Restructuring Searle’s Making the Social World

1. Collective Intentionality

In the 1980s and 1990s Margaret Gilbert and Raimo Tuomela proposed analyses of social groups in terms of collective beliefs or intentions, that is, beliefs and intentions that have a collection of individuals or a “we” as their subject (Gilbert 1987, 1989, 1996; Tuomela 1984, 1995; Tuomela and Miller 1988). Searle proposed his own account of collective intentionality in 1990. He suggested that individuals can have beliefs and intentions with a “we” as their subject, and he maintained that the capacity for such collective intentionality is biologically primitive. Furthermore, he argued that that participating individuals derive their individual intentions from the collective intention. A central example that he considered concerns the collective intention to make a hollandaise sauce. This provides the basis for one individual to intend to stir and for another individual to pour. The collective action is performed by means of the two individuals acting on these individual intentions. Although the details have changed a bit, these core ideas are still there in Making the Social World.

One notable new feature of Searle’s analysis of collective intentionality is this: (i) Searle now requires that those who partake in a collective intention have beliefs about each other’s derived intentions. An agent who has a collective intention, for instance a participant in a football game, needs to believe that his team members will do their parts. Knowledge is not required. Furthermore, the belief does not have to cover the details of the intentions of the other participants. One of Searle’s examples concerns an offensive lineman who does not need to know the routes that the wide receivers follow when he blocks on a pass (55-56). Searle now explicitly argues that an intention cannot cover the action of someone other than the agent who has it. In the sauce example this amounts to the claim that “my intentionality cannot cover your pouring” (56). The upshot is that the other’s action is not covered by my intention but by the belief that the other will do her or his part.

This proposal faces a number of difficulties. Consider the joint action of playing a duet where one person plays the piano part and another the violin part. Each participant in the joint action has the collective intention “that we play the duet” (52). Critics objected that “it looks as if the agent must be having his intention cover the behavior of other people” (56). Searle takes himself to have answered these critics by introducing the belief requirement discussed above. This requirement concerns the individual intentions that people derive from collective intentions, and those intentions only pertain to the respective individual contributions to the joint action. I take it, however, that the critics were concerned with the content of the collective intention and the fact that it covers the behavior of both individuals. Searle does not address this problem.

I do not claim that the problem cannot be solved. Michael Bratman and Raimo Tuomela have put a lot of effort into explaining how one might make sense of that idea. Bratman distinguishes between intending to and intending that, and Tuomela distinguishes between action and aim intentions; both argue that a collective content is problematic only for the first of the two notions they distinguish, respectively (Bratman 1999, 142-61; Tuomela 2005). The point I want to make is rather that Searle does not offer an argument that will satisfy his critics.

A second problem surfaces when Searle specifies the attitude the individuals in his example have towards one another: “I have a collective intention-in-action that we play the duet by way of me playing the piano, in a context where I take it for granted that you are playing the violin” (52). It is not obvious that a belief that you intend to play the violin provides an appropriate basis for me taking it for granted that you will do this. Perhaps you change your mind. This suggests that a belief about the other person’s action is needed as well. It is in fact easy to come up with more objections along these lines. ³ In light of these objections, it seems plausible that some of the more complex analyses on offer are better than Searle’s.

Another major departure from his earlier work is that (ii) Searle now allows for different kinds of collective intentionality and that the collective intentionality that he has come to acknowledge now is reducible to I-intentionality plus mutual belief (58). Searle was opposed to such an analysis of collective intentionality, as is clear from his critique of the account presented by Raimo Tuomela and Kaarlo Miller (1988). ⁴ Now he embraces such a reductive conception as the kind of collective intentionality that provides the basis for institutions. His argument is that institutions do not require cooperation. Hence, collective acceptance or recognition will do (58). ⁵ In his earlier work Searle explicated collective acceptance in terms of the nonreductive kind of collective intentionality he focused on then. Such intentionality plays a role in institutional actions in which people do cooperate, for example when people exchange money in a transaction or when two people get married. However, he now maintains that it does (need?) not play a role with respect to the underlying institutions themselves. In extreme cases, institutions can be supported even by indifference and apathy (57).

