end p.68
 
4 Social Constructionism and the Contribution of Culture to Emotion
Abstract: Social constructionism about emotions is the view that emotions are socially constructed entities. I defend the view that some emotions are pancultural and inherited capabilities against social constructionism. Social constructionism is shown to lack evidence, and to be based on inaccurate characterizations of scientific views. The affect program theory is able to account for the social variation that social constructionists do identify.
Keywords: affect program theory, emotion, inherited emotions, pancultural emotions, social constructionism
Craig Delancey
In chapter 3 I showed that interpretationism fails as a form of cognitivism about emotions, and I argued that it is inadequate as a theory of mind because it fails to account for some emotional actions and their related emotions. Some emotions point us toward a more naturalistic theory of mind, a result that is unsurprising given our common conception of emotions. But there is a view of emotions in some ways very similar to interpretationism that is not vulnerable to the arguments I raised. This is social constructionism, the view that emotions are (in some sense) created by culture. For the social constructionist, the postfunctional actions could have been pursued because the individuals belong to a culture in which that kind of behavior is what one is expected to do:
The experience of passivity may be treated as a kind of illusion. Emotions are not something which just happen to an individual; rather they are acts which a person performs. In the case of emotion, however, the individual is unwilling or unable to accept responsibility for his actions; the initiation of the response is therefore dissociated from consciousness. (Averill 1974, 182)
On this view, Tim would have been taught that anger requires of him that he shoot the rifle until it is empty, and Eric would have been well socialized to know that the proper expression of fear would have him run farther than necessary from a threat. The explanation of the postfunctional actions, and presumably any other features of emotions that appear to fail to fit a reductive cognitive theory of mind, could be found in the culture and socialization of the agent. If a strong form of social constructionism were true, then the postfunctional emotional actions would not be the expressions of emotions that are natural states, but instead they would be socially constructed ways of behaving. An interpretationist, or one who held one of the many other kinds of irrealism about basic emotions or other mental states, could be a social constructionist about the relevant emotions and thus escape the criticisms I raised. Thus, social constructionism is also potentially a form
end p.69
 
of irrealism about emotions: it offers an alternative to the affect program theory in the form of a theory in which cultural factors are used to explain what is just a construction, a kind of role, of a society.
My task in this chapter is thus to confront social constructionism, and show that it is not, at least for the basic emotions, a viable alternative to the affect program theory or any other naturalistic theory of the emotions. The research and ideas that have gone into social constructionist theory have much to offer, and some of the claims that accompany it are likely true. Furthermore, much of it is wholly consistent with the affect program theory. But if it is to be taken as a theory of what the basic emotions are, and a denial of any naturalistic theory of the basic emotions, it is false.
The Problem of Coherence for Strong Social Constructionism
It is difficult to know if social constructionism is a thesis about emotions or about our understanding of emotions and the social role of emotions. For example, in an influential ethnography, Catherine Lutz claims that "emotional meaning is fundamentally structured by particular cultural systems" (1988b, 5). She describes her method as one of paying close attention to emotion terms, arguing, "The complex meaning of each emotion word is the result of the important role those words play in articulating the full range of a people's cultural values, social relations, and economic circumstances" (6). There is nothing in this that the adherent to the affect program theory cannot embrace: the affect program theory is wholly consistent with the cultural diversity of cognitive causes and the expression of basic emotions, and with the claim that in some cultures some emotions are going to receive a great deal of attention and play important roles while in other cultures these same emotions can be suppressed until they seem almost not to exist. And surely it goes without saying that the emotion concepts, the meanings of emotions, and the social roles these emotions play all require a proper placing of the emotion in its culture. Many naturalists today believe that concepts and meaning are constructed by or depend upon the society in which they play a role; and even those who think that meaning is "in the head" will accept that meanings are transmitted and maintained by cultures, and that they play social roles.
