The study of phonotactics is a central topic in phonology. We propose a theory of phonotactic grammars and a learning algorithm that constructs such grammars from positive evidence. Our grammars consist of constraints that are assigned numerical weights according to the principle of maximum entropy. The grammars assess possible words on the basis of the weighted sum of their constraint violations. The learning algorithm yields grammars that can capture both categorical and gradient phonotactic patterns. The algorithm is not provided with constraints in advance, but uses its own resources to form constraints and weight them. A baseline model, in which Universal Grammar is reduced to a feature set and an SPE -style constraint format, suffices to learn many phonotactic phenomena. In order for the model to learn nonlocal phenomena such as stress and vowel harmony, it must be augmented with autosegmental tiers and metrical grids. Our results thus offer novel, learning-theoretic support for such representations. We apply the model in a variety of learning simulations, showing that the learned grammars capture the distributional generalizations of these languages and accurately predict the findings of a phonotactic experiment.

/pdf/a-maximum-entropy-model-of-phonotactics-and-phonotactic-3m0jyalxxn.pdf

A Maximum Entropy Model of Phonotactics and Phonotactic Learning

"The Companion to Phonology" will be the major reference work of the field, drawing together nearly 150 contributions from almost all of the globally recognized, leading scholars in an estimated five-volume set. The international editorial team represents a diverse range of approaches and methodologies to the field, with the objective of creating a truly theoretically balanced work that will be of use to all phonologists and linguists. Different from all handbooks and other reference works for phonology which are available at present, the focus will be on empirical facts, rather than on theoretical insights. The Companion will function as a touchstone for future phonological theories: whenever somebody wants to propose a new theory for some phenomenon, she will be able to find all the relevant facts and insights which need to be accounted for in this work. The work will be organized thematically, and by volume: I. The Segment and Below II. The Segment and Above III. Changes and Processes IV. Interfaces V. Methodologies. In terms of competition, there is no other work currently available with this scope. There are large encyclopedic works covering the field of linguistics, such as "The Encyclopedia of Language and Linguistics", second edition, edited by Keith Brown (Elsevier, 2005); the approach and scope of this sort of work is clearly distinct from the proposed Companion, which is comprised of essays of 8,000- 10,000 words each to allow a fuller exploration of each topic. The Companion is not intended to be encyclopedic, but to offer an overview of the field as it currently stands. Shorter reference works such as the "Blackwell Handbook of Phonological Theory" (edited by John Goldsmith, 1996), or the "Cambridge Handbook of Phonology" (edited by Paul DeLacy, CUP, 2007) differ most obviously in their scope: "Goldsmith's Handbook" has 32 articles in it; DeLacy's has 25; a larger work allows more depth, nuance, and information. The Companion is also theoretically positioned as a work primarily for researchers.

Blackwell Companion to Phonology

A central idea in rule-based phonology is the serial derivation (Chomsky & Halle 1968). In a serial derivation, an underlying form passes through a number of intermediate representations on its way to the surface:[Scheme here]Implementational details can differ: the order of rules might be stipulated or it might be derived from universal principles; the steps might be called ‘rules’, ‘cycles’ or ‘levels’; the steps might involve applying rules or enforcing constraints. But, details aside, the defining characteristic of a serial derivation, in the sense I will employ here, is the pre-eminence of the chronological metaphor: the underlying form is transformed into a succession of distinct, accessible intermediate representations on its way to the surface. I will call any theory with this property ‘serialism’.

/pdf/sympathy-and-phonological-opacity-2vjziub1o1.pdf

Sympathy and phonological opacity

Using our theoretical framework of multimodal processing, we developed and evaluated a computer-animated tutor, Baldi, to teach vocabulary and grammar for children with autism. Baldi was implemented in a Language Wizard/Player, which allows easy creation and presentation of a language lesson involving the association of pictures and spoken words. The lesson plan includes both the identification of pictures and the production of spoken words. In Experiment 1, eight children were given initial assessment tests, tutorials, and reassessment tests 30 days following mastery of the vocabulary items. All of the students learned a significant number of new words and grammar. A second within-subject design with six children followed a multiple baseline design and documented that the program was responsible for the learning and generalization of new words. The research indicates that children with autism are capable of learning new language within an automated program centered around a computer-animated agent, multimedia, and active participation and can transfer and use the language in a natural, untrained environment.

Development and evaluation of a computer-animated tutor for vocabulary and language learning in children with autism.

This article presents a typology of consonant harmony or LONG DISTANCE CONSONANT AGREEMENT that is analyzed as arising through correspondence relations between consonants rather than feature spreading. The model covers a range of agreement patterns (nasal, laryngeal, liquid, coronal, dorsal) and offers several advantages. Similarity of agreeing consonants is central to the typology and is incorporated directly into the constraints driving correspondence. Agreement by correspondence without feature spreading captures the neutrality of intervening segments, which neither block nor undergo. Case studies of laryngeal agreement and nasal agreement are presented, demonstrating the model's capacity to capture varying degrees of similarity crosslinguistically.

