat’ and szëjã ‘neck’), but the most natural reading of (39) is the resultative, considering that the word order in these sentences is used typically to describe a stative situation of a person, not an action. This is a clear equivalent of Polish (40)23: (39) Mùcã miôł wcësnioné glãbòk na łeb, hat.ACC.SG.F have.APP.SG.M pull.PP.SG.N deeply onto forehead.ACC.SG.M a szëjã miôł òpińtóné czerwionym szalã. and neck.ACC.SG.F have.APP.SG.M wrap.PP.SG.N red.INS.SG.M scarf.INS.SG.M
23 This type of word order is particularly characteristic for Kashubian lexical items of inalienable possession: òczë miôł zelonkawé. ‘His eyes were green’ (literally ‘With regard to eyes, he had green’), rãce miôł zëmné ‘His hands were cold’ (literally ‘With regard to hands, he had cold’).
474
Motoki Nomachi
(40) Czapkę miał wciśniętą głęboko na hat.ACC.SG.F have.APP.SG.M pull.ACC.SG.F deeply onto łeb, a szyję miał forehead.ACC.SG.M and neck.ACC.SG.F have.APP.SG.M oplątaną czerwonym szalem. wrap.ACC.SG.F red.INS.SG.M scarf.INS.SG.M ‘He had pulled his hat down and wrapped his neck up with a red scarf.’ Following the developmental model of Heine and Kuteva (2006), we see that semantic reinterpretation occurs first without changing the grammatical forms themselves, as found in (37), and after that a formal unification follows – and not the other way around. Indeed, I could not find any example that would contradict this order in old dialect texts. “Contamination” such as that seen in (39), which shows the opposite situation to (38), that is, a higher degree of grammaticalization from the point of view of its form (i.e., perfect) but a lower degree of grammaticalization according to the semantic criteria (i.e., resultative), can be found particularly among younger speakers across the regions of Kashubia. This contamination could be explained by the fact that those who accept examples such as (39) already cannot always distinguish the have-resultative from the have-perfect, which may occur for three reasons. First, the semantic and formal structures of both grammatical categories are similar and sometimes identical. Second, there is currently no German influence to advance the grammaticalization of the have-perfect, which could be a driving force for differentiating the have-perfect from the have-resultative. Finally, Polish, the presentday dominant language among Kashubs, has only the have-resultative which may influence the reinterpretation of semantics of the Kashubian usage of the have-perfect. This also justifies the assessment made by Heine and Kuteva that the semantic content changes first, but this time in an opposite direction, from the perfect to the resultative. Next, feature (iv) is demonstrated by the fact that sentence (38) is understood without any context: the perfect reading always requires the grammatical subject as an agent, while the resultative reading does not require it. Thus, for the resultative reading, the grammatical subject is a person who is involved in the situation and related to the action expressed by the past participle: the grammatical subject signifies the recipient of the situation (cf. Bunčić 2015). To conclude, although the Kashubian have-perfect is more advanced than the have-resultative in terms of both its semantic and morphosyntactic criteria, one has to admit that its grammaticalization is not complete. After German lost its dominant position in the society, the grammaticalization of the form in Kashubian did not advance. At the same time, one has to notice that the limited
14 Placing Kashubian on the language map of Europe
475
sphere of the have-perfect and its contamination by the have-resultative as in (39) can be interpreted as constituting a new direction for its evolution, which seems to be caused by the influence of Polish which knows only the haveresultative.
5 Negative pronouns and lack of verbal negation: feature (h) Haspelmath (1998: 278) classifies the co-occurrence of verbal negation with negative indefinite pronouns into two major groups – V + NI (verb + negative indefinite) and NV + NI (negated verb + negative indefinite). SAE features the former type, as it is found in French, Occitan, all Germanic languages, and some other languages, while all modern Slavic standard languages (though not all historical ones – see below) belong to the latter group (Vaillant 1977: 198).24 To be more precise, these languages feature strict negative concord, as opposed to languages with non-strict negative concord such as Spanish (van der Auwera and Van Alsenoy 2016: 489). As examples (41) and (42) show, modern Kashubian is a language of strict negative concord. (41) Nicht nie przëchòdzy do mie zazdrzec. nobody.NOM not come.PRS.3SG to I.GEN see.INF ‘Nobody comes to see me.’ (42) Jô nie widzã niżodny białczi. I.NOM not see.PRS.1SG none.GEN.SG.F woman.GEN.SG.F ‘I do not see any woman.’ While the modern Slavic languages all show the aforementioned structure, older varieties of East Slavic, Czech, and Polish, as well as Old Church Slavonic, can be classified as non-strict negative concord languages (van der Auwera and Van Alsenoy 2016: 498) because in them negative concord was not obligatory if a denial word ni stood before the verb (Křížková 1968: 24; Willis 2013). It is thus disputable whether the pattern without negative concord was of
24 According to Jonke (1964: 142), violation of negative concord can happen in modern Croatian when the past passive participle is negated: Što bi značilo to nigdje Ø viđeno čudo. ‘What would it mean that the miracle has not been seen anywhere.’ Lončarić (2015: 123) believes that it was provoked by the influence of German.
476
Motoki Nomachi
Common Slavic origin; for instance, Lunt (2001: 163) is of the opinion that the double negation is “presumably normal in Slavic syntax” and the omission of the negative particle ne as exemplified by (43) is in imitation of Greek syntax.25 (43) i nikto-že Ø daěše emu. give.AOR he.DAT and nobody.NOM-PRT ‘And no one gave him [anything].’ Most older Kashubian texts do not provide any useful data in this respect, as they present examples with strict negative concord only. However, a history of the Slovincian dialects may yield clues. For example, in Symon Krofey (1568), both patterns are found, as in (44) with negative concord, and (45) and (46) without it: (44) Nicht tego nie wie zapewne. nobody that.GEN.SG.N not know.PRS.3SG certainly ‘Nobody knows that for sure.’ (45) w chtorim nam nicht in which.LOC.SG.N we.DAT nobody.NOM spomoc Ø moźe can.PRS.3SG help.INF ‘in which nobody can help us’ (46) Ánÿ Źąden czlowiek Ø národzon. born.PPP.SG.M even none.NOM.SG.M human.NOM.SG.M ‘No man (is) even born.’ However, (45) and (46) are two of only three cases without negative concord, while other 35 examples show negative concord in Krofey (1568) – thus, the variety without negative concord is fairly rare (ca. 8% of Krofey’s examples). Although this text was translated from German and the author presumably consulted Polish books during translation, one can nevertheless conclude that there was then a strong tendency to stick to negative concord. Judging from other religious texts from the 17th and 18th centuries, and dialect texts from the 19th century, this tendency never changed throughout the history of Slovincian; there
25 On the other hand, Savel’eva (1989) who has analysed the Old Russian material regards that the pattern without negative concord is of Common Slavic origin.
14 Placing Kashubian on the language map of Europe
477
are only a small number of examples without negative concord, as in (47), taken from Piotrowski (1981: 57, after Lorentz 1924: 25). (47) Žóŭden čluŏvjek sa Ø móŭg redac. can.APP.SG.M save.INF none.NOM.SG.M human.NOM.SG.M REFL ‘No man could save himself.’ Piotrowski (1981: 56) argues that example (47) is influenced by German. It may be true, as this feature is sporadically found in other Slavic dialects such as Silesian that underwent heavy German influence as evidenced in (48) (Pelka 2015: 146). (48) Łon Ø je nic wert. be.PRS.3SG nothing worth he.NOM ‘He is not worth anything.’ However, one could not deny that this is a relic of the aforementioned old construction. In any case, one can conclude that this SAE feature has not penetrated into Kashubian in general.
6 Subject person affixes as strict agreement markers: feature (k) According to Haspelmath (2001: 1500), SAE has the strict agreement feature as opposed to the reference agreement feature. The strict agreement feature means that subject affixes and overt subject pronouns obligatorily co-occur, which is not the case with the reference agreement feature. Among Slavic languages, according to Haspelmath, Russian belongs to the SAE type in this respect, though this language is not an ideal example (see Fužeron and Brejar 2004). By and large, however, West and South Slavic can be classified as reference agreement languages, though colloquial Sorbian does not belong to the reference agreement type (see Lindseth 1998) which can be clearly ascribed to German influence (Schuster-Šewc 2000); it is, in fact, rather a typical phenomenon in Slavic varieties found in close contact with German (Bayer 2006). In what follows, I use the term “non-pro-drop” instead of “strict agreement,” because in Slavic languages subject–verb agreement is not a focus but the overt use of pronouns is an important criterion to classify them typologically.
