Phonological judgements are often gradient: blick>?bwick>*bnick>**bzick. The mechanisms behind gradient generalisation remain controversial, however. This paper tests the role of phonological features in helping speakers evaluate which novel combinations receive greater lexical support. A model is proposed in which the acceptability of a string is based on the most probable combination of natural classes that it instantiates. The model is tested on its ability to predict acceptability ratings of nonce words, and its predictions are compared against those of models that lack features or economise on feature specifications. The proposed model achieves the best balance of performance on attested and unattested sequences, and is a significant predictor of acceptability even after the other models are factored out. The feature-based model's predictions do not completely subsume those of simpler models, however. This may indicate multiple levels of evaluation, involving segment-based phonotactic probability and feature-based gradient phonological grammaticality.

The intonation of Greek wh-questions consists of a rise-fall followed by a low plateau and a final rise. Using acoustic data, we show (i) that the exact contour shape depends on the length of the question, and (ii) that the position of the first peak and the low plateau depends on the position of the stressed syllables, and shows predictable adjustments in alignment, depending on the proximity of adjacent tonal targets. Models that specify the F0 of all syllables, or models that specify F0 by superposing contour shapes for shorter and longer domains, cannot account for such fine-grained lawful variation except by using ad hoc tonal specifications, which, in turn, do not allow for phonological generalisations about contours applying to utterances of greatly different lengths. In contrast, our findings follow easily from an autosegmental-metrical approach to intonational phonology, according to which melodies may contain long F0 stretches derived by interpolation between specified targets associated with metrically strong syllables and prosodic boundaries.

Optimality Theory explains typological markedness implications by proposing that all speakers possess universal constraints penalising marked structures, irrespective of the evidence provided by their language (Prince & Smolensky 2004). The account of phonological perception sketched here entails that markedness constraints reveal their presence by inducing perceptual ‘repairs’ to structures ungrammatical in the hearer's language. As onset clusters of falling sonority are typologically marked relative to those of rising sonority (Greenberg 1978), we examine English speakers' perception of nasal-initial clusters, which are lacking in English. We find greater accuracy for rising-sonority clusters, evidencing knowledge of markedness constraints favouring such onset clusters. The misperception of sonority falls cannot be accounted for by stimulus artefacts (the materials are perceived accurately by speakers of Russian, a language allowing nasal-initial clusters) nor by phonetic failure (English speakers misperceive falls even with printed materials) nor by putative relations of such onsets to the statistics of the English lexicon.

Morphological concatenation often triggers phonological processes. For instance, addition of the plural suffix /-ən/ to Dutch nouns causes vowel lengthening in some nouns due to the stress-to-weight principle ([xɑt] vs. [ˈxaː.tən] ‘hole’). These kinds of processes often apply only to a subset of words – not all Dutch nouns undergo this process ([kɑt] vs. [ˈkɑ.tən] ‘cat’). Nouns need to be lexically indexed as either undergoing this process or not. I investigate how phonological grammar and lexical indexation are learned when learners are confronted with data like these. Based on learnability considerations, I hypothesise that learners acquire a grammar with default non-alternation, so that novel items are treated as non-alternating. I report the results of artificial language learning experiments compatible with this hypothesis, and model these results in a version of the Biased Constraint Demotion algorithm (Prince & Tesar 2004).

Phonological grammars characterise distinctions between relatively well-formed (unmarked) and relatively ill-formed (marked) phonological structures. We review evidence that markedness influences speech-error probabilities. Specifically, although errors result in unmarked as well as marked structures, there is a markedness asymmetry: errors are more likely to produce unmarked outcomes. We show that stochastic disruption to the computational mechanisms realising a Harmonic Grammar (HG) can account for the broad empirical patterns of speech errors. We demonstrate that our proposal can account for the general markedness asymmetry. We also develop methods for linking particular HG proposals to speech-error distributions, and illustrate these methods using a simple HG and a set of initial consonant errors in English.

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.