Reflections on the Ontology of Money

3.1 Does Constitutive Rule Reveal the Constitution of Money?

An account of money should tell us what money is, what constitutes money. One might therefore expect that John Searle’s idea of constitutive rule would be helpful here, especially given that he suggests that such rules play a central role in understanding institutions, and that he himself uses money as a paradigm example when discussing constitutive rules (Searle 1969, 1995). Constitutive rules have this form:

• [CR] X counts as Y in C.

Searle’s favourite example is this: a piece of paper counts as a dollar bill (or, more generally, money) in the United States.^[9] The question is whether this formula is helpful as an account of the nature of money. I argue it is not. It does not tell us what money is, namely a set of causal powers sustained by systems of certain kinds of practices and institutions. To see the limited services provided by the formula, we need to consider the relationship between institutional reality and non-institutional (or pre-institutional) reality, and the distinction between money as a general type and particular items of money.

Money is often taken to be an institution, but it is seldom made clear what precisely is intended. What does it mean to say that money is an institution? On Searle’s account, applications of constitutive rules of the form [CR] create what he calls institutional facts. It is an institutional fact that a particular piece of paper counts as a dollar bill. Likewise, that a piece of metal counts as a euro coin is an institutional fact. We may say that the rule points out a certain property, or bundle or properties, of the piece of metal: they are social properties that do not reduce to the physical or chemical properties of that piece of metal. But the fact that a particular piece of metal has these properties is not an institution: it is an institutional fact about that piece of metal. Institutional facts presuppose institutions but are not institutions themselves. This leaves it open what institutions are and what it is for money to be an institution. Searle says institutions are “systems of constitutive rules” and that this applies to money, too: “… institutions like money, property, syntax, and speech acts are systems of constitutive rules …” (Searle 1995, 140; see also 2010, 10). But that’s about it, no further elaborations are provided. Searle has a lot to say about institutional facts, but very little about institutions.^[10]

Let us see what we can make of this. A closer look will reveal that Searle’s idea – that institutions are systems of constitutive rules – cannot be quite right. Employing ideas suggested in the previous section, we can say that Y is a bundle of causal properties: these are causal powers of money that economists call its “functions” and regard as its “intrinsic characteristics”. Shouldn’t we say it is this bundle of causal properties sustained by a structure of other institutions that constitutes the institution of money? Whatever else is required for this institution to emerge and persist does not define the institution itself. In particular, rules that govern the way in which the institution is instantiated in particular pieces of physical matter do not define the institution of money. Therefore, we should not say, as Searle does, that the institution of money is a system of such rules.

Consider again Searle’s paradigm example: this piece of paper counts as a dollar bill. Obviously, the particular piece of paper has to have a number of properties in order to qualify: it must be of a certain size and shape and colour and chemical composition, with definite figures printed on it by authorised agencies, and so on. Importantly, none of these properties is essential for the piece of paper to count as money. Rather, they are contingent features of a particular currency such as euro, yen and krona. They are contingent features since they could as well be otherwise, as witnessed by the endless variety of bank notes that there have been across time and place in the history of humankind. By contrast, the causal properties that constitute money are non-contingent, they are necessarily possessed by anything that is money.

Let me propose somewhat more abstract terminology that may help to see the point more clearly. We can say that money is a social universal, while currencies are its generic instantiations, and that certain particular items of matter are instantiations of currencies and money. In philosophers’ jargon, we could put this in terms of the distinction between (general) types and (particular) tokens, but given that ‘token’ is in use with other connotations in various accounts of money, I’ll refrain from using it here. Using the language of universals and particulars, we may now characterize their relationships by saying that the money universal – moneyhood or moneyness – manifests itself in particular items of money and in generic currencies. We may also say that these relationships are characterized by multiple realizability: money has multiple realizations in specific currencies and particular items of money. The money universal is one, its realizations are many. The modal status of these relationships of manifestation and realization is obvious: they are contingent rather than necessary relations.^[11]

It is now easier to see the limited services that Searle’s [CR] might provide. It describes an aspect of the bridge between physical matter and social reality: particular items of matter count as money, that is, have institutionally recognized powers. At most, [CR] tells us what is money, but it does not tell us what money is: it says that certain items of matter are (particular instances of) money, but not what the identity or nature of money is. In even more revealing words, the constitutive rule [CR] does not inform us about the constitution of money. [CR] appears to suggest that certain particular items of matter constitute money, but this would be misleading. At most we might say that the complete set of such items constitutes the extension of a currency, but this is an entirely different thing to say. I have suggested that money – moneyhood – is constituted by a set of institutionally and practically sustained causal powers. [CR] does not help to express this idea, thus in this context should perhaps not be called a “constitutive rule” at all. Calling it the “counts as” rule would be all right.

