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.

Ladd (1988) investigates configurations A[BC] vs. [AB]C of three English clauses containing clause-internal downstep. A sister-node relation between clauses (but not a sequential relation) leads to downstep among clauses, such that C is systematically lower than B in A[BC] but not in [AB]C. These findings are replicated here with German data. In addition, the German phenomenon of upstep (Truckenbrodt 2007b) arguably targets the phonetic reference line that models lowering among clauses (van den Berg et al. 1992). We show that both Ladd's and our results also support Selkirk's (2011) suggestion that root sentences/illocutionary clauses can be interpreted as matched to intonational phrases (not just aligned with them, as in Downing 1970). The results also suggest that, in addition to downstep among intonational phrases, phrase-final lowering takes place.

This paper examines the timing of the L and H tonal targets that constitute the phrase-initial pitch rise in the Seoul Korean LHLH intonation pattern. Two hypotheses, the segmental anchoring hypothesis and the constant duration hypothesis, are examined, by varying speech rate. The segmental anchoring hypothesis states that the beginning and the end of a pitch rise are independently aligned with respect to segments. The alternative hypothesis holds that the interval between the L and H tones should remain stable. It was found that segmental alignment and duration targets simultaneously affect the timing of L and H tonal targets. The model proposed here, the ‘Alignment-Duration’ model, takes these two factors as weighted constraints, which require alignment and duration targets to be realised. The actual timing of the F0 events is determined as the values that minimise the summed cost of violations of these constraints, yielding a compromise between their demands.

Recent work under the theoretical banner of Agreement by Correspondence (ABC) has produced a variety of different – and sometimes contradictory – formulations of the constraints central to this framework. In OT, the effects of such definitional choices come out in the factorial typologies they predict. Yet knowing what languages a theoretical system derives is insufficient unless we know why it does so. This requires analysis of the internal ranking structures of the typology itself. This paper compares the typologies produced under different proposed modifications to the main ABC constraints. We analyse the typologies in Property Theory, a theory of typological organisation in OT. Our analyses show that all variations have a common core structure, and that differences in their factorial typologies reduce to differences in how this common structure expands and iterates for different features. This allows for precise delineation of how and why different ABC constraint definitions affect typologies.

Georgian, a South Caucasian language belonging to the Kartvelian family, is characterised by the ability of its consonants to combine in extensive clusters. Among the possible combinations are a series of two-member clusters which are argued to behave phonologically as single segments (Tschenkeli 1958, Vogt 1958, 1971, Aronson 1982, 1991, Deprez 1988 and others). They are known as ‘harmonic’ clusters, because the laryngeal quality is constant across the cluster. Its two members are both voiced ([dg bg dγ bγ]), both aspirated ([thkh tshkh thχ tshχ]) or both ejective ([t'k' ts'k' p'k' t'q' ts'q']). They can occur either word-initially or in word-medial position. Harmonic clusters do not contrast with identical sequences of segments, except for sequences formed at the junction of two words. There is no evidence that across word boundaries harmonic clusters are derived by some sort of restructuring.

The purpose of the present study is to review the phonological arguments brought in the literature in favour of treating harmonic clusters as single segments, and to look for acoustic evidence that would motivate the distinction made between harmonic clusters behaving as single segments, on the one hand, and simple sequences of consonants, on the other hand. The study uses phonetic data to address the issue of phonological representation. If the difference between a harmonic cluster and a simple sequence of segments is present in the phonology, then it should ideally also be visible in the acoustic signal, for example in the presence or absence of a release burst, or in timing differences, as suggested by previous studies of complex vs. simple segments in various languages (Maddieson & Ladefoged 1989, Maddieson 1989, 1990). The results show that the treatment of Georgian harmonic clusters as complex segments is not supported by the acoustic data.

The paper is organised as follows: § 2 presents the phonological behaviour of consonant clusters in Georgian, § 3 reviews phonetic evidence for complex segments, and spells out the predictions made by the present study. The acoustic study is described in § 4, followed by the presentation and discussion of results in § 5. The conclusions and areas for further study are presented in § 6.

In the past few years a great deal of attention has been paid to the representation of suprasegmental phenomena in phonology, with the resulting development of a number of partly competing theories and models of suprasegmental representation – in particular, various versions of AUTOSEGMENTAL PHONOLOGY (see, for example, Goldsmith 1976; Halle & Vergnaud 1981; Clements & Keyser 1983) and METRICAL PHONOLOGY (Liberman & Prince 1977; Hayes 1980, 1982; Prince 1983; Giegerich 1985). Other frameworks have also been developed which allow for the representation of phenomena in this area, notably that of DEPENDENCY PHONOLOGY (Anderson & Jones 1974, 1977; Ewen 1980; Anderson 1984; Anderson & Ewen 1980, forthcoming).It has, moreover, become obvious that although these theories at first seemed very different, many of the differences are more apparent than real, so that in some respects the proposals are complementary rather than alternative - and in many areas it is clear that we are moving towards a situation where a single model can perhaps be developed from the various frameworks (cf. Leben 1982; Goldsmith this volume).

