The Origins and Abiding Signature of the Ecological Theory of Social Perception in Social Psychology Research: A Tribute to Reuben Baron

Abstract

In this tribute to Reuben Baron, we provide some history regarding how the first author came to collaborate with Reuben on our 1983 publication of Toward an Ecological Theory of Social Perception. We then briefly review the major tenets of the theory and discuss some examples of Reuben’s research that supported those tenets as well as his theorizing that expanded on them. Next, we provide evidence that, in the decades following publication of the ecological theory, its signature can be found in social psychology research in the domains of social perception accuracy and the stimulus information that underpins social perceptions, be they accurate or not. We also discuss implications of implementing paradigms that have active perceivers, a central directive of ecological theory that is not reflected in existing research. We conclude with suggestions for future research necessary to provide better tests of the validity of tenets of the ecological theory of social perception.

The first author of this tribute to Reuben Baron initially met him at a conference in the late 1970s, when her student was presenting a paper they coauthored: “Figural Emphasis and Person Perception” (McArthur & Post, 1977). Drawing on Gestalt principles of perceptual organization that determine which parts of a visual array are perceived as “figure” and which as “grouind” (Rubin, 1915/1958),¹ this research had instantiated principles of figural emphasis, such as the movement and brightness of two people in a social interaction to investigate their effects on causal attributions. A psychologist sitting in front of me challenged my coauthor with questions about “top-down” vs. “bottom-up” stimulus-linked determinants of our effects. Afterward, I tapped him on the shoulder, introduced myself, and told him that I totally agreed with what he had said. That was the beginning of many stimulating and fruitful discussions with Reuben, who had been greatly influenced by his neo-Gibsonian colleagues at the University of Connecticut, Robert Shaw and Michael Turvey, to embrace ecological theory and apply it to social psychology. I had been influenced in the same direction as a graduate student at Yale when I read with great interest J.J. Gibson’s The Senses Considered as Perceptual Systems (Gibson, 1966) in a course taught by Wendell (“Tex”) Garner. Hearing Gibson and Garner engage in a friendly, heated debate about perceptual processes at a department colloquium further sparked my interest in Gibson’s theoretical viewpoint. The mutual interests that Reuben and I shared ultimately led us to collaborate on a paper in Psychological Review: “Toward an Ecological Theory of Social Perception” (McArthur & Baron, 1983). This paper was inspired by J.J. Gibson’s The Ecological Approach to Visual Perception (Gibson, 1979) and his suggestion that the theory also applied to social perception. Reuben brought to our collaboration a much deeper understanding of Gibsonian principles than I had, and I find myself wishing that Reuben could review and comment on this paper. I hope he would approve!

In this tribute to Reuben, we briefly review the major tenets of the ecological theory of social perception derived from Gibson’s (1979) ecological approach to visual perception, followed by examples of Reuben’s work bearing on those tenets. We also argue that the signature of ecological theory can be found in many social psychology research domains over the ensuing decades. These include our discussion of accuracy in social perception and the stimulus information that underpins social perceptions. An important caveat is that the broad range of literature we are reviewing to show the ecological theory signature necessitates that it be selective, not exhaustive, with a high-level summary of findings excluding methodological details.

Tenets of the ecological theory of social perception

Four tenets of the ecological approach to social perception, summarized below, should be understood partly as a corrective to the dominant paradigms of the time in which social “perception” was ignored in favor of social “cognition,” which was largely based on verbal descriptions of people and viewed as highly error-prone. For example, trait impressions were assessed by providing research participants with a list of trait adjectives and determining that the person who is first labeled as industrious and then as stubborn tends to be judged more positively than one who is first labeled as stubborn. Whereas this “primacy” effect reveals something about biases in the cognitive processing of trait labels, it does not tell us what perceptible qualities of people communicate industry or stubbornness or whether these traits can be accurately perceived. As McArthur and Baron (1983) argued:

Most current thinking in the domains of social perception and cognition either takes as axiomatic the proposition that our knowledge of the social environment is highly error-prone (e.g., Nisbett & Ross, 1980; Ross, 1977) or simply ignores the accuracy problem altogether in favor of a process analysis that focuses on the cognitive operations intervening between stimulus and response (e.g., Anderson, 1974; Newtson, 1976, p. 229).

The questions addressed concern the knowledge structures (e.g., schemata, prototypes, scripts) that impose meaning on the blooming, buzzing confusion around us. The questions ignored concern the structured stimulation that exists in our social environment. As such, we have learned much about the processing of information and little about what that stimulus information is (p. 215).

In sharp contrast to such views, the ecological perspective emphasizes the essential accuracy of perception-based knowledge (e.g., Ittelson, 1973). And because Gibson (1979) strongly argues that perception of the social environment is likely to follow the same basic principles as perception of the nonsocial environment, an ecological model challenges the current view that social perceptions are more often flawed than not (p. 230).

Tenet 1

“The ecological approach to perception begins by assuming that perception serves an adaptive function … by informing action, perception is assumed to promote individual goal attainment as well as species survival” (McArthur & Baron, 1983, p. 216). As Gibson (1979) argued, “perceiving is for doing.” This tenet makes the case for the accuracy of social perception. It also argues for the automaticity of at least some social perceptions, what Gibson termed “direct” perception, since cogitating on the meaning of a raging bull or a threatening person is not adaptive.

