Recent work (Clements 1985; Sagey 1986) on the structure of distinctive features has analysed affricates and prenasalised stops as involving branching for the features [continuant] and [nasal] respectively. This analysis explains the edge effects associated with such segments, while simultaneously identifying them as single melodic elements that associate as units to prosodic templates. This paper will argue that some languages have contour tones that show parallel properties to those of affricates: they associate as units, but also exhibit edge effects. Such behaviour can be simply understood if tonal features hang off a tonal root node (Archangeli & Pulleyblank forthcoming), and this tonal root node is allowed to branch. A high rising contour tone will have the structure shown in (i), where [upper] is the tonal root node, and [raised] is free to branch.

A number of phonological laws require adjacent elements to stand a certain distance apart from each other on some prominence scale. For example, according to the Syllable Contact Law, the greater the sonority slope between the coda and the following onset, the better. Languages such as Faroese, Icelandic, Sidamo, Kazakh and Kirghiz select different thresholds for an acceptable sonority slope. This article proposes a theory for deriving hierarchies of relational constraints such as the Syllable Contact Law from prominence scales in the constraint set CON in Optimality Theory. The proposal is compared to two alternative approaches, non-hierarchical constraints and the local conjunction of constraint hierarchies, which are argued to make undesirable empirical and theoretical predictions.

It is generally accepted that there are three major categories of phonological elements in the signs of sign language: (i) the shape of the hand, (ii) the location of the hand on or near the body and (iii) the movement of the hand – either (a) movement of the fingers or palm at a single location or (b) movement of the whole hand along a path from one location to another. It has further been argued that each of these categories consists of hierarchically organised classes of features (Sandler 1987b, 1989a).

One of the most significant challenges in the study of speech production is to acquire a theoretical understanding of how speakers coordinate articulatory movements. A variety of work has demonstrated that articulatory, prosodic and extralinguistic factors all influence speech timing in a complex and interactive way. Models such as Articulatory Phonology that stipulate the relative timing of articulatory units must be revised to allow for this variability. Such a revision is outlined below.

The following work should be viewed as a presentation of a new framework for conceptualising articulatory timing. This approach, meant to be programmatic rather than conclusive, is productive if it motivates research that might not otherwise have been undertaken. §1 overviews Articulatory Phonology. The implementation of articulatory timing in terms of phasing relations is discussed. Speech production data bearing on timing variability are discussed in §2. §3 argues for an alternative to Articulatory Phonology's current rule-based approach to intergestural timing that can allow for linguistic and extralinguistic variables to systematically influence phasing relations. §3.2 introduces the PHASE WINDOW framework, which allows the degree of articulatory overlap between linguistic gestures to vary within a constrained range. Finally, §4 concerns the relation of intergestural timing to the postulation of the segment as a primitive unit in phonology. It is hypothesised that certain intergestural timing relations are stable and lexically specified. Gestures whose coordination is constrained by lexical PHASE WINDOWS seem to bear a close relation to those conglomerates of gestures that constitute what is traditionally considered to be a segment.

In many languages the issue arises as to whether an onglide patters as part of the syllable onset or forms the first part of a (rising) diphthong with the immediately following vowel. If it is part of the syllable onset, the structure of a CGV syllable would be as in (1), but if it forms the first part of a diphthong the structure of a CGV syllable could either be as in (2a) with a monomoraic diphthong (where the glide is ‘co-moraic’ with the following vowel) or as in (2b) with a bimoraic diphthong (C = consonant, G = glide, V = vowel, μ = mora, and σ = syllable).

Russian exhibits several different types of palatalisation. These are exemplified in (1), where we look at voiceless stops and affricates.

(1) a. Velar Palatalisation (velars change into postalveolars): k→č

ruk+a ‘hand (FEM NOM SG)’ −ruč+išč+a (AUG NOM SG), ruč+en’k+a (DIM NOM SG)

b. Affricate Palatalisation (affricates become postalveolar): ts→č

konets ‘end’ −konč+i+t’ ‘to finish’

otets ‘father’ −otč+estv+o ‘patronymic’

c. Iotation (many disparate changes of consonants): t→č

šut ‘joker’ −šuč+u ‘I joke’

d. Surface Palatalisation (consonants become [−back, +high]): t→t’

xvost ‘tail’ −xvost+ik [t’] (DIM), xvost+e [t’] LOC SG)

brat ‘brother’ −brat+j+a [t’] ‘brothers (COLL)’

coherent analysis of these disparate effects is a formidable task, but one process seems to be easy: Surface Palatalisation is a straightforward spreading change. This change is particularly simple in the context of i and j since not only the feature [−back] but also the feature [+high] is spread from the triggering context onto the input consonant. In the following, I will restrict the scope of analysis to this simple case. That is, I will look at Surface Palatalisation applying in the context of i and j. I will demonstrate that standard Optimality Theory (henceforth OT: Prince & Smolensky 1993, McCarthy & Prince 1995), with its insistence on parallel evaluation, cannot offer an adequate analysis of Surface Palatalisation. I will suggest that standard OT needs to be modified and to admit the possibility of a level distinction (a derivational step) in the evaluation of output forms.

