Isabella Poggi, Francesca D’Errico, Laura Vincze,
Alessandro Vinciarelli (Eds.)
Politicalspeech.Multimodalcommunicationtoshapeminds and social action
SSPNet International Workshop.
Rome, Italy, November 10-12, 2010
RevisedSelectedPapers
Onechild, oneteacher, one book and onepen can change the world
We dedicate this book toMalala, a young Afghani woman woundedwhilestrugglingfor the educationofallchildren in the world.
A brightexampleof the politicalrelevanceofspeech.
Introduction
Political speech: words and bodies as a means for social influence
From political communication to social influence
Political discourse has been a subject for the study of communication since the very beginning of western philosophical speculation: [1], Sophists, and the first studies in rhetoric find their roots in Greek democracy, with its need to elaborate and teach techniques for reasoning and argumentation. Later, if [2] set the stage for research in argumentation for the next centuries, [3] and [4] also took into account body communication as an important part of the Orator’s repertoire, by focusing on the use of gestures, voice, posture, gaze and facial expression.
In the last century, the study of political discourse, within studies on persuasion was mainly taken up by the New Rhetoric [5], Argumentation Theory [6] and Sociology [7]. In Psychology the Elaboration Likeliness Model [8] and the heuristic-systematic model [9] particularly influential since the late eighties, distinguished a central and a peripheral route taken by the Receiver of a persuasive message, on the one side its content, on the other its perceptual and affective aspects. But such distinction between the rational side of persuasion and its seemingly marginal features is not so different from the one made by [2] of three aspects of persuasion: logos – the logical argumentation, ethos – the character of the persuader, and his capacity to inspire trust, and pathos – the emotions of the audience; the last two being most typically conveyed by perceptual and affect-inducing features of the message – attractiveness of the source, his voice, body appearance, but also his charisma, as it appears from his physical and mental qualities.
While most literature on argumentation and fallacies has mainly focused on the aspects of logos, more recently the affective aspects of persuasion have been stressed [10; 11; 12] and the bodily features of persuasive behavior have been investigated [13; 14; 15; 16;]
However, during the last twenty years Computer Science has burst on the scene of multimodality, mainly due to interest in the detection and synthesis of Social Signals. In order to build systems for the automatic recognition of gestures, head movements, facial expressions, postures [17; 18], and to simulate them in Embodied Agents and other graphic computer-human interfaces [19], computer scientists and social scientists have started collaborating to enhance each other’s research results: social scientists’ in-depth analyses of communicative signals contribute to build detection algorithms and simulations in ECAs, while system evaluations provide a test of the computational models produced, and a feedback for future research.
On November 10th – 12th, 2010, on behalf of the European Network of Excellence SSPNet (Social Signal Processing Network), aimed at creating a strong connection between Social Scientists and Computer Scientists in research on social signals and in their automatic processing, an International Workshop on “Political Speech” was held at Roma Tre University, Department of Education Sciences, in which important representatives of various disciplines – Social Psychology, Political Science, Linguistics, Theory of Argumentation, Multimodal Communication, Natural Language Processing, Human-Computer Interfaces, etc. – met to investigate and discuss mechanisms, processes and tools of political communication. This volume is a follow-up of that common work.
The book has two main objectives: the first is to widen the study of political discourse from the analysis of bare verbal monological discourse and dialogic interaction to the analysis of the whole multimodal message; the very title “Political Speech” intends to stress that a thorough consideration of political discourse in face-to-face interaction encompasses not only the words and argumentation employed by politicians, but their voice quality, prosody, intonation, their gestures, gaze and facial expressions, posture, head and body movements. The second objective is to consider the impact of technology and Computer Science on political discourse, their effects over people, and their methods of analysis, by stressing the two-ways interactions between them.