Collective intentionality plays a very important role in structuring our relations with one another. When a collection of individuals has a collective belief or intention, they form a social group. Searle mentions ski clubs, nation-states, and corporations as examples. About corporations he claims that the point of bringing them into existence is “to create a rather elaborate set of power relationships between actual people” (98). ⁶ Searle also defines social facts in terms of collective intentionality: “A social fact . . . is any fact that contains a collective intentionality of two or more human or animal agents” (156). Note, however, that on Searle’s account the existence of a society need not be a social fact (156). The reason for this is that the members of a society need not share a collective intentional attitude. Although this makes sense, it is a somewhat striking conclusion for someone who coined the term “Philosophy of Society.” ⁷

2. Institutions

Searle does not regard society as an institution. The existence of a particular society need not be an institutional fact. Searle explicitly states that the state, religion and the church, education, science, and academia are not institutions. Instead, they contain institutions or they are institutional structures (92, 161). His argument is that, as such, these do not come with their own specific rights, obligations, permissions, prohibitions, entitlements, or any other normative or—to use Searle’s term—deontic powers. In his view, a state consists of a number of institutions including legislative, executive, and judiciary institutions, as well as the military and the police. Something similar holds for the church, academia, and, presumably, society. Their constituents are institutions because they do come with their own deontic powers. Think, for instance, of legislative or executive powers. ⁸

The notion of a status function plays a central role in Searle’s account of institutions as well. It is not terribly clear, however, what exactly a status function is. Searle maintains that all functions are relative to an observer and to a (set of) value(s) (58). Something is a screwdriver only if there is someone who might want to use it as such, which depends on the goals she sets herself and thereby on what she regards as valuable. An object has to have a fairly definite physical structure in order for someone to be able to use it as a screwdriver. One thing that distinguishes status functions from other functions is that they are only loosely connected to the physical structure of the entities that have them. The second distinguishing feature of status functions is that something only has a status function if it is collectively accepted or recognized as such. These two features do not add up to anything sensible as such. No positive characterization of status functions is offered. ⁹ As a consequence, it is not clear what work the notion of a function is doing.

In order to make progress regarding this issue, I suggest considering money, one of the few institutional entities that is commonly explicated in terms of its functions. Presumably the key function of money is to serve as a means of exchange. However, this function can easily be cashed out in terms of deontic power, more specifically, in terms of purchasing power. The claim that money functions as a means of exchange means the same as the claim that money is purchasing power. Why not forget about the function element in status functions, use the term “status” instead, and cash statuses out in terms of deontic powers only? ¹⁰ Searle now comes close to doing this. In his 1995 book, Searle claimed that many, but not all, status functions come with deontic powers. In Making the Social World he retains the term “status function” and the characterization presented above. Even though he does not identify the one with the other, he now claims that (iii) all status functions imply deontic powers. Honorific status functions such as the one involved in receiving an honorary degree no longer form an exception (24). ¹¹

3. Constitutive Rules

These claims about status functions and deontic powers can be used to shed some light on the distinction Searle makes between constitutive and regulative rules. Regulative rules, such as the rules of etiquette or traffic rules, regulate actions that can be performed independently of those rules. Constitutive rules, such as the rules of chess and the rules that define the U.S. presidency, constitute the actions to which they pertain. The logical form of constitutive rules is “X counts as Y in context C,” where X can (but need not) stand for a noninstitutional entity such as a piece of wood or a person and Y stands for a status function.

Searle used to believe that institutional facts necessarily exist within institutions and hence that all such facts presuppose constitutive rules. He has now abandoned that claim. Searle’s earlier remarks on these issues were rather confusing. ¹² His current, more sensible view has the striking implication that institutional facts do not require institutions (19-20, 23). People can, for instance, choose a leader on an ad hoc basis. The relevant person acquires the status at issue as well as its concomitant deontic powers even though there is no generally accepted rule for instituting a leader in the context at issue. A constitutive rule would be in place if leadership were determined by means of a systematic procedure, for example, by succession or by voting. Searle still maintains that institutions, and hence constitutive rules, facilitate the creation of institutional facts. In his words, a system of constitutive rules “automatically creates the possibility of institutional facts” (10). However, (iv) not all institutional facts presuppose constitutive rules.

