Context-Dependence in Searle’s Impossibility Argument

Introduction

In Minds, Brains, and Science, John Searle argues that social-scientific laws are impossible (Searle 1984, 81-82). If there were social scientific laws, claims Searle, there would be “some systematic correlation between phenomena identified in social and psychological terms and phenomena identified in physical terms” (Searle 1984, 81-82). No such systematic correlation is possible, says Searle, because social phenomena are physically open-ended, and therefore allow for an infinite number of physical realizations. If, for any social phenomenon, there is an infinite set of physical phenomena to which it might correlate, there can be no systematic correlation between the social and the physical, and therefore no social-scientific laws. Furthermore, claims Searle, we can test a potential social-scientific law by using it to make predictions about a social phenomenon, and then checking these predictions against the predictions of physical laws pertaining to the physical objects associated with that phenomenon (Searle 1984, 75-76). If the physical laws and the social-scientific law yield the same predictions, then the social-scientific law obtains. However, because social phenomena are physically open-ended, the predictions of a social-scientific law cannot reliably match those of physical laws.

William Butchard and Robert D’Amico have recently argued that Searle’s own account of the ontology of social phenomena provides the conceptual apparatus for explaining the systematic connection between the physical and the social that Searle denies, in the form of the “counts as” locution (Butchard and D’Amico 2010). They claim that by incorporating the “counts as” relation into an explanatory bridge principle, Searle’s own account can accommodate the possibility of social-scientific laws. These “counts-as” social-scientific laws are also alleged to pass Searle’s prediction test (Butchard and D’Amico 2010).

The success of Butchard and D’Amico’s argument depends on there being at least one social phenomenon that is not physically open-ended. I will first argue that there is no such social phenomenon on Searle’s view. Then, I will argue that “counts as” social-scientific laws pass Searle’s prediction test only if our attitudes regarding a social phenomenon are held fixed. Since our attitudes about social phenomena are not fixed, the claim that “counts as” social scientific laws pass Searle’s prediction test should be rejected. Butchard and D’Amico’s response to Searle should therefore be rejected as well.

Physical Open-Endedness and Social-Scientific Laws

Butchard and D’Amico argue that at least one social phenomenon, money, is not physically open-ended (Butchard and D’Amico 2010). From this premise, they argue that it is possible to formulate an explanatory bridge principle connecting the physical and social regularities of money. By Searle’s own lights, this implies that social-scientific laws are possible. Butchard and D’Amico’s argument for the possibility of such an explanatory bridge principle is the following:

For this argument to succeed there must be some social phenomenon that is not physically open-ended. Butchard and D’Amico claim that money, by Searle’s own lights, is such a phenomenon (Butchard and D’Amico 2010). In virtue of the function, or purpose that money must serve in society, claim Butchard and D’Amico, there are certain properties that a physical object necessarily must possess if it is to count as money. According to Butchard and D’Amico, if an object lacks these properties, it cannot satisfy the function of money, and therefore cannot count as money. They defend this claim by the following argument:

This conclusion, in combination with the first argument that I discussed, entails the conclusion that it is possible to formulate a bridge principle connecting the physical and the social regularities of money in a systematic way, thereby proving that such explanatory bridge principles are possible. By Searle’s own account, if there can be bridge principles explaining a systematic connection between the social and the physical features of a social phenomenon, this implies that social-scientific laws are possible.

Although one might disagree with several of Butchard and D’Amico’s premises, I am concerned here only with premise 2a. Butchard and D’Amico claim that Searle himself would accept that there are necessary constraints on the physical properties that an object must have if it is to count as money.

When Searle claims that money is physically open-ended, he should not be understood as making the strong claim that absolutely anything could be used as money. He lists some conditions on what properties something has to have for it to be possible for a community to use it as money. (Butchard and D’Amico 2010)

Butchard and D’Amico here suggest a weak interpretation of Searle’s claim that money is physically open-ended. They take Searle to mean only that the physical properties of an object are not sufficient for the object to count as money, rather than the stronger claim that any physical object can count as money, irrespective of its physical properties. Weak physical open-endedness means that a social phenomenon has necessary physical properties, but no sufficient physical properties. Strong physical open-endedness means that a social phenomenon has neither necessary nor sufficient physical properties. Butchard and D’Amico interpret Searle as claiming that social phenomena are physically open-ended only in the weak sense.

Searle’s own view is actually rather obscure on this point. Butchard and D’Amico cite several passages where Searle seems to defend the view that they attribute to him (Searle 1995, 86; Searle 1999, 126). But other passages suggest that Searle regards the physical open-endedness of social phenomena in the stronger sense. For example, in Minds, Brains, and Science, he claims that “(t)he defining principle of such social phenomena set no physical limits whatever on what can count as the physical realization of them” (Searle 1984, 78). Later in the same work, he states “(i)f for some reason money had to be made of ice, then it would be a strict law of economics that money melts at temperatures above 0 degrees centigrade” (Searle 1984, 84). In these passages, Searle seems to be suggesting the stronger claim that Butchard and D’Amico deny is representative of his view.

