wledges the utility of the symbology implied by the form the vocabulary has taken as it has moved toward the status of Modern French. In the general scheme of things, he observes that vowel nasalization occurs in Modern (as opposed to Old or Middle) French. Although it must be said that on the whole Tranel accepts de facto that vowel nasalization is a fundamental element of French, it can also be said that he is far more concerned to explain the rationale behind what he sees as the more or less orderly way in which such nasalization takes place. He asks, for example, under what conditions lexical nasal vowels can be postulated (5); however, his own analysis shows that in large part such a postulation is a function of lexical representation. That is, such representation is a concrete element of, but not dependent upon, the customs of pronunciation. In particular, he criticizes abstract generative analysis, which treats nasal vowels as aspects of grammatical structure, and which seeks to reduce description of such vowels to certain rules of usage and to predict their nasalization based on their interaction with adjacent consonants. The difficulty with predicting whether a French vowel will be nasal or oral in pronunciation based on various grammatical rules is that, inevitably, there must be exceptions to the rules, and that the exceptions are so frequent and so varied in nature that the rules are essentially invalidated. As he puts it,
One of the main generalizations maintained by the abstract
generative analysis is that French nasal vowels are not
lexical. This claim creates two difficulties which, taken
together, constitute somewhat of a paradox because they
reflect antithetic problems concerning phonological
representations. On the one hand, since a given nasal vowel
may come from several underlying /VN/ sequences, the lexical
representation of a nonalternating nas...