Groups and Categories
Groups and categories are the elements of social organization in human societies of any size. Whereas the term ‘group’ primarily refers to sets of identifiable individuals, the term ‘category’ primarily refers the mental representation to the building blocks of the social world. Social organization and its mental representation have co-evolved to striking degrees of complexity; yet, simple – even crude – behavioral patterns such as ingroup-favoritism or parochial morality persist.
Humans are social animals, both in their dependency on interaction with individual conspecifics and in their embeddedness in a social structure comprising groups. The nature of these groups varies greatly with ecological, economical, and historical conditions, making it difficult to uncover the essential characteristics of a group-shaped psychology. Studies of the fossil record and surviving hunter-gatherer societies suggest that hominid social life has long been hierarchically organized. Small units, such as temporarily pair-bonded mates and their close kin, are vital for reproduction; bands comprising a few dozen individuals cooperate in hunting, gathering, and shelter-making; communities comprising multiple bands share a language or dialect, and they may intermittently congregate for ritual activities or common defense. This multilevel social organization has evolved during the Pleistocene, and it distinguishes the only surviving human species from the great apes and other social animals (Layton et al. 2012; cf. Darwin 1871).
Evolutionary psychology seeks to understand how and why this prevalent social organization obtained its present form, and inasmuch as the properties of human social organization are functional, what are the adaptive problems they have solved. Moving to the individual human being as the unit of analysis, evolutionary psychology explores elements of cognition, affect, and behavior that can be properly understood only with reference to the evolutionary legacy of past generations.
With regard to the first question, research highlights the co-evolution of neocortical size, mental capacities, and social organization. Compared with other primates, humans can keep track of a greater number of individual conspecifics, who they are, and, more importantly, how they have acted in the past. Humans can even represent and keep track of those conspecifics who are currently out of sight, often for a long time, and even the deceased. From this knowledge, humans can compute expectations about future behavior and choose between aggression and appeasement, trust and suspicion, pursuit and avoidance. The complexity of these perceptual, memorial, and executive functions is correlated with the size of the neocortex relative to total body size over living species of social animals, and, very probably, over extinct hominid species in evolutionary time (Dunbar 1993). The extraordinary evolution of the human brain is thus intertwined with the complexity of the social organization humans can afford.
A key correlate of advanced psychological functioning is the human capacity to sustain a sophisticated social organization while being able to tolerate very low population density – lower than in any of the great ape species. Bands of hunter-gatherers can roam over large territories, exploiting scarce resources, while occasionally congregating with other bands to affirm shared community identities. Shared language and ritual practice afford community cognition, affect, and behavior. Although some of these shared psychological features reside within the individual and may remain unexpressed (e.g., certain types of affect or elements of the self-concept), many features are expressed in a way that signals shared group identity to others. The evolutionary bedrock of shared identity is genetic relatedness (Dawkins 1976). Social structure is often superimposed on and often correlated with genetic distance, but sometimes it is not. Humans have developed cross-cutting social categories that further complicate social structure. Many hunter-gatherer communities subdivide into clans and totem-defined structures that mimic sympatric bands, by, for instance, imposing taboos on endo- or exogamy, but that are, from the outside observer’s perspective, arbitrary (Stanner 1965).
While the hierarchical and multilevel organization of social life is a human universal, its observed forms vary greatly. The flexibility of the social brain and especially the ease with which it creates, changes, and eliminates categories or groupings spawns ever-new structures. The minimum criterion for a social structure to become mentally represented and socially consequential is a signal that communicates inclusion versus exclusion. The totem-based clan is the evolutionary archetype of the influential minimal-group paradigm (MGP) in social psychological research (Tajfel 1970). The MGP mimics social groups or categories that have no basis in genetic relatedness, differential familiarity, or intended future interaction. Research participants are categorized at random, with the only mark of distinction being a quasi-totemic label, such as “The Greens” or “The Blues.” After the typical hour-long experiment, participants are disbanded. In contrast, many arbitrary marks of group belongingness in the wild are irreversible, such as circumcision.
