(PDF) without Code : The Evolution of the Sofia-Berkeley Archive of Bulgarian Dialectal Speech

Digital Publication without Code: The Evolution of the Sofia-Berkeley Archive of Bulgarian Dialectal Speech Ronelle Alexander and Quinn Dombrowski, UC Berkeley

,

The traditional dialect atlas—providing lists of isolated words collected from various locations, and/or maps that provide a simplified rendering of reflex distribution-- has a number of shortcomings that inhibit a full understanding of dialectal speech. “Bulgarian Dialectology as a Living Tradition”, an ongoing project resulting from a partnership between Ronelle Alexander (UC Berkeley) and Vladimir Zhobov (Sofia University), aims to address some of the shortcomings in previous dialect atlases by presenting transcriptions of vivid, authentic speech from approximately sixty sites. The shift from a print-based workflow to a general-purpose, open- source platform that provides a structured data model for token, lexeme, and text entry has facilitated additional ways of organizing, analyzing and visualizing the material. The project data is derived from the “Sofia-Berkeley Archive of Bulgarian Dialectal Speech”, the result of a series of joint field trips starting in 1991. From this archive, audio clips of up to 10 minutes were selected and digitized from sixty sites. Originally, these audio clips were accompanied by a set of Microsoft Word files that included phonetic transcriptions, English translations, grammatical annotation, analysis of salient linguistic features, as well as a discussion of the village and the content of the text. This traditional approach is optimized for a print-publication workflow, or even for electronic publication whereby text files can be coordinated with streaming audio. Opportunities for using the material in other ways are limited, however. Using Word’s search feature to identify all tokens with a particular set of traits is clumsy; identifying all forms of a lexeme is even more challenging, and neither is at all possible for a scholar who only has a printed copy. Rather than building a custom database to address these and related issues, or using software specifically designed for linguistic corpora or cultural textual material, the project adopted the open-source platform Drupal, which has a large international developer community. Most commonly used as a content management system for websites, Drupal’s flexible architecture and actively-supported feature “modules” allowed the project to implement data models and taxonomies that capture the important characteristics of this material without having to support and maintain custom code. The value of Drupal skills in both the web development and “alternative academic” spheres has also provided an opportunity to involve students from outside the field of Slavic linguistics in technical support. This talk will cover both the project goal of capturing the richness of Bulgarian dialects within the context of authentic speech about culturally relevant topics, and the technical approach that has enabled new ways of interacting with that material, without diverting a significant amount of project funding towards custom programming. 1 The Sociolinguistics of (o) in Ukraine: a Perceptual Approach Yuliia Aloshycheva, The Ohio State University

Studies in perceptual dialectology indicate that based on their “imagined” maps, listeners are able to match speakers to geographic regions (Preston, 1989; Plichta & Preston, 2005). This process is subjective and is sensitive to listeners’ identity and attitudes (Preston, 2010), of which place of origin is one dimension (Benson, 2003; Bucholtz et. al, 2008). In this study, we manipulate a socially marked phonological variable to investigate how listeners from different parts of Ukraine conceptualize the social space in Ukraine. Ukraine represents a complex sociolinguistic situation (Bilaniuk, 2005), with a symbolic confrontation between its South-Eastern (‘pro-Russia/n’) and Western (nationalistic, ‘true’ Ukraine/Ukrainian) parts. Russian is widely spoken in Ukraine; however, under the recent policies of Ukrainianization, more people are using Ukrainian, and sometimes transferring over their Russian accent. One example is /o/ reduction to [a] in unstressed positions – ‘akan'e’ (Monaxova, 2001). While this is the phonetic norm in Russian, akan'e is stigmatized as ‘impure’ in literary Ukrainian, which uses ‘okan'e’, a well-rounded [o]. In this study we analyze Ukrainians’ reactions to [a] use vs. [o] use to reveal beliefs about symbolic social divisions in Ukraine. We look in particular at the stereotypical East-West divide, the Dnieper River. Twelve short texts were recorded by different speakers. Using Praat we resynthesized three versions of each text to contain either unstressed [o], unstressed [a], or intermediate [ə]. After a background questionnaire, 72 participants from Western, Central and Eastern parts listened to one version of each text and matched each speaker with a region of Ukraine. We predicted the same response pattern from all listeners: okan'e samples would be put westwards and akan'e eastwards. Also, the boundary for the stigmatized variant, akan'e, was expected to be pushed westward by Easterners (the pattern shown in Preston (1989) for the U.S.). The basic prediction was upheld. Overall, all three listener groups identified [o] samples with the West and [a] samples with the East. However, listeners from the West were more sensitive to the Dnieper River as a boundary: okan'e samples were put almost exclusively West and akan'e samples almost exclusively East of it. Easterners’ responses were more evenly distributed. Two areas that showed a contradictory pattern were Dnipropetrovs'k and Poltava regions, which were labeled as akan'e by Westerners and okan'e by Easterners. Moreover, Easterners readily associated (nationally) stigmatized variants with their own region. This contrasts with Preston’s (1989) findings in the U.S. In line with third-wave sociolinguistic studies, our results thus emphasize the need to pay attention to locally situated values (e.g. Eckert 1988). What does it mean to be stigmatized on a nationwide vs. local level? Exploration of this question in the context of Ukraine is left for future research. Ultimately, each listener group has its own vision of Ukraine: more homogeneous for Easterners, more segregated for Westerners. The claim that all Ukrainians believe there to be ‘two Ukraines’ (Riabchuk, 2002) thus turns out to be too simplistic. The data show subtle differences in how different groups conceptualize of social divisions in Ukraine; this is not reflected in the broad stereotypes. 2 Hutsul Clitics: A New Look at Frequency and Ordering Erin Coyne, UC Berkeley

While the West and South Slavic languages have furnished most of our evidence on Slavic clitic inventory and ordering , occasional reference is also made to the Southwestern Ukrainian dialect group, and particularly the Hutsul dialect, spoken in the previously isolated Carpathian Mountains. Among the archaisms preserved in Hutsul are clitic strings, which at times include a combination of pronominal, reflexive, conditional, and auxiliary clitics. However, very little is currently known about the inventory, frequency, and ordering of clitics in Hutsul as much of the previous research was based on scant available resources (Franks and King 2000) or focused on structures typical of older forms of the dialect (Budzhak-Jones 1996), rather than the changes presumed to be occurring in modern speech. This paper expands on previous clitic research by taking advantage of the recent upsurge in Hutsul language publications, particularly the recently published Hutsul novel Dido Ivanchik by Petro Shekeryk-Donykiv (2007), which presents an extended text capturing what is thought to be the pinnacle of the old Hutsul dialect. In addition, I will also examine the recently published memoirs of Hutsul poet Paraska Plytka-Horysvit (2008), as well as a series of newspaper articles published during the last decade in Hutsul dialect, the recent transcripts of Hutsul informants interviewed by T. Yastremska, as well as findings from my own fieldwork in Verkhovyna, Ukraine. Through these varied, newly-available sources I will track the changes which have occurred in Hutsul clitic use with a focus on a recent process of attrition in both inventory and overall frequency. By examining a range of works spanning approximately 100 years, I will be able to present both a more complete picture of clitics distribution and use in the old dialect, as well as the much less examined contemporary inventory, frequency, and ordering. 3 Keyword Avnalysis of Czech Political Texts: a Pilot Study Václav Cvrček, Charles University, and Masako Fidler, Brown University

,

This is a pilot study of a larger project on statistical analysis of keywords in political texts in Czech, combined with tools from discourse analysis. The ultimate goal of this project is to provide a maximally objective approach to interpreting political texts across language boundaries. Our presentation will focus on a limited genre (New Year's Address) in Czech at selected points over a span of 20 years (1970-1990), following changes in keywords and discourse strategies. Although the target texts from the socialist period may appear to repeat the same clichés and empty ritualistic expressions, both our preliminary qualitative and quantitative analysis suggest that the texts are not as homogeneous as they seem; in addition to subtle shifts in the assortment of keywords, the recurrent words and phrases are used in slightly different contexts over time and as a result convey different nuances at different points of history. Keyword analysis helps characterize a set of texts or a genre by comparing the frequency of words and phrases within the target text against the background of general language use. It is used for analysis of literature (e.g. Scott and Tribble 2006), language and society (Baker 2010), discourse (Duguid 2010, Baker and Ellece 2011), literary style (Čermák and Cvrček 2009, 2010), and language of a specific time period (Čermák, Cvrček and Schmiedtová 2010). Unlike the previous studies, this project investigates the dynamic shifting functions of keywords in context over time, following keywords with various lengths of lifespan and changes in the focus of attention as reflected by the keywords. The quantitative results will be cross-examined by qualitative and quantitative analysis based on models from discourse-cognitive linguistics and morphosyntax (Individuation hierarchy (Timberlake 1975), Transitivity model (Hopper and Thompson 1980), and conceptual blending (Grady et al. 1999, Fauconnier and Turner 2002). References Baker, P. Sociolinguistics and corpus linguistics . Edinburgh, 2010. Baker, P. and Ellece, S. Key terms in discourse analysis . London, 2011. Čermák, F. and V. Cvrček. Slovník Bohumila Hrabala. [Vocabulary of Bohumil Hrabal] , Prague, 2009. Čermák, F. and V. Cvrček. Author dictionaries revisited: Dictionary of Bohumil Hrabal. In A. Dykstra and T. Schoonheim. Ljouwet: Fryske Akademy Afûk, 2010. 592-8. Čermák, F., V. Cvrček, and V. Schmiedtová, eds. Slovník komunistické totality. Prague. Czech National Corpus (www.korpus.cz), 2010. Duguid, A. Newspaper discourse informalisation: a diachronic comparison from keywords. Corpora, 2(10): 109-138. Fauconnier, G. and M. Turner. The way we think: Conceptual blending and the mind's hidden complexities. New York, 2002 Grady, J. E., T. Oakley and S. Coulson. Blending and metaphor. Metaphor in cognitive linguistics, ed. by R. W. Gibbs Jr. and G. J. Steen, 101-24. Amsterdam-Philadelphia, 1999. Hopper, Paul J. and Sandra Thompson. Transitivity in grammar and discourse. Language 56: 251-299, 1980. Scott, M. and C. Tribble. Textual patterns: Keyword and corpus analysis in language education. Amsterda, 2006. Timberlake, Alan. The hierarchy of the negation. SEEJ 19:123-139, 1974. 4 Another Look at the “ruki” Law in Baltic and Slavic Bill J. Darden, University of Chicago (ret.)

The differences between the reflexes of the “ruki” law in Lithuanian (the only relevant Baltic language) and Slavic have been used to (unjustifiably) argue against a Balto-Slavic protolanguage.. Proto-Slavic must be reconstructed with special phonetics of *s after r, u, k, i. The same is true of Lithuanian, except that it must have had individual differences after k, r, and high vowels. Henning Andersen has argued that this was true of the parent language. The languages differ in how the multiple entities were reduced to two—after they separated. Differences in morphophonemic behavior are relatable to the fate of the palatovelars. In Slavic k’ > s made x(š) a phoneme and eliminated the phonological motivation for s > x. A “rule reversal” allows for an interpretation of alternating morphemes as underlying x with a regular change of x >s before stops or after non high vowels or non-rhotic dentals and labials. The change kx > x allowed x after low vowels and these were removed from the conditioning environment, In oldest OCS, what remains is a rule whereby x > s after nonvelar obstruents and nasals, or before t. In proto-Lithuanian, the palatovelar merged with the back variant š, so morphemes with s/ š alternations could still be interpreted as underlying s with regular s > š . Changes such as d/t > s before t, and d/t > zero before s, and the metathesis of sk, šk > ks, kš before obstruents rendered the rule irregular, and the underlying s was generalized in all productive suffixes.. The phonetic changes in Lithuanian are interesting. The received information is that the change took place after palatovelars and before stops (like Sanskrit, unlike Slavic), and was inconsistent after high vowels. Palatovelars are not a problem. In both languages *k’s yields a single continuant with the reflex of the palatovelar (Lith. š, Slavic s). The behavior before stops may require the separation of velars, r, and high vowels. There are no relevant forms with *sp. I have found examples of z > ž (before d) only after velars (žvaigždė/ žvaizdė ‘star’). This also shows that the change after velars remained possible after “velar insertion”. There are good examples of s > š before t after r and k (mir-št-a with –st- present suffix, aukštas ‘high’ from aug-ti ‘grow’). I have not found any examples of s > š before t after high vowels that are not easily accountable from paradigmatic leveling of stems. There are special problems with *sk, because šk can come from any *sk’ in noninitial position. I have not found any examples where šk is provably due to a preceding high vowel. (I have checked the -*st- examples; I still have some etymologies for – šk- to check.) This opens the possibility of treating Lithuanian as a mixture of two dialectal solutions: one very much like Slavic, except for the merger with the palatovelar, and one where only the back variant after r, k (including the environment before stops) merges with š. The latter variant must have been dominant, with residues of the change after high vowels left in the language. 5 Psycholinguistic Analyses of Six Russian Poets Chad C. Davidson, Johnson County Community College

The purpose of this study was to determine whether the patterns of usage of Russian-written poetry reflect the poets’ psychological states in the same way that the patterns of usage of English-written poetry have been found to reflect their poets’ psychological states. In 2001, psychologists James Pennebaker and Shannon Stirman (P & S) published a study of English- written poetry by eighteen poets, nine of whom committed suicide (Stirman & Pennebaker 2001). P & S found that the linguistic choices of the poet reflected whether or not s/he was suicidal. For example, “suicidal poets used far more first person singular” than non-suicidal poets (p. 520). In his 2011 book, The Secret Life of Pronouns, Pennebaker revealed other linguistic choices that differentiated suicidal poets from non-suicidal poets (e.g. non-suicidal poets use more present tense verb forms and less past tense verb forms than suicidal poets) (Pennebaker 2011). I used the total linguistic predictors of suicide given in Pennebaker’s works as the list of linguistic features analyzed in my study of Russian-written poetry. In conformity to P & S’s study, my study counted a pair of matched poets as one non- suicidal poet and one suicidal poet. Each non-suicidal poet was “matched as closely as possible for nationality, era, education, and sex with one of the suicidal poets” (2001: p. 519). I found a total of six matched poets (three suicidal and three non-suicidal), giving me three pairs. Poems were selected from three different periods of the poets’ lives and compared accordingly. However, P & S compared the poetry written by the suicidal poets’ during their final year of life to the non-suicidal poets’ works written within one to eight years of the same age. Along with this, I compared the poems written in the last years of the non-suicidal poets’ lives. In this way, I was able to ascertain if the heightened awareness of death for the non-suicidal poets was reflected in their poetic language in a way similar to the looming death of the suicidal poets. A total of 11 overarching linguistic predictors of suicide were analyzed. Within some of those features a number of sub-features were analyzed (e.g. under an overarching feature called “self-references,” I counted sub-features “first-person singular” separately from “first-person plural”). Each feature and sub-feature were identified, counted, and divided against the total number of words used in the poems of the same period by the same poet. The composite list of percentages describes that poet’s pattern of usage for that time period. Each finding was compared to those of their matched poet for the same time frame. Finally, the overall findings of the suicidal group were compared to the overall findings of the non-suicidal group. Interesting similarities and differences were noted along with potential for further research. References Pennebaker, James W. The Secret Life of Pronouns. New York City, New York: Bloomsbury Press, 2011. Stirman, Shannon Wilstey & James W. Pennebaker. "Word Use in the Poetry of Suicidal and Nonsuicidal Poets." Psychosomatic Medicine, 2001: 63: 517-522. 6 Actions, Goals, and the “Presuppositional” Use of the Imperfective in Russian Stephen M. Dickey, University of Kansas

The Russian impf general factual (IGF) has been a thorny subject in Russian aspectology for decades. Though there is a consensus on the existential IGF, there has been considerable disagreement on a related use of the impf past in questions and statements regarding particular actions, which has been termed the “actional” meaning of the impf (Padučeva 1996), or the “presuppositional” factual impf (Grønn 2003), and which focuses on some circumstance of the action, as in ex. (1): (1) Devuška, gde vy sapogi pokupali i ? ‘Miss, where did you buy the boots?’ (Israeli 1998: 72) According to Padučeva (1996: 49), the “actional” impf “focuses attention not on the result, but the process [of the action] itself.” Glovinskaja (1981: 183) suggests that in such questions the event is part of the theme, as opposed to the element being asked about. Similarly, Grønn (2003) argues that the impf is appropriate because the event is a presupposition in such utterances. This paper argues that none of these factors is directly relevant to the occurrence of the “actional” or “presuppositional” impf. Rather, the mechanism at work involves knowledge about the original goal/intent of the action. In such questions about the circumstances of some action, there are two possibilities concerning its original goal/intent: in the first case the speaker asks about the action accepting, according to his/her understanding, the terms of its original goal/intent, or asks about that goal/intent; in the second case the speaker asks about the action and (1) implicitly denies that the original goal/intent was realized, or (2) ignores the original goal/intent and asks with some purpose of his/her own. Consider the following examples: (2) a.Kakie u nas neudobnye mesta! Kto pokupali ix? (Forsyth 1970: 84) ‘What uncomfortable seats we’ve got! Who bought them?’ b.Kakie u nas xorošie mesta! Kto kupilp/*pokupali bilety? ‘What good seats we‘ve got! Who bought them?’ (3)— Gde ty kupilap takuju krasivuju štuku? — sprosila Kama, vse ešče vertja v rukax raskrytyj zont. (Iskander, Sandro iz Čegema; RNC) ‘“Where did you buy such a pretty thing?” asked Kama, still twirling the open umbrella in her hands.’ In (2a), the speaker’s view that the action failed to realize the assumed original goal of the action (comfortable seats), triggers the impf; in contrast, the question in (2b) is pf since the speaker accepts that the original goal has been realized. In (3), the speaker asks about the purchase as she admires the particular item purchased, thus acting on her assumption that the agent wanted to buy a nice umbrella. In contrast, in (1) above the speaker is not interested in the agent’s goal for buying the boots, but wants to go and buy boots for herself. References Forsyth, J. 1970. A Grammar of Aspect: Usage and Meaning in the Russian Verb . Cambridge: Cambridge University Press. Glovinskaja, Marina Ja. 2001. Mnogoznačnost´ i sinonimija v vido-vremennoj sisteme russkogo glagola. Moscow: Azbukovnik/Russkie slovari. Grønn, Atle. 2003. The Semantics and Pragmatics of the Russian Factual Imperfective. PhD dissertation, University of Oslo. Israeli, Alina. 1998. Speakers’ Attitudes, Goals and Aspectual Choices in WH-Questions. Le Langage et l’Homme 33.1, 55–77. Padučeva, Elena V. 1996. Semantičeskie issledovanija. Moscow: Jazyki russkoj kul´tury. 7 The Constructional Approach to Lexical Synonymy Dmitry Dobrovolskiy and Ludmila Pöppel, Stockholm University

;

Synonyms are usually defined as words coinciding in their core meanings. The number of common semantic features should prevail over the number of distinctive features (cf. Apresjan 2009: 539). In this research tradition, synonyms proper are defined as lexical units revealing peripheral, insignificant distinctive features. Otherwise we are dealing with near-synonyms. One task of lexical semantics is to identify and describe semantic features that distinguish between synonymic words. Conventional dictionaries are rarely capable of fulfilling this task. Exceptions are special dictionaries of synonyms such as NOSS (2004) based on a solid semantic theory. All methods of synonymy research are based on the analysis of relevant contexts, although each method focuses on specific aspects of the contextual behavior of synonyms. Within the semantic theory developed by Apresjan (1995, 2000, 2009), contexts profiling semantic differences between synonyms play the central role. If a lexical unit cannot be replaced by its (near-)synonym in a given context it proves that the synonyms are not identical semantically. This approach makes it possible to single out all relevant distinctive semantic features of every lexeme in question by analyzing diagnostic contexts. The basic idea of Fillmore’s theory of frame semantics (1982, 1985) is that the meaning of a single word cannot be understood without access both to the essential knowledge that relates to the word and its combinatorial properties. In order to describe a word’s semantics and to distinguish between (near-)synonyms, one has to study the range of its semantic and syntactic valences, i.e. its combinatorial profile. This enables us to fill the slots of corresponding frames, i.e. to postulate all obligatory and facultative participants in the situation pointed to by the lexeme in question. The third method can be labeled the constructional approach. It has been developed since large text corpora became available. Its basic assumption is that (near-)synonyms are sensitive to specific constructions. The method is based on corpus evidence (cf. Janda & Solovyev 2009, Divjak 2010, Divjak & Gries 2008). Examples are the words революция ‘revolution’ and переворот ‘coup’. Both words are sensitive to constructions [во имя N] ‘for the sake of N’ and [на благо N] ‘for the benefit of N’. In the Russian National Corpus the construction во имя революции ‘for the sake of revolution’ occurs 35 times, and на благо революции ‘for the benefit of revolution’ 2 times, whereas there are no hits for на благо переворота ‘for the benefit of coup’, and only one hit for во имя переворота ‘for the sake of coup’ dated 1880, i.e. a context that obviously does not conform to the present-day usage norms. The difference in constructional embedding can be explained by the fact that революция is directed towards noble long-term objectives, which is not the case with переворот. In our talk we will develop this approach to lexical synonymy by analyzing the constructional behavior of the words восстание ‘uprising’, бунт ‘riot’ and мятеж ‘mutiny’. 8 Laterals and Diphthongs in Slavic, Albanian, Turkish, and Aromanian: Convergence and Contact Andrew Dombrowski, University of Chicago (ret.)

Phonology has traditionally played a minimal role in what Joseph (2007: 6) terms “broad” - as compared to “localistic – approaches to the study of Balkan linguistics, due to the absence of solid phonological innovations that characterize the entire scope of the Balkan Sprachbund (although see Sawicka 1998 for a division of the Balkans into three phonological zones). Nonetheless, phonology has played a crucial role in localistic studies of the Balkan Sprachbund, precisely because contact-induced phonological features do not tend to spread across a broad geographical range. As a result, it is often possible to evaluate the linguistic and social dynamics resulting in small-scale contact situations in greater detail than is possible with regard to morphosyntactic features characteristic of the entire Balkan Sprachbund, which in turns enables a more nuanced analysis of the mechanisms of language contact involved in a given case study. In this paper, I analyze convergent developments in laterals and diphthongs between Slavic, Albanian, Turkish, and Aromanian, and show that in these cases, surface similarities in phonological systems have evolved through the pathway of structurally distinct sound changes. The first case study involves diphthongs in Macedonian, Aromanian, and Albanian. Multiple sound changes, some contact-induced and some independently motivated (Dombrowski 2012), have resulted in similar patterns of hiatus resolution and diphthongization in the dialects of these three languages spoken in western Macedonia. The second case study discussed here deals with laterals. Slavic dialects spoken in close contact with Albanian develop systems of laterals that are more similar to the Albanian system than to other varieties of Slavic. Peripheral dialects of Montenegrin and Serbian display a tendency to adjust the distribution of laterals by fronting */l/ to [l] before front vowels, backing */l/ to /ʟ / elsewhere, and adjusting /ʎ / to [l] . The result is a phonemic distinction between /l/ and /ʟ /, which is precisely the phonemic distinction between laterals attested in neighboring Albanian. Analogous developments have also taken place in Serbian dialects spoken in southern Kosovo, such as the urban dialects of Gjakova and Prizren, and the rural dialects of Gora and Sretečka Župa (Remetić 1996: 411-414, see also Vidoeski 1998: 340). Analogous examples from Prizren include poljubim 'I kiss' < poλubim (/λ/ > [lj]), ljek 'medicine' < lek (/l/ > [lj] / _Vfront), and beʟ -a 'white-FEM' < bela (/l/ > [ʟ ] elsewhere) (Remetić 1996: 413). Turning to Macedonian, whereas western central dialects of Macedonian show a tendency to merge /l/ (< λ) with /ʟ / (< l), peripheral dialects spoken closer to Albanian are more conservative, thereby preserving a system more like that found in Albanian. In these dialects, /l/ was velarized like elsewhere in western Macedonian, but /λ/ was shifted to an alveolar /l /. The result is an opposition between /l / and /ʟ / structurally identical to that in Albanian. I also draw upon orthographic evidence in the Turkish text of Gjorgji Puleveski's 1875 Dictionary of Three Languages to argue that a similar change took place in at least some varieties of Balkan Turkish. References Dombrowski, Andrew. 2012. “Diachronic and Areal Aspects of Macedonian Hiatus Resolution.” Balkanistica 25:2.19-32. (=Macedonian Matters: Proceedings from the Seventh Macedonian -North American Conference on Macedonian Studies, ed. by Victor A. Friedman and Donald L. Dyer Joseph, Brian. 2007. “Broad vs. Localistic Dialectology, Standard vs. Dialect: The Case of the Balkans and the Drawing of Linguistic Boundaries.” in Papers from ICLAVE 4 (International Confere nce on Language Variation in Europe), Lefkosia, Cyprus, June 17-19, 2007. Pulevski, Gjorgji. 1875. Rečnik od tri jezika. Belgrade: Državna štamparija. Remetić, Slobodan. 1996. “Srpski prizrenski govor - glasovi i oblici.” Srpski dijalektološki zbornik XLII: 319-614 Vidoeski, Božidar. 1998. Dijalektite na makedonskiot jazik: Tom 1. Skopje: MANU. 9 What Do We Call the Children? Familial Terms in Polish and English: A Corpus -Based Study Katarzyna Dziwirek, University of Washington

