Charisma perception in political speech: A case study

Full text

Charisma perception in political speech: a case study
Rosario Signorello, Francesca D’Errico, Isabella Poggi, Didier Demolin, Paolo
Mairano
To cite this version:
Rosario Signorello, Francesca D’Errico, Isabella Poggi, Didier Demolin, Paolo Mairano.
Charisma perception in political speech: a case study. Heliana Mello, Massimo Pettorino,
Tommaso Raso. International Conference on Speech and Corpora (GSCP 2012), Feb 2012,
Belo Horizonte, Brazil. Firenze University Press, pp.343-348. <hal-00820359>
HAL Id: hal-00820359
https://hal.archives-ouvertes.fr/hal-00820359
Submitted on 18 May 2013
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entific research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destin´ee au d´epˆot et `a la diffusion de documents
scientifiques de niveau recherche, publi´es ou non,
´emanant des ´etablissements d’enseignement et de
recherche fran¸cais ou ´etrangers, des laboratoires
publics ou priv´es.

Charisma perception in political speech: a case study
Rosario SIGNORELLO
1,2
,FrancescaD'ERRICO
2
, Isabella POGGI
2
, Didier DEMOLIN
1
,
Paolo MAIRANO
1
1
GIPSA-Lab, UMR 5216 CNRS, Grenoble INP, UJF, U. Stendhal (Grenoble, France);
2
Dipartimento di Scienze
dell’Educazione,Universita`degliStudiRomaTre(Rome,Italy)
rosario.signorello@gipsa-lab.fr, didier.demolin@gipsa-lab.fr, fderrico@uniroma3.it, poggi@uniroma3.it,
paolomairano@gmail.com
Abstract
The charisma of the leader is conveyed through multiple aspects: his ideas and vision and his perceivable verbal and non verbal
behaviors. Among these perceivable behaviors there are the acoustic characteristics of speech. We present here a study on the
perception of charisma in political speech. We collected speech statements with different illocutionary value taken from two speeches
given by Umberto Bossi, the leader of an Italian party, before and after a stroke which caused him a voice disorder. Stimuli from the
two condition differed significantly in the acoustic-prosodic features. In the first part of the study 40 French listeners rated normal
speech stimuli (20 pre- and 20 post-stroke) and in the second part 22 French (11 pre- and 11 post-stroke) and 31 Italians (15 pre- and 16
post-stroke) rated the de-lexicalizedversionofthesamestimuli.ResultsforthefirstpartofthestudyshowthatpitchcontourinBossi’s
pre-stroke speech positively influence the perception of his speech as charismatic, as opposed to those some years after the stroke.
Results for the de-lexicalized speech confirm for French listeners our hypothesis of the influence of the pitch contour in Bossi’s
charisma perception but they are controversial for Italian participants that seem to perceive Bossi as more charismatic in the
post-stroke condition.
Keywords: charisma; political speech; intonation; illocution; voice disorder; speech synthesis.
1. Introduction
Charisma was firstly described by Weber as an
“extraordinaryquality”ofapersonwhoisbelievedtobe
endowed with superhuman properties thanks to which
s/he gets ac- knowledged as a leader Cavalli, 1995: 5).
Though no specific objective description of the
“extraordinary quality” was given in Weber’s studies,
some works started to study the perceivable behaviors of
charismatic leaders: some, e.g., (Boss, 1976), focus on
whatwecalledthe“charismaofthemind”(Signorelloet
al.,2012),thatdwellsinthestrengthofaleader’sideas,
others, e.g., (Atkinson, 1984) try to find visually or
acoustically perceivable aspects of a leader’s behaviors
thatwecalled“charismaofthebody”(Signorelloet al.,
2012). We suggest that both aspects of charisma, either
jointly or independently, are responsible for its conveying
and perception.
In the present study we focus on one aspect of the
charisma of the body: the speech. We assume here that
some of the perceivable acoustic-prosodic characteristics
of a leaders speech are specifically responsible for
conveying charisma. Our general goal is to characterize
acoustically and distinguish perceptually a charismatic
speech from a non- charismatic one.