But there is a stronger sense in which social constructionism can be understood. Lutz argues that "emotional experience is not precultural but preeminently cultural" (1988b, 5), and that "emotions are cultural concepts" (1988a, 413). Rom Harré claims that "the bulk of mankind live within systems of thought and feeling that bear little but superficial resemblances to one another" (1986a, 12), and so emotions in one culture are only superficially like emotions in another. And James Averill argues that "most standard emotional reactions transcend any biological imperatives related to self- or species-preservation. They are based instead on human capabilities
end p.70
 
above the animal level and, in particular, on the ability of man to create symbolic systems of thought and behavior (i.e., culture)" (1974, 181). These approaches suggest that emotions are only constructs of cultures, not at all the kind of thing that a naturalist approach that draws upon, for example, the neural sciences and the biological history of a species can ever rightly describe. I shall use the term strong social constructionism for the view that the basic emotions are socially constructed and that there are no pancultural features of any of these emotions of the kind biology or another physical science would properly describe.
It is not clear how many of the social constructionists about emotions are actually strong social constructionists; the position of social constructionism about emotions is usually stated as a negation of theories which are not widely, if at all, held, such as the view that all emotions are just feelings, that all the things that we might call emotions are of the same few innate kinds, that all emotions are constructed out of combinations of a small set of simple innate emotions, or that those emotions which are innate are somehow just simple rigid programs akin to reflexes. Thus, although one might get the impression that many social constructionists intend to claim that no emotions are pancultural, this is not usually explicit (of the researchers listed above, it would seem that only Averill explicitly endorses this). But even if no one were to hold strong social constructionism, the work in this chapter will serve both to clarify the consequences of the ambiguities of social constructionism, and to preserve the affect program theory and related kinds of naturalism about affects from one possible interpretation of social constructionism—an interpretation that is quite strongly suggested by most social constructionists at some time or other.
The Problem of Cross-Cultural Evidence
The most compelling evidence for strong social constructionism is found in ethnographic studies of other cultures where supposedly there are emotions with no ready analogue in our (let us say, in the English-speaking world's) emotions. But there is in these approaches an unexamined problem fundamental to strong social constructionism. This is, quite simply, how does the anthropologist recognize emotions in the other culture?
Consider Lutz's intriguing study Unnatural Emotions: Everyday Sentiments on a Micronesian Atoll and Their Challenge to Western Theory. In this ethnographic study of the Ifaluk people, Lutz analyzes our own and the Ifaluk concepts of emotions, and although she criticizes what she considers a typical scientific view of emotions, Lutz never gives explicit identity criteria for emotions. This exposes the incoherence of strong social constructionism. The problem is nothing less than this: if emotions were entirely socially constructed, and none of the emotions (as we refer to them) were pancultural, then what could it mean to investigate the emotions of other cultures? Why presume they even have emotions?
end p.71
 
Lutz's approach includes working with rough translations into the Ifaluk language of English emotion terms, and then discovering the different eliciting conditions, acceptable forms of expression, and social roles of the analogues among the Ifaluk. But differences in any of these things do not indicate that there is nothing shared for any of the emotions under discussion. In fact, the method is wholly consistent with what one would use if we expected some of the emotions to be, or rely upon, inheritable structures that are amendable via learning, and are used in socially specific ways. Lutz's own accounts are always surprisingly unsurprising:
In each cultural community, there will be one or more "scenes" identified as prototypic or classic or best examples of particular emotions. Thus, on Ifaluk the prototypic scene evoked by the concept of metagu (fear/anxiety) might be the encounter with a spirit, a flight from the encounter, and the recounting of that episode to sympathetic others. (1988b, 211)
This prototypic scene would be quite natural for a contemporary American, if we replace spirit with a growling dog or a man with a gun. A similar emotion had by the Ifaluk is rus, which Lutz translates as panic. Both rus and metagu quite recognizably satisfy our own conception of fear:
The two emotions are also conceptualized [by the Ifaluk] as similar in creating flight or avoidance reactions in those who experience them. People may run away from the dangerous object in each case, but rus is often described as freezing its victim in their tracks or causing them to run about in a confused and crazy way. (186)
Not only is this just what we would expect from fear and panic in our own culture, but it is common to other mammals. A scientist studying fear in rats, for example, expects, and can generate reliably, both behaviors—flight and freezing—again and again, by just the kind of stimuli (e.g., the threat of pain) one would expect!