Despite its skeletal and fleeting nature, the MGP produces many of the shifts in cognition, affect, and behavior observed among groups in the field. Most prominent among these phenomena are perceptual contrast and assimilation effects (cognition), ingroup love and positive social identity (affect), and ingroup cooperation versus outgroup competition (behavior). Moving to the second question with the individual as the unit of analysis, we can now consider the three sets of phenomena in turn. Cognitively, the categorization of people into mutually exclusive groups leads to the accentuation of intergroup differences where such differences exist (contrast). If, for example, differences in body height covary with an arbitrary criterion of demarcation, the shorter ones will be collectively viewed as shorter and the taller ones as taller than they would be in the absence of such a criterion. At the same time, individual differences within categories are minimized (assimilation) (Krueger and DiDonato 2008; Tajfel 1969). Affectively, categorization affords identification with those who are like us, even if by label only. In turn, this shared positive affect affords a sense of positive social identity, which extends – usually in mitigated form – positive self-sentiments to the group (DiDonato et al. 2011) and inversely reflects positive features of the group back to the self (Van Veelen et al. 2016). Behaviorally, categorization affords coordination and cooperation with ingroup members, which, in turn may set the stage for group selection effects over evolutionary time (Wilson 1975).
A minimally sufficient theoretical account to explain the correlated patterns of categorical ingroup perception, ingroup love, and ingroup cooperation (Balliet et al. 2014) assumes that most people possess positive self-images, which they selectively project to ingroups but not to outgroups (Robbins and Krueger 2005). This differential projection results in comparatively positive ingroup perceptions and a comparatively high level of trust in ingroup members (Brewer 2008; Krueger et al. 2012). These models, built on assumptions about learning and inductive reasoning, do not require that people actively compare ingroups and outgroups. Learning about ingroups is sufficient so that competition with and discrimination against out groups can begin as a mere byproduct of parochial learning.
Whereas most social psychological theories assume – and show – that the social or arbitrary introduction of criteria for categorization triggers multifaceted psychological discrimination, some basic distinctions arise from even simpler processes. Human infants are highly sensitive to human and human-like faces. An innate positive response is available, providing a stepping-stone for the building of a mental archive of familiar faces. Infants quickly learn individual identities and thus person-to-person differences. The proximal benefit is that they know from whom to expect love and protection. As a result of this identity learning, infants pick up any variable that is correlated with identity, such as sex and race. Computerized simulations of such learning suggest that it is not necessary to introduce infants to ideas of sex or race; they will induce them from observed patterns of variation (Kramer et al. 2017). The appearance of an individual’s face varies somewhat from occasion to occasion, depending on the lighting, the angle of viewing, or the observed person’s mood. The learner’s challenge is to abstract identity from these variable observations. However, a person’s identity does not cross gender or racial boundaries once it is detected – in most cases. Therefore, sex and race information is implicit in identity information and can be easily extracted. Once the infant (or the algorithm) harvests a categorical distinction from observation, the door is open for accepting other distinctions between social groups and categories, including those provided by social learning.
The title of this essay, “Groups and Categories,” signals the unsettled nature of these concepts in the behavioral sciences. Biologists and social psychologists tend to prefer the concept of group as a general way of referring to any nonrandom sorting or the lumpiness of social organization. Cognitive scientists, in contrast, tend to prefer the concept of category, as it supports theories of general category learning, which just happen to include social stimuli. On this view, the lumpiness of the mind is prioritized over the lumpiness of the social world. From an evolutionary perspective, the two concepts are too closely intertwined to allow a strictly separable usage. As we have seen, it seems likely that the physical separation into seen and unseen groups coevolved with the ability to represent seen and unseen categories of greater abstraction. At the present time, there is little reason to assume that one concept preceded the other.
- Brewer, M. B. (2008). Depersonalized trust and ingroup cooperation. In J. I. Krueger (Ed.), Rationality and social responsibility: Essays in honor of Robyn Mason Dawes (pp. 215–232). New York: Psychology Press.Google Scholar
- Dawkins, R. (1976). The selfish gene. New York: Oxford University Press.Google Scholar
- DiDonato, T. E., Ullrich, J., & Krueger, J. I. (2011). Social perception as induction and inference: Anintegrative model of intergroup differentiation, ingroup favoritism, and differential accuracy. Journal of Personality and Social Psychology, 100, 66–83. https://doi.org/10.1037/a0021051.CrossRefPubMedGoogle Scholar