ABSTRACT. Variable patterns of linguistic diversity around the globe suggest that historical processes of diversification, maintenance, and contact have applied in significantly different ways from one region to the next. While this variability is undoubtedly linked to a range of factors, such as geography, latitude, demographics, etc. (e.g. Nichols 1992, Nettle 1999, Collard & Foley 2002, Currie & Mace 2009, Gavin et al. 2013), it is also anchored in local sociocultural practices and ideologies (e.g. Thomason and Kaufman 1988, Epps forthcoming, Di Carlo forthcoming). This point comes into particular focus in small-scale speech communities, where speakers’ attitudes toward interaction with speakers of other languages – and the linguistic outcomes of that interaction – may vary considerably, both across regions and in comparison with larger-scale urban and/or globalized contexts.

            The Amazon basin offers an intriguing illustration of the degree to which processes of diversification and maintenance may vary in global perspective. Despite being the last continent reached by human expansion, Amazonia harbors extreme linguistic diversity on the level of language families; moreover, most of these families have only a few members, and a high proportion are isolates – implying long-term processes of maintenance with constrained diversification. While earlier accounts assumed these outcomes to be a factor of small, isolated populations, recent work suggests that pre-Colombian Amazonia was spanned by intensely interactive multilingual networks.

The Vaupés region of northwest Amazonia offers a glimpse into one region’s version of multilingual interaction, and its linguistic implications. In the Vaupés, where multilingual interaction is enabled in part by the local marriage practice of linguistic exogamy, many of our expectations about the outcomes of contact appear to be turned on their heads: High levels of multilingualism are accompanied by closely constrained code-switching, low rates of lexical borrowing (with more verbs borrowed than nouns), heavy structural diffusion, and the apparent absence and/or invisibility of shift among local languages. While seemingly highly unusual in cross-linguistic perspective, a wider lens reveals that comparable dynamics, and similar linguistic outcomes, are found elsewhere in Amazonia as well. Moreover, a large-scale survey of lexical and grammatical contact effects in some 60 northern Amazonian languages indicates that a close association between language maintenance and language contact has been widespread, despite regional variation. 

Finally, a still wider lens allows us to compare the dynamics of language contact, maintenance, and diversity in Amazonia with those in other regions of the world – such as northern Cameroon, Vanuatu, and Arnhem Land – and consider how these dynamics may be structured by particular sociocultural practices and attitudes associating language and social group membership. Given the likelihood that most of human history was characterized by small-scale, multilingual societies, this exploration leads us to consider the degree to which our understanding of processes of language contact and change may be guided by broad-brush uniformitarian principles, and how our understanding of the limitations of those principles depends on fine-grained investigations of languages spoken in diverse regions and sociocultural contexts.

ABSTRACT. The Muslim conquest of most territories of the Iberian Peninsula in 711 led to a progressive islamization and arabization of the native population, similar to that occurring in other areas of the newly conquered Islamic lands. Despite the fact that the number of native Arabic speakers who took part in the 711 conquest or arrived subsequently was very limited, colloquial Arabic gradually replaced Romance as a spoken language whereas classical Arabic replaced Latin as a written language and language of culture. A tendency to monolingualism with Arabic seems to have dominated in Muslim Iberia since the end of the eleventh century and there is scarce evidence of any use of Romance from the 13^th century until the end of the “Reconquista” in 1492.

In this paper I will explore the evolution of the use of Romance in al-Andalus and, more specifically, its displacement by colloquial Arabic. This linguistic process can only be accounted for in the context of the belligerent situation between the (Arabic) Muslim state of the South and the (Romance) Christian Kingdoms of the North. The identification of Arabic with Islam and that of Romance with Christianity fostered the pro-Arabic and anti-Romance linguistic attitudes that prevailed among the Muslims in al-Andalus in the periods of highest tension between the two worlds, from the 11^th and, very especially, from the 13^th century onwards. In the first two centuries of Islamic rule, however, Romance was not only widely used but also generally devoid of overt political and religious connotations as we can learn from the existing references in the Arabic Muslim sources of that period. The contact of Romance with colloquial Arabic during this early period left a significant linguistic imprint on both languages. This contact came to a minimum as the Reconquista advanced and the Christian population of al-Andalus became the last stronghold for Romance in al-Andalus.

The Muslim conquest of most territories of the Iberian Peninsula in 711 led to a progressive islamization and arabization of the native population, similar to that occurring in other areas of the newly conquered Islamic lands. Despite the fact that the number of native Arabic speakers who took part in the 711 conquest or arrived subsequently was very limited, colloquial Arabic gradually replaced Romance as a spoken language whereas classical Arabic replaced Latin as a written language and language of culture. A tendency to monolingualism with Arabic seems to have dominated in Muslim Iberia since the end of the eleventh century and there is scarce evidence of any use of Romance from the 13^th century until the end of the “Reconquista” in 1492.

In this paper I will explore the evolution of the use of Romance in al-Andalus and, more specifically, its displacement by colloquial Arabic. This linguistic process can only be accounted for in the context of the belligerent situation between the (Arabic) Muslim state of the South and the (Romance) Christian Kingdoms of the North. The identification of Arabic with Islam and that of Romance with Christianity fostered the pro-Arabic and anti-Romance linguistic attitudes that prevailed among the Muslims in al-Andalus in the periods of highest tension between the two worlds, from the 11^th and, very especially, from the 13^th century onwards. In the first two centuries of Islamic rule, however, Romance was not only widely used but also generally devoid of overt political and religious connotations as we can learn from the existing references in the Arabic Muslim sources of that period. The contact of Romance with colloquial Arabic during this early period left a significant linguistic imprint on both languages. This contact came to a minimum as the Reconquista advanced and the Christian population of al-Andalus became the last stronghold for Romance in al-Andalus.

ABSTRACT. Ramón Menéndez Pidal´s (1926 [²1950]) foundational work on the earliest attestations of written Castilian paid due attention to dialectal variation within the Central Northern Iberian space between the C10 and the C12. Unfortunately, Pidal himself helped build the illusion that the surge of highly elaborate literary monuments in the C13 led to the implantation of a solid standard variety of Castilian almost free from identifiable dialectal import and superimposed on neighboring varieties as Castile expanded southwards (cf. Fernández-Ordóñez 2009). Recent research, however, has stressed the ongoing transfer of both Western and Eastern dialectal solutions into literary texts produced in Castile throughout the Middle Ages (cf. a.o. Rodríguez Molina 2010, Fernández-Ordóñez 2011 and references therein, Octavio de Toledo 2015, Bouzouita 2016, Garachana 2016). In my talk, I will offer a few examples (one Medieval and one Modern, mostly Early Modern) to pinpoint the idea that dialectal complexity was of paramount importance for syntactic innovation and diffusion even far beyond the standardization processes fostered by the political reunion of the Iberian territories, the massive spread of printed books and the leap in literary elaboration brought with the Renaissance.

Bouzouita, Miriam (2016): “La posposición pronominal con futuros y condicionales en el códice escurialense I.i.6: un examen de varias hipótesis morfosintácticas”, in Johannes Kabatek (ed.), Lingüística de corpus y lingüística histórica iberorrománica, Boston / Berlin, De Gruyter, 270-298.
Fernández-Ordóñez, Inés (2009): “Los orígenes de la dialectología hispánica y Ramón Menéndez Pidal”, in Xulio Viejo (ed.), Cien años de Filoloxía asturiana: actes del congresu internacional, Oviedo, Universidad de Oviedo / Alvízoras Llibros / Trabe, 11-41.
Fernández-Ordóñez, Inés (2011b): La lengua de Castilla y la formación del español, Madrid, Real Academia Española.
Garachana Camarero, Mar (2016): “La expresión de la obligación en la Edad Media. Influencias orientales y Latinas en el empleo de ser tenudo / tenido {Ø / a / de} + infinitivo”, in Araceli López Serena, Antonio Narbona and Santiago del Rey (eds.), El español a través del tiempo: estudios ofrecidos a Rafael Cano Aguilar, vol. I, Seville, Universidad de Sevilla, 497-514.
Menéndez Pidal, Ramón (1916 [²1950]): Orígenes del español: estado lingüístico de la Península Ibérica hasta el siglo XI, Madrid, Espasa Calpe.
Octavio de Toledo y Huerta, Álvaro S. (2015): “La oculta vida dialectal de bajo + SN”, in José María García Martín (ed.), Actas del IX Congreso Internacional de Historia de la Lengua Española, vol. II, Madrid / Frankfurt a. M., Iberoamericana / Vervuert, 1841-1858.
Rodríguez Molina, Javier (2010): La gramaticalización de los tiempos compuestos en español antiguo: cinco cambios diacrónicos, unpublished PhD, Madrid, Universidad Autónoma de Madrid.

ABSTRACT. In recent years, formal semantic analyses have been applied to historical data, e.g. von Fintel 1995) and Deo (2014). In this paper, I present a formal semantics of the grammaticalization of the durative aspect marker zai from the original existential verb meaning. The most common particle to express the durative aspect in Chinese is the word zai (Li and Thompson 1981). For example:

(1) Wo huilai de shihou, Zhangsan hai zai kanshu.
I return DE moment Zhangsan still DUR read-book.
When I returned, Zhangsan was still reading.

The durative aspect marker zai proceeds the verb “kanshu” (“to read”). Interestingly, this same word zai can be used as a locative preposition. For example:

(2) Zhangsan zai tushuguan kanshu.
Zhangsan at library read-book
Zhangsan reads/read at the library.

Here the word “zai” also precedes the verb “kanshu” (“to read”), but it takes the DP “tushuguan” (“library”) as its direct complement and forms a prepositional phrase. Furthermore, this same word zai can be a verb by itself. For example:

(3) Zhangsan zai tushuguan.
Zhangsan be-at library
Zhangsan is/was at the library.

Historically, the verb “zai” gradually grammaticalized into a durative marker (Wang 1958). Thus we have the following grammaticalization path:

(4) The grammaticalization path of the durative:

Existential verb >> Locative Preposition >> Durative Aspect Marker

Starting from the original use of the word as an existential verb, we may simply use the following semantic representation for example (3).

(5) AT(‘Zhangsan’, ‘library’)

The predicate AT is the meaning of the existential verb. Now building on this predicate AT, and by using the event semantics proposed by Kratzer (1996), we may derive the semantics of the second stage of grammaticalization, i.e. the stage of a preposition in (2).

(6) ∃e. [reading(e) ∧ Agent(e, Zhangsan) ∧ AT(e, library)]

Compared to (4), here the predicate AT takes an event argument e first instead of an individual argument, e.g. “Zhangsan” in (4). Note that the meaning of (6) is: there is an event e, it is an event of reading, the agent of the event e is Zhangsan, and the event e is at the library. This represents a shift in the type of arguments allowed for the predicate AT. Note that I am ignoring the habitual reading of example (2) for now in order to simplify the analysis. Taking our analysis further, we may give the following semantics for example (1), i.e. the third stage of the grammaticalization process.

(7) ∃e. [reading(e) ∧ Agent(e, Zhangsan) ∧ AT(e, t1)]

In this case, the predicate AT shifts its second argument to a time variable t1, which can be assigned the reference time. For example, in (1) the reference time is the time when “I” returned, and the t1 will thus be assigned this time. Therefore the grammaticalization path sketched in (4) represents a shift in the two arguments of the predicate AT. The first stage is a shift of the first argument from an individual to an event, and this leads to the grammaticalization of the original existential verb into a preposition. The second stage is a further shift of the second argument from a spatial location to a temporal location t1 that can be assigned the reference time.

Therefore this formal approach to the grammaticalization of the durative aspect marker shows that grammaticalization can involve the shift of argument domains.

References:

Deo, Ashwini. 2014. Formal semantics/pragmatics and language change. In Bowern and Evans (eds.) 2014, pp393-409.

von Fintel, Kai. 1995. The formal semantics of grammaticalization. In Jill N. Beckman (ed.) Proceedings of the Annual Meeting of the North East Linguistic Society 25. Vol. 2. University of Massachusetts, Amherst: Department of Linguistics, 175-198.

Kratzer, A. 1996. Severing the external argument from its verb, in J. Rooryck and L. Zaring (eds) Phrase Structure and the Lexicon. Kluwer.

Li, Charles N. and Thompson, Sandra A. 1981. Mandarin Chinese: A Functional Reference Grammar. University of California Press.

Wang, Li. 1958. Hanyu Shigao (“A Manuscript of the History of the Chinese Language”), Kexue Chubanshe (Science Press).

ABSTRACT. Afrikaans, an extra-territorial daughter language of seventeenth century Dutch, inherited many characteristics from Dutch, one of which is the 'van'-genitive (‘of’-genitive), exemplified in [1]. However, through the process of creolization many Dutch features were significantly changed or ceased to be used, including the Dutch s-genitive. In its place, a new Afrikaans genitive particle, 'se' [2], developed from the singular male possessive pronoun zijn > sijn > s’n > se.

[1] die melkkamer van ‘n pragtige boerewoning ‘the milk.room of a beautiful farmhouse’

[2] Ketiet se voet ‘Ketiet’s foot’

Early in the twentieth century, when the standardization of Afrikaans just commenced, several different variants were still used in die written language, as could be expected. During the course of the century, the older variants disappeared from use and 'se' remained alongside the inherited 'van'-genitive.

The use of the 'se'-genitive has changed quite significantly during the past century, and in terms of both frequency and function it has made a visible impact on the still more frequent 'van'-genitive. While the 'se'- and 'van'-genitives are not interchangeable in all contexts, they are interchangeable to some extent, which means that speakers regularly make choices between using the one or the other.

In this paper, I investigate the details of the changes in the Afrikaans genitive during the past century through the analysis of corpus data. I will present findings based primarily on the analysis of corpora of written Standard Afrikaans from the decades 1911-1920, 1941-1950, 1971-1980 and 2001-2010, which are all similar in size and categories of text types. Some spoken and informal written data will also be compared to the two more recent written corpora.

The primary focus will be on the factors contributing to the choice of either 'se'- or 'van'-genitives where these two are interchangeable. Some factors influencing this choice include: the type of genitive relation expressed in the construction, the animacy of the possessor, formality, the weight of the noun phrases of the possessor and possession, and the syntactic function or position within which the genitive is used.

Finally, the Afrikaans genitive will be briefly contextualized in the West-Germanic language family, specifically regarding Dutch (from which it originated) and English (with which it has been in continuous contact).

ABSTRACT. Timberlake (1974) argued that Latvian, Lithuanian, and North Russian had modeled the use of the nominative with infinitive construction on the corresponding pattern in some West Finnic language(s). However, the hypothesis is yet open to doubt. Since Potebnja some authors have claimed that both the Baltic and Slavic constructions are of Indo-European origin: the infinitive stands for a purposive dative of the action nominal and the nominative for the former subject (Boris Larin, Jurij Stepanov), cf. Li reikia šienas (hay-nom) grėbti (rake–inf) ‘ it is necessary to rake the hay,’ R nužno voda (water-nom.) prinesti (bring-inf) ‘it is necessary to bring some water.’ According to Ambrazas (2006: 327–344) in East Baltic the oldest layer represents the underlying pattern comprised of the nominative subject together with the infinitive as a reflex of the purposive dative of the action nominal; the second layer might be a product of the influence of the West Finnic nominative object rule. I argue that the nominative with infinitive construction derives from the use of the independent nominative case which was combined with the action nominal encoded by an oblique case. The oblique nominal turned later into a purposive infinitive in the predicate which acquired a modal meaning (possibility, necessity, or obligation). The development of this construction might have progressed along two parallel pathways of grammaticalization. One pathway, discerned in the typologically innovative Indo-European languages, primarily Germanic and Romance, lead to the “passive interpretation” of the infinitive in place of the purposive nominal; hence the use, for instance, of the passive infinitive in English the swarms of locusts were to be seen (= Gr waren zu sehen) (Popov 1881, 48). In Baltic and, to the full extent, Slavic, the infinitive has acquired an “active meaning.” Consequently, the nominative subject in Baltic and Slavic tended to be grammatized as the nominative object, whence its use with other verb forms as well as the appearance of a particular “accusative variety” (feminine in -a/-ja) of the nominative object; Latvian, which retains the nominative object with the debitive only, has deviated from the “active pathway” of the grammaticalization of its infinitive. Contrary to Dunn (1982), the respective construction has a well established areal distribution. It is widely recorded in North Ukrainian as well as South and Southeast Ukrainian (Dnipropetrovs’k, Kharkiv, Nižyn, Sumy, and other Ukrainian places), that is, in particular, in such territories which are far beyond the Circum-Baltic area. Both historical and areal evidence allow me to refute possible replication of this construction, as used in (East High) Lithuanian, North Russian, Ukrainian and some Belarusian dialects, on the nominative object in West Finnic. References Ambrazas, Vytautas. 2006. Lietuvių kalbos istorinė sintaksė. Historische Syntax der litauischen Sprache. Vilnius: Lietuvių kalbos institutas. Dunn, J. A. 1982. The nominative and infinitive construction in the Slavonic languages. Slavonic and East European Review 60 (4): 500–527. Holvoet, Axel. 1993. On the nominative object in Latvian, with particular reference to the debitive. Linguistica Baltica 2: 151–161. Popov, Aleksandr Vasil´evič. 1881. Sintaksičeskija izslědovanija I. Voronež: V. I. Isaev. Timberlake, Alan. 1974. The Nominative Object in Slavic, Baltic and West Finnic. Munich: Verlag Otto Sagner.

ABSTRACT. Expletive subjects in Icelandic: A diachronic study

This paper presents results from an ongoing corpus-based investigation of the diachronic evolution of expletive subjects in Icelandic. The innovation of overt expletive subjects is a historical development common to Germanic, and modern Icelandic exhibits a particularly diverse range of constructions with the expletive subject það: (1) a) Það var samt ljóst að við þyrftum meiri mannskap. (extraposition) EXPL was still clear COMP we needed more crew ‘It was still clear that we needed more crew.’ (IcePaHC, 2008.OFSI.NAR-SAG,.733)

b) Það snjóar á Íslandi á veturna. (weather predicate) EXPL snows on Iceland on winters.DEF ‘It snows in Iceland in the winters.’ (MIM, VISINDAVEFUR-C1L264)

c) Það þarf ekki nema eitt handtak... (modal impersonal) EXPL needs NEG except one grasp ‘It only needs one grasp.’ (IcePaHC, 2008.OFSI.NAR-SAG,.1162)

d) Það eru margir jafnaðarmenn í Hafnarfirði. (existential) EXPL are many socialists in Hafnarfjörður ‘There are many socialists in Hafnarfjörður.’ (MIM, BAEKUR-B0O) e) Það komu nokkrir vopnaðir menn af næstu bæjum. (unaccusative) EXPL came some armed men from next town ‘There came some armed men from the next town.’ (IcePaHC, 2008.OFSI.NAR-SAG,.634)

While a good deal of research has been conducted on the historical development of expletive subjects in e.g. Swedish (Falk 1993), English (Ball 1991; Butler 1980; Ingham 2001; Stowell 1978) and German (Abraham 1993), the Icelandic situation remains only scarcely investigated: notable contributions are Rögnvaldsson (2002), Eythórsson & Sigurðardóttir (2016) and Faarlund (1990) on Old Norse. The previous studies portray overt expletive subjects as a relatively recent development, and það is indeed absent in constructions such as the following in Old Icelandic (1150-1350), where it would be expected in modern Icelandic (cf. 1b, 1d above): (2) a) _______ Líður nú af veturinn og _______ vorar. (weather predicate) ØEXPL passes now of winter.DEF and ØEXPL becomes-spring ‘The winter now passes and (it) becomes spring.’ (IcePaHC, 1350.BANDAMENNM.NAR-SAG,.147)

b) _______ Var fátt manna heima. (existential) ØEXPL was few men.GEN at-home ‘(There) were few men at home.’ (IcePaHC, 1350.FINNBOGI.NAR-SAG,655.1696)

However, I show that expletive það was not completely absent in Old Icelandic, as the following attested constructions demonstrate: (3) a) Það er ráð húsfreyja að taka vel við gestum. (extraposition) EXPL is advice housewives.GEN to take well with guests ‘It is the advice of housewives to receive guests well.’ (IcePaHC, 1310.GRETTIR.NAR-SAG,.658)

b) Það var einn morgun snemma að Grettir kom til hrossahúss. (it-cleft) EXPL was one morning early COMP Grettir came to stable ‘It was early one morning that Grettir came to a stable.’ (IcePaHC, 1310.GRETTIR.NAR-SAG,.89)

c) …en það þarf ekki við mig að tala. (modal impersonal) but EXPL needs NEG with me to talk ‘…but it is not necessary to talk with me.’ (IcePaHC, 1350.BANDAMENNM.NAR-SAG,.687)

In this paper, I discuss data for overt expletive subjects in IcePaHC over 5 time periods, and compare them to ‘null’ expletive subjects (i.e. all contexts in which modern Icelandic would expectedly use an overt expletive but in which það is absent): Table 1: Proportion of overt expletive subjects (versus null) over time Time period All construction types Extraposition only All construction types excluding extraposition Main clauses Embedded clauses Main clauses Embedded clauses Main clauses Embedded clauses 1150-1350 49% 47% 94% 100% 16% 29% 1351-1550 40% 41% 87% 100% 18% 22% 1551-1750 51% 48% 85% 89% 25% 29% 1751-1900 60% 58% 97% 94% 42% 39% 1901-2008 88% 72% 93% 75% 85% 34%

This data allows the following conclusions to be drawn: 1. The proportion of overt expletive subjects in the extraposition construction type (1a and 3a above) is already very high in Old Icelandic (94% in main clauses; 100% in embedded clauses) and remains stable over time. 2. By contrast, removing the relatively stable extraposition type reveals that the other construction types show a much more dramatic increase, particularly in main clauses where the proportion of overt expletive subjects rises from just 16% in Old Icelandic to 85% for 1901-2008. 3. Moreover, the fact that the proportion of overt expletives in constructions excluding the extraposition type is 16% (main clauses) and 29% (embedded clauses) challenges the standard view that Old Icelandic had no overt expletive aside from in extraposition constructions (Faarlund 1990: 66; Rögnvaldsson 2002). 4. Across all construction types in almost all periods, the preference for overt expletive subjects appears to be less strong in embedded clauses than in main clauses. There is a particularly large discrepancy in the non-extraposition construction types, where the increasing proportion of overt expletives in embedded clauses lags significantly behind that for main clauses. 5. Even by the latest period (1901-2008), overt expletive subjects are still not fully obligatory, in either extraposition constructions (93%; 75%) or other construction types (85%; 34%). In sum, this corpus-based investigation sheds new light on the historical development of expletive subjects in Icelandic and challenges a number of standard assumptions stated in previous literature on the topic.

References Abraham, Werner. 1993. Null subjects in the history of German: From IP to CP. Lingua 89(2-3). 117–142. Ball, Catherine N. 1991. The historical development of the it-cleft. PhD Dissertation, University of Pennsylvania. Butler, Milton Chadwick. 1980. Grammatically motivated subjects in early English. University of Texas at Austin, Department of Linguistics. Eythórsson, Thórhallur & Sigríður Sæunn Sigurðardóttir. 2016. A brief history of Icelandic weather verbs: Syntax, semantics and argument structure. Working Papers in Scandinavian Syntax 96. 91–125. Faarlund, Jan Terje. 1990. Syntactic Change : Toward a theory of historical syntax. Berlin ; New York: de Gruyter. Falk, Cecilia. 1993. Non-referential subjects in the history of Swedish. Department of Scandinavian Languages, University of Lund. Ingham, Richard. 2001. The structure and function of expletive there in pre-modern English. Reading Working Papers in Linguistics 5. 231–249. Rögnvaldsson, Eiríkur. 2002. ÞAÐ í fornu máli - og síðar. Íslenskt mál 24. 7–30. Stowell, Tim. 1978. What was there before there was there. Papers from the Fourteenth Regional Meeting of the Chicago Linguistics Society, 458–471.

ABSTRACT. The topic of this paper is clause-final negative particles (FNP) in Swedish dialects. While the common Swedish negation inte may be doubled in a final annex, in standard Swedish as well as in the dialects, many dialects also allow a particle in the shape of e, i or ai after negated clauses. FNP occur in a coherent area around the Baltic Sea, including eastern central Sweden, Åland, Gotland, Åboland, southern Ostrobothnia and western Nyland. FNP differ from doubled negations in some syntactic aspects, occuring after both negations and other negated expressions (such as aldrig ’never’). FNP are also possible in questions and exclamations, contexts that disallow doubling negations. These particles may have developed from the old Swedish negation ej or from the modern negator inte. An argument for the former alternative is that other dialectal phenomena that spread during the late Middle Ages have approximately the same geographic distribution

ABSTRACT. In this paper I will examine and analyze four different word order changes from Old Norwegian (ON – 12th-15th c.), which are no longer found in Modern Norwegian (MN). I will show that one of these changes with necessity entailed the other three. The triggering factor was the loss of OV as a possible word order pattern besides the unmarked VO order. Synchronically in ON this order is derived through a version of scrambling, in this context understood as the left-adjunction of a copy of the complement of the verb to vP. Thus the pattern in (1a) is derived as in (1b).

(1) a. Hælldr mindi hann þessa gripina kiosa rather would he these treasures choose ‘He would rather have those valuable things’ b. hælldr mindi hann þessa gripina vP[kiosa þessa gripina]

The other three changes following the loss of OV as a possibility are: object shift with non-finite verb; preverbal preposition; and topicalization of heads.

Object shift is the tem used about movement of an unstressed object pronoun across the negation or a sentence adverbial, which are assumed to be left-adjoined to vP. The condition on this movement is that the object and the negation/adverbial are phonologically adjacent (‘Holmberg’s Generalization’, Holmberg 1986, 1999). In a VO language, this means that there must be no non-finite verb in VP, consider the MN sentences in (2). The ON sentence (3a), however, can still be considered an instance of object shift, given a (derivationally) previous scrambling to OV. The derivation of (3a) is given in (3b).

(2) a. Vi såg henne ikkje b. *Vi har henne ikkje sett we saw her not we have her not seen (3) a. þurfu þæir þat ekki sægia need they that not say ‘They do not need to say it’ b. þurfu þæir vP[Negekki vP[sægia þat]] Scrambling: þurfu þæir vP[Negekki vP[þat vP [sægia þat]]] OS: þurfu þæir vP[þat vP[Negekki vP[þat vP [sægia þat]]]]

Scrambling creates phonological adjacancy of the negation and the object, which then feeds object shift.

Preverbal preposition. In ON, but no longer in MN, a P may be found in a position preceding the main verb of the sentence, with the complement of P at the end of the sentence.

(4) ok þat hefir mik til rekit svá langrar ferðar and that has me to driven so long journey.G | ‘And that has driven me to such a long journey’

Starting from an OV order, the pattern in (4) is derived by means of extraposition of the complement. (5) Scrambling: þat hefir [mik] [til svá langrar ferðar] VP[rekit [mik] [til svá langrar ferðar]] Extraposition: þat hefir mik til [svá langrar ferðar] VP[rekit] [svá langrar ferðar]

Extraposition of long, complex or focussed elements is very common in ON, independently of the OV/VO pattern.

Topicalized head. In the modern Germanic V2 languages, [SpecCP] is the landing site of maximal projections. This principle seems to be violated in ON, where a non-finite V or a P may be found in the topic position, leaving their complement behind. (6) is an example of a topicalized V.

(6) En binda ma maðr þiof but bind may man thief ‘But a man may tie up a thief’

Rather than being an instance of an undesirable head movement to a Spec position, this is actually remannat movement of the remaining V after scrambling of the complement out of VP.

References Holmberg, A. 1986. Word order and syntactic features in the Scandinavian languages and English. Ph.D. Dissertation, University of Stockholm. Holmberg, A. 1999. Remarks on Holmberg’s Generalization. Studia Linguistica 53. 1-39.

ABSTRACT. The process of anticausativization, as described in the typological literature, consists of three properties: i) valency reduction (transitivity decrease), normally amounting to the deletion of the Agent from the inventory of participants, and ii) addition of an anticausative morpheme to the verbal stem (see e.g. Haspelmath 1987; Paducheva 2003; Alexiadou et al. 2006). However, there is another syntactic phenomenon, closely related from a functional point of view to canonical anticausativization and found in some Indo-European languages, that has received much less attention from typologists and is rarely observed by scholars of Indo-European linguistics (though see, for instance, Siewierska & Malchukov 2011). These structures are dependent-marking instead of head-marking and have been labeled “oblique anticausatives” in recent work by Barðdal (2015). An example from Old Icelandic illustrates this phenomenon: (1a) þeir reiddu they.NOM carry:PAST:3PL ‘They carried him ...’ (1b) örkina reiddi hann ... (Ljósv. 23126) him.ACC um haf (Pr. 704) ark.the.ACC carry:PAST:3SG of ocean ‘The ark got carried over the ocean’ The example in (1b) shows an intransitive variant of the otherwise transitive verb reiða concomitant with a difference in meaning, i.e. ‘get carried’ vs. the transitive ‘carry’. Another property of this alternation is that the object of the transitive verb, marked in the accusative, corresponds to the accusative subject of the intransitive variant. A similar alternation has been referred to as ‘(P-)labile’ or ‘ambitransitive’ in the literature (see Letuchiy 2009, among others), consisting of causative/anticausative pairs like break–break and boil–boil, where no extra morphological (anticausative) marking is found on the verb in the intransitive alternant. However, there is one major difference between labile/ambitransitive pairs of the canonical causative/anticausative type and labile verbs with oblique anticausatives. Specifically, the subject of the intransitive alternant is in the nominative case in the labile/ambitransitive alternation in languages which have case marking, while with oblique anticausatives the subject is in the same case as the object of its corresponding transitive alternant, be it accusative, dative or genitive. This difference in subject case marking across the two types of alternations shows that these really are two different alternations and that oblique anticausatives should indeed be regarded as a specific sub-type of anticausatives, in that the anticausative marker can be viewed as being found on the dependent (the argument) instead of the verb (the head). As far as we are aware, oblique anticausatives of the type shown in (1) above are particularly common in Indo-European languages, while labile/ambitransitive alternations of the canonical causative/anticausative type are found more widely. A further peculiarity of the oblique anticausative alternation is that it also targets ditransitive verbs in the Indo-European languages, while head-marking anticausatives are always discussed in the literature as only targeting transitive verbs, cf.: (2a) fysta ek þik fararinnar. (Egils saga, Ch 49, cited from Ottósson 2013) spurred I.NOM you.ACC trip.GEN ‘I urged you on the trip’ (2b) fýsti Ólaf þess að ... (Ólafs saga Tryggvasonar, Ch. 31, cited from Ottósson 2013) yearn.for Ólafur.ACC it.GEN to ‘Ólafur yearned (for it) to ...’ This paper investigates predicates in Latin and Ancient Greek that occur in the oblique anticausative construction with a Acc-Gen or Dat-Gen argument structure, to show whether or not such verbs also occur as ditransitives (as can be found in Old Icelandic) or if it would be semantically plausible at all to assume a development from ditransitives in these languages. If it is the case that Latin/Ancient Greek oblique anticausative constructions utilize otherwise ditransitive verbs, then an isogloss among Germanic and Latin and/or Greek can be established for oblique anticausativization. However, this does not necessarily imply a shared innovation, so the constructions in Germanic, Latin, and Ancient Greek are compared to constructions like the following in Slavic, where multiple arguments occur in the oblique anticausative structure. (3a) Modern Russian Lodku unes-l-o boat.ACC carry.away-PAST-3SG.N ‘The boat drifted down the stream.’ (3b) Modern Lithuanian Sodą prinešė sniego. garden.ACC brought snow.GEN ‘The garden got filled with snow.’ vniz po tečeniju. down on stream The results of this investigation thus answer the question of whether oblique anticausatives are a shared feature of Germanic and Romance and/or Greek, for example, as a morphosyntactic isogloss as a result of a common innovation among the two branches, or whether this construction should rather be reconstructed for Proto-Indo-European. References: Alexiadou, Artemis, Elena Anagnostopoulou & Florian Schäfer. 2006. The properties of anticausatives crosslinguistically. Phases of interpretation, ed. by M. Frascarelli, 187-211. Berlin: Mouton. Barðdal, Jóhanna. 2015. Syntax and Syntactic Reconstruction. The Routledge Handbook of Historical Linguistics, ed. by Claire Bowern & Bethwyn Evans, 343–373. London: Routledge. Cennamo, Michela, Thórhallur Eythórsson & Jóhanna Barðdal. 2015. Semantic and (Morpho)syntactic Constraints on Anticausativization: Evidence from Latin and Old Norse- Icelandic. Linguistics 53(4): 677–729. Haspelmath, Martin. 1987. Transitivity alternations of the anticausative type. Köln: Institut für Sprachwissenschaft. Letuchiy, Alexander. 2009. Towards a typology of labile verbs: Lability vs. derivation. New challenges in typology, ed. by A.V. Arkhipov & P. Epps, 247-268. Berlin: de Gruyter. Ottósson, Kjartan G. 2013. The Anticausative and Related Categories in the Old Germanic Languages. Diachronic and Typological Perspectives on Verbs, ed. by Folke Josephson & Ingmar Söhrman, 329–382. Amsterdam: John Benjamins. Paducheva, Elena V. 2003. Is There an "ANTICAUSATIVE" Component in the Semantics of Decausatives? Journal of Slavic Linguistics 11.1: 173-198. Siewierska, Anna & Andrej Malchukov, eds, 2011. Impersonal constructions. A cross-linguistic perspective (Studies in language companion series; 124). Amsterdam: Benjamins.

