In Rules and Representations (1980), Chomsky provides an elegantly simple exposition and defense of the concept of psychological reality. This idea, so central to his conception of what linguistics is about, has, Chomsky feels, been seriously misunderstood by its critics. According to Chomsky, to say that a grammatical hypothesis is psychologically real is to attribute to it the same characteristic that one attributes to a physical theory by saying that it is physically real. Chomsky urges that the same scientific realism which is appropriate for physics is appropriate for psychology. The hypotheses constructed within a science should be understood as making claims about underlying causal mechanisms, and not simply as providing fictional instruments which license inferences about observable behavior. For Chomsky, a hypothesis is psychologically real just in case it is psychological in its subject matter and true. There is nothing more to the issue of psychological reality of a hypothesis than the question: Is this hypothesis the best explanation of the observed phenomena? If the hypothesis is, then it is entitled to be called true, and a general scientific realism then leads one to take the hypothesis literally-as describing unobserved processes, properties, and entities. Chomsky thus tries to bring the idea of psychological reality down to earth by anchoring it to the issue of adequate explanation; the psychological reality of a theory is not some occult quality it may have which is still in question after its explanatory adequacy is assessed.
This analogy between psychological reality and physical reality leads Chomsky to group those who have criticized his use of the concept into two classes. First, there are those, like Quine, whom Chomsky sees as thinking that there is something special about psychology which blocks a realist interpretation of it. Chomsky interprets Quine as making demands on psychological theories which he, Quine, would not dream of placing on a physical theory. Quine (1960) is quite happy to allow that two predictively equivalent physical theories may be incompatible, but when two psychological theories are predictively equivalent, Quine will say Linguistics and Philosophy 3 (1980) 395-405. 0165-0157/g010033--0395501.I0.