Conflicting Constraints in Resource-Adaptive Language Comprehension

The primary goal of psycholinguistic research is to understand the architectures and mechanisms that underlie human language comprehension and production. This entails an understanding of how linguistic knowledge is represented and organized in the brain and a theory of how that knowledge is accessed when we use language. Research has traditionally emphasized purely linguistic aspects of on-line comprehension, such as the influence of lexical, syntactic, semantic and discourse constraints, and their time-course. It has become increasingly clear, however, that nonlinguistic information, such as the visual environment, are also actively exploited by situated language comprehenders. The wealth of informational resources which are potentially relevant to situated language comprehension raise a number of important questions. To what extent are the mechanisms underlying comprehension able to exploit linguistic and non-linguistic information on-line, and how do people adapt to the availability or non-availability of contextual information? We begin below, with a brief summary of several important aspects of human language comprehension, including its incremental and even anticipatory nature, and its sensitivity to accrued linguistic experience. We then present a range of experimental findings which reveal the ability of comprehenders to rapidly adapt to diverse linguistic and non-linguistic constraints. To better understand this apparently seamless ability of the human language processing faculty to integrate diverse cues, including linguistic context, intonation, world knowledge, and visual context, many of the experiments are designed so as to better understand the relative priority of these constraints when they are pitted against each other. The findings conspire to paint a picture in which purely linguistic constraints, long thought to identify the core of sentence comprehension mechanisms, can in fact be overridden by highly contextual aspects of the situation, such as the intonation contour of a particular utterance, semantic expectations supported by the visual scene, and indeed events going on in the scene itself.