ABSTRACT. As established by Greenberg (1966) in his universal 38, the usual order of morphemes inside a word is STEM-DERIVATION-INFLECTION or its mirror order. This would be linked to word order in the language, according to Hawkins – Gilligan (1988): OV/Post → STEM-SUFFIX and PREFIX-STEM → VO/Prep. Nevertheless, languages show a preference for suffixation and, in fact, inflection is almost exclusively expressed through suffixes, regardless if languages have a VO or an OV word order (cf. Greenberg 1957, Hawkins – Cutler 1988). Various principles may help to explain the order of the suffixes within a word. According to Baker’s (1985) “mirror principle”, it would reflect the order of syntactic derivation, Rice (1991) underlined the relative scope (semantic and grammatical) of morphemes and Bybee (1985: 4) insisted on the concept of relevance: “The semantic relevance of an affix to a stem is the extent to which the meaning of the affix directly affects the meaning of the stem.” Morphemes would be ordered according to their relevance: the more relevant inside and the less relevant outside. These general tendencies in the order of affixes have a bearing on diachrony. As emphasized by Mithun (2000: 231-232), the order of morphemes is not the result of a spontaneous decision when a speech act occurs. Morphological structures emerge over time and they develop gradually, or more precisely, category by category. Mithun, based the evidence provided by languages of various families, pointed out how the order of affixes usually reflects the actual historical sequence of grammaticalization of the morphemes (cf. Givón 1971). Morphological structures, however, are not inmune to change and, therefore, the order of morphemes can be altered in the history of a language due to various kinds of processes. Mithun (2000) mentioned certain changes in which a derivational morpheme had been reanalyzed as an inflectional morpheme and this had had the consequence that this morpheme had changed its place within the word in the long run. This type of change can occur in languages both with layered and template structures in morphology. The expected order of morphemes can be restored in a language in different ways: loss of morphemes that have been “trapped” between two words in grammaticalization processes; reanalysis of two morphemes as an only morpheme (thus avoiding an unsual order); or analogical extension of a morpheme to a place inside the word that was not its original one (Harris 2005: 140-142). Therefore, if as the result of a given change (e.g. a former clitic that is reinterpreted as part of the word), inflection is “trapped” word-internally (cf. Harris – Faarlund 2006), the expected order of morphemes may be reintroduced by the externalization of inflection (cf. Haspelmath 1993). There are good examples of this type of processes of externalization of inflection and reordering of morphemes in the pronominal inflection of a number of ancient Indo-European languages, as typified by the Latin archaic demonstrative Acc. Sg. masc. eumpse and fem. eampse (Acc. -m plus enclitic –(p)se) vs. Classical Latin ipsum and ipsam. Intermediate forms can emerge, which keep the ancient inflection and at the same time display the externalized one, as in Gk. Dat. Pl. (Homer and Ionian) toîsde(s)si/toísde(s)si and Gen. Pl. (Lesbian) tōndéōn when compared to Attic toîsde and tônde. Beyond those instances, we will show in this paper how this kind of approach can contribute to a better understanding of the role of certan morphemes in the pronominal inflection of Proto-Indo-European and some ancient Indo-European languages. A suffix *-sm- without any apparent semantic content appears in the oblique cases of certain demonstrative pronouns (cf. Mendoza 1998), as in Skt. masc.-neut. Sg. Dat. tásmai, Abl. tásmād, Loc. tásmin; Avest. masc.-neut. Sg. Dat. ǣtahmai, Abl. ǣtahmat, Loc. Ǣtahmi; Goth. Sg. Dat. Þamma; OPrus. Sg. Dat. stesmu; Celtiberian Sg. Dat. somui, Loc. somei; or Umbr. Sg. Dat. pusme and esme. The suffix *-sm- also appears in some cases of the 1st and 2nd plural personal pronouns, as in Skt. 1st Pl. Acc. asmá̄n, Instr. asmá̄bhis, etc., 2nd Pl. Acc. yuṣmá̄n, Dat. yuṣmábhyam, etc. or Gk. 1st Pl. Nom. hēmeîs, Acc. hēmâs, etc., 2nd Pl. Nom. hūmeîs, Acc. hūmâs, etc. Similarly, there is a suffix -sy- in certain femenine forms of the pronouns. After revising the evidence for that kind of suffixes in the Indo-European pronominal inflection, we attempt an explanation in the light of the processes that we have mentioned above. An emphatic particle smā occurs in the Ṛgveda and we can envisage a connection to the the *-sm- found in the pronouns. This is an important clue to understanding how those “empty” suffixes came to be integrated in the morphology of various categories of Indo-European pronouns.

REFERENCES

Baker, M. 1985: “The mirror principle and morphosyntactic explanation”, Linguistic Inquiry 16: 373-416. Bybee, J. 1985: Morphology, Amsterdam – Philadelphia, John Benjamins. Cutler, A. – J. A. Hawkins – G. Gilligan 1985: “The suffix preference: a processing explanation”, Linguistics 23: 723-758. Givón, T. 1971: “Historical syntax and synchronic morphology: an archaeologist's field trip”, Proceedings of the Chicago Linguistic Society: 394-415. Greenberg, J. H. 1957: “Order of affixing: a study in general linguistics”, in Essays in Linguistics, Chicago, The University of Chicago Press, 86-94. ----- 1966: Language Universals, with special reference to feature hierarchies, The Hague, Mouton. Harris, A. C. 2005: “Establishing and mantaining morpheme order”, in W. Østreng (ed.), Convergence, Oslo, Center for Advanced Studies, 139-142 (=http://www.cas.uio.no/Publications/Seminar/Convergence_Harris.pdf). Harris, A. C. – J. T. Faarlund 2006: “Trapped morphology”, Journal of Linguistics 42: 289-315. Haspelmath, M. 1993: “The diachronic externalization of inflection”, Linguistics 31: 279-309. Hawkins, J. A. – A. Cutler 1988: “Psycholinguistic factors in morphological asymmetry”, in J. A. Hawkins (ed.), Explaining Language Universals, Oxford, Blackwell, 280-317. Hawkins, J. A. – G. Gilligan 1988: “Prefixing and suffixing universals in relation to basic word order”, Lingua 74: 219-259. Mendoza, J. 1998: “Morfología de pronombres, adverbios, partículas y numerales”, in F. R. Adrados – A. Bernabé – J. Mendoza, Manual de Lingüística Indoeuropea, vol. 3, Madrid, Ed. Clásicas, 1-139. Mithun, M. 2000: “The reordering of morphemes”, in S. Gildea (ed.), Reconstructing Grammar. Comparative Linguistics and Grammaticalization, Amsterdam - Philadelphia, John Benjamins, 231-255. Rice, K. 1991: “Predicting the order of the disjunct morphemes in the Athapaskan languages”, Toronto Working Papers in Linguistics 11: 99-121.

ABSTRACT. In the present paper, we addressed the function and the distribution of the so-called “accusative of respect” in ancient Greek. To do this, we examined the occurrences of the constructions including an accusative of respect in literary texts, from Homer to the fifth century B.C.E. Among the various functions that the accusative case has in ancient Greek (Luraghi 2003: 52 ff.; Crespo 1988; Chantraine 1942; Schwyzer & Debrunner, 1950; La Roche 1891, among others), the role of the accusative of respect appears to be quite vague: «Il est difficile de définir l’accusatif de relation. Révélatrice à cet égard est la définition de Marouzeau [1969]: “on appelle accusatif et instrumental de relation les cas dont l’emploi répond à l’idée de quant à, pour ce qui est de”» (Jaquinod, 2006: 42). Previous studies almost exclusively focused on the properties of the entity denoted by the accusative of respect (e.g., body parts, qualities, etc.) and – to a lesser extent – on the relationship between this entity and the most animate entity in the construction (e.g., the inalienable possession relationship: Heine, 1997; Aikhenvald & Dixon, 2013): Jacquinod, 2006 and 1989; Chantraine 1942; Schwyzer & Debrunner, 1950; La Roche 1861, among others. However, how the semantic properties of the predicate affect the distribution of the accusative of respect has not yet fully clarified, and the mutual relationship between the predicate and the two main arguments of the construction remains unsettled. In this study, we provide an answer to critical aspects of these questions, by showing how different factors, that involve both the nominal and the verbal domain, co-occur in the accusative of respect and account for its function and distribution in Homeric poems and later texts. In particular, we assess the distinctive role of the following dimensions and the interaction between them: animacy hierarchy (Silverstein, 1976), attention flow (DeLancey, 1981), situation type (Kemmer, 1993), semantic role of the subject (Van Valin & LaPolla, 1997), verb actionality (Vendler, 1967; Bertinetto, 1986; Sorace, 2000).

REFERENCES Aikhenvald, A. & Dixon, R. 2013, Possession and ownership, Oxford: Oxford University Press. Bertinetto, P. M. 1986, Tempo, aspetto e azione nel verbo italiano, Firenze: Accademia della Crusca. Chantraine, P. 1942, Grammaire Homérique, I, Paris: Klincksieck. Crespo, E. 1988, The semantic and syntactic functions of the accusative. In: A. Rijksbaron et al. (eds.), In the Footsteps of Raphael Kühner, Amsterdam: Gieben, pp. 99-121 DeLancey, S. 1981, An interpretation of split ergativity and related patterns, Language 57: 626-658. Heine, B. 1997, Possession, Cambridge: Cambridge University Press. Jacquinod, B. 1989, Le double accusatif en grec d’Homère à la fin du V siècle avant J.C., Louvain: Peeters. Jacquinod, B. 2006, Le domaine de l’accusatif de relation. In, Crespo, E. et al. (eds.), Word classes and related topics in Ancient Greek, Louvain: Peeters, pp. 59-68. Kemmer, S. 1993, The middle voice, Amsterdam/Philadelphia: Benjamins. La Roche, J. 1861, Homerische Studien. Der Akkusativ in Homer, Wien: Gerold. Luraghi, S. 2003, On the meaning of prepositions and cases. The expression of semantic roles in Ancient Greek, Amsterdam/Philadelphia: Benjamins. Marouzeau, J. 1969, Lexique de la terminologie linguistique, 3rd ed., Paris: Geuthner. Schwyzer, E. & Debrunner, A. 1950, Griechische Grammatik, 3 Voll: München, Beck. Silverstein, M. 1976, Hierarchy of features and ergativity, In: Dixon, R.M.W. (ed.), Grammatical categories in Australian languages, Canberra (Australian Institute of Aboriginal Studies), pp. 112-171. Sorace, A. 2000, Gradients in auxiliary selection with intransitive verbs, Language 76: 859-890. Van Valin, R. D. Jr. & La Polla, R. J. 1997, Syntax: structure, meaning & function, Cambridge: Cambridge University Press. Vendler, Z. 1967, Linguistics in Philosophy. Ithaca (NY): Cornell University Press.

ABSTRACT. When working on questions of language change and language variation of the Early Modern Age (for example in Early Modern High German), there is always the challenge of finding an appropriate database. An often neglected possibility to reconstruct not only historical everyday dialogues but also to find metalinguistic reflections about questions like standard, norm and variation in textbooks for teaching modern foreign languages. Our corpus consists of textbooks of different authors (such as Georg von Nürnberg, Noel de Berlement, Nathanael Duez) of the 16th and 17th century containing German but also up to eight other languages like French, Italian and English. They offer rich material on all levels of language (eg. grammar, lexis, pragmatics) from dialogues for the reconstruction of pragmalinguistic and sociolinguistic phenomena. Another advantage of these multilingual textbooks is the possibility to compare similar phenomena of language changes, which only differ in subtle aspects within the different languages. In addition to comparative linguistic analysis of language change the corpus will also allow studies on one particular language. Despite the aforementioned benefits historical multilingual textbooks have not been used to their full potential in research on language change. These often similar dialogues published in different editions allow to observe the changing language and furthermore the changing socio-communicative norms. Due to the transmission and revision of these textbooks across sometimes over hundred years, it is possible to get a new perspective on language change at the level of intertextual grammatical (such as particles) and pragmatic progression (such as pronouns of address, politeness). Since in this period the standardization of the German language was not completed, there are also metalinguistic reflections to claim validity in the epilogues. With the help of this multilingual textbook-corpus it is possible to reconstruct parts of the linguistic knowledge of this period and to provide rich material for analysing change in language.

Bibliography:

Burke, Peter (2004): Languages and Communities in Early Modern Europe. Cambridge. Glück, Helmut (2002): Deutsch als Fremdsprache in Europa vom Mittelalter bis zur Barockzeit. Berlin/New York. Glück, Helmut (2013): Die Fremdsprache Deutsch im Zeitalter der Aufklärung, der Klassik und der Romantik. Grundzüge der deutschen Sprachgeschichte in Europa. Wiesbaden. Jucker, Andreas H. (1995): Historical Pragmatics. Pragmatic Developments in the History of English. Amsterdam/Philadelphia. Langer, Nils/Winifred Davies (2006): The Making of Bad German. Lay linguistic stigmatisations in German, past and present. Frankfurt am Main. Langer, Nils/Winifred Davies (2005): Linguistic Purism in the Germanic Languages. Berlin, New York. Simon, Horst (2006): Reconstructing historical orality in German – what sources can we use? In: Irma Taavitsainen, Juhani Härmä & Jarmo Korhonen (Hgg.). Dialogic language use – Dimensions du dialogisme – Dialogischer Sprachgebrauch. Helsinki. Taavitsainen, Irma/ Jucker, Andreas H. (2010): Trends and developments in historical pragmatics. In: Jucker, Andreas H./Taavitsainen, Irma (ed.): Historical Pragmatics. Berlin/New York.

ABSTRACT. The form one appears in a remarkably wide range of functions (i.e. part of speech classes, contextual positions) and uses (meanings) in English; it appears to lend itself exceptionally strongly to the process of re-allocation, i.e. an existing form adopting new functions. This relates to both the history of English and to innovative uses in several varieties, including New Englishes. One started out as a numeral in Old English; it adopted new functions as indefinite pronoun (one shouldn't behave like this) and a substitution form for noun phrase heads (not a red apple but a green one); and can be found in a wide range of uses in today's varieties (e.g. as a clause-final relativizer in Singaporean English, e.g. the cake John always buy one 'which John always buys', or an indefinite article in Cyprus English: she is one art teacher). The present paper is part of a larger project which investigates the evolution of these uses through space and time and asks for causes which explain these evolutionary trajectories. We adopt a Complex Systems framework (Ellis & Larsen-Freeman 2009; Kretzschmar 2015; AUTHOR1 2015) to account for long-term functional changes in a larger framework. The Complex Systems framework and some of its pertinent principles (systemness and complexity, perpetual dynamism, interaction of order and chaos, emergentism and auto-organisation) are briefly introduced in the first part of this paper. We investigate the emergence and diffusion of innovative uses of one (mainly indefinite pronoun uses and pro-form uses in noun phrases) from the Early Modern English period through modern Standard British and American English to Postcolonial Englishes. We point out developmental trajectories and patterns of continuity and discontinuity by analyzing occurrences of one and ones in four different corpora (or groups of corpora): ARCHER, "A Representative Corpus of Historical English Registers", covering the period from 1600 to 1999 in British and American texts, subdivided into 50-year sub-periods; the "Brown family" of corpora representing American and British English from 1961 and 1991, respectively; two ICE (International Corpus of English) corpora representing Indian and Singaporean English; and the CEDAR corpus (Cyprus English Data Analysis and Research). Occurrences of one(s) in these sources have been tagged according to word class (numeral, determiner, adjective, pronoun – some divided into sub-classes), syntactic function, construction context (i.e. constituent sequences of the NPs containing one), and semantic class of the referent of one. We observe some remarkable frequency shifts and some innovative tendencies across time and between varieties, and interpret these as manifestations of principles of Complex Systems in their trajectories of evolution.

References Ellis, Nick C., and Diane Larsen-Freeman, eds. 2009. Language as a Complex Adaptive System. Malden, MA: Wiley & Sons. Kretzschmar, William R., Jr. 2015. Language and Complex Systems. Cambridge: Cambridge University Press. Rissanen, Matti. 1967. The uses of "one" in Old and early Middle English. Helsinki: Sociéte Neophilologique. AUTHOR1. 2015. "LECTURE TITLE". Plenary given to the conference "CONFERENCE TITLE", University of XX.

ABSTRACT. Annotating Presuppositional Information in Historical Corpora

The present work constitutes a pilot part to a larger project which aims to: A) capture the

informative contribution made by presuppositions (PSPs) and their triggers in historical corpora

of different proveniences; B) use this contribution to explain the evolution of PSPs over time.

Having historical data annotated with PSPs, syntactic subordination becomes one of the

interesting by-products which can be studied with respect to the semantics-pragmatics interface.

It becomes relevant in connection with the projection problem of PSPs across well-established

classes of operators (cf. Schwarz 2016 for an overview). The focus in what follows is

methodological and we offer a punctual flavor of how we proceeded in the annotation process.

Brief theoretical background: PSPs are standardly admittance conditions on contexts, preshared

information by sender/recipient (Stalnaker 1973, Beaver & Geurts 2011, Schwarz 2016

a.m.o). They have well-testable linguistic properties and are taken to be stored in the common

ground (CG, set of incrementally updated propositions). Diachronic work on PSPs is scarce

(e.g. Beck & Gergel 2015 make specific suggestions for again; we capitalize on that strand of

research theoretically; the focus will be on other items and a variety of empirical issues here).

Annotation work: We are evaluating corpus texts of English and German precisely based on

pragmatic information (this means: detailed context-based semantic corpus studies) for the

satisfaction of PSPs. The practical goal is to answer, on a broad scale, the following apparent

simple question, for thousands of tokens: ‘Have PSPs been added to the CG via [1] explicit

textual evidence in preceding context; [2] entailment relationships; [3] scripts (worldknowledge)?’

Else: accommodation must have taken place (on a standard account). In the

longer term, we will develop further metrics as well, but we leave them outside of present scope.

We next illustrate some first results on the PSPs of the German word noch (=‘still’, noh in Old

High German (OHG)) and cleft constructions, respectively. We chose these two PSP triggers

for illustration, not only because their proveniences are necessarily distinct, but the types and

ranges of PSPs they encode show a range of possibilities. While noch is ambiguous between

multiple types of presuppositions (considerably more so than e.g. its ModE counterpart still; cf.

Beck 2016 and references for a synchronic semantic account), clefts typically encode two

presuppositions at the same time, namely existence and exhaustively (e.g. Büring & Kriz 2013).

The major types of presuppositions triggered by noch are temporal, additive, degree (in

combination with comparison constructions) and, in later periods of German, marginal/path.

Consider the following Old High German example:

(1) Tho quad imo ther iungo: alliu thisu gihielt ih fon minera iugundi: uuaz ist

Then said him the youth: All this kept I since my youth, what is

mir noh nu uuan?

me yet/still now missing?

‚Then said the youth, „All these rules I have followed since young age, what else am I

missing now/what am I still missing now?“’

(OHG Tatian, ca. 830 AD1.OHG1.TatianEvHarm.106.3)

(2) a. PSPtem: From the relevant past until reference time, I am overlooking something.

b. PSPadd: There is something else in the context that I am overlooking.

Both (2-a) and (2-b) are licensed by context. A numerically important fact is that an additional

type of noch is available in OHG data, namely a conjunctive one (‘neither… nor…’). For

instance, during the first period of Old High German, OHG1 (until 850 AD), the majority (65%)

of tokens coordinate various kinds of negated constituents. The second largest group in OHG,

was given by temporal uses of noch (20%). The temporal reading is the most prevalent in a

randomized selection of Modern High German data, where the overt satisfaction of the PSPs

involved is at over 40% in our current results.

The usage of clefts as a focus construction has been on the rise since the development

of SVO-order in the English language after the loss of a multifunctional first position (cf.

Jespersen 1937, Los & Komen 2012). However, it-clefts in English offer themselves as a good

candidate for further research in the area of pragmatics, not only because of their increased use,

but also because of their status as a presupposition-triggering construction. A pragmatic feature

of clefts is that they almost invariably project two presuppositions, which are labeled as PSP 1

and PSP 2, illustrated in example (3). The first presupposition is an existential and the second

one is exhausitivity (cf. Büring & Kriz 2013, Hartmann 2016 for further discussion).

(3) And we have the last thing in emptiness and new art. "But you are quite right.” It was

that spirit, only more drawn out, which made progress. We do not do it here, simply

because we can afford so little now, and it seems to us that there are things so much

more important than furniture to spend our few pengo on. "No, it is you who are right,

Margit.” (BLOB, BE 1931: bl_P10_1)

PSP 1: Somebody is right.

PSP 1 the utterance made in the preceding context of “you are right” from the male in

story.

PSP 2: Only you are right.

PSP 2 is accommodated

The initial Modern data gathered on it-clefts was taken from the Brown Family Corpus (1930s,

1960 and 1990s segments). One hundred tokens were selected and sorted based on the

aforementioned categories. The early findings indicate that one-fifth of existential PSPs were

satisfied through either explicit textual evidence in preceding context or entailment

relationships. The data for the classification of exhaustive PSPs, however, shows the majority

of them were not overtly satisfied in the context.

While we capitalize on syntactic corpora in our immediate syntactic work (i.e. the Penn-

Helsinki series of corpora for English), the semantic work and annotation schema is not

restricted to them (cf. above). For clefts, for instance, the complete data set will include the

satisfaction behavior of PSPs across several corpora (cf. Spenader 2002 for an interesting study,

although for less data and without a diachronic dimension). Working diachronically, the

ultimate goal of our larger project will be to provide hands-on information on multiple relevant

presupposition markers, the proportions of their particular readings, and specifically help to lay

a foundation to enable an understanding of their properties at the structure-meaning interface

over time.

Selected references:

Beck, S. 2016. Temporal noch/still and further-to readings of German noch. Proceedings of

     Sinn und Bedeutung 20.

Beck, S., & R. Gergel. 2015. The diachronic semantics of English again. Natural Language

     Semantics 23: 157-203.

Büring, D., & Kriz, M. (2013). It’s that, and that’s it! Exhaustivity and homogeneity

     presuppositions in clefts (and definites). Semantics and Pragmatics 6: 1-29.

Gergel, R., Blümel, A., & Kopf, M. 2016. Another heavy road of decompositionality: Notes

     from a dying adverb. University of Pennsylvania Working Papers in Linguistics 22.

Hartmann, J. 2016. The syntax and focus structure of specificational copular clauses and clefts.

     Habilitationsschrift. Universität Tübingen.

Los, B. L. J., & E. R. Komen. 2012. Clefts as resolution strategies after the loss of a

     multifunctional first position. In T. Nevalainen & E. C. Traugott (eds.), The Oxford

     Handbook of the History of English.

Schwarz, F. 2016. Presuppositions, projection, and accommodation- theoretical issues and

     experimental approaches, Ms. University of Pennsylvania, http://florianschwarz.net/wpcontent/

     uploads/papers/HandbookExpSemPrag_Ps.pdf.

Spenader, J. 2002. Presuppositions in spoken discourse. Ph.D. thesis, Stockholm University.

ABSTRACT. The early Modern English period (ca. 1500–1700, Görlach, 1991) has seen considerable attention from scholars due to it being influential in the shaping of English as we know it today. With reference to early Modern English orthography, scholars have touched upon a large number of topics: examples include epistolary writing (Gόmez Soliño, 1981; Osselton, 1984), gender-related patterns in orthography (Sönmez, 2000), dialectal variation (Nevalainen & Raumolin-Brunberg, 1989) and punctuation (Little, 1984; Cram, 1989; Nunberg, 1990; Salmon, 1962; 1999). While the ‘corpus revolution’ (Beal & Sen, 2014) that has transformed English historical linguistics has expanded our understanding of spelling practice, research on spelling variation in early Modern English still presents methodological and empirical problems. In particular, scholars have generally applied a traditionally qualitative, selective approach to the study of spelling variation, which has prevented a full understanding of the extent of spelling variation in early Modern English. This paper will propose a solution for conducting a more systematic, quantitative study of spelling variation between 1500 and 1700 by showing the potential of a combined application of VARD 2, DICER (Baron & Rayson, 2009; Baron et al., 2009) and a script that automatically detects spelling variants on digitised copies of original texts (ASVI, Schultz, 2016). By drawing on corpus linguistics technology, we intend to identify the extent of spelling variation across different corpora, and compile all of the variants identified together with the number of quotations in which they occur into a preliminary database of early Modern English spelling.

In order to build our database, we will draw on the Corpus of English Dialogues (Culpeper & Kytö, 2006), the Early English Books Online collection (Text Creation Partnership), the Early Modern English Medical Texts (Taavitsainen et al., 2010), the Helsinki Corpus (Rissanen, 1991), the Innsbruck Letters Corpus of ICAMET (Markus, 1999), the Lampeter Corpus of Early Modern English Tracts (Schmied et al., 1994), the Parsed Corpus of Early English Correspondence (Nevalainen et al., 1998) and the Shakespeare First Folio. The methodology used to create spelling variants will involve both manual and automatic processing of the normalisation rules provided by DICER, which will then be integrated with a systematic search conducted by the ASVI tool. Duplicate texts across corpora will be identified systematically and excluded from analysis in order to avoid skewing our results. The analysis will show that our database of early Modern English spelling allows for a quantitative analysis of the extent of spelling variation over two centuries and the identification of the most frequently recurring spelling variants, thus validating or challenging some of the claims made on the nature and the frequency of spelling variation in early Modern English (cf. Scragg, 1974; Görlach, 1991; Salmon, 1999). The analysis will then demonstrate how the database can be used to conduct a comparative, diachronic analysis of spelling variation. Preliminary results from our pilot study have shown that some of the most frequently occurring variants (e.g. the standardisation of word-final , the alternation between and and the alternation between , and ) responded differently to the process of standardisation during the early Modern English period – which poses the question of whether this process was unified throughout the centuries. Our paper could be of interest to specialists in diachronic spelling variation and beyond, across multiple languages and different time periods, thanks to the flexible and versatile methodology proposed.

ABSTRACT.  

The emergence of sentence-internal capitalization in German: New perspectives on the history of a spelling convention 

Sentence-internal capitalization of nouns and nominalizations is a feature of the German writing system that emerged gradually from the 16th to the 18th century (cf. e.g. Bergmann 1999). Previous studies have shown factors like reverence and animacy to be decisive in that process (cf. e.g. Kämpfert 1980; Labs-Ehlert 1993; Bergmann & Nerius 1998). It has been pointed out that the development of sentence-internal capitalization can be seen as a grammaticalization process (Maas 2007), as an originally pragmatic strategy (like reverence) becomes increasingly conventionalized and is not conditioned by pragmatic or (later) semantic factors (such as animacy) any more but rather by the grammatical status of the word in question: If it is the head of a noun phrase, it is obligatorily capitalized. 

Despite the fairly large amount of research addressing the emergence of sentence-internal capitalization, many important questions have not been addressed in sufficient detail so far. In this talk, we present a corpus study that provides answers to two of the most important among these open questions: 

1.) Is there a difference between printed and handwritten texts in the development of capitalization? It has been argued that in handwritten texts, the different steps in the conventionalization process of sentence-internal capitalization occur phase-delayed. However, the only large-scale empirical study taking handwritten texts into account has relied on the production of one single writer (Moulin 1990). 

2.) How do the different factors that have been argued to have an influence on the emergence of capitalization interact, and are there factors that have not yet been addressed? Most of the aforementioned empirical studies have been monofactorial, and they have largely focused on animacy. While animacy is indisputably an important factor, other aspects can be assumed to play a role as well. For example, frequency might either boost or inhibit capitalization, as it seems reasonable to assume that the spelling of highly frequent words is strongly entrenched and will therefore exhibit less variation than the spelling of low-frequency words. Given that animacy interacts with agentivity (Yamamoto 2008), the semantic role and the syntactic function of a noun in the respective sentence could also be a relevant factor. 

In order to address these questions, we compiled a corpus of handwritten texts, which comprises 56 handwritten protocols of witch interrogations edited by Macha et al. (2005). A multi-layer annotation was implemented semi-automatically and checked manually over the course of a three-year project. In addition, animacy, syntactic functions and semantic roles were annotated manually. A multifactorial analysis shows a complex interaction between the different factors. For instance, it shows that capitalization is less likely for high-frequency items that are allocated at the lower end of the animacy scale. In addition, it becomes clear that sociopragmatic factors play a decisive role in the scribes’ choice between uppercase and lowercase initials – for example, terms denoting women are significantly less often capitalized than words denoting men. 

