ABSTRACT. The presentation will discuss current issues in the historical interpretation of inflectional morphology. There are relatively new issues and there are old ones. Among the former is the problem of forming a unified understanding of (i) affixal morphology and the non-affixal techniques of (ii) vowel or consonant mutation and (iii) phonotactic modification (Trommer and Zimmermann 2015). The diversity in techniques calls for a processual understanding of inflection (Mel’čuk 2006), which in turn prompts a renewed discussion of the ontology of inflectional paradigms. Among the old issues are the common deviations from the ideal biuniqueness of morphological signs in inflectional systems—zero exponents, exponents without apparent content, and overlapping and multiple exponence—long discussed in the theoretical literature (e.g. Matthews 1972; Baerman 2015; Harris 2017). Against the background of the predominantly static, synchronic approach in most recent theoretical work I will take up Aronoff’s (1994: 169) suggestion that the complexity of inflectional morphology ”may hold the key to knowledge”. I will argue that a necessary first step is to adopt a consistently dynamic, truly historical perspective on synchronic data, one that takes due account of any observable synchronic variation in inflection. A second step is to push exponent analysis beyond the morphemic level wherever possible (Jakobson 1958), and a third, to interpret the data in semiotic terms (Anttila 1972). It can be shown that a dedication to detailed analysis and an effort to reveal the iconic and, especially, indexical relations that define morphological patterns are not only appropriate to the object of investigation, but fruitful, as Aronoff surmised. For diachronic linguistics such an approach will yield explicit characterizations of the different kinds and degrees of morphological cohesion along the agglutination – fusion – symbolism scale and clear interpretations of individual exponence changes along that dimension, that is, a better understanding of the innovations that give rise to such changes. Anttila, Raimo. 1972. An Introduction to Historical and Comparative Linguistics. New York: Macmillan. Second edition: Historical and Comparative Linguistics. Amsterdam, Philadelphia: Benjamins, 1989. Aronoff, Mark. 1994. Morphology by Itself. Stems and inflectional classes. Cambridge MA, London, England: MIT Press. Baerman, Matthew (ed.). 2015. The Oxford Handbook of Inflection. Oxford: University Press. Harris, Alice C. 2017. Multiple Exponence. Oxford: University Press. Jakobson, Roman. 1958. Morfologičeskie nabljudenija nad slavjanskim skloneniem (Sostav russkix padežnyx form). In: American Contributions to the Fourth International Congress of Slavists, 127–156. ’s-Gravenhage: Mouton. Reprinted in his Selected Writings, vol. 2. Word and Language, 154–183. The Hague, Paris: Mouton. Matthews, Peter H. 1972. Inflectional Morphology. A theoretical study based on aspects of Latin verb conjugation. Cambridge: University Press. Mel ́čuk, Igor’ A. 2006. Aspects of the Theory of Morphology, ed. by David Beck. Berlin, New York: Mouton de Gruyter. Trommer, Jochen and Eva Zimmermann. 2015. Inflectional exponence. Baerman 2015: 47–86.

ABSTRACT. The Slavic Reflexes of the PIE Syllabic Sonorants Daniel E. Collins, The Ohio State University

Henning Andersen has produced a number of major studies (e.g., 1968, 1969, 1996, 1998, 2003, 2009) that have significantly reshaped our understanding of Pre-Slavic and Common Slavic sound changes. Among the problems that he has investigated is the diphthongization of the Proto-Indo-European syllabic sonorants *r̩, *l̩, *m̩, and *n̩, which were reflected in Proto-Slavic (cf. Andersen 1985) as sequences of either *ĭ or *ŭ plus sonorant (see Vaillant 1950; Browning 1989; Schenker 1995). These diphthongizations undoubtedly occurred during the period of shared Baltic-Slavic developments, since the Baltic languages show the same dual outcomes, often in the same roots. While the diphthongization of syllabic consonants is a well-attested type of change (cf. Andersen 1972; Fougeron and Ridouane 2008; Toft 2002; Weise 1995), the Balto-Slavic case has provoked controversy for over a century because no regular phonetic conditioning has been identified for the bifurcating reflexes, despite numerous attempts (see, e.g., Shevelov 1964; Moszyński 1969; Kortlandt 2007, 2008; Matasović 2004). Some of the cases of *ŭ plus sonorant have been explained as onomatopoetic formations. Others have been treated as back-formations from full-grade ablaut alternations, given that the syllabic sonorants were typically zero-grades. Andersen (2003), building on Stang (1966), has plausibly proposed that many of the *ŭ outcomes are the result of prehistoric language contact. My goal in this paper is to offer a complementary explanation that, without negating the contact scenario (see also Andersen 1996, 2009), allows for the possibility that some of the dual outcomes arose by phonological reinterpretation during ordinary language acquisition. The alternative explanation utilizes the principles of Abductive and Deductive change and follows the approach to language change that Andersen has defined in a series of seminal articles (1972, 1973, 1978, 1989).

References Andersen, H. 1968. IE *s after i, u, r, k in Baltic and Slavic. Acta Linguistica Hafnensia 11: 171–90. –––––. 1969. Lenition in Common Slavic. Language 45: 553–74. –––––. 1972. Diphthongization. Language 48: 11–50. –––––. 1973. Abductive and Deductive Change. Language 49: 765–93. –––––. 1978. Perceptual and Conceptual factors in Abductive Innovations. In Recent Developments in Historical Phonology, ed. J. Fisiak, 1–22. The Hague. –––––. 1985. Protoslavic and Common Slavic. International Journal of Slavic Linguistics and Poetics 31–32: 67–80. –––––. 1989. Understanding linguistic innovations. In Language Change, ed. L. Breivik and E. Jahr, 5–27. Berlin. –––––. 1996. Reconstructing Prehistorical Dialects. Berlin. –––––. 1998. The Common Slavic Vowel Shifts. American Contributions to the Twelfth International Congress of Slavists, 239–48. Columbus. –––––. 2003. Slavic and the Indo-European Migrations. In Language Contacts in Prehistory, ed. H. Andersen, 45–76. Amsterdam. –––––. 2009. The Satem Languages of the Indo-European Northwest. First Contacts? In The Indo-European Language Family, ed. by A. Marcantonio, 1–31. Washington D.C..

Browning, T.. 1989. The Diachrony of Proto-Indo-European Syllabic Liquids in Slavic. Doctoral Dissertation. University of Wisconsin, Madison.

Fougeron, C., and R. Ridouane. 2008. On the Phonetic Implementation of Syllabic Consonants and Vowel-less Syllables in Tashlhiyt. Estudios de Fonética Experimental 18: 139–75.

Kortlandt, F. 2007. The Development of the Indo-European Syllabic Resonants in Balto-Slavic. Baltistica 42: 7–12. –––––. 2008. Balto-Slavic Phonological Developments. Baltistica 43: 5–15

Matasović, R. 2004. The Proto-Indo-European Syllabic Resonants in Balto-Slavic. Indogermanische Forschungen 109: 337–54.

Moszyński, L. 1969. K razvitiju praslavjanskix sonantov. Voprosy jazykoznanija 1969, no. 5: 3–10.

Schenker, A. 1995. The Dawn of Slavic. New Haven.

Shevelov, G. 1964. A Prehistory of Slavic. New York.

Stang, C. 1966. Vergleichende Grammatik der baltischen Sprachen. Oslo.

Toft, Z. 2002. The Phonetics and Phonology of some Syllabic Consonants in Southern British English. ZAS Papers in Linguistics 28: 111–44.

Vaillant, A. 1950. Grammaire comparée des langues slaves. Vol. 1. Paris.

Wiese, R. 1996. The Phonology of German. Oxford.

ABSTRACT. Historical linguistics and the study of language change have always dealt with a basic division in the types and causes of linguistic diachrony, namely changes motivated from within the linguistic system and those motivated from outside the system. This distinction has taken various terminological forms, most notably “evolutive” vs. “adaptative” change (Andersen 1973) and “transmission” vs. “diffusion” (Labov 2007), but it is seen also in the use of notions like “internal” vs. “external” change (e.g. in the literature on language contact), “organic” vs. “inorganic” (as found in nineteenth-century writings on language states, especially regarding the effects of language contact), and “endogenous” vs. “exogenous” change (popular now in, for example, phylogenetically inspired linguistic investigations).

Such a distinction is to be expected, given that language is both an individual (psychological/cognitive) entity and a social entity, and recognition of both dimensions to change can be argued to be necessary; as Hamp (1977:279) has observed, they are complementary, not competing—“twin faces of diachronic linguistics”, two key ways of elucidating sources of historical similarities and differences between languages. That is, we can only understand what is due to contact by having a clear picture of what is inherited, and conversely, to determine what is inherited, we need to rule out contact-related similarities.

Given the differing terminology, it is worth asking whether these various terms are nothing more than a reflection of terminological creativity on the parts of the coiners and users, or instead reflect essential conceptual differences or distinct nuances of perspective.

In this paper, we explore this very question, and given the historical priority that Andersen’s terminology has within modern linguistics, we start with the further question of whether, over the now nearly fifty years since he proposed his dichotomy, these notions and these terms have withstood the test of time. Moreover, we explore how Andersen’s conceptualization is similar to or different from those that came before him and those that followed, and thus the extent to which this is just an issue of terminology as opposed to substantively different conceptualizations.

At the same time, though, we delve further into this distinction and interrogate it from different vantage points, asking:

• whether these distinctions (e.g. internal vs. external) apply to the initial motivation for a change or to the spread of a change, or both;

• how to make clear what exactly is being talked about in particular cases, as many who recognize the distinction between initial motivation and spread nonetheless shift back and forth between the two opposed notions or even at times seem to collapse them;

• whether “external” refers just to other speakers in general (including of one’s own dialect) or specifically to speakers of a dialect or language that is in some sense sufficiently different from one’s own;

and finally:

• whether there really is an overall validity or necessity to the distinction between internal and external change, and whether it must be the case that changes are motivated only in one or the other way; cases of enhancement of tendencies present in a language due to contact, as with the spread of evidentiality into Balkan Slavic under Turkish influence, which accentuated existing characteristics emerging in the languages at the time (Friedman 2006), would appear to present a challenge to such a binary view.

Overall, these concerns add up to a rigorous investigation of a key distinction in historical linguistic theorizing and practice, and allow for an assessment of the value of a key element in the oeuvre of Henning Andersen as part of a special session honoring him.

References:

Andersen, Henning. 1973. Abductive and deductive change. Language 49.765–93. DOI: 10.2307/412063.

Friedman, Victor A. 2006. Balkanizing the Balkan Sprachbund. Grammars in contact, ed. by Alexandra Y. Aikhenvald and R. M. W. Dixon, 201–9. Oxford: Oxford University Press.

Hamp, Eric. 1977. On some questions of areal linguistics. Berkeley Linguistics Society 3.279–82. DOI: 10.3765/bls.v3i0.3295.

Labov, William. 2007. Transmission and diffusion. Language 83.344–87. DOI: 10.1353/lan.2007.0082.

ABSTRACT. Previously-described cases of contact-induced changes in constituent order have shown that languages with flexible orders become more rigid due to contact (Heine 2008). This project examines ongoing contact between Malayalam (Dravidian) and English as spoken by Malayalam-dominant bilinguals in Kerala, the Malayalam-speaking state in India. I present results from acceptability judgment tasks conducted in Kerala which show that age, which correlates negatively with measures of English contact, correlates positively with flexibility: increased contact with English results in decreased flexibility in Malayalam.

Malayalam as spoken in Kerala has been in intense contact with English for generations, which has resulted in changes at the level of the lexicon and phonology (Kala 1977, Asher & Kumari 1998). The canonical (basic, discourse-neutral) order in Malayalam is SOV, but all of the other 6 logical orders are grammatical, and changing the order does not result in a change in truth-conditional meaning. Given this, how do we expect experience with English to affect constituent order in Malayalam? I consider two alternatives. It could be that Malayalam speakers with a high level of English contact are borrowing the canonical SVO order from English wholesale, which would result in a preference for SVO. Alternatively, high-contact Malayalam-speakers could be less tolerant of movement in general, resulting in an increased preference for the canonical order in Malayalam, SOV. I used acceptability judgment tasks to measure relative preference.

44 native-speakers of Malayalam residing in the same district of Kerala participated in this study. The task was conducted in participants' homes or places of work, and language background surveys were conducted with each participant. Participants were aged from 18-82, and care was taken to ensure that speakers were from a variety of backgrounds and socio-economic statuses.

Participants rated the six logical variants of transitive sentences on a 7-point scale. Malayalam has no subject-verb agreement and differential object marking, so each experimental stimulus (sample stimuli in 1-6 below) had three constituents: an animate subject, an inanimate object, and a verb. This ensured that the semantic role of each argument would be unambiguous. Stimuli were distributed among six lists pseuorandomly using a Latin Square. Audio stimuli were used, to avoid possible effects of diglossia. Each participant heard five tokens of each condition and 40 filler items of varying acceptability, resulting in 70 total items.

The 1-7 ratings were transformed into z-scores in order to account for individual differences in use of the scale. In Figure 1, the data is split into "Young'' and "Old'' based on a median split in age. The "Young'' participants rate the canonical SOV order significantly higher (p<0.0001 for each) than they do the other non-canonical orders, reflecting less flexibility than ``Old'' participants, who do not rate SOV any differently from other non-canonical orders (Fig. 1). In Figure 2, treating age as a continuous variable, the difference between SOV and OSV is negatively correlated with age; this correlation is small (Pearson's r=-.27) and marginally significant (p<0.064). Age correlates with measures of language contact in this population (medium of grade school instruction, level of engagement with English media, attitude towards code-switching and language mixing), suggesting that language experience could be a source of variation in constituent order.

The data presented here is evidence for a shift from flexible to rigid in a case of ongoing language contact, and it demonstrates that the flexible-rigid pattern cannot be described by wholesale borrowing of the surface word order of the contact language; Malayalam speakers who are English-proficient are not just translating English sentences into Malayalam, rather, contact with English is affecting how speakers treat the canonical order in Malayalam.

ABSTRACT. This presentation explores the operation of child language acquisition as a critical factor in explaining dialect change. It focuses on two well documented but poorly understood changes in the history of Latin American Spanish, one phonological and one morphological, and helps to illuminate the most puzzling aspects of these changes by bringing to bear data from language acquisition. The phonological change in question is the simplification of the system of sibilants in 16th century Colonial Spanish (CS), while the morphological change is the spread of voseo (i.e., the etymological second person plural of medieval Castilian) as the default singular form of address in Río de la Plata Spanish (RPS). Both shifts were nestled in ecological environments characterized by intense contact among L1 and L2 speakers of several varieties of Iberian and non-Iberian Romance varieties, as well as the rapid breakdown and reshaping of social networks. Contact in both situations led to feature pools that were markedly different from those of the contributing demographic groups. We argue that, in the absence of strong normative pressures, the acquisitional advantages of certain forms were crucial in their eventual adoption and generalization.

