John Griffiths
Sociology of Law, Faculty of Law, University of Groningen, Groningen
3. Rules, social control and social order
Every year our nation is invaded by several million ‘things’ … that … threaten total chaos. They hold no brief for the Judao-Christian tradition. They have no modesty. They do not speak our language, nor know our history, nor value our customs. They lack any motive or knowledge that leads them to share, to give and take, to compromise, to accommodate, to cooperate…. They do not respect authority…. Though they do not swear, steal, chew tobacco, fornicate, desecrate the flag, or use four-letter words, they are unaware … that they should not do so….These invaders are human infants…. Something must be done. We need them … but we need them on our terms, not theirs.68
Conscience is the inner voice that warns us somebody may be looking.69
“Law”, I have argued, is for purposes of sociological theory best conceived of as a relatively differentiated form of “social control”. But what is social control? The place to begin, as sociology classically has, is with the “problem of social order”: how is human social behavior possible at all? If the “state of nature” had ever prevailed and human beings really once had been “rational actors” each pursuing his or her own immediate preferences, life would indeed in Hobbes’ words have been “solitary, poore, nasty, brutish and short”,70 and it would have remained so forever. Constraint of individual freedom of choice is a precondition of the coordinated behavior that constitutes social life. Human beings do this to an exceptional extent by subjecting individual behavior to rules.71 The “enterprise of subjecting behavior to the governance of rules”72 has come to be known as social control in the social–scientific literature.73 The concept of social control refers to the fact that the behavior of the members of a group in relation to one another is regulated. Maintenance of the rules resides largely and ultimately in the mutual relationships and interactions of the members of the group. Behavior that is regulated in this way can be called, for short, social behavior.
It is the rules for social behavior that permit the existence of “collective goods”, that is, states of affairs (like civil peace, property, language) that no individual can create on his own and that therefore cannot be brought into existence by individual volition.74 Rules can, of course, and often are produced and maintained by institutions – as in Schwartz’ two settlements – but the most fundamental forms of social control, from which the differentiated forms derive and on which they ultimately depend, take place in an undifferentiated way, in the course of everyday interaction. We learn how to follow and to use rules directly from personal experience: being told what they are, encountering positive and negative responses to our own behavior, and using rules to influence the behavior of others.75 We also learn vicariously, by observing the behavior of others and what happens to them.76 Except in the case of highly differentiated forms of social control – such as “law” – learning and sanctioning tend to be hardly distinguishable aspects of everyday interaction, but certainly not the less important for all that.
As Moore argues,77 it is groups that are the social locus of rule-creation, rule-learning, rule-following and of the social (dis)-approbation required to support rule-following. Some of these groups are known as “states” and their rules as “law”, a form of social control so elaborate that it consists not only of explicit rules, but also of explicit rules about rules (“secondary rules”78), and even of explicit rules about rules about rules (as in the case of a rule of constitutional law that limits the authority of a national legislature to create or modify rules).
3.1. Rules are facts
Anyone who says he doesn’t know what a fact is, is lying, unless he nonchalantly pours water in his gas tank.
(V. Icke)79
Rules of behavior – including “legal” rules – are fundamental to the sociology of law. It does not follow from this that the sociology of law is itself “normative” (in the sense of evaluative). The existence of a rule – whatever anyone thinks about it, such as whether or not it is morally “good” or “legally” binding – can be established in just as empirical a way as the existence of any other fact.80 It is, for example, a statement of fact that is either true or false – one of a rather complicated sort, to be sure, but nevertheless not an expression of a personal opinion or of approval or disapproval – that a man in Saudi Arabia can be married to more than one woman at a time, whereas in England he would be limited to one and in the Netherlands he can also marry a man, but only one at a time. It is also a fact that in the Netherlands, “business lunches” and “professional literature” are no longer deductible from one’s individual income for income tax purposes. And it is a fact that the Supreme Court of the United States (but not the Hoge Raad of the Netherlands) has the authority to assess the constitutionality of legislation. As Vincent Icke might have said, anyone who says he does not know that these are facts is either lying (or trying to be scientifically cute), or he blithely drives on the right when in England.
