Stress in modern Welsh

The first chapter reviews twentieth- century work on stress, considering stress from several points of view: as intensity, pitch and rhythm. The weaknesses of some approaches are pointed out. Some past work on isochrony is summarised and the metrical theory of stress introduced. The problem presented by stress in Welsh is introduced informally. The second chapter describes a series of measurements and experiments carried out to investigate stress in Welsh. Acoustic measurements are made of Fo, intensity, duration, and lengths of rhythmic feet. It is concluded that stress in Welsh is determined by rhythm rather than inherent acoustic cues. Statistical anal ysis of further measurements of feet reveal s the effects of syllabification and also supports the previous findings. Perceptual experiments using native speaker judgements of resynthesised speech are presented and analysed; it is concluded that Welsh stress is linked with rhythm and with consonantal length (the consonant after a stressed vowel being longer than one after an unstressed vowel), while it has no direct link with pitch- prominence as such. The third chapter outl ines a theoretical framework based on the 'metrical' theory of stress, and proceeds to apply it to Welsh . This modified framework makes use of Selkirk's 'prosodic units', but differs from the usual forms of the theory by also taking account of the semantic weight of words in the utterance context. It is agreed that this makes for a more explanatory theory, and some puzzles of English intonation are shown to be described more simply using such a theory. The patterns of intonation in Welsh are also described in this framework. The fourth chapter discusses the implications posed by the experimental results for the general theory of stress. Reference is made to some research in the psychology of stress perception, and evidence is also presented from the historical development of Welsh from ancient British times, and also from the Welsh bardic rules of versification. It is argued that Welsh has a unique contribution to make to our understanding of the nature of and possible forms taken by stress, and the conclusion is that stress in any language, and more particularly in English, may not be as straightforward a matter as has been hitherto assumed