In sum, the results of our study lend further support to previous accounts of the gradual emergence of sentence-internal capitalization but also indicate that the process was probably more complex than previously assumed. In addition, we argue that investigating handwritten texts adds an important new perspective to the study of sentence-internal capitalization as it allows for taking (quasi-)spontaneous situations of production as well as idiolectal variation into account.

References 

Bergmann, R. & D. Nerius. 1998. Die Entwicklung der Großschreibung im Deutschen von 1500-1700. 2 Bde. Heidelberg: Winter. 

Bergmann, Rolf. 1999. Zur Herausbildung der deutschen Substantivgroßschreibung. Ergebnisse des Bamberg-Rostocker Projekt. In Walter Hoffmann, Jürgen Macha, Klaus J. Mattheier, Hans-Joachim Solms & Klaus-Peter Wegera (eds.), Das Frühneuhochdeutsche als sprachgeschichtliche Epoche. Werner Besch zum 70. Geburtstag, 59–79. Frankfurt am Main. 

Kämpfert, Manfred. 1980. Motive der Substantivgroßschreibung. Beobachtungen an Drucken des 16. Jahrhunderts. 99. 72–98. 

Labs-Ehlert, Brigitte. 1993. Versalschreibung in althochdeutschen Sprachdenkmälern: ein Beitrag über die Anfänge der Großschreibung im Deutschen unter Berücksichtigung der Schriftgeschichte. (Göppinger Arbeiten Zur Germanistik.) Göppingen: Kümmerle. Göppingen: Kümmerle. 

Maas, Utz. 2007. Die Grammatikalisierung der satzinternen Großschreibung. Zur schriftkulturellen Dimension der Orthographieentwicklung. In Angelika Redder (ed.), Diskurse und Texte. Festschrift für Konrad Ehlich, 385–399. Tübingen: Stauffenburg. 

Macha, Jürgen, Elvira Topalović, Iris Hille, Uta Nolting & Anja Wilke (eds.). 2005. Deutsche Kanzleisprache in Hexenverhörprotokollen der Frühen Neuzeit. Bd. 1: Auswahledition. Berlin, New York: De Gruyter. 

Moulin, Claudine. 1990. Der Majuskelgebrauch in Luthers deutschen Briefen: (1517 - 1546). (Germanische Bibliothek : Reihe 3, Untersuchungen). Heidelberg: Winter. 

Yamamoto, Mutsumi. 2008. Animacy and reference. A cognitive approach to corpus linguistics. (Studies in Language Companion Series 46). Amsterdam, Philadelphia: John Benjamins.

ABSTRACT.  

 The new Basque Historical-Etymological Dictionary: advancements on the reconstruction of the Basque lexicon 

0 Introduction 

In this talk, we present the new Basque Historical-Etymological Dictionary; we also present some advancements on the reconstruction of the Basque inherited lexicon. 

1 Basque etymological dictionaries 

A collection of Mitxelena’s etymologies (Arbelaiz 1978) has been for years the best available Basque etymological dictionary. Some other later dictionaries can also be mentioned: 

a) Agud-Tovar (1988-1994): unfinished collection of etymological discussions with no critical assessment. 

b) Trask (2008): unfinished, English version of Mitxelena’s etimologies. 

c) Morvan: online-dictionary, written by an amateur advocate of macro-comparison. 

2 A new project 

In 2010, sponsored by the Academy of the Basque Language, our team started to write a Basque historical-etymological dictionary. The foundations of this project are the following: 

a) Words from all the letters (A to Z) were chosen, according to their antiquity, their general character across dialects, prior etymological discussion on them, etc. 

b) This dictionary benefits from the General Basque Dictionary, another project by the Academy that gathers the Basque historical lexicon. 

c) It includes the most recent advances on the reconstruction of Protobasque. 

Important results: 

a) A first edition of the dictionary: 200 families, 2000 words. 

b) Relevant advances on the study of medieval documentation. 

c) New etymological analyses. 

d) Forthcoming English and Spanish versions. 

3 Scientific etymology in Basque studies 

After decades of amateur etymology, a new era in Basque diachronic linguistics starts with Mitxelena. Based on Martinet’s work (1950), Mitxelena extends his reconstruction of occlusives to the whole phonological system (1961-1977), enriching it with a deep study of historical and dialectal data: this reconstruction is shaped by 1200 etymologies. He proposes a full reconstruction of the protobasque phonological system, basing his proposal mainly on loanwords from the latin-romance continuum. This reconstruction has also served as a basis for the phonological reconstruction of the inherited lexicon, but the information pulled from the adaption of loanwords is nearly exhausted. 

A new model for the reconstruction of Basque lexicon has been proposed by Lakarra (1995 and subsequent works), based on the definition of the canonical root: proto-Basque roots have been claimed to have a basic CVC structure, and from this seminal proposal many new analysis have been made of old known data, in which new roots and suffixes from the inherited lexicon have been identified: cf. the *bel ‘black, dark’ root, present in words such as beltz ‘black’, bele ‘crow’, arbel ‘slate’, orbel ‘dead 

leaves’, etc., the *din ‘become’ root (cf. jin ‘come’ < *edin), the hor ‘dog’ root (cf. hortz ‘tooth’), etc.; and comparing words such as hortz ‘tooth’ and beltz, in which a CVC-C syllable-structure has to be proposed, an ancient -tz suffix has been identified. 

4 New etymologies 

Thanks to Mitxelena’s work (1961-1977), we arrived to understand the formal relationship between words such as sari ‘price (> prize)’ and saldu ‘sell’. A *sali protoform is proposed by Mitxelena for sari, which has undergone the regular VlV > VrV change; Mitxelena’s proposal is that the original lateral consonant has been conserved in contexts where it was not between vowels, such as the one in saldu ‘sell’. 

Something similar is proposed for the pair gari ‘wheat’ and galdu ‘lose’, formally kin to the sari/saldu pair, although their semantic relationship seems to be much more obscure than the one between ‘price’ and ‘sell’ (but cf. lat. triticum ‘wheat’ and russian terjat’ ‘lose’, both from the same original indoeuropean *ter- root, see Buck s.v. lose); in this case, Mitxelena proposes a *gali protoform for gari ‘wheat’ too. If we take into consideration the root-theory, we need to identify a *gal root, for which no cognate can be found in the Basque inherited lexicon; however, if we relate gari ‘wheat’ to garagar ‘barley’, which seems semantically and formally plausible (we would deal with a reduplication of the root, an attested phenomenon in Basque), it turns out that a *gar root suits better the general explanation of this family of words, in which the gal- in galdu ‘lose’ has to be treated as an alomorph. 

This is the kind of new etymologies that the new Basque Historical-Etymological Dictionary offers, in which classical phonological reconstructions are enriched with the advancements on root-theory. 

5 Bibliography: 

Agud, M. & Tovar, A. 1988-1994. Diccionario etimológico vasco. Donostia: ASJU. 

Arbelaiz, J.J. 1978. Las etimologías vascas en la obra de Luis Michelena. Tolosa. 

Buck, C.D. 1988. Dictionary of Selected Synonyms in the Principal Indo-European Languages: A Contribution to the History of Ideas. University of Chicago Press. 

Campion, A. 1920. “La primera etimología baska, hasta hoy conocida”, RIEV 11(2), 119-120. 

Lakarra, J.A., 1995. “Reconstructing the Pre-Proto-Basque root”, in Hualde, J. I.; Lakarra, J. A. & Trask, R. L. (eds.) Towards a History of the Basque Language, John Benjamins, 189-206. 

Martinet, A., 1950. “De la sonorisation des occlusives initiales en Basque”, Word 6, 224-233. 

Mitxelena, K., 1961-1977, Fonética Histórica Vasca. Donostia: GFA. 

Morvan, M. Dictionnaire étymologique basque en français-espagnol-anglais. Online. 

Trask, R.L. 2008. Etymological Dictionary of Basque. University of Sussex.

ABSTRACT. Phonological research in historical linguistics typically relies on written source data. While it is generally difficult to infer phonological information from written texts, this is particularly so when grapheme-to-phoneme correspondences that may have held straightforwardly at a particular time are rendered opaque by gradually spreading sound changes. In this paper, we present a methodological work-around that helps us to derive phonological interpretations of written corpus data under precisely such circumstances. Specifally, we examine the development of word-final consonant clusters in Middle English, a period in which the process of schwa loss unfolded gradually (Minkova 1991). Schwa loss, which deleted final schwas (ande > and) as well as checked ones (godes > gods) in final unstressed syllables, clearly increased the frequency of coda clusters. The problem is that schwa loss is not systematically reflected in written prose data, which can therefore be safely interpreted only for the period before schwa loss (in which vowel graphs in final unstressed syllables indicated phonological substance reliably), and for the period after its completion (where vowel graphs, usually ’s were normally ‘silent’). Our paper shows how models of logistic growth can be calibrated to phonologically interpretable poetry data in order to infer time-dependent frequencies of word-final consonant clusters in corpus data of prose writing. Our procedure involved five steps. First, we selected a set of 47 poems covering the period from the 12th to the 18th centuries. From each century, a sample of roughly 70 word tokens with graphemes that potentially represented schwas (normally ) was drawn. The sample size was chosen so that the margin of error was roughly 0.1. Through rhythmic interpretation, we then determined for each token in the sample whether or not actually represented /ə/. Second, we tested – by means of a generalized linear model – to what extent the likelihood of actually representing /ə/ was conditioned by two factors, namely (a) the onset of the following word, and (b) whether the final syllable/consonant was part of the stem or represented a suffix. Predictably, we found that before words starting with vowels, the likelihood of final s representing /ə/s decreased significantly. Morphology, in contrast, yielded no significant effects. This made us distinguish three different environments: checked , final before a consonant and final before a vowel. Third, for each environment and each century, we calculated the probability of s representing /ə/s, resulting in three probability trajectories over the whole observation period. A logistic growth model (Altmann 1983; Kroch 1989; Denison 2003; Wang & Minett 2005; Blythe & Croft 2012) restricted to the unit interval and with a variable growth rate and turning point was fitted to each trajectory using non-linear least squares approximation (R Development Core Team 2013). This allowed for an interpolation of the probability of s representing /ə/s for each point in time in the observation period (Fig. 1).

Fourth, we returned to the (prose) corpus data we were primarily interested in (the Penn Helsinki Parsed Corpora: Kroch & Taylor 2000; Kroch et al. 2004). We extracted all word tokens that could potentially end in consonant clusters, and checked (a) if they contained graphemes that could potentially represent /ə/s and (b) whether or not they were followed by vowel-initial words. Thus, we assigned each token to one of the three environments established in step three. Finally, we derived (by means of the calibrated logistic growth models, and from dates and assigned environments) a probability with which each token actually ended in a cluster. These probabilities were then assigned to the respective tokens as weights. This allowed us to calculate (estimated) cluster-specific token frequencies for each period (e.g. decade or century) as the sums of the weights assigned to the respective tokens. We show that our estimates of the spread of schwa loss largely coincide with point estimates for the onset and offset of schwa loss found in the literature (Dobson 1957; Fisiak 1968; Brunner 1984; Minkova 1991; Mossé 1991). The additional advantage of our method is that it makes it possible to determine (or at least estimate) phonotactic token frequencies for periods between these two point estimates as well. On a more general level, we demonstrate that if well-informed and established formal models of linguistic change are used as what they actually are – namely as functions that describe diachronic trajectories, rather than just diagnostic machines for the delivery of potentially significant statistical effects – much more phonological information can be derived from historical corpus data than has for far been acknowledged.

ABSTRACT. Korandje is a highly divergent, geographically isolated Songhay language spoken in the oasis of Tabelbala in southwestern Algeria. Its inherited Songhay lexicon has been whittled down to a core vocabulary of less than three hundred words, with Berber and Arabic loanwords accounting for the rest. Among these, even a casual glance shows a curious asymmetry: almost all nouns end in a vowel, and almost all verbs in a consonant. Indeed, in some cases the same proto-Songhay root yields two different reflexes: a vowel-final noun, and a corresponding consonant-final verb. Comparative Songhay data reveals that this distribution is neither coincidental nor ancient: in fact, word-final vowels have been systematically lost in verbs that historically ended in *VCV# (except for the subclass discussed below) and retained in other contexts. Intermediate stages in this process, all restricted to verbs, can be reconstructed through comparison with other Northern Songhay varieties.

On the face of it, this looks like a change conditioned by part of speech, a possibility defended by several recent works including Crowley and Bowern (2010:171). Such a conditioning, however, violates the Neogrammarian principle that sound changes are regular and phonetically conditioned, and most or all apparent counterexamples can be explained in other ways (Hill 2014). A Neogrammarian would therefore predict that this apparently regular sound change should in fact reflect analogical change, as in Blevins and Lynch's (2009) analysis of Paamese.

This prediction appears to be borne out. One semantically definable class of verbs is systematically exempt from final vowel loss: verbs of motion and position. This gap can be explained neither phonologically nor in terms neither of part of speech, but follows directly from morphological considerations: the class of verbs in question – like nouns and adjectives – happen not to take vowel-initial suffixes. Bearing in mind that stem-final vowels are elided when a vowel-initial suffix is added, this distribution points points to reanalysis as a key driver for verb-final vowel loss: the process was driven by consonant-final allomorphs being reanalysed as bases, in the context of borrowing-driven relaxation of inherited phonotactic constraints on final consonants. The Neogrammarian assumption that sound changes proper are exclusively phonetically conditioned thus remains fertile and tenable – as long as we bear in mind that morphology can drive analogical change conditioned by part of speech.

References Blevins, Juliette & John Lynch. 2009. Morphological Conditions on Regular Sound Change? A Reanalysis of *l-Loss in Paamese and Southeast Ambrym. Oceanic Linguistics 48(1). 111–129. Crowley, Terry & Claire Bowern. 2010. An introduction to historical linguistic. 4th ed. Oxford: Oxford University Press. Hill, Nathan W. 2014. Grammatically Conditioned Sound Change. Language and Linguistics Compass 8(6). 211–229.

ABSTRACT. Blackfoot (ISO 693-3: bla) is the westernmost Algonquian language. It has also been called ‘[t]he most divergent Algonquian language’ (Goddard 1994: 187), containing phonological sequences and morphology not found in the other languages. The Proto-Algonquian-Blackfoot Hypothesis claims that Blackfoot was the first language to split from Proto-Algonquian, with the remaining languages forming a subgroup, as in (1).

(1) Proto-Algonquian-Blackfoot (Goddard:2015) -- Blackfoot -- Proto-Algonquian -- Cree-Montagnais -- Arapaho-Atsina -- Menominee -- Cheyenne -- etc.

This hypothesis predicts either the existence of shared innovations among the non-Blackfoot languages, or innovations in Blackfoot that are not shared by the other languages. Goddard (2015) discusses two aspects of Blackfoot’s ‘divergent’ morphological nature and claims they are archaisms from Proto-Algonquian-Blackfoot. This paper instead seeks to determine phonological innovations which could corroborate the Proto-Algonquian-Blackfoot Hypothesis.

I focus on reflexes of Proto-Algonquian (PA) clusters in Blackfoot. The Blackfoot forms are draw mainly from the Blackfoot dictionary Frantz & Russell (1995), which uses a standardized orthography. These are compared to cognate forms in the other Algonquian languages found in Proto-Algonquian dictionaries (e.g. Hewson 1993). I include a list of over 40 reconstructions.

I find no phonological evidence for the subgrouping. However, the results do add to our understanding of Blackfoot historical phonology in several ways. First, some phonological distinctions are maintained in clusters which were lost elsewhere. Second, contrary to Berman’s (2006) claim that the first member of true PA clusters always reduces to Blackfoot *h or *ss, I show that PA *? is preserved as a glottal stop in some environments. Third, the first member of clusters ending in *k often merges to [?] in Blackfoot, although the conditioning factors are not yet understood.

ABSTRACT. It is well known that Spanish and Portuguese display comparable patterns of dialectal variation regarding the expression of subject pronouns in wh-interrogatives. In Caribbean Spanish (CS) and Brazilian Portuguese (BP) wh-interrogatives, subject pronouns are typically realized between the interrogative pronoun/adverb and the verb, as in (1a). In contrast, in European Spanish (ES) and Portuguese (EP) wh-interrogatives, subject pronouns are usually omitted, as in (1b).

(1)          a.            CS           ¿Qué tú quieres?

                                BP           O que você quer?

                b.            ES           ¿Qué quieres?

                                EP           Que queres’?

                                              ‘What do you want?’

It is well known that Spanish and Portuguese display comparable patterns of dialectal variation regarding the expression of subject pronouns in wh-interrogatives. In Caribbean Spanish (CS) and Brazilian Portuguese (BP) wh-interrogatives, subject pronouns are typically realized between the interrogative pronoun/adverb and the verb, as in (1a). In contrast, in European Spanish (ES) and Portuguese (EP) wh-interrogatives, subject pronouns are usually omitted, as in (1b).

(1)          a.            CS           ¿Qué tú quieres?

                                BP           O que você quer?

                b.            ES           ¿Qué quieres?

                                EP           Que queres’?

                                                ‘What do you want?’

The fact that overt pronouns are frequently placed after the interrogative pronoun/adverb in CS and BP has often been explained as a corollary of the fact that the likelihood of use of an overt NP before the verb is higher in these dialects than in their European counterparts in any sentence type  (see, e.g., Cameron and Flores-Ferrán 2004 for Spanish; Barbosa, Duarte and Kato 2005 for Portuguese). This finding, in turn, is typically attributed to the influence of African substrates. For instance, there is a higher incidence of the use of overt NPs before the verb in the Antillean bozal, leading to sequences such as “Uté mira, tó mundo, ripiá, facitó, tó mundo camina sobre tiera. Y cuando uté quié ensuciá, uté ensucia la tiera. Y son deuda que uté ta crea con tiera” (Otheguy 2000[1975]: 375-376). In summary, this Contact Hypothesis assumes that the higher likelihood of use of preposed NPs in wh-interrogatives in CS and BP is a direct result of the strong contact between Romance and African languages in the Caribbean and in Brazil.

            Although the Contact Hypothesis has been maintained by many authors (see the summary in Gutiérrez Maté 2013: 35-38), to my knowledge there does not yet exist a quantitative study testing the relevance of the hypothesis for the description of the use of overt subject pronouns in wh-interrogatives. The Contact Hypothesis predicts that (a) the discourse functions of overt pronouns in wh-interrogatives are comparable in CS and BP, and (b) CS and BP differ in the same manner from the respective European dialects regarding their use of overt pronouns in wh-interrogatives.

            I use data from the new two billion-word collection of texts from web pages and blogs in the Corpus del español (Davies 2015-2016) in order to test these predictions. Due to its size and the informality of web texts, the corpus allows for a detailed statistical analysis of (a) the correlations between the usage frequencies of overt and non-overt subject pronoun wh-interrogatives in Spanish and Portuguese dialects and (b) the factors governing the use of overt subject pronouns in CS and BP.

            Preliminary results from the analysis suggest that while the mechanisms governing the use of overt subject pronouns in wh-interrogatives are indeed comparable in CS and BP, its use is conventionalized to a greater degree in BP than in CS, a result that is difficult to explain by reference to the Contact Hypothesis alone.

            Although the Contact Hypothesis has been maintained by many authors (see the summary in Gutiérrez Maté 2013: 35-38), to my knowledge there does not yet exist a quantitative study testing the relevance of the hypothesis for the description of the use of overt subject pronouns in wh-interrogatives. The Contact Hypothesis predicts that (a) the discourse functions of overt pronouns in wh-interrogatives are comparable in CS and BP, and (b) CS and BP differ in the same manner from the respective European dialects regarding their use of overt pronouns in wh-interrogatives.

            I use data from the new two billion-word collection of texts from web pages and blogs in the Corpus del español (Davies 2015-2016) in order to test these predictions. Due to its size and the informality of web texts, the corpus allows for a detailed statistical analysis of (a) the correlations between the usage frequencies of overt and non-overt subject pronoun wh-interrogatives in Spanish and Portuguese dialects and (b) the factors governing the use of overt subject pronouns in CS and BP.

            Preliminary results from the analysis suggest that while the mechanisms governing the use of overt subject pronouns in wh-interrogatives are indeed comparable in CS and BP, its use is conventionalized to a greater degree in BP than in CS, a result that is difficult to explain by reference to the Contact Hypothesis alone.

ABSTRACT. Languages, as we know, are always changing. Over time some innovations manage to get incorporated into the language system as a whole, whereas other changes spread across distinct areas of the speech community and produce regional varieties, which are typical of all languages (Penny 2004). Within this scenario of dynamic instability, linguistic elements differ in terms of their relative propensity for change, although none have been found to be truly immutable (Nichols 2003).

In the present work, we focus on a series of grammatical changes associated with different periods of Spanish history, and yet interrelated to the extent that the category affected by the successive innovations is the opposition between accusative and dative case inherited from Latin.

The phenomena to be examined are:

 (1) a-marked animate direct objects  (Romance origin)

            Vi a_acc mi amigo         vs         Vi Ø la carta

            ‘I saw my friend’                    ‘I saw the letter’

 (2) leísmo = the use of the dative pronoun le for animate direct objects  (early Spanish)

            Le_dat vi                        vs         Lo_acc vi

            ‘I saw him’                              ‘I saw him’

 (3) doubling of the stressed personal pronouns  (renaissance Spanish)

            Lo_acc vi  a él _acc          and      Le_dat hablé  a él_dat

            ‘I saw him’                             ‘I talked to him’

 (4) doubling of the dative nominals  (modern Spanish)

            Le_dat mandé una carta a mi amigo_dat vs        Mandé una carta a mi amigo_dat

            ‘I sent a letter to my friend’                            ‘I sent a letter to my friend’

Our study of these phenomena will evidence the sensitivity of Spanish to the familiar Hierarchy of Topicality (Givón 1976) and will simultaneously bring to the fore how this internal property has operated in detriment of the case distinction (Melis & Flores 2009). We will furthermore be concerned with the dialectal variation hinging on leísmo, whose origin has been traced back to the contact of Spanish with Basque (Fernández-Ordóñez 2001). At present, in very broad terms, the phenomenon of leísmo pits the Old World (presence in Peninsular Spanish) against the New World (absence in American Spanish). We will invoke a scenario of ‘leveling’ through dialect contact (Trudgill 1986) during the early phase of colonization to explain the cause of the linguistic split, and will argue that reflexes of the split can be detected in the ongoing process of diffusion of the doubled dative nominals (Flores & Melis 2004).

In our concluding remarks, looking towards the future, we shall try to imagine what could happen to the case distinction in the newly emerging varieties of U.S. Spanish.

references

Fernández-Ordóñez, Inés. 2001. “Hacia una dialectología histórica. Reflexiones sobre la historia del leísmo, el laísmo y el loísmo”. Boletín de la Real Academia Española 81: 389-464.
Flores, Marcela & Chantal Melis. 2004. “La variación diatópica en el uso del OI duplicado”. Nueva Revista de Filología Hispánica 52(2): 329-354.
Givón, Talmy. 1976. “Topic, pronoun and grammatical agreement”. In Ch. Li (ed.), Subject and Topic, 151-188. New York: Academic Press.
Melis, Chantal & Flores, Marcela. 2009. “On the interplay between forces of erosion and forces of repair in language change”. A case study. Folia Linguistica Historica 30: 271-310.
Nichols, Johanna. 2003. “Diversity and stability in language”. In B. D. Joseph & R. D. Janda (eds.), The handbook of Historical Linguistics, 283-310. Oxford: Blackwell
Penny, Ralph. 2004. Variation and change in Spanish. Cambridge: CUP.
Trudgill, Peter. 1986. Dialects in contact. Oxford: Blackwell.

ABSTRACT. The dialects of Latin American Spanish offer one of the most complex singular address systems in Romance (Rona 1967, Moyna & Rivera-Mills 2016). Considered collectively, they boast three singular forms  (tú, vos, usted), all of which are standard over several national varieties. While usted tends to be reserved for formality and has thus remained distinct, vos and tú have not. Although tú corresponds to the etymological singular, and vos was originally plural and then polite singular, the latter eventually lost deferential value and started to compete with the former. This blurring of pragmatic and semantic values, coupled with a high number of verbal homomorphs (Lapesa 1970a,b, Cuervo 1893), entangled both address forms to such an extent that the resulting pronominal and verbal paradigms share many forms (Fontanella de Weinberg 1976, 1977) (Table 1). This leveling happened independently all over the New World, as evinced by the variability exhibited by epistolary sources from several centuries and distant locations (1, 2).

            Considered at the right level of abstraction, the early variability resulted in three basic outcomes across the continent. In most dialects (e.g., Antilles, Mexico, Peru), voseo was completely eliminated quite early, resulting in an informal second person singular paradigm identical to that of Peninsular Spanish (e.g., tú hablas ‘you talk’) (Páez Urdaneta 1981: 66, Benavides 2003). In other areas, cut off from the Peninsula and from each other, such as Central America and the River Plate, non-diphthongized voseo forms prevailed (cf., vos hablás ‘id.’). In a smaller number of dialects, most of them receding, diphthongized voseo survived (vos/tú hablái(s) ‘id’) (Granda 1978).

            Although the evolution of the three outcomes is of interest, here I will focus solely on the second (non-diphthongized) solution, and try to account for a number of puzzling facts. The first fact that needs accounting for is the virtually identical pronominal and verbal outcome of this non-normative pattern in Central America and the River Plate, two areas that had no direct contact throughout the colonial and postcolonial periods. The second fact to explain is why in both sets of dialects, only a handful of verbal and pronominal slots adopted voseo, while tuteo prevailed elsewhere in the paradigm. The third fact, which I will show through historical and contemporary evidence, is that in both areas the shift to voseo followed the same order (imperative ® present indicative ® present subjunctive; subject ® prepositional object).

            The thesis to be presented is that, absent any evidence that some forms were more frequent in the original dialectal mix, the explanation must be based on the circumstances that were indeed similar throughout Spanish America. Those include: (a) dialectal variability among the settler contingents, resulting in no clear target forms; (b) high numbers of second language learners among the early indigenous populations and later immigrant contingents, (c) high rates of intermarriage and language contact which exposed the children acquiring Spanish to a mixed input. These conditions, I posit, resulted in a process of change led by L1 acquisition (Chambers 2004, Schreier et al. 2010, Trudgill 2010). The clearest demonstration comes from the striking similarities between the natural order of acquisition of verbal and pronominal forms in Spanish and the order of historical adoption of voseo forms (Moyna 2009).

            This paper does not suggest that child acquisition is always the only or the main reason behind the process and outcome of language change. It does suggest that in the right demographic, social, and linguistic circumstances, natural child language acquisition can provide clues for cross-dialectal parallelisms that would otherwise remain unexplained. The study of the history of Spanish would do well to heed the Biblical dictum to ‘let the children come.’

(1) quel marido no era tan malo que otros mas altos que vos se holgaran […] si quieres venirte aqui a estar debaxo de mi mano trae tu mujer e hijos que yo te mantendre […] los onbres de bien an de tener otros terminos que los que vos aveis tenido el tiempo te dara el pago (Mexico City, 1574, letter from Catalina Martin to her son Francisco Marrero, CORDIAM)

      ‘that my husband wasn’t so bad that other people more worthy than you_V would despise him […] if you_T want to come here to be under my wing bring_T your_T wife and children and I will support you_T [...] Good men have to behave in ways different from those you_V have_V had. Time will pay you_T.’

(2)  Visitalo, dadle [...] y sirvelo (Buenos Aires, 1816, Fontanella de Weinberg 1971)

      ‘Visit_V him, give_V him [...] and serve_T him’

Selected References

Benavides, Carlos. 2003. La distribución del voseo en Hispanoamérica. Hispania 86 (3): 612-623.

Chambers, J.K. 2004. Dynamic typology and vernacular universals. Dialectology Meets Typology: Dialect Grammar from a Cross-Linguistic Perspective, Bernd Kortmann (ed.), 127-145. Berlin: Mouton de Gruyter.

Cuervo, R. J. 1893. Las segundas personas de plural en la conjugación castellana. Romania 22: 71-86.

Fontanella de Weinberg, María Beatriz. 1976. Analogía y confluencia paradigmática en formas verbales de voseo. Thesaurus 31: 249-272.


Fontanella de Weinberg, María Beatriz. 1977. La constitución del paradigma pronominal del voseo. Thesaurus 32: 227-241.

Granda, Germán de. 1978. Las formas verbales diptongadas en el voseo hispanoamericano. Una interpretación sociohistórica de datos dialectales. Nueva Revista de Filología Hispánica, 27: 80-92.

Lapesa, Rafael. 1970a. Las formas verbales de segunda persona y los orígenes del ‘voseo.’ Actas del Tercer Congreso Internacional de Hispanistas, Carlos H. Magis (ed.), 519-531.  México, D.F.: El Colegio de México.

Lapesa, Rafael. 1970b. Personas gramaticales y tratamientos en español. Revista de la Universidad de Madrid 19: 141-167.

Moyna, María Irene, and Susana Rivera-Mills (eds). 2016. Forms of Address in the Spanish of the Americas. Amsterdam/Philadelphia: John Benjamins.

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.

Páez Urdaneta, Iraset. 1981. Historia y geografía hispanoamericana del voseo. Caracas: La Casa de Bello.


Rona, José Pedro. 1967. Geografía y morfología del “voseo.” Porto Alegre: Pontifícia Universidade Católica.


Schreier, Daniel, Peter Trudgill, Edgar Schneider and Jeffrey P. Williams (eds). 2010. The Lesser-Known Varieties of English: An introduction. Cambridge: Cambridge University Press.

Trudgill, Peter. 2010. Investigations in Sociohistorical Linguistics. Stories of colonization and contact. Cambridge: University of Cambridge Press.

