I have argued that a prohibition on multiple phonemic stress levels cannot be made to follow from (46a) and/or (46b), as neither are absolute principles cross-linguistically. I have also suggested that while (46c) may be a strong cross-linguistic tendency, it does not follow from any deeper theory-internal principles, but rather, must be stipulated. Thus, the claim that Kinyambo has multiple phonemic stress levels necessitates no substantive change in standard grid theory, but merely represents one member of the typology of cases inherently predicted by the basic tenets of the theory.

Studies of phonological assimilation have played a central role in the development of current phonological theory. As widely discussed in the literature, assimilation is an extremely common phonological process cross-linguistically and therefore an adequate phonological theory should represent it simply and naturally. This has led to the current view of assimilation as spreading (Clements 1976; Goldsmith 1976; Hayes 1986; among others). Much less work has addressed itself to the issue of dissimilation, but recently it has been suggested that dissimilation should be analysed as delinking followed by default fill-in (Odden 1987; Poser 1987; McCarthy 1988; Yip 1988).

In this paper I shall present a theory of phonetic interpretation of headed phonological representations. The phonological representations in question are non-segmental, hierarchical, graphical objects similar to those in common use in autosegmental, metrical, dependency and ‘government and charm’ phonology, although the details of the phonological formalism I employ are different in some respects from each of these. The theory of phonetic interpretation is based on a parametric, dynamic model of phonetic representation. The distinction between ‘head’ and ‘non-head’ constituents is central to the phonetic interpretation model. As well as being formally explicit, I have developed a computational implementation of this theory, constituting a novel speech synthesis program, ‘YorkTalk’. Consequently, although the theory, like any other, is likely to contain certain faults, it goes beyond pencil-and-paper phonological theories, in that it is capable of algorithmically generating quite natural-sounding speech-like signals of a superior quality to other methods of generating synthetic speech, albeit only for isolated words.

Through construction of predictive models, phoneticians are making increasingly sophisticated attempts to account for certain aspects of the phonological structure of languages from very general principles(Stevens 1989; Lindblom 1984; Ohala 1990, 1992).

This paper argues for a prosodic constraint on the formation of diminutive and augmentative forms in Mexican Spanish. Specifically, the stem preceding diminutive (dim) and augmentative (aug) suffixes in this dialect must comprise an absolute minimum of two syllables. When a stem melody cannot satisfy the two-syllable minimum, an epenthetic vowel (V) [e] surfaces at the right edge of the stem (e.g. panesito ‘little bread’ ← pan). In the analysis proposed here, a disyllabic template f[σ σ] is mapped to stems from left to right as part of dim/aug-formation. Whether or not a consonant (C)/s/ surfaces before the suffix will be analysed as a consequence of syllabification in some cases (e.g. koronita/koronota← korona ‘crown’ vs.

Research in moraic phonology has generally assumed that a single representation is sufficient to account for all phonological rules dependent on moraic structure (Hock 1984; Hyman 1984, 1985; McCarthy & Prince 1986; Zec 1988; Hayes 1989). Thus, if a CVC syllable has the moraic structure in (1), all phonological rules will treat it as bimoraic. I term this the MORAIC UNIQUENESS HYPOTHESIS: at any given stage of a derivation, there is only one moraic projection.

In this paper, I examine sonority and place of articulation constraints on the syllabification of consonants in clusters. I argue that (i) both sonority and place of articulation relationships are important in determining the syllabification of consonants in clusters, (ii) well-formed sonority and place relationships are detectable both through phonotactics and through repair strategies that apply if the constraints on sonority and place are not met, (iii) sonority and place constraints are both structurally determined and (iv) place of articulation does not form part of the sonority relationship between consonants as the structural conditions for the licensing of sonority and for the licensing of place of articulation may differ (see Clements 1990a: 313). However, place of articulation restrictions, while independent of sonority restrictions, are, like sonority restrictions, structurally defined.

In order to adequately describe the application of phonological rules across word boundaries, phonologists have appealed to the notion of prosodic domains (Selkirk 1980, 1986; Nespor & Vogel 1982, 1986; Kaisse 1985; Inkelas & Zee 1990). This research has suggested that the domains within and across which rules apply cannot be defined in purely syntactic terms, but rather that a domain structure consisting of prosodic entities such as the phonological word, phonological phrase and intonational phrase must be built up from the syntactic structure. It is to these prosodic categories that phonological rules refer. Prosodic domains are derived from but not necessarily coextensive with syntactic or morphological domains. In fact, some of the best evidence for the necessity of a prosodic structure in addition to a syntactic structure comes from cases where the two do not match, and where the correct phonological generalisations can only be captured in terms of the prosodic structure. Igbo presents just such a mismatch. This paper will examine two rules - ATR vowel harmony and vowel assimilation - that make the mismatch clear. These two rules apply in complementary situations: harmony applies only within the word, assimilation only between words. Both rules delineate the same domain, although one operates within the domain boundaries, one across them.

In SPE (Chomsky & Halle 1968), stress was formalised as a distinctive feature, on a par with [consonantal], [continuant] and so on. Serious problems with this approach were pinpointed in Liberman & Prince (1977). Building on Liberman (1975), these authors conceived of stress as the product of a syllable-grounded network of hierarchical relations. In particular, they argued that, in any given domain (say, a word), syllables are prosodically organised into layers of binary constituents, each constituent made up of a strong element, construable as the ‘head’, and its weak sister. A path linking heads uninterruptedly leads from the tree root to the most salient, and thus primary stressed, syllable of the domain, which they named the ‘designated terminal element’.

In loanword phonology we seek to uncover the processes by which speakers possessing one phonological system perceive, apply native representational constraints on, and ultimately produce forms which have been generated by a different phonological system. We are interested in how speakers instantiate segmental and prosodic structure on an input which may or may not abide by native rules. Crucial to this assumed strategy is the idea that loanwords do not come equipped with their own phonological representation. For any phonetic string, it is only native speakers for whom a fully articulated phonological structure is present; as we will see, the input to loanword phonology is merely a superficial non-linguistic acoustic signal. Thus as host-language speakers perceive foreign forms in accordance with their indigenous phonological system, they instantiate native phonological representations on the acoustic signal, fitting the superficial input into the native phonological system as closely as possible.

Black (1991) proposes that the failure of Velar Glide Deletion in some contexts motivates an iambic typology in which a disyllabic iamb is better than a monosyllabic iamb. Velar Glide Deletion in Axininca fails when a demotion on the iambic scale would result from its application. However, empirical findings from epenthesis, the regressive marker, augmentation and comparative data from Caquinte Campa argue against this analysis. It is important to note the relationship between the assumption that the root + suffix constitutes the prosodic domain of Velar Glide Deletion, as argued by the typological analysis, and the conclusion that Velar Glide Deletion fails when a monosyllabic foot would result from a disyllabic one. If the suffix is abandoned as part of the prosodic domain for computing Velar Glide Deletion, the root alone is the prosodic domain prohibiting Velar Glide Deletion. In the latter case, the minimal word blocking Deletion is a regular, minimal quantity-sensitive foot.