Lutz is concerned to ensure that a naturalistic view of fear gives proper place to the social roles of fear; but she also tries to argue that fear for the Ifaluk is primarily social because it primarily concerns social relations. This is consistent with the affect program theory. But it is also not established by her own evidence. She grants that the Ifaluk can have "rus (panic/fright) in the face of an approaching typhoon," but unconvincingly suggests that this is not an exception to rus being primarily concerned with social relations by noting that the Ifaluk talk about it: "Emotion is surely also experienced in response to overtly non-social events. . . . In most of the cases, however, it can be argued that the social world plays a significant part" (212). But no one could deny this; anything can be discussed and can play a significant social role.
end p.72
 
The picture that we get from Lutz's study of the Ifaluk is that the Ifaluk have the same basic emotions we have, but that these emotions play different social roles, have different specific eliciting conditions, and have different expressions. We learn that the Ifaluk have emotions like song (justifiable anger), rus (panic/fear), and metagu (fear/anxiety). It is difficult to understand why such results are not taken to be good evidence that some pancultural, and therefore probably inheritable and biologically based, elements underlie a rich cultural diversity. The Ifaluk are surely quite different from us—they admire fearfulness, they believe strongly in active spirits—but they also share much with us.
Other often-cited ethnographies, although always fascinating, fail to establish a lack of, or radical difference in, the basic emotions. Michelle Rosaldo (1980) has studied the Ilongot people of the Philippines. The Ilongot are undoubtedly very strikingly different from us. From our perspective, they seem obsessed with anger and killing. Males are considered immature and unprepared for adulthood if they do not murder someone, preferably an outsider, and decapitate the body. But the very motivation given for this behavior is anger, liget. There is perhaps something special about the importance and role of anger in this society, but it is hardly surprising that anger is seen as a motivator for a murderous attack.
Another advocate of social constructionism is Rom Harré. Harré argues that emotions are properly understood through the "proper understanding of how various emotion vocabularies are used" (1986a, 5). He explicitly attempts to give some identity conditions for the use of emotion terms:
1.  
Many emotion words are called for only if there is some bodily agitation. . . .
2.  
All emotions are intentional—that is, they are "about" something, in a very general sense. . . .
3.  
Finally, the involvement of the local moral order, both in the differentiation of the emotions and in the situationally relative pre- and proscription of the emotions, includes that there is a third set of conditions for the use of emotion words—namely, local systems of rights, obligations, duties and conventions of evaluation. (8)
Unfortunately, Harré also adds that not all conditions must be met; thus, what we have here are just loose guides, since not all of them, nor any one of them, is claimed to be necessary. But, taken independently, none of these criteria is either sufficient or noncircular for even our own English emotion terms. First, we have plenty of bodily agitation (and I shall grant some intuitive notion of "bodily agitation") which are not emotions. Stomachaches, headaches, illness, exhaustion, and many other bodily states would seem to be agitations and yet are not on the usual lists of emotion that a native speaker of English would form. Second, without a theory of intentionality being given, it is difficult to know what is not intentional; but,
end p.73
 
regardless, there are lots of things that are intentional but that we do not call emotions; belief is one of them. Third, many philosophers and some social scientists believe that emotions are necessary for the existence or force of moral codes (for example, some hold that emotions are necessary in order for us to make evaluations). Unless this is denied in some way, then observing that emotions are associated with the moral order threatens to be circular. And, just as with intentionality, there are nonemotional states (as I use the terms) that are also involved in the moral order. These include beliefs about the law.
Harré and Grant Gillett more recently emended Harré's previous list of potential identity criteria. They claim that the " 'rules' for the correct use of an emotion word fall naturally into four groups": a felt bodily disturbance, a characteristic display, a judgment that is expressed, and illocutionary force or an intended result of the emotion word's use (1994, 149). And, though they assume that "these four conditions exhaust the rules for the use of emotion words," these conditions do not, they argue, tell us what the "components" of emotion are: "One can't just say that the obtaining of these four conditions constitutes the having or being of an emotion. emotions are brought into being in the interaction between actual or imagined persons in well-structured episodes and in specific historical conditions" (150). Thus again, the criteria are mere guidelines, and we must wonder what emotions are supposed to be.
