Processing Fluency and Aesthetic Pleasure: Is Beauty in the Perceiver’S Processing Experience?

Processing Fluency and Aesthetic Pleasure: Is Beauty in the Perceiver’s Processing Experience?

Rolf Reber Department of Psychosocial Science University of Bergen, Norway Norbert Schwarz Department of Psychology and Institute for Social Research University of Michigan Piotr Winkielman Department of Psychology University of California, San Diego

We propose that aesthetic pleasure is a function of the perceiver’s processing dynamics: The more fluently perceivers can process an object, the more positive their aesthetic response. We review variables known to influence aesthetic judgments, such as figural goodness, figure–ground contrast, stimulus repetition, symmetry, and prototypicality, and trace their effects to changes in processing fluency. Other variables that influence processing fluency, like visual or semantic priming, similarly increase judgments of aesthetic pleasure. Our proposal provides an integrative framework for the study of aesthetic pleasure and sheds light on the interplay between early preferences versus cultural influences on taste, preferences for both prototypical and abstracted forms, and the relation between beauty and truth. In contrast to theories that trace aesthetic pleasure to objective stimulus features per se, we propose that beauty is grounded in the processing experiences of the perceiver, which are in part a function of stimulus properties.

What is beauty? What makes for a beautiful face, kiewicz, 1970). This objectivist view inspired many appealing painting, pleasing design, or charming scen- psychological attempts to identify the critical contrib- ery? This question has been debated for at least 2,500 utors to beauty. Among the identified features were bal- years and has been given a wide variety of answers ance and proportion (Arnheim, 1974; Birkhoff, 1933; (Feagin, 1995; Tatarkiewicz, 1970). However, one can Fechner, 1876; Gombrich, 1995), symmetry (Arnheim, broadly distinguish three main positions. 1974; Birkhoff, 1933; Gombrich, 1984; Humphrey, Many theorists, dating back at least to Plato, saw 1997), informational content and complexity (Berlyne, beauty as a property of an object that produces a plea- 1971, 1974; Eysenck, 1941; Garner, 1974), as well as surable experience in any suitable perceiver (Tatar- contrast and clarity (Gombrich, 1984, 1995; St. Thom- as of Aquinas, see Maritain, 1966; Solso, 1997). The objectivist view of beauty was so dominant in the The reported research was supported by the Swiss National Sci- 16th century that artists introduced pattern books, of- ence Foundation (Grant 61–57881.99 to Rolf Reber) and the Na- fering pictorial elements that artists could copy and tional Science Foundation (Grant BCS–0217294 to Piotr Winkiel- combine with each other to create beauty (see Gom- man). Preparation of this article was supported through a fellowship from the Center for Advanced Study in the Behavioral Sciences to brich, 1995). Norbert Schwarz. We thank Marilyn Brewer, Tedra Fazendeiro, Other theorists, dating back at least to the Sophists, Greg Feist, Jamin Halberstadt, Kent Harber, Christian Jakob, Jack proposed that anything could be beautiful if it pleases Katz, Bruce Morton, John Skowronski, Bruce Whittlesea, Bernd the senses (Tatarkiewicz, 1970). From this perspective, Wittenbrink, and an anonymous reviewer for comments on the arti- beauty is a function of idiosyncratic qualities of the cle and helpful discussions. Requests for reprints should be sent to Rolf Reber, University perceiver and all efforts to identify the laws of beauty of Bergen, Department of Psychosocial Science, Christiesgate 12, are futile. This subjectivist view, reflected in expres- N-5015 Bergen, Norway. E-mail: [email protected]; Norbert sions like “beauty is in the eye of the beholder” or “de Schwarz, Institute for Social Reseach, 426 Thompson Street, Ann gustibus non est disputandum” (taste cannot be de- Arbor, MI 48106–1248. E-mail: [email protected]; Piotr Win- kielman, Department of Psychology, University of California at San bated), underlies the social constructivist emphasis on Diego, 9500 Gilman Drive, La Jolla, CA 92093–0109. E-mail: the historically changing and culturally relative nature [email protected] of beauty (see Kubovy, 2000).

364 FLUENCY AND AESTHETIC PLEASURE

Most modern philosophical analyses, however, re- Beauty and Judgment ject the objective versus subjective distinction. Instead, Research participants in experimental aesthetics they suggest that a sense of beauty emerges from pat- and psychology studies are rarely asked to judge terns in the way people and objects relate (e.g., In- “beauty” per se. Instead, most studies have focused on garden, 1985; Merleau-Ponty, 1964). In this article, we judgments like figural goodness, pleasantness, liking, adopt this interactionist perspective and seek to iden- and preference. One may wonder whether these more tify those patterns. As detailed later, we propose that modest judgments capture the grand realm of beauty. beauty is grounded in the processing experiences of Most researchers, including us, believe that by study- the perceiver that emerge from the interaction of stimu- ing such simple judgments, one can identify basic pro- lus properties and perceivers’ cognitive and affective cesses underlying the aesthetic experience. Further, processes. several lines of research, for example on the mere exposure effect (addressed later in this article), demonstrate that different kinds of evaluative judgments op- Beauty, Experience, and Judgment erate via similar processes (see Bornstein, 1989). Thus, there are reasons to believe that judgments of prefer- Before we present our proposal, a few clarifications ence, liking, and beauty are closely related. are in order. We first introduce our use of the concept of Another issue raised by the focus on basic processes beauty and relate it to the concept of aesthetic pleasure. and the requirements of experimental control is that Subsequently, we discuss the relation between beauty most psychological studies explore fairly mild aes- and various aesthetic judgments, such as judgments of thetic experiences, such as those likely to accompany preference and judgments of aesthetic value. the perception of a vase, a simple melody, an abstract shape, or a human face. Clearly, much of what most humans call “beautiful” on a daily basis falls into the cat- Beauty and Aesthetic Pleasure egory of such mild experiences. However, we hope that future research will also explore stronger subjective The philosopher George Santayana described three experiences and examine to what extent our analysis defining features of beauty: “beauty … is value posi- can inform such research. tive, intrinsic, and objectified” (Santayana, 1896/1955, Finally, we note that beauty as explored in this arti- p. 31). By value positive and intrinsic, Santayana cle is unrelated to aesthetic value (see, e.g., Beardsley, meant that beauty provides pleasure without any rea- 1981). Since the emergence of modern art, a piece of soning about expected utility. This is similar to Thom- art can have aesthetic value without being beautiful as of Aquinas’ definition of beauty as what gives plea- (i.e., without producing an experience of aesthetic sure at sight (“id quod visum placet”), suggesting pleasure). Conversely, a painting that “pleases the immediate joy without intermediate reasoning (see eyes” may be without any artistic merit. The judgment Maritain, 1966). Similarly, the art historian Read of aesthetic value, in contrast to beauty or aesthetic (1972) did “not believe that a person with real sensibil- pleasure as defined here, often involves substantial rea- ity ever stands before a picture and, after a long process soning about the piece of art under consideration. We of analysis, pronounces himself pleased. We either like return to these issues in the discussion. at first sight, or not at all” (p. 38). Finally, beauty is objectified. For example, the experience of having a cold drink on a hot day is both value positive and intrinsic, but this immediate pleasure lies exclusively in a Overview positive sensation of the body and has little to do with aesthetic appreciation of the object. In contrast, per- Our core proposal is straightforward. We suggest ceivers look at a painting not to please their body, but to that aesthetic experience is a function of the perceiver’s enjoy the painting’s beauty. Hence, people experience processing dynamics: The more fluently the perceiver beauty as something that lies in the object. Therefore, can process an object, the more positive is his or her beauty is not an “objective,” but an “objectified” prop- aesthetic response. This proposal entails four specific erty (see also Feagin, 1995). assumptions. First, objects differ in the fluency with In our analyses we follow this philosophical tradi- which they can be processed. Features that facilitate tion and define beauty as a pleasurable subjective expe- fluent processing include all the core features identi- rience that is directed toward an object and not medi- fied in the objectivist tradition, like goodness of form, ated by intervening reasoning. This definition closely symmetry, figure–ground contrast, as well as variables resembles the definition of aesthetic experience used that have not received attention in traditional theories in empirical aesthetics (e.g., Kubovy, 2000; Martindale of aesthetic pleasure, like perceptual and conceptual & Moore, 1988). Accordingly, we use the words priming procedures. Second, processing fluency is it- “beauty” and “aesthetic pleasure” interchangeably. self hedonically marked and high fluency is subjec-

365 REBER, SCHWARZ, WINKIELMAN tively experienced as positive, as indicated by psy- tlesea, 1993), and even truth (e.g., Begg, Anas, & chophysiological findings. Third, processing fluency Farinacci, 1992; R. Reber & Schwarz, 1999). feeds into judgments of aesthetic appreciation because Although fluency may characterize mental process- people draw on their subjective experience in making es occurring at various levels, our discussion is primar- evaluative judgments, unless the informational value ily concerned with perceptual fluency (i.e., the ease of of the experience is called into question. Finally, the identifying the physical identity of the stimulus). Per- impact of fluency is moderated by expectations and ceptual fluency is influenced by variables such as per- attribution. On one hand, fluency has a particularly ceptual priming, clarification, presentation duration, strong impact on affective experience if its source is repetition, or figure–ground contrast, as discussed unknown and fluent processing comes as a surprise. shortly. However, our arguments apply as well to con- On the other hand, the fluency-based affective experi- ceptual fluency, or the ease of mental operations con- ence is discounted as a source of relevant information cerned with stimulus meaning and its relation to se- when the perceiver attributes the experience to an irrel- mantic knowledge structures (e.g., Whittlesea, 1993; evant source. Winkielman, Schwarz, Fazendeiro, & Reber, 2003). In the remainder of this article, we elaborate on this We use the more general term processing fluency to proposal and discuss supporting evidence from re- capture these commonalities. search in social and cognitive psychology as well as empirical aesthetics. We first introduce the concept of Hedonic Marking of Fluency processing fluency, discuss its relation to evaluative Multiple theoretical notions converge on the as- processes, and review relevant experimental findings. sumption that high fluency is positively marked. The Next, we provide a selective review of research that at- basic idea in all these notions is that high fluency says tempted to identify key contributors to beauty, drawn something about a positive state of affairs, either from different research traditions. We propose that all within the cognitive system or in the world (see of the previously identified variables share a common Winkielman et al., 2003, for a more comprehensive underlying characteristic, namely the ability to facili- treatment). Specifically, high fluency may elicit posi- tate stimulus processing. Following this review, we tive affect because it is associated with progress toward highlight how attributional processes and processing successful recognition of the stimulus, error-free pro- expectations moderate fluency effects on perceived cessing, or the availability of appropriate knowledge beauty, and discuss challenges to our proposal. We fi- structures to interpret the stimulus (Carver & Scheier, nally show that a perceptual fluency theory of beauty 1990; Derryberry & Tucker, 1994; Fernandez-Duque, can account for phenomena that are difficult to concep- Baird, & Posner, 2000; Schwarz, 1990; Simon, 1967; tualize in the context of other theories. Specifically, a Ramachandran & Hirstein, 1999; Vallacher & Nowak, perceptual fluency theory helps explain the interplay 1999). High fluency may also feel good because it sig- between developmentally early preferences and social- nals that an external stimulus is familiar, and thus un- ization, the apparent contradiction between prefer- likely to be harmful (Zajonc, 1968, 1998). ences for average versus exaggerated forms, and why artists and scientists have often considered truth and beauty as two sides of the same coin. Fluency and Evaluations: Empirical Evidence

