Voice and Identity

In my doctoral thesis I have focused on the suprasegmental level of
language in order to find out which combinations of prosodic
features convey not only content information but also information
about our personality, our attitudes and our emotions, that is,
information about our identity. After considering different
definitions of the term “identity”, I come to the conclusion that
“identity” is a relational term that establishes a parallelism
between two domains: the domain of individuality and the domain of
collectiveness. In this regard, theories from cognitive
linguistics, radical constructivism and radical experientialism are
studied and applied to show how our identity is a necessary
construction expressed through different means. Moreover, the
theoretical part of the dissertation bases itself on the notion of
embodiment from the cognitive field and gives great importance to
the role played by perception in the use of our voice. The
theoretical implications are then verified in an empirical part: 60
female voices (20 German, 20 American, 20 Spanish) are recorded in
interviews and analysed phonetically with Praat. After classifying
the voices in terms of nationality and in terms of personality
groups (introversion/extroversion) vocal aspects such as length,
average pitch, intensity and speech rate are measured. Although the
results show nationality differences in pitch and speech rate there
are no concrete tendencies observed for the personality parameters
of introversion/extroversion, even though we do have common
expectations about how an introvert or an extrovert speaks. These
results are also checked statistically. In a second step, all the
voice labels used by the candidates interviewed are collected,
classified according to their frequency and compared with results
given by certain corpora tools. This shows clear vocal stereotypes:
for the three nationalities studied, deep and raspy voices are
judged to be sexy and high and squeaky voices are considered
unpleasant. Phonetic voice labels are preferred to impressionistic
voice labels. Certain voice labels such as sweet or soft seem more
language dependant. A comparison between judgements made on the own
voice (in terms of frequency, intensity and speech rate) and the
results from the phonetic analysis proves that non-experts have an
accurate self perception of voice, which is then confirmed in a
perception test by using high and deep voices as stimuli. This test
proves that independently of whether we understand the language
heard or not, we are capable of guessing correctly basic emotions
and attitudes by interpreting the prosodic cues available. For
certain voice correlations between vocal cues and identity cues a
verticality schema has been found. As a conclusion, by contrasting
theory and practice it can be stated that our voice encodes the
dichotomy of our identity very well (as unique beings and as
members of a collectivity) and that voice production cannot be
understood without voice perception and vice-versa. Both phenomena
are linked to our embodied condition and our cognition.