FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W109969595> ?p ?o ?g. }

• W109969595 abstract "Modelling the similarity of discourse connectives Ben Hutchinson (B.Hutchinson@sms.ed.ac.uk) School of Informatics, University of Edinburgh 2 Buccleuch Place, Edinburgh, EH8 9LW, UK tivation is that knowledge of which connectives are sim- ilar can complement theoretical linguistic analysis and so inform theories of discourse coherence. Eliciting sim- ilarity ratings for all pairings of discourse connectives is infeasible, however. We therefore explore the hypothe- sis that the semantic similarity of discourse connectives correlates with their distributional similarity. Abstract Discourse connectives enable discourse coherence rela- tions to be studies empirically. This paper presents two experiments on the semantic similarity of discourse con- nectives. Subjects are found to agree significantly on the similarity of pairs of connectives, and their similarity judgements are related to the ability of the connectives to be used to paraphrase each other. Subjects’ simi- larity judgements are also found to correlate positively with the distributional similarity of the connectives. Discourse connectives Introduction This paper contributes to the empirical study of dis- course connectives by considering the problems of mea- suring and modelling the similarity of discourse con- nectives. The concept of semantic similarity occupies an important role in psychology, artificial intelligence, and computational linguistics. However its applicability to discourse connectives has not been previously studied. If two discourse connectives are found to be similar, this may have implications for theories of the coherence re- lations that they signal. Discourse coherence relations contribute to the meaning of texts by specifying relationships between se- mantic objects such as events and propositions. They also assist in the interpretation of anaphora, verb phrase ellipsis and lexical ambiguities (Hobbs, 1985; Kehler, 2002; Asher & Lascarides, 2003). Som of the many theo- ries of discourse coherence, have been motivated on cog- nitive grounds. For example, Sanders, Spooren, and No- ordman (1992) propose that coherence relations be de- composed into cognitive primitives such as polarity and source of coherence. It is argued that this approach is more likely to be psychologically real than theories which posit relations as indecomposable complex objects. Knott (1996) argues for the empirical study of the dis- course connectives (e.g. because) that explicitly signal coherence relations, on the basis that relationships be- tween discourse connectives correspond to relationships between discourse coherence relations. Knott argues that if people really do use coherence relations when pro- cessing texts, then it is likely that languages will develop ways of signalling these relations explicitly. Discourse connectives thus provide a means of studying coherence relations empirically. This paper investigates the application of the concept of semantic similarity to discourseconnectives. Our mo- This section discusses the relationship between discourse connectives and coherence relations. It also introduces Knott’s substitution methodology for studying discourse connectives. Two distinct functions of discourse connectives have been distinguished by Cohen (1984): (1) enabling the faster recognition of coherence relations by the reader, and (2) allowing the recognition of coherence relations which could not be inferred in the absence of a connec- tive. This implies that in some situations the use of a connective by the writer is optional, whereas in others it is required. Moser and Moore (1995) point out that the writer has to decide which connective to use to signal a given coherence relation, as the correspondence between connectives and relations is not one-to-one. For example, both because and seeing as can be used in (1). (1) Seeing as/because we’ve got nothing but circumstantial evidence, it’s going to be difficult to get a conviction. (Knott, 1996, p. 177) This question of whether two discourse connectives can be used to signal the same relation is explored by Knott (1996), who proposes a Test for Substitutability for connectives. The test can be summarised as follows: 1. Take an instance of a discourse connective in a corpus. Imagine you are the writer who produced this text, but that you need to choose an alternative connective. 2. Remove the connective from the text, and insert an- other connective in its place. 3. If the new connective achieves the same discourse goals as the original one, it is considered substitutable in this context. For example, because is substitutable for seeing as in (1), but not in (2)." @default.
• W109969595 created "2016-06-24" @default.
• W109969595 creator A5071599724 @default.
• W109969595 date "2005-01-01" @default.
• W109969595 modified "2023-09-24" @default.
• W109969595 title "Modelling the Similarity of Discourse Connectives" @default.
• W109969595 cites W1483817654 @default.
• W109969595 cites W1534872089 @default.
• W109969595 cites W1535015163 @default.
• W109969595 cites W1574440611 @default.
• W109969595 cites W1574901103 @default.
• W109969595 cites W1586232984 @default.
• W109969595 cites W1595555765 @default.
• W109969595 cites W1604136638 @default.
• W109969595 cites W1605174196 @default.
• W109969595 cites W1659833910 @default.
