By BERNARD SUITS, University of Waterloo

By means of a critical examination of a number of theses as to the nature of game-playing, the following definition is advanced: To play a game is to engage in activity directed toward bringing about a specific state of affairs, using only means permitted by specific rules where the means permitted by the rules are more limited in scope than they would be in the absence of the rules, and where the sole reason for accepting such limitations is to make possible such activity.

Prompted by the current interest of social and behavioral scientists in games and encouraged by the modest belief that it is not demonstrably impossible for philosophers to say something of interest to scientists, I propose to formulate a definition of game playing.

1. Game-Playing as the Selection of Inefficient Means. Mindful of the ancient canon that the quest for knowledge obliges us to proceed from what is knowable to us to what is knowable in itself, I shall begin with the commonplace that playing games is different from working. Games, therefore, might be expected to be what work, in some salient respect, is not. Let me now baldly characterize work as “technical activity,” by which I mean activity in which an agent (as rational worker) seeks to employ the most efficient available means for reaching a desired goal. Since games, too, evidently have goals, and since means are evidently employed for their attainment, the possibility suggests itself that games differ from technical activities in that the means employed in games are not the most efficient. Let us say, then, that games are goal-directed activities in which inefficient means are intentionally (or rationally) chosen. For example, in racing games one voluntarily goes all around the track in an effort to arrive at the finish line instead of “sensibly” cutting straight across the infield.

The following consideration, however, seems to cast doubt on this proposal. The goal of a game, we may say, is winning the game. Let us take an example. In poker I am a winner if I have more money when I stop playing than I had when I started. But suppose that one of the other players, in the course of the game, repays me a debt of a hundred dollars, or suppose I hit another player on the head and take all of his money from him. Then, although I have not won a single hand all evening, am I nevertheless a winner? Clearly not, since I didn’t increase my money as a consequence of playing poker. In order to be a winner, a sign and product of which is, to be sure, the gaining of money, certain conditions must be met which are not met by the collection of a debt or by felonious assault. These conditions are the rules of poker, which tell us what we can and what we cannot do with the cards and the money. Winning at poker consists in increasing one’s money by using only those means permitted by the rules, although mere obedience to the rules does not by itself insure victory. Better and worse means are equally permitted by the rules. Thus in Draw Poker retaining an ace along with a pair and discarding the ace while retaining the pair are both permissible plays, although one is usually a better play then the other. The means for winning at poker, therefore, are limited, but not completely determined by, the rules. Attempting to win at poker may accordingly be described as attempting to gain money by using the most efficient means available, where only those means permitted by the rules are available. But if that is so, then playing poker is a technical activity as originally defined.

Still, this seems a strange conclusion. The belief that working and playing games are quite different things is very, widespread, yet we seem obliged to say that playing a game is just another job to be done as competently as possible. Before giving up the thesis that playing a game involves a sacrifice of efficiency, therefore, let us consider one more example. Suppose I make it my purpose to get a small round object into a hole in the ground as efficiently as possible. Placing it in the hole with my hands would be a natural means to adopt. But surely I would not take a stick with a piece of metal on one end of it, walk three or four hundred yards away from the hole, and then attempt to propel the ball into the hole with the stick. That would not be technically intelligent. But such an undertaking is an extremely popular game, and the foregoing way of describing it evidently shows how games differ from technical activities.

But of course it shows nothing of the kind. The end in golf is not correctly described as getting a ball into a hole in the ground, nor even, to be more precise, into several holes in a set order. It is to achieve that end with the smallest possible number of strokes. But strokes are certain types of swings with a golf club. Thus, if my end were simply to get a ball into a number of holes in the ground, I would not be likely to use a golf club in order to achieve it, nor would I stand at a considerable distance from each bole. But if my end were to get a ball into some holes with a golf club while standing at a considerable distance from each hole, why then I would certainly use a golf club and I would certainly take up such positions. Once committed to that end, moreover, I would strive to accomplish it as efficiently as possible. Surely no one would want to maintain that if I conducted myself with lesser efficiency in pursuit of this end I would not be playing a game, but that I would be playing a game just to the extent that I permitted my efforts to become sloppy. Nor is it the case that my use of a golf club is a less efficient way to achieve my end than would be the use of my hand. To refrain from using a golf club as a means of sinking a ball with a golf club is not more efficient because it is not possible. Inefficient selection of means, accordingly, does not seem to be a satisfactory account of game playing.

