Beckman and Edwards' paper has two parts: a concise synthesis of an emerging phonological view of prosodic structure, in which the prominent element or head at each level of the prosodic constituent hierarchy is cued by a level-specific set of phonological and phonetic variables, and a description of an experiment aimed at clarifying the articulatory correlates of the levelspecific contrasts in stress described in the theory. I will focus here, briefly, on the phonological model and, more extensively, its account of perceived stress shift in terms of the placement of phrase-level pitch accents.

A model of stress

Beckman and Edwards (henceforth B & E) see prosody as a mechanism for organizing utterances, and “stress” as a term for a number of different types of prominence. They concentrate on two kinds of prominence contrast, presence/absence of a nuclear pitch accent, and presence/absence of a full vowel. Each type of prominence corresponds to the head of a different constituent level in the prosodic hierarchy:

1. the nuclear-pitch-accented syllable (the last and strongest prominence in the constituent “intermediate phrase”) versus

2. the non-pitch-accented full-vowel syllable (the first and strongest prominence in the constituent “stress foot”) versus

3. the reduced syllable (a nonhead element in the stress foot).

A fourth type of generally recognized prominence, the prenuclear accented syllable, is not included in the discussion of the model but is mentioned by B & E.