478
Motoki Nomachi
According to Cybulski and Wosiak-Roża (2001: 186), in Kashubian pronouns are usually expressed anywhere they can be. Duličenko (2005: 401) is of a similar opinion: the use of pronouns with verbs is obligatory in this language. The following fragment from Janusz Mamelski (2010: 36–37) in (49), in comparison with the Polish translation made by the author himself in (50), seems to confirm this view: (49) – Jô doch wiém, të jes òpiekùnką I.NOM but know.PRS.1SG you.NOM.SG be.PRS.2SG guardian.INS.SG.F tich wòdnych lelijów. those.G.PL water.G.PL lily.G.PL – Jô nie chcą zmieniwac miona. Wiész co, I.NOM not want.PRS.1SG change.INF name.GEN.SG know.PRS.2SG what pò smiercë jô sã zamieniã w after death.LOC.SG.F I.NOM REFL change.NPST.1SG into jeden z tich kwiatów. one.ACC.SG.M from those.GEN.PL flower.GEN.PL – Pò smiercë të bãdzesz piãkniészô after death.LOC.SG.F YOU.nom be.FUT.2SG beautiful.COMP.NOM.SG.F jak ten kwiat. A jô bãdã przë how that.NOM.SG.M flower.NOM.SG.M and I.NOM be.FUT.1SG. by cë, całi szczestlëwi. you.LOK whole.NOM.SG.M happy.NOM.SG.M – Mòże më sã zamienimë w taczé can.PRS.3SG you.NOM REFL change.NPST.1PL into such.ACC.PL kwiatë i òbsëpiemë wszëtczich flowers.ACC.PL and shower.PRS.3PL all.ACC.PL dobrich bògów? good.ACC.PL god.ACC.PL (50) – Ø Wiem, opiekunko lilii. know.prs.1sg guardian.VOC.SG.F lily.GEN.PL – Ø Nie chcę zmieniać imienia. Wiesz, not want.PRS.1SG change.INF name.GEN.SG know.PRS.2SG po śmierci Ø zamienię się w jeden after death.LOC.SG.F change.NPST.1SG REFL into one.ACC.SG.M z takich kwiatów. from such.GEN.PL flowers.GEN.PL – Po śmierci Ø będziesz piękniejsza be.FUT.2SG beautiful.COMP.NOM.SG.F after death.LOC.SG.F
14 Placing Kashubian on the language map of Europe
479
niż ten kwiat. A ja przy than that.NOM.SG.M flower.NOM.SG.M and I.nom at tobie szczęśliwy. you.LOC happy.NOM.SG.M – Ø Zamienimy się w kwiaty i change.NPST.1PL REFL into flowers.ACC.PL and obsypiemy wszystkich dobrych bogów, zgoda? shower.NPST.1PL all.ACC.PL good.ACC.PL god.ACC.PL concent.NOM.SG.F ‘– I know, you are a guardian of these water lilies. – I do not want to change my name. You know what, after death I will turn into one of these flowers. – After death you will be more beautiful than this flower. I will be with you, happy. – Let’s turn into those flowers and shower all the good gods, all right?’ According to Lorentz (1959: 651), Kashubian personal pronouns tend to become proclitics by losing distinctive stress. Indeed, other dialect texts also show the same tendency; based on the analysis of texts from the 16th to the 20th century, I showed elsewhere (Nomachi 2014) that Kashubian does have four of the five typological criteria of non-pro-drop languages advanced by Haiman (1974).26 However, in some registers of Kashubian, the use of personal pronouns is not always obligatory even among those writers who are bi- or trilingual in Kashubian, German, and/or Polish. In (51) we see an excerpt from Aleksander Labuda (1902–1981) who uses pro-drop forms particularly in narratives: (51) Ø Jidã so wedle szkòłë i Ø czëjã hear.PRS.1SG go.PRS.1SG REFL near school.GEN.SG.F and wiôldżi rozgòwôr i trzôsk. Ø Widzã see.PRS.1SG big.ACC.SG.M voice.ACC.SG.M and noise.ACC.SG.M
26 Namely: (i) allow no deletion of unstressed personal pronoun subjects, (ii) have subjects for impersonal verbs, (iii) have a special indefinite pronoun subject like French on or German man, (iv) have dummy subjects to replace extraposed sentences, and (v) have a dummy pronoun (or some equivalent) to take the place of logical subjects that have been displaced from sentence-initial position (Haiman 1974: 91). Kashubian does not have (iii) as a grammaticalized form.
480
Motoki Nomachi
jak czile szkólnëch i szkólninëch na how some male teacher.GEN.PL and female teacher.GEN.PL on miónczi wrzeszczą i pòdskakiwają. competition.ACC.PL scream.PRS.3PL and jump.PRS.3PL ‘I go across the school and I hear a loud voice and noise. I see some male and female teachers screaming and jumping on the racetrack.’ Since Labuda uses the non-pro-drop forms very often as well, especially for conversations and other colloquial language, one could conclude that two systems – non-pro-drop and pro-drop – somehow already coexisted in the 20th century. Today, some speakers/writers of the younger generation do not generally seem to have a non-pro-drop preference. Example (52) is a fragment from a text written by Grzegorz Schramke (born in 1978 in Brusy, in Southern Kashubia), in which personal pronouns are totally omitted. (52) Ju Ø nie mëszlôł. Ø Szedł. W głowie go.APP.SG.M in head.LOC.SG.F already not think.APP.SG.M Ø miôł blós zapisóné, że Ø mùszi jic. must.PRS.3SG go.INF have.APP.SG.M only impute.APP.SG.N that Ju Ø nie wiedzôł nawetka, czemù Ø mùszi, why must.PRS.3SG already not know.APP.SG.M even Ø nie przëbòcziwôł so, że wëlôzł z not remember.APP.SG.M REFL that get off.APP.SG.M from banë, że dzysô je Wilëjô. train.GEN.SG.F that today be.PRS.3SG Christmas Eve ‘He already did not think. He went on. He had just imputed in his head that he has to go. Already he even did not know why he should go. He did not remember that he got off from a train and that it was Christmas Eve.’ Of course, there might be dialectal, age, and personal stylistic differences in the use of non-pro-drop or pro-drop forms, but these differences appear to be rather recent, hardly found in texts recorded in the 19th and the first half of the 20th centuries. The tendency to pro-drop is most probably of a supra-dialectal nature, as one can see it in various (younger) writers from different regions. For instance, (53) and (54) are respectively taken from Tomasz Fopke (born in 1973 in Gdyni and lives in Wejherowo, in Central Kashubia) and from Krystyna Lewna (born in 1977 in Puck, in Northern Kashubia):
14 Placing Kashubian on the language map of Europe
481
(53) – Bez– Ø òdknypsnął jem maszinie. Zapòwiôdelë without type.APP.SG.M be.PRS.1SG typewriter.DAT.SG.F predict.APP.PL wczora galakticzny wiater w całi yesterday galactic.ACC.SG.M wind.ACC.SG.M in whole.LOC.SG.F Ùni. union.LOC.SG.F Ø Jesz bëm sã przeznobił, jak w łoni . . . still AUX REFL catch a cold.APP.SG.M how in last year –Ø Dôl jem pòlét wëstawieniô give.APP.SG.M be.PRS.1SG suggestion.ACC.SG.M issue.GEN.SG.N miks-wanodżi. traveler.DAT.SG.M – Czas ùrësznieniô: 3 minutë – Ø ùczuł jem hear.APP.SG.M be.PRS.1SG time.NOM.SG.M start.GEN.SG.N 3 minute s kòmùnykat. Ø Zazdrzã tej w stôri, dobri look.PRS.1SG then in old.ACC.SG.M good.ACC.SG.M information.ACC.SG.M Jinternét. Mòże co nowégò Ø sã dowiém . . . REFL know.PRS.1SG internet.ACC.SG.M perhaps what new.GEN.SG.N ‘– No– I was typing, making a noise. Yesterday they predicted the galactic wind in the whole Union. I might catch a cold as in the last year . . .. – I suggested that we go out traveling immediately. – The starting time: 3 minutes – I heard the information. I have a look into the old but good Internet. Perhaps I will find out something new.’ (54) Ø Pòdszed do szpinie i approach.APP.SG.M to medicine box.GEN.SG.F and wëjąl wszëtczé péle jaczé take.APP.SG.M all.ACC.PL.F tablet.ACC.PL.F which.ACC.PL.F Ø miôl doma. Ø Zdrzi i czëtô. see.PRS.3SG and read.PRS.3SG have.APP.SG.M home Ø Chcôl mù pòdac Piralginã. want.APP.SG.M he.DAT give.INF metamizole.ACC.SG.F ‘He approached his medicine box and took all the tablets that he had home. He sees and reads. He wanted to give him the metamizole.’ Although these authors also use non-pro-drop forms, particularly in dialogues and colloquial registers of Kashubian, considering the supra-dialectal nature of the phenomenon in question one could conclude that the source of this new
482
Motoki Nomachi
pro-drop tendency should be sought in the influence of Polish, a pro-drop language, which is the dominant language today in Kashubia.
