Abstract

Abstract

Linguists build theories of grammar based largely on acceptability contrasts. But these contrasts can reflect grammatical constraints and/or constraints on language processing. How can theorists determine the extent to which the acceptability of an utterance depends on functional constraints? In a series of acceptability experiments, we consider two factors that might indicate processing contributions to acceptability contrasts: (1) the way constraints combine (i.e., additively or super-additively), and (2) the way a comprehender’s working memory resources influence acceptability judgments.

Results suggest that multiple sources of processing difficulty combine to produce super-additive effects, but multiple grammatical violations do not. Furthermore, when acceptability judgments improve with higher working memory scores, this appears to be due to functional constraints. We conclude that tests of (super)-additivity and of differences in working memory can help to identify the effects of processing difficulty (due to functional constraints), but that such tests are limited in contexts of extreme processing difficulty.