2. The Inseparability of Rules and Ends In Games. The objection advanced against the last thesis rests upon, and thus brings to light, consideration of the place of rules in games: they seem to stand in a peculiar relation to ends. The end in poker is not simply to gain money, nor in golf simply to get a ball into a hole, but to do these things in prescribed (or, perhaps more accurately, not to do them in prescribed) ways; that is, to do them only in accordance with rules. Rules in games thus seem to be in some sense inseparable from ends. To break a rule is to render impossible the attainment of an end. Thus, although you may receive the trophy by lying about your golf score, you have certainly not won the game. But in what we have called technical activity it is possible to gain an end by breaking a rule, for example, gaining a trophy by lying about your golf score. Whereas it is possible in a technical action to break a rule without destroying the original end of the action, in games the reverse appears to be the case. If the rules are broken the original end becomes impossible of attainment, since one cannot (really) win the game unless he plays it, and one cannot (really) play the game unless he obeys the rules of the game.

This may be illustrated by the following case; Professor Snooze has fallen asleep in the shade provided by some shrubbery in a secluded part of the campus. From a nearby walk I observe this. I also notice that the shrub under which he is reclining is a man-eating plant, and I judge from its behavior that it is about to eat the man Snooze. As I run across to him I see a sign which reads KEEP OFF THE GRASS. Without a qualm I ignore this prohibition and save Snooze’s life. Why did I make this (no doubt unconscious) decision? Because the value of saving Snooze’s life (or of saving a life) outweighed the value of obeying the prohibition against walking on the grass. Now the choices in a game appear to be radically unlike this choice. In a game I cannot disjoin the end, winning, from the rules in terms of which winning possesses its meaning. I of course can decide to cheat in order to gain the pot, but then I have changed my end from winning a game to gaining money. Thus, in deciding to save Snooze’s life, my purpose was not “to save Snooze while at the same time obeying the campus rules for pedestrians.” My purpose was to save Snooze’s life, and there were alternative ways in which this might have been accomplished. I could, for example, have remained on the sidewalk and shouted to Snooze in an effort to awaken him. But precious minutes might have been lost, and in any case Snooze, although he tries to hide it, is nearly stone deaf. There are evidently two distinct ends at issue in the Snooze episode: saving Snooze and obeying a rule, out of respect either for the law or for the lawn. And I can achieve either of these ends without at the same time achieving the other. But, in a game, the end and the rules do not admit of such disjunction. It is impossible for me to win the game and at the same time to break one of its rules. I do not have open to me the alternatives of winning the game honestly and winning the game by cheating, since in the latter case I would not be playing the game at all and thus could not, a fortiori, win it.

Now if the Snooze episode is treated as an action which has one, and only one, end- (Saving Snooze) ampersand (Keeping off the grass)-it can be argued that the action has become, just by virtue of that fact, a game. Since there would be no independent alternatives, there would be no choice to be made; to achieve one part of the end without achieving the other part would be to fail utterly. On such an interpretation of the episode suppose I am congratulated by a grateful faculty for my timely intervention. A perfectly appropriate response would be: “I don’t deserve your praise. True, I saved Snooze, but since I walked on the grass it doesn’t count,” just as though I were to admit to kicking the ball into the cup on the fifth green. Or again, on this interpretation, I would originally have conceived the problem in a quite different way: “Let me see if I can save Snooze without walking on the grass.” One can then imagine my running as fast as I can (but taking no illegal shortcuts) to the Athletic Building, where I request (and meticulously sign out for) a pole vaulter’s pole with which I hope legally to prod Snooze into wakefulness, whereupon I hurry back to Snooze to find him disappearing into the plant. “Well,” I remark, not without complacency. “I didn’t win, but at least I played the game.”

