Segmental Phonology in Optimality Theory (review)

This is the first collection of a comprehensive body of research articles in which the constraint-based framework of optimality theory (OT) begins to ‘attack the rich range of phenomena found in segmental alternations’ (jacket inset). It contains an introduction and three sections. Section 1 is ‘The content of representations’. The question Linda Lombardi poses in ‘Why place and voice are different’ addresses a ‘too-many-solutions’ problem in OT: why coda [voice] markedness undergoes neutralization (pig → pik) but not epenthesis or deletion (*pigi, *pi) while coda place markedness does undergo epenthesis and deletion, in addition to debuccalization. The analysis pivots on the fact that [voice] is privative, hence its deletion from a segment is optimal, while place specification is inescapable. One of the repairs to place markedness rests on the proposal that pharyngeal is least marked, a characterization framed, crosslinguistically, more in terms of alternations than inventories. Cheryl Zoll’s article, ‘Constraints and representations in subsegmental phonology’, unifies the treatment of ‘floating features’ (e.g. Inor palatalization, which ‘docks’ on the rightmost noncoronal) and latent segments (e.g. the vowel in Yawelmani -m(i) which shows up only with consonant-final stems) through the suggestion that the two differ only in the relative ranking of Dep(Root) (the cost associated with introducing an independent segment) and constraints on syllabification, alignment, and feature-cooccurrence. Robert Kirchner, in ‘Phonological contrast and articulatory effort’, introduces the notion of minimization of effort (articulatory displacement, constriction) in the explanation of various seemingly unrelated lenition processes (spirantization, flapping, elision, [de]voicing) in Tümpisa Shoshone. His representational upheaval admits the full range of nondiscrete phonetic detail into phonological representations, with contrast preservation dictating binarity and effort costs imposing additional demands on the articulatory trajectory. The strict locality hypothesis is raised by Máire NíChiosáin and Jaye Padgett in ‘Markedness, segment realization, and spreading’, where they discuss the effects of Turkish vowel harmony on putatively transparent intervening consonants. An architectural revision is proposed: Gen does not produce structures in which segments are skipped in a spreading domain.

Section 2 is ‘The content of constraints’. ‘Austronesian nasal substitution’, by Joe Pater, addresses the formation of nasal-obstruent clusters in Indonesian, resolved by fusion in the case of voiceless obstruents; and in Muna, where similar clusters are variably resolved by deletion of either the nasal, obstruent, or the entire um- affix. The analysis has reasonable ingredients (aimed to replace Pater’s earlier *NC treatment): prosodic edge alignment, identity of the feature pharyngeal expansion (which groups nasals and voiceless obstruents to the exclusion of voiced obstruents), and a ban on multiple labials. However, evaluation inconsistencies (Max in 16 vs. 17), unexplained symbols (‘M’ in 2), and erroneous tableaux (fatal violations in 4’s winner) may detract readers from being effectively convinced. Chip Gerfen offers ‘A critical view of licensing-by-cue’ through an investigation of s-licensing in Andalusian Spanish. He argues that the distribution of [s] must be stated in terms of syllabic constituency as it is not always perceptually cued by adjacent vowel transitions, particularly in forms such as abs.trak.to. While Gerfen’s own analysis is rather abbreviated, the phonetic data (e.g. the durational behavior of [s.l] and [t.l]) amply support the intuition that ‘coda’ is relevant to the distributional generalization. Moira Yip discusses the identity requirements alliterate (e.g. li-lom) and rhyme (e.g. bui-lui) in ‘Segmental unmarkedness versus input preservation’, focusing on a fascinating range of data from Chinese languages. Whether or not the constraints, in particular the prosodically and segmentally faithful structrhyme and featrhyme, are intended to replace standard base-reduplicant faithfulness remains an intriguing question.

Section 3 is ‘Approaches to opacity’. ‘Local conjunction and extending sympathy theory’ is Haruka Fukazawa’s discussion of stop and affricate dissimilation before homorganic stops and affricate in Yucatec Maya. The change of...