[Show abstract][Hide abstract] ABSTRACT: Phonological constraints can, in principle, be classified according to whether they are natural (founded in principles of Universal Grammar (UG)) or unnatural (arbitrary, learned inductively from the language data). Recent work has used this distinction as the basis for arguments about the role of UG in learning. Some languages have phonological patterns that arguably reflect unnatural constraints. With experimental testing, one can assess whether such patterns are actually learned by native speakers. Becker, Ketrez, and Nevins (2007), testing speakers of Turkish, suggest that they do indeed go unlearned. They interpret this result with a strong UG position: humans are unable to learn data patterns not backed by UG principles. This article pursues the same research line, locating similarly unnatural data patterns in the vowel harmony system of Hungarian, such as the tendency (among certain stem types) for a final bilabial stop to favor front harmony. Our own test leads to the opposite conclusion to Becker et al.: Hungarians evidently do learn the unnatural patterns. To conclude we consider a bias account—that speakers are able to learn unnatural environments, but devalue them relative to natural ones. We outline a method for testing the strength of constraints as learned by speakers against the strength of the corresponding patterns in the lexicon, and show that it offers tentative support for the hypothesis that unnatural constraints are disfavored by language learners.
[Show abstract][Hide abstract] ABSTRACT: 1 Introduction 1 This paper proposes a unified analysis of diachronic and synchronic coalescence in West Greenlandic (WG) based on perceptual information. There are no consonant clusters in WG. Derived or historical clusters are resolved by coalescence to a geminate: C 1 C 2 C: 1,2 . In both diachronic (1a) and synchronic (1b) clusters containing a uvular stop /q/, pharyngealization (indicated by an underdot) is preserved in the output. (1) a: upinraaq up:aaq 'spring' b: /imq+it/ [:it] 'water' pl. Pharyngealization may be preserved as a primary articulation (on a uvular), or as a secondary articulation on a labial or coronal geminate. 'N' is a uvular nasal. (2) a: upinraaq [upN:aaq] or [up:aaq] 'spring' b: /imq+it/ [N:it] or [:it] 'water' pl. This talk argues that preservation of place features (secondary pharyngealization as well as primary place) is best explained by reference to perceptual information computed in a relevant citation form. The consistent preservation of 1 I am grateful for discussion and comments from audiences at MIT's Phonology Circle, the MIT-UMass Meeting, the 2007 Annual Meeting of the LSA and WIGL5. All remaining errors are my own. Gillian Gallagher 96 pharyngealization and the asymmetry in primary place preservation across manners of articulation is explained by considering the hypothetical articulation of the uncoalesced cluster. The properties of the articulated cluster are deduced from surface properties of the language. The paper is organized as follows. Section 2 gives background on WG and Section 3 presents the data to be accounted for. Section 4 discusses pharyngealization in WG and how perceptual information is relevant to coalescence. Section 5 analyzes pharyngeal preservation in coalescence and Section 6 analyzes preservation of primary place. Section 7 concludes.
[Show abstract][Hide abstract] ABSTRACT: The paper discusses phonologically motivated gaps in inflectional paradigms. A model is offered in which the appearance of gaps is based on a tension between markedness constraints, faithfulness constraints, and constraints which require the expression of morphological categories. After presenting the model, additional implications are analyzed. Situations in which the same problem has different solutions in different morphological contexts are predicted insofar as constraints requiring the expression of different categories can vary in their ranking relative to some faithfulness constraint. Hence, the same phonotactic problem can yield a gap in one situation and a repair in another. This prediction is illustrated and further details of the prediction are explored, including the identification of a situation requiring a more restrictive version of the model. This is achieved by drawing on Smith's (2001) proposal that faithfulness constraints can be indexed to lexical categories to model this situation. 1. I N T R O D U C T I O N Any word generated by the morphology of a language must be reviewed by the phonology to determine whether it is well-formed. When the output of the morphology is not phonologically well-formed, the phonology will most often kick in to repair the word, perhaps by changing a feature specification, or by inserting a segment, or perhaps by shifting the location of stress. But there is another option which the grammar might take. Instead of repairing the offending output, the grammar may simply leave the word unformed, declaring instead that a particular combination of morphemes is for phonological reasons unutterable. The attempt at word-formation fails  For helpful feedback on various aspects of this work, I'm grateful to audiences at the Manchester Phonology Meeting 13 (May 2005) and NELS 36 (November 2005). This work has also benefitted from helpful conversations or correspondence along the way with. I'm also grateful for the input of two anonymous Journal of Linguistics reviewers and Nigel Fabb. This paper supersedes Rice (2006b).