As our review of the empirical literature will indi- Processing Fluency cate, high fluency is reliably associated with more positive evaluations. Historically, this possibility has re- The Concept of Processing Fluency ceived most attention in the context of debates on the The processing of any stimulus can be characterized nature of the mere-exposure effect (i.e., the observa- by a variety of parameters that are nonspecific to its tion that repetition enhances liking for an initially neu- content, such as speed and accuracy of stimulus pro- tral stimulus; for reviews, see Bornstein, 1989; Zajonc, cessing (see R. Reber, Wurtz, & Zimmermann, 2004). 2000). Several authors proposed that the mere-expo- These parameters tend to lead to a common experience sure effect might reflect increases in perceptual flu- of processing ease or “fluency” (for reviews see Clore, ency (e.g., Bornstein & D’Agostino, 1994; Jacoby, 1992; Jacoby, Kelley, & Dywan, 1989; Whittlesea, Kelley, et al., 1989; Seamon, Brody, & Kauff, 1983, Jacoby, & Girard, 1990). A large number of studies Whittlesea, 1993). If so, we may expect that any vari- show that people draw on fluency to make a variety of able that facilitates fluent processing similarly results nonaesthetic judgments, including judgments of loud- in increased liking, even under conditions of a single ness (e.g., Jacoby, Allan, Collins, & Larwill, 1988), exposure. Studies from our lab support this possibility. clarity (e.g., Whittlesea et al., 1990), duration (e.g., Some of these studies relied on variables that have a Witherspoon & Allen, 1985), familiarity (e.g., Whit- long tradition in theories of aesthetics, like figure–

366 FLUENCY AND AESTHETIC PLEASURE ground contrast, and will be reviewed in a later section. stead, the same pattern of effects was obtained when Other studies relied on manipulations that are uniquely participants studied a list of concept primes before they suggested by a fluency account of aesthetic pleasure, were exposed to the pictures. namely visual and conceptual priming procedures. Affective Response Priming Procedures Importantly, the influence of processing fluency is R. Reber, Winkielman, and Schwarz (1998, Study not limited to explicit judgments of preference and 1) presented participants with pictures of everyday ob- beauty, but can also be captured with psychophysi- jects, such as a desk, bird, or plane (taken from Snod- ological measures. This was demonstrated by Win- grass & Vanderwart, 1980). The quality of the pictures kielman and Cacioppo (2001), who assessed partici- was slightly degraded and processing fluency was pants’ affective responses to fluent stimuli with facial manipulated through a visual priming procedure. De- electromyography (EMG). EMG relies on the observa- pending on conditions, the target picture was preceded tion that positive affective responses increase activity by a subliminally presented, highly degraded contour over the region of the zygomaticus major (“smiling of either the target picture or a different picture. We ex- muscle”), whereas negative affective responses in- pected that a matching contour would facilitate pro- crease activity over the region of the corrugator su- cessing (high fluency), consistent with research show- percilli (“frowning muscle;” e.g., Cacioppo, Petty, ing that subliminal visual primes enhance the accuracy Losch, & Kim, 1986; Lang, Greenwald, Bradley, & with which a stimulus can be identified (Bar & Bieder- Hamm, 1993). As expected, high fluency was associ- man, 1998). Some participants were asked to indicate ated with stronger activity over the zygomaticus region how much they liked the target pictures. Other partici- (indicative of positive affect), but was not associated pants were asked to press a key as soon as they could with the activity of the corrugator region (indicative of recognize the object in the picture, thus providing a negative affect). Moreover, the observed differences measure of recognition speed, an indicator of fluency. occurred in the first 3 sec after stimulus presentation The data were consistent with predictions: Pictures and several seconds before participants made overt primed by matched contours were recognized faster, judgments, indicating a spontaneous affective response indicating higher fluency, and were liked more than to processing fluency. A subsequent study replicated pictures preceded by mismatched contours. Moreover, this effect, using presentation duration as a manipula- participants were unaware of the fluency manipulation, tion of fluency. thus eliminating the possibility of strategic responding The Mediating Role of Affect to pictures preceded by various primes. Extending this work with a cross-modal priming Theoretically, we assume that the spontaneous af- task, Winkielman and Fazendeiro (2003) showed par- fective response observed in the Winkielman and ticipants a series of unambiguous pictures of common Cacioppo (2001) studies mediates the impact of flu- objects and animals. Each picture was preceded by a ency on evaluative judgments. Presumably, perceivers letter string consisting either of a word or a nonword. interpret the positive affect elicited by processing flu- Participants first indicated, as fast as possible, if the let- ency as their response to the target, resulting in more ter string was an actual English word. Subsequently, positive evaluations. This assumption is consistent they reported their liking for the picture. The letter with the feelings-as-information model, which holds strings served as the fluency manipulation. Some pic- that feelings serve as a source of information in their tures were preceded by matching words (e.g., word own right, unless their perceived informational value “dog”—picture of a dog), introducing the highest level for the judgment at hand is undermined through of fluency. Other pictures were preceded by associa- (mis)attribution manipulations (e.g., Schwarz & Clore, tively related words (e.g., word “key”—picture of a 1983; for a review see Schwarz & Clore, 1996). lock), introducing a medium level of fluency. Yet other Winkielman and Fazendeiro (2003) tested this pre- pictures were preceded by an unrelated word (e.g., diction by replicating the conceptual priming study de- word “snow”—picture of a desk), introducing the low- scribed earlier with a misattribution manipulation. est level of fluency. The results showed a robust effect Before they made their liking judgments, they specifi- of concept priming on participants’ evaluation of the cally told participants that their reactions to the stimuli target pictures. As expected, pictures preceded by might be influenced by background music played to matching words were liked significantly more than them. The music was an ambiguous new-age piece re- pictures preceded by related words, which, in turn, corded at half-speed (see Schwarz et al., 1991, for de- were liked significantly more than pictures preceded tails on this manipulation). Some participants were by unrelated words. Follow-up studies indicated that told that the music might bias how easily stimuli come these fluency effects do not require that the concept to mind (i.e., their fluency experience), whereas other primes immediately precede the target pictures. In- participants were told the music might influence how

367 REBER, SCHWARZ, WINKIELMAN they feel about the various stimuli (i.e., their affective ing on two classes of factors that influence perceived experience). The results were highly informative. Spe- beauty. We first discuss “objective” features of stimuli, cifically, attributing subjective fluency to music did not such as amount of information, symmetry, figure– eliminate the effect of processing facilitation on liking. ground contrast, and clarity. Subsequently, we address That is, participants who were informed that the music the role of perceivers’ previous experience with stim- might influence how easily things come to mind still uli, such as repeated exposure to the stimulus, implicit judged pictures as more likeable when they were pre- learning of rules that underlie stimuli, and the pro- ceded by related rather than unrelated words, replicat- totypicality of the stimulus. We propose that the influ- ing the previous findings. In contrast, attributing the af- ence of all of these variables derives from their ability fective response to the music did eliminate the effect of to facilitate fluent stimulus processing. processing facilitation on liking. That is, participants who were informed that the music might influence Objective Features of Stimuli their feelings toward various stimuli no longer judged new pictures as more likeable when they were pre- Amount of Information ceded by related rather than by unrelated words. The idea that the amount of information is an im- This pattern of results suggests that it is the flu- portant determinant of beauty has a long history in aes- ency-based affective reaction, and not the fluency ex- thetics (e.g., Arnheim, 1974; Gombrich, 1984). In psy- perience itself, that serves as a basis of judgment, chology, relevant research has mostly been conducted resulting in enhanced liking. This interpretation is con- in the Gestalt tradition. Early researchers focused on sistent with research into the use of moods (e.g., stimulus organization and proposed that perceived Schwarz & Clore, 1983) and other phenomenal experi- “goodness” of the stimulus depends on the relation be- ences (e.g., ease of recall; Schwarz, 1998) as experien- tween stimulus organization and psychological mecha- tial sources of information. It suggests that the affec- nisms (e.g., Koffka, 1935). In their view, stimuli iso- tive reactions to fluency, captured in EMG studies morphic to physiological mechanisms are easy to (Harmon-Jones & Allen, 2001; Winkielman & Caciop- process and are rated as “good figures.” This work’s po, 2001) as well as self-reports of mood (Monahan, emphasis on processing ease anticipated later interest Murphy, & Zajonc, 2000), serve a crucial mediating in perceptual fluency by several decades.1 role. Subsequent research in the Gestalt tradition has pri- marily focused on the amount of information rep- Summary resented in a stimulus, consistent with the traditional In combination, the reviewed studies demonstrate assumption that beauty resides in the object (e.g., Att- that variables that facilitate the processing of a stimu- neave, 1954; Hochberg & McAlister, 1953). Garner lus result in more positive affective reactions, as cap- (1974) expressed the amount of information extracted tured by psychophysiological measures (Winkielman from a stimulus in terms of inferred R × R subsets, & Cacioppo, 2001), as well as more favorable judg- which represent the number of different shapes after ments of preference (R. Reber et al., 1998). The impact Reflecting and Rotating a shape. He found that judg- of fluency on evaluative judgments is apparently medi- ments of figural goodness were higher the less infor- ated by the elicited affective reactions and hence elimi- mation people had to extract from a stimulus to per- nated when the informational value of the affective receive it (i.e., when the stimulus had high redundancy). action is called into question through misattribution This finding is consistent with people’s preference for manipulations (Winkielman & Fazendeiro, 2003). symmetric shapes, because they contain less informa- Not surprisingly, the key variables used in these ex- tion than asymmetric, but otherwise identical shapes periments—visual and conceptual priming—have not (Garner, 1974). Further research has demonstrated that received attention in traditional theories of aesthetic this principle extends beyond judgments of figural judgment. Yet, they share with the variables commonly goodness to evaluative judgments. For example, Nicki, considered in these theories—like goodness of form, Lee, and Moss (1981) found that ambiguity in cubist symmetry, figure–ground contrast or prototypicality— paintings, defined in informational terms, was nega- that they facilitate fluent processing of the target stimu- tively related to pleasantness judgments. lus. We propose that this shared characteristic is at the Important for our proposal, stimuli with less in- heart of aesthetic pleasure. formation are not only more pleasing, but also easier to process, as measured, for example, by recognition

Determinants of Beauty: A Review 1Another historical precursor is Eysenck’s observation that “the pleasure derived from a percept as such is directly proportional to the decrease of energy capable of doing work in the total nervous sys- Next, we review research from social psychology, tem, as compared with the original state of the whole system” cognitive psychology, and experimental aesthetics bear- (Eysenck, 1942, p. 358).

