Institutional Externalism

1. Does Social Ontology Rest on a Mistake?

Social entities are “things” like promises, contracts, presidents, associations, money, debts, corporations, and financial crises. Although there may be social kinds that have a natural basis (candidates are sex, gender, race, sexual orientation ¹ ), most social kinds seem to be essentially dependent on human activities and thoughts. However, the way even the most institutional and conventional aspects of social reality depend on human activities is no trivial matter. Institutional entities are neither flimsy (as individual thoughts and ways of behaving typically are), nor are they always “easier” to modify than physical ones. It may take very little effort to smash a stone into small pieces, but it does not seem to be within the power of any single individual to alter the value of a 5 euro note. A widespread take on social ontology has it that the subjective and yet resilient status of social and institutional reality is given by its dependence not on individual mental states but rather on collective intentions with shared contents. Even though there is no consensus on how collective intentionality should be analyzed, there seems to be substantial agreement on the fact that collective intentions play a central role in explaining why social objects are, on one hand, dependent on people, and on the other hand, largely independent from individual will and whim. ²

In what follows, I will focus on Searle’s version of collective intentionality, and criticize one of its central points. That will allow me to introduce a radically different picture of social reality and the role of intentionality in its constitution. In Searle’s (1995) well-known account of money, the collective intention of attributing a function to certain pieces of paper—the function of enabling their bearer to exchange them in return for commodities up to a certain established value—is what makes those pieces of paper currency. In this picture, the whole of social reality rests on collective intentions of this sort. In particular, the shared content of the intention determines the nature of the social object to which it gives rise—roughly, what Searle calls the “status function,” that is, the function that an object has in virtue of being collectively attributed a certain status.

I think the explanatory model based on collective intentions with a shared content is flawed. ³ It works (to a certain extent) for simple and informal cases. However, it cannot be applied to the complex variety of social entities that we find in modern societies, in particular, to the ones we find in institutional contexts. In the alternative approach that I propose here, contents of shared collective intentions do not determine the fundamental features of institutional entities; collective intentions play a more “neutral” role with respect to the features of institutional reality. Roughly, institutional contexts are characterized by a shared intention to defer to elements of the context with regard to what determines the nature of institutional reality. Such elements are laws, contracts, and the like (what I will call “enactments”), which are in principle publicly accessible, and are regarded as valid and binding because they are produced according to an established procedure; thus, not in virtue of their content. In what follows, I start with section 2 highlighting Searle’s implicit assumption of continuity (AC), which I take to embody a widespread stance in contemporary social ontology. I then present the principles of my alternative view in section 3. In the following two sections (4 and 5), I discuss two problems that follow from the continuity assumption, and then in sections 6 and 7, I argue that institutional externalism resolves them and has further merits. In section 8, I draw my conclusion.

2. The Assumption of Continuity

John Searle’s theory of status function is one of the most developed forms of social ontology on the market, and one that shares many widely accepted assumptions with other theories. According to Searle, social entities depend on people in the sense that they have properties that they would not have if people did not exist. To pick up on the example of money: what makes it the case that a piece of paper can function as a 5 euro note? It is certainly not its microphysical structure or some other intrinsic property. It is the fact that there is a community of people that recognize it as having precisely that function (although there may be physical constraints on the social function: a 12-ton coin would be very impractical). Ontologically speaking, then, social objects depend on people for both their existence and their features. If there were no people, there would be no society, and if there were not certain behaviors and thoughts (performed or entertained by people), a society with such and such features would not exist. According to this theory, we find this sort of dependence all the way down and all the way up in social reality:

Collective mental phenomena of the sort we get in organized societies are themselves dependent on and derived from the mental phenomena of individuals. This same pattern of dependence continues higher as we see that social institutions such as governments and corporations are dependent on and derived from the mental phenomena and behaviour of individual human beings. (Searle 2010, 4)

In a sense, this continuity of dependence on individual intentions from simple collective behavior to complex institutions is obvious. What would be the alternative? Maybe some caricature of the “Hegelian spirit,” which is embodied in the higher tiers of social reality and whose objectivity is utterly independent from what the individuals composing the society think and do? It seems more in line with a naturalist methodology (even a very moderate or “liberal” one) to accept that there is no substantial difference between the informal collaboration among people in small groups and the kind found in much more complex situations, such as those involving governments issuing laws, contracts, and institutions. However, this trivial reading of Searle’s passage is not what is meant by the idea that collective intentionality is the ground of social reality.