The social constructionist James Averill does offer a definition of emotions: "An emotion is a transitory social role (a socially constituted syndrome) that includes an individual's appraisal of the situation and that is interpreted as a passion rather than as an action" (1980a, 312). However, here again, the problem of cross-cultural identification arises. For example, Averill cites a behavior by the Gururumba, a people living in New Guinea, as an example of "an emotionlike syndrome" (1980b, 44). The behavior in question is called by the Gururumba "being a wild pig," and typically occurs in young men. For a short while, they loot, shoot arrows at bystanders, and perform other aggressive acts. This behavior either ends spontaneously, or a kind of re-domestication is undertaken by the tribe.
Averill claims that " 'being a wild pig' and related syndromes are not emotions in the ordinary sense; nevertheless they exhibit many of the features of standard emotional reactions. For example, these behaviors are experienced passively" (1980b, 46). Averill's point here is that there is a behavior which is claimed to be experienced passively, but which from our perspective is a kind of social role and not some kind of necessary (say, from a biological perspective) behavior. However, even if we grant this, it does not establish that emotions are just social roles. That there are pseudo-passive states does not in any way establish that all purportedly passive states are actually pseudo-passive, any more than that some claims are lies establishes that all claims are lies. Furthermore, there are many passive experiences which we don't consider emotions. Suppose that the Gururumba have a disease that they conceptualize in a way similar to how we
end p.74
 
conceptualize disease: namely, as a thing that overcomes its victim. Since they are experienced passively, why are not diseases emotions? In other words, it is again unclear how we are to distinguish emotions from other kinds of behaviors. Averill does add, "Every emotional reaction is a function of a particular kind of appraisal" (1980b, 64), but interestingly he gives different kinds of appraisals that could underlie the "being a wild pig" behavior; these include the appraisal that the social expectations on the individual are too great, or that quarreling with a person who then dies leaves behind a terrible guilt. If appraisal were necessary to make the (seemingly) passive experience an emotion, one would expect that not just any appraisal would do.
The weakness of these many conditions points to the central problem in the use of cross-cultural evidence for strong social constructionism: on what grounds can we identify in another culture a term as an emotion term or a behavior as emotional behavior? This problem even reappears within a culture. It is striking that Harré makes claims about what in our own language is an emotion and what isn't. Here is a sample passage:
But reliance on unexamined common sense can have an unfortunate effect on research methods which the linguistic turn can help to prevent. . . . For example, one well known textbook mentions only depression, anxiety, lust and anger [as paradigm cases of emotions]. Lust and depression are not emotions. Depression is a mood and lust a bodily agitation. (1986b, 5)
How does the linguistic turn settle this issue? How does it entitle Harré to criticize other native speakers of English? We are not told. I grant that many people in our culture would call depression a mood (I would; and my reasons would be that our normal usage pegs depression as an affective state that is longer lasting than, say, anger). Studies have been made of how such terms are used in our own culture, and they reveal much variation, but they also reveal some normal uses (for review see Plutchik 1994, 45-73). But if this is the method to settle such questions, it cannot allow us to criticize our own emotion concepts, nor can it allow for cross-cultural generalizations about emotions.
Strong social constructionism faces a dilemma. Either emotions are just social constructs of our own culture, amounting to nothing more than a tradition like baseball or voting; or the social constructionists need to explain how it is that they identify emotions in other cultures in any sense other than just identifying analogous traditions. The irony is that the cross-cultural evidence is usually taken as the primary evidence for social constructionism. But without some criteria aside from our own use for the emotion terms for identifying any of the emotions, we should well ask what a social constructionist could mean by claiming that other cultures even have emotions. Why do not the categories of Ifaluk emotion, for example, include what we would call illnesses? These are bodily disturbances, and,
end p.75
 
since they can be caused by evil spirits, they may have some important connections to the moral order, satisfying two of Harré's criteria; and they are passive, satisfying one of Averill's criteria. Why is it that instead we find striking evidence of quite similar emotions in these cross-cultural studies?