368 FLUENCY AND AESTHETIC PLEASURE speed (Checkosky & Whitlock, 1973). Hence, our pro- preferred because symmetry facilitates fluent process- posal suggests that controlling for the amount of infor- ing (see Reber, 2002). mation should reveal a preference for easy-to-process stimuli. Several studies bear on this issue. Some of Contrast and Clarity these studies explored processing ease of different Theorists of aesthetics have long considered con- forms of symmetry, such as symmetry involving re- trast and clarity as “objective” determinants of beauty flection around a vertical axis (as in capital “A” or (Gombrich, 1984, 1995; Maritain, 1966; Solso, 1997). “V”), symmetry involving reflection around a horizon- Again, empirical research suggests that the influence tal axis (as in capital “E” or “D”), or symmetry involv- of these features may be mediated by their effect on ing reflection around a diagonal axis (as in the Nordic processing fluency. character “Ø”). Palmer and Hemenway (1978), as well Studies show that recognition speed, a standard as Royer (1981), used reaction times to show that verti- measure of fluency, is faster for stimuli high in fig- cal symmetry is easier to detect than horizontal sym- ure–ground contrast (e.g., Checkosky & Whitlock, metry, which in turn is easier to detect than diagonal 1973). Similarly, Whittlesea et al. (1990) showed that a symmetry. Building on this work, Palmer (1991) pre- related variable, visual clarity, influences perceptual sented dot patterns in vertically, horizontally, or diag- fluency as measured by memory misattribution. In onally symmetrical arrangements, manipulating the their study, participants saw short and rapidly pre- ease of processing while controlling for the amount of sented lists of words. After each list, a target word was information presented (in terms of R × R subsets, as presented within a visual noise mask, resulting in discussed earlier). Consistent with our perceptual flu- higher or lower visual clarity of the target word. As ex- ency analysis, the same stimuli received the highest pected, higher clarity increased the likelihood that an ratings of figural goodness when presented in a vertical item was erroneously recognized as having been pre- arrangement and the lowest ratings when presented in a sented earlier. This finding indicates a misattribution diagonal arrangement, with horizontal arrangements of high perceptual fluency, induced by visual clarity, to falling in between. Given that the amount of presented previous exposure. Consistent with this interpretation, information was identical in all conditions, these find- the effect disappeared when participants knew that vi- ings strongly support the hypothesis that figural good- sual clarity was manipulated. ness is a function of perceptual fluency rather than Building on this work, we explored the influence amount of information per se. of figure–ground contrast on liking (R. Reber et al., 1998). In one study we manipulated perceptual fluency Symmetry by varying the figure–ground contrast of circles presented for 1 sec. As expected, circles with high fig- The aforementioned findings also bear on proposals ure–ground contrast were judged as prettier (or less that try to locate beauty in objective symmetry. Sym- ugly, depending on the framing of the judgment task) metry has been found to influence the perceived attrac- than circles with low figure–ground contrast. More- tiveness of human faces (Gangestad, Thornhill, & Yeo, over, the impact of processing fluency did not depend 1994; Rhodes, Proffitt, Grady, & Sumich, 1998; on the framing of the judgment task. Specifically, we Rhodes, Sumich, & Byatt, 1999). Moreover, symmet- asked some participants to report how “pretty” the ric patterns are preferred even if they do not serve any stimuli were and others how “ugly” the stimuli were. biologically relevant function, both in humans (e.g., Increased fluency resulted in judgments of higher Humphrey, 1997; R. Reber & Schwarz, in press) and in “prettiness” as well as judgments of lower “ugliness,” animals (Rensch, 1957, 1958). These observations are indicating that fluency did not simply facilitate ex- often explained by postulating an innate preference for treme judgments per se.2 symmetry (e.g., Etcoff, 1999; Pinker, 1997). Given that Of course, one may argue that in the R. Reber et al. symmetry is indicative of mate value in several species (1998) study, it is the high figure–ground contrast per (e.g., Thornhill & Gangstead, 1993, 1999), this is a se (i.e., the objective feature), not perceptual fluency, plausible hypothesis, although the evidence for hu- that contributes to more positive evaluations. If so, mans is mixed (see Kalick, Zebrowitz, Langlois, & high-contrast stimuli should be judged more favorably Johnson, 1998, for a discussion). More important for our purposes, symmetrical patterns also have less information and are hence easier to process (Garner, 2This result, related findings by Seamon, McKenna, and Binder 1974). In addition, computer-modeling work suggests (1998), as well as studies using psychophysiological measures (Har- that a perceptual system designed to recognize objects mon-Jones & Allen, 2001; Winkielman & Cacioppo, 2001) speak from different viewpoints will process symmetrical against the possibility that the influence of fluency on preferences is reducible to misattribution of nonspecific activation to a salient judg- patterns more efficiently than any other pattern (En- ment dimension (Mandler, Nakamura, & Van Zandt, 1987). Instead, quist & Arak, 1994; Johnstone, 1994). Based on these they suggest that the affective reaction induced by fluency is hedoni- findings, we propose that symmetrical patterns may be cally positive.

369 REBER, SCHWARZ, WINKIELMAN than low-contrast stimuli, regardless of stimulus pre- Perceiver’s History With the Stimulus sentation time. Conversely, if perceptual fluency is the Psychological research on factors underlying beau- key to preference, figure–ground contrast should be ty has also investigated the history of a perceiver’s ex- most influential during short presentation times, when perience with the stimulus. This research points to the high contrast can facilitate processing, and least influ- role of repeated exposure, implicit learning of stimulus ential during long presentation times, when stimuli can structure, and prototypicality. Again, the influence of be easily processed regardless of their contrast level. these variables can be traced to their influence on pro- R. Reber and Schwarz (2001) tested this reasoning in a cessing fluency. study where figure–ground contrast was manipulated for circles presented for 0.3, 1, 3, and 10 sec. As pre- Repeated Exposure dicted by the fluency hypothesis, figure–ground contrast influenced aesthetic judgments only at short expo- Repeated exposure to a stimulus results in more sure durations, but not at the duration of 10 sec. Note favorable evaluations, a phenomenon known as the that a theory that assumes that some objective feature mere exposure effect (Zajonc, 1968, 1998). Mere ex- of the stimulus per se, such as figure–ground contrast, posure effects have been obtained with a variety of determines preference cannot account for this finding. stimuli (faces, ideographs, words, melodies) and a va- After all, different exposure times do not change the riety of measures (judgments of preference, behavioral objective features of the stimulus. If anything, the ob- choices, physiological responses), indicating the ro- jective features of the stimulus are most clearly percep- bustness of the phenomenon (for a meta-analysis see tible at long exposure times—yet, they fail to influence Bornstein, 1989; for mere-exposure effects with works judgments under this condition. of art, see Leder, 2001). Previously seen stimuli differ from novel stimuli Summary with regard to at least three fluency-related parameters. First, familiar stimuli are processed faster than novel In sum, a diverse body of research shows that objec- stimuli (e.g., Haber & Hershenson, 1965; Jacoby & tive features of stimuli—like amount of information, Dallas, 1981). Second, familiar stimuli elicit less at- symmetry, or figure–ground contrast3—influence per- tentional orienting than novel stimuli (Desimone, Mil- ceptual fluency as well as preference judgments. More ler, Chelazzi, & Lueschow, 1995). Third, familiar stim- important, knowledge of the objective features per se is uli have more organized processing dynamics than not sufficient to predict an object’s evaluation. Instead, novel stimuli (Lewenstein & Nowak, 1989; Norman & the crucial variable appears to be processing fluency. O’Reilly, 2001; E. R. Smith, 2000). Based on such Although fluency typically covaries with objective fea- findings, several researchers suggested that perceptual tures, fluency is the predictive variable when both are fluency is central to the mere exposure effect and pro- juxtaposed. Thus, identical patterns are rated more fa- vided evidence consistent with this account (e.g., vorably when presented with vertical rather than hori- Bornstein & D’Agostino, 1994; Jacoby, Kelley, & zontal symmetry (Palmer, 1991), consistent with the Dywan, 1989; Seamon, Brody, & Kauff, 1983; observation that vertical symmetry facilitates process- Whittlesea, 1993; Whittlesea & Price, 2001). ing (e.g., Royer, 1981). Similarly, high contrast en- In our reading, a perceptual fluency account of mere hances liking for patterns shown briefly, but not for exposure effects is compatible with Zajonc’s (1968, identical patterns shown longer (R. Reber & Schwarz, 1998) original account. Specifically, Zajonc proposed 2001). Moreover, objectively identical stimuli are eval- that the mere exposure effect reflects a precognitive uated more favorably when their processing is facili- mechanism that ensures caution in encounters with tated through priming procedures (R. Reber et al., novel, and potentially harmful, stimuli and the gradual 1998; Winkielman & Fazendeiro, 2003). In light of extinction of this caution over repeated encounters, these results, we propose that the influence of objective consistent with the observation of a “fear of the un- features is mediated by their influence on processing known” in a variety of species (for a review see Hill, fluency. 1978). Given that processing fluency can often be assessed before a stimulus is identified with certainty, 3 One well-known objective feature that we did not address is the fluency may indeed serve as a very early indicator golden section. The possibility that people have a preference for the golden section is one of the oldest and most studied phenomena in of stimulus novelty (see Curran, 2000, and Rugg & empirical aesthetics (e.g., Fechner, 1876; for a review, see Green, Yonelinas, 2003, for a discussion of specific mecha- 1995). For example, rectangles are assumed to be most pleasing if nisms). Moreover, the spontaneous affective responses the ratio of the short side to the long side of the rectangle equals the that accompany fluency may serve as a biologically ratio of the long side to the sum of short and long side, which is about functional signal (see Winkielman et al., 2003, for a 1:1.618. However, evidence for preference of the golden section is mixed (Green, 1995; Kubovy, 2000). Further, we are not aware of discussion). We therefore consider both accounts com- any work testing how golden section influences processing fluency. patible, but note that a fluency account provides a more Accordingly, we do not address the phenomenon. general conceptualization that can accommodate the

370 FLUENCY AND AESTHETIC PLEASURE influence of priming procedures as well as the influ- Buchner (1994) first exposed participants to grammati- ence of repeated exposure. In addition, a fluency ac- cal letter strings and subsequently presented strings count predicts under which conditions mere exposure with a perceptual clarification procedure, asking par- effects should not be observed, namely when partici- ticipants to identify the strings. Specifically, the test pants are aware of the source of fluency (cf. Bornstein, strings were presented in a black mask, from which 1989). We return to this issue in a later section, when pixels were gradually removed at random until the par- we discuss the moderating role of attributions and ticipant could identify the letter string. Participants re- expectations. acted faster to grammatical rather than ungrammatical letter strings, indicating that grammatical letter strings are easier to process. Implicit Learning In combination, the available findings again indi- of Stimulus Structure cate that increased preference is associated with in- Research into the effects of repeated exposure is creased fluency: “Regular” or “grammatical” stimuli closely related to research into the relation between are not only preferred over “irregular” or “ungrammat- preferences and the implicit learning of stimulus struc- ical” ones, but also easier to process. ture. In one experiment, Gordon and Holyoak (1983) used letter strings constructed in accordance with an Prototypicality artificial finite state grammar (see A. S. Reber, 1967, Martindale (1984) proposed that prototypical forms 1993). In a learning phase, participants were exposed are preferred over nonprototypical forms—a proposal to grammatical letter strings. In a subsequent test that is closely related to the idea that people prefer “av- phase, participants received novel grammatical and un- erage” stimuli (Rhodes & Tremewan, 1996).4 Numer- grammatical strings and indicated their liking for these ous studies confirm that prototypical and “average” strings. As predicted, grammatical letter strings were forms are preferred over nonprototypical ones. For liked more than ungrammatical ones (see also Manza, example, several studies found a positive relation Zizak, & Reber, 1998). In a second experiment, between prototypicality and aesthetic evaluations in Gordon and Holyoak (1983) were able to show that color patches (e.g., Martindale & Moore, 1988), paint- participants’liking of complex visual patterns was neg- ings (e.g., Hekkert & van Wieringen, 1990), and furni- atively related to pattern distortion from an acquired ture (e.g., Whitfield & Slatter, 1979). Preference for standard; the less distorted the pattern was in compari- prototypicality has also been found in music. For ex- son to the standard, the more it was liked. ample, J. D. Smith and Melara (1990) showed that pro- In a series of experiments, Newell and Bright totypical chord progressions were preferred by novices (2001) presented grammatical letter strings in an en- (though not by experts, as we discuss later). Similarly, coding phase and then assessed grammaticality and Repp (1997) mixed music performances of different liking ratings for both grammatical and ungrammatical performers into an averaged performance, and found items at test. In all three experiments, there was a con- that such performance was highly rated. sistent effect of grammaticality on liking, but only if A number of studies show that prototypical faces encoding and test conditions stayed the same, ensuring are preferred over nonprototypical faces (e.g., fluent processing. In contrast, grammaticality judg- Langlois & Roggman, 1990; Rhodes & Tremewan, ments were not influenced by inconsistencies between 1996). Theoretical explanations for this robust effect encoding and test. Newell and Bright concluded that have typically focused on the idea that organisms are grammaticality judgments reflected attempts to explic- biologically predisposed to interpret prototypicality itly recall information about training items, whereas as a cue to mate value (Symons, 1979). For example, the effect on liking was based on an (mis)attribution of Thornhill and Gangestad (1993, 1999) suggested that fluency. facial prototypicality signals health, lending individu- Sollberger and Reber (2004) demonstrated effects als with a preference for facial prototypicality in mate of implicit learning on liking in the domain of music. They constructed tonal music sequences in accordance with a finite state grammar. During a learning phase, 4Although the concept of “averageness” can refer to an objective participants were exposed to grammatical tone se- feature (i.e., the stimulus represents the arithmetical mean of all exemplars in the population), in most studies averageness is closely re- quences. During the test phase, they were given old lated to prototypicality (i.e., perceiver’s mental representation of the grammatical, novel grammatical, and novel ungram- best or most typical instance of a category). That is, when partici- matical sequences. Participants liked old and novel pants are given stimuli that are average (e.g., they were created by grammatical sequences more than ungrammatical se- morphing multiple exemplars together), usually those stimuli are quences, supporting the notion that grammatical stim- prototypical (e.g., participants also see the contributing exemplars and form a mental representation of the category). Though it is theo- uli are pleasant. retically possible to distinguish the “objective” population average Importantly, research shows that grammatical stim- from a more idiosyncratic psychological prototype, few studies ex- uli are processed with higher fluency. For example, amine distinct contributions of these two factors.