In the current nonlinear model of CV phonology (Clements & Keyser 1983), phonological representations are organised into several levels or tiers, of which the following three are the most significant: the central CV tier (skeleton), whose units, C and V, are autosegmentally associated with the units of the syllable tier on one end and with the units of the segmental tier on the other. The segmental tier itself is organised into subtiers of distinctive features; the internal structure of the syllable tier is subject to considerable debate in the literature. In addition, CV-level units may be linked to melodies on separate autosegmental tiers in the case of suprasegmental phenomena like tone and vowel harmony.

The Australian language Arrernte has been argued by Breen & Pensalfini (1999) and Evans & Levinson (2009) to present a case of VC syllabification with coda maximisation, rather than CV syllabification with onset maximisation. In this paper we demonstrate that greater insights into a number of phenomena are achieved when they are analysed with CV syllabification and onset consonants that are moraic, a possibility independently proposed for a wide range of languages by Topintzi (2010). We review a range of evidence from phonetic studies, acquisition and musicology that points towards CV syllabification in Arrernte, and analyse allomorphy, stress assignment, reduplication and the transpositional language game ‘Rabbit Talk’ in terms of reference to moraic structure. The results lend themselves to new directions in the analysis of Arrernte, and provide further evidence for moraic onsets in prosodic morphology.

Norwegian retroflexion exhibits some phonetic properties that do not seem to ‘make sense’. In Standard East Norwegian, an alveolar /ɾ/ causes a following alveolar coronal to become postalveolar, and in the Frogner and Arendal dialects of Norwegian, the same postalveolarisation process is triggered by a uvular /ʁ/. Comparative analyses of Norwegian dialects reveal that these properties are the results of historical changes and phonological diffusion across dialects. Theories attempting to analyse Norwegian retroflexion as phonetically ‘natural’ can neither fully account for these properties of Norwegian retroflexion nor capture the typological generalisations found across Norwegian dialects.

Although a growing body of literature in formal phonology has espoused the view that phonological knowledge is gradient and probabilistic, this perspective remains somewhat controversial. This paper provides further empirical support for this strand of work: it offers an analysis of the gradient deletion of word-final schwa in Southern French, using a corpus containing 7787 data points obtained from 45 subjects spread over three dialectal areas (the Basque Country, Languedoc and Provence). In addition to confirming or nuancing previous findings about the role of several phonological and non-phonological variables, the study demonstrates the influence of lexical frequency, grammatical category, sonority and the feature specification of the consonant before schwa.

We document a fronting process in Latin that is difficult to model as syntactic movement but fairly easy to model as phonological movement. Movement with similar properties has been observed elsewhere in Classical Greek, Russian, Irish and Japanese; we suggest that the Latin movement is of the same type and takes place in the phonological component of the grammar, following the mapping from syntactic to prosodic structure.

In nasal harmony systems, any major consonant class (excluding laryngeal glides) may fail to undergo nasalisation. This paper proposes to derive the neutrality of stops, fricatives, liquids and glides from the satisfaction of a special set of faithfulness constraints; each member of the set commands the presence of a particular segment class in a consonant inventory. The analysis requires that the constraint commanding the presence of stops be active in every grammar and invariably take precedence over the demands of nasal harmony. The explanation of stop neutrality forces a rethinking of the nature of constraints and how they interact. The paper argues that the constraints regulating the cross-linguistic manifestation of segment neutrality are either universal grammatical features (i.e. principles) or language-specific choices (i.e. parameters of variation). It also postulates that constraint interaction is regulated by a set of metaconditions that determine ordering; constraint ranking is not random.

Prosodification processes like syllabification and foot parsing are often limited to a particular morphological constituent; i.e. they do not apply across a morpheme boundary. Prince & Smolensky (1993) present a new interpretation of this phenomenon which identifies the edge of a morphological constituent, rather than the morphological constituent as such, as opaque to prosodification. On their view, the morphological boundedness of processes that erect metrical structure reflects the desire for a grammatical constituent edge to coincide with the edge of a prosodic constituent, formalised as ALIGN(GCat-Edge 1, PCat-Edge 1) in their theory. Since prosodification across a morpheme boundary results in a mismatch between the edges of a prosodic and a morphological constituent, it is blocked whenever alignment is obeyed.

In this paper I claim that not only prosodification processes but also feature spreading is subject to pressure from an alignment constraint, and so is avoided. The case in point is dorsal fricative assimilation in Modern Standard German. Dorsal fricative assimilation does not apply across a compound boundary or to the dorsal fricative of the diminutive suffix -chen, which, though morphologically a suffix, is prosodically a separate phonological word of German (Noske 1990, Iverson & Salmons 1992, Borowsky 1993, Wiese 1996). Bringing Itô & Mester's (in press) notion of crisp alignment to bear on the analysis, I argue that the application of fricative assimilation is constrained by CRISPEDGE(PrWd), which requires the prosodic word to have sharply defined boundaries. Since spreading from a word-final back vowel to the initial dorsal of the following word results in a blurred word edge, it is ruled out, because CRISPEDGE(PrWd) is ranked higher than the constraint governing spreading.

