Linguistic Variation - Volume 12, Issue 2, 2012

In this paper, I propose that switch-reference in Quechua can best be analyzed as agreeing tense. Given the properties of the Quechuan switch-reference system and the clause structure of Quechua, I assume that a switch-reference adverbial clause does not have a valued tense feature and must agree with its superordinate clause in tense. Tense agreement is only possible if the subjects of both clauses are identical. The same subject marker is analyzed as the realization of successful tense agreement, while the different subject marker is the realization of failed agreement. I argue that this approach to switch-reference is conceptually as well as empirically compelling and raises the interesting question as to whether switch-reference is a true morphosyntactic category or not. Keywords: switch-reference; Quechua; Agree; minimalist framework; non-local dependencies

In this paper I discuss three empirical generalizations about local scenarios: (i) four different realization strategies found cross-linguistically, (ii) an asymmetry in the number of arguments encoded on the verb in languages with person hierarchy-effects in non-local scenarios and (iii) the fact that person portmanteaux are particularly prominent in local scenarios. I claim that all three generalizations can be derived if Agree is relativized to target only positively valued person features on a goal. (ii) falls out directly from the Agree mechanism. (i) is a purely morphological phenomenon arising from the specification of local person exponents. Person portmanteaux are analyzed as inclusive markers in a derived inclusive context. Such a context can only emerge in local scenarios, which derives (iii). In general, the paper addresses the question how morphological theories that rely on discrete slots can handle portmanteaux by vocabulary insertion, without additional mechanisms like e.g. fusion. Keywords: Agree; relativized probing; person agreement; local scenarios; portmanteau morphemes; hierarchy effects; underspecification

While it is clear that some languages have a grammatical mass/count distinction (e.g. English), in other languages (e.g. Inuktitut) it is not so obvious. In this paper, I show that Inuttut (Labrador Inuktitut) has a subtle grammatical mass/count distinction: while number, numerals, and most quantifiers do not disambiguate between mass and count nouns, in a few places, the morphology or the semantics disambiguates between mass and count. Thus, Inuttut is not a counterexample to Doetjes (1997) or Chierchia (2010), who both argue that all languages distinguish between mass and count. I further argue against Borer (2005) who claims all nouns in all languages are underlyingly neutral, and are assigned mass interpretation by default in the absence of individuation. I show that Inuttut nouns cannot all be underlyingly neutral and/or mass by default. Keywords: mass; count; classifiers; Inuktitut; number; numerals; quantifiers

By comparing two sign languages of approximately the same age but which arose and developed under different social circumstances, we are able to identify possible relationships between social factors and language structure. We argue that two structural properties of these languages are related to the size and the heterogeneity versus homogeneity of their respective communities: use of space in grammatical structure and degree of lexical and sublexical variability. A third characteristic, the tendency toward single-argument clauses appears to be a function of a different social factor: language age. Our study supports the view that language is not just a structure in the brain, nor is it strictly the domain of the individual. It is very much a socio-cultural artifact. Keywords: community and language structure; sign languages; ISL; ABSL; variation; space; argument structure

The reduction of vowels is a popular topic for research, but little has been said about the effects of vowel harmony on vowel reduction. Gendrot and Adda-Decker (2006, 2007) claim that phonetic reduction is linked to the phonetic duration of the vowel so that in short syllables, the vowel converges towards a schwa-like quality. I support this claim in general, but I add the claim that reduction is blocked in vowel harmony domains. Within a harmony domain, even in vowels of short duration, the quality of the vowel retains the quality of the feature which is spreading. This paper demonstrates this with data from Kera, a Chadic language spoken in Chad. I examine fronting, rounding, and height harmony in Kera and confirm the fact that harmony (but not accidental agreement) in some way blocks the reduction. This has implications for the role of harmony and claims concerning privative features. Keywords: phonology; reduction; Kera; vowel harmony; acoustic; formants; duration; blocking; privative