This paper addresses cross-cultural differences in the use of familial forms of address, that is, kinship terms (mother, father), kinship diminutives or hypocoristic forms (mom, daddy), nicknames, first names and their diminutives, and terms of endearment (sweetie). I argue that the differences in the use of familial forms of address between English and Slavic stem from their different cultural values. The data discussed in the paper comes from the Polish National Corpus (NKJP), British National Corpus (BNC), the Corpus of Contemporary American English (COCA), as well as from a web questionnaire. In particular, I focus on child-directed familial terms and the major contrast between English and Slavic in this regard: English speakers typically do not address their children as son or daughter, while Polish speakers do so frequently. There are also no diminutives of these words in English (other than sonny), even though English parents do use other diminutives in child- directed speech (doggie, blankie, auntie, etc.) as evidenced in the BNC and COCA. In Polish, there are 8 common hypocoristic forms based on córka ‘daughter’ and 9 such forms based on syn ‘son’. These forms are frequently found in the vocative in NKJP showing that they are used as forms of address. I discuss these forms briefly, focusing on their meanings, frequencies, and the fact that they exhibit both archaic and some innovative morphological properties. I also discuss the gender difference in the use of terms for son and daughter: some of the diminutives of syn are used primarily as forms of address as they are considered too intimate to be used in discourse about one’s sons. In English, in contrast, there are no hypocoristic forms of daughter and English speakers tend not to use daughter as an address form (apart from religious contexts). We do find examples of English son, sonny and sonny boy as forms of address, but they have a special flavor: son, when used to the speaker’s child is most often found in the context of offering manly advice, and when son, sonny and sonny boy are directed at other people’s children they can be patronizing and condescending. Dziecko “child” and its 6 common diminutives are also frequently used as forms of address in Polish. In English child (especially in the collocation my dear child) can be used appellatively, but not necessarily at the speaker’s own offspring. I argue that the reason for the lack of adult-source familial terms for children in English stems from the enormous importance of the English cultural value of individuality, That is, calling someone son or daughter assigns them to a category of people and denies their individuality, and that is why speakers of English address their children by their names or family nicknames. In Polish, as argued by Wierzbicka (1999), one of the key cultural values is warmth, which is conveyed by, inter alia, expressive derivation of both first names and family terms. This is confirmed by the fact that using son appellatively in English (to other people’s children) often has a denigrating sense, while in Polish it does not. Reference Wierzbicka, Anna. 1999. Emotions Across Languages and Cultures: Diversity and Universals. Cambridge: Cambridge University Press. 10 Allomorphy via Borrowing? The Status of the Prefixes PRE- and PERE- in Modern Russian Anna Endresen, Tromsø University

Can a borrowed prefix become an allomorph of a native morpheme? Addressing this theoretical question, I present the first corpus-based study of two prefixes that are diachronically and semantically related in Modern Russian: the native East Slavic prefix PERE- and the parallel prefix PRE- borrowed into Russian from Old Church Slavonic. As opposed to Modern Ukrainian and Belorussian, where the prefix PRE- has almost completely disappeared (Adamovič et al. 1980; Pugh & Press 1999), Russian preserves both prefixes and shows an interesting pattern of distribution. The use of PERE- and PRE- in Modern Russian have received dramatically different amounts of attention in literature. While the native prefix PERE- is well described (Janda 1986; Dobrušina & Paillard 2001), the borrowed PRE- has been largely neglected due to its “vagueness of meaning” (Isačenko 1960) and low productivity. I question both of these statements and show that PRE- is a highly productive intensifier which operates on adjectives and adverbs (e.g. bol’šoj ‘big.ADJ’ > prebol’šoj ‘very big.ADJ’) and that its diverse semantics is compatible with that of the prefix PERE-. I focus on the mutual relations between PERE- and PRE- and examine whether they satisfy the two crucial criteria of regular allomorphy – identical meaning and complementary distribution (Haspelmath 2002). The data are extracted from the Russian National Corpus (www.ruscorpora.ru) and cover both verbal derivation as well as non-verbal uses of PERE- and PRE-. First, I provide a contrastive radial category analysis of the two prefixes and argue that their diverse semantics can be captured within a single network of fourteen submeanings, in eight of which they overlap. The large degree of semantic overlap suggests that PERE- and PRE- might represent a single morpheme. On the other hand, minimal pairs like pere-stupit’ ‘step over’ – pre- stupit’ ‘violate (a law)’ question the complementary distribution of the prefixes across verbs. Thus, PERE- and PRE- do not fully satisfy either of the two criteria of regular allomorphy: their semantics is similar but not identical; their distribution is distinct, but not exactly complementary. Since the absolute criteria fail, the data has been subjected to statistical analysis (Exact Fisher’s Test), which measures the degree of semantic similarity and divergence, as well as estimates mutual attraction and repulsion between the submeanings of the network on the one hand and each of the two prefixes on the other hand. The distribution of PERE- and PRE- in five submeanings is found to be statistically significant. I argue that the complexity of this nontrivial case can be best captured by extending the traditional notion of allomorphy. I suggest that PERE- and PRE- should be viewed as non- prototypical allomorphs of a single prefix rather than as different morphemes. The Cognitive linguistic framework and statistical analysis of corpus data make it possible to account for both the subtle differences and the important similarities of the two prefixes. References Adamovič, Ja. M. et al. (1980) Sučasnaja belaruskaja literaturnaja mova. Marfalogija. Vyšejšaja škola. Dobrušina, Je. R., Je. A. Mellina and D. Pajard. (2001) Russkije pristavki: mnogoznačnost’ i semantičeskoje edinstvo: Sbornik . Moscow. Russkije slovari. Haspelmath, M. (2002) Understanding Morphology. Oxford University Press. London. Isačenko, Aleksandr V. (1960) Grammatičeskij stroj russkogo jazyka v sopostavlenii s slovackim. Morfologija II. Bratislava: Slovak Academy of Sciences. Janda, Laura A. (1986) A Semantic Analysis of the Russian Verbal Prefixes ZA-, PERE-, DO- and OT- (= Slavistische Beiträge, Band 192). Munich: Otto Sagner. Pugh, Stefan M. and Ian Press (1999) Ukraininan: A Comprehensive Grammar. Routledge Publishers. 11 Impoliteness and Mock Impolitenes s: A Descriptive Analysis Michael Furman, The Ohio State University

This paper offers an empirical analysis of mock-impoliteness in colloquial Russian conversation by examining the ways in which interactants produce and display an orientation to mock- impolite utterances. The corpus consists of video recorded naturally occurring Russian talk-in- interaction gathered from the Russian reality television show ‘Dom Dva’. Rather than adhering to a strict first order participant constructed (Lachenicht 1980, Culpeper 1996, 2005) or second order analyst constructed (Terkourafi 2005, 2008 Locher and Watts 2005, 2008) conception of impoliteness, the paper argues that one can be used to inform the other and utilizes both first and second order methodological approaches to classify impolite and mock-impolite utterances. The paper utilizes second order approaches to identify potentially impolite or mock- impolite utterances; yet, argues that the actual classification of impolite or mock-impolite utterances is a first order concept discursively constructed through the ways in which participants orient to a given utterance. The identification of impolite utterances consists of three separate but compatible approaches; Culpeper’s (2005) model of impoliteness, the presence of taboo or foul language within an utterance and those occasions where interactants publicly orient to an utterance as impolite through the usage of such words as grubij ‘rude’, or oskorbitel’nij ‘offensive’. The paper analyzes the specific ways mock-impolite turns are designed both linguistically (lexical items, syntax) and non-linguistically (prosody, laughter, pauses) and examines how mock impolite turns are discursively co-constructed by the interactants. Additionally, the paper provides an account of the effect of context on the production and perception of impolite utterances through the use of Levinson’s (1992) notion of activity types. Despite the increase in scholarship devoted to impoliteness, mock-impoliteness has received relatively little attention and by providing a detailed account of the design of mock- impolite utterances the current paper contributes to a largely unexplored phenomenon. 12 A Not So Iron Curtain between Activity and Knowledge Exchange: Grammatical Metaphor, Logic and Aspect in the Battle against Alcoholism Michael Furman, The Ohio State University

What exactly are the linguistic and semiotic tools that allowed for Soviet political posters to be effective in spreading propaganda and how were they used to build the ideological underpinnings of the Soviet regime? Using tools from Critical Discourse Analysis I present an analysis of pro- temperance Soviet Political posters (from 1920-1955). Central to the analysis of the posters is the distinction (or lack thereof) between two types of speech function: activity exchange (getting people to do things) and knowledge exchange (dissemination of information facts etc.) (Fairclough 2003). Yet, even those posters that foreground knowledge exchange also construct the very reality of which they purport to merely represent. The values selected are particular values that encourage the viewer to align themselves with the representational image and thereby bring the represented image into reality. In this way the posters are similar to Wernicke's notion of a promotional image (Wernicke 1991). Further, I argue that there is a functional concord between the visual and the textual elements of the text and that a full understanding of the posters requires the simultaneous analysis of both the textual and the visual elements. The example to the right illustrates the conflation between knowledge and activity exchange. The poster not only represents a negative portrayal of a drunken father, but advocates on the behalf of temperance and anticipates that the viewer will act temperate so as to avoid the negative consequences depicted in the poster. Here, knowledge is exchanged for the purpose of acting in a particular way. Interestingly, each poster exhibits a functional concord between visual and textual elements. Posters foregrounding P'janyj otec – gore knowledge-exchange utilize atemporal, 0-state constructs, while sem'i A drunken father is posters foregrounding activity exchanges utilize perfective 2-state grief to his family verbs (Klein 1995). 0 state constructs are not limited to a particular timeframe and are atemporal (e.g. a drunk father is grief to his family), while 2-state verbs have both a source state and a target state. Unlike the poster above, the poster to the right projects 2 states; one where alcoholism persists (source state) and another where alcoholism does not persist (target state). This is reflected both visually through the implied motion of the worker and textually through the use of the perfective dolbanem 'we will strike'. Utilizing both types of posters allows for the projection of not only an unassailable current world (depicted through the 0 state constructions) but also that of a projected better world (depicted through the 2 state constuctions). Dolbane By paying careful attention to how the visual and the textual elements m 'we will function together to project current and future worlds, this paper offers a strike' detailed understanding of the linguistic and visual tools that the Soviets used to both inform the citizens and prompt them to act. Further, by using tools from Critical Discourse Analysis to analyze historical, rather than contemporary social transitions this paper extends the domain of Critical Discourse Analysis and argues that this methodology can be a useful tool to analyze any social change for which there is significant textual documentation. 13 Slavic P-compounds as Non-canonical Adjectives Cammeron Girvin, Johanna Nichols, Elizabeth Purdy, Malgorzata Szajbel-Keck, UC Berkeley

;

;

;

Slavic languages rarely allow adjectives as resultatives (Spencer & Zaretskaya 1998), though the resultative function has been described as a defining criterion for adjectives as a part of speech (Baker 2003:219). There has been little work on the kind of form that regularly fulfills this function in Slavic: e.g. Russ. raskalit’ dobela ‘heat white-hot’, Pol. wymyć do czysta ‘wash clean’, Bulg. obrŭštam naopaki ‘turn upside-down’. We call this class of forms P-compounds (“bipartites” per Goeringer 1998) because they consist of two lexical roots, the first of which is etymologically a preposition. At first glance they look like frozen PP’s (e.g. Russ. dobela, Pol. do czysta ‘until clean’, though Slavic PP’s do not and historically never could have adjectives as complements), but on closer investigation their semantics is often non-compositional, their internal morphosyntax is unconnected to their syntactic function, and their morphology is often unmotivated (e.g. the acc. sg. fem. suffix in Russ. vkrutuju ‘hard-boiled’). Therefore we consider them compounds. (They are written as single words regularly in Russian and less regularly in the other languages.) So defined, P- compounds comprise a large set of forms attested throughout the history of Slavic. They sometimes have cognates in two or more daughter languages, and they serve a range of syntactic functions. In dictionaries they are usually classified as adverbs; in OCS handbooks they are described as either adverbs or uninflected adjectives depending on their attested syntactic functions. This paper reports on our ongoing survey and analysis of P-compounds in several Slavic languages (so far, Russian, Polish, Bulgarian, OCS). We have collected at least 30 sentences for a number of P-compounds in each modern language (44 different pcompounds), using national corpora or (for OCS) print resources. The preferred functions of the individual p-compounds depend primarily on their semantics. Though some P-compounds function as adverbs (see example 1), in our survey they more often resemble adjectives. Generalizing from Corbett 2004 on Russian adjectives, thepredicative (2), depictive (3), and (especially) noun-modifying functions (4) are those of adjectives; resultative (5) is an adjective function in many languages, though not often in Slavic. Adjuncts and verb modification are the two adverb-like functions of P-compounds, but they are distinctly less frequent than the canonically adjectival function of adnominal modifier. A P -compound in one language often corresponds to an adjective in another (e.g. Bulg. na staro, Pol. używany, Russ. poderžannyj for ‘used’ (6)). Also, P-compounds have several grammatical properties of words typically used as predicates: they can be specified for aktionsart (cf. Bulg. kriv ‘crooked’ with na krivo ‘askew’); the stress patterns of Pcompounds in Russian are like those of short adjectives and past-tense verbs. To conclude, P-compounds are property words with some verbal characteristics (aktionsart, stress) and function mostly as predicates; some are dedicated resultatives or depictives, and most were probably coined for depictive or resultative functions. But, where semantics permits, they can also have canonical adjective functions. Examples (1) Pol. Często jem na szybko. ‘I often eat in hurry.’ Russ. Xotja by načerno prosčitat’ vse sxemy ‘If only to roughly figure out all the designs’ Bulg. Izkazvam se dokraj. ‘I express myself completely.’ (2) Pol. Ta wycieczka jest na lekko. ‘This trip is a light-gear one.’ (lit. ‘is lightly equipped’) Russ. Glaza u nego byli na vykate i slezilis’. ‘His eyes were bulging and teary.’ 14 Bulg. Najasno sme s pravilata. ‘We’re clear on the rules.’ (3) Pol. Często podróżuję na lekko. ‘I often travel lightly equipped.’ Russ. Oni pošli nalegke, bez polušubov ‘They traveled light, without their coats.’ Bulg. Prekarvam noštta naštrek. ‘I spend the night on guard.’ (4) Pol. Jedzenie na wynos jest tańsze. ‘Take-out food is cheaper.’ Russ. Vnezapno končilos’ jajco vkrutuju. ‘They suddenly ran out of hard-boiled eggs.’ Bulg. Izpolzvam kakao na prax. ‘I use powdered cocoa.’ (5) Pol. Wytarłem stół do czysta. ‘I wiped the table clean.’ Russ. Rybki obgolodali trup dočista. ‘The fish gnawed the corpse clean.’ Bulg. Neštata se obŭrnaxa naopaki. ‘Things turned around.’ (6) Bulg. Kupix kola na staro. ‘I bought a used car.’ Pol. Kupiłem używany samochód. ‘I bought a used car.’ Russ. Ja kupil poderžannuju mašinu. ‘I bought a used car.’ References Baker, Mark C. 2003. Lexical Categories: Verbs, Nouns, and Adjectives. Cambridge: Cambridge University Press. Corbett, Greville G. 2004. The Russian adjective: a pervasive yet elusive category. In Adjective Classes: A Cross-Linguistic Typology, eds. R. M. W. Dixon and Alexandra Y. Aikhenvald. Oxford: Oxford University Press. 199-222. Goeringer, Keith Eliot. 1998. The Syntax and Semantics of Adverbs in Russian: Scope, Orientation, and Word Order. Ph.D. dissertation, UC Berkeley. Spencer, Andrew, and Marina Zaretskaya. 1998. Verb prefixation in Russian as lexical subordination. Linguistics 36: 1-39. 15 Some Problems with Inflectional Verbal Aspect Frank Gladney, The University of Illinois at Urbana-Champagne

Treating verbal aspect as inflectional entails positing ±Perfective as a feature of sentence predicates for which verbs are inflected, rather than as a feature for which verbs are specified in the lexicon. Aspect inflection is regular only with prefix-verb compounds. In P-V compounds verbs regularly show their unmarked, lexical shape in +Pfv contexts, e.g., za-pisa-, and a modified shape, e.g., za-pisyva-, in –Pfv contexts. Into this aspectual system, which originated in prefixed verbs as several layers of iterative thematization were grammaticalized, the integration of unprefixed verbs has been incomplete. While most verbs, like /pros/ ‘request’ and /klad/ ‘lay’, are regular, occurring unprefixed only in –Pfv sentences and requiring a prefix for +Pfv use, there are also verbs like /pad/ ‘fall’ and /bros/ ‘throw’, which occur unprefixed in +Pfv environments (padët, brósit). The problem is how to specify /pad/ and /bros/ in the lexicon. Not as +Pfv because they occur in both +Pfv and –Pfv forms. Perhaps as +Determinate. This feature figures prominently in diachronic studies of Slavic verbal aspect, but in synchronic descriptions of Slavic verbal morphology it has come to be restricted to oppositions among pairs of –Pfv verbs like polzët/polzáet ‘crawl’ and taščít/taskáet ‘pull’—the verbs of motion. Yet padët denotes the motion of an agent no less that polzët, and brósit denotes motion imparted to an object the same as taščít. Van Wijk (cited by Němec 1956:497) wrote that +Det verbs denote actions which are minimally complex and lead directly to a goal, whereas –Det verbs denote actions composed of plural acts or actions which are prolonged or repeated. This definition, as well as the morphology of the verbs in question, suggests that despite the distribution of + and –, it is the –Det forms that are marked (+Iterative), the +Det forms unmarked (–Iterative). This revised marking, perhaps less convincing for unprefixed verbs, is clearly seen in prefixed verbs, where výtaskaet and perebrosáet are +Iterative for výtaščit and perebrósit. But van Wijk’s claim that +Det verbs lead to a goal, i.e., are +Telic, is questionable. Russian +Det verbs, which denote running, leading, conveying, driving, riding, walking, rolling, climbing, flying, carrying, swimming, crawling, and dragging, denote unidirectional motion but none imply a goal; these are all atelic activities. This is the particular semantic quality which, as Dickey (2010:101) notes, prevents their integration into the aspectual opposition. On the other hand, falling (/pad/) and throwing (/bros/) are telic: falling culminates when the body hits the ground and throwing when the projectile is launched. ±Telic shoud be added to the feature set that determines aspect. Unprefixed verbs which are +Det and +Telic (/pad/, /bros/) occur unprefixed in +Pfv sentences; +Det, –Telic verbs (/polz/, /task/) do not. How this criterion accounts for aspectual usage must be tested in individual contexts: dal has the goal of transfer of possession, with lišíl it is termination of possession, with kupíl it is completion of transaction. But blagoslovíl and blagodaríl are problematic: why should blessing be +Telic, +Pfv and thanking –Telic, –Pfv? Biaspectual verb forms are not a problem: kazníli and organizováli are +Telic, +Pfv or –Telic, –Pfv according to whether the speaker assumes attainment of a goal. The +Det feature has diachronically been associated also with infixed and suffixed /n/. and /n/ verbs straddle the aspect divide the same as polzët/padët, occurring with both +Pfv forms (sjádet, prýgnet) and –Pfv (mólknet stýnet). The contrast is between semelfactive verbs denoting an action and inchoative verbs denoting a state. Thus /drog/ ‘jerk’ occurs in +Pfv drógnet, while /drog/ ‘chill’ occurs in –Pfv drógnet. A ±Telic contrast might be at play also here. References Dickey, Stephen M., “Common Slavic ‘indeterminate’ verbs of motion were really manner-of-motion verbs,” New Approaches to Slavic Verbs of Motion, Victoria Hasko and Renee Perelmutter, eds., Amsterdam: Benjamins, 2010 , pp. 67–109. Němec, Igor, “Kategorie determinovanosti a indeterminovanosti jako základ slovanské kategorie vidu”, Slavia 45 (1956), 496–534. 16 Morphological Adaptation of English Borrowings in Slovak Nina Haviernikova, The Ohio State University

In my paper I discuss the patterns of linguistic variation in the adaptation of English lexical borrowings to the Slovak declensional system. Analyzing data obtained from a survey, I examine how native Slovak speakers decline loanwords with stems ending in <l> and <r>, where variation can be observed, in particular, in the locative singular. Nouns of this type take the masculine inanimate (Mi) gender in Slovak based on the morpho-phonological criterion of consonant ending, which is perceived as the “strong” ending for masculine nouns. In my study only Mi are investigated. According to declensional rules in Standard Slovak, nouns with this stem ending, considered in Slovak as “hard” should take the locative singular ending <-e>, but some nouns take the ending <-i>, which is more common for a different phonological environment, specifically after a consonant classified in Slovak as “soft”. Actual usage shows considerable variation. In my study I am interested in how the geographical affiliation of the speakers affects the use of one of the variants. I use both quantitative and qualitative approaches to data analysis. The data was elicited from informants who are native speakers of Slovak, using a socio-linguistic questionnaire. The informants were asked to evaluate the acceptability of declined loanwords on a scale from 1 to 4 (with 1 being fully acceptable and 4 being unacceptable). The informants were also asked to evaluate native words with the same stem ending. Comparison of both sets of words is also presented in the paper. 17 Dialect/Standard Variation in Slovakia Nina Haviernikova, The Ohio State University

In my paper I examine switching between a dialect and the standard language in Slovak. The data is from the speech of inhabitants of Smolenice, a town of 3,500 located in western Slovakia. I am particularly interested how the patterns of variation of older generation speakers differ from those of younger generation. Code switching between dialects and/or languages has been investigated mainly after 1972, when a study on code switching in Norway was published by Jan Blom and John Gumperz. Other scholars who have studied this phenomenon include Carole Myers-Scotton (1993), Shanna Poplack (1993), Jeanine Traffers-Daller (1992), Penelope Gardner-Chloros (1985; 1991), Francois Grosjean (1982), and Peter Auer (1984; 1992), among others. Most studies that have been published, investigated switching between different languages. Code switching between a dialect and the standard in Slovakia has not, to my knowledge, been previously examined. In Smolenice, the local people mostly speak the western Slovak ‘Trnava’ dialect, especially in the family environment and with acquaintances. The local variety of standard is normally used in places such as school and church and in conversations with non- locals. Many inhabitants of Smolenice are becoming more educated, moving away from the town, and changing social circles; this may cause a shift from the dialect to standard and different patterns of variation. In my study I investigate code switching during a conversation with the same interlocutor. The triggers for code switching in which I am particularly interested, are the topic and presence of an overhearer. I am also interested in any cases where the dialect poses influence on the local variety of standard and whether code mixing occurs with any speakers. The data was collected from nine speakers from three different generations and recorded on a computer. To elicit speech, I have designed series of questions, concerning various topics, such as childhood experiences, religious events, school and education, describing simple tasks, work and employment, and Slovak and European politics and culture. Some of these topics are related to local environment and events and were to hypothetically elicit use of the dialect, particularly with the older generation. Other topics, especially related to Slovak and European politics, Slovak culture and media, were to potentially cause standard language use. I also hypothesize that the younger generation speakers may use the local standard variant of Slovak more often than the older generation based on the fact that most of the younger inhabitants of Smolenice have reached a higher level of education than the older speakers and many of them also commute out of the town for school and work. 18 A Circum-Pannonian Word-Prosodic Parallel: The Case of Doudleby Paroxytonic Accent Zbyněk Holub, Slezská univerzita v Opavě and Marc L. Greenberg, University of Kansas

;

Slavists generally think of the Czecho-Slovak dialect area as having a word-prosody configuration of fixed word-initial stress with contrastive length possible in any syllable. A notable exception is the Silesian dialect area, which typically has the Polish configuration with fixed penultimate stress and no contrastive length. Much less familiar is a small area in SW Bohemia (Doudleby and environs) with “paraoxytonic stress” and contrastive length, which at first blush may be considered akin to the Silesian configuration. Recent fieldwork expeditions in the Doudleby area (see Holub 2011) carried out by the authors of the paper have helped shed new light on the phenomenon. While superficially similar to the Silesian border phenomenon, the Doudleby type is characterized by length contrasts and also carries a distinctive word melody that is a hallmark of the local dialect. The profile of the Doudleby word melody matches a prosodic pattern that has also been observed in the Slovene dialect of the Zilja Valley (Gailtal, Austria) (Pronk 2009, Greenberg 2010, Holub 2011). This pattern in turn can be viewed as part of a wider areal pattern of word-prosodic configurations in Slavic that reflect differing outcomes of the inherited pitch-accent system of Common Slavic (Greenberg 2007). Whether this is a chance outcome or a function of a more recent migration of Slavic speakers remains to be determined. More important, however, is to take a closer look at the variety of word-prosodic types and understand their structural elements, which will help us to understand the full complexity of variation found in the Slavic languages. The Doudleby phenomenon, which will be discussed in this paper, will advance Slavists’ knowledge of this complexity. References Greenberg, Marc L. 2007. Phonetic Evidence for the Development of the ‘Acute’ Tone in Slavic. Mate Kapović and Ranko Matasović, eds. Tones and Theories: Proceedings from the International Workshop on Balto-Slavic Accentuation: 95–108. Zagreb. Greenberg, Marc L. 2010. Nova opisna slovnica ziljščine. Slavistična revija 58/4: 489-494. Holub, Zbyněk. 2011. K problematice současného úzu paroxytonického přízvuku v mluvě západních a jižních Čech. Slavia Centralis IV/2: 5–31. Pronk, Tijmen. 2009. The Slovene Dialect of Egg and Potschach in the Gailtal, Austria (= Studies in Slavic and General Linguistics, vol. 36). Amsterdam, New York: Rodopi. 19 Grammatical Profiles and the Aspect of Old Church Slavonic Verbs Laura A. Janda and Hanne Martine Eckhoff, Tromsø University