371 REBER, SCHWARZ, WINKIELMAN selection a selective advantage in the evolutionary and Vrana (1998) observed that repeated exposure past. However, recent research casts doubt on this as- increased liking up to nine exposures. After 27 expo- sumption. Some studies failed to document a relation sures, liking increased if the salience of the repetition between facial prototypicality and health (Kalick et scheme was low, but decreased if repetition salience al., 1998). More important, innate preferences for was high. Other studies found that mere-exposure prototypical faces should not necessarily lead to pref- effects were more easily obtained for complex than for erence for other prototypical objects. Yet, several simple visual stimuli (Bornstein, Kale, & Cornell, studies using different methodologies show that peo- 1990), although the latter are easier to process. ple also prefer prototypical dogs, watches, and birds To account for such findings within a perceptual (Halberstadt & Rhodes, 2000, 2003). Thus, there are fluency account, it is useful to consider factors that de- reasons to look for another explanation of the pro- termine (a) the elicitation of fluency-related subjective totypicality preference. Again, perceptual fluency is a experiences and (b) the use of these experiences in promising candidate. judgment. A large body of literature in cognitive psychology indicates that prototypical stimuli are processed more Processing Expectations easily than nonprototypical ones (see Posner & Keele, 1968; Rumelhart, McClelland, & the PDP Research According to the discrepancy–attribution hypothe- Group, 1986; E. E. Smith, Shoben, & Rips, 1974). Fol- sis (Whittlesea & Williams, 1998, 2000), fluency asso- lowing up on this literature, we conducted a study in ciated with processing a certain event is more likely to which participants first studied several visual dot pat- elicit a subjective experience (pleasure, familiarity, terns. The patterns were random, but were constructed etc.) if the fluency is unexpected in light of the person’s to converge on a prototype. In a later test phase, partici- processing expectations, which constitute a “norm” for pants (a) preferred the prototype over other previously the event (Kahneman & Miller, 1986). For example, unseen patterns; (b) showed stronger EMG responses seeing your dentist (a fluent stimulus) at the dental of- from the zygomaticus region, indicating more positive fice does not generate a strong feeling of familiarity, affect; and (c) falsely recognized the prototype as hav- whereas unexpectedly seeing the same dentist at the ing been shown previously, indicating high fluency airport does.5 According to the discrepancy–attribu- (Fazendeiro & Winkielman, 2003). Moreover, P. J. tion hypothesis, salient causes of fluency (e.g., an obvi- Reber, Stark, and Squire (1998) showed that exposure ous repetition scheme, a very simple pattern, predictive to several exemplars of dot patterns from the same cat- context) allow participants to formulate accurate ex- egory resulted in more fluent visual processing of the pectations regarding the processing fluency of the tar- prototype, as reflected in decreased activity in the pos- get stimuli and thus reduce the likelihood that fluency terior occipital cortex (see Aizenstein et al., 2000, for a will elicit a subjective experience (Whittlesea & Wil- related finding). liams, 2000). Nevertheless, many fluent stimuli may In sum, we propose that the robust relation between continue to elicit a pleasant experience even when flu- prototypicality and attractiveness may be traced to pro- ency of processing is expected. For example, we sus- cessing fluency: Prototypical stimuli are easier to pro- pect that people continue to enjoy prototypical faces, cess and hence evaluated more positively, consistent symmetrical patterns, harmonious chords, and high with the observation that any other variable that in- clarity drawings even after they formed fairly accurate creases processing fluency also increases liking. processing expectations for these stimuli. Schwarz (2004a, 2004b) suggested that these di- verging intuitions reflect that expectations have a dual Moderating Variables: influence: On the one hand, unexpected fluency is Expectations and Attributions more likely to capture attention, resulting in a con- scious experience as suggested by the discrepancy-at- Next, we turn to variables that may moderate this re- tribution hypothesis. On the other hand, expectations lation and address a number of potential complica- entail an attribution to the variable that gave rise to the tions. To date, moderating variables have received expectation in the first place. When this variable is ir- most attention in research on the mere exposure effect relevant to the evaluation of the stimulus (as is the case (Zajonc, 1968). As reviewed earlier, this research showed that repetition elicits a positive mood (Mona- han et al., 2000) and increases preference for stimuli 5This possibility is elegantly illustrated by experiments showing (Bornstein, 1989). However, this research has also that pseudohomophones (letter strings that sound like real words, identified important qualifications. Some studies such as “phraug”) and orthographically regular nonwords (e.g., “hension”) elicit illusions of familiarity and are rated as pleasant found that liking initially increased with the number of (Whittlesea & Shimizu, 2001). Presumably, these stimuli elicit cog- presentations, followed by a decline in liking (e.g., nitive and affective reactions because they are processed faster than Kail & Freeman, 1973). For example, Van den Bergh participants expect.

372 FLUENCY AND AESTHETIC PLEASURE for repetition schemes or predictive contexts), the ex- Stimulus Complexity perience is considered uninformative and not used An important challenge for our fluency account is informing a judgment, as discussed later. When the why people sometimes prefer complex over simple variable is relevant to the evaluation (as is the case for stimuli. After all, a simple stimulus (e.g., a single line symmetry, prototypicality, and so on, which are char- or one musical note) should be easier to process than a acteristics of the stimulus itself), the experience is con- complex stimulus (e.g., a pattern of lines or a chord), sidered informative and used as a basis of judgment. but the latter are commonly preferred. We propose that This conjecture, derived from the feelings-as-informa- several processes may contribute to a relative prefer- tion model (Schwarz & Clore, 1996), awaits empirical ence for complexity. testing. Expectations and Attributions Attributional Processes As discussed in the preceding section, the ability of fluency to elicit an experience, and the perceiver’s will- Once an affective experience is elicited by a fluent ingness to use that experience in forming a judgment, stimulus, its impact on preference judgments is moder- depends on expectations and attributions. Simple stim- ated by attributional processes (e.g., Bornstein & uli allow the perceiver to form more accurate process- D’Agostino, 1994; Van den Bergh & Vrana, 1998). In ing expectations, reducing the strength of the fluency- general, individuals only rely on experiential informa- based experience. Further, simple stimuli provide the tion when it seems to bear on the target of judgment, perceiver with an obvious attribution for the experi- but not when they are aware that it may reflect the in- ence. Consistent with these ideas, a common view in fluence of an unrelated variable (Schwarz & Clore, aesthetics holds that highest beauty is attained by “uni- 1996). Thus, people should only rely on fluency-based formity in variety,” or “simplicity in complexity,” as experiences when their informational value is not when a complex theme is presented in an accessible called into question. When it becomes apparent that way (see Dickie, 1997). This phenomenon can be un- their experience may merely reflect the influence of derstood by conceptualizing uniformity or simplicity high repetition or a preceding prime, the experience is as a source of fluency, and variety or complexity as a discounted and perceivers switch to alternative inputs source of processing expectations. That is, when pro- to form a judgment (e.g., Bornstein & D’Agostino, cessing is expected to be difficult, yet turns out to be 1994; Van den Bergh & Vrana, 1998). This assumption easy, it creates a particularly strong experience of aes- is consistent with (mis)attribution effects observed in thetic pleasure (see Gombrich, 1984, for a related the exploration of other sources of experiential infor- discussion).6 mation, like moods (e.g., Schwarz & Clore, 1983; for a A similar logic may explain why stimulus complex- review see Schwarz & Clore, 1996), ease of recall (e.g., ity is often related to preference by an inverted U- Schwarz et al., 1991; Winkielman, Schwarz, & Belli, shaped function (e.g., Berlyne, 1971; Vitz, 1966). With 1998; for a review see Schwarz, 1998), or familiarity low levels of complexity, the source of fluency is very (Fazendeiro, Winkielman, Luo, & Lorah, in press). salient. As complexity increases, the salience of the Consistent with these earlier findings, we observed source of perceptual fluency decreases, enhancing the that the impact of processing fluency on judgment is misattribution of fluency to beauty. However, further eliminated when participants attribute their affective increases in complexity will eventually reduce pro- reactions to an irrelevant source, such as a background cessing fluency, leading to a decrease in perceived music (e.g., Winkielman & Fazendeiro, 2003), as re- beauty. These mechanisms would combine to form a viewed earlier. U-shaped relation between complexity and beauty, as predicted and found by Berlyne (1971).

Challenges and Conjectures Different Sources of Processing Facilitation In this section, we address various challenges that a Perceptual versus conceptual fluency. Complex- fluency account of aesthetic pleasure needs to handle. ity may sometimes be preferred because it facilitates We first discuss why people sometimes prefer complex access to the meaning of the stimulus. That is, a de- rather than simple stimuli, including preference for crease in perceptual fluency due to complexity may be complex stimuli by experts. We then turn to the ques- outweighed by an increase in conceptual fluency due to tion of how stimulus valence may moderate fluency effects on preference. The empirical evidence bearing on these challenges is often sparse and sometimes contra- 6Similar processes may underlie Eysenck’s (1941, 1942) obser- dictory. We review what is known and offer some con- vations that order (source of fluency) and complexity (source of ex- jectures and recommendations for future research. pectations) interact in determining aesthetic pleasure.

373 REBER, SCHWARZ, WINKIELMAN meaningfulness. Consistent with this idea, Martindale, Melara, 1990). Do these observations imply that on- Moore, and Borkum (1990) found that meaningfulness ly novices’ preferences are influenced by processing was a better predictor of aesthetic preference than fluency? complexity per se. In our reading, two different processes are likely to Future research may fruitfully address the relative contribute to these differential preferences. First, train- contribution of fluency to the subjective experiences of ing in the arts presumably increases the fluency with the perceiver at various stages of processing. In a first which complex art stimuli can be processed. Repeated step toward this question, R. Reber et al. (2004) exam- exposure to complex stimuli results in higher percep- ined how processing speed contributes to the subjec- tual fluency, and training in arts gives meaning to com- tive experience of “ease” at various perceptual stages. plex structures in paintings, poems, or music, which Participants were presented with words in high and low results in an additional increase in processing ease. Ac- figure–ground contrast, printed in more and less read- cordingly, the observation that experts have higher able fonts. The participants’ task was to either detect preference for complex stimuli than novices may itself the appearance of these words, or to identify them as reflect fluency effects. Second, experts are more likely quickly as possible. As expected, high contrast sped up than novices to consider aesthetic value, the ideas be- performance on a detection task, reflecting early pro- hind the work, and the norms of “good” and “bad” taste cessing stages, but not on the identification task, re- (Bourdieu, 1979/1987; Gombrich, 1995). As a result, flecting later processing stages. On the other hand, they may evaluate simple stimuli more negatively than high font readability sped up performance on the iden- novices, despite the pleasure they receive from easy tification task, but not on the detection task. Interest- processing. ingly, both high figure–ground contrast and high font As Wilson et al. (1993) demonstrated, such an ana- readability enhanced judgments of the subjective ease lytic approach to aesthetic preference bears the risk of of perception. This finding suggests that facilitation of making poor choices. In their studies, participants who the early stimulus detection stage and the later stimulus had to explain why they liked a poster subsequently se- identification stage of processing can jointly contribute lected posters from which they derived less pleasure to the experience of ease. Future studies should address once they had it at home than participants who could how facilitation at various processing stages contrib- rely on their gut preference. Similarly, experts may run utes to liking. the risk of acquiring art that they find less enjoyable than expected, once they no longer interact with it in Objective simplicity versus cognitive ease. Fi- the thoughtful manner that led to their choice. How- nally, objectively simpler stimuli are not always easier ever, our processing fluency account provides some on the cognitive system. In the visual domain, complex hope for these experts: Extended exposure to their new shapes often have higher redundancy and thus are rec- treasure may eventually result in a preference that is ognized faster than simple shapes (Biederman, Hilton, driven by processing fluency instead of aesthetic value. & Hummel, 1991). In the auditory domain, recognition In a similar vein, in the history of western music, more of a melody is facilitated through multiple embedding complex musical intervals were gradually introduced of redundant patterns (Gentner & Hulse, 1998). Empir- and then accepted by an initially reluctant audience, ical studies and connectionist models suggest that rec- potentially reflecting increased fluency through re- ognition is often easier when a pattern is embedded in a peated exposure. In fact, Roederer (1973) suggested larger context (e.g., the letter “r” by itself or in the con- that more complex musical intervals require more text of a word), with different levels of analysis sim- complex information processing, a skill that the audi- ultaneously supporting each other (Rumelhart & ence acquires with increasing familiarity. McClelland, 1982). In short, simplicity per se does not necessarily imply ease of processing. Hence, it is al- Stimulus Valence ways necessary to empirically examine the actual ease of processing for simple and complex stimuli. Central to our proposal is that processing fluency is experienced as positive and that this experience, in turn, results in more favorable judgments. As our re- Experts and Novices view indicates, this effect has been observed for stimuli Numerous studies demonstrated profound differ- that are largely content-free (abstract designs, novel ences in the aesthetic preferences of novices and ex- melodies, nonwords, etc.), as well as for stimuli that perts. In general, people without art training prefer are highly meaningful (everyday objects, faces, dogs, simple and symmetric visual elements, whereas people birds, watches, guns, words, sentences, etc.). This ob- with art training prefer complex and asymmetric visual servation, however, does not imply that stimulus con- elements (e.g., McWhinnie, 1968). Similarly, music tent is irrelevant, and we would certainly expect that novices prefer prototypical chord progressions, where- stimuli with positive content are evaluated more favor- as experts do not show this preference (e.g., Smith & ably than stimuli with negative content. But how does