This paper gives a computational characterisation of opaque interactions in phonology. Specifically, a range of opaque interactions are shown to be Input Strictly Local (ISL) maps (Chandlee 2014), which are string-to-string functions that determine output based only on contiguous sequences of input symbols. Examples from Baković’s (2007) extended typology of counterfeeding, counterbleeding, self-destructive feeding, non-gratuitous feeding and cross-derivational feeding, as well as a case of fed counterfeeding from Kavitskaya & Staroverov (2010), are all given ISL analyses, which show that these interactions can be computed based on sequences of bounded length in the input. It is discussed how ISL maps are restrictive in their typological predictions, have guaranteed learning results and known methods for generation and recognition, and thus compare favourably to rule-based and constraint-based approaches to these interactions.

Vowel deletion in Yine crucially refers to both morphological and phonological information. It has been argued that the process is only analysable in a theory where the phonology has access to morphology, either on the assumption of different morphological domains of constraint evaluation (Lin 1987, 1997a, b) or on the assumption of morphologically indexed optimality-theoretic constraints (Pater 2009). In contrast, I propose a phonological analysis of vowel deletion in Yine in a parallel Optimality Theory model. The phonology, I assume, has only limited access to morphological information, and can only distinguish between affix and stem material. I argue that the morphemes that trigger deletion of a preceding vowel have a defective underlying representation: they lack a mora, and ‘usurp’ the mora of a preceding vowel.

This article re-examines ‘conflicting directionality’ in Japanese mimetic words, a distributional pattern in which palatalisation is preferentially realised on the rightmost of two coronal consonants, but on the leftmost consonant in a word without coronals. Analysis of the original dictionary evidence given in support of this generalisation and an exhaustive search of the Japanese mimetic stratum reveal both several counterexamples to conflicting directionality and the fact that the datasets are far too small to support linguistic generalisation. The theoretical assumptions employed to account for Japanese mimetic palatalisation are thus re-examined, with a focus on clarifying the predictions for future valid examples of conflicting directionality.

Speech errors have often been used to support the psychological reality of phonologically dependent allomorphy in inflectional rules. The phenomenon of morphological accommodation to phonological errors is the most compelling evidence of this sort. We investigate reduplication in Ewe, experimentally inducing phonological errors to see whether reduplicated forms show accommodation. This was the case. Implications for reduplication in Ewe and for models of language production are discussed.

Dene (Athabaskan) verbs are widely known for their complex morphophonology. The most complex patterns are associated with two conjugation markers, /s/ and /n/, which are associated with a floating H tone to their immediate left. In this paper, we provide an analysis of /θe/ and /ɲe/, the reflexes of the /s/ and /n/ conjugations in Tetsǫ́t’ıné. Whereas previous accounts of these conjugations have relied heavily on morphological conditioning, we show that, once level ordering, autosegmental phonology and metrical phonology are brought to bear on the problem, morphological conditioning is not required. Within the framework of Stratal OT, we propose the Domain Reference Hypothesis, by which phonological constraints may only refer to morphological domains and their edges. In addition, we show that in Tetsǫ́t’ıné there is a correlation between phonological opacity and morphological structure, as predicted by the Stratal OT model.

This paper reports on a pattern of voicelessness dissimilation in the Kordofanian language Moro. Voiceless stops and affricates become voiced before a voiceless obstruent in a transvocalic configuration. The dissimilation is robust, and productive across morphological contexts. Phonetically, voicing in Moro is realised as a difference between prevoicing and short-lag voice onset time. This makes [voice] the most realistic featural characterisation; using another feature like [spread glottis] in lieu of [–voice] doesn't explain the contrast. Consequently, dissimilation of voicelessness in Moro is strong evidence that [voice] is a binary feature, and that […voice] may be phonologically active despite being ‘unmarked’. We show that when [–voice] is admitted, the Moro pattern is straightforwardly analysed on a par with other cases of dissimilation. Our analysis uses the theory of surface correspondence, which carries no crucial assumptions about markedness; other theories of dissimilation are considered in an online supplement.

Adequately accounting for patterns of dissimilation has challenged more than one linguistic theory. This paper brings together evidence for certain recurrent patterns of dissimilation in Austronesian and Australian languages. It does not claim to have found a definitive solution to why these patterns occur, but it does provide indications that avoidance of multiple markedness may be causally implicated. Although the emphasis is different, the proposal offered here thus has fundamental similarities with Alderete (1997) in arguing that where it applies to dissimilation the Obligatory Contour Principle is inseparably connected with marked elements. Its primary contributions are to provide further empirical support for this claim that may not be readily accessible to non-specialists, to generalise the claim to a larger class of data, to suggest that the explanation for such patterns may be cognitive rather than phonetic and particularly to draw attention to conditions under which markedness-triggered dissimilation is suppressed.