FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2397986080> ?p ?o ?g. }

• W2397986080 abstract "Constraints on Abstraction: Generalization Across Languages Sara Finley (finleysr@plu.edu) Department of Psychology, Pacific Lutheran University Tacoma, WA, 98447, USA Abstract Recent research has suggested a division between lexical representations and phonological patterns; lexical items are stored with talker-specific information, while phonological patterns are represented at a separate, abstract level of representation (Finley, 2013; Smolensky & Legendre, 2006). The present paper provides further evidence for this proposal, demonstrating that learners will extend a novel phonological pattern (vowel harmony) to speakers of a novel dialect when the words are familiar, but not when the words are unfamiliar, further supporting a distinction between the representations of lexical items and the representations of phonological patterns. Keywords: statistical learning, vowel harmony, phonological representations, lexical representations. Introduction One of the major issues in the cognitive science of language is how the phonological representations that make up words are stored in the human mind. Specifically, there is a debate between whether the phonological representations that govern word formation are represented in the mind in terms of abstract, rule-governed processes, or as exemplars of specific instances, based on the phonetic forms of lexical items. While classic phonological theory has proposed highly abstract, symbolic representations for phonological patterns (Chomsky & Halle, 1968), experimental research has questioned the abstract nature of phonological processes, and suggested that phonological rules are epiphenomenal (Port & Leary, 2005). Because listeners encode highly detailed information from the individual speakers they encounter (Nygaard & Pisoni, 1998), researchers have proposed that representations of lexical items are not based on normalized or abstract representations, but based on individual, stored exemplars (Johnson, 1997). Because the generalizations that can be made about phonological restrictions for word formation (e.g., phonotactics and morphophonology) necessarily come from examination of lexical items, it is reasonable to assume that these phonological restrictions share the same constraints on representations as lexical items (i.e., fine- grained, talker-specific representations) (Pierrehumbert, 2001). However, it is also possible that language makes use of multiple levels of representation, with fine-grained exemplar-based representations at one level, and abstract representations at another, higher level (Smolensky & Legendre, 2006). One of the difficulties in distinguishing between an abstract model of phonological representations and an exemplar model of phonological representations is in teasing apart the representations of individual words and the representations of phonological patterns, since phonological patterns are instantiated through word formation. In order to tease apart the distinction between lexical representations and phonological representations, the researcher must look to how the language handles phonological patterns in novel words. While researchers have investigated questions of phonological representations using a phonological version of a wug test (Berko, 1958), these investigations typically assume an abstract, generative view of phonological representations (Becker, Ketrez, & Nevins, 2011; Becker, Nevins, & Levine, 2012; Gouskova & Becker, 2013), and therefore do not address the question of a distinction between lexical representations and phonological rules. Another approach to testing the differences between lexical items and phonological patterns is by testing how learners treat novel and familiar forms following a brief exposure to a novel language. If there is a distinction between lexical representations and abstract phonological patterns, then learners should treat familiar, known words differently than unfamiliar, unknown words. Recent research used learning to test the hypothesis that lexical representations and abstract phonological patterns are stored under distinct representations (Finley, 2013). Participants in an artificial grammar learning experiment were exposed to a novel phonological pattern (vowel harmony, where vowels in a word shared the same value of a phonological feature, back and round), and then were tested on their knowledge of that phonological pattern using both familiar words and unfamiliar, novel words. In addition, participants were tested on the same set of familiar and novel items, but these items were spoken by an unfamiliar talker of the opposite gender. Participants were able to extend the newly learned vowel harmony pattern to both novel and familiar items, for both novel and familiar talkers. In addition, the analyses tested for a ‘transfer deficit’ from familiar to novel talkers. The more speakers make use of talker-specific representations in learning, the larger the transfer deficit should be when accepting items spoken in an unfamiliar voice. There was a clear transfer deficit for familiar items, but not for novel items. This difference suggests that learners store known items in terms of talker-specific phonetic details, but make use of abstract phonological patterns when making grammaticality judgments for novel words. These abstract representations are independent of the specific talker heard during training. The ability generalize a newly learned phonological pattern to unfamiliar talkers for both familiar and novel" @default.
