The purpose of perception is action (to paraphrase Bruner Reference Bruner1957b), and so it is important that our percepts be reasonably accurate. And evidently they are, or else we would not have survived so long as a species and as individuals (or maybe the Universe is just very forgiving). Nevertheless, over the last few decades many social psychologists have come to embrace the view that social perception is riddled with error and bias – a framework that I have dubbed the “People Are Stupid School of Psychology” (Kihlstrom Reference Kihlstrom2004b; see also Kihlstrom Reference Kihlstrom2004a; Reference Kihlstrom2004c; Reference Kihlstrom, Baer, Kaufman and Baumeister2008). The tenets of “stupidism” may be summarized as follows:

1. 1. People are fundamentally irrational: In the ordinary course of everyday living, we do not think very hard about anything, preferring heuristic shortcuts that lead us astray; and we let our feelings and desires get in the way of our thought processes.

2. 2. We are on automatic pilot: We do not pay much attention to what is going on around us, and to what we are doing; as a result, our thoughts and actions are inordinately swayed by first impressions and immediate responses; free will is an illusion.

3. 3. We don't know what we are doing: When all is said and done, our behavior is mostly unconscious; the reasons we give are little more than post-hoc rationalizations, and our forecasts are invalid; to make things worse, consciousness actually gets in the way of adaptive behavior.

4. 4. We don't know what we want: We are extremely poor at predicting how we will feel about various eventualities, and we are so poor at making choices that we might just as well let others choose for us – largely because, again, we don't have accurate introspective access to our beliefs, feelings, and desires. One is reminded of the joke about the two behaviorists who had sex: one said to the other: “It was good for you, but was it good for me?”

5. 5. We don't even know how stupid we are: Because of the limitations on our cognitive abilities, we fail to appreciate when our judgments and behaviors are less than optimal.

Stupidism – to the extent that it is not just a figment of my imagination – was in some respects an unanticipated consequence of a very reasonable program of research which employed evidence of errors to produce a more realistic description of how people actually make judgments and decisions. But there are even deeper roots of social psychology's preference for the thoughtless, unconscious, automatic, biased, and error-prone. Somehow, fairly early on, social psychology got defined as the study of the effect of the social situation on the individual's experience, thought, and action (G. W. Allport Reference Allport, Lindzey and Aronson1954a; see also Kihlstrom Reference Kihlstrom and Carlston2013). And, perhaps in a quest for institutional approval, it got tied to the functional behaviorism of Watson and Skinner (Zimbardo Reference Zimbardo, Rodrigues and Levine1999). Think, for example, of the classic work on the “Four A's” of social psychology: attitudes, attraction, aggression, and altruism; think, too, of the history of research on conformity and compliance, from Asch and before, to Milgram and beyond. In each case, the experimenter manipulates some aspect of the environment, and observes its effect on subjects’ behavior. Sometimes there were inferences about intervening mental states, but not very often – otherwise, the cognitive revolution in social psychology wouldn't have been a revolution.

Occasionally there have been attempts at correction (e.g., Gigerenzer et al. Reference Gigerenzer and Todd1999; Hastie & Dawes Reference Hastie and Dawes2001; Krueger & Funder Reference Krueger and Funder2004; Malle Reference Malle2006). For example, the self-other difference in causal attribution appears not to occur, at least in the form that is usually claimed for it; and, by extension, the “Fundamental Attribution Error” turns out to be problematic, too (someone, not me, once quipped that the Fundamental Attribution Error isn't an error, but it is fundamental). Still, errors and biases are so much a part of the current social-psychological Zeitgeist that these critiques have not, seemingly, had much impact on how psychologists think about social interaction. Now comes Jussim (Reference Jussim2012; and present BBS target article) with the heavy artillery, systematically dismantling most of the canonical claims for the power of error and bias. And pretty convincingly, too.

But it is one thing to argue for the fundamental accuracy of social perception, and quite another thing to argue for a particular view of perceptual realism, and against a particular view of constructivism. Social constructivism shouldn't be abandoned entirely – not least because, despite the exaggerations of so much constructivist theory (Hacking Reference Hacking1999), so much of the social world is a social construction (Searle Reference Searle1995; Reference Searle2011). But Jussim seems to opt for some version of perceptual realism, which is not the only alternative.

Historically, the study of perception has been framed by two competing paradigms (Epstein Reference Epstein1979; Epstein & Park Reference Epstein and Park1964; for a complete review, see Palmer Reference Palmer1999).The most influential approach, beginning in the 19th century with Helmholtz and continuing in the 20th with Hochberg, Gregory, and Rock, is, indeed, constructivist in nature. Helmholtz and the others argued that stimulus information is vague, fragmentary, and ambiguous, and that the perceiver must, in Bruner's (Reference Bruner, Bruner, Brunswik, Festinger, Heider, Muenzinger, Osgood and Rapaport1957a) phrase, “go beyond the information given” by the stimulus by drawing on knowledge, memory, expectations, and inferences (even unconscious inferences) to form a mental representation of that is the most likely interpretation of stimulus information – an interpretation that may be inaccurate in important respects. Perceptual constructivism has been challenged by Gibson's theory of direct perception, or ecological optics, which holds that all the information needed for perception is provided by the stimulus environment, and that our perceptual apparatus evolved to pick up just that information which allows us to perceive the world the way it really is. Some former constructivists were persuaded by this point of view (Neisser Reference Neisser1976a; Reference Neisser1976b), and some advocates have gone so far as to argue that there are no “top-down” cognitive influences on perception at all (Firestone & Scholl Reference Firestone and Scholl2016).