Searle’s approach requires not only (a) that all intentionality should be in individual minds or brains, and, thus, that also collective intentionality is at bottom a psychological phenomenon that concerns the individual (independently from the issue of whether its form is irreducibly plural or not), but also (b) that collective intentionality is “the fundamental building block of all human social ontology” (Searle 2010, 43, emphasis added). That is to say, it is in virtue of collective intentions and their content that there is a domain of social objects that possesses characteristic features and has, in some sense, objective existence. I will label these two basic requirements the thesis of individualism (TI) and the AC.

Many philosophers agree with the gist of Searle’s position, at least with respect to AC. ⁴ Others, such as Ruben 1989, Thomasson (2002, 2003, 2009), Khalidi (2013) and Guala (2010), disagree on whether the intentions involved in the construction of a social kind K must always involve the concept of K, and argue that social reality may show much more “opacity” than Searle’s account entails. Here, I intend to criticize the grounding role of collective intentionality for institutional reality, which is elicited by AC. I will remain neutral with respect to TI. I start by presenting my alternative proposal, which I term institutional externalism. I then discuss the flaws of AC and argue that institutional externalism can overcome them.

3. Institutional Externalism

I take AC to fail because in institutional contexts, the role of collective intentions is not that of determining the nature of institutional phenomena. Insofar as institutional contexts are part of complex social situation, thus, AC fails. Examples of social facts that require institutional contexts are the fact that a certain person is the chief executive officer (CEO) of a certain company, that a certain action is a tax crime, that a certain corporation exists, or that I have acquired certain duties by signing a contract. In other terms, facts about the existence of CEOs, tax crimes, corporations, marriages, and the like are not grounded on the shared content of collective intentions. The nature of institutional entities depends rather on enactments that are collectively regarded as valid. I will use the term enactment to cover things such as laws, declarations, and contracts, whose contents are usually written in documents, and are always external to the content of collective intention, in the sense of being elements of the context that are in principle publicly accessible. Institutional externalism can be summed up in two theses:

Consider two examples: the case of a tax crime, and that of a marriage. Suppose that an individual S carries out at a time t a certain financial operation that in an institutional context C is a tax crime. According to Institutional Realism, the fact that there is a tax crime (committed by S at t) is grounded on facts about what S did at t (the financial operation that she carried out), and facts about the tax law (an enactment) in context C. According to Content Deference, in C, the enactment that establishes that S has committed a tax crime (i.e., a certain tax law) has been produced according to an established procedure that is collectively regarded as producing valid enactments. That is to say, the content of the enactment (which says that whoever carries out certain financial operations commits a tax crime) is collectively regarded as binding not because it is the content of a shared intention, but because people in the institutional context defer to what enactments produced according to established procedures declare to be legally binding. The case of marriage is slightly more complex, because it (usually) requires two enactments: a marital law and a marriage certificate. Suppose individuals S and S’ do what it takes to get married in context C: let’s say, undergoing a marriage ceremony and signing a marriage certificate. According to Institutional Realism, the fact that they are married is grounded on facts about what they did in the ceremony, and facts about both the marital law and the marriage certificate they signed in context C (viz. facts about two enactments). According to Content Deference, in C, the marital law has been produced according to an established procedure that is collectively regarded as producing valid enactments. The binding content requires there to be a further enactment (the marriage certificate), which is produced according to an established procedure. The content of the certificate, usually in the terms specified by the law, is thus also collectively regarded as binding. Therefore, an enactment can depend on others, and the detail of how the content that is regarded as binding is specified may involve several different enactments (often contracts and declarations refer to laws instituting the possibility of issuing them). In the following paragraphs, I will spell out Content Deference in more detail, and argue that it is the correct way to understand the role of collective intentionality in institutional contexts. ⁷

