Andries W. Coetzee

What it Means to be a Loser: Non-optimal Candidates in Optimality Theory

WHAT IT MEANS TO BE A LOSER: NON-OPTIMAL CANDIDATES IN OPTIMALITY THEORY SEPTEMBER 2004 ANDRIES W. COETZEE, B.A., POTCHEFSTROOM UNIVERSITY M.A., POTCHEFSTROOM UNIVERSITY M.A., UNIVERSITY OF MASSACHUSETTS AMHERST Ph.D., UNIVERSITY OF MASSACHUSETTS AMHERST Directed by: Professor John J. McCarthy In this dissertation I propose a rank-ordering model of EVAL. This model differs from classic OT as follows: In classic OT, EVAL distinguishes the best candidate from the losers, but does not distinguish between different losers. I argue that EVAL imposes a harmonic rank-ordering on the complete candidate set, so that also the losers are ordered relative to each other. I show how this model of EVAL can account for non-categorical phenomena such as variation and phonological processing. Variation. In variation there is more than one pronunciation for a single input. Grammar determines the possible variants and the relative frequency of the variants. I argue that EVAL imposes a harmonic rank-ordering on the entire candidate set, and that language users can access more than the best candidate from this rank-ordering. However, the accessibility of a candidate depends on its position in the rank-ordering. The higher a candidate appears, the more often it will be selected as output. The best candidate is then the most frequent variant, the second best candidate the second most

Variation as accessing ' non-optimal ' candidates*

This paper argues that rather than just select the best candidate, EVAL imposes a harmonic rank-ordering on the full candidate set. Language users have access to this enriched information, and it shapes their performance. This paper applies this idea to variation. The claim is that language users can access the full candidate set via the rank-ordering imposed by EVAL. In variation, more than one candidate is well-formed enough to count as grammatical. Consequently, language users will access more than just the best candidate from the rank-ordering. However, the accessibility of a candidate depends on its position on the rankordering. The higher the position a candidate occupies, the more likely it is to be selected. In a variable process, variants that appear higher on the rank-ordering (i.e. are more well-formed) will therefore also be the more frequent variants. This model is applied to variation in the phonology of Faialense Portuguese and Ilokano.

Grammar is both categorical and gradient

In this paper, I discuss the results of word-likeness rating experiments with Hebrew and English speakers that show that language users use their grammar in a categorical and a gradient manner. In word-likeness rating tasks, subjects make the categorical distinction between grammatical and ungrammatical – they assign all grammatical forms equally high ratings and all ungrammatical forms equally low ratings. However, in comparative word-likeness tasks, subjects are forced to make distinctions between different grammatical or ungrammatical forms. In these experiments, they make finer gradient well-formedness distinctions. This poses a challenge on the one hand to standard derivational models of generative grammar, which can easily account for the categorical distinction between grammatical and ungrammatical, but have more difficulty with the gradient well-formedness distinctions. It also challenges models in which the categorical distinction between grammatical and ungrammatical does not exist, but in which an ungrammatical form is simply a form with very low probability. I show that the inherent comparative character of an OT grammar enables it to model both kinds of behaviors in a straightforward manner.

Phonological Variation and Lexical Frequency

Lexical usage frequency is known to influence the application rate of some variable processes. Specifically, variable lenition processes typically affect frequent lexical items more often than infrequent lexical items. For instance, variable t/d-deletion in English is more likely to apply to a frequent word (just) than to an infrequent word (bust). Existing grammatical models of phonological variation do well at accounting for the influence of grammatical factors on variation, but these models cannot account for the contribution of non-grammatical factors such as lexical frequency. In this paper, I propose a model of phonological variation that can simultaneously account for the influence of both these factors on variation. I assume an Optimality Theoretic grammar with lexically indexed faithfulness constraints. Variation arises as a result of variable lexical indexation - a single lexical item can be assigned to different lexical classes on different evaluation occasions, and will hence not always be evaluated by the same faithfulness constraints. Each lexical item is associated with a probabilistic distribution function that determines the likelihood of it being assigned to each of the lexical classes. The shape of a lexical items distribution function is determined by its usage frequency, so that frequency influences the likelihood of the lexical item being assigned to specific lexical classes, and hence the likelihood of it being evaluated by specific faithfulness constraints. I apply the proposed model to variable t/d-deletion in English, and show that it succeeds in accounting for the way in which usage frequency influences this process. In two appendices, I show how the model can be implemented by interpreting the lexical distribution functions as instantiations of the beta distribution (Evans et al. 2000). This implementation of the model is then used to determine the expected t/d-deletion rates in a corpus of American English. I also use this implementation to model the acquisition path of a variable process, showing that the model predicts more variation during earlier acquisitional stages.