The one consistent alternative is that we can start with our own emotional categories, find analogues in other cultures, and then observe how those analogues fit or fail to fit with our own. But several things must be noticed about this possible approach. First, it is not clear why the strong social constructionist should then expect to find emotions in other cultures. We don't expect to find significantly similar analogues to bond trading among the Ifaluk, or backgammon tournaments among the Ilongot. Why should we find fear or anger, if these too are just social constructs? Second, this is not what the social constructionists claim to be doing. They freely talk about the emotions of other cultures, and not about the analogues of our emotions. Third, suppose that Averill or Harré and Gillett give some rough identity criteria such that one could actually find analogues of our emotions in other cultures, while denying any form of "essentialism"; but then why are the criteria so surprising? On such a view, the criteria would be the distillations of our own native speaker proficiencies. But not only is this prima facie inconsistent with the criticism of our own cultural concepts of emotions, which are ubiquitous in social constructionist literature; it is also a complete mystery why the criteria are so unlike our folk psychology. That is, how is it that conceptual analysis finds that our emotions are just social constructs, while at the same time the typical view of our own culture, as the social constructionists are eager to point out, sees them as pancultural and inheritable biologically based capabilities? This is at best a very fragile position. It would amount to simultaneously criticizing, amending, and reporting our own views. In order to do this, we need some reason to doubt our own emotion concepts: that is, we need some reason to reject what our immediate conceptual analysis will find that our folk psychology supposes emotions to be (biologically based, pancultural capabilities). More important, we will ultimately need some reason to doubt our scientific account of the relevant affects. The only reason that the social constructionists have offered for taking such a position is to refer to cross-cultural evidence that supposedly establishes that there is little like our own emotions in other cultures, and thereby supposedly casts doubt on the idea that these emotions are anything like what our folk psychology tells us they are. But, as we have seen, the cross-cultural evidence fails to establish this.
The Social Constructionist Critique of Naturalism
The instability of this position is revealed by the claims made in common by Lutz, Harré and Gillett, and Harré that the opposition to social constructionism has been misled by their use of emotion terms. Naturalists are said
end p.76
 
to naively assume that since we have the terms, there must be a kind of object that is the referent to the terms.
Two aspects of the Western approach to language—as something which primarily refers to or even is a series of things—act together to predispose us toward a particular view of the words used to talk about emotion, such as "anger," "fear," "happiness," or "emotion" itself. At best, these words are seen as labels for emotion "things"; at worst, the words become the things themselves rather than human, cultural, and historical inventions for viewing self and relations with others.
The problem of the referential and reified view of language is found in even more extreme form in the domain of emotion words than it is elsewhere in language. This is so because the Western approach to language reinforces the already existing view of emotions as primarily physical things. "Anger," "fear," and "happiness" are treated, through the process of reification, not as concepts used to do certain kinds of things in the world but as labels for concretized psychophysical states or objectivized internal "event-things." (Lutz 1988b, 9)
Similarly, Harré:
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. . . . But in the case of the emotions, what is there is the ordering, selecting and interpreting work upon which our acts of management of fragments of life depend. (1986a, 4)
But it is widely understood that a crude referential theory of language is not adequate, and that naive realism is false. Furthermore, even a cursory review of the history of the sciences of emotion reveals that there have often been debates about what kinds of things emotions are, and even whether they exist (e.g., Skinner in Science and Human Behavior: "The 'emotions' are excellent examples of the fictional causes to which we commonly attribute behavior"). And it is wholly obscure what Lutz means when she says that at worst we treat the emotion words as the things, presumably the emotions, themselves. Surely no naturalist does this.25
And, contrary to Harré's warning against the "illusion" that there is "something there" in an emotion, it is safe to say that there is something there. Harré grants that emotion terms are used foremost when there is "bodily agitation," surely a non-linguistic state, and one which is quite there. This ambivalence is repeated when Harré admits that: "There can be little doubt that, even if there are some universal emotions, the bulk of mankind live within systems of thought and feeling that bear little but
end p.77
 
superficial resemblances to one another" (1986a, 12). Much about this sentence is objectionable. It is not clear how there can be universal emotions, or even any emotions that occur in two different cultures, while at the same time the individuals in these different cultures have only superficial resemblances to one another. Are the striking similarities between rus and panic just a miraculous coincidence? But what is most notable about this claim is that Harré is willing to admit there is nothing inconsistent about the observations both that different cultures have different affect terms, and that some of the emotions (that we identify with our emotion terms) are universal. But of course a naturalist would want to start research right here, at what appears universal (or, at least, pancultural)!
