Ten years ago, Paul Amselek (1988: 187) said that "philosophy of law should feel particularly challenged by the theory of speech acts". If anything, this is more true today, after the publication of John Searle’s recent The Construction of Social Reality. Most of Searle’s examples of institutional facts are legal concepts. As Amselek said, "[in the legal world] philosophers of language felt that they had found a domain particularly suited to illustrate and test their ideas" .
If the law is a ‘domain particularly suited to illustrate’ those ideas, then the application of those ideas to the law must be a fruitful enterprise. This is what Neil MacCormick and Ota Weinberger showed in their joint work An Institutional Theory of Law (1986).
In his inaugural lecture, MacCormick argued that "if the law exists at all, it exists not on the level of brute creation (...) but rather (...) on the plane of institutional facts". What makes propositions of law true or false, he tells us, is not "merely the occurrence of acts or events in the world, but also the application of rules to such acts or events" (MacCormick, 1974: 51). Contracts, for example, are legal institutions. But legal institutions are not identical with rules, since the institution of a contract is one thing, the contract I have with Edinburgh University another. MacCormick’s claim is that institutions are ‘concepts’, concepts that are regulated by rules in the sense that instantiations of them can be brought about, have consequences and be terminated according to those rules:
The term ‘institution of law’, as I shall use it, is therefore to be understood as signifying those legal concepts which are regulated by sets of institutive, consequential and terminative rules, with the effect that instances of them are properly said to exist over a period of time, from the occurrence of an institutive act or event until the occurrence of a terminative act or event (MacCormick, 1974: 53).
According to MacCormick, institutive rules are those that "lay down that on the occurrence of a certain (perhaps complex) act or event a specific instance of the institution in question comes into existence" (1974: 52); consequential are those rules that provide for the consequences the existence of an instance of a given institution has. Lastly, rules are terminative when they provide for the termination of the particular instance of the institution under consideration (53).
Contrary to what we may think from the title of his article, MacCormick claims that it does not follow from the fact that legal concepts (or at least some of them) are ‘institutions’ (and hence that the existence, effects and termination of instances of them are determined according to rules) that the law itself is an institution: "there is an almost overwhelming temptation (...) to treat the concept ‘law’ like the concept ‘contract’ as denoting an institution which is defined and regulated by the relevant set of institutive consequential and regulative rules" (MacCormick, 1974: 57). This temptation must be resisted, for
But MacCormick is giving up too much ground here. Because (if) the institutive, terminative and consequential rules relating to the common law as a source of law do not render sufficient and necessary criteria of validity, then the common law cannot be an institution (as MacCormick says in 1974: 57, just quoted ). But if that is the case, then all legal concepts MacCormick is willing to call ‘institutions’ and whose institutive, consequential and terminative rules are (at least partly) to be found in rules of the common law cannot be institutions because of the very same reason, i.e. because those rules would not completely regulate the (creation, consequences and termination of instances of the) concept. A contract, for example, can exist even if the institutive rules have not been followed (MacCormick gives an example of this: 1974: 68), and it can fail to exist even if the institutive rules have been followed (MacCormick and Weinberger give also an example of this: 1986: 12) Thus it seems that either the law is an institution along with the others, or none of them is.
Later in his lecture, MacCormick returns to this subject. He accepts (as did Hart, 1948) that institutive, consequential and terminative rules are defeasible, with the consequence that they cannot specify necessary and sufficient conditions for the existence of an instance of an institution of law:
MacCormick’s answer is not straightforward: "we neither have criteria of validity for legal principles, nor therefore a distinction between valid and invalid principles of law" (MacCormick, 1974: 73). Though it is possible to give an account of what makes true the statement ‘the principle ‘no one may profit from his or her own wrong’ is a principle of English law’, how those conditions actually work is something that cannot be understood without considering the values and purposes of the law.