4. Against the Assumption of Continuity: The Explanatory Gap

In his 2010 book, Searle claims that collective intentionality manifests itself in two ways. The first, cooperative behavior, involves action; the second, collective recognition, involves belief. Cooperative behavior does not require that a complete intentional content be shared: the only part of the intentional content that has to be shared for cooperation to occur is the goal. If we cooperate to play a duet, I may play the violin while you play the piano: our intentional contents will be different, but we will share the same goal (viz. playing a duet). Collective recognition requires a more complete shared content: an attribution of a particular function to something or someone. We have collective recognition when people in a community recognize that something has a certain function, as, for instance, when everybody agrees that pieces of paper produced in certain ways have a given monetary value. The two forms of collective intentionality can work in tandem. Consider a simple case. Four friends decide to go camping for the weekend. They cooperate to assign responsibilities for setting up the tent (two people), gathering wood for the fire (one person), and cooking (one person). As a result of this cooperative assignment, there is collective recognition of the role that each person has. According to AC, much more complex situations than this one work basically in the same way, namely, through cooperation to accomplish shared goals, and collective recognition of status.

Although cooperation is the fully fledged form of collective intentionality, according to Searle, collective recognition is somewhat more fundamental to social reality, in that it yields the minimal requirement for there to be a social domain. And, given AC, it should come as no surprise that collective recognition is also the minimal condition for there to be institutions, which are the most complex forms of social organizations:

. . . the existence of an institution does not require cooperation but simply collective acceptance or recognition. (Searle 2010, 58)

Let us then focus on collective recognition and institutions. Given AC, what is claimed about institutional reality is just an extension of what happens at less complex levels of social organization. We say that there is an institution X in a given society when we find institutional facts, to the effect that certain people (and/or objects) have properties that they would lack if there were no X. The thesis is that institutional facts are grounded in collective beliefs: in every case in which we have an institution, people and objects possess the associated social properties in virtue of a collective recognition of such properties.

I have two main objections to this thesis. First, taken literally, it leads to implausible consequences regarding the shared content of such grounding beliefs. Assuming that in the toy example of the four friends on the camping trip the thesis works, it is highly implausible to maintain that most people (in our society) hold detailed beliefs about the social role or function of most of the objects and persons surrounding them. (If you doubt this, ask people what are the duties and powers of the CEO of a company, or what they have agreed to by clicking on the last update of their music software.) Still, if the social function is grounded on a collective attribution, then only a feature of the shared content can account for a feature of the social entity. If we acknowledge that the social entity has a property that does not result from the collective attribution (and is not just grounded on its physical constitution, if there is one), then we need to revise the standard account somehow, because its generalization from simple cases fails. The challenge that any approach based on AC has to face, then, is to fill the explanatory gap between the rough and imprecise beliefs (if, in fact, there are any at all) that most people have with respect to most institutional entities around them, and the very fine-grained system of roles, statutes, duties, and rights that we find in institutional contexts.

One may have the feeling that this critique is overhasty. It is true that we cannot extend the standard account from simple cases to more complex ones without some adjustments. However, how these adjustments should be made is quite straightforward. The collective intentional contents on which social reality depends according to the standard account are not necessarily wholly explicit. Even though it is often convenient to describe these intentional contents in terms of explicit propositional contents, as attributions of a certain function to a certain entity, their psychological status is not necessarily that of a propositional belief. It is feasible to view them as a set of dispositions and capacities (cf. Searle’s 1995 notion of “background capacities”). In particular, in institutional contexts, I do not need to have detailed knowledge of the functions that I attribute implicitly to the CEO of a company in respect to his status as CEO, insofar as other people have it, and I have a disposition to defer to them for such knowledge. We can call the people with the relevant knowledge concerning a given status function “the experts.” ⁸