The social constructionists are caught in an awkward position. If all they mean to point out is that there are some states which we may call "emotions," or which are analogous to things we call emotions, but which seem to be culturally specific, and that the roles of these and other "emotions" vary with different cultures, then the social constructionist position is important but it is merely consistent with emotions being socially constructed, and is equally consistent with a naturalist view like the affect program theory. The naturalist who believes that some emotions are pancultural and that a theory of them will have to account for their biological substrates will be happy to grant that culture plays significant roles, and eager to learn from the findings of social constructionists. The important point is that the naturalist generally begins by looking for what is common, in the hope of uncovering deep structures, ideally natural kinds; thus, Harré's admission that there may be universal emotions is where the naturalist will start work. If it turns out that our emotion terms refer to some things which are not natural kinds, nor otherwise stable kinds, this is no disaster; one will amend the science of emotion to take this into account. And if the social constructionist position is a more substantial one—that there are no natural kinds or stable and biologically based kinds of emotions, nor any pancultural identity criteria for them—then the social constructionists are contradicting themselves, or at least are making claims with no clear meaning, when they freely make claims about different emotions in other cultures which supposedly reveal that there is no easy analogue in our own.
Finally, it is worth nothing that strong social constructionism that draws upon cross-cultural study is victim to the very "essentialism" it claims to oppose. For the very idea that there is "emotion" in all these different cultures is one highly open to doubt. Our best scientific evidence points toward, on the one hand, there being a host of pancultural capabilities which in our culture are called "emotions"—fear, anger, disgust, sadness, joy, and others—but, on the other hand, there is unlikely to be any interesting theory that finds significant shared features of all of these capabilities and the many such other states that we group under the term "emotion." Thus, our best understanding is that "emotion" is a useful term for a family resemblance of things; and that our best theories of emotions will be just that:
end p.78
 
different theories of different affects. Yet, these social constructionists who seek cross-cultural evidence always talk about looking for instances of "emotion." And perhaps this is to be expected, since after all, to be looking for fear, or anger, in all these different cultures would be to admit the possibility that these are pancultural. Thus, instead, the social constructionist is looking for "emotion," presumably because this is seen as an appropriately general category to allow for their relativist stance.
Strong social constructionism based upon cross-cultural observations is therefore not a coherent position.
The Problem of Scientific Evidence
I began by suggesting that social constructionism, because it is immune to the criticisms raised in chapter 3 to irrealism, might itself be part of a coherent form of irrealism. I have just shown that strong social constructionism that draws upon cross-cultural study can be shown to be incoherent on conceptual grounds. But we should also remember that all of the evidence for significant biological determination of some emotional capabilities—the pancultural facial expressions, emotions in other species, the neuropsychological evidence that emotions are potentially independent of cognition, the neuroanatomical evidence that separates some emotion and cognitive structures, and so on—is incompatible with strong social constructionism.