There is a long-standing debate on the status of aspect in Old Church Slavonic. Some scholars have claimed that aspect was nascent, but not fully grammaticalized at the time of OCS. Others have argued that aspect was already a full-fledged grammatical category in the oldest attested Slavic texts. Our key questions are: Was there an aspectual distinction between imperfective vs. perfective verbs in OCS? If so, how and to what extent was the aspectual distinction in OCS different from that in modern Russian? We take a fresh, objective approach to this debate. Instead of engaging in further polemics, we use quantitative methods to determine whether there was an aspectual distinction between verb lemmas in OCS. This study takes as its point of departure the fact that the aspectual distinction in modern Russian is associated with a difference between the grammatical profiles (frequency distributions of grammatical forms as attested in a corpus) of perfective and imperfective verbs. Janda & Lyashevskaya (2011) show that this difference is statistically significant. Thus in modern Russian we know that there is aspect, we know which verbs are of which aspect, and we know that aspect is correlated grammatical profiles. We start from an agnostic view of aspect in OCS: we do not know for certain whether there is an aspectual distinction, nor how all verbs should be classified. Our strategy is to run the analysis from Janda & Lyashevskaya (2011) in reverse for OCS and ask: Given the grammatical profiles of OCS verbs, can statistical models separate them into two groups, and do these groups resemble imperfective vs. perfective verbs? We use 15,683 attestations of verbs from the PROIEL corpus (http://foni.uio.no:3000/) to obtain the grammatical profiles of OCS verbs. For example, the grammatical profiles of the verbs tvoriti ‘make’ and jęti ‘take’ are attested in our database as shown in Table 1 (with both raw and relative frequencies): aorist imperative imperfect infinitive/supine present past present tTotal participle participle tvoriti 0 14 12 23 99 0 26 174 ‘make’ 0% 8% 7% 13% 57% 0% 15% 100% jęti 25 7 0 10 28 20 0 90 ‘take’ 28% 8% 0% 11% 31% 22% 0% 100% Table 1: Sample grammatical profiles for tvoriti ‘make’ and jęti ‘take’ We apply two statistical methods to sort the verbs: a correspondence analysis and a divisive clustering analysis. Both methods yield a division of the verbs into two groups, and, remarkably, the results are nearly identical. These results also correspond 96% to the aspectual designations of OCS verbs posited by Dostál 1954. While our results support the hypothesis that there was aspect (or an aspect-like distinction) in OCS, there are “mismatches” and other tangible differences between the situation in OCS and in Russian. References Dostál, Antonín. 1954. Studie o vidovém systému v staroslověnštině. Prague: Státní pedagogické nakladatelství. Janda, L. A. and O. Lyashevskaya. 2011. “Grammatical profiles and the interaction of the lexicon with aspect, tense and mood in Russian”, co-authored with Olga Lyashevskaya. Cognitive Linguistics 22:4, 719-763. 20 Influence of Speech Rate on the Rhythmic Structure of Polish and Russian Sylwester Jaworski, Szczecin University

Most human activities, such as walking, breathing, running or heart beat, are rhythmical. Rhythm can be defined as “the pattern of intervals between movements, or between their beginnings or peaks, or the pattern of durations of movements” (Allen 1973: 97). When rhythm is taken as a criterion, the world’s languages are divided into stress-timed (e.g. English, Russian, German, Arabic), syllable-timed (e.g. French, Spanish, Italian) and mora-timed ones (e.g. Japanese). The most distinguishing characteristic of stress-timed languages is the relative regularity with which stressed syllables occur in utterances regardless of how many unaccented syllables separate them. The foot, i.e. the amount of time taken between each two stressed syllables, is the unit of rhythm in stress-timed languages. By contrast, the rhythmic organisation of syllable-timed languages is based on syllable isochrony which, by analogy with the foot, is thought to take the same amount of time irrespective of its structure. However, in languages, such as Polish, that have very complex clusters on the one hand and are not stress-timed on the other, syllable- timeness appears to be a purely theoretical concept as one can easily find a large number of monosyllabic words, e.g. kot [kɔ t] 'cat' vs. chrząszcz [xʃ ɔ nʃ ʧ ] 'beetle' that differ in duration due to considerable structural differences. Therefore, languages that have the properties of syllable- timed and stresstimed languages are referred to as mixed-type languages. According to Waniek- Klimczak (2005: 32), “[c]ross-linguistically, we can expect the same articulatory timing to be preferred on the basis of a shared articulatory mechanism”. The author further argues that stress- timing is the prefered rhythmic organisation of language and, therefore, in fast speech the stress- timing effect should be observable even in syllable-timed languages. The present paper reports the results of an experiment whose primary objective was to determine to what extent the rhythmic stucture of Polish and Russian is influenced by speech rate. In order to establish it, five native speakers of the two languages were asked to read ten syntactically matched sentences in two different tempos: natural and fast. Since Russian and Polish represent different rhythm classes, interstress intervals (henceforth ISIs), i.e. the time measured from the onset of one stressed vowel to the onset of the next, were used as units of rhythm as suggested by Dauer (1983: 54). The results show that in both cases the stress-timing effect takes place in the sense that Polish rhythm does become more stress-timed, whereas in Russian the ISIs differ to a lesser extent. The conclusion drawn from the data is in keeping with that of Dauer (1983) who measured ISIs in several stress and syllable time languages and did not find significant differences between the two groups. References Allen, Sidney W. 1973. Accent and rhythm. Prosodic features of Latin and Greek: A study in theory and reconstruction. Cambridge: Cambridge University Press. Dauer, Rebbeca, M. 1983. “Stress -timing and syllable-timing reanalyzed”, Journal of Phonetics 11: 51- 62. Waniek-Klimczak, Ewa. 2005. Temporal parameters in second language speech . Łódź: Wydawnictwo Uniwersytetu Łódzkiego. 21 On Embedded Non-Finite Complement Patterns in Old Polish Łukasz Jędrzejowski, Zentrum für Allgemeine Sprachwissenschaft

The aim of this talk is twofold. First, I will briefly re-examine clause-embedding predicates selecting non-finite clause arguments in Modern Polish (= MP) and propose a new classification of control/raising patterns. Second, it will be shown to what extent such a re-classification is related to and motivated by the Old Polish (= OP) non-finite complementation system. To begin with, I will explore the most frequent embedded pattern of OP: modal verbs. According to Hansen (2001), this group contains the following MP predicates: móc, umieć, potrafić (‘can’), musieć (‘must’), chcieć (‘want’), mieć (lit. ‘have’, ‘have to’/‘be supposed to’), (po-)winien (‘should’) (cf. 1). Additionally, it has been assumed that some of these modal verbs have undergone a grammaticalization process over time and have acquired a more subjective, i.e., an epistemic or evidential meaning. Following the modal approach advocated by Kratzer (1981, 2012) and the collected OP data, I will show that (i) umieć, potrafić (‘can’) and chcieć (‘want’) are not modal verbs but subject control predicates like, for example, próbować (‘try’), (ii) there is no significant evidence for grammaticalization of modals since their epistemic/evidential use can be attested in OP. As pointed out by Šimík (2011), Polish also possesses hidden patterns of modality. One of them pertains to so called modal existential wh-constructions (=MECs), which in MP can be embedded under the possessive predicate mieć (‘to have’) (cf. 2). The pattern in question consists of the matrix predicate mieć, a wh-phrase, and an embedded infinitive. In this connection, I will present data indicating that MECs were present in OP. OP MECs, however, differ from the MP MECs to the extent that the former embed l-participles which are inflected for number and gender, and do not embed infinitives. I will briefly examine how this difference affects the control configuration of OP MECs. Finally, I analyze subject and object matrix predicates licensing primarily infinitival complements in MP. Combining the classifications proposed in Bondaruk (2004) and Słodowicz (2008) and comparing them with the OP data, I will show that (i) OP non-finite indirect questions resemble OP MECs mainly formed with l-participles, and not with infinitives, (ii) diachronically Polish does use infinitives (contra Dziwirek 2001’s approach to MP). As for (ii), main evidence will come from different predicate groups: phase predicates (e.g. jąć ‘begin’), volitional predicates (e.g. chcieć ‘want’), complex modal predicates (e.g. być trudno ‘be difficult’), causative/manipulative predicates (e.g. kazać ‘command’/‘cause’), etc. Accordingly, I argue that some ill-formed infinitival constructions in Polish which are grammatical in English and extensively discussed in Dziwirek (2001) are traceable to the diachronic competition l-participle vs. infinitive in Polish that by contrast has not taken place in the history of English. What it means is that Polish and English are different from one another and not that Polish doesn’t use/like infinitives. (1) Kamila może/umie/potrafi/musi/chce/ma/(po-)winna grać na gitarze K.NOM may,might,can/can/must/wants/has,is said/should play.INF on guitar.LOC ‘Kamila is allowed/can/has/wants/should (to) play the guitar.’ ‘Kamila may/might/must/should be playing the guitar.’ (2) Andrzej ma gdzie spać A.NOM has where sleep.INF ‘There is a place where Andrzej can sleep.’ 22 References Bondaruk, Anna (2004): PRO and control in English, Irish and Polish: A Minimalist Analysis. Lublin: Wydawnictwo KUL. München: Otto Sagner. Dziwirek, Katarzyna (2001): Why Polish doesn’t like infinitives, in: Journal of Slavic Linguistics 18: 57- 82. Hansen, Björn. (2001): Das Modalauxiliar im Slavischen. Grammatikalisierung und Semantik im Russischen, Polnischen, Serbischen/Kroatischen und Altkirchenslavischen. Kratzer, Angelika (1981): The notional category of modality, in: Words, Worlds, and Contexts. New Approaches in Word Semantics ed. by Hans-Jürgen Eikemeyer & Hannes Rieser. Berlin: de Gruyter, pp. 38-74. _____. (2012): Modals and Conditionals. New and Revised Perspectives. Oxford: Oxford University Press. Šimík, Radek (2011): Modal existential wh-constructions. Ph.D. thesis, Rijksuniversiteit Groningen. Słodowicz, Szymon (2008): Control in Polish Complement Clauses. München: Otto Sagner. 23 Verbal Aspect in Old Church Slavonic Jaap Kamphuis, Leiden University

Since the comprehensive study of Dostál (1954) on verbal aspect in Old Church Slavonic (OCS), the study of verbal aspect has developed considerably. Verbal aspect is no longer considered to be the same in all Slavic languages (cf. Dickey 2000) and verbal aspect in e.g. Old Russian has been shown to differ considerably from the verbal aspect in Modern Russian (Bermel 1997). These insights make a new study of verbal aspect in OCS necessary, using modern technological tools for empirical research. The starting point of our study is not to classify each OCS verb as either imperfective or perfective, but to go one step back by addressing the question whether there is verbal aspect in OCS at all, and if so, what its different functions are. As a tool for our research a database has been compiled in which all verbal lexemes and their attestations in OCS are included. Although research into an ancient language like OCS has to be conducted without informants, the database has been set up to act like one. This results in four 'informants', each giving their version of the same text. Our four 'informants' are the main OCS gospel codices: Marianus (M), Zographensis (Z), Assemanianus (A) and Savvina kniga (Sk). These texts have been integrated into a parallel corpus, making comparison between them possible. The Greek original has been included as well. This enables us to ask our 'informants' questions of the type: "How would you translate this verb in your language ...?" All forms, both OCS and Greek, from all four gospel texts are fully annotated, so that even the slightest grammatical or contextual differences can be detected. My presentation at the SLS conference will consist of a demonstration of the database and a discussion of some salient examples. One such example is the following general statement where the Zographensis uses a derived form (traditionally labeled 'imperfective), while the Marianus uses a simple prefixed form (traditionally labeled 'perfective'): MNi vь zemi ni vъ gnoi trěbě estъ . vonъ isypljǫtъ jǫ . (Luke 14:35) Z Ni vъ zemi ni vъ gnoi . trěbě estъ . vъnъ isypajetъ sę . ENIt is useless either for the soil or for the manure pile; it is thrown out. There are also cases in which the Greek original and the OCS translations show considerable difference, like in the case of the Greek historical, which in OCS is normally rendered with an aorist: GRΤότε λέγει αὐτῷ ὁ Ἰησοῦς (Matth 4:10) Ztogda gl[agol]a emu i[su]s[ъ] Ass/Sktogda reče emu i[su]sъ ENThen Jesus said to him Dickey (2000) uses the historical present in the modern Slavic languages as one of the parameters for his typology, so the relative absence of it in OCS has to be taken into account in the study of the development of OCS verbal aspect. In my paper I will address the benefits as well as the limitations of a sophisticated database which includes a wide variety of parameters which are crucial for answering the question about the specific nature of verbal aspect in OCS. References Dickey, S.M. (2000) Parameters of Slavic Aspect. Stanford. Bermel , N. (1997) Context and the Lexicon in the Development of Russian Aspect . Linguistics Volume 129. University of California Press: Berkely, Los Angeles, London. Dostál, A. (1954) Studie o vidovém systému v staroslovĕnštine. Prague. 24 Language Attitudes toward Surzhyk in Ukraine Kateryna Kent, University of Minnesota

Centuries of colonization by the Russian Empire and political, cultural, and economic connections in the Soviet and Post-Soviet eras created favorable conditions for Russian-Ukrainian language contact in Ukraine. The product of such millennia-long contact, linguistic variety consisting of Ukrainian grammar with admixture of Russian morphemes, is known as Surzhyk. Current public discourse associates Surzhyk with parochialism, lack of education, and a low culture (Bilaniuk, 2004). The Ukrainian government does not recognize Surzhyk as a separate spoken language. Surzhyk speakers do not receive any recognition or support as a distinct linguistic group. The purpose of this paper is to measure language attitudes toward Surzhyk and to compare the place of Surzhyk with respect to the Ukrainian and Russian languages. A few studies on language attitudes conducted in Ukraine focused mainly on Ukrainian, Russian, and English (Bilaniuk, 2005; Goodman & Lyulkum, 2010). There is no research on attitudes toward Surzhyk. To measure attitudes a survey using the method of semantic differential was conducted in Zaporizhzhia Region in Eastern Ukraine. The sample consisted of 168 (N=168) students attending three educational establishments: a university located in a large city, a technical vocational school located in a mid-size city and an agricultural vocational school located in a small town. The survey asked about the participants’ native language, their use of Ukrainian, Russian, and Surzhyk in various domains, and the Semantic Differential Scale (Osgood et al., 1957), which elicited their attitudes toward three concepts: UKRAINIAN, RUSSIAN, and SURZHYK. The scale consisted of 10 bipolar adjectives (the same for each concept) placed at opposite ends of a four-point continuum. The analysis of the survey aimed to corroborate or disprove the following hypotheses: 1) Students from a large city have more negative attitudes toward Surzhyk, since Surzhyk is more common in semi-urban and rural areas 2) Mean score for concept SURZHYK is predicted by one’s native language, language spoken in the family, and the institution attended An analysis of variance showed that the effect of city size was significant for Surzhyk mean score, F(2, 177) = 6.28, p=.002.1 Post hoc analysis using the Bonferroni post hoc criterion for significance indicated that the Surzhyk mean score was significantly lower for students from the mid-size city and a small town than for those from a large city. Multiple regression analysis was used to test if one’s native language, language spoken in the family, and the institution attended predicted participant’s score of Surzhyk (F(3, 174)=6.0, p=.001). Language spoken in the family (β=.215, p=.011), and the institution attended (β=.176, p=.002). Native language of the participants was not a significant predictor. The results also showed that people who self- identified as Surzhyk speakers ranked Surzhyk significantly higher (M=2, SD=.85), than those who reported not speaking Surzhyk (M=1.7, SD=.62), t(23)= -2.3, p=.022. References Bilaniuk, L. (2004). A typology of Surzhyk: Mixed Ukrainian-Russian language. International Journal of Bilingualism, 8(4), 409-425. Bilaniuk, L. (2005). Contested tongues: Language politics and cultural correction in Ukraine . Ithaca: Cornell University Press. Goodman, B. A., & Lyulkun, N.A. (2010). Ukrainian, Russian, English: Language Use and Attitudes of Students at a Ukrainian University. Working Papers in Educational Linguistics, 25 (1), 77-93. Osgood, C. E., Suci, G.J., & Tannenbaum, P.H. (1957). The Measurement of Meaning. Urbana: University of Illinois Press. 1 Notably, the effect of city size was not significant for the Ukrainian mean score F(2, 177) = 1.10, p>.05 and for the Russian mean score F(2, 177) = 1.54, p>.05. 25 Interference of Korean on Constructing Czech Sentences (based on teaching Czech at HUFS) Inchon Kim, Hangkuk University of Foreign Studies

Learning complicated and complex Czech syntactic rules presents one of the most serious problems and biggest challenges for Korean students when they start to acquire language so completely different as Czech is from Korean. Providing a short introductory overview of syntactic structure of (standard and neutral) Korean sentences, the author of the paper considers and discusses the possible interference of processes and stereotypes in constructing grammatically correct Korean sentences brought over to Czech sentences (e.g. the final position of the verb in Korean sentence – unlike in Czech where the word-order is basically free). Authentic students’ utterances from Czech grammar classes (esp. for beginners and intermediate levels) at HUFS will be used to illustrate situation and results of studying Czech based on Korean. As a many-year experience of the author shows the source language interference seems the strongest on the syntactic level. Apart from such system interference, some other factors occur as reasons of making mistakes in Korean students’ Czech sentences: their strong habit to memorize grammar and its patterns and paradigms, effort to translate word after word following Korean sentence structure and no understanding of the function of grammar in the text. Korean students learn the system (isolated knowledge of langue), but they lack experience and practice in speech (parole). 26 The Actional Imperfective in Russian and Czech: A Comparative Approach John Korba, University of Kansas

Among the large body of literature on Slavic aspect, several seminal works in recent years (Glovinskaja 2001; Grønn 2004; Padučeva 1996; Šatunovskij 2009) have focused, in whole or in part, on the importance of the imperfective general-factual (IGF). The body of works that consider aspect within a comparative Slavic tradition are rarer (cf. Petruxina 1978, Dickey 2000) and, in any case, have not considered in full the aforementioned studies. This paper begins to explore the issues examined in Russian in the former works by examining the Actional Imperfective in Czech using the approach of the latter works. Padučeva (1996) offers what is arguably the most fine-grained typology of the IGF in Russian. In her study, Padučeva compares the Existential IGF type in (1) and Concrete IGF in (2) with the Actional Imperfective (Impf) in (3). (1) Moj djadja vosxodili na Everest (Padučeva 1996: 43) ‘My uncle (has) climbed Mount Everest’ (2) Ty otkryvali okno? (Padučeva 1996: 46) ‘Did you open the window? (The window is now closed)’ (3) Gdje apel'siny pokupali ? (Padučeva 1996: 49) ‘Where did you by (these/those) oranges?’ Rassudova (1968: 38) notes that the Actional Impf is used when “the attention of the speaker is direct at where, when, why, and who completed an action.” Some scholars classify the former Actional Impf as IGF (Bondarko 1971, Grønn 2004) while others do not (Rassudova 1968). Padučeva (1996) perhaps offers the most fine-grained exploration of the nature of the Actional Impf in relation to the Resultative (Existential and Concrete) types of IGF. She determines the semantic notions communicated by the use of each type, as summarized in a chart below. Existential IGF Concrete IGF Actional Impf i. Articulation of   factuality/existentiality ii. Resultativity (reaching    some limit) iii. Disassociation from the   moment of speech iv. Vagueness concerning    the time when the action was/is completed v. Multiplicity  (кратность) vi. A retrospective    reference point Table 1. Padučeva’s (1996: 32–52) analysis of the Existential and Concrete types of IGF and Actional Impf according to semantic features My analysis expands upon both types of works by comparing the semantic and pragmatic factors involved in Russian to those relevant to Czech. While the Existential IGF is attested with accomplishment predicates in Czech (Dickey 2000: 125), the Concrete GF is not (cf. also Cummins 1987). However, the Actional Impf is well attested in Czech, as in (4). 27 (4) Ahoj Jardo, mám na vagu závadu Hallova snímače polohy a auto moc netáhne… Mám stejnej motor, kde jsi kupovali ten snímač??? (http://forum.skodahome.cz/topic/13403-vrakoviste-skoda/) ‘Hello Jardo, I have a problem in my Volkswagon with the position of the Hall effect sensor and my car barely runs… I have the same motor, where did you buy that sensor?’ The speaker in (4) is referring to a completed event, but the speaker’s focus is more directed to the question of where the interlocutor purchased the sensor rather than on the result. Using the data gathered from the internet as well as from parallel corpora like ParaSol (http://parasol.unibe.ch), I discuss comparative examples of the Actional Impf and consider the semantic and pragmatic conditions that motivate its use in further detail. References Bondarko, A. V. 1971. Vid i vremja russkogo glagola (značenie i upotreblenie). Moscow: Prosvešenie. Cummins, George M. III. 1987. “Preliminary Notes on the Constantive Imperfective in Czech.” International journal of Slavic linguistics and poetics 35–36: 29–50. Dickey, Stephen M. 2000. Parameters of Slavic Aspect: A Cognitive Approach. Stanford, Calif., Center for the Study of Language and Information. Glovinskaja, Marina J. 2001. Mnogoznacnost' i sinonimija v vido-vremennoj sisteme russkogo glagola. Moskva: Azbukovnik. Grønn, Alte. 2004. The Semantics and Pragmatics of the Russian Factual Imperfective . Doctor Artium Thesis. University of Oslo. Padučeva, Elena V. 1996. Semantičeskie issledovanija: semantika vremeni i vida v russkom jazyke, semantika narrativa. Moskva: Škola Jazyki Russkoj Kul'tury. Petruxina, E. V. 1978. “O funkcionirovanii vidogo protivopostavlenija v russkom jazyke v sopostavlenii s češskim (pri oboznačenii povtorjajuščixsja dejstvij).” Russkij jazyk za rubežom 1: 57–60. Rassudova, O.P. 1982. Upotreblenie vidov v russkom jazyke. Moscow: Izdatel'stvo Moskovskogo universiteta. 28 Aspectual Features of Light Verb Constructions in Russian Julia Kuznetsova, Tromsø University

This paper investigates a well-known topic among Slavic studies – aspect – applied to a less well- known area – light verbs in Russian. The term “light verb”, usually ascribed to Jespersen (1965, Volume VI: 117), was introduced in reference to English constructions such as have a rest, take a drive, give a sigh. Even though they follow standard grammatical structure for the verbs have, take or give, from the semantic point of view, these constructions cannot be seen as full predicates. A light verb only provides verbal support for a noun, which carries the semantic weight. Since Jespersen’s establishment of the term, light verbs have been studied in various languages, including Japanese, Hindi, Urdu and Romance languages (see Butt 2003 and references wherein). In Russian, studies have identified the following light verbs: brat’ ‘take’, vesti ‘lead’, davat’ ‘give’, delat’ ‘do’, imet’ ‘have’, nesti ‘carry’, položit’ ‘put (down)’, prinimat’ ‘take/get’, stavit’ ‘put (up)’ (Mudraya et al. 2008, Babych et al. 2009). These verbs serve as light verbs in contexts such as 0 and 0: Irina ne mogla vesti urok. (RNC1) ‘Irina was not able to give a lesson.’ Esli zakon ne vypolnjaetsja, činovnik objazan nesti nakazanie. (RNC) ‘If the law is not obeyed, then an official must be punished.’ One might argue that the phenomenon of light verbs is irrelevant to aspect, since the verbs do not show truly verbal characteristics in such examples. However, our research shows that both imperfective and perfective counterparts of each light verb appear in light verb constructions. Although some nouns are compatible with both aspectual variants (vesti/provesti urok ‘give a lesson’), most nouns in the light verb construction have strong preference for either an imperfective or a perfective light verb, see vesti/?? provesti torgovlju ‘trade’, vesti/?? provesti rasskaz ‘tell a story’ vs. provesti/?? vesti reformy ‘carry out reforms’, provesti/# vesti parad ‘parade’. This strong preference shows us that aspect in light verb constructions is defined by the noun. This is an unusual situation for Russian, since it is often claimed that “[i]n Slavic, the direct object does not influence the (a)telicity of the VP” (MacDonald in print:2, see also Schoorlemmer 1995). In this paper I explore the properties of the nouns that lead to a preference for either an imperfective or a perfective light verb. Analysis of aspectual preferences of the light verb constructions can serve as an independent source of information about the internal aspectual structure of nouns describing situations, a topic that has long interested Slavists (see for example Pazelskaya & Tatevosov 2006). References Babych, Bogdan, Anthony Hartley & Serge Sharoff. 2009. Evaluation -guided pre-editing of source text: improving MT-tractability of light verb constructions. Proceedings of the 13th Annual Conference of the EAMT, pages 36–43, Barcelona, May 2009. Butt, Miriam. 2003. The Light Verb Jungle. Harvard Working Papers in Linguistics. Volume 9. Papers from the Harvard/Dudley House Light Verb Workshop 2003. Jespersen, Otto. 1965. A Modern English Grammar on Historical Principles, Part VI, Morphology. London: George Allen and Unwin Ltd. Mudraya, Olga, Scott S. L. Piao, Paul Rayson, Serge Sharoff, Bogdan Babych, and Laura Lofberg. 2008. Automatic extraction of translation equivalents of phrasal and light verbs in English and Russian. In S. Granger and F. Meunier, editors, Phraseology: an interdisciplinary perspective, pages 293–309. John Benjamins. 1 Examples are taken from the Russian National Corpus (www.ruscorpora.ru). 29 MacDonald, Jonathan E. In press. Nouns affect aspect syntactically. For proceedings of 29th West Coast Conference on Formal Linguistics, University of Arizona. Available at: https://netfiles.uiuc.edu/jonmacd/www/MacDonald_files/MacDonald%20- %20Nouns%20affect%20aspect%20syntactically.pdf Pazelskaya, Anna and Sergei Tatevosov. 2006. “Deverbal nouns, uninflected VPs, and aspectual architecture of Russian”. In Formal Approaches to Slavic Linguistics 14. The Princeton Meeting 2005. Ann Arbor: Michigan Slavic Publications, pp. 258-276. Schoorlemmer, Maaike 1995. Participial passives and aspect in Russian. Ph.D. Dissertation, Utrecht University. 30 An Argument against Configurational Case Assignment James E. Lavine, Bucknell University