Within previous work investigating the relationship
be- tween the acoustic-prosodic characteristics of a
political leader’s speech and the perception of his/her
charisma, Rosenberg and Hirschberg, 2009) studied the
correlation between acoustic, prosodic, and
lexico-syntactic characteristics of political speech and the
perception of charisma; Touati (1993) investigated the
prosodic features of rhetoric utterances in French political
speech in pre and post- elections discourses. Other works
examined the relationship between prosodic features
andtheperceptionofaspeakerasa“goodcommunicator”
(Strangert & Gustafson, 2008) or analyzed the pitch
contour of French political leaders’ speech and its
idiosyncratic and contextual variations (Martin, 2009).
2. A hypothesis about charisma
According to Poggi (2005), in persuasive discourse the
speaker tries to convince the audience to do some action by
exploiting the three strategies posited by Aristotle (2011):
Logos (the rational argument), Pathos (the appeal to the au-
diences emotions), and Ethos (the character of the speaker).
According to the theory of Poggi (2005) and Poggi et al.
(2011), the dimension of Ethos also includes, for the
political leader, three sub-dimensions: Benevolence (the
tendency to act in the interest of the audience), Competence
(the capacity for rational foreseeing and planning), and
Dominance (the power to prevail in a competition).
The notion of charisma we proposed in (Signorello
et al., 2012) is based on this theoretical framework. We
de- fined charisma as a set of characteristics of a leader
that include his “having a vision” (a goal towards which
he wants to lead his followers), a “high level of
dominance” (look strong, persistent and fighting) and
“emotional intelligence”(the ability to feel and transmit
emotions, and to be and look empathic). The combination
of these features makes a leader charismatic, and is
displayed by his/her non-communicative and
communicative behavior..
3. What makes a speech charismatic?
To investigate the perception of charisma in political
speech we analyzed the acoustic and prosodic
characteristics in the speech of Umberto Bossi, an Italian
politician who in 2004, during his political career, had a
stroke that resulted in severe speech impairment. We
collected two samples taken from two speeches
performed, respectively, in 1994 (the pre-stroke condition,
PRE) and in 2011 (the post-stroke condition, POST). Our
hypothesis was that the important differences in
Heliana Mello, Massimo Pettorino, Tommaso Raso (edited by), Proceedings of the VIIth GSCP International Conference : Speech and Corpora
ISBN 978-88-6655-351-9 (online) © 2012 Firenze University Press.

344 ROSARIO SIGNORELLO, FRANCESCAD’ERRICO,ISABELLA POGGI, DIDIER DEMOLIN, PAOLO MAIRANO
DIMENSION
PRE
POST
Pathos
passionate, empathetic,
enthusiastic, reassuring
cold, indif-
ferent
Ethos
Benevolence
extraverted, positive,
spontaneous, trustwor-
thy, honest, fair, friendly,
easygoing, makes the
others feel important
untrustworthy
dishonest,
egocentric,
individ-
ualistic,
introverted
Ethos Com-
petence
visionary, organized,
smart, sagacious, cre-
ative, competent, wise,
enterprising, deter-
mined, resolute, who
propose, seductive,
exuberant, sincere, clear,
communicative
inefficient,
inadequate,
uncertain,
faithless,
unclear,
menacing
Ethos Domi-
nance
dynamic, calm, active,
courageous, confident,
vigorous, strong, leader,
authoritarian, captivat-
ing, who persuade, who
convince
apathetic,
timorous,
weak, con-
formist,
unimpor-
tant, who
scare
Emotional
Induction
Effects
charming, attractive,
pleasant, sexy, bewitch-
ing, eloquent, influential
boring
acoustic-prosodic characteristics of Bossi’s speech, in
samples of political speeches preceding and following the
stroke, give rise to a different perception of charisma. If
this hypothesis is validated perceptually we might
conclude that information about charismatic qualities are
borne by the acoustic-prosodic characteristics that differ
in the two samples.
