Information Technology Reference
In-Depth Information
with speci
c patterns of autonomic activation (Cacioppo et al.
2000
). Moreover,
listeners may experience both sad and happy feelings at the same time depending
on the stimulus (Hunter et al.
2008
). It was also proposed that each basic emotion is
associated with a characteristic facial expression. This assumption, however, has
not been con
rmed. Certain facial expressions are often associated with more than
one emotion (smile, for example, is associated with both happiness and pride).
It was argued that the basic emotions of the discrete model are not adequate to
capture the richness of musical emotion (Zentner et al.
2008
). Therefore, a new
model was proposed based on a study that used self-reports of the listeners. The
listeners were asked to list how frequently perceived a group of affective terms
related to music. Principal component analysis of the results showed that affective
responses can be grouped into nine categories. The resulted Geneva Emotion Music
Scale model (GEMS) includes wonder, transcendence, tenderness, nostalgia,
peacefulness, power, joyful activation, tension, and sadness. Zentner et al. (
2008
)
compared GEMS model with the discrete and dimensional emotion models by
asking listeners to rate music-induced emotions in a list containing all emotions.
The results of this comparison showed that the listeners preferred to describe the
emotions induced in terms of GEMS rather than the other two. Moreover, the most
effective discrimination of musical excerpts was obtained using the terms provided
by GEMS. Although it was reported that GEMS model outperformed both discrete
and dimensional models, the results have to be further investigated and tested on
larger collections including various music genres.
6.2.2 Dimensional Emotion Models
The alternative to discrete models is the dimensional approach (Fig.
6.1
). While the
discrete approach focuses on the distinct characteristics that distinguish emotions
from each other, in the dimensional models, emotions are expressed on a plane
along two axes such as valence and arousal. In contrast to basic emotion concept,
dimensional models suggest that an interconnected neurophysiological system is
responsible for all affective states.
The circumplex emotion model (Russel
1980
) proposes that emotions can be
expressed in terms of two dimensions, one related to arousal (activation-deactiva-
tion) and valence (pleasure-displeasure) that are orthogonally situated in the
affective plane. Thus, all emotions can be considered as varying degrees of both
valence and arousal. Although Russell
'
is model has found wide application and is
the dominant model in emotion research other potential variants of two-dimensional
models have been proposed. Thayer (
1989
) proposed a different two-dimensional
model. He suggested that the two affective dimensions are two separate arousal
dimensions: energetic arousal and tension arousal. According to this model, valence
can be explained as varying combination of energetic and tension arousal.
Another variant of the circumplex model is the Tellegen
Watson model (Watson
-
and Tellegen
1985
). This model extends
the two-dimensional models by
