Micro-variation in the Andes: Two Case Studies
Neil Myler (BU)—Morphology, Syntax.
This talk reports on two ongoing projects which at first sight might seem to have nothing in common except one of their subject languages. However, they are united by their use of variation as a window onto the workings of the morphosyntactic component of the grammar.
The first part of the talk looks at micro-variation in the interpretation of high applicatives in the Quechua family. The standard approach to high applicatives (Pylkkänen 2008:17, her (13)) would involve postulating a range of lexically distinct Appl heads with different meanings from which different languages “choose”. I argue instead that the micro-variation seen in Quechua is better analyzed by assuming (i) that High Appl itself is a semantically vacuous piece of syntactic scaffolding, which serves only to license an oblique argument; (ii) that variation in the apparent meaning of Appl instead arises from variation in the oblique case markers/adpositions that can accompany the applied argument in spec-ApplP; and finally (iii) that cross-linguistic (and language-internal) variation in what High Appl can mean reduces to (micro-)parametric variation in what is allowed to occupy spec-ApplP. I go on to show how this perspective solves a number of enduring puzzles in the literature on High Appl, including a novel explanation for Mirror Principle-violations involving these heads.
The second part of the talk reports on joint work with the sociolinguist Daniel Erker. We are currently carrying out a pilot study of Quechua-Spanish bilinguals in Cochabamba, Bolivia using a new corpus of sociolinguistic interviews, grammaticality judgment data, and reading passage data to examine a range of phonological and morphosyntactic variables. Quechua-Spanish bilinguals, as well as some monolingual speakers of Andean Spanish, exhibit word-order variation in each of their languages induced by contact between Spanish (a head-initial language) and Quechua (a predominantly head-final language). I outline an approach to such variation inspired by the notion of “combinatorial variability” (Adger 2006 et seq.), which yields quantitative predictions about the frequencies of surface orders purely from the interaction of independently existing parameter settings. I then show that the quantitative predictions change if the Final-over-Final Constraint (Biberauer, Holmberg, and Roberts 2014), a proposed universal ban on certain kinds of disharmonic word-order, is assumed to hold. This is an interesting conceptual result, because it shows that categorical UG constraints can, in principle, make probabilistic predictions about usage.