In order to describe the charisma phenomenon
through common language adjectives we conducted a
qualitative study collecting adjectives describing what
charisma is and what it is not (a brief summary is
presented in section 3.1.. For the extensive study see
(Signorello et al., 2012).  We then analysed Bossi’s
acoustic-prosodic features in the PRE and POST and
conducted a language-independent perceptual study on
French participants (section 3.2.4.). We then
de-lexicalized our stimuli by synthesis only preserving
the pitch contour, the duration and the intensity and
conducted a perceptual study on French and Italian
listeners. In isolating the pitch contour we could verify if
this is the aspect that influences the perception of
charismainBossi’sspeech(section3.3.).
3.1 Describing charisma
In a previous work (Signorello et al., 2012) we
constructed a questionnaire aimed to assess the perception
ofcharismainthesamplesofBossi’sspeechrequiredto
previously make up a list of adjectives that express
charismatic and non-charismatic qualities. To find out
such adjectives in an empirically grounded way, we
administered a questionnaire through Internet to 58
French participants (42 female, 16 male, mean age 30),
asking to freely generate adjectives connected to the idea
of what charisma is and what it is not. We obtained a list
of French adjectives, 106 describing charisma positively
and 105 describing what charisma is not. In order to make
a manageable questionnaire, we further selected 67
adjectives (Table 1) retaining only those occurring more
than once, 42 positively and 20 negatively related with
charisma. We then classified those adjectives in a
multidimensional scale of charisma under five
dimensions describing this phenomenon. An extended
report of this multidimensional scale of charisma and on
how adjectives describing charisma are classified in it can
be founded in (Signorello et al., 2012).
3.2 Normal Speech
3.2.1. Stimuli
Previous works about the perception of a speaker as a
good (Strangert & Gustafson, 2008) or charismatic
speaker (Rosenberg & Hirschberg, 2009) rely on the
acoustic analysis and the perceptual evaluation of stimuli
classified per speaker, topic and genre of speech. Our
approach is different. We chose 3 stimuli per condition
(PRE and POST) according to their illocutionary value:
an assertion, an in- citation and a rhetorical wh- question.
As we know the speaker shapes prosody differently in
relation to different speech acts (Firenzuoli, 2001). Our
hypothesis is that all three types of speech acts are
perceived as more charismatic in the PRE condition
thanks to prosodic features. Further we argue that
incitation might be perceived as more charismatic than
rhetorical question which in turn might be perceived as
more charismatic than assertion. Below we describe the
acoustic-prosodic features of our stimuli.
Table 1: The 67 positive and negative adjectives related
with charisma collected amongthena¨ıveFrench
participants (in English for clarity purposes). Reprinted
from Signorello et al., 2012)
3.2.2. Overall F0 measures
The PRE speech presents higher F0 means than the POST
speech: PRE (F0 mean 178.89 Hz; min 101.84 Hz; max
241.10 Hz), POST (F0 mean 120.20 Hz; min 91.78 Hz;
max 155.99 Hz). All means from the PRE differ
significantly from the POST (p<0.0001). Our findings
confirm and extend (Murry, 1978)’s findings on
significant differences in F0 measures between normal
and disordered voice. We argue that F0 values might be
positively correlated to charisma perception.
3.2.3. Pitch contour description
The assertion in the PRE condition (Figure 1a below)
presentsasyntacticfocuson“questo”[this],emphasized
by a high fall and separated by a pause from the rest of the
sentence. The right-side part of the tonal unit presents a
falling contour with a small peak on the last tonic syllable.
Instead, in the POST condition (Figure 2a below) the

CHARISMA PERCEPTION IN POLITICAL SPEECH: A CASE STUDY 345
sentence presents a moderate falling and flat pitch contour
with a peak on the third lexical word. The incitation in the
PRE condition (Figure 1b) includes two parts, each with a
(a)
(b)
(c)
Figure 1: Intonation contour, transcription, translation,
du- ration and F0 measures of PRE stimuli per speech act.
(a):Assertion.“Questo amici ereditiamo”[This,my
friends, is what we inherit]. 3,51s. F0 mean 52.62 Hz; SD
12.40 Hz; min 95.25 Hz; max 210.94 Hz; range 13 ST. (b):
Incitation.“Siritornaall’attacco,fuoridalletrincee”
[Let’stakeupagaintheoffensive,getoutofthetrenches].