Competing proposals on the syllabification of initial consonants in Moroccan Arabic are evaluated using a combination of experimental and modelling techniques. The proposed model interprets an input syllable structure as a set of articulatory landmarks coordinated in time. This enables the simulation of temporal patterns associated with the input syllable structure under different noise conditions. Patterns of stability between landmarks simulated by the model are matched to patterns in data collected with Electromagnetic Articulometry experiments. The results implicate a heterosyllabic parse of initial clusters so that strings like /sbu/ comprise two syllables, [s.bu]. Beyond this specific result for Moroccan Arabic, the model reveals the range of validity of certain stability-based indexes of syllable structure and generates predictions that allow evaluation of a syllabic parse even when stability-based heuristics break down. Overall, the paper provides support for the broad hypothesis that syllable structure is reflected in patterns of temporal stability and contributes analytical tools to evaluate competing theories on the basis of these patterns.

Phonologies are characterised by regularity, from the stereotyped phonetic characteristics of allophones to the contextually conditioned alternations between them. Most models of grammar account for regularity by hypothesising that there is only a limited set of symbols for expressing underlying forms, and that an independent grammar algorithm transforms symbol sequences into an output representation. However, this explanation for regularity is called into question by research which suggests that the mental lexicon records rich phonetic detail that directly informs production. Given evidence for biases favouring previously experienced forms at many levels of production and perception, I argue that positive feedback within a richly detailed lexicon can produce regularity over many cycles of production and perception. Using simulation as a tool, I show that under the influence of positive feedback, gradient biases in usage can convert an initially gradient and variable distribution of lexical behaviours into a more categorical and simpler pattern.

We asked whether invariant phonetic indices for syllable structure can be identified in a language where word-initial consonant clusters, regardless of their sonority profile, are claimed to be parsed heterosyllabically. Four speakers of Moroccan Arabic were recorded, using Electromagnetic Articulography. Pursuing previous work, we employed temporal diagnostics for syllable structure, consisting of static correspondences between any given phonological organisation and its presumed phonetic indices. We show that such correspondences offer only a partial understanding of the relation between syllabic organisation and continuous indices of that organisation. We analyse the failure of the diagnostics and put forth a new approach in which different phonological organisations prescribe different ways in which phonetic indices change as phonetic parameters are scaled. The main finding is that invariance is found in these patterns of change, rather than in static correspondences between phonological constructs and fixed values for their phonetic indices.

This article is concerned with an interesting upstep phenomenon in the intonation of some speakers of Southern dialects of German. It experimentally establishes the main properties of this upstep phenomenon, and discusses the theoretical consequences. The upstep occurs on the nuclear pitch accent of a non-final intonation phrase. It targets the phonetic height of the utterance-initial peak, regardless of downstepped peaks that intervene between the initial peak and the upstepped peak. The findings are argued to provide unexpected support for a model of intonation in which downstep among accents can be embedded inside downstep among larger prosodic domains (Ladd 1988, van den Berg et al. 1992). In a combination of that model with an extension of Pierrehumbert & Beckman (1988), it is suggested that the choice between downstep and upstep is conditioned by association to higher prosodic constituents in a systematic way.

If we compare the contrastive potential of different phonological contexts in any language, it usually does not take long to establish that the distributional spoils are unevenly divided. Each context typically displays its own subsystem of oppositions which may be bigger or smaller than those associated with other contexts (cf. Twaddell 1935). The traditional term NEUTRALISATION describes the relation between a defective subsystem and one that is distributionally better endowed.

The failure of a position to sustain a particular contrast can manifest itself in one of two ways, as Trubetzkoy was among the first to point out (1939: 209ff). Under ASSIMILATIVE NEUTRALISATION, the phonetic interpretation of the position with respect to the relevant contrast is determined by the melodic content of an adjacent position. This type of pattern is evident in vowel harmony, where the quality of a harmonising vowel is wholly or partially dependent on that of the dominant vowel within the domain. It is also to be seen in the assimilative suspension of consonantal contrasts. For example, in coda–onset interludes consisting of full or partial geminates, the phonetic interpretation of one position is wholly or partially dependent on that of the other.