Background. The “configurational” or “global” strategy for structural case assignment states that structural case is assigned by reference to the properties of both structurally-case-marked arguments taken together. Accusative is assigned, for example, in contrast to a higher Nominative-marked NP. This mechanism for structural case assignment can be traced back (at least) to Marantz 1991 (note also Yip et al. 1987). We find its most current manifestation in the constraint-based proposals of Woolford 2003 and Malchukov & de Hoop 2011. In all such approaches, coding is subject to the functional notion of “distinguishability” as the relevant factor in case assignment. This provides a straightforward account for the distribution of nominative objects: nominative occurs on the object, as the unmarked case, if there is no higher, structurally- marked argument in contrast to which the object must be identified. So on the “global calculation” nominative objects should appear naturally in constructions with quirky subjects. Individual functional heads on the global strategy play no role in realizing structural case. Proposal. In point of fact, nominative objects are often “costly” and unstable. A novel set of data is introduced from North Russian and standard (East High) Lithuanian impersonals with oblique subjects, in which nominative on the object should appear normally on configurational accounts, but fails syntactically, since nominative on the object is not related to any feature of Tense, which, on a local theory of case assignment (Chomsky 2001), is taken to be responsible for nominative on the object, as in the well known case of the oblique subject–nominative object configuration in Icelandic. While the nominative object in Icelandic agrees with Tense for number (but not person), there is no agreement whatsoever in the North Russian and Lithuanian impersonals below. It is precisely here that we can test the global strategy against one that relies on local checking of formal features (cf. Baker & Vinokurova 2010, in which both global and local strategies are argued to operate in Sakha). On the global strategy, the oblique subject in the North Russian and Lithuanian impersonals sets up a configuration in which unmarked (non- dependent) nominative on the object is the least costly case. This is because the global calculation rests on distinguishability rather than feature checking. However, it is shown that while nominative indeed occurs on the object of certain North Russian and Lithuanian constructions, as in (1–2), it is dispreferred: accusative appears on the object instead in much greater frequency in the North Russian Perfect (Kuz’mina & Nemčenko 1971), while in Lithuanian the impersonal (“Inferential Evidential”) is virtually limited to intransitives (Holvoet 2007), even though the evidential is in no way incompatible with two-place, agentive predicates. The object position in such cases is projected, but not case marked in the syntax. The “variable effability” of the nominative object (Baker 2011), marked by Accusative replacement in North Russian and a transitivity restriction in Lithuanian, is entirely unpredicted by the configurational rules of the global strategy. (1) a. North Russian Perfect: Nom Objectb.North Russian Perfect: Acc Object U nix byl postavlen konjušnja. U bat’ki u tvoego saženo berezku. at them:GEN was built stable:NO M at father at yours planted birch:ACC ‘They had built a stable.’ Your father planted a birch tree.’ [Kuz’mina & Nemčenko 1971: 79] [Kuz’mina & Nemčenko 1971: 38] (2) a. Lith. Inferential Evidential: Transitiveb.Lith. Inferential Evidential: Intransitive Ingos nuraminta vaikas. Mano užmigta. Inga:GEN calmed-down child:NOM me:GEN fell-asleep ‘Inga must have calmed the child down.’ ‘I must have dozed off’ [Holvoet 2007] 31 The Balkan Slavic Definite Article: What Role is Left for the Sprachbund? Jouko Lindstedt, University of Helsinki

The enclitic definite article has been regarded as one of the central Balkanisms ever since the first studies on the Balkan Sprachbund, and it is considered to be one of the most striking areal features distinguishing Balkan Slavic from other Slavic languages (Asenova 2002: 123). On the other hand, Mladenova (2007) has shown that the rise of the Bulgarian definite article can reasonably well be accounted for as a mainly language-internal development, perhaps with some Pre-Ottoman Turkic influence, making comparisons with the Albanian and Balkan Romance systems (as in Stölting 1970 and other studies) seem misguided. The grammatical category of definiteness was not new for Balkan Slavic, since the new enclitic articles came into being before the old system of definite (long) adjectives had disappeared completely. Thus, we can observe two partly overlapping definite article cycles (in the sense of Lyons 1999: 326–330) from Proto- Slavic to Balkan Slavic. The purpose of my paper is to contribute to the chronology of the rise of the enclitic definite article in Balkan Slavic and to defend the view that this diachronic process must, after all, be explained in the context of the Balkan linguistic area. As for chronology, I will try to explain, for example, the different outcomes of the jers in Bulg. dnes ‘today’ < dьnь sь and deńǝt ‘the day’ < dьnь tъ. My four main arguments for the importance of the areal aspect are as follows: 1. If a language is known to have been in intense contact with several other languages, the burden of proof cannot always be laid only on those proposing a contact-induced change rather than a language-internal explanation; in other words, we cannot let methodology override ontology. 2. If, for the sake of argument, we assume that Balkan Slavic acquired an enclitic definite article by internal development only, the enclitic articles in Albanian and Balkan Romance must then have been strongly influenced by Slavic, for coincidental parallel development in these neighboring languages is too improbable. This means that contact-induced language change needs still to be accounted for, though in the reverse direction. 3. Balkan Slavic has certainly developed the category of definiteness also for internal reasons, but in the Balkan feature pool (in the sense of Mufwene 2001: 4–6; 2008: 117–132) definiteness marking had several sources, and the other languages may have mutually reinforced the development in Slavic itself (Asenova 2002: 126; cf. Lindstedt 2000). 4. The multilingual communication situation in the Balkans seems to have acted as a selective environment which favored overt marking of the information structure of the sentence. Intense language contact forms a selective environment for seemingly language-internal changes even without outright structural borrowing. The challenge is not, as Mladenova (2007: 347) maintains, to explain how a language without overt marking of definiteness can borrow it from a language that has it, but rather to explain how two or more contacting languages with partially similar systems of demonstratives can develop partially similar systems of articles through a shared drift. References Asenova, Petja. 2002. Balkansko ezikoznanie. (2nd edn.) Veliko Tărnovo: Faber. Lindstedt, Jouko. 2000. Linguistic Balkanization: Contact-induced change by mutual reinforcement. In: Dicky Gilbers, John Nerbonne, and Jos Schaeken (eds.), Languages in Contact (Studies in Slavic and General Linguistics 28), 231–246. Amsterdam—Atlanta, GA: Rodopi. Lyons, Christopher. 1999. Definiteness (Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press. Mladenova, Olga M. 2007. Definiteness in Bulgarian: Modelling the processes of language change (Trends in Linguistics, Studies and Monographs 182). Berlin—New York: Mouton de Gruyter. 32 Mufwene, Salikoko S. 2001. The Ecology of Language Evolution. Cambridge: Cambridge University Press. Mufwene, Salikoko S. 2008. Language Evolution: Contact, Competition and Change. London—New York: Continuum. Stölting, Wilfried. 1970. Beiträge zur Geschichte des Artikels im Bulgarischen (Slavistische Beiträge 44). München: Otto Sagner. 33 The Belarusian ‘Classical Standard’ and Its Contemporary Use Lidia Federica Mazzitelli, Universita Roma Sapienza

Contemporary Belarusian has two standardized variants: the so-called taraškevica, codified in 1920 by Branislaŭ Taraškevič (hence its name), and revised for the last time in 2005, and the so- called narkamaŭka, codified by a special Commission of the Soviet Department of Education (Narodny Kamisaryjat Asvety) in 1933, and revised in 1957 and 2008. Though its current status of official standard in Belarus’, the codification of the narkamaŭka represented, in 1933, an open attempt to russify Belarusian (Mayo1975, 1978; Bieder 2000:653), on the orthographic, morphosyntactic and lexical level: some case endings were ‘aligned’ with the Russian ones (the Gen.Sg. of feminine adjectives in –aja is –ae in taraškevica and –aj, like in Russian, in narkamaŭka), as well as the case taken by some prepositions (for instance, pa takes the Locative case in taraškevica and the Dative case in Singular and the Locative in Plural in narkamaŭka); Russian loanwords, phonetically adapted, were introduced instead of existing Belarusian words, etc. Nowadays, after a period of debate about which one of the two standards was the most suitable to become the official one (Sjameška 1995), the taraškevica (who had been used by Belarusian writer in the emigration since 1918 until 1991, as in Soviet Belarus the narkamaŭka was used) is not allowed in public acts and publications on the territory of Belarus’: its use is limited to on-line resources (mostly newspapers or single authors who publish on the web) and to publishing houses operating abroad (Lithuanian, Czech Republic). Its use, however, has known a certain revival in the last years, especially among young people, who, being formerly Russian- speaking, decided to become Belarusian-speaking. As it is known, in fact, Belarusian is far not the dominant language in both everyday practice and public life: both these roles are fulfilled by Russian. According to recent data, just 23% of the population habitually speaks Belarusian at home and only 19% of the pupils (both in primary and high school) is taught in this language. The sociolinguistic situation in present-day Belarus’ cannot be thus described either as ‘diglossy’ or ‘bilingualism’ (Gutschmidt 2000:77): rather, most Belarusians have just a passive knowledge of Belarusian (as it is obligatorily taught in schools) and there are no communicative situations or given sphere of use which would require an active knowledge of it (ib.). In the last years a certain ‘revival’ of Belarusian can be observed, namely among the youngest generation (usually, the decision of beginning to use Belarusian in everyday life is accompanied by a political engagement), and it has its privileged expression in the Internet. A consistent number of young people have decided, in the last years, to begin speaking Belarusian in everyday life (of course, theirs is a somehow ‘artificial’ language, as it has been normally acquired at school, and not by family practice). Some of them, moreover, choose to begin using the taraškevica, an even more unknown standard – as it is not taught in schools nor the publication of books or periodicals following this codification is allowed. The aim of this paper is to present the use of the Belarusian classic standard on the Internet, examining three main points: 1) the motivations who led young people to choose taraškevica instead of the more familiar narkamaŭka and their opinion about the “question of the standard” 2) the deviations from the taraškevica standard, due to the influence of the narkamaŭka (or even directly of Russian); 3) the metalinguistic debates about specific norms in taraškevica and narkamaŭka. 34 References Bieder, H. 2000 Die weissrussische Standardsprache am Ende des 20.Jahrhunderts, in: Zybatow, L. (ed.) Sprachwandel in der Slavia. Die slavischen Sprachen an der Schwelle zum 21. Jahrhundert , Peter Lang, Frankfurt a.M.; Gutschmidt, K. 2000 Sprachenpolitik und sprachliche Situation in Weißrussland seit 1989, in: Panzer, Baldur (Hrsg.) Die sprachliche Situation in der Slavia zehn Jahre nach der Wende. Beiträge zum Internationalen Symposion des Slavischen Instituts der Universität Heidelberg vom 29. September bis 2. Oktober 1999, Frankfurt/M., Berlin, Bern, Bruxelles, New York, Oxford, Wien; Mayo, P. 1975 Recent development in the norms of the byelorussian literary language , The Journal of Byelorussian studies, III,XI, London; Mayo, P. 1978 Byelorussian orthography: from the 1933 reform to the present day , The Journal of Byelorussian studies, IV,XIV, London; Sjameška, L. Prablemy norm i kadyfikacyi belaruskaj litaraturnaj movy, in Gajda, S.(ed.) Języki słowiańskie 1945-1995: Gramatyka-Leksyka-Odmiany, Wydawnictwa Św.Krzyża, Opole 35 Semantic Prototypes of Se-verbs in Russian and Bosnian/Croatian/Serbian: The Case of Across-the-Board Dependencies Erin E. Moulton, University of Kansas

This paper is a comparative study presenting a difference in the semantic prototypes of Slavic se- verbs (commonly termed “reflexive” verbs) in Russian and Bosnian/Serbian/Croatian (BCS) by considering across-the-board dependencies that occur with BCS se-verbs but do not occur with Russian sja-verbs. For example, within a given phrase, two BCS verbs can share one se clitic. (1) Poljubili smo se i zagrlili.2 ‘We kissed and embraced each other.’ The clitic se in (1) appears only once for the reciprocal verbs poljubiti se and zagrliti se, which are equal in semantic function. The single se represents the two agents who act upon each other in both events. The frequency of this kind of construction is expected according to the parallelism requirement of relative prominence presented by Franks (1993). However, the across- the-board dependency can also exist when the relative prominence is less equal. (3) Pogledali smo se i poceli smijati. [sic]3 ‘We looked at each other and began to laugh.’ Pogledati se is a reciprocal verb. This is evident according to a diagnostic in which true reflexive verbs or reciprocal verbs can be rendered with the heavy reflexive pronoun4 (sebe/jedan drugog) in a different context, often for emphasis. This verb is found as brijati sebe in other contexts. The reflexive se/sebe is anaphoric, referring back to the subject. By contrast, kupati se is a middle verb in which se is not anaphoric. Similar across-the-board dependencies with se in BCS indicate that reflexive and reciprocal se-verbs are syntactically equivalent to middle and intransitive se-verbs in that they can co-occur with the same clitic se, such as in (3) where pogledati se is a true reflexive verb and smijati se is a middle verb. These across-the-board dependencies in BCS help to show that that the semantic prototype for BCS se-verbs is reflexive, and other semantic functions are related as kinds of the prototypical reflexive. A different situation exists in Russian, where -sja is affixed to the verb and must appear uniquely with each verb. The lack of across-the-board dependencies, along with other semantic tendencies in Russian, indicates that the prototype for Russian sja-verbs is not reflexive, but middle. The semantic prototypes and functions for Slavic se-verbs vary significantly between the two languages. References Franks, Steven. “On Parallelism in Across -the-Board Dependencies.” Linguistic Inquiry Vol 24, No. 3, The MIT Press, 1993. Kemmer, Suzanne. The Middle Voice. Philadelphia: John Benjamins Publishing Company, 1993. 2http://hogwartslife.blog.hr/ 3 http://www.blogger.ba/komentari/5978/350274/str2 4 The notion of heavy versus light reflexive pronouns/middle markers is developed in Kemmer’s (1993) work as cited. 36 Normal Exceptionality? Russian Temporal Adverbials with v ‘in’ Tore Nesset and Anastasia Makarova, University of Tromsø

Contemporary Standard Russian has a peculiar rule for temporal adverbials, whereby the case government of v ‘in’ depends on the length of the time span in question. Simplifying somewhat, if the time span is shorter than a week, the accusative is used (cf. v ètu minutu ACC, v èto utro ACC, vo vtornik ACC), while for longer time spans v governs the locative case (hence v janvare LOC, v ètom godu LOC, v dvadcatom vekeLOC etc.). This rule is theoretically interesting for two reasons. First, the rule seems to be a counterexample to Talmy’s (1977 & 2000) claim that closed-class elements such as prepositions and case endings are magnitude neutral, in the sense that they do not include reference to absolute quantity. As Talmy (2000:26) points out, the ant crawled across my palm (a very small distance) and the bus drove across the country (a long journey) are equally felicitous, since the semantics of across is not sensitive to the magnitude of the path in question. The second reason why the Russian rule is of theoretical interest is the fact that v is only magnitude sensitive in the domain of time. In spatial expressions, however, both the accusative and the locative are compatible with small and vast spaces, as predicted by Talmy (cf. v karman ACC/v karmane janvare LOC and v Rossiju ACC/v Rossii janvare LOC). In cognitive linguistics, time is traditionally analyzed as a metaphorical extension of space (Haspelmath 1997), and time-space asymmetries of the kind described above are not expected. The present paper addresses two questions: How has this typologically unusual situation developed historically and to what extent does Russian differ in this respect from the other Slavic languages? With regard to diachrony, on the basis of examples from dictionaries and historical corpora (Regensburg Russian Diachronic Corpus http://rhssl1.uni-regensburg.de/SlavKo/korpus, beta version of the historical part of the Russian National Corpus www.ruscorpora.ru) we show that in Old Russian the accusative was the preferred case after v’’ in temporal adverbials, and that the locative has expanded over time. However, there is considerable variation and the situation is complicated by other competing constructions, such as the genitivus temporis (Grannes 1986). As for the cross-Slavic situation, we propose a North Slavic continuum, whereby v+locative is more widely used in temporal adverbials in the west than in the east. For instance, data from the Parasol corpus (www.parasol.unibe.ch) indicate that locative constructions like ve volné chvilce LOC are used as Czech equivalents of Russian temporal adverbials with v + accusative, such as v svobodnuju minutu ACC. Further complicating the picture, the genitivus temporis is more prominent in the west; in the Parasol corpus, for instance, toho dne GEN is attested as a Czech translation of Russian v tot den’ACC. While the magnitude sensitivity of Russian temporal adverbials is exceptional in a cross- Slavic context, time-space asymmetries seem to represent a normal state of affairs; in all the Slavic languages we find differences in the distribution of the accusative and locative cases in the domains of time and space. Thus, in our paper we demonstrate that Russian temporal adverbials with v ‘in’ are both typologically exceptional and normal at the same time. References Grannes, Alf (1986): “Genitivus Temporis” in Early 18th Century Russian. Russian Linguistics 10.1: 53-60. Haspelmath, Martin (1997): From Space to Time. Munich: Lincom Europa. Talmy, Leonard (1977): Rubber-sheet cognition in language. CLS 1977: 612-628. Talmy, Leonard (2000): Toward a cognitive semantics. Cambridge: MIT Press. 37 Slavic Lessons for Computational Phylogeny Johanna Nichols, UC Berkeley and Julie McAnallen, Instituto de Empresa University Segovia-Madrid

Very little computational phylogenetic work has been done within Slavic, though the family offers rich resources and is an excellent test case in view of its mix of inheritance and contact situations and its well-studied differences between the modern three-branch family tree and the known history of dispersal and spread. This paper tests the usefulness of several different kinds of characters for capturing the modern phylogeny, the early dispersal, and contact situations: sound changes, lexicostatistics (cognacy), morphological type, and combined morphological type and cognacy. We test: (1) The 12 major post-Proto-Slavic prehistoric sound changes. These replicate the 3-branch tree, distinguish northern (Lechitic + East Slavic) from southern, and most clearly distance West Slavic from the rest. (2) Loss of intervocalic *-j- in 14 exemplar wordforms (Marvan 1979): reflects the geography of central vs. peripheral languages. (3) Lexicostatistics (pilot studies, extension underway): the 22 most stable Swadesh wordlist items and the 25 least stable (Holman et al. 2008), separately graphed; both sets essentially give star phylogenies with no usable internal structure (though both distinguish Slavic from Baltic). (4) Morphological typology, specifically derivational types in two standard published verb lists for typological elicitation: the causative alternation and posture verbs (each with 15-20 verbs). These reflect the 3-branch structure imperfectly; a few individual languages are displaced in ways reflecting early contact (Slovene affinities with West Slavic, the rest of South Slavic fairly distant, the northern/southern split visible). Within all of Eurasia, Slavic is discrete and compact, and clusters more closely with Romance, Albanian, and Greek than with Baltic, and closer to Baltic than to Germanic. These morphological typological phenomena are known to be fairly stably inherited and somewhat contact-susceptible, and the Slavic behavior is as expected. (5) Cognacy plus typology: the same two wordlists, where the characters are not just the derivation types but also, for each verb, the cognacy of the root, ablaut grade, derivational suffix, conjugation type, and prefix if any. These give by far the most accurate reflection of branch and subbranch structure, plus some reflection of broad geographical configurations (central/peripheral, northern/southern). All in all the features tested distinguish northern/southern more clearly than eastern/western, and most distinguish center/periphery. The northern and southern peripheries are quite different from each other, an effect evidently due not to pure linguistic geography (as then both peripheries should be conservative and similar) but to the different settlement and migration histories (migration into sparsely inhabited lands in the linguistically conservative northwest, much language shift and contact in the more innovative south). The main lesson for computational phylogeny is that neither sound changes, simple root cognacy, nor typology yield fully accurate phylogenies; for that one needs to consider not just roots but derivational and inflectional classes, and not just cognacy but also type. Also, to relate phylogenies usefully to population history one needs to know not just what is shared but also what is innovative or conservative. (Sample Splitstree diagrams on following page) 38 (1) Post-Proto-Slavic sound changes (4b) Typology: Causative alternation (Baltic and Slavic) (5b) Cognacy plus typology in the causative alternation 39 40 Mapping Kashubian in Europe: Which Linguistic Area(s) does Kashubian B elong to? Motoki Nomachi, Hokkaido University

,

This paper analyses areal and typological features of Kashubian that have yet to receive a full discussion in the context of the several linguistic areas, including Standard Average European (SAE), in which languages of Europe participate. Although there has been almost no serious attempt to classify Kashubian as a member of a particular linguistic area, a few scholars have mentioned Kashubian in the context of linguistic areas. These areas are as follows: 1) the Baltic linguistic area, 2) the Rokytno linguistic area, and 3) Peripheral languages (out) of the SAE linguistic area 1) The Baltic linguistic area: In his famous paper on the Eurasian linguistic area, Jakobson (1931) firstly placed the northern dialect of Kashubian into the Baltic linguistic area, only according to its phonological features, which was later properly dismissed by Lehiste (1988). Later on, Dahl and Koptjevskaja- Tamm (2001), and Wälchli (2011) taking into consideration grammatical features of the area’s languages, have included Kashubian into the group of Circum-Baltic languages, but, again, almost without any factual materials of Kashubian. 2) The Rokytno linguistic area: This linguistic area was first proposed and later developped by Déscy (1973, 2000), although its phonological, morphologial and sytactic features seem to be of genetic origin, rather than of convergence due to language contact. 3) Peripheral languages (out) of SAE linguistic area: Although Kashubian has never been analysed in the framework of studies on SAE, it was Heine and Kuteva (2006) who first dealt with the language as a peripheral language in Central Europe from the view point of morphosyntactic criteria. They mention the use of the demonstrative pronoun as article, the so-called possessive perfect, and dative passive construction in Kashubian as a instances of contact-induced grammatical change in the framework of grammaticalization theory, but again almost without any examples or analysis. Taking these facts into consideration, in this paper the author will place Kashubian into proper linguistic area(s), paying special attention to its morphological and syntactic features on the basis of gramamticalization theory introduced by Bernd Heine (2005, 2006), in the context of language contact and changing sociolinguistic emviroments of Kashubian, such as the trend from Kashubian-German billingualism Kashubian-Polish billingulism. References Breza, Edward and Treder, Jerzy. 1980. Gramatyka kaszuska. Gdańsk: Zrzeszenie pomorsko-kaszubskie. Dahl, Östen and Koptjevskaja-Tamm, Maria (eds.). 2001. Circum-Baltic Languages I-II. Amsterdam: John Benjamins Publishing Company. Déscy, Gyula. 2000. The Linguistic Identity of Europe. Bloomington: Eurolingua. Haspelmath, Martin etc (eds.) 2001. Language Typology and Language Universals. Berlin: Walter de Gruyter. Heine, Bernd and Kuteva, Tania. 2006. The Changing Languages of Europe. Oxford: Oxford University Press. Heine, Bernd and Kuteva, Tania. 2005. Language Contact and Grammatical Change. Cambridge: CUP. Jakobson, Roman. 1931. K xarakteristike evrazijskogo jazykovogo sojuza. Paris: Izdatel’stvo evrazijcev. Kortmann, Bernd and van der Auwera, Johan (eds.). 2011. The Languages and Linguistics of Europe. Berlin: Walter de Gruyter. Lehiste, Ilse 1988. Lectures on Language Contact. Cambridge: The MIT Press. Lorentz, Friedrich, 1958-62. Gramatyka pomorska I-III. Wrocław: Ossolineum. Popowska-Taborska, Hanna. 1998. Szkice z kaszubszczyzny: Leksyka, zabytki, kontakty językowe. Gdańsk: GTN. 41 Pests Before Pets: Digging up the Dirt on Old Slavic homҌk ~ homҌstoru/ ‘hamster’ ~ ‘insect(?)’ Josh Pennington, The Ohio State University