Many people have been puzzled by the distinction between constitutive and regulative rules or have criticized it for reasons other than its fuzziness (Giddens 1984; Ransdell 1971; Ruben 1997; Warnock 1971). Searle now introduces a new characterization: (v) regulative rules are standing directives, whereas constitutive rules are standing declaratives (97). A (standing) directive is meant to influence behavior (in the relevant circumstances). A (standing) declaration is made in order to bring about a certain state of affairs or to change the world (when the appropriate conditions obtain). In response to a directive, someone is supposed to do something. A declaration does not call for any response except, as Searle notes, to accept its consequences (97). This is a perfectly legitimate distinction. It is not obvious, however, that this establishes that constitutive rules make a significant difference to the way the world is and affect it in a way that regulative rules do not.

We can make progress regarding this issue by considering how Searle has come to think of the relation between status functions and deontic powers. Constitutive rules specify the conditions X under which an entity has a certain status function Y in a particular context C. When a constitutive rule is collectively accepted, we can express this in terms of the locution “X counts as Y in C.” As indicated above, Searle now holds that all status functions imply deontic powers. He makes this claim more precise in the following way: when a status function Y exists in a context C, some subject S who bears the appropriate relation R to Y in C has the power to do A, where A represents “the acts determined by the Y status function” (103). All this implies that when a constitutive rule is in place and some entity in context C satisfies the conditions X some person(s) will have the deontic powers that come with the status function at issue.

Searle has explicated the structure of regulative rules as “If B, do A.” ¹³ This could be taken to suggest that all regulative rules are commands. However, traffic rules and the rules of etiquette also involve prohibitions and permissions, and there is no reason for systematically excluding other normative statements from the scope of regulative rules. Looking at regulative rules from this perspective suggests that they specify deontic powers that someone has in circumstances B. Furthermore regulative rules can be introduced by declaration (think of traffic rules). All this suggests that regulative rules are akin to constitutive rules in more respects than Searle recognizes. The main difference is that (only) when a constitutive rule is in place we have a special term for someone who satisfies the relevant conditions. The difference does not lie in deontic powers. Although it is not clear from the surface structure of constitutive rules X counts as Y, such rules also involve deontic powers. Whereas they feature explicitly in regulative rules, they figure implicitly in constitutive rules. It may well be, then, that Searle overestimates the difference between these two kinds of rules. The distinction between regulative and constitutive rules marks (primarily) a linguistic difference. ¹⁴

4. Language and Status Function Declarations

As mentioned, Searle characterizes constitutive rules as standing declarations. Consider institutional facts that do not depend on constitutive rules, such as cases in which someone is appointed leader in an ad hoc manner. Such a person can become a leader by some people declaring her to be their leader. This suggests that declarations can play a crucial role in a wide range of institutional facts. Searle in fact claims that declarations are important for all institutional facts: “The main theoretical innovation of this book . . . is . . . a very strong theoretical claim. All institutional facts, and therefore all status functions, are created by speech acts of a type that in 1975 I baptized as ‘Declarations’” (11). He introduces the term “Status Function Declaration” for these declarations and goes on to claim that “all of human institutional reality is created and maintained in existence by (representations that have the same logical form as) SF Declarations, including the cases that are not speech acts in the explicit form of Declarations” (13). In other words, (vi) all institutional facts involve Status Function Declarations. This claim can easily be taken to suggest that Searle grossly overestimates the role of speech in the creation of institutional reality.