3.2 Is Money a Convention?

I have proposed distinguishing between levels of abstractness or generality – money as a social universal, generic currencies, and particular instantiations. This will help to clarify a few other issues as well. It is sometimes suggested that money is a social convention. This can now be questioned. Conventionality generally understood combines the properties of being customary and being arbitrary. The concept of convention I have in mind focuses on the arbitrariness or contingency of a rule or practice or fact: things could as well be otherwise (but naturally being contingent and having a customary status may be conjoined). Let us then introduce the idea of degree of conventionality. The important observation to make is that currencies are characterized by a high degree of conventionality, while the money universal is not. The properties of the dollar bill are conventional, while the causal properties of money as a social universal are not. Any given currency is dependent on various local-natural and local-cultural conventions and constraints. This applies to the material aspects of money particulars, and the symbolism and numerics used. The Netherlands nowadays uses coins and bank notes for 1, 5, 10, 20, 50 euros, whereas a couple of decades ago it used to have them for 1, 2.5, 10, 25 guilders, with different symbolism and colours.

These are recognitional properties, that is, properties in terms of which we recognise tokens of a currency as instances of money. Using the terminology of essence, those recognitional properties belong to the nominal essence of money, but they do not constitute money’s real essence (while they may be taken to constitute the real essence of a currency – supposing one wishes to entertain such an idea). The nominal essence of money is more conventional than its real essence: it may vary from country to country, from historical period to another, and this variation is dependent on local contingencies that do not contribute to the essence of moneyhood, the money universal. In contrast to many other cases – such as the metal gold – the relationship between the real essence and the nominal essence in the case of money is relatively less stable across time and place. The connection between the atomic constitution of gold (presumably being its real essence) and its colour, malleability and other observable features (its nominal essence) is more stable than, say, the connection between money and gold.

The degree of conventionality is relational. In relation to the money universal, currencies are highly conventional as just noted: it is up to a set of conventions what counts as a currency relative to money. Likewise, in relation to a given currency, any particular items of money are conventional. Like a currency, a particular item of money is highly conventional in relation to money.

We may conclude that money as a social universal is not a social convention. We may grant that certain traffic rules are contingent conventions in that they could as well be otherwise. Driving on the right hand side of the road is such a rule, and there is nothing necessary about it, at least not in the same sense as it is necessary for money to serve as a unit of account and as a means of payment. The traffic rule has been shaped by historical contingencies just as currencies and money particulars. They are rightly characterized as conventions. This means, among other things, that David Lewis’s account of convention does not apply to money, while it has a better chance with respect to currency (see Lewis 1969, 48–49).

4.1 Exercise and Immanence

We may employ the universal/generic/particular division suggested above to make a couple of further observations. The first is about how the existence and exercise of money’s causal powers are related. My suggestion is that in this respect the money universal is different from its instantiations. In the case of the money universal, existence and exercise are linked in that existence requires exercise: money’s powers need to be exercised in order for them to exist. In this respect money’s powers are similar to many human powers, such as physical strength and various cognitive skills at given levels: without exercise, they tend to deteriorate. They are dissimilar to the (active) power of landmines to explode and the (passive) power of glass windows to break. These latter powers may exist even if never exercised: even though no landmine ever explodes and no window ever breaks. The same applies to particular items of money: a particular item may never be used for any purpose, thus its powers remain unexercised without being deteriorated. But the money universal is different: if the powers constituting it are never exercised, they cannot possibly exist. In this sense the existence of money’s powers are dependent on their exercise.