In Searle’s later works, he explicitly expresses the view that money is physically open-ended in the strong sense. In Mind, Language, and Society, Searle states that “the move from commodity money to fiat money is a move from the assignment of a function in virtue of physical structure to a pure case of a status function” (Searle 1999, 128). This passage indicates a distinction between commodity money as having certain necessary physical properties, and fiat money, which has no necessary physical properties. Searle’s more recent discussion of electronic money further demonstrates his commitment to the strong physical open-endedness of money. Electronic money fulfills the function of money despite having no physical manifestation at all (Searle 2010, 20-21). If something can fulfill the function of money despite having no physical properties, surely there cannot be any necessary physical properties of money.

Butchard and D’Amico suggest, in a footnote, that electronic money does not count as money, citing the phrase “electronic money” as a kind of misnomer for the practice of transmitting records and recording receipts. This misses Searle’s point. Despite the fact that electronic money has no physical manifestation, it still counts as money in virtue of our collective recognition of the status function. “Electronic money” does not refer only to a kind of transaction; it is a status function that imposes obligations upon us. Despite this, it has no necessary physical properties, which makes it a problematic case for Butchard and D’Amico.

Although Searle’s own discussion of the physical open-endedness of money is somewhat unclear, the details of his social ontology serve to eliminate further confusion about his view. On Searle’s account, a social entity fulfills its function in virtue of the collective recognition of the object as having the relevant status, irrespective of the object’s physical properties. It is only within a certain social context that the set of physical objects that can count as money is finite. In principle, any physical object can count as money if the community collectively recognizes its status as money. To see that this is the case, I will briefly turn to the details of Searle’s social ontology.

Searle’s Social Ontology and Strong Physical Open-Endedness

Searle argues that all of social reality depends on a single type of speech-act, the Declaration, for its existence (Searle 2010, 13). Declarations are a unique type of speech-act that has both the mind-to-world and world-to-mind directions of fit. Declarations change reality so that it matches the content of the declarative speech-act, which gives Declarations the world-to-mind direction of fit. For example, my promising to pick you up at the airport on Thursday makes it the case that I have made a promise to you. However, the speech act of promising also represents reality as being so changed, which gives Declarations the mind-to-world direction of fit. In virtue of having promised, I have both brought a promise into existence where it did not exist before and represented a state of affairs that includes my promise to you.

All social phenomena, claims Searle, are brought into existence by a specific logical form of Declarations, aptly named the Status Function Declaration. Status Function Declarations are typically explicit speech-acts with the logical form “X counts as Y in context C.” The Status Function Declaration gives rise to a constitutive rule, which makes it the case that some X just is some social phenomenon Y. The X term specifies the conditions of satisfaction for something to count as Y, where the X term is satisfied by an object (or a state of affairs) and Y names a social phenomenon. Searle refers to the Y term as a status function.

The notion of a function is particularly important for both Searle’s social ontology and Butchard and D’Amico’s argument. Functions, says Searle, are intentionality-relative (Searle 2010, 59). Humans impose functions on objects in order to serve some specific purpose. Whether some object counts as that type of object depends on whether the object fulfills the imposed function. For example, in order to count as a screwdriver, an object must serve the specific purpose for which a screwdriver is intended. Status functions also serve a specific purpose, which is imposed upon them by humans. Therefore, status functions are also intentionality-relative. As Searle puts it, “they are functions that a person or other entity has, not in virtue of physical structure, or at any rate not solely in virtue of physical structure, but in virtue of collective imposition and recognition of a status” (Searle 2010, 59). It is not the physical properties of an object which make it the particular status function that it is. Rather, a community imposes a status function in order to serve some purpose, and the object counts as that status function in virtue of satisfying this intended purpose.

Status functions are able to fulfill their intended purpose, claims Searle, because they give rise to deontic powers. Deontic powers impose obligations, irrespective of agent desires. For example, when I promise to pick you up at the airport, I have imposed an obligation on myself to be at the airport at a certain day and time in order to pick you up. This obligation gives me a reason to go to the airport when I have said that I would, even if when the day comes, I desire to stay in bed and watch television instead. Promising has deontic power because it creates reasons for action that are independent of agent desires. In general, this is how status functions accomplish their intended function. They impose desire-independent reasons on people, which lead them to act in a way that satisfies the purpose of the status function.

