The Definition of a Problem:
Emotion Theory in the Nineties

Leslie J. Morgan
Departments of Women's Studies and Psychology
Simon Fraser University


Despite the recognition of the importance of emotion for human experience evidenced in the writings of both of psychology's founders, Wundt and James, "research and theory about emotions in mainstream psychology is strikingly impoverished" (Waters, 1992). For behaviourism, which ignored both consciousness and cognitivism, emotions fell within that black box of 'mentalisms' it was unnecessary to explore in order to explain (predict and control) human behaviour. More recently, cognitive psychology, with its emphasis on computer-based information processing models, likewise rejects the importance of emotions in favour of rational thought and logical errors in those thought processes. Starting slowly during the 1970's and 1980's, the 1990's has seen a resurgence of emotion theory and research, both within and outside of mainstream psychology. This paper will critically examine some of the major, current theoretical approaches to emotion with a view to articulating the basic concepts of my preferred theoretical framework.

Normally, when evaluating a variety of theories, a list of criteria is developed which may include: comprehensiveness, coherence, parsimony, pragmatic value, predictive validity, and explicitness, among others. In this paper I want to engage primarily with the issue of definition, a method which lends itself well to a consideration of the ontological, epistemological, and ethical underpinnings of theories. The adequate definition of terms forms the basis of both philosophical and scientific inquiry, and so can be used to engage with even those theories which do not take a traditionally scientific stance. One of the primary advantages to this method however, lies in its potential to reveal how the basic assumptions of theorists may be disguised in how terms are operationalized.

Despite White's (1990) assertion that psychologists need not be concerned with what causation is (rather, they are concerned with how people understand and perceive causation), I maintain that psychologists must be aware of the implications of their theories in terms of causality. I argue that one of the essential ingredients of different theorists interpretations of emotion involves their particular stance on causal variables. Following Hume, I assert that ideas about causation are a construction of the human mind (White, 1990), and that these ideas penetrate our definitions of emotions. Thus my investigation into the definitional terms theorists of emotion are using will necessarily include an examination of their claims about causes.

Historical Background

The major theoretical approaches to emotion I will discuss here are the biosocial, the cognitive, and the social constructionist perspectives. This choice, although atypical, appears to be the most representative of current research practices. As always, there is a historical root to this division, one which is well revealed by an appeal to the introductory level texts of psychology. At present, Schacter's two-factor theory (arousal must be labelled before emotion experience proper), the James-Lange theory (the arousal is or signifies the emotion), and the Cannon-Bard theory (that we experience arousal and the labelled emotion concurrently), and the controversies they raise, are often the only theories of emotion presented in introductory psychology texts (Waters, 1992; examples of this are Myers, 1989; Baron & Byrne, 1991). Recently, the Cannon-Bard theory has been largely refuted by evidence favouring the James-Lange theory (Baron & Byrne, 1991), leaving mainstream psychology only the controversy between those theories derived from Schacter's model (cognitive approach) and those derived from the James-Lange (the biosocial). Briefly stated, this controversy is the question of whether cognition must always precede affect, or whether "we experience various emotions because of bodily changes" alone (Baron & Byrne, 1991) .

The Definition of a Problem

In previous incarnations (the dark ages of emotion theory), emotion theory could be seen as subsumed by motivation theories and evolutionary theories of behaviour. The designation 'biosocial theories' used in this paper reflects Izard's (1992) use of the term, who acknowledges their direct lineage from James (Izard, 1990). Biosocial theories would appear now to subsume evolutionary theories of emotion, along with motivational theories which pertained to physiological drives. Cognitive theories of emotion, on the other hand, have a direct connection to Schacter (Nisbett & Ross, 1980), and subsume those motivational theories which pertained to appraisal, preferences, and beliefs. Thus, although emotion theories vary, it seems to make sense to categorize the mainstream theories into these two camps, the biosocial and the cognitive.

Standing in sharp contrast to these theories, is the social constructionist perspective, which asserts that emotions are not merely individual, but culturally diverse and culturally derived. Although this perspective is not reported in the introductory textbooks, it has many proponents (using slightly different foci) from diverse disciplines, including: psychology (Harre, 1986), anthropology (Lutz, 1990), sociology (Denzin, 1984), and women's studies (Burack, 1994). The social constructionist theory of emotion may not yet exist as a comprehensive theory, however, using Harre (1986) as a guide, this position acts as a good counterpoint to mainstream theory.