How strong exactly is the claim Searle makes? At the heart of it lies the idea that institutions depend on symbols or linguistic representations for their existence. Even if we abstract from the somewhat unclear role that speech acts are supposed to play, there can be no doubt about the fact that Searle regards language as crucial for all institutional facts. This claim about the constitutive role of language is so strong, I suggest, that Searle faces a dilemma: either it is trivial, or it is implausible. ¹⁵

Institutions obviously involve beliefs about behavior. The notion of a linguistic representation might be so weak as to imply that having such a belief necessarily involves linguistic representations. In that case, the claim that language is constitutive of institutions is trivially true. If, on the other hand, a stronger notion of linguistic representation is at issue, the claim is implausible. To see whether Searle offers a way out of this dilemma, we need to determine what role language plays exactly in his social ontology. As Searle recognizes, for instance, we do not have special terms for all status functions. Searle presents a more sophisticated version of the claim in chapter 4 of his new book. His central claim is that “you cannot have institutional facts without language” (63). He makes clear, however, that he does not believe that all thought requires language. This would obviously trivialize the claim. He goes on to argue that language structures experiences in discrete segments and thereby facilitates experiences unavailable to prelinguistic thought. This does not appear to be unique to institutional vocabulary, so it does not provide the basis for defending a substantial version of the claim.

Searle then moves on to discuss the double direction of fit that declarations have on his view (68-69). The notion direction of fit concerns the relation between thoughts and words on the one hand and the world on the other. The mind-to-world direction of fit characterizes intentional states such as beliefs. Such states should fit the world in the sense that someone needs to adjust his belief if it turns out that it does not correspond to the facts. The world-to-mind direction of fit is characteristic of intentional states such as desires and intentions. The way to achieve a fit between these states and the world is by changing the world. The idea is simply that you need to act in order to satisfy your desires and to execute your intentions. The word-to-world and the world-to-word directions of fit apply to speech acts. They concern speech acts that come with a claim to truth on the one hand (e.g., assertions) and speech acts that purport to induce action on the other (e.g., commands). As mentioned above, declarations have the double direction of fit.

Consider promising as an example. When you make a promise, you are obligated to do as you promised. As this requires action, promises have the world-to-word direction of fit. However, by promising something you also create a fact, the fact of your promise and the obligation involved in it. This implies that promises also have the word-to-world direction of fit. The very making of a promise licenses the belief that a promise has been made and the person who made it is under the obligation to do as promised. More generally, declarations create facts by representing them as already existing. Searle maintains that declarations require language. In his view, this is true even when they are performed without any words. This is due to the fact that declarations have the double direction of fit. Searle adds that language facilitates “conventionally encoded commitments” (84). Status Function Declarations create a deontology (85). And a “publicly recognized deontology . . . requires language” and “the concept of an obligation” (95).

It would be interesting to see whether these statements provide a way out of the dilemma I sketched above. I am not yet convinced that they do so. Consider Searle’s example of a wall that decays and turns from a physical structure into an institutional boundary (94-96). It is not obvious that any linguistic communication is required in the process. People can observe each other’s behavior, including sanctioning behavior such as frowning when someone crosses the boundary. At some point it is true that the stones that are left form a boundary, and this fact involves the obligation not to cross it. The stones form a boundary because the relevant people recognize it as such. These people believe that it is a boundary and recognize the deontic powers that come with being a boundary. In light of this, it seems fair to say that the collective intentional states that are involved in the constitution of the boundary have the double direction of fit. Language needs not enter, neither to account for the double direction of fit nor to explain the normative nature of this institutional fact. I am not sure what Searle has in mind when he mentions conventionally encoded commitments, but I see no reason to believe that I have left anything out of the picture that essentially involves linguistic conventions. So it seems that Searle overestimates the role language plays in institutions when he claims language is constitutive of institutions.

The notion of a Status Function Declaration and the central role it plays in Searle’s social ontology can easily be taken to suggest that all institutional facts involve speech acts. Now Searle considers virtually all acts involving some form of symbolism as speech acts. Even in this broad reading, however, the claim that all institutional facts involve Status Function Declarations is striking. One commitment that is involved in this claim is that a particular action performed at a particular time is required for an institutional fact to come into existence. Searle’s own example of the decaying wall seems to form a counterexample.