Journal of Linguistics 01/2007; 43(01). DOI:10.1017/S0022226706004488 · 1.18 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: this article, unless otherwise indicated, Arabic words are given in their stem form, which abstracts away from the effects of phonological rules and the addition of inflectional affixes from the agreement, mood, and case-marking systems): (1) Shape Invariance in Arabic Form 2 (fa..ala) /ktb/ /drs/ /.lm/ /sm/ 'write' 'study' 'know' 'poison' kattab darras .allam sammam perfect active kuttib durris .ullim summim perfect passive Cu+ kattib darris .allim sammim imperfect active Cu+ kattab darras .allam sammam imperfect passive Taken in the context of the fuller analysis of the verbal system in McCarthy (1981), this small array of facts is sufficient to demonstrate the property of shape-invariance in Arabic templatic morphology. Moving across the columns of (1) changes the consonantal root, the fundamental lexical unit of the language. Despite this change in the consonants, the canonical pattern remains the same. Similarly, moving down the rows of (1) changes the vocalism: voice goes from active to passive or aspect from perfective to imperfective. Again the canonical pattern remains the same. Similar regularities are met with throughout the Arabic system of verbal conjugations
[Show abstract][Hide abstract] ABSTRACT: It is pointless to define evaluation measures until we know what kinds of phonological descriptions we want them to select. Kiparsky (1968:151) 1 Introduction The facts concerning the distribution of] in Eastern Massachusetts and other dialects of English have resisted explanatorily adequate analysis primarily because they involve the insertion of], a generally unexpected epenthetic consonant. 1 The fact that underlying ] is also deleted in a complementary set of environments is clearly relevant, as originally noted by Vennemann (1972), but no synchronic analysis of these facts to date has satisfactorily explained this connection. I present here an analysis of the facts of Eastern Massachusetts English within Optimality Theory (OT; Prince & Smolensky 1993) that establishes the relevant connection in an explanatory way. Crucial aspects of the proposed analysis provide support for the view that featural faithfulness constraints are symmetrical IDENT[Ä] constraints, as originally proposed by McCarthy & Prince (1995). for helpful discussion. Remaining errors are all mine. 1 I employ the IPA symbol] (coronal approximant) for perspicuity, but assume (following Delattre & Freeman 1968) that this segment involves both a coronal gesture and a pharyngeal constriction. See ¤3.2.