This revised view will not suffice. It is implausible to maintain that each institutional status requires specific experts (or people otherwise worthy of being deferred to). Maybe for the category of permanent residents, the appointed officers count as the experts, and we defer to their attitudes to establish who counts as a permanent resident in a certain area, but what about all the cavils involved in my status as contractor that I obtain by clicking at the bottom of a long agreement on a website, or those involved in my job contract? Who are, in those cases, the experts who possess the relevant attitudes toward me (or even a more generic attitude toward whoever signs or clicks)? It also seems implausible to maintain that each institutional status requires specific deferential beliefs. To whom am I deferring when I sign a contract? It might be claimed that the deferential beliefs that ground institutional kinds are very general (indeed, they are non-specific dispositions to defer) so it is plausible to maintain that nearly everyone has them. Yet, this would not be an answer, but merely a way of rephrasing the explanatory problem facing any view according to which institutional entities are grounded on the content of shared intentions: if the shared content is general, how can it determine the specific variety of institutional reality? A shared intention with a general content of defer may be a necessary condition for the existence of institutional entities, I agree, but that simply means that the account according to which institutional entities are determined by the shared content of collective intentions requires a (substantive) integration. Institutional statuses are complex sets of functions, constraints, rights, and obligations that cannot be real if their only purported ground comprises general beliefs or dispositions. Elaborating an account of how the contents to which we defer play a role in grounding social reality seems a promising way to overcome this predicament. However, as will soon become clear, by so doing, we will discover that the fundamental AC has to be drastically reconsidered, and in a way that leads us to institutional externalism.

5. Against the Assumption of Continuity: Collective Mistakes

The second objection to AC is that it does not seem to leave room for mistakes about social statuses and about social facts in general. Even mistakes on the part of individuals are problematic for a naïve reading of the thesis. Consider the following example: I promise to take you to the cinema tomorrow. Tomorrow, though, you think that the day before I promised to take you to a high-class French restaurant. How are we to straighten things out? We seem to have the intuition that I am right (I did promise to take you to the cinema and I did not promise to take you to the restaurant) and you are wrong (you have misremembered the promise or you are making some other sort of epistemic mistake). However, Searle’s account predicts a stalemate here. The status of my obligation is unclear, because there is no shared collective intentional content (assuming you are not just trying to tease me), and so nobody is right.

One could counter-argue that insofar as the existence of rights and obligation is concerned, a stalemate is what we should expect. If you and I are the only ones who have partaken in the act of promising, then as soon as you (or I) forget it or fail to recognize it, the promise is no longer binding and so no longer exists. That may be true for informal contexts involving only a few people. The situation is different if many people have heard me promising, and even more different if many people are ready to see to it that we keep the promise. In that case, a plausible reading of AC predicts that you would be mistaken because the majority (or the relevant “experts”) think that you are mistaken; exactly as, in general, you have a certain social status only because the majority (or the relevant “experts”) thinks that you have it.

I agree that for non-institutional, informal, social situations, that is correct, and that in such contexts, individual mistakes are compatible with AC. But what about institutional contexts? Consider the case of a collective mistake, for instance, a football match of the national league between team A and team B. Team A wins the match 3:1, but for some strange reason, everybody comes to believe that team B won (perhaps everyone in the stadium suffered a drug-induced flashback to the previous year’s game, when team B won 3:1, and then the incorrect result was broadcast nationwide over the television and radio). Would not everybody be mistaken? Yet how can this be the case, if the social status of the match depends only on collective attribution and everybody believes that B won? If AC holds, they are not mistaken; team B has won. Yet this is wrong. If everybody came to know that A scored three goals while B scored only one, but they believed the reverse happened because their beliefs were drug-induced, everybody would realize that they were wrong, and not that the result had changed! In an institutional context, the final results would be corrected after the mistake has been discovered.

