Showing papers in "The Linguistic Review in 2000"
TL;DR: This paper proposed a constraint-based version of Lexical Phonology and Morphology (LPM), in which stems, words, and sentences are subject to separate, serially related OT constraint systems.
Abstract: Cyclic phonology-morphology interactions and opacity have been dealt with in strictly parallelist OT by introducing new constraint types, including Base-Output constraints and Sympathy. An alternative OT approach to the phonology-morphology interface is a constraint-based version of Lexical Phonology and Morphology (LPM), in which stems, words, and sentences are subject to separate, serially related OT constraint systems. I show that OT-based LPM provides a superior account of the cyclic morphology/phonology interaction in Levantine data from which Kenstowicz and Kager drew support for Base-Output constraints, and that the same account explains the derivational opacity phenomena for which Sympathy theory provides a purely descriptive solution.
433 citations
TL;DR: This paper showed that Progressive Voicing Assimilation is characteristically restricted to the interword environment (i.e. it occurs at the WORD level) and is the consequence of WORD-faithfulness.
Abstract: This paper proposes that differences in the direction of application of phonological rules can be attributed to the differences in the observed patterns of faithfulness at the WORD and ROOT-levels. Using data from English and Dutch I show that Progressive Voicing Assimilation is characteristically restricted to the inter-word environment (i.e. it occurs at the WORD-level) and is the consequence of WORD-faithfulness. I consider whether the same kind of faithfulness effect can account for assymetrical patterns observed with other phonological processes such as vowel harmony, vowel elision and nasal place assimilation
33 citations
TL;DR: In this paper, it was shown that the process of predicate formation in this construction could not be the result of the familiar syntactic mechanism of operator-variable binding, rather, predicate formation is made possible by a simultaneous analysis (coanalysis) of this construction as a morphological construction, thereby making possible the application of certain word-internal processes.
Abstract: The adjectival genitive construction in Modern Hebrew is similar in its internal syntactic structure to the construct-state construction. Being adjectival, it functions as a predicate which involves, semantically, the binding of an implicit possessor argument of a structure internal noun. It is shown that the process of predicate formation in this construction could not be the result of the familiar syntactic mechanism of operator-variable binding. Rather, predicate formation is made possible by a simultaneous analysis (coanalysis) of this construction as a morphological construction, thereby making possible the application of certain word-internal processes
30 citations
TL;DR: In Japanese, the locational verbs iru (animate) and aru (inanimate) can express locative-existential meanings when they are intransitive, and possessive meanings if they are transitive as mentioned in this paper.
Abstract: In Japanese, the locational verbs iru (animate) and aru (inanimate) can express locative-existential meanings when they are intransitive, and possessive meanings when they are transitive. The animacy alternation that occurs between aru and iru is conditioned by the subject of an intransitive locative-existential verb, and by the object of a transitive possessive verb. In Japanese, animacy agreement is generally obligatory, but when the inanimate verb aru is used transitively, agreement does not obtain (unless it is forced by other factors). This peculiar behavior of animacy agreement in Japanese is shown to follow straightforwardly on the assumption that the light verb associated with the animate verb iru has a strong D-feature to attract its nominative phrase overtly, while the light verb associated with the inanimate verb aru has a weak D-feature, which does not require its overt movement, and that instantiation of agreement is contingent upon whether or not the nominative phrase is overtly raised to [Spec, v]. This paper demonstrates that Japanese is a language which selectively induces overt object shift
28 citations
TL;DR: The authors surveys some recent OT work surrounding these issues in light of the data provided by the Marshallese vowel system, finding much of that work wanting in rigor and coherence. But they do not consider the relationship between phonetics and phonology.
Abstract: It is apparent from much recent work in Optimality Theory, as clearly indicated in Kager's (1999) survey, that the restructuring of the architecture of phonological theory has presented phonologists with the opportunity for a fundamental reconceptualization of many of the intriguing issues arising from the relationship between phonetics and phonology. This article surveys some recent OT work surrounding these issues in light of the data provided by the Marshallese vowel system, finding much of that work wanting in rigor and coherence.
20 citations
TL;DR: For instance, the authors argues that the grammar of a language can only specify (in addition to the properties of its lexical items) a language particular ranking of the universal constraints, and that some of the language particular constraints may be expressed as an assembly of universal ingredients.
