those which are especially important to us, meaningful to us, concerningmatters in which we have invested our Selves” (Solomon 1976, 188). Else-where he characterizes emotions as “urgent” judgments: “Emotional re-sponses are emergency behavior” (Solomon 1980, 264). An emotion is “anecessarily hasty judgment in response to a difficult situation” (265).

Robert Solomon has suggested that an emotion is a specialkindof judg-ment: emotions are “self-involved and relativelyintenseevaluative judg-ments....Thejudgments and objects that constitute our emotions are

Emotionsare indeed distinguished by the particular wants and interests that are atstake in some situation and by the way that situation is cognitively evalu-ated in light of those wants and interests. And these wants and interestsand evaluations may be specific to some particular culture.

In anyevent I can have only the dimmest of ideas about the sequence of affectiveappraisals and cognitive reappraisals that actually occurred in temporalsequence. Later, however, when I have had a chance to reflect upon myvarious emotional reactions, I can catalog them in recollection. I can de-scribe my state in terms of the emotion concepts familiar to me in English:I was resentful, grieving, depressed, jealous, frightened, and so on. How Icatalog this stream of events in reflection is largely a function of the emo-tion concepts at my disposal in the language and culture that I inhabit.

The psychologist Phoebe Ellsworth (1994) has pointed out that thevarious events that occur in an emotion process are themselves processesthat to some extent unfold independently and run their own temporalcourse. Thus autonomic arousal typically continues even after cognitivemonitoring has changed my appraisal of a situation. One consequence ofthis more complicated picture of emotion is that Solomon is wrong to saythat once my evaluation that I have been wronged (say) changes and Irealize I have not been wronged after all, my emotion of anger will vanish.In fact the physiological changes associated with anger will not—indeed,cannot—vanish as soon as a change in judgment occurs.Ellsworth also stresses that our emotional life occurs in “streams” thatchange all the time in response to ever changing appraisals, ever evolvingactions and action tendencies, ever changing bodily states and feelings. Itis only after the process is over that we can catalog the process “in recol-lection” (as she puts it) as belonging to “one of the familiar emotions cat-alogued by ancient and modern taxonomists” (228, 1994).

Lyons believes thatan emotional response is a physiological response caused by an evaluativejudgment.

As we saw earlier, Solomon claims that emotions are “self-involved andrelativelyintenseevaluative judgments” that are “urgent” and “especiallyimportant to us, meaningful to us, concerning matters in which we haveinvested our Selves” (1976, 188); emotional responses are “emergency be-havior” (1980, 264). The notion of a noncognitive affective appraisal ex-plains very well the features of emotional “judgment” that Solomon indi-cates. An affective appraisal operates very fast or “urgently” in situationswhere the person or other organism has a vital stake in what’s going on;it concerns matters “which are especially important to us.”

We might still wonder, however, why I sometimes respond emotionallyto an event or situation—and make an affective appraisal of it—and some-times not. Why do I sometimes get angry with my small son and sometimesnot, even when he is doing exactly the same thing on both occasions? Thereare a number of reasons, I think. One is that an affective appraisal is prob-ably more likely when the wants and interests that are at stake in an en-counter are intensely important to us. Another is the vividness of the per-ception or mental image that prompts the affective appraisal: we are morelikely to get emotional about refugees when we see them on TV than if weread about them in the newspaper. A third relevant fact is our bodily state,whether we are fatigued or hyped up on caffeine or under the influence ofalcohol. Our current mood will also be a factor. The tired and hungrymother beset with anxiety about other things is much more likely to getupset with her small son. Affective appraisals occur more readily in somebodily and mood states than others. And then there are individual differ-ences among people: some people are more emotionally labile than othersand readily change mood and emotional state.

It is important to remember that in an emotion process there is con-tinuous feedback of various sorts from one event in the process to another.For example, physiological changes may help to fix our attention on what-ever it is that is important to us in the situation. Autonomic arousal mayprompt us to action. Subjective feelings may also reinforce attention and inaddition can serve as a source of information about what I am reacting toand how important it is to me. Cognitive appraisals and reappraisals canmodify action tendencies, physiological changes, and subjective feelingsalike. After responding instinctively by freezing or tensing, I may see thatthe situation is no longer threatening, so I relax, my heart gradually stopspounding, and my feelings gradually become calmer. And action tendenciesinduced by an affective appraisal, such as freezing in fear or tensing inanger, often alter the situation itself—perhaps the enemy retreats—and soindirectly modify my appraisal of the situation and thus in turn affect mysubsequent behavior.