Definitional Arguments


And so we come to the critical, and age-old, question: What is an emotion? According to Waters (1992), "a good definition of emotion distinguishes emotion from other mental phenomena, such as sensing, judging, believing or thinking" (p.11). Waters further asserts that definitions of emotion should be neither too inclusive nor too exclusive. The definition should "account for all experiences which are normally termed 'feelings' or 'emotions', and should not include experiences that are not ordinarily included in that concept" (Waters, 1992, p.12). This criterion for definitions sounds innocent enough, but as the remainder of this paper will show, and in agreement with Amelie Rorty, it may be that "that emotions cannot be conceptualized as a single, discrete class of phenomena" (Waters, 1992, p.12).

It may be that our attempts to define emotions as discrete phenomena, compounded by problems relating to the observability of 'internal' states, have created the poverty of emotion theory in psychology. Although Waters (1992) insists that definitions of emotion must not violate our everyday understandings of the term, the question remains whether or not we experience emotions as discrete from our beliefs, values, and ideas. How we define the phenomena of emotions has a direct impact on what questions we may then investigate. This is a central paradox of scientific inquiry, reminding us of the searchlight and the street lamp analogies, that we often can only find what we are looking for.

Traditional Arguments

Traditionally, the three most common axes of the definition of emotion, which are derived from James, are: expression/behaviour, physiological arousal, and subjective experience (Izard, 1990; Myers, 1989; Waters, 1992). The splitting of evidence of emotions into physiological arousal, expression/behaviour, and subjective experience allows different theorists to privilege one aspect over another, and this privileging is evident in most emotion theory in one way or another. Although there has been some debate over what the necessary and sufficient components of emotion are (Lazarus, 1991; Clore & Ortony, 1991), most theorists admit evidence from each of these categories, despite their stance on which is more primary.

One of the best illustrations of the controversial nature of definitions of emotion, and of their tendency to capture the basic assumptions of the theorist, may be found in the ongoing debate over the status of 'basic' emotions between emotion theorists Turner, Ortony, Izard, Ekman, and Panksepp (Izard, 1992; Turner & Ortony, 1992). In 1990, Turner and Ortony set out to question the evidence, criteria, and utility of the 'basic' emotion position adopted by many leading emotion theorists. In 1992, Psychological Review [Vol.99(3)] published the responses of Ekman, Izard, and Panksepp along with a further rebuttal from Turner & Ortony. The definitional nature of this debate, essentially between biosocial vs. cognitive definitions, is well illustrated in Izard's reply:

Of particular relevance is the difference pertaining to the definition of emotion experience. Cognitive theories hold that emotion experience includes various cognitive components including the activating appraisals, subsequent desires, and intentions. ... [Biosocial theory] views emotion experience as a feeling state or motivational condition, a direct and immediate product of the particular neural processes associated with that emotion. (Izard, 1992, p.561 - emphasis added)

Izard's framing of this problem emphasizes definitional differences which indicate fundamental assumptions about what causes emotion experience in each of these perspectives. For the cognitive approach, following Schacter, physiological arousal initiates a search for causes which, when found, determines the label we put on the arousal, which then determines the emotion we experience. The biosocial perspective, on the other hand, following the James-Lange model, emphasizes the causal nature of physiological activity, whereby the emotion is the arousal.

The Biosocial Perspective

The way these perspectives define emotion determines what are considered the necessary and sufficient causes of emotion. For the biosocial perspective, facial feedback, the subject of a great deal of ongoing research on emotion, is a necessary component of basic emotions, although there is ongoing debate as to whether it is also a sufficient cause (Izard, 1990; Baron & Byrne, 1991). Objections that facial expression is not a necessary component of emotion, especially the more subtle ones, are answered in two main ways. Firstly, that

micromomentary expressions and nonobservable patterns of muscle action potentials may remain as a functional component . . . [And finally, in an appeal to] . . . a different perspective on the issue of basicness [which] leads to a set of criteria for basic emotions that can be subjected to empirical validation . . . . [which holds that] . . . particular emotions are called basic because they are assumed to have innate neural substrates, innate and universal expressions, and unique feeling states (Izard, 1992).