7 Comitative–instrumental syncretism: feature (p) As discussed by Stolz (1996) and Heine and Kuteva (2006), cross-linguistically, comitative and instrumental marking can be classified into three types: the incoherent type (different marking of comitative and instrumental), the coherent type (same marking of comitative and instrumental), and the mixed type (two different case marking systems coexist in a given language). The majority of Western European languages are of the coherent type (Heine and Kuteva 2006: 185). The contemporary Slavic languages, however, represent all three types. Historically, Common Slavic seemingly had the coherent type, and both comitative and instrumental seem to have been marked by the bare instrumental case (Bernštejn 1958: 33).27 However, as one can see in Old Church Slavonic, the comitative is more often marked by the instrumental case with the preposition sŭ ‘with,’ while the instrumental meaning is expressed by the bare instrumental case (Staniševa 1958: 42). Among the contemporary Slavic languages, Russian, Ukrainian, Belarusian, Polish, Czech, Slovak, and Serbo-Croatian, at least in their standard varieties, are of the incoherent type. Their comitative and instrumental case marking developed in the direction of Old Church Slavonic, marking the comitative meaning with the equivalent of the preposition sŭ on the one hand and avoiding case syncretism by not marking the instrumental meaning with the equivalent of sŭ on the other. Another direction of development is toward a syncretism of comitative and instrumental by marking both with the instrumental case (with the equivalent of the preposition sŭ). This phenomenon is rather typical in those Slavic languages that have had close contact either with Romance or with German. Both varieties of Sorbian, Slovene, Kajkavian Croatian, Čakavian Croatian, Molise Slavic, and extinct Polabian are of the coherent type, which, unlike Common
27 According to Mrazek (1964: 22), it is difficult to conclude that the bare instrumental case was a regular grammatical means to express the comitativity, because the number of examples in Old Church Slavonic, older varieties of Czech and East Slavic are rather scanty.
14 Placing Kashubian on the language map of Europe
483
Slavic, always takes the equivalent of the preposition sŭ with the instrumental case (for details, see Nomachi and Heine 2011). According to Lorentz (1962: 1131), because of German influence the instrumental often takes z(e), the equivalent of the preposition sŭ, in Kashubian as in (55), in addition to the existing pattern without the preposition, as in (56): (55) Jô gò bił z cziją. I.NOM he.ACC hit.APP.SG.M with stick.INS.SG.F ‘I hit him with a stick.’ (56) Òn sã przeżegnôł Ø tim that.INS.SG.M he.NOM REFL farewell.APP.SG.M swiãtim krziżã. holy.INS.SG.M cross.INS.SG.M ‘He bid his farewell with the holy cross.’ In Lorentz’s time, the syncretism was not complete, and Kashubian then showed the “mixed” feature. According to Stone (2002: 768), “there is a strong tendency for the instrumental to acquire the preposition z (s)/ze (se) ‘with,’ when used with its basic function as an expression of instrument”; however, one can hardly agree with this statement, as the strong tendency that allegedly existed in the times of Lorentz does not seem so evident. Even Breza and Treder (1981: 181), who did their field work in the 1970s, claimed that the prepositional instrumental had become rarer since Lorentz, and Jerzy Treder mentioned to me that the construction is now very rare.28 Today, among my younger consultants from various regions of Kashubia, not a single one uses the prepositional way of marking the instrumental as in (55), but also, none accepts it as a proper form for the instrumental in Kashubian. Thus, it seems that comitative and instrumental marking has changed from the mixed type to the incoherent one. Since this is most probably a supra-dialectal feature, one can assume that the completion of syncretism of comitative and instrumental following the German model has been hindered by the presence of dual factors. On the one hand, there has been the complete loss of German influence after World War II. On the other hand, judging from the present patterning of comitative and instrumental that is identical with that of Polish, the influence of the Polish language has been significant.
28 In my personal communication made in 2015.
484
Motoki Nomachi
8 Conclusion The foregoing analysis allows me to conclude that the Kashubian language became typologically closer to the SAE languages particularly between the 19th and the first half of the 20th century, acquiring non-Slavic features as a result of heavy contact with German. However, the degree of penetration of each feature varied significantly: the have-perfect was relatively highly grammaticalized and rather stable in the grammatical system of Kashubian, while the negative concord was very marginal in Slovincian, and there was no visible trace of it in other Kashubian dialects. In addition, the non-pro-drop feature seems to have been grammaticalized and degrammaticalized rather quickly, as does the syncretism of instrumental and comitative. One can also note that none of the five features analyzed in this chapter became fully established in the system of Kashubian. When German influence stopped completely after World War II, Polish influence became the strongest force of change in the linguistic structure of Kashubian. This is why all non-Slavic (German) features in Kashubian that were not shared with Polish tended to disappear or be replaced with equivalent Polish features. Consequently, all this moved the position of Kashubian on the language map of Europe away from the core of SAE toward its periphery, like the other Slavic languages, based on the criteria provided by Haspelmath. This shows the degree of shift that can occur in this regard even in less than one century in Kashubia. Acknowledgments: I am grateful for the very important comments and suggestions on the earlier version of this article provided by Bernd Heine, Johan van der Auwera, Sarah Thomason, Yaroslav Gorbachov, Dieter Stern, Gerhard Neweklowsky, Evangelia Adamou, and Brian Joseph. In addition, I would like to express my sincere gratitude to all my consultants of the Kashubian language.
Abbreviations ACC AOR APP AUX COMP DAT F FUT GEN IMPER
accusative aorist active past participle auxiliary comparative dative feminine future genitive imperative
14 Placing Kashubian on the language map of Europe
INF INS LOC M N NOM NPST PL PP PPP PRS PRT PST Q REFL SAE SUBJ VOC
485
infinitive instrumental locative masculine neuter nominative non-past tense plural past participle past passive participle present tense particle past tense question particle reflexive Standard Average European subjunctive vocative
References Bayer, Markus. 2006. Sprachkontakt deutsch-slavisch. Eine kontrastive Interferenzstudie am Beispiel des Ober- und Niedersorbischen, Kärntnerslovenischen und Burgenlandkroatischen. Frankfurt am Main: Peter Lang. Benacchio, Rosanna. 2014. Grammatikalizacija v situacijax jazykovogo kontakta: razvitie artiklja v rez’janskom dialekte [Grammaticalisation in situations of language contact: development of the article in the Resian dialect], In: Motoki Nomachi, Andrii Danylenko, and Predrag Piper (eds.), Grammaticalization and Lexicalization in the Slavic Languages, 205–217. Munich: Otto Sagner. Berger, Tilman. 1993. Das System der tschechischen Demonstrativpronomina (Textgrammatische und stilspezifische Gebrauchsbedingungen). Munich: LMU (unpublished postdoctoral thesis). Bernštejn, Samuil Borisovič. 1958. Metody i zadači izučenija istorii značenij i funkcij padežej v slavjanskix jazykax [Methods and problems of studying the history of meanings and functions of cases in Slavic languages]. In: Samuil Borisovič Bernštejn (ed.), Tvoritel’nyj padež v slavjanskix jazykax, 5–40. Moscow: Akademija nauk SSSR. Boryś, Wiesław and Hanna Popowska-Taborska. 1999. Słownik etymologiczny kaszubszczyzny [Etymological Dictionary of Kashubian]. Vol. 3. Warsaw: Polska academia nauk/Instytut slawistyki. Breu, Walter. 2012. The grammaticalization of an indefinite article in Slavic micro-languages. In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 275–322. (Trends in Linguistics 242.) Berlin/ Boston: Mouton de Gruyter. Breza, Edward and Jerzy Treder. 1981. Gramatyka kaszubska [Kashubian Grammar]. Gdańsk: Zrzeszenie Kaszubsko-Pomorskie.