LeDoux’s model of what goes on in a simple episode of emotion sug-gests an answer to our dilemma. In the simple cases he has studied, thereis an affective appraisal that causes autonomic and motor changes and issucceeded by cognitive monitoring. In other words, even a simple episodeof emotion is aprocess, involving a number of different events, and, inparticular, involvingbothaffectiveandcognitive evaluations. On this modelit is the affective appraisal that generates autonomic and motor changesand puts the emotionality into emotion, as James might have said.

Ame ́lie Rorty (1980) has an example of a man who is resentfulof women in authority even though he sincerely believes he isn’t. Rortyexplains the situation as due to his ambiguous feelings about his mother,of whose love he has never been sure. As Rorty tells the story, the manautomatically reacts with resentment to a new female boss before he knowsanything about her. Regardless of the details of the case, what is interestingfor my purposes is that a highly complex cognition—“Here is a woman inauthority!”—results immediately and automatically in an unfavorable af-fective appraisal.

In short, I am suggesting that it is always an affective appraisal thatinitiatesan emotional response, but that this affective appraisal may itselfbe an appraisal of some cognitively complex information (“My boss insultedme”). Furthermore, although it causes physiological changes willy-nilly, itwill give way to further cognitive activity, which will modify my responses.

In the remainder of this chapter I would like to show that the picture ofemotion I have painted is consistent with folk psychological accounts ofemotion in terms of our everyday emotion concepts, so that we do not needtwo entirely different accounts for basic emotions and for more cognitivelycomplex emotions (as Paul Griffiths [1997] has recently argued). My sug-gestion is that there is a set of inbuilt affective appraisal mechanisms, whichin more primitive species and in neonates are automatically attuned toparticular stimuli, but which, as human beings learn and develop, can alsotake as input more complex stimuli, including complex “judgments” orthoughts.

How we categorize our emotions will be largely a matter of the con-cepts available to us in our culture. In Japanese culture one can experiencethe emotion ofamae, a propensity to “depend or presume upon another’slove,” which according to some “social construction” theorists is not anemotion that is experienced—or experienced in the same way—in the West(Morsbach and Tyler 1986, 290).

At the same time we seem to be faced with a dilemma. On the onehand, we have evidence that emotional responses are caused by noncog-nitive affective appraisals, but these affective appraisals are simple appraisalsperhaps conceptualizable as “I don’t like this!” or “This is a threat!” or even“Yuk!” These appraisals cannot make fine discriminations. And on the otherhand many of the emotions for which we have names in our languageseem to be definable in terms of complex cognitions.

Once in the lateral nu-cleus, the signal is dispatched to the central nucleus of the amygdala, whichcontrols the characteristicfear responses: the freezing response, the releaseof stress hormones into the bloodstream, and the various cardiovascularand other visceral responses, including increased blood pressure, sweating,and increased galvanic skin response. If information from the thalamo-amygdala route is coordinated with information from the cortico-amygdalaroute in the lateral nucleus of the amygdala, then presumably it would bepossible for the information derived from the auditory cortex to confirm ordisconfirm the information received directly from the thalamo-amygdalaroute; it could confirm whether the “affective appraisal” is appropriate ornot, and whether the responses generated should be maintained or aborted.Significantly, the auditory thalamus cannot make very fine discrimina-tions in a stimulus. It is in the auditory cortex that fine discriminations aremade. LeDoux concludes that he has discovered two different pathways forprocessing the same sound. On the one hand, there is a “quick and dirtyprocessing system,” which responds very fast, warns the organism thatsomething dangerous may be around without identifying it very carefully,and gets the organism to respond appropriately to whatever it is. And onthe other hand, there is a slower, more discriminating processing system,which operates through the cortex and figures out whether the thalamo-amygdala’s “affective appraisal” is appropriate or not (LeDoux 1989,1996).

The amygdala com-putes the affective significance of the stimulusbeforethe more precise in-formation about what the stimulus is gets received from the auditory cortex.Significantly, for a rat it takes about 12 milliseconds (12 one-thousandthsof a second) for an acoustic stimulus to reach the amygdala through thedirect pathway from the thalamus, whereas it takes almost twice as longfor the cortical route. The thalamo-amygdala and cortico-amygdala routesconverge in the lateral nucleus of the amygdala.

ditory cortex,where the sound is “cognitively processed”—for example, thesound is identified. The auditory cortex then sends signals to the amygdala,where the emotional impact of the buzzer sound is assessed. However—and this is the important point—the auditory thalamus also sends signalsdirectlyto the amygdala,bypassing the cortexaltogether.

