1 Introduction

The notion of constitutive rules has been addressed by many prominent analytical philosophers, including Searle (1969: 33–41, 1995), Midgley (1959), von Wright (1977), and Rawls (1955). More recently, that notion has been applied quite frequently to specific topics in the philosophy of language (e.g. Lewis 1979; Williamson 1996) or in the philosophy of law (MacCormick and Weinberger 1986; Ruiter 1997; Marmor 2009). Constitutive rules are traditionally distinguished from regulative rules. What exactly are these two kinds of rules? Let me start with regulative rules, typically conceived as prescriptions or imperatives – they tell us what to do in certain circumstances. Usually the rules of etiquette are presented as a paradigm example of regulative rules (e.g. “One should use a spoon to eat one’s soup”). As Searle puts it “Regulative rules regulate a pre-existing activity, an activity whose existence is logically independent of the rules” (1969: 34). The possibility of eating a soup does not depend on any rules. Constitutive rules, traditionally opposed to regulative rules, are those that have a creative function – they make it possible to perform particular actions or to participate in a particular practice. Constitutive rules are metaphysically prior to the practices they create. Let me provide you with an example: suppose that two people are moving wooden pieces on a black and white board in such a way as is typical for chess players. But playing a game, in this case chess, is not a matter of, to use Searle’s words, “brute facts” – it is not merely a matter of moving wooden pieces, because the role of rules is absolutely critical:

In general, social behavior could be given the same specifications even if there were no rules of etiquette. But constitutive rules, such as those for games, provide the basis for specifications of behavior which could not be given in the absence of the rule. Of course, regulative rules often provide the basis for appraisals of behavior, e.g., “He was rude”, “He was immoral”, “He was polite”, and perhaps these appraisals could not be given unless backed up by some such rules. But appraisals are not specifications or descriptions as I am now using those phrases. “He voted for Willkie”, and “He hit a home run”, are specifications which could not be given without constitutive rules, but “He wore a tie at dinner”, “He held his fork in his right hand”, and “He sat down”, are all specifications which could be given whether or not any rules requiring ties at dinner or right-handed fork use, etc., existed at all (Searle 1969: 36).

It appears that the basic intuition concerning constitutive rules is quite appealing, but it is also a little vague, as Searle said, self-critically: “I am fairly confident about the distinction, but do not find it easy to clarify” (1969: 33). In my brief paper, I will not try to evaluate if constitutve rules capture the essence of all institutional or social phenomena (cf. Ruben 1997; Giddens 1984), I rather intend to shed some light on the issue as to how to define constitutive rules. Answering such critique is important task but we should take one step at the time. It is not controversial that games, such as chess or basketball or rugby, are paradigm examples of practices that are constituted by rules. I will examine the rules of such games in order to acquire some knowledge of constitutive rules in general.

2 The Formal Account of Constitutive Rules

The most influential account of constitutive rules was produced by Searle, who used two formulae in order to clarify what constitutive rules are:

  1. 1.

    “The creation of constitutive rules, as it were, creates the possibility of new forms of behavior”;

  2. 2.

    “constitutive rules often have the form: X counts as Y in context C” (1969: 35).

Let me begin with the second formula that somehow became the received view of constitutive rules. In such a traditional reading “constitutive rules often have the form: X counts as Y in context C” – typically X stands for some natural phenomenon, such as moving a round object into a basket, and Y gives some “institutional” description of X: in this case, scoring points. Obviously, not all constitutive rules take that form, but according to Searle, all constitutive rules are easily transferable to such a form, or to put this differently, it is somehow their “natural form” (1969: 36). Typical examples of constitutive rules might be, in the case of football, “Moving the ball into the space between the posts and the bar counts as scoring a goal” or in the case of chess “A situation in which one player’s king is checked, and there is no possibility of removing that threat, counts as checkmate”. In the orthodox formal account of constitutive rules “they are almost tautological in character, for what the ‘rule’ seems to offer is part of a definition of ‘checkmate’ or ‘touchdown’ (Searle 1969: 34)”. Constitutive rules provide a clue to the “translation” of brute facts into institutional ones - a purely natural phenomenon in the light of constitutive rules has an institutional interpretation of that phenomenon.