The argument that children have a role to play in language change processes is not new. Generative approaches to language change (Kiparsky 1968, Lightfoot 1999, Kroch 2005) have identified reanalysis by children acquiring a language as the primary trigger to the generation and transmission of grammatical innovations. Within this view, language change results from an entire generation of children building a grammar different from that of their parents. Elsewhere, however, first language acquisition is usually not recognized as a major factor in explaining language change. This alternative argument can be summarized as follows: in the presence of consistent adult input, children are able to reconstruct the grammatical system of their caregivers. Analogical errors or reanalyses typical of child language are usually discarded as children fine-tune their grammars and approximate those of the older generations (Aitchison 2001 [1981]: 201-210). Within this view, the reasons for language change must be sought instead in later stages of life, either in late childhood and early adolescence, as social networks are developed beyond the home (Eckert 1988, Kerswill 1996), or in the incremental changes in the frequency distribution of language variants in competition throughout the lifespan (Tagliamonte and D’Αrcy 2007).

Within this second line of thought, the driving assumption is that the input that children receive is consistent enough to provide a target grammar, as commonly described in studies of monolingual L1 acquisition (Clark 2003). However, the input that children receive in order to build their grammars is always variable (Yang 2002). In demographically stable situations, child acquisition has been argued to operate as a mechanism of language change via the incremental advancement of changes in progress already present in adult input (cfr. transmission, Labov 2007). For less demographically stable settings (i.e., situations of sudden and/or pervasive language or dialect contact), the literature has been even more open to the possibility that L1 acquisition may operate as a significant factor in determining the direction and progression of language change. These situations include large-scale resettlement events (Tuten 2003), the formation of colonial koinés (Trudgill 2004), the speciation of colonial languages into creole languages (DeGraff 1996) and the development of new sign languages (Senghas & Coppola 2001). In these situations, children grow up in the presence of multiple variants of heterogeneous provenance (often within the household) and, consequently, the target of acquisition can be assumed to be less clear than in more stable demographic situations. The lack of a fixed prestige norm also means that children do not have the same motivation to conform to any given models of linguistic behavior once their social networks start including peers beyond the immediate household environment.

The unstable situation described above obtained in both of the cases we explore in this paper. In CS, the pool included a wide array of articulatory realizations for the original sibilant classes, representing several degrees of merger across the etymological voicing and place of articulation distinctions. In RPS, there was variation between at least two main options for informal second person singular address, with additional phonetic and morphological complications resulting in a mixed input. Quantitative data from written corpora will be presented to illustrate the initial mixture and the progression of both changes. These data will then be interpreted in light of acquisition studies showing that contemporary children favor these same alternatives when they are presented with options similar to those available in these historical situations.

For the case of sibilant mergers in CS, the evidence comes from the path of acquisition of complex sibilant contrasts in monolingual children learning Mandarin (Li 2007) and Polish (Buckmaier and Harrington 2016). In both of those languages, the adult systems feature a three-way system of sibilants very reminiscent of that of pre-colonial Spanish. Children acquiring Mandarin or Polish have been documented to be at pains to master these contrasts, with full phonemic distinctions arising very late in the acquisition process. Additional data from bilingual children in situations of contact between sibilant systems with different degrees of complexity will provide supporting evidence that these contrasts are especially unstable when children have abundant evidence of more than one system of sibilants in their adult input (e.g., Taiwanese-Mandarin; Shih 2012; English-Polish, Marecka et al. 2015). Sociodemographic information from the early Spanish colonial period will be used to argue that CS developed in precisely such an environment, with a variety of sibilant systems available as acquisition targets to children and no social mechanism in place to enforce any dialectal norm.

For the second person variation in the RPS verbal paradigm, what needs to be accounted for is the progression of voseo, whose historical spread from imperative to present indicative to subjunctive has been robustly corroborated (Fontanella 1989, Moyna 2009, Moyna & Ceballos 2008). This sequence is supported by modern Latin American dialects, since voseo variants are cross-dialectally more frequent in the imperative than in the present indicative, and least frequent in the subjunctive. The parallel between historical and dialectal data can be accounted for elegantly by invoking the acquisition of the verbal paradigm, which has been shown to proceed in the same sequence. It will be argued that the extreme level of inflectional variation present in the RPS mix, together with the absence of normative pressures in a marginal area of the Spanish Empire subject to massive subsequent waves of non-Hispanic immigration, gave children several options to choose from and allowed for the process of acquisition itself to operate in the selection of target variants.

To summarize, in both changes presented here, children exhibited an acquisitional bias that favored the options that ultimately prevailed. In turn, this bias can be explained as a function of saliency as a universal principle of child language acquisition, at both the phonological and the morphological level (Kehoe 2001, Dressler 2005). Although these data do not imply that children are the exclusive agents in these changes, they demonstrate that child acquisition can play a critical role in the actuation of language change. Finally, this study supports the view that children are not innovators, but propellants of changes already in progress in adult input, via the stabilization of the options available in their linguistic environment (Mufwene 2008).

Selected References

Bukmaier, Véronique, and Jonathan Harrington. 2016. The articulatory and acoustic characteristics of Polish sibilants and their consequences for diachronic change. Journal of the International Phonetic Association 46: 311-329. Fontanella de Weinberg, María Beatriz. (1987). El español bonaerense. Cuatro siglos de evolución lingüística (1580–1980). Buenos Aires: Hachette. Li, Fangfang. 2008. The Phonetic Development of Voiceless Sibilant Fricatives in English, Japanese, and Mandarin Chinese. Unpublished PhD Dissertation. Ohio State University. Marecka, Marta; Wrembel, Magdalena; Otwinowska-Kasztelanic, Agnieszka; and Dariusz Zembrzuski. 2015. Phonological development in the home language among early Polish-English bilinguals. In The Scottish Consortium for ICPhS 2015 (eds.), Proceedings of the 18th International Congress of Phonetic Sciences. Glasgow: The University of Glasgow. Online publication (accessed 11/1/2015) Moyna, María Irene. 2009. “Child Acquisition and Language Change: Voseo Evolution in Río de la Plata Spanish.” Proceedings of the 2007 Hispanic Linguistics Symposium. Joe Collentine, Barbara Lafford, MaryEllen García, and Francisco Marcos Marín (eds), 131-142. Somerville, MA: Cascadilla. Also available at: http://www.lingref.com/cpp/hls/11/index.html Moyna, María Irene, and Beatriz Ceballos. 2008. Representaciones dramáticas de una variable lingüística: Tuteo y voseo en obras de teatro del Río de la Plata (1886-1911). Spanish in Context 5 (1): 64-88. Senghas, Ann, and Marie Coppola. 2001. Children creating language: How Nicaraguan Sign Language Acquired a Spatial Grammar. Psychological Science 12 (4): 323-328. Shih, Ya-ting. 2012. Taiwanese-Guoyu Bilingual Children and Adults’ Sibilant Fricative Production Patterns. Unpublished PhD dissertation. Ohio State University.

ABSTRACT. As many other languages, the Spanish language displays a lot of devices or strategies to express possession. Typologically, we know the adnominal possession where possessor and possessed are coded both in the same Noun Phrase, exemplified in (1), the possessive predication, as the one in (2) characterized by the presence of a possessive verb as tener ‘to have’, and the one known as external possession (3), in which the relation is coded without a possessive verb and without the concurrence of possessor and possessed in a Noun Phrase: (1) [El auto de María] es nuevo. [Mary’s-Possessor] [car-Possessed] is new. (2) [María] tiene [una casa nueva] [María-Possessor] has [a new house-Possessed]

(3) a. Lei lastimaron [la mano] a Maríai Heri (they) hurt [the hand-Possessed] [to Maryi-Poseedor]= Mary’s hand b. Los soldados lei mataron a[l hijo] [a Maríai] The soldiers heri killed [the son-Possessed] [to Mary- Possessor] c. Los ladrones lei destruyeron [la casa] [a Maríai] The thiefs heri – destroyed [the house-Possessed] [to Mary-Poseessor] The paper analyses the diachronic evolution from one of the basic type of external possession predication in the languages of the world: the ditransitive o bitransitive predications in which both objects, direct and indirect are coded by Noun Phrases, as the ones exemplified in (3). In all of the constructions in (3), the possessor is codified as an argument of the construction to Mary ‘a María’, correfential with the clitic lei, whereas the possessed is codified by other argument, the direct object the hand, the son, the house in (3a), (3b) and (3c), respectively. These are examples of the so named external possession (Payne y Barshi, 1999: 3). The io is a possessive dative, that is, a constituent, not required by the argumental structure (Payne y Barshi, 1999: 3), that express the possessor of the direct object (Alcina y Blecua, 2001/1975: §7.2.13; rae-asale, 2009: §35.7f, 2010: §35.4.1b). The corpus is made up by data of the Spanish language from 12th to 20th centuries. The analysis shows that bitransitive construction with possessive dative is documented in all centuries and gives information about the properties of the predication, the verbs and the possessor and possessed that concurs as indirect and direct arguments, respectively. The analytic approach assumes that transitivity is a scalar phenomenon (Hopper and Thompson 1980), in which the distinction between actantial dative and non-actantial dative is irrelevant. In our perspective, ditransitive constructions with possessive dative are an extension of the prototypic ditransitive (Goldberg 1995). Our proposal supports that the construction fulfills a specific communicative need, that of giving relevance to the possessor over the possessed (Company 2001 and Ortiz 2006, 2011).

Bibliography Alcina Franch, Juan y José Manuel Blecua. 2001/1975. Gramática española, 11th edition, Ariel: Barcelona. Company Company, Concepción. 2001. “Multiple dative-marking grammaticalization. Spanish as a special kind of primary object language", Studies in Language, 25, no. 1, pp. 1-47. Goldberg, Adele E. 1995. A Construction Grammar Approach to Argument Structure, Chicago/London: The University of Chicago Press. Hopper, Paul J. y Sandra A. Thompson. 1980. “Transitivity in grammar and discourse”, Language 56, pp. 251-99. Ortiz Ciscomani, Rosa María. 2006. “La Bitransitividad”, Cap. 7, in Sintaxis Histórica de la lengua española, Primera parte: la frase verbal, Vol. I, Concepción Company (ed.), México: UNAM/FCE, pp. 575 - 668. _______________________. 2011. Construcciones bitransitivas en la historia del español, México: Universidad Nacional Autónoma de México/Universidad de Sonora. Payne, Doris L. e Immanuel Barshi. 1999. “External possession. What, where, how, and why”, en External Possession, Doris L. Payne e Immanuel Barsh (eds.), Amsterdam/Philadelphia: John Benjamins, pp. 3-29. [RAE-Asale]= Real Academia Española-Asociación de Academias de la Lengua Española. 2009. Nueva gramática de la lengua española, Madrid: Espasa-Libros.

ABSTRACT. In this talk I will investigate to what extent our knowledge of Old French grammar has been determined by the choice of texts that we generally use in our analyses. This topic is partly motivated by recent research in Latin linguistics (Adams 2013) that has revealed that our focus on so-called Vulgar Latin has affected our concept of the sociolinguistic and diachronic varieties of Latin, their interrelationship, and their subsequent development. Moreover, investigating compound verb forms in informal Old French, Bauer (In Prep.) has found a number of remarkably modern characteristics that in fact predate linguistic change as identified in the handbooks, warranting further research. In "Social variation and the Latin language" (2013) Adams identifies and analyzes the different social varieties of Latin, convincingly showing that language change was not the prerogative of informal language. A number of structures that traditionally have been ascribed to “Vulgar Latin” turn out to be exclusive to Classical Latin and several of the 30 linguistic phenomena that Adams investigates, did not originate in informal, but rather in educated varieties of Latin (e.g ‘have’ + perf. participle). Adams’ findings confirm earlier research into the development of -mente adverbs in Romance, which originated in Classical texts as well, especially in poetry (Bauer 2010). These findings challenge the conventional wisdom that the Romance languages exclusively trace back to “Vulgar Latin” and they raise questions as to what language variety we find in the documents that we use for our analysis of linguistic phenomena and how this affects our interpretation of the data. Mutatis mutandis these questions are also relevant to the study of Old French grammar, but perhaps in a different way. Where for the diachronic analysis of Latin, the concept of Vulgar Latin has predominated, for Old French emphasis has been on literary texts. The picture of Old French grammar as presented in the handbooks indeed is based to a large extent on texts that may be qualified as “literary” rather than “informal”. The grammatical patterns that emerge from these documents include: pro-drop, brace constructions, and single and partially double negation, to mention just a few (e.g. Foulet [1930], Marchello-Nizia [1979], Picoche & Marchello-Nizia [1998], Buridant [2000]). It is noteworthy that these phenomena indeed are attested in literary sources, but patterns in informal texts or texts authored for ordinary people rather than the elites, seem to show a high incidence of subject pronouns, double negation, and relatively few brace constructions. In this paper I will analyze these phenomena in two types of Old French non-literary texts, the 13th century Old French Gospels and the Lapidaires en prose. I will compare my findings to the patterns as defined in the handbooks and evaluate how they relate to what we know about long-term language change in Latin-Romance.

References Adams, James N. 2013. Social variation and the Latin language. Cambridge: Cambridge University Press. Bauer, Brigitte L.M. 2010. “Forerunners of Romance -mente adverbs in Latin prose and poetry”. In Eleanor Dickey & Anna Chahoud (eds), Colloquial and literary Latin, 339-353. Cambridge: Cambridge University Press. Bauer, Brigitte L.M. In Prep. “Compound verb forms in the Old French Gospel of Marc”. Buridant, Claude. 2000. Grammaire nouvelle de l’ancien français. Paris: Sedes Foulet, Lucien. 1930. Petite syntaxe de l’ancien français. Paris: Champion. Marchello-Nizia, Christiane. 1979. Histoire de la langue française aux XIVe et XVe siècles. Paris: Dunod. Picoche, Jacqueline & Christiane Marchello-Nizia. 1998. Histoire de la langue française. Paris: Nathan.

ABSTRACT. This study explores whether quantifier yixie ‘some’ in Chinese has undergone reinforcement or, assuming a constructional perspective (Traugott and Trousdale 2013, inter alia) it can be interpreted as a case of constructional realignment.

Xie ‘some; a small amount of’ developed from xiezi and its variants such as xier and xiexie. What characterised the variants was diminutive suffixes (-zi and -er, originally ‘son’) and reduplication (xiexie) that intensified xie’s diminutive semantics. Xie later acquired an optional yi ‘one’. There was no obvious semantic distinction between yixie and xie at any stage; both could quantify concrete or abstract nouns (e.g. xie jiu ‘some alcohol’; xie daoli ‘some reason’; yixie zhulu ‘some dew’). Frequencies of xie and its variants quantifying common nouns in Early Modern Mandarin (12^th to 18^th c.) and Modern Mandarin were calculated, based on data extracted from the Academia Sinica Tagged Corpora, which shows that the most frequent diachronic form was xie, but synchronically it is yixie. Yixie is also taken as the ‘basic’ form by grammarians (Chao 1968, Li & Thompson 1981) from which xie is derived. 

This pattern of formal loss (xiezi > xie) followed by formal gain (xie > yixie), without change in meaning, would be interpreted as a case of reinforcement in the grammaticalisation literature (cf. Lehmann 1995: Ch. 2.5), which happens to ‘[compensate] for and [check] the decay’ (ibid: 20) of phonological or semantic substance. This seems especially plausible if yi is considered as reinforcing the diminutive semantics previously encoded by -er, -zi, or reduplication, or when the fact that some Sinitic languages have undergone the ‘diminutive cycle’ (Cao 2006; similar to Jespersen’s cycle) is taken into account. However, while reinforcement has been invoked to explain the accretion of form in negation markers (e.g. Fr. ne V > ne V pas), indefinite pronouns (e.g. Lat. aliquis ‘someone’ + unus ‘one’ > It. alcuno ‘someone’), or prepositions (e.g. Fr. dans ‘in’ > dedans ‘inside’; Lehmann 1995), there seems to be no report that any quantifier has undergone reinforcement. Moreover, even though xiezi > xie certainly experienced phonological loss, it is also not clear if this loss should be ‘compensated’ for, or any semantic ‘decay’ needed ‘checking’, calling into question a reinforcement account.