Studies designed to show that facial feedback is a sufficient cause of emotion (the two approaches to this are muscular [Laird] and vascular [Zajonc]) have not distinguished between particular emotional states (i.e.: basic) but do show positive vs. negative affect differences. This result hardly begins to explain the range and depth of human emotionality, as defined colloquially, nor is it intended to do so within the narrow definition this perspective uses.

Before turning to the cognitive side of this debate, there is an aspect of scientific practice which must be discussed. My portrait of the biosocial, as it will be of the cognitive, is narrowly focused on definitional terms and causal attributions. However, this narrowing of focus does not do justice to the depth and scope of Izard's theory of emotion, which can be considered quite comprehensive in many ways (Waters, 1992). I have instead concentrated on one aspect of the theory, its biological focus, in hopes of making one point clear: Theories might be grand, but the research on them may be narrow, and this relates to causal assumptions and the nature of empirical research in its present incarnation in scientific psychology. Theories are used to generate hypotheses which are testable. Izard states, as quoted above, that his perspective on basicness leads to the development of criterion for basic emotions which can be tested. Despite the complexity of his complete theory, which includes developmental insights and a discussion of the role of emotion in personality, what remains the focus of contemporary research, and what is attributed to this perspective in textbooks, are facial feedback studies!

The Cognitive Perspective

In a similar fashion, despite complex and comprehensive theories of emotion such as those of Lazarus (1991), Arnold (1960), and Mandler (Waters, 1992), experimental psychology tends to focus on how positive vs. negative affect act on such things as memory, perception, task performance, creativity, and attributions (Baron & Byrne, 1991). Again, given the range and depth of human emotionality, studies which reduce emotion to [poorly defined] positive vs. negative affect variables in experiments seem to violate what we mean when we say emotion.

Cognitive models of emotion privilege cognition over physiological arousal, but often use expressiveness or behaviour to measure emotionality. According to Oatly (1989), "emotions happen when certain events affect our goals" (p.34) [cognitive causal inference]. Believing that true emotions are discrete, Oatly asserts that emotions have five salient characteristics: the urge to act, bodily perturbations, conscious feeling, recognizable expressions, and involuntary thoughts. Oatly further argues that: "most theorists now agree that emotion must be understood in relation to action. Emotions involve readiness and involuntary predispositions to act" (p.34). Again, this violates common conceptions of emotion, for it is possible to generate countless examples of colloquial definitions of emotion which are not associated with particular actions, or indeed to urge to act at all. If I am sad, this might colour my thoughts, but it probably does not involve bodily perturbations, nor involuntary predispositions to act.

The Social Constructionist Argument

The search for bodily states or recognizable expressions which coincide with discrete basic emotions characterizes much of mainstream psychological research on emotions. These reductionistic approaches are ridiculed by Harre who asserts:

Psychologists have always had to struggle against a persistent illusion that in such studies as those of the emotions there is something there, the emotion, of which the emotion word is a mere representation. This ontological illusion, that there is an abstract and detachable 'it' upon which research can be directed, probably lies behind the defectiveness of much emotion research. (Harre, 1986, p.4)

Harre (1986) asserts that, when considering definition and causation, that our answer to the question of what an emotion is "is, as likely as not, liable to reflect the unexamined commonsense assumptions of our local culture" (p.4). Harre advocates linguistic analysis as a method of understanding what emotion is, and as a necessary precursor to studies of emotion. He sets out the conditions of word use as a theory of emotion, and argues for the cultural relativity of emotions. In Harre's view, emotions are related to the language games, local moral orders, and social functions, which make sense of both emotion displays and emotion talk in a given culture or sub-culture, and can only be understood in the context of these. Thus Harre claims:

When, and only when, all this hard work has been done, we are likely to profitably engage in the tracking of the physiological details of the various bodily perturbations ... The extent to which perturbations are important and how they are managed varies along dimensions that only the above methodology can reveal. For some cultures emotions are seen as 'located' in social relations, in others as residing in individual bodily reactions. (Harre, 1986, p.13)

