Syntactic island constraints are generally viewed as paradigmatic evidence for the autonomy of syntax. However, the existence of exceptions has been known for some time. Close examination reveals that these exceptional phenomena resist purely syntactic explanations, requiring instead an account in terms of semantic relations like attribution, various kinds of semantic framing effects, and discourse variables like topic and focus. It will be argued that the syntactic and extrasyntactic factors which limit extraction can be subsumed into a general account based on a cognitive theory of attention. According to the analysis presented below, extraction phenomena represent a situation in which the language user must attend simultaneously to two parts of the syntactic structure; a situation which strains the limited working memory available for automatic syntactic processing. Long-range extraction takes place under conditions which reduce that processing strain, that is, when both the extracted element and the matrix for extraction command attention anyway. In other words, the factors which control the distribution of long-range extraction do so indirectly through their impact on the distribution of attention.
No takes yet. Share an insight, caveat, or question.
Paul Deane (1991) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: