/\-,Motivation and Emotion. Vol. 15. No.4, 1091

Contradictions in the Study of Contempt:What's It All About? Reply to Russell

Paul Ekman 1

University of California, San Frallcisco

Maureen O'SullivanUniversity of Sail Francisco

David Matsumoto

San Francisco State University

A number of methodological problems make it difficult to draw any concluionsfrom Russell's .'itudiesof contcmpt, including a task which may maximize theinfluence of unfamiliarity with the ta.<>:k.alld instructiolL"which may encourageobservers to rate many rather titan few enIotlorlS. We rabie questioll.<t;al<t;oaboutecological validity, and the appropriatene.rs of using still photographs to studythe influenceo! context on the jugmcilis of cmotioll.

;'I.

. --- -- - .--

We agree that many factors may influence the inferences observers makewhen vjewing a facial expression. It is evident by examining differences ob~

tained jn Russell's various studies, and from our findings as well, t~at still. I:

photographs.oLa.uniiateral tightening .of-the lip corner is judged.as con~ - .tempt undcr one circumstance and not undcr another. Observcrs do infercontempt if they make that judgement when judging many different ex-pressions, and wheJ:l given a choice of at Icast seven different cmotions,including both positive and negative tcrms. Obscrvcrs arc more likcly toinfer disgust than contempt if they are shown only one or two photographs.What are we to make of this?

lAddrcs... all corrc5pondcncc 10 Paul Ekman, Human Inleraclion Laboratory, Univcr.;ily ofCalifornia, 401 Parnassus, San Francisco, California 94143.

293

0I4('-72~f)J')III200~2QJSO(,.'i(ltfl f'I 1<t<!1"I..num "utolid,;nr C"nr",..,,;nn

---294 . Ekman, O'Sullivan, and Matsumoto

-

First let us consider two possible methodological problems in Russell'sseries of experiments. In the first experiment Russell gave his subjects onlyone photograph and asked them "to describe the emotion expressed withwhatever single label thc) chose", While this procedure has the virtue ofseeing how people freei.. bhcl photographs of expression when they havenot had prior cxposure I.' ..,ther photographs, it may have the disadvantageof maximizing thc influ. II .e of task unfamiliarity on the results. The factthat nearly one-third of t11('responses were completely idiosyncratic sug-gcsts thc possibility that hi'>.;ubjects might have becn confused about whatwas expected of them, In 0ur.early studies (Ekman, 1972) we found unre-liability in initial responses when subjects had to judge exprcssions, Forthat rcason we have since always thoroughly explained the task and thenencouraged subjects to ask questions about the task before proceeding.Even if some subjects might still have been uncertain after we answeredtheir questions, this probably had little influence on our results, as we havefound that subjects better understand what is expected of them after tryingit a few rimes. Typically we collect judgments on 30 or more photographs,not on one or two, and (he effect of possible confusions on initial responsesdoes not count for much.

A different problem may have occurred in the second experiment andin a number of Russell's (199Ia) other recent studies of contempt in whichhe had his subjects "rate the degree to which the face shown expressedeach of six emotions," rating each emotion on a 4-point intensity scale,Wedding this judgment task with a design in which the subject sees onlyonc or two photographs. intended to depict only one or two different emo-tions, may crcatc dcmand charactcristics to rate many rather than fewemo-tions. OthclWise why would the experimenter have given them so manydiffcrent labels whcn thcy are seeing only one or two expressions? Whenwc use the multiple-emotion rating procedure we show the subjects manyphotographs. Our subjects may develop a different set, believing that manylabels ~lrc provided because many faces are to be judged, and thus maynot bc impiicitly encouraged to use many labels on each photograph.

. Even if we dismiss these possible design problems. we regard the dif-ferences-bctween Russell's findingsaod Q~IJ~as Jri~~'ll fo~~~e following rea- . .

sons. First, the problem may be limited only to the English language.--Our +-----findings have been replicated in. many cultures and languages; Russell hasonlystudicd English speakers in North America. It may be that in Englishthe lexical distinction between disgust and contempt is less clear than it isin most other languages. Such an interpretation would be consistent withour finding that there is higher agreement about contempt judgments in anumber of other languages than we h.tVe found for EngJish speakers(Ekman, O'Sullivan and Matsumoto, 1991). Second, Russell has not deter-

ji

Ii

I

"- .