However, in Construction Grammar (Goldberg 2006), this change can be interpreted as xie’s realignment with the mensural classifier construction schema, which has the form [numeral classifier]. That is, xie became a classifier, made possible by xie’s functional and formal compatibilities with the mensural classifier schema. Functionally, xie and other mensural classifiers all quantify objects, similar to measure words in English. Formally, xie bears more resemblance to classifiers than other non-classifier quantifiers: it is monosyllabic. Typical quantifiers in Chinese are disyllabic, such as quanbu ‘all’, henduo ‘many’, and renhe ‘any’. Based on the analogy between xie and mensural classifiers, users then sanctioned yi + xie based on the template [numeral classifier].

Construction Grammar can also further explain why xie took on only yi, while typically mensural classifiers can allow any numeral; that is, why xie is sanctioned only partially by the classifier schema. For example, yi xie shui ‘some water; lit. one some water’ is acceptable, but *liang xie shui ‘lit. two some water’ is not, while yi/liang ping shui ‘one/ two bottle(s) of water; lit. one/two bottle water’’ is grammatical. Xie’s partial sanction is motivated by the convergence of yi’s role in the schema and xie’s semantics. [yi classifier] can be used to signify a holistic amount of things, reflecting the construal of an object, or objects, as a unified whole (Li & Thompson 1981). This corresponds to xie’s semantics, ‘a small amount of’. Yi is also the most frequent numeral in the schema, and when a numeral in the schema is unexpressed, the default reading is yi ‘one’ (Sun 2006), making yi the most likely numeral with which xie was associated.

A morpheme-based analysis might interpret xie > yixie as reinforcement, but a constructional analysis more clearly accounts for why and how xie acquired yi. This study thus suggests that a diachronic constructional approach (e.g. Traugott & Trousdale 2013) looking at a higher level of generalisations may provide different, insightful explanations.

References: Cao, F. 2006. On Grammaticalization Cycle. Hanyu Xuebao 14(2): 2-15.Chao, Y. R. 1968. A grammar of Spoken Chinese. Berkeley: University of California Press. Goldberg, A. 2006. Constructions at work: the nature of generalization in language. Oxford: OUP. Lehmann, C. 1995. Thoughts on grammaticalization. Munich: Lincom Europa. Li, C., & S. Thompson. 1981. Mandarin Chinese: A functional reference grammar. Berkeley: University of California Press. Traugott, E. C., & G. Trousdale. 2013. Constructionalization and Constructional Changes. Oxford: OUP. Sun, C. F. Chinese: A linguistic introduction. Cambridge: CUP.

ABSTRACT. This paper discusses the development of the indefinite article in English and analyzes the phenomenon of category emergence from a diachronic construction grammar perspective (DCxG). It is argued that the development of the OE numeral 'an' and the OE demonstrative 'se' are linked and should not be told as a story of two individual morphs. The emergence of the article category follows from changes on the abstract, procedural level in the linguistic networks of OE speakers. What has traditionally been interpreted as a case of grammaticalization on the morphosyntactic level is at the same time a case of ‘grammatical constructionalization’ (i.e. the emergence of a new complex schematic form-meaning pairing which previously did not exist in the constructicon) and constructional competition. Articles which only express indefiniteness or non-specificity are not frequent cross-linguistically (cf. Himmelmann 1997; van Gelderen 2007). Thus, what needs to be explained is why and how English changed from an ‘articleless’ language into a language with obligatory overt indefiniteness marking (at least with singular count nouns) which recruits a free-form article in the prehead of the NP to do so. Although many formalist and functional explanations have been put forward (cf. Christophersen 1939; Diessel 1999; Greenberg 1978; Hopper & Martin 1978; Lyons 1999; Osawa 2007; Philippi 1997; Rissanen 1967; Stark, Leiss & Abraham 2007; Traugott 1982; van Gelderen 2007; Yamamoto 1989), a cognitive constructional network approach seems to provide the most promising answers to WHEN, WHY and HOW the indefinite article emerged. Construction grammar is a useful tool for diachronic analysis because its architecture invites us to think “about change in form and meaning equally, as well as the creation of and changes to links between constructions in a network” (Traugott & Trousdale 2013: 231; Barðdal & Gildea 2015). In particular, it will be argued that the development of the articles in English reflects the establishment of two complex schematic lexically underspecified constructions with a fixed determination slot in the prehead of the NP (Traugott & Trousdale 2010: 12; van de Velde 2010: 291). As a first step, the high frequency of some definite OE NP constructions (e.g. [DEM+ CN], [POSS+CN],…) leads to the emergence of a new ‘form-meaning pairing’ on a more abstract level. The speakers interpret definiteness marking to be obligatory and attribute it a locally fixed slot. The emerging construction is formalized as the [[Xdefinite determinative]DETERMINATION + [Zcn]HEAD]NP{def}– construction. This construction is a form-meaning pairing in the following sense: On the construction’s ‘semantic side’ we find ‘the concept of definite reference’ which is formally expressed in the grammar by an NP with a fixed determination slot which has to be filled obligatorily. Such a lexically underspecified slot is an attractor position and has the potential of recruiting fillers. This has triggered the grammaticalization of the OE demonstrative se. The form is co-opted for acting as the default definiteness marker, or slot filler (= definite article). This reanalysis of the linguistic system (from definiteness marking being optional to definiteness marking becoming obligatory) increases processing efficiency (Hawkins 2004) and happens due to complex frequency and analogy effects (Fischer 2007; Sommerer 2012, 2015). The cognitive entrenchment of this construction is sanctioned by the high frequency of specific NP constructs, which are compatible with the schema. Later in time, another abstract construction is added to the family of NP constructions (via analogy). Namely an indefinite NP construction where indefinite reference also has to be marked overtly in the determination slot (at least with singular count nouns): [[Xindefinite determinative]DETERMINATION + [Zcn/sing]HEAD]NP{indef} Semantics/meaning:Indefinite singular reference; Form:Expressed in grammar by an overt indefinite determinative in NP construction with a determination slot. This construction competes with and ousts another indefinite NP construction, in which indefiniteness does not have to be marked by an overt element in grammar and where leaving the slot empty (zero marking) is indicative of indefinite reference (Sommerer, forthc.). As a result the numeral an is recruited to mark indefiniteness for singular count nouns, thereby becoming the indefinite article. So the two elements did not grammaticalize on their own but in the context of emerging schematic constructions with a slot. From that perspective, an abstract construction is a “grammatical primitive” which can be “the source […] of grammaticalization” (Traugott & Trousdale 2010: 13). Both constructions, which are embedded into the constructional network of OE/ME NP constructions, are being linked to and influenced by other existing constructions on more or less abstract levels. For example, it is very likely that both constructions will be linked to an even more abstract ‘referential DP construction’: [[_X_]DETERMINATION + [Zcn]HEAD]NP{ref}– construction. The paper is a first attempt to sketch the network/constructional family of Old English NP constructions. Empirically, the argumentation rests on a large quantitative and qualitative analysis of definite and indefinite NP constructions in 13 Old English prose texts in the York-Toronto-Helsinki Parsed Corpus of Old English Prose (YCOE). Corpus Search, AntConc, and R were used for analysis.

ABSTRACT. In grammaticalization research, it has become a common assumption that the formal and semantic properties of the source construction determine the trajectory of its grammaticalization (see Bybee et al. 1994 on the semantic side of things in particular). Taken literally, this means that a grammaticalizing construction develops in a predetermined way that is uninfluenced by the language system. While a very useful concept, this idea conflicts with evidence from the paradigmatization of classes of function words out of morphologically, syntactically and semantically heterogeneous sources (e.g. van de Velde 2009 on the paradigmatization of determiners in Dutch, or Reinöhl 2016 on the paradigmatization of postpositions in Indo-Aryan). Reinöhl & Himmelmann (to appear) argue therefore that, while functional as well as formal source determination is the main factor driving grammaticalization, paradigmatization is a constraining factor. In this talk, I explore this proposal further, applying it to a test case that has not been studied so far, the paradigmatization of a pair of imperfective and perfective participles in the development from Old Indo-Aryan to modern Hindi. As in the case of the Dutch determiners and the Hindi postpositions, the participles in question, too, start off showing morphological, syntactic, and semantic asymmetries, and only over time converge into the paradigm they form today. Nowadays, one of them is found in almost every Hindi clause as the predicative nucleus. They are largely complementary semantically as imperfective and perfective forms, optionally combining with auxiliaries for tense and modal marking. They show the same inflectional behaviour and are only partially distinct with regard to the suffixal forms, but not otherwise transparently different on a morphological level (with very few exceptions involving suppletive forms). This homogeneity, however, only developed over time. For illustration I show some morphological and syntactic differences that place them apart in Sanskrit. First of all, we are dealing with two unrelated morphological formations. The Hindi imperfective participle derives from a present active participle in Sanskrit, based on the present tense stem (see Lowe 2015). The perfective, by contrast, derives from the so-called ta-form (often referred to as past passive participle). In contrast to the present active participle, the ta-form is not based on a tense stem, but formed directly from the verbal root, which presents the first difference between the two formations. The ta-form has resultative meaning and is frequently found also in “adjectival” and “nominalized” usages. Syntactically, the ta-form appears to be ahead of the present active participle in the historical development of the characteristics they show today. There are very few examples in a corpus of 911 clauses of the Vedic prose text Maitrāyaṇī-Saṁhitā studied for this paper where the present active participle appears to be used in predicative function (one example can be seen in (1)). By contrast, about 25 cases show the ta-form in predicative usage, both in relative and main clauses as illustrated in (2), foreshadowing the default usage as main predicate in modern Hindi. (1) abhikrā́mantī vā́ ékā́hutir abhikrā́mantī vā́ ékā ā́hutir approach.PPA.NOM.SG.F PRT one.F offering.NOM.SG.F ‘One offering is the approaching one’ (MS 1.4.12)

(2) áchinno divyás tántur iti not_cut_off.ta.NOM.SG.M heavenly.NOM.SG.M thread.NOM.SG.M QUOT ‘The heavenly thread is not cut off’ (MS 1.4.7)

The semantic, morphological and syntactic differences between the present active participle and the ta-form in Sanskrit are at odds with the homogeneity of the perfective and imperfective participles in Hindi. The apparent convergence calls for an approach that allows for interactions between grammaticalizing constructions above and beyond the trajectory set out for them by their source forms and meanings. This talk thus aims at testing the claim put forward in Reinöhl & Himmelmann (to appear) that grammaticalization is a process that is not always exhaustively accounted for by source determination. Instead, paradigmatization has to be taken into account. At the same time, paradigmatization seems unlike typical cases of analogical change (see e.g. Harris & Campbell 1995 for case studies of analogical change in historical syntax under the heading of “extension”), as we are not only dealing with the transferral of one formal pattern from one element to the next, but with convergence both formal and functional. The study of paradigmatization is thus complementary to grammaticalization in such instances where the development of a construction seems to be not exhaustively explained by its source meaning, but influenced by the development of other grammatical constructions.

References Harris, Alice C. & Lyle Campbell. 1995. Historical syntax in cross-linguistic perspective. Cambridge: Cambridge UP. Lowe, John J. 2015. Participles in Rigvedic Sanskrit. The Syntax and Semantics of Adjectival Verb Forms. Oxford: Oxford University Press. Reinöhl, Uta. 2016. Grammaticalization and the rise of configurationality in Indo-Aryan. Oxford: Oxford University Press. Reinöhl, Uta & Nikolaus Himmelmann. “Renewal: A figure of speech or a process sui generis?” To appear in Language. van de Velde, Freek. 2009. “The emergence of modification patterns in the Dutch noun phrase”, Linguistics 47 (4), 1021-1049.

ABSTRACT. In this paper, I argue that word order changes can be cases of grammaticalisation, involving gram-mation, regrammation and degrammation in the sense of Andersen (2006). As a consequence of the prevailing perspective on grammaticalisation, this is not the general position. In the Oxford Handbook of Grammaticalization, Sun & Traugott presuppose general agreement that “word order change per se has not been regarded as a case of grammaticalization” (2011: 378). By contrast, they take syntagmatic dependency to be a generally accepted parameter, and “the development of obligatory subject in Romance and Germanic languages is regarded as a hallmark of grammaticalization construed as increased dependency”, referring to Lehmann 1995a [1982]. The fixation of the subject in French and English is part of grammaticalisation, but the fixa-tion of a word order alternation for subjects in the Mainland Scandinavian V2 languages is not, in spite of its central role in distinguishing interrogative and declarative clauses. Sun and Traugott quote Lehmann for the syntagmatic dependency part of his view, but not for his emphasis on the paradigmatic aspect of grammaticalisation; cf. also Diewald (2010). To include word order in grammaticalisation, we need a theory with a focus on the paradigmatic dimension, as emphasized by Nørgård-Sørensen et al. (2011). Word order alternations form paradigms in a gener-alised sense, and to identify these, we need to describe both the linear expression of the word order systems and their content. To cope with linearity, we need a methodology that does not axiomatical-ly intertwine linearity and syntactic constituency (Danish topological tradition offers this, cf. Faar-lund 1989, 1990); to cope with the description of the content, we need a Peircean distinction be-tween symbolic and indexical meaning (Andersen 1980; Nöth 2008). In Old Scandinavian constituent shift (often erroneously called ‘object shift’), we find a word order contrast between background position (preceding negation) and focus position (immediately following negation, but preceding the non-finite verb). This symbolic contrast was retained in 16-17th century Danish, restricted to pronouns. At this stage, an indexical function had been added to the positions for background and focus, that of pointing to the pronominal material that may fill them. Example (1) shows a pronominal object in focus position. (1) du skalt icke saadant tencke Thou shalt not such think you must not think like that A major challenge is the word order of Old English where OV (finite) and V2 are both found: (2) Þær wearð ofslægen Lucumon cynges gerefa (V2) there was slain Lucumon king’s sheriff there Lucumon, the king’s sheriff, was slain (3) hira þær tu sæ on lond wearp (OV) of those there two sea ashore threw ’of those the sea threw two ashore there’ (4) he hi him eft ageaf (OV) he them him back gave ’and he gave them back to him’ These two word order types are hard to relate in models that axiomatically combine linearity and hierarchy. Ans van Kemenade (2009) suggests a new functional projection to cope with them: [CP [XP] C [AgrP PronounSubj/Obj Agr [NegP Neg Adv Neg [TP DPSubj T ... ]]]. Instead, I shall argue that given a sharp distinction between word order and hierarchical syntax, the bold face constituents can all be claimed to occupy the second position (as a late reflection of Wackernagel’s law in OE). The two word order types form sets of paradigmatic contrasts that can be described semantically in terms of symbolic contrasts and indexical relations, and as such, they form a grammatical system. Such synchronic exercises must be tested through their diachronic returns. The analysis proposed will suggest a scenario for the formation of Scandinavian V2; and attested sentence types of Early Runic are compatible.