On the one side, technology may serve as a tool for the analysis of political speech, but the analysis in its turn may provide hints to the construction of systems for automatic recognition or simulation in Embodied Agents. The ultimate goal of such work is to turn the knowledge achieved in Sociology, Political Science, Linguistics, Psychology into an effort to implement more persuasive technologies, for example by evaluating the persuasive impact of some words, prosodic or gestural features, argumentative or affective strategies, or by providing hints for the synthesis of persuasive agents.
Issues in political communication
Among the various relevant topics in the study of political communication, one is its heavy intertwining with technology: social networks may count both as a tool and as a database for data mining and sentiment analysis research [20]. The use of social networks by Barack Obama in the States or BeppeGrillo in Italy revealed a shift from TV to interactive media as the main route to electoral consensus. Was the application of sentiment analysis to these media simply a consequence or possibly even a cause of Obama’s victory or political activism? [21; 22; 23; 24].
These phenomena require a novel look at some classical results of social psychology research. Should social networks have existed at the time of Moscovici [25], would he have phrased the construct of minorities’ influence the way he did? Are social networks more a case of majority or minority influence? Models of persuasion in the first part of the 20th century – since the bullet theory [26] on – see the receiver as a passive subject, while those since the sixties (e.g. [27]) credit him/her with a more active attitude. Nowadays a person navigating in the new media might look more like the latter than the former: s/he can participate more, open a new group, feel in a peer relationship with others; s/he is not subjected to information but may search for it. Therefore a topic that the study of political communication must take into account is the web revolution.
Another relevant issue in political discourse refers to its being a case of persuasive communication, hence a way to influence people’s action through influencing their beliefs. This raises, among others, the issue of knowledge manipulation. Within the many possible ways to conceal, withdraw, distort information, are there some verbal or bodily strategies that are most typically used in politics? Are there particularly subtle ways to manipulate an addressee’s mind, like for instance, the use of fallacies, obscurity, ambiguity or vagueness? Are there ways to train laypeople to defend themselves from such manipulation strategies? How might educational programs – for example, courses in Critical Thinking – take advantage of new findings to enhance political self-consciousness in young and older citizens? Might technology be of help in this, not only through the immediate spreading of news and action decisions, e.g., by Twitter, but also, for instance, by implementing systems for the automatic recognition (or for training human recognition) of deceptive messages, vague information, or fallacious argumentations?
Actually, since not only the clarity or sincerity of the message, but also its source is of the utmost importance for its efficacy, a relevant issue in research on political speech is the management of the politician’s image, in which body behavior has a great role. What aspects of a politician’s multimodal communication are mainly responsible for the impression s/he gives to the audience? The notion of charisma, along with the physical aspects in which it is manifested (voice, expression, style of behavior) is presently a subject of investigation [28; 29] but given its being a multidimensional construct encompassing mental, affective, perceptual, social aspects, only an interdisciplinary effort might be able to disentangle its multifaceted nature. Furthermore, besides caring their own self-presentation, politicians often try to undermine their opponents’ image, using the weapons of discredit and denigration. What are, at present, the means of political delegitimization? Is political discourse more blatantly aggressive than it was in the past? How does the use of new media affect the public spreading of bad reputation?
Finally, a topic presently investigated also as to its multimodal direct and indirect signals is conflict and its dynamics of escalation, negotiation, reconciliation [30; 31]. Since politics is a prototypical case of conflict, studying conflict and their signals in the paradigmatic scenario of political debates might provide new tools also for recognizing conflict dynamics in other fields, like between couples, or at the workplace. Moreover, while conflict has been studied more as to the internal psychological issues – for example, studies on reconciliation [32; 33] have highly stressed the importance of internal feelings of victims and perpetrators of violence – what has still to be investigated in depth are the signals that most specifically indicate or trigger negotiation and reconciliation processes. And once you master signals of negotiation or reconciliation you might manage these processes better, once you can read signals of escalation, you might learn to prevent escalation.
The tangled net of political speech
The papers in this book analyze political speech in various modalities, while adopting a number of approaches and disciplines and dealing with various topics.