• W109969595 cites W1950936596 @default.
• W109969595 cites W1985661114 @default.
• W109969595 cites W2071094139 @default.
• W109969595 cites W2080100102 @default.
• W109969595 cites W2103318667 @default.
• W109969595 cites W2109783296 @default.
• W109969595 cites W2125742935 @default.
• W109969595 cites W2165841172 @default.
• W109969595 cites W2167702024 @default.
• W109969595 cites W2208685668 @default.
• W109969595 cites W97242617 @default.
• W109969595 hasPublicationYear "2005" @default.
• W109969595 type Work @default.
• W109969595 sameAs 109969595 @default.
• W109969595 citedByCount "6" @default.
• W109969595 countsByYear W1099695952015 @default.
• W109969595 crossrefType "journal-article" @default.
• W109969595 hasAuthorship W109969595A5071599724 @default.
• W109969595 hasConcept C103278499 @default.
• W109969595 hasConcept C105795698 @default.
• W109969595 hasConcept C115961682 @default.
• W109969595 hasConcept C130318100 @default.
• W109969595 hasConcept C138885662 @default.
• W109969595 hasConcept C154945302 @default.
• W109969595 hasConcept C15744967 @default.
• W109969595 hasConcept C184337299 @default.
• W109969595 hasConcept C199360897 @default.
• W109969595 hasConcept C204321447 @default.
• W109969595 hasConcept C2780876879 @default.
• W109969595 hasConcept C2780922921 @default.
• W109969595 hasConcept C2781181686 @default.
• W109969595 hasConcept C33923547 @default.
• W109969595 hasConcept C41008148 @default.
• W109969595 hasConcept C41895202 @default.
• W109969595 hasConcept C542102704 @default.
• W109969595 hasConceptScore W109969595C103278499 @default.
• W109969595 hasConceptScore W109969595C105795698 @default.
• W109969595 hasConceptScore W109969595C115961682 @default.
• W109969595 hasConceptScore W109969595C130318100 @default.
• W109969595 hasConceptScore W109969595C138885662 @default.
• W109969595 hasConceptScore W109969595C154945302 @default.
• W109969595 hasConceptScore W109969595C15744967 @default.
• W109969595 hasConceptScore W109969595C184337299 @default.
• W109969595 hasConceptScore W109969595C199360897 @default.
• W109969595 hasConceptScore W109969595C204321447 @default.
• W109969595 hasConceptScore W109969595C2780876879 @default.
• W109969595 hasConceptScore W109969595C2780922921 @default.
• W109969595 hasConceptScore W109969595C2781181686 @default.
• W109969595 hasConceptScore W109969595C33923547 @default.
• W109969595 hasConceptScore W109969595C41008148 @default.
• W109969595 hasConceptScore W109969595C41895202 @default.
• W109969595 hasConceptScore W109969595C542102704 @default.
• W109969595 hasIssue "27" @default.
• W109969595 hasLocation W1099695951 @default.
• W109969595 hasOpenAccess W109969595 @default.
• W109969595 hasPrimaryLocation W1099695951 @default.
• W109969595 hasRelatedWork W2019160848 @default.
• W109969595 hasRelatedWork W2074376769 @default.
• W109969595 hasRelatedWork W2083778102 @default.
• W109969595 hasRelatedWork W2099265759 @default.
• W109969595 hasRelatedWork W2152197045 @default.
• W109969595 hasRelatedWork W2170496269 @default.
• W109969595 hasRelatedWork W2229981823 @default.
• W109969595 hasRelatedWork W2250862451 @default.
• W109969595 hasRelatedWork W2252159385 @default.
• W109969595 hasRelatedWork W2278298482 @default.
• W109969595 hasRelatedWork W2354342850 @default.
• W109969595 hasRelatedWork W2383671540 @default.
• W109969595 hasRelatedWork W2402752494 @default.
• W109969595 hasRelatedWork W2545174273 @default.
• W109969595 hasRelatedWork W2695048786 @default.
• W109969595 hasRelatedWork W2801262615 @default.
• W109969595 hasRelatedWork W2810725560 @default.
• W109969595 hasRelatedWork W2962225313 @default.
• W109969595 hasRelatedWork W3181334767 @default.
• W109969595 hasRelatedWork W36098481 @default.
• W109969595 hasVolume "27" @default.
• W109969595 isParatext "false" @default.
• W109969595 isRetracted "false" @default.
• W109969595 magId "109969595" @default.
• W109969595 workType "article" @default.