4.27s. F0 mean 225.51 Hz; SD 38.58 Hz; min 107.74 Hz;
max 270.36 Hz; range 16 ST. (c): Rhetorical wh- question.
“Ecomefacevamoafarlo?” [How could we have done
it?]. 1.81s. F0 mean 138.28 Hz; SD 27.98 Hz; min 96.07
Hz; max 189.39 Hz; range 11.72 ST. Spectrogram and
pitch contour graphics obtained with WinPitch software
(Martin, 2011)
pitch contour starting with high frequency and falling
sharply in the last tonic syllable. In the POST condition
instead the incitation (Figure 2b) presents two
rising-falling contours in the first part and falls gradually
in the right part of the tonal unit. The rhetorical wh-
question in the PRE condition (Figure 1c) presents two
contiguous pitch contour movements: the rising part
corresponds to the wh- element, the falling part
corresponds to the verb. A gradual falling movement
comes on the right side of the tonal unit. In the POST
statement (Figure 2c) a falling contour corresponds to the
wh- element and a rising contour to the verbal element,
with a gradual falling movement on the right side of the
tonal unit.
3.2.4. Perception experiment
Forty French participants with no knowledge of Italian
rated the stimuli presented in the section above via a
HTML/PHP browser-based interface. Twenty of them
listened to the PRE condition and twenty to the POST
condition stimuli. The test took place in an anechoic
chamber and participants wore a Sennheiser HD 25-13
headphone. After listening to each stimulus a participant
had to answer to some check questions to verify that the
perception of the acoustic signal was optimal and that the
semantic content was not understood. Then they had to
express their judgment about the stimuli through our
67-adjective inventory on a 7-point Likert scale
(a)
(b)
(c)
Figure 2: Intonation contour, transcription, translation,
du- ration and F0 measures of POST stimuli per speech
act. (a):Assertion.“Noi siamo schiavi del centralismo
romano”[WeareslavesoftheRomancentralism].2,46 s,
F0 mean 116.77 Hz, SD 10.74 Hz, min 86.64 Hz, max
146.45 Hz, range 9 ST. (b):Incitation.“La Lega e`
pronto per conquistare la liberta` della padania”[The
Lega is ready to conquer the freedom of padania]. 6.61s,
F0 mean 142.02 Hz, SD 38.58 Hz, min 86.2 Hz, max
182.08 Hz, range 12 ST. (c): Rhetorical wh- question“E
come fanno a lavorare questa gente?”[How can these
people work?]. 1.89 s, F0 mean 117.93 Hz, SD 15.54 Hz,
min 90.56 Hz, max 192.99 Hz, range 13 ST. Spectrogram
and pitch contour graphics obtained with WinPitch
software (Martin, 2011)
(0 = “totally disagree”, 7 = “totally agree”), with some
adjectives from the list substituted by their reverses (e.g.,
warm instead of cold) to avoid answer habituation. The
average duration of the test was of
20 minutes.
3.2.5. Results
From our check questions it resulted that perception was
good and there was no semantic comprehension. Hence,
the differences between PRE and POST, that are mostly
significant (t-test, p<0.05), must be due only to acoustic
and not to semantic features. Out of the 67 adjectives used
to measure the perception of charisma, about 33
adjectives obtained significantly different values (t-test,
p<0.05) between PRE and POST speech, and most of
them were rated higher for the PRE condition (Table 2
below). This is consistent with our hypothesis about the
PRE speech as more charismatic than the POST thanks to
its acoustic features. The PRE speech is positively
correlated with most adjectives describing charismatic
qualities (Table 1 below). In the dimension of Pathos the
speaker is perceived as passionate, eloquent and
enthusiastic in the PRE and as indifferent in the POST. As

346 ROSARIO SIGNORELLO, FRANCESCAD’ERRICO,ISABELLA POGGI, DIDIER DEMOLIN, PAOLO MAIRANO
DIMENSION
PRE
POST
Pathos
passionate (5.02), enthu-
siastic (3.25)
indifferent
(2.83)
Ethos Benev-
olence
egocentric (4.51), dis-
honest (3.95), makes the
others feel important
(3.68), individualistic
(4.29)
trustworthy
(3.51), in-
troverted
(2.41)
Ethos Compe-
tence
competent (4.83), smart
(4.52), organized (4.75),
determined (5.51), exu-
berant (4.57), faithless
(3.57), clear (4.65), com-
municative (4.25), se-
ductive (3.17)
wise
(3.90),
unclear
(3.37)
Ethos Domi-
nance
dynamic (5.13), author-
itarian (5.73), confident
(5.89), leader (5.87),
captivating (3.57),
convincing (4.40),
captivating (4.78)
calm
(4.29)
Emotional
Induction
Effects
attractive (3.10), elo-
quent (4.68), charming
(4.78)
boring
(3.63)
to Ethos Benevolence results are quite inconsistent: the
adjectives attributed to the PRE speech include egocentric,
dishonest and individualistic, which in our previous
qualitative study (Table 1) are non-charismatic qualities.