374 FLUENCY AND AESTHETIC PLEASURE stimulus content interact with fluency? Does fluency Finally, some studies investigated the effects of rep- enhance the evaluation of positive as well as negative etition on stimuli presented in differentially valenced stimuli? contexts. In a positive context, mere exposure gener- To date, the role of stimulus valence has mostly ally increases preference for old compared to new received attention in mere-exposure research. As re- stimuli (Burgess & Sales, 1971; Saegert, Swap, & viewed earlier, most studies indicate that repetition of Zajonc, 1973). In a negative context, however, the pic- initially neutral or positive stimuli enhances preference ture is mixed. Saegert et al. (1973) found that personal for these stimuli (see Bornstein, 1989, for a review). attraction increased, but Burgess and Sales (1971) ob- However, studies with initially negative stimuli pro- served that increased exposure to nonsense words in a duced mixed findings. Some authors found that re- negative context resulted in a decrease in preference. peated exposure enhanced preference for initially These findings presumably reflect that the previous negative stimuli (Zajonc, Markus, & Wilson, 1974), pairings of the stimulus with a negative context re- whereas others observed that repeated exposure to ini- sulted in more negative associations, which were suffi- tially negative stimuli decreased preference for these cient to override the positive influence of fluency per stimuli (Brickman, Redfield, Harrison, & Crandall, se. Unfortunately, the procedures used do not permit 1972; Grush, 1976; Klinger & Greenwald, 1994; the separation of these effects. Swap, 1977). In our reading, mixed findings are to be As the reviewed examples illustrate, it is difficult to expected for several reasons. separate the relative contributions of fluency-based af- First, fluency manipulations like repetition or high fective reactions and fluency-related changes in per- figure–ground contrast are likely to have multiple ef- ceived stimulus meaning and associations. Developing fects. On one hand, they facilitate fluent processing, methodologies that can shed some light on this issue is which elicits a positive affective response (Winkiel- therefore an important next step. One promising strat- man & Cacioppo, 2001). On the other hand, they en- egy was used by Halberstadt (personal communica- hance the mental representation of the stimulus and fa- tion, March 2004), based on relative attractiveness cilitate the extraction of stimulus meaning. Suppose, judgments. Specifically, Halberstadt showed partici- for example, that you are repeatedly exposed to a nega- pants a variety of negative objects and assessed judg- tive item, like a somewhat blurred picture of a rotting ments of the relative attractiveness of a given object. carcass. With each repetition, the picture is easier to His results showed that highly prototypical negative perceive, which by itself is experienced as positive. objects, such as guns, were rated as more “attractive” But each repetition also increases access to the item’s than nonprototypical negative objects, paralleling the negativity, resembling a shift from viewing the rotten prototypicality effect observed for neutral and positive carcass through a fogged window to viewing it through objects (Halberstadt & Rhodes, 2000). Thus, it appears a clear window. This increased access to the stimulus’ that fluency due to prototypicality increases attractive- negativity, and the resulting affective response may ness irrespective of stimulus valence. eclipse any initial positive reactions due to higher fluency. Similarly, when a sinner views a perceptually fluent image of hell, the fear from the prospect of eternal Implications of a Processing Fluency damnation may well override the pleasure from ease of Theory of Beauty perception. Second, fluent processing of initially negative stim- We now turn to a discussion of the implications of uli may also elicit negative evaluations via meta-cogni- our theoretical framework and show how it can accom- tive inferences (Skurnik, Schwarz, & Winkielman, modate preference phenomena that are problematic for 2000). For example, multiple repetitions of the name of other theoretical approaches. a notorious murderer may increase perceivers’ feeling of name familiarity, as observed in Jacoby, Kelley, Nature Versus Nurture Brown, and Jasechko’s (1989) “false fame” effect. This in Aesthetic Preferences “false fame” may then lead the participant to infer that the criminal must be particularly vicious, or why else Our theoretical proposal allows for a conceptual in- would the name seem so familiar (Klinger & Green- tegration of findings bearing on the aesthetic prefer- wald, 1994)? Again, this inference may override any ences of young children and the influence of socializa- positive reaction due to the higher fluency of the stimu- tion. This issue has received some attention in the lus. One paradoxical, and not yet tested, implication of psychology of music. this conjecture is that names and pictures of highly Medieval music theorists described tone combina- negative individuals (e.g., Stalin or Hitler) may tions with simple frequency ratios as naturally pleasing initially elicit brief fluency-based positive responses, rather than displeasing (see Schellenberg & Trehub, which are quickly overridden by responses based on 1996). In line with this assumption, Zentner and Kagan meaning. (1996) observed that 4-month-old infants looked lon-

375 REBER, SCHWARZ, WINKIELMAN ger at the source of music, and showed less motor ac- dimentaryface-likefeatures(Turati,2004).Becauseav- tivity, when the music involved consonant sounds (i.e., eragefacesbestmatchtheinnateprocessingbiases,they those with simple numerical frequency ratios, like 1:2 are processed most fluently and as a result are preferred or 2:3) rather than dissonant sounds, presumably indi- by newborns (see Bednar & Miikukulainen, 2000, for a cating an early preference (see Trainor & Heinmiller, comprehensive discussion and a computational model). 1998, for related findings). However, the emergence of However, as infants mature, they are exposed to many very different musical tastes in adults indicates a pow- different exemplars of a category and form new proto- erful role of socialization and experience. For example, types, often very quickly (Walton & Bower, 1993). This people without musical training prefer consonant to acquisition of new prototypes results in preferences for dissonant sequences, whereas people with musical new types of “average” faces (Rubenstein, Langlois, & training, who had more exposure to dissonant se- Kalakanis, 1999). What will eventually end up being quences, do not show this preference (Francès, 1987). “prototypical” for an adult will depend on the person’s A processing fluency account can accommodate culture and social class, resulting in individual differ- both sets of findings. From this perspective, one source ences and beauty standards that vary somewhat across of fluency is found in biologically given perceptual cultures. Hence, from the perspective of a perceptual mechanisms that are more attuned to certain stimulus fluency account, the nature versus nurture debate in fa- organizations. In a study by Schellenberg and Trehub cial preferences (as in music) can be resolved by assum- (1996), infants showed better detection of subtle ing that aesthetic preference is a function of fluency, changes to harmonic intervals if the combinations of which, in turn, is determined by both biological equip- tones were related by simple rather than complex fre- ment and socialization. quency ratios. This finding is in line with the assump- In sum, a perceptual fluency account can provide a tion that simple frequency ratios are easier to process conceptual integration of the apparent contradiction than complex frequency ratios (Roederer, 1973). A between early and acquired preferences by tracing second source of fluency is stimulus repetition and the both to fluency of processing. It therefore has an ad- implicit learning of stimulus regularities (e.g., Buch- vantage over theories that emphasize objective features ner, 1994; Haber & Hershenson, 1965; Jacoby & Dal- of beauty, which cannot account for effects of social- las, 1981; A. S. Reber, 1993; Whittlesea et al., 1990). ization and suggest fixed preferences across the life Given these two sources of fluency, one can under- span. Fluency theory also has an advantage over social- stand why there are few interindividual differences in ization theories, which cannot account for the system- music appreciation by infants (Schellenberg & Trehub, atic nature of early preferences. 1996; Zentner & Kagan, 1996), yet quite remarkable interindividual differences in music appreciation by adults Prototypical Versus Exaggerated (Francès, 1987). Specifically, infants come equipped and Abstracted Forms withsimilarperceptualmechanismsthatresultinhigher In apparent contradiction to a common preference processing fluency for specific classes of stimuli. As for prototypicality, two popular techniques in art are ex- they mature, repeated exposure to various examples of aggeration and abstraction (Ramachandran & Hirstein, musicoftheircultureandsocialclassresultsinmoreflu- 1999). A perceptual fluency theory may again shed light ent processing for familiar and regular musical ele- on this paradox. On the one hand, prototypical stimuli ments. In fact, Tillmann, Bharucha, and Bigand (2000) are easier to process than nonprototypical ones, result- presented a model showing how people may implicitly ing in higher preference. On the other hand, under some learnmusicalstructuresoftheirownculture,facilitating conditions, exaggeration and abstraction may facilitate the perception of music that obeys the rules of harmony processing even more by emphasizing central features within this culture. In sum, the debate about nature ver- and discarding others. This is suggested, for example, sus nurture in music appreciation is mistaken from the by the initially paradoxical finding that caricatures of perspective of a perceptual fluency account. Instead, faces are sometimes better recognized than undistorted this account suggests that aesthetic preferences always images (e.g., Mauro & Kubovy, 1992; Rhodes, Bren- depend on fluency, which, in turn, is determined by both nan, & Carey, 1987). Future research may address this biological equipment and socialization. hypothesis directly and may fruitfully relate it to the dis- A similar discussion emerged in research on facial cussion of preferences for prototypical versus exagger- preferences in infants. It has been shown that newborns ated facial stimuli (see Perrett, May, & Yoshikawa, prefer attractive faces (Slater et al., 1998). As discussed 1994; Rhodes & Tremewan, 1996).7 earlier, attractive faces are close to the mathematical average of the population (Langlois & Roggman, 1990). But how can infants prefer a population average if they 7Some effects of exaggeration on attractiveness may reflect the have seen only very few exemplars of the category? This operation of other processes, such as appeal of indicators of sexual finding can be explained by assuming the existence of dimorphism and sex-typical traits (Perrett et al., 1998). For example, innateprocessingbiasesthatfacilitaterecognitionofru- heterosexual men tend to prefer women with fuller than average lips.

