What Chess Can Teach Us About the Nature of Society
The 2021 FIDE World Chess Championship between the reigning champion Magnus Carlsen and his challenger Ian Nepomniachtchi has just started. As I’m writing, the players have drawn their first three games. Eleven games are remaining for one of them to make a difference. While I’m playing occasionally, I’m not a particularly good player so I’ll not dare make comments about the first games and the rest of the match. Instead, I shall briefly explain how we can use chess to learn about the nature of our social institutions.[1]
Chess is obviously a game. But at a more general level, it can be characterized as a social practice (“playing chess”) that consists in following a determined set of rules. All games are related to a social practice in this way, though most social practices do not consist of playing a game. I define a social practice as a set of activities corresponding to rule-following behaviors. In other words, participating in a social practice consists in following a determined set of rules. This set of rules is usually defined as an institution. In other words, an institution implements a social practice by establishing the rules that participants in the social practice are following qua-participants in the social practice. Given these definitions, the game of chess is an institution that implements the social practice of playing chess.
I’m not the first to use chess to reflect on the nature of society. The philosopher John Searle for instance has been using the example of chess to illustrate his distinction between constitutive rules and regulative rules. The latter only regulate a preexisting practice, in the sense that the practice exists (or could exist) independently of the specific regulative rules which apply to it. Traffic rules such as driving on the right side are thought to be examples of regulative rules. Constitutive rules quite the contrary are defining the social practice. To participate in the practice is to follow the constitutive rules. What is it to play chess? It consists in moving pieces on a board in such a way that, e.g., pieces called “bishops” only in diagonal, that a game ends when one of the kings is checkmated, and so on. Using a Wittgensteinian terminology, constitutive rules correspond to a form of life; if the rules change, the form of life also changes. A very similar idea has been developed for instance by David Lewis, though Lewis uses baseball rather than chess to make his point. Interestingly, Searle’s, Wittgenstein’s, and Lewis’s explicit or implicit accounts of constitutive rules have all been made first about language, which is arguably one of the core institutions of human societies. But it is easy to see how the concept can be used for any institution, as Searle’s writings on social ontology show.
The constitutive/regulative rules distinction has been criticized recently, however. The main point is that the distinction is merely linguistic rather than referring to an underlying characteristic of the social reality. In particular, it can be shown that any constitutive rule formulated using Searle’s famous formula “X counts as an Y in condition C” can be reframed as a “If…, then…” regulative rule. Whatever the validity and implications of this linguistic argument, it remains true that there is an ontological link between sets of rules corresponding to institutions and sets of activities corresponding to social practices. The exact nature of this link thus remains to be explored.
Rather than the controversial constitutive/regulative rules distinction, I propose to investigate another, less common distinction, between essential and peripheral rules. Intuitively, an essential rule Re of a social practice P is a rule that cannot be substituted by a rule Re’ (or simply eliminated) without altering the nature of P in such a way that P becomes P’.[2] A peripheral rule Rp does not have this property: it can change without affecting the nature of P. We can use chess to illustrate the distinction. Consider the following situation:
You’re meeting a friend to play a series of chess games. Before the start of the first game, you’re exchanging about the appropriate time limit. You would prefer to play blitz games (5 minutes per player), but your friend insists on playing rapid games (15 minutes per player). A few minutes later, you find an agreement on a limit of 10 minutes per player. As you’re about to start, you make however another suggestion: what about arranging randomly the pieces on the first row of each side of the board, i.e., playing “Fisher chess”? This time, your friend categorically rejects the proposal. As you insist, he gets up, exclaiming “this is not how we play chess!”, before leaving.
Though a bit caricatural, this example illustrates well the distinction between peripheral and essential rules in a social practice. The rule regarding time limit is regarded by the players as being amenable to compromises and concessions. None of the players has a definitive view about which time limit is appropriate for a chess game. Of course, each player may have a preference and there is no reason that their respective preferences converge. But they both acknowledge that, within the practice of chess, a range of time limits are possible. In this sense, the rule about the time limit is only peripheral. Matters are different regarding how pieces should be positioned at the start of the game. In the example, one of the players has a definitive view about this. It reflects the fact that, among the community of chess players, while Fisher chess is recognized as a proper variant, it does not tend to be regarded as belonging to chess stricto sensu.
This example highlights two or three things about the essential/peripheral rules distinction. First, the distinction is not absolute and rigid. One could argue about my illustration of the positioning of pieces as an essential rule. For some rules in a social practice, we may expect that it is indeterminate whether they are essential or peripheral. Moreover, the classification of a rule as essential or peripheral is time relative. Second, whether a rule is peripheral or essential is subjectively dependent on the participants’ attitude with respect to the practice. This is the most important point. There have been many attempts to characterize institutions objectively, for instance in terms of their functions in society.[3] While the search for a general classification of institutions (e.g., “marriage”, “slavery”, “racism”, “money”, …) may be valuable from a scientific point of view, it should not be forgotten that this classification remains a (scientific) image of a more complex reality. The point is that while it may make sense to label different historically situated social practices in the same way, we should not forget that these practices remain different under a range of significant aspects. What is common, for instance, between the 14th century marriage institution in Florence and the 21st century French “mariage pour tous”? Are they really belonging to the same institution? Most of the time, the answer to this kind of question will be indeterminate and not relevant.
My point is thus that what the essential rules of an institution are is settled by the participants in the corresponding social practice, not by social scientists using ideal types to order their thinking and discoveries. Of course, that does not mean we cannot account for the reason why participants at some specific moment consider that a rule is essential or peripheral. I would suggest in particular that this depends on the value that the participants attach to the practice as a whole. It may be valuable for a person to engage in a practice P for a series of reasons. Playing (competitive) chess may be valuable because it makes it possible to compare players’ specific abilities, or because of its inherent aesthetics. These reasons serve as justifications for engaging in the very practice. Some (but not all) of the rules that define the practice may be regarded as tightly related to these reasons. To take the aforementioned example, introducing randomness in the positioning of the pieces may be regarded as being against one of the main points of engaging in the practice of chess. On the other hand, the normative importance of a rule may also depend on the justification for following it within the practice. Fairness considerations may for instance have independent value, irrespective of the value of the practice as a whole.
The bottom line is that we cannot account for the nature of an institution and the related social practice without considering the participants’ attitudes toward the rules that are constitutive of it. Moreover, to understand these attitudes, we have to locate the social practice within the web of social institutions that constitute a society and the values that ground individuals’ reasons for actions. This is true for chess, as well as for any other social institution.
[1] The remaining of this essay is largely based on a paper I’ve published this year in the Journal of Social Ontology.
[2] This definition implies that the relationship between Re and P is a relation of supervenience. But it is stronger: if P evolves toward P’, that indeed implies that at least one of the Re has changed, as supervenience implies. However, on top of supervenience, we add the condition that if one the Re is changed into Re’, P becomes P’.
[3] This is the approach taken by Francesco Guala in his book Understanding Institutions. I’ve commented on this book here.