Table 2: Adjectives describing the perception of charisma
intheBossi’sspeechbyconditionwithratingvalues
(t-test, p<.001)
Table 3: Adjectives describing the perception of charisma
intheBossi’sspeechbyspeechact(A=assertion,
I=incitation, Q=rethorical wh- question) and condition
with rating values and one-wayANOVA’svalues
(p<.001). Higher rates in bold
As for the dimensions of Ethos Competence and
Ethos Dominance our hypothesis is almost completely
validated: the speaker is perceived as competent, smart,
clear, seductive, etc. in the PRE and as unclear in the
POST; as dynamic, authoritarian, confident, leader in the
PRE and as boring in the POST speech. These results
validate our hypothesis on the attribution of charismatic
qualities to the PRE as opposed to the POST speech.
Taking into account the different types of speech act
both in the PRE and in the POST speech the different
illocutionary act elicitates a different perception. The
incitation is the one that influences the most the
perception of charisma. In particular for the dimension of
Ethos Competence the incitation elicitates adjectives as
competent (F(2, 123)=3.114; p<0.048), resolute (F(2,
123)=6.767; p<0.002), enterprising (F(2, 123)=8.515;
p<0.001), clear (F(2, 123)=3.046; p<0.05), exuberant
(F(2, 123)=4.232; p<0.017) and communicative (F(2,
123)=2.705; p<0.05). More than other speech acts the
incitation has a significant effect on the perception of the
speaker’s emotional state (see adjectives as passionate
(F(2, 123)=2.999; p<0.05), influential (F(2, 123)=9.359;
p<0.001) and enthusiastic (F(2, 123)=4.765; p<0.010)).
The assertion on the other hand evokes more
non-charismatic qualities like indifferent (F(2,
123)=3.459; p<0.035) and unclear (F(2, 123)=3.662;
p<0.029). Finally the rhetorical question seems to not
influence a specific dimension of charisma. However, if
we consider effect of both condition and a particular
speech act the results are quite different. Through a
one-way ANOVA we crossed the results of the condition
(PRE vs. POST) and the different types of speech act
(assertion, incitation and rhetorical wh- question) to study
the influence of the different illocutionary acts on the
perception of Bossi’s charisma (see Table 3).  The
incitation makes Bossi to be perceived as more dynamic,
authoritarian, active, menacing, and energic in the PRE
condition and as extraverted and wise in the POST
condition. Through the Assertion he has been perceived
as individualistic in the PRE speech and as calm and
introverted in the POST speech. As for the rhetorical wh-
question the only significantly results is timorous in the
PRE speech.