486
Motoki Nomachi
Bunčić, Daniel. 2015. “To mamy wpajane od dziecka” – a recipient passive in Polish? Zeitschrift für Slawistik 60 (3): 411–431. Bybee, Joan, Revere Perkins and William Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago/London: The University of Chicago Press. Comrie, Bernard. 1976. Aspect: An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: Cambridge University Press. Cybulski, Marek and Róża Woisiak-Śliwa. 2001. Składnia kaszubska [Kashubian Syntax]. In: Edward Breza (ed.), Kaszubszczyzna, 185–201. Opole: Uniwersytet Opolski. Dahl, Östen and Maria Koptevskaja-Tamm (eds.). 2001. Circum-Baltic Languages: Typology and Contact, Vol. 1: Past and Present; Vol. 2: Grammar and Typology (Studies in Language and Companion Series 54, 55.) Amsterdam/Philadelphia: John Benjamins. Décsy, Gyula. 2000. The Linguistic Identity of Europe: The 62 Languages of Europe Classified in Functional Zones. Bloomington, IN: Eurolingua. Diessel, Holger. 1999. Demonstratives: Form, Function, and Grammaticalization. (Typological Studies in Language 42.) Amsterdam/Philadelphia: John Benjamins. Drinka, Bridget. 2017. Language Contact in Europe: The Periphrastic Perfect through History. Cambridge: Cambridge University Press. Duličenko, Aleksandr Dmitrievič. 2005. Kašubskij jazyk [The Kashubian language]. In: Aleksandr Mixajlovič Moldovan et al. (eds.), Jazyki mira: slavjanskie jazyki, 410–433. Moscow: Academia. Friedman, Victor. 2014. Grammatical Categories of the Macedonian Indicative. (Second edition.) Bloomington, IN: Slavica Publishers. Fužeron, Irina and Žan Brejar. 2004. Mestoimenie “ja” i postroenie diskursivnyx svjazej v sovremennom russkom jazyke [The pronoun “I” and the construction of discourse connectors in contemporary Russian]. In: Tat’jana Mixajlovna Nikolaeva (ed.), Verbal’naja i neverbal’naja opory prostranstva mežfrazovyx svjazej. Moscow: Jazyki slavjanskoj kul’tury. Giger, Markus. 2003. Resultativa im modernen Tschechischen. Unter Berücksichtigung der Sprachgeschichte und der übrigen slavischen Sprachen. Bern: Peter Lang. Givón, Talmy. 1981. On the development of the numeral ‘one’ as an indefinite marker. Folia Linguistica Historica 2 (1): 35–53. Gołąb, Zbigniew. 1984. The Arumanian Dialect of Kruševo in SR Macedonia SFR Yugoslavia. Skopje: Macedonian Academy of Science and Arts. Hawkins, John. 1978. Definitness and Indefiniteness: A Study in Reference and Grammaticality Prediction. London/New York: Routledge. Haspelmath, Martin. 1998. How young is Standard Average European? Language Sciences 20 (3): 271–287. Haspelmath, Martin. 2001. The European linguistic area: Standard Average European. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, 1492–1510. Berlin/New York: Walter de Gruyter. Heine, Bernd and Tania Kuteva. 2006. The Changing Languages of Europe. Oxford/New York: Oxford University Press. Heine, Bernd. 1997. Cognitive Foundations of Grammar. Oxford/New York: Oxford University Press.
14 Placing Kashubian on the language map of Europe
487
Heine, Bernd. 2003. Gramaticalization. In: Brian D. Joseph and Richard D. Janda (eds.), The Handbook of Historical Linguistics. Malden, MA: Blackwell. Heine, Bernd. 2012. On polysemy copying and grammaticalization in language contact. In: Claudia Chmoreau and Isabelle Léglise (eds.), Dynamics of Contact-Induced Language Change, 125–166. Berlin/Boston: Mouton de Gruyter. Hilferding, Aleksander. 1965. Ostatki słowian na południowym brzegu bałtyckiego morza [The Remnants of Slavs in the Southern Seashore of the Baltic Sea]. Poznań: Polskie Wydawnictwo Muzyczne/Ludowa Spółdzielnia Wydawnicza. Himmelmann, Nikolaus. 2001. Articles. In: Martin Haspelmath, Ekkehard König, Wulf Oesterreicher, and Wolfgang Raible (eds.), Language Typology and Language Universals: An International Handbook, 831–842. Berlin/New York: Mouton de Gruyter. Hinze, Friedhelm. 1967. Altkaschubisches Gesangbuch. Berlin: Akademie. Jakobson, Roman. 1931. K xarakteristike evrazijskogo jazykovogo sojuza [On the Characteristics of the Eurasian Linguistic Area]. Paris: Société anonyme imprimerie de Navarre. Jonke, Ljudevit. 1964. Književni jezik u teoriji i praksi [Literary Language in Theory and Practice]. Zagreb: Nakladni zavod Znanje. Knoll, Vladislav. 2012. Kašubština v jazykověm kontaktu [Kashubian in Language Contact]. Prague: Filozofická Fakulta Univerzity Karlovy. Křížková, Helena. 1968. K voprosu o tak nazyvaemoj dvojnoj negacii v slavjanskix jazykax [Toward the so-called double negation in Slavic languages]. Slavia 37 (1): 21–39. Koneski, Blaže. 1966. Istorija makedonskog jеzika [A History of the Macedonian Language]. Belgrade: Prosveta. Krámský, Jiří. 1972. The Article and the Concept of Definiteness in Language. The Hague: Mouton. Lehr-Spławiński, Tadeusz. 1929. Gramatyka połabska [Polabian Grammar]. Lviv: Nakład i Własność K. S. Jakubowskiego. Lindseth, Martina. 1998. Null-Subject Properties of Slavonic Language: With Special Reference to Russian, Czech and Sorbian. (Slavistische Beiträge 361.) Munich: Otto Sagner. Lončarić, Mijo. 2015. Gramatički utjecaj njemačkoga na hrvatski jezik [The grammatical influence of German on the Croatian language]. In: Sonja Wölke and Hauke Bartels (eds.), Einflüsse des Deutschen auf die grammatische Struktur slawischer Sprachen, 121–135. (Schriften des Sorbischen Instituts 62.) Bauzen: Domowina. Lorentz, Friedrich. 1903. Slovinzische Grammatik. Sankt-Peterburg: Imperatorskaja akademija nauk. Lorentz, Friedrich. 1905. Slovinzische Texte. Sankt-Peterburg: Imperatorskaja akademija nauk. Lorentz, Friedrich. 1924. Teksty pomorskie (kaszubskie) [Pomeranian (Kashubian) Texts]. Cracow: Polska Akademia Umiejętności. Lorentz, Friedrich. 1958. Pomoranisches Wörterbuch I: A-P. Berlin: Akademie. Lorentz, Fryderyk. 1958–1962. Gramatyka pomorska I-III [Pomeranian Grammar] (Second edition.). Wrocław: Zakład Narodowy im. Ossolińskich. Lunt, Horace. 2001. Old Church Slavonic Grammar. (Seventh edition.) Berlin/New York: Mouton de Gruyter. Maslov, Jurij Sergeevič. 1988. Resultative, perfect and aspect. In: Vladimir Nedjalkov (ed.), Typology of Resultative Constructions. Amsterdam/Philadelphia: John Benjamins.