Non-“legal” rules are also facts, things you can bump into just as easily as you can bump into a door. Anyone who does not believe this is invited to perform a small experiment: behave in a socially unacceptable way at a job interview for a white-collar job – show up in shabby, dirty clothes, for example – and see whether doing this reduces your chances of getting the job. If that is indeed the case, then the dress code in question (however silly or offensive you may think it to be) seems to be a pretty “hard” fact in your society.
As far as the sociology of law is concerned, it is irrelevant whether a legal (or a social) rule is good or bad (although the fact that local actors judge it one way or the other can be very important). The question is, rather, how it works, what people do with it, and why? Not only desirable behavior (environmental protection, for example) but also morally neutral behavior (getting married?) and repellent behavior (slavery, genocide) takes place pursuant to rules. How law works to realize evil is, sociologically considered, the same question – and every bit as interesting and theoretically important – as how it works to accomplish good.
3.2. The definition, ontology, epistemology, methodology and pluralism of rules
One consequence of the growing acceptance of the social fact of legal pluralism has not received the attention it requires. Abandoning the idea that “law” is taxonomically distinct from other social rules cuts us loose from the comfortable (apparent) legal certainty of formal “law” and requires that we have at least roughly serviceable answers to some fundamental questions about the concept of a rule. What counts as a rule (the definition of a rule)? When does a rule exist (the ontology of rules)? How can one know whether rule R exists in a particular social field (the epistemology of rules)? And how can such knowledge be acquired by a researcher (the methodology of rules)? Social scientists are in practice often simply unreflecting legal positivists who take the corpus of “legal” rules to be a distinct entity whose nature is obvious, whose validity is settled by legislatures and higher courts, and which can be known by applying the rules of legal reasoning or, if one is not a lawyer, by consulting a lawyer. Even anthropologists and others who take the idea of “legal pluralism” seriously, generally tend to apply a more or less watered-down version of legal positivism to the rules of “indigenous (i.e. non-state) law”,81 and both sociologists and anthropologists often treat the existence and social significance of non-“legal” social rules as a matter of unproblematic social fact.82 A non-taxonomic approach to the concept of “law” requires a more careful approach to the concept of a “rule”.
3.2.1. Definition
What counts as a rule? To begin with, we are of course concerned with rules of behavior – not with rules of thumb or of craftsmanship (how-to-do-it rules for things like gardening or photography), nor with the names of observed natural regularities (like the “law” of gravity). Second, rules of behavior are social facts: it is groups that have them.83 Individuals may “make it a rule” to brush their teeth every morning for five minutes, but this is something different.84
“Rules” and “norms”
Some people prefer the term “norm” for (non-“legal”) rules. I suppose this is because they think social rules are in some way vaguer, diffuser, ontologically or epistemologically more problematic, less binding, more “normative” (closer to “values”), or whatever than “legal” rules. I do not believe any of this, and I think it important not to fall into the taxonomic trap by using different terminology for certain social rules which differ from (some of) the rest only in their degree of differentiation.
Caveat: The discussion here pays no explicit attention to propositions other than rules (i.e. principles, values, policies) that, as Dworkin (1977) has shown, play a critical role in “legal” reasoning. Dworkin’s analysis applies equally to less highly differentiated cases of social rules than “law”.
Rules are formulations of expectations that the members of a group have with regard to each other’s behavior85 that generally are (or can be) socially enforced; a rule of behavior limits a group member’s freedom to choose how to behave.86 These expectations are not a matter of prediction nor of the exercise of purely private (i.e. not rule-based) pressure between individuals.87 They are social expectations in the sense that breaching the rule entails the risk of a socially approved sanction. Such social sanctions consist of changes in the behavior of others, ranging from expressions of disapproval (which can be as subtle as the turning of a lip or the raising of an eyebrow), through temporary or longer lasting reduction of social interaction (“I expect him to stand behind his product if he ever wants to do business with meagain”88), all the way to effective expulsion from the social group concerned (and hence the loss of the benefits of group membership).89
There are also secondary social rules (rules about rules) that regulate the nature and extent of appropriate sanctions, who is authorized to react in such ways, what the proper procedural approach is to a case of violation, and so forth.90 These rules authorize the sanctions, in the sense that those who react in this way, can justify their behavior (otherwise itself a breach of the rules of social interaction) by pointing to the rule violated and to the relevant secondary rules, and can expect support from other members of the group. Such social sanctioning can, as we know from our discussion of Schwartz and the two Israeli agricultural settlements, be more or less differentiated.