There are some doubts with the formal account of constitutive rules. Most importantly, one might wonder why we should accept the claim that constitutive rules have the form “X counts as Y in context C”. It appears that there are many rules of games that do not “fit naturally” (whatever that might mean) into this form – for instance the crucial rule of basketball that forbids taking three or more steps without bouncing the ball on the floor seems to be more of a prescription than anything else (cf. Conte 1988). Similarly, in case of football, the rule that tells us that kicking members of the opposite team is forbidden also appears to be a prescription. There is no clear reason to claim that such important rules of games do have the natural form of “X counts as Y in context C”. Secondly, it appears that Searle used the “X counts as Y in context C” formula as a useful tool, it had mostly mnemonic character. He explicitly writes: “‘X counts as Y in context C’: This is not intended as a formal criterion for distinguishing constitutive and regulative rules (Searle 1969: 36)”. It appears that in the whole debate there is too much focus on the form of constitutive rules, and too little on their function.

There are other accounts of constitutive rules which put emphasis on the form of constitutive rules. Famously, Ransdell (1971) argued that we should be aware that institutional terms have two aspects: “connotation” and “import”. The former is basically equivalent to a definition; it gives us a description of terms that need to be met in order for some institutional term to apply. For instance, it could be the specification of what counts as the ball used in basketball matches. The latter is what can and cannot be done with the object, for instance, that the ball can be touched with hands and not with legs during basketball matches. Ransdell claims that constitutive rules have the form “X counts as Z”, where Z stands for expression of the import of an institutional term.

Alternatively, Hindriks presented what can be called an “XYZ” account of constitutive rules. As he states, “XY or constitutive rules specify the conditions that have to be met within a particular context in order for something to have the relevant status” (2009: 261) and these rules should be combined with “status rules” which link institutional Y terms with Z that is specification of the normative consequences that Y has.

I am not going to discuss details of these accounts, rather I intend to challenge all the theories of constitutive rules that appeal to the form which constitutive rules are supposed to have. It is without question that rules of games are constitutive rules (if anything is constitutive rules, rules of games most certainly are). As Ransdell (1971: 387) correctly noted, it is not the case that rules of games are analogous to constitutive rules but that games are systems of constitutive rules. Hence, if one finds a single rule of any game in the world (basketball, chess, Gaelic football, cricket, bridge, water polo...) that does not fit into the form that a proponent of any formal account assumes it should, then it makes a genuine counterexample of such an account of constitutive rules. In such a case, the proponent of any formal account needs to “catch the bullet in her teeth and”, for example, make a convincing argument that such a rule, in fact, is having some other form than it appears to have. Any theory of constitutive rules that puts emphasis on their form is vulnerable to multiple counterexamples, and any opponent of such a theory and its proponent might easily end up playing cat and mouse.

3 The System Reading

The idea of how some alternative account of constitutive rules might be formulated could be taken from another passage in Speech Acts:

First, since constitutive rules come in systems, it may be the whole system which exemplifies this form and not individual rules within the system. Thus, though rule I of basketball – the game is played with five players to a side – does not lend itself to this form, acting in accordance with all or a sufficiently large subset of the rules does count as playing basketball. And secondly, within systems the phrase which is the Y term will not in general simply be a label. It will mark something that has consequences. Thus “offside”, “homerun”, “touchdown”, “checkmate” are not mere labels for the state of affairs that is specified by the X term, but they introduce further consequences, by way of, e.g., penalties, points, and winning and losing (Searle 1969: 36).

What might be crucial is the fact that constitutive rules are interconnected – and it appears that any formal account, including those of Hindriks and Ransdell, downplays that fact. Moreover, it appears that there are rules which, by themselves, have no normative “import” or the Z element. For instance, in the case of football, the rule that specifies what the goal (understood as an object) is, states that it is made of such and such materials, has certain shape and dimensions, yet that rule alone does not say what we can or cannot do with this object. That single rule must be accompanied by other rules - the ones that specify what the ball is and how it is permissible to use it (with legs or head, not with hands) and that it is possible to score a goal by putting the ball into the space between the bar and the posts but not possible to score by throwing in from the sideline or from an indirect free kick. The point is that the single rule may not have any “import”, and a single rule might be insufficient to establish any institution within the game; it must be accompanied by some other rules. It seems reasonable to think, therefore, that constitutive rules form a system, and within that system, the form of a single rule might be irrelevant to its constitutive nature.

