FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2767996926> ?p ?o ?g. }

• W2767996926 abstract "Distributional Statistics and Thematic Role Relationships Jon A. Willits (willits@wisc.edu) Department of Psychology, 1202 W. Johnson Street Madison, WI 53706 USA Sidney K. D’Mello (sdmello@memphis.edu) Department of Computer Science, University of Memphis Memphis, TN 38152 USA Nicholas D. Duran (nduran@mail.psyc.memphis.edu) Department of Psychology, University of Memphis Memphis, TN 38152 USA Andrew Olney (aolney@memphis.edu) Institute for Intelligent Systems, University of Memphis Memphis, TN 38152 USA Abstract cop ), eliminating simple bidirectional association strength as the locus of the effect, and (3) the failure of corpus-based approaches to account for the degree of fit between a prototypical agent or patient for a specific verb (event) and the specific NP (entity or object). Rather than associative strength of nouns and verbs based on distributional information, McRae and colleagues propose that nouns and verbs have prototype representations defined in terms of semantic features, with the fit of nouns as arguments for verbs a probabilistic function of how well its features satisfy the constraints for that verb. And while McRae and colleagues assert that this prototype information is learned through both nonlinguistic conceptual experiences with objects and actions and linguistic descriptions of those objects and actions, they also claim that linguistic experience is insufficient for learning the proper roles for verbs, and that a critical part of this knowledge is in the form of conceptual, non-linguistic representations of relations between objects, actions, and events. The main goal of this research was to determine how far simple distributional statistics can go towards capturing thematic role relationships, and to contrast the successes and failures of these statistics with McRae and colleague’s feature-based thematic fit model. The success of distributional measures is relevant to assessing whether a more complex model like that of McRae et al is necessary. Conversely, any limitations of distributional statistics would also be informative, insofar as they suggest that other types of information must be learned as well. The present research assessed the sufficiency of corpus- based distributional statistics for establishing association strengths between verbs and thematically related nouns (argument 3). If such measures can account for goodness of thematic fit, this would obviate concerns about the failures of normative association strength (argument 1). Distributional statistics may simply be more powerful predictors than association norms (see also Willits & Past research (McRae et al., 2005) has claimed that distributional statistics do not have enough structure to support representational relationships between thematically related nouns and verbs. We directly investigated this claim, using measures of distributional similarity. We found that several distributional statistics are sufficient not only to distinguish related from unrelated noun-verb pairs, but also more graded differences like obligatory vs. non-obligatory pairs. The consequences of these results for lexical association vs. feature-based thematic fit models are discussed, and suggestions are made for how future research might test feature-based and lexical-based versions of probabilistic constraint models of syntactic processing. Keywords: distributional statistics; thematic roles; event knowledge; language comprehension Syntactic Processing and Thematic Roles Syntactic processing is a crucial component of language comprehension that guides the integration of linguistic elements (e.g., words, phrases, sentences) into coherent, meaningful representations. One particular area of syntactic processing that is of current interest deals with verbs and their arguments (like agents, patients, instruments, and locations). Specifically, how are these relations represented, and are simple associations between nouns and verbs sufficient to establish this relationship? McRae and colleagues (Ferretti, McRae, & Hatherall, 2001; McRae, Hare, & Tanenhaus, 2005; McRae et al, 1997; McRae et al, 2005) cite three arguments regarding the insufficiency of direct lexical association strength as the basis for thematic role comprehension: (1) lack of normative association strength between noun-verb pairs that nonetheless prime each other or facilitate reading times; (2) experimental evidence that this facilitation only occurs when nouns are in proper, role-fitting constructions (e.g. facilitation for the cop arrested the woman but not the woman arrested the" @default.
• W2767996926 created "2017-11-17" @default.
• W2767996926 creator A5006905045 @default.
• W2767996926 creator A5008316506 @default.
• W2767996926 creator A5010483444 @default.
• W2767996926 creator A5028699253 @default.
• W2767996926 date "2007-01-01" @default.
• W2767996926 modified "2023-09-23" @default.
• W2767996926 title "Distributional Statistics and Thematic Role Relationships" @default.
• W2767996926 cites W1972170138 @default.