The second and related observation has to do with rival accounts of universals. Transcendental (Platonic) accounts claim that universals reside in a separate realm independent of their instantiations in particular things, while immanent (Aristotelean) accounts claim that universals only exist as instantiated in particular things (see Armstrong 1978; Mäki 1990, 1997). What I have suggested above implies an immanent view about the money universal: moneyhood only exists as instantiated in particular currencies and money items rather than independently of them. Viewing this thought from the point of view of philosophy’s own debates, here we have an argument for a “naturalized” position in the theory of universals: the dispute between the transcendental and immanent views about the money universal is being resolved by relying on specific factual information about money as part of social reality. This is much in line with how I see the role of factual (preferably scientifically acquired) information in relation to many philosophical issues.

4.2 In Virtue of Collective Acceptance

In their accounts of social ontology, philosophers often use the case of money to illustrate some general ideas about institutional reality. One such general idea is that social facts exist in virtue of being generally believed to exist. A somewhat weaker formulation is to say social facts exist in virtue of collective acceptance. This gives us an answer to the question, How does money exist? The answer is: in virtue of general acceptance or belief. This gives us the rudiments of “idealist” accounts of social reality. Note that this view does not go quite as far as the extreme individualist version put forth by one social thinker: “Money is money, a word is a word, a cosmetic is a cosmetic, if and because somebody thinks they are.” (Hayek 1948, 60)^[12]

There are many things that should be said about this view, but I will only take up two issues, one briefly, the second more extensively. First, to say that money exists just in virtue of collective acceptance is to suggest that money is highly conventional. It is conventional in the sense of being arbitrary or contingent. It is contingent upon such collective belief or agreement or acceptance: the relevant population of people could as well have agreed otherwise, had they had different perceptions and convictions; and whenever they were to change their minds, money would cease to exist. From the point of view of the ontology of existence, there would be no major difference between money, currency, and particular instantiations thereof.

The second issue has to do with existence of money more directly. Does money exist? Does it exist in any sense that would satisfy the realist? To address this question, I will next discuss David-Hillel Ruben’s argument about realism and money (Ruben 1989).^[13] The key question in the argument is whether we can be collectively mistaken about money. One important implication of realism is that even our most perfect theories and beliefs may be in error about reality since reality exists independently of them, whereas the relevant version of antirealism implies that beliefs or theories meeting our most stringent methodological standards cannot be in error – such beliefs are true of some facts simply because they somehow create those facts (see Mäki 2002). If there is no possibility of being in error in the case of money, then money does not exist independently of beliefs about it, and we are advised to take an antirealist stance about it.

Optimism about realism about money can be fostered by considering situations like this: A person passes over a physical item, believing it to be a valid bank note, with an intention to purchase a piece of cheese, but fails to achieve her goal since the note is found to be counterfeit by the cashier. The person turned out to be in error concerning the causal powers of the item. Such physical items really do or do not have such institutionally supported powers regardless of whether an individual believes so. The general epistemological observation is this: “Even if we cannot be wrong about what function something was supposed to have, we can certainly be wrong about what function something actually can have.” (Ruben 1989, 71) This could be taken to suggest the ontological claim that we are entitled to be realists about money as real because of the possibility of error: “the social sciences rely on a system of classification which is open for correction” – thus fallible (ibid.).

But things are not quite as simple as this. Two distinctions are required:

1. Classificatory beliefs on the part of one person or a few personsversus Classificatory beliefs on the part of everyone in a society

2. Singular classificatory belief [“Some specific item i is of social kind s”] versusGeneral classificatory belief [“There are things of social kind s”]

Singular classificatory beliefs – for example beliefs about particular money tokens – pose no threat to realism: those tokens exist since we may be in error about them. Indeed, everybody might be wrong about them. If one person can be mistaken in believing that a given item is a valid bank note, then everybody in the society can have that false belief. “Whether or not something is a dollar bill depends on its not being counterfeit, on where it was made and under what circumstances, and this is something that we may all get wrong. The dollar bill can be counterfeit, even though forever undetected as such.” (Ibid., 72)

General classificatory beliefs are more problematic. Ruben argues that if a general classificatory belief is held by one person or a few persons, it can indeed be wrong. But if a general classificatory belief is held by everyone in a society, it cannot be wrong. The generally held general belief cannot be mistaken since that belief creates the fact it is about, rather than being about any independent fact. “Here the intuition is that we reach a point at which generalized thinking really does make things so, that there is no distinction at this point to be drawn between a general illusion about the social world and the reality of such a world.” (Ibid.) This feature of general classificatory beliefs held by everyone in society implies anti-realism about the things those beliefs are about: such things don’t exist independently of beliefs about them but rather result from such beliefs.