Searle writes: “acting in accordance with all or a sufficiently large subset of the rules does count as playing basketball” (1969: 36). Some people might find that plausible, but I will try to show that the quantitative criterion of a “sufficiently large subset of the rules” is not satisfactory. First of all, when one adopts the quantitative criterion, one implies that we gradually stop playing the game in question with every rule we violate. But this is not the case, because there are many rules of games that could be violated, one after another, without taking even one step towards the resultant termination of the game. Suppose that the basketball team players perform the following consecutive actions: travelling, double dribble, lane violation, five-second violation, and carrying. During these five consecutive actions they have violated five different rules of basketball, but it is not the case that with every such violation they stopped playing basketball and started to do something else. They are just appallingly bad players. But if they started to play with their feet, rather than with their hands, it would mean that they were no longer playing basketball.

There is another challenge against the quantitative account as presented above (cf. Gluër and Pagin 1999: 221) that concerns the identity of games. Suppose that two teams are playing a game, for instance rugby. If playing rugby consists of acting in accordance with a sufficient number of the rules of rugby, then there is some minimal number n of rules that a team must act in accordance with, in order to play rugby. Now suppose that the players of one team are acting in accordance with n rules and one additional rule r1. The second team players act in accordance with n rules and an additional rule r2. Hence, the first team plays some variant, rugby1 (because they are acting in accordance with n + r1) and the second team plays at the same time another variant, rugby2 (because they are acting in accordance with n + r2). Two teams that are playing a rugby match with each other are in fact playing two different variants of rugby, and that is implausible.

4 Two Types of Constitutive Rules and Two Types of Games

First of all, I would like to note that games differ greatly from one another. Take a look at chess –. there are relatively few rules constituting the game. This is different in the case of popular team sports like football, rugby or basketball, which are much more complex games because they are composed of a large number of rules. It is worth noting that the official football rulebook is 212 pages long,Footnote 1 the rugby rulebook is 144 pages long,Footnote 2 and the basketball rulebook is 90 pages long.Footnote 3 The chess rulebook is just 25 pages long.Footnote 4

Chess is also unforgiving when it comes to the violation of its rules. It appears that violating the rules of chess, for instance moving a rook diagonally or moving a pawn backwards is something “outside the game”, or even automatically terminates the game.Footnote 5 Such a move is simply not a part of the game of chess. But things are much different in the case of complicated team games. It is possible to violate numerous rules of football, rugby or basketball without terminating the game (e.g. holding the opponent by her shirt is considered a foul, an “incorrect” move in football; obviously if you perform this action you do not simply stop playing football and start participating in some other activity). Hence we need to explain what the difference between these games is; why violation of the rules of the game results in such a variety of consequences, from just a fault (like travelling or offside) to termination of the whole activity (as in the case of moving a rook diagonally).

It seems reasonable to assume that constitutive rules could differ from one another (they can be prescriptions or definitions) and, what can be equally important, the system of constitutive rules of a game could have some internal structure. A system of constitutive rules could be composed of rules of different importance - some rules play a more significant role than others. Hence, constitutive rules that form a system that creates such a complex team game as football or basketball might be divided into two categories. Let me call these “central constitutive rules” and “peripheral constitutive rules”.

The rules of the first type are the essence of the game, they form a game’s hard core. For example, it is essential for football that we do not play it with our hands (if we are not a goalkeeper in the penalty area), but with our legs and head and we try to put the ball into a goal; it is essential for basketball that we play it using our hands and try to put the ball into the basket. One might ask what the central role of these rules consists of. The question is important because the claim that they are essential to the game might be intuitively quite appealing, but not particularly strict. What makes them central is the fact that violation of these types of rules is something “outside the game”; when we violate the central constitutive rule of a certain game we risk the termination of the game in question.

