Testing Two Basic Predictions of PM

PM, Propositional Model, is a highly interactive model of language comprehension which can be contrasted with language processing models which assume an autonomous syntactic component. Models assuming an autonomous syntax can be divided into two basic types: (a) Strong Autonomy Models which assume that only the part of speech of lexical items is available to the syntactic analyzer, and (b) Weak Autonomy Models which assume that information other than the part of speech, but of a purely syntactic nature, is available to syntactic analyzer. This paper presents the results of two experiments using the cumulative self-paced reading task which are intended to distinguish PM from both Strong and Weak Autonomy Models. The first experiment considers the influence of verb argument preferences on the initial determination of structure. The existence of such influences would argue against the Strong Autonomy Model and in favor of Weak Autonomy or Interactive Models. The second experiment considers the influence of object schematicity on the initial determination of structure. An influence of object schematicity argues against the Weak Autonomy Models and in favor of Interactive Models like PM. The results of the two experiments do indeed show effects of verb argument and object schematicity preferences. If the cumulative self-paced reading task is a legitimate measure of immediate, on-line processing, then Interactive Models are supported by these experiments and Autonomous models are not.