There are two distinct approaches in the modeling and classification of hum
an emotions. One is the identification of the fundamental variables or dime
nsions of the affective system, and the other is the recognition of primary
emotions by the analysis of universal facial expressions. In this paper a
system of classification of the human affective system is proposed which in
tegrates the foundations of both approaches using a method derived from the
Munsell color system. The resulting architecture is a topological model ba
sed upon a methodic, constructive and progressively empirical analysis of t
he emotional vocabulary in Spanish. Since the words used to designate parti
cular emotions in natural languages seek to express discrete affective stat
es and effectively serve as communication devices among individuals, we asc
ertain that the identification of such words and their mutual meaning relat
ionships constitutes a valid route to understand the structure of the affec
tive system.
The exercise consists of six consecutive stages: 1) The compilation of a vo
cabulary of words in Spanish which designate particular emotions (n=328); 2
) The grouping of these terms in 28 clusters or "semantic fields" of 6 to 1
7 related ideas (mean = 12); 3) The selection of a representative term for
each cluster or set of kindred words; 4) The ordering of the words within e
ach set in a sequence according to the intensity of the designed emotion; 5
) The identification of pairs of antonymous sets (pleasant and unpleasant)
and, thereby, the establishment of 14 bipolar axis of human emotion (calm-t
ension, certain-uncertainty, compassion-anger, fun-boredom, pleasantness-un
pleasantness, happiness-sadness, pleasure-pain, satisfaction-frustration, d
esire-reject, love-hate, courage-fear, strength-tiredness, enthusiasm-apart
hy, arrogance-humiliation); 6) The elaboration of the final model of the hu
man affective system as an circle or wheel where the 14 axis are located in
reference to an orthogonal two-dimensional system. A vertical dimension of
hedonic value (pleasant above and unpleasant below) and a horizontal dimen
sion of activation (excitation to the right and relaxation to the left) con
stitutes the two reference variables. In this way each one of the selected
emotional terms falls within a specific space.
Even though the model is preliminary and requires further empirical validat
ion it is presented as plausible, particularly concerning the theoretical a
nd practical strategy used to obtain it.
Moreover, the model suggests the existence of specific bipolar and mutually
inhibitory neural systems involved in the expression of each identified ax
is of human emotion.