3.3 Synthesized speech
3.3.1. Stimuli
We decided to carry out a perceptive test on de-lexicalized
stimuli in order to further validate our hypothesis that the
pitch contour is a relevant element for the perception of
charisma. In fact, our de-lexicalization procedure enables
us to isolate the pitch contour of a sentence from the
semantic content, segmental features and voice quality
characteristics. In this way, listeners are therefore forced
to give their judgments solely on the basis of intonation,
all other linguistic information being eliminated. The
de-lexicalized procedure we chose has been developed for
the AMPER (Atlas Multimedia Prosodique de l’Espace
Roman) project developed by Albert Rilliard on the basis
of scripts originally elaborated by Antonio Romano (see
Contini et al., 2002 for details). It consists in synthesizing
a periodic waveform with the original pitch, intensity and
duration values of the actual sentence (this is done by
taking three measures per vowel, respectively at the onset,
PRE
POST
AD J EC T I VE S
A
I
Q
A
I
Q
dynamic
5
5.45
4.42
2.09
3.09
2.14
authoritarian
6.19
6.42
4.57
3.61
4.04
3.61
calm
2.66
1.76
3.42
4.61
4.14
4.09
extraverted
1.66
1.23
2
2.14
3.14
2.9
timorous
3
2.38
3.95
3.76
3.38
3.42
wise
2.85
2.28
3.38
3.95
4.23
3.52
individualistic
4.81
4.61
3.42
3.28
3.33
3.71
active
4.9
5.66
4.28
2.28
2.81
3.52
introverted
1.52
1.14
2.33
2.81
2.19
2.23
menacing
4.57
5.33
3.33
3
2.9
2.85
energic
5.14
6.09
4.52
2
2.9
3.38

CHARISMA PERCEPTION IN POLITICAL SPEECH: A CASE STUDY 347
middle and offset-consonants are replaced with silence).
This procedure has been used by several authors working
on the AMPER project and has already proved its
efficacy.
3.3.2. Perception experiment
Twenty-two French (11 PRE, 11 POST) and thirty-one
Italian (15 PRE, 16 POST) listeners participated to a
perception analysis with the same methodology described
in section 3.2.4. Thus the only differences were the
de-lexicalized stimuli.
3.3.3. Results
The first results for the de-lexicalized stimuli perception,
compared to results for normal speech perceptions,
confirm in one hand our hypothesis of the influence of the
pitch con- tourinBossi’scharismaperceptionforFrench
participants but they are, in the other hand, controversial
for Italian participants. In fact French listeners describe
Bossi as charming, who propose, timorous, confident,
pleasant, introverted in the PRE speech and as inadequate,
spontaneous, active, leader in the POST speech (t-test,
p<0.05). For Italian participants we only performed the
perceptual test of de-lexicalized stimuli in order to avoid
semantic and ideology influence on the perception of
Bossi’sspeech.Italianlistenersperceivedthespeakeras
boring, indifferent and unimportant in the PRE speech
and as attractive, visionary, sexy, cold, passionate,
seductive in the POST speech (t-test, p<0.05). From these
preliminary results it seems that the pitch contour-only
stimuli elicit a different type of of Bossi’s charisma for
Italians listeners. In fact the POST speech is described
with adjectives positively related with charisma and the
PRE speech with adjectives describing charisma
negatively, a trend in results that goes against our theory
of pre-stroke speech as more charismatic than the
post-stroke.
4. Conclusion
In this study we aimed to demonstrate that the perception
of charisma in political speech is partly determined by the
acoustic characteristics of speech. To do so, we first
analyzed samples from the speech of the Italian politician
Umberto Bossi before and after a stroke; through a
qualitative study we singled out 67 adjectives describing
charismatic and non-charismatic qualities. finally we run
a perception study asking participants to rate Bossi’s
samples in terms of those adjectives. As resulted from the
acoustic analysis, the PRE speech, with its intonation
features as focus words, tonal jumps, and higher values,
dramatically differs from the POST. And since the results
of the perception study validate our hypothesis that
Bossi’s speech after the stroke is perceived as less
charismatic than before, we may reasonably conclude that
the characteristics of intonationthatdifferentiateBossi’s
PRE and POST speeches are an important factor in the
perception of charisma. And this hypothesis has been
validated once more through a perceptual experiment in
which we only tested the intonation contour influence on
the Bossi’s charisma perception. We also de-lexicalized
stimuli and preserved original pitch, intensity and
duration values and we tested French and Italian
participants. Results validate our hypothesis on the
intonation contour relevance on charisma perception of
the PRE speech for French participants but are
controversial for Italians. In any case our results on
synthesized speech are preliminary and they will be
statistically analyzed more in depth. Naturally we are
aware that the acoustic characteristics of speech also
include voice quality, which we think is relevant too. In
future work we will investigate the importance of voice
quality in determining the perception of charisma, while
trying to distinguish it from the contribution of intonation,
also through synthesis of speech fragments.