References and sources: Andersen, H. (1980). Morphological change: towards a typology. Historical Morphology, ed. by J. Fisiak, 1-50. The Hague, Paris, New York: Mouton Publishers. Andersen, H. (2006). Grammation, Regrammation, and Degrammation. Tense Loss in Russian. Diachronica XXIII, 2, 231-258. Campbell, A. (1959). Old English Grammar. Oxford University Press. Diewald, G. (2010). On some problem areas in grammaticalization studies. Grammaticalization: Current views and issues, ed. by K. Stathi, E. Gehweiler & E. König, 17-50. John Benjamins. Fourquet. J. (1938). L’ordre des éléments de la phrase en germanique ancien. Paris: Les Belles Lettres. Faarlund, Jan Terje (1989). Discourse Functions and Syntax. CLS 25 (= Papers from the 25th Annual Regional Meeting of the Chicago Linguistic Society). Part Two: Language in Context, 30-40. Chicago Linguistic Society. Faarlund, Jan Terje (1990). Syntactic change. Toward a theory of historical syntax. Trends in linguistics. Studies and monographs 50. Berlin and New York: Mouton de Gruyter. Kemenade, A. van (2009). Discourse relations and word order change. Information Structure and Language Change. New Approaches to Word Order Variation in Germanic, ed. by R. Hinterhötzl & S. Petrova, 91-118. Mitchell, B. (1985). Old English Syntax. Oxford: Oxford University Press. Nørgård-Sørensen, J., L. Heltoft & L. Schøsler (2011). Connecting Grammaticalisation. John Benjamins. Nöth, W. (2008). Semiotic foundations of natural linguistics and diagrammatic iconicity. Natural-ness and Iconicity in Language, ed. by K. Willems & L. De Cuypere, 73-100. John Benjamins. Sun, C. & E. Traugott (2011). Grammaticalization and Word Order Change. The Oxford Handbook of Grammaticalization, ed. by H. Narrog & B. Heine, 378-88. Oxford: Oxford University Press.

ABSTRACT. The island of New Guinea is one of the most linguistically diverse regions on earth, with over 900 languages representing over 20 language families (Foley 2000; Ross 2005). Yet it is also one of the most poorly documented in terms of both the languages and their historical relationships (Hammarstöm & Nordhoff 2012). At the heart of these historical-comparative challenges is the third-largest language family in the world: Trans-New Guinea (TNG). Generous early proposals regarding this family were heavily revised and qualified in the following decades, and the most well-known account of the family to date is based largely on the reconstruction of pronoun paradigms (Ross 2005), although there also exist reconstructions of some basic vocabulary (e.g., Pawley 2005). Significant progress in these endeavors largely awaits advances in two areas: the documentation and description of individual languages and the collection of language data in comprehensive and accessible databases for comparative purposes. Progress in large-scale database creation seems to be underway. Among other projects, an online database of Papuan languages was recently established by Simon Greenhill for use in both traditional comparative and advanced computational research (Greenhill 2015). Additionally, while hardly new, Timothy Usher’s “Newguineaworld” has remained somewhat under the radar despite featuring an ever-growing array of bottom-up reconstructions of language families across New Guinea (Usher 2016). The creation of these large databases for comparative and computational purposes is, in turn, dependent on data made available by language documentation and description. My own research is focused on the documentation and description of two TNG languages of Southern New Guinea: Bitur and Abom. This work is part of a renewed attention in recent years on Southern New Guinea, a region whose linguistic diversity has often been overlooked (Evans 2012; Evans et al., to appear). Bitur [ISO 639-3: mcc] is one of four members of the Tirio language family, which has only very recently been shown to belong to the Anim stock (Usher & Suter 2015). As the “least-known of all Southern New Guinea groups” (Evans et al. to appear), descriptive work on the Tirio languages is crucial for our understanding of how TNG languages spread southward into the lowlands and how these languages can be affected by Southern New Guinea areal influences and non-TNG languages. I began a long-term project with the speakers of Bitur in 2015 which includes the Boasian trilogy of lexical documentation, grammatical description, and the collection, transcription, and translation of connected speech in various genres. Abom [aob] was only discovered by a survey team in 2002 (Jore & Alemán 2002), and it was assumed to be a member of the Tirio language family until very recently (Usher & Suter 2015). With only three elderly speakers, documentation of Abom is time-sensitive and of the highest priority; I was able to do some preliminary documentation—lexical, grammatical, and based on connected speech—in the Summer of 2016. The pronouns of Abom suggest TNG affiliation, and some apparent cognates in basic vocabulary reinforce this hypothesis. However, it is yet to be determined whether a direct connection to another TNG subgroup can be made or if Abom represents a primary branch of this very large family. The answer will be borne out through continued collaboration with Timothy Usher and his comparative database. The efforts of various individuals in documentation and description, in concert with the growing access to comparative databases, have the potential to yield muchneeded advances in our understanding of the relationships among the languages of New Guinea and the histories of its peoples. References Evans, Nicholas. 2012. Even more diverse than we thought: the multiplicity of Trans-Fly languages. In Nicholas Evans & Marian Klamer (eds.) Melanesian Languages on the Edge of Asia: Challenges for the 21st Century. Language Documentation and Conservation Special Publication No. 5: 109-149. Evans, Nicholas, Wayan Arka, Matthew Carroll, Christian Döhler, Eri Kashima, Emil Mittag, Kyla Quinn, Jeff Siegel, Philip Tama and Charlotte Van Tongeren. The languages of Southern New Guinea. To appear in Bill Palmer (ed.), The Languages and Linguistics of New Guinea: A Comprehensive Guide. Berlin: Walter de Gruyter. Foley, William A. 2000. The Languages of New Guinea. Annual Review of Anthropology, 29: 357–404. Greenhill Simon J. 2015. TransNewGuinea.org: An Online Database of New Guinea Languages. PLoS ONE 10(10): e0141563. doi:10.1371/journal.pone.0141563 Hammarström, Harald and Sebastian Nordhoff. 2012. The languages of Melanesia: Quantifying the level of coverage. In Evans, Nicholas and Marian Klamer (eds.), Melanesian Languages on the Edge of Asia: Challenges for the 21st Century (Language Documentation & Conservation Special Publication 5), 13-34. Honolulu: University of Hawaii Press. Jore, Tim and Alemán, Laura. 2002. Sociolinguistic survey of the Tirio language family. SIL, Ukarumpa: Unpublished MS. Pawley Andrew. 2005. Trans New Guinea hypothesis: Recent research and its implications. In Andrew Pawley, Robert Attenborough, Robin Hide, Jack Golson (eds.), Papuan Pasts: Cultural, linguistic and biological histories of Papuanspeaking peoples, 67-107. Canberra: Pacific Linguistics. Ross, Malcolm. 2005. Pronouns as a preliminary diagnostic for grouping Papuan languages. In Andrew Pawley, Robert Attenborough, Robin Hide, Jack Golson (eds.), Papuan pasts: cultural, linguistic and biological histories of Papuanspeaking peoples, 15-66. Canberra: Pacific Linguistics. Usher, Timothy. 2016. “Newguineaworld” Classification and reconstruction of Papuan language families. Online manuscript. < https://sites.google.com/site/newguineaworld/> Usher, Timothy, and Edgar Suter. 2015. The Anim Languages of Southern New Guinea. Oceanic Linguistics, Vol. 54, No. 1, 110-142.

ABSTRACT. Historical-comparative studies of South American languages and in particular of Amazonian languages is an area of investigation still in its infancy and those studies that have been undertaken are largely focused on the identification of larger groupings (e.g., Rodrigues’s (1985) TuCaJê) or on the reconstruction of large language families (e.g., Gildea’s (1998) reconstruction of Cariban morphosyntax), with comparative work in smaller families lagging behind (Campbell 1997; Epps 2009). This situation is largely due to a lack of linguistic documentation and descriptive studies but with the increase in awareness about language endangerment and the rise of documentary linguistics as a new subfield in linguistics, there has been an exponential increase in the number of documentation projects and descriptions of Amazonian languages in the past three decades (Epps 2009; Epps and Salanova 2013; Everett 2010) and this has allowed researchers to undertake historical-comparative studies of small language families and isolates (e.g., see Adelaar 2000; Ribeiro and van der Voort 2010). However, the investigation of the prehistory of small language families and isolates in Southern Venezuela and Eastern Colombia—a region with multiple small genetic groupings interspersed throughout alongside languages from large families—has received little attention. Jodï [ISO: yau] is a small language spoken in the Venezuelan Amazonas and Bolivar states by approximately 1,000 people (INE 2013). The language has generally been treated as an isolate or left unclassified in the language classification literature (e.g., Campbell 2012, Kaufman 2007, Loukotka 1968 inter alia). However, genetic relationships between Jodï and four different families of the region, namely Cariban, Yanomaman, “Makúan”, and Sáliban, have been postulated in the literature (see Wilbert 1963:125-126; Migliazza 1975, 1985; Henley et al. 1994- 1996; and Coppens 1983:253, Mosonyi 2000:660, Zent & Zent 2008:503 respectively). In this paper, I investigate in depth the proposed Jodï-Sáliban relationship by means of a comparison of lexical and grammatical material. Based on numerous regular sound correspondences as well as grammatical correspondences—some of which are too idiosyncratic to be anything but the product of inheritance—I conclude that Jodï is related to the Sáliban languages Sáliba, Mako and Piaroa. The lexical comparison of 189 etyma in a modified-Swadesh list shows that, out of 34 lexical items reconstructable to Proto-Sáliban (i.e. present in all daughter languages), 21 have a Jodï cognate. Additionally, 24 other etyma show some degree of cognacy between Jodï and one or two of the three Sáliban languages. This represents about 24.3% of cognate lexical material. Further, the cognate sets identified show numerous regular sound correspondences; the two sets in (1) for instance show the regular correspondence of Sáliba and Piaroa d with Mako t and Jodï jt. (1) SÁLIBA PIAROA MAKO JODÏ õdete ɑ̃dẽ õte ojte ‘fat/oil’ duda duɑʔɑ tuba jtuwëdo ‘hot’ In addition to these regular sound correspondences, there is at least one regular process of sound change attested, namely the deletion in Jodï of a hV syllable where the V is identical to the vowel in the adjacent syllable as shown by the sets in Table 1. TABLE 1 Cognate sets supporting -hV deletion process in Jodï SÁLIBA PIAROA MAKO JODÏ 13 soil sẽxẽ ɾẽhẽ nihi ne 62 grass oda mæ̃hĩjæ̃-tsˀe mehe-ˀʤe me-jtej̱ te ̱ 69 hear (v) ĩse æ̃ hũkũ ãhãkʷɨ aḵ u 111 person hoho tʰɨ̃hɑ̃ ĩsɑ̃ hoho jo Comparison of morphological material also yields numerous sets of cognate morphemes; these include a proximate deictic root; an adverbial formative; an object, a sociative and an instrumental case markers; three plural markers; a topic marker; an imperative suffix; and a past and future tense markers. Some of these are “idiosyncratic, peculiar, arbitrary morphological correspondences […], instances so distinctive that they could not be easily explained by borrowing or accident” (Campbell 2008:177). For instance, while most animate nouns form their plural with the PL.ANIM1 marker, a handful of nouns use instead the PL.ANIM2 suffix with the form -mu in Sáliba, Piaroa and Mako and -mo in Jodï. One such noun is ‘child’, which takes this less frequent plural marker in all four languages: SINGULAR PLURAL (2) ĩtʰĩ ‘child, son’ vs. ĩtʰĩ-mu ‘children, sons’ MAKO (3) chitti ‘my son’ vs. chittimu ‘my children’ PIAROA (Overing 1974:361-362) (4) jui ‘son’ vs. juimu ‘sons’ SÁLIBA (18TH CENTURY) (No author 1790:156) (5) iṉ i ‘child’ vs. iṉ imo ‘children’ JODÏ (Quatra 2008a:318) Summing up, the investigation of a possible genetic relationship between Jodï and Sáliban shows that there are numerous lexical cognate sets which display regular sound correspondences as well as regular process(es) of change, and abundant cognate morphology. This paper thus demonstrates the existence of a Jodï-Sáliban language family and contributes to our understanding of the prehistory of the languages of Northwestern Amazonia. Methodologically, I emphasize the link between language documentation and language classification and how the documentation of endangered languages can lead to new insights in historical linguistics. Selected references: CAMPBELL, LYLE. 2008. How to show languages are related: the methods. In Language classification: history and method, eds. Lyle Campbell and William J. Poser, 162-223. Cambridge: CUP. CAMPBELL, LYLE. 2012. Classification of the indigenous languages of South America. In The indigenous languages of South America: A comprehensive guide, eds. Lyle Campbell and Verónica M. Grondona, 55-166. Berlin; Boston, MA: De Gruyter Mouton. ZENT, STANFORD AND EGLEÉ MARIANA LOPEZ ZENT. 2008. Los jodï: Notas sobre su situación presente y actualización bibliográfica. In Los aborígenes de Venezuela., ed. Miguel Ángel Perera, 499-570. Caracas: Fundación La Salle/Monte Avila Editores/Ediciones IVIC/Instituto Caribe de Antropología y Sociología.

ABSTRACT. In his article “Abductive and Deductive Change” (1973), Henning Andersen stated that two types of inference, “abduction” and “deduction,” are operative in the phonological change of language, and both “evolutive” and “adaptive” change, which are mostly equivalent to what have been traditionally referred to as “internal” and “external” change, respectively, involve these two types of inference, but in a different manner. The evolutive change in phonology occurs as a result of abductive innovation followed by deductive process, whereas the adaptive change requires two-fold abduction before an innovation reaches the deductive level of inference. Andersen’s model was succinct and broadly recognized in the field of linguistics for its explanatory power. With respect to the study of historical linguistics, however, his proposal, at least its adequacy for explaining various aspects of language change, is open to reevaluation. Just because Andersen’s model may adequately account for the structural process in phonological change, it does not necessarily mean that other phenomena of language change, such as the emergence and decline of particular grammatical categories, word-order change, and semantic shift of lexical items, can be explained by his model in the same manner. In particular, the availability of the two-fold abduction model, proposed for the adaptive change of phonology, should be examined in relation to different manifestations of the effect of language contact. Regarding the long dispute in contact linguistics on the borrowability of structure and its sociolinguistic relevancy, a question arises whether or not this two-fold abductive model could be integrated into the development of an explanatory model for structural borrowing. With these thoughts in mind, this presentation will deal with diachronic change in Croatian idioms, especially in such idioms formed and being formed in language contact. Objectives of examination are Croatian dialects spoken in present-day Croatia and those spoken outside Croatia, namely, idioms currently used in Austria's Burgenland and an idiom once spoken in southernmost Moravia. Language phenomena taken as illustrations will be morphosyntactic as well as semantic, such as nominal case markers, clitic alignment, expression of the future perfect, and the functional shift of prepositions. Our examination will inevitably include also an evaluation of the validity of Andersen’s model itself. By introducing two sets of terms and notions, namely, abductive and deductive innovation, alongside evolutive and adaptive change, Andersen succeeded in proposing a well-formed explanation for phonological change. However, a question remains concerning the essential difference between Andersen’s model and the traditional way of understanding historical change, articulated by such terms as “reanalysis,” “analogy,” and “generalization.” What is the benefit of using the terms of logical operation to explain diachronic change in a language system which always includes in itself variability and probability? This question will be addressed against the background of a peculiar phenomenon, the use of the third person plural form to denote the third person singular human object found in the Čakavian as well as Kajkavian dialects.