It would be wrong to overemphasize the negative aspect of rules, assimilating all of them to the prohibitions of criminal law. Rules provide for facilities (how to do it rules, important among them: powers91), cooperation and coordination with others, and so forth, all of which greatly increase the range of behavioral choice, but which necessarily do so subject to particular constraints. Many rules do not themselves require any particular behavior: they simply provide, for example, for how a particular result is to be accomplished (you don’t have to get married, but if you want to, here is how to do it).92 The behavioral constraint imposed by such rules lies in the fact that if you do not follow them, you will not accomplish the desired result. If your “marriage”, for example, is not valid, you will not be entitled to the special privileges enjoyed by married people – in, for example, the areas of taxation, inheritance, kinship, parental power – and if you are caught exercising them anyway, your social group will make life difficult for you in one way or another.93 Similarly, someone seeking to exercise a power to which a group’s rules do not entitle him will simply not be obeyed (or worse: if the power in question is that of a king, for example, he risks being treated as a usurper, possibly at the cost of his head).
Finally, there is the problem of the “unit” of a rule. Behavioral requirements can be expressed at very different levels of generality. A rule can be so general (the requirement that a person take “reasonable care” to avoid injury to others) that the whole panoply of subrules it entails are themselves generally referred to as rules (e.g. the rule requiring a property owner to take adequate precautions to prevent small children from falling into his swimming pool). I do not believe there is any way to “solve” this problem94 – that is, any way of identifying the essential unit of a rule. Whether something is described as one rule or as a part of a rule or as many rules is simply a matter of convention and convenience: the observer does the characterization, it is not latent or inherent in the regulatory material itself. This conclusion requires that in comparative research, in particular, one be very careful to formulate the rules to be compared at a comparable level of generality.95
3.2.2. Ontology
The existence of a rule in a social group is what Durkheim calls a “social fact”, that is to say, it is a fact about the group. When can we say that R exists in a social group such as the New York garment industry (Moore 1973), an Israeli agricultural settlement (Schwartz 1954), the tuna court of Tokyo (Feldman 2006), or the farming community of a rural county in northern California (Ellickson 1991)? The matter is simple: to say that rule R “exists” in the group is the same thing as saying that the members of the group generally treat it as a binding rule of the group. And being able to say this means that one can observe – or could observe, if the question were raised – the moral argument and secondary rules required, in the group concerned, to establish the rule’s existence and proper interpretation.
The question what it means to say that a rule exists in a particular social group has received considerable (indirect) attention among ethologists who concern themselves with “moral” behavior among animals. There is, of course, a great deal of animal (and human) behavior that can be described in terms of regularities: the behavior normally occurs in particular circumstances, and in some cases is learned (e.g. a rat can learn to avoid food that once made it ill, or an ape to use a twig to get ants out of an anthill). In a more limited sense, there are apparently animals with behavioral expectations that are socially shared and that elicit remorse and sanctions when broken. De Waal’s96 findings on chimpanzees imply, according to Mackor,97 that chimpanzees have “social rules” because they follow learned, socially shared rules in their own behavior and in reacting to the behavior of others.
It is important, however, to go one step further and insist that a social rule can only be said to exist in a given group if it is a potential subject of reflection – of conscious thought about rules. If asked to explain their behavior, members must give the rule as a reason for it. And seeing the group’s rules as reasons for behavior implies that the rules could in principle be different: if there were not a rule against keeping one’s hat on in church, it would be OK to do so. A statement that simply warns of a danger (“No swimming: sharks!”), is only in the marginal sense of the way it is formulated a “rule”: it is not the rule but the sharks that are the reason for not swimming. Similarly, apes do not have a “rule” requiring the use of a stick to extract ants from an ant-hill. Such cautionary or how-to “rules” simply call attention to the natural constraints to which members of the group are subject. One does not really “follow” such a rule, one simply avoids unpleasantness or impossibility imposed, not by one’s group, but by the environment.
