Bana l l , N . K . , Lernm, K.M., and Carpenter, S.J. (2001). The social unconscious In A. Tesser and N. Schwarz (Eds.), Blackwell Handbook of Social Psychology: Intraindividual Processes. Malden, MA: Blackwell Publishers, Inc.

The Social Unconscious

Contemporary social psychologists are aware that long before concepts of cognitive media- tion were admissible in scientific psychology, their prcdecesso~had been sufficiently en- tranced with matters of mind to study them even at risk of marginalization by the then dominant antimentalists. The first social psychologists were bold not so much in their recog- nition that thinking, feeling, and motivation were fundamental mental systems - for hun- dreds of years, thinkers even in the western world, had known the same. The unique audacity of these psychologists was in the belief that thought, feding, and motive could be scruti- nized, manipulated, and subjected to experimentation in a manner not unlike the treatment accorded to particles, ions, and bacterii It should be of little surprise then, that a field so confident a century ago that processes of conscious mental life could indeed be measured is no;;. equally confident about measuring m e n d life that lies beyond consciousness.

Johnson-Laird's (1983) question "What should a theory of consciousness'exp1ain?'' pro- duced four components, awareness, conttol, intmtion, and self-r&mon, that a tractable theory of consciousness must explain. The focus of this chapter is on the hidden side of consciousness, which leads us to focus on the inverse of these components: thoughts, feel- ings, and actions performed outside conscious awareness, without conscious control, or without intention. At this stage, research on unconscious processes reflects a basic artempt to demonstrate that particular unconscious processes occur at &to trace the boundary conditions of their operation, to document the fid richness of the systems that are engaged (cognition, affect, motivation) and the levels of social objects they include (e.g. self, other, social group). With a strong emphasis on developing robust and replicable methods for investigation, researchers have asked: can knowledge that resides outside conscious aware- ness influence social thinking, feeling, and action? Is unawareness actually a precondition for observing particular effects? How should we characterize attitudes that are fully within

This resevch was supported by grant number SBR 9709924 from the Naciond Science Foundation. We thank Richard Hackman and Aiden Gregg for their helpfirl comments and suggestions.

The Social Unconscious 1 3 5

,warencss but relatively outside conscious conuol? Is it possible to consider the uncon- g i o ~ activation of go& and motives as we have come to accept unconsciou cognition md sect? What do these investigations imply regarding the notion of free-will, panicu- lasly as it concerns assumptions about the heedom or constraints to think, feel, and act toward one's self and others? T h e are among the questions that have mattered to con- temporary psychologists interested in the analysis of the social unconscious.

~esearch on unconscious processes does not, unfortunately, reflect a sensible obsewancc of terminology of constructs and processes. Terms like auwrnatic, implicit, unconrcious,

indzrect arc often used interchangeably, and sometimes to refer to divergent underly- ing processa (eg. awareness versus control). In recognition of our own complicity in cre- ating this conhion, Wartempt to restore some order for h w e discussions by using the term unconrciow to refer to the family of processes that occur outside conscious awareness, withour conscious control, or without intention to The use of the term uncon- ~riow reflects a deliberate attempt to capture i s usage from still largdy psychodynamic meaning. In addition, we the term implicit to refer to those processes that operate without the actor's conscious awucness, and the term automtic to refer to processes that operate withoat the actor's conscious control. In time, these issues and concerns about

validity and relationships among fiunilies of measures will be resolved. How- ever, progress will be greatly speeded-up by charting a dear research agenda and encourag- ing greater collaboration across laboratories with divergent perspectives and methodological allegiances.

Much social psychologid research can be *said to be, in essence, the study of processes that operate outside conscious awareness and intention. After all, experiments must rou- tinely create circumstances in which the behavior that is observed and measured is free of the concerns of social desirability and demand characteristics, and in that sense unaware- ness about the source of influence on behavior is the norm. Yet, it is only more recently that unconscious processes in social behavior have been examined in their own right, rather than as a methodological by-product of social psychological experimentation. This review revolves around those experiments that bring a deliberate focus to unconscious processes becaw of a genuine interest in the limits on introspection, in understanding the extent and nature of the social unconscious, and in using the study of unconscious processes as the basis to challenge commonplace notions of individual responsibility on the pan of s d actors, and assumptions of justice in interpersonal treatment of social targets.

Over two hundred yean ago, Immanud Kant wrote:

We a n reduce all the powers of & human mind, without acepaon, to these three: the cophiw powr, thejdng ofp lwwr and a!zsphm, and the powr of dcn'm. It is m e chat philosophers who otherwise desezve unlimited praise fbt the thoroughness in their way of h k m g have asserted that this distinction is only illusory, and have mcd to bring all powers under no* but the cognirive power. Yet it is quite easy to establish, and has in fict been d izcd for some time, that this attempt to bring miry into chat diwrsity of powers, though otherwise undertaken in the genuine philosophic spirit, is futile. ( h t , 179011987, p. 394; italics in original)

Two centuries later, we find it worthwhile to retain this mumvintc, and are bemused b rhe similar dominance of the "cognitin pond then, as it is now. We include raeych

on unconscious brms of &kt, and now that rcsarch on unconscious motives is at such an early smge of development that it would be hard to provide a responsible review of such work at this time.

This chapter brings together selected samples of research on unconscious processes as they inform cognition and a f k a ( b r reviews see B e , 1997; Bornstein & Pitunan, 1992; Granwald, 1992; Kihlstrom, 1990; Kihlsuom, Mulvaney, Tobias, & Tobis, in press; Ulcman & Bargh, 1989; Wegcner & Bargh, 1998). By necessity, the treatment here is not historical, and the attention to any srngle era is superficial. Should the review reflect a sense of the potential pemiveness and emerging lawfulness of unconscious pro- as they are revealed in social life, it will have succeeded.

Unconscious Cognition

By far the most research attention has been devoted to the study of unconscious social cognition. This is not surprising because of disproportionate attention given to the study of social perception, attention, memory, categorization, and judgment more generally. In this section, we devote our attention to three aspects of unconscious social cognition: the study of self, other, &d social group. Mminally, che demarcation is somewhat arbitrary, but it will allow us to build the case that a wide bandwidth of learning has become possible within a relatively short time on how humans think about themselves, others individuals, and the social groups of their species.

Self

Proposals to study the self as a cognitive structure have appeared since the 1970s (see Banaji & Prentice, 1994; Markus, 1977; Greenwald & Pratkanis, 1984; Ws t rom, Can- tor, Albright, Chew, Klein, & Niedenthal, 1988; Kldn & Lob, 1990; ~ i n d - e & Carlston, 1994) and these set the stage for contemporary research on unconscious processes involv- ing the self. These position papers and experimental accounts placed the study of self firmly in cognitive space, often using the dominant language of models that allowed a connection to established constructs such as memory (e.g. Greenwald & Banaji, 1989; Klein, Lofius, & Kihlstrom, 1996). The language of these m o d e h d the demystification of self that emerged out of these accounts permined unconscious~~lf p r o ~ ~ e s to also be measured. The research may roughly be separated into analyses of the unconscious manner by which sh ih in self-related processes such as self-presentation and self-evaluation occur, and the role of unconscious self-related processes in guiding an understanding of the social world.

The most illustrative findings chat show that unconscious activation of significant oth- ers can have implications for sdf-cvaluation come from studies in which the priming pro- cedure uses subliminal presentation. Baldwin (1994) used subliminal primes to activate representations of a significant other (one's adviser) who is cr i t id or accepting in orienta- tion and showed parallel sh& in v i m of self; also, that approval or disapproval from

unc~&~usly activated others can influence one's d u a t i o n of one's work ( ~ d d k ~ , cmell, & LOP, 1990). Besides the role accorded to sigd-t ohm, he idea hat membership in social groups has repercussions for individd psychologi~ functioning h s been of perennid interest to social xiendso (see Wakh & Banaji, 1997, for a roriov). In its most recent form, chis idea has led to research suggesting a potentid link bemen activated knowledge of beliefs about one's group and performance on ability tw. Steelc and Aronson (1 995, p. 808) point out that "the existence of a negative stercorype about a group to which one belongs. . . means that in situations where the stereotype is appliuble, one is at risk of confirming it as a self-characterization, both to one's self and to others who know the stereotype. That is what is meant by stereotype threat." In their experiments, ~1ackArnericans underpaformed on tests of intellectual ability, and women underperformed on tests of mathemadd ability (Steele, 1997) when subtly made aware of their group membership or the link berween the group and the negative attribute. Several addiriqd demonstrations of this finding now exist. For example, Lcvy (1 996) showed that sublirni- n d y activated negative stereotypes about old age creates decrements in memory perforrn- mce among elderly subjects; Croizet & Claire (1998) showed that eliciting information about parents' h e 1 of education led to a decrement in verbal ability among low SES stu- dents; Shih, Pittinsky, & Arnbady (1 999) showed that activating gender identity or ethnic identity among Asian women shifted performance to be respectively inferior or superior on a math test. Evidence about the robustness and ease of replication of these effects is only just beginning to be determined, and the mechanisms by which such effects are produced are not yet identified. Yet their implications foc the ease with which quality and fairness in treatment can be compromised by group membership are suEiciently shocking to rquire particularly intensive study by investigators with mrying theoretical perspectives.

In other research we learn that who one is and how one assesses oneself can implicitly influence views of others, just as we observed previously that significant others and social groups can influence judgments of self. Spencer, Fein, Wolfe, Fong, and Dunn (1998) showed that threat to self-image can automatically activate stereotypes of social groups even under conditions that otherwise do not produce such activation. And Sedikides & Skowronski (1993) showed the role of self in forming impressions of others more gener-

- ally, by demonstrating that central dimensions of the selfconcept were influential in judg- ments of others. Perhaps the most impressive corpus of research showing the role of one's s i g d a n t others in shaping social perception has becn obtained by Andean and her colleagues (Anderscn & Glassman, 1996). In providing the first aperimend evidence for transference, they show that activation of information pertaining to significant others im- plicitly lead m inferen& h u t new individuals that mimic representations of signXcant others and seIf. Moreover, such a c t h i o n a n a h elicit &id dm that captures the evaluation of the signrficant other and produces bchzvionl confirmation in interpersonal interaction.

T o d e m ~ ~ l ~ ~ a t e the role of unconscious processa in short-cum to sclfduation, S w a ~ , Hkon, Stein-Seroussi, & Gilbert (1990) demonstrated &at under conditions of limited cognitive apacity, pvticipants showed a simpler p r d k n c for selhkhmcing d whereas the a&ilicy of rrsourccs led to more infbmative self-vurfym% suategies. It also qpars that processes of s o d comparison occur with m h h d cognitive resources or intention to compare, and even d e n the souroc of comparison is nondiagnostic for

t

assessment. Social comparison can l a d to decrements in sdf-evdutcion in nrch cases when -mom arc unavailable to adjust for cheinapproprktcampuison (Giibm, Gicsler, &

1

i Morris, 1995). I

Studies showing the involvun~~t of unamcious sdf" pmccsscs arc numerous, and these examples arc sdectcd to show that iack of both awareness and aenaol play a role in ;rssessments of self, and that self-knowla& and personal relationships can unconsciously in-

I

t

fluence assessments of the s u r r o u n h social world. The bre;lkdown of simplidistiiictions a

between thdung and f d n g uc quite obvious in many on?lyses of self, and arvnining the role of unconscious processes shows such interrelations among mend systems to be funda- mtntal, and ddying of our Mpcrscd separation of these sysums fbr expository purpose..

