Past Event

Learning to change

Dr. Ailís Cournane
(New York University)

Child language development necessarily involves grammatical change over time. I argue this process of change is the same one that creates the innovative variants that may then diffuse into the speech community and become the language change we see in the historical record.

Historically, modal expressions (/must, might, have to,…/) are known to change unidirectionally from main verbs to auxiliaries (e.g., Lightfoot 1979; Roberts 1985), and from root meanings (ability, obligation, permission) to future or epistemic meanings (knowledge-based inferences) (e.g., Traugott 1989; Bybee et al. 1994). For over 100 years, linguists have appealed to the child innovator to explain these systematic language changes (Paul 1920 [1880]; Meillet 1958 [1912]; Andersen 1973; Halle 1964,i.a; c.f., Dasher & Traugott 2005; Diesel 2011, i.a). However, the child innovator proposal has not been tested in the domain where it makes predictions: child language.

In this talk, I provide a testable linking hypothesis based on the syntax-semantics mapping problem (see Clark 1977, 1993; Bloom 2000). The mapping problem refers to a fundamental question in language acquisition: How do children learn which forms in their input associate with which meanings? The child must make (and continually revise) hypotheses about which meanings map to which forms. In doing so she relies on evidence from her input, but also on internal properties of her emerging grammar, including mechanisms of extension and generalization. I hypothesize that essential child meaning extensions sometimes /over/extend (see Clark 1977; Bowerman 1982, i.a.) and create new mapping relationships directionally consistent with historical change.

I provide evidence from longitudinal and experimental acquisition studies (Cournane 2014, 2015a,b, 2016; Cournane & Pérez-Leroux, in prep; Cournane & Hacquard, in prep) showing that: (a) child innovative mapping relations (=overextensions) for modal expressions are directionally consistent with predictions from modal change (i.e., Root > Epistemic), and (b) development in both the child and the language involve /grammatical/ re-mappings, consistent with formally constrained syntax-semantics mapping relations for modal expressions (e.g., Hacquard 2006, 2013; c.f., dominant /conceptual/ views, which I show make inaccurate predictions; e.g., Moore et al. 1990; Diessel2011).

About the Presenter

Dr. Ailís Cournane is a recent PhD graduate from the University of Toronto, where she worked on child language acquisition and its relationship to linguistic change. She is now Assistant Professor at NYU.