The notion of collective recognition is the final feature of Searle’s social ontology that is relevant to understanding the failure of Butchard and D’Amico’s argument. As previously mentioned, Status Function Declarations are speech-acts that bring status functions into existence. They provide constitutive rules for an object to count as a status function under specific social conditions. But not all Status Function Declarations alter social reality. For example, I cannot make it the case that I am the president of the United States simply by declaring it to be so. The Status Function Declaration in virtue of which some X counts as the president is backed by the authority of numerous institutional structures, and the community over which the status function of president has power collectively recognizes that these institutions have the authority to impose this status function. In the absence of collective recognition, status functions have no deontic powers and therefore cannot fulfill the function for which they were created. If no one recognizes any obligations as following from a status function, then no one has any desire-independent reason to act in a way that satisfies the intended purpose of the status function. The capacity of a status function to serve its intended purpose is also dependent on the community continuing to collectively recognize the status function. If the American citizenry suddenly ceased to recognize some person as the president, this person would cease to be able to fulfill the function of the president, and thereby cease to have the status function of president.

We are now in a position to explain the social phenomenon of money according to Searle’s social ontology, and to demonstrate why it is physically open-ended in the strong sense denied by Butchard and D’Amico. First, Status Function Declarations specify a constitutive rule by which something counts as money in a particular social context. Some object X counts as money in virtue of a Declaration, which makes this the case. The X term refers to some object, such as a dollar bill, while the Y term names the status function, money. The relevant context is specified by the set of propositions that must be true for a dollar bill to count as money. Among these propositions will be the fact that the community collectively recognizes the status function as authoritative, as well as collective recognition of certain obligations that follow from the status function. For example, recognizing a dollar bill as money obligates me to accept it as payment of a private debt, and it obligates others to accept it from me in exchange for goods that I wish to purchase. Any X can count as money, so long as the community collectively recognizes both the authority of the Status Function Declaration as giving rise to a constitutive rule, and the obligations that follow from the status function. The collective recognition of a status function is both necessary and sufficient for some object X to count as some social phenomena Y in context C, while no physical properties are necessary or sufficient for X to count as Y in C. The physical properties cited by Butchard and D’Amico are necessary conditions for what can count as money only given an intentionality-relative social context.

Given a social context where the function of money is such that it must be durable, handleable, transportable, noncounterfeitable, and recognizable, these physical properties necessarily constrain what can count as money. But these necessary constraints are themselves context-dependent. For example, durability is not a necessary physical property for something to count as money in a social context where all monetary transactions take place on a single day of the week, whereafter all money is effectively deemed valueless and discarded. In a social context where all monetary transactions involve electronic money, handleability and transportability are not necessary physical conditions for an object to count as money. And so it goes for any physical property that we might take to be necessary for something to count as money. The necessity involved is a consequence of the social context where the status function is collectively recognized. Prior to the imposition of necessary physical constraints, which follow from the intended function of money within a particular social context, any physical object can count as money. Searle’s social ontology therefore does imply the strong sense of physical open-endedness that Butchard and D’Amico must deny for their argument to succeed, both in general and with regard to the social phenomenon of money.

Returning to premise 2a, even if it is true that an object must fulfill the function of money to count as money, it is not the case that an object must have certain physical properties in order to fulfill the function of money. Any physical object can fulfill the function of money if the object is collectively recognized as having the status function of money. While there often are physical constraints on what can count as money in a given social context, this does not imply that money is physically open-ended only in the weak sense. The function of money depends on the intended purpose for which a status function is imposed upon some object. Function constrains the physical manifestations of a status function, as Butchard and D’Amico claim, but only within an intentionality-relative social context.

Passing Searle’s Prediction Test

One might find this response to Butchard and D’Amico unconvincing, for at least two reasons. First, they explicitly state their intention to account for a systematic explanatory connection between the physical and the social, rather than a systematic causal connection (Butchard and D’Amico 2010). The physical features of money, given a social context, do seem to explain something about the function of money within that social context, even if the necessity of these features is context-dependent. Furthermore, Butchard and D’Amico stipulate that “counts as” social-scientific laws only yield the same predictions as physical laws if our attitudes are held fixed (Butchard and D’Amico 2010). If we take this to include holding all relevant contextual factors fixed, then “counts as” social-scientific laws perhaps would yield the same predictions as the physical laws that pertain to a particular status function. Butchard and D’Amico’s argument for this claim is the following:

Butchard and D’Amico contend that the systematic connection between the physical and social regularities of money is an explanatory connection, captured by the “counts as” locution. But what can Butchard and D’Amico mean when they stipulate that our attitudes about a social phenomenon are fixed? It seems that, for “counts-as” social-scientific laws to reliably produce the same predictive results as physical laws, at least two sets of attitudes must remain fixed. First, the collective recognition of the status function for which an object satisfies the X term must remain constant. For some X to count as money in the future, it must continue to be acknowledged as money by the community, which includes the collective recognition of the obligations implied by this status function. This collective recognition would also have to include the collective recognition of the authority of whatever institution is responsible for the Status Function Declaration that gave rise to a constitutive rule whereby some X counts as money in context C. In addition, the attitudes that comprise the relevant social context must remain fixed. If all of social reality is maintained by the collective recognition of Status Function Declarations and the obligations that they imply, then our collective recognition of the social institutions that comprise social reality must remain fixed if we are to make reliable predictions about phenomena within that social reality.