One might defend AC by claiming that in the case of a football match, the status of the final result is not entirely a matter of what people believe. It also depends on how many goals have been scored by each team (even though, of course, what counts as a goal is, in turn, a matter of what people believe). Thus, people may have false beliefs concerning the historical facts and thereby have collective false beliefs about who won. It is quite plausible to think that there is a sort of implicit deference to the relevant historical facts here, in the sense that what the historical facts are partly determines the nature of the institutional facts. However, collective mistakes may involve basically every kind of social status, at least insofar as we are in an institutional context in which we can collectively decide to defer to what is written on an external accessible support to establish what is socially binding. For instance, by signing a contract, I (usually) accept a large number of duties and assume a large number of rights, regarding which, I, just like everybody else, may entertain false beliefs. Yet, it is not my beliefs that establish what my duties and rights are, nor the beliefs of the collective I am part of, but the content of the written text of the contract. I defer my knowledge to it, because it contains all the information that determines the status function relevant in the situation. And I defer my knowledge, because I defer to its content as to what determines institutional reality. In general, it should always be possible for there to be a discrepancy between what someone (or everyone) believes about the institutional status of an entity x, and its actual institutional status. Yet, an account of social reality based on AC has difficulty providing an explanation of such a possibility, unless it is augmented with the idea that in institutional contexts, we defer to external contents. This latter consideration takes us beyond any account based on AC toward institutional externalism. ⁹

6. Intentions and Established Procedures

According to Searle’s widely influential model of social ontology, the collective recognitions on which social reality is grounded are given by cooperation among individuals. Collective recognitions require a shared intentional content: the assignment of a status function to something or someone. As I have argued, the complexity of many status functions in our modern societies and the possibility of collective mistakes about the social status of something cast doubts on the tenability of such a model. For simple situations of collaborative behavior among members of small groups, the model may work, but to account for more complex situations, it needs to be revised. I think that the problem here is deep, and that the revisions that need to be made go far beyond small adjustments that would make AC still tenable.

The intentions that are required for there to be complex social entities such as institutions do not have the form of the collective recognitions that they possess in simpler cases. Consider again the simple example of a promise: I promise you that I will pay you 5 euros by tomorrow. By this act, which requires a community of at least two persons, a social entity is brought into existence: a promise. You and I are affected by the promise in certain ways: I have an obligation to pay you, and you have the right to request 5 euros from me. In other words, the promise has a binding power. Now, what is it that makes the promise binding? The fact that the parties jointly agree to keep the promise is a precondition for the existence of the social constraints that arise from the act of promising. However, facts regarding who is bound, and how and with respect to what they are bound, depend on the content of the social act that the two parties have performed. The content tells us what has been promised to whom, and thus determines the nature of the bond between the parties.

In simple situations, the relevant content usually coincides with that of the intentions of the parties involved in the promise (or their community, or a relevant part thereof), and those intentions have the form of a recognition of a status function. Yet, even in the case of a promise between you and me, if we write down that I promised to pay you 5 euros by tomorrow and we agree that the content of this inscription is binding, even if we both forget what has been promised, there will be a way, in principle, to settle the issue. Therefore, if we both agree to defer to an external content, the persistence of the social bond and the nature of the social constraints at issue will be determined by this external content. According to institutional externalism, this is what usually happens in more complex situations. In institutional contexts, the content that determines all the social constraints and socially relevant facts is not that of the actual intentions of the participants of the social act that gives rise to the social object. As we have seen, people often do not have detailed beliefs about many status functions, and they can be mistaken about them. Institutional situations require deference to publicly accessible contents, and not a specific collectively shared content. Note that here, I am not trying to substitute a kind of continuity based on shared intentions with the idea that external supports are essentially involved in any social situation, no matter how informal. ¹⁰ Suppose that I make the promise to pay you 5 euros by tomorrow, and we both agree to write it down. Suppose I do the inscription with a computer and it ends up saying I pay you 500 euros (while we thought it says 5.00 euros). Do I now owe you 500 euros as that is what the inscription says? If the context of the promise is non-institutional, it does not seem so, even if we said that the content of this inscription is binding. When the external content is not essential to the existence and nature of the social entity, we almost never give up our interpretative power to documents. However, in a context where the writing of the promise is part of an institutionalized procedure, then—once (and if) we realize the mistake—we must undertake some further action to change the external content to change the institutional facts. It is not that our interpretive power or intentions are irrelevant, but they are not what constitute the features of the institutional reality that we inhabit. In other words, I do take discontinuity between informal contexts and institutional contexts very seriously. The difference between the two kinds is not just a difference in the way social reality comes to be, it is a difference in the kind of ground that it has.