The social constructionist position has gained a wide currency in other disciplines for whom physiological studies of emotion entirely sidestep their need to have an emotion theory that speaks of embodied human actors. Thus sociologist Arlie Hothschild (1976) examined models of conscious, rational actors vs. unconscious, emotional actors using the psychological theories available at that time, and asserted that "between these two lies the missing image of the sentient actor, to which I tie the possibility for a sociological study of emotion and feeling" (p.299). Her starting point for this theory is that "feelings take on meaning only in relation to a specific sociohistorical context" (p.288). The social constructionist position is also being used to generate gender specific theories of emotion, a consideration sadly lacking in psychology. Thus, Lutz (1990) asserts that "The cultural construction of women's emotion can thus be viewed not as the repression or suppression of emotion in men ... but the creation of emotion in women" (p.87).

Future Directions

As we have seen, both the definition of emotion and causal inferences about them are rendered problematic in the social constructionist perspective. That the problem of defining emotions is fraught with difficulties is obvious. Mainstream theorists tend to foreclose on definitional questions in the [misguided] belief that operationalizing their terms will protect them from having to deal with such ambiguities, but this is not the solution. Not all questions about emotion will be answered in an appeal to linguistic evidence. Some questions must answered by studying expressive behaviours (as in the case of pre-verbal development of emotions) or by appeals to physiological evidence (perhaps in the case of innate gender differences). However, it remains to be seen if we can find definitions which appeal to evidence other than the linguistic which do not violate our everyday understandings of what emotions are. The problem of many contemporary definitions of emotion is that they write out most of the subtle, complex, and everyday experiences of emotion which, as James insisted (Markus, 1990), are an essential component of our personal selves. That emotions might be better studied as components of the whole self, a self embedded in specific socio-historical relations, highlights the futility of reductionistic studies of human actors, and points to psychological studies of tripartite beings whose qualities of acting, thinking and feeling are necessarily intertwined.

My Current Research Interests

My own research interests centre on the gendered reason/emotion dichotomy, thus the issue of gender and emotion is a central concern for me. The failure of virtually every emotion theorist I am aware of to deal with the issue of gender, even the majority of social constructionists, is in my view highly problematic. Does the reluctance of emotion theorists to treat gender as factor in emotionality imply that these theorists believe that the difference in the emotionality of women and men is more apparent than real? Is it a difference in quantity, and not quality, and therefore not effectively a difference? Given Harre's (1986) hypothesis that there are culturally diverse emotions, might women have different emotions than men? Do physiological brain differences account for differences in emotionality? And would this mean that brain structure causes emotionality? Does Lutz' (1990) thesis that discourses of power create emotion make sense?

I am in agreement with Harre (1986) that it is only after sociocultural investigation into local languages and local moral orders that physiological studies of emotion will make sense. This is because observing physiological 'disturbances' alone can never reveal the experiential quality of emotions. Likewise, studies of behaviours and expressions of emotion can never claim to describe 'internal' experience. However, if Harre's sociocultural analysis results in a position parallel to my reading of his claim that self is a construct, or merely a concept, which is wholly socially determined (Harre, 1987), the question of who it is that experiences emotion arises. The distinction between public and private, and between internal and external, is an issue which confounds psychology. Empirical methods, and even Harre emphasizes empiricism, tend to be reductive. My question is - Is it ethical, in theory, in research, or in practice, to attempt to reduce the complexity of emotional experience so that we might study so-called discrete emotions?

Relying on everyday understandings, however difficult it might be to derive relatively stable definitions of phenomena such as emotion, is a method which takes its justification from the subject themselves. If emotions are private, internal states, and the study of linguistic practices reveals the nature of those states, then it should be included in the methodological practices of psychology. I am in agreement, also, with Izard (1992) that differences in theoretical approaches serve a useful function, as it fosters "a kind of hybrid vigor in research" (p.565), but I am wary of scientific psychology given its tendency to attempt to 'reduce' phenomena into its constituent 'parts', 'basic elements', or operational definitions. Consonant with the methodological hermeneutics approach, I assert that "appropriate methods for the human sciences will diverge from those of the natural sciences" (Woolfolk, et al, 1988). This hermeneutic approach to definition suggests that meanings are always a process of negotiation, essentially dialectical in nature, not indicative of a search for universal truths which correspond with some fixed reality, but instead form a 'dance' of relatedness and meaning dependent on context and vantage point (Woolfolk, et al, 1988). This is not to assert the unbridled relativity of terms and definitions, but rather to emphasize, in this case, the context dependent nature of definitions of emotion.