When I say that violation of central rules entails termination of one’s participation in practice I do not mean that a referee, when such an event occurs, necessarily excludes the player from the game or one team automatically loses the game. Participants in the game internalise rules that are central to the game and in fact violation of such type of rules occur extremely rarely, but if one violated central rules of the game she would face the shock from other participants in the practice. For instance, during an amateur match, players would face disbelief and criticism from their opponents (“Are you insane?!”), and after that players would probably go home or pretended that such a violation did not occur and went back to the moment just before the breach. In case of professional game, the spectators would probably start to boo and ask questions “What the hell is going on?”, “Is this a joke or maybe some crazy artist is making a provocation?!”. Summarising, the termination of the game or move outside the game means not necessarily a disqualification but also a situation when the game ends on the tacit understanding that “we are not playing the game anymore.”

Note that I am not saying that every single violation of central rule will necessarily result with the termination of the game, I am rather trying to say that central rules, as opposed to peripheral ones, have the game-termination potential. An isolated touch of the ball by hand can be accommodated within the game of football, and isolated kicking the ball during basketball match either - especially, but not exclusively, when these violations are unintentional. Hence, the violation of the central rule means that we terminate participation only when it is made with the relevant game-specific standard of persistence or excessiveness (such as when members of a football team start to grab the ball and throw it to each other).

Now let me move on to “peripheral constitutive rules”. These rules are also constitutive of certain practices, but they are not as important as the central ones. Violation of rules of this type is considered as a fault, an “illegal move” within the game. Note that the offside rule in case of football or 3 s in the key or 12 s rule in case of basketball do not have the same game-termination potential, as central rules mentioned above. One can be in an offside position or 3 s in the key multiple times during the match and it would not entail the conclusion that “this is not football/basketball anymore”.

At this point, one might wonder why peripheral rules are constitutive for a game if they could be, and in fact usually are, frequently violated. Are they not just an insignificant addition to the central rules? I believe the answer must be negative. Although many rules of team sports (football, basketball, ice hockey, rugby, whatever) are open to violation, they are still constitutive for that game because these rules, together with the central ones, determine the identity of the game. Suppose that there were no rules in the football rulebook to forbid holding on to members of the opposite team or knocking them to the ground. Would that kind of game still be football? It seems not: it would instead be something new and different, perhaps a kind of hybrid of football and rugby. Hence, central rules alone cannot create such complicated practices as football games. Peripheral rules are necessary in order to determine the identity of the practice. Their peripheral role consists of the fact that they could be violated without risk of termination of the practice.

So there are two types of constitutive rules - central and peripheral. These types of rules could form different types of systems that constitute a practice. The system of rules that is constitutive for such complex team games as basketball or football consists of many rules. The majority of them are peripheral rules; central rules appear to be in minority. But there could be other types of systems of constitutive rules, and therefore other types of games. The obvious example is chess. As already noted, chess is unforgiving when it comes to the violation of its rules, which means that the system of rules constitutive for chess is different from the system of rules constitutive of, for instance, basketball. It appears that there are exclusively (or almost exclusively) central constitutive rules for chess. The system of rules for chess is evidently more homogeneous than the system of rules for basketball or rugby.

I have distinguished two types of constitutive rules – central and peripheral – and now it is time to formally distinguish two types of game. Let me call them “rigid” and “flexible”. The games of the former type are those which are created through a system of rules that consists of only (or almost exclusively) central constitutive rules; the games of the latter type are those which are created by a system of rules that consists of a small number of central rules and many peripheral rules. As has been noted, chess is a paradigm example of a rigid game, and football and basketball are paradigm examples of flexible games.Footnote 6

Marmor has suggested a slightly different idea of constitutive rules. He emphasised the role of “values that the constitutive rules give rise to. Some of the values we associate with the activity of following the constitutive rules are such that they are more essential to the kind of practice that it is” (2009: 44). He states that the tennis dress code is not essential to the game, but any change made to the rules that determine the height of the net or the size of the ball will change aspects of the game that we do value or appreciate (2009: 44). Let me express some concerns.

Firstly, it looks as if that particular example might be a little unfortunate. Actually, there were some changes to the parameters of tennis balls in the past: notable modifications were made in 2002 when new types of balls were introduced. With that change, the differences between matches played on different surfaces have decreased, and there were some changes in the game, for instance, almost nobody plays regularly so-called “serve & volley” at Wimbledon where it was a dominant style of play before. Yet tennis before and after 2002 is not a completely different game, despite the fact that there have been some important changes in the game play (it became less offensive and more physical). Marmor could respond that the changes were not significant enough and if the tennis ball were replaced with a much bigger one (like the one that is used in volleyball), then the change would be significant. This is true, but leaves us with a rather vague criterion of distinguishing more essential rules from less important ones.