The difference between a view based on the shared content of collective intentions and institutional externalism is profound. On the first view, what determines the nature of social entities tout court—hence, also in the case of the institutional sphere—is a shared intentional content. The elaborate social statuses that we find in complex social structures are grounded in shared beliefs about such social statuses. As I argued, there is an explanatory gap here between the rough grounding of shared intentions and the actual fine-grained complexity of the grounded institutional phenomena: generally, people have only rough and general beliefs about complex status functions. In contrast, on the proposed view, in institutional contexts, the intention that is required for the existence of a status function F is not necessarily a collective recognition of the status function F—not even by a designated group of “experts.” Rather, the shared intentional content consists in deferring to the content that has been produced by an established procedure. Thus, social intentions of complex structures have to be general, because we need a non-specific intention to regard the production of a certain publicly accessible content as valid and binding, so that the status function will be determined by the specific content produced, regardless of whether or not it is entertained as a shared content by the parties involved (or any other person in the social context, for that matter).

On the proposed view, the collective recognition that grounds institutional reality is a form of cooperation, but in a much weaker sense than Searle proposes. In his view, the social world, composed of social entities, exists only if there are collective intentions with a shared substantive content, such as that a piece of paper possessing certain features is a 5 euro note. Even if, strictly speaking, a common goal is not required for collective recognition, the fact that the shared content has to do with the specific features of the social entity makes collective recognition and cooperation two sides of the same coin, as it were. In a sense, a shared assignment of status function is a limiting case of a common goal (Searle 2010, 65-68). In the proposed view, the existence of institutions requires only that the individuals in a collective share the general belief that the enactments produced according to certain established procedures determine social constraints, rights, duties, and the like. The shared part of the intentional content is general, not only in the sense that it does not contain any specific attribution of social functions, but also in that it does not concern any specific procedure. For there to be institutional kinds and entities, it is required only that we defer to enactments produced by the procedures that are established in the context in which we find ourselves. How the content of this shared intention may be described in detail is probably at least partly an empirical issue, which is connected to the wider problem of social coordination. According to the general working hypothesis, which procedures are regarded as producing valid enactments is largely a contextual matter, and it may depend on the cultural background. What the thesis predicts is that the relevant shared intention is essentially deferential in nature, and that its target is the content produced by a contextually established procedure.

7. Institutions and Non-collaborative Behavior

Collaborative behavior of a more substantive kind than collective recognition of the validity of certain procedures to produce documents or other forms of binding contents is a widespread phenomenon in society, and it is very important and valuable in many social contexts. Searle’s view takes cooperation (especially in its degenerative form of the collective recognition of status function) as the basic building block of social reality. However, as I have argued, while this view may work for simple cases, it is wrong to generalize it to complex cases. A further objection to any view based on AC is that collective recognition of status functions seems to require acceptance of the assignment of status functions. Discussion of this point will allow me to clarify why deference is an essential aspect of institutional reality.

According to Searle, there is an association between recognition and acceptance: the recognition is shared by being accepted by at least some individuals of the relevant group. However, there certainly can be cases in which we are involved in social situations which we do not accept. A defender of AC can reply that shared recognition may come with apathy, distress, or even hatred toward the attributed status at least insofar as there is some higher-order goal that is shared (e.g., the cohesion of society). ¹¹ If I am playing chess with you and you are winning, I will recognize that you are winning even if you are humiliating me, insofar as I am willing to play with you (i.e., I aim to win, but I also have the higher-order goal of playing with you, and the latter is what we share). Or, another example, if I am willing to live in a democratic society, I will recognize the social status and power of the elected president even if I strongly disagree with their political views.

Now, is it really plausible to suppose that a specific higher-order collaborative intention is always at work in every case in which people behave non-collaboratively toward the institutions? Take a prima facie simple case, such as getting fined for parking. Suppose that I think that the officer has made a mistake in fining me. I can refuse to consider this particular fine as justified, while continuing to regard the practice of fining (and other “useful” forms of social punishment) as generally justified. However, lack of justification with respect to a specific fine is not tantamount to believing that the fine lacks validity. I may not at all be happy to recognize its validity, but this is what I am doing if I protest—it would be pointless to protest if I thought it to be invalid.