Abstract: At the substantive level, OT represents a research program that seeks to explain the invariance, recurrence, and variability of linguistic patterns across languages by postulating a set of universal requirements on linguistic representations, and accounting for complex patterns as interactions among simple requirements. At the formal level, OT employs constraints to express requirements on representations, and constraint ranking to express interactions among the requirements. This article tries to unearth the meanings of the formal devices, so as to facilitate a productive critical evaluation. I make three points. First, constraint ranking in OT is a device to capture the formal property of non-monotonicity. It is only one of the different types of non-monotonic formalisms for the expression of conflict resolution, an obvious alternative being weighted constraints. By overlooking weight assignment as an alternative, OT focuses exclusively on competition among constraints, ignoring other interesting aspects of interaction. Second, OT makes the strong claim that there are no language particular constraints; the grammar of a language can only specify (in addition to the properties of its lexical items) a language particular ranking of the universal constraints. By denying that at least some constraints are language particular, OT fails to see that some of the language particular constraints may be expressed as an assembly' of universal ingredients. The result is a proliferation and duplication of the constraints in UG. Third, the two problems above have led the research program to an unhealthy preoccupation with the formal expression of the universal requirements and their interactions, to the detriment of a serious substantive inquiry into what these requirements are.
18 citations
TL;DR: In this paper, a unified approach was proposed to solve a large number of empirical problems for OT, that go from cyclicity to non-derived environment blocking effects to opacity, revealing that outputs are not calculated only from unique 'inputs', but are rather also influenced by other outputs.
Abstract: Work on segmental contrast and work on allomorphic variation within OT has been converging on the conclusion that outputs are not calculated only from unique 'inputs', but are rather also influenced by other outputs, in a way that seeks to eliminate weak contrasts. I propose a unified approach to this effect and note that it thus unifies solutions to a significant number of empirical problems for OT, that go from cyclicity, to 'Non-Derived-Environment Blocking' effects to 'opacity', revealing that this is a crucial development.
16 citations
TL;DR: The authors compared Optimality Theory of Phonology as Human Behavior of the Columbia School of linguistics from the point of view of the interaction between the opposing forces of markedness (the human factor) and faithfulness (the communication factor) integral to both theories.
Abstract: In this article, Optimality Theory is first compared and contrasted with the theory of Phonology as Human Behavior of the Columbia School of linguistics from the point of view of the interaction between the opposing forces of markedness (the human factor) and faithfulness (the communication factor) integral to both theories. The theories are contrasted based on their differing theoretical and methodological (sentence-oriented versus sign-oriented) research paradigms. Despite these basic differences, similarities are found in their shared functional basis which are discussed in the context of the phonotactic functional processes of Natural Phonology with which they are both compared. The theories' syntactic (grammatical) analyses are then further compared and contrasted based on the alternative ways they analyze the phenomenon of do-support (the auxiliary do) as being either semantically empty or not.
12 citations
TL;DR: The appeal to Richness of the base in Optimality theory may undermine many OT analyses, as well as OT's status as a psychological, and not merely computational theory as mentioned in this paper.
Abstract: The appeal to Richness of the Base in Optimality Theory may undermine many OT analyses, as well as OT's status as a psychological, and not merely computational theory. Constraint violability in OT theorizing is critiqued as arising from a confusion between epistemological and ontological issues. Following Kiparsky (1973), conspiracies are explained as epiphenomena due to the nature of language change and language acquisition.
12 citations
TL;DR: The authors review evidence that the major empirical advances of Optimality theory over earlier work result from its theory of constraints and constraint interaction, rather than from its theories of parallel candidate evaluation, and suggest that it may be time to ask whether classical OT can be reconfigured in such a way as to consolidate its empirical advances while reexamining the nature of the input-output mapping, incorporating some of the desirable features of serialist approaches to phonology.
Abstract: These remarks review evidence that the major empirical advances of Optimality Theory over earlier work result from its theory of constraints and constraint interaction, rather than from its theory of parallel candidate evaluation. It is argued that while Optimality Theory's theory of constraints has clarified many areas of phono¬logical and morphological pattern¬ing, its implementation in terms of parallel evaluation has led to few, if any, empirical insights in these domains, while raising a number of important and largely unsolved problems. It suggests that it may be time to ask whether classical OT can be reconfigured in such a way as to consolidate its empirical advances while reexamining the nature of the input-output mapping, incorporating some of the desirable features of serialist approaches to phonology.
11 citations
TL;DR: The authors assess phonological optimality theory (OT), arguing that this theory is formulated within the conceptual framework of standard generative phonology and thus has not freed itself from its SPE-heritage.
Abstract: In this article, we start out discussing the issue of how many phonological levels one might like to distinguish. In addition, we discuss how each level is characterized and how the mapping between levels is handled. Then, against the background of a more general discussion of constraint-ba sed approaches to phonology, we assess phonological Optimality Theory (OT), arguing that this theory is formulated within the conceptual framework of Standard Generative Phonology (SGP) and thus has not freed itself from its SPE-heritage. By implicitly adopting many aspects ofSPE (such as an intra-lavel input/output distinction and one holistic mechanism), OT compares unfavorably with other available constraint-based models in needing the notion of extrinsic ordering and failing to characterize the notion of what a possible language is.
