he bese we wae bee ee a ae oh ee mal of Rea esl aby Facial Signs of Er Paul Ekman and Wallace V. Friesen University of California, San Francisco motional Experience Sonia Ancoli University of California, San Diego facial expressions were (ound to provide accurate information about erence than just the pleasant versus us Pleasant distinction. Videotape records were gathered while subjects viewed ‘motion picture films and then reporte od on their subjective expecience. A new technique for measuring facial movement isolated a particular type of smile that was related to differences in report this action and those who did not, to Of two happy experiences was reported facial actions hypothesized to be signs ted happiness between those who showed the intensity of happiness, and to which ‘38 happiest. Those who showed a set of ff various negative afects reported x- Detiencing more negative emotion than those who did not show these actions How much these facial actions were sh of negative affect. Specific facial actor gust were identified, Current emotion theorists disagree about whether different emotions are characterized by distinctive bodily cesponse system changes (eg, autonomic nervous system [ANS] or facial expression patterning). Those who fol- low Schachter and Singer (1962) claim that cognitive expectations are the only important determinants of which emotion is subjectively experienced. ANS activity isnot patterned but rellects the extent, not the type, of emotion that is aroused. This viewpoint’ has largely ignored the possibility that facial expressions might be a differentiated response system distinguishing among emotions. Those who disagree with Schachter and Singer do not share the same theory, but each has empha- sized response system’ changes that are dis- tinctive for each emotion: the ANS by Lazarus (1966) and facial expressions by Plutchik (1962) and Tomkins (1962, 1963) ‘his reseteh was supported in pare by grants feom the Harry H. Cugsenieim Foundation and the Nalonal Inttote of Mental Health (MH 11976) ‘We thank Linda Camras for scoring some of 8 videotapes forthe realty study Requets for eepints should be sat to Paul Ek- sman, 401 Parnassus Avenue, San Francia, Call Tomia 9815, own was related «0 the reported intenslty ns associated with the experience of dis If facial expressions are distinctive for ‘each emotion, this would have important con- sequences both intrapersonally, coloring the subjective experience, and interpersonally, signaling to others how one feels. Although both roles have been recognized, those theoriz- ing about the role of the face in the experience of emotion have emphasized intrapersonal functions. For example, Tomkins (1962) defined the subjective experience of emotion as the feedback from the facial muscular changes. Recent experiments have investi- sated how a subject’s performance of differ- fent muscular movements influences his or het subjective experience of emotion (Tzard, 1977: Laird, 1974: Tourangeau & Ellsworth, 1979}. However, the face may also influence 4 person's emotional experience by providing signals to others about how the person feels IE B perceives As facial expression of emo- tion, B’s behavior toward A may change, and A's notice of this may iniluence or determine As experience of emotion, The viability of proposals that facial ex- pression plays important, perhaps multiple, roles in the experience of emotion depends on the capability of the face to (a) show dis- tinetive expressions for each of a number of emotions. such as fear, anger, disgust, bappi- ness, and so forth, and (b) vary with the as1126 PEKMAN, W. FRIES! felt differences in the intensity of emotion. Such evidence does not exist for spontane- ously occurring emotion. The data only show that the face can provide information about the much simpler, grosser distinction between whether an emotion is pleasant or unpleasant, Much of the current renascence of interest in the face was generated by cross-cultural studies that found universality. Most of these experiments, however, did not examine spon- taneous behavior. Instead people in different ‘cultures were asked to label contrived expres- sions (Ekman & Friesen, 1971; Ekman, Sorenson, & Friesen, 1969; Tard, 1971). The fone study that examined’ spontaneous facial expression in two eultures (Ekman, 1972) focused only on the gross distinction between pleasant and unpleasant emotions. Since then ‘a number of studies within a single culture hhave documented but not extended the find- ing that the face ean show whether an emo- tion is pleasant or unpleasant. (See review of such studies from 1970 to 1979 in Ekman & Oster, 1979, and of such studies from 1914 to 1970 in Ekman, Friesen, & Ellsworth, 1972, chapters 