• W2397986080 created "2016-06-24" @default.
• W2397986080 creator A5051453745 @default.
• W2397986080 date "2014-01-01" @default.
• W2397986080 modified "2023-09-23" @default.
• W2397986080 title "Constraints on Abstraction: Generalization Across Languages" @default.
• W2397986080 cites W1562911371 @default.
• W2397986080 cites W1598851216 @default.
• W2397986080 cites W2037202300 @default.
• W2397986080 cites W2049521441 @default.
• W2397986080 cites W2067965259 @default.
• W2397986080 cites W2088443227 @default.
• W2397986080 cites W2118219228 @default.
• W2397986080 cites W2129087637 @default.
• W2397986080 cites W2136231143 @default.
• W2397986080 cites W2149774313 @default.
• W2397986080 cites W2186780112 @default.
• W2397986080 cites W3178156975 @default.
• W2397986080 cites W3216401400 @default.
• W2397986080 hasPublicationYear "2014" @default.
• W2397986080 type Work @default.
• W2397986080 sameAs 2397986080 @default.
• W2397986080 citedByCount "0" @default.
• W2397986080 crossrefType "journal-article" @default.
• W2397986080 hasAuthorship W2397986080A5051453745 @default.
• W2397986080 hasConcept C111472728 @default.
• W2397986080 hasConcept C124304363 @default.
• W2397986080 hasConcept C134306372 @default.
• W2397986080 hasConcept C138885662 @default.
• W2397986080 hasConcept C148934300 @default.
• W2397986080 hasConcept C151670969 @default.
• W2397986080 hasConcept C15744967 @default.
• W2397986080 hasConcept C177148314 @default.
• W2397986080 hasConcept C180747234 @default.
• W2397986080 hasConcept C2779581591 @default.
• W2397986080 hasConcept C2781280609 @default.
• W2397986080 hasConcept C33923547 @default.
• W2397986080 hasConcept C41895202 @default.
• W2397986080 hasConceptScore W2397986080C111472728 @default.
• W2397986080 hasConceptScore W2397986080C124304363 @default.
• W2397986080 hasConceptScore W2397986080C134306372 @default.
• W2397986080 hasConceptScore W2397986080C138885662 @default.
• W2397986080 hasConceptScore W2397986080C148934300 @default.
• W2397986080 hasConceptScore W2397986080C151670969 @default.
• W2397986080 hasConceptScore W2397986080C15744967 @default.
• W2397986080 hasConceptScore W2397986080C177148314 @default.
• W2397986080 hasConceptScore W2397986080C180747234 @default.
• W2397986080 hasConceptScore W2397986080C2779581591 @default.
• W2397986080 hasConceptScore W2397986080C2781280609 @default.
• W2397986080 hasConceptScore W2397986080C33923547 @default.
• W2397986080 hasConceptScore W2397986080C41895202 @default.
• W2397986080 hasIssue "36" @default.
• W2397986080 hasLocation W23979860801 @default.
• W2397986080 hasOpenAccess W2397986080 @default.
• W2397986080 hasPrimaryLocation W23979860801 @default.
• W2397986080 hasRelatedWork W1512292014 @default.
• W2397986080 hasRelatedWork W1557334090 @default.
• W2397986080 hasRelatedWork W1979804675 @default.
• W2397986080 hasRelatedWork W2016216792 @default.
• W2397986080 hasRelatedWork W2057925568 @default.
• W2397986080 hasRelatedWork W2089938479 @default.
• W2397986080 hasRelatedWork W2121300535 @default.
• W2397986080 hasRelatedWork W2364747197 @default.
• W2397986080 hasRelatedWork W613881584 @default.
• W2397986080 hasVolume "36" @default.
• W2397986080 isParatext "false" @default.
• W2397986080 isRetracted "false" @default.
• W2397986080 magId "2397986080" @default.
• W2397986080 workType "article" @default.