Secondly, since we can value many different things in a game, would not it mean that almost every rule is essential to the game in question? We value the teamwork, technical skills, tactical cleverness and mental strength, etc. of football players. The rules that give rise to all these things are almost all rules of football. The rules that determine the character of the outfit seem to be obvious candidates for being classified as not central. But this also could be a matter of controversy, because if the rules that determine outfit of football players were such that they oblige players to wear, for instance, something similar to medieval armour, would not that mean that the game is entirely different and rules give rise to very different set of things we value about the game (we start to value almost exclusively physical strength of the players)? If the answer is affirmative, then even the issue of outfits is central to the game on Marmor’s account. Since every rule (or, at best, almost every rule) seems to be central, then why bother distinguishing between central and non-central rules?

5 Beyond Constitutive Rules

We have distinguished two types of constitutive rules – central and peripheral – that can have multiple forms and, importantly, they are interconnected. These rules always form a system that could consist of (1) central rules (rigid games such as chess) or (2) central and peripheral rules (flexible games: rugby, football, basketball). But that is not the end of the story. Now consider the notion of a “tactical foul”, which is an intentional violation of the rules of a game in order to interrupt an opponent’s action. In the case of playing football, sometimes it is beneficial to foul an opponent by, for instance, holding her by her shirt, thus preventing the opponent’s team from scoring a goal. It is clear that playing “correctly” (that is, in accordance with all the rules of the game) is sometimes less efficient than playing “incorrectly” (i.e. violating the peripheral rule of the game in order to obtain a beneficial result). Hence there are rules of the game and there are “rules” for playing the game efficiently (e.g. that the defender should prevent the opponent’s team from scoring a goal by committing a foul). The rules that are constitutive of a particular practice should not be confused with “rules” that recommend how to behave in order to achieve a favourable result. The matter of playing efficiently is something logically different from the very possibility of playing the game. That is, it is not possible to participate in a practice without the system of rules constitutive for that practice, but it is perfectly conceivable that a football match could be played without “rules of efficiency” that tell us how to obtain a favourable result. These “rules” might advocate some types of moves (for instance, in the case of football, that a defender should kick a ball out of the defensive area rather to the sideline than to the centre of the pitch) or even encourage the violation of some peripheral rules, but they are merely recommendations about how to play in order to achieve success. These rules are instrumental, not constitutive. Without rules of that kind, it would be a much worse game but it would still be a football match. Perhaps children, or generally amateurs, play football without any sophisticated tactics or other “rules” of efficiency. The reason why their matches are so sloppy is not only the difference in skill between amateurs and professionals but also the lack of knowledge of “rules” of efficiency (cf. Roversi 2014).

But there is still something missing from that picture. As was noted by Schwyzer (1969), the rules alone are not sufficient to establish a practice for playing games. One might feel slightly confused at this point because it has been said that games are paradigm examples of practices constituted by rules. I uphold that claim, but there is something that should be added to the picture presented above. Let me start with an example. Suppose that two people are sitting at a table and they are moving wooden pieces on a black and white board in the exact way as is typical for chess players. But at the end of the “game” they look utterly terrified. Despite following all the rules of chess, these people are not playing chess, they are treating their actions as some sort of religious ritual that determines if their society will be saved or doomed by the gods (for example, when black “loses” it means flood and famine). Constitutive rules do determine what moves are allowed, necessary or forbidden, but they do not determine the socio-cultural sense that a society ascribes to their behaviour. The mere fact that participants in some sort of practice, in this case chess, act in accordance with its constitutive rules does not necessarily mean that they are participating in the game of chess. They could conceive of their activities not as playing a competitive game, but rather as a religious ritual that will determine their future or afterlife, or something like that.