Consider an example to help clarify precisely what Butchard and D’Amico are proposing. During the Civil War, the Confederate government issued their own currency. This act amounted to a Status Function Declaration that gave rise to the following constitutive rule: a Confederate dollar counts as money in the Southern states that comprise the Confederacy. Suppose I were to predict at the beginning of the Civil War that the Confederate dollar would still count as money when the war ended. Suppose I were to simultaneously predict the same result from the set of physical laws. A number of attitudes would have to remain fixed for these predictions to coincide. First, the Confederate dollar would have to be collectively recognized as money by the members of the Confederacy. This collective recognition would have to include things like being willing to accept Confederate dollars as repayment of private debts, as well as collectively recognizing the constitutive rule whereby Confederate money is created as authoritative. Finally, the collective attitudes that comprise the relevant social context would have to remain fixed. In the absence of the collective recognition of the Confederacy as an existent social institution, the Status Function Declaration that created Confederate money ceases to obtain, because the relevant social context ceases to exist.

From this historical example, one can immediately see the problem with stipulating that our attitudes are fixed. It is because our attitudes are not fixed that the Confederate dollar ceased to count as money when the Civil War ended, irrespective of the predictions of the physical laws that pertain to Confederate dollars as physical objects. Even if the physical analogue of the predicted social result were to obtain in this case, the social-scientific prediction would only coincide if the relevant attitudes were held fixed. If attitudes are not fixed, then the predicted physical result can obtain even where the predicted social result does not. As Searle points out, this is precisely what makes the social-sciences resistant to law-like explanations (Searle 1984, 77-78). While the gas laws and the laws of molecular motion will always yield the same predictive results for a given phenomenon, the predictions of physical and social-scientific laws will not yield the same predictive results in cases where our attitudes change (Searle 1984, 76-77). As Searle might put it, even if all the relevant molecules move exactly as predicted, without collective recognition of the status function, there is no Confederate money (Searle 1984, 75). Even if the systematic connection between the physical and the social is an explanatory one, the physical regularities of a social phenomenon do not explain anything about that social phenomenon if the relevant attitudes fail to hold. Butchard and D’Amico’s “counts-as” social-scientific laws pass Searle’s prediction test only by stipulating a premise that trivially guarantees the truth of their conclusion.

Conclusion

Butchard and D’Amico defend the position that, by Searle’s own lights, the “counts-as” locution can support an explanatory bridge principle connecting the physical and social regularities of money in a systematic way. The possibility of such a principle implies that social-scientific laws are possible. Their argument for this position relies on money not being physically open-ended in the strong sense that allows for any physical object to count as money. I have attempted to demonstrate that, while Searle’s own discussion of the physical open-endedness of money is somewhat ambiguous, his account of social ontology implies a commitment to the strong physical open-endedness of all social phenomena, because any necessary physical constraints on what can count as a particular social entity are themselves intentionality-relative and context-dependent.

To bolster their argument, Butchard and D’Amico claim that “counts as” social-scientific laws would yield the same predictions as the physical laws for a given social phenomenon, which Searle takes to be a litmus test for the legitimacy of a proposed social-scientific law. However, they achieve these results only by stipulating that our attitudes are fixed. But our attitudes regarding social phenomena are not fixed, which is precisely why Searle rejects the possibility of social-scientific laws. Even if the physical world behaves exactly as predicted, a predicted social phenomenon obtains only if the relevant attitudes obtain. Butchard and D’Amico pass Searle’s prediction test only by failing to grasp the significance of our attitudes to the existence of social phenomena.

Even if Butchard and D’Amico’s challenge to Searle were successful, their argument would not have much bearing on the explanatory methodology of the social sciences. As Searle notes, the interesting thing about social phenomena is that they are both intentionality-relative and dependent on human attitudes for their continued existence (Searle 1984, 82-85 and Searle 1991). Even if we could come up with social-scientific laws, it’s doubtful that explanations in terms of such laws are the sorts of explanations that the social scientist ought to seek out, because such laws would be silent on the most intriguing features of social reality—the attitudes and intentions of human beings.