According to LeDoux, the key to the fear system in the brain is theamygdala,a small almond-shaped formation where the emotional signifi-cance of threat is registered. LeDoux found that when rats that have beenconditioned to fear the sound of a buzzer hear that sound, theauditorythalamusis activated. He notes that the thalamus has been called a “waystation” in the brain, because it receives incoming stimuli and sends themon to different parts of the cortex that are responsible for “higher” proc-essing in the various different sense modalities (vision, hearing, etc.). Theauditory thalamus receives auditory signals and sends them on to theau-

LeDoux thinks that whatever your theory of emotion, the “core”of an emotion system is “a mechanism for computing the affective signifi-cance of stimuli” (1989, 271).

Emo-tions are not peculiar to humans or “higher” animals; thebasicemotionsystems are to be found in many “lower” species, even insects and fish.These basic emotion systems are designed so that the organism can copewith fundamental life encounters. It is important to think of a basic emo-tion system as adapted through evolution to particular sorts of importantinteraction between the organism and its environment. LeDoux thinks thata list of basic emotions would correspond to a list of “special adaptivebehaviors that are crucial to survival” (1996, 126). His “working hypoth-esis” is that “different classes of emotional behavior represent different kindsof functions that take care of different kinds of problems for the animaland have different brain systems devoted to them” (127).

Joseph LeDoux has argued that there is no unitaryphenomenon called “emotion,” but only a variety of emotion systems.

From an evolutionary perspective, it seems plausible to speculate thatfast, automatic appraisals are adaptive just because they take place morequickly than “higher” cognitive processes and instantly get the organisminto a state of action readiness. It may be very important to us to knowimmediately whether something in the environment is friendly or hostile,is a threat or a poison, and to get ready to respond accordingly. Rememberthat emotional appraisals concern matters that are of significance to meand mine—to our survival and/or our well-being—and that emotions focusour attention on those aspects of the environment that we appraise ascrucial to our survival and well-being.

at least some emotions involve primitive evaluations that occur belowconsciousness and independently of “higher” cognitive processing.

cedes cognition and that cognitive evaluation is therefore not necessary toemotion.

Zajonc himself interprets his results as evidence that affect pre-

3.Among the most compelling evidence that Zajonc cites are someclassic experiments carried out in the early 1950s by Lazarus and McCleary,called the “subception” experiments (Lazarus 1991). In these experimentspeople were presented with ten five-letter nonsense words. The subjects’galvanic skin responses were then conditioned to five out of the ten, usingelectric shock as the unconditioned stimulus. In other words, five out ofthe ten “words” were associated with a shock, and when the subjects sawthose “words” they responded with a heightened galvanic skin response.Such a response is part of the “emergency” response studied by Cannon.All ten stimuli were presented an equal number of times, so that the sub-jects would be equally familiar with all ten. After the subjects had beenconditioned in this way, a test was administered in which the “words” werepresented at speeds too fast for recognition and the subjects were asked tosay which “word” they had “seen.” In the test itself no shocks were ad-ministered, but the subjects were not told this. The remarkable result ofthis experiment was that the galvanic skin response was much higher for“words” associated with shock than for those associated with nonshock,even when the subjects were unable to identifythe words that had been pre-sented to them. Lazarus called this effect “subception,” indicating a kind ofperception that occurs below awareness: the subjects did not know which“word” had appeared to them, but their skin did. We can infer that thesubjectsfearedordislikedcertain “words” even though they did not knowwhat they were.

2.More recently, Zajonc has undertaken a series of experiments de-signed to show that new affective reactions can be induced apparently with-out the intervention of cognition. In these experiments he and his col-leagues have studied what he calls “nonconscious affective priming,” thatis, inducing an affective reaction to a “neutral” stimulus such as a Chineseideograph (presented to someone who does not know Chinese) by “priming”the stimulus with an “affective” picture, such as a picture of a happy oran angry human face. The “prime” is exhibited to the experimental subjectsso briefly that they are unconscious they have seen anything, and yet theChinese ideographs that have been primed by a smiling face are muchpreferred to those primed by an angry face. However, when the ideographsare presented more slowly so that they can be consciously recognized, thispriming effect vanishes (Murphy and Zajonc 1993).