Few topics in social psychology can be regarded as more mad to chc field's mission of understanding the stuff of human relations than the processes i n v o i d in one person observing, understanding, and assessing another. Although Katz and Braly (1933) and Icheiser (1 949) explicitly recognucd the role of unconscious processes in person and group perception, it was a later generation of mpcrimentus who with their newfingled technolo- gies studied the unconscious operation of person perception: to what degree and in what manner, they asked, are awareness, control, and intention components of the pervasive act of judging others? It is now dear that spontaneous, fluid, and effortless acts of person perception, when brought under scientic scrutiny, reveal the operation of a vastly intri- cate thought system, able to perform social gymnastics of incredible speed and elegance. The social gymnast, however, docs not alw& land on the balance beam. The research we review shows also a more clumsy side of person perception: susceptibility to situational intrusions, the constraints of routinized thought patterns, of errors in computation and application that create costs of varying magnitude.

Implcitpmrption ofothm ~ m r f m m the conmvrtr in tbe pmrivni m i d Among the highlights of this research literature are experiments conducted by Higgins, Rholes, & Jones (1977) whkh appeared without heralding the study of unconscious processes in person perception. Yet it ushered in a wave of research that has produced what some regard to be a law of social perception: constructs that are active in a perceiver's mind implicitly shape perception and judgment of others (Sedikides 8r Skowronski, 1990). Participants read material that served to activate knowledge about peraonali~qdities such as "srub- bornn or "persistent." Later, in a test of reading comprehension ostensibly unrelated to the previous session, participants judged an ambiguously described target to be more in line with the previously activated knowledge; those who had been primed with "stubborn" were inclined to find the target to be relatively stubborn, and others who had been primed with "persistent" judged the identical target to be more persistent. Participants in such experiments were not aware of the influence of previous experience in shaping their judg- ment, and in the absence of such awareness, there was no opportunity to controi judgment. Grtainly, we assume that participants intended to provide an unbiased judgment, on the basis of the actions of the target. Yet, as this experiment and the countless others using

The S o d Unconscious 1 39

I vyiationr of this procedure suggest, person perception can be guided by hon tht may emanate elsewhere, outside consciously accessible cognition (see Higgins, 1996, and sedikides & Skowronski, 1990, for reviews).

The many experiments that followed on the heels of Higgint, Rhola, & Jona (1977) served as more than just the dean-up crew. Thae experiments, continuing up to the prmnt,

a rich understanding of unconscious person perception. Additionally, theoretical be- works of various 1mL of specificity have been proposed that offer working explanarions, ruggat useful metaphors, and spec@ mechanisms (Higgins, Bargh, & Lombardi, 1985; Higgins, 1989; Herr, Sherman, & Fazio, 1983; Martin, 1986; Wyer & S d , 1980). While we cannot review the theoretical models here, it is dearly the case that the aperimend findings and rhmretitalZiempts to understand unconscious person perception (e.g. recency, frequency, awareness, specificity, chronicity, contrast) have allowed hidden aspects of un- conscious processes themselves to be revealed. Together, they have created a vim of person

that is altogether more compla and complete, and more troubling in its implia- [ions: perceivers belim that their judgments of others reflect properties of the target, and not of the thoughts that are implicitly active in their mind. That such influences on judg- ment occur witfiout the intention to create bias in the judgment process, and without a m - ness that such bias may even exist, starkly raises the question of the extent to which "mend due processn (Banaji & Bhaskar, in press) in interpersonal interaction can be assumed.

The robustness of a theoretical construct is evident when a diversity of applications provide supporting evidence tbr the principle. The activation of constructs, either t e m p rary or chronic, have been shown to influenci behavior in a variery of domains: desire to work with a gay person (Johnson, Bryant, Jadcson, Gatto, Nowak, & MacVittie, 1994); reducing risk of pregnancy (Norris & Devine, 1992); increasing the assessment of "ah- holicn ( Southwick, Stecle, & Lindell, 1986); explaining the cognitive states of depressives (Gotlib & McCann, 1984); priming aggression by sports (Wann & Branscombe, 1989); explaining individual diflkrences in aggression (Graham & Hudley, 1994); increasing judg- ments ofwomen as sexual objects ( R u b & Borgida, 1995); implicating television viewing as a vehicle of priming (Shrurn & O'Guinn, 1993); the role of chronic accessibility in electoral choice (Lau, 1989); and the role of priming self-interest in political reasoning (Young, Thornsen, Borgida, Sullivan, & Aldrich, 199 1).

Impliritpnrrption of o t b m f i k * spontaneous trait i n h e s The construct acccssi- bility literature shows thu our judgments of others arc influenced by the concepts that are active in our own minds at the time of perception. But what exactly is perceived when we obsem others' behavior? Knowing that there may be multiple causes for behavior, to what do we attribute a particular action? Following decades of rrstvch in person perception w i n g with Lewin and Heider, we know that the most common inference made is a trait arcribution - we encountu a behavior, and infer that some trait about the actor must be associated with its occurrence.

Ulunan and colleagues (Newman & Ulunan, 1989) kidred off a controversy in the field of person perception by nrggaiag that tnirs uc i n f d spontaneously, or possibly automatically, upon encountering a behavior. In an evfy demonstration, Winter and Uemvl(1984) had participants s d y descriptions of people puforming behaviors that Mplicd traits, such a 'the d o r laves his wifi with 20 pounds of laundry." ha, parcia-

pvlts~rskcdm~d~~nrmar~Uprcviourtgrrpdgi~~~dcucsth;zt were either tnip that-had bsm iinpiad By& stntt~cr (ca. inoonsidaru), n o n - d t semantic associates of the senunccs sea or d), xu no art. T u i t m c s frciiitated r e n t e n o c d bcmrrhuraoore,andud as or bermthanscrong#mp~~ticPSSOCiUt~, N g g a d n g ~ t 2 # L T i u p m o h s d u r m d d y d e m i t ~ ~ n ~ d m o f l e u n i n g b u t h e M o r . T k o& 5TI &a bas been replicated in many imataons, -providing convincing

evidence that traits arc inkffdd outside of ~~IlSCious rwvcncss (Moskowitz & Roman, 1992) and without oonscious impress ion-fbdn goals (eg. Skowronski, Culston, Mae, & Cnwfbrd; 1998; Uleman,lNewmul, &Winter, 19%; Whimy, Waring, & Zqmark, 1992). Although the apit-cud mall p a d g m has becn chdcngcd methodologically (D'Agostino & Becgle, 1996), evidence that uaics arc inftrnd sponweously at encoding has been provided by rtsczfch using methods other than cued d, i n d h probe rtc- ognition (Uleman, Hon, Roman, and Moskowitz, 6996), savings in a deafning task (Cariston & Skowfonski, 1994), and using blatant or sub& priming at mooding to in- hibit or facilitate STI (Newman & Uleman, 1990).

Spontaneous trait in* may pmvidr input to drrpan'tionaf i ~ e p r v c ~ u c s The bulk of the evidence suggests that most trait inf'nccs made without intention are inferences about behavior, not dispositional llerencn d i d y lin4ccd to the actor (e.g. Carlsron, Skowronski, & Sparks, 1995; Moskowitz, 1993; Ulernan, Moskowim, Roman, & Rhee, 1993; Whitney, Davis, & Waring, 1994). However, SII may play an essenad tole in the formation of personality inferences. Several models of person perception h v e proposed that dispositional inference proceeds in mulaple ~pga, the first of which requires minimal cognitive resources or control, and may thus be ansidered spontaneous or automatic (e.g. Bumstein & Schul, 1982; Brewer, 1988; Nggm, Strauman, & Klein, 1986; Ross & Olson, 1981).

Trope's rwo-stage infcrencc model (Trope, 1986; Trope & Liberman, 1993) and Gil- bert's threestage model (Gilbert, Pelharn, & W, 1988; Gilbert & I h U , 1988; Gilbert & Osborne, 1989) propose that observed behavior, the situation in which it occurs, and prior information about the actor are automatically identified in terms of underlying traits (e.g. "chis is a fiiendly behavior"). The output of this automatic identification stage in turn .

becomes the input for the dispositional inference, in which behavioral, situational, and prior information that has been identified in terms of traits is combined to form a trait attribution about the actor. Experimental evidence shows thatpeople make behavioral identifications even under conditions of diminished cognitive resources, whereas conscious correction for situational contributions to behavior may be inadequate if perceivers do not have adequate cognitive resouras (Trope & ffieri, 1997; Gilbert, Pelharn, & W, 1988). The implications about human nature and n u m e from thae models are also troubling: because people are so often engaged in concurrent activities, behavior characterizations are ofien not appropriately adjusted for situational contributions to behavior. This can tilt toward trait- (rather than situation-) correspondent inferences, a phenomenon also termed the fundamental attribution error (Ross, 1977).

What early attribution theorists had predicted, research over the last m n r y yevs has confirmed about the swift and remarkably sophisticated inferences chat are made about

jndiGdd others in one's social world. The methods that are used are reliable, rhis dowed a healthy exporting of methods outside the labontoria in which hey were devel-

The theories of human inference processes in social conten that have emerged ue ,--rive and continuously generative of research. AU in all, resevch on unconviour prm- cwa in understanding, and judging orhen shows how intelligent but hllible v tems operate within the constraints of the cognitive a r c h i t w e rhat evolution a d learning allows and the demands of daily social life.

social group ._ . I-

perhaps the most distinguishing mark of social psychological research on unconscious proc- esses is its interest in the social group as a legitimate unit of analysis. In the previous sec- tions, we discussed how judgments about individual personality qualities can arise fiom unconsciously perceived sources. In this section, we discuss research on the unconscious activation of stereotypes and their application in judgments of individuals and groups. In the next sectionon unconscious a i k , we will review related research on attitudes of preju- dice that reside outside of conscious conuol or awareness.

As has often been argued, stereotypes about social groups arc heuristics that simplify and organize perception of the social yorld. In so doing, M i d i about social groups and their use in individual judgment merely reveal ordinary processes of learning and generalization. Our discussion of these. parriculv short-cuts will show the various ways in which uncon- scious processes r d their presence. Our discussion will also point out the moral ques- tion that emerges from this rather ordinary discove'y about catcgory learning, generalization, and inferences. Stereotypes exact a toll by subsuming individuals into the larger social categories and by giving to individuals privileges and punishments that are not their due. We noted previously that social judgment may not reflect the actions of the target but of unconsciously applied consum in the perceiver's mind It is unsettling, at least in socie- ties that consciously & that judgments ought to be based on the "content of one's character," to discover the extent to which judgments of individuals may reflea beliefs about their d group. -- Unconsriovs -types atr rooted in s d c a q o h d o n A rich literature on social catego- r k a o n proccsJes indiatcs that such processes arc automatidy prompted by the mere presence of a stimulus rugcr (Banaji & Hudin, 1996; Blair 8 Banaji, 1996; Blascovich, Wycr, Swan, & Kibler, 1997; Brewer, 1988; Eckes, 19%: Fis& & Neuberg, 1990; Ford, Smgor, & Dun, 1994; Hvnilton & Skrman, 1994; Perdue, Dovidio, Gurunan, & Tyler, 1990; Pendry & Macrac, 1996; Sangor & Lulgc, 1994; Strocssner, 1996; Z h t e , Bonilh, & -0, 1995). And in the mind of the social puceiver, stereotypes that ac- company a particular category automatically accrue to its members.

hconknu is wbisyitous In an i d 4 demonstration of unconscio~s nucayping, Dcvinc (1989) found that subhmindy pncra* race idbrmation i d * d h o w pvticipano subscqudy judged the pmbipw behnior of a ncr-ulls@ed ~ B Q Q ~ . n d l o r - p r r i u d i d p ~ i d ~ n t c d r b r ~ r l p ~ m o ~ h o n i l e w h m

chey had been presented with a list containing 80 percent stawrypicllly black words (e.g. jazz, basketball, Africa) than whm the list contained only 20 perourt stereotypically black words. Devine's evidence that rtercotypcs could be mornat idy activated by presenting cues about a stercoryped group inspired research on how stereotypes operate without con- scious awaxcness, conuol, and intention (see Fiskc, 1998, and Grunwald & Banaji, 1995, for reviews). This body of research provides strong cvidmcc that beli& about social groups are readily activated, and influence pctccption of the cugct. What's more, the tcscarch suggests that unconscious processes not only ficilimte stereotyped responding but also inhibit counterstereotypical associations, perhaps maiung stereotypes additionally resistant to changing in the fie of atypical group exemplars (Trope & Thompson, 1997; Van Knippenberg & Dijksterhuis, 1996).