It must be pointed out, however, that this example is seriously misleading. Saving a life and keeping off the grass are, as values, hardly on the same footing. It seems likely that the Snooze episode appears to support the contention at issue (that games differ from technical actions because of the inseparability of rules and ends in the former) only because of the relative triviality of one of the alternatives. This pe-culiarity of the example can be corrected by supposing that when I decide to obey the rule to keep off the grass, my reason for doing so is that I am a kind of demented Kantian, and thus regard myself to be bound by the most weighty philosophical considerations to honor all laws with equal respect. So regarded, my maddeningly proper efforts to save a life would not appear ludicrous but would constitute moral drama of the highest order. But since the reader may not be a demented Kantian, a less fanciful though logically identical example may be cited.

Let us suppose the life of Snooze to be threatened not by a man-eating plant but by Professor Threat, who is found approaching the snoozing Snooze with the obvious intention of murdering him. Again I want to save Snooze’s life, but I cannot do so (let us say) without killing Threat. However, there is a rule to which I am very strongly committed which forbids me to take another human life. Thus, although (as it happens) I could easily kill Threat from where I stand (with a loaded and cocked pistol I happen to have in my hand), I decide to try to save Snooze by other means, just because of my wish to obey the rule which forbids killing. I therefore run toward Threat with the intention of wrestling the weapon from his hand. I am too late and he murders Snooze. This seems to be a clear case of an action having a conjunctive end of the kind under consideration, but one which we are not at all inclined to call a game. My end, that is to say, was not simply to save the life of Snooze just as in golf it is not simply to get the ball into the hole, but to save his life without breaking a certain rule. I want to put the ball into the hole fairly and I want to save Snooze morally. Moral rules are perhaps generally regarded as figuring in human conduct in just this fashion. Morality says that if something can be done only immorally it ought not to be done at all. What profiteth it a man, etc. The inseparability of rules and ends does not, therefore, seem to be a completely distinctive characteristic of games.

3. Game Rules as not Ultimately Binding. It should be noticed that the foregoing criticism requires only a partial rejection of the proposal at issue. Even though the attack shows that not all things which correspond to the formula are games. It may still be the case that all games correspond to the formula. This suggests that we ought not to reject the proposal, but that we ought first to try to limit its scope by adding to it an adequate differentiating principle. Such a differentiation might be provided by noticing a striking difference between the two Snooze episodes. The efforts to save Snooze from the man-eating plant without walking on the grass appeared to be a game because saving grass strikes us as a trifling consideration when compared with saving a life. But in the second episode, where KEEP OFF THE GRASS is replaced by THOU SHALT NOT KILL, the situation is quite different. The difference may be put in the following way. The rule to keep off the grass is not an ultimate command, but the rule to refrain from killing is. This suggests that, in addition to being the kind of activity in which rules are inseparable from ends, games are also the kind of activity in which commitment to these rules is never ultimate. For the person playing the game there is always the possibility of there being a non-game rule to which the game rule may be subordinated. The second Snooze episode is not a game, therefore, because the rule to which the rescuer adheres, even to the extent of sacrificing Snooze for its sake, is, for him, an ultimate rule. Rules are lines that we draw, but in games the lines are always drawn short of a final end or a paramount command. Let us say, then, that a game is an activity in which observance of rules is part of the end of the activity, and, where such rules are non-ultimate; that is, where other rules can always supersede the game rules: that is, where the player can always stop playing the game.

However, consider the following counter-example. Suppose an auto racer. During a race a child crawls out on the track directly in the path of his car. The only way that he can avoid running over the child is to turn off the track and by breaking a rule disqualify himself. He chooses to run over the child, because for him there are no rules of higher priority than the rules of the game. I submit that we ought not for this reason, to deny that he is playing a game. It no doubt strikes us as inappropriate to say that a person who would do such a thing is (only) playing. But the point is that the driver is not playing in an unqualified sense, he is playing a game. And he is evidently playing it more wholeheartedly than the ordinary driver is prepared to play it. From his point of view a racer who turned aside instead of running over the child would have been playing at racing; that is, he would not have been a dedicated player. But it would be paradoxical indeed if supreme dedication to an activity somehow vitiated the activity. We do not say that a man isn’t really digging a ditch because his whole heart is not in it.