376 FLUENCY AND AESTHETIC PLEASURE

Beauty and Truth psychological processes. In fact, there is growing empirical evidence that people use a common source for Finally, our processing fluency proposal sheds new evaluations of both beauty and truth—processing flu- lightonwhyartistsandscientistsalikehavelongconsid- ency. For example, stimulus repetition increases judg- ered beauty and truth as two sides of the same coin. In ments of liking (e.g., Bornstein, 1989; Zajonc, 1968) as Greek and medieval art, it was considered impossible well as judgments of truth (e.g., Arkes, Hackett, & that something untruthful could be beautiful (see Eco, Boehm, 1989; Begg et al., 1992; Brown & Nix, 1996; 1988; Tatarkiewicz, 1970). Even in 19th century poetry, Hasher, Goldstein, & Toppino, 1977). Similarly, fig- beauty and truth were seen as two sides of the same coin, ure–ground contrast increases judgments of liking asexpressedbyKeats’famousassertion“beautyistruth, (e.g., R. Reber et al., 1998) as well as judgments of truth is beauty” (McGlone & Tofighbakhsh, 1999, truth. For example, R. Reber and Schwarz (1999) ob- p. 240). It was only later that beauty and truth were served that the same statements were more likely to be viewed as separate attributes in art. judged true when presented with high rather than with A similar identification of beauty and truth can be low figure–ground contrast. Similarly, McGlone and seen in the philosophy of science. For a long time, an Tofighbakhsh (2000) demonstrated that the same prop- important aspect of the acceptability (i.e., truth) of a ositions are more likely to be perceived as true when scientific hypothesis has been its simplicity (cf. Hem- presented in a rhyming rather than a nonrhyming form. pel, 1966). This view has a long tradition and some the- Interestingly, a rhyming form also enhances the flu- orists suggested that the aesthetic value of simplicity ency with which statements are understood (e.g., motivated ancient cultures to begin to classify their Meyer, Schvaneveldt, & Ruddy, 1975, Rubin, 1995) flora and fauna, regardless of the survival needs of the and is used in poetry to increase the reader’s pleasure tribe (see Lévi-Strauss, 1962/1966). Hence, aesthetic (McGlone & Tofighbakhsh, 1999; Vendler, 1997). In criteria seemingly stood at the very beginnings of sci- combination, these findings suggest that judgments of entific inquiry. Recently, philosophers of science, par- beauty and intuitive judgments of truth may share a ticularly those concerned with the complexities of evo- common underlying mechanism. Although human rea- lutionary theory, began to cast doubt on the value of son conceptually separates beauty and truth, the very universally applying the simplicity principle because it same experience of processing fluency may serve as a may preclude the acceptance of more accurate, but less nonanalytic basis for both judgments. “elegant” theories (Sober, 1994). Despite these concerns, it appears that at least some scientists use the beauty of a theory as a guide to its Summary and Conclusions truth. The Nobel price-winning physicist Chandrasek- har has exemplified this thinking by quoting the mathe- In sum, we propose that aesthetic pleasure is a func- matician Weyl: “My work always tried to unite the true tion of the perceiver’s processing dynamics: The more with the beautiful; but when I had to choose one or the fluently the perceiver can process an object, the more other, I usually chose the beautiful” (Chandrasekhar, positive is his or her aesthetic response. This proposal 1987, p. 65). Chandrasekhar continued: entails four specific assumptions. First, objects differ in the fluency with which they can be processed. Fea- “the example which Weyl gave was his gauge theory tures that facilitate fluent processing include all the of gravitation, which he had worked out in his Raum- core features identified in objectivist theories of Zeit-Materie. Apparently, Weyl became convinced that this theory was not true as a theory of gravitation; beauty, like goodness of form, symmetry, and fig- but still it was so beautiful that he did not wish to aban- ure–ground contrast, as well as variables that have not don it and so he kept it alive for the sake of its beauty. received attention in traditional theories of aesthetic But much later, it did turn out that Weyl’s instinct was pleasure, like perceptual and conceptual priming pro- right after all, when the formalism of gauge invariance cedures. Second, processing fluency is itself hedoni- was incorporated into quantum electrodynamics.” cally marked and high fluency is subjectively experi- (Chandrasekhar, 1987, p. 65f) enced as positive, as indicated by psychophysiological findings. Third, the affective response elicited by pro- Of course, this anecdote does not imply that beautiful cessing fluency feeds into judgments of aesthetic ap- theories are true—there may have been as many beau- preciation, unless the informational value of the expe- tiful theories that turned out to be wrong after all. Yet, it rience is called into question. Finally, the impact of does illustrate that some scientists believe that when a fluency is moderated by expectations and attribution. scientific proposition cannot be decomposed analyti- On the one hand, fluency has a particularly strong im- cally, its truth can be revealed nonanalytically or “intu- pact when its source is unknown and fluent processing itively,” by assessing its beauty. comes as a surprise. On the other hand, the fluency- Given this parallelism of beauty and truth in art and based affective experience is discounted as a source of science, one may ask if both judgments involve similar relevant information when the perceiver attributes the

377 REBER, SCHWARZ, WINKIELMAN experience to an irrelevant source. One of the strengths Beardsley, M. C. (1981). Aesthetics: Problems in the philosophy and nd of our proposal is its ability to integrate distinct phe- criticism (2 ed.). Indianapolis, IN: Hackett Bednar, J. A., & Miikkulainen, J. A. (2000). Self-organization of in- nomena under a common theoretical framework. nate face preferences: Could genetics be expressed through Whereas the basic fluency–liking relation is robust learning? In Proceedings of the Seventeenth National Confer- and reliably replicable, several complications may ence on Artificial Intelligence (pp. 117–122). Cambridge, MA: emerge that deserve further investigation. First, vari- MIT Press. ables that influence fluency of processing can also in- Begg, I. M., Anas, A., & Farinacci, S. (1992). Dissociation of pro- fluence the accessibility of the pre-existing semantic cesses in belief: Source recollection, statement familiarity, and the illusion of truth. Journal of Experimental Psychology: Gen- and affective content of the stimulus. This may in- eral, 121, 446–458. crease the impact of stimulus content, potentially over- Berlyne, D. E. (1971). Aesthetics and psychobiology. New York: riding the influence of the fluency-based affective re- Appleton-Century-Crofts. sponse. Second, we conjecture that “pure” fluency Berlyne, D. E. (1974). Studies in the new experimental aesthetics. effects are most likely observed at an early stage of Steps toward an objective psychology of aesthetic appreciation. Washington, DC: Hemisphere. processing and that stimulus content exerts a subse- Biederman, I., Hilton, H. J., & Hummel, J. E. (1991). Pattern good- quent influence. Third, the extent to which people rely ness and pattern recognition. In J. R. Pomerantz & G. R. on their initial fluency-based gut response or on stimu- Lockhead (Eds.), The perception of structure (pp. 73–95). lus content is a function of their processing motivation, Washington, DC: American Psychological Association. as observed with other sources of experiential informa- Birkhoff, G. (1933). Aesthetic measure. Cambridge, MA: Harvard University Press. tion (for a review, see Schwarz & Clore, 1996). In gen- Bornstein, R. F. (1989). Exposure and affect: Overview and meta- eral, experiential information is more influential under analysis of research 1968–1987. Psychological Bulletin, 106, the heuristic processing style adopted under low moti- 265–289. vation conditions than under the systematic processing Bornstein, R. F., & D’Agostino, P. R. (1994). The attribution and dis- style adopted under high motivation conditions. Dif- counting of perceptual fluency: Preliminary tests of a perceptual fluency/attributional model of the mere exposure effect. ferences in the preferences of experts and novices may Social Cognition, 12, 103–128. be plausibly traced to this variable. These conjectures Bornstein, R. F., Kale, A. R., & Cornell, K. R. (1990). Boredom as a await systematic testing. limiting condition on the mere exposure effect. Journal of Per- Let us turn to a final question. We have seen that sonality and Social Psychology, 58, 791–800. beauty does not rest in the objective features of an ob- Bourdieu, P. (1987). Distinction: A social critique of the judgment of taste (R. Nice, Trans.). Cambridge, MA: Harvard University ject, but derives instead from the processing experi- Press. (Original work published 1979) ences of the perceiver. Is beauty therefore in the eye of Brickman, P., Redfield, J., Harrison, A. A., & Crandall, R. (1972). Drive the beholder? In a sense it is, if folk wisdom thinks of andpredispositionasfactorsintheattitudinaleffectsofmereexpo- the eye as the perceptual processes of the beholder. sure. Journal of Experimental Social Psychology, 8, 31–44. However, beauty “in the eye of the beholder” has often Brown, A. S., & Nix, L. A. (1996). Turning lies into truths: Referen- tial validation of falsehoods. Journal of Experimental Psychol- been contrasted to “objective beauty.” The fluency hy- ogy: Learning, Memory, and Cognition, 22, 1088–1100. pothesis resolves this apparent contradiction: Beauty is Buchner, A. (1994). Indirect effects of synthetic grammar learning in in the processing experiences of the beholder, but these an identification task. Journal of Experimental Psychology: processing experiences are themselves, in part, a func- Learning, Memory, and Cognition, 20, 550–566. tion of objective stimulus properties and the history of Burgess, T. D., & Sales, S. M. (1971). Attitudinal effects of “mere the perceiver’s encounters with the stimulus. Hence, exposure”: A reevaluation. Journal of Experimental Social Psy- chology, 7, 461–472. beauty appears to be “in the interaction” between the Cacioppo, J. T., Petty, R. E., Losch, M. E., & Kim, H. S. (1986). stimulus and the beholder’s cognitive and affective Electromyographic activity over facial muscle regions can dif- processes. ferentiate the valence and intensity of affective reactions. Jour- nal of Personality and Social Psychology, 50, 260–268. Carver, C. S., & Scheier, M. F. (1990). Origins and functions of positive and negative affect: A control-process view. Psychological References Review, 97, 19–35. Chandrasekhar, S. (1987). Truth and beauty. Aesthetics and motiva- Aizenstein, H. J., MacDonald, A. W., Stenger, V. A., Nebes, R. D., tions in science. Chicago: University of Chicago Press. Larson, J. K., Ursu, S., et al., (2000). Complementary category Checkosky, S. F., & Whitlock, D. (1973). The effects of pattern learning systems identified using event-related functional MRI. goodness on recognition time in a memory search task. Journal Journal of Cognitive Neuroscience, 12, 977–987. of Experimental Psychology, 100, 341–348. Arkes, H. R., Hackett, C., & Boehm, L. (1989). The generality of the Clore, G. L. (1992). Cognitive phenomenology: Feelings and the relation between familiarity and judged validity. Journal of Be- construction of judgment. In L. L. Martin & A. Tesser (Eds.), havioral Decision Making, 2, 81–94. The construction of social judgments (pp. 133–163). Hillsdale, Arnheim, R. (1974). Art and visual perception: The new version. NJ: Lawrence Erlbaum Associates, Inc. Berkeley: University of California Press. Curran, T. (2000). Brain potentials of recollection and familiarity. Attneave, F. (1954). Some informational aspects of visual percep- Memory and Cognition, 28, 923–938. tion. Psychological Review, 61, 183–193. Derryberry, D., & Tucker, D. M. (1994). Motivating the focus of at- Bar, M., & Biederman, I. (1998). Subliminal visual priming. Psycho- tention. In P. M. Niedenthal & S. Kitayama (Eds.), The heart’s logical Science, 9, 464–469. eye (pp. 167–196). San Diego, CA: Academic.

