Underspecification in spoken and written discourse: interpretation, compensation and cognitive implications

Informations projet

Selfish discourse

N° de convention de subvention: 794575

DOI

10.3030/794575

Projet clôturé

Date de signature de la CE 23 Février 2018

Date de début 1 Février 2019

Date de fin 31 Janvier 2021

Financé au titre de

• EXCELLENT SCIENCE - Marie Skłodowska-Curie Actions

Coût total

€ 195 454,80

Contribution de l’UE

€ 195 454,80

Coordonné par

THE UNIVERSITY OF EDINBURGH
United Kingdom

Periodic Reporting for period 1 - Selfish discourse (Underspecification in spoken and written discourse: interpretation, compensation and cognitive implications)

Période du rapport: 2019-02-01 au 2021-01-31

Résumé du contexte et des objectifs généraux du projet

This project examines weak connectives in relation to their stronger alternatives (e.g. 'and' vs. 'so' vs. 'therefore'), to their linguistic environment and to their communicative context, in spoken and written French and English. Their importance and pervasiveness in everyday discourse calls for a better understanding of the contextual and cognitive constraints to their use and interpretation, focusing in particular on their relation to cognitive load, planning pressure and genre standards. In other words, when and why are we ambiguous when we connect our discourse? The project is rooted in usage-based linguistics insofar as it considers corpus frequency, association patterns and the role of context as crucial in the analysis of authentic data and how they reflect cognitive processes. It adopts a synchronic multi-method approach to the production and comprehension of connectives in spoken and written English and French, on a continuum from weakest to strongest markers of coherence relations and other pragmatic functions. I describe the conditions for the use of weak connectives, the types of elements that compensate for their weakly encoded processing instruction, and their impact on comprehension. To do so, I have developed a solid theoretical and methodological framework to analyze weak connectives, combining corpus-based and experimental paradigms. The results refine the divide between explicit and implicit relations by introducing weak and strong connectives in the continuum.

Travail effectué depuis le début du projet jusqu’à la fin de la période considérée dans le rapport et principaux résultats atteints jusqu’à présent

"During this project, I have carried out a large corpus study on the distribution of underspecified discourse markers and their co-occurrence with contextual cues that compensate for their weak instruction. The analyses revealed that text genre (spoken, written or computer-mediated) has little effect on the production of underspecified discourse markers or on the presence of contextual cues. However, the results showed that these reinforcing cues are more frequent in some discourse relations than others, and tend to co-occur with underspecified connectives rather than stronger markers, as expected. These results were disseminated in two journal articles (one is published in Discourse Processes, the other is still under review).

I went to Germany for three months to carry out my secondment. During that time I ran a series of experiments on the offline comprehension of weak connectives and compensation strategies. The results show that compensation strategies indeed help make discourse relations more easily interpretable, but no interaction with different types of connectives was found. This study is currently in press (accepted) in Pragmatics & Cognition.

In the final 15 months of the project, I strived to complement the corpus-based and offline production data with online processing experiments to further check for interactions and comprehension effects. I first tested the interaction between contrastive connectives and structural parallelism (i.e. repetition of the verb phrase across the two clauses), and found that reading times were faster when the ambiguous ""and"" was used in a parallel sentence rather than a sentence with two different structures. This study was published in a second Discourse Processes paper.
I then designed a new set of experiments investigating the interaction between concessive connectives and negative polarity (e.g. do not, is not), and found that reading times were again faster when the ambiguous ""and"" was used in a negative sentence rather than an affirmative one. This study has been submitted as a journal article, still currently under review."

Progrès au-delà de l’état des connaissances et impact potentiel prévu (y compris l’impact socio-économique et les conséquences sociétales plus larges du projet jusqu’à présent)

For this project, I developed a methodology for the coding of contextual discourse cues, which was applied to a large dataset. This method goes beyond previous accounts of discourse marking and offers an independent approach to contextual cues regardless of the type of discourse relation. In particular, it provides a new perspective on parallelism as a cue in contrastive relations, against the standard view of parallelism as an additive cue. In addition, the project focused on the processing of ambiguous connectives, which are often left out of psycholinguistic studies. I showed corpus-based and experimental evidence that “and” is polyfunctional rather than a monosemous additive marker. Overall, I integrated corpus-based and experimental findings on the interaction between connectives and other cues in a systematic and controlled approach.

The results of the Action have a number of direct and indirect implications for society. First of all, it allows us to understand how much information is sufficient when we write or speak and when we can allow ourselves to remain ambiguous. Secondly, the project could inform language teaching policies by showing that weak connectives are not particularly detrimental to interpretation, especially when they are compensated in context, and that very frequent connectives such as 'and' or 'so' successfully trigger adequate inferences in discourse, which nuances the stigmatisation of these expressions as “lazy speech” features.

"Reading time pattern for ""and"" vs. ""but"" in parallel and non-parallel sentences"