Tenet 2

“The information that guides perception is typically revealed in objective physical events—dynamic, changing, multimodal stimulus information” (McArthur & Baron, 1983, p. 215). This focus on the what of perceptual processing contrasts with what was at that time a focus on the how of processing (e.g., various cognitive processes that influence the meaning we extract from the social environment).

Tenet 3

“The stimulus information available in events specifies, among other things, environmental affordances, which are the opportunities for acting or being acted upon that are provided by environmental entities’ (McArthur & Baron, pp. 215–216). For example, as Gibson quoted from Koffka (1935, p. 7), “Each thing says what it is … a fruit says ‘eat me,’ ‘water says drink me’ …” (Gibson, 1979, p. 138). This tenet emphasizes the sufficiency of the stimulus information to enable accurate perceptions. It also focuses attention on the intrinsic connection between action and perception since the exploratory behaviors of an active perceiver can reveal affordances. The affordance concept was a departure from the traditional focus on inferring stable dispositional properties, like abilities and personality traits.

Tenet 4

“… the perception of affordances depends upon the perceiver’s attunement, that is, the particular stimulus invariants to which the perceiver attends” (McArthur & Baron, p. 216). Some attunements are innate, while others are influenced by the perceiver’s behavioral capabilities and goals as well as by perceptual learning, what Gibson (1966) called the “education of attention.”

Reuben Baron’s work bearing on the ecological tenets

In a 1997 interview, Reuben made the astute observation that “after so many years on the independent variable side, we are still rather primitive on the dependent variable side. Here is where I quarrel methodologically with some of the ecological work, which is strong in independent variables, but still relies on some sort of verbal judgment as the dependent measure” (Baron, 1997, p. 264). He remains absolutely correct more than 15 years later, as the following sections of this paper will attest, and our own research is guilty. However, Reuben’s research is notable and laudable for efforts to assess the actual behavioral affordances of people and how they influence interpersonal interactions.

One outstanding example is Reuben’s research on the accurate perception of intrapersonal and interpersonal physical behavioral affordances for people in dyadic interactions. Individuals could accurately judge whether they needed one hand, two hands, or a tool to remove planks of different lengths from a shelf, revealing accurate intrapersonal affordance perception (Richardson et al., 2007). Consistent with the ecological argument that affordances are directly perceived, the behavioral responses to the affordances of this task required minimal cognitive resources (Lopresti-Goodman et al., 2009). Moving to more social perceptions, an extension of this paradigm found that people were sensitive to the interpersonal affordances of a similar task designed to elicit cooperation (Isenhower et al., 2010). Individuals who were matched or mismatched on arm span moved wooden planks off a conveyor belt by holding the ends of the plank. The results showed that they transitioned between one- and two-handed grasping, depending on the plank size and their own hand span, and between two-handed solo action and cooperative action, depending on the plank size and the arm spans in the pair, with cooperative behavior emerging when needed by the actor with the smaller arm span. Thus, the physical capabilities of the individuals and the pair of individuals influenced the behavioral affordances of the planks, and their actions showed accurate perception of an interpersonal affordance, yielding cooperative vs. independent behavior.

The examination of the emergence of cooperative behavior from the joint affordances of two people illustrates Reuben’s concern with understanding how the perceived structural and dynamic qualities of one person affect another. It also points to the importance of investigating embodied social interactions that require an understanding of the emergent influence of each person’s actions. Reuben and his colleagues have addressed this issue in insightful theoretical papers. For example, they argued that in embodied social interactions, there can be emergent environmental affordances that are absent when someone is acting alone, a social synergy perspective (Marsh et al., 2006). As a prototypical example of an emergent perception-action system, they cite Asch’s (1952) account of two boys moving a log during which time they are co-ordinating their actions to bring a common force to the task with a unity of action that includes the actors and the log—and is not merely the sum of their separate actions. Neither boy would act the same way in the absence of the other (pp. 173–174).

Reuben not only sought to expand the affordance concept to encompass social interactions but also to refine it. Considering Gibson’s statement that “I have been moving to a psychology of values instead of a psychology of stimulus” (Locker, 1980 cited in Reed, 1988, p. 296) and Heft’s (1989) observation that affordances provide both the possibilities for actions and the consequences of these actions, Hodges and Baron (1992) argue that affordances do not cause actions to occur in a mechanistic manner, but rather they are the behavioral opportunities that stimuli provide for consequential goal attainment. Although Gibson’s discussion of affordances does state that perceiving them depends on the perceiver’s goals as well as behavioral capabilities and perceptual learning, it does not make explicit how and when perceiving them leads to action. Hodges and Baron (1992) make an important theoretical contribution by highlighting the role of perceiver values in this next step. An example they discuss is a young child eating. The child perceives the affordances of fingers for picking up the food and is taught to perceive the affordances of forks. Either action can satisfy the goal attainment of eating, but the child learns to value actuating the affordance of a fork. Similarly, young children may perceive that a smaller child provides the behavioral affordance of grabbing to obtain a coveted object, but they are taught to value the alternative behavioral affordance of asking or trading. Thus, perceiving particular stimulus information does not simply reveal various opportunities for action. It also can realize values through the selection of a one action among many possibilities.