However, it is not very clear what exactly Searle’s position on this is. Consider the following passage:

There need not be any specific moment at which there is a speech act of Declaring, but there must be some speech act or set of speech acts and other sorts of representations that constitute representing the line of stones as a boundary in a way that makes it into a boundary. When the representations are collectively recognized or accepted, the line of stones acquires a new status: it now is a boundary. (96)

At the beginning of this passage, Searle seems to disavow the idea that time-specific speech acts need be involved. But then he goes on to commit himself to the idea that speech acts have to be involved after all. The phrase “and other sorts of representations” makes me wonder why these other representations are not sufficient. Why not say that the line becomes a boundary when nonlinguistic representations to that effect are collectively recognized? ¹⁶ The upshot is that Searle’s notion of a Status Function Declaration raises more questions than it answers. As it is confusing rather than enlightening, Searle would be better off without it. Given that it features in the main thesis of his latest book, it follows that what Searle claims to be his main theoretical innovation does not advance his social ontology significantly beyond his earlier work.

5. Political Power and Human Rights

Searle explicates political power in terms of deontic power (169). In his view (vii) all political power is deontic power. ¹⁷ Political power, he argues, “requires some degree of acceptance” (163). One might want to object that political power involves the use of physical force. In practice, however, it often requires only the threat of physical force. What is more, as Searle points out, the physical force that police and military might use also depends on deontic powers. The point, I take it, is that the police and the military are institutions that involve offices that come with certain deontic powers that depend on collective recognition. Exercising political power often is a matter of getting people to want what the person in power wants them to do, that is, of influencing their reasons for action.

Searle also claims that the exercise of power “is always an intentional act” (148). This is not very plausible. Imagine a president signing a document that she has not carefully read. It seems that the deontic consequences follow irrespective of the fact that the act was unintentional with respect to them. It is also unclear how the claim that power is always exercised intentionally fits with Searle’s assertion that one can exercise power unconsciously (158).

Searle tries to make sense of the claim that society can exercise power over its members (and that social pressure can be a form of power). For this purpose he introduces the notion of Background power, deontic power that does not require a special status:

The basic concept of Background power is that there is a set of Background presuppositions, attitudes, dispositions, capacities, and practices of any community that set normative constraints on the members of that community in such a way that violations of those constraints are subject to the negative imposition of sanctions by any member of the community. (160)

Searle insists that sensible power talk can always be made more precise by explicating who has power over whom to do what. In effect he argues that talk of a society having power over its members can be explicated in terms of each member being in a position to sanction other members for a violation of a societal constraint.

Unsurprisingly, Searle also analyzes the notion of a human right in terms of deontic powers. He argues that a human right is an institutional status that we have imposed on all human beings. (viii) Human rights, then, are status functions. This implies that, as such, a human right is a social right rather than a moral right. We could refrain from assigning such rights, and if we did, people would not have them. However, they can be justified in terms of moral values. In spite of the fact that he provides an institutional account of human rights, Searle tries to make sense of the idea that human rights remain in existence in some sense even when they are no longer recognized. This is surprising because the institutional account entails that human rights exist due to the fact that they are collectively recognized. Searle suggest that the claim that unrecognized rights still exist means that the relevant agent still is a human being, which implies that she meets the (only) condition that has to be met in order to have such a right (182). It is not obvious, however, that the fact that this condition is met implies that the right exists. Given the institutional account he provides of them, Searle’s existence claim about human rights in the absence of recognition is problematic.

More needs to be said in order to get a good idea of what exactly Searle might mean. The claim that the human rights of some people are not recognized can mean a number of things. There might be disagreement about the rights or the status function Y that is involved. Think of those who contest the Universal Declaration of Human Rights on the grounds that it favors Western values over Eastern ones. Alternatively, people might have different ideas about the condition X that has to be met in order for someone to have human rights. Searle favors assigning the status to all members of our species, where species is defined in biological terms; others might want to ascribe such rights only to men, property owners, or people above a certain level of intelligence. A third way in which some people might fail to properly recognize human rights concerns their practices. Imagine that all agree about who possesses which rights. In such a situation, some might still go ahead and violate those rights, the deontic powers that persons S1, . . . , n have. They might, for instance, torture prisoners of war. In this third case, the existence of the rights is not challenged. This suggests that Searle is concerned with one of the first two cases. Unfortunately, it is not clear which of these two is Searle’s target. The good news, however, is that Searle’s framework facilitates a discussion of all three. The bad news is that it is unlikely that Searle’s existence claim about human rights can be sustained on any of them.