REDUCTIVE NEUTRALISATION, on the other hand, refers to a situation in which restrictions on the melodic content of a position operate independently of contrasts in neighbouring positions. In vowel systems, for example, it is quite usual to find that the maximal inventory of oppositions is restricted to prosodically prominent nuclei, while shrunken subsystems of various shapes and sizes show up in weak positions. In its most extreme form, syncope, this results in a nuclear position being gutted of all melodic content. In the case of non-nuclear positions, contrastive potential can be curtailed by sonority sequencing constraints and by consonantal lenition processes which neutralise distinctions of manner (as in vocalisation and spirantisation) or place (as in debuccalisation).

In order to account for the accentual and rhythmical structure of English, a binary-branching prosodic constituent structur is assumed, in which minimally the syllable and the foot must be headed. Feet are potentially marked as accented. This representation makes it possible to describe the prominence patterns of word groups as resulting from three accent deletion rules, the Compound Rule, the Initial Accent Deletion Rule and the Rhythm Rule. It was shown that the structural change effected by Initial Accent Deletion cannot be expressed in theories which represent stress as a relative concept. Moreover, this rule, which like the Compound rule is a lexical rule, provided evidence for the existence of a stratum in the lexical phonology of English in which compounding and so-called ClassII derivation take place. The Rhythm Rule is a postlexical rule, which was shown to apply to the output of the other two rules. Without the aid of any conditions or constraints, it accounted effortlessly for the stress-shift data presented in the recent literature. It could moreover be shown that apparent cases of stress shift in unaccented speech (in which the Rhythm Rule does not apply) should not in fact be viewed as the output of any stress-shift rule at all, but should be explained as the effect of preboundary lengthening as applying to the different constituents in the prosodic hierarchy. It was argued that an analysis of sentence accentuation whereby focused constituents have to be assigned accents can run into problems that do not exist in a ‘deaccenting’ analysis, in which nonfocused constituents are deprived of their accents. Finally, it was argued that English, unlike Dutch, lacks phonological rules that refer to primary word stress, and that, at best, primary stress may reveal itself in low-level timing distinctions.

Recent work by Clements (1985), Sagey (1986), Steriade (1987a) and others has shown clearly that distinctive features are hierarchically organised, and that the hierarchy includes a Place of Articulation constituent. Proposals differ, however, as to the organisation below this Place node. Clements (1985) suggests that there is a Secondary Place node dominating the vowel features, but that [anterior], [coronal] and [distributed] are directly dominated by the Place node itself. Sagey (1986) has argued that there are distinct Articulator nodes, Labial, Coronal and Dorsal, each of which dominates certain binary features, respectively [round]; [anterior] and [distributed]; and [high], [back] and [low]. Dorsal is thus present for both velar consonants and vowels. Steriade (1987a) modifies the Sagey model by adding a Velar node for velar consonants, distinct from the Dorsal node for vowels.

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.

A corpus of more than 500 speech errors that involve a vowel or syllabic nucleus is examined for evidence that bears on the nature of the processing representation that is in force when such errors occur. Evidence is obtained from the patterns of similarity between target segments and the intrusion segments that replace them in errors, on the assumption that target– intrusion similarity arises from characteristics of the processing representation. Findings include (1) a distinctive feature similarity between vowel targets and intrusions, (2) evidence that complex syllabic nuclei can function as error units and (3) evidence that vowel errors are constrained by lexical stress. Finally, the error patterns in both vowels and consonants, and the processing representations they suggest, are evaluated in the light of recent theoretical proposals about the phonological component of the grammar.

The group of guttural consonants consists of those consonants articulated at the back of the vocal tract or with some constriction of the pharynx: uvulars, pharyngeals, laryngeals and emphatics, or pharyngealised/uvularised consonants. These consonants pattern together in terms of several phonological properties, as extensively documented by McCarthy (1989, 1991, 1994). He argues that this natural class can be captured in a model of feature geometry by a node dependent on the Place node: the Pharyngeal node. However, the status of the laryngeal consonants within this class of segments is unresolved. While it is clear that the laryngeals [? h] pattern with the other guttural consonants in Semitic languages, these sounds have also been argued to lack Place specification altogether (Steriade 1987; Bessell 1992). McCarthy (1991, 1994) suggests that languages may in fact stipulate whether laryngeal consonants are specified with a Pharyngeal node, or are Placeless. Bessell & Czaykowska-Higgins (1992) and Bessell (1992), based on data from Interior Salish languages, argue that this selection is independent of the inventory of guttural segments within a language.