378 FLUENCY AND AESTHETIC PLEASURE

Desimone, R., Miller, E. K., Chelazzi, L., & Lueschow, A. (1995). Hasher, L., Goldstein, D., & Toppino, T. (1977). Frequency and the Multiple memory systems in the visual cortex. In M. S. Gaz- conference of referential validity. Journal of Verbal Learning zaniga (Ed.), The cognitive neurosciences (pp. 475–490). Cam- and Verbal Behavior, 16, 107–112. bridge, MA: MIT Press. Hekkert, P., & van Wieringen, P. C. W. (1990). Complexity and Dickie, G. (1997). Introduction to aesthetics. Oxford, England: Ox- prototypicality as determinants of appraisal of cubist paintings. ford University Press. British Journal of Psychology, 81, 483–495. Eco, U. (1988). Art and beauty in the middle ages. New Haven, CT: Hempel, C. G. (1966). Philosophy of natural science. Englewood Yale University Press. Cliffs, NJ: Prentice Hall. Enquist, M., & Arak, A. (1994). Symmetry, beauty and evolution. Hill, W. F. (1978). Effects of mere exposure on preferences in nonhu- Nature, 372, 169–172. man mammals. Psychological Bulletin, 85, 1177–1198. Etcoff, N. (1999). Survival of the prettiest: The science of beauty. Hochberg, J., & McAlister, E. (1953). A quantitative approach to New York: Anchor Books/Doubleday. figural “goodness.” Journal of Experimental Psychology, 46, Eysenck, H. J. (1941). The empirical determination of an empirical 1953. formula. Psychological Review, 48, 83–92. Humphrey, D. (1997). Preferences in symmetries and symmetries in Eysenck, H. J. (1942). The experimental study of the “good ges- drawings: Asymmetries between ages and sexes. Empirical talt”—A new approach. Psychological Review, 49, 344–364. Studies of the Arts, 15, 41–60. Fazendeiro, T., & Winkielman, P. (2003). Effects of conceptual Ingarden, R. (1985). Selected papers in aesthetics (P. J. McCormick, fluency on affective judgments and psychophysiological re- Ed.). Washington, DC: Catholic University of America Press. sponses. Manuscript in preparation. Jacoby, L. L., Allan, L. G., Collins, J. C., & Larwill, L. K. (1988). Fazendeiro, T. A., Winkielman, P., Luo, C., & Lorah, C. (in press). Memory influences subjective experience: Noise judgments. False recognition across meaning, language and stimulus for- Journal of Experimental Psychology: Learning, Memory, and mat: Conceptual relatedness and feelings of familiarity. Mem- Cognition, 14, 240–247. ory and Cognition. Jacoby, L. L., & Dallas, M. (1981). On the relationship between au- Feagin, S. F. (1995). Beauty. In R. Audi (Ed.), The Cambridge dictio- tobiographical memory and perceptual learning. Journal of Ex- nary of philosophy (p. 66). Cambridge, England: Cambridge perimental Psychology: General, 110, 306–340. University Press. Jacoby, L. L., Kelley, C. M., Brown, J., & Jasechko, J. (1989). Be- Fechner, G. T. (1876). Vorschule der Ästhetik [Elements of aes- coming famous overnight: Limits on the ability to avoid uncon- thetics]. Leipzig, Germany: Breitkopf & Härtel. scious influences of the past. Journal of Personality and Social Fernandez-Duque, D., Baird, J. A., & Posner, M. I. (2000). Execu- Psychology, 56, 326–338. tive attention and metacognitive regulation. Consciousness and Jacoby,L.L.,Kelley,C.M.,&Dywan,J.(1989).Memoryattributions. Cognition, 9, 288–307. InH.L.Roediger&F.I.M.Craik(Eds.),Varietiesofmemoryand Francès, R. (1987). The perception of music (W. J. Dowling, Trans.). consciousness: Essays in honour of Endel Tulving (pp. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. 391–422). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Gangestad, S. W., Thornhill, R., & Yeo, R. A. (1994). Facial attrac- Johnstone, R. A. (1994). Female preference for symmetrical males tiveness, developmental stability, and fluctuating asymmetry. as a by-product of selection for mate recognition. Nature, 372, Ethology and Sociobiology, 15, 73–85. 172–175. Garner, W. R. (1974). The processing of information structure. Poto- Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparing re- mac, MD: Lawrence Erlbaum Associates, Inc. ality to its alternatives. Psychological Review, 93, 136–153. Gentner, T. Q., & Hulse, S. H. (1998). Perceptual mechanisms for in- Kail, R. V., & Freeman, H. R. (1973). Sequence redundancy, rating dividual vocal recognition in European starlings (Sturnus dimensions and the exposure effect. Memory & Cognition, 1, vulgaris). Animal Behavior, 56, 579–594. 454–458. Gombrich, E. H. (1984). A sense of order (2nd ed.). London: Kalick, S. M., Zebrowitz, L. A., Langlois, J. H., & Johnson, R. M. Phaidon. (1998). Does human facial attractiveness honestly advertise Gombrich, E. H. (1995). The story of art (16th ed.). London: health? Longitudinal data on an evolutionary question. Psycho- Phaidon. logical Science, 9, 8–13. Gordon, P. C., & Holyoak, K. J. (1983). Implicit learning and gener- Klinger, M. R., & Greenwald, A. G. (1994). Preferences need no in- alization of the “mere exposure” effect. Journal of Personality ferences?: The cognitive basis of unconscious emotional ef- and Social Psychology, 45, 492–500. fects. In P. M. Niedenthal & S. Kitayama (Eds.), The heart’s Green, C. D. (1995). All that glitters: A review of psychological re- eye: Emotional influences in perception and attention (pp. search on the aesthetics of the golden section. Perception, 24, 67–85). New York: Academic. 937–968. Koffka, K. (1935). Principles of Gestalt psychology. London: Rout- Grush, J. E. (1976). Attitude formation and mere exposure phenom- ledge & Kegan Paul. ena: A nonartifactual explanation of empirical findings. Jour- Kubovy, M. (2000). Visual aesthetics. In A. E. Kazdin (Ed.), Ency- nal of Personality and Social Psychology, 33, 281–290. clopedia of psychology (Vol. 8, pp. 188–193). New York: Ox- Haber, R. N., & Hershenson, M. (1965). The effects of repeated brief ford University Press. exposures on growth of a percept. Journal of Experimental Psy- Lang, P. J., Greenwald, M. K., Bradley, M. M. & Hamm, A. O. chology, 69, 40–46. (1993). Looking at pictures: Affective, facial, visceral, and be- Halberstadt, J. B, & Rhodes, G. (2000). The attractiveness of nonface havioral reactions. Psychophysiology, 30, 261–273. averages: Implications for an evolutionary explanation of the at- Langlois, J. H., & Roggman, L. A. (1990). Attractive faces are only tractivenessofaveragefaces.PsychologicalScience,11,285–289. average. Psychological Science, 1, 115–121. Halberstadt, J. B. & Rhodes, G. (2003). It’s not just average faces Leder, H. (2001). Determinants of preference. When do we like what that are attractive: Computer-manipulated averageness makes we know? Empirical Studies of the Arts, 19, 201–211. birds, fish, and automobiles attractive. Psychonomic Bulletin Lévi-Strauss, C. (1966). The savage mind. Chicago: University of and Review, 10, 149–156. Chicago Press. (Original work published in 1962) Harmon-Jones, E., & Allen, J. J. B. (2001). The role of affect in the Lewenstein, M., & Nowak, A. (1989). Recognition with self-control mere exposure effect: Evidence from psychophysiological and in neural networks. Physical Review, 40, 4652–4664. individual differences approaches. Personality and Social Psy- Mandler, G., Nakamura, Y., & Van Zandt, B. J. S. (1987). Nonspe- chology Bulletin, 27, 889–898. cific effects of exposure on stimuli that cannot be recognized.

379 REBER, SCHWARZ, WINKIELMAN

Journal of Experimental Psychology: Learning, Memory, and Reber, A. S. (1993). Implicit learning and tacit knowledge: An essay on Cognition, 15, 646–648. the cognitive unconscious. New York: Oxford University Press. Manza, L., Zizak, D., & Reber, A. S. (1998). Artificial grammar Reber, P. J., Stark, C. E. L., & Squire, L. R. (1998). Cortical areas learning and the mere exposure effect. In M. A. Stadler & P. A. supporting category learning identified using functional MRI. Frensch (Eds.), Handbook of implicit learning (pp. 201–222). Proceedings of the National Academy of Science USA, 95, Thousand Oaks, CA: Sage. 747–740. Maritain, J. (1966). Beauty and imitation. In M. Rader (Ed.), A mod- Reber, R. (2002). Reasons for the preference for symmetry. Behav- ern book of Esthetics (3rd ed., pp. 27–34). New York: Holt, ioral and Brain Sciences, 25, 415–416. Rinehart & Winston. Reber, R., & Schwarz, N. (1999). Effects of perceptual fluency on Martindale, C. (1984). The pleasures of thought: A theory of cogni- judgments of truth. Consciousness and Cognition, 8, 338–342. tive hedonics. The Journal of Mind and Behavior, 5, 49–80. Reber, R., & Schwarz, N. (2001) The hot fringes of consciousness: Martindale, C., & Moore, K. (1988). Priming, prototypicality, and Perceptual fluency and affect. Consciousness and Emotion, 2, preference. Journal of Experimental Psychology: Human Per- 223–231. ception and Performance, 14, 661–670. Reber, R., & Schwarz, N. (in press). Perceptual fluency, preference, Martindale, C., Moore, K., & Borkum, J. (1990). Aesthetic prefer- and evolution. Polish Psychological Bulletin. ence: Anomalous findings for Berlyne’s psychobiological the- Reber, R., Winkielman, P., & Schwarz, N. (1998). Effects of perceptual ory. American Journal of Psychology, 103, 53–80. fluency on affective judgments. Psychological Science, 9, 45–48. Mauro, R., & Kubovy, M. (1992). Caricature and face recognition. Reber, R., Wurtz, P., & Zimmermann, T. D. (2004). Exploring Memory & Cognition, 20, 433–440. “fringe” consciousness: The subjective experience of percep- McGlone, M. S., & Tofighbakhsh, J. (1999). The Keats heuristic: tual fluency and its objective bases. Consciousness and Cogni- Rhymeasreasoninaphorisminterpretation.Poetics,26,235–244. tion, 13, 47–60. McGlone, M. S., & Tofighbakhsh, J. (2000). Birds of a feather flock Rensch, B. (1957). Aesthetische Faktoren bei Farb- und Form- conjointly (?): Rhyme as reason in aphorisms. Psychological bevorzugungen von Affen [Aesthetic factors in color and form Science, 11, 424–428. preferences in monkeys] & [The effects of aesthetic factors in McWhinnie, H. J. (1968). A review of research on aesthetic measure. vertebrates]. Zeitschrift für Tierpsychologie, 14, 71–99. Acta Psychologica, 28, 363–375. Rensch, B. (1958). Die Wirksamkeit ästhetischer Faktoren bei Wir- Merleau-Ponty, M. (1964). The primacy of perception, and other es- beltierent. Zeitschrift für Tierpsychologie, 15, 447–461. says on phenomenological psychology, the philosophy of art, Repp, B. (1997). The aesthetic quality of a quantitatively average history, and politics. Chicago: Northwestern University Press. music performance: Two preliminary experiments. Music Per- Meyer, D. E., Schwaneveldt, R. W., & Ruddy, M. G. (1975). Loci of ception, 14, 419–444. contextual effects on visual word recognition. In P. M. Rabbit & Rhodes, G., Brennan, S., & Carey, S. (1987). Identification and rat- S. Dornic (Eds.), Attention and performance (Vol. 5, pp. ings of caricatures: Implications for mental representations of 98–118). London: Academic. faces. Cognitive Psychology, 19, 473–497. Monahan, J. L., Murphy, S. T., & Zajonc, R. B. (2000). Subliminal Rhodes, G., Proffitt, F., Grady, J. M., & Sumich, A. (1998). Facial mere exposure: Specific, general, and diffuse effects. Psycho- symmetry and the perception of beauty. Psychonomic Bulletin logical Science, 11, 462–466. and Review, 5, 659–669. Newell, B. R., & Bright, J. E. H. (2001). The relationship between Rhodes, G., Sumich, A., & Byatt, G. (1999). Are average facial con- the structural mere exposure effect and the implicit learning figurations attractive only because of their symmetry? Psycho- process. Quarterly Journal of Experimental Psychology, 54A, logical Science, 10, 52–58. 1087–1104. Rhodes, G., & Tremewan, T. (1996). Averageness, exaggeration, and Nicki, R. M., Lee, P. L., & Moss, V. (1981). Ambiguity, cubist works facial attractiveness. Psychological Science, 7, 105–110. of art, and preference. Acta Psychologica, 49, 27–41. Roederer, J. G. (1973). Introduction to the physics and psycho- Norman, K., & O’Reilly, R. (2001). Modeling hippocampal and physics of music. New York: Springer. neocortical contributions to recognition memory: A comple- Royer, F. (1981). Detection of symmetry. Journal of Experimental Psy- mentary learning systems approach (Tech. Rep.). Boulder: In- chology: Human Perception and Performance, 7, 1186–1210. stitute of Cognitive Science, University of Colorado. Rubenstein, A. J., Langlois, J. H., & Kalakanis, L. (1999). Infant Palmer, S. E. (1991). Goodness, Gestalt, groups, and Garner: Local sym- preferences for attractive faces: A cognitive explanation. Devel- metry subgroups as a theory of figural goodness. In G. R. Lockhead opmental Psychology, 35, 848–855. & J. R. Pomerantz (Eds.), The perception of structure (pp. 23–39). Rubin, D. C. (1995). Memory in oral traditions: The cognitive psy- Washington, DC: American Psychological Association. chology of epic, ballads, and counting-out rhymes. Oxford, Palmer, S. E., & Hemenway, K. (1978). Orientation and symmetry: England: Oxford University Press. Effects of multiple, near, and rotational symmetries. Journal of Rugg, M. D., & Yonelinas, A. P. (2003). Human recognition mem- Experimental Psychology: Human Perception and Perfor- ory: A cognitive neuroscience perspective. Trends in Cognitive mance, 4, 691–702. Sciences, 7, 313–319. Perrett, D. I., Lee, K., Penton-Voak, I., Burt, D. M., Rowland, D., Rumelhart, D. E., & McClelland, J. L. (1982). An interactive activa- Yoshikawa, S., et al., (1998). Sexual dimorphism and facial at- tion model of context effects in letter perception: Part 2. The tractiveness. Nature, 394, 884–886. contextual enhancement effect and some test and extensions of Perrett, D. I., May, K. A., & Yoshikawa, S. (1994). Facial shape and the model. Psychological Review, 89, 60–94. judgments of female attractiveness. Nature, 368, 239–242. Rumelhart, D. E., McClelland, J. L., & the PDP Research Group. Pinker, S. (1997). How the mind works. New York: Norton. (1986). Parallel distributed processing: Explorations in the Posner, M. I., & Keele, S. W. (1968). On the genesis of abstract ideas. microstructure of cognition. Cambridge, MA: MIT Press. Journal of Experimental Psychology, 77, 353–363. Saegert, S., Swap, W., & Zajonc, R. B. (1973). Exposure, context, Ramachandran, V. S., & Hirstein, W. (1999). The science of art: A and interpersonal attraction. Journal of Personality and Social neurological theory of aesthetic experience. Journal of Con- Psychology, 25, 234–242. sciousness Studies, 6, 15–51. Santayana, G. (1955). The sense of beauty. New York: Dover. (Origi- Read, H. (1972). The meaning of art. London: Faber & Faber. nal work published in 1896) Reber, A. S. (1967). Implicit learning of artificial grammars. Journal Schellenberg,E.G.,&Trehub,S.E.(1996).Naturalmusicintervals:Evi- of Verbal Learning and Verbal Behavior, 6, 855–863. dence from infant listeners. Psychological Science, 7, 272–277.

