Processing instructions, markedness, and congruity effects in a sentence verification task.

Abstract

Subjects were required to decide whether sentences like treaty passed were true or false, given the number of votes cast for the bill and the criterion that determined its status. There was variation across experimental conditions in whether the votes shown to subjects were positive (votes for) or negative (votes against) and in whether pass or fail was stressed in defining the status of the bill. An additive-stages model was applied to the verification times from the present experiment and previously reported studies, and was used to describe certain markedness and congruity effects. It is suggested that congruity effects can be attributed to the similarity between critical stimulus values and (a) noncritical or intrusive stimulus values or (b) the content of the processing instructions given to subjects. These effects may be different for affirmative and negative stimuli, and may be observed early or late in processing. The use of the sentence verification task, in which subjects determine whether a sentence correctly describes a picture (Clark, 1969), has been an effective method for studying the influence of variables such as lexical markedness and negation (e.g., Carpenter & Just, 1975; Clark & Chase, 1972; Trabasso, Rollins, & Shaughnessy, 1971). The study of these effects has led to a description of some aspects of sentence comprehension (e.g., how subjects represent sentences) and the processes by which the information in a sentence is related to its referents (Carpenter & Just, 1976). This description is model based inasmuch as typical research strategy has been to fit subjects' data, usually the time taken to respond true or false to a sentence or picture,