The English word hamster has quietly lived in its etymological burrow for the past century, not making too much of a ruckus. In the majority of etymological dictionaries of English (Weekley 1921, OED 1969, Klein 1971, to name a few), the word is said to have been initially borrowed from OHG hamustro (OSaxon hamstra), initially entering Europe through OSlavic homěstoru. This is typically where the chase ends, so to speak. However, recently having uncovered an interesting twist in the etymology from BDE (Barnhart Dictionary of Etymology 1988), I attempt to shed light on hamster via a view that takes hamster as a compound going back (at least) to Avestan hamæstara ‘oppressor, hurling to the ground’, and, I believe, ultimately has its roots in Indic (Sanskrit). In British English, one term is surprisingly used as a synonym for hamster: corn weevil. “Weevil” is a term denoting a variety of insects (usually beetles) which damage crops. While hamsters clearly are not insects, they certainly are notorious for damaging crops and hording grain. Moreover, living as crepuscular beings, they are less likely to be seen doing their damage, so to speak. I believe the stealthy nature of hamsters led farmers to categorize them as another pest to their crops under the general rubric ‘insect’. I tie together related etymologies for hamster (including Russian hamjak and Lithuanian staras, both meaning hamster), from the various IE languages in which it is attested in some form or another to reveal the original semantics behind the naming of the creature. 42 Slavic Languages in Contact: How Can This be M odelled? Achim Rabus, University of Freiburg

Abstract: In recent years, various approaches, hierarchies and models have been developed which claim to adequately describe and explain the lexical and structural effects of language contact. Thomason’s Borrowing Scale (e.g. 2001) has been frequently – and critically – cited. However, works which should also be noted in this field include, among others, Johanson’s Code-Copying Model (e.g. 2002) and Van Coetsem’s Two Tranfer Types (e.g. 2000), which have received little attention in Slavic studies up until now. The aim of this article is to critically discuss the above-named models with the help of various Slavic contact combinations, as well as to present possible amendments and syntheses. In doing so the aspiration is to aplly a panchronic approach which should also incorporate historical contact configurations. The contact combinations examined include non-Slavic – Slavic contact, for example German – Czech; inter-Slavic contact, for example Czech – Polish or Church- Slavonic – Russian; as well as the Slavic Balkan languages that evolved as a result of convergence processes within a Sprachbund. Linguistic data is provided by a corpus as well as relevant studies. References Berger, Tilman (2009): Tschechisch-deutsche Sprachbeziehungen zwischen intensivem Kontakt und puristischer Gegenwehr. In: Christel Stolz (ed.): Unsere sprachlichen Nachbarn in Europa. Die Kontaktbeziehungen zwischen Deutsch und seinen Grenznachbarn . Bochum: Brockmeyer (Diversitas linguarum, 24), pp. 133–156. Johanson, Lars (2002): Contact-induced change in a code-copying framework. In: Mari C. Jones und Edith Esch (eds.): Language change. The interplay of internal, external and extra -linguistic factors. Berlin: Mouton de Gruyter (Contributions to the sociology of language, 86), pp. 285–313. Thomason, Sarah G. (2001): Language contact. An introduction. Edinburgh: Edinburgh Univ. Press. Van Coetsem, Frans (2000): A general and unified theory of the transmission process in language contact. Heidelberg: Winter (Monographien zur Sprachwissenschaft, 19). 43 Reciprocity and Temporality in Russian: a Closer Look at drug druga Lauren Ressue, The Ohio State University

An utterance containing the English expression each other can convey different relations given the context. In the sentence, Artists painted each other, possible configurations of painting include, for example (i) each artist paints every other artist, (ii) each artist painted one other artist. Typically, an expression is called reciprocal if one of its possible interpretations is a strong reciprocal meaning, as defined in (1) and exemplified in (i). (1) (x,y A)(x≠y→xRy) Every member of set A is in the same relationship to every other member of A. In addition to (1), Dalrymple et al. (1998) find four other meanings for each other. Since this expression also conveys relations other than strong reciprocity, we must ask whether all these relations are “reciprocal”, or simply expressed with the same form. One way to determine whether all the relations compatible with each other are “reciprocal” is to explore whether the same set of relations is expressed by a single marker in other languages. If so, it suggests a fundamental commonality to this set of meanings, and we can label it “reciprocal”. If there are two meanings that are only coincidentally expressed by each other, we would not expect to find them expressed by one expression in different languages. Indeed, Majid et al. (2011) show that although languages differ with regard to what meanings utterances containing reciprocal expressions can have, a set of meanings exists common to most languages. In this study, I explore the meanings the Russian anaphoric pronoun drug druga conveys and demonstrate that the form-meaning mapping in regard to reciprocity differs from that of English’s each other, specifically in terms of temporality. The acceptability of drug druga is partially dependent on whether a relation happens simultaneously or sequentially, whereas such a distinction is not significant for each other. I examined sentences from the Russian National Corpus and elicited acceptability judgments from consultants. Both tasks show that although drug druga can convey Dalrymple et al.’s (1998) relations, further temporal distinctions are needed to account for Russian data- felicitous usage of drug druga in many contexts requires a simultaneous relation between participants. For example, (2) Context: Tanja looks at Vanja then looks away. Then Vanja looks at Katja and looks away. You say: #Oni smotreli drug na druga. they look.P ST each on other (Intended: ‘They looked at each other.’) The relation between Vanja and Katja is strongly reciprocal and so should be expressible with drug druga, just as each other is in this context. The sentence in (2) is fine in a context when participants look at each other simultaneously. Therefore, temporality interferes in (2). In a context in which participants hit each other sequentially, however, drug druga may be used felicitously. I argue the difference is in the temporal distinction of the actions. If intervals of the relation can overlap, such as hitting, then drug druga may describe it; otherwise, it cannot. These results suggest that a fine-grained analysis of temporality must be considered when exploring the meaning of drug druga. I also provide further evidence that, cross-linguistically, reciprocal expressions convey different sets of relations between participants, and to understand “reciprocity”, we must look at a large sample of these expressions to find the overlap between these sets. 44 Multiple Aspectual Correlations in Russian George Rubinstein, University of North Carolina at Chapel Hill

1.Russian aspectual correlations can be divided into paired (двучленные) when one form of each aspect correlates with only one form of the opposite aspect, and multiple (многочленные), when one form of an aspect correlates with two (or more) forms of the opposite aspect (Tixonov 1981: 301). There is an opinion that multiple aspectual correlations, including triplets, cannot exist because there are only two aspects in Russian – imperfective (impf) and perfective (pf). However, the idea of aspectual triplets has been given support and much attention lately. Enough is to list the names of scholars who treated this topic in one form or another: Tixonov 1959, 1998; Apresjan 1995; Emel’ianova 1972; Golovin 1999; Petruxina 1990; Jasai 1991; Zalizniak, Mikaelian 2010; Feldstein 2007; Mеrenkova 2003; Xrakovskii 2005, among others. 2.In multiple correlations (aspectual triplets ‘тройки', quadruplets 'четвёрки', quintuplets 'пятёрки', etc.), each link of an aspectual chain (Golovin 1999) can form an aspectual pair with every link of the opposite aspect in the same chain. Thus, aspectual triplets demonstrate a chain of three links representing both impf and pf aspects, which are morphologically related and have a shared meaning. Triplets are subdivided into imperfective and perfective dependent on the prevalence of the related aspect in the chain. The traditional impf triplet contains two imperfectives and a perfective: читать (impf) – прочитать (pf) – прочитывать (impf). Impfective quadruplets, quintuplets, etc., are formed similarly: готовить (impf) – приготовить (pf) –приготовлять (impf) – приготавливать (impf); резать (impf), резнуть (pf), разрезать (pf), разрезать(impf), разрезывать (impf), etc. Multiple perfective correlations contain two or more perfectives. They are subdivided into prefixed pf triplets (quadruplets, quintuplets, etc.): белить (impf) – выбелить (pf) – побелить (pf); мутить (impf) – помутить (pf) – замутить (pf) – взмутить (pf); and suffixed pf triplets, quadruplets, etc.: настигать (impf) – настичь (pf) – настигнуть (pf); болтать (impf), болтнуть (pf), взболтать (pf), взболтнуть (pf), взбалтывать (impf). Despite the abundance of studies of aspectual triplets, that topic has not been exhausted. Almost all these studies concentrate on problems related to impf triplets. Multiple pf correlations have not received due attention. Therefore, suffixed pf correlations, as least studied, will be the focus of this presentation. 3.Verbs of the same root can form triplets of two types: morphological and aspec - tual (Zalizniak, Mikaelian 2010). The former are links which have no shared meaning, as in Tixonov’s example холостить –выхолостить – выхолащивать (1998: 24). The latter are morphological correlations having at least one shared meaning. Such triplets exhibit multiple aspectual correlations in which “one form of a certain aspect correlates with two (or more) forms of the opposite aspect” (Tixonov, ibid.). 4.A decisive factor in impf triplets is the availability of derived imperfective (DI), without which no impf triplet can exist. The predominant number of impf triplets is the result of the productivity of DI suffixes (-ива-, -ыва-), which are the source of derivation of such type of tripltes. In pf correlations, a decisive factor is the availability of a second perfective, which can be partially or fully synonymous with either impf simplex or DI or both. Multiple pf derivation is not a result of one productive source of derivation but uses various sources. These sources include: 1) phonetic variations of roots: выстрогать/выстругать (Axmanova 1957:197); 2) phonetic variations of prefixes: обтесать/отесать (ibid.: 200); 3) phonetic variations with a change of stress: исчеркать/ исчёркать (ibid.: -4) 4) prefixation of paired verbs of motion: нанести / наноси ть 5) prefixation of aspectual correlates: наро и ть – нарожа ть; 6) prefixation of same-root verbs suffixed with –и- and -a-: взломить /взлома ть 7) prefixation of Imperfectives: ослабеть/ослабнуть; 8) prefixation of Imperfectives and Semelfactives: впихать 45 /впихнуть); 9) de-etymologized verbs with a desemantized prefix: насти чь /насти гнуть. Other sources of multiple perfectives with shared senses will be described in the presentation. 5.Aspectual triplets can be presented in the form of a paradigm or a chain (Golovin1999:3). Presentation in a chain is more usual, and the arrangement of links in the chain is more or less stable: imperfectives as extreme members and a perfective in the center, e.g: 1 st link – unprefixed imperfective; 2 nd link – prefixed perfective; 3 rd link – DI: копить –скопить – скапливать. Zalizniak, Mikaėlian (2010) proposed to change the tradition and arrange such triples by the aspect: PF – IMPF1/ IMPF2 (созреть – зреть / созревать). This kind of arrangement is used in the present study. 6.Different semantic relations can be observed between components of aspectual correlati- ons. Following V.G. Golovin (1999:5) and A.G. Merenkova (2003:28), I distinguish three types of relations 1) identical; 2) partially identical and 3) partially partial. In identical relations, semantic identity can be observed between links, which is characteristic of monosemous verbs. Such are four links with the root -кап-: IMPF1 ка пать/ IMPF2 вка пывать – PF1 вка пать/ PF2 вка пнуть, which mean «Лить медленно, по капле внутрь ч-либо», as in «… лекарство в рюмку» (капать, вкапывать, вка пать, вка пнуть). Synonymy of ка пать is contextual but it passes the Maslov test. Partially identical relations characterize the triplet including three prefixed verbs IMPF1 (схлёстывать-1 sense) / IMPF (схлёстывать-2 senses) – PF1(схлестать-1 sense) / PF (схлестнуть-2 senses), which form two aspectual pairs: схлестать/ схлё-стывать and схлестнуть/ схлёстывать. In this triplet, the impf components are expressed by a syncretic verb which has a different number of senses in both impf links because it is a correlate for pf partners with a different number of senses. Besides, the prefixed verbs have no shared senses with their non-prefixed motivators. Thus none of the six senses and five sub-senses of the simplex хлестать or of the two senses of хлестнуть has the sense «Сбить, сорвать, хлестнув чем-л», which is shared by схлестать and схлестнуть, as in Схлестать листья с куста or Схлестнуть цветок. Therefore, “the aspectual correlate of full value for a pf verb is a derived rather than the primary impf verb” (Golovin 1999:6). Partially partial aspectual correlations include triplets, whose all members have different number of senses. They can be observed between five verbs of the same polysemeous root: ки а ть (4 senses), ки нуть (2 senses), выки ать (1 sense), выкинуть (6 senses), выки ывать (2 senses), which form three pairs of correlates: ки а ть/ ки нуть выки ать/ выки ывать and выкинуть/ выки ывать, which form a pf quintuplet (a five-link multiple chain): IMPF1 (ки а ть)/IMPF2 (выки ывать) – PF1(ки нуть) /PF2 (выкинуть)/ PF3 (выки ать). There is partial synonymity between ки а ть and ки нуть, both of which denote throwing but ки нуть implies a one-time action, while the former implies iteration. Other examples of semantic relations between members of multiple correlations, as well as quantitative data related to suffixed pf correlations will be described in the presentation. References Apresjan, Ju.D. (1995) “Traktovka izbytočnyx aspektual’nyx paradigm v tolkovom slovare”. The Language and Verse of Russia. In Honor of Dean S. Worth On his Sixty-fifth Birthday. Moscow: “Vostočnaia literatura” RAN, 31-41. Axmanova , O.S.(1957) Očerki po obščej i russkoj leksikologii. Moscow: Gos. uchebno-pedagog. izd-vo. Emel’ianova S.A. (1972) “Mnogočlennye vidovye korreliacii v “Slovare russkogo iazyka” S.I.Ožegova” Trudy Samarkandskogo gosuniversiteta, vyp.209, 23-31. Feldstein, Ronald (2007) “Russian Dual Stem Aspectual Syncretism and the opposition of Phase and Determinacy.” Glossos 9: Fall 2007. http://seelrc.org/glossos Golovin, V.G. (1999) “Cepočečnye vidovye trojki”. In V.M.Mjakotina (ed).Sbornik naučnyx statej i materialov po russkomu jazykoznaniju i metodike prepodavanija russkogo jazyka. Elec, 3-8. Jasai, L. (1991) “Zametki ob odnom tipe vidovyx troek russkogo glagola” // Studia Slavika: Academual 46 schiant. hungarikant. Budapest, 1991/ 1992. T. 37, fask.1/4. Marenkova A.G. (2003) Vidovye trojki v sovremennom russkom iazyke. Kand. dissertaciia. Elec: Eleckij gos. universitet. Petruxina, E.V. (1990) “K voprosu o konkurencii pervičnyx I vtoričn yx imperfektivov v sovremennom russkom jazyke”. Russkij jazyk za rubežom, 4:82-87. Tixonov, A.N. (1959) “O stilisticheskoj i smyslovoj differenciacii variantnyx form nesoveršennogo vida glagolov s grammatikalizovannymi prefiksami (Kratkie soobščeniia Uzbekskogo universiteta im. A. Navoi. Samarkand. Tixonov, A.N. (1981) “Kategorija vida”. In N.M.Šanskij (ed.). Sovremennyj russkij literaturnyj jazyk. Leningrad: Prosveščenie, 292-310. Tixonov, A. (1998) Russkij glagol; problemy teorii i leksikografirovanija . Moscow:Academia. Xrakovskij,V.S. (2005) “Aspektual’nye trojki i vidovye pary”. Russkij jazyk v naučnom osveščenii. #1 (9). Moscow: Jazyki slavjanskoj kul’tury, 46-59. Zalizniak, Mikaėlian (2010) “O meste vidovyx troek v aspektual’noj sisteme russkogo jazyka”. Komp’juternaja lingvistika i intellektual’nye texnologii. Trudy Meždunarodnoj konferencii Dialog 2010. Bekasovo, 26-29 maja.Moscow:130-135. 47 Structural Attraction in Croatian: Evidence for Morphological Structure in Syntactic Processing Andrea Sims, The Ohio State University

This paper uses structural attraction in Croatian to explore the relationship between morphological structure and the cognitive processing of syntactic agreement. Structural attraction is false agreement in which a target agrees with the local (nearest) noun, rather than the proper but more distant controller: (1) How much correctionSUBJ-SG of syntactic errorsLOCAL-PL are there, anyway? (Bock et al. 2006) In a local context, syntactic errors is a viable subject for the verb; the ambiguity results in „false‟ plural agreement. Structural attraction is widely considered to be a language processing effect (e.g. Badecker & Kuminiak 2007; Franck et al. 2006). Attraction effects are less common in case-inflected languages (Hartsuiker et al. 2001; Nicol & Wilson 2000) because the local noun is usually not a viable subject. However, [reference removed] showed experimentally that in Croatian, a local noun induces more agreement from a following verb when that noun is homophonous (syncretic) with nominative (e.g. hapšenja in (2a)) than when it is not (e.g. spomenike in (2b)). (2) a. NalogNOM.SG za njihovaACC.PL hapšenjaACC.PL (morati) biti u pisanoj formi. „The order for their arrests (should) be in written form.‟ b. StazaNOM.SG uz spomenikeACC.PL (nastavljati se) do samog podnožja gradske stijene. „The path by the monuments (continue) to the very foot of the city wall.‟ This suggests that syntactic processing is sensitive to morphological structure. The present paper expands this analysis along two lines. First, I show that attraction effects in Croatian are not solely a laboratory-induced phenomenon. Exploration of Croatian-language Wikipedia confirms that naturally-occurring attraction errors are rare but existent. In (3), tjerala su („droveN.PL‟) does not agree with either of its conjoined subjects, nor do we find masculine plural resolution. Neuter most likely occurs because ova plemena („these tribes‟) is neuter. This accusative NP is also syncretic with nominative, making (3) structurally similar to (2a). (3) Vječiti ratovi(M).NOM.PL i opasnosti(F).NOM.PL od uvijek mogućih napada… tjeralaN.PL su3PL ova plemenaN.ACC.PL u saveze. „Perpetual wars(M).NOM.PL and dangers(F).NOM.PL from ever-possible attacks… drove N.PL these tribesN.ACC.PL into alliances.‟ This and other corpus examples accord with the experimental results. Second, I tease apart false agreement effects driven by surface identity of form, and by deep morphological identity. There are four experimental conditions for the local noun: (a) true syncretism with nominative, (b) accidental homophony with nominative, (c) “hypothetical nominative” (e.g. koncerta „concert.GEN.SG‟ is unambiguously non-nominative only if inflection class is known), and (d) never same as nominative. I use an acceptability judgment task and a production task to investigate how easily these conditions induce false number/gender agreement on verbs. While data collection is underway, I prediction that true syncretism will produce more attraction errors than the other conditions. The results are expected to indicate that syncretic forms are „visible‟ during syntactic processing, but accidentally homophonous forms and other surface similarities are not. Previous research failed to distinguish between superficial formal identity and deep morphological identity. Doing so will shed light on exactly what is happening during syntactic processing to trigger structural attraction effects. 48 Is Having Rhythm a Prerequisite for Being Slovak? The Case of Fishers and Kingfishers in Czech and Slovak Roman Sukač, Slezská univerzita v Opavě

It is argued that Czech rybařík ‘kingfisher’ and Slovak rybárik ‘fisher’ are representatives of derivative quantity (both derived from rybář and rybár. Derivative quantity in Czech can be observed in certain derivation categories only and its distribution is controlled by the bisyllabic maximally trimoraic domain (in Bethin’s sense). Both phenomena can be labeled the “Rhythmic Law” (RL). The paper summarizes the basic principles of derivative quantity, arguing that both Czech and Slovak rhythmicity arose after the transformation of Proto-Slavic accentual paradigms into West Slavic quantitative paradigms. The author also tries to answer a question why the Rhythmic Law has been traditionally described for Slovak and scarcely observed in Czech. It seems that Czech “rhythmicity” is limited to certain derivational categories, e.g., disyllabic -ař/- ář derivatives (rybář, mísař, kravař) or -ík derivatives in the process of secondary derivation from -ař/-ář nouns (e.g. rybařík). On the other hand, the Rhythmic Law in Slovak was originally described as a phenomenon without exceptions. After the standardization of Slovak many exceptions have been adduced, yet they are hardly explained anywhere. Since the 1940s Slovak scholars have attempted to explain not so much the origin of the RL, but mainly the process of codification. Codified data have then been used as a source of material for analysis. Most Slovak authors have dealt with RL from synchronic and standard viewpoints, but few have tried to explain it historically (e.g., Pauliny) or comparatively in the context of West Slavic. It seems that Slovak linguistics takes the RL as a source “national pride” without acknowledging the possibility of the existence of the similar process in the rest of the West Slavic territory. References Bethin, C.Y. (2003) Prosodic effects in Czech morphology. American contributions to The thirteenth international congress of Slavists, Ljubljana 2003, eds. Maguire, R.; Timberlake, A., Slavica Publ., 2003, 9–22. Kvantita v spisovnej slovenčine a slovenských nárečiach . (ed. Považaj, M.)., Bratislava:Veda 2005. SUKAČ, R. (2011) Old Czech rhythmic law. In Baltische und slavische Prosodie. (Hrsg. Stadnik Holzer, E.), Peter Lang Verlag, 2011, 141–152 49 Syntactic Evidence for vP inside the Verbal Nominals in Polish Malgorzata Szajbel-Keck, UC Berkeley

Polish verbal nominals (e.g. czytanie ‘reading’, mycie ‘washing’) often behave syntactically like verbs rather than nouns, even though they have external morphosyntax of NPs (they have grammatical gender, take nominal case and number inflection, can be modified by adjectives and pronouns) and function as NPs in the greater context of the clause (as subjects, objects, PP complements, etc.). It is already generally agreed for many other languages (e.g. English in Fu et al. (2001), Bowers (2011); Irish in Carnie (2011); Sakha in Baker (2011); Quechua in Cole (2011)), that this is due to a “dual” or “mixed” structure of such nominals, which contain both a nominal and verbal projection (see (1) for a schematic representation). Polish verbal nominals are similar to verbal nominals in other languages in this respect = (Rappaport (2001)) and this paper supports the thesis that this is a cross‐linguistic phenomenon, and that languages differ only in details. One of these details is how much of the verbal/clausal projection is included in the verbal nominal, as this seems to be language specific, and various sizes have been proposed, from “high” nominalizations, where the whole CP is postulated inside the nominal (e.g. in Sakha – Baker (2011)) to “low” nominalizations, where only verbal head is included (e.g. nonevent nominals in English – Bowers (2011)). In my presentation, I will provide structural evidence that the verbal nominal in Polish selects the whole vP (including the full argument structure, i.e. both the internal and external arguments), which puts it in the middle of the scale. The evidence from zrobić to anaphora (Eng. do so) resolution, which can be bound only by a VP (Kehler and Ward (1995)) indicates clearly that there must be a verbal projection inside the verbal nominal that binds the anaphora (2). This is also supported by the observation that Polish verbal nominals can participate in control constructions and can, in fact, serve as a reparation mechanism for Accusative object control which is impossible in Polish (cf. (3) to (4)). Further the positioning of the VP‐ and sentential adverbs shows that it is a vP that is selected by the nominal head. Finally the binding of reflexive pronouns inside the verbal nominal (5) is identical to the binding inside an infinitival phrase (6) and different to the binding inside an NP headed by an underived noun (7). Since the difference in binding between infinitives and nouns ((6) and (7)) has been attributed to the presence of the subject in the VP containing the infinitive (Reinders‐ Machowska (1991)), I argue that the subject position must also be retained in the verbal nominal, and it is headed by PRO if the subject is not expressed overtly (8). This provides yet another argument that the verbal nominal must select for no less than a vP in Polish. (1) 50 (2) Adam tak długo namawiał Anię do zanurkowania, aż wreszcie zgodziła Adam‐N so long persuade‐PAST.3.SG Ania‐A to diving‐PF.G until finally agree‐ PAST.3.SG się to zrobić REFL this‐A.SG do‐INF ‘Adam was trying to persuade Ania to dive for such a long time, that she finally agreed to do it’ (3) *Adam poprosił Anię napisać list Adam‐N ask‐PAST.3.SG Ania‐A write‐INF letter‐A (intended) ‘Adam asked Ania to write a letter’ (4) Adam poprosił Anię o napisanie listu Adam‐N ask‐PAST.3.SG Ania‐A about writing‐PF.A letter‐G ‘Adam asked Ania to write a letter’ (5) [Adam1 nakazał Piotrowi2 odwiedzenie swojej1/2 / jego1/*2 cioci] Adam‐N1 order‐PAST.3.SG Piotr‐D2 visiting‐PF.A self‐G1/2 / his‐G1/*2 aunt‐G ‘Adam ordered Piotr the visiting of his own / his aunt’ (6) [Adam1 nakazał Piotrowi2 [PRO2 odwiedzić swoją1/2 / jego1/*2 ciocię]] Adam‐N1 order‐PAST.3.SG Piotr‐D2 PRO2 visit‐INF self‐A1/2 / his‐A1/*2 aunt‐A ‘Adam ordered Piotr to visit his own / his aunt’ (7) [Adam1 nakazał Piotrowi2 wizytę u swojej1/*2 / jego*1/2 cioci] Adam‐N1 order‐PAST.3.SG Piotr‐D2 visit‐A at self‐G1/*2 / his‐G*1/2 aunt‐G ‘Adam ordered Piotr (to make) a visit to his own / his aunt’ (8) Adam1 nakazał Piotrowi2 [DP [vP PRO2 odwiedzenie swojej1/2 / jego1/*2 cioci]] Adam‐N1 order‐PAST.3.SG Piotr‐D2 PRO2 visiting‐A self‐G1/2 / his‐G1/*2 aunt‐G ‘Adam ordered Piotr the visiting of his own / his grandmother’ References Baker, Mark C. (2011). Degrees of nominalization: Clause‐like constituents in Sakha. Lingua 121 (7): 1164‐1193. Bowers, John. (2011). Non‐event nominals and argument structure. Lingua 121: 1194‐1206. 51 Carnie, Andrew. (2011). Mixed categories in Irish. Lingua 121 (7): 1207‐1224. Cole, Peter, and Gabriella Hermon. (2011). Nominalization and case assignment in Quechua. Lingua 121 (7): 1225‐1251. Fu, Jingqi, Thomas Roeper, and Hagit Borer. (2001). The VP within process nominals: evidence from adverbs and the VP anaphor do ‐so. Natural Language & Linguistic Theory 19: 549‐582. Kehler, Andrew, and Gregory Ward. (1995). "On the anaphoric status of do so." In LSA Annual Meeting. New Orleans, LA. Rappaport, Gilbert C. (2001). "The geometry of the Polish nominal phrase: problems, progress, and prospects." In Generative Linguistics in Poland: Syntax and Morphosyntax, eds. Piotr Bański and Adam Przepiórkowski. Warszawa: Polish Academy of Sciences. 173‐189. Reinders‐Machowska, Ewa. (1991). "Binding in Polish." In Long ‐Distance Anaphora, eds. Jan Koster and Eric Reuland. Cambridge [England]; New York: Cambridge University Press. 137‐150. 52 Mood in Serbian Nataša Todorović, University of Illinois at Chicago