ABSTRACT. In grammaticalization research it has become generally accepted that the semantic development of a grammaticalizing construction can be derived from its source meaning (e.g. Bybee et al. 1994, followed by many others). While this approach has been a highly useful concept in explaining semantic trajectories, what has been overlooked is that the development of a construction is not only determined by the overall constructional meaning, but also significantly by the function words which are part and parcel of many grammaticalizing constructions. I present in this talk a typology of what I refer to as old grams (cf. Reinöhl 2010, Reinöhl & Himmelmann, to appear), i.e. function words already present from the start in a grammaticalizing construction, such as the -ing or the to in be going to VP. I highlight that old grams may play widely different roles within constructions and that investigating them helps not only i) to understand individual grammaticalizations better, but also ii) to explore the variation space in the semantic and formal development across different types of grammaticalization phenomena. In addition, a study of old grams unveals that iii) grammaticalization and reinforcement, considered distinct processes in some writings, span up a continuum rather than present a simple bifurcation. Already after a brief look at what are possibly the most oft-quoted examples for grammaticalization – English be going to, and French chanterai and ne… pas – it becomes clear that old grams can play very different roles in the constructions they grammaticalize in. The two French examples alone show how different these roles can be: While the old gram in chanterai, the infinitive morpheme in cantare habeo (lit. ‘I have to sing’), has hardly left a functional and/or formal trace, the old gram ne was the core of the construction that it came to form with pas, the latter having been added to ne as a reinforcer. Ne determined form (e.g. position) and function (negation of verbs) of the new construction and pas only became a negator in its own right long after having ceded its lexical meaning to that of the overall construction, negation. Cases such as ne…pas have been described by some authors as a separate phenomenon from grammaticalization, namely as reinforcement (e.g. Lehmann 2002 [1982]: 20). On the other hand, at least as many consider it a prototypical grammaticalization (e.g. Meillet 1975 [1912], Hopper & Traugott 2003). However, ne… pas shows a trajectory quite unlike that of chanterai, for instance. While the future meaning of chanterai can be traced back to the modal meaning of habeo (Reinöhl & Himmelmann, to appear), it is hard to claim the same for the relation between the original meaning of pas, ‘step’, and that of negation. This is due to the fact that the construction habeo cantare is a grammaticalization of the lexeme habeo within a construction with an infinitive, while ne… pas was not built around the lexeme pas, but around the old gram ne. Consider two more examples where old grams play yet other roles: The Hindi postposition mẽ goes back to a locative Sanskrit madhye (middle.LOC.SG.N) ‘in the middle’. Today, its functional space shows an almost uncanny resemblance to that of the old Sanskrit morphological locative (Reinöhl 2010). Madhye itself, of course, was a locative. While the new construction of an oblique noun governed by the postposition mẽ was not “built around” the old locative in the same way as ne… pas was built around ne, the new gram does contain the old locative in its original morphological make-up. Are we dealing here with grammaticalization or with reinforcement? Also within the domain of “reinforcement” do we find variation. For instance, in spoken German, some spatial adpositions often do not occur on their own any more, but get reinforced by partially cognate adverbs, all in one intonation contour, e.g. Ich bin im Haus drinnen (‘I am in the house inside’), Ich sitze auf dem Dach drauf (‘I sit on the roof on top.’). Here, the old gram and the reinforcing element have almost the same meaning. By contrast, pas had by no means almost the same meaning as ne when it started being added to ne. Thus, is the case of the German adposition and adverb a more prototypical case of reinforcement than ne… pas? On the basis of a selected number both of well-known grammaticalization examples as well as of some lesser known ones I will develop a typology of old grams that takes into account their semantic and formal roles inside the respective grammaticalizing constructions, and assesses the different ways in which they influenced the development of the latter. To take into account the roles played by old grams helps understand pathways of change as well as disentangle differences between sub-types of grammaticalization. To distinguish between such sub-types is a necessary step in order to identify the appropriate level of generalization for what is and what isn’t defining for grammaticalization. To give only one example, the well known bleaching metaphor applies well only to such grammaticalizations that are centrally built around a content word and not around an old gram, while there may also be intermediate cases, as for instance when a content word which grammaticalizes contains an old gram in its morphological structure.

References Bybee, Joan, Revere Perkins & William Pagliuca. 1994. The Evolution of Grammar. Tense, aspect, and modality in the languages of the world. Chicago, London: The University of Chicago Press. Hopper, Paul J. & Elizabeth C. Traugott. 2003. Grammaticalization (2nd edition). Cambridge: Cambridge University Press. Lehmann, Christian. 2002 [1982]. Thoughts on Grammaticalization (2nd, revised edition). (Arbeitspapiere Seminar für Sprachwissenschaft 9). Erfurt: Seminars für Sprachwissenschaft der Universität Erfurt. Meillet, Antoine. 1975 [1912]. “L’évolution des formes grammaticales”, in: Antoine Meillet. Linguistique Historique et Linguistique Générale. Paris: Champion, 130-148. Reinöhl, Uta. 2010. Zum Begriff der Renovation im Rahmen der Grammatikalisierungstheorie. Magisterarbeit. Westfälische Wilhelmsuniversität Münster. (unpublished) Reinöhl, Uta & Nikolaus Himmelmann. “Renovation – A figure of speech or a process sui generis?”. To appear in Language.

ABSTRACT. In this paper, the micro-variation in Germanic pseudo-partitives is investigated, focusing on what appears to be a ‘failed’ or ‘arrested’ change in English. Partitives consist of two nominals, which are usually designated N1 (a portion or container) and N2 (that which is contained or portioned). A distinction has long been made between two different construction types: true partitives and pseudo-partitives (Selkirk 1977, Jackendoff 1977, Koptjevskaja-Tamm 2001, Alexiadou, Haegeman & Stavrou 2007, Rutkowski 2007). Essentially, the embedded nominal (N2) in a true partitive must be definite whereas in the pseudo-partitive N2 is either non-count or plural (i.e. it cannot be singular and count). Partitive binominal constructions of the form N1 of N2 (a cup of tea) have a straightforward syntactic analysis with N1 as the head, and [of N2] as a postmodifying prepositional phrase. The structure of pseudo-partitives is less clear and within the Germanic languages there appears to be considerable microvariation. Partitives have long been identified as a locus of semantic and syntactic change. For example, Traugott (2008:27) building on Denison (2002) identifies 5 general stages: (1) lexical noun>partitive (binominal) >complex determiner/quantifier > degree adverb > free adverb One of the well-documented examples of this change is English lot, which in (one of) its lexical meanings refers to ‘a portion or share of property’ but is also a complex determiner/quantifier, ‘a lotta’. The change focused on here involves the cognates of English pair in other Germanic languages, e.g. Dutch (paar) and Danish (par). The word has grammaticalised to a quantifier in these languages, and cognates of pair may mean ‘few’, as in (2) and (4) while also retaining the only meaning available in English ‘a set of two’ as in (3): (2) Du. a. Er staan een paar schoenen op de tafel quantification reading there stand a few shoes on the table ‘A few shoes are on the table’

(3) Du. b. Er staat een paar schoenen op de tafel partitive reading there stands a pair shoes on the table ‘A pair of shoes is on the table’ (van Riemsdijk 1998:17)

(4) Da. så lod hun hænderne glide søgende gennem de sidste par jordbærplanter as let she hands-the glide, searching through the last few strawberry plants ‘As she let her hands glide searching through the last few strawberry plants’ (Korpusdk; Familie-Journalen) However, historically English pair could mean ‘a few, some’ as evidenced in (5) from two different manuscripts of the same text, Cursor Mundi. (5) a. Þe king a pair o letters writte did (1400 Cursor Mundi (Vesp.) b. Þe king did sone lettris to write (1400 Cursor Mundi (Gött.) ‘The king wrote some letters.’

An additional related question to be addressed is whether there is historical evidence for Direct Partitive Constructions in English. In the West Germanic languages (other than English) and mainland Scandinavian languages, two different types of pseudo-partitive have been discussed (van Riemsdijk 1998:11 for Dutch, Hankamer & Mikkelsen 2008:318 for Danish). These two strategies for forming the pseudopartitive are referred to as Direct Partitive Constructions (DPC), which do not use a preposition and Indirect Partitive Contructions (IPC) which do. (6) Da. a. en gruppe turister (unrestricted set: pseudopartitive, DPC) a group of tourists

Da. b. en gruppe af turister (unrestricted set: pseudopartitive, IPC) a group of tourists (Hankamer & Mikkelsen 2008:318)

If the presence or absence of of is taken to be the only indicator for distinguishing between an IPC and a DPC, then, as Hankamer & Mikkelsen (2008:319) state, English has only IPCs. However, historically a linking of was frequently not used: (7) I bequethe to Marie Tendall, my goddoughter, my peir bedys of calcidenys gaudied with siluer and gilt. ... (1482: copy of will of Margaret Paston) Selected References Alexiadou, Artemis, Lilianne Haegeman, & Melita Stavrou. 2007. Noun phrase in the generative perspective. Berlin: Mouton de Gruyter. Denison, David. 1998. Syntax. In The Cambridge history of the English language, ed. Suzanne Romaine, 92-329. Cambridge: Cambridge University Press. Hankamer, Jorge & Line Mikkelsen. 2008. Definiteness marking and the structure of Danish pseudopartitives. Journal of Linguistics. 44. (317–346). Riemsdijk, Henk van. 1998. Categorial feature magnetism: the endrocentricity and distribution of projections. Journal of Comparative Germanic Linguistics. 1.1–48. Rutkowski, Pawel. 2007. The Syntactic Structure of Grammaticalized Partitives (Pseudo-partitives). In T. Scheffler, J. Tauberer, A. Eilam and L. Mayol, eds. University of Pennsylvania Working Papers in Linguistics, vol. 13.1: Proceedings of PLC 30, 337-350, Philadelphia: Department of Linguistics, University of Pennsylvania. Selkirk, Elizabeth. 1977. Some remarks on noun phrase structure. In Akmajian, Adrian, Peter Culicover & Tim Wasaw eds. Formal Syntax. New York: Academic Press. Jackendoff, Ray. 1977. X-bar syntax: A study of phrase structure. Cambridge, Mass: MIT Press. Koptjevskaja-Tamm, Maria. 2001. "A Slice of the Cake" and "a Cup of Tea": Partitive and Pseudo-Partitive Constructions in the Circum-Baltic Languages. In Dahl. Östen. & Maria Koptjevskaja-Tamm (eds.). The Circum-Baltic Languages: Typology and Contact, v.1 - 2. Amsterdam /Philadelphia: John Benjamins. Traugott, Elizabeth. 2008. The grammaticalization of NP of NP patterns. In Bergs, Alexander & Gabriele Diewald eds.Constructions and Language Change. Berlin: de Gruyter.

ABSTRACT.  

 Grammaticalization and the Emergence of Personal Pronouns 

Recent typological studies by Heine and Song (2010, 2011) use four grammaticalization parameters (extension, desemanticization, decategorialization, and erosion) to analyze five major conceptual sources of personal pronouns: (a) nominal concepts, (b) spatial deixis (i.e. demonstratives), (c) reflexives/intensifiers, (d) plurification, and (e) shift in deixis. This study argues that while the development of noun-based forms can be considered the canonical case of grammaticalization, the developmental processes concerning other sources fall outside the scope of grammaticalization, thus requiring separate treatments. More specifically, I show that (i) based primarily on data from Asian languages which they cite as examples where demonstratives and reflexives gave rise to personal pronouns, there is little functional reason to treat them as grammaticalized personal pronouns and (ii) while plurification and shift in deixis contribute to the emergence of new pronoun usage, their development is distinct from grammaticalization. 

The general thrust of this paper will partly be negative in that I will demonstrate unnecessities and inadequacies of grammaticalization in accounting for the development of forms of non-lexical origins and offer basic suggestions for future approaches without proposing a complete framework. The positive contribution of this paper consists in laying out the developmental processes involving person forms of non-lexical origins that are not consistent with grammaticalization and should therefore be accounted for in different ways. 

Though demonstratives and reflexives are frequently used for person referents, examinations indicate that their usage does not go beyond the scope of the demonstrative and reflexive functions. In many languages, demonstrative-based forms are used in the way that is predictable from their spatial semantics: speaker-proximal forms for the speaker and non-speaker-proximal forms (i.e. speaker-distal and addressee-proximal) for the addressee, as in (1). 

(1) a. Japanese: konata ‘this way’ for first person and sonata ‘that way’ for second person 

b. Korean: i jjog/pyon ‘this side’, i gos ‘this place’, and yeogi ‘here’ for first person, and jeogi ‘there’ for second person 

c. Thai: nîí ‘this (one)’ for first person 

d. Vietnamese: đây ‘here’ for first person and đó ‘there’ for second person 

e. Chinese dialect of Huojia: zher ‘here’ for first person plural (Hagège 1993) 

These usages are not innovative in that it is a simple metonymic extension of the demonstrative function. In addition, parameters other than extension do not apply to these forms, including Japanese sonata which is a second person pronoun according to Heine and Kuteva (2002). With the absence of change observable by the parameters, however, one cannot in principle distinguish demonstratives from various items such as kinship terms often used pronominally in these languages. 

Parameters also face difficulty when applied to reflexives in other Asian languages, including Japanese jibun which Heine and Song treat as a first person pronoun. The same can be said for forms like Korean caki and Mandarin ziji (Zubin et al. 1991), and Thai tua and Vietnamese mình (Cooke 1968). Being morphologically invariant, reflexives in these languages are used for all person categories, and serve the function of (self-)objectification in discourse where the use of reflexives instead of personal pronouns presents a referent as an entity that should be seen and evaluated objectively: e.g. Japanese jibun wa kore de ii no daroo ka ‘(I) wonder if this is how self(=I) should be’ in which the speaker treats his/her ego as if it were another entity, creating a sense of self-reflection. I argue that the perception of reflexives as personal pronouns is due to this discourse behavior of reflexives. The question of grammaticality has to be considered as well: (assuming that demonstratives/reflexives do develop into personal pronouns) are personal pronouns more grammaticalized than demonstratives/reflexives? There is no obvious answer to this question, offering no clear reason to treat demonstratives and reflexives in these languages as personal pronouns. 

In the case of plurification (e.g. French vous) and shift in deixis (e.g. German Sie) to which only extension applies, semantic features ordinarily reserved for another member of the same paradigm are given the new usage, which is analogous to what Joseph (2005) calls “lateral shift”. Heine and Song try to deal with this problem by proposing “grammaticalization in a wide sense” where only the parameter of extension and the principle of unidirectionality are relevant. They claim that lateral shift falls within the domain of grammaticalization in a wide sense. However, this position is problematic in that extension is involved in most, if not all, cases of language change. The relevance of unidirectionality is equally controversial in that Joseph shows the results of verbal inflectional change in Ancient and Modern Greek are neither less nor more grammatical. Similarly, it is not clear if there is a change in the grammatical degree when the erstwhile French second person plural vous and German third person plural Sie are used as a second person singular form. Additionally, these cases also pose difficulty for grammaticalization parameters conceptualized in other ways in that the pronominal paradigm has become looser and the speaker’s choice has expanded with the addition of a new item, Lehmann’s (1995) parameters of paradigmaticity and paradigmatic variability. 

How then should one account for plurification and shift in deixis if they are not grammaticalization? I propose that they are socio-pragmatic strategies motivated by politeness in the sense of Brown and Levinson (1987). The speaker’s displacement of semantic features symbolically manipulates conceptual distance toward the second person, yielding opposite politeness effects with respect to the addressee. On the one hand, the use of second person plural forms and third person (singular and plural) forms for the addressee signifies increasing conceptual distance between the speaker and the addressee, creating what Brown and Levinson call negative politeness. On the other, the use of first person (singular and plural) forms for the addressee decreases conceptual distance, which produces their positive politeness. I finally show examples of each strategy from a variety of languages, and argue that these uses are not necessarily semanticized (e.g. English medical-we), but may potentially conventionalize, leading eventually to semanticization. 

References 

Cooke, Joseph Robinson. 1968. Pronominal reference in Thai, Burmese, and Vietnamese. Berkeley: University of California Press. 

Hagège, Claude. 1993. The language builder: An essay on the human signature in linguistic morphogenesis. Amsterdam/Philadelphia: John Benjamins. 

Heine, Bernd, and Kuteva, Tania. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press. 

Heine, Bernd and Song, Kyung-An. 2010. On the genesis of personal pronouns: Some conceptual sources. Language and Cognition 2: 117-147. 

Heine, Bernd and Song, Kyung-An. 2011. On the grammaticalization of personal pronouns. Journal of Linguistics 47: 587-630. 

Lehmann, Christian. 1995. Thoughts on grammaticalization. München: LINCOM Europa. 

Joseph, Brian. 2005. How accommodating of change is grammaticalization? Logos and Language 6: 1-8. 

Zubin, David, Chun, Soon Ae, and Li, Naicong. 1990. Misbehaving reflexives in Korean and Mandarin. Berkeley Linguistics Society:338-352

ABSTRACT.
The standard view on weather verbs in Modern Icelandic is that they are “no-argument predicates” (Thráinsson 2007:267). This was essentially also the view of earlier scholarship on Old and Modern Icelandic (Nygaard 1905:6–7, Smári 1920:21). In this paper we argue against this view and show that weather verbs in Modern Icelandic do in fact have arguments, both quasi-arguments (i.e. non-referential arguments, Chomsky 1981:325, Rizzi 2000:43–44) and overt NPs, occurring in the nominative, accusative or dative case (1). Moreover, we show that this state of affairs involves a continuity from Old to Modern Icelandic.

(1) a. Vindurinn kólnar. b. Vindinn hvessir. c. Eldi rignir.
the-wind.nom gets-cold the-wind.acc sharpens fire.dat rains
‘The wind gets cold.’ ‘It gets windy.’ ‘It rains fire.’

Note that for the examples in (1), in all instances the relevant verb can occur without an NP; compare (1c) to (2b). The expletive það, attested from the 16th century onwards, is purely a “placeholder” and not an argument. It occurs clause-initially in certain types of declarative clauses in the absence of another phrase (2a) and is not found elsewhere (2b–c). Since það does not have an argument status we will not discuss it further.

(2) a. Það rignir mikið í dag. b. Í dag rignir (*það) mikið. c. Rignir (*það) mikið í dag?
it rains much today today rains it much rains it much today
‘Today it rains a lot.’ ‘Today it rains a lot.’ ‘Does it rain a lot today?’

Quasi-arguments with weather verbs are mostly covert (unexpressed), but they can also be overtly expressed by the non-referential “weather-hann” (homonymous with the masculine singular pronoun hann ‘he’), which emerged in the 17th–18th centuries and today mainly occurs in fixed expressions with an archaic flavor. Weather-hann inverts with the verb, which is different from the behavior of the expletive það.

(3) a. Hann rignir mikið í dag. b. Í dag rignir hann mikið. c. Rignir hann mikið í dag?
he rains much today today rains he much rains he much today
‘It rains a lot today.’ ‘Today it rains a lot.’ ‘Does it rain a lot today.’

The fact that weather verbs occur with NP arguments and the quasi-argument hann in Icelandic constitutes clear evidence, hitherto overlooked, for the claim that these verbs can have arguments. Moreover, weather verbs in Icelandic can occur in control infinitives (4a), just as in English (4b), where this fact has been considered as evidence for the quasi-argument status of weather-it (e.g. Chomsky 1981:323–325).

(4) a. Stundum rignir eftir að hafa snjóað. b. It sometimes rains after snowing.
sometimes rains after to have.inf snow

By contrast, impersonal passives in Icelandic cannot occur in control infinitives, presumably because they do not have an (external) argument.

(5) *Í samkvæminu var sungið án þess að vera dansað.
at the party was sung without it to be.inf danced
Intended meaning: ‘At the party people sang without dancing.’

A diachronic study based on various corpora, including Icelandic Parsed Historical Corpus (IcePaHC) (Wallenberg et al. 2011), Icelandic Text Collection (ÍT) and Dictionary of Old Norse Prose (ONP), shows that the lexical items constituting weather verbs are essentially the same in Modern and Old Icelandic. In Old Icelandic weather verbs could also occur with overt argument NPs

(6a). Furthermore, it is reasonable to assume that weather verbs in Old Icelandic occurred
with a quasi-argument, even if this was always covert

(6b); after all, Old Icelandic was an argument-drop language (Sigurðsson 1993, Kinn, Rusten & Walkden 2014), so an unexpressed argument with weather verbs is expected.

(6) a. Þá lægði storminn. b. Áður en þeir fóru á brott snjóaði mjög á fjöll.
then subdued the storm.acc before than they went away snowed much on mountains
‘Then the storm subdued.’ ‘Before they left it snowed heavily on the mountains.’

Thus, there is no qualitative difference in weather verbs between Old and Modern Icelandic (although there may be a quantitative difference). The only exception is the fact that the overt non-referential quasi-argument weather-hann emerged in the language in the 17th–18th centuries. Originally, weather-hann was a referential pronoun which was later reanalyzed as a quasi-argument. It is clear that the emergence of weather-hann coincided with the loss of null arguments in Icelandic, which occurred along with several other syntactic changes in early Modern Icelandic (mid-16th century) and was completed in the 18th–19th centuries.

The emergence of weather-hann is expected; in fact, it has parallels in Faroese and some Mainland Scandinavian dialects (Thrainsson et al. 2012, Bandle 1973). However, whereas this element always remained rather marginal, the covert quasi-argument persisted throughout the history of Icelandic. So this is the real puzzle: Why does Modern Icelandic, a non-argument-drop language, have covert quasi-arguments with weather verbs at all, rather than an overt element comparable to Mainland Scandinavian det, English it and German es? We propose that the covert quasi-arguments have not been replaced by overt elements altogether because they are not pronouns, but belong to a different category as non-referential arguments.
In summary, our research shows that weather verbs in Icelandic can have arguments, both full NPs and (overt and covert) quasi-arguments and that this situation reflects a continuity from Old to Modern Icelandic. Thus, even though we have been able to detect some surface changes in weather verbs in the history of Icelandic, notably the emergence of weather-hann and the placeholder það, we conclude that they are basically the same today as in Old Icelandic.

References
Bandle, Oskar. 1973. Die Gliederung des Nordgermanischen. Beiträge zur nordischen Philologie 1. Basel: Heilbing and Lichtenhahn. [Reprinted 2011, Tübingen: Francke].
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris.
Kinn, Kari, Kristian A. Rusten and George Walkden. 2016. Null subjects in early Icelandic. Journal of Germanic Linguistics 28(1):31-80.
Rizzi, Luigi. 2000. Comparative Syntax and Language Acquisition. London and New York: Routledge.
Sigurðsson, Halldór Ármann. 1993. Argument-drop in Old Icelandic. Lingua 89:247–280.
Thráinsson, Höskuldur. 2007. The Syntax of Icelandic. Cambridge: Cambridge University Press.
Thráinsson, Höskuldur, Hjalmar P. Petersen, Jógvan í Lon Jacobsen and Zakaris Svabo Hansen. 2012. Faroese: An Overview and Reference Grammar. Second edition. Tórshavn and Reykjavík: Faroese University Press and Linguistics Institute, University of Iceland.

ABSTRACT. Noun phrase word order in Old Swedish – from pragmatic fronting to determiner-first word order

Viking Age Scandinavian before c.1000 is a language with no formal means to express definiteness or indefiniteness. The mandatory distinction between indefinite and definite noun phrases develops during the following centuries. This development comprises the emergence of a definite infection of nouns as well as free articles, both an indefinite and a definite. At the same time, the noun phrase word order undergoes a considerable change, resulting in the modern pattern, where noun phrases are (essentially) headed by determiners. In contrast, it seems as “noun first” was a widely used alternative for the Vikings, as we find numerous instances of phrases such as stæin þenna (acc.), lit. “stone this”, or faður sinn goðan (acc.), lit. “father his good” in the runic inscriptions.

Although it is likely that the word order change is a consequence of the language developing (in)definiteness, this circumstance has not, so far, been paid much interest by scholars. However, the shifting word order may reveal what is going on on the abstract level.

A question that has recently been brought to the fore within the minimalist enterprise is whether noun phrases are DPs also in article-less languages. The DP analysis, proposed for argumental noun phrases in Abney 1987, was earlier taken more or less for granted in article-less languages as well, but Željko Bošković (2012) argues on empirical grounds that languages with and without articles differ on specific points related to the presence/absence of a DP on top of the NP. In addition, when testing Old Norse on some of the criteria put forward by Bošković, Lander & Haegeman (2013) conclude that this language behaves like an NP language.

On the other hand, it is not evident how a fairly simple NP structure makes sense when considering the word order variation of the Scandinavian Viking Age noun phrases. Possessive attributes, for instance, could appear before as well as after the noun; cf. Karls sonr (nom.) “Karls son” – son sinn (acc.) “his son” (lit. son his). The same kind of variation applied to adjectival attributes; cf. goðan dræng (acc.) “a good man” – dræng goðan. Furthermore, it is not evident how an NP language becomes a DP language when developing (in)definiteness.

These problems are dealt with in Stroh-Wollin 2015. Here, it is argued that Viking Age Scandinavian noun phrases are EPs (with E as a dummy for “edge”), i.e. headed by a functional projection other than the DP on top of the NP (= the N domain). Developing an idea, put forward by Börjars, Harries & Vincent (2016), of a “discourse prominent” first position in Old Scandinavian noun phrases, the EP is taken to host either a prominent modifier or the noun in the unmarked case (in both cases moving there from below). This suggestion was also supported empirically by an investigation on noun phrases searched for in the Scandinavian Runic Text Database.

As for the further development of the noun phrase, it is suggested in Stroh-Wollin 2015 that the noun phrase of the modern language has a double–DP structure, with a big DP on top of a little dP, and, furthermore, that the transition from the EP structure to the DP structure comes in at least two steps. In a first step, an optional dP is sandwiched in between EP and NP. Even though the EP-(dP)-NP structure prevails for a long time, this first step is important since it sets the ball rolling. Possibly, there is only one more step; it may be that the dP becomes obligatory and the EP is reinterpreted as a DP in one sweep. The first step was supported empirically by an investigation of non-contrastive “this- demonstratives”, that were normally post-posed in the Viking Age inscriptions, but appear pre-nominally to an increasing degree from the 12th century and on.

In my talk, I will take the model in Stroh-Wollin 2015 as my theoretical point of departure and present an empirical investigation on the word order in Viking Age Scandinavian and medieval Swedish noun phrases containing various kinds of modifiers: demonstratives, possessive attributes, adjectival attributes and numerals. It will be shown that pragmatic fronting is still used to some extent in the earliest medieval manuscripts. Possessives, for instance, which are obligatorily pre-posed in modern Swedish, appear most often post-nominally in the early provincial laws; see e.g. the first instance of “her”, sinu (neutr. sing. dat.) in (1) below (from Äldre Västgötalagen, c.1225). However, the second instance of the same word, sinum (neutr. plur.dat.), is pre-posed, which is a way to stress that it is question of the woman’s “own” children.

(1) Sva ær an kona firiger stiupbarni sinu. vill sinum barnum arf vnnæ. so is if woman deserts stepchild her wants her children heritage grant “So is if a woman deserts her stepchild (and) wants her (own) children to have the heritage.”

Numerals, on the other hand, most often appear pre-nominally (for obvious reason) already in the earliest manuscripts. However, if the number is less important than the lexical content of the noun, then a numeral may follow the noun; see e.g. tver (mask. plur. nom.) “two” in (2) below (from the same law as before). The next word, ena (fem. sing. acc.) “one”, is also a numeral, but pre-posed, as it stresses that it is question of one and the same woman.

(2) Egho bröþær tver ena kono. […] þæt ær firnærværk owns brothers two one woman. that is incest “If two brothers have (a relation with) one (and the same) woman, that is incest.”

A more confusing result concerns the adjectival attributes, which seem to take a pre-nominal posi-tion early on. In this context, it may be of importance that noun phrases with both possessives and adjectival modifiers show three predominant diachronic patterns. A Viking Age phrase such as faður sinn goðan (acc.) represents the first pattern: noun–possessive–adjective. Early medieval texts, on the other hand, often have adjective–noun–possessive, whereas the modern pattern is possessive–adjective–noun. (Delsing 1994) Possible implications of the second pattern will be further discussed in my talk.

References Abney, S.P. (1987): The English Noun Phrase and Its Sentential Aspect. PhD dissertation, MIT. Äldre Västgötalagen. Accessed at: Fornsvenska textbanken: http://project2.sol.lu.se/fornsvenska/ Börjars, K., P. Harries & N. Vincent (2016): “Growing Syntax: The development of a DP in North Germanic”. Language 92:1, e1-e37. Bošković, Ž. (2012): “On NPs and clauses”. G. Grewendorf & T.E. Zimmermann (eds.), Discourse and Grammar: From Sentence Types to Lexical Categories. Walter de Gruyter, Boston. 179–242. Delsing, L.-O., 1994: ”Hans siukt ben” – Om starka och svaga adjektiv i fornsvenskan. Nils Jörgensen, Christer Platzack & Jan Svensson (eds.), Språkbruk, grammatik och språkförändring: En festskrift till Ulf Teleman 13.1.1994, Institutionen för nordiska språk, Lunds universitet. 99–107. Lander, E. T. & L. Haegeman (2013): “Old Norse as an NP language: With observation on the Common Norse and Northwest Germanic runic inscriptions”. Transactions of the Philological Society. Vol. 00. 1–40. Scandinavian Runic Text Database. (Samnordisk runtextdatabas.) Department of Scandinavian Languages, Uppsala University. www.nordiska.uu.se/forskn/samnord.htm. Stroh-Wollin, U. (2015): “Understanding the gradual development of definiteness marking: the case of Swedish”. Working Papers in Scandinavian Syntax 95. 11–32.

ABSTRACT. One of the main differences between Greek-Latin parts of speech theory and the parts of speech theory of so-called traditional linguistics lies in the presence of the adjective as an independent word class, but hitherto the literature on the topic has not discussed this question specifically. The paper therefore analyses the definitions of the noun, the verb and the epithet-adjective class from Dionysius Thrax to the Port Royal grammar with the aim of demonstrating that the birth of the adjective as an independent word class, as well as the stabilization of the labels nomen substantivum and nomen adjectivum with reference to the common noun and the adjective, depend on the reinterpretation of Aristotle’s metaphysics in the light of Neo-platonic ontology in the Middle Ages.