As is evident from his account of institution generally and his accounts of political power and human rights in particular, the notion of deontic power lies at the heart of Searle’s social ontology. It is not obvious, however, that Searle’s conceptualization of the source of deontic power is adequate. He maintains that people have deontic powers because people collectively recognize status functions together with the deontic powers that come with them. Recall that, in Searle’s view, the collective intentionality involved in institutions and institutional facts is of the kind that is reducible to individual attitudes. Margaret Gilbert and Raimo Tuomela, however, have forcefully argued that social rights and obligations require a stronger kind of collective intentionality. Only collective attitudes that involve a joint or collective commitment, they argue, issue in social rights and obligations (presumably their arguments generalize to other deontic powers). If their arguments are adequate, collective recognition of the Searlean kind does not provide an adequate basis for deontic powers. You just do not get deontic powers without collective commitments. ¹⁸

This line of reasoning suggests that Searle should recognize a stronger kind of collective acceptance or recognition than he has done so far. How does this relate to the point discussed in section 1, that Searle now allows for collective intentional states that can be reduced to individual intentional states and that such states form the basis of institutions? My claim that he needs to recognize a stronger kind of collective acceptance in order to properly account for institutions is in fact consistent with his claim that a weaker conception is required. Recall that Searle distinguishes between the attitudes people have toward institutions as such and those that are involved in institutional actions. He maintains that only the latter are a matter of cooperation. This implies that only they need to involve the stronger kind of collective acceptance that involves collective commitment.

It is unlikely that things divide up so neatly. Tuomela (2011, n10) notes in this context that withdrawing money from an ATM does not involve cooperation in the same sense as getting married does. Perhaps the idea that all institutional actions require commitment-involving collective acceptance is rather farfetched. At the same time, however, it is rather plausible that institutional phenomena are in fact based on a range of attitudes, some stronger than others. So I welcome Searle’s weakening of his collective acceptance requirement for institutions. I just add two qualifications. First, it should be combined with a stronger conception of collective acceptance. Only commitment-involving collective acceptance can adequately ground deontic powers. Second, the distribution of the stronger and the weaker forms of collective acceptance in relation to actual institutions may be more an empirical matter than a matter that can be decided on the basis of reflection.

6. Conclusion

The eight claims (i-viii) that Searle makes in Making the Social World make reading his new book very worthwhile. They reveal that the book clearly is an improved version of The Construction of Social Reality. Making the Social World also extends this earlier book by including discussions of political power and human rights, which are also noteworthy. The fact that Searle now allows for ad hoc institutional facts such as improvised leaders that do not presuppose constitutive rules is a definite improvement. The development of his account of collective acceptance is also more than welcome. How different forms of collective acceptance are involved in institutions is a difficult issue, but that more than one form is needed to adequately account for the existence of institutions can hardly be doubted.

My criticisms suggest, however, that there is substantial room for further improvement. I have argued that the notion of a Status Function Declaration raises more questions than it answers and that Searle would be better off without it (section 4). The term “function” in “status function” does not serve a clear and useful purpose either (section 2). Furthermore, Searle overestimates the role that language plays in institutions when he claims that language is constitutive of institutions (section 4). And he fails to appreciate that the distinction between constitutive and regulative rules is primarily a linguistic one (section 3). In addition to this, Searle’s account of collective acceptance is too weak to account for the normative nature of institutions that takes center stage in his own account of them. Deontic powers require collective commitment. Furthermore, actual institutions are likely to involve both weak and strong collective attitudes (section 5). Finally, his account of human rights as status functions is flawed because it cannot adequately account for the existence of human rights that are not recognized (section 5).