The puzzle of social norms
In the analytical tradition, norms are thought of as devices that allow social dilemmas to be overcome (Ullmann-Margalit 1977; Coleman 1990). A social dilemma exists when the behavior of individuals, even if rational, may produce an undesired, suboptimal social outcome. Rousseau’s parable of the stag hunt, where two men must cooperate to hunt a stag, but either of them may be easily tempted to trap a hare, is the classical illustration. Extending this reasoning to the topic of this chapter, all workers of a production plant may have an incentive to follow some norm of resistance; however, workers may be tempted to defect from this common interest either to avoid possible punishments from a supervisor or to fulfill upward mobility expectations.
In order to preclude these temptations, at least three possible mechanisms could operate in real societies. First, individuals may hold moral preferences towards the collective good; meaning that “moral norms” guide their behaviors. Second, individuals may be frightened by an external authority (the state), which holds the capacity to punish them if they do not pursue the collective good; meaning that individuals are ruled through “legal norms”. Third, individuals may mutually encourage or discourage each other’s behavior by promoting the collective good; meaning that they are ruled by a “social norm”. Of course, these three mechanisms actually operate concurrently, making it difficult to empirically ascertain whether the observed behavior is due to one, two or all three of them. Nevertheless, by analytically isolating one of them, we can better understand how it works. Obviously, it is impossible to do so in an empirical setting (you cannot, for instance, “switch morality off” in real individuals), but it can be done (clearly, under simplifying assumptions) in an artificial society (see below).
The first task to accomplish is to answer a simple, but not easy, question: what exactly is a social norm? The question is not an obvious one because, turning to sociological literature, we find definitions of social norms with very different implications: punishment-based definitions, expectation-based definitions and emotion-based definitions.
Thus, for the political scientist Axelrod (1986, 1097 - emphasis added): “A norm exists in a given social setting to the extent that individuals usually act in a certain way and are often punished when seen not to be acting in this way”. This definition assumes that the existence of a certain norm can be empirically tested because it involves two kinds of behavior: the usual behavior performed by individuals, and the sanctioning behavior also performed by them. On other grounds, for the analytical philosopher Bicchieri (2006, 2 - emphasis added), “the existence of a social norm depends on a sufficient number of people believing that it exists and pertains to a given situation, and expecting that enough other people are following it in those kind of situations”. Like the preceding definition, this one assumes that a social norm is something shared by a number of people, but stresses a component which is more difficult to test empirically: beliefs and expectations. For Bicchieri, sanctions may or may not be a component of “social norms,” which is to say that they are not a necessary condition for them to exist. Finally, for the economist, Ostrom (2005, 121-122 - emphasis added), “Norms are prescriptions held by an individual that an action or outcome in a situation must, must not, or may be permitted [...] The changes may occur as a result of intrinsic motivation such as pride when keeping a norm or guilt when breaking a norm” In this case, as in Axelrod’s definition, the emphasis is placed on sanctions; but the nature of the sanctioning process is somehow different, since it is assumed to be an internal (and emotional) one. Of course this process may be triggered (and usually will be triggered) by the awareness of not fulfilling others’ expectations, and therefore implies that the behavior is performed under the possible scrutiny of others (Elster 2009).
What of these components (external sanctions, expectations, internal emotions), if any, are necessary conditions to account for the existence of a social norm? In this chapter I will support a “combined definition” (Opp 2001; see also Horne 2001), and assume that:
- (1) A social norm is a non-intended process, whose outcome is the emergence of a certain pattern of behavior held by a large part of a population,
- (2) where the behavior of individuals is controlled in a decentralized way (meaning, by other individuals),
- (3) by means of an influence process based on the progressive spreading of the belief of what is expected in that population, and
- (4) where that process is backed by informal sanctions which may (but not necessarily) produce certain emotions such as pride or guilt.
For most sociologists, the existence of “informal sanctions” is a necessary condition for a social norm to exist. Other kinds of regular behavior, such as conventions or social routines produced by mutual expectations are not, in fact, social norms because they do not require social sanctions in order to be maintained over time. The role played by sanctions is, however, mainly as a deterrent. Both in case studies (Ostrom 2005) and experimental settings (Fehr and Catcher 2000), it is a well-established empirical fact that sanctions are rarely performed. Any understanding of social norms must therefore account for this fact, and provide an explanation for why so little sanctioning behavior is required to maintain prosocial behavior. I will assume (Linares 2012) that the mechanisms which produce this fact work as shown in diagram 1, which represents a cycle where:
- (1) The initial triggering of sanctions pushes the generalization of beliefs about expected behavior (influence-spreading mechanism).
- (2) The more generalized the expected behavior, the less triggering of sanctions is required (sanctioning-reduction mechanism).
Diagram 1: Mechanisms producing a social norm