ABSTRACT. In this paper we study the split between case marking and referential properties for a class of arguments that, we argue, qualify as non-canonical subjects in Ancient Greek. These experiencer arguments are marked with an oblique case but have some subject properties: notably, they can be the referential antecedent of the subject of an adjunct participle. A vast number of languages, as is well known, attest constructions in which an argument not formally encoded as a subject (as far as, e.g., morphological case, verb agreement patterns, word order are concerned) nevertheless shares behavioural and semantic properties with "canonical" subjects (i.e. with arguments formally encoded as subjects). Phenomena of this type have increasingly attracted, in recent years, specialists of historical and comparative linguistics. Investigations of non-canonical subjects have been carried out on different ancient and modern languages, producing considerable results, but also opening up many unresolved questions (cf. the recent surveys in Aikhenvald, Dixon, Onishi eds., 2001, Bhaskararao & Subbarao eds., 2004, Eythórsson & Barðdal 2005, Seržant & Kulikov eds., 2013). A crucial issue concerns the definition of subject behavioural properties (Keenan 1976): control phenomena, coreferential deletion, constraints on reflexivization etc. have proven to be valid criteria for subjecthood in a variety of languages; but, of course, they do not apply to all languages (and anyway not in the same way and with the same manifestations). Therefore, a careful investigation on individual languages is needed prior to possibly identifying candidates for non-canonical subjecthood. For some ancient Indo-European languages - Ancient Greek among them - the research has not gone beyond a pioneering stage, and an in-depth definition of criteria for subjecthood is still missing. The present study contributes to this issue, by focussing on a phenomenon of Ancient Greek hitherto uninvestigated within the landscape of canonical/non-canonical subjecthood, namely argument coreference and agreement patterns in participial adjunct clauses. In the unmarked case, in participial adjunct clauses the participle agrees in case (as well as in gender and number) with its referential anchor in the main clause. So, the subject of a nominative participle usually corefers with the nominative subject (expressed or pro) of the main clause: cf. (1), with the subject of boulómenos 'wishing' being referentially anchored to the nominative (pro) subject of ékploun poieîtai 'makes a sailing out':

(1) βουλόμενοςi τὰ ἐπὶ τούτοις παρασκευάζειν boulómenos tà epì toútois paraskeuázein [..] wish- PTCP.NOM.M.SG. DET-ACC.N.PL. after DEM-DAT.PL. prepare- INF.PRES. ἔκπλουν ποιεῖται (proi) ékploun poieîtai sailing out- ACC.SG. do-IND.PRES.3SG. 'wishing to make preparations for the next actions, he sales out' (T. 1.65.1)

There are, however, also more complex cases, like (2), which represent the topic of our research. In (2) the subject of the nominative boulómenos is coreferent with the dative argument of édoxen '(it) seemed', i.e. autō̂i 'to him' (and not with its nominative (raised) argument, Aráspas!):

(2) Βουλόμενοςi δὲ κατάσκοπόν τινα πέμψαι boulómenos dè katáskopón tina pémpsai wish-PTCP.NOM.M.SG. PTC spy-ACC.M.SG. someone-ACC.SG. send-INF.AOR. ἐπὶ Λυδίας [..] ἔδοξεν αὐτῷi ἐπιτήδειος εἶναι epì Ludías édoxen autō̂ii epitḗdeios eînai into Lydia-GEN.SG. seem-IND.AOR.3SG. he-DAT.SG. apt-NOM.M.SG. be-INF.PRES. Ἀράσπας ἐλθεῖν ἐπὶ τοῦτο Aráspas eltheîn epì toûto Araspas-NOM.M.SG. go-INF.AOR. for DEM-ACC.SG 'Now, wishing to send somebody as a spy into Lydia [..] Araspas [..] seemed to him to be the proper person to go on this mission.' (X. Cyr. 6.1.31)

Examples like these, with a participle showing up in the nominative, despite being co-referent with a non-nominative argument of the main clause, have occasionally been pointed out in historical grammars, and treated as anacolutha. Our study aims at investigating this phenomenon in depth, collecting data from an extensive corpus of Ancient Greek and proposing a more accurate description in syntactic and semantic terms. We note, in particular, that this phenomenon typically involves oblique experiencers, i.e. a class of arguments which frequently appear associated with some subject properties (Luraghi 2010, Barðdal et al. 2012, Dahl & Fedriani 2012, Verbeke, Kulikov, Willems 2015). Also in view of this, we suggest that the oblique argument in this construction displays behavioural subject properties and that the phenomenon of coreference and case mismatch with the subject of participial clauses may provide a valid test for non-canonical subjecthood in Ancient Greek. Bibliographical references: Aikhenvald, A.Y., Dixon, R.M.W., Onishi, M. (2001), (eds.), Non-canonical Marking of Subjects and Objects, Amsterdam/Philadelphia: Benjamins. Bhaskararao, P. & Subbarao, K. W. (2004), (eds.), Non-nominative Subjects. Volume 1 and 2. Amsterdam/Philadelphia: Benjamins. Barðdal, J., Smitherman, Th., Bjarnadóttir, V., Danesi, S., Jenset, G.B., McGillivray, B. (2012), "Reconstructing constructional semantics. The dative subject construction in Old Norse-Icelandic, Latin, Ancient Greek, Old Russian and Old Lithuanian", Studies in language 36, 511-547. Dahl, E. & Fedriani. C. (2012), "The Argument Structure of Experience: Experiential Constructions in Early Vedic, Homeric Greek and Old Latin", Transactions of the Philological Society 110(3): 342–362. Eythórsson, Th. & Barðdal, J. (2005), "Oblique subjects: A common Germanic inheritance", Language 81, 824–881. Keenan, E. (1976), Towards a universal definition of ‘subject’. In: Li, Ch. N. (ed.), Subject and topic, New York: Academic Press, 303-34. Luraghi, S. (2010), Experiencer predicates in Hittite. In R.I. Kim, E. Riecken, N. Oettinger & M.J. Weiss (eds.), Ex Anatolia lux, Ann Arbor: Beech Stave Press, 249–264. Seržant, I. & Kulikov, L. (2013), (eds.), The Diachronic Typology of Non-Canonical Subjects. Amsterdam/Philadelphia: Benjamins. Verbeke, S., Kulikov, L., Willems, K., (2015), "Oblique case-marking in Indo-Aryan experiencer constructions: historical roots and synchronic variation", Lingua 163, 23-39.

ABSTRACT. This paper aims at demonstrating how information on the paths of semantic extensions undergone by content words may be incorporated into semantic maps. For this purpose, particular changes that affected the meanings of words in the course of the Ancient Egyptian and of the Ancient Greek language history will be investigated. The semantic map model was initially created in order to describe the polysemic patterns of grammatical morphemes (e.g., Haspelmath, 2003). However, recent studies by François (2008), Perrin (2010), Wälchli and Cysouw (2012), and Georgakopoulos et al. (2016) have drawn attention to the lexical domain, showing that the model can be extended to lexical items. Although it may be stated that ‘the best synchronic semantic map is a diachronic one’ (van der Auwera 2008: 43), the big bulk of research has been adopting a strictly synchronic perspective and, importantly, the limited research that has added the diachronic dimension, has focused almost exclusively on the grammatical domain (van der Auwera & Plungian, 1998; Narrog, 2010; Luraghi, 2014; cf. the articles in Juvonen and Koptjevskaja-Tamm, 2016). Situated within the emergent tradition of the semantic map methodology, this paper addresses the question of how the tool makes predictions about long-term language change at the lexical level. We analyze the diachronic evolution of the polysemy network of lexemes in order to produce ‘dynamicized semantic maps’ (Narrog & van der Auwera, 2011) of lexical items. More specifically, we study 20 concepts from the semantic domains of natural objects and phenomena (e.g., FOG, EARTH), position and movement (e.g., WALK, FALL), body sensations and activities (e.g., BREATHE, EAT), and from more abstract domains (e.g., TIME). The empirical basis of the paper consists of two languages with significant diachronic material: Ancient Greek (8th – 1st c. BC) and Ancient Egyptian (26th c. BC – 10th c. AD). The data are extracted from dictionaries, grammars, and the Perseus digital library (http://www.perseus.tufts.edu/hopper/) for Ancient Greek, and from the Thesaurus Linguae Aegyptiae (http://aaew.bbaw.de/tla/), the Ramses corpus (http://ramses.ulg.ac.be), and Coptic etymological dictionaries for Ancient Egyptian. In the design of our sample of concepts, one criterion is that the concept appears in databases of basic vocabulary, such as the Swadesh list (see, e.g., Swadesh 1950), the Leipzig-Jakarta list (Haspelmath & Tadmor, 2009) or in CLICS (List et al., 2014), which is an online database containing tendencies of meaning associations. In CLICS, concepts are represented as nodes in the network and instances of polysemy are visualized as links between the nodes. Fig. 1 exemplifies how the diachronic dimension of meaning extension may be added to such a network. On the basis of a diachronic analysis of TIME in Ancient Greek (lexical unit: hṓra), which reveals that the meaning ‘time’ is historically prior to the meaning ‘hour,’ we may add a directed arrow representing directionality of change. Similarly, Ancient Egyptian data seems to point to an extension of the polysemy network of the lexical unit tr – originally meaning ‘time,’ ‘moment in time’ – to ‘season’ (cf. Coptic ⲧⲏ tê ‘time, season’).

season <-- time --> hour Fig. 1 | Polysemy network of time with directionality of meaning extension (cf. CLICS)

However, historical priority is not a sufficient criterion for an arrow to be added. Rather, one should be able (1) to show that meaning extensions have clear motivations and (2) to identify the bridging contexts allowing such extensions of the polysemy networks. As for (1), we suggest describing the cognitive (e.g., metaphor, metonymy, etc.) and the cultural factors that lie behind the observed evolutions. For example, in the case of the Greek concept TIME, one could establish a metonymic motivation between TIME and HOUR, which arises due to the correlation between the canonical time periods and the time these take to unfold. Regarding (2), in-depth corpus analysis is needed in order to identify the contextual parameters and the period during which the changes actually take place. The present study will provide answers to the question of the directionality of change in the lexicon of two particular languages, namely Ancient Greek and Ancient Egyptian. However, our expectation is that by looking at diachrony in this fashion, significant dimensions of directionality of change with cross-linguistic extensions can be revealed.

References van der Auwera, J. (2008). In defense of classical semantic maps. Theoretical Linguistics, 34(1), 39–46. van der Auwera, J., & Plungian, V. (1998). Modality’s semantic map. Linguistic typology, 2, 79–124. François, A. (2008). Semantic Maps and the Typology of Colexification: Intertwining Polysemous Networks across Languages. In: M. Vanhove (Ed.), From Polysemy to Semantic Change. Towards a Typology of Lexical Semantic Associations (pp. 163–215). Amsterdam/Philadelphia: John Benjamins. Georgakopoulos, Th., Werning, A.D., Hartlieb, J., Kitazumi, T., van de Peut, E.L., Sundermeyer, A., & Chantrain, G. (2016). The meaning of ancient words for ‘earth’. An exercise in visualizing colexification on a semantic map. eTopoi. Journal for Ancient Studies 6, 1–36. Haspelmath, M. (2003). The geometry of grammatical meaning: Semantic maps and cross-linguistic comparison. In: M. Tomasello (Ed.), The new psychology of language, Vol. 2 (pp. 211–242). Mahwah/ New Jersey: Lawrence Erlbaum Associates. Haspelmath, M., & Tadmor U. (Eds.). (2009). Loanwords in the world’s languages: A comparative handbook. Berlin: De Gruyter. Juvonen, P, & Koptjevskaja-Tamm, M. (Eds.) (2016). The Lexical Typology of Semantic Shifts. Berlin: De Gruyter. List, J.-M., Mayer, Th., Terhalle, A., & Urban, M. (2014). CLICS: Database of Cross-Linguistic Colexifications. Marburg: Forschungszentrum Deutscher Sprachatlas (Version 1.0, online available at http://CLICS.lingpy.org, accessed on 2016-27-10). Luraghi, S. (2014). Plotting Diachronic Semantic Maps: The Role of Metaphor. In: S. Luraghi & H. Narrog (Eds.), Perspectives on Semantic Roles (pp. 99-150). Amsterdam: Benjamins. Narrog, H. (2010). A Diachronic Dimension in Maps of Case Functions. Linguistic Discovery, 8(1), 233–254. Narrog, H., & van der Auwera, J. (2011). Grammaticalization and Semantic maps. In: H. Narrog & B. Heine (Eds.), The Oxford Handbook of Grammaticalization (pp. 318-327), Oxford: OUP. Perrin, L-M. (2010). Polysemous qualities and universal networks, invariance and diversity. Linguistic Discovery, 8(1), 259–280. Swadesh, M. (1950). Salish Internal relationships. IJAL, 16, 157–167. Wälchli, B., & Cysouw, M. (2012). Lexical typology through similarity semantics: Toward a semantic map of motion verbs. Linguistics, 50(3), 671–71.

ABSTRACT. It is a well-established fact in the field of historical and comparative Indo-European linguistics that several ancient Indo-European languages exhibit a construction where the subject-like argument is not canonically marked in the nominative, but occurs with an accusative, dative or genitive marking (Conti 2008, 2010, Luraghi 2010, Barðdal et al. 2012, 2013, 2016, Dahl & Fedriani 2012, Matasović 2013, Danesi 2014, Fedriani 2014, Viti 2016a, 2016b, inter alia).

The existence of such argument structure constructions raises the question of how old such structures are in the respective language branches of the Indo-European languages: if it is archaic, and whether or not it can be reconstructed back to the proto-stage of these specific branches. Our goal here is to add to the growing body of evidence showing that oblique subjects can be reconstructed for Proto-Germanic, which in turn raises interesting questions with regard to the Proto-Indo-European pre-stage of the relevant verbal roots.

In a recent study Barðdal and Eythórsson (2012) have shown that in cases where a lexico-syntactic match extending over several ancient daughter languages can be found, projecting the argument structure back into the proto-stage of the language family certainly is both cogent and compelling. We illustrate several more of these lexico-syntactic matches and show how taking the non-canonical argument structure of these verbs into account may shed light on their Proto-Indo-European etymology and development. As an example, consider the Proto-Germanic *þreutan- ‘to strain, exhaust, grieve, vex’ below:

(1a)      Old High German

des                  ne        mag                             sie                   irdriêzzen

this.GEN.        NEG    may.3SG                    them.ACC       vex.INF

‘they should not be vexed by this’ (N. Ps. 307.15)

(1b)      Middle Dutch

            Also                 seer                 verdriet           mi                    der                tijd

            just.as              much               vex.3SG          mi.ACC.SG     the.GEN.SG time

            ‘just as much I am grieved by the time’ (Hs. Moll 5, 4c., Holland, 1440-1460)

There is no doubt that the original PIE meaning of this stem is ‘to shove, push’ (cf. LIV 651; IEW: 1095–6). The same etymon is also present in Latin trūdō ‘to thrust, push’ and Middle Welsh cythrudd ‘to vex’ (Proto-Celtic *kom-trūdit), the latter arguably deriving from PIE *troud-éie- (= caus. “cause thrust, push” or intensive). Indeed, one even comes close to a lexico-syntactic match when considering the etymologically matching noun kythrud ‘distress’ and its accompanying experiential construction in the Middle Welsh collocation ny bydd kythrud gynnyt ‘you will not be distressed’.

            How, then, can one account for a development involving the agentive ‘shove, push’ changing into the attested ‘to strain, exhaust, grieve, vex’ meaning in Germanic? This is where argument structure constructions involving oblique subjects enter the equation, as their meaning is always less agentive than that of corresponding nominative subject constructions (cf. Bauer 2000, Barðdal 2004, Barðdal & Eythórsson 2009, Danesi 2014). Assuming that the stem *treud- could also occur in an oblique subject construction at a stage before Proto-Germanic and Proto-Celtic explains how an agentive meaning can develop into the non-agentive, or experiencer, counterpart. On this basis, we put forward a reconstruction of the PIE root *treud- together with an argument structure involving non-canonical case marking, as in (2):

(2)        Proto-Indo-European Reconstruction

*h₁me     tréudeti                                

*mike      þreutedi

 me.OBL shoves.3SG

‘I feel shoved’ > Germanic ‘I am strained, vexed’, Celtic ‘I am vexed, distressed’

We present several examples of this type, from which there can be no doubt about the existence of non-canonical subject marking in Proto-Germanic, as argued by Barðdal & Eythórsson (2012). High-quality lexico-syntactic matches between Germanic and other Indo-European branches are hard to find, but they certainly exist. Such matches form the basis for the reconstruction of oblique subject constructions between Proto-Germanic and Proto-Indo-European. As such, they are the key to understanding many otherwise obscure semantic shifts documented across the Indo-European phylum. We believe that such lexico-syntactic matches are not confined to Germanic, but are likely to be found for other Indo-European branches as well.

References

Barðdal, Jóhanna. 2004. The Semantics of the Impersonal Construction in Icelandic, German and Faroese: Beyond Thematic Roles. Focus on Germanic Typology, ed. by Werner Abraham, 105–137. Berlin: Akademie Verlag.

Barðdal, Jóhanna, Carlee Arnett, Stephen Mark Carey, Thórhallur Eythórsson, Gard B. Jenset, Guus Kroonen & Adam Oberlin. 2016. Dative Subjects in Germanic: A Computational Analysis of Lexical Semantic Verb Classes Across Time and Space. STUF: Language Typology and Universals 69(1).49–84.

Barðdal, Jóhanna, Valgerður Bjarnadóttir, Serena Danesi, Tonya Kim Dewey, Thórhallur Eythórsson, Chiara Fedriani & Thomas Smitherman. 2013. The Story of ‘Woe’. Journal of Indo-European Studies 41(3–4).321–377.

Barðdal, Jóhanna & Thórhallur Eythórsson. 2009. The Origin of the Oblique Subject Construction: An Indo-European Comparison. Grammatical Change in Indo-European Languages, ed. by Vit Bubenik, John Hewson & ’öSarah Rose, 179–193. Amsterdam: John Benjamins.

Barðdal, Jóhanna & Thórhallur Eythórsson. 2012. "Hungering and Lusting for Women and Fleshly Delicacies": Reconstructing Grammatical Relations for Proto-Germanic. Transactions of the Philological Society 110(3).363–393.

Barðdal, Jóhanna, Thomas Smitherman, Valgerður Bjarnadóttir, Serena Danesi, Gard B. Jenset & Barbara McGillivray. 2012. Reconstructing Constructional Semantics: The Dative Subject Construction in Old Norse-Icelandic, Latin, Ancient Greek, Old Russian and Old Lithuanian. Studies in Language 36(3).511–547.

Bauer, Brigitte L.M. 2000. Archaic Syntax in Indo-European: The Spread of Transitivity in Latin and French. Berlin: Mouton de Gruyter.

Conti, Luz. 2008. Synchronie und Diachronie des altgriechischen Genitivs als Semisubjekt, Historische Sprachforschung 121.94–113.

Conti, Luz. 2010. Análisis del dativo en construcciones impersonales: Los conceptos de sujeto y de semisujeto en griego antiguo. Emerita LXXVIII 2.249–273.

Dahl, Eystein & Chiara Fedriani. 2012. The Argument Structure of Experience: Experiential Constructions in Early Vedic, Homeric Greek and Early Latin. Transactions of the Philological Society 110(3).342–362.

Danesi, Serena. 2014. Accusative Subjects in Avestan: ‘Errors’ or Noncanonically Marked Arguments. Indo-Iranian Journal 57(3).223–260.

Fedriani, Chiara. 2014. Experiential Constructions in Latin. Brill: Leiden.

IEW = Indogermanisches Etymologisches Wörterbuch. 1959. Julius Pokorny. Tübingen: Francke Verlag.

LIV = Lexikon der indogermanischen Verben: die Wurzeln und ihre Primärstammbildungen. 2001. Helmut Rix, Martin Kümmel, Thomas Zehnder, Reiner Lipp & Brigitte Schirmer (eds.). Reichert: Wiesbaden.

Matasović, Ranko. 2013. Latin paenitet me, miseret me, pudet me and active clause alignment in PIE. Indogermanische Forschungen 118.93–110.

Viti, Carlotta. 2016a. The morphosyntax of experience predicates in Tocharian. Cahiers de Linguistique Asie Orientale (CLAO) 45(1): 26–70.

Viti, Carlotta. 2016b. Areal distribution of argument marking of Indo-European experience predicates. Journal of Indo-European Studies 44: 1–84.

ABSTRACT. Crosslinguistically, languages tend to mark grammatical relations either via case, word order or agreement (cf. Kiparsky 1988). Icelandic is known to have both a relatively fixed word order and a rich case morphology with complex agreement patterns (see e.g. Thráinsson 2007) and thus constitutes an interesting object of study in this respect. This paper focuses on the marking of subjects in Icelandic and presents results from ongoing corpus investigations using the Icelandic Parsed Historical Corpus (IcePaHC, Wallenberg et al. 2011).

Icelandic is standardly assumed to be an SVO-language exhibiting the verb-second constraint (e.g. Thráinsson 2007). Nevertheless, Icelandic allows V1 in matrix declarative sentences, mainly confined to narrative texts (so-called ‘narrative inversion’). According to Sigurðsson (1990), V1 structures appear throughout the history of Icelandic and are still common in the present-day language. Butt et al. (2014) present a corpus linguistic study of V1 matrix declarative sentences in IcePaHC and confirm that V1 constructions are found in all attested stages of Icelandic and mainly appear in narratives, but they also show a marked decrease as of 1900. One explanation for the decrease in V1 post-1900 would be a simultaneous increase in the overt expletive subject það (as suggested in Franco 2008: 152). Það is restricted to the clause-initial prefinite position: one would expect that the occurrence of expletive það would render some of the old V1 structures lacking an overt subject (1a) into V2 constructions (1b): 1a) _______ Var fátt manna heima. ØEXPL was few men.GEN at-home ‘There were few men at home.’ (1350.FINNBOGI.NAR-SAG,655.1696) b) Það var fátt manna heima. EXPL was few men.GEN at-home ‘There were few men at home.’ However, this cannot fully account for the decrease in V1 as the V1 constructions in Butt et al. (2014) were not solely confined to clauses with ‘null’ expletives in the older texts.

Schätzle et al. (2015) used IcePaHC to investigate the diachrony of dative subjects in Icelandic. They show that while dative subjects generally appear throughout the history of Icelandic, their distribution has changed significantly over time (contra e.g. Barðdal & Eythórsson 2009, Barðdal et al. 2012). Specifically, an increase in the frequency of dative subjects was found, resulting from an increasingly systematic association of dative case with experiencers and goals, and driven by a striking rise of dative subjects with verbs carrying middle morphology in the data. Interestingly, the observed changes peak around 1900.

Given these facts, it seems reasonable to investigate the interaction between the decrease in V1 c.1900 and the apparently simultaneous reorganization of the case system, the establishment of a fixed prefinite subject position and the requirement for this to be overtly filled (as shown by the increase of expletives). In this paper, we present the results of a corpus study examining more closely the diachrony of V1 constructions, subject position, dative subjects and overt expletive subjects in matrix declarative sentences:

Table 1: Percentages of V1 constructions, prefinite subjects and overt expletive subjects in IcePaHC Time period V1 Prefinite subjects (overall) Prefinite dative subjects Overt expletive subjects (overall) Overt expletive subjects (information structural type) 1150-1350 9% 52% 25% 49% 19% 1351-1550 9% 55% 20% 40% 24% 1550-1750 9% 54% 29% 51% 26% 1750-1900 11% 57% 35% 60% 71% 1900-2008 2% 73% 56% 88% 96% All time periods 8% 58% 34% 58% 60%

With respect to V1, we do indeed find that when compared to all non-V1 matrix declarative sentences, the decrease in V1 as of 1900 is statistically significant (p<0.001). This contrasts starkly with the relative stability of V1 throughout previous periods. Moreover, our results show that V1 sentences do not differ from this pattern within dative subject or middle constructions. Interestingly, comparing the data for prefinite subjects, our results show that the preference for dative subjects to occur in the prefinite position is weaker than that shown for all types of subject overall. The data for all subject types and for dative subjects alike, however, shows an increasing preference for subjects to occur in the prefinite position post-1900, leading us to draw the hypothesis that subjecthood was increasingly marked by word order at this time.

Regarding the connection between V1 and expletive subjects, the decrease in V1 does appear to coincide with an increase in overt expletive subjects post-1900. Moreover, a refinement of the expletive subject data to exclude those constructions with an extraposed clausal subject (e.g. 2a, where overt expletive subjects were already highly frequent in Old Icelandic, see Rögnvaldsson 2002), hence focusing solely on those with a post-verbal logical subject or ‘expletive associate’ (e.g. 2b) shows an even more dramatic increase post-1900. This is relevant to the V1 issue, as expletive subject constructions like those in 2b can be regarded as having an information-structural motivation: they are ‘presentational’ in that they feature a subject-like element occurring postverbally which represents new information (the postverbal elements are almost always indefinite), an information structural property which is generally favoured later on in the clause. Motivated by the increasingly strong association of the prefinite position as a subject position as outlined above, the prefinite position is hence occupied by the expletive subject það. 2a) Það er ráð húsfreyja að taka vel við gestum. (extraposition) EXPL is advice housewives.GEN to take well with guests ‘It is the advice of housewives to receive guests well.’ (1310.GRETTIR.NAR-SAG,.658) b) Það voru liðnir margir dagar síðan sýningar hófust. (information structural type) EXPL were passed many days since exhibitions began ‘Many days had passes since exhibitions began.’ (1985.MARGSAGA.NAR-FIC,.1124)

The conclusions our data allows us to draw about this historical stage of Icelandic are as follows: over time, subjecthood becomes increasingly associated with a particular structural position (clause-initial). This fixing of a structural position for subjecthood coincides with and likely motivates an increase in overt expletive subjects as a way of filling this position, which in turn could explain the observed decrease in V1 constructions.

References Barðdal, J. & Eythórsson, T. (2009). The origin of the oblique subject construction: An Indo-European comparison. In Bubenik, V., Hewson, J. & Rose, S. (eds.): Grammatical Change in Indo-European Languages, 179-193. Amsterdam: John Benjamins. Barðdal, J., Smitherman, T., Bjarnadóttir, V., Danesi, S., Jenset, G. B. & McGillivray, B. (2012). Reconstructing constructional semantics: The dative subject construction in Old Norse-Icelandic, Latin, Ancient Greek, Old Russian and Old Lithuanian. Studies in Language 36(3), 511-547. Butt, M., Bögel, T., Kotcheva, K., Schätzle, C., Rohrdantz, C., Sacha, D., Dehe, N. and Keim, D. (2014). V1 in Icelandic: A multifactorical visualization of historical data. In: Proceedings of VisLR: Visualization as added value in the development, use and evaluation of Language Resources, Workshop at the 9th edition of the Language Resources and Evaluation Conference (LREC 2014), Reykjavik, Iceland. Franco, I. (2008). V1, V2 and criterial movement in Icelandic. Studies in Linguistics 2, 141-165. Kiparsky, P. (1988). Agreement and Linking Theory. Unpublished manuscript. Stanford University. Rögnvaldsson, E. (2002). ÞAÐ í fornu máli og síðar. Íslenskt mál 24, 7-30. Schätzle, C., Butt, M. & Kotcheva, K. (2015). The diachrony of dative subjects and the middle in Icelandic: A corpus study. In: Proceedings of LFG15 Conference, 357-377. Stanford: CSLI Publications. Sigurðsson, H. A. (1990). V1 declaratives and verb raising in Icelandic. In Maling, J. & Zaenen A. (eds.): Modern Icelandic Syntax, 41-69. San Diego: Academic Press. Thráinsson, H. (2007). The Syntax of Icelandic. Cambridge: Cambridge University Press. Wallenberg, Joel, Anton Karl Ingason, Einar Freyr Sigurðsson & Eiríkur Rögnvaldsson. (2011). Icelandic Parsed Historical Corpus (IcePaHC). Version 0.9. http://www.linguist.is/icelandic_treebank.

ABSTRACT. Discussions of comparative reconstruction usually focus on lexicon (e.g. Ross, Pawley & Osmond 2016) or morphology (Berg & Boerger 2011) with more recent forays into syntax (Barðdal & Smitherman 2013; Walkden 2014). There has been very little discussion about when it is possible to reconstruct prosody for unattested protolanguages.

In this paper we argue that, with the right data, it is possible to make meaningful generalizations about the prosodic structure of protolanguages. We illustrate this with a case study that concerns an instance of reconstructed borrowing between two language families of Papua New Guinea: the Ramu languages and the Sogeram languages.

These two families are completely unrelated: the Ramu languages belong to the Lower Sepik–Ramu family and the Sogeram languages belong to the Trans New Guinea family. Comparative evidence suggests that a pragmatic particle *=a was borrowed from Ramu into the common ancestor of three western Sogeram languages: Mand, Nend, and Manat. This particle is an enclitic that attaches to a prosodic phrase and serves one of two functions, each of which has its own intonational properties. The first is to link a dependent clause or other unit to what follows it, as illustrated below with Chini (1) and Nend (2). The second function is exclamative: =a attaches to the end of an utterance to convey different kinds of emphasis, as illustrated with Chini (3) and Manat (4).

The reflexes in Ramu languages and Sogeram languages have a very high degree of functional parallelism, and also have a meaning and a phonological form that is very borrowable (Matras 1998, Fuller 2001). Since the languages are unrelated, it is most plausible to suppose that *=a was borrowed from one family into the other. This clitic is widespread in the Ramu family, being found in languages as far afield as Bore at the mouth of the Ramu River (Parrish 1989), but is restricted to the three western Sogeram languages and absent in the eastern ones which have no known historical contact with Ramu languages. Thus the distributional evidence suggests it was borrowed into Sogeram from Ramu. There is additional evidence that *=a was borrowed into the Sogeram languages, though, which comes in the form of peculiar effects that *=a has had on the lexicon of the languages it was borrowed into.

Proto-Sogeram has been reconstructed in some detail (Daniels 2015). In the Sogeram languages that have borrowed this enclitic, word-final *a was lost from pronouns and verbs but retained in nouns, adjectives, and adverbs. Sound change is not normally sensitive to word class, so this pattern demands an explanation. The explanation can be found in the interaction between the pragmatic and phonological properties of *=a and certain syntactic facts about the Sogeram languages. The Sogeram languages are verb-final, meaning that the verb often occurs at the end of a prosodic unit. They also exhibit frequent zero anaphora, meaning that pronouns are used most often for emphasis and are thus more likely to receive their own intonation contour. Nouns, adjectives, and adverbs, however, are not likely to occur at the end of a prosodic unit.

We thus reconstruct the following scenario. The enclitic *=a was borrowed from a Ramu language into the common ancestor of Mand, Nend, and Manat. After it was borrowed, tokens of the phoneme *a that happened to occur at the end of a prosodic unit were reanalyzed as being tokens of the enclitic *=a. After reanalysis was complete, these words began to occur in non-emphatic contexts without their former final phoneme. So the pronoun *ara ‘1pl.sbj’ was reanalyzed as *ar=a ‘1pl.sbj=excl’, which then resulted in the modern Nend and Manat pronoun ar ‘1pl’.

This is a kind of edge-aligned reconstruction (Round 2010), in which we leverage the occurrence of linguistic change at the edge of a linguistic unit (in this case, a prosodic unit) to infer the existence of associated structure. Because the linguistic change in our case occurred at the edge of a prosodic unit, we are able to reconstruct some prosodic facts about the western Sogeram dialect into which *=a was borrowed: namely that that verbs usually ended prosodic units and that pronouns often occurred under their own prosodic contour. We thus demonstrate that it is possible, in the right circumstances, to reconstruct some of the prosodic structure of protolanguages.

References

Barðdal, Jóhanna & Thomas Smitherman. 2013. The quest for cognates: A reconstruction of oblique subject constructions in Proto-Indo-European. Language Dynamics and Change 3(1). 28–67.

Berg, René van den & Brenda H. Boerger. 2011. A Proto-Oceanic passive? Evidence from Bola and Natügu. Oceanic Linguistics 50(1). 221–246. doi:10.1353/ol.2011.0000.

Daniels, Don. 2015. A Reconstruction of Proto-Sogeram: Phonology, Lexicon, and Morphosyntax. University of California, Santa Barbara Ph.D. dissertation.