Another contribution to the utility of the affordance concept is the discussion by Baron and Misovich (1993) of how to make dispositional qualities perceptible, which would allow people to accurately detect them as well as to enable empirical tests of ecological theory tenets. They proposed that embodiment in relational activities could reveal dominance or cooperativeness or other dispositions. They further argued for event-activity tests that would control for insufficient information or deception, because only a truly dominant or cooperative person would show the behavior. The prototypical example given is King Solomon’s test of proposing to cut the disputed baby in half to discern who is and is not the baby’s real mother. Less extreme examples would be observing people at a party rather than when watching a movie to discern how extraverted they are or observing people making a speech rather than when watching a movie to discern how intelligent they are.

Reuben’s research has also considered the ecological theory tenet that an understanding of social perception requires attention to the nature of the available stimulus information. One study modified the classic Heider–Simmel film to either disrupt the dynamic aspects of the original animation, with structural properties preserved (large triangle, small triangle) or vice versa. The results revealed that it was the dynamic movement patterns that had the most influence on perceivers’ tendency to anthropomorphize the events in the original display (Berry et al., 1992). Another study showed that perceivers use diagnostic stimulus information about a person when it is available, rather than relying on their social category. Perceivers’ judgments of a child’s ability were elicited by observing a child’s high or low performance on an IQ test, but not by knowing the child’s socioeconomic background, which had a significant effect only when there was no behavioral information (Baron et al., 1995).

In summary, among Reuben’s contributions since the publication of the 1983 paper is evidence that perceivers favor dynamic as well as diagnostic stimulus information, which not only is consistent with ecological theory tenets but also is rarely addressed in research examining social perception accuracy. Most striking is Reuben’s singular examination of people’s accuracy in perceiving actual behavioral affordances. Moreover, his research showed accurate perceptions of behavioral affordances not only by an individual but also by pairs of individuals who perceive emergent affordances that inform adaptive cooperation. Reuben’s work has also refined the affordance concept to incorporate values that guide the selection of actions among the many possible behavioral affordances that are perceived. Furthermore, he has considered what prompts a person’s “hidden” affordances to become perceptible to perceivers, an endeavor that is essential to expanding research investigating whether or when people accurately perceive behavioral affordances in their social interactions.

Research demonstrating accurate perception of qualities that can inform adaptive action

In the years following the publication of the ecological theory of social perception, there was a striking increase in research examining the accuracy of perceptions in contrast to the almost singular focus on errors and biases in earlier research.² In fairness, we cannot attribute this paradigm shift solely to ecological theory, as other researchers at about the same time also spoke up for accuracy (Funder, 1987; Kenny & Albright, 1987: Swann, 1984), and the closely related study of interpersonal sensitivity that predated the ecological theory of perception (Rosenthal et al., 1979) was also fueling more interest in accuracy. That research focused more on individual differences in sensitivity, what ecological theory might call “perceiver attunements,” than on the stimulus information that conveyed what was accurately perceived. For this reason, and in the interest of space limitations, we do not discuss perceiver attunements. (See Hall and Bernieri [2001] for several excellent chapters on individual differences.) Notably, some accuracy research has been motivated by considerations of those qualities that are adaptive to quickly perceive, which is consistent with Gibson’s ecological canon that “perceiving is for doing” (Gibson, 1979). Most significant is research inspired by evolutionary psychology, which focuses on perception in the service of evolutionary adaptation (e.g., Cosmides & Toobey, 1997; Maner et al., 2005). Thus, many researchers examined accuracy from a functional perspective with the assumption that it should be most evident for attributes whose correct or incorrect identification had implications for social actions relevant to survival and reproduction in our evolutionary past (Gangestad et al., 1992).

Research investigating social perception accuracy has often deviated in important ways from ecological theory tenets, and this is equally true for research examining the stimulus information reviewed in the following section. Accuracy has been typically assessed by perceivers’ ratings, with no opportunity for perceivers’ actions toward a stimulus person to inform their perceptions. Also, the accuracy criteria are not whether perceived affordances are realized in social interactions with the perceiver. Rather, they include validation of perceptions by personality and ability tests, health records, self-ratings, ratings by people who knew them, consensus of strangers, or behavior and social outcomes in other settings.

Although the reliance on trait ratings is a shortcoming, it is noteworthy that their meaning does capture behavioral affordances even if these ratings do not actuate them. Indeed, Baron and Misovich (1993) argued that traits are actually dispositional affordances. Moreover, research shows that trait ratings in fact capture the behaviors that others direct toward a target, behavioral affordances, more than the behaviors that the target shows (Beauvois & Dubois, 2000). Trait ratings also capture the states that are elicited in others by a target, emotional affordances, more than the emotional states that the target shows (Mignon & Mollaret, 2012).