Furthermore, the applicability of even the simplest social rule to a concrete situation always requires interpretation of the rule and often the consideration of other rules as well (for example, those concerning justifications and excuses). Moral reasoning is necessarily involved both in following rules and in assessing behavior (one’s own or that of others) in terms of rules. The members of a group whose behavior is regulated by rules must thus be able to engage in moral argument about what the group’s rules are and what they require, and the observable fact that such argument takes place is part of the ontology of a rule.
Moral argument in turn implies that a group’s collection of rules consists not just of what Hart calls “primary rules” of behavior, but also of “secondary rules” – rules about the ascertainment, interpretation, application and change of primary rules – and it is this that makes social order through the governance of rules among humans so complicated.98
Crucial among Hart’s secondary rules is the “rule of recognition”, which specifies which rules are binding. In Hart’s view, it is the observable fact of acceptance of its secondary rules by a group’s officials that distinguishes “law” from non-legal social control. This seems wrong: a system of non-legal social control – without differentiated officials – will have secondary rules governing decision-making concerning its primary rules. For example, if there is a difference of opinion in a religious group over whether a rule requiring men not to wear hats in church applies in a given situation and what sort of informal sanctions are appropriate, participants in a discussion will appeal to much the same sort of considerations as do lawyers and judges: precedent, authoritative sources (e.g. texts, and the opinions of respected members of the social group), practical matters (e.g. how cold it is in the church concerned), whether the case is one for a differentiated “official” (priest or minister) to handle and so forth. In short, the validity, applicability and interpretation of the primary rules of both relatively undifferentiated social control and of a highly differentiated “legal” system is determined through reasoning in which appeal is made to a group’s secondary rules, and the validity of the latter is in both situations a matter of social fact: acceptance in practice. The difference between “informal” social control and “law” depends not on the existence of secondary rules, but on whether the group’s secondary rules recognize the authority of differentiated (“legal”) control.
Moral reasoning implies that those who engage in it can formulate rules in language and use language in the process of applying secondary rules to the interpretation and application of primary rules. Possession of a language that permits such moral reasoning therefore seems to me an essential condition for the existence of rules of behavior, and hence of social order, in a social group. I am unaware of evidence that chimpanzees engage in moral reasoning – that they can make, interpret, compare, apply and change their rules of behavior or contemplate the possibility of doing so. I am, therefore, inclined to deny that rules of behavior can be said to exist in chimpanzee social groups. The behavioral regularities observable among chimpanzees therefore, it seems to me, cannot be included within the scope of the sociology of law.
The foregoing paragraphs imply that in practice the existence of a rule in a group will rarely be a matter of all or nothing. In every group, some rules enjoy more general support and more effectively limit behavior than others; some have a more secure foundation in the group’s secondary rules than others. Many formerly effective religious rules – e.g. the rule among Roman Catholics prohibiting the use of contraceptives, or the rules among Muslims requiring specific clothing for women – may be losing support in (some of) their respective social groups. A rule that once was effective may, like an old soldier, not die but just fade away, so that the question whether R is a rule in a given group can be contested in the group itself and therefore pose problems for an external observer – problems that can only be dealt with by careful observation and precise formulation. New rules may similarly emerge in a more or less gradual way (e.g. social rules limiting the freedom to smoke). How widespread and vigorous the social support for a rule, and how effective a limitation of behavioral choice it entails, is a matter for careful observation, as we will see in the next section.
3.2.3. Epistemology
How can we know that R is a rule of behavior in a social group G in which we happen to be interested? Epistemology follows ontology: rule R exists in a social group if its formulation is recognizable to the group’s members,99 if the group’s members use R in social interactions, if R is accepted as a reason for behavior (rule-following, criticism, sanctioning, etc.), if the reactions authorized by R (and the relevant secondary rules) are accepted as appropriate by the group. In other words, R exists in G if behavior in G, in relevant respects, is observably different from what it would be if R did not exist. In short, the question whether R is or is not a rule in G, is the same sort of question as whether French is spoken in France, to be answered with the same sort of evidence. The proposition that R exists in G is falsified if R is in fact not generally accepted in G; if, when confronted by the proposition that R is a rule in their group, its members do not accept it as such and behavior in the group does not exhibit or reflect its use. The assertion that, in written French, all nouns must be capitalized (as in German) is easy to falsify by observing French practice. That some social rules may be contested is no different from the fact that there are always differences of opinion about peripheral rules of a language. In neither case is this a reason for doubting the real existence of a social group’s rules any more than of its language.