Gender, as a category, has received much attention, in part because of its fundamental nature and presence in all human societies and in part Eor its convenience in not attracting attention to social category as the focus of the experiment (Banaji & Greenwald, 1995; Banaji & Hardin, 1996; Blah& Banaji, 1996; Dunning & Sherman, 1997; Lamben, 1995; Macrae, Bodenhausen, Milne, & Thorn, 1997; Nelson, Adrer, & Manis, 1996; Pratto & Bargh, 1991). This research has shown that information about one's gender, - -. whether conveyed through names, pictures, or gender stereotypical words, exerts an un- conscious influence on judgment. For example, Banaji and Grtcnwald (1995) found that more male names than female names were identified as hmous under conditions &memory uncertainty, suggesting that stereotypical beliefs about fame were implicitly applied in assigning fame to people. In other research, using traditional semantic priming proce- dures, participants were found to more quickly idenufy male and female target names (Blair & Banaji, 1996) or pronouns (Banaji & Hardin, 1996) when the names matched the gender stereorypicaliry of the primes than when they were incongruent with the primes.

Support is also found for the unconscious operation of race stereotypes (Bodenhausen, Schwarz, Bless, & W a c , 1995; Dovidio, Evans, & Tyler, 1986; Gilbert and Hixon, 1991; Glaser & Banaji, 1998; Kawakami, Dion, & Dovidio, 1998; Lepore & Brown, 1997; Spencer, Fein, Wolfe, Fong, & Dunn, 1998; Von Hippel, S e k a q y t m , & Vugas, 1997; Wittenbrink, Judd, & Park, 1997). These studies show, for example, that race stere- otypes are easily activated upon encountering members of stereotyped groups (e.g. Gilbert & Hixon, 199 1; Gaertner & McLaughlin, 1983; Lepore & Brown, 1997). Other research has indicated that activating unconscious stereotypes can influence not only individuals' judgments of others but also their overt behavior (Bargh, Chen, & Burrows, 1996). Such experiments starkly reveal that perceivers may have less control over the knowledge they use in social interaction than they or even the scientists wh&udy them may have as- sumed. When knowledge about the social groups to which one belongs enters into the equation of social judgment early and with force, it can shape the cumulative record of social interaction without the hindrance of awareness or hence responsibility.

Is unconscious stereotyping umvoidablr! Despite the preponderanci of evidence that un- conscious stereotypes hold a tight grasp over everyday thinking, the extent to which they are related to explicit beliefs and attitudes, the circumstances under which they are activated, and ,- the . dsgree to which unconscious stereotypes a n be brought under deliber- ate control remain less certain. The question of the relationship between conscious and

The Social Unconrciour 143

unconscious measures emerged early. Are those who hold weaker forms ofconscious stere- otypes also likely to evidence weaker forms of unconsciou~ stereorlpa?-lfly a m p a to

this question suggested that unconscious stereorypa, ass& indirectly by cum- ining nonverbal behavior, social perception, memory, and speeded reactions to social nimuli, are ofien unrelated or only slightly related m explicitly expressed stereotypes assessed by self-report measures dBmaji & Greenwald, 1995; Banaji & Hardin, 1996; Devine, 1989; Dunning & Sherman, 1997; Gaertner & McLaughlin, 1983; Hense, Penner, & Nelson, 1995; Von Hippel, Sekaquaptewa, & Vargas, 1997; Wittenbrink, Judd, & Park, 1997). However, there is also a body of recent evidence that suggests the convary (Augousrinos, Innes, & Ahrens, 1994; Dovidio, Brigham, Johnson, & Gaertner, 1996; Hense, Penner, & Nelson, 1995; Kawdwni, Dion, & Dovidio, 1998; Lepore & Brown, 1997; Locke, MacLeod, & Walker, 1994; Wittenbrink, Judd, & Park, 1997).

Lepore and Brown (1 997, 1999) argued that individual differences in consciously ex- pressed prejudice should predict unconscious stereotyping. Using a procedure similar to that used by Dwine (1989), Lepore and Brown subliminally primed high- and low-preju- diced participants either with waluatively neutral words that connoted the social category Blacks (withoutsonnoting particular stereotypes), or with nonsense syllables. Then, par- ticipants read behavioral descriptions of a race-unspecified target person and rated the target on a number of traits stereotypic of Blacks. Participants who had scored high in prejudice against Blacks employed more negative stereotypes and fewer positive stereotypes in the prime condition than in the no-prime condition. In contrast, low-prejudiced par- ticipants used more positive stereotypes in the prime condition than in the no-prime con- dition, but showed no difference on the negative d e s . Lepore and Brown concluded that when race category information is primed but race stereotypes are not, unconscious race stereotyping is contingent upon how much one explicitly endorses prejudice.

Recently, Kawakami, Dion, & Dovidio (1998) proposed an additional explanation for the miuky.relationship between implicit and explicit beliefs. They noted that even in re- search in which implicit and explicit measures are associated, the association is relatively

. - weak, and proposed that highly sensitive procedures may be necessary to pick up relation- ships berween implicit stereotyping and explicit beliefs and attitudes. In addition, Cunningham, Nalak, & Banaji (1999) have shown that a general ethnocentric personal- ity disposition is related to specific unconscious prejudices (toward foreigners, Black Ameri- cans, the poor, Jews, and gays). Such efforts represent initial strides in identifylng the conditions under which implicit and explicit beliefs converge and diverge, by identifylng methodological, statistical, and theoretical hurdles that need to be set aside before a more complete picnuc regarding the relationship between conscious and unconscious stereotypes or prejudice may be observed.

Can unconrrious stereotypes be controllrd? The controllability of unconscious stereotypes has sparked considerable theoretical debate and empirical research. There is abundant evi- dence that s t e r e o w that operate unconsciously defcnd their territory fiercely, influenc- ing social interactions even when perceivers are consciously vigdant and motivated to defeat them (Banaji & Greenwald, 1995; Bargh, in press; Blair & Banaji, 1996; see Greenwald & Banaji, 1995). Indeed, conscious attempts to purge stereotypic thoughts can easily back- fire, bringing stereotypes to the fore with redoubled force (Macrac, Bodenhausen, Milne,

& Ford, 1997; Nelson, Acker, & Manis, 1996; Ndson, Bicmat, & Manis, 1990, Sherman, Suocssner, Infnrs, %r Deguzmul, 1997). Bargh (in press) has proposed a metaphor to characterize unconscious stereotyping, comparing it to a monster whose influence annot be restmined once it is set into motion. The solution may Lie in rnotktcd individuals' ability to devdap, over h e , chronidy accessible rgalitr3rian be&& that can counter the effects of unconscious stereotypes.

Our assessment of the issue of controbg automatic processes is in line with Bargh (in press). When a process operates unconsciously, there is little, if anything, that can be done to react , revoke, or rescind. If this message from basic rcscvch on unconscious stere- otypes is to makc contact with the world it seeks to improve, the responsible suggestion at the present time is not the simplistic one to "just say no." Automatic stereotypes can and will influence perception, memory, and judgment. If the goal of judgmg individuals by the content of their character is one that this and other societies wish to take seriously, this body of social psychologid research suggests two radical strategies. First, aeate the social conditions that allow n m associations and new leaning about s o d groups that blur the bright line that demarcates socia groups. Second, generate individual and groupbased strategies for compensation in conscious recognition of the stark and pervasive uncon- scious biases that operate in social judgment.

Unconscious Affect

Whereas no uncertainty is expressed about the existence of an unconscious form of cogni- tion and whether it can be reliably prsasCd, there is still active debate regarding the existence of unconscious affect (Clore, 1994; Clore & Kctelaar, 1997; Kihlstrom, Mulvaney, Tobias, & Tobis, in press; LeDoux, 1994, 1996; Zajonc, 1994, 1998). Research on unconscious &ea (and related concepts better recognized by the labels emotion, evaluation, attitude, and prejudice) has aquiml increased prominence in social psychology in part from the desire to provide more complete models of social behavior, and in part f;om the availability of tractable methods to measure these warm and wet constructs. Perhaps a rigorous analysis of unconscious affect is naturally located in social psychology because of the field's long- standing interest in constructs that tap feeling, most obviously that of attitude and esteem (see Greenwald & Banaji, 1995). Research on unconscious evaluation, attitudes, and affect has its origins in a multitude of experimental traditions necessitating tough choices regard- ing selection for review. However, no attempt is made here to &inguish between the vari- ous terms that an used to refer to slightly differing aspects of the basic construct. For additional coverage and differing emphases see Kihlstrom, et al. (in press) and Zajonc (1 998).

Pbysiologcal measures of twal~~~t ion and attitude

Among the rewns to probe evaluation and attitudes in their physiological form, Gcioppo, Critcs, Gudncr, & Bernuon (1 994, p. 12 1) offer the following rationale: "Unfortunately, the attitudes that individuals are last willing to repon arc often those that are most impor-

The Social Unconsciow 145

ant to measure accurately, as they may differentiate individuals along &ret idy impor- dimension*" Autonomic measurer of unconscious evaluation were initidy l y d

d& hope, but such assessments hiled to separate intensity and valence of amrude (xe p e w and Cuioppo, 1983; Zanna, Dmeiler, 8L Olson, 1984). F a c l EMG rerponss have dso b a n obtahed ( h i o p p o , MaI'dcC, Petty, dr T ~ S S ~ U Y , 1988) but +, their diudvanqCS have been noted, including the inability of such m a s u m to protea -s

and distortion (see Gcioppo, Petty, Laxh 8c Kim, 1986). Recently, are- menu in which a late positive potential (LPP) of the went-related brain potential (Em)

related to evaluative categorization have been reported (Cacioppo, Crita, Bernrxm, & C&S, 1993). Funher, the amplitude of LPP incrmses as a b c t i o n of the mismatch be- meen evaluative catqpkation and expectation of nnluarive significance through salient c o n t d cues (Gcioppo a al., 1994). The amplitude of LPP is larger when a n-tive (rather than positive) attitude stimulus is presented within a sequence of positive stimuli, and such measures also appear to show sensitivity to intensity of negative stimuli. The &vious utility of such a masure to provide a marker of individual di&ences (e.g. feu rtsponses to rocid situations implicated in phobia or negative responk to members of

groups as-revealed in prejudice) will be realized in fume experiments to test the construct and predictive validity of the measure.