Fuller, Janet M. 2001. The principle of pragmatic detachability in borrowing: English-origin discourse markers in Pennsylvania German. Linguistics 39(2). 351–369.

Harris, Kyle. n.d. Nend texts. Electronic files, Pioneer Bible Translators.

Matras, Yaron. 1998. Utterance modifiers and universals of grammatical borrowing. Linguistics 36(2). 281–331.

Parrish, David. 1989. The Bore grammar essentials: Calibration of the Bore language. Unpublished ms, Pioneer Bible Translators.

Ross, Malcolm, Andrew Pawley & Meredith Osmond. 2016. The Lexicon of Proto Oceanic. Vol. 5 – People: Body and mind. Canberra: Asia-Pacific Linguistics.

Round, Erich R. 2010. Syntactic reconstruction by phonology: Edge Aligned Reconstruction and its application to Tangkic truncation. In R. Hendery & J. Hendriks (eds.), Grammatical Change, 65–81. Pacific Linguistics.

Walkden, George. 2014. Syntactic Reconstruction and Proto-Germanic. Oxford University Press.

Examples

Chini (Ramu)

(1)       Nu    chindata                                    gavɨgɨ.

nu     chi=ndatɨ=a                               ga=avɨ-gɨ

2sg   ascend.opt=med.irr=lnk          after=descend-opt

‘Once you’ve gone up then come back down.’

Nend (Sogeram)

(2)       O-e-m            mɨra    ikŋɨ-z=a                      ntɨ        na-ma-r.

go-ss-cont     pig       shoot-3sg.ds=lnk       blood   eat-hpst-3sg

‘He went and shot a pig and it drank the blood.’                                              (Harris n.d.)

Chini (Ramu)

(3)       Avɨnɨ        mɨndara!

av-ɨnɨ       mɨ=nda-rɨ=a

rain-pc     it=stop-opt=excl

‘The rain needs to stop already!’

Manat (Sogeram)

(4)       A,     amɨŋ=a!                        Inɨ-n       pɨ         krɨs=a!

ah     mother.1.poss=excl      nd-acc   house  bad=excl

‘Ah, Mom! This is a bad house!’

Abbreviations

acc      accusative

cont    continuative

ds        different subject

excl     exclamative

hpst     historic past tense

irr       irrealis

lnk      linker

med      medial verb

nd        near deictic

sg        singular

ss         same subject

obj       object

opt       optative

pc        paucal

pl         plural

poss     possessive

sbj su

ABSTRACT. The Emergence of a New Phoneme: the Vietnamese Case

Studies have shown that in order to understand sound changes, it is important to look at not only the destination and route of change, but also the ‘origin’ of the change, i.e., the dialects spoken in places where these immigrants originated (e.g., Britain 2008, Clopper and Pisoni 2006, Preston 1993). This paper is a synchronic account of the vowel /ɑ/ found in Quảng Nam Vietnamese in south central Vietnam. This vowel is not seen in any major dialects of Vietnamese. Data from fieldwork in 2016 show that the vowel originated from two sub-dialects, Hến and Kẻ Chay, spoken in remote areas of Ha  Tĩnh province, north central Vietnam. Ha  Tĩnh is 300 miles north of Quảng Nam, separated with mountains and rivers. The paper traces the vowel /ɑ/ to its earliest stages and shows that this vowel was brought with early immigrants from Ha  Tĩnh to Quảng Nam, where it was integrated into a completely new dialect and phonemicized to become a new phoneme. The three dialects are compared with Hanoi, a northern, standard dialect. The vowel /ɑ/ is equivalent to the Hanoi /a/. Table 1 shows that /ɑ/ emerges in the Hến and Kẻ Chay dialects at a cost: the loss of /a/ in open syllables in these dialects. When /ɑ/ was introduced to the Quảng Nam dialect, however, the vowel /a/ was restored in open syllables. Comparative examples in Table 2 show that [ɑ] appears to be more stable in the Hến dialect than in the Kẻ Chay dialect. In the Hến dialect it consistently appears as [ɑ] in all syllable types (open syllables, with final consonants or glides). However, in open syllables in the Kẻ Chay dialect, the vowel could be realized either as a labialized [a], [ɑ] or [ɔ]. In the Quảng Nam dialect the vowel /ɑ/ systematically appears as [ɑ] in closed syllables. However, in open syllables, like in the Kẻ Chay dialect, two variations, [wa] and [ɑ], can be found. The vowel /a/, lost in the Hến and Kẻ Chay dialects in open syllabes, appears again in the Quảng Nam dialect (i, j). Certain phonological process (simplification of initial clusters) also supports the phonemic status of /ɑ/ in the Quảng Nam dialect, where the labialized vowel, [wa], can be treated as phonetic diphthongization. In summary, the vowel /ɑ/ in the Quảng Nam dialect is a contact-induced innovation, but also a result of an internal structuring process. Finally, this is one of cases where a sound change was spread by speakers of non-dominant dialects.

References Britain, David. 2008. When is a change not a change? A case study on the dialect origins of New Zealand English. In Language Variation and Change, 20, 187-223. Clopper, Cynthia G. and David B. Pisoni. 2006. Effects of region of origin and geographic mobility on perceptual dialect categorization, in Language Variation and Change, 18, 193-221. Hoa ng Thị Châu. 1989. Tiếng Việt trên các miền đất nước [Vietnamese dialectology]. Hà Nội: Nhà Xuất Bản Khoa Học Xã Hội. Kiparsky, Paul. 1995. The phonological basis of sound change. In The handbook of phonology theory, ed. John A. Goldsmith, 640–670. Oxford: Blackwell. Li, Tana. 1998. Cochinchina, Southern Vietnam in the Seventeenth and Eighteenth centuries. New York: Southeast Asia Program publications. Pham, A. H. 2014. Ngôn ngữ biến đổi và số phận của nguyên âm /a/ trong giọng Quảng Nam [Sound change and the destiny of /a/ in the Quang Nam dialect]. Ngôn Ngữ Học [Journal of Vietnamese Linguistics], 6, 10-18. Hanoi. Pham, A. H. 2016. Sự biến âm trong vâ n tiếng Việt: thổ ngữ la ng Hến, huyện Đức Thọ, tỉnh Ha  Tĩnh [The Hến dialect of Đức Thọ District, Ha  Tĩnh]. Ngôn Ngữ Học, 11. Hanoi. Preston, Dennis R. 1993. Folk dialectology. In D. R. Preston (ed.), American dialect research. Philadelphia: Benjamins. 333–378. Trudgill, Peter. 1988. Dialects in Contact. Oxford: Blackwell. Weinreich, Uriel. 1953. Languages in Contact: Findings and Problems. The Hague: Mouton. Winford, Donald. Contact-induced changes: classification and processes. In Diachronia, 22:2, pp. 373-427. 3

ABSTRACT. The origin of the lateral affricate tl in the development of Nahuatl has been a major point of discussion in Uto-Aztecan historical phonology. In general, the phoneme is found in central Nahuatl dialects and corresponds to a simple t and in some cases an l in other areas.

Examples of the ƛ/t/l correspondences in different dialect areas are shown as follows: (1) tl dialects (Central Nahuatl, some Western Nahua Consider following simple examples showing dialect correspondences of tl ~ t ~l Table 1: 'mountain ''smoke' 'speaks' Colonial central Nahuatl: tepe:-tl po:k-tli tla-htoa Eastern Nahuat Zacapoaxtla, Pue tepe:-t po:k-ti ta-htoa Western periphery Nahual Ostula, Mich tepe:-l po:kli lahtoa

Sapir (1914-15: 456-461) searched for a possible Uto-Aztecan sound law to account "for N[ahuatl] tl as developed from original t according to certain phonetic circumstances," since he then "could dispense with a Uto-Aztecan tl." However, he could identify no systematic evidence for its unique development as a phoneme, concluding that both tl and t should be reconstructed for proto-Uto-Aztecan.

Then in his paper of 1937 Whorf argued that the change was an innovation of the Central Nahuatl dialect area and that all cases of tl --which he established as a single phoneme /ƛ/-- could be traced statistically most frequently to a change in the pUA sequence *ta, reconstructing a number of stem forms, and in detail as an older form of the absolutive suffix that subsequently lost the final *a. Based on Whorf's proposal, the distribution of the ƛ/t/l variants was considered a basic isogloss for Nahua dialectology.

Subsequent publications by Campbell and Langacker (1978), Canger (1988), Manaster Ramer (1996) while basically accepting Whorf's analysis, have added details that include proposed explanations for exceptions as well as arguments that the change should be reconstructed to proto-Nahua and as a result is not an important isogloss chronologically for the languages diversification. In contrast with Whorf's proposal and following Sapir's initial concerns to find the roots of the tl in proto-Uto-Aztecan, this paper focuses on comparative evidence identified in the languages in the family that suggests that at least some tl's developed through vowel loss in morphological sequences of pUA *-LVtV- and *-tVLV- at morpheme boundaries of various kinds. The resulting t-L and L-t word-final consonant clusters in some Nahuatl dialects fused, converging to create the lateral affricate. Proto-UA affixes beginning with *-L- or *-t that formed -l-t and -t-l- sequences yielding ƛ include: a. *-Li suffix, a much-used derivational morpheme that can be reconstructed for pUA. Among its various functions is as the principle UA nominalizing or resultative suffix, a suffix that can be reconstructed in deverbal as well as in other noun formations and which precedes the absolutive suffix. b. *-La suffix, reconstructed in a number of morphological formations of interest here including the accusative case marker following the absolutive suffix on nouns, the possessed noun suffix, and as a locative suffix. c. *-ti, pUA absolutive suffix used in both nominative and accusative case marking on nouns.

It is possible to reconstruct parallel systematic developments from the same pUA sequences found in Nahuatl in other members of the family, in particular for Takic languages, and for Cora, Eudeve, Tubar, and Tarahumara as included in examples under (2) below.

Key in this evidence is a Cora vocabulary from Ortega's 1732 word list in which a number of nouns end with the morpheme sequences -ri-ti and -ri-t. In many of the UA languages in northern Mexico, cognate forms with those carrying final -tl in Central Nahuatl end in -ri, -r or corresponding forms. These include Tarahumara and Guarijío, in which only a -ri or -li suffix is found. Others, such as present-day Huichol show a -ri in singular nouns, but in the case of some plural formations, a -te suffix reappears now functioning as plural marker.

(2) Examples of *-Li(-ti) morpheme sequence in Uto-Aztecan languages in northern Mexico: Eudeve (Lionnet): 'frost, ice'; 'red oak'; 'kind of fish'; 'bathed' ; Cora (Ortega 1732) 'heart'; 'atole'; Tubar (Lionnet): tini-pu-rí-r 'lip'; Tarahumara Re’cho-rí 'pheasant'

There will not be time to discuss arguments based also on word-order change in Nahuatl and Coracholan languages that point to similar parallel developments of ƛ's from pUA -lV-tV and -tV-lV sequences in word-initial and intervocalic positions.

References Campbell, L. & R. W. Langacker. 1978. "Proto-Aztecan Vowels", International Journal of American Linguistics 44. Pt. I, 85-102; II, 197-210; III, 262-279. Canger, U. 1988. "Nahuatl dialectology: A survey and some suggestions”, International Journal of American Linguistics. 54. 1. 28-72. Manaster-Ramer, Alexis. 1996. On Whorf's law and related questions of Aztecan phonology and etymology. International Journal of American Linguistics 62: 176-187. Ortega, José de. 1732. Vocabulario en lengua castellana y cora. México. D. F. [Reprinted in the Boletín de la Sociedad Mexicana de Geografía y Estadística, 1a época, 8:561-605, 1860; also reprinted in Tepic, 1888]. Sapir, E. 1915. “Southern Paiute and Nahuatl, A Study in Uto-Aztecan, Part II", American Anthropologist 17. 98-120, 306-328. Also published in 1914-19. Journal de la Société des Americanistes de Paris, n. s. 11:443-488. Steele, S. M. 1976. “A law of order: word order change in Classical Aztec”, International Journal of American Linguistics 42. 1. 31-45. Whorf, B. L. 1937. “The origin of Aztec tl ", American Anthropologist, 39:265-274.

ABSTRACT. In 1785, James Hutton proposed that we can (and should) account for historical change only using forces we see at work in the present time. Hutton’s proposal was addressed to geology, but later Darwin applied the same reasoning to give a gradualist account of biology. This gradualism, applied to language, seeks explanations of language change in everyday processes of language interpretation, internalisation and production.

In this talk, we use a model of bilingual lexical selection (see [1]). This model combines Grosjean’s account of bilingual lexical selection in terms of language activation and mode with the use of a language monitor to ensure that only within-language terms are expressed. The key finding concerns vocabulary items shared between a bilingual’s languages, items we call doppels (in the psycholinguistic literature these are called cognates confusing them with the overlapping category of the same name in historical linguistics). We find that speakers in bilingual mode, but engaging strong monitoring, avoid doppels in comparison to monolinguals speaking the same language.

Arnal [2] describes an avoidance of doppels in Catalan by native speakers of Spanish who learn the language as adults. These non-dominant speakers of Catalan revive archaic terms for, e.g. bust ́ıa (not buzo ́n) for ’letter-box’, in place of forms identical. Arnal suggests that the lexicon is more consciously accessible than other aspects of language structure, making it an easy prop for maintaining the distinction between the bilinguals’ two languages. We suggest this is implemented via the language monitor, and that the conscious accessibility of the lexicon correlates highly with our ability to monitor for the language-appropriateness of word forms. Arnal notes that the large proportion of learner-speakers of Catalan (upto 40% of the language community) means that peculiarities in bilingual speech can become the linguistic norms for the language.

The Model

Mathematically, our model takes the form of an assessment of the probability P( f |s, t ; b, m) of using a form f while trying to express semantics s in a target language t. The other two parameters are the language mode b expressing how much activation is shared between the target and non-target languages, and m the effort expended in monitoring lexical selections to ensure they come from the target language. As shown in (1), this probability is derived from two values: the activation PAssoc ( f |s, t ; b) from associative memory alone, which reflects form distribution in past experience and the impact of priming, along with PMon(l|f,s,t;b,m) which is the probability assigned by the monitor that this form should be ascribed to language l. When monitoring for our target language, we are of course only interested in the probability that the form originates in our target language. k, here as elsewhere, is a normalising constant, depending on the subscripted parameters.

P(f|s,t;b,m) = k_{s,t;b,m} P_{Mon}(t|f,s,t;m) P_{Assoc}(f|s,t;b)

The definitions for PMon and PAssoc are given in (3) and (2) respectively. In (2) F(f,s,l) is the raw frequency with which the speaker has encountered f to express semantics s in language context l, L is the set of languages the bilingual speaks, and δ is the Kronecker delta (δ_t^l = 1 if l = t, and 0 otherwise).

P_{Assoc}(f|s,t;b) = \sum_{l\in L} (\frac{b}{|L|}+(1-b)\delta^{l}_{t}) \frac{F(f,s,l)}{\sum_{s,l} F(f,s,l)} \eqnlabel{Passoc}

We assume that the monitor implements Bayesian reasoning, with a flat prior over the languages they might be speaking. In (3), the strength m of monitoring acts as a linear coefficient combining an agnostic distribution with the Bayesian estimation of the source language responsible for the meaning-form pairing.

P_{Mon}(l|f,s,t;m) = \frac{1-m}{|L|} + m k_{s,f} P_{Assoc}(f|s,l;0.0) \eqnlabel{Pmon}

This model is given strong support over a 2-monolinguals-in-one-head model in the results of a bilingual lexical selection task (Bayes’ Factor > 106).

In the following sections, we consider the predictions of the model for four different scenarios. We assume that speakers of a non-dominant language will perforce always be in bilingual mode to some extent. In what follows, we consider the impact of the bilinguals level of L2 frequency on their lexical selection, and - if there are enough of these bilinguals in the language community - their impact on the language at large.

Lexical Gaps - Non-Maximal Monitoring

At the lowest level of proficiency, non-dominant speakersoften find themselves lacking a form in their L2 for the semantics they wish to express. As they are in bilingual mode, the corresponding word-form in their L1 is activated.

What is realised depends on their level of monitoring. At anything less than maximum monitoring effort, the form from the bilinguals’ L1 (if there is only one activated) is selected. If more than one L1 available for this meaning, then the most activated is selected.

With non-maximal monitoring, we predict that large numbers of non-dominant speakersat low levels of proficiency are likely to introduce nonce-loans or borrings from their L1 into their L2. Figure (1), blue line shows how the rate of intrusions responds to synset size.

Lexical Gaps - Maximal Monitoring

Maximal monitoring is likely to occur when the non-dominant speakerand their interlocutors do not share their L1. This maybe because social dynamics strongly encourage asymmetric acqusition, or because the speakers are interacting in their L2 with bilinguals with a range of dominant languages.

At maximum monitoring effort, all forms from the non-target language are blocked, and consequently the speaker must resort to circumlocutions or other devices - other than their L1 - to express their meaning. With this level of monitoring, and large numbers of non-dominant speakers, we expect many forms in the richer target language to be replaced by compositional periphrastic constructions.

Doppel Synonyms

Strong monitoring has a very different effect in non-dominant speakerswhen they are selecting between a number of synonyms for a particular situation, and one of the options is a doppel. Assuming that their frequencies are similar, the model predicts (and the effect has been seen directly in experiment) that bilinguals will avoid the shared vocabulary item in favour of one which is distinctive to the target language. Over time, with enough non-dominant speakerswho are this proficient, the target language can expect to see a differential loss in frequency of shared forms and increase in frequency of distinctive forms within synsets. Over generations, this is likely to lead to a relatively rapid divergence in lexical forms between the two languages.

Even between unrelated languages, we may see this effect. If there is a mix of proficiency levels among the non-dominant speakers, then low-proficiency speakers with lexical gaps may introducing borrowings into their L discourse. Monitoring among strong speakers may result on these forms being filtered, so that even if often exposed to them, they do not use these forms frequently themselves (compare red and green lines in Figure (1)).

Conclusion

In summary, based on our model, we see two very different effects of non-dominant speakerson a language, depending on their level of proficiency. If the non-dominant speakersare weak in their second language, then we expect many borrowings into the target language or periphrastic expressions introduced into it. In contrast, if they are strong speakers, and command polyvalent synsets for many meanings, we expect a loss of doppels (shared lexical forms) between the two languages.

References

[1] T. Mark Ellison and Luisa Miceli. Language monitoring in bilinguals as a mechanism for rapid lexical divergence. Language, forthcoming. [2] Antoni Arnal. Linguistic changes in the Catalan spoken in Catalonia under new contact conditions. Journal of Language Contact, 4(1):5–25, June 2011.

FIGURE Caption: The x-axis shows the parameter λ defining a Poisson distribution over synset size. The blue curve shows the expected proportion of intrusions for a given distribution of synset size - as λ increases, there are fewer empty synsets, and so fewer intrusions from the dominant language. The green curve shows the frequency of use of doppels (which may be lucky intrusions) from the dominant language into the non-dominant language, in contrast to their expected rate of use by first-language acquirers with a similar distribution of synset size (red). The bilingual simulations assumed a language mode of 0.667 and monitoring effort at 90%; synsets had Zipf distribution; doppels were assumed to be the most frequent elements of their synsets.

ABSTRACT. With about 300-350 languages, the Bantu family presents a rich testing ground for the study of language variation and language change. While comparative and historical linguistic work in Bantu has traditionally focused on phonological, morphological and lexical data, the use of morphosyntactic data in language comparison has more recently been pursued more widely. Morphosyntactic variation is often thought to largely fall within the domain of typology, but there is also a clear relation to language change and language contact.

The present paper presents an overview of recent studies in morphosyntactic micro-variation in Bantu languages and focuses on an approach based on surface parameters first developed in Marten et al. (2007). We will introduce results from a large-scale comparative project involving 142 parameters, with an empirical base of 30 Bantu languages (Gibson et al. 2016), addressing the question to what extent similarity and differences between Bantu languages are due to

1) common historical origin, at some fairly remote point in time, 2) more recent periods of language and cultural contact, or 3) cognitive or functional constraints on the way language is structured and used.

Initial results from the project support the hypothesis that Bantu languages are both a genetic unit and a linguistic contact zone, and that different processes of transmission and contact lead to higher morphosyntactic similarity on the centre of the Bantu area (‘centripetal convergence’), and decreasing morphosyntactic similarity at the periphery (‘centrifugal convergence’).

The overall findings of the project will be illustrated with specific results from three case studies in three different grammatical areas:

1) verbal morphology: negation 2) nominal morphology: diminutive formation 3) morphosyntax of (non-verbal) predication: copula constructions

While differing in a number of respects, the three case studies also show common features – a central, more conservative conversion zone, as opposed to innovative zones in the northeast and the south of the Bantu-speaking area, as well as a few outlier examples, which do not fit in the overall picture, and which can be related to more recent migration patterns. Each case study thus shows how patterns of change in the relevant system are reflected in the geographic distribution of the relevant languages.

The talk contributes to a better understanding of the historical and typological relations between Bantu languages and helps to trace effects of inheritance and contact with recourse to morphosyntactic data. More generally it explores the contribution parameter-based comparative analyses can make to wider concerns of historical and contact linguistics.

References Gibson, Hannah, Rozenn Guerois, Lutz Marten, and Francisca Everduim. 2016. Morphosyntactic variation and convergence in Bantu. Ms, SOAS. Marten, Lutz, Nancy C. Kula, and Nhlanhla Thwala. 2007. Parameters of morphosyntactic variation in Bantu. Transactions of the Philological Society 105: 1-86.

ABSTRACT. This work provides new theoretical and methodological insights about intersubjectivity and intersubjectification (Traugott & Dasher, 2002; Traugott, 2003, 2010, 2012; Nuyts 2012; Verhagen, 2005; Narrog, 2010, 2012; Dancygier & Sweetser, 2012; Traugott & Trousdale 2013). The theoretical framework of this enquiry is based on Tantucci’s (2013, 2015, 2016a, 2016b) distinction between immediate and extended dimensions of intersubjectivity. While the former is bound to the mutual awareness of speaker/writer and addressee/hearer (immediate intersubjectivity, I-I), the latter includes an assumed third party (specific or generic) who has a social bearing on the utterance (extended intersubjectivity, E-I) (cf. Tantucci, 2013, 2015, 2016a, 2016b). I propose that along a unidirectional cline of change (i.e. Traugott & Dasher, 2002; Traugott, 2003, 2010, 2012), extended intersubjectification constitutes a further stage of semantic and/or grammatical reanalysis with respect to its immediate counterpart. This approach is compatible with the most recent and influential accounts of intersubjectivity and intersubjectification, yet it provides new operational tools to address the relationship between semasiological change and intersubjective construing. The diachronic continuum from immediate to extended intersubjectification that I will discuss in this paper is thus aimed at providing a theoretical and methodological complement to the studies proposed in the literature. I provide corpus evidence about the diachronic continuum from I-I to E-I through the analysis of the construction gànma ‘do what’ in Mandarin and the constructions [you don’t want X] and believe it or not in American English.

References:

Dancygier, Barbara, and Eve Sweetser. 2012. Viewpoint in Language: A Multimodal Perspective: Cambridge University Press.

Narrog, Heiko. 2010. "Modality and Speech-act orientation." Grammaticalization and (inter-)subjectification, Brussel.

Narrog, Heiko. 2012. "Beyond intersubjectification: Textual usages of modality and mood in subordinate clauses as part of speech orientation." English Text Construction 5 (1):29–52.

Nuyts, Jan. 2012. "Notions of (inter) subjectivity." English Text Construction 5 (1):53–76.

Tantucci, Vittorio. 2013. "Interpersonal Evidentiality: The Mandarin V-过 guo construction and other evidential systems beyond the ‘source of information’." Journal of Pragmatics 57:210–230.

Tantucci, Vittorio. 2015. "From immediate to extended intersubjectification: A gradient approach to intersubjective awareness and semasiological change." Language and Cognition FirstView:1-33. doi: doi:10.1017/langcog.2015.26.

Tantucci, Vittorio. 2016a. "Immediate and Extended Intersubjectification in Language Change: Beyond the Opposition between “Theory-Theory” and “Simulation-Theory”." In Dualism, Platonism and Voluntarism: Explorations at the Quantum, Microscopic, Mesoscopic and Symbolic Neural Levels, edited by Sean O'Nuallain, 81–107. Newcastle: Cambridge Scholars Publishing.

Tantucci, Vittorio. 2016b. "Textual factualization: The phenomenology of assertive reformulation and presupposition during a speech event." Journal of Pragmatics 101:155-171.

Traugott, Elizabeth C. 2003. "From subjectification to intersubjectification." In Motives for Language Change, edited by R. Hickey, 124–139. Cambridge: Cambridge University Press.

Traugott, Elizabeth C. 2010. "Revisiting subjectification and intersubjectification." In Subjectification, Intersubjectification and grammaticalization, edited by K. Davidse and L. Vandelanotte, 29–70. Berlin: De Gruyter Mouton.

Traugott, Elizabeth C. 2012. "Intersubjectification and clause periphery." English Text Construction 5 (1):7–28.

Traugott, E. C., and R. B. Dasher. 2002. Regularity in semantic change. Cambridge: Cambridge University Press.

Traugott, Elizabeth Closs, and Graeme Trousdale. 2013. Constructionalization and constructional changes. Oxford: Oxford University Press.

Verhagen, Arie. 2005. Constructions of Intersubjectivity: Discourse, syntax and cognition. Oxford: Oxford University Press.

ABSTRACT. Subjectification — the process of semantic/pragmatic change whereby ‘meanings tend to become increasingly situated in the speaker’s subjective belief state or attitude toward the proposition’ and ‘meanings with largely propositional (ideational) content can gain either textual (cohesion-making) and expressive (presuppositional, and other pragmatic) meanings, or both’ (Traugott 1982; 1989; 2003; 2010; Traugott & Dasher 2001), leading to ‘strengthening of the expression of speaker involvement’ (Traugott & König 1991:191) — has normally been approached from the point of view of semasiology: the object of study has been the meaning of individual words. In this paper I suggest that subjectification may also be approached fruitfully from an onomasiological perspective: here, the object of study is the word that is the exponent of a particular meaning. I present a case study from Spanish which suggests that onomasiological subjectification is a significant mechanism of linguistic change. Spanish has two copular verbs, ser and estar. The distinction between them is complex, but is often defined in terms of individual-level predicate vs. stage-level predicate, essential vs. contingent properties, or characteristic vs. state (for a detailed recent survey, see Camacho 2012) — compare: (1) Eres joven ‘You are young’ (ser) (2) Estás joven ‘You are young-looking; you look young’ (estar). However in several varieties of Spanish, chiefly in the New World — Mexico (Cortés Torres 2004; Gutiérrez 1992; Juárez-Cummings 2014), Cuba (Alfaraz 2012), Costa Rica (Aguilar-Sánchez 2012), Venezuela (Díaz-Campos & Geeslin 2011), Puerto Rico (Ortiz López 2000; Brown & Cortés-Torres 2012), New Mexico (Salazar (2007), and Los Angeles (Silva-Corvalán 1986) — but also in Spain itself (Icardo Isasa 2014; Guijarro Fuentes & Geelin 2006), the use of estar is encroaching on that of ser. This development is often ascribed to contact with languages which have only a single copula — English (Silva-Corvalán 1986) or some varieties of Basque (Icardo Isasa 2014) — or which have two copular verbs, which may be etymologically the same as those of Spanish but with a different distribution (Galician: Guijarro Fuentes & Geelin 2006) or etymologically distinct (other varieties of Basque: Icardo Isasa 2014). Nonetheless, it is far from clear that language contact is at work in all instances of this widespread phenomenon. Moreover, the language-contact hypothesis fails to explain why estar (the marked term of the opposition: see Leonetti 1994) should replace ser, rather than the contrary. I shall demonstrate that, regardless of language contact, this change hinges on the use of estar in evaluative contexts (compare (1) and (2) above). The verb which encodes salience of speaker-attitude is preferred to the alternative in such circumstances and then spreads to other contexts as well; a more ‘subjective’ item replaces a more ‘objective’ item. Onomasiological subjectification yields different surface effects from semasiological subjectification: individual lexical items appear to extend their meaning into less subjective contexts. However, this is an epiphenomenon: underlyingly, in both processes, semantic change shifts in the direction of speaker-attitude. In a tentative outline of future work, I shall suggest that onomasiological subjectification may account for a number of other changes, such as the extension of the definite article in Romance, the replacement of the preterite by the present perfect in several languages, and the conventionalization of some diminutive nouns and frequentative and inchoative verbs in Late Latin.