380 FLUENCY AND AESTHETIC PLEASURE

Schwarz, N. (1990). Feelings as information: Informational and mo- Symons, D. (1979). The evolution of human sexuality. New York: tivational functions of affective states. In E. T. Higgins & R. M. Oxford University Press. Sorrentino (Eds.), Handbook of motivation and cognition: Tatarkiewicz, W. (1970). History of aesthetics. The Hague, The Foundations of social behavior (Vol. 2, pp. 527–561). New Netherlands: Mouton. York: Guilford. Thornhill, R., & Gangestad, S. W. (1993). Human facial beauty: Schwarz, N. (1998). Accessible content and accessibility experiences: Averageness, symmetry and parasite resistance. Human Nature, The interplay of declarative and experiential information in judg- 4, 237–269. ment. Personality and Social Psychology Review, 2, 87–99. Thornhill, R., & Gangestad, S. W. (1999). Facial attractiveness. Schwarz, N. (2004a). Meta-cognitive experiences in consumer judg- Trends in Cognitive Sciences, 3, 452–460. ment and decision making. Journal of Consumer Psychology, 14, Tillmann, B., Bharucha, J. J., & Bigand, E. (2000). Implicit learning 332–348. of tonality: A self-organizing approach. Psychological Review, Schwarz, N. (2004b). Meta-cognitive experiences: Response to 107, 885–913. commentaries. Journal of Consumer Psychology, 14, 370–373. Trainor, L. J., & Heinmiller, B. M. (1998). Infants prefer to listen to Schwarz, N., Bless, H., Strack, F., Klumpp, G., Rittenauer-Schatka, consonance over dissonance. Infant Behavior and Develop- H., & Simons, A. (1991). Ease of retrieval as information: An- ment, 21, 77–88. other look at the availability heuristic. Journal of Personality Turati, C. (2004). Why faces are not special to newborns: An alterna- and Social Psychology, 61, 195–202. tive account of the face preference. Current Directions in Psy- Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and judgments chological Science, 13, 5–8. of well-being: Informative and directive functions of affective states. Vallacher, R. R., & Nowak, A. (1999). The dynamics of self-regula- Journal of Personality and Social Psychology, 45, 513–523. tion. In R. S. Wyer, Jr. (Ed.), Perspectives on behavioral Schwarz, N., & Clore, G. L. (1996). Feelings and phenomenal expe- self-regulation (pp. 241–259). Mahwah, NJ: Lawrence rience. In E. T. Higgins & A. W. Kruglanski (Eds.), Social psy- Erlbaum Associates, Inc. chology: Handbook of basic principles (pp. 433–465). New Van den Bergh, O., & Vrana, S. R. (1998). Repetition and boredom York: Guilford. in a perceptual fluency/attributional model of affective judg- Seamon, J. G., Brody, N., & Kauff, D. M. (1983). Affective discrimi- ments. Cognition and Emotion, 12, 533–553. nation of stimuli that are not recognized: Effects of shadowing, Vendler, H. (1997). Poems, poets, poetry. An introduction and an- masking, and central laterality. Journal of Experimental Psy- thology. Boston: Bedford. chology: Learning, Memory and Cognition, 9, 544–555. Vitz, P. C. (1966). Affect as a function of stimulus variation. Journal Seamon, J. G., McKenna, P. A., & Binder, N. (1998). The mere expo- of Experimental Psychology, 71, 74–79. sure effect is differentially sensitive to different judgment tasks. Walton, G. E., & Bower, T. G. R. (1993). Newborn form “prototypes” Consciousness and Cognition, 7, 85–102. in less than 1 minute. Psychological Science, 4, 203–205. Simon, H. A. (1967). Motivational and emotional controls of cogni- Whitfield, T. W. A., & Slatter, P. E. (1979). The effects of categoriza- tion. Psychological Review, 74, 29–39. tion and prototypicality on aesthetic choice in a furniture selec- Skurnik, I., Schwarz, N., & Winkielman, P. (2000). Drawing infer- tion task. British Journal of Psychology, 70, 65–75. ences from feelings: The role of naive beliefs. In H. Bless & J. P. Whittlesea, B. W. A. (1993). Illusions of familiarity. Journal of Ex- Forgas (Eds.), The message within: The role of subjective expe- perimental Psychology: Learning, Memory, and Cognition, 19, rience in social cognition and behavior (pp. 162–175). Phila- 1235–1253. delphia: Psychology Press. Whittlesea, B. W. A., Jacoby, L. L., & Girard, K. (1990). Illusions of Slater, A., von der Schulenburg, C., Brown, E., Badenoch, M., immediate memory: Evidence of an attributional basis for feel- Butterworth, G., Parsons, S., et al., (1998). Newborn infants ings of familiarity and perceptual quality. Journal of Memory prefer attractive faces. Infant Behavior and Development, 21, and Language, 29, 716–732. 345–354. Whittlesea, B., & Price, J. (2001). Implicit/explicit memory versus Smith, E. E., Shoben, E. J., & Rips, L. J. (1974). Structure and pro- analytic/nonanalytic processing: Rethinking the mere exposure cess in semantic memory: A feature model for semantic deci- effect. Memory and Cognition, 29, 234–246. sions. Psychological Review, 81, 214–241. Whittlesea, B. W. A., & Shimizu, Y. (2001). The SCAPE framework Smith, E. R. (2000). Subjective experience of familiarity: Functional and subjective experience, Part II: Pleasantness, interest, pre- basis in connectionist memory. In H. Bless & J. P.Forgas (Eds.), dictability and alertness. Manuscript submitted for publication. The message within: The role of subjective experience in social Whittlesea, B. W. A., & Williams, L. D. (1998). Why do strangers cognition and behavior (pp. 109–124). Philadelphia: Psychol- feel familiar, but friends don’t? The unexpected basis of feel- ogy Press. ings of familiarity. Acta Psychologica, 98, 141–166. Smith, J. D., & Melara, R. J. (1990). Aesthetic preference and syn- Whittlesea, B. W. A., & Williams, L. D. (2000). The source of feel- tactic prototypicality in music: ‘Tis the gift to be simple. Cogni- ings of familiarity: The discrepancy-attribution hypothesis. tion, 34, 279–298. Journal of Experimental Psychology: Learning, Memory, and Snodgrass, J. G., & Vanderwart, M. (1980). A standardized set of Cognition, 26, 547–565. 260 pictures: Norms for name agreement, image agreement, fa- Wilson, T. D., Lisle, D. J., Schooler, D., Hodges, S. D., Klaaren, K., miliarity, and visual complexity. Journal of Experimental Psy- & LaFleur, S. J. (1993). Introspecting about reason can reduce chology: Human Learning & Memory, 6, 174–215. post-choice satisfaction. Personality and Social Psychology Sober, E. (1994). From a biological point of view: Essays in evolu- Bulletin, 19, 331–339. tionary philosophy. Cambridge, England: Cambridge Univer- Winkielman, P., & Cacioppo, J. T. (2001). Mind at ease puts a smile sity Press. on the face: Psychophysiological evidence that processing fa- Sollberger, B., & Reber, R. (2004). Artificial grammar learning as a cilitation leads to positive affect. Journal of Personality and So- potential mechanism for the acquisition of music preference. cial Psychology, 81, 989–1000. Manuscript submitted for publication. Winkielman, P., & Fazendeiro, T. (2003). Effects of conceptual flu- Solso, R. L. (1997). Cognition and the visual arts. Cambridge, MA: ency on memory and liking. Manuscript in preparation. MIT Press. Winkielman, P., Schwarz, N., & Belli, R. F. (1998). The role of ease Swap, W. (1977). Interpersonal attraction and repeated exposure to of retrieval and attribution in memory judgments: Judging your rewarders and punishers. Personality and Social Psychology memory as worse despite recalling more events. Psychological Bulletin, 3, 248–251. Science, 9, 124–126.

381 REBER, SCHWARZ, WINKIELMAN

Winkielman, P., Schwarz, N., Fazendeiro, T., & Reber, R. (2003). Zajonc, R. B. (1998). Emotions. In D. T. Gilbert, S. T. Fiske, & G. The hedonic marking of processing fluency: Implications for Lindzey (Eds.), The handbook of social psychology (pp. evaluative judgment. In J. Musch & K. C. Klauer (Eds.), The 591–632). Boston: McGraw-Hill. psychology of evaluation: Affective processes in cognition and Zajonc, R. B. (2000). Feeling and thinking: Closing the debate over emotion (pp. 189–217). Mahwah, NJ: Lawrence Erlbaum Asso- the independence of affect. In J. P. Forgas (Ed), Feeling and ciates, Inc. thinking: The role of affect in social cognition (pp. 31–58). New Witherspoon, D., & Allan, L. G. (1985). The effects of a prior pre- York: Cambridge University Press. sentation on temporal judgments in a perceptual identification Zajonc, R. B., Markus, H., & Wilson, W. R. (1974). Exposure effects task. Memory & Cognition, 13, 103–111. and associative learning. Journal of Experimental Social Psy- Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal chology, 10, 248–263. of Personality and Social Psychology, Monograph Supplement, Zentner, M. R., & Kagan, J. (1996). Perception of music by infants. 9, 1–27. Nature, 383, 29.

382