488
Motoki Nomachi
Mladenov, Stefan, Hristov Kodov, and Hristov Vakarelski. 1935. Bitǎ i ezikǎ na trakijskite i maloazijskite bǎlgari [The Life and Language of the Thrakian and Anatolian Bulgarians]. Sofia: Trakijskijat naučen institut. Mrazek, Roman. 1964. Sintaksis russkogo tvoritel’nogo [Syntax of the Russian Instrumental Case]. Prague: Státní pedagogické nakladadelství v Praze. Nau, Nicole. 1995. Möglichkeiten und Mechanismen kontaktbewegten Sprachwandels unter besonderer Berücksichtigung des Finnischen. Munich: LINCOM Europa. Neweklowsky, Gerhard. 1978. Die kroatischen Dialekte des Burgenlandes und der angrenzenden Gebiete. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Nomachi, Motoki. 2006. Polska konstrukcja rezultatywno-posesywna mam to zrobione a kaszubskie jô móm to zrobioné [The Polish resultative-possessive construction mam to zrobione and Kashubian jô móm to zrobioné]. Język polski 86 (3): 173–183. Nomachi, Motoki. 2008. On a periphrastic perfect in the Kashubian literary language. Slavia Occidentalis Iaponica 11: 4–23. Nomachi, Motoki. 2014. On the Kashubian past tense form jô béł ‘I was’ from a languagecontact perspective. In: Motoki Nomachi, Andrii Danylenko, and Predrag Piper (eds.), Grammaticalization and Lexicalization in the Slavic Languages, 218–242. Munich: Otto Sagner. Nomachi, Motoki. 2015. Possessive constructions in the South Slavic languages: Some implications for areal typology. Balkanistica 28: 337–360. Nomachi, Motoki. 2016. On the second be periphrasis (BE-2) in Kashubian: Its grammatical status and historical development. Slavia 84 (3): 268–283. Nomachi, Motoki and Bernd Heine. 2011. On predicting contact-induced grammatical change. Evidence from Slavic languages. Journal of Historical Linguistics 1 (1) 48–76. Pelka, Daniela. 2015. Deutsche Einflüsse auf die Grammatik des polnischen Schlesisch. In: Sonja Wölke and Hauge Bartels (eds.), Einflüsse des Deutschen auf die grammatische Struktur slawischer Sprachen, 136–156. (Schriften des Sorbischen Instituts 62.) Bauzen: Domowina. Piotrowski, Jan. 1981. Składnia słowińska wobec wpływów języka niemieckiego [Slovincian Syntax in the Context of Influence of the German Language]. Wrocław: Zakład Narodowy im. Ossolińskich. Popowska-Taborska, Hanna. 1980. Kaszubszczyzna: Zarys dziejów [Kashubian: An Outline of History]. Warsaw: Państwowe Wydawnictwo Naukowe. Rudnicki, Mikołaj. 1913. Przyczynki do gramatyki i słownika narzecza słowińskiego [Contributions to the Grammar and Lexicon of the Slovincian Dialect]. Cracow: Polska Akademia Umiejętności. Savel’eva, Lidia Vladimirovna. 1989. Formy otricanija v russkom jazyke donacional’nogo perioda [The Forms of Negation in the Russian Language in the Pre-National Period]. Leningrad: Izdatel’stvo Leningradskogo universiteta. Scholze, Lenka. 2012. On the grammaticalization of the definite article in Colloquial Upper Sorbian (CUS). In: Björn Wiemer, Bernhard Wälchli, and Björn Hansen (eds.), Grammatical Replication and Borrowability in Language Contact, 323–353. (Trends in Linguistics 242.) Berlin: Mouton de Gruyter. Schroeder, Christoph. 2006. Articles and article Systems in some areas of Europe. In: Giuliano Bernini and Marcia L. Schwartz, Pragmatic Organization of Discourse in the Languages of
14 Placing Kashubian on the language map of Europe
489
Europe, Vol. 8, 545–611 (Empirical Approaches to Language Typology). Berlin: Mouton de Gruyter. Schuster-Šewc, Heinz. 2000. Das Sorbische im slawischen Kontext: Ausgewählte Studien. Bauzen: Domowina. Staniševa, Dina Sergeevna. 1958. Tvoritel’nyj sociativnyj [Sociative instrumental]. In: Samuil Borisovič Bernštejn (ed.), Tvoritel’nyj padež v slavjanskix jazykax, 41–75. Moscow: Akademija nauk SSSR. Stieber, Zdzisław and Hanna Popowska-Taborska (eds.). 1964–1978. Atlas językowy kaszubszczyzny i dialektów sąsiednich [Linguistic Atlas of Kashubian and Neighboring Dialects]. Wrocław: Zakład Narodowy im. Ossolińskich. Stolz, Thomas. 1996. Some instruments are really good companion–some are not: On syncretism and the typology of instrumentals and comitatives. Theoretical Linguistics 23 (1/2): 113–200. Stolz, Thomas. 2009. Europe as a linguistic area. In: Keith Brown and Sarah Ogilvie (eds.), Concise Encyclopedia of Languages of the World, 388–405. Amsterdam: Elsevier. Stone, Gerald. 2002. Cassubian. In: Bernard Comrie and Greville G. Corbett (eds.), The Slavonic Languages, 759–794. London/New York: Routledge. Sychta, Bernard. 1968. Słownik gwar kaszubskich na tle kultury ludowej, Vol. 2 [Dictionary of the Kashubian dialects on the background of the folk culture]. Wrocław: Zakład narodowy im. Ossolińskich. Tambor, Jolanta. 2006. Mowa górnoślązaków oraz ich świadomość językowa i etniczna [The Language of the Upper Silesians and Their Linguistic and Ethnic Identity]. Katowice: Wydawnictwo Uniwersytetu Śląskiego. Tetzner, Franz. 1899. Die Slowinzen und Lebakaschuben. Land und Leute, Haus und Hof, Sitten und Gebräuche, Sprache und Literatur im östlichen Hinterpommern. Berlin: Emil Felber. Tommola, Hannu. 2000. On the perfect in North Slavic. In: Östen Dahl (ed.), Tense and Aspect in the Languages of Europe. Berlin/New York: Mouton de Gruyter. Treder, Jerzy. 2005. Historia kaszubszczyzny literackiej [A History of the Kashubian Literary Language]. Gdańsk: Wydawnictwo Uniwersytetu Gdańskiego. Trovesi, Andrea. 2004. La genesi di articoli determinative. Modalità di espressione della definitezza in ceco, serbo-lusaziano e sloveno. Milan: Franco Angeli. Vaillant, André. 1977. Grammaire Comparée des langues slaves, Vol. 5: La syntaxe, Paris: Klincksieck. van der Auwera, Johan. 2011. Standard Average European. In: Bernd Kortmann and Johan van der Auwera (eds.), The Languages and Linguistics of Europe: A Comprehensive Guide. Berlin/Boston: Mouton de Gruyter. van der Auwera, Johan and Lauren Van Alsenoy. 2016. On the typology of negative concord. Studies in Language 40 (3): 473–512. Wackernagel, Jacob. 2009. Lectures on Syntax: With Special Reference to Greek, Latin, and Germanic. Oxford: Oxford University Press. Wiemer, Björn and Markus Giger. 2005. Resultativa in den nordslavischen und baltischen Sprachen. Bestandsaufnahme unter arealen und grammatikalisierungstheoretischen Gesichtspunkten. Munich: LINCOM Europa. Willis, David. 2013. Negation in the history of the Slavonic languages. In: David Willis, Christopher Lucas, and Anner Breitbarth (eds.), The History of Negation in the Languages of Europe and the Mediterranean, 341–398. Oxford: Oxford University Press.