References:

Aguilar-Sánchez, Jorge. 2012. ‘Formal instruction and language contact in variation: the case of ser and estar + adjective in the Spanishes of Limón, Costa Rica.’ Selected Proceedings of the 14th Hispanic Linguistics Symposium, edited by Kimberly Geeslin & Manuel Díaz-Campos, 9-25. Somerville MA: Cascadilla Press. Alfaraz, Gabriela G. 2012. ‘The status of the extension of estar in Cuban Spanish.’ Studies in Hispanic and Lusophone Linguistics 5, 3-25. Brown, Esther L. & Mayra Cortés-Torres. 2012. ‘Syntactic and pragmatic usage of the [estar + adjective] construction in Puerto Rican Spanish: ¡Está brutal!’ Selected Proceedings of the 14th Hispanic Linguistics Symposium, edited by Kimberly Geeslin & Manuel Díaz-Campos, 61-74. Somerville MA: Cascadilla Press. Camacho, José. 2012. ‘Ser and estar: the individual/stage-level distinction and aspectual predication.’ The Handbook of Hispanic Linguistics, edited by José Ignacio Hualde, Antxon Olarrea & Erin O’Rourke, 453-476. Chichester: Wiley–Blackwell. Cortés-Torres, Mayra. 2004. ‘Ser or estar? Linguistic and social variation of estar plus adjective in the Spanish of Cuernavaca.’ Hispania 87, 788-795. Díaz-Campos, Manuel & Kimberley L. Geeslin. 2011. ‘Copula use in the Spanish of Venezuela: Is the pattern indicative of stable variation or an ongoing change?’ Spanish in Context 8, 73-94. Franco, Fabiola & Donald Steinmetz. 1983. ‘Ser y estar + adjetivo calificativo en en español.’ Hispania 33, 176-184. Guijarro-Fuentes, Pedro & Kimberley L. Geeslin. 2006. ‘Copula choice in the Spanish of Galicia: the effects of bilingualism on language use.’ Spanish in Context 3, 63-83. Guitiérrez, Manuel J. 1992. ‘The extension of estar: a linguistic change in progress in the Spanish of Morelia, Mexico.’ Hispanic Linguistics 5, 109-141. Guitiérrez, Manuel J. 2003. ‘Simplification and innovation in US Spanish.’ Multilingua 22, 169-184. Icardo Isasa, Ane. 2014. ‘Ser and estar variation in the Spanish of the Basque Country.’ Studies in the Linguistic Sciences: Illinois University Working Papers 39, 1-20. Juárez-Cummings, Elizabeth. 2014. ‘Tendencias de uso de ser y estar en la Ciudad de México.’ Indiana University Linguistics Club Working Papers 14, 120-137. Leonetti, Manuel. 1994. ‘Ser y estar: estado de la cuestión.’ Pliegos de la Ínsula Barataría 1, 182-205. Ortiz López, Luis A. 2000. ‘Extensión de estar en contextos de ser en el español de Puerto Rico: ¿evaluación interna y/o contacto de lengua?’ Boletín de la Academía Puertorriqueña de la Lengua Española 28, 99-118. Salazar, Michelle L. 2007. ‘Está muy diferente a como era antes: ser and estar + adjective in New Mexico Spanish.’ Spanish in Contact: policy, social and linguistic inquiries, edited by Kim Potowski & Richard Cameron, 343-353. Amsterdam & Philadelphia: John Benjamins. Silva-Corvalán, Carmen. 1986. ‘Bilingualism and language change: the extension of estar in Los Angeles Spanish.’ Language 62, 587-608. Traugott, Elizabeth Closs. 1982. ‘From propositional to textual and expressive meaning: some semantic-pragmatic aspects of grammaticalization.’ Perspectives on Historical Linguistics, edited by Winfred P. Lehmann & Yakov Malkiel, 245-271. Amsterdam & Philadelphia, PA: John Benjamins. Traugott, Elizabeth Closs. 1989. ‘On the rise of epistemic meanings in English: an example of subjectification in semantic change.’ Language 65, 31-55. Traugott, Elizabeth Closs. 2003. ‘From subjectification to intersubjectification.’ Motives for Language Change, ed. Raymond Hickey, 124-139. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs. 2010. ‘Revisiting subjectification and intersubjectification.’ Subjectification, Intersubjectification and Grammaticalization, edited by Kristin Davidse, Lieven Vandelanotte & Hubert Cuyckens, 29-70. Berlin: De Gruyter Mouton. Traugott, Elizabeth Closs & Richard B. Dasher. 2001. Regularity in Semantic Change. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs & Ekkehard König. 1991. ‘The semantics-pragmatics of grammaticalization revisited.’ Approaches to Grammaticalization: Volume I, Focus on Theoretical and Methodological Issues, edited by Elizabeth Closs Traugott & Bernd Heine, 189-218. Amsterdam & Philadelphia, PA: John Benjamins.

ABSTRACT. The origins of the Afro-Hispanic Languages of the Americas (AHLAs), the languages that developed in Latin America from the contact of African languages and Spanish in colonial times, are extremely intriguing, since it still has to be explained why we do not find creole languages in certain regions of Spanish America, where the socio-demographic conditions for creole languages to emerge appear to have been in place in colonial times. Nowadays, in contrast, we can find such contact varieties in similar former colonies, which were ruled by the British, the French or the Dutch (McWhorter 2000). Despite the fascinating implications of this phenomenon, our knowledge of the AHLAs remains extremely limited. Several hypotheses have been proposed to account for this situation, but no common consensus has yet been achieved (Chaudenson 2001; Mintz 1971; Laurence 1974; Granda 1968; Schwegler 1993, 2014; Lipski 1993; etc.). The pull of different views on the issue has been labelled in the literature as the “Spanish creole debate” (Lipski 2005: ch.9).

The current study is aimed at casting new light on the Spanish creole debate by relying on a comparative analysis of slave laws in the Americas. This article highlights the role that legal differences played in shaping colonial societies and the Afro-European languages that developed in the New World.

Findings indicate the presence of a highly heterogeneous legislation, whose origins must be sought back in Europe, where the bases of slave law were originally laid down—by the Romans. This research shows that the legal figure of the ‘serf/slave’ had been received by the Spanish legal system in ancient times, from the Roman Corpus Juris Civilis; it had been gradually modified and progressively softened into the medieval Spanish code, called Siete Partidas, and then further smoothed in the Leyes de Indias ‘colonial laws’. In particular, the Spanish slave, unlike the Roman one, was granted legal personality and a series of legal rights that derived from it.

By contrast, the legal concept of ‘serf’/‘slave’ followed a significantly dissimilar evolutionary path in the other European codifications, which did not receive it in ancient Roman times. Thus, by the time the Americas were “discovered”, the English, the French and the Dutch found themselves borrowing directly from the Corpus Juris Civilis to fill such a legal gap and introduced slaves into their overseas plantations. As a consequence, English, French and Dutch slaves did not have legal personality and the living conditions set by these legal systems for black captives were much more brutal than the ones dictated by the Spanish Crown (Watson 1989). The Portuguese, on the other hand, had received Roman slave law in ancient times but over time did not modify it to the extent the Spaniards did. As a result, Brazilian slaves were not considered legal persons, and had many more restrictions constraining their freedom than Spanish saves did.

The Legal Hypothesis of Creole Genesis highlights the impact that these legal differences had on the development of black-white relations and therefore on the evolution of Afro-European contact varieties in the Americas. In particular, it stresses the importance of the reception of Roman slave law in Europe as a significant factor for understanding the evolution of creole languages overseas. The point here conveyed might be summarized as follows: if certain colonial societies in the Americas were more or less conducive to creolization than others, it is in great part due to the degree of legal Romanization their homeland countries went through in ancient times.

ABSTRACT. Recently, historical linguists have started to discuss the potential influence of asymmetric priming on long term linguistic change. For example, it has been suggested that this psycholinguistic mechanism might explain the strong tendency of unidirectionality in grammaticalization (Jäger & Rosenbach 2008). This paper will present a mathematical model which in addition corroborates the notion that asymmetric priming can account for the phenomenon of ‘layering’ in language change. Asymmetric priming “describes a pattern of cognitive association in which one idea strongly evokes another, while that second idea does not evoke the first one with the same force” (Hilpert & Saavedra, forthcoming). In a nutshell, the hypothesis asserts that more explicit items (e.g. semantically and phonologically richer forms) are more likely to prime less explicit items (e.g. semantically bleached and phonologically reduced forms) than the reverse (Bock 1986; Shields & Balota 1991). Although these neurological/cognitive effects operate on a very short time scale, they are said to have potential long-term effects (Kaschak 2007). Asymmetric priming supports the successful entrenchment of certain reduced linguistic forms and favors their repeated production, which increases their frequency diachronically. The asymmetric priming hypothesis with its postulated influence on language change has rightfully been criticized for various reasons (some of which we will discuss). For example, it remains unclear whether semantic changes (e.g. semantic bleaching, metaphorical extensions) can suffciently be explained by it (Thompson-Schill et al. 1998; Traugott 2008; Eckhard 2008; Hilpert & Correia Saavedra forthc.). Although we also have our doubts about the explanatory power of the phenomenon with regards to semantic change, we believe that asymmetric priming on the formal level (Brooks & MacWhinney 2000, Bard et al. 2000) can account for reductionist tendencies in phonological change. We call this ‘asymmetric formal priming’. By using a mathematical model on the interaction of linguistic items which differ in their formal substance, we show that asymmetric priming results in layering, i.e. the stable coexistence of two formally similar albeit not entirely identical variants. The mathematical analysis unfolds in two steps (along the lines of Kisdi 1999). First, we formulate a population-dynamical Lotka-Volterra model of the competition between linguistic items with different degrees of formal substance. In order to account for asymmetric priming, the model crucially also features an asymmetric competition term (Figure 1a), so that formally explicit items have a less severe competitive impact on formally reduced items than the other way round. In a second step, we conduct an evolutionary invasion analysis of the model (Geritz et al. 1998); that is we investigate whether new and formally less explicit variants replace their more explicit counterparts. This procedure allows for a simulation of the diachronic long-term development of linguistic items with respect to their formal substance. We show that if full and reduced forms are sufficiently different from each other, layering, i.e. the coexistence of both forms, is stably established (Figure 1b). We empirically apply our model to the (mor)phonotactic domain (Dressler & Dziubalska-Kołaczyk 2006). Our argumentation is based on a study by Plag et al. (2015) which implies that word-final sound sequences exhibit a longer duration, and thus more formal substance, if they occur morpheme internally than if they span a morpheme boundary (e.g. /nz/ in lens vs. run+s). Drawing on diachronic data, we use our model to explain the semiotically unexpected coexistence of – apart from duration – phonemically identical morphonotactic and lexical consonant clusters in languages such as English and Afrikaans (cf. Dressler et al. 2010). We will conclude that asymmetric formal priming can help to explain phonotactic reductions better than the traditional ‘ease of effort’ argument and at the same time accounts for the layering of similar patterns.

ABSTRACT. The lexicon of any natural language consists of a series of phonological matrices, each of which is associated with a particular set of semantic and procedural features. Within this series of discrete lexical items, the relation between phonological and content features is highly unstable, and the very origin and evolution of this relation is the cause of such instability. Human lexicons are sociohistorical entities which can be synchronically and diachronically studied, inasmuch as they undergo, in the course of time, loss and addition of lexical items, and reconfiguration of the relation between the phonological matrices and the semantic or procedural features they may encode.

This workshop aims to provide an outline of certain syntactic phenomena present in Early Modern English. From the semantic-syntactic framework proposed by Jaume Mateu i Fontanals (2000, 2002), we consider argument structure, as defined in terms of relational-semantic construals, to be an essential component, indeed the primary one, of any morphosyntactic derivation. These construals can be represented by means of relational-semantic tree diagrams, which render visible the way the arguments are related to one another by means of a number of primitive predicates encoding relations such as cause, telicity (i.e. change and state) and direction (i.e. terminal and central coincidence). These diagrams stand for the existence of a conceptual structure which is relevant to syntax; thus, the very formal device used to represent argument structure conveys meaning, that is, meaning is inherent to the formal relations established between the elements which merge on X-bar theory tree structures.

By taking into account that a whole verb typology can be derived from this framework, we propose to extend these theoretical principles to a number of phenomena in Early Modern English (EME). Special attention is dedicated to the behaviour, in morphosyntactic contexts found in the original Elizabethan English version of William Shakespeare’s works, of verbs which are classified in Present-day English into categories which do not correspond with their EME counterparts. These are mainly unaccusative, ergative, unergative and transitive verbs.

The phenomena included seem to reveal certain similarities between Early Modern English and other languages, while it also presents a number of differences from the way certain Present-day English verbs behave. These differences become clear in the behaviour of certain verbs which are now classified as “intransitive” and which were used transitively. More generally, both the syntactic and the semantic features of the verbal constructions in Elizabethan English resemble those of Romance languages like Spanish. This is made clear through the use of double negatives, the richer agreement features of that variety of English, and the movement of the verb to T°, thus allowing direct negation without auxiliary support. In Relational Semantics terms, the presence of a dative clitic and the reversed argument structure with some unergative verbs are certainly comparable to their equivalents in other languages as well.

Ultimately, we hope the analysis and the theoretical framework used here will be valuable tools in other formal diachronic studies which aim to shed light on the nature of semantic-syntactic change.

ABSTRACT. This study is an empirical analysis of Light Verb Constructions (LVCs) in Old Swedish. Following diachronic studies on light verbs in a variety of languages (Brinton and Akimoto 1999, Claridge 2000, Elenbaas 2013, Butt and Lahiri 2013), I undertake a comparative analysis of Old Swedish and focus on complex predicates that contain a semantically light verb, such as giva 'give' or göra 'make', and an abstract nominal object, as in giva radh 'give advice' or giva hiälp 'give help'. The main research questions of the study are:

1. Does the frequency of LVCs increase or decrease in Old Swedish? 2. Does the diversity of light verb + nominal object pairings increase or decrease in Old Swedish? 3. What are possible causes for the changes in frequency and diversity of LVCs in Old Swedish?

Using a corpus of 800,000 words in twelve Old Swedish texts written between 1225 and 1525 available through the Fornsvenska Språkbanken and the Korp online concordance tool, I track the frequency of LVCs with five common light verbs (fa 'get/receive', giva 'give', göra 'do/make', hava 'have', and taka 'take') and analyze the diversity of transitive light verb + nominal object pairings. Texts were selected from a variety of text types (legal texts, religious prose, secular prose) and from different sub-periods of Old Swedish, including a mixture of translations (from Low German or Latin) and original Swedish texts. The corpus search with Korp and manual filtering for alternative spelling of variant forms yielded 1,457 tokens (light verbs + nominal objects) that contain 352 unique nominal objects with which light verbs are paired. These include examples such as those in (1), with gifwa hiälp 'give help', or (2), with göra dom 'make/do judgment':

(1) at thin gudh ok herra skal gifwa thik hiälp that your God and Lord shall give-INF you help-ACC 'That your Lord God shall give you help.' (Birg) (2) Ok ey gör iak dom vtan miskund and not do.PRS I judgment-ACC without mercy 'And I do not pass judgment without mercy.' (Birg)

In the empirical portion of the paper, I consider multiple variables that affect the frequency and variety of light verb and NP object pairings. In particular, I divided the texts into sub-periods (1225-1375, 1375-1450, 1450-1526) in order to be able to track changes in frequency over time. Other variables include text type (laws, secular prose, religious prose), and the type of modification to the nominal object in the quantitative analysis (e.g., presence of definite or indefinite articles, pre-nominal adjectives, possessives, or quantifiers). To measure the frequency of LVCs, I calculated the overall frequency of each of the five light verbs with various NP objects over all three periods and also calculated the normalized frequency of LVCs per 10,000 words for each individual text. In the same way, I also measured the Type-Token Ratio (TTR) as a ratio of the number of unique NP objects (types) to the overall number of LVCs (tokens) both in the aggregate for the three periods and as a normalized ratio per 10,000 words, adjusting for the varying length of texts. Results indicate that the frequency of LVCs and the number of unique NPs with which light verbs occur increase in Old Swedish as whole. However, data on each individual text indicate that text type is an important factor to consider. Law texts show a higher frequency but lower number of unique light verb-NP object combinations while secular and religious prose texts exhibit a lower frequency and higher level of diversity. Quantitative analysis also reveals that there is an interaction effect between text type and time, as the diversity increases in law texts and the frequency increases in religious prose over the entire Old Swedish period. Data on the type of modification to the NP indicate that this variable does not play a role in the patterns of LVC use. These results are discussed in terms of previous diachronic studies that propose different reasons for changes in LVCs over time. In particular, I examine the possible influence from other languages that may result in calques or borrowings during periods of language contact, as suggested in Akimoto and Brinton (1999), Iglesias-Rabade (2001), and Ronan (2014) for the history of English. Based on the strong text-type effect that differentiates the use of LVCs in secular and religious prose from the legal texts, we suggest that the LVC frequency and the increase over time may be due to foreign influence, either indirectly through the multilingual context in the writing process or directly through translation. I follow Höder's (2010) analysis of syntactic effects of multilingualism during the Old Swedish period and suggest that because of the written culture that emerges with the elite cultural group of speakers/scribes during this time, there may be significant differences in word order and vocabulary that come about through grammatical transfer. In addition, I evaluate the possibility that lexicalization may play role in the increase of LVCs during the Old Swedish period. I build on the discussion in Brinton (2008) in which she suggests that some LVCs may be the result of becoming fixed idiomatic phrases. Citing evidence from the Old Swedish data, I suggest that many of the LVCs are the products of lexicalization, especially in the legal texts, showing high frequency and repetition of the same phrases and little modification within the NP. Moreover, verbs other than the five common light verbs are often included in these idioms, such as bära 'bear' in bära vitn 'bear witness' or bära skuld 'bear shame'. I contrast the data on lexicalized phrases with data in studies on early English and argue against the possibility that has been proposed (Brinton 2008) that LVCs exhibit some characteristics of grammaticalization.

References BRINTON, LAUREL J., and MINOJI AKIMOTO (eds.) 1999. Collocational and idiomatic aspects of composite predicates in the history of English. Amsterdam: John Benjamins Publishing.

BRINTON, LAUREL J. 2008. Where grammar and lexis meet. Theoretical and empirical issues in grammaticalization, ed. by Elena Seoane, and María José López-Couso, 33-53. John Benjamins Publishing, 2008.

BUTT, MIRIAM and ADITI LAHIRI. 2013. Diachronic pertinacity of light verbs. Lingua 135.7- 29. CLARIDGE, CLAUDIA. 2000. Multi-word verbs in early Modern English: A corpus-based ELENBAAS, MARION. 2013. The synchronic and diachronic status of English light verbs. Linguistic Variation 13.48-80.

HÖDER, STEFFEN. 2010. Sprachausbau im Sprachkontakt: syntaktischer Wandel im Altschwedischen. Vol. 35. Heidelberg: Universitätsverlag Winter.

IGLESIAS-RÁBADE, LUIS. 2001. Composite predicates in Middle English with the verbs nimen and taken. Studia Neophilologica 73.143-163.

RONAN, PATRICIA. 2014. Light verb constructions in the history of English. Corpus interrogation and grammatical patterns, ed. by Kristin Davidse, Caroline Gentens, Lobke Ghesquière, and Lieven Vandelanotte, 15-34. Amsterdam: John Benjamins Publishing.

ABSTRACT. [Research question] The diachronic development of *rs in the prehistory of Umbrian (Indo-European, Italic branch, Sabellic group) has puzzled a number of scholars (see, e.g., von Planta 1892, Buck 1928, Poultney 1959 and Meiser 1986). While the original *rs is preserved in Umbrian written as or (bold type indicates it was written in the native alphabet and italics indicate that the Latin alphabet was used), secondary *rs sequences which arose via syncope or assimilation are not retained but are written as or . The research question I take up in this paper is how to account for their different development both in terms of synchronic and diachronic point of view.

[Data] 1. Proto-Sabellic *rs > Umb. , : e.g., Umb. tuseṭu, tursitu ‘scare’ [3.sg. imp. act.] < *torsētōd (cf. Lat. torreō ‘to terrorize’ < *tersējō, see de Vaan 2008:617), Umb. fasiu, farsio ‘(an offering cake made from a kind of grain[?])’ [acc.pl.] < *farsejo- (see Untermann 2000:266). 2. Proto-Sabellic *rtt > ?*rs > Umb. : e.g., Umb. trahuorfi ‘in crosswise fashion’ [adv.] < *-wort-to- (cf. Lat. transverse <*-wort-to-, for *we < *wo in Latin, see Weiss 2009:140). 3. Proto-Sabellic *rVs > ?*rs > Umb. , : e.g., Umb. çerfie, śerfie ‘(theonym)’ [dat.sg.] < *keresijāj or *kerasi(j)āj (see Untermann 2000:390).

[Previous claim] Poultney (1959:77): explains that original */-rs-/ did not undergo the diachronic change of */rs/ > */rf/ because it had already become */-rz-/ by the time when */rs/ > */rf/ occurred. According to his analysis, */rtVs/ > */rs/ and */rtt/ > */rs/ had completed by this time and therefore they underwent */rs/ > */rf/, resulting in the spelling of . Meiser (1986:172, 174), on the other hand, assumes that */s/ in */rVs/ was first voiced to */z/, and that */rVz/ developed to */rf/ after syncope by a sound change of */rz/ > */rf/ (thus */rVs/ > */rVz/ (voicing) > */rz/ (syncope) > */ rf/). For */rtt/, he explains it became */rss/ via assibilation and changed to */rf/ by a different sound change (namely, */rss/ > */rf/). According to his analysis */rs/ did not change to */rf/ because its structural description did not meet with these two sound changes.

[Problem] Both Poultney and Meiser’s analyses suffer several drawbacks. Since according to Poultney’s analysis, original */rs/ must have become */rz/ at a certain prehistorical stage. Therefore, in order to account for Umb. tuseṭu, tursetu ‘scare’ whose and point to [rs] (since if they represent [rz], tuseṭu would have been written as †), we would have to assume either a synchronic devoicing rule /rz/ → [rs] or a diachronic devoicing rule */rs/ > */rz/ > Umbrian /rs/. However, not much motivation is found for supporting the former, and it is preferable to dispense with the latter which would reverse the diachronic change. If we accept Meiser’s analysis, we need two synchronic phonological changes at a certain prehistoric stage of Umbrian: */rz/ > */rf/ and */rss/ > */rf/. However, he is not explicit about the relationship between these two changes which produced the same *[rf] cluster. Nor does he give any possible reason which kept *[rs] apart from these changes. Plausible phonetic motivation which lead */rz/ → */rf/ but */rs/ → †/rf/ is difficult to find. Moreover, */rss/ was not likely to be stored as an underlying representation at any prehistorical stage of Umbrian given that /rtt/ is always realized as [rs] both in Latin and Oscan (a Sabellic language sister to Umbrian; e.g., Latin versus [wersus] ‘turned’ < /wert-to-/ and Osc. ϝερσορει [werso(ː)re(ː)j] ‘(epithet of Jupiter)’ < / wert-tōr-/). In view of these synchronic derivations, the most natural reconstruction would be to assume that */rtt/ was also realized as *[rs] in the prehistorical stage of Umbrian.

[Proposal] In this paper, I propose that the synchronic phonological rule which gave birth to Umbrian , was neither */rss/ → *[rf] nor */rz/ → *[rf], but */rs/ → *[rf] (as in Poultney 1959). According to my analysis, */rtt/ had been synchronically realized as *[rs] by the stage when this rule was operated. Also, syncope of a medial vowel had given rise to */rVs/ → *[rs] by this stage. I analyze that synchronically these two rules were in the feeding order to */rs/ → *[rf]. As the result, derivation was synchronically */rt-t/ → rss → rs → *[rf] and */rVs/ → rs → *[rf]. On the other hand, */rs/ → *[rf] did not take place for words such as tursitu. I propose that this is because their /rs/ cluster was not in a synchronically derived environment. It has been known as Non-Derived Environment Blocking (see, e.g., Kiparsky 1993) in which certain synchronic rules operate only in a derived environment (e.g., In Finnish, /t/ becomes [s] when it is followed by /i/ across morpheme-boundary, but it remains when followed by a morpheme-internal /i/. e.g., / halut-i/ → [halusi] ‘want-PAST’ vs. /koti/ → [koti] ‘home’ †[kosi]). I argue that it is this effect that blocked the non-derived [-rs-] to undergo this rule. My proposal solves the problem of unnatural two phonological changes (*/rz/ > */rf/, */rss/ > */rf/), and provides a unified account of the origin of Umbrian and .

References: Buck, Carl D. 1928. A Grammar of Oscan and Umbrian. 2nd ed. Boston: Ginn and Company. de Vaan, Michiel. 2008. Etymological Dictionary of Latin and Other Italic Languages. Leiden/ Boston: Brill. Kiparsky, Paul. 1993. Blocking in non-derived environments. In Sharon Hargus and Ellen M. Kaisse (eds.), Phonetics and Phonology 4: Studies in Lexical Phonology, 277–313. San Diego: Academic Press. Meiser, Gerhard. 1986. Lautgeschichte der umbrischen Sprache. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck. Poultney, James. W. 1959. The Bronze Table of Iguvium. Baltimore: American Philological Association. Untermann, Jürgen. 2000. Wörterbuch des Oskisch-Umbrischen. Heidelberg: Carl Winter. von Planta, Robert. 1892. Grammatik der oskisch-umbrischen Dialekte. Vol 1. Strassburg: Karl J. Trübner. Weiss, Michael 2009. Outline of the Historical and Comparative Grammar of Latin. Ann Arbor: Beech Stave Press.

ABSTRACT. The development of electronic linguistic corpora, together with the assistance of Corpus Linguistics and Social History, is allowing Historical Sociolinguistics to immerse the researcher into remote periods of a language and explore its internal functioning and its users’ sociolinguistic behaviour in social interaction more accurately. The preservation of collections of English private correspondence involving writers of different personal circumstances offers a very useful source to carry out quantitative and qualitative sociolinguistic analysis. The aim of this paper is to show results of the HiStylVar Project, which explores the motivations and mechanisms for stylistic variation in historical corpora of English written correspondence in connection with the contemporary theoretical models developed for its study. The extension and extrapolation of conclusions obtained from sociolinguistic studies on patterns of stylistic variation of current English situations to Late Middle English and Early Modern English communities allow us to test the validity of these current theoretical models of intra-speaker variation assuming: (i) that the evolution of linguistic and social systems always occurs in relation to the socio-historical situations of their speakers, (ii) that the past should be studied in order to understand and explain the present (and viceversa), and (iii) the feasibility of universal and temporal validity of the Uniformitarian Principle. The study is carried out through the analysis of the behaviour of variable (TH) in members of the Paston family from the archival source of The Paston Letters. The results show that, in addition to tracing language variation and change throughout a speech community, private letters from historical corpora may also shed light onto the mechanisms and motivation(s) for variability in individuals and their stylistic choices in remote societies such as those of the late Middle and early Modern English periods.

ABSTRACT. The development of degree modifers has been the object of recent study in Germanic languages (Claridge & Kytö 2014, a.o.), but has been understudied in Ibero-Romance, particularly in Portuguese. This paper focuses on the syntactic and semantic change undergone by the noun bocado, a derived noun meaning ‘a piece that fits in the mouth, a piece of bread’ (Corominas & Pascual 1980) in Old Portuguese, as in (1), and traces its development to the degree modifier um bocado ‘a bit’ in contemporary Portuguese. The degree adverbial may modify verbs, nouns, and adjectives (an example of the latter being [2]).

(1) dando a cada hûû seu bocado vïîdo ordinhadamët hûû depos outro. (Afonso X, Primeyra Partida, 1300s, CdP) ‘giving each person his/her piece of bread, orderly, one after the other’

(2) Os últimos dois anos foram um bocado frustrantes. (Newspaper interview, 1900s, CdP, Portugal) ‘The last two years have been a bit frustrating.’

Two aspects are studied: (i) the relation between the lexical meaning of the noun and its use as a minimizer in Old Portuguese; and (ii) the reanalysis of the phrase indefinite determiner + bocado as a degree modifier, which was completed in the 1700s. I argue that both aspects can be explained by the scalar semantics of the noun; bocado denoted a small portion of food or drink and hence is used to denote a low value on a scale of quantity. This provides both the conditions for the scalar reasoning underlying its use as a minimizer (Fauconnier 1975, Israel 2011) and the semantic source for the development of the cross-categorial modifier meaning ‘a bit’. In Old Portuguese bocado is found as a bare noun occurring under the scope of negation (either the negative adverb non or the scalar negator nem ‘not even’), as in (3), to reinforce a negative statement (Hoeksema 2001):

(3) que pan non comeu bocado nen beveu agua (Cantigas de Sta. Maria, 1200s, CdP) ‘that she didn’t even eat a piece of bread or drink water’

When bocado is a “partitive minimizer” (Martins 2000), as in (3), the noun denotes a minimal amount (Llorens 1929). While in other Romance languages minimizers often undergo Jespersen’s cycle and become negation markers, in Portuguese this development is rare and minimizers tend to disappear from the language (Pinto 2015). However, the phrase um bocado remained, with its scalar meaning being the basis for the adverbial use. At an intermediate stage of the development, we find um bocado as a minimizer with a verb that is neither a transitive verb nor a verb of consumption, the gradable verb gostar ‘to like’, and with modification by the adjective só ‘a single bit’:

(4) tam maa vida passou aqui jaz quem nom gostou deste mundo hû soo bocado. (Cancioneiro, Garcia de Resende, 1516, CdP) ‘He had such a bad life; here lies someone who didn’t like this world (even) a little bit’.

While in Old Portuguese bocado displays nominal properties (e.g. it could be preceded by both definite and indefinite determiners, it could be modified by adjectives and it could be pluralized), as an adverbial it must be preceded by the indefinite determiner and no modifier may occur between determiner and noun. Evidence for the recategorization of the sequence um bocado appears in the 1700s; we observe a widening of the distribution of the phrase, that now occurs with any class of verbs (not just verbs of consumption) and also with adjectives and nouns (in this case, in a pseudo-partitive construction with mass nouns, similarly to what happened with algo in Spanish, see Amaral 2016). The reanalysis of um bocado as a degree modifier is favored by intersective gradience in the sense of Aarts (2007): elements of a syntactic category may occur in environments that are typical of another category (i.e. their properties intersect). This is common at the phrase level: “This type of gradience typically occurs in the syntax in locations where we can distinguish ‘slots’ that are normally filled by a particular class of elements” (Aarts 2007: 162). In the case of um bocado, this process was facilitated by word order: in Old Portuguese um bocado is a direct object of a verb of consumption meaning ‘to eat’, ‘to drink’, ‘to swallow’, and hence typically occurs post-verbally. Given this position in the clause, it may be reanalyzed as a VP-modifier with gradable verbs. Note that although bocado has nominal properties, it is not a prototypical noun because of its non-referential functions (e.g. as a minimizer). Its potential to operate on a scale creates the conditions for the sequence um bocado to have a quantificational meaning and hence occur in positions that are associated with adverbials, which provides cues for its recategorization as an adverb. That is to say, its scalar meaning underlies intersective gradience, which historically can lead to reanalysis. The syntactic and semantic change undergone by bocado not only exemplifies the development of degree modifiers from nominal expressions in Ibero-Romance but it also sheds light on the connection between adverbials that express degree and the semantics and pragmatics of negation and minimizers.

References Aarts, B. 2007. Syntactic gradience. Oxford: Oxford University Press. • Amaral, P. 2016. When something becomes a bit. Diachonica 32(2): 151-186. • Claridge. C. & Kytö, M. 2014. • “You are a Bit of a Sneak”: Exploring a degree modifier in the Old Bailey Corpus”. In M. Hundt (ed.), Late Modern English Syntax. Cambridge: CUP, 239-268. Corominas, J. and Pascual, J. 1980. Diccionario crítico etimológico castellano e hispánico. Madrid: Gredos. • Hoeksema, J. 2001. Rapid change among expletive polarity items. In L. Brinton (ed.), Historical Linguistics 1999: Selected papers from the 14th ICHL. Amsterdan: John Benjamins, 175-186. • Fauconnier, G. 1975. Pragmatic scales and logical structures. Linguistic Inquiry 6: 353-375. • Israel, M. 2011. The grammar of polarity. Oxford: OUP. • Llorens, E.L. 1929. La negación en español antiguo con referencia a otros idiomas. Madrid (Revista de Filología Española, anejo XI). • Martins, A. M. 2000. Polarity items in Romance: Underspecification and Lexical Change. In S. Pintzuk et al. (eds.), Diachronic Syntax: Models and Mechanisms. Oxford: OUP, 191-219. • Pinto, C. 2015. Para a história da negação: o minimizador homem no português antigo. Estudos de lingüística galega 7: 109-123.