Jussim, by emphasizing realistic accuracy over constructivist error and bias, seems to incline toward the Gibsonian view. A Gibsonian approach has also been embraced by some other social psychologists, (e.g., McArthur & Baron Reference McArthur and Baron1983), and indeed there is a great deal about social perception that can be studied from the ecological point of view. There is a lot of information in the stimulus field, and its background context, and it seems particularly appropriate when analyzing facial emotion, lie detection, and other aspects of person perception which may be largely based on physical appearance and gesture. At the same time, there is a lot of evidence favoring the (Helmholtzian) constructivist view, and some of it even comes from errors on these very tasks. It seems that person perception is prone to inaccuracy, after all.

For example, people do not seem to be particularly accurate at detecting deception, largely because their naive theories of deception lead them to pick up on the wrong cues (e.g., Bond & DePaulo Reference Bond and DePaulo2006; Reference Bond and DePaulo2008; Hartwig & Bond Reference Hartwig and Bond2011; Reference Hartwig and Bond2014). Our “gaydar” does not appear to be that good, either, once we take account of base-rates (e.g., Bruno et al. Reference Bruno, Lyons and Brewer2014; Lyons et al. Reference Lyons, Lynch, Brewer and Bruno2014; Poderl Reference Poderl2014) – a problem that bedevils the detection of deception as well. Even our accuracy at reading emotion from facial expressions – which seems the likeliest candidate, in the social domain, for an evolved, hard-wired, perceptual module of the Gibsonian sort – seems to be inflated by such method factors as the use of a forced-choice response format (e.g., Hassin et al. Reference Hassin, Aviezer and Bentin2013; Nelson & Russell Reference Nelson and Russell2013).

Although Jussim is right to be skeptical of a radical social constructivist approach which denies the existence of an independent reality, it would seem that the nature of social reality invites a perceptual-constructivist approach. Bruner and Tagiuri (Reference Bruner, Tagiuri and Lindzey1954), in an early analysis of person perception, listed a number of factors that influence perceptual organization, including the stimulus array itself (a prescient nod toward Gibson), but also selective attention, linguistic categories, and especially the internal state of the perceiver – his mental set, or expectations, and his own emotional and motivational state. Much as the stimulus array for nonsocial perception consists of the energy (light waves, sound waves, etc.) that radiates from the distal stimulus, falls on the sensory surfaces, and is transduced by receptor organs into neural impulses, the stimulus array for person perception also consists of the person's appearance and behavior, as well as the language that others use to describe the person. Much more so than the nonsocial case, the interpersonal stimulus is almost inherently vague, fragmentary, and ambiguous, affording a great deal of room for divergent interpretations. Often, the environment provides conflicting cues as to the nature and activity of the stimulus person, increasing the difficulty of forming an accurate perceptual representation of reality. Moreover, the social situation provides plenty of leeway for emotion and motivation to bias perceptual-cognitive processes (AbelsonReference Abelson, Tomkins and Messick1963; Bruner Reference Bruner1992; Bruner & Goodman Reference Bruner and Goodman1947; Bruner & Klein Reference Bruner, Klein, Wapner and Kaplan1960). While all theories of perception, including Gibson's, assume that the context makes a great deal of difference to perception, context effects are arguably even more salient in the social world, so that the same person, or behavior, may be perceived differently, depending on the situation – which is itself inherently vague, fragmentary, and ambiguous. For all these reasons, the social perceiver must fill in the gaps, and resolve the ambiguities, by making inferences about the stimulus given his knowledge, expectations, and beliefs. This is the expressly cognitive contribution of the perceiver to perception; and in this constructive activity lies the possibility for error and bias to occur.

Brunswik's (Reference Brunswik1955a; Reference Brunswik1955b) lens model offers one framework for conceptualizing these constructive processes. The stimulus may provide ecologically valid cues as to its nature, but the perceiver has to utilize those cues in order to form an accurate mental representation of the stimulus; if the perceiver utilizes the wrong cues, or weights valid cues incorrectly, the representation will be inaccurate or biased. Neisser's (Reference Neisser1976a) idea of the perceptual cycle offers a similar framework. The stimulus provides information to the perceiver, but the perceiver's exploration of the stimulus is guided by internal cognitive schemata; eventually, the cycle of assimilation and accommodation should result in an accurate mental representation of reality – provided, of course, that the stimulus is richly informative in the first place, and the cycle is allowed to run to completion. Neither is always the case, especially in the social domain – hence, the intrusion of error and bias.

Jussim is right to offer a corrective to the current emphasis on error and bias in social perception – though, as my examples indicate, there remain plenty of opportunities for error and bias as well. This is the price we pay for living in a world in which perception and cognition occur under conditions of uncertainty. As with the literature on bounded rationality exemplified by the program of research on judgment heuristics, anomalies of perception and cognition can tell us a great deal about how social perception actually works. More important, though, the choice Jussim offers between perceptual realism and social constructivism is a false one, because these are not the only choices available. There is at least a third way of cognitive constructivism, which allows us to understand both accuracies and inaccuracies in perception, where and when they occur.