When social control is highly differentiated and secondary functionaries such as lawyers and judges are charged with identifying and interpreting a group’s rules, acceptance by these functionaries that R is a binding rule, and secondary behavior consonant with such acceptance, generally settles the matter. In the case of language, the secondary rules in some countries similarly make the decision of a differentiated authority definitive. Hence, in France, as far as the proper spelling of a French word is concerned, a decision by the Académie Française settles the matter; the secondary rules of English spelling are much less clear-cut and depend largely on the country one is in.100
The matter can, of course, be rather more complicated than I have suggested. The question whether R is a rule in G, and in particular, what its proper interpretation is, can be contested. Dworkin has analyzed with unequalled sophistication how, in a situation of differentiated social control, the generally accepted values, policies and principles of G (knowable in the same ways as are the rules of G) determine for lawyers and judges the “right answer” to such a question.101 When maintenance of a group’s rules is undifferentiated, “legal reasoning” about the group’s rules proceeds among ordinary people in essentially the same way, making use of the same sorts of arguments. In either case, fundamental and insoluble differences concerning a group’s important rules will (ultimately) entail the dissolution of the group, often a gradual process (as in the case of the drifting apart of descent groups within families or of language communities). When maintenance of rules is differentiated and a group’s functionaries cannot satisfactorily settle disagreements about its important rules, a dramatic schism may ensue, as is the case from time to time in religions, states and so forth. Whether the split be gradual or abrupt, the epistemological puzzle gets simplified, the original “group” having ceased to exist.
3.2.4. Methodology
In establishing the existence and specific contents of R in a group, individual-level variation among one’s informants is often almost irrelevant. If you want to know whether to drive on the right or the left in Indonesia, asking one Indonesian will generally suffice. Since knowing the applicable rules – whether written and differentiated (‘legal’) rules, or relatively undifferentiated social rules on subjects ranging from non-verbal interaction (such as staring at another person102), clothing,103 table manners,104 the division of labor in a household105and even “illegal” or immoral behavior such as “bribery”106 – can be of vital social importance, all normal members of society have learned how to acquire such knowledge.107
Knowing the rules of a group as an external, social scientific observer is not fundamentally different, although – especially when the group concerned is not one’s own – it does entail some methodological care. If you want to know what the rule on queuing is in, say Armenia, how would you go about finding out? The obvious first thing to do would be to ask someone with local knowledge, for example, someone standing at a bus stop. Or (for example, if you do not speak Armenian) you might simply observe behavior at a bus stop: is there a queue or just a random-looking mass of people? And what happens if someone who arrives later than others tries to get on the bus ahead of them? Do other people react negatively,108remonstrate, invoke a queuing rule, either explicitly or implicitly (e.g. “I was here before you”)?109 What reaction does this produce (e.g. does the queue-jumper retreat sheepishly to the end of the line, does he seek to justify his behavior in terms of other rules, such as an exception in cases of urgency?) Finally, if the social group you are interested in has a system of differentiated control, you will look to see what position its officials take toward the queuing rule (e.g. which side of the dispute gets their support).
You will not necessarily take all of this talk at face value, but on the other hand, people will not on the whole invoke rules or interpretations of rules for which there is no local support: doing so simply will not get them anywhere. Here, however, you must keep a crucial caveat in mind: you may have stumbled upon an example of regulatory pluralism, the queue-jumper (and his supporters) may belong to a different subgroup from those who remonstrate with him – a subgroup, for example, with an idiosyncratic rule that men go first in such situations. If this turns out to be the case, your initial conclusion that there is a first come-first served queuing rule in Armenia will have to be qualified to the effect that certain subgroups do not share it, at least where different sexes are involved. Having done all this, you should know with some assurance what the rule on queuing is in the social group concerned.110
But what about the normal methodological requirements for sociological research? If we really want to know what the rule is on queuing, don’t we need a proper survey of the population: how else can we know whether the population at large or – in the case of a differentiated social rule, lawyers and similar professionals – actually do “accept” the rule? The answer to this sort of question is: no. Rules – unlike attitudes, preferences, behavior and the like – are not individual characteristics but social facts. Most members of the relevant population will be familiar with the group’s most important rules. Just as mutually intelligible language would not be possible if there were not general agreement on the everyday rules, so social life in G would be anarchic (the “problem of social order” would be acute) if there were not the same sort of general agreement. Hence, when seeking to ascertain how one says “toilet” in Arcadian or how one asks for a woman’s hand in marriage among the Inuit, asking one “normal” member of the relevant population may often suffice, these not being matters of personal opinion.