15 and 16.) ‘The primary purpose of the present experi- ment was to venture beyond the pleasant versus unpleasant dichotomy to determine Whether spontaneous facial expressions can provide information about more specific as- pects of emotional experience: Does facial behavior vary with the felt intensity of a pleasant or of an unpleasant emotion? Are subtle differences between pleasant experi- fences shown on the face? Is there one type of smile more than other smiles that relates to a pleasant experience? Can a particular nega- tive emotion be read from the face, not just ‘the more general information that the emo- tion is unpleasant? A second purpose of this study was to obtain evidence about how spon- taneous facial actions might be related to a person's subjective experience of emotion, Most previous studies instead have focused fon how spontaneous facial behavior varies with some manipulations in experimental con- dition, or signals information to others. Either of two complementary methods could be used to score facial behavior: di- rectly measuring facial actions when different SN, AND S. ANCOLE emotions occur or asking observers to make emotion judgments when viewing samples of such facial behavior. We chose direct. mea- surement of facial actions to accomplish the third purpose of this experiment—to test hypotheses about the particular facial actions that signal particular aspects of emotion (e.., Ekman & Friesen, 1975; Ekman, Friesen, & Tomkins, 1971; [zard, 1971). So far these formulations have been tested only in terms of their ability to identify contrived facial expressions, not spontaneously occurring emo- tion. We sought to provide a sounder basis about what different facial actions may signify. ‘Method A new technique for measuring facial bebavioe, te Facial Action Coding Sytem (Ekman & Friesen 1976, 1078a), wes. applied to. videotape records ratheed while subject viewed. postive and ste Indicing motion picture Sims and then reported on ‘heir subjective experience Sub jects ‘Thirty-five right-handed, female Caucasians, rang- ing ia age ffom 18 to 38 years (W=25.1), Folune laored as subjects. Subject of only one gender were recruited, since Hal (978) has shown Some difere nce ie acal encoding asocated with gender. Tt (Was rhoupht that female subjects ght feel more Comfortable chan_ males with the female” exper fmenter (SA). ‘Through "advertaements subjects ‘Were reruted to. parcipate in an exptiment om payehophysiology in which brain waves, Beart rate Tnusee” tenon, basal skin restance (BSR), and fespraton would be recorded. while they watched fms. ‘They were aot told that we were interested In emotion or that they would be videotaped. This information wat witheld until alter the erpetiment 1 void sefseonscious behavior, Stimuli In the positive condition, subjects saw Alms in- tended "to induce postive aft; inthe negative ondition, subjects Saw 1 Sim intended to induce regative aoc, The posive Alms canted of three tinct tomin. Sins produced by Ekman and Feisen (ior) that yielded similar sllrepors of TONE pastve afeet” One fm showed ori paying in zoo, another showed ocean waves, andthe third Showed 2 puppy playing with a Bower, The reactions the ocean fm ae not reported since thi lm sr designed 10 yield Iitle olvecrable facial te Sponses indeed, this was the case Only seven of the =e ge Pe Pes. eee ee p= =he kee we wae & ee ee ee oS eS FACIAL SIGNS OF EMOTION ar Subjects showed. any facial response to the ocean ‘lm. This Bm hid been included for the. purpose St other investigation ot the paychopbysiologial Fesponses of thee. subjecs (Anco, 1979)- The three pesitve Alms were siways shown im the came order, with 2 l0see blank pecod between eh orl, ocean, and then puppy. The’ im intended to produce negative fect was an edited version of workshop sesdent Alm Sst Sed in reaarch by ‘Birnbaum (1968). Otber ine NGitirton (Lataruy 1906; Eliworts, ‘Note 1) ound that subjects reported experiencing lear and lsgust. The mwas cited to eliminate most of the scenes other then thote rely Tending to and Sowing two accidents: Aman has the up of bis finger cut of by a saw, and a man dis after plane of wood i thrast through his chest by 3. coal Saw. These two seenes were always shown in the Same order, with only a few seene between then fist he finger ext, then the death Subjective Experience of Emotion Subjects reported thir emotional reactions on questionnaire composed of 4 set of point scales that previous studies (Ekman & Fresn, 1974) had found sensitive to. aiferental reactions tothe simul fms, Separate scales were included. for interest, anger, gust, fear, happiness, pain, sad ‘es surprise, and azousl. Each sesle wat unipolar, with 0 tepresenting no emotion and 8 the strongest