Adverbial subordination at the peripheries of the Andean and Chaco linguistic areas

GOLLUSCIO, LUCÍA; HASLER, FELIPE; WILLEM DE REUSE

Minnesota

Encuentro; The Society for the Studies of the Indigenous Languages of the Americas (SSILA)/Linguistic Society of America Annual Meeting; 2014

SSILA / Linguistic Society of America

Following Comrie et al. (2010), we assume that linguistic areas are the result of dynamic processes with shifting boundaries. In this paper we first examine features of adverbial subordination in languages at the peripheries of two South American linguistic areas: the Chaco area and the Andean area: Vilela (considered part of Lule-Vilela, although this family is not completely established), at the margins of the Chaco area, and Mapudungun (language isolate), at the margins of the Andean area. We then consider Lule (Lule-Vilela) a language with an unusual typological position with respect to both the Andean and Chaco areas. Vilela, a extremely endangered language, was spoken in the Argentine Chaco. However, recent work (Golluscio and Nercesian 2011; van Gijn et al. 2012) has highlighted differences between Vilela and other Chaco languages regarding adverbial subordination: while core Chaco languages use finite verbs, Vilela prefers a converbal construction (Golluscio 2009-10). This strategy links it to the Andean languages in which nonfinite forms predominate. Consider the contrast between Wichí (1) (Mataguayan, a core Chaco language) and Vilela (2) in the following temporal adverbial clauses: (1)wit kanu takasit [tox Takfwax t?ospe? kan] CONJ needle 3.SUBJ:stand up CONJ Takfwaj 3.SUBJ:step on needle ?And the needle stood up when Takfwaj was about to step on it?. (Nercesian 2011:443) (2)[Hate na-l] nie xope umo-e man come-CONV fire pot-CL put-3.SUBJ ?When the man came, he put the pot on the fire?. (Golluscio 2009-10:267) Mapudungun shows a significant difference in the formation of adverbial clauses in relation to the rest of the languages of the Andean area: while these exhibit a switch-reference system, Mapudungun has no switch-reference at all. For example, the Quechua suffix -pti (3) indicates that the subject of the subordinate clause differs from that in the main clause, whereas the Mapudungun suffix -lu (4) does not indicate anything about the identity of the subjects involved, as can be seen in the following temporal adverbial clauses: (3)[kay-pi ka-pti-y=qa] allin-ta=m llamka-n this-LOC to.be-DS-1.POS=TOP good-ADV=ASS work-3.SUBJ ?He works well when I am here?. (Soto Ruiz 1976:157, cited in Muysken 2011:135) (4)[aku-lu iñche], Pedro amu-tu-y arrive.here-FNF2 I Pedro go-RE-IND.3.SUBJ ?When I arrived, Pedro returned (Héctor Mariano, personal communication) We will also describe adverbial clause formation strategies of manner, purpose and instrument in Vilela and contrast them with what is found in core Chaco languages: Toba (Guaycuruan), Wichi (Mataguayan) and Ayoreo (Zamucoan). We compare Mapudungun in a similar fashion with core Andean languages of different families: Quechua (Quechuan), Uchumataqu (Uru-Chipaya), Puquina (unclassified) and Millcayac (Huarpean). Finally, we focus on the special case of Lule, an extinct language, originally spoken right on the boundaries of the Andean and Chacoan areas (although considered to belong to the Inca Sphere by Adelaar with Muysken (2004)). Lule adverbial clause formation strategies show a mix of core Chaco areal features, as well as features of Mapudungun, but unlike Mapudungun (and other Andean) Lule has very few nonfinite forms. We conclude that Lule, even more than Vilela and Mapudungun, does not fit well in the areal typology of South America, and that it is thus a useful corrective to overly essentializing views (e.g. ?lowland? versus ?highland?, or ?Amazonian? versus ?Andean?) of the complex interrelationships of linguistic areas in South America.