Here is the example of money: “Suppose, in society s, everyone believes that there is money, some medium of exchange. What might these people be wrong about? They can be wrong concerning some particular item that it is a bona fide example of money. But what would it be like for them all to entertain a false general belief that there is money? They act just as if there is. They transfer title by handing over physical tokens of some sort. They regard those tokens in this way, just as if they were money. What more would have to be the case for there to be money? The plausible answer is: nothing more.” (Ibid., 72–73) Thus, generally held belief in the money universal ensures that money exists: it brings money into existence.

I am somewhat puzzled by the phrasing of this argument. It seems that the argument is not merely about general beliefs but also and importantly about singular beliefs. It talks about people “handing over physical tokens of some sort” and it says that people “regard those tokens in this way, just as if they were money”. Maybe the idea is that holding such singular beliefs amounts to holding the general belief that there is money? But this is not clear at all. What is the difference between everybody being “wrong concerning some particular item that it is a bona fide example of money” (supposed to be possible) and everybody being wrong in regarding “those tokens in this way, just as if they were money” (supposed to be impossible)? Ruben starts talking about people acting “just as if there is money” in general, but then he goes on talking about people using certain “physical tokens” “just as if they were money”, that is, as if they were instantiations of money. I have a difficulty seeing clearly what exactly the general classificatory belief here is. Of importance has to be a distinction between something like “this particular token” and “just any token”: the latter would somehow constitute general beliefs.

Be as it may, Ruben’s claim is that if everybody believes there is money, then there is money: it is not possible for everybody to be mistaken at the same time, thus money exists just if everyone believes so. But this means money is very straightforwardly dependent on human belief for its existence, indeed to the extent of removing the possibility of error, and this serves as an invitation to antirealism. I have two arguments that speak against, or at least aspire to qualify, such an account of social reality.

4.3 Beliefs of Social Actors and of Social Scientists

Even if money were completely dependent on people’s beliefs for its existence, there is a way in which it could exist objectively. This is a sense of existence that is relevant to scientific realism about economics and other social sciences. Scientific realism is usually defined in terms of something existing mind-independently, but this is not a very good idea when considering realism about the social sciences: social objects don’t so exist. Social objects are mind-dependent, thus a realism capable of accommodating the social sciences must take this into account. The solution is simple: the social world is dependent on the minds and beliefs of social actors, but it may be independent of the social sciences, including their theories and inquiries. Indeed, this suggestion can be made more strongly: the relevant notion of existence for scientific realism is not mind-independent existence, but science-independent existence (Mäki 2005). This creates room for scientific realism about money. This suggestion contrasts with Ruben’s argument that fails to make a distinction between beliefs and theories held by social agents on the one hand and those held by social scientists on the other. His argument is mainly in terms of the former, even though the title of his essay (“Realism in the social sciences”) refers to the social sciences.

An obvious objection to this solution would refer to the various ways in which the social sciences may contribute to the shaping of social reality – such as by way of policy advice, institutional design, and the so-called self-fulfilling and self-destructive prophecies.^[14] Indeed, we may grant that some currencies (such as the euro) might not be science-independent since economists have contributed to their design. However, admitting this does not imply that money as a social universal is not science-independent, thus scientific realism about money is not thereby threatened.

To fortify the case, we can make a further distinction between causal dependence and constitutive dependence. We may grant an extent of causal dependence between the social sciences and social reality: the contents of social scientific theories may shape the beliefs of social actors prompting them to take action that has consequences for social reality. At the same time, we may deny constitutive dependence: the theories and beliefs held by the social scientists do not constitute, create or give rise to the social world just by being held or uttered. Little more is required by a scientific realism about economics and other social sciences. (For more detailed discussions of these concepts and arguments, see Mäki 2002, 2005, 2008, 2012, 2013)

The argument can be formulated as the claim that even if the money universal were fully dependent on the collective beliefs by social agents, this would not imply a denial of scientific realism: money is not constitutively dependent on economic and other social scientific theories and inquiries. In short, regardless of whether or not some generic notion of realism applies to money, scientific realism has no trouble with it: money is in an appropriate sense science-independent.

