SOME COMMENTS ON COMPETENCE AND PERFORMANCE

If you are ever asked to be a discussant, I’ll tell you how you d o it. You go off and gather your own thoughts; then you hope that somebody says something that will make one of those thoughts relevant to the occasion. It will turn out that some of the things I’ll have to say now will be relevant to what you’ve already heard, but that’s just my good luck. How can we best understand the notion of linguistic competence in order to account for linguistic performance? In collecting my thoughts about competence and performance, I began with a strong suspicion that these words mean different things to different people, and that much of the disagreement that this distinction has inspired in recent years could be understood if we thought in terms of people who believed they were talking about the same thing but really weren’t. So, as a first step, I began listing alternative uses of the terms “competence” and “performance.” Without even trying very hard, 1 got up to eight different versions of the distinction. I will review them for you briefly. Two of them, 1 believe, are historically established. At the time Chomsky formulated his version, and cast it in terms of competence and performance, these were precedents that he could (and in fact did) cite. The principle precedent was a linguistic one. Sometime before 1910 de Saussure had distinguished between langueandparole, i.e., language and acts of speaking, in terms of a distinction between a social norm and individual manifestations of that norm. His purpose was to give primacy to the study of language as a way to unify a diverse range of physical, social, and psychological facts. I’ve always wondered to what extent de Saussure was influenced by the sociological distinction, current in Paris in those days, between collective representation and individual representation. But I don’t know any historian of science who has dug into that. Another precedent that existed was what I would call a rognifive version. Lashley, as early as 1929, and Tolman many times through the 1930s, distinguished knowledge and performance in terms of what an organism had learned versus what he might be motivated to d o with his knowledge on any particular occasion. That is, a rat could have learned where the food bin was and, in some sense of the word “know,” could know where the food bin was, but not go there if he wasn’t hungry. 1 think their purpose was to deepen some rather superficial behavioristic theories that were popular in their day by introducing theoretical variables that could not be directly observed in the behavior but, perhaps, could be abstracted or inferred from it. Chomsky’s inspiration was largely from Saussure, but he was aware of the cognitive distinction as well, and I think he tried to subsume them both in his own view of the matter. 1 would call Chomsky’s view a rafionalistic view of the distinction. This distinction was initially developed in our chapters in the Handbook of Mathematical Psychology ( 1963). In those chapters Chomsky distinguished between competence and performance, between the language that a person knows and the speech acts that a language user performs. This was done in terms roughly analogous to the distinction between, say, an axiom system and various realizations of such a system.