Hungarian Vowel Harmony

A system of phonological rules and representations has at least two aims: one is to describe the way in which words (and larger stretches of utterances) are realized in speech, and to state which logically possible streams of speech sounds are legitimate in the language under description – this is part of observational adequacy. The other aim is to capture the pattern of sounds characteristic of the language in question – this is part of descriptive adequacy. Phonological feature systems play a pivotal role in the realization of these two aims: on the one hand, they contribute to observational adequacy by virtue of their phonetic interpretation, which ties the abstract units (phonemes) employed in phonology to articulatory (and perhaps perceptual) correlates. On the other hand, they contribute to descriptive adequacy by classifying these units into ‘natural classes’ according to the phonemic patterning of the language. In Section 1 I will argue that in the case of Hungarian vowels, the aims of observational and descriptive adequacy are in conflict, because the qualitative differences between short and long vowels are only phonetic: the phonological patterning of Hungarian vowels makes it necessary to treat long vowels as geminates. This result will not only call into question the existing treatments of Hungarian vowel harmony, but also raises a larger question: if phonetic facts do not serve as an infallible guide in feature analysis, what is the content of the features? I will argue that the substantive content of phonological theory is largely independent of the intrinsic content of the features, and that the predictive power of the theory lies with the restrictions it puts on the use of features. In Section 2 I will provide the groundwork for a discussion of the use of features by presenting the pattern of vowel alternations in Hungarian in a pre∗I would like to thank Don Churma, Sharon Inkelas, László Kálmán, Paul Kiparsky, Will