4.4 In Virtue of the Pressures of a Complex Evolving System of Economic Institutions

Purely idealist accounts of social reality make that reality entirely dependent on people’s attitudes and acceptances. An institution or institutional fact exists just if people believe or collectively agree that it does. This means that if the relevant population of individuals were to stop believing that an institution exists, in consequence the institution would cease to exist. Thus, on the idealist conception, money exists because people believe or collectively agree that it does, and it would cease to exist if people were to stop believing that it does.

My second argument against purely idealist accounts remains at the level of collective belief amongst social actors, thus does not make appeal to the second-order beliefs of social scientists. My argument is based on a systemic perspective, while the purely idealist account tends to be separatist and to downplay the systemic character of social reality. The separatist account suggests a counterfactual speculation: if people were to lose their faith in money, money would cease to exist. But it is easy to see that this counterfactual speculation is indeed just that: mere imaginary speculation of a strongly counter-to-the-facts sort. It is conceivable that a currency would cease to exist in case people no more believed in its powers, in consequence of a collapse of an economy or its monetary system for example; or in consequence of an administered transformation such as from guilder to euro in The Netherlands. It is not similarly conceivable that there would be a separate loss of collective belief (or any other volitional attitude) in money, and that due to this loss of belief, money would cease to exist. This is because money is a deeply rooted essential component in a larger complex system of institutional and organizational structures and practices. There is no way money can be made to vanish separately just by way of collectively agreeing so: it is that larger complex system that would have to disappear in order for money to disappear. Money as a social universal is far more robust than currencies and specific monetary institutions.^[15],[16]

A qualification is needed. Any talk about collective belief must not exaggerate the cognitive powers of the collective and its members, and must have the targets of those beliefs right. Money is a case in point. As the old wisdom has it, people are typically able to use money without understanding what money is – without having the correct concept of it. The mystery of money is a tough enough challenge to theoretical social scientists (and to philosophers doing social ontology), and surely ordinary money users are in no better position on this matter. Therefore, I don’t want the above argument to be dependent on collective belief in money having to be based on having the correct concept of money. This would be too strong a requirement. Some weaker idea of collective acceptance is needed to meet this demand. But there is a limit beyond which the weakening of the notion of acceptance should not go. Suppose we take ‘acceptance of X’ to mean going along with X in one’s practices – rather than violating the practical requirements of the institution. Given this meaning of collective acceptance, it would become trivially true that were collective acceptance of money to be removed, money would cease to exist. So this notion would be too weak. The appropriate idea of collective belief or acceptance for our purposes must lie somewhere in between the above two extremes.

It may therefore be undeniable that money exists in virtue of people collectively believing in it (without necessarily having the right concept of it) or collectively accepting it (involving more than just blindly following the relevant practices). But it is equally obvious that money exists in virtue of the pressures of a complex and comprehensive system of monetary and other social institutions. Rather than thinking of voluntary collective belief somehow hanging in the air and serving as an autonomous source of institutional existence, it makes more sense to regard it as itself dependent on the system of institutional interconnections. Indeed, I suggest we consider the situation as a triple whose components are mutually dependent on one another. On this view, money, collective belief or acceptance, and the larger system of institutional dependencies mutually shape and condition one another. Within this triple, we find another manifestation of de re necessity in contemporary social reality: the three components mutually necessitate one another in that it is not possible to do without any one of them.

This shows how the case of money can be used to make a more general argument in social ontology. What I have emphasized is the systemicity of social reality, the synchronic and diachronic interlocking of institutions. This has deep implications regarding the way in which institutional reality exists. It is surely misleading to say it exists independently of people’s beliefs about it: it would not exist if people did not have beliefs about (at least parts of) it. It is equally misleading to say it exists just in virtue of people believing it does: it is too large, complex, non-transparent, inherently interdependent, and historically staggered to be reducible to collective belief. Institutional reality, money as a social universal included, is both dependent on and independent of any conceivable collective belief at any given time and space.