CdP: Corpus do Português (http://www.corpusdoportugues.org/)

ABSTRACT. As in other linguistic fields, within the area of historical semantics several authors have been arguing for the need to pursue corpus-linguistic methods in order to facilitate a more principled way of verifying the results. However, the application of empirical, quantitative methods to the study of semantics is not straightforward: how can meaning, an intrinsically non-observable phenomenon in our mind, be investigated by means of quantitative methods (e.g. Geeraerts 2010: 64; Glynn 2010: 240, 2014: 7)? Moreover, apart from this challenge of the study of meaning in general, the study of meaning change adds several additional challenges. Indeed, semantic change is an area in which quantitative methods face specific challenges due to the nature of the data (bias towards specific registers, authors and genres, discontinuity of genres, sparseness of data, etc., cf. Hilpert 2013). In this talk, we use a dynamic ‘behavioral profile’ (BP) (e.g. Gries/Divjak 2009; Gries 2010) in order to disentangle the diachronic evolution of the polysemy of the Spanish perception verb sentir (‘to feel’).

The BP approach starts from a very fine-grained manual annotation of a large number of syntactic, morphological, semantic and pragmatic variables. More precisely, we annotated 4488 matches of the lemma sentir, according to five chronological cutoff points (1270-90, 1470-90, 1670-90, 1870-90, beginning 21st century), randomly selected from the Peninsular Spanish part of the CORDE corpus for the diachronic data and the CREA, PRESEEA and COLAm corpus for the contemporary data.

Methodologically, this study presents the first application of the BP approach to historical data and proposes some methodological innovations both within the current body of research in historical semantics and with regard to previous applications of the BP approach. As such, this study provides an extension of the methodological apparatus of the BP approach by complementing the traditional Hierarchical Agglomerative Cluster analysis with a dynamic BP approach derived from Multidimensional Scaling maps.

Semantically, the results show that the diachronic evolution of sentir turns out to match the general tendency of polysemy extending from more concrete meanings towards more abstract meanings. As such, it has evolved from a true physical perception verb (13thc.) towards a verb with a clearly dominant emotional meaning in contemporary Spanish. Syntactically, the dynamic BP shows a syntactic continuum between the middle voice use (SENTIRSE) and the other uses of the verb (SENTIR), which remains fairly constant in the course of time.

Finally, the dynamic BP also contributes to a comprehensive perspective on the process of constructionalization and the nature of networks by visualizing the rise and development of the Discourse Marker (DM) lo siento (‘I’m sorry’):

(1) Lo siento, señor, pero […] me encontraba un poco despistada. ‘I am sorry, Sir, but I was a bit distracted’

By examining the DM within the bigger picture of the changing polysemic profile of the verb through the course of history, the dynamic BP approach provides both a holistic and a very detailed interpretation of the change. More precisely, our analysis shows that, instead of a direct lineage from 'sentir' towards 'lo siento', two different paths converged giving rise to the DM: in its present use, it is essentially a hybrid, one that combines a cognitive form with an emotional meaning.

References: - Geeraerts, D. (2010). The doctor and the semantician. In D. Glynn & K. Fischer (Eds.), Quantitative methods in cognitive semantics: corpus-driven approaches (pp. 63-78). Berlin: Mouton de Gruyter. - Glynn, D. (2010). Testing the hypothesis: Objectivity and verification in usage-based Cognitive Semantics. In D. Glynn & K. Fischer (Eds.), Quantitative methods in cognitive semantics: corpus-driven approaches (pp. 239-269). Berlin: Mouton de Gruyter. - Gries, S.T., & Divjak, D. (2009). Behavioral profiles: A corpus-based approach to cognitive semantic analysis. In V. Evans & S. Pourcel (Eds.), New directions in cognitive linguistics (pp. 57-75). Amsterdam: John Benjamins. - Gries, S.T. (2010). Behavioral profiles: A fine-grained and quantitative approach in corpus-based lexical semantics. The Mental Lexicon, 5(3), 323-346. Hilpert, M. (2013). Constructional change in English: Developments in Allomorphy, Word Formation, and Syntax. Cambridge: Cambridge University Press.

ABSTRACT. There are four autochthonous Portuguese-based Creole languages in the Gulf of Guinea: Santome, Angolar, both spoken in the Island of São Tomé, Lung’ie, spoken in the Island of Príncipe, and Fa d’Ambô, spoken in the Island of Ano Bom (Ferraz 1979, Hagemeijer 2009). All languages derive from the Proto-Creole of the Gulf of Guinea (PGG), which emerged in the beginning of the 16th Century due to the contact between Portuguese settlers and African people taken as slaves to São Tomé. Isolation and new inputs from African languages helped to promote language speciation: on the one hand, PGG evolved to Santome (ST) in the 16th Century in the Island of São Tomé, whereas Angolar (AN) is the language of the descendants of runaway slaves who created a local maroon community; on the other hand, speakers of the Proto-Creole colonized the Islands of Príncipe and Ano Bom, where there were conditions for the speciation to Lung’ie (LI) and Fa d’Ambô (FA), respectively.

After a phonological and lexical reconstruction of the PGG based on 536 sets of cognates from the four daughter languages (data from our own field-work notes and Segorbe 2007, Araujo & Hagemeijer 2013 and Agostinho 2016), the aim of this presentation is to show how a proto-language that allowed liquid consonants in the onset and coda evolved to languages that treated them differently: none of these languages has rothics and laterals in codas, LI has rothics in onsets, while ST, AN e FA do not have them. ST allows complex onsets whose second element is /l/, while the same cognates in LI, AN e FA have no complex onsets, but rather long vowels. Thus, in this presentation, we will show how liquid consonants in complex onsets and codas of the PGG were managed in the daughter languages .

Our reconstruction proposes the following: (1a) three liquid consonants, *l, *ʎ and *r, appear at the onset of PGG, (1b) *l occurs as the second element in complex onsets, and (1c) *r and *l are allowed in codas. Based on reflexes in ST, AN, FA and LI, there is no indication that *r was allowed at the end of words in the proto-language, although that happened in Portuguese, its superstrate language.

(1) a. Onset: *liN.pu ‘clean’, *zu.ʎu ‘July’, *ri.zu ‘hard’ b. Complex onset: *gle.za ‘church’ c. Coda: *pɛr.tu ‘close to’, *mal.da.di ‘wickedness’

As a matter of fact, in onset, the proto-phoneme *r remained only in LI (2), while lambdacism (*r > [l]) was the general pattern for ST, AN and FA. Additionally, *l and *ʎ survived in onsets, except for Angolar, where *ʎ generally evolved to [l].

(2) a. *ra.bu > [ˈla.bu] (ST, AN, FA), [ˈra.bu] (LI) ‘tail’ b. *sɛ.ra > [sɛˈla] (ST, AN, FA), [sɛˈra] (LI) ‘to smell’

However, when complex onsets are considered, PGG and its four daughter languages show the following: if the complex onset is on the first syllable of disyllabic words, ST keeps /l/ in situ, whereas AN, FA and LI delete it and the vowel is lengthened (3a). If it is on the second syllable, ST keeps the /l/ in situ, AN and LI remove the /l/ with no vowel lengthening and, finally, FA creates a new syllable in which /l/ moves to the onset and a copy vowel is inserted (3b).

(3) a. *gle.za > [ˈgle.za] (ST), [ˈgeː.za] (FA), [ˈgeː.za] (LI), [ˈŋɡeː.ða] (AN) ‘church’ b. *fɛ.blɛ > [ˈfɛ.blɛ] (ST), [ˈfi.bi.li] (FA), [ˈfɛ.bi] (LI), [ˈfɛ.bɛ] (AN) ‘fever’

In the coda position, liquid consonants makes ST different from its sister languages: *r in coda was removed before a [coronal] consonant in all languages, FA has further vowel lengthening. Nonetheless, before a [labial] or [dorsal] consonant, FA, LI e AN removed *r (with no vowel lengthening), whereas ST changes *r into /l/ followed by a metathesis, generating a complex onset.

(4) a. *kar.ni > [ˈka.ni] (ST, AN), [u.ˈka.ni] (LI), [ˈxaː.ni] (FA) ‘meat’ b. *kur.tu > [ˈku.tu] (ST, AN, LI), [ˈkuː.tu] (FA) ‘short’ c. *bar.ga > [bla.ˈga] (ST), [baː.ˈga] (AN, LI, FA) ‘tear apart’ d. *por.ko > [ˈplo.ko] (ST), [poː.ˈko] (AN, LI), [poː.ˈxo] (FA) ‘pig’

If *l is the consonant before a [coronal] in the coda, ST, AN and LI remove it. FA, in turn, shows vowel lengthening after its removal (5a). Before a [labial] or [dorsal] consonant, ST creates a complex onset from the metathesis of /l/, while AN and LI remove it. In both cases, there is no vowel lengthening. Notwithstanding, FA removes *l and the vowel is lengthened (5b).

(5) a. *falta > [fa.ˈta] (ST, AN, LI), [faː.ˈta] (FA) ‘to lack’ b. *salva > [ʃla.ˈva] (ST), [sa.ˈva] (LI), [θa.ˈva] AN, [saː.ˈva] (FA) ‘to save’

In this presentation, we aim to show how the interaction between the Obligatory Contour Principle (Leben 1973, McCarthy 1986) and specific rules of each language led to such changes.

References

AGOSTINHO, Ana. 2016. Fonologia do Lung'ie. Lincom Studies in Pidgin & Creole Linguistics 15. Lincom:

ARAUJO, G. & HAGEMEIJER, T. 2013. Dicionário Santome-Português/Português-Santome. São Paulo: Hedra.

FERRAZ, L. I. 1979. The creole of São Tomé. Johannesburg: Witwatersrand University Press.

HAGEMEIJER, T. 2009. Initial vowel agglutination in the Gulf of Guinea creoles. In: ABOH, Enoch & SMITH, Norval (Ed.). Complex processes in new languages. Amsterdam/Philadelphia: John Benjamins Publishing Company. p. 29-50.

LEBEN, William. 1973. Suprasegmental Phonology. MIT Thesis, Massachusetts.

MCCARTHY, John J. 1986. OCP effects: Gemination and antigemination. Linguistic Inquiry, v. 19, p. 451-475.

SEGORBE, A. 2007. Gramática descriptiva del fa d’ambô. Barcelona: CEIBA Ediciones.

ABSTRACT. Analyses of Latin/Romance historical morphology commonly state that sound change caused the morphological collapse in of various parts of the morphological systemareas, especially case, the future, and the passive (Juge 2009). Romance reflexes of the third person genitive masculine/neuter plural illōrum “their” show that this view fails to consider all the relevantimportant data. I argue that eliminating flawed explanations constitutes progress even when no satisfactory explanation is yet available. Otherwise, incorrect analyses may receive repeated approving citations, as in the case of Aski (1995) on suppletion in Romance motion verbs. Conventional wisdom holds that the syncretism caused by the loss of accusative singular -m and vocalic mergers, especially of -o and short -u-, made the Latin case system unsustainable. This view, however, does not fully consider the plurals, especially the genitive. The se 2nd declension ending s in the 1st (-ārum) and 2nd declensions (‑ōrum) are is longer and less subject to syncretism caused by sound change. Unsurprisingly, some Romance personal pronouns languages feature reflexes of come from these genitive plurals among their personal pronouns (Table 1). Whether these forms constitute shared retentions is not clear. Maiden (1995: 113, 170), citing Rohlfs (1968: 122, 164), suggests a Gallo-Romance source for both the possessive and the indirect object uses in Italian. This path, however, would not explain why Italian shows the greatest expansion of these reflexes beyond the genitive, now serving three distinct plural functions — nominative, dative, and genitive plural. Two of the other three languages with reflexes of illōrum also show expansion to the dative plural. In Catalan, meanwhile, the form has moved the form out of the pronominal system and into the possessive system, whereas shown by its plural marking, like some other members (e.g., nostre “our”), it allows plural marking (but not gender marking). No adequate explanation of for the expansion of the reflexes of illōrum to the dative plural and nominative plural functions has been proposed. Another puzzle is the absence of Romance reflexes of the Latin feminine genitive plural, illārum. Goyette (2000) argues that the Romance tendency toward analytic structures, including the loss of illārum, is a sign of creolization. He cites such patterns as like Spanish de las puertas versus *laro portaro < illārum portārum, vis-à-vis the supposedly higher degree of synthetic expression in Greek. Joseph (2005) argues concisely but convincingly that Goyette disregards significant Greek developments, especially the periphrastic future and the periphrastic subjunctive constructions, that render Goyette’s argument untenable. The traditional explanation, sound change, does not account for the collapse of the Latin case system. Other approaches must be similarly meticulous; for example, a comparative solution suggesting creolization has to display employ great care when evaluating in the evaluation of comparative data and in the applyicationg of the current state of the art ininsights from creolistics to diachronic problems. When satisfactory answers are not immediately forthcoming, a rigorous approach requires persistent analysis that allows us to eliminate hypotheses in a principled way.

Table 1. Genitive plural relics in Romance pronouns (reflexes of illōrum in bold).

third person masculine case/number Latin French Romanian Italian Catalan nominative singular ille il el lui ell dative singular illī lui lui gli li accusative singular illum le el lo el/l’/-lo/’l nominative plural illī ils ei loro ells dative plural illīs leur lor loro li/els/-los/’ls genitive plural illōrum leur/leurs lor loro llur/llurs

ABSTRACT. In Modern German bare infinitives and those formed with the particle zu are both used as independent main clause predicates, where they each have illocutionary force. Contrary to previous claims that deny the zu-infinitive its potential to be used independently (Reis 1995, Deppermann 2006, Rapp & Wöllstein 2009, to name but a few) I argue that it has in fact emerged as a construction in its own right in interactive discourse. The following sentences exemplify the contrast in terms of illocutionary potential: (1) Die Zwiebeln braten! DIRECTIVE/EXCLAMATIVE the onions fry.INF (possible paraphrases: a. Braten Sie die Zwiebeln. 'Fry the onions!' b. Es ist absurd/lächerlich, die Zwiebeln zu braten. 'It's absurd/ridiculous to fry the onions!') (2) Die Zwiebeln zu braten! *DIRECTIVE/EXCLAMATIVE the onions PARTICLE fry.INF (possible paraphrase: Es ist absurd/lächerlich, die Zwiebeln zu braten. 'It's absurd/ridiculous to fry the onions!') Two points will be discussed in particular to support my claim: - The functional spectrum of the zu-infinitive is a subset of that of its sister construction, the independent bare infinitive, where zu is a disambiguator rather than a carrier of illocutionary force. - The independent use of zu-infinitives is constrained by formal requirements in terms of argument structure, which is not the case in bare infinitives, so that conventionalisation is further advanced. My study focuses on the specific illocutionary force of independent zu-infinitives and seeks explanation in the emergent conventionalisation of a disambiguating strategy, which appears to be a recent development during the New High German period. I propose that free-standing zu-infinitives in Modern German are characterised by special prominence of the semantic feature INDIGNANT (sp, prop), which expresses indignation on the part of the speaker with respect to the state of affairs encoded in the proposition. This meaning component is located on the attitudinal level of semantic structure, not the propositional level (cf. Jacobs 2016: 48). Bare infinitives may also express indignation, but unlike their zu-marked counterpart this is not the only possible reading. The restriction to speaker indignation is therefore not a feature of the particle zu, but rather of the entire construction. The particle as such can only be associated with disambiguating function. Another necessary condition for syntactically independent zu-infinitives with exlamative function is the presence of a constituent in pre-verbal position. (3) a. *Zu niesen!,*Zu lachen! PARTICLE sneeze.INF, PARTICLE laugh.INF 'How can you sneeze!', 'How can you laugh!' b. (Einfach) auf den Tisch zu niesen! (just) on the table PARTICLE sneeze.INF 'How can you sneeze on the table!' c. Ohne vorgehaltene Hand zu niesen! without held.before hand PARTICLE sneeze.INF 'How can you sneeze without covering your mouth with your hand!' This reflects the typical constituent order of verb-final subordinate clauses in German, but in independently used zu-infinitives it is less variable than in their embedded counterparts, as postposition of the additional constituents results in ungrammaticality in most cases. (4) a. *Zu kochen die Jeans! PARTICLE boil.INF the jeans 'How can you boil the jeans!' b. *Zu niesen (einfach) auf den Tisch! (see (3b)) c. *Zu niesen ohne vorgehaltene Hand! (see (3c)) Data extracted from the DeReKo (German Reference Corpus) shows that expressions such as Kaum zu fassen! ('That's hard to grasp!) or Nicht zu verstehen! ('That's incomprehensible!') display the highest degree of conventionalisation; here the propositional argument of INDIGNANT (sp, prop), i.e. the state of affairs the speaker expresses indignation about, is not explicitly spelled out. As a result the only possible use available for such expressions is an exclamation of general, unspecified indignation. Preliminary findings suggest constructionalisation through use of increments in interactional discourse (cf. Ford, Fox & Thompson 2002, Traugott in press). The specific illocutionary force results from an increasingly conventionalised form-function-relation between the structural pattern [X+zu+Inf] and the feature INDIGNANT (sp, prop), where X can be a variety of elements ranging from particles such as, for instance, the clausal negator nicht to a full-fledged NP or PP in either argument (as in (2b) and (3b) respectively) or adverbial function. A comparison with other Germanic languages will bring to light parallels and support the development of similar construction types. I will demonstrate that independent exclamatives in Modern English such as To think that I once was a millionaire! (Foolen 2012: 354) likewise contain the feature INDIGNANT (sp, prop) and are also subject to language-specific structural constraints. This warrants a closer investigation of the emergence and development of syntactically independent infinitive constructions within as well as outside of the Germanic subfamily and their potential to acquire specific indignant meaning.

References: Deppermann, A. (2006). Deontische Infinitivkonstruktionen: Syntax, Sematik, Pragmatik und interaktionale Verwendung. In: Günthner, S. & Imo, W. (eds.) Konstruktionen in der Interaktion. Berlin: de Gruyter. 239-262. Foolen, A. (2012). The importance of emotion for language and linguistics. In: Foolen, A. et al. (eds.) Moving ourselves, moving others – motion and emotion in intersubjectivity, consciousness and language. Amsterdam: John Benjamins. 349-368. Ford, C. E., Fox B.A & Thompson, S.A. (2002). Constituency and the grammar of turn increments. In: Ford, C.E., Fox, B.A. & Thompson, S.A. (eds.) The language of time and sequence. Oxford: Oxford University Press. 3-38. Jacobs, J. (2016). Satztypkonstruktionen und Satztypsensitivität. In: Finkbeiner, R. & Meibauer, J. (eds.). Satztypen und Konstruktionen im Deutschen. Berlin, Boston: de Gruyter. 23-71. Rapp, I. & Wöllstein, A. (2009). Infinite Strukturen: selbständig, koordiniert und subordiniert. Linguistische Berichte, Sonderheft 16. 159-179. Reis, M. (1995). Über infinite Nominativkonstruktionen im Deutschen. Sprache und Pragmatik, Sonderheft. Traugott, E. C. (in press). “Insubordination” in the light of the uniformitarian principle. In: Hoffmann, T. & Bergs, A. (eds.) Special issue of English Language and Linguistics.

ABSTRACT. The structural asymmetry between V2 main clauses and V-final subclauses is well attested in all historical stages of German, in spite of V1 and V3 main clauses in Old High German (Axel 2007, Donhauser & Petrova 2009, Tomaselli 1995) and Middle High German (Coniglio 2012). The rise and loss of V-final root clauses throughout the period of Early New High German (1350 – 1650) is hence unexpected from a diachronic point of view. Maurer (1926) was the first to point out the word order pattern in question, followed by Behaghel (1932), Ebert (1986) and Lötscher (2000). They agree that the emergence of V-final root clauses in Early New High German is due to Latin influence, in particular due to clause linkage patterns frequently involving relative pro-nouns. Demske-Neumann (1990) and Senyuk (2014) on the other hand argue that examples like (1) are instances of subclauses, considering the position of the finite verb as an unambiguous marker for the syntactic dependency of the clauses at issue, whereas the pronominal element may receive either a relative or a demonstrative reading. They show that the ambiguous d-adverbs are replaced by w-adverbs in V-final clauses in the course of the 17th century, resolving the ambiguity. In the present talk, I will consider V-final clauses introduced by a demonstrative, either by a pronominal adverb as in (1) or by a demonstrative DP as in (2), suggesting that we have to dis-tinguish different levels of clause linkage: Regarding the pragmatic level, the V-final clauses un-der consideration are independent of their matrix clause, providing their own information unit. As regards the syntactic level, however, the V-final clauses are dependent on their matrix clau-se. Evidence for their syntactic dependency comes from examples like (3), where the finite au-xiliary in the clause introduced by darauff 'then' does not appear in final position but is drop-ped. Assuming that V-final clauses introduced by a demonstrative are pragmatically indepen-dent, the question arises what triggers the verb placement in clauses like (1) and (4), the latter exhibiting the finite verb in second instead of final position. I will show that both word order patterns serve different discourse functions best accounted for in terms of foreground and background (Fleischmann 1973). In the present talk, I will secondly address the question of the alleged Latin influence. Contrary to the widely held view, it will be argued that word order patterns as (1) and (2) are motivated language-internally. Focussing on the pronominal adverb darauf 'then', we will see that its use as a temporal connector is fairly younger than its use as a pronominal adverb calling for the fini-te verb in final position when used as a relative (5a) or the finite verb in second position when used as a demonstrative (5b). In addition, the extremely frequent temporal connector do 'then' introduces clauses which are either V2 or V-final in Early New High German. As a temporal connector, darauf 'then' is used as a demonstrative in spite of the final position of the finite verb, just as the historical record attests V2 relative clauses (cf. Gärtner 2001 for V2 relative clauses in modern German). Support for this assumption comes from V-final clauses introduced by demonstrative DPs (2). In contrast to V2 relative clauses, V-final patterns as exemplified in (1) and (2) are no longer available in modern German. Their loss during the 18th century was presumably triggered by further sharpening the structural asymmetry between V2 main clauses and V-final subclauses, including the rise of w-adverbs.   Examples (1) Unnd do sy also bey ein ander in allermengklichs abwesens warend, empfieng herczog Wilhalmen Agleyen. Darauff sy czů im sprach:
 ’And when they finally met without any other people, Duke Wilhelm welcomed Agley. She then spoke to him:' (2) Do sprachen die freydigen zů der klaren junckfrawen, das sy sich mit kurczen worten erklagte, wie sy wo:elt, und den tod darnach lytte. Des selben urlaubs die magt auß der maßen fro ward. ’The fugitive people then spoke to the pure virgin that she might briefly mourn, how ever she wanted. She would suffer death afterwards. The maid was rather pleased about this delay.’ (3) Zu Parma hat es ein erschrecklich Wetter gehabt/ so Tag vnnd Nacht gewehret/ welches viel Bäum vnnd Kemich nidergerissen/ darauff es auch 6. Tag vnd Nacht ohne vnterlaß geregnet/ Sonst habe derselbe Hertzog ein Curier nach Span. vnnd viel Hauptleut vnnd Soldaten zur Guardia nach Prasenza geschickt/ 'There was an awful storm in Parma, which lasted day and night. Many trees and chimneys were destroyed, then it did rain for 6 days and nights incessantly. Otherwise the very same duke has sent a courier to Spain and many officiers and soldiers as guards to Praesenza' (4) "Nun helffent unnß auch, das wir fürbaß bey dem unsern bleiben und nit so la:esterlich darvon getrengt mu:essen werden rc.“ Darauff antwurten sy ir beyd, sy wo:elten ir helffen und ir leib und leben mit ir wagen. '"Please, help us now that we keep our possessions and that nobody makes us leave." They both answered then that they wanted to help her and to risk their lives with her' (5) a. Besunder sahe er eyn gezelt, darauf die baner Frigia stůnd. 'Especially, he saw a tent on which the flag of Frigia was shown' b. er hett eins menschen angesicht, darauf stůnd im ein kron von einem rubin. 'he had the face of a human being, over that was a crown made of rubies'

References Axel, K. 2007. Studies on Old High German Syntax. Linguistics Today 112. Amsterdam. | Behaghel, O. 1932. Deutsche Syntax: eine geschichtliche Darstellung. Bd. IV: Wortstellung. Periodenbau. Heidelberg. | Coniglio, M. 2012. On V1 Declarative Clauses in Middle High German. Linguistische Berichte 229, 5–37. | Demske-Neumann, U. 1990. Charakteristische Strukturen von Satzgefügen in den Zeitungen des 17. Jahrhunderts. In Neuere Forschungen zur historischen Syntax des Deutschen, ed. A. Betten, 239–252. Tübingen. |Donhauser, K. & S. Petrova. 2009. Die Rolle des Adverbs tho bei der Generalisierung von Verbzweit im Deutschen. In Gesprochen – geschrieben – gedichtet: Variation und Transformation von Sprache, ed. M. Dannerer, P. Mauser, H. Scheutz & A. E. Weiss, 11–24. Berlin.| Ebert, R. P. 1986. Histori-sche Syntax des Deutschen II: 1300-1750. Frankfurt/M.| Fleischmann, K. 1973. Verbstellung und Relieftheorie. Ein Versuch zur Geschichte des deutschen Nebensatzes. München. | Gärtner, H.-M. 2001. Are there V2 Relative Clauses in German? Journal of Comparative Germanic Linguistics 3:97–141. | Lötscher, A. 2000. Verbendstellung im Hauptsatz in der deutschen Prosa des 15. und 16. Jahrhunderts. Sprachwis-senschaft 25:153–191. | Maurer, F. 1926. Untersuchung über die deutsche Verbstellung in ihrer ge-schichtlichen Entwicklung. Germanische Bibliothek II/21. Heidelberg. | Senyuk, U. 2014. Zum Status rela-tivähnlicher Sätze im Frühneuhochdeutschen: Ein Beitrag zum Wesen der Subordination im älteren Deutsch. Doctoral Dissertation, Universität Potsdam. | Tomaselli, A. 1995. Cases of verb third in Old High German. In Clause Structure and Language Change, ed. A. Battye and I. Roberts, 345–369. Oxford.

ABSTRACT. Synchrony meets diachrony: Reconsidering convergence 

Rena Torres Cacoullos, Penn State University

 A long-standing issue in historical linguistics is the distinction between internal and external factors in change (Bybee 2015: 237-255; Labov 2007). The bilingual speech community provides the optimal window on contact-induced change, since languages come into contact through the people who speak them. Their speech regularly features a substantial complement of variability. By examining distribution and co-occurrence patterns, the variationist comparative method incorporates systematic variation into the traditional comparative method of historical linguistics (Poplack 2017). The working hypothesis is that integral to grammatical principles and language users’ knowledge is the covariation between linguistic elements that is probabilistic (Bresnan et al. 2001; Cedergren & Sankoff 1974; Labov 1969). The structure of variation—the way linguistic forms are used in discourse—then becomes the tool for measuring grammatical similarity or difference in order to assess first the existence of change, and then its source.

In this talk I capitalize on a new corpus, the New Mexico Spanish-English Bilingual corpus (NMSEB; Torres Cacoullos & Travis 2017) to contextualize candidate contact-induced changes socially and linguistically, i.e. with respect to both the speech community and the grammatical system in which they are embedded (Weinreich et al. 1968; Poplack & Levey 2010). The speakers are members of a longstanding community in New Mexico, USA, where Spanish and English have been in intense contact for over 150 years (Bills & Vigil 2008). Equal portions of Spanish and English are represented in the corpus, allowing for study of both languages, as speakers juxtapose them. Code-switching, illustrated in the excerpt in (1), is copious enough to qualify as a community discourse mode. These conditions make NMSEB the ideal locus to probe contact-induced change: Are grammars in contact different from grammars not in contact?

(1)

I don't have that ... energy ya para hacer aquí en la casa como Ø tenía más antes.... porque yo venía del trabajo,

I don't have that ... energy anymore to do things here at home like (I) had before.  ... because I would come back from work, (NMSEB, 04 Piedras y Gallinas, 48:44-48:51)

The linguistic variable of interest is subject pronoun expression (see (1)), a poster child for convergence of Spanish toward English (cf., Heine & Kuteva 2005:70; Silva- Corvalán 1994:145-165). A commonly entertained prediction is the extension of subject pronouns in the null-subject language on the model of the non-null subject language. However, to assess contact-induced change, similarities due to crosslinguistic tendencies, such as accessibility effects, must first be ruled out. Here we look to language-specific patterns, which enable us to diagnose convergence. Such is the prosodic-initial position restriction on unexpressed (null) subjects that we find for English. And, while the direction of effect for accessibility holds across both languages, accessibility manifested as clause linking turns out to be a relatively stronger probabilistic constraint in English than in Spanish. Variation patterns are compared across four datasets (Figure 1): monolingual benchmarks of both languages and the contact varieties as spoken by the bilinguals themselves. A total of nearly 700,000 words and 10,000 tokens of the variable are analyzed. The results show that bilinguals’ Spanish and English line up with their respective monolingual counterparts and most remarkably, are different from each other. Consideration of not only rates but also the conditioning of subject expression reveals no evidence of convergence: no movement of bilinguals’ Spanish towards English grammar, and no change in their English towards Spanish. Morphosyntactic change is thus far from an inexorable outcome of contact in bilingual communities, where speakers can independently apply language-specific grammatical principles.

References

Bills, Garland D., & Vigil, Neddy A. (2008). The Spanish language of New Mexico and southern Colorado: A linguistic atlas. Albuquerque, NM: University of New Mexico Press.

Bresnan, Joan, Dingare, Shipra, & Manning, Christopher D. (2001). Soft constraints mirror hard constraints: Voice and person in English and Lummi. In: M. Butt & T. Hollaway (eds.), Proceedings of the LFG01 Conference. Stanford: CSLI Publications. 13-31.

Bybee, Joan. (2015). Language change. Cambridge: Cambridge University Press.

Cedergren, Henrietta, & Sankoff, David. (1974). Variable rules: Performance as a statistical reflection of competence. Language 50:333-355.

Heine, Bernd, & Kuteva, Tania. (2005). Language contact and grammatical change. Cambridge: Cambridge University Press.

Labov, William. (1969). Contraction, deletion, and inherent variability of the English copula. Language 45(4):715-762.

Labov, William. (2007). Transmission and Diffusion. Language 83(2):344-387.

Poplack, Shana. (2017, To Appear). Borrowing: Loanwords in the speech community and in the grammar Oxford: Oxford University Press.

Poplack, Shana, & Levey, Stephen. (2010). Contact-induced grammatical change: A cautionary tale. In: P. Auer & J. E. Schmidt (eds.), Language and Space: An international handbook of linguistic variation, vol. 1. Berlin: De Gruyter. 391-419.

Silva-Corvalán, Carmen. (1994). Language contact and change: Spanish in Los Angeles. Oxford: Clarendon Press.

Torres Cacoullos, Rena, & Travis, Catherine E. (2017, To Appear). Bilingualism in the community: Code-switching and grammars in contact. Cambridge: Cambridge University Press.

Weinreich, Uriel, Labov, William, & Herzog, Marvin I. (1968). Empirical foundations for a theory of language change. In: W. P. Lehmann & Y. Malkiel (eds.), Directions for historical linguistics: A symposium. Austin, TX: University of Texas Press. 95-188.