Attitudinal
meaning: prosodic and visual aspects (João
Antônio de Moraes – UFRJ/CNPq)
The
concept of prosodic attitude generally refers to the expression of social
affects, voluntarily controlled by the speaker (Fónagy 2000). Their acoustic
manifestation is linked to the culture and the language of the speaker, and
differs on that point from basic emotional expressions, which may be seen as
more spontaneous and universal (Ekman
& Davidson 1994, Scherer & Bänziger 2002). However, inside the set of
attitudinal expressions, two cognitively different categories of attitudes should
be set apart: (i) propositional attitudes, which interfere with the propositional content of the
sentence being presented to the interlocutor (e.g. with irony, incredulity, obviousness
etc.) and (ii) social attitudes, which refer to the social interpersonal
relationship established by a speaker with his/her interlocutor (e.g. with
politeness, arrogance etc.) (Moraes et al. 2010).
Our
goal is to investigate whether the difference between propositional (PA) and
social attitudes (SA) is reflected in their manifestation, both on the
production and perception levels.
Six
social attitudes (arrogance, authority, contempt, irritation, politeness and
seduction) along with nine propositional attitudes, of which 5 assertive (doubt,
irony, incredulity, obviousness, surprise) and 4 interrogative (incredulity,
rhetoricity, confirmation and surprise ) were analyzed from a production and a
perception perspective, based on
audio-visual performances of these attitudes by two native speakers of
BP as well as on their recognition by Brazilian
listeners. Each attitude was performed on the same semantically neutral 6
syllable long Portuguese sentence Roberta
dançava (Roberta was dancing).
A
recognition test was set up in order to validate the speakers’ performances.
Attitudes, grouped in both categories (social or propositional) and separated
by modality as well (assertive and interrogative), were presented to 30 subjects
in three possible conditions: audio-only,
visual-only and audio-visual. Subjects had to recognize the presented attitude
during a forced-choice paradigm, amongst a list of 5 to 7 possible answers,
which included all the attitudes of the given category, plus the neutral
expression.
The
results show that propositional and social attitudes presented different perceptual and acoustic behaviors. As the
former are directly linked to the linguistic content of the utterance, they
were expected to be more strongly related to acoustic variations – and the
results support this expectation: in the perceptual tests, subjects rely
clearly on both audio and visual cues for the identification of propositional
expressions, while they mainly use visual cues for socially motivated attitudes.
However, audio cues have an important role in perception, if not a primary one:
they are used to disambiguate some visual expressions, as well as to construct
the detailed meaning of each expressivity.
From
a production point of view, the acoustic analysis reveals that the melodic
contours of utterances, conveying social attitude, show
narrower and global F0 variations, which may explain the less important role
played by audio presentation in expressing these
attitudes. The propositional attitudes, on the
other hand, present more important local modifications in their F0 contour, which
in fact changes their basic melodic configurations and are likely to be
represented by a phonological system such as the ToBI, proposed by autosegmental
phonology (Moraes 2008).
References:
Ekman
, P. and Davidson, R. (eds.) The Nature
of Emotion: fundamental questions. Oxford: Oxford University
Press, 1994.
Fónagy,
I. Languages within Language. Amsterdam: John Benjamins,
2000.
Moraes,
J. The Pitch Accents in Brazilian Portuguese: analysis by synthesis”, In:
Barbosa, P., Madureira, S. and Reis, C. (eds.) Proceedings of the Speech Prosody 2008.
Moraes,
J., Rilliard, A. Mota, B. e Shochi, T. “Multimodal perception and production of
attitudinal meaning in Brazilian Portuguese”. Proceedings Speech Prosody 2010.
Scherer,
K. and Bänziger, T. Emotional expression in prosody: a review and an agenda for
future research. Proceedings Speech
Prosody 2002.