13e Journée de Linguistique suisse: some conference notes to showcase corpus-based research on spoken language

Last Friday, June 28th, ShareTIGR participated in the yearly meeting of the Swiss Society for Linguistics, which this time was hosted by the University of Neuchâtel. The meeting’s program consisted in a single session with 32 short project presentations and two poster sessions with 26 posters in total, most of which provided more detailed information to complement the corresponding project pitches. ShareTIGR chose this double modality, too, and you find our poster and slides attached to this blog post.

Among the many inspiring collaborative research endeavours and PhD dissertation projects that were presented, I would like to comment today on those that used spoken corpora in one way or another. They show the great vitality, in Switzerland, of corpus-based research on spoken language and interaction in various perspectives, from language acquisition to variational sociolinguistics, pragmatics, interactional linguistics and argumentation.

Laura Patrizzi (University of Basel) provided evidence from the Leo corpus (a longitudinal case study that was conducted by Heike Behrens in Leipzig and is available in Talkbank) for the hypothesis that, in language acquisition, “complex constructions build on simpler constructions”. More precisely, Patrizzi studied the German ditransitive construction, that is, uses of verbs such as geben ‘to give’, which combine with an accusative complement to encode the object given and with a dative complement to encode the recipient. The longitudinal data show that the emergence of this construction with two complements, which often have the form of pronouns, is preceded by constructions with one single reflexive pronoun. Moreover, the young child made case errors treating dative pronouns similarly to reflexive pronouns. Patrizzi argued that, jointly, these observations suggest that the child used reflexive constructions as a sort of building blocks to compose the more complex ditransitive construction.

Mathieu Avanzi and Laure Anne Johnsen, who are among the responsibles of the corpus Oral de Français de Suisse Romande OFROM, and their collaborators Yannick Emery and Aylin Pamuksaç (University of Neuchâtel) presented two initiatives by which their teams are currently inquiring about spoken French. The first is the campaign “Vos vocaux pour la science”, which aims at building a corpus of WhatsApp voice messages in the French speaking part of Switzerland. That corpus will help study a very frequent, but under-researched spoken communicative practice mediated by digital devices. It adopts a participatory approach, letting study participants decide which messages they want to donate to science and giving them the possibility to withdraw their messages at any moment. The second initiative is a set of smartphone applications called “Dis voir”. It has been designed for the general public and includes games, a tool to record one’s own realization of a set of sentences, a tool to guess the geographical origin of sentences pronounced by others, and a map that allows to listen to the collected productions associated to various places. It aims at documenting language variation in the French speaking part of Switzerland, at investigating the speakers’ knowledge and opinions at this regard and at sharing the results of the investigation with the participating community, in the spirit of citizen science.

Another investigation regarding Swiss French varieties is Suzanne Lesage’s (University of Fribourg) study of intensifiers. Lesage presented a morpho-syntactic classification of intensifiers and gave preliminary evidence showing that this class of linguistic resources varies a great deal among regions not only as far as the lexical choices are concerned, but also when it comes to the frequency of construction types. She intends to corroborate this hypothesis and to describe variation in detail studying various available corpora of spoken French.

Two of the showcased corpus-based investigations adopted the perspective of interactional linguistics. Jérôme Jacquin and Ana Keck (University of Lausanne) gave an overview over the POSEPI project, whose empirical core is the systematic annotation of epistemic and evidential expressions in a video-recorded corpus of spoken French. The corpus documents the use of French in public political debates, in TV-broadcasted political debates and in meetings at the workplace. The team consistently adopted a FAIR data approach by making the corpus accessible via an instance of the CLAPI platform, by publishing detailed information about the performed annotation in the form of an annotation manual, and by publishing the annotations as secondary data in a database designed and hosted by DaSCH. Tiziana Kowalczuk (University of Neuchâtel), on the other hand, showed first results of a study focused on address terms in recent presidential debates in France, insisting on the functional difference between these terms’ use in moves that attack the interlocutor and their use as mere regulatory devices.

Let me conclude by mentioning four studies on argumentation in spoken dialogue. Elisa Angiolini (University of Neuchâtel) investigated interactions between children and adults, in particular the phenomenon of subdiscussions. She analysed this phenomenon by reconstructing problem-solving activities as argumentative discussions with a hierarchical structure (main issues and sub-issues) and by looking at the specific ways in which children manage such emerging argumentative structures. Chiara Jermini (USI) examined dispute mediation in French and Italian. She explained how mediators recur to the strategy of naming the parties’ emotions in order to make relational issues explicit and to show the parties trajectories of change with the potential to resolve conflict. Finally, Costanza Lucchini and Giulia D’Agostino (USI) spoke about a particular type of dialogue, namely earnings conference calls, that is, press conferences in which listed companies communicate their quarterly results and respond to questions asked by financial analysts. In that institutional activity type, argumentation plays a key role, for example when managers and analysts justify their assessment of the company’s performance or make predictions that are relevant for investors. Lucchini looked at argumentation through the lens of evidential expressions, which she treated as indicators of argumentative structures and of epistemic positioning in an institutional setting that is characterized by strong epistemic asymmetry (managers, as insiders, know much more about their company than do analysts and investors). D’Agostino’s approach focused on argument mining, that is the automated extraction of arguments from a large transcribed corpus, which required adaptations of common mining techniques to the specific genre of earnings conference calls.

Johanna Miecznikowski