Phonetically grounded phonology and sound change: The case of Tswana labial plosives

Abstract A widely held assumption in phonology is that phonology should be phonetically grounded. Under a strict version of this view, productive phonological processes that counter phonetic naturalness should not be possible. Traditional grammars of Tswana describe it as having a productive process of post-nasal devoicing of voiced plosives (/mb/→[mp]), counter to phonetic expectations. The vocal fold settings during the nasal should promote, not inhibit, voicing in a following consonant. Leakage through the velic valve during the initial part of the post-nasal oral closure should inhibit buildup of intra-oral air pressure and hence promote voicing. The unexpectedness of the Tswana pattern leads to questioning the accuracy and reliability of the traditional descriptions of this language. In this paper, we report an acoustic study of 12 Tswana speakers, showing that there is evidence that at least some speakers have an active, productive process of post-nasal devoicing. However, our data also show evidence that this phonetically unnatural system is unstable, and is in fact in the process of changing towards a more natural system. We consider the relevance of these results for views about the relationship between phonetics and phonology, and for theories of language change.

Gestural reduction, lexical frequency, and sound change: A study of post-vocalic /l/

Abstract The magnitude of anterior and dorsal constrictions for laterals in /(C)(C)VlC/ words produced by eight American English speakers was measured using ultrasound imaging. The results replicate previous findings that laterals have weaker anterior constrictions when followed by labial or velar consonants than when followed by alveolar consonants. The main novel finding is that, in words with /VlClabial/ or /VlCvelar/ sequences, this anterior constriction was weaker in high-frequency words (help , milk ) than in low-frequency words ( whelp , ilk ). Although high-frequency words also showed slight reduction of the dorsal constriction, dorsal reduction was stable, small in magnitude, and not correlated with anterior reduction, consistent with alveolar reduction not being simply a consequence of overall weaker lingual constrictions in more frequent words. Acoustic measures for laterals showed that the degree of anterior constriction correlated with the frequency separation between F1 and F2: more reduced alveolar constrictions – especially likely in high-frequency words – were linked with greater formant proximity. These articulatory and acoustic patterns are interpreted as potentially contributing to the initiation and lexical diffusion of historical /l/ lenition. It is proposed that gestural reduction in high-frequency words in which the anterior gesture for laterals must be coordinated with another supralaryngeal constriction serves as a precipitating factor in /l/ vocalization and possibly (although to a lesser extent) /l/ loss.

Learning lexical indexation

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 ).

Global and local durational properties in three varieties of South African English

Abstract This article investigates the relationship between the global and local durational properties of an utterance. We show that languages that are similar in terms of their global durational properties are also similar in terms of their local durational properties. However, languages that differ globally also differ locally. We illustrate this with three varieties of South African English. We show that South African English L1 and Afrikaans English both pattern with stress-timed languages and both apply phrase-final lengthening. Tswana English, however, patterns with syllable-timed languages, and does not apply phrase-final lengthening.

DIE AKOESTIESE EIENSKAPPE VAN STEMLOSE EKSPLOSIEWE VAN AFRIKAANS

The voiceless plosives of Afrikaans: a first investigation. In this article the characteristics of the voiceless plosives of Afrikaans are investigated, with special focus on the affrictionjaspirat...

Emergent tonogenesis in Afrikaans

Afrikaans is usually described as having a contrast between prevoiced and voiceless unaspirated plosives ([b]-[p], [t]-[d]). This study documents an ongoing change in the pre-vocalic realization of the contrast. Preliminary data from five speakers show that all speakers produce some phonologically voiced plosives without prevoicing, with frequency of devoicing ranging from 30% to 85% across speakers. This devoicing is nearly categorical: VOTs of devoiced plosives average 12 ms and those of phonologically voiceless plosives average 17 ms. (By comparison, VOTs of voiced plosives average around Ő130 ms.) The contrast appears to be preserved in the F0 contour of the following vowel, which is 50 to 100 Hz lower (depending on the speaker) after phonologically voiced than after phonologically voiceless plosives. The F0 difference continues through at least 70% of the vowel. However, post-plosive F0 variation is not contingent on devoicing, e.g., F0 contours after phonological /b/ are the same regardless of whether the production is [p] or [b]. In these preliminary data, the magnitude of the F0 difference is linked to speaker age, with younger speakers showing a larger difference. Data from a larger group of speakers are being analyzed and will be presented.