Studies in Abstract Phonology

The issue of abstractness is currently one of the central questions in the ongoing research into the theoretical foundations of modern phonology. Linguists are divided on the degree of abstractness admissible in phonological representation. While some would impose far-reaching restrictions on admissible abstractness, others have felt that such restrictions are unrealistic. This monograph offers extensive arguments for admitting a high degree of abstractness in phonology. The empirical material on which the monograph is based is the phonology of modern Polish. A substantial portion of the phonological rules of the language are studied and their theoretical implications are investigated in detail. Although there exists a rich literature on Polish phonology, very little of it has been written with clear theoretical issues in mind, and practically none of it is in English. The results obtained in the monograph, especially with regard to the vowel-zero alternations hold not only for Polish, but with some straightforward modifications can be shown to hold for all other Slavic languages as well. The book consists of four chapters. The first chapter examines critically and rejects the restrictions on abstractness in phonology that have been advanced by proponents of Natural Generative Phonology, in particular Hooper's &"true generalization condition.&" The second chapter considers the vowel-zero alternations in Polish and shows that the facts require an abstract analysis including the postulation of two abstract vowels that do not appear directly in phonetic forms of the language. The proposed vowel inventory is exploited in the third chapter which deals with one of the most complex problems of Polish phonology, the alternations of the nasal vowels. The proposed vowel inventory also plays a central role in the final chapter, which deals with the [o-u] alternations.