| Summary: | Usually, politicians have to give the image of a moral, benevolent and competent individual in order to be convincing and increase their chance to be elected. For contemporary psycho-sociologists (Poggi & D’Errico, 2010), politicians should also appear to be dominant, especially in political debates, where it becomes an essential quality. Dominance signals may include a strong displaying of self-control and more or less tacit aggressive expressions such as contempt, sarcasm, irritation, and indignation. Questions are raised about the role of vocal expressiveness in such challenges and situations. This study shows a comprehensive set of recommendations to address the methodological issues facing the phonetic analysis of political speech, beginning with perceptive evaluations. Yet voice occupies an important part of the expressive function of language (Jakobson, 1963/81), phonetic studies on political speech are, in our view, still less done compared to lexical, pragmatic and even nonverbal studies. Such a gap may be explained by the fact that expressive natural speech, produced within a particular context, raises problems in experimental protocol of the study. Despite methodological difficulties, perceptual evaluation is an essential prerequisite step in the definition and validation of the data that will be submitted to acoustical analysis. In addition, perceptual evaluation is particularly important in the study of political speech as intended ethos does not equal perceived ethos (Amossi, 1999). It is also imperative to take into account that in a natural setting, expressivity conveyed by the linguistic message (i.e. semantic content) and by paralinguistic cues most of the time can hardly be dissociated. A possible and easily conducted additional precaution could therefore be to conduct a “stand-alone” perception test of the orthographic transcriptions of the original sound stimuli which are analyzed in the auditory perception tests. This study focuses on political debates extracted from video archives of Montreuil’s city council sessions in 2013, during the last months of the mayor’s term, Dominique Voynet (Europe Ecologie Les Verts). At that time, and since a couple years already, sessions were marked by tensions of politically charged climate of the city. Five speakers were selected and extract of their speeches were submitted to perceptive evaluation (naïve speakers) and individual interviews: the mayor and four members of the opposition parties (Gaylord Le Chequer, Frédéric Molossi, Jean-Jacques Serey, Alexandre Tuaillon). A first perceptual evaluation campaign was launched on 200 stimuli (40 audio fragments x 5 speakers) concerning the perceived hostility (from none to strong) and the estimated addressee (from the audience to a specific interlocutor). Two perception tests were conducted simultaneously with 11 French participants for each: one regarding the original spoken data (audio stimuli condition), and the other regarding the orthographic transcriptions of the stimuli only (written form condition). The results of this first evaluation show that in both conditions, speakers’ extracts were widely perceived as expressive a light hostility. The scores obtained in the written form condition show that semantic content is already conveying the expression of hostility by itself. This first evaluation enabled the creation of a balanced subset of 125 stimuli (25 audio stimuli for 5 speakers) expressing variable degrees of hostility. This smaller corpus served as working material for individual interview conducted with each of the speakers. Interviews included particularly a self-evaluation task of their own stimuli. Speakers were given the task to describe their own stimuli through three 5-points Likert scales representing the dimensions of Valence (very dissatisfied/very satisfied), Arousal (very calm/very agitated), and Control (none/absolute). The information given by the politicians were then compared to a similar perceptual evaluation conducted online with 24 native French listeners (16 women, 8 men, average age= 22 years old). In the grid presented to the naïve listeners, the control dimension has been replaced by dominance (does not dominate the situation at all / absolutely dominates the situation). Such a difference is justified by the fact that it was felt appropriate to question politicians on their degree of control over their emotions in parallel and at the same time to gain information on how they were perceived by naïve listeners in terms of control of the situation.The results of the naïve dimensional evaluation show a rather homogeneous profile for all five speakers’ extracts as they were all perceived as expressing light dissatisfaction, a neutral degree of arousal (neither calm nor agitated) and a light form of dominance. Comparison with the self-evaluation illustrates some discrepancies between intended and perceived ethos: even though Dominique Voynet was the only speaker declaring being mostly expressing a light form of satisfaction, she was perceived as much unsatisfied as her opponents. Also, the light agitation related by the large majority of the speakers was perceived by the listeners in the Arousal evaluation, offering some evidence of the effectiveness of the politicians’ self-control. As a whole, the combination of the different perception tests conducted and analyzed enabled us to obtain a quantifiable description of the degree of hostility, dissatisfaction, agitation, control and dominance perceived. The collected information will serve as a guide for further acoustic analysis of the data. The acoustic parameters (F0, intensity, spectral configuration, duration, syllabic rhythm, pause duration) should be analyzed in the light of the perceptual scores and dimensional profiles obtained. Constituting a multidisciplinary team seems to be quite essential for obtaining the most fine-grained analysis possible of what underlies the power relations in an interaction such as a political debate. Indeed, as argued by Charaudeau (2005), to approach this issue, it is much necessary to work on creating a gateway between disciplines despite their different terminologies and theories. In our view, the relevance of further research on vocal expressivity in political debates can only be found in multidisciplinarity and by supporting a dialogue between different disciplines such as discourse analysis, speech sciences and social psychology.
|
|---|
