Situated Action: Reply to William Clancey

All but one of our commentators stress (as we did) the diversity of forms taken by the doctrine of situated action. The exception, William Clancey (1993), embraces and reifies a holistic version of the SA doctrine that brings together a trinity, consisting of (1) the whole environment, (2) the brain depicted as a network of neurons, and (3) the "agent," into a unified process called "human activity." This environment-brain-agent trinity somehow achieves an "immediacy" of activity that bypasses any need for symbolic mechanisms. It converts situation and intension instantly into action, avoiding thereby the inflexibilities and limits to which symbol systems, according to Clancey, are subject. In particular, Clancey argues that the actors in this drama need no symbolic models to convert their intensions into actions, and that models are not required for the occurrence and reproduction of social interactions. We will show that there are three major and fatal flaws in Clancey's position, as well as a number of lesser ones. First, he consistently misconstrues the meaning of the word "symbol," restricting that term to linguistic objects, wholly contrary to the practice in symbolic modeling of cognitive processes. Second, he commits a fundamental "category error," the exact converse of the one he mistakenly accuses us of. Third, in his preoccupation with the "immediacy" of action, he deprives SA of any operationally defined and testable set of mechanisms that might account for the system's behavior and be susceptible to confirmation or falsification through observation and experiment. The definitional error persists throughout Clancey's article. He nowhere refers to anything as symbolic that is not linguistic. Clancey has the usual right to define terms in any way he pleases; but this does not entail the right then to characterize the symbolic systems we describe in our article as also "symbolic" in his much more restricted sense of the term. Most of his