Research using the aforementioned methods has revealed significant accuracy for trait perceptions that can facilitate adaptive social interactions. These include people’s agreeableness, conscientiousness, dominance, extraversion, intelligence, and openness to experience (Hall et al., 2008). Accuracy has also been documented for perceptions of people’s aggressiveness (Carré et al., 2009), health (Tskhay et al., 2016; Zebrowitz et al., 2014; Zebrowitz & Rhodes, 2004), political affiliation (Rule & Sutherland, 2017), sexual availability (Ambady et al., 1999; Gangestad et al., 1992), and sexual orientation (Rule & Alaei, 2016).

Consistent with the ecological theory argument that social affordances may be directly perceived from stimulus information, accurate perceptions have been demonstrated with short exposures to nonverbal information. For example, perceivers’ ratings from “thin slices” of videotaped behavior (Ambady et al., 2000) accurately predict whether someone is a good teacher, therapist, or physician, and accuracy is as great for thin slices shorter than 30 seconds as for those lasting up to 5 minutes (Ambady & Rosenthal, 1992). Consistent with the ecological theory emphasis on the adaptive function of perception, thin slices of a videotaped interaction as brief as 5 seconds were sufficient for the accurate perception of negative affect, which could require fast responses, whereas accurate perception of positive affect required a 20-second duration. At the same time, however, a wide range of traits whose recognition has varying utility were accurately perceived from only 5 seconds of exposure (Carney et al., 2007). Thin slices of vocal cues can also yield accurate impressions. Perceivers’ ratings of sales managers’ interpersonal traits from 20-second content-masked audio clips accurately distinguished average from outstanding managers (Ambady et al., 2006). Consistent with the ecological theory assumption that dynamic stimulus information is more diagnostic than static, thin slices of videotaped behavior yielded greater accuracy in identifying sexual orientation than did still photographs, although accuracy was above chance for both (Ambady et al., 1999). Accurate perceptions of sexual orientation are also provided by inherently dynamic voice cues (Tskhay & Rule, 2013).

While the level of accuracy based on only a few seconds of exposure to a videotape is striking, even more surprising is that accurate social perceptions emerge from even briefer exposures to still facial photographs. A mere 39-millisecond exposure to faces was sufficient for the accurate perception of aggressiveness (Carré et al., 2009). Perceptions of the relative competence of two opposing politicians when their faces were shown together for as little as 100 milliseconds accurately predicted who won the election (Todorov et al., 2005). Moreover, the predictive accuracy of competence judgments was better when perceivers were given only 100 milliseconds than when they were asked to deliberate for a longer duration and make good judgments (Ballew & Todorov, 2007). Similarly, perceptions of men’s and women’s sexual orientation were accurate when their faces were shown as briefly as 50 milliseconds, with no increase in accuracy for longer durations (Rule & Ambady, 2008). Furthermore, asking perceivers to think about their decisions before responding resulted in perceptions of sexual orientation that were no better than chance (Rule et al., 2009). Such results are consistent with the ecological theory tenet that behavioral affordances are directly revealed in perceptible stimulus information as well as evidence that deliberative processing can interfere with automatic processes (Wilson & Schooler, 1991).

An important caveat to the aforementioned accuracy findings is that, while accuracy was statistically significant, there is room for increases in effect sizes. Perhaps research paradigms that adhere more closely to tenets of the ecological theory of perception would provide stronger evidence for accuracy.

Research identifying the stimulus information for accurate social perceptions

Consistent with ecological theory tenets, we focus here on three sources of stimulus information for which there is a theoretical basis to claim utility for adaptive action: emotion expressions, attractiveness, and immaturity. As ecological theory advocates, the stimulus information provided to perceivers captures some of what one would encounter in a real social interaction, rather than in written descriptions of a person, which were formerly favored. Thus, research has provided nonverbal stimulus information about those being judged (e.g., facial photographs, voice recordings, videotapes that were silent or devoid of meaningful content). Also, consistent with ecological theory tenets, the stimulus information presented is often multimodal and sometimes includes the dynamic information that ecological theory holds to be most diagnostic (see Hall et al., 2008 for a review). On the other hand, contrary to ecological theory tenets, the research rarely allows active perceivers to extract the stimulus information in an actual social interaction. Moreover, research often identifies the stimulus information that influences perceptions without ascertaining whether those perceptions are accurate. Nevertheless, there is some evidence that utilizing emotion expressions, attractiveness, and/or immaturity cues can inform accurate social perceptions.

Expressions of emotion

Predating ecological theory, a large body of research had demonstrated accuracy in perceiving emotions from facial cues (Ekman et al., 1972), with later evidence of accuracy from vocal cues (Bachorowski, 1999) and body cues (Van den Stock et al., 2007). Moreover, responsiveness to facial expressions in early infancy suggests an innate preparedness that serves an adaptive function (Walle et al., 2017).

Although we do not consider accuracy in emotion perception per se, that research is relevant to the perceived affordances and traits that we do consider. Specifically, drawing on Darwin’s (1965) evolutionary view of the adaptive value of emotion expressions for social communication, some emotion researchers have suggested that expressions of emotion reveal not only people’s affective states but also their behavioral intentions (Ekman, 1997; Fridlund, 1991), behavioral intentions would reflect the affordances that the perceiver detects. A related perspective proposes that expressions of emotion give rise to stable trait impressions, akin to stable affordances, via the process of temporal extension whereby “the perceiver regards a momentary characteristic of the person as if it were an enduring attribute” (Secord, 1958).