References Andersen, Henning. 1973. Abductive and Deductive Change. Language, Vol49, No.4.　　 765-793. Campbell, Lyle. 1993. On Proposed Universals of Grammatical Borrowing Sources. in Aertsen, H. and Robert J. Jeffers (eds.) Historical Linguistics 1989: Papers from the 9th International Conference on Historical Linguistics, Rutgers University, 14–18 　　August 1989. Amsterdam: J. Benjamins, 91–110. Hickey, Raymond. 2010. The Handbook of Language Contact. Chichester, West Sussex: Wiley-Blackwell. Kučerová, Kvetoslava. 1998. Hrvati u Srednjoj Europi. Zagreb–Bratislava: Matica hrvatska, Matica slovačka. Posebna izdanja. Neweklowsky, Gerhard. 1978. Die kroatischen Dialekte des Burgenlandes und der angrenz- enden Gebiete. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Popovic, Ivan. 1960. Geschichte der serbokroatischen Sprache. Wiesbaden: Harrassowitz. Sankoff, Gillian. 2001. Linguistic Outcomes of Language Contact. in Chambers, J. K., Trudgill, Peter and Natalie Schilling-Estes (eds.) The Handbook of Language Variation and Change. Malden, MA: Blackwell. 638–668. Thomason, Sarah and Terrence Kaufman, 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press. Vážný, Václav. 1934. Mluva charvátských osad v republice Československé. Československá vlativěda. Díl III, Jazyk. Praha : Sfinx, 518–523. Weinreich, Uriel. 1953. Languages in Contact. Findings and Problems. New York: Linguistic Circle of New York.

ABSTRACT. Henning Andersen’s article entitled “Abductive and Deductive Change” (1973) has multiple notable contributions at various levels. On the level of general historical linguistics, he first introduced into linguistic analysis the notion of abduction advanced by a famous philosopher Ch. Pierce and by applying this notion, Andersen has convincingly shown its validity in his analysis of the enigmatic phonological changes in the Litomyšl’s “teták“ dialect of Czech. This alone was a unique contribution to Slavic diachronic dialectology. Further, though, in spite of the fact that the article has dealt with a phonological change only, Andersen’s approach has drawn significant attention as one of the possible accounts for a linguistic change in general, including (morpho-)syntactic change (cf. Harris and Campbell 1995: 384-385) and it has been indeed successfully applied in various contexts, for instance, by Timberlake (1977) but many others as well. Another important aspect of Andersen’s article is the theoretical claim of the distinction between evolutive and adaptive changes. With all this in mind, Andersen pointed out the importance of recognizing the sociolinguistic dynamics that a group of speakers of a given language was facing. Although there have been discussions about pros and cons of Andersen’s approach (cf. Deutscher 2002), it certainly helps us understand the nature of various linguistic changes particularly in small languages that have been in close contact with other languages and whose sociolinguistic dynamics may have significant impact on grammatical change. As one such case, I show that Andersen’s framework allows important light to be shed on morpho-syntactic changes in Kashubian. This language has experienced two different language contact contexts, namely, with German, until the end of World War II, and then with Polish, and these situations have resulted in the contact-induced changes of various grammatical structures. The degree of grammaticalization in them naturally varies, but it is important to note that there is often a significant discrepancy between highly grammaticalized forms and low frequency of their usage on the one hand, for instance, as with the case of the recipient passive with the auxiliary dostac ‘to get’ with PPP (Nomachi 2012). On the other hand, there are cases in which the degree of grammaticalization in forms does not correspond to its semantics as in periphrastic constructions with the auxiliaries miec ‘to have’ verb to be and bëc ‘to be’ with PPP (Nomachi 2016). In both cases, one has to pay attention to the fact that there is a wide range of acceptability of given forms and their sphere of usage, depending on the sensibilities of individual speakers of Kashubian, which have not been studied yet. In this presentation, I aim to show that these discrepancies can be most probably explained by Andersen’s abductive mode of innovation.

References: Andersen, Henning. 1973. Abductive and deductive change. Language 49. 765–93. Deutscher, Guy. 2002. On the misuse of the notion of ―abduction‖ in linguistics. Journal of Linguistics 38. 469–85.Harris, Alice C, Lyle Campbell. 1995. Historical syntax in cross-linguistic perspective. Cambridge: Cambridge University Press. Nomachi, Motoki. 2012. The Kashubian recipient passive and its grammaticalization. In Andrii Danylenko, Serhii Vakulenko (eds.), Gedenkschrift für George Y. Shevelov zum 100. Geburtstag, 109–135. München: Verlag Otto Sagner. Nomachi, Motoki. 2016. On the second Be periphrasis (BE-2) in Kashubian: its grammatical status and historical development, Slavia 84, No. 3. 268–83. Timberlake, Alan. 1977. Reanalysis and actualization in syntactic change. In Charles N. Li (ed.), Mechanism of syntactic change, 141–77. Austin: University of Texas Press.

ABSTRACT. The aims of this paper are twofold: First, we discuss why the borrowing of verbs, i.e. its argument structure, has never been truly dealt with in models of language contact including borrowing scales and hierarchies. Second, we propose a way to deal with the borrowing of argument structure by using Johanson’s (2002) terminology and Eisenberg’s model of gradual integration, and his notion of integration conflicts (2001; 2011). Since verbs can be seen as the core of sentences as they determine the number and shape of arguments, it is likely that the borrowing of verbs has a considerable impact on the grammar of the recipient language. Thus, if we want to come closer to an understanding of contact-induced change we must take into account this type of borrowing.

Concerning the first point, it can be assumed that the reason for the neglection of verbs in hierarchies and borrowing scales is the use of the long-standing traditional dichotomy between lexical and structural borrowing. Since verbs have both lexical and structural properties there is no question of either or, and this is why we should rethink our terminology when we define borrowing.

One of the most frequently cited works in the field of language contact clearly is Thomason & Kaufman’s 1988 monograph, the reason being that the authors for the first time dealt with language contact in the historical dimension in a systematic fashion. Scrutinising their borrowing scale which aims at measuring the intensity of contact reveals that although the authors thoroughly defined five stages going from casual contact to heavy structural borrowing by giving very detailed information as to what is likely to be borrowed, the authors do not mention the borrowing of the argument structure of verbs at all. If we extend the survey to other borrowing scales or hierarchies that have been proposed in the literature we find a slightly different picture. The hierarchies suggested for example by Haugen (1953), Muysken (1981) and Matras (2007) predicting the borrowability of word-classes from a quantative perspective, include verbs and state that they are less likely to be borrowed than nouns but more likely than for example subordinating conjunctions or inflectional affixes (Muysken and Matras). Nevertheless, again the distinction between lexical and structural borrowing is implicit, and crucially, it is not explicitly defined what verb borrowing actually means.

Interestingly, even if we consult works that focus on the borrowing of verbs like Wohlgemuth (2009) we do not find any information as to how argument structure might be borrowed. The author dedicates his work solely to the accommodation or formal adaptation of verbs in the recipient language (cf. Heine and Kuteva 2005). This seems to be surprising if not paradoxical since we know from contemporary studies of contact and acquisition (e.g. Schmitt, 2000, Myers-Scotton, 2002) that argument structure can be and is borrowed. So the question arises why this phenomenon has never been included in borrowing scales and hierarchies, and in more general terms, in the study of borrowing.

The second point addressed in the talk deals with an approach to classify and analyse the borrowing of argument structure. Our assumptions are based on studies of the contact situation between Old French and Middle English in medieval times (ca. 1150-1500). Using Johanson’s (2002) model of language contact and code copying, we interpret this situation as a case where we had an asymmetrical dominance relation between the foreign model code Old French and the base code Middle English. Old French was the sociolinguistically dominant and prestigious code influencing Middle English, the less prestigious basic code. Approximately 2000 verbs came into ME in this contact situation as global copies. Results from our studies so far suggest that (i) the copying of verbs is semantically driven, (ii) matching/mismatching of the syntactic realisation of arguments in the foreign model code and basic code is constrained. Concerning the latter point, three different cases can be distinguished:

1. The structures in the foreign model code and basic code match, and thus no change on the syntactic level will occur (see example (1)).

(1) a. Il donne [NP l’épée] [PP au roi] b. He donates [NP the sword] [PP to the king] (Basic code: He gives the sword to the king.)

2. There is a mismatch between the foreign model code and the basic code, and the structure is adapted to an existing structure of the basic code (see example (2)). This case can be evidenced by higher frequencies of this particular structure.

(2) For God wasted þe bones of hem þat plesen to men; for God wasted the petitions of them that please to men. ‘Because God destroyed the petitions of those who please men.’ EARLPS,63.2771,M2 (Basic code: He quemes men.)

3. There’s a mismatch which leads to the adoption of a new structure in the basic code. This case implies a change of the grammatical system and is supposed to require a strong motivation, either due to social dominance and/or to the attractiveness of linguistic structure (see example (3)).

(3) a. Middle English Men feeren hem in al the toun (reflexive) ‘men were afraid in the whole town’ b. Anglo-Norman Meis mult se dute de la gueite. (reflexive) ‘But (he) was very afraid of the lookout’ (ex. from Richard Ingham)

The most interesting cases are mismatches which can be interpreted as integration conflicts in Eisenberg’s terms because they show that the copied verbs build a subsystem in the grammar of the recipient language subject to their own rules on both the level of semantics and syntax. They therefore address the question of whether subsystems of this type reflect the L2 status of ME speakers/writers learning OF.

Eisenberg, P. 2001. “Die grammatische Integration von Fremdwörtern. Was fängt das Deutsche mit seinen Latinismen und Anglizismen an?” In Neues und Fremdes im deutschen Wortschatz. Aktueller lexikalischer Wandel, G. Stickel (ed), 183–209. Berlin: de Gruyter. Eisenberg, P. 2011. Das Fremdwort im Deutschen. de Gruyter Studium. Berlin: de Gruyter. Haugen, E. 1953. The Norwegian language in America: A study in bilingual behavior. Indiana. Heine, B. and Kuteva, T., (eds). 2005. Language contact and grammatical change. Cambridge approaches to language contact. Cambridge: Cambridge University Press. Johanson, L. 2002. “Contact-induced change in a code-copying framework”. In Language change: the interplay of internal, external and extra-linguistic factors, M. C. Jones and E. Esch (eds), 285–313. Berlin: de Gruyter. Matras, Y. 2007. “The borrowability of grammatical categories”. In Grammatical borrowing in cross-linguistic perspective, 31–74. Berlin: de Gruyter. Muysken, P. 1981. Generative studies on Creole languages. Dordrecht: Foris. Myers-Scotton, C. 2002. Contact linguistics: bilingual encounters and grammatical outcomes. Oxford linguistics. Oxford: Oxford University Press. Schmitt, E. 2000. “Overt and Covert Codeswitching in Immigrant Children from Russia”. International Journal of Bilingualism 4: 9–28. Thomason, S. G. and Kaufman, T. 1988. Language contact, creolization and genetic linguistics. Berkeley: University of California Press. Wohlgemuth, J. 2009. A Typology of Verbal BorrowingBorrowings. Trends in Linguistics. Studies and Monographs [TiLSM] 211. Berlin, New York: de Gruyter.

ABSTRACT. In this study, I investigate diachronic changes in the assignment of grammatical gender to nouns borrowed into Bosnian/Croatian/Serbian (BCS) from different languages at different time periods.

Previous research on contemporary BCS, a language with three grammatical genders, has shown that the neuter class is closed to new lexical items. Recent loans from western European language which fit the phonological profile of neuter stems (those with final /e/ or /o/ vowels) are instead borrowed as masculine (Simonović 2009).

The analysis of Turkish loanwords introduces a diachronic dimension to this issue, as the extensive Turkish lexical contribution to BCS dates from an earlier period (Mønnesland 1973). In a quantitative study of Turkish loanwords in BCS, I find that assignment of neuter gender is avoided for Turkish /e/-final forms as well. However,

1) Turkish /e/-final loans are typically borrowed as feminine, not masculine, 2) In these cases there is typically a change in final vowel quality (from /e/ to /a/, in keeping with the phonological profile for feminine nouns), and 3) Avoidance of neuter is not categorical, but mirrors the lexical distribution

I conclude that this behavior resembles a pattern previously documented for other South Slavic languages (Walter 2015, in press) and give an account based on gradient ranking of gender constraints (Rice 1996, Boersma & Hayes 2001). Moreover, the data demonstrate a qualitative change in the status of the grammatical category NEUTER in BCS, from open (though a minority) to closed – a change which is complete in Croatian before the other two varieties. The question of what triggered this change must be left unresolved, though it is possibly related to a perceived association of the neuter class with words from the language of the imperial oppressor.

Methods and data After establishing the baseline gender percentages in the BCS lexicon via dictionary counts (Cahen 1920, Škaljić 1966; see Table 1), mini-corpora of vowel-final non-human noun loans from Turkish to BCS were compiled from three different sources (see Table 2).

1) Source 1: For BCS forms compared to Turkish source forms, the number of /e/-final (~neuter) nouns halves, and the number of /a/-final (~feminine) nouns doubles. These changes are primarily due to final /e/-to-/a/ vowel shifts. The resulting ratio of neuter-to-feminine nouns in the lexicon is close to but less than the ratio in the lexicon as a whole (.22 vs .28)

2) Source 2: The number of /e/-final nouns almost halves again, and the number of /a/-final nouns almost triples. As before, the resulting neuter-to-feminine ratio is a close approximation of the overall lexicon’s (.26 vs .28)

3) Source 3: /e/-final nouns are categorically converted to /a/-final feminine BCS forms, with one exception.

masc fem neuter % 45 43 12 N 2016 1898 513 Table 1. Grammatical gender in the Serbo-Croatian lexicon (Cahen 1920)

Turkish -/a/ BCS -/a/ Turkish -/e/ BCS /e/-final N (Friedman) 17 37 17 8 N (Sjoberg) 34 96 42 25 N (B&M) 13 31 12 1 Table 2. Vowel and gender changes in BCS Turkish loans (Friedman 2003, Sjoberg 1972, Ban & Matovac 2012)

References Ban, Lidija and Darko Matovac. 2012. On Turkish loanwords in Croatian language. Trakya Üniversitesi Edebiyat Fakültesi Dergisi, 2:4, 155-166. Boersma, Paul and Bruce Hayes. 2001. Empirical tests of the Gradual Learning Algorithm. Linguistic Inquiry 32.45-86. Cahen, Louis. 1920. Serbian-English and English-Serbian pocket dictionary. Kegan Paul, London. Friedman, Victor. 2003. Turkish in Macedonia and beyond: Studies in contact, typology and other phenomena in the Balkans and the Caucasus. Turcologica 52. Otto Harrassowitz Verlag, Wiesbaden, Germany. Mønnesland, Svein. 1973. Loanwords in the two variants of Serbo-Croatian. Scando-Slavica, 19:1, 197-205. Rice, Curt. 2006. Optimizing gender. Lingua 116.1394-1417. Simonović, Marko. 2009. Immigrants start on the periphery: A unified approach to loanword phonology. MA thesis, Utrecht University. Sjoberg, Marilyn Joel. 1972. Turkish loanwords in the language of Aleksa Šantić. PhD dissertation, The Ohio State University. Škaljić, Abdulah. 1966. Turcizmi u srpskohrvatskom jeziki. Svjetlost Izdavačko Preduzeće, Sarajevo. Walter, Mary Ann. In press. Frequency Distributions as Faithfulness Targets: Or, Why Bulgarians Feminized Turkish Nouns. In Formal Approaches to Slavic Linguistics 24, The New York University Meeting, Yohei Oseki, Masha Esipova, and Stephanie Harves (eds.). Michigan Slavic Publications: Ann Arbor, MI. Walter, Mary Ann. 2015. Probability matching and Macedonian gender morphology: The case of Turkish loanword adaptation. Presentation at the 11th Conference on Formal Description of Slavic Languages (FDSL). Potsdam, Germany.

