An overview of the historical phonology of the Austronesian languages shows certain recurrent patterns of change that resemble the synchronic notion of a conspiracy. Over 90% of all lexical bases in Proto-Austronesian and other early Austronesian proto-languages are disyllabic. This dominant pattern, which was transmitted to most of the 1200-plus Austronesian languages spoken today, has tended repeatedly to reassert itself in forms that have come to have other than two syllables. As a structurally defined target that is satisfied by diverse historical changes, this preferential disyllabism can be considered an ‘attractor’ in the sense of Kelso (1995). Perhaps the most interesting consequence of disyllabic attractors in Austronesian historical phonology is the widespread occurrence of syncope only between identical consonants, a pattern that Odden (1988) has characterised in synchronic systems as one of ‘anti-antigemination’.

Romance languages show hiatus and diphthongal realisations of inherited iV sequences of rising sonority (e.g. ia). We study five Romance varieties with different degrees of contrast between the two realisation types: Romanian, with a diphthong–hiatus contrast, Spanish, with a weaker contrast, French, with no contrast (all diphthongs), and European and Brazilian Portuguese, with no contrast (all hiatus). We show that the different degrees of synchronic contrast are related to three independent factors: (i) a general articulatory tendency for [iV] hiatus to resolve to diphthongs, due to the relative stability of diphthongal articulations; (ii) a structural ‘attractor’ effect of pre-existing [jV] diphthongs in a language, from different historical sources; and (iii) prosodic lengthening effects which inhibit the shift from hiatus to diphthong, supported by phonetic studies of durational patterns across the five languages.

Consonant harmony involves long-distance featural assimilation, or agreement, of consonants across intervening segments. Current correspondence-based analyses of such sound patterns assume that they originate in the cognitive exigencies of articulatory planning, either synchronically, through the functional grounding of the constraints responsible, or diachronically, whereby processing factors incrementally shape the lexicon over time. This paper challenges the validity of this assumption as an all-purpose functional explanation for the full range of long-distance consonant agreement patterns by demonstrating that a variety of diachronic trajectories underlies their emergence and evolution. Focusing on the comparatively rare phenomenon of secondary articulation agreement, the evolutionary histories of three cases are examined: (labio)velarisation agreement in Pohnpeian (Oceanic), palatalisation agreement in Karaim (Turkic) and pharyngealisation agreement in Tsilhqot'in (Athabaskan). These histories provide explanations for a range of synchronic properties of the systems in question, some of which are problematic for restrictive typologies of consonant harmony.

Laryngeal realism (Honeybone 2005) holds that thoroughly voiced stops in a language like Dutch will be represented phonologically with the feature [voice], leaving the voiceless unaspirated stops laryngeally neutral, whereas the typically aspirated stops of a language like German are marked with the feature [spread glottis], rendering the passively voiced stops in this language neutral. These two languages also merge laryngeal oppositions in final environments, Dutch undergoing final devoicing but German final fortition. We apply the findings of Evolutionary Phonology (Blevins 2004, 2006b) to these distinctions in final laryngeal neutralisation, underscoring that the evolutionary approach to phonological alternation allows for non-assimilatory feature addition as well as loss. We examine in particular the known history of final fortition in German and find that the reference standard form of the language has evolved an alignment condition to the effect that a fortified syllable edge must match up with the morpheme edge.

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.

A growing body of work on exemplar-based theories of learning suggests the possibility of formal models of phonological representation which will offer deeper explanations of basic phonological properties than current models allow. The main purpose of this paper is to shed light on near merger, a recalcitrant problem in sound change and in phonological theory, with this newer perspective, through a case study of tonal near merger in Cantonese.