However, the rejoinder may be made that, to the contrary, that is just the mark of a game: it, unlike digging ditches, is just the kind of thing which cannot command ultimate loyalty. That, it may be contended, is just the force of the proposal about games under consideration. And in support of this contention it might be pointed out that it is generally acknowledged that games are in some sense essentially non-serious. We must therefore ask in what sense games are, and in what sense they are not, serious. What is believed when it is believed that games are not serious? Not, certainly, that the players of games always take a very light-hearted view of what they are doing. A bridge player who played his cards randomly might justly be accused of failing to take the game seriously; indeed, of failing to play the game at all just because of his failure to take it seriously. It is much more likely that the belief that games are not serious means what the proposal under consideration implies: that there is always something in the life of a player of a game more important than playing the game, or that a game is the kind of thing that a player could always have reason to stop playing. It is this belief which I would like to question.

Let us consider a golfer, George, so devoted to golf that its pursuit has led him to neglect, to the point of destitution, his wife and six children. Furthermore, although George is aware of the consequences of his mania, he does not regard his family’s plight as a good reason for changing his conduct. An advocate of the view that games are not serious might submit George’s case as evidence for that view. Since George evidently regards nothing in his life to be more important than golf, golf has, for George, ceased to be a game. And this argument would seem to be supported by the complaint of George’s wife that golf is for George no longer a game, but a way of life.

But we need not permit George’s wife’s observation to go unchallenged. The cor-rectness of saying that golf for George is no longer merely a form of recreation may be granted. But to argue that George’s golf playing is for that reason not a game is to assume the very point at issue, which is whether a game can be of supreme impor-tance to anyone. Golf, to be sure, is taking over the whole of George’s life. But it is, after all, the game which is taking over his life, and not something else. Indeed, if it were not a game which had led George to neglect his duties, his wife might not be nearly as outraged as she is; if, for example, it had been good works, or the attempt to formulate a definition game playing, she would no doubt still deplore rich extra-domestic preoccupation, but to be kept in rags because of a game must strike her as an altogether different order of deprivation.

Supreme dedication to a game, as in the cases of the auto racer and George, may be repugnant to nearly everyone’s moral sense. That may be granted; indeed, insisted upon, since our loathing is excited by the very fact that it is a game which has usurped the place of ends we regards as so much more worthy of pursuit. Thus, although such behavior may tell us a good deal about such players of games, I sub-mit that it tells us nothing about the games they play. I believe that these observa-tions are sufficient to discredit the thesis that game rules cannot be ultimately binding on game players.

4. Means, Rather than Rules, as Non-Ultimate. I want to agree, however, with the general contention that in games there is something which is significantly non-ultimate, that there is a crucial limitation. But I would like to suggest that it is not the rules which suffer such limitations. Non-ultimacy evidently attaches to games at a quite different point. It is not that the rules which govern a game must be short of ultimate commands, but that the means which the rules permit must be short of ultimate utilities. If a high-jumper, for example, failed to complete his jump because he saw that the bar was located at the edge of a precipice, this would no doubt show that jumping over the bar was not the over-riding interest of his life. But it would not be his refusal to jump to his death which would reveal his conduct to be a game; it would be his refusal to use something like a ladder or a catapult in the attempt. The same is true of the dedicated auto racer. A readiness to lose the race rather than kill a child is not what makes the race a game; it is the refusal to, inter alia, cut across the infield in order to get ahead of the other contestants. There is, therefore, a sense in which games may be said to be non-serious. One could intelligibly say of the high jumper who rejects ladders and catapults that he is not serious about getting to the other side of the barrier. But one would also want to point out that he could be deadly serious about getting to the other side of the barrier without such aids; that is, about high-jumping. But whether games as such are less serious than other things would seem to be a question which cannot be answered solely by an investigation of games.