Thus, my preferred theoretical framework is eclectic, incorporating hermeneutical commitments and social constructionist approaches, but also including an examination of cultural politics, such as Lutz (1975) does. I favour a return to phenomenological methods, where we can learn from each others experience, rather than those that are purportedly objective and value-free, yet reduce subjects to objects while embedding within theory deep ideological commitments. This is what I believe scientific psychology does when it does not attend to gender and other socially defined differences between people. Rather than assuming that emotions are innate, universal, uncontrollable processes, I believe we need to look at what creates differences in emotionality, how reason and emotion interrelate and are interdependent, and how we might begin to deconstruct the public/private divide. In other words, we need to define the problem, and the problem is not the emotions themselves, but the theories of emotion.


Arnold, Magda. 1960. Emotion and Personality. New York: Columbia University Press.

Baron, R.A. & Byrne, D. 1991. Social Psychology: Understanding Human Interaction (6th Ed). Boston: Allyn and Bacon.

Burack, Cynthia. 1994. The Problem of the Passions: Feminism, Psychoanalysis, and Social Theory. New York: New York University Press.

Clore, G.L. & Ortony, 19?? A. "What More is There to Emotion Concepts Than Prototypes?". Journal of Personality and Social Psychology 60(1) 48-50.

Denzin, Norman. 1984. On Understanding Emotion. San Francisco: Jossey-Bass.

Harre, Rom. 1986. The Social Construction of Emotions. Oxford: Basil Blackwell.

Harre, Rom. 1987. "The Social Construction of Selves" In K. Yardley & T. Honess (Eds). Self and Identity: Psychosocial Perspectives, New York: John Wiley & Sons.

Hothschild, Arlie. 1975. "The Sociology of Feelings and Emotions: Selected Possibilities" In M.Millman & R. Kanter (Eds). Another Voice: Feminist Perspectives on Social Life and Social Science, New York: Octagon Books.

Izard, Carroll. 1992. "Basic Emotions, Relations Among Emotions, and Emotion-Cognition Relations". Psychological Review 99(3) 561- 565.

Izard, Carroll. 1990. "The Substrates and Functions of Emotion Feelings: William James and Current Emotion Theory". Personality and Social Psychology Bulletin 16(4) 626-635.

Lazarus, Richard. 1991. "Progress on a Cognitive-Motivational- Relational Theory of Emotion". American Psychologist 46(8) 819-834.

Lutz, Catherine. 1990. "Engendered emotion: gender, power, and the rhetoric of emotional control in American discourse" In C. Lutz & L. Abu-Lughod (Eds). Language and the Politics of Emotion, Cambridge: Cambridge University Press.

Markus, Hazel. 1990. "On Splitting the Universe". Psychological Science 1(3) 181-184.

Myers, David. 1989. Psychology. New York: Worth.

Nisbett, R. & Ross, L. 1980. Human Inference: Strategies and Shortcomings of Social Judgement. Englewood Cliffs: Prentice- Hall.

Oatly, Keith. 1989. "The Importance of Being Emotional". New Scientist 19 (Aug.).

Turner, T. & Ortony, A. 1992. "Basic Emotions: Can Conflicting Criteria Converge?". Psychological Review 99(3) 566-571.

Waters, Janet. 1992. A Comparison and Evaluation of Major Theories of Emotion: Issues and Problems. Unpublished doctoral comprehensive paper, Simon Fraser University, B.C.

Woolfolk, R.L. Sass, L.A. & Messer, S.B. 1988. Introduction to Hermeneutics. In Messer, Sass, & Woolfolk (Eds). Hermeneutics and Psychological Theory: Interpretive perspectives on personality, psychotherapy, and psychopathology, New Brunswick: Rutgers University Press.

White, Peter. 1990. "Ideas About Causality in Philosophy and Psychology". Psychological Bulletin 108(1) 3-18.