Index of subjects actant 267, 268 adjective, definite 122 adstrate 262, 320 agglutinative language 266 agreement marker 113, 136, 137, 145, 306, 454, 477–481 Aktionsart 159, 170–172, 175 analyticity 284, 351, 368, 373, 374, 414 analytic language 267, 282 aorist 11, 32, 34, 152, 153–155, 159–163, 169–172 aorist, complexive 170–172, 179 areal cline 23, 24, 35, 36, 53 areal linguistics 4, 8, 10, 54, 63–80, 187–189, 263, 264, 284, 319, 323, 324 areal-typological profiling 2, 5, 11, 187–214 article, definite 456–463 article, indefinite 463–467 article, zero 412, 413, 443–445 aspect 396–400 aspect, viewpoint 11, 160, 173–179 auxiliary 11, 13, 135, 154, 157–160, 166–168, 264, 274–276, 278, 280, 282, 283, 296, 298–300, 302, 304, 305, 307, 325, 327, 355, 366, 389, 392, 393, 395, 400–402, 418, 419, 421, 428, 468, 472 auxiliation 149, 156 Balkanism 6, 13, 75, 127, 263, 295, 332, 347–375 Balkanism, Carpathian 13, 347–375 Balkanization 336, 350, 351, 366, 372 Balkan linguistics 2, 13, 75, 79, 264, 317–320, 324, 335, 339, 347–350, 352, 353, 356–358, 372 Balkan Sprachbund 2, 7, 8, 12, 35, 95, 128, 187, 212, 262, 264, 283, 295, 296, 317, 319, 320, 324, 339, 347, 348, 350–354, 356–359, 361, 371, 423 ‘be’-language 202, 203 ‘be’-perfect 44, 149, 154, 156, 157, 158, 300, 301, 467, 468 bilingualism 128, 228, 229, 252, 262, 297, 318, 385, 435 https://doi.org/10.1515/9783110639223-016
borrowability 337, 351, 373 borrowing 2–4, 13, 14, 74, 75, 100, 103, 104, 132, 197, 249, 265, 315, 317, 330, 331, 333, 334, 336–339, 348, 349, 354, 356, 357, 359, 361, 365, 366, 371, 373–375, 386, 388, 399, 410 calquing 13, 14, 39, 163, 315, 318, 333, 335, 385, 394, 400, 417, 420 Carpathian-Balkan macroarea 7, 8, 348, 349, 357, 363, 372–375 Carpathianism 13, 347–375 Carpathian linguistic area 356, 359, 367 Carpathian linguistics 347, 348, 353, 356–359, 362, 367, 372, 375 case –accusative 266, 269, 328, 369, 427 –comitative 14, 306, 407, 482, 483 –dative 266, 363 –genitive 266, 328, 405 –instrumental 351, 408, 482 –locative 46, 136, 332, 403, 406, 427 –vocative 403, 406 case syncretism 482 Central European linguistic (convergence) area 79, 353 Charlemagne Sprachbund 4, 145 Circum-Baltic area 5, 39, 453 clitic 296–299, 363, 364, 371, 386, 389, 417–420, 423–427 comitative-instrumental syncretism 14, 306, 454, 482–483 comparative 4, 66, 68–70, 79, 113, 136, 320, 327, 352, 354–356, 360, 364, 365, 414, 415, 445, 454 contact-induced grammatical change 337, 433 contact-induced grammaticalization 2, 3, 6–8, 13, 347, 352, 371, 372, 433, 438, 439 convergence 4, 5, 7, 8, 10, 11, 13, 24, 30, 35, 37, 45, 54, 55, 64, 69–71, 78–80, 96,
492
Index of subjects
203, 315, 324, 347–349, 357, 358, 361, 363, 371, 372, 374 convergence area 4, 12, 261–282 Danube Basin 12, 74, 78, 291–304 dative external possessor 113, 136, 145, 454 de-andative future 37 debalkanization 263, 284 decategorialization 366 declension 26, 36, 374, 405, 406, 409–411 definiteness 13, 115–117, 122, 123, 128, 139, 295, 325, 337, 338, 411–414, 426, 434, 435, 456, 466 demonstrative pronoun 115, 117, 118, 120, 122, 128, 139, 293, 424, 456–459, 461–463 diachronic constant 386, 387, 389–415 diagrammar 385, 403, 417, 426 donor language 104, 331, 334, 388, 456 double negation 420, 421, 426, 427, 476 doubling of pronouns 262, 283 dummy subject 269, 479 enclitic 32, 164, 338, 351, 354, 358, 363, 364, 369 erosion 366, 457 Eurolinguistics 1 EUROTYP 1, 145, 187, 453 evidential 36, 41, 42, 333 external determinant 8, 373, 374 family tree 68, 94–99 formula of equivalence 442–445 future –de-obligative 390, 401, 418 –de-venitive 37 –de-volitive 28, 352, 366, 367 –perfective 279, 281, 298, 302 –periphrastic 275–279, 281–283, 293, 296–298 gender 13, 36, 49, 147, 294, 299, 331, 405, 406, 409–411, 472 Germanoidism 225, 227, 335 gram 28, 33, 34, 37, 41
grammaticalization 2, 3, 6–9, 34, 39, 40, 132, 139, 179, 277, 283, 294, 347, 349, 364, 365, 367, 370–375, 419, 428, 433–446 grammaticalization area 352 ‘have’-language 203 ‘have’-perfect 113, 139, 264, 428, 454, 467–475 Hebraism 234, 237, 243, 252, 253 Hebraization 237 Hebroid 231–233, 235–237, 239, 243, 249, 250 Hebroidism 247, 248 Hebroidization 238 high-contact language 6, 435 historical linguistics 1, 64, 79 imperfect 34, 153, 157, 164, 174–179, 274, 275, 278, 298–300, 302–303, 374, 386, 389, 392–403, 417, 427 imperfective 24, 26, 33, 34, 48, 159, 173–177, 275–282, 296–298, 302, 397–399, 401, 402 imperfective future 298, 302 impersonal 43, 267–269, 371 inceptive 275, 276, 278, 280, 283 inchoative 28, 278 infinitival complementation 270 infinitive 269, 270, 276–278, 280, 283, 296–298, 305, 333, 338–339, 351, 389, 394, 395, 418, 419 inflectional type 396 ingressive 275–281 interference 265, 276, 336, 337, 373 internal determinant 373–375 intransitive 50, 90, 113, 133, 159, 263, 276–278, 283, 371, 468, 472 Iranianism 228, 230, 238, 243, 245–247, 249 Isomorphism 387, 390, 403, 441, 442 Judaism 12, 226, 227, 230, 231, 239 Koine 155–157
Index of subjects
language contact 2, 3, 6–10, 13, 14, 63–80, 98, 113, 197, 209, 212, 265, 283, 318, 373, 385–425, 434, 440, 441, 445, 456 language contact studies 63, 65, 80 language isolation 8 linguistic area 1–9 loan translation 14, 385, 394, 400, 406, 417, 427 low-contact language 6 l-participle 11, 42, 158–160, 171, 175, 178, 277, 296, 297, 299, 302, 389, 392, 393, 395, 428, 468 l-perfect 11, 159–160, 166, 167, 169–171, 173, 174–179 matter borrowing 13, 348, 349, 359, 361, 365, 371, 386 micro-language 3, 13, 385–425 model language 6, 371, 390, 396, 406–408, 412, 417–419, 423, 427, 436, 438, 441, 442 mood 400–403 Moravianism 151 morphosyntactic agreement 284 multilingualism 78, 79, 251, 318, 372, 454 negative concord 14, 475–477 nominal inflection 26, 34, 266, 267, 406 nominative subject 40, 41, 47, 192, 267 non-pro-drop language 479 non-standard variety 6, 204 number –dual 25, 240 –pseudo dual 251, 252 numeral 76, 124, 240, 251, 294, 295, 332, 351, 352, 355, 360, 366, 411, 412,434, 435, 437, 439, 448, 463–466 object doubling 117, 127, 332, 408, 423–425 palatal 212, 213 palatalization 27, 76, 77, 103, 104, 204, 206, 208–210, 212, 213, 361, 362
493