ABSTRACT. Presentatives can be defined as “constructions which serve to introduce a new element into a discourse” (Trask 1993: 216). Their functions can be summarized as follows: 1) they introduce an element into the reality (here I am), or 2) into the discourse (there came a man). Presentatives are found in numerous languages. Some well-known examples are Latin ecce, French voici/ voilà, Spanish he aquí, English lo and behold, Russian вот/ вон or a number of constructions in Semitic languages (Follingstad 1995 & 2001; Cohen 2014).

Common Basque has two main presentatives based on the allative forms of locative adverbs: hona ‘here.ALL’ & horra ‘there.ALL’. In addition, Souletin (Eastern dialect) has developed a specific presentative construction. Thus, for the Latin presentative clause in example (1a), common Basque employs the presentative hona (1b) and Souletin its own construction (1c), in which haur is thought to be the first degree demonstrative:

(1a) Ecce ancilla Domin-i. (Luc 1, 38) Behold handmaiden.NOM Lord-GEN

(1b) Huna Jaun-aren neskato-a. (Leiçarrague, 1571) here.ALL Lord-GEN handmaiden-DET

(1c) Haur naizü-la Jinko-aren neskato-a. (Belapeire, 1696) this († here) be.1SG.ALOC-COMP God-GEN handmaiden-DET ‘Behold the handmaiden of the Lord’

In this paper I will study Souletin presentatives in the light of a corpus of old texts (16th–19th centuries) and show how the change in the deixis system and intensive contact with Gascon Occitan contributed to the form of presentative constructions.

From a diachronic point of view, the study of the Souletin presentative formula reveals that haur should not be analyzed as the first degree demonstrative, but the homophone ancient proximate locative adverb haur ‘here’, which was an extraordinary fossil already in the 17th century. Ancient haur is the origin of historical hor ‘there’; in fact, Basque deixis evolved from a proximate vs distal distinction to a three-degree system similar to those of the neighboring Romance languages (cf. Irigoyen 1997 and Martínez-Areta 2013). In the Souletin dialect, in particular, speakers reanalyzed the old proximate adverb haur as the first degree demonstrative. Subsequently new presentatives were created: they are based on the second degree demonstrative hori ‘that’ or plural forms such as hauk ‘these’ and horik ‘those’.

Furthermore, Souletin presentatives show another interesting element: the complement clause marker -la in the main clause. Apart from some modern imperative constructions, in the rest of Basque varieties -la cannot appear in main clauses. I shall argue that this peculiarity of Souletin is due to contact with Gascon. In particular, I will propose that the inclusion of the marker -la into the presentative is an adaptation of the Gascon expletive que, which is one of the five so-called “particules énonciatives” of this language and can be used in affirmative clauses (Rohlfs 1970 § 524).

References

Cohen, Eran, 2014, “Presentatives in comparative view: Biblical Hebrew and Neo-Aramaic”, in Tal Davidovich, Ablahad Lahdo & Torkel Lindquist, From Tur Abdin to Hadramawt. Semitic Studies. Festschrift in Honour of Bo Isaksson on the occasion of his retirement, Harrassowitz Verlag, Wiesbaden, 23-38. Follingstad, Carl M., 1995, “Hinnēh and focus function with application to Tyap”, Journal of translation and textlinguistics 7 (3), 1-24. Follingstad, Carl M., 2001, Deictic Viewpoint in Biblical Hebrew Text: A Syntagmatic and Paradigmatic Analysis of the Particle kī, SIL International, Dallas. Irigoyen, Alfonso, 1997 [1981], “Haur hola zirola. Elementos deícticos en la lengua vasca”, Opera selecta, University of Deusto, Bilbao, 131-170. Martínez-Areta, Mikel, 2013, “Demonstratives and Personal Pronouns”, in Martínez-Areta, Basque and Proto-Basque. Language-Internal and Typological Approaches to Linguistic Reconstruction, Peter Lang, Frankfurt, 283-321. Rohlfs, Gerhard, 1970 [1935], Le Gascon. Études de philologie pyrénéenne, Max Niemeyer Verlag, Tübingen & Pau. Trask, Robert L., 1993, A dictionary of grammatical terms in linguistics, Routledge, London & New-York.

ABSTRACT. This study examines the emergence of the negative construction ne…pas jamei ‘never’ in the Béarnais variety of Gascon, using data from a number of sources which extend from the 15th century into the 21st century. In recent texts, the indefinite temporal adverbial jamei ‘never’ appears in a variety of constructions ranging from expressions (1a) and simple negations (1b) to bipartite (1c) and tripartite negative constructions (1d). (1) a. jamei ‘forever,’ ara o jamei ‘now or never.’ b. L’ as jamei vist aidà ’ns him you.have never seen help us ‘You have never seen him help us.’ (Miquèu de Camelat Vita Vitanta, 1937) c. Mes n’ a jamei ensenhat la medecina but NEG he.has (n)ever taught the medicine ‘But he never taught medicine.’ (Miquèu de Camelat La literatura gascona, 1950) d. Ne l’ èi pas jamei prometut arren NEG him I.have NEG (n)ever promised anything ‘I never promised him anything.’ (Albert Peyroutet De la pèth de Cohet, 2000) Although the tripartite construction is recorded in grammars of different Gascon varieties (Hourcade 1986: 50, 247; Romieu & Bianchi 2005: 139), it has not been analyzed in detail, especially in relation to the other negative constructions in which jamei appears; moreover, previous accounts of Occitan negation in general are few (Jagueneau 2007; Medina Granda 2000, 2007; Price 1965). I argue that the Gascon case adds another dimension to the discussion on indefinites and negative polarity contexts, since jamei appears in both bipartite and tripartite constructions. My main objective is to trace the emergence of this tripartite negative construction, while taking into account the contexts and functions of Gascon jamei. I examine its contexts in terms of (non-)specificity and negative polarity (Haspelmath 1997) and propose to trace a pathway of emergence for ne…pas jamei, which appears to be a recent innovation. The sources of this study include a 15th century legal document, several poems from the 17th– 19th century, and lastly the Còrpus Textuau Occitan, which is an online collection of written texts, both literary and non-literary, ranging from 1898 to 2012. Seven authors produced the 18 literary texts, which include plays, poems, novels, and short stories, while the non-literary texts have an unknown number of authors. A striking feature of this corpus is the diversity of the functions of jamei. Of the 647 occurrences of jamei in the corpus, 124 appear without a negative adverb, 298 in bipartite negative constructions, and 225 in tripartite negative constructions. While this could be attributed to synchronic and diachronic dialectal variation, closer examination of the literary texts reveal variation in one text by one author. The emerging patterns suggest that overall the environment determines the choice of construction in which jamei is embedded. Cases in which jamei is not accompanied by a negative adverb are limited to particular expressions (a jamei ‘forever’), the negation of a phrase (AdvP, AdjP), or negative polarity contexts represented by questions, conditionals, having irrealis and non-specific readings, as in (2). (2) Avetz jamei vist un camp de lin florit? you.have.PL ever seen a field of flax blooming ‘Have you ever seen a flax field in bloom?’ (Simin Palay Petita vita e vitanta, 1961) Lastly, jamei, as is typical of indefinite pronouns, is found in negations and, in this corpus in particular, in bipartite and tripartite negative constructions. In De la pèth de Cohet (Peyroutet, 2000), bipartite constructions are limited to verbs in the conditional and in the future or to contexts in which other negative emphasizers are already present, indicating degrees of non-specificity. These constructions are the most frequent in the literary texts (234/419), while tripartite constructions as illustrated in (1d) are less common (67/419). This is in contrast with the non-literary texts (all from the 21st century) where tripartite constructions are prevalent (158/262). However, literary texts from only the end of the 20th and the beginning of the 21st centuries show that tripartite constructions are preferred (37/75) over bipartite constructions (17/75) suggesting that the tripartite construction has only recently become more common, possibly triggered by the development of jamei and the preverbal negator. This proposal seems to be supported by data from earlier sources, since the earliest texts of the study, from the 15th century, lack the tripartite structure entirely. Instead, as (3) illustrates, a bipartite structure is preferred, where both elements precede the conjugated verb. (3) et james non ha podut ren crubar and never NEG he.has can.PTCP anything retrieve ‘and he could never retrieve anything (i.e. money)’ (Lespy et al. Un baron béarnais au quinzième siècle, 1878) Due to the nature of the sources, there is a gap in the data for the 16th century, which might present a critical period in the development of the construction. However, the data provides an example of the tripartite as early as the 17th century, as seen in (4), while the most frequent construction remains a bipartite construction until the end of the 20th century. (4) Car lou rey lou digou, quoan fini sa harengue,/ Que n' abé pas aüdit because the king him told when finished his harangue That NEG had NEG heard.PTCP jamès ta bère lengue. never your beautiful language. ‘Because the king told him, when he had finished his harangue, that he had never heard your beautiful language.’ (Fondeville, from Lagrèze Essai sur la langue et la littérature du Béarn, 1856). The increased use of the tripartite construction in recent texts underlines the importance of considering the data in their diachronic distribution. While the earlier texts illustrate instances of bipartite constructions, the later non-literary texts contain the tripartite construction almost exclusively. Moreover, data from the Linguistic Corpus of Old Gascon (Thomas Field) will be considered to further examine the emergence of the functions of jamei and the question of the directionality of negative polarity and negative readings (Larrivée & Ingham 2012: 9).

References Camelat, M. de. (1950). La literatura gascona. Pau: Escole Gastou-Febus. Camelat, M. de. (1937). Vita Vitanta. Pau: Edicions Reclams. Haspelmath, M. (1997). Indefinite pronouns. Oxford;New York; Clarendon Press. Hourcade, A. (1986). Grammaire béarnaise. Pau: Association Los Caminaires. Jagueneau, L. (2007). Négation simple et négation discontinue en occitan limousin. In F. Floricic (Ed.), La Négation dans les langues romanes (Vols. 1–Book, Section, pp. 99–116). Benjamins. Lagrèze, G. B. (1856). Essai sur la langue et la littérature du Béarn. Gounouilhou. Larrivée, P., & Ingham, R. (2012). The evolution of negation: Beyond the Jespersen cycle (Vol. 235). Walter de Gruyter. Lespy, V. R., Lespy, J. D., & Raymond, P. (1878). Un baron béarnais au quinzième siècle. Ribaut. Medina Granda, R. M. (2000). Expresiones de valor mínimo y polaridad negativa en occitano antiguo: Elementos de comparación con otros romances medievales. Archivum: Revista de La Facultad de Filología, 50–51(Journal Article), 279–362. Medina Granda, R. M. (2007). Occitano antiguo ge(n)s: Su ausencia en ciertos contextos negativos. In F. Floricic (Ed.), La Négation dans les langues romanes (Vols. 1–Book, Section, pp. 279–362). Benjamins. Palay, S. (1961). Petita vita e vitanta. Pau: Edicions Reclams. Peyroutet, A. (2000). De la pèth de Cohet. Pau: Edicions Reclams. Price, G. (1965). Aspects de l’emploi des particules négatives en occitan. Linguistique et Philologie Romanes. Xe Congrès International de Linguistique et Philologie Romanes, 265–271. Romieu, M., & Bianchi, A. (2005). Gramatica de l’occitan gascon contemporanèu. Bordeaux: Presses universitaires de Bordeaux.

ABSTRACT. Language comparison and population history

Nina Radkevich, Andrea Ceolin, Guido Cordoni, Cristina Guardiano, Dimitar Kazakov,

Shin-Sook Kim, Monica Irimia, Giuseppe Longobardi, Dimitris Michelioudakis

Molecular anthropology and its quantitative models have changed our understanding of the

peopling of entire continents. However, it is still unclear how demographic and cultural changes

have interacted in shaping patterns of biological and linguistic diversity across such wide and

geographically diverse areas (Cavalli-Sforza and Feldman 1981, Boyd and Richerson 1985). To

understand such processes, we want to address the possible parallelism between language and gene

transmission on a broad scale, a question anticipated by Darwin’s (1859) prediction of a global

congruence of biological and linguistic variation.

1. Mathematical modeling. Through the pioneering work of Ringe, Taylor and Warnow (2002), and

Gray and Atkinson (2003), mathematical procedures of evolutionary biology have been introduced into

historical and taxonomic linguistics and applied to traditional phonological and lexical data. However,

most existing comparisons of linguistic and genetic diversity (Sokal 1988; Cavalli-Sforza et al. 1988;

Barbujani and Pilastro 1993; Belle and Barbujani 2007; de Filippo et al. 2012) are still undermined in

areal scope and quantitative resolution by their use of traditional linguistic classifications, based on lexical

cognacy: the latter fail to safely establish relatedness beyond relatively time-shallow language families

(Nichols 1996, Ringe 1996, Heggarty et al 2005), and tend to undergo non-discrete variation and

environmental selection (Levinson and Gray 2009). Therefore, more universally applicable characters

are necessary for linguistic classifications to achieve a scope comparable to population-genetic

classifications (Atkinson 2011, Creanza 2015, Hunley 2015, Jaeger 2015).

2. Formal language phylogenies. Since Longobardi and Guardiano (2009), methods importing the

deductive style of formal grammar into the study of language phylogenies and assessing the historical value

of (controversial) parametric hypotheses have been explored (Longobardi et al. 2013). The Parametric

Comparison Method (PCM) introduced in Longobardi and Guardiano (2009) employs universally

definable syntactic characters, opening the possibility of long-range language relationships comparisons.

3. Experiments and results. Combining the insights of the PCM with mathematical procedures

from evolutionary biology and genetics, we compared gene/language diversity in a sample of

populations spanning across Eurasia and belonging to several traditionally irreducible linguistic

phyla (Indo-European, Uralic, Altaic, Semitic, Sinitic, Dravidian, North Caucasian and others).

We built a database of over 3000 parametric states (90 each for 35), which deductively define

several thousands of syntactic phenomena. This database is meant to preserve the deductive depth

and predictive intricacy of highly axiomatized syntactic theories, while exceeding the average

empirical coverage of most current formal syntactic analyses by more than one full order of

magnitude (Baker 2012 calculated that the average number of languages compared in formal

analyses involving parameters rarely exceeds 4 per article).

Language distances were compared with those defined through whole-genome autosomal SNP

markers (~300,000 SNPs) in several representative individuals from corresponding populations.

The extent and significance of the obtained overall gene-language correlation (Mantel r = 0.50)

stands out as strong prima facie evidence for a non-trivial association between genomic and

linguistic diversity. However, given the well-established effect of geography on human

diversification (Prugnolle et al 2005, Novembre et al 2008), it is critical to control for its impact

on the observed pattern of gene-language congruence. Thus, the effect of spatial relationships

between populations was further modeled through five different types of geographical

measurement: the straightforward Great Circle Distance (GCD), and four using metrics which

attempt to account for the complexity of the territory and possible migration routes available:

Roads Map (RM), GCD with waypoints (GCDWP), Least Cost Path (LCP), and Commuting (CO).