But exactly the same can be said of contract, testaments and other legal concepts. MacCormick himself acknowledges that "‘law’, even in its purely normative significance, spills over the edges of the valid rules. Rules as institutional facts are not the whole of the law, though they are singularly important of it" (MacCormick, 1974: 74). This is true not only concerning the common law, but also regarding the institutive, consequential and terminative rules that regulate all other legal concepts.
I believe that the solution for this problem is to accept the institutional nature of the law even if there are no rules stating sufficient and necessary conditions of validity for rules or principles of the common law.
This would require that the notion of institutional fact be adjusted to cater for the fact that the law ‘spills over the edges of valid rules’. MacCormick believes that a theory of law cannot go that far. It is at this point, he argues, where "the philosopher may still pose questions, but he will have either to become a sociologist to answer some of them, or alternatively, have to wait for his sociological colleagues to give him the answers" (1974: 74).
In the next section, I would like to offer an explanation for the defeasibility of legal rules based on the kind of institution the law happens to be. An adequate explanation of the law’s nature will help us to see that legal rules (as opposed to, say, rules of games) are defeasible not because they are rules, but because they are legal.
Searle on Constitutive And Regulative Rules
In this paper I want to comment upon Professor Searle’s famous distinction between constitutive and regulative rules. His recent articulation of that distinction in a more ambitious ‘general theory of institutional facts’ may be used as an initial justification for my going through this familiar subject.
I will try to show that Searle’s distinction has to be seen as a distinction not between rules but institutions (or systems of rules) of different kinds.
This is particularly important, I shall argue, when Searle’s distinction is applied to institutions like the law. It explains what is special about the law when it is compared to, say, games that makes disagreement so pervasive in the former and so rare in the latter. Here, however, I will have space only to point out the problems that a ‘general theory of institutional facts’ faces when this is not recognised.
In an often-quoted passage, Searle introduces his now famous distinction:
I do not know whether Searle would agree with Giddens in that all rules are regulative, though at times he seems to hint at an affirmative answer (he speaks of ‘purely regulative’ rules, and says that ‘constitutive rules do not merely regulate’). If this is the case, then all rules can not also be constitutive, since in that case there would be no distinction whatsoever. Are, then, all rules constitutive? Searle answers:
Tony Honoré has claimed that a satisfactory theory of individuation of laws must allow for the following kinds of laws:
2. Rules of inference provide how facts may or must or should preferably be proved and what inferences may or must or should preferably be drawn from evidence.
3. Categorizing laws explain how to translate actions, events, and other facts into the appropriate categories.
4. Rules of scope fix the scope of other rules.
5. Position-specifying rules set out the legal position of persons and things in terms of rights, liabilities, status, and the like.
6. Directly normative rules (which are a few in number but important) guide the conduct of the citizen as such (Honoré, 1977: 118).
This leaves out only 6, ‘directly normative rules’. Are they not purely regulative? They certainly do not constitute what they regulate: if they did, they would fall into another category. The problem is, these rules cannot be purely regulative since they are expressed in institutional terms. These rules are almost tautological (Searle seems to believe that they are, in fact, tautologies: 1969, at 191).
A rule like ‘perform contracts’ is not purely regulative, since rules are required for some descriptions of the an action in accordance with it. Nor is it constitutive, since it is not required for such a description.
If this is correct, we can see a more interesting problem concerning the two criteria Searle has offered to distinguish regulative from constitutive rules. That problem is that those criteria do not necessarily overlap, because they answer different questions. As Geoffrey Warnock said:
The second criterion was
Searle’s new general theory of institutional facts is still liable to this problem. This time the distinction makes its appearance in the book (1995: 27) with the help of the following pair of examples: ‘drive on the right-hand side of the road’ (regulative) and ‘the rules of chess’ (constitutive). Here we can see Searle using the first criterion. The ‘drive on the right’ rule is said to be regulative because it regulates driving and driving is an antecedently existing form of behaviour, while rules of chess are constitutive because they ‘create the very possibility’ of playing chess .
When we try to apply Searle’s second criterion to the first rule, however, we run into troubles. Imagine that it is section (1) of the highway code:
(1) Drive on the right hand side of the road.