Sensory-motor ptoccsses in evdkztion and attiflcdc

Evidence suggesting the involvement of motor processes and their sensory consequences in attirude formation merits attention here, because it points to y a another path that can reveal the role of unconscious evaluative pmcesscs in social cognition. Wells and Petty (1980) showed that the motor action of shalung versus nodding one's head while listening to persuasive messages resulted in lesser or gram agreement with th message. Likewise, Strack, March, & Stepper (1988) showed th motor activity that fiditatcd smiling in- creased rating of the humot of cartoons compared with slightly diff;cring motor activity that inhibited s m d q . Such &&a emerged in spite of subjects' being unaware of the meaning of the contractions of the ygomaaam musde. Other research supports these findings that manipulations of kcid expressions create ;Bxttvc states or influence amtudinal responses outside conscious awareness (Martin, Harlow, & Smck, 1992; Zajonc, Murphy, & Inglehut, 1989). Stepper & Strack (1993) have shown that proprioceptiw cues from body posture (upnght v m slumped) can influence the afkrive experience of pride, just as it can influence nolllffl#tiw judgments of dl& and Fiirster and Strack (1996) have shown that h a d nodding versus shaking increases memory for dcncc-consistent words. A disrinaion has beto proposed between q m m d knowiadge, in which "fwling ut 'immedidy given' to the individual and have a dihctphenomenal quality" (Stepper & Strack, 1993, p 2 1 8) v m nocrir representations which-r&cet inf;cmd indirect knowl- edge, with the former being impliated in the i a f ; o d o n that is obtained h m bodily posture o t kcid apression without conscious -esr F i e , there is suggestive evidence that stmatic manipulatioru involved ia prm &rion v a n u arm acemion a n bzvt s d but reliable e f k s on the dup t ioas of m i d i d l y nausai stimuli such as idcognphr, lading the autho~oo conch& thatu-

. . &us involve active motor

pmnter and thnt a pcnon docs not need to know &e duativc or motivational rignifi- a c e of thc motor proccss f$r it to have attitudinal cffcas" (Gcioppo, Priester, & k c s o n , .

1993, p. 16). This intriguing research n d to be n d and d d o p a d further, for it has the potential to infbrm about the role of unconscious processes in a most fundamental association of body and mind, and the potential for the products of such unconscious operation a> influenot feeling and social behavior.

With the increased usage of indirect measures ofpercqtion and mnnory, a welcome blurring of the sharp distinction between these two processes has occumd. Viewing unconscious processes of memory and perception as they inform about the nature of &ective experience and expression has contributed to a broadening of our understanding of consciousness.

Mew exposure Among the most influential ideas linking perception and affect comes from the discovery that exposure to a stimulus leads to enhanced liking for it (Zajonc, 1968). There have been over two hundred published experiments testing this hypothesis (Bornstein, 1989; Bornstein & D'Agostino, 1994) that verify the reliability and robustness of the basic effek across a wide variety of stimulus forms, dependent variable formats, methods of exposure, and experimental settings. The finding that mere exposure produces liking has also been extended to research on interpersonal interaction (Bornstein, Leone, & Galley, 1987). The literature on category accessibility (see previous section on uncon- scious cognition) showed the peculiar effect of awareness on unconscious thought and social judgment (i.e. the influence of the priming m n t is most visible when that event is least available to conscious recollection). Rcsearch on the mere exposure effea has pointed up a p d e l finding regarding unconscious affect: the magnitude of the effect is greater under conditions of subliminal rather than supraliminal exposure (Bornstein, 1989, 1992; Bornstein & D'Agostino, 1992). Theoretically, the mere ucposure effect has shifted from king considered a phenomenon unique to the expression of affect t o n e that most pani- moniously fits into the broader landscape of familiarity and its effect on judgment more generally (Bornstein & D'Agostino, 1994; Jacoby & Kelley, 1987; Jacoby, Toth, Lindsay, & Debner, 1992; Mandler, Nakarnura, and Van Zandt, 1987). Whatever the interpreta- tional leaning, the mere exposure effea will remain among the most important discoveries of twentieth-century psychology. Here, its importance is in having identified a dissocia-

-c-C. -'

tion between what is consciously known and what is unconsc~ously felt.

Automatic eva1wtion In the early 1970s the discovery was made that meaning is automati- cally activated upon the mere presentation of a word (Meyer & Schevaneveldt, 1 97 1 ; Posner & Snyder, 1975). Efforts to resist activation of default meaning are moot when conditions do not permit the exerting of conscious control (Neely, 1977). The evidence to be reviewed here pertains to the finding that just as semantic meaning is extracted auto- matially upon presentation of a word, the evaluative meaning of information is also grasped without conscious control. Fazio, Sanbonmatsu, Powell, & Kardes (1986) showed that judgments of a target were ficiliuted when its valence was congruent rather than incon-

The Social Unconsciow 147

f3" ent with that of the prime. Bargh, Chaiken, Govender, & Prano (1992; B&, Gun, Raymond, & Hymes, 1996) replicated and mended this finding, additiondy ppmpoSing hat automatic evaluation occw regardless of the strength (extremity) of prime d n c c , a claim about which there is debate (Chaiken 8~ Bargh, 1993; Fazio, 1993), and m n in [he absence of a focus on rhc wahative properties of information in judgment. Glwr & Banaji (1998) have reported a series of studies in which contrast effects in automatic evalu- ation appar when primes are of extreme valence, and they interpm this finding as a automatic correction for the implicitly perceived biasing influence of the prime.

In an effort to test the reliability, robustness, and boundary conditions of automatic evaluation, Greendd, Draine, & Abrarns (1 996) and Grccnwald, Klinger, & Liu (1 989)

showed that-the evaluative meaning of words is automatically registered by pre- senring the prime subliminally. In their most recent research they did this by inventing a

of the technique called the "response window" that reliably produces the cffea. this research reveals, experiments have relied on time (measured in milliseconds) to

respond to the target as an indicator of automatic evaluation. A second prpcedure has also been used in which evaluative primes, usually in the form of evaluative ficial expressions, are briefly flashed (on the order of 4 milliseconds to prevent conscious registration) follow- ing which a neutral target (e.g. a Chinese ideograph) is to be rapidly judged. The replicated finding is that judgments of the neutral stimuli shifi in the direction of the evaluative position of the prime (Murphy & Zajonc 1993; Niedenthal, Setterlund, & Jones, 1994). Pratto & John (1991) used a Sump color-naming task with evaluative stimuli in place of color names to extend the generality of automatic evaluation, showing that automatic evaluation can intedere with a conscious cognitive msk.

Together, these experiments on automatic evaluation have changed our thinking about the existence and tractability of unconscious affect. First, they demonstrate that the afftctive quality of information registers without conscious awareness of the stimulus (as in the sub iirninal presentation studies) and without conscious conuol over the response (Murphy, Monahan, & Zajonc, 1995). Second, it appears that automatic evaluations are sensitive only to gross distinctions of polarity and not to anydung that can be considered to be a more fine- grained evaluative assessment (see Prarto, 1994; Zajonc, 1994). Finally, although the experi- ments have examined unconscious perception and memory for evaluative material, they have been interpreted as revealing an attitude. This is a noteworthy shift in social psycholo- gy's understanding of the concept of attitude. In commenting on research on unconsciously activated attitudes, Cacioppo, et 4. (1 993, p. 16) which one? note that "Indeed, the day may come when we q u d attitudes as being 'duat ive perceptions' . . . aroused by stimuli." Becaw conceptions of attitudes as necessvily accessible to conscious awareness and control arc difficult to shake off, the rcsearch summarized here will comc'to be viewed as historically important - as the first robust and reliable demonstrations that permitted a sufficient break- though to d o w us to conccptuah attitudes as automatic evaluations.

Jhpcrs who study attitudes and beiidi toward social groups haw anphasztd the need to trut attitude (prejudice) with the same irnporunoe as has been accorded to belief

(ctatotypc), urd hnn reEiRcd rhe raaging ofthue two consaua;r b t h in theory and in apcl;imcnal prPaicc In put, this deEke iaas armmrd born the conviction that attitudes

1 of prejudice reprc~cnt r aniqueanci rcpunblcauqxmau &om stereotypic bclicf$. The organhtion of rhir dqter d m rhu distinceion ro aontinurt eves at thc c x p s c of rcpvocing rcsevah on u n d prejudicthm ita cousin, 111~~)nscious scercocypes (see previous d o n ) . Jum us she study of un~~nsciout cognition has m e i d greater attention &an the study of unconscious p&a, cht pdld OO~UPUCXS of ul~~onsciou~ i e r d b ~ and unconscious prejudice have received rimillrly&mdpl traanent. For evidence of this, ree the greater coverage dowed zo rcscvch on ubconscious ttemtypes c a m p a . with unconscious prejudice in a rcccnt rcvicw (Fiskc, 1998). Even since that review, however, -tion to the study ofuhoonscious prcjdce has in& W y from su;ligh&rward actensions of rcchniqws to study automatic d tudcf mon g e n d y . As with the study of srereocypes, such research is already challenging aaxptcd notions of h t prejudice means and raising troubling questions rcgarbg the impliations f6r how to regard human nature and human nurture (see Banaji & Bhaskar, 1999).

Indirect measures of prejudice have been dimrest Cbr over awo decades (Crosby, Bromley, & Saxe, 1980), with continuing in- in related issues such as physiological - indicators of prejudice (Vanman, Paul, Im, & Millo, 5997; V- & Rollock, 1998), the- '

relationship berween public and private expressions of prejudice ( h b e r t , Cronen, Chasteen, & Lickel, 1996), and the impact of ringlc direct or indirect exposGe'to negative behavior on judgments of groups and members of groups (Henderson-King & Nisben, 1996). Yet a rigorous analysis of the role of consciousness and the disjunction between the unconscious roots of prejudice and its conscious md%S~tionS has only just become pos- sible. W ~ t h methods to measure automatic evaluation and automatic stereotypes in place, it was only a matrer of time before such techniques were used to study applications to prejudice. In fkr, research to show that priming of race stereotypes produced evidence of linking evaluatively positive information with Whirc compared with Black has been avail- able for some time (Dovidio, Evans, & Tyier, 1986), in addition to evidence of a more general liking for one's own group. Associating neutral syllables with "we," "us," 'ours' versus "they," 'them," 'thein" p r o d u d greater Wring for syllables attached to ingroup compared with oucgroup primes (Perdue, Dovidio, Gurunan, & Tyler 1990).

With the fitst publication in 1998, the Implicit Association Test (MT; Greenwald, M c G k 8r Schwartz, 1998) has already a t u a a d anention as a measure of automatic asxrciation, most notably in the investigation of automatic evaluative associations toward rocid groups and self. The attraction of the method lies in w o of its most salient proper- ties: (a) the ability to obtain large e&cu compared with priming methods of automatic dua t i on , and (b) the ability to compehngly reveal a lack of control over automatic cvdu- arive associations. Like related measures of automatic association (e.g. semantic priming) the t e h i q u e is based on the assumption that if two concepts have come to be associated in memory, they will be associated more quickly when they are encountered. The IAT procedure operationalizes this assumption by requiring participats to mifdy wocia~e "mplm of categoria such as "oldn and "young" along with exunpLn of the evaluative category 'bad" or 'unpleasant." The sped with which old-ood and young-bad are clas- sified compued to the speed with which old-bad and young-good are classified produces a robust mdrure bf the relative automatic evaluation of young and old. The original re-

The Social Unconscious 149

sear& demonstrated that the method is upable of detecting robust positive automadc toward flowers compared with insects, toward White compvcd Bla&

~ ~ ~ r i c a l l S (among non-Black subjms), and automatic ingroup positivity mong Korm md Japan= Americans. (Repom of ongoing explorations with the technique are av&le at ~ . ya l e . edu / imp l i c io or <depuwashington.eddia~, showing the wide appliuGon ofthe technique to investigate the attitudinal basis of depression or smoking, and attitudes toward a variety of social groups, e.g. TurkslGermans, JewslChristians, EastIWest Ger- mans, old/young, omnivore/vegetarian, male/female, overweighdthin.)

