Vanishing vowels and nasal enigmas

Linguists have long been aware that Portuguese pronunciation is "difficult", and that major differences exist between European and Brazilian Portuguese (EP and BP), but very little exists in English to explain the complexities of Portuguese phonology. Maria Helena Mateus and Ernesto d'Andrade, dominant figures in phonological studies in Portugal since the publication of their doctoral theses in 1975 and 1977 respectively, have finally broken through the language barrier with this volume.

The sound pattern of Portuguese is dominated by a small number of major processes with repercussions throughout the language. The placement of word stress, itself governed by a combination of morphological and phonological factors, affects vowel quality in that unstressed vowels are modified and, particularly in EP, reduced and even lost completely. The syllable is a key unit, as there are complex principles of basic syllable structure that limit consonant clusters, only for the loss of unstressed vowels in EP to create a wide range of consonant sequences at the phonetic level. The few consonants able to appear both initially and finally in syllables have a wide range of variants conditioned by their position in the syllable, and there are large numbers of diphthongs, both present in underlying structure and created by syllabic processes. Nasalised vowels and diphthongs, one of the most salient features of the Portuguese sound system, result from the incorporation of an enigmatic nasal element (consonant? semivowel? autosegment?) in a range of syllabic structures. Numerous vowel quality alternations, particularly involving the mid vowels, are conditioned by morphology, stress, nasality and syllable structure.

Mateus and d'Andrade divide their treatment of this material according to levels of phonological structure, with separate chapters on segmental phonology and phonological features, syllable structure, inflectional morphology, derivational morphology, word stress and phonological processes. This logical arrangement has the effect of dispersing the discussion of central topics such as nasality and atonic vowel reduction, and emphasising the different theoretical frameworks used at each level of structure, particularly in chapter six, which attempts to reconcile two rather different theories of word stress. The final chapter, on phonological processes, which could have drawn all these threads together, is disappointingly slight.

With just 162 pages of text, the authors do not give themselves enough space to do justice to the complexities of Portuguese phonology. Let us hope that Portuguese and Brazilian linguists will be encouraged by their initiative to take us further into the sound systems of this major world language.

Stephen Parkinson is lecturer in Portuguese language and linguistics, University of Oxford.