Some papers analyze political communication in the verbal modality, by taking only or mainly monological discourse into account (Cedroni;Longobardi;Conoscenti;Catellani et al.;Sensales et al.; Bongelli et al.;ZurloniAnolli) and studying their lexical, textual or rhetorical patterns. Other works, relying on a conversation analysis approach (Koutsombogera & Papageorgiou), take the structure of turn-taking, overlaps and interruptions during political discussions as a cue to the social and communicative relationships of power and dominance. Two papers take into account the acoustic modality investigating phonetic aspects of political speech (Martin andSalvatiPettorino) or analyzing the lexical and syntactic structures that trigger laughter or applause (Guerini et al.), while others focus on the visual aspects (gaze, gestures, facial expressions) of communicative interaction during political talk shows or parliamentary speeches (D’Errico et al.; Leone;Maricchiolo et al.; Paggio & Navarretta; Shaw) or on linguistic as well as bodily, visual and acoustic aspects of political discourse in presidential rallies (Gelang) and TV spots (Pellegrino et al.).
Contributors come from different disciplines and research areas, from political science (Cedroni) to social psychology (Catellani et al.; D’Errico, et al.; Leone;Sensales et al.;Maricchiolo et al.), linguistics (Longobardi; Bongelli et al.), argumentation theory (Zurloni), rhetoric (Gelang) phonetics (Martin;SalvatiPettorino), computational linguistics (Conoscenti; Guerini), conversation analysis (Paggio & Navarretta;KousombogeraPapageorgiou), and their approaches range from traditional and lexicographic text analysis (Cedroni;Longobardi;ZurloniAnolli; Sensales et al.) to data mining (Conoscenti; Guerini et al.), from experimental research to observation and analysis of corpora. The fragments of political communication taken into account involve politicians from diverse political tendencies and diverse countries: U.S.A., Ireland, Greece, France, Italy.
Although the papers in this book might be clustered in very different ways, according to even other criteria beside the ones above, here we distinguish them following a classic of political discourse, Aristotle’s “Rhetoric”, that devotes the first book to the Orator, the second to the Audience, the third to Discourse proper. In Part I we then deal with the Sender of the message, that is, with the aspects of his/her multimodal behavior that contribute to the persuasiveness of political discourse; Part II deals with the effects of the Orator’s behavior on the Audience; and Part III with the content and structure of the Discourse: on the one side, what are the topics specifically chosen by the Orator in his/message, on the other its lexical, discursive, rhetorical, argumentative structure.
The Orator
KoutsombogeraPapageorgioustudy persuasion and interruption attempts during political interviews as cues to conversational dominance [34]. They describe the contribution of participants’ multimodal behavior to the management of interruptions and, in so doing, to the achievement of their persuasive goals. All instances of overlaps accompanied by non-verbal activity were automatically extracted, distinguishing collaborative overlaps, such as turn-completing or feedback, from competitive overlaps (pure interruptions) aimed at taking the floor to restrict the conversational rights of the other speaker. KoutsombogeraPapageorgiou notice the speakers' tendency to make use of more than one modality to interrupt: debaters make use of facial expressions, gestures, body posture movements. Interruptions are assessed in terms of success and contribution to dominance and persuasiveness (seen from the perspective of the reactions of the co-locutor to the interruption). Although an interruption is generally considered successful and an interruptor as dominant when the interruptee withdraws and the interruptor completes the turn, this is not always the case. The interruptee, by managing the interruption and advancing counter-arguments, or by facial expressions communicating annoyance, disconfirmation or surprise, might lessen the persuasive effect of the interruptor and his temporary dominance. KoutsombogeraPapageorgiou's analysis gives us a broad picture of the multimodal behavior of both interruptor and interruptee, emphasizing the power of the facial expressions of the latter in diminishing the successful impact of the interruption.
