Inferring Indirect Objects in Sentences: Some Implications for the Semantics of Verbs

Abstract

Certain verbs imply specific instruments, and readers appear to make such inferences when reading the verb (Garrod and Sanford, 1981). For other verbs the implications seem much weaker and no specific inferences are made when reading the verb (Singer, 1979). The present studies explore the reasons for this disparity. A study of dictionary entries for the two kinds of verbs suggested that instruments for Garrod and Sanford verbs are a more inherent part of the definition than they are for Singer verbs. However, two measures of verb-instrument relation, (instrument generation tasks), showed no difference between the two sets of verbs. Experiment I, using a mixture of verb types, excluded methodological factors as the reason for the different findings of Garrod and Sanford and Singer. It pointed to there being verb differences which are tapped by a reading task and not by questionnaires, where there are no time pressures. Experiments II and III measured the strength of the verb-instrument relation using a "word-recognition" task and found that one set of verbs (Garrod and Sanford) seemed to be more closely linked with particular indirect objects than did the other set (Singer). In Experiment IV, reading was slowed by the introduction of an uncommon instrument only after mentioning a verb from the set having stronger verb-instrument links (Garrod and Sanford verbs). Experimental results, together with the dictionary study, supported the idea of "semantic-relatedness" for Garrod and Sanford verbs, and their instruments contrasted with "association" for Singer verbs. Experiment V examined the relationship between a category and its noun exemplar. Reading was slowed by the introduction of an uncommon exemplar suggesting that a category is instantiated at the time of reading. This effect was stronger than that found previously for verbs suggesting that verbs and nouns differ in the structure of their semantic representations.