feeling, Tstrucuons explained that the scength of 2 etl should be viewed a coma- bination of (a) the aumber of times You felt the ‘motion—its frequency; (b) the Tenge of ime You felt the emotlon—its duration; tad (©) how Intense or extreme the emotions wari Intensity, Only two of the emotion terms were explsned. Pain was std to refer to the expesience of emps- thetic pain Arousal was sid to be an index of the {otal emotional state Procedure Each subject was sen individually, She was seated In ad mz m clecsicaly shielded. room, Leads for recording electoencephalogram, heart rate, repie ration, BSR, and skeletal muscle electromyogram were attached. (Results on these autonomic met sures are reported. in Anca 1979) The. video: ‘camera was concealed, The lights were leit on io ‘he room, and the subject was tld that abe had been randomly selected tbe in 4 BHEAeight cone tt tatty Note All teste are one-tailed except for anger and interest, which are wo-taled, well as other facial actions in which the lip comers are not pulled up. These scores were not. significantly cortelated with happiness self-reports in either film and did not predict which films the subject had liked best To summarize, measurement of a specific facial action produced by the zygomatic major ‘muscle, Action Unit 12, accurately differenti- ated reported differences in positive affective experience. The more Action Unit 12, the more happiness the subjects subsequently reported. The amount of Action Unit 12 pre- dicted which of two films would be rated as having produced the strongest happy feelings. And subjects who never showed Action Unit 12 reported less happiness than those who did, Similar questions were asked about facial activity and negative affect. Facial Signs of Negative Affect Unhappy or not. Among the variety of actions scored by FACS, there is a subset that Ekman and Friesen hypothesized as being signs of anger, fear, disgust, sadness, contempt, or blends of those emotions. Forty. fone percent of the facial movements that occurred when the subjects watched the acci- dent film involved such facial actions. The first test of the hypotheses about which facial actions show negative affect was whether those subjects who never showed such facial activity reported experiencing less negative affect than those who did show such facial actions. Twelve subjects never showed the facial actions hypothesized as negative affect signs, Most of these subjects showed facial behavior interpreted by Ekman and Friesen as positive or ambiguous. The alfective self- reports averaged across the two accidents were compared for these 12 persons and 21 persons showing negative facial actions. (Two ‘more subjects, whose facial behavior was ambiguous, were excluded because negative action units were shown considerably before the appearance of the first accident.) Table 3 shows that those who showed ac- tions hypothesized as signs of negative affects reported significantly more fear, pain, sadness, surprise, and arousal than those who did not ‘The trend was in the predicted direction for reports of interest and disgust, but not for anger. How unhappy. The next test of the by- potheses about those facial actions that show reyative affect was to determine whether these actions correlated with the self-reports of the intensity of specific negative emotions. Table 4 shows that these facial actions correlated with the self-reports (averaged across the two accidents) for some of the negative emotions as well as with the report of generalized arousal How disgusted. The next issue was whether the specific predictions about the facial signs of particular negative emotions are correct. These hypotheses could be tested by determining if variations in facial signs of a particular negative emotion were associated ‘with variations in the subjective experience reported for that emotion. So, for example, actions predicted to be signs of disgust should correlate with the report of disgust more than with reports of sadness, fear, o pain; actions Predicted to be signs of fear should correlate with the report of fear more than with the ee ee eee ee ee ee ee)he bee we ee ob ee ee ae es he FACIAL SIGNS OF EMOTION nt Table + Correlations Between the Sum of All Negat Facial stions and Reports of Specific Emotions Facial measure Diagust Anger 08 = Frequency scores Duration sores Note, N= 35.4 p 10, p <.05. 