Another study which emphasizes the importance of multimodal behavior in presidential debates is Gelang’s comparative analysis of Barack Obama’s and Hillary Clinton’s actio. With a rhetorical approach to multimodal communication, Gelang examines the politicians’ actio – their set of body behaviors in delivering a discourse – in relation to the concept of ethos and its possible argumentative dimensions. Starting from the importance of actio in the rhetorical training of classical orators, Gelang finds out two rhetorical actio strategies used in political debates: enacted (active energetic, dynamic actio) and restrained (moderated, limited degree of expressiveness and energy). These two rhetorical strategies can function as ethos-related argumentation, and be used by debaters to acclaim or defend their own ethos and/or to attack the ethos of the opponent, influencing the way politicians are perceived by the audience. In her comparative study of Hillary Clinton’s enacted, passionate ethos and Obama’s restrained ethos, Gelang puts forward an interesting possible reason why electors voted for Obama: exhibiting a restrained ethos may be perceived as being more fit to becoming a president, since it is seen as an evidence of being in control of the situation.
An opposite case of a negative image projected by a politician’s multimodal behavior is illustrated by Shaw in her analysis of theunpopular Irish female Minister of Education. Shaw analyzes both the Minister’s discourses in the Northern Ireland Assembly and the interviews released by her colleagues and opponents about the Minister. Her linguistic and body communication style (finger wagging, planting both feet on the floor, aggressive sentences, no mitigation and in general her “confrontational stance”) is stigmatized as a definitely “masculine” style, that according to the Author contributed to her becoming very unpopular, so much so as to be the most likely cause for people not to vote for her in subsequent elections.
The next two chapters provide synchronic and diachronic analysis of the acoustic features of three important exponents of French and Italian politics. Martin, after presenting the notions of Prosodic Events and Prosodic Structure, the general principles of prosody and the constraints to prosodic structures in read and spontaneous speech, applies the software Winpitch to analyze fragments of speech by Nicolas Sarkozy e Ségolène Royal. In Royal he finds a high quantity of text prefixes, and melodic contours very often falling but never conclusive, that do not allow the listener to finally process the whole sentence, and make her sentences very similar to the long sentence prescribed by the canons of the ÉcoleNationaled’Administration; this, characterizing her speech as typical of the dominant class, might have possibly backfired against her pretense to be seen as a leftist leader. To the contrary, Sarkozy splits single kernels into smaller units ended by conclusive contours, allowing the listener to process his sequence of text faster and more easily, which may result in a more comprehensible and appealing speech.
SalvatiPettorino present a diachronic analysis of the suprasegmental aspects of Berlusconi’s speech from 1994 through 2010. Taking into account his spontaneous speech during discourses, debates and interviews, in a corpus focused on common themes, they measure duration of speech chains, number of syllables per chain, maximum and minimum pitch, and calculate the prosodic features of articulation rate, speech rate, average duration of silent pauses, fluency and tonal range. Their results show a fair consistency of Berlusconi’s speech across time for all parameters, along with a high level of adaptation to different interlocutors and situations. For example his speech rate – the number of words per time unit – that is considered a cue to dominant versus submissive relationship with the interlocutor, is lower during an interview with a journalist quite prone to his will than with others he is afraid of, before whom he tries to avoid silences, probably not to let them take the turn. The same effect is found for tone range, typically lower in prepared discourses in which Berlusconi feels in a dominant position and wants to convey authority and self-confidence. By calculating duration of his silent pauses, the Authors find that Berlusconi makes more use of emphatic pauses in discourses than in interviews, probably, again, because a longer silence might give a journalist the opportunity to take the floor. In conclusion, even from the acoustic point of view, Berlusconi’s skills as communicator – his clarity, witnessed by low articulation rate, and his capacity of adapting his speech to different power relationships – show a remarkable constancy overtime.
The Audience
A second set of papers examines verbal or body communicative behavior while also focusing on their effects over the audience.