Considerable research supports the aforementioned views. Angry expressions elicit perceptions of high dominance and low affiliation, happy expressions elicit perceptions of high affiliation as well as more physical movement toward a smiling person, and expressions of sadness and fear elicit perceptions of low dominance and moderate affiliation (Hess et al., 2000; Knutson, 1996; Montepare & Dobish, 2003). Interestingly, although fear expressions are judged as unpleasant, they facilitate approach behaviors (Hammer & Marsh, 2015; Marsh et al., 2005), suggesting that trait impressions are not isomorphic with behavioral affordances. Consistent with the ecological theory emphasis on multimodal stimulus information, the research demonstrating trait impressions from emotion expressions reveals effects for facial cues (Montepare & Dobish, 2003) and body movement cues (Thoresen et al., 2012). Although vocal cues also influence impressions (cf. Lavan, 2023), to our knowledge, research has not specifically focused on the emotional quality of voices.

In addition to the influence of people’s transient emotion expressions on perceptions of their more stable traits, the resemblance of their permanent facial structure to an emotion expression also influences trait impressions. This effect has been labeled emotion overgeneralization (Zebrowitz & Collins, 1997). Specifically, it was hypothesized that the adaptive value of responding appropriately to emotional expressions, such as avoiding an angry person and approaching a happy one, has produced a strong preparedness to respond to the facial qualities that communicate emotion. This response is then overgeneralized to individuals whose facial structure resembles a particular emotional expression. Research using computer modeling to quantify the extent to which people’s neutral-expression faces resembled various emotion expressions revealed impressions of traits that are consistent with each emotion. Those with a greater resemblance to angry faces were perceived as more threatening, hostile, dominant, and untrustworthy, while those with a greater resemblance to happy faces were perceived as less threatening as well as more sociable and emotionally stable (Said et al., 2009; Zebrowitz et al., 2010).

The finding that emotion expressions and emotion resemblance significantly influence perceptions of behavioral dispositions makes an important contribution to identifying the stimulus information for social perceptions. Moreover, there is also some evidence that emotion expressions and emotion resemblance contribute to accurate social perceptions. The greater perceived and actual sadness of men with sexually transmitted diseases than without them mediated accuracy in perceiving who was sick or healthy (Tskhay et al., 2016). The greater perceived and actual happiness of gay men’s faces than of straight men’s faces and the greater perceived and actual happiness of Democratic politicians’ faces than of Republican politicians’ faces also mediated accuracy in perceiving sexual orientation and political affiliation (Tskhay & Rule, 2015). In addition, people who reported more fear when relating emotional experiences scored higher on a measure of neuroticism, and accurate perceptions of fear were positively correlated with accurate perceptions of this trait (Hall et al., 2017). Finally, one study relevant to emotion resemblance found correlations between the emotion that older women’s neutral expressions conveyed and their personality traits (Malatesta et al., 1987). For example, those whose neutral-expression faces were most frequently confused with anger tended to show a personality profile high in aggressiveness.

Attractiveness

Attention to perceptions of people who vary in attractiveness also predates the ecological theory of perception by many years (Berscheid & Walster, 1974). However, in contrast to ecological theory tenets, this work was primarily concerned with the social consequences of attractiveness and did not attempt to identify the underlying stimulus information. It also largely assumed that responses to attractiveness were culturally driven rather than having a universal adaptive value. Later research—motivated by evolutionary as well as ecological theories, which share common assumptions (Zebrowitz & Montepare, 2006)—better reflects ecological theory tenets. Consistent with the emphasis on multimodal stimulus information, later investigators differentiated effects of facial, vocal, and body attractiveness. Also consistent is the functional approach taken by many who have argued that high attractiveness signals health or “good genes,” which makes it adaptive to choose more attractive people as mates (Thornhill & Gangestad, 1993). This assumption motivated researchers to test hypotheses concerning perceptions of particular stimulus qualities that may signal high or low fitness (Kościński, 2013; Park et al., 2003; Rhodes, 2006) as well as perceiver attunements that may enhance the adaptive value of detecting these qualities (e.g., Little et al., 2002). There are theoretical explanations for the effects of variations in attractiveness besides the possibility that they signal fitness, and these are not mutually exclusive (see Langlois et al., 2000; Zebrowitz & Collins, 1997). Nevertheless, positive reactions to more attractive faces by infants and young children (Langlois et al., 1987) are consistent with the argument that the influence of attractiveness on perceptions reflects, at least in part, an innate preparedness with an adaptive value.

Research has provided strong evidence for an attractiveness “halo effect.” More positive traits are perceived in those with attractive faces, including social and intellectual competence, dominance, and psychological adjustment (Eagly et al., 1991), and a positive halo generalizes across cultures (Dion, 2002; Zebrowitz et al., 2012). Voices considered more attractive also elicit more positive trait perceptions (Berry, 1992; Zuckerman et al., 1995), as do body weights considered more attractive (Ryckman et al., 1989). Although tallness is related to several positive social outcomes, evidence for a positive halo effect in trait perceptions is mixed (Chu & Geary, 2005; Jackson & Ervin, 1992).