Imagine further that section (2) is
(2) failure to comply with section (1) of this statute shall constitute a misdemeanour, punishable by fine of $50.
In this case (1) would be constitutive since "behaviour which is in accordance with the rule can receive specifications or descriptions which it could not receive if the rule or rules did not exist" (Searle, 1969: 35).
One way out of this problem is to say that the literal formulation of the particular rule under consideration is decisive. In this case, (1) would be purely regulative because it does not mention the word ‘misdemeanour’. (1’), however, would be constitutive:
(1’) Not to drive on the right hand side of the road counts as a misdemeanour.
This, however, would demote the distinction between regulative and constitutive rules to the category of a stylistic distinction, and I do not think Searle would agree with that. In brief, unless Searle wants to say that the distinction is a stylistic distinction, it is not enough to know the content of a particular rule to know whether it is regulative or constitutive.
And we can see how the problem reappears as the ‘general theory of institutional facts’ is developed. I submit that Searle developed that theory without having in mind the crucial distinction between systems of rules suggested by the first criterion: the distinction between criminal law (regulative of independently existing forms of behaviour) and chess (constitutive of chess). Some of his remarks, accordingly, fit institutions like the law, but fail to fit, as we shall see shortly, systems of rules whose point is to create new possibilities of behaviour rather than to regulate some antecedently existing form of it (e.g. games). For reasons of space, I will focus upon two moments of the theory, in which the problem is, I believe, particularly noticeable.
Systematic relationships between institutional facts
One feature of institutional facts, according to Searle, is that they "cannot exist in isolation but only in a set of systematic relationships to other facts" (1995: 35). For money to exist a system of exchange has to exist beforehand, and for a system of exchange there has to be a system of property and property ownership. "Similarly, in order that society should have marriages, they must have some form of contractual relationships. But in order that they can have contractual relationships, they must have such things as promises and obligations".
Generally speaking, the existence of institutions that are as a whole regulative presupposes the existence of the practice that the institution is there to regulate. The reason for this is rather obvious: the point of the institution is the regulation of the practice. But institutions have these relationships because of their regulative character, not because they are institutions. Concerning games, for example, the point is less straightforward.
Searle, however, thinks that games are not counterexamples to his claim, though "it might seem" that they are, "because, of course, games are designed to be forms of activity that do not connect with the rest of our lives in a way that institutional facts characteristically do" (1995: 36). When this point is looked at carefully, Searle claims,
Institutions and their Consequences
So let us go back to the quotation in page 36, where Searle claims that "It might seem that games are counterexamples to this general principle, because, of course, games are designated to be forms of activity that do not connect with the rest of our lives in a way that institutional facts characteristically do" (Searle, 1995: 36).
The sentence that immediately follows seems to imply that the way in which institutional facts characteristically connect to our lives is that the former have consequences for the latter: "Today’s philosophy department softball game need have no consequences for tomorrow, in a way that today’s wars (...etc) are intended precisely to have consequences for tomorrow".
Searle thinks that it is a peculiarity of games that they do not have consequences that are characteristic of institutions. I believe that that is precisely the mark of constitutive institutions, shared not only by games but also by, for example, geometry or mathematics and the like. And precisely because they do not have ‘consequences for tomorrow’ they need have no systematic relationships with other facts.
I do not want to object to the thesis that some institutions stand in systematic relationships with other institutional and non-institutional facts. In fact, this is an extraordinarily important and usually disregarded feature of institutions like the law. But this is not a characteristic of institutions as such, but only of institutions that are, according to my version of Searle’s first criterion, regulative, because they are regulative.
This might seem beside the point, since Searle explicitly accepts in the statement quoted above that institutions need not have these consequences. What Searle misses is that institutions have ‘systematic relationships to other facts’ because they have ‘broader consequences’: games do not have consequences, hence they need not have those relationships.
The Evolution of Institutions
The second problem in the general theory is related to the issue of the evolution of institutions. Can institutions evolve without the participants’ being aware that they are evolving one?