Apart from a body of fairly simple and straightforward rules, about whose existence and contents there is rarely any serious problem (not for members of the group concerned nor for an external observer), most groups have some rules whose status or interpretation is more or less contested, or could be if an appropriate case presented itself. This fact of life raises difficult methodological problems, but these are for an external observer neither more nor less insuperable than they are for members of the group concerned. Dworkin’s analysis of “the necessary and sufficient conditions for the truth of a proposition of law”,111 gives an empirical researcher the help he needs on how to proceed in case of doubt. Stripped to the bare bones and applied to the question at hand, Dworkin’s argument would be that when a group’s rule on a particular matter is not clear to everyone in the group, the participants in a conflict invoke the group’s principles to support their respective claims. The existence and weight of a principle is established by referring to the characteristic practices, rules and institutions of the group. Dworkin does not argue that the outcome of such a procedure is always noncontroversial, but he does claim that in a functioning legal/moral system there will in fact be sufficient agreement within the community on what counts as good or bad argument that, although different people will reach different judgments, these differences will be neither so profound nor so frequent as to make the system unworkable.112 A social scientist would want to add, that where its normative order has proven unworkable, a group will generally be well on the way to extinction.
3.2.5. Regulatory pluralism113
What is more orderly, a jungle or a garden?114
Even if we know what a rule is and how to go about establishing the truth of the proposition that R is a rule of group G, we are not done with the difficulties that rules give rise to when we give them their (inevitable) place in sociological theory. The problem is that, unless we restrict our attention to entirely “pristine” social groups (groups whose members are not members of any other social group) – a notion that for practical purposes will almost always be beyond the vanishing point115 – we will always be confronted with “too many” rules. That is to say, every group’s members will also be members of a number of other groups, and the behavioral demands of these overlapping groups will often not be entirely consistent. The consequences of such regulatory pluralism for rule-following – and hence for sociology – are profound.116
The regulatory pluralism characteristic of social life at first sight looks very disorderly and it is in practice a major challenge for a serious researcher in the sociology of law. But before giving in to the urge to streamline social theory by factoring out social complexity, one should reflect on Flannery’s question (above). The idea that a single, exclusive system of social regulation is (or would be) more orderly than one of regulatory pluralism is comforting to legal positivists (among them, of course, lawyers, but also many social scientists), just as the idea that manicured gardens and “endless waves of grain” are more orderly than a jungle has its adherents. And the wish for a world of simple, uniform order seems often to be father to the thought that social life really is like that: that is, made possible by a single set of binding rules. But if by “order” we mean the sustainability of a self-regulating system, the jungle wins hands-down. Similarly, “orderly” social institutions are often not sustainable over time,117 whereas the resilience of pluralistic social order, over time and through catastrophic social disturbances (natural and man-made disasters, war and terrorist attacks), is sometimes astonishing.118
3.2.6. Conclusion
The foregoing remarks on the empirical status of social rules are rather general and rudimentary. Nevertheless, they seem to me sufficient for present purposes, that is, for making plausible the claim that rules exist and can be studied as observable social facts. The common notion that rules are things that inhabit a rarified “normative” sphere and that “existence” therefore cannot be predicated of them, has on closer analysis little to be said for it. Having got over that false obstacle, what remains are some truly difficult problems of operationalization for purposes of empirical research. These problems have hardly been discussed in the literature, even the best researchers seeming to make the opposite mistake of treating the rules they are interested in as empirically unproblematic.119