Notably, traits captured by the attractiveness halo effect are also accurately perceived, and there is considerable support for the hypothesis that facial attractiveness may provide the stimulus information for accurate perceptions. Comprehensive meta-analyses that correlated behavioral data and trait measures with facial attractiveness (Langlois et al., 2000) supported the hypothesis that using attractiveness as a cue could yield accuracy in judging intelligence, mental and physical health, extraversion and social skills, and sexual experience, all of which were positively correlated with attractiveness. Attractiveness was also positively correlated with self-confidence, which may enable it to accurately signal dominance, which is perceived with significant accuracy. No evidence to date has found that vocal or body attractiveness is correlated with traits that are accurately perceived (Berry, 1990; Lassek & Gaulin, 2018); this is in contrast to findings regarding facial attractiveness.

Although facial attractiveness elicits trait perceptions overlapping those that are accurately perceived and is also significantly related to those traits, the question remains as to whether variations in attractiveness actually mediate accurate trait perceptions. Some studies have examined the relationship of attractiveness to both perceived and actual traits, but most have not tested whether attractiveness mediates accuracy. Attractiveness judged from still photographs was positively related to both perceived and actual health and intellectual competence (Zebrowitz et al., 2002; 2014; Zebrowitz & Rhodes, 2004), unaggressiveness (Boshyan et al., 2014), and power (Berry, 1991), and attractiveness judged from silent or sound videotapes was positively related to perceived and actual intelligence as well as to extraversion, agreeableness, conscientiousness, and emotional stability (Borkenau & Liebler, 1995). The latter authors also tested several possible mediators of accuracy for each trait perception and found that attractiveness alone was not a significant mediator. However, Zebrowitz et al. (2002) found that accurate perceptions of intelligence were lost when attractiveness was controlled.

An important caveat to the these findings is relates to evidence that accurate perceptions of health and intelligence based on facial attractiveness are limited to people in the bottom half of the attractiveness continuum. Although greater attractiveness was correlated with perceptions of greater intelligence and health across the whole continuum, attractiveness and actual intelligence or health were positively related only among people who ranged from the categories of unattractive to average and not among those who ranged from the categories of average to attractive (Zebrowitz & Rhodes, 2004). Moreover, intelligence and health were perceived accurately for faces in the lower but not the upper half of the attractiveness distribution. Parallel results were found in research that examined only perceptions and not their accuracy. The groundbreaking study of the attractiveness halo effect (Dion et al., 1972) found a greater difference between low and medium attractiveness than medium and high attractiveness for a variety of perceptions. This pattern was confirmed by Griffin and Langlois (2006), who reported that the attractiveness halo effect was driven more by the perception that “ugly is bad” than by the perception that “beautiful is good.” There is also more evidence that shortness is bad than that tallness is good (Chu & Geary, 2005; Jackson & Ervin, 1992).

The finding that the level of attractiveness provides a more valid cue to low than high intelligence or health is consistent with a “bad genes” explanation for trait perceptions associated with attractiveness as well as evidence that people at greater risk for disease are more influenced by attractiveness (White et al., 2013; Zebrowitz et al., 2014). Notably, subtle negative deviations from average attractiveness can signal not only poor health or low intelligence but also maladaptive social traits (Paulhus & Martin, 1986; Streissguth et al., 1978). An interesting unexamined question is whether the accurate perception of such traits is also more likely when attractiveness is low.

Immaturity

Early ethologists demonstrated that visual and auditory immature stimulus qualities elicit caretaking responses and inhibit aggression in humans as well as infrahumans (Eibl-Eibesfeldt, 1989). Lorenz (2010) coined the term kindchenschema to capture the differences between infants and adults in head shape and feature placement that is true for many animals. Years later, neo-Gibsonians, who were Reuben’s colleagues, showed that perceivers’ age judgments were highly sensitive to changes in schematic human profiles produced by a mathematical transformation that simulates the remodeling of the skull with growth (Pittenger & Shaw, 1975). Building on this finding, research further demonstrated increased feelings of protectiveness as profiles became less mature (Alley, 1983) as well as increased perceptions of weakness and decreased perceptions of threat (Berry & McArthur, 1986).

These findings stimulated the babyishness overgeneralization hypothesis, which holds that the evolutionary importance of identifying babies may have produced such a strong tendency to respond to their stimulus qualities that these responses are overgeneralized to those who merely resemble babies, regardless of their age (Berry & McArthur, 1986; Zebrowitz & Collins, 1997). As discussed below, researchers testing this functional hypothesis have examined facial, vocal, and body cues to immaturity, consistent with the ecological theory emphasis on multimodal stimulus information. Also, consistent with the hypothesis that a universally adaptive response to babies underpins babyishness overgeneralization are infants’ and young children’s reactions to babyfaced adults (Keating & Bai, 1986; Kramer et al., 1995; Montepare & Zebrowitz-McArthur, 1989).