0p <0 reports of disgust, sadness, or pain, and so forth. It was possible to test hypotheses only. about dispust, since very few subjects showed action units relevant to any of the other negative emotions. ‘Thirteen subjects showed facial actions hy- pothesized as signs of disgust without showing actions relevant to any of the other negative emotions. (These appearances involve pulling ‘upward the central portion of the upper lip. raising and stretching the nostril wings, and eepening the nasolabial fold, They are pro- duced by levator labii superioris; caput in- fraorbitali, Action Unit 10; and alaeque nasi, Action Unit 9, which also wrinkles the bridge and sides of the nose.) The top portion of Table $ shows that these facial actions were correlated with the self-reports of dis- gust and not at all, or negatively, with the Table S Selereport Sadees Fear Pain ‘Arousl 1 rs 0 12 reports of anger and sadness. Contrary to predictions, one of the disgust facial action measures was significantly correlated with the report of fear, and there was a trend with the pain report. Suspecting that these findings on fear and pain might be due to general arousal rather than the specifies of these emotions, we calculated correlations partialing out the self- report of arousal. The middle rows in Table 5 show that the correlations between disgust facial actions and the self-report of disgust survived, whereas the relationship between these facial actions and the report of fear or of pain weakened. The pattern of partial cor- relation thus supports the hypothesis that these particular facial actions are specific to the subjective experience of disgust. Te might be argued that when people watch fan unpleasant film, any facial activity, not Parliol Correlations Between Facial actions and Reports of Specific Emotions, Controlling on the Seif Report of Arousal Saltreport Type of correlation Facial measure Diogest Anger Sadness Fear Pain Taw Diagust frequency a 18 Diagute duration a a Panta, controlling on arousal Dingo frequency BL st 82S Diagere duration su a a Ambiguous frequency 10 Ambiguous duration 2» Noe N= 35. ¢p < 10. p <08 p< OL32 P. ERMAN, W. FRIESEN, AND 5. ANCOLI Just Action Unit 9 oF 10, would be associated with the experience of disgust. This was not the ease. The bottom two rows of Table 5 show that the ambiguous facial actions corre- laced less with the report of disgust than did the disgust facial actions. Discussion Facial action was found to provide accurate information about a number of different as- pects of the subjective experience of emotion. Variation in specified facial actions was re- lated to the intensity of reported emotions, to the extent of happiness, and to the extent of unhappiness. Even when a person reported enjoying two experiences, facial measurement discriminated which was’ enjoyed most. And there was evidence to suggest that facial actions “may be specific to each negative ‘Although others (e.g, Birdwhistell, 1970) have claimed that smiling occurs with nega- tive as much as with positive emotions, this study supported Ekman and Friesen’s hypathe- sis that a particular type of smile—Action Unit 12 produced by the zygomatic major rmuscle—is associated with the experience of happiness. Some support. was also obtained for Ekman and Friesen’s hypothesis that only this action is a positive sign. Smiling actions in which the upward curve of the lip corners is produced by risorious, buccinator, oF zyg0- ‘matic minor muscles (predicted to be signs of fear, contempt, and sadness) occurred al- ‘most exclusively during the accident film. Since Action Unit 12 was not totally ab- sent during the accident film, the argument ‘ould still be made that even this type of smile may be a sign of negative emotion. Although not conclusive, the evidence sug- ests this is not so. At least it would be a rare sign of negative emotion, for Action Unit 12 occurred only about one tenth as much during the accident as during the gorilla/puppy film. Furthermore, the occurrence of Action Unit 12 during the accident ilm was not related to the self-report of negative emotions ot posi- tive emotions. Ekman and Friesen (1975, chapter 11) have outlined a number of reasoas why Action may occur during negative affect, jzh they maintain that this action is fof necative affect. Quite the com trary. Signaling positive emotion, Aetion Unit 12 may be deplaved in a number of negative contexts: to comment on the negative feelin (ext, afin and bear it): if there is a blend of positive and negative feelings (eg, scom, hietersweet, etc): to.mask a neyative feeling: fo to cleliberately, falsely simulate the ap: pearance of positive feeling. Two further fnd- ings about Action Unit 12 expressions during the accident film were consistent with these interpretations. First, almost half of them were part of expressions that also included actions that signify negative emotions. This virtually never occurred for Action Unit 12 expessions during the gorilla/puppy film. Second, preliminary analyses suggest that dur- ing the accident film, Action Unit 12 expres- sions