As noted by Marmor (2006, 2007) or Lorini (2014) there is something that might be called deep conventionsFootnote 7 that underpin games (and also other practices such as theatre). What is meant by that is that in order to play the game the very concept of playing a competitive game must be shared among the members of a certain society. These deep conventions typically emerge as responses to some basic human social and psychological needs. Of course, different societies may play different games, but it is not so easy to imagine that some societies might not play any games, although Martians may not have similar psychological or sociological needs to our own and in effect it is possible that they do not have even the concept of playing a competitive game, and do not participate in any games. Although it is not so easy to specify precisely what people should know in order to be able to play a game, I will try to put forward at least a rough formulation of the deep convention of playing a competitive game. A person could be considered as knowingFootnote 8 the deep conventions that underpin any competitive game if they understand the following things:

  1. (1)

    games are established by rules – rules govern the practice and they specify what it takes to win or lose, what moves are forbidden, permitted and necessary, who is viewed as a participant etc.;

  2. (2)

    games are something different from ordinary life or other types of practices such as art or religious rituals (cf. Marmor 2007: 595);

  3. (3)

    the whole rivalry makes sense only when the rivals are in some specific respect similar (this is why there are weight categories in combat sports);

  4. (4)

    that games have certain aim - obtaining favourable result.Footnote 9

It seems that (1), (3) and (4) are pretty clear, but one might wonder what (2) means. Suppose that a member of a football team makes a tackle in order to take the ball away from her opponent but that the tackle is very poor and there are unfortunate consequences – the player of the opposite team is seriously injured and she needs to undergo surgery and long-term rehabilitation. Now ask yourself why the player who made the unfortunate tackle is not subject to criminal charges concerning damage to his opponent’s body. Why does the injured player have no possibility of pressing charges against the unfortunate defender for the reimbursement of rehabilitation costs? The answer is quite simple – because a practice such as a football game is detached from ordinary life, we normally have no difficulty in distinguishing moves within the game and things outside the game. The same applies to art: for instance, one can appear, without consequences, completely naked in front of a whole bunch of people in a theatrical production but not in ordinary life.

I am not saying that the account of the deep conventions of playing competitive games as presented above is meant to explain how games evolved historically. It is not the case that we somehow invented deep conventions and then started to play games. It is more likely that we started doing some activities for fun and then tried to formulate rules, and finally came to an abstract notion of playing a competitive game. What I am trying to say is that it is impossible to realise what is going on when someone who does not understand the relevant deep conventions observes a game. If a Martian that does not understand the deep conventions of playing a competitive game was a spectator at a boxing match she would be clueless as to what was going on. She would not know why we allow boxers to punch each other but do not allow other people to do the same. For us, people who do understand the deep conventions of playing competitive games, the different treatment of a boxing match and a street brawl is something obvious.

6 Conclusions

I have sketched out the most influential account of constitutive rules and some obvious problems it faces. An alternative version of a system reading of constitutive rules has been proposed. Thanks to adopting qualitative criterion for distinguishing central and peripheral constitutive rules I was able to explain how it is possible that sometimes violation of a rule terminates our participation in a given practice and why sometimes violation of a rule is just an “incorrect” move within the game. Moreover, that very account makes it possible to distinguish between rigid and flexible games that are determined by other types of systems of rules. Finally, it was made clear that constitutive rules should not be confused with deep conventions that make the emergence of games possible and with the “rules” of efficiency that tell us how to win.

Lastly, I would like to address the issue of how the framework elaborated in this paper could be applied to the non-games institutions. Let me take a property as an example. It appears that all “layers” of games also occur in the case of property. First, it is essential for property that an owner can use, consume, dispose and abandon things - this seems to be central rules that determine what property is and without rules allowing that kind of actions for the owner of a thing, there would be no property. Next, there are many special (peripheral) rules that determine very specific entitlements and rights (for instance, who is entitled to collect the fruit that falls from the tree that is a part of neighbour land, but its branches enter to our land), that can vary a lot in different countries and in time. There is also a deep convention; this is a social and cultural sense that we ascribe to property (for example, that it enables subject to satisfy her needs; it guarantees subject some material security and the possibility of becoming wealthy and developing herself, etc.) and lastly there are rules of efficiency that tell us, given certain circumstances, what to do in order to fulfil one’s aims concerning property (“If you want to get profit without much risk, you should invest in government bonds”). This is obviously very sketchy, but I believe that it shows framework elaborated within that article could be useful when discussing non-game institutions.