• W2767996926 cites W1983578042 @default.
• W2767996926 cites W1986707196 @default.
• W2767996926 cites W2011999325 @default.
• W2767996926 cites W2050092301 @default.
• W2767996926 cites W2051534348 @default.
• W2767996926 cites W2058616517 @default.
• W2767996926 cites W2094931100 @default.
• W2767996926 cites W2103748294 @default.
• W2767996926 cites W2106480521 @default.
• W2767996926 cites W2152478872 @default.
• W2767996926 hasPublicationYear "2007" @default.
• W2767996926 type Work @default.
• W2767996926 sameAs 2767996926 @default.
• W2767996926 citedByCount "4" @default.
• W2767996926 countsByYear W27679969262012 @default.
• W2767996926 crossrefType "journal-article" @default.
• W2767996926 hasAuthorship W2767996926A5006905045 @default.
• W2767996926 hasAuthorship W2767996926A5008316506 @default.
• W2767996926 hasAuthorship W2767996926A5010483444 @default.
• W2767996926 hasAuthorship W2767996926A5028699253 @default.
• W2767996926 hasConcept C121934690 @default.
• W2767996926 hasConcept C138885662 @default.
• W2767996926 hasConcept C154945302 @default.
• W2767996926 hasConcept C15744967 @default.
• W2767996926 hasConcept C204321447 @default.
• W2767996926 hasConcept C2776397901 @default.
• W2767996926 hasConcept C2778987988 @default.
• W2767996926 hasConcept C41008148 @default.
• W2767996926 hasConcept C41895202 @default.
• W2767996926 hasConcept C59822182 @default.
• W2767996926 hasConcept C86803240 @default.
• W2767996926 hasConceptScore W2767996926C121934690 @default.
• W2767996926 hasConceptScore W2767996926C138885662 @default.
• W2767996926 hasConceptScore W2767996926C154945302 @default.
• W2767996926 hasConceptScore W2767996926C15744967 @default.
• W2767996926 hasConceptScore W2767996926C204321447 @default.
• W2767996926 hasConceptScore W2767996926C2776397901 @default.
• W2767996926 hasConceptScore W2767996926C2778987988 @default.
• W2767996926 hasConceptScore W2767996926C41008148 @default.
• W2767996926 hasConceptScore W2767996926C41895202 @default.
• W2767996926 hasConceptScore W2767996926C59822182 @default.
• W2767996926 hasConceptScore W2767996926C86803240 @default.
• W2767996926 hasIssue "29" @default.
• W2767996926 hasLocation W27679969261 @default.
• W2767996926 hasOpenAccess W2767996926 @default.
• W2767996926 hasPrimaryLocation W27679969261 @default.
• W2767996926 hasRelatedWork W1017558726 @default.
• W2767996926 hasRelatedWork W1534993021 @default.
• W2767996926 hasRelatedWork W1981617416 @default.
• W2767996926 hasRelatedWork W1983578042 @default.
• W2767996926 hasRelatedWork W1984251878 @default.
• W2767996926 hasRelatedWork W1986707196 @default.
• W2767996926 hasRelatedWork W2134601421 @default.
• W2767996926 hasRelatedWork W2140690392 @default.
• W2767996926 hasRelatedWork W2201206544 @default.
• W2767996926 hasRelatedWork W2242645069 @default.
• W2767996926 hasRelatedWork W2273636673 @default.
• W2767996926 hasRelatedWork W2576450548 @default.
• W2767996926 hasRelatedWork W264385502 @default.
• W2767996926 hasRelatedWork W2738473203 @default.
• W2767996926 hasRelatedWork W2804541399 @default.
• W2767996926 hasRelatedWork W3012685695 @default.
• W2767996926 hasRelatedWork W33386844 @default.
• W2767996926 hasRelatedWork W657303157 @default.
• W2767996926 hasRelatedWork W836146308 @default.
• W2767996926 hasRelatedWork W993336757 @default.
• W2767996926 hasVolume "29" @default.
• W2767996926 isParatext "false" @default.
• W2767996926 isRetracted "false" @default.
• W2767996926 magId "2767996926" @default.
• W2767996926 workType "article" @default.