5. Acknowledgements
We would like to thank Christophe Savariaux (GIPSA-
Lab, CNRS, Grenoble, France) for his technical support
in acoustic analysis. Research partly supported by a
Doctoral Fellowship from the French Ministry of
National Education and Research (years 2010-2013) and
by the 7th Framework Program, European Network of
Excellence SSPNet (Social Signal Processing Network),
Grant Agreement N. 231287.
6. Acknowledgements
Aristotle. (2011). Rhetoric. The Internet Classics Archive.
Atkinson, M. (1984). OurMasters’Voices:Language
and Body Language of Politics. Routledge, London,
UK.
Bligh, M.C., Robinson, J.L. (2010). Was
Gandhi ”charismatic”? Exploring the rhetorical
leadership of Mahatma Gandhi. In The Leadership
Quarterly, 21(5), pp. 844--855.
Boss, P. (1976). Essential attributes of charisma. In
Southern Speech Communication Journal, 41(3), pp.
300--313.
Cavalli, L. (1995). Carisma: la qualita` straordinaria del
leader. Laterza, Rome, Italy.
Contini, M., Lai, J.-P., Romano, A., Roullet, S., de Castro
Moutinho, L., Coimbra, R., Bendiha, U. P. and Ruivo,
S.S. (2002).Unprojetd’atlasmultime´diaprosodique
del’espaceroman.InB.Bel,L.Marlien(Eds.),Speech
Prosody 2002, Aix- en-Provence, France, pp. 227--230.
Firenzuoli, V. (2001). Le forme intonative di valore
allocu- tivo dell’italiano parlato. Analisi sperimentale
di un cor- pus di parlato spontaneo (LABLITA). PhD
thesis, Uni- versita` degli Studi di Firenze.
Goleman, D.P. (1995). Emotional Intelligence: Why It
Can Matter More Than IQ for Character, Health and
Lifelong Achievement. Bantam Books, New York,
USA.
Martin, P. (2009). Autres styles: quelques variantes id-
iosyncratiques. In Intonation du Français, Armand
Colin, pp. 228--232.
Martin, P. (2011). WinPitch W7 by Pitch Instruments Inc.
Computer program. Available at:
< http://www.winpitch.com/>.

348 ROSARIO SIGNORELLO, FRANCESCAD’ERRICO,ISABELLA POGGI, DIDIER DEMOLIN, PAOLO MAIRANO
Mohammadi, G., Mortillaro, M., and Vinciarelli, A.
(2010). The voice of personality: Mapping nonverbal
vocal behavior into trait attributions. In International
Workshop on Social Signal Processing, pp. 17--20.
Murry, T. (1978). Speaking fundamental frequency
characteristics associated with voice pathologies. In
Journal of Speech and Hearing Disorders, 43, pp.
374--379.
Poggi, I. (2005). The goals of persuasion. In Pragmatics
& Cognition, 13(2), pp. 297--336.
Poggi,I.,D’Errico,F.,andVincze,L.(2011).Discrediting
moves in political debate. In P. Ricci-Bitti (Ed.),
Proceedings of UMMS 2011, Springer LNCS, pp.
84--99.
Rosenberg, A., Hirschberg, J. (2009). Charisma
perception from text and speech. In Speech
Communication, 51(7), pp. 640--655.
Signorello, R., D’Errico, F., Poggi, I., and Demolin, D.
(2012). How charisma is perceived from speech. A
multidimensional approach. In ASE/IEEE International
Conference on Social Computing 2012, Amsterdam,
The Netherlands.
Strangert, E., Gustafson, J. (2008). What makes a good
speaker? subject ratings, acoustic measurements and
per- ceptual evaluations. In Interspeech 2008, pp.
1688--1691.
Touati, P. (1993). Prosodic aspects of political rhetoric. In
ESCA Workshop on Prosody, Lund, Sweden, pp.
168--171.