1.In the “mere exposure effect” experiments, Zajonc and others haveconvincingly established that people prefer stimuli to which they have beenexposed more often, even when the stimuli are presented so fast that thesubjects cannot consciously recognize what they are seeing. For example,in one experiment, Zajonc and a colleague flashed slides of irregular poly-gons to a group of subjects, which—it was independently established—appeared too fast for recognition. The subjects were asked to discriminatewhich of two polygons they liked better and which they had seen before.In this study the rate of recognition was virtually equivalent to chance, butthe liking responses reliably distinguished between those polygons that were“old” or (in some sense) “familiar” and those that were “new” or “unfa-miliar” (Kunst-Wilson and Zajonc 1980).

Zajonc claims that at least in some cases, “affect” occurswithout any prior “cognition” or cognitive appraisal.

Nico Frijda has studied what he calls “unlearned stimuli” for emotionsamong several different species. He points out that many species—includinghumans—respond “instinctively” to the strange and weird. For example,dogs and apes are afraid of people dressed or behaving in unfamiliar ways,horses are afraid of flapping plastic bags, and Rhesus monkeys are terrifiedof mechanical moving monsters. Moreover, other fear responses, while notinbuilt or present at birth, are easily acquired with very little learning, suchas human fear of spiders and snakes (Frijda 1986).

The founder of behaviorism, John Watson(1929), performed some notorious experiments on newborn babies. Hefound that restraining the newborn’s ability to move its head was a uni-versal stimulus for rage, that sudden loss of support (dropping the baby!)was a universal stimulus for fear, and that gentle caresses universally elic-ited pleasure (although Watson dubbed this response “love”). Yet newbornbabies do not seem to have the cognitive resources necessary to make cog-nitive evaluations such as “That was an offense!”

Interestingly, Ekman has also found evidence that physiologicalchanges characteristic of a particular emotion can be induced “directly” bymanipulations of the facial musculature. When subjects moved their facialmuscles without knowing what expression they were being asked to con-figure, there was activity in both skin temperature and heart rate distinctiveof the particular emotion (Ekman 1984).

The only really good evidence that there are distinct physiologicalchanges associated with distinct emotions comes from studies of facial ex-pressions across cultures. Paul Ekman has conclusively demonstrated thatfor several emotions or emotion families there are universal facial expres-sions. He has shown that there is remarkable agreement among differentresearchers using different methods about the facial expressions of anger,fear, enjoyment, sadness, and disgust.This evidence is based not just on high agreement across literate andpreliterate cultures in the labeling of what these expressions signal, but alsofrom studies of the actual expression of emotions, both deliberate and spon-taneous, and the association of expression with social interactive contexts(Ekman 1992, 176).

There is some evidence of systematic autonomic differences amongsome of the emotions. Thus anger, fear, and sadness all produce a largerincrease in heart rate than does disgust, and anger produces a larger in-crease in finger temperature than fear (Levenson 1994, 255). But such re-sults are hardly enough to justify the claim that each emotion has auniquely identifying physiological profile.

William James, the father of modern psychology, held that without a con-sciousness of physiological changes, an emotion would reduce to a “coldand neutral state of intellectual perception” (James [1890] 1981, 1067).

But Lyons’s suggestion is useful because it suggests that even ifemotion requires an evaluative judgment, it also requires physiologicalchanges.

The trouble with it is that it does not explain why sometimes anevaluative judgment leads to physiological change and hence emotion,while at other times what appears to be the very same evaluative judgmentfailsto lead to physiological change and emotion

William Lyons (1980) suggests that while an evaluative judgment allby itself is not sufficient for an emotion to occur, if the evaluation causesan “abnormal physiological response,” thisissufficient for emotion to occur.

To see that one can make the appropriate evaluative judgment yet failto be in the corresponding emotional state, it is enough to note that I canjudge that you cut me off in traffic and that this was offensive and insulting,without getting angry: I may be resigned or saddened or even cynicallyamused. Or I may simply judge dispassionately that I have been offendedwithout getting emotional about it at all. Likewise I can judge that I havemistreated my children and that this is bad, without being ashamed. I maybe making an emotionless judgment, or I may experience another emotion:resignation at my bad character or heartless rejoicing in it.

that the judg-ments involved in emotion areevaluativejudgments about a situation interms of one’s ownwants, wishes, values, interests, and goals

If emotions were noth-ing but feelings, argument would be beside the point: normally you wouldnot try to argue somebody out of a pang or a pain.

emotion is a cognitive state:

A change in the belief or judgment seems to entail a change in the emotionand/or the abandonment of the emotion (Solomon 1976, 1980).

emotion is a cognitive state