Research with the technique has explored attitudes toward self and social groups: fe- male/male, fernininelmasculine, or female leaderlmale leader, and the relationship between &-identity and g e n d c ~ i t u d e (Carpenter & Banaji, 1998; Lemm & Banaji, 1998; Mitchell, Nosek, & Banaji, 1998; Rudman & Glick, 1998; Rudman & Kilianski, 1998; Rudman, Greendd, & McGhec, 1998); attitudes toward mathlscience versus arts and the relationship among automatic gender identity, gender stereotypes about mathlscience, self-math identity, and performance (Nosek, Banaji, & Greendd, 1998); race identity, groupesteem, and self-esteem (Rosier, Banaji, & Greenwald, 1998); k' iated attitudes toward multiply-sategorizable objects (Mitchell, Nosek, & Banaji, 1998); attitudes regard- ing age, nationality, and religion (Rudman, Greenwald, Mdot, & Schwam, 1998); and the role of personality in automatic prejudice (Cunningham, Nezlak, & Banaji, 1999). New designs for research have been suggested, based on a unified view of social cognition that draws on consistency theories (especially the Heiderian notion of balance) and associationist networks (Grcenwald, ct al., in press). However, several questions regarding its construct validity are only k n n i n g to be addressed. The theoretical questions of ut- most interest concern the predictive validity of this and other measures of automatic asso- ciation (see Fazio, Jackson, Dunton, & Williams, 1995; Bessenoff & Sherman, 1998), developing measures of motivation to control prejudice (Dunton & Fazio, 1997; Plant & Devine, 1998), the relationship between automatic and controlled prejudice (Dovidio, Kawakami, Johnson, Johnson, & Howard, 1997; Kawakami, Dion, & Dovidio, 1998; Lepore & Brown, 1997; Von Hippel, Sekaquaptewa, & Vargas, 1997; Wittenbrink, Judd, & Park, 1997), and the malleability of automatic evaluative associations (Carpenter & Banaji, 1999; Dasgupta & Greenwald, 1999). Research on unconscious forms of prejudice elicits attention, in pm because it sp& to a problem of great social significance. Because of this, and because of the potential to challenge many assumptions about the propensity to create harm without intention and awareness, this research requires the attention of a diversity of methodological and theoretical perspectives.

Conclusion

&ine with consciousness have the luxury to speculate that their own mind and behavior m y also operate in a strikingly different mode, detached from consciousness. For hun- dreds of years, lay people and a p e m have believed that not only is t h m a mental world that remains hidden fiom consciousness, but that the workings of this world have impor- tant and far-reaching consequences for understanding who we are and who we aspire to be.

Ya ir is only in* lrstbundd ycus, beginnrng with experiments on Btsmau and other a n i d s , th?t areicncedahemamsciousum atccmpd a n d d in chif chapter, r v t a ~ d c d o o & e ~ r k o f t h ~ ~ e r s h o g r o ~ P h t i r i a v ~ ~ i n & e ~ world, the ether in which mental Lih: opaptcb.

T h e ~ u m p t i a n ~ h ~ ~ b c h z v i o r c o n l o n l y b e u n d d b y ~ c h ~ s ~ +le of LnguP%c to say, phf;cmbly in grammatical Engiish, what Ithey zhink, kl, and intend to do about tbanschzs and~odms in their world is a Limiting one. In the last two decades, d psychology has shown thc advances that are possible when such an issump- don is mom^ set aside. In another oontext, we made the point tbat it is not difficult to imagine why it is that social perceivers and social psychologists h?vc trouble inugining and investigating those processes that lie outside conscious awareness (Banaji, Blair, & Glaser, 1997). We argued that d e n the soura of an action emanarcs in h e and space unconnected to the observed aaion, it is difficult to p p the conncaion between the source and target of influence. It took Newton's genius to discover dm W t , a source unattached to physical objeces, was responsible tbr producing the subjaaive experience of color. Likewise, sources of iduencc on thoughts, &kt, and motives arc not iikly to be discerned easily because their causes lie in places that are unreachable by conscious aware- - ness. In addition, as we observed, even under conditions that pumit awareness, the ability to wnuol thoughts, feelings, and motives may be weaker than ?ssumd The problem here is more complex than contemplating an undemding of the physid world; for unlike the physical world, the object of inquiry (unconscious mind) is a part of the thinking system that must wndua the inquq. The limits on being able to look inward are serious, and here the social world offers a solution for theory and praxis: a rich array of events, situations, and o p p o d t i e s to explore the manner in which unconscious processes oper- ate, in contexts in which they have stgruficant impact on happiness, liberty, and justice. It is perhaps the case that as we discover the extent to which unconscious processes control social thought, f&ng, and behavior, we will arrive at a fuller appreciation of the unique role played by consciousness in a species with the capability to evaluate the nature of the social unconscious.

Andersen, S. M., & G h m a n , N. S. (1996). Responding to sigruficant others when they are not there: Effects on interpersonll interference, motivation, and affect. In R M. Sorrentino & E. T. Higgins (Eds.), Handbook of motivation andcognition Vol. 3 (pp. 262-321). New York Guilford Press.

Augoustinos, M., Imes, J. M., & Ahtens, C. (1994). Stereotypes and prejudice: The Australian experience. British j o u m l of Social P&ology, 33, 125-1 4 1 .

Bddwin, M. W. (1994). Primed relauonll schemas as a source of seLf-evaluative rcaaions. J o u m f of & & f a d Clinical PsychobgyJ 13.380403.

Bdd*nJ M. W., Carrell, S. E., & Lopez, D. F. (1990). Priming relationship.schemas: My advisor a d Pope are watching me from the back of my mind. Journal of Eprimmtaf Sorial Pychof- ogy, 26 4 3 5 4 5 4 .

Bmaji, M. R, h Bhaskar, R (1999). Implicit stereotypes and memory: The bounded rationaliry of IodJ Mefr. In ?. L. Schacter and E. Scarry (U.), Bdi$andmmroty. Cambridge, MA: H w u d University Pks.' *

The Social Uncon~rious 1 5 1

M. R, B h , I. V., & G b , J. (1997). Environments and unconscious pr-. In R S. Wyer (Ed.), Tk drrto*tin'ty of eyydzy h$i: Advances in s o d cognition, VOI. 10 (pP. 6>74). Mahwah, NJ: Erlbaurn.

~anaji, M. R , Hudin, C., & Rothmul, k 1. (1993). Implicit stereotyping in person judgment. ~ O U & ofPmonrrlity and Social PsycbobgyS 65,272-28 1.

Bar&, J. A. (1997). The automaticity of everyday lift. In R S. Wycr, Jr. (Ed.), The arrtomrtfity of w h y h?: Advances in s o d cognition Vol. 10 (pp. 1-6 1). Mahwah, NJ: Lawrence Erlbaum Associates.

Bar&, J. A (1999). The cognitive monster: The case against the conuollability of automatic stereotype effects. In S. Cidccn & Y. Trope (Eds.), Dvrrlpmcess thcorics in socirzlpsych~b~. New York: Guiiford Press.

Bar&, J. A, Chen, M., 8r Burrows, L (1996). Automaticity ofsocial behavior. Direct effects of trait construct and stereotype activation on action. J o u d of Pmonakty and So& Psychology, 71, 230-244.

Bargh, J. A , Chaiken, S., Govender, R , & Pntto, F. (1992). The genaality of the automatic attitude activation effect. journal of Pmonafiy and Sorial PryhlOgyB 62,89>9 12.

Bargh, J. A, Chaikcn, S., Raymond P., & Hymcs, C. (1996). The automatic evaluation effect: Unconditional automatic attitude activation with a pronunciation task Journal of l3pmrnrncntal Social Psycbob~, 32, 104-1 28.

Beck, A. T. (1 976). Gpitiw &aPI urdthc -tioddiror&rx New York Penguin BooUMerid- ian.

Bessenoff, G. R , & Sherman, J. W. (1998). Automaac and controlled components of prejudice toward the overweight Evaluation versus stereotype &tion. Unpublished manuscript, North- western Univasiry.

Blair, I. V., & Banaji, M. R (1996). Automatic and controlled pmcews in stereotype priming. Journal of Pnronality and S o d Psybobgy, 7 0 , l 142-1 163.

Blucovich, J., Wyer, N. A, Swart, L A , & Kibler, J. L (1997). Racism and racial categorization. J o v d of P m o d t y and S o d Psychabg, 72, 1 364-1 372.

. - Bodenhawn, G. V., Schwuz, N., Bless, H., & Winkc, M. (1995). E.fkts of atypical exemplars on racial belid%: Enlightened racism or gencnlized appraisals? Jovmal of &erimmtal Social PychoG ogy,31,48-63.

Bornstein, R F. (1989). Exposure and & Overview and m_+-adys~~ of resevch, 1968-1 987. PsycholOglCd B&n, 106 265-289.

Bornstein, R F. & D'Agostino, P. R (1992). Stimulus recognition and the mere exposure effect. JovmalofPmonrrlity andSocial~hoIogVB 63,545-552.

Bornstein, R F., & D'Agostino, P. R (1994). The attribution and discounting of perceptual flu- ency: Prekrmnvy tcsa of a puocptuaJ flwncyiattribuaod modd of the mere exposure effect. Social Cognition, I2,10%128.

Bornstein, R F., & Pitanan, T. S. (1992). Pcrrrpih without awramrrr: Cognitive, cl ink4 and s o c i a l p m p ~ ~ w . New York: Guilford Press

Bornstein, R F., Leone, D. R, & Galley, D. J. (1987). The gamahbility of subbind mere acposurr e&cts: Mmcc of stimuli p e r d witbout awareness on social behavior. Journal of Pmonaliy and Social Psycho&gy, 53,1070-1079.

Brewer, M . B. (1988). A dud ptooas modd of impression fbrrmtion. In 'FL K S d & R S. Wyer (Eds.), A&ace in JaciJ @tier, V d 1 (pp 1-36). HilLdrk, NJ: Erlbaum.

Burnstdn, E, & Schul, Y. (1982). The hfbnrmtiod bash of d judgmcnp: O p t i o n s in forming an impression of motha p a s o n . J m n n J O f ~ d S o c i J ~ , 18,217-234.

Gdoppo, J . T., P r i e ~ f ~ , J. R, & Bemoon, G. G. (1993). Rudimcnwy c i a a d x m m of attitudes

11. Arm fkrion and atension have di&rential e&as on attitudes. /d ofPCrUrmilizy and Social Pphr (m, 65,517.

CMoppo, J. T., Giter, S. L, & n o o n , G. G., & Coles, M. G. (1993). If attitudes rffict how stimuli ue p r d , rbwld thy not &kt the event-dated brain potential? Prychbbgd Sci- mcc, 4, 108-112.

