290 Emotion, Affect, and Mood in Social Judgments
for any word in any language. She developed a contemporary
version of an ancient philosophical dream: the creation of a
universal language based on fundamental concepts indis-
pensable for thought. Wierzbicka developed a list of univer-
sal semantic primitives (I, you, someone, something, know,
good, bad, maybe, feel, etc.). These, together with a mini-
grammar specifying the rules of their combination, constitute
a universal language. This universal language can then be
used to analyze any emotion word in any language. The in-
teresting result of her analyses so far is that emotion words
(anger, liget, ningaq) and emotionitself have all been found
to be culture-specific but, nevertheless, definable in terms of
her universal semantic primitives, especially feel, good,and
bad. Even words with a similar etymology, such as emotion,
the Italian emozione,and the Spanish emoción,have not been
found to be equivalent.
Psychologists have also offered formal analyses of the
emotion lexicon, although they are limited to the English lan-
guage. Oatley and Johnson-Laird (1990) defined emotionas a
disjunctive set of five semantic primitives: happy, sad, fear,
anger, and disgust. Any emotion term is then defined by ref-
erence to one or more of these five. This approach clearly
also has ties to the position of ontological realism. Ortony,
Clore, and Collins (1988), in contrast, define all English emo-
tion terms as referring to a valenced reaction to an event. Dif-
ferences between terms are defined by cognitive differences
in interpreting that event (along the lines of appraisal theo-
ries). Their approach has ties to a conceptualist perspective.
An alternative formal analysis began with Wittgenstein
and entered psychology largely through the work of Rosch
(e.g., Rosch, 1978, 1987). This analysis is skeptical of the
classical search for necessary and sufficient features to define
such everyday words as emotionoranger. Various nonclassi-
cal alternatives have been proposed (e.g., Fehr & Russell,
1984), but all share the idea that membership in the category
labeled by a word is a matter of degree and that the border be-
tween members and nonmembers is fuzzy.
TOWARD INTEGRATION
No one research program has been able to achieve consensus.
The persistence of competing and possibly incommensurate
programs is frustrating, but at the same time fascinating and
potentially useful. Differences force us to question assump-
tions and to notice ignored questions. Competing approaches
thus create the grounds for a qualitative shift in our under-
standing of emotion. This shift might take the form of an in-
tegration of two or more of the various paradigms or even of
a revolutionary change in our understanding of emotion. In
this chapter we offer neither a revolutionary theory nor even
a complete integration of available paradigms, but we do
offer the beginnings of one possible integration. We describe
a new descriptive framework deliberately built on all of these
paradigms.
So far, we have perhaps overemphasized the limitations of
these various paradigms. Their longevity indicates that each
addresses some aspect of the topic. All of them have made sub-
stantial contributions to the understanding of emotion. Indeed,
all of them are necessary for raising—if not answering—
essential questions about those very important events that
are labeledemotion.
We also believe that all can be integrated within a com-
mon framework. What has prevented integration in the past is
the assumption that each of these research programs is deal-
ing with the same thing, namely, emotion. If the word emo-
tiondenoted a homogeneous, well-defined set of events, then
different theories of emotion would, indeed, be in direct con-
flict with each other over the same territory. Scientific analy-
sis would long ago have settled major disputes. If, instead,
emotion is a heterogeneous, poorly defined mix of qualita-
tively different events (originally grouped together by our
hunter-gatherer ancestors and modified with each era to suit
cultural concerns), then different theories could be about dif-
ferent topics within that loose domain. Selected evidence
could easily find support for each such theory.
We therefore begin by abandoning emotionas a scientific
term. It remains here only as an everyday term and as a fig-
urehead, a convenient symbol for the general domain of
study, but it is not allowed to set the boundary for the set of
events that any theory in this domain must explain. In fact,
our proposed integrative framework extends beyond the tra-
ditional boundaries of emotion by including such states as
fatigue, drowsiness, and calm. It is especially important to
underscore that abandoning emotionas a scientific term does
not mean abandoning the study of those very real and very
important events now called emotion.
Abandoningemotionas a scientific term allows us to bor-
row from each of the established research programs on emo-
tion (as diagrammed in Figure 12.1). Programs based on an
ontological realist position embody the traditional scientific
search for basic entities that underlie all the varied manifest
differences in a domain. They rightly emphasize empirical
examination of physiological and behavioral details. Pro-
grams based on a nominalist position emphasize the unique-
ness and complexity of each emotion event and of emotional
experience. They also emphasize the role of meaning systems
shared by members of a culture. Unique events are under-
stood (both by a scientist and a nonscientist) through the
mediation of concepts (which are mental processes that