Babyfaced people ranging in age from infancy to older adulthood are perceived to have more childlike traits than their more maturefaced peers, including less social autonomy and greater naivete, physical weakness, and warmth (Zebrowitz & Montepare, 1992). Notably, this research included ratings tapping perceptions of behavioral affordances, such as “listen to parents,” “give in to friends,” “not know right from wrong like an adult,” and “not able to follow complicated instructions.” Vocal immaturity also yields perceptions of childlike traits in both children and adults (Berry et al., 1994; Zuckerman et al., 1995). Evidence for effects of body immaturity is provided by the relationship of physical stature to perceived traits. Consistent with the babyishness overgeneralization hypothesis, shorter women were perceived to be warmer, less intelligent, and less assertive than taller women (Chu & Geary, 2005). Children as young as 3 to 6 years also perceived shorter men and women as weaker and more submissive than taller peers (Montepare, 1995), and even preverbal infants use height as a cue to social dominance (Thomsen et al., 2011). Other research has shown that childlike trait perceptions of babyish adults generalize across diverse cultures (McArthur & Berry, 1987; Montepare & Zebrowitz-McArthur, 1987; Zebrowitz et al., 2012).

As is true for attractiveness, some traits whose perception varies with the maturity of stimulus information are also accurately perceived. Research has demonstrated correlations between vocal babyishness and measures of actual warmth and low dominance (Berry, 1990) as well as aggressiveness (Berry, 1991). Research investigating links between shorter stature and accurately perceived traits found that shorter people were more submissive (Melamed, 1992) and less intelligent (Humphreys et al., 1985). Babyfaceness was correlated with greater warmth/agreeableness and less dominance in college students (Berry, 1991; Berry & Brownlow; 1989; Berry & Landry, 1997; Borkenau & Liebler, 1995). On the other hand, research investigating a representative sample of individuals across the lifespan found that only for women in later adulthood was there a positive relationship between babyfaceness and childlike traits. In contrast to the accuracy of perceptions for older women, they were inaccurate for babyfaced boys and young men, who showed contrary traits and behavior, including greater assertiveness, hostility, and likelihood of winning a military award (Collins & Zebrowitz, 1995; Zebrowitz et al., 1998). Also, in a large at-risk sample of adolescent boys, those who were more babyfaced were more likely to be delinquent (Zebrowitz et al., 1998). The mixed results regarding the accuracy of perceptions related to immature stimulus qualities may implicate perceiver actions, as discussed in the next section.

Whereas immature stimulus qualities elicit trait perceptions overlapping those that are accurately perceived and are also significantly related to those traits, none of the aforementioned studies directly tested whether variations in babyishness mediate accurate trait perceptions.

Perceiver action

A close connection between action and perception is central to ecological theory. If perceiving is for doing, then it is the accurate perception of behavioral affordances that enables the perceiver to engage in adaptive behaviors. In addition, the theory holds that accurate perception will be enhanced when a perceiver can actively explore the stimulus information. An example of such exploration is Reuben’s reference to “event-activity tests” that can improve the accuracy of perceiving someone’s affordances. These are situations in which the observed behavior is more diagnostic of a person’s traits. Some experimental results support this possibility, although the “event tests” were provided by the experimenter, not the perceivers. Perceivers viewing someone perform the single task of reading newspaper headlines aloud exhibited more accurate perceptions of intelligence than those viewing various other tasks. On the other hand, watching someone perform the single task of pantomiming various uses for a brick yielded the most accurate perceptions of the personality trait “openness to experience” (Borkenau et al., 2004). Thus, different situations elicit stimulus information that is diagnostic of different traits, akin to the person × situation effects in Mischel’s (1968) reconceptualization of personality. Perceivers in social interactions may spontaneously or deliberately create event activity tests that can advance their accuracy in detecting the presence or absence of affordances consistent with their goals or behavioral capabilities. However, the research we have reviewed has rarely examined the affordances perceived by someone who is doing anything besides completing paper-and-pencil ratings of a target person.

Despite Gibson’s suggestion that the ecological theory of perception applies equally to social perception, an important difference is that, unlike properties of the physical world, those of the social world are more likely to vary as a function of the perceiver’s actions. This highlights the necessity of understanding when perceivers’ actions might influence rather than merely reveal stimulus information that is diagnostic of affordances. One significant example is actions that are influenced by perceivers’ expectations, which can elicit stimulus information that fulfills or defeats those expectations.

Self-defeating prophecies may be relevant to the mixed results regarding the accuracy of perceptions related to immature stimulus qualities. More specifically, it may be that some babyfaced or short men, but not girls and women, react against being treated as if they will exhibit childlike warmth and weakness in the physical, social, and intellectual domains by behaving contrarily to the undesirable expectations engendered by babyishness overgeneralization (Zebrowitz et al., 1998). Although this reactive behavior might be evident to active perceivers in a social interaction, where additional stimulus information is available, the passive perceivers in the existing research had no opportunity to acquire more stimulus information than static facial appearance, and babyishness overgeneralization produced erroneous perceptions. Some evidence consistent with the suggestion that actively exposing oneself to stimulus information can improve accuracy is provided by the finding that those who were urged to form accurate perceptions of people interviewing for a job gathered more stimulus information and avoided an impact of prior expectations (Neuberg, 1989).