The results indicate that across Eurasia; 1) syntactic diversity distributes within and across ancient

language families in a robust tree-like form, accounting for nearly all the expected clades (cf. the

phylogenetic tree Fig. A below), 2) languages correlate with geography only as a byproduct of

their congruence with genetics, while 3) the gene/language congruence is salient and significantly

independent of geography in Europe and in east Asia, (regardless of the methods used to calculate

geographical distances); a mismatch (rich linguistic diversity corresponding to strong genetic

uniformity) is instead found in Western Asia/Near East. This suggests that genes and languages

spread across Eurasia mainly as a result of the same demographic processes at the two (Eastern

and Western) extremes of the macro-continent, with a remarkable area of genetic admixture

without comparable linguistic admixture in the intermediate area.

4. Conclusions. Overall, our experiments revealed a high correlation between syntactic variation

and genetic diversity. Furthermore, when controlled for geography, the correlation turned out to

be higher (and statistically more significant) than those previously observed using phonological

variables (Creanza et al. 2015). Thus, the results largely fulfill Darwin’s expectation, suggesting

that grammars and genes prevailingly diffused along similar routes, and mostly through the same

demographic events, with limited cases of elite-dominance language replacement (Renfrew 1992),

and no evidence for the reverse trend (massive genetic introgressions without relevant linguistic

changes). This general pattern is only seriously disrupted in an area at the crossroads of three

continents where very different languages (Turkic, IE, Semitic, North East Caucasian) seem to

have been adopted by populations characterized by conspicuous and continuous genetic

uniformity.

The comparisons of linguistic and genetic variation at such a large scale have crucially been made

possible by the adoption of properties from a theory of linguistic diversity rooted in a strong

conception of universal grammar.

To conclude, our experiments show that characters provided by a formal theory of syntactic

variation can complement lexical and phonemic data in historical linguistics and anthropological

reconstruction, and support demographic history with previously unavailable linguistic insights.

In fact, the correlation between syntactic and genomic diversity provides evidence that the

parametric format is a realistic way of modeling diachronic transmission of language.

[Figure A & Footnote References included in PDF version of the abstract from the ICHL program index]

ABSTRACT. German modal particles (MPs) show all the signs of traditional grammaticalization on the paradigmatic dimension, but none of those on the syntagmatic dimension, based on Lehmann's (1985) influential definitions. I argue that this unexpected mismatch is ultimately due to the fact that MPs are restricted to the "middle field" of the German clause, in which scrambling is an option. Data from natural spoken language show that the collocation patterns of MPs are essentially unconstrained. This syntagmatic freedom explains why MPs have clausal rather than narrow scope, and also that they do not show signs of morphophonological dependence (cf. the Linear Fusion Hypothesis; Bybee 2002). One of the few MPs restricted to questions, denn, does cliticize, however, and I argue that this is because the middle field in questions is less elaborated and offers fewer options for scrambling, and hence more for fusion. Taken together, these data suggest that a diachronic constructional approach, which is capable of explaining the observed data by taking the various contexts of MPs into account, is superior to traditional grammaticalization, with its focus on elements in isolation.

ABSTRACT. Since Kiparsky & Kiparsky (1970), Spanish el hecho de (que) (‘the fact of/that’) has been used as a standard test for factivity (1). It has traditionally been considered as semantically equivalent to the clausal nominalization el que (lit. ‘the that’) (2) (Demonte 1977; Leonetti 1999):

(1) María lamenta el hecho de que Pedro esté triste ‘Mary regrets the fact that Peter is sad’ (2) María lamenta [D el [ CP que Pedro esté triste ] ] ‘Mary regrets (lit. the) that Peter is sad’

Serrano (2015) has proved that el que is not semantically restricted to factivity, but el hecho de (que) continues to be typically analyzed as factive (Palmerini 2006; Becker 2014; Serrano 2015; a.o.). Furthermore, Becker (2014: 304) claims that it is a late arrival in Spanish, emerging in the 19th century, whereas el que is attested in the 16th century (author 2013, a.o.). Thus, recent studies analyze el hecho de (que) as a modern factive construction. However, we prove that this conclusion is untenable.

In this paper we examine the diachrony of el hecho de (que) with two main goals. Firstly, we show that the N hecho (or fecho) modified by a clause –whether finite or non-finite– is already attested as early as the 16th century (3). This development patterns with a broader syntactic change in Spanish in the evolution of clauses selected by nouns (author 2013, a.o.):

(3) Se pregunta si es causa justa de hacer la guerra la infidelidad de los bárbaros, y el hecho de que rechacen el evangelio (José de Acosta, 16th c.) ‘S/he wonders whether the unfaithfulness of the barbarians and the fact that they reject the gospel are fair causes for war’

Secondly, while (3) is clearly factive, we provide historical evidence that el hecho de (que) could also receive non-factive interpretations from the beginning. We find that non-factive readings are linked to the non-veridicity of the whole sentence through (i) the negation of positive implicative verbs – in the sense of Kartunnen (1971) – like consentir ‘allow’ (4) and through (ii) non-veridical selecting predicates (5). Non-factive contexts continue to be attested beyond the 16-17th centuries (6-7) and even to this day (authors 2016):

(4) el hecho de degollar a su hijo de ninguna suerte lo consintió (José de Acosta, 16th c.) ‘The fact of slitting his son’s throat, he didn’t allow it at all’ (5) En el hecho de que fuese muerto por nuestros segovianos hay mucha duda (Diego de Colmenares, 17th c.) ‘It is highly doubtful that he was killed by our people from Segovia’ (6) Este discurso supone el hecho de que la agua fría es remedio de la hydrophobia, lo qual es mui dudoso, o falso (Feijóo, 18th c.) ‘This speech supposes the fact that cold water is the remedy for hydrophobia, which is very doubtful or false’ (7) ¿De modo que consideras como cierto el hecho de que ese hombre ame a doña Catalina y a doña María? (Álvaro Carrillo, 19th c.) ‘So you consider it to be true that that man loves lady Catalina and lady María?’

In sum, our paper provides new evidence proving that el hecho de (que) is attested much earlier than previously acknowledged and also challenge existing semantic analyses of el hecho de (que) as factive. Our findings also highlight the role of the veridicality properties of the whole sentence in determining factivity across languages (Schulz 2003).

More broadly, our paper contributes to the discussion on factivity and clausal nominalization, both synchronically and diachronically. In situating the history of el hecho de (que) within the general pattern of clausal subordination to nouns in Spanish, our paper contributes to the exploration of the compositionality of constructions or collocations and the nature of light nouns (Pustejovsky 1995; Mendívil 1999; Panagiotidis 2015; a.o.). Finally, our paper brings a historical perspective to debates on veridicity (Karttunen & Zaenen 2005), an area that has been understudied.

References

Author 2013. Anonymized reference. Authors 2016. Anonymized reference. Becker, M. 2014. “Informational status and the semantics of mood in Spanish preposed complement clauses”. In A. Dufter & A. Octavio de Toledo (eds.), Left Sentence Peripheries in Spanish, 283-308. Amsterdam: John Benjamins. Demonte, V. 1977. La Subordinación Sustantiva. Madrid: Cátedra. Karttunen, L. 1971. The Logic of English Predicate Complement Constructions. Bloomington: Publications of the Indiana University Linguistics Club. Karttunen, L. & Zaenen, A. 2005. “Veridicity”. In G. Katz, J. Pustejovsky & F. Schilder (eds.), Annotating, Extracting and Reasoning about Time and Events. Dagstuhl (Germany). http://drops.dagstuhl.de/opus/volltexte/2005/314/ Kiparsky, P. & Kiparsky, C. 1970. “Fact”. In M. Bierwisch & K.E. Heidolph (eds.), Progress in Linguistics, 143-173. The Hague: Mouton. Leonetti, M. 1999. “La subordinación sustantiva: las subordinadas enunciativas en los complementos nominales”. In I. Bosque & V. Demonte (eds.), Gramática Descriptiva de la Lengua Española, 2083-2104. Madrid: Espasa-Real Academia Española. Mendívil Giró, J.L. 1999. Las Palabras Disgregadas. Sintaxis de las Expresiones Idiomáticas y los Predicados Complejos. Zaragoza: Prensas Universitarias de Zaragoza. Palmerini, M. 2006. “Factividad en la nominalización: sobre la estructura [el hecho/il fatto + subordinada] en español e italiano”. In C. Mourón & T. Moralejo (eds.), Studies in Contrastive Linguistics, 761-768. Santiago: Universidad de Santiago de Compostela. Panagiotidis, P. 2015. Categorial Features. A Generative Theory of Word Class Categories. Cambridge: Cambridge University Press. Pustejovsky, J. 1995. The Generative Lexicon. Cambridge: MIT Press. Schulz, P. 2003. Factivity: its Nature and Acquisition. Tübingen: M. Niemeyer. Serrano, S. 2015. Subordinación y Determinación: Completivas precedidas de Artículo Definido en Español. PhD diss., Universidad Autónoma de Madrid & Instituto Universitario Ortega y Gasset.

ABSTRACT. Most of the computational approaches in research on paradigm leveling make use of artificial neural networks, machine learning or the direct simulation of paradigm leveling by the computer (cf. Albright 2005). In contrast to these, the purpose of this paper is to present the results of the research on the typology of paradigm leveling based on the computational simulation by 'Phono' (Hartman 2003) of the four sound changes that occurred in the history of four different Indo-European languages: the Latin rhotacism, the Indo-Iranian Brugmann's law, the development of the Proto-Indo-European labiovelars in Ancient Greek and the Polish vowel metaphony. The computational simulation of sound change consists of the application of rules that are modeled on the sound changes that actually occurred in the development of a language on a dataset that is the earlier stage of its development (e.g. simulating the sound changes from Latin to Spanish). The results of the application can then be compared with the actual attested data in order to uncover the rate of the regularity of the changes and the correct order in which they are applied. Most of the approaches to the simulation of phonological changes operated on small, selected lexicons and usually with only a selection of rules (cf. Smith 1969, Burton-Hunter 1976, Eastlack 1977, Hartman 2003, Mittmann 2009). The authors of the programs were mostly interested in the regularity of the sound changes and their relative chronology. In this talk we will discuss the use of the most advanced of the programs simulating phonological change created so far - 'Phono' by Steven Lee Hartman (Hartman 2003) - which makes use of the IPA alphabet and the matrix of distinctive features rather than simple letter substitution rules. Our aim will be to discuss the direction of the paradigm leveling in the data driven from the results of the application of the four sound changes to the pre-stages of the languages in which those changes occurred (cf. Kümmel 2007): 1) the Latin rhotacism (from pre-Latin) 2) the Indo-Iranian Brugmann's law (from Proto-Indo-Iranian) 3) the development of the Proto-Indo-European laryngeals in Ancient Greek (from Proto-Greek) 4) the Polish vowel metaphony (from pre-Polish), cf. Stieber 1973: 24-26 The mechanical application of the sound changes mentioned above to the lexical datasets of the pre-stages of the respective languages creates allomorphy in their paradigms which is then leveled out according to seemingly different principles and in different directions. However, with the use of extensive datasets it is possible to point out important factors concerning the directions of the levelings, their frequency and the dominating patterns. It has been observed 50 years ago by Warren Cowgill that „at present each linguist judges the plausibility of a newly proposed solution pretty much by what he happens to remember of the morphologic innovations which during his career he has been led, for one reason or another, to accept as plausible. A reasonably objective standard of plausibility should make it easier for historical linguists to agree on solutions for problems of historical morphology that at present are still disputed (…). A sufficiently large collection of such individual changes, appropriately classified, should give linguists a measure of the relative plausibility of different solutions for problems in historical grammar” (Cowgill 1966: 115). It is our hope that this work might advance our knowledge of the typology of paradigm leveling and analogical change in general by the use of extensive datasets and a wide variety of data driven from various languages.

References

Albright, Adam C. 2005. “The morphological basis of paradigm leveling”. In: Laura J. Downing, T. Alan Hall, Renate Raffelsiefen (eds.) Paradigms in Phonological Theory. Oxford: University Press. 17-43.

Burton-Hunter, Sarah K. 1976. “Romance etymology: a computerized model”. Computers and the Humanities, 10, 217-220.

Cowgill, Warren. 1966. “A Search for Universals in Indo-European Diachronic Morphology”. In: Greenberg, Joseph (ed.) Universals of Language. 2nd ed. Cambridge: MIT Press. 114-141.

Eastlack, Charles L. 1977. “Iberochange: a program to simulate systematic sound change in Ibero-Romance”. Computers and the Humanities, 11, 81-88.

Hartman, Steven Lee. 2003. "Phono (Version 4.0): Software for Modeling Regular Historical SoundChange". In: Leonel Ruiz Miyares, Celia E. Álvarez Moreno, and María Rosa Álvarez Silva(eds.), Actas: VIII Simposio Internacional de Comunicación Social: Santiago de Cuba, 20-24 de Enero del 2003, I, 606-609.

Kümmel, Martin Joachim. 2007. Konsonantenwandel. Wiesbaden: Reichert Verlag.

Mittmann, Roland. 2009. “Ein Verfahren zur Ermittlung der relativen Chronologie der vorgotischen Lautgesetze“. In: Chiarcos, Christian / Eckart de Castillo, Richard / Stede, Manfred (ed.): Von der Form zur Bedeutung: Texte automatisch verarbeiten. From Form to Meaning: Processing Texts Automatically. Proceedings of the Biennial GSCL Conference 2009. Tübingen: Narr, 199–209.

Smith, Raoul N. 1969. “A computer simulation of phonological change”. ITL: Tijdschrift voor Toegepaste Linguistiek, 1(5), 82-91.

Stieber, Zdzisław. 1973. A historical phonology of the Polish language. Heidelberg: Carl Winter.