palatalized 76, 208, 213, 362 parataxis 368, 371 passive –participial 113, 135, 136, 138, 145, 454 –periphrastic 263, 264, 273–274, 282, 284 pattern replication 3, 365, 373 perfect 169–172 –double 303, 304, 393 –past/preterit 11, 165, 167, 171, 175, 179, 264, 273, 275, 298–301, 305, 363, 389, 392–394, 401, 402 perfective 24, 26, 32, 33, 157, 159, 169, 170, 173, 174–177, 179, 274, 275, 278–281, 297, 298, 302, 338, 397–399, 470 Persianism 239, 247 phonetic protensity 362 pluperfect 134, 153–155, 163, 165, 169, 274, 353, 393 pluperfect, supercompound 293, 302–304 polisemization, see polysemy copying polysemy 14, 47–53, 306, 390, 397, 401, 403–406, 412, 417, 427, 441 polysemy copying 14, 438, 441 possessive construction 91, 133, 196, 203, 269, 334 predicate 135, 192, 199, 267, 268, 274, 409, 445 prefixation 88, 169, 174–176, 267, 397, 399 preterit 11, 149, 165, 167–169, 171, 175, 176, 178, 179, 272–275, 278, 282, 293, 298–302, 304–306, 353, 359, 468 preterit decay 300, 306 preverbation 271–273, 282 proclitic 338, 418, 420, 479 pro-drop language 263, 479, 481 puristic intervention 306 recipient language 73, 330, 331, 417 reciprocal 47–53 reflexive 47–53, 77, 136, 138, 243, 263, 264, 268, 274, 282, 468 relative clause 113, 114, 128, 129, 138, 145, 367, 368, 425, 454 relativization strategy 368 relativizer 128, 129, 270, 354, 367–369
494
Index of subjects
relexification 225, 226, 234–236, 244, 252, 253 replica language 4, 373, 385, 386, 390, 404, 406, 418, 423, 436, 438, 441, 442, 444–447 replication 2, 3, 5, 8, 13, 154, 179, 349, 352, 373–375, 441, 445 resultative 5, 39–45, 132–135, 139, 158, 170, 171, 274, 275, 283, 428, 455, 467, 469–471, 473–475 resumption 98, 128, 367–369 resumptive pronoun 128, 129, 367, 369 roofing 11, 146–148, 150, 178, 179 secondary articulation 188, 189, 203–214 Slavic type 22, 24, 25, 321, 324, 337, 395, 421 societal factor 7, 8, 13, 349, 371, 373, 375 sociolinguistic typology 8, 13, 349, 373 standard variety 6, 204, 414, 422 stem-inflection 266, 282 stratification 22, 165–169, 179, 264 stratified convergence zone 5, 24 stratum 167, 171 substrate 262, 276, 320, 406 superlative 117, 119, 136, 355, 356, 364, 365, 414, 415
syncretism 14, 34, 46, 172, 179, 306, 369, 454, 482–483 syntactic complexity 371 synthetic inflection 266 systemic typology 8, 349, 372 three-tense system 293, 302, 305 transitive 51, 89, 90, 92, 130, 131, 134, 154, 276, 283, 371, 468, 472 typological parallel 88, 89, 369 typology 4, 8, 11, 13, 21, 35, 38, 63, 64, 80, 95, 189, 203, 225, 228, 232, 253, 254, 291, 292, 297, 305, 319, 349, 372, 373, 375, 387, 433, 453–455 verb modifier 271 vocal harmony 266 voluntative 277 wh-adverbial marker 368, 369 word-inflection 266 word order 89, 115, 117, 123, 127, 137, 139, 269, 386, 416–418, 420, 427, 466, 473
Index of languages Albanian 4, 6, 13, 87, 137, 145, 187, 212, 304, 325, 327–334, 337, 350, 355, 358, 366 Arabic 89, 229–230, 233, 235–236, 238–240, 243–250, 253 Aramaic 234, 235, 244, 288 Aramaic (Modern East) 234 Armenian 146, 204, 209, 215 Aromanian 6, 325–326, 330–335, 350 Avestan 254 Balkan Slavic 315–339, 406, 408, 414, 420, 424 Baltic 3, 5, 22, 39–43, 46, 49, 90, 97, 98, 103, 104, 146, 187, 188, 207, 267, 293, 301, 306, 350, 353 Belarusian 3, 46, 94, 97–99, 128, 194–197, 199–200, 202, 203, 213, 240, 267, 296, 297, 299, 302, 306, 358, 439, 482 Bosnian-Croatian-Serbian 388–390, 393, 411, 414, 416, 423 Bulgarian 13, 74, 97, 116, 117, 124, 127, 133, 150–152, 169, 208, 211, 212, 228, 266, 295, 316, 323, 325, 330, 333, 335, 338, 350, 354, 355, 358, 359, 361–363, 365, 368, 369, 371, 392, 406, 423 Bulgarian (Rodopian) 326–329, 339, 363 Bulgarian (South Thracian) 134 Bulgarian (Standard) 32 Celtic 10, 11, 65, 66, 73, 87, 89–94, 96, 100, 101, 105, 113, 129, 130, 138, 207, 208, 242, 243, 306 Celtic (Hispano-Celtic) 129 Celtic (Proto-Celtic) 129 Church Slavonic 11, 26, 91, 104, 130, 131, 145–182, 276, 277, 296, 297, 302, 303, 390, 475, 482 Chuvash 242, 245 Croatian 28, 31, 34, 45, 46, 73, 102, 120, 122, 123, 128, 136, 137, 197, 200, 211, 214, 246, 252, 265, 292, 294, 295, 296
https://doi.org/10.1515/9783110639223-017
Croatian (Burgenland) 292, 296, 297, 301, 303, 306, 385, 399, 414, 435, 456, 467 Croatian (Čakavian) 28, 298, 322, 435, 482 Croatian (Kajkavian) 28, 101, 270, 292, 294, 295, 296, 298, 299, 301, 303, 306, 310, 353, 335, 482 Croatian (Molise). See Molise Slavic Croatian (Standard) 297, 389 Czech 12, 27, 91, 92, 94, 98, 117, 118, 119, 120, 128, 187, 197, 203, 211, 226, 229, 232, 236, 245, 251, 261, 265–270, 273–275, 277–280, 282, 292–303, 305–307, 428, 455, 456, 475, 482 Czech (Middle) 182 Czech (Standard) 27, 293, 294 Danish 94, 96, 276 Dutch 4, 90, 98, 145, 147, 236, 270–272, 276, 367 English 37, 38, 48, 49, 92, 93, 96, 97, 98, 100, 103, 123, 136, 146, 147, 191, 263, 266–271, 276, 277, 300, 351, 367, 436 Esperanto 236 Estonian 42, 56, 101, 227, 271, 298 Faroese 10, 92, 96, 276 Finnish 53, 125, 137, 192, 207, 215, 271, 298 Finno-Ugric 5, 6, 12, 42, 54, 94, 100, 227, 261, 267, 271, 284, 293, 301 French 2, 4, 40, 43, 48, 49, 53, 90, 93, 97, 98, 103, 104, 128, 137, 145, 147, 149, 226, 230, 233, 234, 263, 266, 267, 269, 275, 300, 301, 304, 428, 433, 455, 475 German 2–6, 11–12, 14, 37, 38, 43, 48–49, 51, 52, 90, 92, 93, 97, 115, 117, 119, 120, 123, 132, 138, 145, 147, 149, 150, 203, 225–254, 261, 262–263, 265–285, 291,
496
Index of languages
294, 297, 300, 302, 304–306, 367, 399, 407, 411, 413, 414, 416–417, 421, 428, 433, 435–436, 454–456, 460, 463, 465, 467, 468, 470–471, 474–477, 479, 482–484 German (High) 6, 92, 226, 234, 236, 239, 251, 454 German (Low) 5, 276, 454 German (Old High) 92 Germanic 3, 7, 10, 12, 21, 36, 37, 48, 54, 65, 73, 87–90, 98, 100, 103, 116, 119, 130, 137, 139, 147, 149, 159, 187, 188, 207, 225, 226, 232, 251, 261, 266, 270, 271, 274, 276, 291, 293, 306, 350, 367, 475 Greek 4, 6, 11, 13, 35, 74, 79, 87, 98, 99, 102, 116, 131, 132, 134, 135, 145–177, 233, 235, 278, 293, 300, 316, 325, 329–336, 350, 362, 364, 366, 367, 369, 476 Greek (Classical) 146, 147 Hungarian 10, 12, 92, 94, 104–105, 137, 215, 225, 229, 236, 241, 245, 261–262, 265–267, 269–283, 292, 294, 295, 297, 303, 305, 321, 351, 356, 358, 359, 361 Icelandic 10, 92, 96, 191, 199, 202, 203, 266, 276, 300 Indo-European 4, 10–12, 23, 68, 71–72, 79, 88, 95, 100–101, 106, 113–114, 116, 138–139, 159, 176, 178, 194, 207–209, 261, 271, 284, 293, 295, 296, 315, 351, 367–371, 393 Iranian 12, 65, 66, 72, 87, 98, 100, 225–253 Iranian (Middle) 243 Italian 2, 4, 6, 49, 73, 76, 93, 119, 120, 122–126, 136, 138, 145, 147, 150, 233, 234, 249, 263, 301, 304, 306, 387–404, 406–428, 433–437, 439–448 Judeo-Arabic 235, 236, 239, 240, 246, 247, 249, 250 Judeo-Belarusian 244