Gcioppo, J. T., Critcs, S. L,Gardna, W. L, & &moon, G.G. (1994). Bioelearial echoes from evaluative categorizntions: 1. A late positive brain potential that vvia as a function of a i r nega- tivity and extremity. j o u d of P m n d r y and S o d PIycho&gy, 67,115-125.

Gdoppo, J. T., Muake, J. S., Petcy, R E., & Tusinnry, L G. (1988). S p d c forms of kid EMG response index emotions during an intaviw. From Dvwin to the continuous flow hypothesis of &kt-laden &mation pfoccssing. Jd ofPcrs0naldy and Social pslrcholog)l, 5 4 5 924O4.

Gcioppo, J. T., Petry, R E, Zbsch, M. E, dr Kim, H. S. (1986). Elec~omyognphic activity over fici;L1 muscle regions a n difirentiate the valence and intensity of &ective reactions. / o u d of Pmonality and S o d *tholog: 50,260-268.

Carlston, D. E., & Skowronski, J. 1. (1994). Savbgs in the relearning of tnit information as evi- dence for spontaneous inference geaention.Jonmal of Pnronality a d S o d Pzychobgy, 66,840- 856.

Carlston, D. E., Skowronski, JJ., & Sparks, C. (1995). Savings in deuning: 11. On the formation of behavior-based trait associadons and infmnces. jonmal of P d t y and S o d Psychology, 69, 420436.

Cupenter, S. J., & Banaji, M. R (1998). Implicit attitudes and behavior toward f a a l e lead&. ' Paper presented at Midwestern Psychologid Association mating, Chicago.

Chatken, S., & Bugh, J. A. (1 996). Occurrence vmus moderation of the automatic attirudc activa- tion effect: Reply to Fazio. Journal of Pmonality rurd Sorial P~ychology, 6X 759-765.

Clore, G. L. (1994),Why emotions arc never unconscious. In P. Ekman & R J. Davidson (Eds.), The nature ofmotion: Funrhmmzaf qumions (pp. 285-1 90). New York: Oxford University Press.

Clore, G. L., & Ketch , T. (1997). Minding our emotions: On the role of automatic, unconscious afFea. In R S. Wyer (Ed.), Thc aut~matin'ty Ofcvyaky hi: Advances in s o d cognition Vol. 1 0 (pp. 105-1 20). Mahwah, NJ: Lawrence Erlbaum Associates.

Croizet, J., & Claire, T. (1998). Enending the concept of stereotype threat to social class: The intellectual underperformance of students from low socioeconomic backgrounds. Pmodty and S o d P ~ y c h b Bdain, 24,588-594.

Crosby, F., Bromley, S., & Saxe, L. (1980). Recent unobtrusive studies of Black and White dis- crimination and prejudice: A literature review. P~chological BuUm'n, 87, 546-563.

Cunningham, W. A , Nezlak, J. B., & Banaji, M. R (1999). The roots of prejudice. Unpublished manuscript, Yale University.

D'Agostino, P. R , 8r Beegle, W. (1996). A reevaluation of the evidence for spontaneous trait infer- ences. journal of4m'mmtal S o d ~ ~ y c h o l o ~ , 32, 1 5 3- 1 64.

Dasgupta, N., & Greenwid, A. G. (1999). Exposure to admired group members reduces implicit prejudice. Paper presented at the annual meetings of the American Psychological Society, Denver, CO.

Devine, P. G. (1 989). Stereotypes and prejudice: Their automatic and controlled components. jour- rial ofPmona& and Social P~ycholo&~, 56 5- 1 8.

Dovidio, J. F., E v a , N., & Tyler, R B. (1986). Racial stereotypes: The contents of their cognitive representauons. j o u d of ~ m m m c n t d Social P~ycbokyy, 22,22-37.

Dovidio, 1. F.9 B r i e , J., Johnson, B., & Gaertner, S. (1996). Stereotyping, prejudice, and dis- crimination: Another look. In N. Macne, C. Stangor, & M. Hewstone (Eds.), Stereotypes and smeotyping (pp. 12761 31%. New York: Guilford Press.

Dovidio, J. F., K a w h , K, Johnson, C., Johnson, B., & ~owuh,x . (1997) . On the nature of prejudice: htomatic and controlled p r o c e s s e s . j o u ~ l o f ~ m m m m ~ S o c i a 1 P ~ c h o l o ~ , 33,5 10- 540.

Dunning, D., *& Sherman, D. A (1 997). Stereotypes and tacit inference. j o u m l of Pmonrzltty and S o d II$ch'u@y, 73,459-47 1.

Dunton, B. C., & F ~ o , R H. (1997). An individual difference measure of motivation to control

The Social Unconscious 1 53

P rcjudiced d o n s . P m d q and S o d Pzycho& Bullm'n, 23,3 16326. w, T. (1994). Explorations in gender cognition: Content and svucture of female de subrypes. S o d cipition, 12 (I), 3740. .. '. -*

F ~ O , R H. (1993). Variability in the likelihood of aummatic attitude activation: Dau d F k cornmenmy on Bargh, Chaiken, Govender & Prano (1992).jo~d0fPm0nrrli~~~d~~

psychology, M, 75S758. Fazio, R H., Jadrsoh,']. R, Dunton, B. C., & Wihns, C. J. (1995). Variability in auto-c

activation as an unobtrusive measure of racial attitudes: A born fide pipeline? journal ofPmo& ity and S o d P ~ y ~ h o b ~ , 69, 1 0 13- 1027.

Fazio, R H. Sanbonrmrsu, D. M., Powell, M. C., & Kardes, F. R (1986). On the automatic activation of attitudes. jounrrrl of Pmonality and Sod Psychok, 50,229-238.

Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook o f d p s y c h o ~ Vol. 2 (pp. 35741 1). Boston: Mcgraw-Hill.

Fiske, S. T., & Neuberg, S. L. ( 1 9 ~ c o n t i n u u m of impression formation, from category-bd to individuating processes: Influences of information and motivation on attention and interpre- tation. In M. P. Zanna (Ed.), Mvanccs in eapmmmentafsorirslpsycboology Vol. 23 (pp. 1-74). New York: Academic Press.

Ford, T. E., Sungor, C., & D u n , C. (1994). Influence of social category accessibiliry and category- associated trait accessibility on judgments of individuals. S o d Cognition, 12 (2), 149-168.

Forster, J., & Sack, F. (19%). Influence of overt head movements on memory for vdenced words: A case of concepdmotor compatibility. Journal ofl)morrcrlir). and Social Psychology, 71,42 1-430.

Gacrtner, S. L, & M c L a w , J. P. (1983). R;lci;zl stcreorypes: Associations and ascriptions of positive and negative characteristia Sorirrl Psycho&gy Qurtrtb, 6 2530.

Gilbert, D. T., & Hixon, J. G. (1991). The trouble of Activation and appliation of stereocypic belid%. J m d ofPmonolity and Social Piychob, 60,509-5 17.

Gilbert, D. T., & KNU, D. S. (1988). Seang less and knowing more: The benefits of perceptual ignorance. Journal of P m d t y and Socrisl Psycboology, 54,93- 102.

Gilbert, D. T., & Osborne, R E. (1989). Thinking baclcward: The curable and incurable conse- quences of cognitive busyness. j o u d of Pmonrsliy and Social Psychology, 551.940-949.

Gilbert, D. T., Giesler, R B., & Morris, K k (1995). When comparisons arise. j o u d ofPmon- d t y and Social Psychologr 69,227-236.

Gilbert, D. T., Pelham, B. W., & KNU, D. S. (1988). On cognitive busyness: When person perceivers m m persons perceived. j m d of P d t y and Social P~hoology, 54,733-740.

> -- Glascr, J., & Banaji, M. R (1998). Asshilation and contrast in automatic evaluation and prejudice. Paper presented at the annual meeting of the Society fbr the Psychologid Study of Social Issues, Ann Arbor, MI.

Gotlib, I. H., & McCann, C. D. (1984). COnstruct accessibility and depression: An examination of cognitive and a f k a h ficton. J d of P m o d t y and S o d Psychology, 93, 19-30.

Graham, S., & Hudly, C. (1994). Amibutions of @ve and nolllggrrssive Afiian-American male early adolescents: A study of consrma accessibility. h b p m m t a l Psychok, 30,365-373.

Grecndd, k G. (1992). New look 3: Unconscious a p i t i o n reclaimed. Anmican Psycholop, 47, 766-779.

Grecnwald, k G., & Banaji, M. R (1989). The d u a memory system: Powerful, but ordinary. j o u d of Pmonahy and SocirJ wb 5 Z 4 1-54.

Grcenwald, A. G., & Banaji, M. R (1995). Implicit social cognition: Amtudes, self'cem, and stucotypes. l%yrholosrrJ RrYiou, 102 (I), 4-27.

Grccnwdd, A. G., & Prathnis, A. R (1984). The seif. In R S. Wya & T. K Srull (Eb.), H a d bool ofsodqninbr , Vol. 3 (pp. 129-178). Hillsd?k, NJ: Erlbaum.

Grecnwdd, A. G., Banaji, M. R, Rudrmn, L A, Funhun, S. D., NOS& B. A, & Rosier, M. (in press). Prologue to a unified theory of attitudes, es,m and Jelf-conccpt In J. P. Fo- (Ed.), F&g and thinkiv Thr rrlr of e in s o d tvgnihm and beibwia. New Yo& b bridge University Press,

Grcenwdd, k G., Dnine, S. C., & Abnms, R L (1996). Thee cognitin markers of unconscious d c a c c k b ~ . Srir#, 283 1699-1 702

Gntnwald, h C., Kknger, M. R, & Iiu, T. J. (1989). Unconscious proceasing of dichopudy & rids. Memy &cpsl.* 17,3547.

Grccnwald, A. G., McGhec, D. E, & Schwacu, J. K (1998). Masuringindividunl di&rtnce in implicit aognitiaa: The implicit Assuiation Ta t J o w d o f P ' d S u d mb.

Hamilton, D. L,& Shcmun, J. W. (1994). S t e ~ . In R S. Wyer, jr., & T. K - S d (Eds.), Handbobk .fsDciolcognition Vol. 2 (pp. 148). Hillsctle, NJ: Erlbaum.

Huxlcrson-King, E. I., & Nisbat, R E. (1996). ha-black prejudice as a funaion of exposure to che negatiyc behavior e f a angle black puson. J o u d of Pmnality and S& ~ b g y , 71, 654-664.

Hense, R L, Pennu, L A, & Ndson, D. L. (1995). Implicit memory b r age nntotypes. Social Cognirion, 13 (4). 3 M l 5 .

Herr, P. M., Sherman, S. J., & Fazio, R H. (1983). On the conscquenoa of priming Assi.rnilation and contrast &m. J o u d of ErpmmtntrJ Social p s v t h o ~ , +9,333-340.

Wns, E. T. (1989b). Knowledge lcmsibiity and activation: Subjeaivity and suffering from unconscious sources. In J. S. Ulanan & J. A. Bvgh (Eds.), Uninanrdcdhmgbt (pp. 75-123). New York: GuiVord Prcss.

Higgins, E. T. (1996). KnowIadgt activation: Accessibility, appliability, and nlience. In E. T. Higgins &A. W. Krugknrki (Eds.), Socialpsychology: Handbook ofbaricpn'mpks (pp. 13S168). New York GuiVord Press.

Higgins, E. T., Bargh, J. A, & Lombudi, W. (1985). The nature of priming &cts on ategoriza- tion. journal of Erpmmcntrrl Psychbgy: Learning, Mewwry, a d +ition, 11.5949.