In the proposed view, the general intention on whose ground non-collaborative social behavior can occur does not need to be a higher-order one. Generally speaking, complex situations require only the first-order intention to consider validating the established procedures we find around us. Consider again the case of the fine. If I think that the officer has made a mistake, I will try my best to have the fine dismissed; that is, I protest the validity of a certain document, the official piece of paper that the officer has issued to me upon which the fine and the circumstances that warranted it are recorded. Such a protest can only be upheld by appealing to another validating procedure, which typically will require me to provide evidence for my claims. Indeed, even if I consider the whole practice of fining to be unjust, I do not consider it as lacking validity if I engage in some form of “protest” against it. Again, it would be pointless to protest if I do.

Yet one may have the feeling that the proposed view is too minimalist. People do not always behave so apathetically toward institutions and society. Even granting that the content of collective recognition is often too simplistic and general to ground sophisticated institutional functions, and that we need publicly accessible records to flesh out the details of status functions, the content of the shared intentions that we find in society is relevant for the existence of the corresponding social object or function. If nobody cared about a certain social function, it is difficult to see how this function could affect our social life, even when there is a document that describes it in full detail and has been issued by an established procedure. ¹²

This objection is barking up the wrong tree. The thesis that institutional functions are grounded on enactments produced by established procedures to which we defer is compatible with the fact that we do care about what these functions are, and, indeed, also about the established procedures as such. The reason we are interested in them is that they can, and often do, have important consequences for us. To be sure, in modern societies, the established procedure of producing laws in parliament affects the lives of everyone. The picture is also compatible with collectively shared beliefs having a crucial causal role in explaining social changes over time. For instance, if nearly everybody disagreed with a certain content produced in an established procedure, which has far-reaching consequences (e.g., a law), it is likely that people will start to act to change this content (either inside or outside the boundary of what is institutionally permitted). Or, more dramatically, if at a certain point, the majority of the population start to share a hostile belief against an established procedure (like the parliamentary procedure of issuing laws of a given state), it is likely that they will collectively aim to disestablish the institution in question. Besides, procedures can be established through their coercive imposition, and small minorities can affect directly established institutions (as a consequence, for instance, of terrorist actions).

Hence, I am not denying that shared beliefs with a substantive content (as to what the social and institutional features of a given situation are or should be) or substantive collaboration with respect to a common goal are often crucial aspects of social reality. However, common goals and shared beliefs emerge more decisively in the account of the (broadly speaking) causal mechanisms that underlie many aspects of society, rather than in the account of the peculiar ontological status of the institutional sphere. The existence of an institutional sphere requires only the presence of established procedures; it does not require the sharing of more substantial goals or beliefs. Besides, within an institutional framework, we find both cases of harmonious cooperation between people sharing a common goal, and cases of difficult negotiation over conflicting interests. Modern societies allow for both. Sometimes there is substantive collaboration toward achieving a common goal, such as when we want to play a duet and we decide that I will be playing the violin and you the piano. However, at other times, there is agreement only on the very “general rules of the game.” Substantive collaboration can be a valuable thing in certain circumstances precisely because it is not the basic form of social behavior to be found in every circumstance. A duet is a wonderful thing, just as society in many cases is not.

8. Conclusion

In this article, I argued that most of contemporary social ontology rests on the mistake of providing an “internalist” account of institutional reality. The thesis that the content of shared intentions determines social functions, roles, and the like has some plausibility with respect to informal and non-complex social contexts, but faces two serious problems when it comes to more institutionalized aspects of society. First, it has great difficulty explaining how often very imprecise and general shared content can ground complex and often detailed institutional profiles; and second, it cannot account for the possibility of collective mistakes about social facts. I propose an alternative view, which I term institutional externalism, according to which the only relevant shared intention in an institutional context for the existence of institutional entities is that of deferring to established procedure as to what contents are socially binding, namely, as to what grounds facts about institutional reality. The view does not suffer from those problems, and it also provides a framework in which the role of substantive collaboration and substantive shared intentions in society can be correctly assessed.