[Show abstract][Hide abstract] ABSTRACT: this paper, I demonstrate that this is indeed the case. Section 1 provides an account of the regular patterns of pretonic stress, a necessary prerequisite to the analysis of stem stress preservation and lexical exceptionality in section 2, which incorporates prosodic faithfulness constraints (It, Kitagawa, and Mester 1996 and McCarthy 1996) into the hierarchy established in -3section
[Show abstract][Hide abstract] ABSTRACT: 2.1 The Theory is also an Object of Analysis * Common sense is often a poor guide to methodology. Any theory presents us with two fundamental and often difficult questions: — What is it? — How do you do it? The first of these arises because a theory is the totality of its consequences. It must be given as the set of its defining conditions, and we may polish them, ground them, tailor them to meet various expectations, but unless we have mapped out what follows from them, the theory remains alien territory. Newton's theory of gravitation can be written on a postcard, and we might like to think of it as nothing more than what makes apples fall straight to earth and planets follow simple repetitive paths, but its actual content is strange beyond imagining and still under study hundreds of years after it was stated. 1 Once formulated, a theory has broken definitively with intuition and belief. We are stuck with its consequences whether we like them or not, anticipate them or not, and we must develop techniques to find them. The second question arises because the internal logic of a theory determines what counts as a sound argument within its premises. General principles of rigor and validation apply, of course, but unless connected properly with the specific assumptions in question, the result can easily be oversight and gross error. Here's an example: in many linguistic theories developed since the 1960's, violating a constraint leads directly to ungrammaticality. A parochial onlooker might get the intuition that violation is somehow ineluctably synonymous with ill-formedness, in the nature of things. A grand conclusion may then be thought to follow: (1) "… the existence of phonology in every language shows that Faithfulness [in Optimality Theory] is at best an ineffective principle that might well be done without." (Halle 1995b). 'Phonology' here means 'underlying-surface disparity'. Each faithfulness constraint forbids a certain kind of input-output disparity: case closed. But no version of Optimality Theory (OT) has ever been put forth that lacks a full complement of Faithfulness * I'd like to thank for interactions which have shaped and re-shaped my views on the matters addressed here. 1 Saari (2005) is a recent study. To get a sense of what can happen, see Ekeland (1988), esp. pp. 123-131.
[Show abstract][Hide abstract] ABSTRACT: Many sorts of ambiguity are tolerated in language. Pronouns may be ambiguous as to their referents, quantifiers may have ambiguous scopes, and structural ambiguity can arise, as in (1) from Japanese (see Inoue and Fodor (1995) for more examples of ambiguity in Japanese). (1) Sumiko-to Jiroo-no okaasan = Sumiko-to [Jiroo-no okaasan] Sumiko-CONJ Jiroo-GEN mother 'Sumiko and [Jiroo's mother]' = [Sumiko-to Jiroo]-no okaasan '[Sumiko and Jiroo]'s mother' But ambiguity is not always tolerated. Occasionally a syntactic process appears to be blocked or triggered in order to prevent ambiguity. One such process is Japanese scrambling, which cannot occur in sentences where subjects and objects are morphologically identical (i.e. are not distinguished by case morphology). If scrambling were allowed in such sentences, the sentences would be ambiguous as to their subjects. This word order freezing occurs both in sentences like (2), where the subject and object both receive nominative morphology, and also in sentences like (3), where both case particles are dropped. In both (2) and (3), the (ungrammatical) scrambled structure with subject Hanako would sound exactly like the unscrambled structure with subject Taroo; when the scrambled sentences are blocked, subject-related ambiguity is prevented., and the audience at HUMDRUM 2005. Special thanks to Ellen Woolford and Shigeto Kawahara, both tireless sources of very different kinds of information, judgements, and general support. A version of this paper appears in UMOP 32: Papers in Optimality Theory III.
[Show abstract][Hide abstract] ABSTRACT: 1. Lateral distribution and possible analyses Australian languages have a large number of contrastive coronal places of articulation; three or even four coronal stops and nasals are common. Australian languages often have laterals articulated at each of these coronal places, and as such have more contrastive voiced lateral approximants than any other language group (Ladefoged and Maddieson 1996: 185). Panyjima's four laterals represent the maximal lateral inventory; other Australian languages have all or a subset of these laterals. (1) Panyjima phoneme inventory (Dench 1991) Labial Dental Apico-alveolar Retroflex Palatal Velar Stop p t t c k Nasal m n n Lateral l l Rhotic Glide w j Laterals are often subject to phonotactic restrictions which prevent them from surfacing word-initially or postconsonantally in many Australian languages, as in (2)., the participants in UMass Second Year Seminar and Phonology Group, and the audience at NELS 35 for suggestions and helpful discussion; also to Tim Beechey, Claire Bowern, Gavan Breen, and Rob Pensalfini for language data and advice.