Judeo-French 234 Judeo-German 234, 245 Judeo-Iranian 232, 237, 242 Judeo-Slavic 242, 244, 251 Judeo-Spanish. See Judezmo Judezmo 230, 235, 236, 249, 250, 350 Karaite 239, 240 Kashubian 3, 8, 14, 94, 132–133, 211, 295, 299, 301, 306, 322, 435, 453–483 Khazar 227, 230, 231, 232, 234, 244, 245 Lachian 265 Latin 6, 11, 73, 74, 79, 98–101, 103, 130–131, 133, 135, 139, 146–147, 149, 150, 152, 154, 156, 178, 235, 263, 270–271, 273, 277, 293, 329, 333, 420 Latvian 4, 41, 44, 53, 306, 351 Lithuanian 36, 40–41, 43–44, 46, 49–53, 56, 94, 137, 202–203, 208, 214, 268, 306, 351 Macedonian 6, 13, 99, 116–117, 123, 124, 127, 133, 134, 139, 151, 195, 211–213, 266, 293, 300, 315, 317, 322, 325–338, 350, 362–363, 366, 392, 406, 420, 423, 436, 440, 469 Molise Croatian. See Molise Slavic Molisean. See Molise Slavic Molise Slavic 3, 8, 13, 14, 119, 122, 294, 385–425, 433–446, 456, 482 Norwegian 96, 276 Old Church Slavonic 11, 26, 91, 130, 131, 145–177, 296–297, 302, 475, 482 Old Prussian 104 Ossete 227, 234, 240, 245, 251 Persian 233–234, 235, 238–244, 247–250 Polabian 226, 435, 467, 482 Polish 4, 26–27, 31, 36, 39–51, 94, 97–99, 102, 118–120, 133, 195, 197, 211, 228, 236, 240, 246, 248, 252, 265, 267, 270, 274, 277, 292, 294, 296–303, 354, 356,
Index of languages
359, 361, 453–455, 461–463, 468, 473, 474, 475–476, 478–479, 481–483 Polish (Old) 257 Polish (Silesian). See Silesian Polish (Standard) 27, 36, 40, 265, 394 Polissian 251, 365 Polissian (Kyiv-) 251 Romance 3, 7, 10, 13, 21, 36–37, 42, 48, 54, 64–66, 72–80, 87–90, 119, 122, 128, 130, 134, 137, 147, 149, 159, 187–188, 207, 225–226, 233, 235–236, 249, 263, 270–274, 284, 291, 293, 304–306, 322, 335, 367, 386, 395–397, 408, 411, 415, 420, 425, 456, 463, 482 Romance (Ibero-) 233, 235, 236 Romani 137, 208–209, 211, 228, 253, 292, 294–295, 350, 362 Romanian 6, 10, 13, 27, 74–78, 102, 146–147, 293, 304, 350, 354–356, 359, 361–362, 364, 366, 368–369, 371 Russian 2, 21, 27, 31, 33, 36, 41–53, 75, 91–94, 97–98, 100, 124–126, 128, 133–134, 195–200, 208, 211–213, 245–246, 264, 267, 293, 296–297, 299, 302–303, 306, 322, 351, 355, 367, 369, 397, 399, 404, 411, 414, 433, 455, 477, 482 Russian (North) 5, 45, 124, 126, 133, 211, 293, 322 Russian (Northwest) 31, 33, 42, 44, 46 Russian (Standard) 31, 36, 41, 44, 48, 124, 125 Rusyn. See Ukrainian (Rusyn) Scandinavian 6, 37–38, 53, 92–94, 96, 98, 100–101, 105, 135, 137, 306, 351 Scythian 227, 232 Semitic 54, 88–89, 95, 225, 234, 236 Serbian 13, 31, 34, 46, 74, 102, 128, 203, 211, 214, 246, 292–293, 295–296, 299–303, 322, 325–330, 332, 334–335, 338, 350, 353, 388–389, 390, 393, 411, 414, 416, 423, 426, 428 Serbian (Eastern) 13 Serbian (Prizren-Timok) 350, 353 Serbian (Torlak) 362
497
Silesian 477 Slavic 1–14, 21–55, 63–80, 87–105, 113–138, 146, 150–153, 155, 158–160, 172–174, 177, 187–215, 225–250, 291–304, 315–339, 385–425, 433–447 Slavic (Balkan) 7, 12–13, 27–29, 46, 123, 267, 275, 283, 294–296, 298–299, 301–302, 315, 317, 320, 324, 329, 332, 335–337, 350, 359, 361–362, 365, 368–369, 374, 406–408, 420, 424 Slavic (Carpathian) 374 Slavic (Common) 10, 24–28, 33, 42, 63–80, 100–101, 129, 151, 159, 173, 178, 197, 230, 291, 296–298, 358–359, 363–364, 366, 371, 386, 456, 476, 482 Slavic (East) 2, 4, 6, 21, 27, 40, 43–45, 47–49, 53, 97, 101–104, 126–128, 152, 172–173, 177, 195, 211, 213, 267, 277, 283–284, 296–299, 301–303, 316, 358–359, 371, 475 Slavic (North) 3, 28, 33–34 Slavic (Old East) 98, 126–127, 177, 303 Slavic (South) 27–28, 33–34, 46–47, 73, 99, 152, 177, 211–212, 229, 245, 277, 299, 316–317, 320–325, 330, 332–335, 337, 357, 364, 369, 388, 390, 436, 467, 469, 471, 477 Slavic (West) 5, 26, 39, 42, 47–49, 90, 103, 125, 137, 177, 207, 211, 214, 251, 299, 316, 321, 467 Slovak 12, 99, 102, 211, 236, 245, 248, 261, 265–270, 273–280, 282, 292, 294–295, 298–299, 301–303, 354, 360, 364, 482 Slovene 28, 33–34, 46, 73, 128–129, 137, 265, 270, 273, 277, 292–296, 298, 300–303, 306, 353, 397, 404, 411, 420, 435, 439, 455–457, 482 Slovene (Carinthian) 385 Slovene (Resian) 119–123, 456 Slovene (Standard) 312 Slovincian 435, 453, 456, 457, 468, 476, 484 Sorbian 27, 34, 92, 99, 119–120, 122–123, 207, 211–214, 226, 236, 245, 251, 253, 265, 275, 294–295, 299–305, 322, 359, 385, 399, 409, 411–414, 426, 435–436, 439, 456–457, 467, 477, 482
498
Index of languages
Sorbian (Upper) 92, 119–120, 123, 211–212, 236, 245, 294, 300–303, 385, 399, 411, 413, 426, 435–436, 439, 456, 457 Standard Average European 1, 9–11, 21, 79, 89, 93, 113–139, 145–179, 187–214, 291, 367, 387, 428, 433, 453 Standard Average Indo-European 23, 284 Standard Average Western European 4 Standard Central European 10, 90 Štokavian 292, 295–296, 298–300, 302–303, 353, 388, 406, 436 Turkic 6, 12, 87, 97, 105, 188, 209, 225, 227–236, 240, 242–245, 251–253 Turkish 137, 146, 192, 199, 202, 211–212, 215, 239, 317, 329, 350, 358, 362, 372, 436 Ukrainian 2–3, 21, 40, 45–46, 56, 94, 97–99, 102, 128, 133, 136, 194–197, 212–213, 215, 236–237, 240–241, 244–245, 250–251, 267, 270, 276, 278,
283–284, 292, 296–297, 299, 301–303, 306, 347–348, 353–371, 439, 455, 482 Ukrainian (Bojkian) 363 Ukrainian (Bukovyna) 356, 361 Ukrainian (East) 361 Ukrainian (Hucul) 355 Ukrainian (Lemkian) 363 Ukrainian (Rusyn) 296 Ukrainian (Southeast) 278, 366–367 Ukrainian (Southern Bessarabian) 380 Ukrainian (Southwest) 21, 40, 102, 296, 300, 347, 348, 353–357, 359, 362–364, 366, 370 Ukrainian (Transcarpathian) 254, 366 Ukrainian (West) 46, 301, 361 Uralic 65–68, 73, 87–88, 90, 94, 98, 106, 188, 193, 195, 198, 201, 205, 207, 209 Yiddish 11–12, 225–250, 262, 292–295, 302–304 Yiddish (Slavic) 236, 239