were more asymmetric than during the orilla/puppy film. This fits with other ind- ings (Ekman, 1980; Lynn & Lyna, 1938: Ekinan, Hager, & Friesen, in press) that when Action Unit 12 is a sign of spontaneous hap- Piness, it is more symmetrical, and when it is deliberately performed, itis more asym- metrical Finally, it should be acknowledged that some of the Action Unit {2 expressions dur- ing the accident iil could have been in- stances in which positive affect was felt There are probably few experimental manipu- tions of negative affect that totally succeed for all subjects continuously and that do not strike at least a few, at least for a moment, 44s ridiculous or amusing. And, of course, subjects may at points show a happy expres- sion in relief once they realize chat the nega- tive affect manipulation is not going to be as terrible as they feared ‘The findings on the facil actions relevant to negative aifect, and in particular for dis- ust, were encouraging. They are consistent with findings on the meaning of particular facial actions in studies of posed behavior. They should be replicated with other stimulus films and vith other events apart from film as the source of negative afect, Similar re- search is needed also to test hypotheses about the facial actions signaling anger, fear, sad- ee Fe Se ses Sees ee Ps es oe See ee PPheb ew Ses we ee be we ee ee a el FACIAL SIGNS OF EMOTION ns ness, pain, and surprise: variations in the in- tensity af each of these emotions: and blends these emotions. To demonstrate more con- clusively the speciicity of the emotional in- formation signified by a particular set of facial aetions—chat they are associated with one ut rot with another negative emotion—a com parison is needed that could not be made in this study. It is necessary to show not only that disgust actions correlate with disgust ‘more than with anger ceports (as found in this study) but also that anger actions cor- relate with anger more than with disgust reports. To make such a comparison, emotion- inducing stimuli would have to be used that would elicit facial actions relevant to anger and actions relevant to disgust. More generally, this experiment demon- strated that facial expressions are different ated for the spontaneous occurrence of par- ticular emotions. [t thus strengthens the em pirical base of the theoties of emotion that emphasized differentiated response systems, particularly those that dealt extensively with the face (Plutchik, 1962; Tomkins, 1962, 1963; and theorists who derive from Tom skins, such as Ekman, 1977, and Izard, 1977). ‘These findings should challenge the cognitive theorists of emotion to expand their formula- tion to accommodate differentiated facial ex- pressions of emotion and to consider measur- ing facial expressions in their experiments, How important spontaneous facial expres- sions of emotion may be as social signals was ‘not addressed in this study, although these findings raise such a possibility. Before pre- suming that facial expessions provide clues to ‘others about how a person feels, let alone the further assumption that feedback of how others reacc may influence the expressot's ex: perience, two research steps must be taken, First, it! must be shown that people who ab- served expressions, such as those shown in this experiment, would be able to make the dlifferentitions ‘that were achieved by the fine-grained, slowed motion measurement with FACS. Second, it must be determined whether factal expression will remain as rich a source of differentiated information about emotion when the subject is not alone but in the pres- ence of others, Ekman and Friesen (1969) have theorized that in social situations, people wittingly and unwittingly manage their facial expressions of emotions following culturally based display rules specifying who can show What emotion to whom and when, Friesen (1972) began such work by showing difer- ences between Japanese and Americans when they” watched “stressinducing films in the presence of another person. It is Skely that in some social situations, at least some people may amplify sthile others may conceal or dis- ise their expressions, but this remains to he studied, ‘This experiment also served to show the utility of FACS, Since all observable facial behavior was measured, not just some actions presumed to signify emotion (as in Tzard's, Note 2, facial measurement technique), it was pustible to test whether the presumptions were cortect or not, Observers’ global juda- ‘ments of emotion, the most popular alterna- tive for measuring information from the face, ‘cannot isolate the particular actions that do and do not relate to a particular emotion. Reference Notes 1. Ellsworth, P. Personal communication, 