Consider a third variant of Snooze’s death. In the face of Threat’s threat to murder Snooze, I come to the following decision. I choose to limit myself to non-lethal means in order to save Snooze even though lethal means are available to me and I do not regard myself to be bound by any rule which forbids killing. (The author has argued for the possibility that life itself is a game In “Is Life a Game We Are Playing?” Ethics. Vol. 77, No. 3, April 1967). In auto racing, for example, the infield would not be filled with land mines. And I make this decision even though it may turn out that the proscribed means are necessary to save Snooze. I thus make my end not simply saving Snooze’s life, but saving Snooze’s life without killing Threat, even though there appears to be no reason for restricting myself in this way.

One might then ask how such behavior can be accounted for. And one answer might be that it is unaccountable, that it is simply arbitrary. However, the decision to draw an arbitrary line with respect to permissible means need not itself be an arbitrary decision. The decision to be arbitrary may have a purpose, and the purpose may be to play a game. And it seems to be the case that the lines drawn in games are not actually arbitrary at all. For not only that the lines are drawn, but also where they are drawn, has important consequences not only for the type, but also for the quality, of the game to be played. It might be said that drawing such lines skillfully (and therefore not arbitrarily) is the very essence of the gamewright’s craft. The gamewright must avoid two extremes. If he draws his lines too loosely the game will be dull because winning will be too easy. As looseness is increased to the point of utter laxity the game disappears altogether, since there are then no rules proscribing available means. Thus a homing propellant device could be devised which would insure a golfer a hole in one every time he played. On the other hand, rules are lines that can be drawn too tight, so that the game becomes too difficult. And if a line is drawn very tight indeed the game is squeezed out of existence. Suppose a game in which the goal is to cross a finish line. One of the rules requires the contestant to stay on the track, while another rule requires that the finish line be located at a position such that it is impossible to cross it without leaving the track. The present proposal, therefore, is that games are activities in which rules are inseparable from ends (in the sense agreed to earlier), but with the added qualification that the means permitted by the rules are smaller in scope than they would be in the absence of the rules.

5. Rules are Accepted for the Sake of the Activity They Make Possible. Still, even if it is true that the function of rules in games is to restrict the permissible means to an end, it does not seem that this is in itself sufficient to exclude things which are not games. When I failed in my attempt to save Snooze’s life because of my unwillingness to commit the immoral act of taking a life, the rule against killing functioned to restrict the means I would employ in my efforts to reach a desired end. What then distinguishes the case of the high jumper and of the auto racer from my efforts to save Snooze morally, or the efforts of a politician to get elected without lying? The answer lies in the reason for obeying rules in the two types of case. In games I obey the rules just because such obedience is a necessary condition for my engaging in the activity such obedience makes possible. But in other activities —e.g., in moral actions— there is always another reason, what might be called an external reason, for conforming to the rule in question; for a moral teleologist, because its violation would vitiate some other end, for a deontologist because the rule is somehow binding in itself. In morals conformity to rules makes the action right, but in games it makes the action.

Further to illustrate this point, two other ways in which rules function may be contrasted with the way in which rules function in games. Rules can be directives to attain a given end (If you want to improve your drive, keep your eye on the ball), or they can be restrictions on the means to be chosen to a given end (Do not lie to the public in order to get them to vote for you). In the latter way morals, for example, often appear as limiting conditions in a technical activity, although a supervening technical activity can also effect the same limitation (If you want to get to the airport in time, drive fast, but if you want to drive safely, don’t drive too fast). Consider a ruled sheet of paper. I conform to these rules, when writing, in order to write straight. Now suppose that the rules are not lines on a sheet of paper, but paper walls which form a labyrinth, and while I wish to be out of the labyrinth, I don’t wish to damage the walls. The walls are limiting conditions on my coming, to be out. Returning to games, consider a third case. Again I am in labyrinth, but now my purpose is not to be outside (as it might be if Ariadne were waiting for me), but to get out of the labyrinth, so to speak, labyrinthically. What is the status of the walls? Clearly they’ re not means for my coming to be outside the labyrinth because it is not my purpose to (simply) be outside. And if a friend suddenly appeared overhead in a helicopter I would decline the offer of a Lift although I would accept it in the second case. My purpose is to get out of the Labyrinth only by accepting the conditions it imposes. Nor is this like the first case. There I was not interested in seeing whether I could write a sentence without breaking a rule (crossing a line), but in using the rules so that I could write straight.

