In English, it is possible for a morphologically singular collective noun like government to control both singular (syntactic) agreement and plural (semantic) agreement in the same sentence (e.g. The government has praised themselves). It has been claimed that sentences with the opposite pattern of agreeing elements are ungrammatical (e.g. *The government have praised itself), and there is a corresponding asymmetry in corpus frequencies of these two configurations. Across two acceptability judgement experiments, we show that the acceptability contrast is affected by the relative order of the two agreeing elements, with degraded acceptability in the case where the first agreeing element shows plural agreement and the second shows singular agreement, relative to the opposite configuration. This pattern is found both when the agreeing verb precedes the reflexive, and when the reflexive precedes the verb. Overall, the results suggest that the initial formation of a semantic agreement dependency between an agreement target and a collective controller makes subsequent morpho-syntactic agreement with the same controller less accessible. We argue that any theoretical account of these results would require an important role for incremental processing.  

In incremental parsing, infinite local ambiguity occurs when the input word can be combined with the syntactic structure built so far in a infinite number of ways. A common example is left recursion (e.g. "'railway station clock" or "his sister's boyfriend's shirt"), where local information cannot tell us the depth of embedding of the left descendent chain of nodes. From the processing point of view, infinite local ambiguity causes a technical problem, which a model must solve in order to implement incrementality fully. This paper provides a general solution to the problem of infinite local ambiguity, by introducing the concept of Minimal Recursive Structure. We give two examples of parsers in which the solution is used.

This study reports one eye-tracking and one sentencecompletion study investigating the antecedent biases ofTurkish-speaking L2 speakers of English, for anaphor it anddeixis this. Our results show L2 speakers displayed native-like sensitivity to the type of antecedents while using it andthis in sentence completion, but this sensitivity was notreplicated in our online reading experiment. This showslimitations in L2 speakers’ use of information in onlinereading, and poor performance in making use of pragmaticchanges in context to track the antecedents of it and this.

Recent work has emphasised the importance of thematic domains in sentence processing. T w o questionnaire studies examined whether thematic domains influence attachment of relative clauses to complex NPs in Japanese. The results suggest that definitions of thematic domains should be revised to cover left-branching stmctures, but do not support a distinction between domains associated with clauses and adpositional phrases.