Higgins, E. T, Rholes, W. S., & Jones, C. R (1977). Category accessibility and impression forma- tion. joumal of Erpmmmtntal Social P'cho&gy, 13. 14 1-1 54.

Higgins, E. T., Straurnan, T., & Klein, R (1986). Standards and the process of self4uation. In R M. Sorrentino & E. T. Higgins (Eds.), Handbook of motivation and cognition: Foundation of mcial behavior Vol. 1 (pp. 2-3). New York: Guilford Press.

Icheiser, G. (1 949). M i s u n d e r s t v l ~ in human relations: A study in filsc social perception. Amnican j o u d of Sociology, 55, Put 2.

Jacoby, L. L. & Kdley, C. M. (1987). Unconscious influences of memory for a prior event. Pmon- rzlity and Social PsychorOgY BuWn, 13,3 14-336.

Jacoby, L. L., Toth, J. P., Lindsay, D. S., & Debner, J. A (1 992). h r c s for a layperson: Methods for revealing unconscious processes. In R F. Bornstein and T. S. Pitunan (Eds.), Perc'ption with- out awareness: Cognitive, clinic4 andsocirrlpmpectives (pp. 81-120). New York: Guilford Press.

Johnson, J., Bryant, M., Jackson, L A, Gatto, L., Now& A , & MacVittie, T. (1994). Construct accessibiliry, AIDS, and judgment. journal ofsocial Behavior and Pmonality, 9, 19 1-1 98.

Johnson-Laird, P. N. (1 983). A computltiod analysis of consciousness. Cognition and Brain Theory, 6,499-508. -

Kant, I. (179011987). C n ' W ofjdgmcnt. Trans. W . S. Pluhar. Indianapolis, IN: Hackett Pub- lishing.

Katz, D., 8~ Braly, K. (1933). Racial stereotypes of one hundred college students. journal ofAbnor- ma( and Social Psychology, 28,280-290.

Kawakami, K , Dim, K L, & Dovidio, J. F. (1998). Racial prejudice and stereotype activation. P m o d t y and S~cial P~ychoIogy BuIlrtin, 24 (4), 4 0 7 4 1 6.

Kihlsuom, J. F. (1990). The psychological unconscious. In L. A. Pcrvin (Ed.), Handbook ofpmon- d i g : Theory and research (pp. 445464). New York: Guilford Press.

Kihlstrom, J. F., Mulvancy, S., Tobias, B. A , & Tobis, I. P. (in press). The emotional unconscious. Kihlsuom, J. F., Cantor, N., Albright, J. S., Chew, B. R, Klein, S. B., & Niedenthal, P. M. (1988).

I~f~rmat ion processing and the study of the self. In. L. Berkowitz, et al. (Eds.), Advance* in ~ m r n ~ ~ s 0 ~ l p ~ c h o ~ Vol. 2 1 (pp. 145-1 78). San Diego: Aademic Press.

Klein, S. B., & L o b , J. (1990). The role of abstract and exemplar-based knowledge in self-judg- mms: Implications for a cognitive model of the self. In T. K Srull& R S. Wyer (Eds.), Content andpmcess ~ ~ @ n ' t y in &cts ofprior ncprrirnces: Advances in socirzl cogniti4n Vol. 3 (pp. 13 1- 139). Hillsdale, NJ: Lawrence Erlbaum Associates.

Mein, S. B., bfrus , J., 8r Kihlsuom, J. F. (1 996). Self-knowledge of an amnesic patient: Toward a

The Social Unconsciow 1 5 5

m b e n , A. (1995). Stereotypes and social judgment: The contcquenar of group e i l i r g . loup & ofPmonrJity and Social Psychology, 68,388-403.

m b e n , k J., Cmnen, S., Chasteen, k L., & Lickd, B. (1996). Private vs. public apressions of racial prejudice. journd of &wnmrntd Sock/ Psycho&gy, 32,437459,

~ ~ u , R R (1 989). Consuucr accessibility and electoral choice. Political Behavior, I I, 5-32. h ~ o u x , J. E. (1994). Emotional processing, but not emotions, a n occur unconsciousIy. In P.

Ekman & R J. Davidson (Eds.), T . nutun ofemotion: Fundomend paions (pp. 291-292). New York Oxford University Press.

~ D O U , J. E. (1 996). The motionul brain. New York: Simon & Schustcr. kmm, K, & Banaji, M. R (1998). Implicit and explicit gender identity and attitudes to-d

gender. Paper presented at the Midwestern Psychological Association meeting, Chicago. kporc, L., & Brown, R (1997). Category and stereotype activation: Is prejudice inevitable? journal

of P m o d t y and S o d P@ology, 72,275-287. kpore, L., & Brown, R (1999). Exploring automatic stereotype activation: A challenge to rhe

inevitability of prejudice. In D. Abrams & M. A. Hogg (Eds.), Sotialidnttity andsodcognition (pp. 14 1-1 63). W o r d : Bladrwell Publishers.

Levy, B. (1 996). Improving memory in old age through implicit self-stereotyping. JoutruJofPmon- alrty and Social Psychlog: 71, 1 092- 1 1 07.

Linville, P. W., & Carkron, D. E. (1994). Social cognition of the self. In P. G. Devine, D. C. Hamilton, & T. M. Osmm (Eds.), Social cognition= Impact on rocialpsycho~ (pp. 14S193). New York Academic Pms.

Lodtc, V., MacLeod, C., & W&r, I. (1994). Automatic and controlled activation of stereotypes: Individual di&ences associated with prejudice. British J o v d of Socirrl Plychbgy, 33,29-46.

Macnt, C. N., Bodenhausen, G. V., Milne, A. B., & Ford, R L (1997). On the rcguIation of recollection: The intentional forgetting of stereotypical memories. J o d of P m d g and Sc cial Psychhgy, 72,709-7 19.

Macrae, C. N., Bodenhawn, G. V., Milne, A. B., Thorn, T. M. J., & C a d i , L (1997). On the activation of social stereotypes: The moderating role of processing objectives. J o u d of hpni- mend SocirJ PsycborogT, 33,471489.

Mandler, G., Nakamun, Y., & Van Zandt, B. J. S. (1987). Nonspecific effens of aposure on stimuli that cannot be r e a p z e d . J w d of -mend Psychobgy: Lamins Mnnoly, & Gp nition, 13, 646-648.

Markus, H. (1 977). Self-schemata and processing information about the self. J o d of Pmom(iry and Sorirsl PsycbolOgr 3,44445-450.

Martin, L L (1986). Sedrrser: Use and disuse of concepts in impression formation. j o u d of P m o d t y and S d P ~ b o b , 51,49S506.

Martin, L L, Harlow, T. F., & Sack , F. (1992). The role of baddy d o n s in the evaluation of social events. P d ~ & S o l c i a l I s y h o ~ B u U r n ' n , 18,412419.

Meya, D. E., & Schcy;u~evddfr R W. (1971). Facilitation in rroognizin% pairs of words: Evidence of a dependence betw#n mrievlt operations. j d of&mmord Psyw, 90,227-234.

Mitchell, J., N o d , B., & Banaji, M. R (1998). A rose by any 0 t h name! D i d a t t i t u b toward social group mernberr Paper presented at the Arnericui Psychological Society meeting, wvhrngon,DC

Moskowitz, G. B. (1993). Person organization with a memory set: Arc sponpn#w trait i n h c e s personality h m a a k i c s or belcrviour lab&? Europcrur j o d t f P d t y , 7,191-208.

Moskowia, G. B., & Roman, R J. (1992). Spontaneous ni t idkxes as s e i f w primes: Implications br conscious social judgment. J m d of P d t y and SaciJBycholog(i 62.728- 738. . % -, '<t. .- -

Murphy, S. T., & Zajonc, R B. (1993). A f k , cognition, and rwuares. A f l k c h priming with suboptimal and optimal stimulus. jmtd of PCrSOllrPIiq and Socjaf I+hob 64 723-739.

Murphy, S. T., M o b , J. L, & Zapnc, R B. (1995). Additivity of p~u)ZUCious ;&a: Cant b i d effscts of priming and exposw. jowd of Pcmvsu&y d SoriJ* 69,589-602.

N d y , J. H. (1977). Sanvrtic priming and rarkvd f)om laid memorg: Roles of inhibitionlw sp+ aaivation and 1imitedcppacit-y artention. jovnrolof Wmnd S o d Pqchohgy, 26 505-527.

Nelson, T. E., AcLer, M., & Mulis, M. (1996). Irrt~rcssible cfctco- j o v d ofErprnmcntrrl Socirrl P@&, 32,l S 3 8 .

N b n , T. E., Bicmat, M. B., Bt Manis, M. (1990). Everyday h e ntcs (sa stereotypes): Potent and resilient. j d of Pmonaiig and Sorial M, 59,664475.

Newman, S. L, & Uieman, J. S.41989). Spontaneous trait infercna. In J. Ulanan & J. A. Bargh (Eds.), U n i n d e d thought (pp. 155-188). New York: Guilford Press.

Ncwrnan, L. S., & Ulemur, J. S. (1990). A d i t i o n and contrast e&ms in spontaneous trait inference. Pcllonality and Socirrl Psycbobgy BuUm'n, 16 224-240.

Niedcnthal, P. U, Smcrlund, M. B., &Jones, D. E. (1994). Emotional o r p i a t i o n of perceptual memory. In P. M. Niedenth;rl & S. Kitayama (Eds.), TiK kcorti eye: Enrotionuf inp...Ces in pctuption a d m m h (pp. 87-1 13). San Diego: Academic Press.

Norris, A E., & Devine, P. G. (1992). hdung pregnancy conarm to pregnancy risk avoidant action: The role of construct accessibility. P d t y a n d S o r i a l P q d w ~ B a n , 18,118-127.

Nosek, B., Banaji, M. R, & Gtecnwdd, k G. (1998). Math = Bad + Male, Me = Good + Female, therefore Math ~t Me. Paper presented at the American Psychologid Society meeting, Washing- ton, DC.

Pendry, L. F., & Macrae, C. N. (1996). What the disinterested perceiver wcrloob: Goaldirected - social atcgorization. P m o n d t y and Social Psycbobgy BIrllm'n, 22,249-256.

Perdue, C. W., Dovidio, J. F., Gurtman, M. B., &Tyler, R B. (1990). 'Us" and 'Them": Social categorization and the process of intergroup bias. j o w ~ l of P m o d t y and S o d Psychology, 59, 475-486.

Petty, R E., & Gcioppo, J. T. (1983). The role of bodily responses in attitude measurement and change. In J. T. M o p p o and R E. Petty (Eds.), Sociafpsychopbyswhgy: A zourcebook (pp. 5 1- 10 1). New York: Guilford Press.

Plant, E. A, & Devine, P. A (1998). In terd and arternid motivation to respond without preju- dice. Journal of PmmtrZli~y and S o d P~cbology, 75,8 1 1-832.

Posner, M. I., & Snyder, C. R R (1975). Attention and cognitive control. In R L. Solso (Ed.), Infitmrsrionpto~essin~andcognition: Tbcl;yoLu Symposium (pp. 55-85). H i W e , NJ: Lawrence Erlbaum Associates.

Pmrto, F. (1994). Consciousness and automatic evaluation. In P. M. Nicdentha and S. Kitayama (Eds.), The hcarri qc : Emotional injluenccs in pmrption andonention (pp. 1 1 5-1 43). San Diego: Academic Press. -

Pntto, F., & Bargh, J. A (1991). Stereotyping based on apparently individuating information: Trait and global components of sex stereotypes under attention overload. J o u m l ofE.pmimm~l Social P~cbobgy, 227,2647.