As proposed by Progovac (2005) the clausal structure in Serbian comes in two layers: PolSP (the subject layer) and PolO P (the object layer). Analyzing subjunctive and indicative da-complements in Serbian within Progovac’s (2005) clausal proposal and Giannakidou’s (2009) notion of (non)veridicality, I claim that indicative da-complements are introduced by da [+veridical] projected in PolSP while subjunctive da-complements are introduced by da [-veridical] projected in PolO P. In contrast to Progovac (2005), I propose that the lower PolO P domain lacks TO P projection, and that PNP 5 (perfective non-past) and INP (imperfective non- past) anchored by da [-veridical] in subjunctive complements have no tense projections. On the contrary, I claim that the higher PolSP domain does not lack TSP, which is responsible for checking tense features associated with the auxiliary clitics and the INP anchored by da [+veridical]. Pol sP As psP da [+veri dical] Agrs P Asps Ts P Agrs Pol oP Aux As poP da [-veridical] AgroP As po VP Agro Evaluation of data suggests that the da [+veridical] can select for any tense, it requires clitics to follow it, it allows negation to separate it from the complement verb, and it creates a domain for negation licensing of NI-NPIs (ni-negative polarity items) and negation anti-licensing of I-NPIs (i-negative polarity items). On the other hand, the da [-veridical] disallows any other tense but the Non-Past restricting its complement selection to PNP and INP, it allows clitics to precede it, and it does not allow negation to separate it from the verb. The notion of different da projections allows for a more uniform explanation of clitic placement, tense restrictions, and licensing of NPIs in the indicative and subjunctive da- 55 The notions of (non)veridicality and PNP/INP are adopted from Giannakidou (2009). 53 complements. Finally, one of the most important claims this analysis provides is that mood interpretation (indicative vs. subjunctive) in Serbian is co-dependent on the matrix verb as well as the selection of the embedded da-complement. (1) a. Znam da pravim tortu. know-1sg da make-1sg.imprf. cake i. I know that I am making the cake. ii. I know how to make a cake. b. Znam da napravim tortu. know-1sg da make-1sg.perf. cake i. I know how to make a cake. ii. It happens that I (on occasion) make a cake. As observed in (1a-b) factive znati (know) can select either da [+veridical] or da [- veridical]. The interpretation of (1a) or (1b) is dependent on the semantic properties of the matrix verb, da selection, and the aspect/tense selection of the embedded verb. As noted by Quer (2010: 168) “the verb meaning is not the only factor determining the choice of mood” since “[verb’s] interaction with other elements like operators, aspect, or presupposition can be decisive in determining the mood in the embedded clause.”It is exactly this observation that provides an answer for Serbian: da projected in PolSP, which selects for INP, past, or future tenses, allows for the indicative interpretation while another da projected in PolO P, which only selects for PNP or INP, allows for the subjunctive interpretation. References Giannakidou, Anastasia. 2009. On the temporal properties of mood: the subjunctive revisited. Lingua. 119:1-27 Progovac,Ljiljana. 2005. A Syntax of Serbian-Clausal Architecture. Bloomington: Slavica. Quer, Josep. 2010. On the (un)stability of the mood interpretation in Romance. In M.G Becker and E. Remberger eds., Modality and Mood in Romance, 163-179. Berlin/New York. De Gruyter. 54 Linguistic purity: how linguistic ideology affects speakers of Surzhyk, Ukrainian, and Russian in Ukraine Susan Vdovichenko, Washington & Jefferson College

The way that language and power interact affects every person in society. On one hand, language can be overtly connected to power, with only one language (and it speakers) given access to desired positions in society through the use of an official language. In a much more subtle way, ideas about “proper” or “clean” language also allow speakers of a specific type of language to maintain positions of power over those who speak “incorrectly.” In Ukraine, after nearly a century of widely varying linguistic policies and customs, Ukrainian was declared the sole official language in 1991. Two decades later, language issues remain critical. Even as politicians argue about the use of language in an official capacity, an ideal is being held up about pure language, both Russian and Ukrainian, and those that speak a mixture between the two are similarly at a disadvantage. This paper looks at the results of surveys and interviews conducted by the author in Simferopol, Kherson, Kyiv, and Kharkiv, Ukraine, to help understand the way that people relate to each other based on their speech. This paper adds to available scholarly research about the situation in Ukraine, for example Laada Bilaniuk’s Contested Tongues: Language Politics and Cultural Correction in Ukraine (Ithaca, NY: Cornell University Press), by using quantitative and qualitative data to fully illustrate the current linguistic situation. 55 Early Edition: Linguistic features of the Proglas in Hilandar 23 Monica Vickers, The Ohio State University

The 13th -century manuscript known as Hilandar 23 contains the earliest extant copy of the Proglas, the Slavic verse prologue to the Gospels, attributed to St. Cyril (Konstantin Filosof). The manuscript also includes all four Gospels, prayers, a synaxarion, and a menologion; most of the manuscript is written by the same scribe. Most previous research on the Proglas has been restricted to reconstructions of the original Old Church Slavonic text, literary discussions, and debates over its authorship. The central goal of this project, however, is to provide a precise linguistic description of this manuscript, including variant features of orthography, phonology, and morphology. Although the language of Hilandar 23 has been classified as the Serbian (Raška) recension of Old Church Slavonic, the Proglas exhibits unexpected spellings in certain grammatical forms. Close reading revealed cases of spelling etymological back jer as o and front nasal and jat as a, which suggests influence from East Slavic or the central Balkans (modern-day Macedonia). Close reading of the Proglas and sampling of other portions of the manuscript, including parts written by different scribes, has shed light on the distribution of variant features as well as their possible origins, whether due to dialects or personal scribal influence. Sampling shows that these features are generally limited to the Proglas and restricted to certain stems and grammatical forms. These features most likely appear due to either East Slavic or central Balkan influence; both are possibilities, although certain factors, including the history of the Proglas as a South Slavic text, decrease the likelihood of an East Slavic provenance. 56 Approaching variability in Slovenian vs. Czech aspect from a pan-Slavic perspective Ruprecht von Waldenfels, Universität Bern

Previous research into cross-Slavic variation in the use of verbal aspect has pointed to a rather exceptional position of Slovenian among the Slavic languages (Galton, 1981; Dickey, 2000, 2003). In the paper this is approached from a parallel corpus perspective. Using a corpus-driven approach based on the ParaSol corpus, von Waldenfels (2012) could confirm Dickey’s (2000) broad division of Slavic aspect use into an Eastern and Western Group as well as Benacchio’s (2010) compatible analysis for aspect use in the imperative. A similar finding for aspect use in negated past events is reported in von Waldenfels (forthc.). However, in both these studies the languages of the Western Group, primarily Slovene and Czech/Slovak, seem to be much less similar to each other than the languages of the Eastern Groups. The present study builds on these results and examines aspect use in Slovenian and Czech against the background of aspect use in equivalent translated text chunks across other Slavic languages. On the basis of three di_erent environments of aspect usage, namely the imperative, negated past forms, and non-past finite verb forms, the variability of aspect in the Slovenian and Czech texts is assessed by determining the frequency of the actual forms and their aspectual partners in large reference corpora. From preliminary investigations it is expected that aspect choice in Slovenian is much more restricted than in Czech, pointing to important di_erences between aspect use in these two languages. References Benacchio, Rosanna (2010) Vid i kategorija vežlivosti v slavjanskom imperative. Sravnitel’nyj analiz. München, Berlin: Kubon und Sagner. Dickey, Stephen M. (2000) Parameters of Slavic aspect: a Cognitive Approach. Stanford: CSLI Publications. Dickey, Stephen M. (2003) Verbal Aspect in Slovene. Sprachtypologie und Universalienforschung 56(3):182–207. Galton, H. (1981) The specific Position of Slovene in the Slavic Verbal Aspect. Slovene Studies Journal 3(2):49–58. von Waldenfels, Ruprecht (2012) Aspect in the Imperative Across Slavic - a Corpus Driven Pilot Study. In: Atle Grønn and Anna Pazelskaya (Eds.), The Russian Verb. Oslo Studies in Language 4, 141–154. von Waldenfels, Ruprecht (forthc.) Explorations into Slavic Variation: Taking a Bottom-up Approach. In: Benedikt Szmrecsanyi and Bernhard Wälchli (Eds.), Aggregating Dialectology and Typology: Linguistic Variation in Text and Speech, within and across Languages. De Gruyter Mouton. 57 New Challenges of Serbian as Non-Native Language Julijana Vučo University of Belgrade, School of Philology

Democratization in public life, heightened self-awareness of minority groups across Europe and simultaneous processes of globalization, glocalization and localization have all intensified the urge to adapt and adjust to the new global/glocal context and have imposed in Europe a specific, plurilinguistic view on the importance of foreign languages, and special importance of developing national strategies for language learning, which could contribute to preservation, or to obtaining a “linguistic ecological harmony” at the global level. Linguistic variety of the continent is fully recognized as a key factor for strengthening European unity, and its maintenance and valorization is strongly supported in the face of the ever sounding alarm about the overarching presence of English as the global lingua franca (Seidlhofer, 2005: 340, House, 2003; Bugarski, 2005). The implicit and explicit objectives as well as planned outcomes of a specific language policy and planning vary significantly in accordance with the socio-political, epistemological and strategic attitudes of the language planners in different socio-political and cultural contexts around the world (Ricento, 2000, 2006; Filipović, Vučo, Djurić, 2007; Filipović, 2009). Council of Europe and the European Union continuously search for alternative ways to create new learning and teaching practices, which would take into account a number of individual and communicative needs6 . However, the role of small languages is equally important in that they may retain and maintain their own linguistic expression in the territorial, cultural and historical sense. Following historical changes in the Western Balkans, Serbian language is gaining a specific importance in the linguistic market over the past few years, having entered a common social, economical and political space among the languages of communication. A new role of the Serbian language in plurilinguistic European context includes also new aspects of teaching and learning this language in its multiple global and regional paradigms: as the language of a milieu, a foreign language, non-native language, or a national language and a language of education. The paper refers to the criteria for creating the internationally recognized proficiency levels of Serbian L2 related to the Common European Framework of Reference for Languages, to formulate by academia and State agencies to support the Serbian as an L2 (Filipović, 2007:226). The criteria should be applied to the tests of Serbian language as L2, along with the requirements which need to be met if Serbia is to become a member of international organizations of L2 testing and evaluation as Association of Language Testers of Europe, http://www.alte.org, to design the database of standardized tests, programs, courses, online courses and didactic materials (Vučo 2007:108, 2010:280). The standardized criteria will recognize to its speakers a functional value in international communication (educational systems abroad, labor market, international organizations etc.). References AA.VV. Rewarding challenge: How the multiplicity of languages co uld strengthen Europe”, Proposals from the group for intellectual Dialogue set up the Initiative of the European Commission, Brussels, 2008. http://ec.europa.eu/education/policies/lang/doc/maalouf/report_en.pdf [30.06.2012] Bugarski, Ranko. 2005. Jezik i kultura (Language and Culture). Beograd: Biblioteka XX vek,. 6 The linguistic objective of Europe should be trilingual individuals who would speak 1. their L1 (language for identification), 2. a lingua franca (language of global or continental communication, such as English, French, Spanish, Chinese, or other), and 3. a personal adoptive language, which should be developed as an L2 to the full extent of an individual’s capacities. (European Commission, 2008). 58 Council of Europe (2001). Common European framework of reference for languages: Learning, teaching,assessment. Cambridge: Cambridge University Press. Filipović Jelena (2009). Moć reči: Ogledi iz kritičke sociolingvistike. (The social power of words: Essays on critical sociolinguistics). Beograd: Zadužbina Andrejevi�. Filipović Jelena, Vučo Julijana. & Djurić, Ljiljana (2007). Critical review of language education policies in compulsory primary and secondary education in Serbia. Current Issues in Language Planning , 8(1), 222– 242. Filipović Jelena (2007) Vrednovanje znanja i organizacija ispita iz srpskog za strance prema kriterijumima međunarodnih organizacija, Srpski kao strani jezik u teoriji i praksi, Filološki fakultet, Beograd, 217 - 226 House, J. 2003. English as lingua franca: A threat to multilingualism? Journal of Sociolinguistics, 7(4): 556-578. Ricento, Thomas (2000). Historical and theoretical perspectives in language policy and planning. Journal of Sociolinguistics 4(2), 196–213. Ricento Thomas (Ed.) (2006). An introduction to language policy. Theory and method. Oxford, UK: Blackwell Publishing. Seidlhofer, Barbara (2005) Key concepts in ELT. English as a lingua franca, ELT Journal, Volume 59/4, Oxford University Press. Vučo, Julijana (2007) Srpski kao strani u ranom i adolescentnom uzradtu, Srpski kao strani jezik u teoriji i praksi, Filološki fakultet, Beograd, 99 – 110. Vučo, Julijana (2010). Srpski jezik i jezici obrazovanja, Književni (standardni)jezik i jezik književnosti, M. Kovačević (Ed.), FILUM, Kragujevac, 279 – 287. 59 The Effect of Teaching Foreign Language Vocabulary in Semantic Groups: A Study of Russian Language -Learners and Vocabulary Acquisition Kate White, The Ohio State University

A long-standing assumption in the field of second language acquisition research is that learning words in semantic groupings has a positive effect on acquisition and retention, though this assumption was not based on empirical evidence (Finkbeiner & Nicol, 2003). This assumption is common among researchers and instructors of second languages, as it seems to fit intuitively with one of the most popular current teaching methods for foreign languages, the communicative method. Previous results have shown that grouping words semantically for word-learning is misguided (Erten & Tekin, 2008; Finkbeiner & Nicol, 2003; Papathanasiou, 2009). The use of this method was initially based on results from studies on monolingual memory. The most important issues with this explanation are: 1. Memorization and word-learning are two different cognitive processes, and 2. Monolinguals and bilinguals cannot be compared in this way (Grosjean, 1998). The methodologies of previous studies have not been consistent in treatment of various processing variables that may influence word-learning. These include proficiency level and processing depth. Speakers at higher proficiency levels may be less influenced by semantic groupings because of greater control of the language (Altarriba & Mathis, 1997). However, few have investigated proficiency level, perhaps due to the use of non-words rather than second- language vocabulary in some studies. As for processing depth, semantic groupings may actually be learned better, though slower, due to being processed more deeply (Schneider, Healy, & Bourne, 2002). The purpose of this study was to explore the issue of semantic groupings while considering these processing issues. This study also contributes data to the literature from a new language, Russian. A word-learning task was presented to Russian language learners at two levels of proficiency (first- and third-year) over a two-week period. The task was presented in the classroom in a close simulation of classroom vocabulary learning. In the first session they learned words in semantic groupings; in the second they learned words grouped randomly. In order to test the variable of processing depth, early and late post-tests were used. The post-tests include a production task (picture-naming) and a recognition task (word-decision). The reaction time and accuracy of the participants were measured and analyzed. The results demonstrate that semantic groupings influence the processing of language- learners. In agreement with previous studies, the assumption about semantic groupings in word- learning is not supported by this study. This study contributes to the body of literature on the use of semantic-groupings in word-learning. The purpose was to more strictly control the variable of processing depth and to introduce the variable of proficiency level. Finally, this study contributes data from learners of Russian, who were enrolled in Russian at the time of the experiment and learned the words in a classroom setting. 60 Impersonal Modals and Face -Threatening Acts Renee Perelmutter, University of Kansas

This talk investigates the usage of impersonal modals with or without a dative argument in contexts of potential face-threatening acts (Brown and Levinson 1987), especially advice-giving, in Modern Russian. According to (Timberlake 2004, 365), modal dative-with-infinitive constructions developed from historically free dative-with-infinitive constructions (c.f. as nam rabotat’! ‘to.us.DAT work.INF’ in Modern Russian). Impersonal modals involve such modal predicates such as nužno/nado ‘necessary’, možno ‘possible’, nel’zja ‘forbidden, impossible’, sleduet ‘appropriate’, stoit ‘worth it’ and short adjectives such as trudno ‘hard’, xorošo ‘good’ (with a subjunctive variant xorošo by ‘it would be good’). These modals express a general obligation; they can include a dative of the goal (a specific individual) at whom the obligation is directed: [mne] nado dopisat’ pismo ‘[I.DAT] have to finish writing the letter’. Not surprisingly, impersonal modals are frequent in contexts of potential face-threatening acts, especially in advice-giving. Contrasting usages with and without a dative argument, I show how these constructions can remove reference to both the advice-giver and the addressee of the advice. By using impersonal predicates such as nužno/nado ‘necessary’ without a dative argument, the adviser signals that the advice offered is objective; there is an element of inevitability and necessity in following this advice; this redresses the face-threatening aspect of the advice, since the illocutionary force is marked as universal, proposing that an action is objectively necessary. Though the dative argument used to be a required part of the DAT+INF construction, I question whether the dative argument is a required component of the [DAT]+modal predicate+INF construction in Modern Russian. Statements such as nado rabotat’ ‘[I] must work’ or nado rozhat’ ‘[you] have to give birth!’ in my data attest to the productivity of this construction with a modal predicate +INF realization and a universalizing meaning. The ubiquity of dative-less realizations might point at the increasing the impersonality of this construction and its interpretation as a statement of universal/objective truth that applies to all referents; nevertheless, a referent can still be specified with the dative argument. Examining this evolving usage, I discuss how constructions may change over time under the influence of pragmatic motivations. References Brown, Penelope, and Stephen C. Levinson. 1987. Politeness: some universals in language usage, Studies in interactional sociolinguistics, 4. Cambridge; New York: Cambridge University Press. Timberlake, Alan. 2004. A reference grammar of Russian. Cambridge, U.K.: Cambridge University Press. 61 Associations between Conditionals and Habituality in Slavic Languages Jurij Knjazev, Saint Petersburg State University

1. In Bulgarian, common Slavic iterative verbs with the suffix -va-/-iva- (e. g Russian byvat’ ‘be (habitually)’ from < byt’ ‘be’ or xaživat’ ‘go (habitually)’ < xodi t’ ‘go’) now are used mainly as the forms of so called ‘conditional mood of willingness’ (Bulg. uslovnoto naklonenie za gotovnost): (1) Bulgarian (Stankov 1969: 135) Ne sǎm gladen, ama ako počerpite, pijvam edno vince ‘I am not hungry, but if you entertain me, I would like to drink some wine’. 2. In Serbian and Upper Sorbian, the meaning of the Past Habitual may be expressed by an analytical form of the conditional mood consisting of a special form of the verb be (or an invariable verbal particle of the same origin) and the main verb in the past tense form in -l: (2) Serbian (Mønnesland 1984: 72) a. Kad god bi popila čašu rakije, napila bi se ‘When(ever) she drank a glass of brandy she would get drunk’; b. Kad bi popila čašu rakije, napila bi se ‘If she drank a glass of brandy she would get drunk’; (3) Upper Sorbian (Fasske, Michałk 1981: 266) K narodinam by nan małemu jubilarej stajnje całty z kołbasku přinjesł ‘For the son’s birthday, father used to bring rolls with sausage for the small jubilant’. 3. Similar meaning may be rendered by Russian conditional mood in universal concessive constructions with the particle ni: (4) Kogda by ona ni pila, vsegda napivaetsja’ ‘Every time she drink she gets drunk’. 4. Macedonian forms of the Future-in-the-past, consisting of the main verb in the form of imperfect accompanied by the future marker ќe, may render, besides, unreal conditional meaning and iterativity in the past: (5) Macedonian (Usikova 1977: 263) a. Ќe dojdev, ako imav vreme ‘I would come in time if I had enough time’; b. Ќe dojdeše, ќe sedneše, pa ќe staneše i ќe izlezeše bez da prozbori ‘He used to come in, take a seat, then get up and go out without a word’. 5. Ability of constructions of different origin in various not-adjacent languages to express both habituality and irreality may hardly be treated as accidental. Verbal forms with a similar combination of meanings are also found in Armenian, English, the Modern Greek and a number of other languages (Cristofaro 2004). The point of connection of these two meanings lies, perhaps, in the fact that both categories refer to situations which do not occur strictly within the reference time. The iterative meaning excludes the specification of the single occasion in which the situation took place, whereas conditional describes situations that yet desired or, on the contrary, cannot occur (Knjazev 1997: 269). References Cristofaro, S. 2004. Past habituals and irrealis. Ju. A. Lander, V. A. Plungjan, A. Ju. Urmančieva (eds.). Issledovanija po teorii grammatiki. Vyp 3. Irrealis i 62 irreal’nost’. Moscow: Gnosis, 256-272. Fasske, H. and S. Michałk. 1981. Grammatik der obersorbischen Schriftsprache der Gegenwart. Bautzen: Domowina. Knjazev, Ju. 1997. Expression of situational plurality in Russian and other Slavic languages. V. S. Xrakovskij (ed.). Typology of iterative constructions. München; Newcastle: LINCOM EUROPA, 241-270. Mønnesland, S. 1984. The Slavonic frequentative habitual. C. de Groot, H. Tommola (eds.). Aspect bound. A voyage into the realm of Germanic, Slavonic and Finno - Ugrian aspectology. Dordrecht: Foris, 53-76. Stankov, V. 1969. Bǎlgarskite glagolni vremena. Sofia: Nauka i izkustvo. Usikova, R. P. 1977. Makedonskij jazyk. A. G. Širokova, V. P. Gudkov (eds.). 63 Slavic Aspirates Revisited Ondřej Šefčík, Masaryk University

The traditional reconstruction of the IDE phonemic system leads to the structure: T DDh That system is often reinterpreted with the use of the glottalic theory, but for the purpose of our paper we will keep the traditional terminology. The Balto-Slavic languages have only an opposition based on the voice-feature, the former IE voiced aspirates having merged with the voiced counterparts. T D( Dh) The general opinion is that Slavic data offer no possibility of internal reconstruction both of the IE voiced and voiced aspirated, but with an application of the Winter’s law we can at least sometimes distinguish a different origin of the Balto-Slavic voiced stops. Winter’s law, remarkably reworked by Kortlandt, proposes that if in an acute syllable vowel preceded an original voiced stop, the resulting vowel is a long acute vowel: cf. OIA ádmi, átti vs. Lith. jsti, Latv. êst, SCr. j sti, OCS jasti < * Sl. ěsti “eat”. More interesting is a question of the unvoiced aspirated stops. As known, in the Indo- Iranian languages the unvoiced aspirated stops are a regular result of the development of clusters T+H (cf. OIA ratha- “chariot” vs. Lat. rota “wheel”, OIA khidati “to tear” vs. Lat. caede “I kill” etc.). There is no internal proof for the existence of the unvoiced aspirated in the Baltic languages yet because the cluster T+H is realized in Baltic as T only (OIA sthX- “stand” vs. Lith. stoti “to stand”). The reflexes of the cluster T+H in Slavic are more generous but only for the cluster *k+H. See the following examples: OIA śXkhX “branch, twig”, Ru. soxá “plough”, Cz. socha “sculpture”, SCr. sòha “forked stick” vs. Lith. šakà “branch”, Goth. hoha “plough”; OCS xrab(ъ)rъ “brave”, OIA khára- “sharp, hard” vs. Latv. skârbs, Germ. scharf “sharp” but OCS pěny “foam”, Lat. spūma, OE fXm, Lith. spáinė “foam” vs. OIA phena- “foam”; OCS stati, Lith. stoti, Lat. ste “stand” vs. sthX- “stand”. From the examples above we can suppose that at least the clusters *k+H phonologized in Slavic as x, which merged with the x resulting from borrowings (OCS xlěbъ “bread”) or (and more importantly) with a post-ruki-rule back variation of š (< *s). In my paper I would like to sketch both the detailed history of the aspirated in Balto-Slavic and Slavic and the internal chronology of the whole series of developments and to discuss general validity of my observations for both branches of the Balto-Slavic languages. 64 “Exempt” Anaphors of the Balkans Steven Franks, Indiana University