TL;DR: In this paper, the authors focus on the mechanism of constraint re-ranking which is central to OT analyses of sound change and propose an approach potentially providing new constraints for each change is difficult to reconcile with the cross-linguistic evidence for phonological phenomena usually provided in OT.
Abstract: Although Kager (1999) does not discuss sound change, it now figures increasingly in the OT literature. In any appraisal of OT, sound change provides a useful perspective, since it raises a number of problems, and brings other, more general issues into sharp focus. This article is primarily concerned with processes of historical segment loss, and with the mechanism of constraint reranking which is central to OT analyses of sound change. These changes can certainly be modelled in OT terms, but the analyses are only as insightful as the constraints they use. An approach potentially providing new constraints for each change is difficult to reconcile with the cross-linguistic evidence for phonological phenomena usually provided in OT. The status of reranking is also problematic: if it explains sound change, we must accept some form of teleology, assuming that a change takes place in order to satisfy a constraint, or to allow a constraint to be reranked; but if it is post hoc and descriptive, the motivation for the change must lie outside the theory.
TL;DR: A rule-based account of Hixkaryana is offered and is shown to be superior to the Optimality Theory account and to predict incorrect results for certain words.
Abstract: An Optimality Theory analysis of the stress and length patterns of Hixkaryana is criticized. The analysis is shown to predict incorrect results for certain words, and is shown to be incompatible with the standard Optimality Theory account of acquisition. A rule-based account of Hixkaryana is offered and is shown to be superior to the Optimality Theory account.
TL;DR: This article explored the phonology of HH sequences in Kikerewe as a case study in how the theory might account for rule ordering, showing that two-level constraints are required to even approach a description of the facts of the language.
Abstract: One of the properties held to distinguish Optimality Theory from derivational theory is that Optimality Theory has no sequential derivations, and instead correct outputs are selected on the basis of surface properties of various pronunciations, given a particular underlying form As has been made clear from many years of research within a derivational framework, the relation between underlying and surface form can be quite complex and resistant to simple surface oriented statements This article explores the phonology of HH sequences in Kikerewe as a case study in how the theory might account for rule ordering, showing that two-level constraints are required to even approach a description of the facts of the language, and ultimately the two-level approach does not work In addition, it is shown that some degree of actual derivationality is required, whereby distinct cycles of constraint evaluation are needed at the word and phrase levels
TL;DR: In this paper, the authors illustrate how optimality theory can account for a functional conspiracy by considering in detail syllable contact phenomena in Korean and show that the analysis of phonological opacity remains a challenge to the theory.
Abstract: One important breakthrough of optimality theory involves the formal analysis of the functional unity of phonological processes. Rule-based theories have no formal means of accounting for why diverse processes in a single language may conspire to have output forms meet a specific phonological shape or characteristic. I will illustrate how optimality theory can account for a functional conspiracy by considering in detail syllable contact phenomena in Korean. While the account of functional conspiracies is one of the major analytical achievements of optimality theory, there are, nonetheless, analytical challenges confronting it. I briefly touch on two such challenges, one concerning a specific prediction regarding blocking segments in harmony systems and the other concerning the analysis of phonological opacity. The overall conclusion of this article is that while optimality theory offers analytic insight and does make specific empirical predictions, the analysis of phonological opacity remains a challenge to the theory.
TL;DR: In this paper, the authors examine OT from the point of view of two related science-theoretical criteria: its formalism / notation and its status as a formal theory.
Abstract: This article begins by critically examining OT from the point of view of two (related) science-theoretical criteria: its formalism / notation and its status as a formal theory. This examination is oriented towards usual linguistic standards, rather than the stricter standards of harder sciences (which, possibly, no linguistic theory could yet achieve). The result, namely that OT is not a formal theory, even though it has a quite elaborate formalism, is explicated in science-sociological terms. The further question is posed: if OT is not a formal theory, is it at least a theory of (a part of) linguistics? The answer is that it is not, and that, given the general aims of the linguistic theory, OT should be abandoned in favor of theories which do attempt to achieve these aims.
[...]
TL;DR: The authors argued that generality is still necessary for empirical reasons, regardless of its non-role in the evaluation of competing grammars, even with the advent of Optimality Theory, which has no comparable evaluation metric.
Abstract: The Occam's Razor evaluation metric of generative phonology drove analyses to their most general statement possible, and had far-reaching effects on representations as well as on rules. With the advent of Optimality Theory, which has no comparable evaluation metric, the question arises of whether generality is still prized in formal accounts of language. In this short work, I explore this question, coming to the conclusion that generality is still necessary for empirical reasons, regardless of its non-role in the evaluation of competing grammars.