Towards an Operational Definition of Discourse Contrast

Abstract

Linguists have often invoked the concept of contrast in attempting to explain the use of certain constructions without explicitly defining this term. The present paper proposes an operational definition of contrast which can be applied to naturally occurring data so as to provide a more precise account of the relationship between form and function. The specific problem addressed is word order in Biblical Hebrew and Chinese. It is shown that contrast is one factor affecting deviations from canonical word order in these languages, but that the effect of contrast differs in a number of respects in the two languages, and many deviations from canonical word order cannot be accounted for by reference to contrast.

What do subject pronouns do in discourse? Cognitive, mechanical and constructional factors in variation

“…In the next section we present the corpus and variable context. In Section 3 we apply three operationalizations of contrast independently elaborated in previous literature, namely "double contrast" (Myhill and Xing 1996), juxtaposition of references to the speaker and interlocutor (Travis and Torres Cacoullos 2012), and Potential Referential Interference (Sun and Givón 1985), and nd that none of these measures o ers an account of yo expression in these data. Abandoning the attempt to individually classify tokens as contrastive, in Section 4 we use multivariate analysis and nd that speakers are more likely to express yo in the presence of intervening human subjects between coreferential 1sg subjects, a cognitive e ect; and when the previous coreferential rst singular subject was realized as yo and when the subject of the immediately preceding clause was realized pronominally, two distinct manifestations of priming, a mechanical e ect.…”

Stress onI

“…In sum, coding tokens one by one directly for contrast as a unitary global concept, even when done in a replicable manner through the operationalization of converse predicates (cf. Myhill & Xing 1996), proved a fruitless exercise and the results of the indirect tests for "contrasting opinions" (Dehé & Wichmann 2010a, 2010b did not go in the predicted direction.…”

“…Here we test operationalizations of contrast applicable to natural speech. We begin with an operationalization of contrast in a semantic sense, that of "converse predicates", based on Myhill & Xing (1996), and then consider contrast in an interactional sense, related to the expression of speaker opinion vis-à-vis the interlocutor (Dehé & Wichmann 2010a, 2010b, by examining turn position in relation to cognition verbs and the preceding clause subject.…”

“…In order to address these issues, we first configure two operational definitions of contrast (one related to converse predicates (cf. Myhill & Xing 1996), and the other related to the marking of contrastive attitudes (e.g., Dehé & Wichmann 2010a), which we apply to the subject pronoun I in a corpus of conversational American English. We find that neither explains the patterning of stressed I, and we therefore turn from such unitary contrast accounts to exploration of the set of factors jointly conditioning the variation, from which we are able to offer a more complete account of actual usage.…”

Entre syntaxe, prosodie et discours : les topiques sujets en français parlé

“…Il faut remarquer d'autre part que les données issues d'expériences de laboratoire permettent de limiter les possibles variables en jeu. Par exemple, nos données pouvaient varier par rapport aux facteurs suivants : la présence ou non des alternatives dans le discours; le fait que le topique se trouve dans le premier ou bien dans le deuxième énoncé d'un pair d'alternatives, et donc que le contraste soit ou non déjà explicité au moment de l'énonciation du topique (l'importance de ce dernier facteur a été souligné par Myhill & Xing, 1996 à propos de l'hébreu biblique); la présence ou non de mots ou expressions qui explicitent le contraste (mais, d'autre part, etc.). La quantité de données que nous avons à disposition pour le moment ne nous permet pas de tester ces différents cas séparément.…”