Uninterpretable features in learning and alternative grammars?

Abstract

Walkden and Breitbarth employ historical data in order to test a conjecture expressed by Trudgill (2011) regarding a link between linguistic complexity and the language-contact situation: namely, “short-term contact involving extensive adult second-language (L2) use is predicted to lead to simplification” (W&B). Specifically, the authors address this conjecture with respect to syntactic complexity. They operationalize syntactic complexity as universal (i.e. L1-independent) L2learner difficulty. Drawing on L2 research in the generative tradition, W&B use a specific theoretical concept supposed to faithfully capture L2-difficulty: Interpretability Hypothesis, which states that (assuming a broadly MinimalistFramework analysis) uninterpretable features are not easily available to L2learners. This should lead to syntactic analyses involving uninterpretable features being harder to learn, which in turn should facilitate patterns of language change where uninterpretable features get lost. W&B discuss a number of cases, primarily concerning Jespersen’s cycle in Germanic and Romance, that suggest that their explanation is compatible with the data. The program outlined by the authors is ambitious and promising; the shortcomings of the current argument as it stands should not detract from that. Though the authors employ a syntactic characterization relative to a single kind of syntactic theories, uninterpretable features in their framework would likely correspond to other specific types of syntactic properties in other traditions; briefly put, an uninterpretable feature indicates the presence of a purely formal syntactic relation, such as agreement, movement or movement-like relation, etc. It is reasonable to argue that the presence of such (obligatory) relations corresponds well to the intuitive notion of syntactic complexity. The authors only provide cases where uninterpretable features got lost, under their analyses, and cite roughly contemporaneous contact circumstances that approximately agree with the circumstances described in Trudgill’s conjecture. This does provide tantalizing support for W&B’s theory, in that the data are compatible with their explanation. It does not at this point provide