We may therefore say that games consist in acting in accordance with rules which limit the permissible means to a sought end, and where the rules are obeyed just so that such activity can take place.

6. Winning Is Not the End with Respect to which Rules Limit Means. There is, however, a final difficulty. On the one hand, to describe rules as operating more or less permissively with respect to means seems to conform to the ways in which we invent or revise games. But on the other band it does not seem to make sense at all to say that in games there are means for attaining one’s end over and above the means permitted by the rules. Consider chess. The end sought by chess players is, it would seem, to win. But winning means putting a chess piece on a square in accordance with the rules of chess. But since to break a rule is to fail to attain that end, what other means are available? It was for just this reason that the first proposal was rejected: using a golf club in order to play golf is not a less efficient, and thus alternative, means for seeking the end in question; it is a (logically) indispensable means.

The objection can be met, I believe, by pointing out that there is an end in chess analytically distinct from winning as an end. Let us begin again, therefore, from a somewhat different point of view and say that the end in chess is, in a very restricted sense, to place one of your pieces on the board in a position such that the opponent’s king is in terms of the rules of chess, immobilized. Now, without going outside the game of chess we may say that the means for bringing about this state of affairs consist in moving the chess pieces. The rules of chess, of course, state how the pieces may be moved; they distinguish between legal and illegal moves. Since the knight for example, is permitted to move in only a highly restricted manner, it is clear that the permitted means for moving the knight are of less scope than the possible means for moving him. It should not be objected at this point that other means for moving the knight —e.g., along the diagonals— are not really possible on the grounds that such use of the knight would break a rule and thus not be a means to winning. For the present point is not that such use of the knight would be a means to winning, but that it would be a possible (though not permissible) way in which to move the knight so that he would, for example, come to occupy a square such that, according to the rules of chess, the king would be immobilized. A person who made such a move would not, of course, be playing chess. Perhaps be would be cheating at chess. By the same token I would not be playing a game if I abandoned my arbitrary decision not to kill Threat while at the same time attempting to save Snooze. Chess, as well as my third effort to save Snooze’s life, are games because of an “arbitrary” restriction of means permitted in pursuit of an end.

The chief point is that the end here in question is not the end of winning the game. There must be an end distinct from winning because it is the restriction of means to this other end which makes winning possible, and also defines, in any given game, what it means to win. In defining a game we shall therefore have to take into account these two ends and, as we shall see in a moment, a third end as well. First there is what might be called the end which consist, in a certain state of affairs: a juxtaposition of pieces on a board, saving a friend’s life, crossing a finish line. Then, when a restriction of means for attaining this end is made with the introduction of rules, we have a second end, winning. Finally, with the stipulation of what it means to win, a third end emerges: the activity of trying to win, that is playing the game. It is noteworthy that in some cases it is possible to pursue one of these ends without pursuing the others and that in some cases it is not. Thus, it is possible to pursue the end of getting as many tricks at bridge as you can without pursuing the end of winning, since you may seek this goal, and also achieve it, by cheating. But it is impossible to seek to win without seeking to take a certain (relative) number of tricks, nor is it possible to seek to play without seeking both of the other ends.

7. The Definition. My conclusion is that to play a game is to engage in activity directed toward bringing about a specific state of affairs, using only means permitted by specific rules, where the means permitted by the rules are more limited in scope than they would be in the absence of the rules, and where the sole reason for accepting such limitation is to make possible such activity.

PHILOSOPHY OF SCIENCE, What is A Game?

1967, Vol. 34, Page 148, Reprinted by permission