/pdf/a-typology-of-consonant-agreement-as-correspondence-g381eoyv2a.pdf

A typology of consonant agreement as correspondence

/pdf/nasalization-neutral-segments-and-opacity-effects-17sbuejqgr.pdf

Nasalization, Neutral Segments and Opacity Effects

This paper examines vowel harmony initiated by a weak trigger. Height harmony in Veneto Italian dialects, wherein a post-tonic high vowel triggers raising of preceding mid vowels, forms a case study. Veneto presents two variable patterns: stress-targeted harmony, in which harmony propagates to the stressed syllable, and maximal extension harmony, in which raising persists to pretonic vowels. The conditions under which weak vowels trigger and control harmony are examined. It is argued that weak trigger harmony is motivated by perceptual disadvantage: harmony improves exposure of the spreading feature, accomplished either by extending to a stressed position or maximizing duration in the word. The apparent primacy of the vowel quality in weak position in Veneto is accounted for by markedness factors that independently prevent the stressed vowel from overriding it. In regard to positional privilege, it is argued that positional licensing (markedness) drives stress-targeted spreading – positional faithfulness cannot be responsible for strong targets. A weak trigger pattern emerges when licensing constraints operating over perceptually marked structure dominate positional faithfulness. Crosslinguistic applications are also explored.

Weak Triggers in Vowel Harmony

This paper examines long distance voicing agreement between consonants (Cs). Two related patterns are observed. In the first one [voice] agreement is restricted to pairs of oral stops that match in place of articulation, as seen in Ngbaka. In the second, observed in Kera, [voice] agreement occurs among all pairs of stops. I argue that these agreement patterns come about through a correspondence relation that is established between Cs in the output. The notion of intersegmental correspondence will be important in explaining two key properties of the phenomena: (i) the potential for interaction between Cs at a distance, and (ii) the preference for voicing agreement to occur between similar Cs. From a wider perspective, this analysis is supported by work on other consonantal agreement patterns that display similar characterizing properties (Walker 1999, Rose & Walker in prep.). In addition, I propose that the correspondence approach has the potential to extend to cases of voicing dissimilation. The analysis is couched within Optimality Theory (OT; Prince & Smolensky 1993). The paper is organized as follows. In §2 I present the data illustrating voicing agreement between Cs at a distance. §3 diagnoses the agreement as arising through the mechanism of segmental correspondence rather than feature spreading. In §4 I lay out a theoretical overview of the correspondence approach to long-distance agreement, and then develop the details of analysis of Ngbaka and Kera. §5 discusses an extension to voicing dissimilation phenomena, and §6 gives the conclusion.

Long-distance Consonantal Identity Effects

Many Altaic languages restrict round vowel distribution. This paper exam- ines round harmony in Classical Manchu and Oroqen, where round spreading occurs only when the first two syllables of a word are round, that is, it requires a bisyllabic trigger (Zhang 1996). It is argued that the binary threshold emerges from conflict between well- established phonological demands - numeric reference is neither necessary nor desirable. The study isolates two distinct restrictions on rounding in bisyllabic trigger languages: initial round licensing and round spreading - requirements occurring independently in Classical Mongolian and Ulcha. Separating these restrictions is key: each is active in languages with bisyllabic triggers, but they are ranked asymmetrically with respect to a conflicting constraint that restricts features to a tautosyllabic domain. Ranking the tautosyl- labic constraint between round licensing and spreading prevents cross-syllable spreading except when violations of tautosyllabicity are independently necessitated by round li- censing. As a result, spreading is initiated only when the first two syllables are round. Implications are identified for the characterization of faithfulness. Positional faithfulness constraints play a key role in realizing the privileged status of the root-initial syllable in round licensing and harmony. In addition, the analysis supports the separation of IDENT(F) into IO and OI constraints, which distinguish between the loss and gain of privative fea- ture specifications, respectively. The distinction proves essential in the case of bisyllabic triggers. The constraint interaction that produces the two-syllable trigger threshold is an instance of a general phenomenon explored here, termed Parasitic Constraint Satisfaction. This kind of interaction arises when there are two constraints or constraint sets, α and β, whose satisfaction each necessitates violating a constraint, γ, and they are ranked α � γ � β. When satisfaction of α compels violations of γ that also permit satisfaction of β, then β is described as parasitic on α. Two outcomes for Parasitic Constraint Satisfaction are discussed. The first is an emergence of the unmarked, occurring when β is a marked- ness constraint whose activity emerges in contexts where it is parasitic on α. The second outcome, where β is a faithfulness constraint, is an emergence of the faithful.

http://www.ai.mit.edu/projects/dm/featgeom/walker-altaic-round.pdf

Rachel Walker

Papers

A typology of consonant agreement as correspondence

Nasalization, Neutral Segments and Opacity Effects

Weak Triggers in Vowel Harmony

Long-distance Consonantal Identity Effects

Round licensing, harmony, and bisyllabic triggers in altaic