ABSTRACT. Introduction: Various impersonal pronouns (Imp) in Romance are derived from the Latin noun HOMŌ, “man.” For example, Modern French (MF) has on (1) and Catalan has hom (2). (1) On ne parle pas anglais ici. (2) Hom no parla anglès aquí. Imp Neg speaks Neg English here Imp Neg speaks English here “One does not speak English here.” “One does not speak English here.” Old Spanish (OS) omne could be used to express a generic subject, as in (3) and (4), until other impersonal constructions replaced it in the sixteenth century (Brown 1931, Pozas Loyo 2004). (3) Quando omnei trabaja...proi non siente tenptaçión (4) Cuando omne non puede dormir... When Imp works Imp Neg feels temptation when Imp not can sleep “When one works one does not feel temptation.” “When one cannot sleep....” Why did this pronoun disappear in Spanish but persist in French? I propose that omne was subject to a linguistic cycle, in the sense of Van Gelderen (2011), that failed to reach completion; i.e., a “stalled” cycle. In MF, the cycle is complete with on fully grammaticalized. I take on to be a functional head that licenses generic pro (Mendikoetxea 2008). Some researchers attribute the disappearance or emergence of HOMŌ pronouns to language contact (Giacalone Ramat & Sansò 2007, D’Alessandro 2014) or lack of frequency (Brown 1931). I claim that omne disappeared due to competition from another method of generic subject expression, impersonal se (Impse), which was already grammaticalized as a head in the sixteenth century (Maddox 2016), when omne was a full DP. Chomsky’s (1995) Late Merge Principle (LMP) states that it is more economical to merge as a head rather than merge and then move. This also applies diachronically (Van Gelderen 2011). In this case, when posed with two options to express a generic subject, omne or Impse, a speaker was more apt to “choose” Impse for purposes of economy, which also explains the lack of frequency of omne compared to Impse (Brown 1931). Diffrences in distribution of omne and on is evidence that they were subject to a linguistic cycle. Degrees of deficiency: Following Van Gelderen (2011), tests of coordination and modification can diagnose the status of an element as a head (X) or a phrase (XP). Thus, phrases tend to be reanalyzed as heads diachronically because it is more economical. MF on has the properties of a head since it cannot be coordinated (5), modified (6), or separated from the verb (7) by XPs. (5) *On et elle parlent beaucoup. (6) *On intelligent parle français. (7) *On souvent arrive tard. Imp and she speak a-lot Imp intelligent speaks French Imp often arrives late Egerland (2003) proposes two classes of impersonal HOMŌ pronouns. Class 1 pronouns like MF on can have variable agreement (other than masculine singular) (8), can have either a generic (1) or arbitrary reading (9), and cannot be objects (10). Class 2 impersonal pronouns cannot have variable agreement, can only have generic interpretation, and can be subjects and objects. (8) Quand on est belles... (9) On a travaillé deux mois pour résoudre le problème. when Imp is pretty.F.Pl Imp has worked two months to resolve the problem “When one is pretty...” “Someone worked two months in order to fix the problem.” (10) *Ils on voient. (11) ca non ensuzia a omne comer con las manos non lavadas. they Imp see for not soils DOM Imp eat with the hands not washed Intended: “They saw one.” “For it does not soil a person to eat with their hands unwashed.” Omne belongs to Class 2. A search in the CORDE yielded no evidence of variable agreement or arbitrary interpretation. Omne can, however, function as an object, as in (11) above. Omne additionally qualifies as a weak pronoun DP following Cardinaletti & Starke (1999) since it can be separated from the verb (12), it can occupy XP positions and it can be modified (13). (12) si omne bien non cata... (13) ... todo omne que deva deuda a otro... if Imp well not observe every Imp that owes debt to another “In one does not observe well...” “...every one that owes debt to another...” Thus, omne was a weak Class 2 impersonal pronoun. It was deficient but not as deficient as MF on since it retained agreement features, had only a generic interpretation, and displayed pronominal distribution. Omne stalled at an earlier stage in the cycle than MF on has reached. Coreferentiality: Maddox (2016) shows that another impersonal pronoun, Spanish generic uno, licenses a generic subject pro across clauses via topic-identification while Impse does not. This difference is due to uno being a full DP and thus eligible for topichood. Since Impse is a functional head, it cannot serve as a topic. If omne was a full DP in OS and on is a functional head in MF, similar patterns are predicted where omne should license pro crossclausally while on will only license it clause internally. A factor that complicates this is that MF is a non-null subject language lacking bare subject pro. In fact, MF requires on in each clause (Cabredo-Hofherr 2004, 2010). Nevertheless, Old French (OF) had null subjects and so the diagnostic can be applied to OF which also had on, though with orthographic variation. Historical data support the prediction: OF on licenses generic pro crossclausally (14) as does OS omne, as in (3) above. (14) ne se peüst oni   porpenser de richece ...en cel leu trover ne proi peüst. not Pronse could Imp imagine of richness in that place find   not pro could “One could not imagine any richness...in that place that one could not find.” Since OF on licensed pro in the same way as omne and uno, two full DPs, it may not have been fully grammaticalized as a functional head; i.e., in OF it may have still been a weak Class 2 impersonal pronoun like omne. Supporting evidence for this is in (15) and (16) from Welton-Lair (1999), showing that on can be separated from the verb by XPs; i.e. interpolation. (15) Sainz Boneface que l’um   martir apelet. (16) Quant l’en en la meson Dieu entre. Saint Boniface whom the-Imp martyr call when the-Imp in the house God enters “Saint Boniface, whom people call martyr.” “When one enters the house of God.” The parallel with (12) is expected if OF on and OS omne were less grammaticalized than MF on. Cyclic Replacement: In MF, the impersonal pronoun cycle is complete. The clitic on, originally a full DP, is now a functional head that licenses pro locally, like Impse in Spanish. This cycle failed to reach completion in OS with omne because Impse was already grammaticalized as the same generic pro-licensing head while omne was still a full DP, albeit a weak pronoun. Since it is more economical to merge as a head rather than merge and then move as a DP to satisfy feature-checking per the LMP, speakers selected Impse more often than omne, which eventually disappeared. Generic uno has now replaced omne in the impersonal pronoun cycle. This explains why Impse does not exist in French (Dobrovie-Sorin 1998); i.e., on was grammaticalized as the head marking impersonality early on and, consequently, se was never reanalyzed as such. References: Brown, C.B. 1931. The Disappearance of the Indefinite Hombre from Spanish. Language 7:265-277. Gelderen, E. van. 2011. The Linguistic Cycle. Oxford: UP. Maddox, M. 2016. Null Generic Subjects in Spanish and the Typology of Null Subject Languages. Paper at LSRL 46 at Stony Brook University. Mendikoetxea, Amaya. 2008. Clitic Impersonal Constructions in Romance: Syntactic Features and Semantic Interpretation. Transactions of the Philological Society 106:290-336. Diachronic data: (3) Vicente Ferrer, Sermones; (4) Juan Manuel, El Conde Lucanor; (11) Alfonso X, Setenario, p. 248; (12), Gonzalo de Berceo, Vida de San Millán de Cogolla, p. 249; (13), Fuero de Burgos, Philadelphia Ems.245, para. 516; (14), Roman d’Énéas, 456-458; (15), Anonymous, Alexis, .566; (16), Rutebeuf., Le Miracle de Théophile, 275.

ABSTRACT. Eastern Basque dialects in France such as Zuberoan (also Souletin) and the Mixean variety of Low Navarrese (Camino 2016) are far from being transmitted in the way Standard Basque or some Western Basque varieties are. Historically, both dialects have shared a diglossic situation with Bearnese Gascon, but now all these languages are disappearing to French. Ethnologue (Gordon 2005) lists 8700 speakers of Zuberoan in 1991, but the number of speakers has decreased drastically in the region during the last decades (Barreña et al. 2013: 14), and only one out of ten children speak Zuberoan today (Camino, p.c.). The situation is similar in Mixe. This paper, as well as other previous studies (Egurtzegi 2014), aims to show that these two lesser studied eastern dialects of Basque are linguistically very interesting. Among many other interesting features not found in other Basque varieties, Zuberoan and Mixean show a contrast between /h/ and a nasalized /h̃/ (compare ehe ‘wash water’ vs. eh̃e [ẽh̃ẽ] ‘no’, cf. Hualde 1993; Egurtzegi 2014), which continues Proto-Basque intervocalic *n (Michelena 1977; Igartua 2015). This cross-linguistically rare contrast between oral and nasalized aspirates has been lost in different ways in the other Basque varieties: while western Basque dialects have lost aspiration altogether, the eastern Basque dialects Lapurdian and Low Navarrese have merged /h/ and /h̃/ in /h/ (1). This contrast shows traces of recession in Zuberoan and Mixean as well, but it seems to be following different fading patterns: While in Zuberoan some etymologically nasalized /h̃/s have lost nasalization in the modern language (compare Zuberoan ahal ‘to be able to’ to the derivative form ah̃alke ‘embarrassment’, both with a nasalized /h̃/ in Mixean Basque), recordings of speakers of Mixean Basque attest a tendency towards nasalizing aspiration in contexts where no nasalization is historically found, including the examples in (2). This tendency has not been described for the Zuberoan dialect, but can be found in some speakers in the towns neighboring the region of Mixe. In this paper, I analyze the loss of the contrast between /h/ and /h̃/ as an instance of equation of aspiration with nasalization: at some point, the speakers of Mixean Basque have begun to produce instances of the aspirate /h/ (<*h) with audible phonetic nasalization (cf. Ohala 1975), as occurred in other languages such as Thai (/haː/ [hãː] ‘five’, cf. Matisoff 1975), the Nepalese language Hayu or Scottish Gaelic (àtha [ãːhə] ‘fore (gen. sg.)’, Ó Maolalaigh 2003), in a process usually known as rhinoglottophilia (Matisoff 1975; Blevins 2004). In contrast to what happened in these other languages, given that Mixean Basque contrasted /h/ and /h̃/, the overgeneralization of phonetic nasalization in aspirates may yield the loss of a scarcely attested phonological opposition in this variety of the language. In short, this paper presents a description of the loss in progress of a cross-linguistically rare phonological opposition between /h/ and /h̃/ in the endangered Mixean and Zuberoan Basque, showing that the study of endangered languages and dialects (Jones 1998) can illustrate typologically unusual sound patterns. In addition, a detailed description of this opposition strengthens the reconstruction of previous stages of Basque by offering the possibility of measuring with modern means an otherwise lost segment that was common to all varieties of the language. (1) Modern reflexes of Proto-Basque *h and intervocalic *n Reconstruction Western dialects Lapurdian and Low Navarrese Zuberoan and Mixean Gloss *hasi /as̺i/ /has̺i/ /has̺i/ ‘to begin’ *aho /ao/ /aho/ /aho/ ‘mouth’ *anari /aaɾi/, /aɾi/ /ahaɾi/ /ah̃ai/ ‘ram’ Lat. anate(m) /aate/, /ate/ /ahate/ /ah̃ate/ ‘duck’ (2) Non-etymological /h/ nasalization in Mixean Basque Zuberoan Mixean (sporadically) Gloss behar beh̃ar ‘must’ behi beh̃i ‘cow’ bihi bih̃i ‘seed’ zahar zah̃ar ‘old’ hau h̃au ‘this’ hotz h̃otz ‘cold’ heuskaldun h̃euskaldun ‘Basque’ hogei h̃ogei ‘twenty’ References Blevins, J. 2004. Evolutionary Phonology. The emergence of sound patterns. Cambridge: CUP. Camino, I. 2016. Amiküze eskualdeko heskuara (Mendaur saila 11). Bilbao: Euskaltzaindia. Egurtzegi, A. 2014. Towards a phonetically grounded diachronic phonology of Basque. PhD dissertation, University of the Basque Country. Gordon, R.G., Jr. (ed.). 2005. Ethnologue: Languages of the World, 15th ed. Dallas: SIL. Hualde, J.I. 1993. Topics in Souletin phonology. J.I. Hualde & J. Ortiz de Urbina (eds.), Generative Studies in Basque Linguistics, 289-327. John Benjamins. Igartua, I. 2015. Diachronic effects of rhinoglottophilia, symmetries in sound change, and the curious case of Basque. Studies in Language 39: 635-663. Jones, M.C. 1998. Language Obsolescence and Revitalization. Linguistic change in two sociolinguistically contrasting Welsh communities. Oxford: Clarendon Press. Matisoff, J.A. 1975. Rhinoglottophilia: The mysterious connection between nasality and glottality. In C.A. Ferguson, L.M. Hyman & J.J. Ohala (eds.), Nasálfest: Papers from a symposium on nasals and nasalization, 265-287. Stanford: Stanford University Press. Michelena, L. 1977. Fonética histórica vasca. San Sebastián: Diputación Foral de Guipúzcoa. Ohala, J.J. 1975. Phonetic explanations for nasal sound patterns. In C.A. Ferguson, L.M. Hyman & J.J. Ohala (eds.), Nasálfest: Papers from a symposium on nasals and nasalization, 289-316. Stanford: Stanford University Press. Ó Maolalaigh, R. 2003. Processes in nasalization and related issues. Ériu 53: 109-132.

ABSTRACT. Muya (Minyag/ Miyao, Qiangic, Tibeto-Burman) is an endangered language spoken by approximately 9,000 Tibetan people in Shade (Kangding County) 康定沙德乡, Shimian County (Ya’an) 雅安石棉 and Jiulong County九龙县 at the Ganzi Tibetan Autonomous Prefecture in Western Sichuan, China. According to the UNESCO statistics, this language is severely in danger since the 1980s. Like its Qiangic counterparts, Muya is notable for the spatial expressions, with four individual settings to encode spatial reckoning. To illustrate: • Landscape vocabulary, where speakers use ‘sun rise’ and ‘sun set’ to encode the only two members of the absolute frame as EAST and WEST: (1) ȵɯ53 qhɯ31=tsyi31-ʐɔ31khu31 mɯ31 sun EAST: sun rise place be.at ‘The sun rises from the east [lit. The sun is at the place where sun rise].’

(2) ȵɯ53 nu31=ɕie31 -ʐɔ31khu31 mɯ31 sun WEST: sun set place be.at ‘The sun goes down from the west [lit. The sun is at the place where sun goes down].’

• Spatial relators, which locate the relative spatial relations with reference to the upriver vs. downriver system as well as the conceptualized distance distinctions:

(3) khyi55=mɛ55 tɕo55=tsie55 ɣæ55 ȵɯ55ʐiɹ55khu53 mɯ53 bag-TOP desk GEN front: far to upriver be.at ‘The bag is in front of the desk. [lit. The bag is in the front of the desk, where the space of bag is conceptualized to be included in that of the desk far to the upriver].’

(4) khyi55=mɛ55 tɕo55=tsie55 ɣæ55 qɑ55ʐe31 mɯ53 bag-TOP desk GEN front: close to downriver be.at ‘The bag is in front of the desk. [lit. The bag is in front of the desk, where the space of bag is conceptualized to be close to the downriver which is very far to the desk].’

(5) ndõ55mbi33 ɣæ55 pu55 wa33tsi53 khi55=ɡɔ33 ni33 bottle GEN top: close to upriver socks big COP ‘There is a pair of big socks on the top of the bottle [lit. the location of socks is close to upriver].’ (6) ndõ55mbi33 ɣæ55 tsiɛ55muɐ33 ȵɯ53 ni33 bottle GEN above: far to river sun COP ‘The sun is above the bottle [lit. the sun is far to the riverside].’

• General relators, which locate the left and right sides. Since Muya is in short of the spatial orientation of LEFT and RIGHT, this couple of spatial concepts is described by an ambiguous relator tɑ33=tɕi53:

(7) ɦo55-tsi33 ja33-pi53 tʂɑ55-nbuɯ33 tɑ33=tɕi53 mɯ33 DEM CL pen book side: right vs. left ? be.at ‘This pen is close to the right/left of the book [lit. This pen is close to the sides of the pen, right or left].’

• Morphologically, directional prefixes bounded with predicates to denote the directional/spatial meanings. Collocation between the two items is flexible and arbitrary in some cases. Unstable combining may trigger out derived aspectual and modality meanings when connecting with the dynamic and static predicates separately:

Directional prefixes tə31- upward/ general direction nɐ31- downward ɣə31- to upper stream/ to opposite ɦɑ31- to downstream/ turn back ɡə31- towards the addresser thɑ33- far from the addresser rə31- circling khə31- general direction ʑi31- upward and downward, repeated

The primary goal of this paper is to test, refine and reformulate the spatial systems in Muya based on a broad spatial frame of reference proposed by Brown 1994 and Levinson 2003. In particular, I confine myself to the historical development of the spatial categories, focusing especially on issues of grammaticalization of spatial grams (Svorou 1993, Heine 1997) and the genesis of the quadripartite spatial expressions in this language. Finally, I argue that this multiple spatial expressions are derived from the internal development of the Qiangic language on the one hand, and the external factors caused by language contact within the western Sichuan corridor on the other (Shirai 2009), which demonstrates a pretty areal typology of this region.