This study deals with syllable structure in Polish. The central theme is the question of when and how syllabification rules apply in the lexical phonology of Polish. In § i we lay the ground for our subsequent discussion by giving the basic syllable patterns of Polish. We also propose here a first version of the syllabification algorithm for Polish. In §2 we show that syllabification applies cyclically, because certain cyciic phonological rules make crucial use of information about the prosodic structure of their potential inputs. § 3 then shows that the syllabification algorithm has to apply both before and after the application of cyclic phonological rules on one cycle, and that syllabification is therefore a continuous process. In § we argue that the syllabification algorithm proposed in § i must be modified to enable us to predict whether a high [-consonantal] segment will surface as a vowel or as a glide. Since the distinction between vowels and glides is crucial for the application of certain cyclic phonological rules of Polish, this again shows that syllabification has to apply cyclically. § defends the hypothesis that resyllabification is restricted to Coda Erasure (and the subsequent syllabification of the desyllabified consonants). Again, the (un)predictability of the vowel/glide distinction plays a crucial role here. We summarise our conclusions in §6

It has been proposed that Tashlhiyt is a language which allows any segment, including obstruents, to be a syllable nucleus. The most striking and controversial examples taken as arguments in favour of this analysis involve series of words claimed to contain only obstruents. This claim is disputed in some recent work, where it is argued that these consonant sequences contain schwas that can be syllable nuclei. This article presents arguments showing that vowelless syllables do exist in Tashlhiyt, both at the phonetic and phonological levels. Acoustic, fibrescopic and photoelectroglottographic examination of voiceless words (e.g. [tkkststt]) provide evidence that such items lack syllabic vocalic elements. In addition, two types of phonological data, metrics and a spirantisation process, are presented to show that in this language schwa is not a segment which can be independently manipulated by phonological grammar and which can be referred to the syllable structure.

Markedness distinctions can be ignored. For example, in some languages stress avoids central vowels, and falls on high peripheral vowels, yet in the Uralic language Nganasan central and high peripheral vowels are treated in the same way: stress avoids both types equally. Such ‘conflation’ of markedness categories is not only language-specific, but also phenomenon-specific. In contrast, dominance relations in markedness hierarchies are universal; e.g. stress never seeks out a central vowel when a high peripheral vowel is available. This article argues that both language-specific conflation and universal markedness relations can be expressed in Optimality Theory. Constraints that refer to a markedness hierarchy must be freely rankable and mention a contiguous range of the hierarchy, including the most marked element. The empirical focus is sonority-driven stress in Nganasan and Kiriwina. In addition, Prince & Smolensky's (1993) fixed ranking theory of markedness hierarchies is shown to be unable to produce the full range of attested conflations.

Ternary rhythmic systems differ from binary systems in stressing every third syllable in a word, rather than every second. Consider the following examples from Cayuvava (Key 1961), where stress is on every third syllable counting from the end of the word:

(1) a. à.ri.hi.hí.be.e ‘I have already put the top on’

b. ma.rà.ha.ha.é.i.ki ‘their blankets’

c. i.ki.tà.pa.re.ré.pe.ha ‘the water is clean’

Ternary rhythm is well-established for only a small group of languages, including Chugach Alutiiq, Cayuvava and Estonian, and possibly Winnebago. Nevertheless the stress patterns of these languages are sufficiently complex to warrant an ongoing debate about the implications for metrical theory (see Prince 1980, Levin 1985, 1988, Halle & Vergnaud 1987, Halle 1990, Hammond 1990, Dresher & Lahiri 1991, Rice 1992, Hewitt 1992, Kager 1993, 1994, Halle & Idsardi 1995, Hayes 1995, Ishii 1996, Elenbaas 1999, among others).

The reason for a fresh look at ternarity is the rise of Optimality Theory (henceforth OT; Prince & Smolensky 1993, McCarthy & Prince 1993a), a theory abandoning most devices on which rule-based accounts of ternarity were based. It abandons serial derivations and together with it directional foot assignment, a core device in parametric theories of word stress, as well as special parsing modes for ternary rhythm (Weak Local Parsing; Hayes 1995). Derivational mechanisms and parameters are replaced by universal and violable constraints, stating well-formedness on output forms, and ranked in language-particular hierarchies.

The issue then arises whether OT is able to predict the ternary patterns in a descriptively adequate fashion. The first goal of this paper is to argue that adequate and insightful analyses are indeed possible in OT for two ternary stress languages: Cayuvava and Chugach Alutiiq. We argue that these analyses require no ternarity-inducing mechanisms, such as ternary feet or special parsing modes. Instead ternarity emerges by LICENSING, involving interactions of the anti-lapse constraint *LAPSE (banning long sequences of unstressed syllables; Selkirk 1984) with standard foot- alignment constraints (ALL-FT-X, ALIGN-Y; McCarthy & Prince 1993b). Our analysis incorporates Ishii's (1996) insight that ternarity is a kind of underparsing, which is licensed by an anti-lapse constraint, and induced by standard foot alignment.