This paper seeks to understand the non-canonical properties of the colloquial Balkan Slavic anaphor nego si in traditional binding theoretic (BT) terms. This item differs from the anaphors found in other Slavic languages in that it expresses phi-features (person-number-gender). However, it is nothing like English himself, since, as described by Schürcks (2003, 2006, 2008), it flouts standard structural requirements which restrict the set of antecedents to sufficiently local c- commanding expressions. These forms are thus not reflexive in the standard sense (cf. Reuland and Reinhart 1993) of marking a predicate as reflexive. Schürcks provides numerous examples which show nego si to be “exempt” (in the sense of Pollard and Sag 1992) from the binding theory. Compare well-behaved sebe si in the (a) examples with apparently BT exempt nego si in the (b) examples: (1) a. Ivanovijat bašta kritikuva sebe si. ‘Ivan’s father criticizes SEBE SI.’ [= Ivan; ≠ bašta] b. Ivanovijat bašta kritikuva nego si. ‘Ivan’s father criticizes NEGO SI.’ [= Ivan; = bašta] (2) a. Ivan razkazva na doktora istorii za sebe si. ‘Ivan tells the doctor stories about SEBE SI.’ [= Ivan; ≠ doktora] b. Ivan razkazva na doktora istorii za nego si. ‘Ivan tells the doctor stories about NEGO SI.’ [= Ivan; = doktora] (3) a. Ivan kazva če doktorǔt mrazi sebe si. ‘Ivan says that the doctor hates SEBE SI.’ [≠ Ivan; = doktorǔt] b. Ivan kazva če doktorǔt mrazi nego si. ‘Ivan says that the doctor hates NEGO SI.’ [= Ivan; = doktorǔt] Schürcks claims that what makes nego si special is that it forces an interpretation taken from the speaker’s point of view (POV), in the sense of Kuno 1987 and, for Slavic, Yokoyama 1986. In this paper I expand on her insight, arguing that the “anaphor extending” possessive clitic si enforces the imposition of the speaker’s (“[+speaker]” perspective more generally (e.g., in sebi si as well), but rejecting her need for [±subject], since this is derivable from the anaphor’s morphological properties. I also argue that this effect of adding si is a Balkanism: Turkish has both kendi ‘self’ and kendisi ‘self’s self’, for example, and the data suggest that si is playing a similar POV role in that language. I also argue that the POV anaphor phenomenon is connected to other typological hallmarks, such as (i) grammaticalization of an “reported” vs. “witnessed” distinction (4a), (ii) the prominence of preverbal topics (4b), and (iii) clitic doubling, with its implication not only of overt topics (4c), but also wholesale silent ones (4d). (4) a. Ivan (e) nameril knigata. ‘(I attest that) Ivan found the book.’ b. Ivan nameri li knigata? ‘As for Ivan, did he find the book?’ c. Knigata ja nameri Ivan. ‘As for the book, Ivan found it.’ d. PRO/pro nameri ja Ivan. ‘It was found by Ivan.’ Finally, a formalization of anaphoric POV is developed in terms of a syntactic position for a discourse prominent element, leading the way to unification of these seemingly disparate typological phenomena. References Kuno, Susumu. 1987. Functional Syntax: Anaphora, Discourse, and Empathy. Chicago: The University of Chicago Press. Pollard, Carl and Ivan Sag. 1992. “Anaphors in English and the Scope of Binding Theory”. Linguistic Inquiry 23: 261–303. Reinhart, Tanya and Eric Reuland. 1993. “Reflexivity”. Linguistic Inquiry 24: 657–720. Schürcks, Lilia. 2006. “Binding and Point of View in Bulgarian”. eitschrift f r Slawistik 51: 386–405. 65 Possible and potential syllables in Czech Aleš Bičan, Masaryk University

The paper will discuss syllables in modern standard Czech. It will be argued that they can be described upon a model consisting of nine positions of which five are syllable-initial, one nuclear and three syllable-final. Each position can be either occupied by a phoneme belonging to a certain set, or empty (see the figure below; the capitals stand for consonants for which certain features such as voicing are predictable from the context). If a phoneme belongs to a position stretching over several other positions, the difference between those positions is irrelevant for it. For example, whereas /j/ can be preceded by four other consonants (cf. /FSkvjeT/ vzkvět “prosperity”), /ď/ can be preceded by two consonants only (cf. /FSďelāňī/ vzdělání “education”). By different ways the positions are filled with a phoneme or left empty, we get syllables in Czech. For example, in /a/ a “and” all syllable-initial and syllable-final positions are empty (i.e. /a/ = /a/), while in /FSkvjeT/ all syllable-initial positions are filled with a phoneme (i.e. /FSkvjeT/ = /FSkvjeT/; cf. also /Stan/ stan “tent” = /Stan/, /PŠtroS/ pštros “ostrich” = /PŠtroS/). ‘pre2’ ‘pre1’ ‘e3’ ‘e2’ ‘e1’ ‘n’ ‘i1’ ‘i2’ ‘i3’ kgxh rl PTK PTK v PTK T tdsz mnň jrl SŠ FSŠX M SŠ SŠ šž jř ieaou PTK FX Mjr l  Ť ř  rl īēāōū mnň ř vth   pbf  ëäö    ťď mnňrl However, not all combinations the model allows for correspond to attested and possible syllables in Czech, for example /tvjeT/, i.e. /tvjeT/ or /FŠtřīT/, i.e. /FŠtřīT/ or /Tpmel/, i.e. /Tpmel/. We can intuitively say that /Tpmel/ is not a well-formed syllable in Czech, whereas /TvjeT/ is (cf. attested /dvje/ dvě “two” or /kvjeT/ květ “flower”), though in the case of /FŠtřīT/ the intuition need not be sufficient. There are obviously certain restrictions limiting the combinability of phonemes. Thus, /Tpmel/ is ruled out because the labial occlusives cannot occur before the labial nasal syllable-initially in Czech, and so is /FŠtřīT/ because /ř/ cannot combine with the palatal fricatives within a single combination. The author has compiled a comprehensive list of these restrictions which will filter out ill- formed combinations out of the combinations the model allows. It will be argued that the result will be two kinds of combinations, i.e. two kinds of syllables: (1) all combinations and syllables attested in Czech (N.B. there are over 400 syllable-initial and 100 syllable-final consonantal combinations attested; the model generates all of them), and (2) combinations and syllables not attested but having such a structure that they are potential combinations and syllables Czech. That is, should there ever be such a need, the Czech grammar could make use of these combinations in brand new words. A simple computer program can be created for listing these combinations. 66 Futures and Futurates in Polish: A Force -Theoretic Account Joanna Błaszczak and Dorota Klimek-Jankowska, University of Wrocław

The issue As it is the case in many other languages, in Polish there are “real future forms”, which have only future time reference (see (1a,b)) and so-called “futurates” (see (2)), which as such are not future forms, but nonetheless can (at least in some contexts) refer to future-oriented eventualities (see Copley 2008, 2011). (1) a.Ajax zagra z FC Porto.simple future (SF) Ajax play.prs.perf.3sgwith FC Porto ‘Ajax will play with FC Porto.’ (it cannot mean: ‘Ajax is playing with FC Porto.’) b.Ajax będzie grał / grać z FC Porto. periphrastic future (PF) Ajax be.aux.3sg play.prt.impf.sg.m /play.inf.impf with FC Porto ‘Ajax will be playing with FC Porto.’ (it cannot mean: ‘Ajax is playing with FC Porto.’) (2) Ajax gra z FC Portofuturate Ajax play.prs.impf.3sg with FC Porto ‘Ajax is playing with FC Porto (at the moment of speaking or in the future).’ Given that language usually does not tolerate redundancy, the question is why there are so many ways of referring to future eventualities in Polish. In this talk we will show that even if the future forms in (1a) and (1b) are not mutually exclusive, there are nevertheless clear preferences as to when the form in (1a) and when the form in (1b) is used. The most important difference between these forms is that in contexts which express pre-planned or pre-arranged future eventualities only the latter form (i.e., (1b)) is good. Interestingly, the notion of a PLAN is also relevant in the case of futurates. As will be shown in the talk, these forms can only refer to future eventualities in contexts in which the existence of a PLAN is presupposed. Questions to be answered: If both the future form in (1b) and the futurate in (2) have something to do with PLANs, the question is whether these forms differ semantically, whether their interaction with a PLAN is the same, and what the nature of this interaction is. More specifically, we will address the following questions: Q1: Why is it so that if the existence of a PLAN is presupposed, futurates are preferred over PF ( = periphrastic future) and SF (= simple future)? Q2: Why is it so that if the speaker wants to convey the meaning that a future eventuality is consistent with some PLAN, but the existence of a PLAN as such is not presupposed (or it is not immediately recoverable from a context) PF is preferred over futurates and SF? Q3: Why is it so that in contexts in which a future eventuality cannot be pre-planned since we want to have an option of changing or preventing the future outcomes SF is preferred over futurates and PF? The goal of the paper is to answer these questions and in particular to clarify the role of plannability / PLANs in the cases under discussion. We will base our analysis on a force-theoretic model developed by Copley (2011) and Copley and Harley (2011). References:  Copley, B. 2008. The plan’s the thing: deconstructing future meanings. Linguistic Inquiry 39/2: 261-174.  Copley, B. 2009. The Semantics of the Future. NY: Routledge.  Copley, B. 2011. Causal chains for futures and futurates. To appear in: Kissine, M., De Brabanter, P., Sharifzadeh, S. (eds.), Proceedings of the Societas Linguistica Europaea Workshop “Future Tense(s) / Future Time(s)”, Oxford University Press, Ms., 2011.  Copley, B. and H. Harley. 2011. Force dynamics for event semantics: reifying causation in event structure, Ms. 67 Gender Stereotypes in Georgian and Russian Paremies: Concepts: deda /мать and mama/отец Irine Chachanidze, Akaki Tsereteli State University

For folk linguistics empirical material is presented in the form of the folk legends, epistolary pieces of writing, phraseologisms and etc. Paremiology has been chosen intentionally as the research subject _ it is located at the intersection of phraseology and folklore. Therefore, the study of proverbs and figurative expressions is very valuable for linguoculturology in general. Paremies are the texts that are accumulated in folk culture. The abovementioned texts clearly and distinctly present those segments of life and consciousness of the mankind that are connected to the perception of the laws and traditions of the society. Differences between men and women have always been the subject of interest since the ancient times. Certain stereotypes have been formed a lot of which is still preserved in paremies. There is one more issue in Language and Gender Studies concerning deda /мать and mama/отец (the words for mother and father in Georgian), and their cognitive range in language and culture. Cognitive linguistics has proved that both the process and outcomes of metaphorization have both special peculiarities and unique features. (Kikvidze,2002). When studying gender issues in paremies cultural differences are manifested more interestingly. Thus, we focused on two words denoting the gender in Georgian and Russian languages: deda/мать/ and mama/отец/. It is significant that apart from denotative meaning they possess quite interesting and noteworthy reference that in fact, indicates cultural differences. In terms of the research question and the material to be analyzed these concepts and their denoting lexemes in paremies provide extremely valuable data. (Кирилина, 1997). It is clear that: deda/мать and mama/отец can’t be analyzed along with the other words denoting a man and a woman. Due to the specificity that is conditioned by the social connotation of these concepts the above-mentioned words are the subject of metaphorization in various languages, but it does not necessarily mean that the metaphors derived from these words have one and the same shape in every language. For example: in Georgian: ‘deda naxe, mama naxe, švili ise gamonaxe’ [Like mother, like child]; in Russian, ‘Каковы родители, таковы и детки’. ‘ded-mama švilistvis, švili k’i tavistvisao’; [One father can better nourish ten children than ten children can nourish one father]; in Russian, ‘ Один отец прокормит десять детей, нежели десять детей одного отца’. The following materials have been used as the source for the presented paper: ‘Georgian- Russian and Russian-Georgian Proverbs’ (1976), Н.С. Ашукин, М.Г. Ашукина‘Крылатые слова’ (1960), ‘Пословицы русского народа’, сборник В. Даля (1984). References Coates, J. Women, Men and Language. A sociolinguistic account of gender differences in language. London: Longman, 1993. Hoenigswald, H. M. A proposal for the study of folk-linguistics. Sociolinguistics (Proceedings of the UCLA Sociolinguistics Conference, 1964), ed. by W. Bright. The Hague & Paris: Mouton, 1966: 16-20. Kikvidze, Z. - Language - made Paradox: Gender Preferences in Georgian. Women and Language, Volume XXV, Number 1, Spring 2002. Кирилина А. В. Женский голос в руссой паремиологии // Женщина в российском обществе, No 3, 1997. 68 L1 & L2 Processing of Clause-Internal Phrasal Scrambling in Russian Tatiana Luchkina, University of Illinois, Urbana-Champaign

Word order variability is used in morphologically rich languages to encode information status and discourse prominence (Watson 2010). While some languages (Italian) discourage special acoustic marking of ex-situ discourse-prominent constituents, others (Georgian) recruit additional prosodic licensing of prominent constituents independently of their position in the sentence (Skopeteas 2010). Psycholinguistic studies (Erdocia et al. 2009) find that processing scrambled sentences is context-restricted and often resource-intensive, which may explain why focus-driven constituent displacement may preclude concurring acoustic prominence. In the absence of prior experimental work, it is unclear whether positioning the prominent constituent ex-situ and marking it acoustically results in a yet higher processing load or, on the contrary, facilitates processing non-canonical word orders. Turning to L-2 acquisition, it is unknown if L2 learners of a ‘free’ word order language with no/little L1 experience of word order flexibility show a processing advantage when one vs. both prominence cues are present (scrambling only vs. scrambling & acoustic marking). The goal of our study is to test the relative accessibility of acoustically emphasized in- and ex-situ focused discourse constituents in Russian with both native Russian speakers and L1-English L2-Russian learners, and to compare the processing costs of the prominence cues in these populations. A Probe Recognition Task crossed canonical SVO and non-canonical OVS sentences (1) with 3 context types, matching, non-matching, and generic and 2 prosodic conditions, with or without emphasis, for a total of 12 conditions (8 tokens per condition) counterbalanced across two test versions. Probes were 1-3-syllable long lexical nouns matching or not matching the discourse-prominent NP. To date, 12 native Russian speakers and 13 upper-intermediate– advanced L2-Russian learners have been tested. A screening ensured that learners are familiar with all verbs used in the stimuli and can correctly interpret OVS sentences. All respondents were equally good at recognizing lexical probes independently of the test condition, indicating that they understood the task (Table 1). Both groups showed greater mean RTs in the generic context condition (t=2.63, p<0.01)7 . The L2 group only showed greater mean RTs in the OVS condition (t=1.92, p=0.05), independent of the context or emphasis (Figures 1,2). Figure 2 illustrates a dissociation of the context & emphasis effects for L1 and L2 respondents evident from greater mean learners’ RTs to the no-emphasis stimuli in the OVS condition (t=- 2.03, p<0.05) and faster OVS but slower SVO RTs to the no-emphasis stimuli in the L1 group (t=2.34, p<0.05, t=-2.12, p<0.05). In the matching context condition, mean RTs to accented SVO and unaccented OVS stimuli are nearly identical for the L1 group (Figure 2), which contradicts the Scrambling Complexity hypothesis (Sekerina 1999, Mishra 2011). Our findings reveal that acoustic cues were used opportunistically by the L2 speakers of Russian and made the OVS stimuli more accessible. Selective application of the acoustic cues by the L1 group provides novel evidence for a complementary relationship between acoustic prominence and constituent dislocation in Russian, invisible to even the proficient L1 English L2 speakers. We will discuss these findings in light of different theories of scrambling processing. 7 A follow-up acceptability judgment task has so far shown that the L1 and L2 respondents were sensitive to both, generic and non-matching context conditions. While the non-matching context condition received the lowest acceptability ratings (z=4.62, p<0.001) it did not slow down the reaction times in the Probe Recognition task because of introducing the first mention of the probe. 69 On Conditional Argument Clauses in Polish Lukasz Jedrzejowski & Kerstin Schwabe, Zentrum für Allgemeine Sprachwissenschaft

In this talk, we will deal with constructions where the propositional argument of the matrix pre- dicate is realized non-canonically, i.e., by a conditional clause (cf. 1.). We dub them (C)onditional (A)rgument (C)lauses and will discuss the syntactic and semantic status of CACs as well as the aspectual and semantic properties of the verb classes licensing them. First, we will show that a CAC is an adverbial clause, which restricts the matrix clause on the one hand and provides the propositional argument clause of the matrix predicate on the other (cf. Williams 1974 and Pullum 1987 for English, Fabricius-Hansen 1980 and Hinter-wimmer 2012 for German, and Quer (2002) for Spanish and Catalan). We treat CACs as adverbial clauses for the following reasons: (i) unlike canonical that-argument clauses, CACs cannot be adjoined to PPs (cf. 2). (ii) CACs cannot occur as fragmentary answers to complement questions relating to a given event as exemplified in (3). (iii) The paraphrase of constructions with CACs can contain a that-clause in the complement position of the matrix predicate (cf. 4). Accordingly, similar to Pesetsky (1991) we consider CACs to be a part of the VP being base-generated as VP- or vP- Adjuncts. Second, it seems to be common sense that CAC-licensors are predicates allowing a factive interpretation – e.g. żałować ‘regret’ in (6). The CAC σ in a construction with such a predicate is either regarded as a protasis of an implication σ  τ(σ) (cf. Williams 1974, Fabricius-Hansen 1980) or as the restrictor of an operator in the matrix clause: [ OPe: e involves σ]{e involves τ(σ)} (cf. Kratzer 1986 for ordinary conditionals and Quer 2002 and Thomson 2012 for CACs), respectively. We will show that both approaches do not account for constructions in which the matrix verb occurs as a preference predicate (cf. 5). From that, it follows that the potential factivity is not a necessary condition for a predicate to generate a CAC. Moreover, it is not a sufficient condition either since there are factive predicates like rozmyślaċ nad/o czymś ‘cogitate about’ that are factive, but do not admit CACs. In this connection, we will propose a new classification of CAC-licensors showing that CACs are, roughly speaking, licensed by verbs that (i) can be factive, but do not embed questions (ii) embed question extensions, (iii) are veridical and relate to question intensions, and (iv) embed contingent propositions. In the last part of our talk, we will link the interpretation of embedded CACs to the aspect category. Following Dahl (1995), who postulates a strong correlation between the imperfective aspect and generic sentences, and based on data extracted from the National Corpus of Polish, we will argue that if both perfective and imperfective opponents c-select CACs and if both of them are productively used as clause-embedding predicates, then the imperfective verb triggers a generic interpretation of the embedded CAC. This aspect-driven correlation does not hold, however, if the matrix verb is modalized by a modal verb, the future auxiliary będzie (cf. 6), an imperative form, etc. Finally, we will demonstrate that if both, the matrix verb and the embedded verb are imperfective, CAC never gets an episodic reading and it is always generic (cf. 7). (1) Duchy uwielbiają [jeśli się je czci] spirits.NOM adore.3PL if REFL they.ACC worship.3SG ‘Spirits adore being worshiped.’ (2) [P P Z tego, [że/*jeśli pojedziesz ze mną]], cieszę się bardzo from that that/if go.2SG with me be.happy.1SG REFL very ‘I’m very happy that you will go with me.’ (3) Co nie podoba się Magdzie? Że/*Jeśli Ania jest chora 70 what NEG like.3SG REFL M.DAT that/if A.NOM is ill ‘What doesn’t Magda like? That/*If Ania is ill.’ (4) Jeśli Magda jest chora, Ania nienawidzi tego, że ona jest chora if M.NOM is ill A.NOM hates that.GEN that she is ill ‘If Magda is ill, Ania hates it that she is ill.’ (5) Wolę jeśli/jeżeli/jak/gdy/kiedy pada prefer.1SG if falls ‘I prefer it when it rains.’ (6) Nie będzie żałował [jeśli mu zapłacimy] OK episodic/*generic NEG will.3SG regret.l-PTCP .M .SG if him.DAT pay.3PL ‘He will not regret it, if we pay him (for it).’ (7) Żałuję [kiedy nie wybieram morza] *episodic/OK generic regret.1SG.IPFV if NEG choose.1SG .IPFV sea.GEN ‘I (usually) regret (it), if I don’t choose the sea.’ References Dahl, Östen (1995): The marking of the episodic/generic distinction in tense/aspect systems, in: The Generic Book ed. by Gregory N. Carlson & Francis Jeffry Pelletier. Chicago: University of Chicago Press, pp. 412-416. Fabricius Hansen, Cathrine (1980): Sogenannte ergänzende wenn-Sätze. Ein Beispiel syntaktisch-semantischer Integration, in: Festschrift für Gunnar Bech zum 60. Geburtstag am 23. März ed. by Gunnar Bech. København: Institut for germansk filologi, pp. 61-83. Pesetsky, David (1991): Zero Syntax: vol. 2: Infinitives, Manuscript, MIT, 196. Quer, Josep (2002): Non-logical If, in: Current issues in Romance languages: selected papers from the 29 th Linguistic Symposium on Romance Languages (LSRL), Ann Arbor, 8 -11 April 1999 ed. by Teresa Satterfield, Christina Tortora & Diana Cresti. (Current Issues in Linguistic Theory 220). Amsterdam: John Benjamins, pp. 241-254. Williams, Edwin (1974): Rule Ordering in Syntax. Unpublished PhD thesis, MIT. 71 Rhetorical Strategies for the Discourse of Ethnicity in the Bulgarian Press Traci Lindsey, UC Berkeley

One interest of discourse analysis is the way in which power structures are perpetuated in language, including the language of the media; Teun van Dijk notes that people not only associate crime and violence with minority group members, but also “refer to the media to substantiate their prejudices” (1989: 220). Van Dijk (2008) speaks to the role of discourse in both racism and the denial of racism, making discourse analysis a fertile field for Bulgaria, contending with new interpretations of integration and non-discrimination fostered by membership the European Union. My current research focuses on the construction of discourse around ethnicity in the Bulgarian press, specifically regarding the two primary ethnic minorities, the Turks (9.4% of the population) and the Roma (4.7%). My data thus far speak more to the lexical choices made in discussing the Roma. Rhetorical strategies available to characterize Roma neutrally or negatively include the use of the preferred term ром and its variants rather than the pejorative циганин; my early research shows that lexical forms derived from ром now predominate in the press, a relatively recent phenomenon. The question remains, however, whether this necessarily correlates with a trend toward neutral discourse. My preliminary research does not indicate that journalists using terms derived from the politically correct ром are necessarily more neutral in their portrayals of Romani than those using the disfavored циганин (e.g., the term цигански барон ‘Gypsy baron’ has largely been replaced by the more “correct” ромски барон ‘Romani baron’, but “correctness” does nothing to improve the image of the subject and may function ironically). Another arguably pejorative term is the adjective мургав ‘swarthy’ used to indicate Romani ethnicity; for example, a recent article in Standart refers to мургави мла ежи ‘swarthy youths’ who steal everything suitable to be sold for recycling—whether or not explicitly stated, the youths’ Romani ethnicity is evident to the Bulgarian reader. These observations raise the question of whether journalistic language employed to report on Roma has indeed become more neutral, or whether other lexical choices substitute for pejorative language to convey negative attitudes. Stereotypes of the Turkish minority are less prevalent and less uniformly negative than those of Roma; instead, media reports on Turks tend to index metaphors of Bulgarian victimization and slavery conditioned by Bulgaria’s Ottoman past, or images of fundamentalist Islamic terrorists, referencing modern Western concerns. Oftentimes, little in the language of the newspaper articles seems to suggest these images, beyond the mere reference to problems associated with Turks or Moslems. The inferences made by readers are, however, clarified in comments posted with such articles. My research will employ several major Bulgarian online news sources, including both online version of the most widely circulated print newspapers (Тру , 4 Часа, Стан арт) and as sources available only online (Dariknews.bg). I will use both routine reports of certain types of events (e.g., crime, cultural events, developments with political parties) and systematic analyses of reporting on particular events associated with a minority group to analyze how Bulgarian newspapers mediate discourses of identity. References Van Dijk, Teun A. (1989). “Mediating racism: The role of the media in the reproduction of racism.” In Language, power and ideology: Studies in political discourse, ed. R. Wodak. Amsterdam: John Benjamins. (2008). Discourse and Power. New York: Palgrave MacMillan. 72 Regressive Palatalization and Contrast in Russian Jeff Parker, The Ohio State University

