Experiments on competence and performance

Studies in experimental linguistics must take into account the difference between a speaker´s competence (his ability with regard to the grammar of his language) and his performance (his ability with regard to the production and perception of particular utterances). Where these differences have been ignored, wrong assumptions have led to inadequate models. This paper discusses some of the important distinctions between the two, and suggests that experiments can be designed to test hypotheses relating to either or both. Three experiments are described. The first is concerned with how subjects would pronounce nonsense words; the results are predictable from a model of their linguistic competence. The second involves the auditory perception of temporal intervals in speech; the data show that a subject´s responses are influenced by both his competence (his stored grammatical knowledge) and his performance (his immediate perception in the particular circumstances). The third experiment investigates aspects of the control of the vocal organs which are independent of the language being spoken; it is therefore entirely concerned with the speaker´s performance.