Searle’s answer is: indeed they can. Consider the example of money: people can go around buying and selling and exchanging, without their thinking that the particular goods they use as a medium of exchange is ‘money’:
The evolution may be such that the participants think, e.g. ‘I can exchange this for gold’, ‘this is valuable’, or even simply ‘this is money’. They need not think ‘we are collectively imposing a value on something that we do not regard as valuable because of its purely physical features’, even though that is exactly what they are doing (...). In the course of consciously buying, selling, exchanging, etc., they may simply evolve institutional facts (Searle, 1995: 47).
Patrick Atiyah (1981) has made a similar point:.
But it is difficult to see how this can happen concerning constitutive institutions. When I was a boy we used to play football in a park. As there were, of course, no goal posts in the park, we had to use our jumpers and bags as goalposts. The first time my friends did this I could not understand what were they up to, until one of them said: ‘this is your goal, and that is ours’. We could not have played football in the park had we not been aware of the fact that by placing those bags and jumpers where we placed them we where collectively assigning meaning to them, a meaning that was not exhausted by the physical properties of the bags and jumpers. In other words, there are no implied goals or goal posts on the back of which football can grow up.
Constitutive and Regulative Institutions
I think that the first of the two criteria offered by Searle to characterise institutions is the important one. According to it, the law is a regulative institution, since its point is to regulate antecedently existing forms of behaviour . Games, on the other hand, are constitutive institutions, that is, systems of rules whose point is to create new possibilities of behaviour rather than to regulate antecedently existing forms of it.
The reason why I call both of them ‘institutions’ is because both of them characteristically allow the creation of institutional facts, like offences and goals: thus, it turns out that the first of Searle’s criteria distinguishes types of institutions, while the second can be used to establish the existence of institutional facts.
If this is correct, a crucial conclusion follows (though the full argument for it has not been presented here): the defeasibility of legal rules, probably the single most important phenomenon around which most of the literature on legal reasoning has been produced, is to be explained by the systemic features of the institution of law. Institutional reasoning is not necessarily defeasible: against authors like H. L. A. Hart, the above argument would lead us to realise that legal rules are defeasible not because they are rules, but because they are legal. One could, for example, use the reading of Searle presented above to explain why we develop what Dworkin (1986) calls ‘an interpretative attitude’ concerning some of our practices and some of our practices only.
—Atiyah, PS (1981): Promises, Morals and Law. Oxford: Clarendon Press.
—Dworkin, R (1967): "Is Law a System of Rules", reprinted in Dworkin (1977).
—Dworkin, R (1977): Taking Rights Seriously. Cambridge, Mass: Harvard University Press.
—Dworkin, R (1986): Law’s Empire. London: Fontana.
—Flew, A, Ed. (1951): Logic and Language. Oxford: Basil Blackwell.
—Giddens, A (1984): The Constitution of Society. Cambridge: Polity Press.
—Hart, HLA (1948): "The Ascription of Responsibility and Rights", in Flew (1951).
—Hart, HLA (1994): The Concept of Law. Oxford: Clarendon Press.
—Honoré, AM (1977): "Real Laws", in J Raz and PMS Hacker Raz and Hacker (1977).
—MacCormick, N (1974): "Law as Institutional Fact", in MacCormick and Weinberger (1986).
—MacCormick, N and O Weinberger (1986): "Introduction", in MacCormick and Weinberger (1986).
—MacCormick, N and O Weinberger (1986): An Institutional Theory of Law. Dordrecht: D. Reidel.
—Raz, J and PMS Hacker, Eds. (1977): Law, Morality and Society. Essays in Honour of H.L.A. Hart. Oxford: Clarendon Press.
—Searle, J (1969): Speech Acts. Cambridge: Cambridge University Press.
—Searle, J (1995): The Construction of Social Reality. London: Allen Lane; The Penguin Press.
—Warnock, GJ (1971): The Object of Morality. London: Methuen & Co.