Thus, the strong social constructionists are in an uncomfortable position as regards not only the status and import of cross-cultural evidence, but also as regards the neuroscientific and biological evidence. How are they to explain the many precognitive aspects of affect that we saw in chapters 1 and 2 if they claim all emotions are socially constructed? If smiling is just a socially learned behavior that is part of our socially constructed emotion of joy, why is there an independent neural pathway for facial control that allows for spontaneous smiling in the hemiplegiac? Of course, it could be that this separate track is also socially trained to a significant degree, but then why do Irenäus Eibl-Eiblesfeldt's blind, deaf, and brain-damaged subjects spontaneously smile and laugh when playing, or cry and shout when placed in unfamiliar situations? And unless we are wrong to believe that nonhuman animals can show emotions like fear and anger, then it would seem that there is something to fear and anger which is not socially constructed in the relevant sense, since it is shared by organisms that are not only outside our culture, but outside all culture: they do not in the relevant sense have a culture.
Averill has taken up the challenge presented by some of this scientific evidence, and tried to respond to it. He considers four kinds of evidence, concerning biological foundations, physiological correlates, localization in the central nervous system, and cognitive role. Averill's target in reviewing these bodies of evidence is "the association of emotional with physiological
end p.79
 
processes on the basic of extrinsic symbolic relationships" (1974, 151). For Averill, extrinsic symbolic relationships are those relationships which (in this context) are not supported by the scientific evidence (hence they are "extrinsic"). They are instead just prejudices, carried along uncritically.
Averill's objection to the view that there are biological foundations to emotions is to claim that humans are the most emotional of animals. I do not know what this means, but let us grant the claim. It does not establish that our emotional capabilities are therefore not biologically based products of the evolutionary past of humans. It is a parody of evolutionary theory to suppose it means that no species can have unique capabilities, or have capabilities exercised to some greater degree. Furthermore, for those emotions that he grants we do share with nonhuman animals, Averill conflates having different cognitive contents with the existence of taxonomic distinctions of affects: "no animal has as many [fears] as man, not only of concrete, earthly dangers, but also of a whole pantheon of spirits and imaginary evils as well" (1974, 175). But surely this only establishes that we can have different objects of fear, not that we have many different kinds of fearlike affects. We can explain many of the added features of basic emotions that we share with nonhuman animals by properly accounting for the cognitive contribution to the emotion.
Averill's arguments against the importance of physiological correlates of emotions also fail. First, he argues that in research on emotions, there has been a focus on affects, such as fear and anger, that are correlated with "vigorous muscular exertion." Thus, we should not be surprised that these lead to physiological changes. But note that this response has force only if we assume that emotions are somehow normally disconnected from action. For example, on the affect program theory, with the added supposition that some emotions carry action programs as part of their syndrome, for these emotions it is part of what they are that they are tightly connected to "vigorous muscular exertion." That is the very point of affect program theory in this regard. Furthermore, Averill argues that many nonemotional cognitive states lead to physiological changes, and so emotions are not special in this regard. But, even if it were true that some nonaffective cognitions exhibit some, or even all, of the physiological changes of the kind in dispute here, it would establish only that emotions are not alone in having physiological correlates, a result that neither supports social constructionism nor refutes a naturalist theory like the affect program theory.
Third, Averill denies that we can identify emotion systems in the brain, and so achieve a neuroanatomical separation between emotions and cognitions. But his argument is to suppose that lesion studies, in which we infer systematic roles based on deficits that arise from brain damage, cannot distinguish necessary from sufficient structures. He suggests that the neuroscientist acts like a person who removes a resistor from a radio, finds the volume decreases, and supposes that the resistor is an amplifier. This does not describe contemporary neuroscience. Researchers carry out extensive studies, not only of lesioned brains, but of normal brains, using MRIs, CAT
end p.80
 
scans, and electrical probes; of the biochemistry of brains, using a multitude of techniques; and of the effects of direct electrical stimulation of various parts of the brain. Most important, hypotheses about a systematic role are made as part of a theory, and must answer as such, so that in contemporary neuroscience a very large body of evidence is pieced together with the goal of coherent and interrelated theories. Averill argues that "the workings of any of [the brain's] parts can only be understood in relationship to other parts" (1974, 178). If this just means that the brain is a system with interrelated parts, then all properly done systems-level neuroscience takes this into account since any theory is going to be related to other theories of brain function in related structures (e.g., if I argue that fear is generated in the thalamus, I must explain the results about the role of the amygdala in fear and so respond to theories about amygdala function, and so on). If it means that no brain system can have a function that can be described alone, then it is false; we can describe a brain system independently when we specify the role independently (that is, we can discuss the role of amygdala in some clearly defined aspect of fear conditioning because we may be specifying the responses that this nucleus has to certain inputs; we don't need to also explain in this context how the organism is using that fear to act, and so explain motion control and so on). Thus, neuroscience, and specifically neuropsychology, is nothing like the approach that Averill critiques.
