Why some beliefs of the transformational linguists are unbelievable

Abstract

On occasion, transformational linguists write about the psychological reality of the rules that they produce for their grammers. Assuming such rules have psychological reality, some curious facets of the theory come to light. In particular, the well-known and accepted series of rules which constitute the lexicalist hypothesis would require belief in the fact that the human brain stores such obviously related pairs as careful/carefully as separate entities and does not relate them. Similarly, the transformationalist's position would require that lexical units manifesting varying degrees of polysemy would also be stored separately. This would then make it impossible to derive ambiguous strings since the individual lexical units are subcategorized in such a way that their distribution is mutually exclusive. Both of these beliefs seem to be incorrect since, on the one hand, the brain must store and then produce items such as careful/carefully in a maximally efficient manner and, on the other hand, it can obviously produce intentionally ambiguous strings such as puns. At both of these points, then, the rules produced by a transformational grammar cannot be psychologically real.