Abstract
In the target article, Heyes (2022) delivered a forceful critique of the nativist cognitive-evolutionary accounts that have dominated norm psychology to this point (e.g., Sripada & Stich, 2006). However, in developing her cultural cognitive-evolutionary account, Heyes repeated a mistake made by the cognitive-evolutionary accounts by overlooking a core explanandum of normative psychology: normative motivation, or a feeling that one “ought to,” is “obligated to,” or “should” do something, as opposed to a nonnormative feeling of desire/aversion (for review, see Theriault, 2023; Theriault et al., 2021a, 2021b; see also, Asch, 1952/1962, Chapter 12). Nonnativist accounts of normative motivation, like Heyes’s, must thread a needle in describing how domain-general processes could produce distinct kinds of motivation, and unfortunately, nothing in Heyes’s account appears to specify how distinctly normative motives develop.
Despite its problems, the Sripada and Stich (2006) model makes an important distinction between the acquisition and implementation of social norms. Norm acquisition has been analogized to Chomsky’s work on grammar, in which it might involve identifying the “deep structure” or “syntactic quality to social norms” (Shaffer, 1983, p. 287), and Heyes made a convincing case against poverty-of-the-stimulus arguments that follow from this line of thought. Norm implementation, however, involves an affective response that motivates norm-consistent behavior (i.e., normative motivation). Although the causes of normative motivation have remained unclear, as an empirical phenomena, it has long been recognized in social psychology.
The importance of normative motivation has been obscured by changes in terminology and focus. The modern distinction between descriptive and prescriptive/injunctive norms (Cialdini et al., 1990, 1991) stems from work extending the Asch (1951, 1955) conformity experiments. But this work was not interested in distinctions among kinds of norms; rather, it was interested in distinguishing sources of influence—that is, informational and normative influence (Deutsch & Gerard, 1955). Informational influence created a nonnormative motivation to accept evidence about reality, whereas normative influence created a normative motivation to conform to expectations. The distinction between descriptive and prescriptive/injunctive norms (Cialdini et al., 1990, 1991) conflated norm acquisition with the affective experience of norm implementation. That is, Cialdini assumed that different kinds of social rules were paired with different kinds of affective experience: Descriptive norms were “norms of ‘is,’” and prescriptive/injunctive norms were “norms of ‘ought’” (Cialdini et al., 1991, p. 203). Recent evidence, cited by Heyes, has refuted this assumption and shown that descriptive norms can produce prescriptive judgments (e.g., Bear & Knobe, 2017; Eriksson et al., 2021; Roberts et al., 2017); but this work bypassed the more fundamental question of why normative and nonnormative motivations feel different.
One problem for any account of norm psychology, then, is to explain how normative motivation, as a distinct feeling, is produced. Nativist accounts attempt to make short work of this problem by proposing that innate emotional predispositions are elicited by norm violations (e.g., anger, contempt, disgust; Sripada & Stich, 2006; but see Munch-Jurisic, 2022). But this does not really solve the problem: Feeling angry or disgusted is distinct from feeling obligated. Indeed, even feeling anger intensely enough to make you hurt someone is different from feeling obligated to hurt them (Fiske & Rai, 2014). Furthermore, normative motivation can apply to behaviors outside of the third-party enforcement that Sripada and Stich (2006) focused on. Nativist accounts of norm psychology, then, do not clarify how normative and nonnormative motives come to feel distinct.
Like the nativist account, Heyes’s cultural account cannot distinguish normative and nonnormative motivation. On her account, domain-general reinforcement learning creates an implicit repertoire of compliance and enforcement behaviors. This reinforcement is what initially gives the implicit behaviors motivational force. To flexibly specify norm content, implicit behaviors are combined with explicit expectations (i.e., commentary); but again, the implicit processes are the engine of motivation:
A child without a repertoire of compliance and enforcement behavior, acquired by implicit processes, would struggle to get the message about others’ normative expectations . . . because there would be little of personal relevance to be explained by the expectations of others. (Heyes, 2022, p. 45).
This can explain the flexibility of social norms, and it can explain how norms acquire a general motivational force, but it cannot explain why people feel obligated to follow them.
Two aspects of Heyes’s account might recover the distinction between normative and nonnormative motivation, but closer inspection shows that they cannot. First, like Sripada and Stich (2006), Heyes proposed that norms are intrinsically motivated such that “people are motivated to comply with norms as ultimate ends, rather than as a means to other ends” (Heyes, 2022, p. 8). However, intrinsic motivation actually says nothing about normative phenomenology. Heyes showed that intrinsic motivation can develop through domain-general mechanisms such as conditioned reinforcement (Heyes, 2022, p. 36; Kruglanski et al., 2018), but nothing about this mechanism necessarily produces feelings of obligation. Indeed, the examples of means–end fusion given by Kruglanski and colleagues (2018) are clearly nonnormative: “‘Running’ when fused with the goal of ‘fitness,’ is subjectively experienced as (the attainment of) fitness. ‘Having a drink at a bar’ when fused with the goal of ‘unwinding’ becomes the experience of unwinding” (pp. 167–168).
Second, normative and nonnormative motivation may be distinguished by emotional construction, the categorization of sensory signals (including interoceptive signals, which underlie affect; Barrett, 2017b, 2022; Hutchinson & Barrett, 2019) “into full-blown emotions such as shame, guilt, and moral rage—the intrinsic motivators of explicit normativity” (Heyes, 2022, pp. 45–46; although such categorization need not be explicit, Barrett, 2017a, pp. 104–107). Kohlberg (1971, pp. 189–190) made a similar constructivist point about the categorization of affect and observed that distinctions between nonnormative (preconventional) and normative (conventional) motivation reliably emerge across cultures around the ages of 10 to 13 (Kohlberg, 1971; Table 4). However, that a learned distinction between normative and nonnormative motivation reliably develops suggests that development is scaffolded. Emotional constructivism on its own does not provide that scaffold. For example, a recent cultural-evolutionary account of emotional constructivism (Lindquist et al., 2022) addressed how cross-cultural similarities in emotion concepts might emerge through (a) forces of cultural attraction, (b) social learning, or (c) environmental shaping. Likewise, we proposed elsewhere (Theriault et al., 2021a) that normative motivation (i.e., a sense of “should”) develops through a special case of social environmental shaping: When others’ expectations are violated, they react in ways that are less easily predicted, and their unpredictable reactions produce a reliable link from behaviors that violate others’ expectations (i.e., violating norms) to the metabolic, interoceptive, and affective consequences of prediction error reflected back by the social environment. Heyes (2022) acknowledged that “prediction error, produced by the occurrence of unexpected events, is associated with increased arousal which is typically aversive” (p. 37), but she did not elaborate on how particular patterns and sources of social reinforcement might structure distinctions that constructivism can magnify. Without such attention to such latent structures of reinforcement, Heyes’s account risks being reduced to a domain-general account of social learning and missing its target of normativity completely.