Fourth, Averill's claims that emotions essentially have cognitive contents construes cognitivism so weakly that it can be shared by nonhuman animals. For example, he endorses the view that some emotions are object directed. But this is shared by other species of animals also: a cat is afraid of a dog, or angry at another cat. Furthermore, he claims that emotions necessarily have an appraisal element, and therefore are essentially cognitive. The notion of appraisal is not clear here; but, more important, I will describe in a coda to chapter 8 how the affect program theory is actually able to explain what some kinds of appraisal are, whereas cognitive theories of emotion merely take appraisal as a mysterious primitive. And in this regard, nonhuman animals also appraise with their emotions: a frightened cat has appraised a stimulus as dangerous. Finally, Averill argues that reason and affect cannot be separated; but the dependence of reason upon affect is wholly consistent with a naturalist theory of emotion and the affect program theory.
We can see from these responses that social constructionism is often based on the intuition that humans are qualitatively different from other animals because of their culture and increased cognitive abilities—especially the ability to use language. But unless one rejects evolution, this can only mean that we have additional capabilities to other kinds of animals. It does not remove humans from the evolutionary chain of life. Some emotions could be basic, and any abilities special to humans could allow for additional mental states that we group together with the basic emotions to form variations that are not pancultural or, in turn, that we use to create
end p.81
 
emotionlike syndromes which are culturally specific (more on these possibilities below). And this is what we should expect given the neuropsychological evidence. Recall that much evidence is consistent with the view that there is a hierarchy of systems, and as a result there are instances of basic emotions that are distinguishable based on the degree to which they participate in "higher" cognitive functions. Cognitive instances of basic emotions allow for a biological capability to have complex and variable eliciting conditions. Averill and the other social constructionists betray a kind of inversion of the prejudice they attack. Averill argues that we have a long history of associating emotions with "animality"; but what he accepts is the idea that somehow animality is "lower" (in an evaluative sense of the term), and that nonhuman animals are subject to "biological imperatives." This is the true extrinsic symbolism in play, and it is just another version of the cognitive autonomy fallacy (see chapter 13).
Ultimately, strong social constructionism construes naturalism too simplistically. The naturalist is not in the position of seeking some simplistic mechanism that comes fully packaged at birth—because the naturalist knows, from the empirical evidence, that affects are just far too complex for this. Rather, the naturalist seeks the grammar of affects, including of basic emotions. These underlying structures of individual affects are expected to be capable of yielding immense varieties of experiences and behaviors. To reject naturalism about emotions with the erroneous claim that it treats them as simplistic universals is tantamount to rejecting the view that the capability for language and some language structures are inheritable by observing that there are different languages.
Irrealism, One Last Time
Averill holds that social constructionism is the view that "there is no invariant core to emotional behavior which remains untouched by sociocultural influences. The latter view (that there is an invariant core) is essentially a reification of emotion into a biological given" (1980b: 57). As we saw with most social constructionist pronouncements, this can be read in at least two ways. Suppose we identify the neural circuits that underlie fear conditioning and more complex fear behaviors. If Averill's position is that these neural structures are not a "biological core" of fear because there is nothing of interest to say about fear that is not wholly determined by society, then any naturalist who has reviewed the scientific evidence must reject this position. But if we interpret Averill's claim to be that this "biological core" is not unaltered by learning, or that it does not under all circumstances act in a way independent of cognitive capabilities which have been trained in a socially specific way, and which normally will result in socially significant behavior, then not only can the naturalist embrace this position, but she can offer neuropsychological evidence in support of it. This view is not only consistent with the naturalist program—it is actually revealed by the scientific evidence.
end p.82