- I

Contradictions In the Study ur Contempt " 1

mined whether his results are specific to contempt or common to other em"'>tions. Third, his procedure--each person sees one or two faces-has leclaim to ecological validity than our procedure-each person sees many dferent expressions in many different cOntexts. While neither of our desigcan claim to be very similar to actual life, certainly secirr. many differeexpres.':>ionsis a bit morc life-like than forming impres.';i('fls about onetwo isolated expressions. Fourth, we believe the v~.: of still phO1graphs-even multiple expressions as we have done-to :.tudy the issuehow context influences observers' inferences is too artiticial to learn muof value. Most of our inferences arc drawn from seeing moving faces rstatic ones, attached to bodies not disembodied, with sound and words coing out of the mouth not silent, and with observers having informationexpectations about the situation in which the expression occurs. It is qupossible to utilize videotape to manipulate many of those contextual v,abIes and measure their influences on observers' inferences. We and oth

(Berry, 1991~iJugental, 1986; Ekman, Friesen, O'Sullivan and Scherer, )9:have done such work. Russell (1991b) terms such work the study of Iexpresser's context, and says that he, instead is interested in the variabother than the expresser which affect the observer, what he terms the jwment context. Granted his interest, we still believe it would be far preferato utilize more realistic, robust stimuli than still photographs. It is not tlwe believe still photographs are useless-we continue to use them to ansvsuch questions as whether the apex of an expression can elicit agree meacross a group of observers-but we think it is the wrong medium to addr,the more complex social psychological questions which Russell focuses upe

Russell believes he has established the importance of the observt:context on the judgment of facial expressions of emotions. We thinkhas failed to demonstrate this because of design flaws, and because 1

use of still photographs to represent expression and to manipulate cont,has little relevance to important issues in the socia! psychology of intpersonal perception. We hope in the future Russe!! examines these imp'tant questions in more 'robust instantiations of the phenomenon.

_m m_.REFERENCES.' ~---~----------------

Berry, D. S. (1991). Accuracy in social perception: Contributions of facial and v(information. Journal of PenOlwlity and Social Psychology. 61. 298-307.

Bugental, }). (1986). Unmasking thc "polite smilc": Situational and personal dcCermin:mtmanaged affect in adult-child interaction. PcrsOlwlity alld Social P!'YC:/lOlogyBullet;n,7-16.

Ekman. P. ft972). Universal and cultural differences in facial expressions of emotion. ICole (Ed.). Nebm.d:lI .\ymfJos;ulII 011 11/o{;",lI;OIl.!1J71 (pp. 207-2K\) Lincoln: Univclof Nebraska Press.

I'

296 Ekman, O'Sullivan, und Matsumoto

Ekman, P., Friesen, W. V., O'Sullivan. M., & Scherer. K. (1980). Relative importance of face,body and speech in judgments of personality and affect. Journal of Personality and SocialP.o;ycholofO',38, 270-277.

Ekman, P., O'Sullivan, M., & Matsumoto, D. (1991). Confusions about context in thejudgment of facial expressions: A rcp1y to "The contempt expression and the relativitythesis." Motil'otiOlI and Emotion, 15, 169-176.

Russell. J. A. (1991a). The contempt expression and the relativity thesis. Motivation andEmotion. lJ, 149-168-

Russell, J. A. (I991b). Rejoinder to Ekman, O'Sullivan, and Matsumoto. Motivatioll alldEmotion, 15, 177-184.

~- -~-- -~ ~ ~ ~ -~_. .- ~.- - -~ -~ -.- ~ ~ .. --'---'-_.' '-'---' '---'---~ ... '..

I.

IIeI -L ~-.-

!iI'

:iIIF't:,

"Ii!'I'iIil.,

~ IIIii!

Ii

Motivatioll and Emotioll, Vol. 15, No.4, 1991

ACKNOWLEDGMENT

The Editors wish to thank the following people. who contributed by re-viewing manuscripts for Volume 15:

Robert AbelsonLynn Y. AbramsonLauren AlloyCraig AndersonKristen Joan - AndersonSusan AndersonJames Averill

Richard J. DavidsonEd Deci

Douglas DerryberryEdward F. Diener

David Dunning

P. Christopher EarleyPaul Ekman

. --

Anita Barbee

John A. BarghRobert A. BaronRobert S. BaronLeonard BerkowitzP. M. Biner

Virginia BlankenshipJames BlascovichElliott BlassSteven J. BrecklerMarilyn BrewerJoel--.Brockner-------

Daphne EugentalBrad Bushman

Michael FanselowNorman T. Feather

Tiffany FieldJoseph ForgasAlan FridlundIrene Frieze

John J. Foredy

Thomas Gilovich

Tony Greenwaldm mm ~-----------------

Jennifer CampbellRobert B. CialdiniLee Anna CJark

John C. CondryMichael ConwayMichael Cunningham

Judith A. HaUJudith M. HarackiewiczSusan D. HarterJeannette Haviland

Bill Ickes

Dan ligenRick Ingram

--- - -- '297