Co-speech Gesture Generation for Embodied Agents and its Effects on User Evaluation - Coverbal Synchrony in Human-Machine Interaction - page 225

the speaker on the right-hand side depicts the window by drawing
its shape in the air.
So the same stimulus is depicted in quite different ways, showing
that there is no one-to-one mapping between a gesture and the object
it depicts. McNeill and Duncan (2000) termed this phenomenon
'idiosyncrasy' implying that gestures are not held to standards of good
form, but are rather created locally by speakers while speaking. This
view receives support by Gullberg (1998, p. 51) who stated: “One of
the most salient aspects of gesture is that people differ in their use
of it”. This becomes particularly obvious in gesture frequency: while
some speakers rarely move their hands at all, other use gestures all the
time (Krauss et al., 1996; Jacobs and Garnham, 2007). Marsh and Watt
(1998) found that individuals vary in their preferences for particular
gestural techniques of representation. Also, regarding handedness in
the use of gestures, a right- or left-hand preference was reported, and
there exist speaker-specific handshape preferences (Bergmann, 2012).
Widely unexplained is, however, the question of how these differences
among individuals arise. Although a number of correlates are possible,
such as cultural background, personality traits, or cognitive skills,
only very few studies have addressed this question so far. Regarding
gesture frequency, Hostetter and Alibali (2007) found that individual
differences in gesture rate are associated with the speakers' verbal and
spatial skills. Individuals with low verbal fluency and high spatial
visualization skill were found to gesture the most.
The fact that people are able to recover and interpret the meaning
of iconic gestures, as shown by Cassell and Thórisson (1999), suggests
that there is at least some systematicity in the way speakers encode
meaning in gestures. The examples above clearly indicate that there is a
certain degree of iconicity in the gestures since the circular shape of the
window becomes apparent in all of them. There are, however, differences
in how the physical form of the gestures corresponds to the object they
depict: in the gestures of the first two speakers, the hand(s) adopt the
shape of the window to be depicted, that is, there is a resemblance
Figure 1. Gestures from three different speakers all of which depicting the same object, a
round church window.