1976 2 aard, C. The maxially disriminatioe fit ‘movement coding sytem. Unpubined manisept, Truteuclonal Resources Center, University of Delaware, 1978. References Ancol, $. Piochophasiolorcal response patters 10 ‘msi. Unpublished doctoral disertati, Unie ‘erty’ of Calforna, San Franesco, 1979 Birdwhiel, R. L Kiseriee and contest, Philadel- fla: Universy of Pensiyivania Pras 1970. Bimnbaum, Re Autonomic vention to vhret. ond Confrontation conditions of payeholopicl ret Unpublished” doctoral daseration, Univenity of California, Berkeley, 1966, Burton Jone, N. G. Cates for we in deveibiog facial ‘expresion in chldren Human Diology, 1071, #1, 365-415, Brannigan, C. R, & Humphries, D. A Human none ‘eval behavior, 2 means of communiciuen. In NB Blucton Jones (Ed), Etholoial studer 3f ‘child behavior. London: ‘Cambridge Univesity Pres 1972 Darin, C. The exprsion of the emotions im mon ‘and animals New York Polosoptieal Libry, 1985. (Originally published, 1872.)1s P. EKMAN, W. FRIESEN, AND S. ANCOLE Ekman, P. Universals and cultural diferenes in Tacal expeesion of emotion. Tn. J. Cole, (Ed), Vebraske Symposium om Mowvation (Vol. 19) Lincoln: University of Nebraska Pres, 1972. Ekman, P. Biologie! snd eltural contributions to ‘Body’ and facial movement. Tn J. Blackiog, (E4, Mnthropelogy of the body.” London: Academie Pree, 107? Ekman, P. Asymmetry in facial expresion. Science, 1980, 209, s83-856 Ekman, P, Frien, W. V. The repertoire of non- ‘erbal behavior: Categories, ign, usage, and eng. Semiotica, 1968, 1(1), 9-98 Ekman P, & Friesen, W. V. Constants across cule tre inthe face and emotion. Journal of Per- onal ond Socal Pryehology, 197%, 17, 12-129, Ekman, P, & Frisen, W. V. Detecting “deception ‘rom the body or face. Journal of Prtonaty end Social Poychoogy, 1974, 2, 288-198 Ekman, P, & Frccen, W. V. Unmasking the Jct ‘Englewood Clie, NJ. Prentice Hall 1975. Ekman, Py & Friesen, W. V. Measuring faal move- tment Environmental Peycholory and Nonverbal Behavion 1976, 11), $675. Ekman, Py & Feisen, W. V, Inestiqators Guide 19 the Facial Action Coding System, Part IT. Palo Ato, Calf Comulting Psychologate Pres, 1978, tC) Ekman, P. & Friesen, W. V. Monaal Jor the Pedal ‘dion Coding Sytem. Palo. Alto, Call: Come sulting Poyeholorits Press 1978 Ekman, P, Feasen, W. V, & Ellsworth, P. Emo- on i ihe humm face Guidelines for research Gnd am integration of findings. New York: Per~ samen Pres, 1972, Ekman, Py Frisen, W. V,, & Tomkins, S. 5. Facial fect scoring technique (FAST): Aes vablty dy. Semiotig, 197, 30), 37-38. ‘Ekman, P Hager, J. C., & Frsen, W. V. The wym- retry of emotonal and deliberate facal actions Prychophysology, in press Ekmaa, P, & Oster, H. Facial exresions of emo- Won Annual Review of Psychology, 1979, 30, S27 ssi. Ekman, P. Sorenson, E, R, & Feissen, W. V. Pan- altura ‘ements ia facial apays of emotions ‘Science, 1968, 16403875), 86-88. Fensen, W. V. Cultral ferences in facil expete sions in a socal situation: Am experimental est, of the concept of eispay rales. Unpublished doce Coral ditertation, Univetsty of California, San Francaco, 1972. Grant, NG. Human fadsl expresion. Man, 1965, 525-806 al, J. Gender eects in decoding nonverbal exe Piycholopca! Buti, 1978, 35, 46-857. Inara, C. The fae of emotion, New ork: Ap (Cenerary- Crofts, 197) lard, C. Human emotions, New York: Plenum Pres, 1977 Kiineberg, 0. Social prvchology, New York: Hot, 1% Laird, J. D. Selt-atibution of emotion: The eflets ‘of eapeesive behavior on the quality of emotion ‘ioerenee Journal of Personally and Social Pye chology, 1978, 29, 47586. LaBare, W. ‘The ‘cultural bass of emotions and aesturcs, Journal of Personality, 106, 15, 4-68 Lassnss, R. 5. Psyehologieal tres and the coping proces. New York: McGeaw-Hil, 1966 Leach, E. The infence of cltral Context on none erbal communication ia man. In Ry Hinde (Ed), Nonverbal communication. London: Cambridge Univenity Pres, 197 Lynn, JG, & Lyas, D. R. Facechand laterality in ‘elation to personality. Journal of Abnormal and Social Psychology, 1938, 33, 29% Nig Ny Hull, CH, Jenkins, J, Steinbrenes, K. © Bent, D. Statistical package for the soil sie fences (2nd od). New Yorks MeGraw-H. 1975, Plutchiky R. The emotions. New York: Random House, 1962, Schachter, Wy & Singer, J. E. Cognitive, socal, and ‘ysiolopial determinants of emotional Sat, Prychologcal Review, 1982, 69, 19-399, ‘Tourangeau, R, & Ellsworth, P. C. The rle of facial Tesporse in the experince of emotion. Journal of Peronaity and Socal Psychology, 1999, 37, 1810 1831 Tombins, SS. A¥eet, imagery, conciousness: Vol Tr The postive afectr New Works Springer, 1982 ‘Tomkins, 5S, ABec!, imagery, comtcowsnes: Val The negative aes. New York: Springer, 1963, Received January 18, 1980 pe Pe Ss se ees Pee Pee ee ee