In Russian phonemically unpalatalized (hard) consonants may palatalize when directly followed by palatalized (soft) consonants. Previous studies (Drage 1967, Avanesov 1972, Krysin 1974) suggest that whether a consonant will regressively palatalize is affected by, at least, the articulatory similarity of the consonants and morphological position. For example, the initial consonant in [zj dj ]elat’ ‘to do-PERF’ is palatalized but in [vdj ]elat’ ‘fix into’ it is not (Timberlake 2002); /z/ and /d/ share place of articulation whereas /v/ and /d/ differ. It has also been noted that regressive palatalization is losing ground as a regular process (e.g. Timberlake 2004), though this has not been directly tested. In this paper I present new empirical data on the synchronic status of regressive palatalization. Employing an information theoretic approach, I also discuss increased phonemic contrast as a structural consequence of regressive palatalization becoming less regular. Binary categorization of sounds into phonemes and allophones is traditional to phonological theory, but this binary view has long been known to be problematic. A classic example is the status of soft velars in Russian. Soft velars only appear before non-front vowels in foreign borrowings and one Slavic root (e.g. tkёš’ - ‘you weave’). Velars thus exhibit some degree of contrast for palatalization, but not as systematically as other consonants. Recently, Hall (2009), using information theory, argued that contrast should be conceptualized in gradient terms. Another environment in which there is (traditionally) minimal contrast for consonant palatalization in Russian is before consonants. Labials and velars never exhibit palatalization before hard consonants, i.e. there is no possibility of phonemic contrast in this position for these sounds. And while some examples of soft nasals and dentals occur before hard consonants, 88.5% of all words with soft-hard consonant clusters have a lateral as the first consonant (Parker 2012). Laterals thus have greater potential for phonemic contrast before hard consonants. Clusters ending in a soft consonant are even more complicated because where regressive palatalization occurs the hard/soft contrast is neutralized for the first consonant in the cluster. Importantly, as the process loses ground in the language the degree of contrast increases. In this paper I present data from recordings of five native Russian speakers. Each speaker produced 206 word types (1030 word tokens). Word types vary by cluster type (combinations of /d/, /t/, /p/, /n/, /l/), palatalization (hard-hard, hard-soft, etc.) and token frequency (based on a 24 million token corpus of Russian). I establish an acoustic correlate of palatalization (change in F2 in the preceding non-high vowel) by comparing potential targets for regressive palatalization to unambiguously hard-hard clusters. I show that regressive palatalization is more likely for dental obstruents than for nasals than for laterals and labials. Some factors previously investigated (phonetic similarity of consonants and morphological boundary) also affect the likelihood of the process. I use frequency-based information theoretic measures to calculate the degree of contrast possible in clusters that end with a soft consonant and suggest that this contrast is increasing (and will continue to increase) as regressive palatalization declines in the language. 73 Sententional clitic hosts in Slavic languages Anton Zimmerling, Moscow State University for the Humanities

The paper proposal is focused on the problem of clitic hosts and discusses the correlation between types of clitic hosts vs types of clitics (second-position clitics vs verb-adjacent clitics). I argue that Slavic languages pattern with other world’s languages which have the option for attaching clause-level clitics and clitic clusters to an initial sententional constituent (IP or CP). This option is characteristic of temporal and conditional clauses in Slovene, cf. (1) and Bulgarian, cf. (2). (1) Slov [CP ko = smo =se vrnili] # =se = je =  e stemnilo. ‘We came back when it was already dark.’ (2) Bg [CP Dokato=se=čudeše kak da postŭpi], #=se =vŭrna žena=mu. ‘While he was wondering what to do his wife came back.’ Slovene examples like (1) can be explained in a different way without postulating sententional hosts for the main-clause clitics like =se=je=že in (1), since Slovene allows for clitic fronting in main clauses and seemingly does not obey the Tobler-Mussafia law, cf. (3b). (3) Slov a. #Videl=sem=ga. ‘I have seen him.’ (Default variant) b. #Sem=ga=videl. ‘I have indeed seen him.’ (Marked variant). However, this explanation is dubious, since there is no communicative emphasis and no motivation for clitic fronting in structures like (1). Therefore, there is no independent evidence that the preposed CP in (1) is merged after the main-clause clitics have been fronted to the clausal left edge. An alternative explanation of Bg examples like (2) is that Bg clusterizing clitics like the reflexive se in (2) are verb-adjacent proclitics and syntactic constituents to the left of them do not count, cf. [Franks 2008]. However, main-clause structures like *se=vŭrna žena=mu are ill- formed, while structures with a preposed CP like (2) are well-formed. This confirms that Bg obeys the Tobler-Mussafia law and suggests that the preposed sententional constituent may belong to the domain of the main-clause clitics, though this fact is not always recognized in the minimalist syntax. Sententional hosts (CPs and IPs) are marginally possible in other Slavic languages, notably in Czech and Upper Sorbian. These West Slavic languages also have parameters that license clitics after direct oration and after a pause (the latter option is sometimes explained as a ‘silent’ host). This is exactly the opposite of the situation in Serbo-Croatian where clause-level clitics immediately following a preposed CP are only possible after restrictive relative clauses that are pronounced without a pause, according to [Radanović-Kocić 1996]. I claim that the parameter licensing sententional clitic hosts is independent from parameters licensing clitics after a pause, clitics at the clausal left edge and clitic fronting, and argue that both languages with 2P-clitics and languages with verb-adjacent clitics can have sententional hosts. Recognizing sententional hosts is an economic solution for the description of Slavic languages, since the elimination of such hosts is not possible on a uniform basis. Slavic languages with sententional hosts taking main-clause clitics seem to falsify the claim raised in [Agbayani, Golston 2010] that categories hosting fixed-position enclitics always lie clause-external to them. References Agbayani В., Golston С. Second-position is first-position. Wackernagel’s law and the role of clausal conjunction. Indogermanische Forschungen. 2010, 115. Franks S. Clitic placement, prosody and the Bulgarian verbal complex. Journal of Slavic linguistics. 2008. 16/1. Radanović-Kocić V. (1996). The placement of Serbo-Croatian clitics: A prosodic A.L. Halpern, A.M. Zwicky (eds.). Approaching second: Second position clitics and related phenomena. Stanford, 1996. 74 Kosta P., Zimmerling A. (2012). Slavic clitic systems in a typological pers pective. Lilia Schürcks, Urtzi Etxeberria, Anastasia Giannakidou and Peter Kosta (eds.) The Structure of NP and Beyond (Studies in Generative Grammar). Berlin, London: de Gruyter. (to appear). 75 Acquisition of morphophonemic alternations in Russian verb system. Elena Kulinich, Université de Montréal

The present work discusses the acquisition of the Russian verbs, which involves, in most cases, the acquisition of the correlation between two stems: the past and the non-past tense (e. g., plakatj ‘cry’: plak-/plach-). Previous studies (i.e. Ceytlin 2009; Slioussar 2003) distinguished different types of verb overregularizations in Russian child speech, which imply stem modifications (e.g., risovaj-u instead of risuj-u) or inflection substitutions (e.g., posplj-ut instead of pospj-at ‘sleep’). In early stages of language acquisition (3-4 year olds), the «j-correlation» model corresponding to the correlation between the past and non-past stems of the most productive Russian verb class (e.g., chita-tj ‘read’ –chitaj-u) is preferred. This model is transparent (e.g., plaka-/plakaj-, risova- /risovaj-), i.e., it does not involve any morphophonemic alternations. The question then arises as to how children acquire these alternations. According to Ceytlin (2009), the main mechanism responsible for children’s overregularizations in Russian is the analogy based on two types of associations: one between stems and one between inflexions. However, this type of analogy does not give any explanation on how the connections between alternating stems are carried out. Another way to treat words with morphophonemic alternations is suggested by the dual-processing model (Marcus et al. 1992): the alternating stems should be simply memorized. A recent study on the acquisition of Russian verbs (Slioussar 2003) demonstrates that the dual-processing model proposed for the acquisition of English past tense does not work in Russian due to the more complex verb system including a range of productive, semi-productive and non-productive verb classes. The connectionist model (Rumelhart & McClelland 1986) and network model (Bybee 1995, 2001) were not tested on Russian verb morphology. The present study of new and nonce verbs is an attempt to show that Bybee’s network model implying connections between discrete units (phonological and semantic features) can account for the acquisition of morphophonemic alternations in the Russian verb system. Taking into account only palatalization (d/zh, s/sh, z/zh, etc.), I propose that alternating consonants create an independent type of connections relying on morphological context and distinct from stem connections, i.e., connections between stems become less important than connections between stem final consonants and a specific morphological context (which corresponds to Bybee’s “product-oriented schema”). The data suggest that the acquisition of new verbs (in adult language) that include alternating consonants may follow several possible scenarios, as shown below for the new verb frenditj ‘make friends’ in the 1sg. non-past form: (1) frend’u – no alternation (morphological context is irrelevant); (2) frenzhu – correct d/zh alternation (zh occurs in the 1sg. non-past); (3) frendzhu – incorrect: two members of connection are phonologically realized; (4) frendl’u – incorrect: connection to a wrong phonological class (here, a labial instead of a dental). From the above examples, I conclude that there are intermediate stages between overregularizations (when the morphological context is irrelevant) and the correct form with the strong connection of a consonant to a specific morphological context. The process of acquisition of morphophonemic alternations consists then in the reinforcement of associations between alternating consonants and morphological context. References Bybee, J. (1995). Regular morphology and the lexicon. Language and Cognitive Processes. Vol. 10 (5), 425-455 76 Bybee, J. (2001). Phonology and language in use. Cambridge. Cambridge University Press. Ceytlin, S.N. (2009). Ocherki po slovoobrazovaniju i formoobrazovaniju v detskoj rechi. Moskva, Znak (Studia Philologica). Marcus, G.F., Pinker, S., Ullman, M., Hollander, M., Rosen, T.J. & Xu, F. (1992). Overregularization in language acquisition. Monographs of the Society for Research in Child Development, 57(4) Rumelhart, D., & McClelland, J. (1986). On learning the past tenses of English verbs: Implicit rules or parallel distributed processing? In J. McClelland, D.Rumelhart & PDP Research Group (Eds), Parallel distributed processing: Explorations in the microstructure of cognition. Cambridge, MA: MIT Press. Slioussar N.A. (2003). Psiholingvisticheskoe issledovanie struktury mentaljnogo lexicona na materiale russkih glagolov. MA Thesis, Saint Petersburg State University, Department of General Linguistics. 77 Animacy and differential object marking in Old Church Slavonic Hanne Martine Eckhoff, University of Oslo

All case-marking Slavic languages have some version of the animacy category. Old Church Slavonic (OCS) provides its earliest attestation, displaying a situation where animacy does not constitute a (sub)gender, as argued for the modern Slavic languages. Instead, we have a situation where masculine singular o-stems denoting male persons may get genitive-accusative (GA) marking, as may masculine singular pronouns and nominalised adjectives/participles, but considerable variation is attested, and the nominative-accusative (NA) is found fairly frequently as well. (1) ašteženapouštšimǫža posagnetъ if woman released man.GA marries za inъ after other.NA “If a woman divorces her husband and marries another …” (Mk. 10:12) (2) idi prizovi mǫžь tvoi go summon man.NA your.NA “Go, summon your husband.” (Jh. 4:16) The extensive literature has argued that the GA-marking correlates with degree of social prominence (high-status adult men preferred), but also with pragmatic properties (definiteness, specificity). It is, however, not entirely clear what these pragmatic properties are. In this paper I argue that the animacy category in OCS is a fairly typical case of differential object marking: “overt case-marking of direct objects to mark some objects, but not others, depending on semantic and pragmatic features of the object” (Aissen 2003). To support this claim, I use parallel OCS and Greek data from the PROIEL corpus,8 which is provided with detailed morphosyntactic annotation, and also information status annotation. information status NA GA old 11 64 accessible 3 27 anchored 12 7 new 10 10 non-specific new 8 7 non-specific 1 10 old/accessible Case-marking of human masculine singular o-stem direct objects in the Codex Marianus by information status Data for OCS common masculine o-stem nouns denoting human beings show that there is a good correlation between GA-marking and the occurrence of Greek definite articles, but no strong tendency for indefinite Greek nouns to be rendered as nominative-accusatives. However, the correlation with information status is considerably better. Old and accessible objects strongly favour the GA, whereas new and anchored objects show variation. Specificity appears to play no role. There is a good deal of lexical variation, which supports the social prominence account, but many of the nouns occurring predominantly or exclusively in the GA are uniques and thus always at least accessible from world knowledge. GA- only nouns also have much higher average saliency than NA nouns. I therefore conclude that OCS had discourse-based GA marking which could easily be reinterpreted as lexicalised. 8 http://foni.uio.no:3000 78 Contrastive Focus in Child Russian and English Lydia Grebenyova, Baylor University

Many studies have been devoted to acquisition of the properties of focus operators like ‘only’, as in Crain et al. (1994), Philip & Lynch (1999), Patterson et al. (2003), and Zhou & Crain (2009) among many others. However, acquisition of contrastive focus without the presence of such operators has not been explored to the same extent. The present study attempts to bridge that gap and explore how children acquire the syntactic properties of contrastively focused R-expressions in English and Russian. The syntax of contrastive focus without overt focus operators differs across languages in a way that presents an intriguing learnability question. Contrastively focused R-expressions are fronted in Russian (and Slavic in general) while they remain in situ in English, as shown in (1) - (2). (Capitalization represents contrastive focus intonation.) This cross-linguistic variation raises a question of when and how this language-specific information is acquired by children. Using Elicited Production Test (Thornton, 1990), contrastively focused expressions were elicited from children and adults in carefully controlled contexts. The participants included 20 monolingual children acquiring Russian with the age range 3;3 – 5;7 (mean 4;4), 20 monolingual children acquiring English with the age range 3;6 – 5;6 (mean 4;5), as well as 20 adult controls for each language. The results show that English-speaking children have acquired the in-situ property of contrastive focus in their language successfully, which was evident from the fact that they never fronted a focused Rexpression. However, Russian-acquiring children seem to take longer in acquiring the fronting property of contrastive focus in Russian. These children were found leaving a contrastively-focused R-expression in situ 33% of the time, which significantly contrasted with the behavior of the Russian-speaking adult controls. To account for these findings, a parametric analysis is developed where the non-fronting parameter value for contrastive focus is the default value. This explains the early success Englishacquiring children demonstrate with the syntax of this construction. To account for the Russian data, I we explore a possibility of two syntactic focus positions in Russian (above and below TP) and their interaction with verb raising, suggesting that the seemingly post-verbal focus position in child Russian is not actually an in-situ position but rather is obviated by verb-raising. We also relate our results to the fact that Russian, unlike English, is a multiple wh-fronting language, as demonstrated in (3)-(4). There is an intriguing parallel between the results of this study and previous studies on the acquisition of multiple wh-questions in Russian, where it was found that Russianacquiring children fail to front one of two wh-phrases in multiple wh-questions 18% of the time (Grebenyova, 2006). We argue that contrastive focus and multiple wh-fronting are likely to be a result of the same process, following the theoretical accounts of Stjepanović (1998), Bošković (1998, 2002), and Stepanov (1998) developed on independent grounds. (1) a. IVANA ja uvidela. [Russian] IvanACC I saw ‘I saw IVAN.’ b. *Ja uvidela IVANA. (2) a. I saw JOHN. b. *JOHN I saw. (3) What did Smurf put t where? (4) Čto kuda Smurf položil t t? [Russian] what where Smurf put References Bošković, Željko. (1998). Multiple wh-fronting and economy of derivation. Proceedings of the 79 16th West Coast Conference on Formal Linguistics. E. Curtis, J. Lyle and G. Webster. (eds.), Stanford, CSLI Publications: 49-63. Bošković, Željko. (2002). On multiple wh-fronting. Linguistic Inquiry 33: 351-384. Crain, Stephen, Ni,Weijia & Conway, Laura. 1994. Learning, parsing and modularily. In C. Clifton, L. Frazer, & K. Rayner (eds.), Perspectives on Sentence Processing, 443-467. Grebenyova, Lydia. (2006). Multiple Interrogatives in Child Language. Proceedings of the Boston University Conference on Language Development (BUCLD 30). Paterson, Kevin B., Liversedge, Simon P, Rowland, Caroline & Filik, Ruth. (2003). Children’s comprehension of sentences with focus particles. Cognition 89, 263-294. Philip, William & Lynch, Emily. (1999). Felicity, relevance, and acquisition of the grammar of every and only. In S. C., Howell, S. A. Fish, & T. Keith-Lucas (eds.), Proceedings of the 24th annual Boston University Conference on Language Development. Somerville: MA: Cascadilla Press. Stepanov, Arthur. (1998). On Wh-Fronting in Russian, in Proceedings of the 30 th Meeting of the North-Eastern Linguistic Society. Stjepanović, Sandra. (1998). Short-distance movement of wh-phrases in Serbo-Croatian matrix clauses. Paper presented at the Comparative Slavic Morphosyntax Workshop, Spencer, Ind., June 1998. Thornton, Rosalind. (1990). Adventures in long-distance moving: The acquisition of complex whquestions. Doctoral dissertation, University of Connecticut, Storrs. Zhou, Peng and Crain, Stephen. (2009). Focus in Child Language: Evidence from the Acquisition of Chinese. In Proceedings of the 3 rd Conference on Generative Approaches to Language Acquisition North America (GALANA 2008), ed. Jean Crawford et al. 336-346. Somerville, MA: Cascadilla Proceedings Project. 80 Morphosyntactic Feature Economy in the Slavic Dual Tatyana G. Slobodchikoff, University of Arizona

I analyze the problem of disappearance and preservation of dual number in Slavic pronouns. Previous studies (Iordanskij 1960, Derganc 1988, Žolobov 1998, Nevins 2011) do not address the underlying mechanism responsible for the loss of dual pronouns in the majority of Slavic languages including Contemporary Standard Russian (CSR) and Contemporary Kashubian (CK), but excluding Contemporary Standard Slovenian (CSS), Contemporary Upper Sorbian (CUS), and Contemporary Lower Sorbian (CLS) (1). In my analysis, I provide a principled account of (i) why the dual was replaced by the plural in CSR and CK; (ii) why it was ‘renewed’ as a bi- morphemic dual in CSS, CUS, and CLS. In the framework of Distributed Morphology (Harley & Noyer 1999), I propose a new principle of Morphosyntactic Feature Economy (MFE) which applies to a marked feature combination [-singular, -augmented] of the dual to reduce its markedness. The morphosyntactic markedness of the dual is resolved in two different ways. In CSR and CK, a marked [-augmented] feature is deleted via impoverishment, which results in the syncretism of the dual with the plural. In CSS, CUS, and CLS, a marked [-singular, -augmented] feature combination is split via fission into two separate nodes – [-singular] and [-augmented] realized by two morphemes in a bi-morphemic dual. This difference in strategy resulted in the preservation of the dual in Slovenian and Sorbian, but in its subsequent reanalysis as plural and disappearance in Russian and Kashubian. (1) The 1st and 2nd Person Nominative Pronouns in CSR, CK, CSS, CUS, and CLS CSR CK CSS CUS CLS Singular ja, ty ja, të jaz, ti ja, ty ja, ty Dual - - mi-dva, vi- mó-j, wó-j me-j, we-j dva Plural my, vy më, wa mi, vi, oni my, wy my, wy 81 Two types of relative pronouns in Serbian Ivana M Mitrovic, Stony Brook University

This paper (i) presents novel empirical facts on the distribution of relative pronouns (RP) in Serbian relative clauses (RC): koji ‘which’ and ko/˘sto ‘who’/‘what’, and (ii) provides an analysis that captures the observed distributional patterns. I claim that the two sets of RPs occupy different syntactic positions. Koji is in SpecNP and exhibits the behavior of adjectival elements (agrees in '�features and case with an N it modifies). It introduces RCs with a (pro)noun relative head or an element showing the '�feature agreement with an unpronounced noun. Ko/˘sto is an N head (Fiengo 1980) and triggers a default 3rd person m/n sg agreement. The m/n gender reflects the [+/-human] feature specification. These pronouns strictly appear with relative heads that are demonstrative or indefinite pronouns (e.g. to, ovo, onaj, ne˘sto, niko) or entire clauses. Note that there is also a complementizer ˘sto (Gooduck and Stojanovic 1996, Bo˘skovic 2009, Gra˘canin-Yuksek 2010) which is an invariant word, allows the occurrence of resumptive pronouns and appears with (pro)noun relative heads. ˘Sto-RP, besides triggering the agreement with a verb, requiring certain types of relative heads and disallowing resumptive pronouns, is also case-marked and as such, exhibit all the properties of a pronoun (unlike the homophonic complementizer). I claim that demonstrative pronouns are base-generated in their own functional projection (DP) while indefinite pronouns are base-generated in QP (ne-, ni-, etc.). QPs take NPs ko/˘sto as arguments which raise to Q head due to Stranded Affix Filter (Lasnik 1981; Citko 2004). DPs take QPs as complements (and they can be phonologically overt or non-overt). The existence of functional projections in a nominal domain predicts that LBE, Adjunct Extraction and long-distance Scrambling out of them should yield ungrammaticality (Bo˘skovic 2005). This prediction is borne out. I adopt the matching analysis of RCs (Sauerland 1998). The internal and external relative head are related by ellipsis, the latter must be an antecedent of the former and an internal relative head is complex. The external relative head appearing with demonstrative and indefinite pronouns (ko/˘sto) have a matching internal relative head: the N ko/˘sto. The deletion of the internal relative head does not take place and in fact must not take place for the following reasons: (i) the internal RP is not similar enough to the external head (the external N raises to Q) and, (ii) it has a [wh]-feature that the RC CP requires. On the other hand, the (pro)noun external relative head has a matching internal relative head which has koji-RP in its Spec. The internal relative head hence can delete since (i) the external and internal relative heads are similar enough, (ii) the internal N undergoes '�feature agreement with koji-RP which inherits the features and case and, (iii) the internal relative head does not bear [wh]-feature (koji-RP does). It is in fact possible, though uncommon, that the internal relative head does not delete, as in roman koji roman ˘citam... ‘the novel which novel I am reading...’ (Browne 1986). The paper makes several theoretical proposals based on novel empirical data: (i) koji-RP and ko/˘sto-RP are in different syntactic locations, (ii) external relative heads match the internal ones (iii) there are two functional projections (DP and QP) in nominal domain in Serbian, which host elements that do not allow LBE, Adjunct Extraction or Scrambling. References Bo˘skovic, ˘Z. (2005). On the locality of left branch extraction and the structure of NP. Studia Linguistica 59:1-45. Bo˘skovic, ˘Z. 2009. On relativization strategies and resumptive pronouns. FDSL 7 Proceedings. Browne, W. (1986). Relative clauses in Serbo-Croatian in comparison with English. Zagreb: Institute of Linguistics, University of Zagreb. Citko, B. 2004. On Headed, Headless, and Light-Headed Relatives. NLLT 22:95-126. Fiengo, R. (1980). Surface Structure: The Interface of Autonomous Components. Cambridge, Mass: Harvard University Press. Goodluck, H. and D. Stojanovic (1996). The structure and acquisition of relative clauses in Serbo-Croatian. Language Acquisition 5(4):285-315. Gra˘canin-Yuksek, M. 2010. On a matching effect in headed relative clauses. In FASL 18 Proceedings, 193-209. Lasnik, H. (1981). Restricting the theory of transformations: a case study. Explanation in Linguistics , 152- 173. London:Longman. Sauerland, U. 1998. The meaning of chains. Diss, MIT. 82 Registering Survival: Bilingual Sorbs and Multistylism in Germany Elizabeth Spreng, University of Illinois, Urbana-Champaign / University of Kansas

Whether the Sorbs still speak the Sorbian language fifty years from now pales in importance compared to their current battles to maintain the ways that they mix registers. A Sorbian stagehand explained to me, “The Sorbs have been crying wolf for a thousand years, but the language is still not dead.” After “surviving” nationalism in the mid-nineteenth century, Nazism, and socialism, Sorbs overcame struggles to assert their linguistic identity in the face of German/Sorbian structural inequalities. Now living in EU multiculturalism, bilingual Sorbs experience what Mikhail Bakhtin described as the "contradictions in a multi-languaged world.” Through an anthropological approach focused on linguistic questions, I consider how Sorbs (currently numbering 10,000 Upper Sorbian speakers), erect, erase, and blur boundaries between linguistic codes/varieties. Without losing sight of the problematics of language death, my paper explores how Sorbs vitalize the Sorbian language using an array of linguistic practices via register variation. By upholding two complementary registers, bilingual Sorbs are not experiencing monostylism, but rather maintaining linguistic virtuosity and variation through, what I identify as, multistylism or register mixing. In translating Munro Leaf’s classic children’s tale, The Story of Ferdinand; Sorbs simultaneously used markers of a village and a Budyšin/urban register. On one hand, Sorbs preserve a village register envisioning a spoken idealized form, affiliated with Sorbian-speaking villages, learned as a child, and strengthened by elementary/secondary Sorbian education. On the other hand, the Budyšin register entails an imagining of modernized written Sorbian language use, developed through Sorbian higher education, and associated with institutional sites located in Budyšin (Bautzen, Germany). Furthermore, users of the Budyšin (urban) register may critique the village register as impure or backward, but villagers may assess speakers of the Budyšin register as not sounding Sorbian enough. Aware of the possibility of sounding too German or not like the Sorbian practices used in the villages, bilingual Sorbs constructed their translations of Ferdinand according to these ever-present ideologies. Through analysis of their translations of Ferdinand, I detail how bilingual Sorbs negotiate notions of a village vernacular and expert Sorb, Sorbian/German grammatical structures, orthographic possibilities, Sorbianized declinations, and German borrowings. Thus, I argue that Sorbs use two registers that complement one another using intralingual/Sorbian distinctions and challenge an analytical perspective focused on code-switching between German and Sorbian resources. From a broader perspective, a consideration of multistylism offers another perspective on language contact, bilingualism, and language endangerment. As a corollary, multistylism, as an interdiscursive strategy, is characterized by mono-/bilingual sensibilities, dual enactments of authority and expertise, and tensions among standards. By drawing on discussions of blurred codes, syncretic practices, and enregisterment, my paper exposes the linguistic and cultural details of register mixing via multistylism. Moreover I argue that since the collapse of socialism, linguistic tensions in Eastern and Central European national communities including the Sorbs warrant more attention. 83