(Salle F106)
Balthasar, Lukas
Mondada, Lorenza
Dep. of Linguistics, Univ. Lyon2 & Laboratoire ICAR (CNRS)
Multiscope video and the continuous access
to relevant details
In this paper, we argue that conversation analysis - with its requirement
to work on naturally occurring activities and on the details to
which members are actually orienting - entails particular ways of
conceiving and of realizing video recordings. We show that single
camera views are often limited in offering a continuous access to
the relevant details organizing the interaction. In this case, multiple
source recordings (with various video cameras as well as various
microphones) are necessary : we will show how multiscope videos,
i.e. videos edited with a split screen containing 3 or 4 views on the
same event, allow a better detailed understanding of action.
The paper will be based on a series of corpora we realized at the ICAR
lab, consisting in multi source recordings of ordinary conversations
during meal time. These recordings were edited in order to synchronize
the various views in a multiscope video : we will show the relevance
of such edited data for the analysis of particular phenomena
such as exchanges of gazes, trajectories of gestures and body movements.
Multiscope video secure a continuous access to these details
which would otherwise be covered by the relative positions of the
participants or by other perspective effects. In this sense, it allows a
transcription and a description of the relevance of glances, gestures
and bodily postures during entire stretches of talk, indipendently
from recording limitations.
In the paper we will stress the importance, for this kind of analysis,
of a specifi c « professional vision » of the analyst, refl exively
embodied in the way in which interaction is recorded and records are
edited and then read-interpreted during the enquiry.
L’analyse conversationnelle - avec son exigence de travailler sur des
données dites « naturelles », i.e. sur des interactions telles qu elles
se déroulent dans leur contexte social ordinaire, sans avoir été
orchestrées par le chercheur - implique une manière spécifi que de
concevoir et de réaliser les enregistrements vidéo qui constituent
son fondement empirique.
Nous allons donc discuter de différentes manières de fi lmer une activité,
en insistant sur les limitations des prises de vues uniques et
sur les potentialités des prises de vues multiples. Celles-ci, une fois
montées en une « vidéo multiscope » offrant simultanément plusieurs
perspectives sur une même scène, permettent de travailler les
détails pertinents de l’interaction d’une manière continue, non interrompue
par des problèmes de cadrage ou de changement de posture
des participants. Nous insisterons dans cette communication sur le
type d’analyse spécifique que ces vidéos multiscopes permettent de
réaliser.