Although perceiver actions might reveal valid stimulus information in some instances, there are others in which they may influence it. In a classic study of the self-fulfilling prophecy effect, men were shown a photo of an attractive or unattractive woman before a “getting acquainted” phone conversation allegedly with that woman but, in fact, with a woman not selected based on appearance (Snyder et al., 1977). The men spoke more warmly to a woman whom they thought was attractive and these women responded in kind. Thus, the woman’s behavioral affordances were influenced by the perceivers’ actions. Although the men may thus accurately perceive the affordances they elicited, their perceptions are unlikely to show what Swann (1984) called “global accuracy,” which enables accurate perception of how the target will generally behave. Rather, perceivers will show “circumscribed accuracy,” which is accurate perception of how the target will behave in the presence of a particular perceiver.

Integrating person × situation effects with self-fulfilling prophecy effects yields an interesting thought experiment. What if the intelligence of the women’s behavior on the telephone was assessed rather than their sociability, as it was by Snyder et al. (1977). Even though more attractive people are generally perceived as more intelligent, and they are, it seems unlikely that the men’s differential behavior toward women perceived to vary in attractiveness would elicit stimulus information indicative of differences in intelligence. This is because a “getting acquainted” phone conversation is not a particularly good event-activity test for discerning intelligence. Now imagine that the men want to decide whether to choose the woman as a partner on an intellectually challenging task. Perhaps in this case, their conversations will elicit stimulus information pertinent to intelligence. Then what happens? Will there be a self-fulfilling prophecy whereby the women perceived to be more attractive evidence greater intelligence? Perhaps not insofar as people who are not actually intelligent may have difficulty displaying high intelligence. Thus, the presence of a good diagnostic situation (event-activity test) may yield global accuracy, with perceptions of the woman’s intelligence unbiased by expectations and accurate beyond that particular social interaction.

Some research has demonstrated that perceivers’ actions can influence the stimulus information available to detect social affordances quite apart from actions guided by expectations that yield self-fulfilling or self-defeating prophecies. A nonverbally supportive physician led people to display pain cues that were more valid indicators of their experienced pain than did an unsupportive physician. In addition, uninvolved perceivers showed greater pain assessment accuracy in the former case (Ruben & Hall, 2016). Presumably, accuracy would also be higher for the nonverbally supportive physicians. Whereas the actions in this study were spontaneous, the study mentioned above (Neuberg, 1989) reveals that a pereceiver’s deliberate goals can also affect the information available to detect social affordances.

In summary, active perceivers in the social domain may be more accurate than inactive ones insofar as they are able to elicit diagnostic stimulus information for the affordances they are motivated to perceive. On the other hand, active perceivers in interpersonal interactions may vary in the stimulus information that their activity makes available, making circumscribed accuracy more common than in nonsocial perception. This is not necessarily problematic for ecological theory, since affordances are the action opportunities for a particular perceiver, which may or may not generalize to other perceivers. This is also true in the nonsocial domain. What differs is that active perceivers in the social domain are more likely to influence the available stimulus information.³

Conclusions

The signature of the ecological theory championed by Reuben Baron can be seen in a broad swath of social psychological research on person perception since publication of Toward an Ecological Theory of Social Perception (McArthur & Baron, 1983). An interest in the accuracy of social perceptions has flourished, as has an interest in uncovering the stimulus information that gives rise to the perceptions, whether they are accurate or not. Moreover, much of this research is grounded in theoretical assumptions regarding which stimulus information will reveal affordances that can inform adaptive perceiver actions. While this work is consistent with ecological theory tenets, there are several shortcomings that future research needs to address. One is in the research on stimulus information. We know that very short durations of information are sufficient for statistically significant accuracy. What is needed are more systematic variations in the modalities of stimulus information to determine whether (and which) additional modalities increase accuracy. Research is also needed to determine whether (and when) active perceivers elicit information that improves accuracy. Finally, as Reuban lamented more than a decade ago, the research reviewed here has taught us much about the independent variable of stimulus information but has rarely used paradigms that elucidate which behavioral affordances it conveys to a particular active perceiver. Although addressing this question will take more methodological effort than current research paradigms, we are confident that Reuben would want researchers to take on the challenge.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. For a more current discussion of Gestalt principles, see Wagemans, J., Elder, J. H., Kubovy, M., Palmer, S. E., Peterson, M. A., Singh, M., & von der Heydt, R. (2012). A century of Gestalt psychology in visual perception: I. Perceptual grouping and figure–ground organization. Psychological Bulletin, 138(6), 1172–1217. https://doi.org/10.1037/a0029333.

2. We recognize that very early research in the 1930s and 1940s focused on accuracy, but it unfortunately fell out of favor, partly in response to Cronbach’s (1955) argument that scores geared to identify perceivers with greater accuracy were compromised by several possible artifacts.

3. We must acknowledge that our discussion of perceiver action remains stuck in what Marsh et al. (2006) called the individual-level approach, which examines how others influence either perceptions or actions using the individual as the unit of analysis rather than the social synergy approach that these authors, including Reuben, favored.