Pntto, F., & John, 0. P. (1991). Automatic vigdance: The attention-grabbing power of negative social information. j o u d of P m o d i t y and Social P~chobgy, 66 380-39 1.

Rosier, M., Banaji, M. R, & Gracnwald, A. G. (1998). Implicit and explicit self-esteem & group membership. Paper presented at the Midwestern Psychological -ation meetings, Chicago.

Ross, L. (1 977). The intuitive psychologist and his shortcomings. In L. Berkowia (Ed.), Advances in ~ ~ m r r r t a l s o d p s y d o b g y Vol. 10 (pp. 173-220). New York Academic Prw.

Ross, M.. & Olson, J. M. (1981). An apemcy-attribution modd of the effects of placebos. P.y cho&d ~nrmu, 88,408437.

R u d m , L. A, & BorgiL, E. (1995). The lhcrglow of construct accessibility: The behavioral co"x9uenw of priming men to view women as sexual objects. jovnyrl,of ,Gpm'rnrnrol Social P F h o k , 31,493-517.

L. & Glick P. (1998). Implicit gender stereotypes and backlash toward agcntic women: The hidden costs to women of a kinder, gender image of managers. Unpublished manuscript. Rutgem University.

A* & -&, S. (1998). Implicit and explicit attitudes toward female authority. Unpublirhed - e p t , Rutgcrs University.

The So&[ Unconsciow 157

~ ~ & x m n , L. A, G r n n d d , A G., 8~ McGhce, D. E. (1998). Sex di&tences in ender types rmded by the Implicit Association Test. Unpublished manuscript, Rurgen UnivefSiry.

L A, Grnnwald A G., Mdot, D. S., & S c h w ~ m J. L K (1996. XuttLmmdc preju- &a: Flexibility and genedry of the Impkir Associ;rtion T a Unpublished hrgm University.

~ ~ d i k i d ~ , C., & Skowronski, J. J. (1990). Towards reconciling personality and social psychology: A construct acccssibilir); a'pproach. j o u d of S o d Bebavior and Pmonaliy, 5,53 1-546.

sedikidcs, C., & Skowronski, J. J. (1993). The self in impression formation: Trait centdry and social perception. J o u d of Erpmmmmtd S o d Pzychbgy, 29.347-357.

Sherman, J. W., SuOeS~ner, S. J., IAfrus, S. T., & Deguunan, G. (1997). Stereotype suppression and recognition memory for stereotypid and nonstereotypical information. Social Copirion, 15 (3)' 205-21 5.

Shih, M., Pitrinsky, T. L., &Ambady, N. (1999). Stereotype sweptibility: Identity salience and shifts in quantitative performance. PzycholOgrcal S k e , 10,8 1-84.

Shrum, L. J., & O'Guinn, T. C. (1993). Processes and effeca in the construction of social rdily: - Construct accessibility as an explanatory variable. Communications Re~arch, 20,43647 1.

~kowronski, J. J., Carlston, D. E., Mae, L, & Crawford, M. T. (1998). Spontaneous uait transfer- ence: Communicators iakc on the qualities they describe in others. joumal of P m d t y and Social Psycbobgy. 74,837-848.

~outhwidc, L., Stnk, C., Lindell, M. (1986). The roles of historical experience and construct acces- sibility in judgments about alcoholism. Cognitiu Thcrazpr a d Research, 10. 167-1 86.

Spencer, S. J., Fein, S., Wolfe, C. T., Fong, C., & Dunn, M. A. (1998). Automatic activation of stereotypes: The role of self-image threat. P m d t y andSocialPsycbobgy Bullctn, 24(1 I), 1 139- 1152.

Stangor, C., & Lange, J. E. (1994). Mental representations of social groups: Advances in under- standing stereotypes and stereotyping. Advances in &&mend Social P ~ b o b g y , 24 357-4 16.

Steele, C. M. (1997). A threat in the ak How stereotypes shape the intellatual identities and performance of women and Afiian Americans. A k a n Psycbo&gzst, 52.6 l S 2 9 .

Stecle, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. j o u d of Pmonolity and S o d Psychbgy, 69,797-8 1 1.

Stepper, S., & Strack, F. (1993). Proprioceptive determinants of emotional and nonemotional &I- ings. /'ournaf of P m o d t y and So& F s ~ ~ h o k , 6V., 2 1 1-220.

Suack, F., Martin, L L , & Stepper, S. (1 988). Inhibiting and ficilitating conditions of the human smile: Unobtrusive test of the fkiaI feedback hypothesis. journal of P m d t y and S o d P~chol- ogy, 54, 768-777.

Suoessner, S. J. (1996). Social ategorizarion by race or sac E&co of perceived non-normalcy on response times. Social *tion, 14 (3), 247-276.

Swum, W. B., Hixon, J. G., Stein-Seroussi, A, & Gilbert, D. T. (1990). The flaring gleam of pnise: Cognitive processes undcriyq behavioral racrions to self-relevant M b a & Journal of P m o d t y and Socirrl Psycbob, 59,17-26.

Trope, Y. (1986). Idcnrifiation and inferential pprocessc~ in dirposttiod amibution PsycholOgKlZl RrYinu, 93. 237-257.

Trope, Y., & Aifieri, T. (1997). Effbrdulncsd and f l d i t y of dispsciaonal judgment procmes. Joumai of Personality and Social psvCbolsg: 13.662474.

Trope, Y., & Libcrman, A. (1993). The use of trait concepaons to idcnafl otha people's behavior and draw inkrences about their personalities. Pmonrrlty r m d S o c i r J P + ~ Bullrn'n, 55j562.

Trope, Y., & nempson, E. P. (1 997). Loohg fbr truth in dl the wrong places? Asymmetric search of in&d&ting information about rmcotyped group manbas. Jd O f P d t y and Sorinl P~yckbgy, 73,229-241.

Ulunan, J. S., & Bargh, J. A. (Eds.) (1989). Unirr trndcdthqk New York: Gu&rd Press. Uleman, J. S., Newman, L, & Winter L (1992). G n personality a i r s be intirrrd urtornatidy!

Spontaneous bees quire cognitive apaciry at encodq. Gnrciwmm and Cognition, I, 77-90.

Ulemul, J. S., Hon, A, Roman, R J., & Moskowicz, G. B. (1996). On line evidcna b r spontane-

4nKaritiahrrnarstaxxdlng. P d t y # I d S o c i r l - B u l k i r r , r 2 2 377-394. Ulanan, J. S., Moskowitz,G. 0.. k r m , R j.,&Rhec, E (1993) T6ei~mmiks~ind intentional h How rpontancous trait i n h a s ref& to persons. SosiJ 11,321-35 1.

V m hippenbag, A, & Dijiwrhuis, A. (1996). A posteriori . . T h e p r w v a -

cion of ncnocypes through memory distortion. S o d -n, 14 (I), 21-53. Vulm~n, E. J., Pad, 8. Y., loo, T. A, & Miller N. (1997). The madan fia of p@dioe and

rauaunl fcrt~rt~ th;uudaatc the e&a of ooopcnrion on a f k c ~ jbvrnol ufPmnlluy and SocirrlP*@, 73,941-959. - ; -*

Von Hippel, W., Sckaqurptcwl, D., & Vugzs, P. (1997). The kng\usoc intergoup bias as an impfiat indicator of prejudia. j d o f E*pmmcntrrl Seciol Prynkob , 33,490-509.

V ~ M , S. R, & Roll& D. (1998). Physiologid re~ponsc to r minirn;21 d encounter: Efkts of gender, ethniaty, and 4 contat. Psycix,pbyswlog,, 35,462469.

Walsh, W. A, & h j i , M. R (1997). Thc coil& sdf. i+G. Snodgmss br R L Thompson (Eds.), Thr s e I f ~ ~ ~ ~ p ~ k o l o g V : &&api&m se~-muss, d nbr ufroncrpt. Annab ofthe New YorkAm&q of* Vol. 818 (pp. 193-214). New York: New Yurkhduny of Sciences.

Wann, D. L , & Branscornbe, N. R (1990). Pemn perception when e v e or nonaggressive sports are primed. &r&u Behavior, I6 27-32.

Wegner, D. M., & Bargh, J. A (1998) Control and automaticity in d We. fa D. T. Gilbert, S. T. Fish, & G. Lindzey (Eds.), The bandbook of socialpsycbobgy Vol. 2 (pp. W 9 6 ) . Boston: McGraw-Hill.

Wells, G. L, & Petty, R E (1980). The e k of overt h a d movement on persuasion: Compatibil- ity and incompatibility responses. &uic and Applied S o d Psychlog: 1, 2 19-230.

Whitney, P., Davis, P. A, & Wuing, D. A. (1994). Task effixzs on trait inkncc: Distinguishing categorization from characterization. S o d Ggnirion, 12, 1 !3-35.

Whitney, P., Waring, D. A, & Zmgmuk, B. (1992). Task on the spontaneous activation of trait concepts. S o d Cbgnmon, 10,377-396.

Winter, L , & Uleman, J. S. (1984). When arc social judgments made? Evidence for the spontane- ousness of tnit infrrentcs. J o u d ofPmnczlicy and Socirzl Psychology, 47,237-252.

Wittenbrink, B., Judd, C. M., & Park, B. (1997). Evidence tbr racial prejudice at the implicit level and its relationship with questionnaire me;rsures. J o u d of Pmnrzl i4 and Social Psychology, 72, 262-274.

Wyer, R S., Jr., & Srull, T. K. (1980). The processing of social stimulus infbrmation: A conceptual integration. In R M e , E. B. Ebbesen, T. M. Osuom, R S. Wyer, D. L. H d t o n , & D. E. Carlton (Eds.), Pmon w r y : Thr cognitive basis of s o ~ p m ~ z i o n (pp. 227-300). Hillsdale, NJ: Erlbaurn.

Young, J., Thornsen, C. J., Borgida, E., Sullivan, J. L, & Aldrich, J. A. (1 99 1). When self-interest makes a difference: The role of consuuct accessibility in politid reasoning. j~urmalofExp~mcn- tlzr Sociul P.luch'obgy, 227,271-296.

Zajonc, R B. (1968). The atcitudinaJ effects of mere exposure. j o u d of Personality and Social . Psychology, Monograph Supplement, 9, (2, Pt. 2). Gone, R B. (1994). Evidence for nonconxious emotions. In P. E h a n and R J. Davidson (Ms.),

The nature Ofmt ion: F ~ n ~ ~ u J q r r m i o n r (pp. 293-297). New York Oxford University Press. Zajonc, R B. (1998). hotions. In D. T. Gilbert and S. T. Fiskc (Eds.), The handbook of sorial

psycho& Val. 2,4th. edn. (pp. 591-632). Boston: McGraw-Hill. Zajonc, R B., Murphy, S. T., & Inglchan, M. (1989). Feeling and ficid efferenct: Implications of

the vascular theory of emotions. P~cbological h, %, 395-4 I 6. -9 P.9 D m i l e r , R A, & Olson, J. M. (1984). Physiological mediation of attitude mainte-

~ c c , fomtion, and change. In W. M. Waid (Ed.), Sociopbysiology (pp. 163-196). New York Springer-Verlag .

Zhte1 M. A, Bonilla, S., 8~ Luhrano, M. (1995). Ethnic influenos on exemplar r e u i d and srereoryp& S o d Cognition, 13 (2), 145-1 62.