The Chess Mind

Q. Is chess a sport?

A. It might at first seem that chess is a sport. First of all, it's clearly a competitive activity, which seems to be a necessary if not sufficient condition for something's being a sport. Second, the same sorts of general mental and physical disciplines needed by the sportsman (e.g. mental toughness, strong self-confidence, endurance, etc.) are required for chess players to succeed. To take a prominent example, Karpov's (then-) frail physique nearly cost him twice in big matches against Korchnoi (one for the world championship, the other in a final candidates match) and quite possibly did cost him the title to Kasparov when he lacked the endurance to finish him off in 1984.

Yet despite the above, I think that chess is not a sport. Here's why:

1. I take the following to be necessary conditions of being a sport:

a. That it's a competitive activity.
b. That the performance of the activity have an intrinsically physical component.

2. Chess fulfills (a) but not (b). As far as the nature of chess is concerned, it could be played by disembodied spirits using mental telepathy or by conscious computers.

(Whether either exists is a question for another time; I'm inclined to think the former do exist and to be skeptical about the possibility of the latter, and I'm sure some of my readers think I have it exactly backwards. No matter; the point here is just that either sort of being could play chess either without any physical activity whatsoever, or without the physical activity's being an intrinsic part of the fulfillment of the exercise.)

What I mean by an "intrinsically physical component" is easy to grasp by considering a paradigmatic case: in football, players score touchdowns by using their bodies to move the football across the field and into the end zone, field goals or extra points by sending the ball through the goal posts using only their feet. A physical object must be moved through physical space using particular bodily means.

Not so with chess. Moving the wood or plastic pieces isn't an intrinsic part of the game - one could play an online game by moving one' s mouse or better still, not move anything to play a blindfold game. (One has to move something to state one's move, but the expressing of a move isn't itself a move.) What counts is the production of a move, and that is not an intrinsically physical activity.

3. Thefore, chess isn't a sport.

Now, if one chooses to define a sport merely as some sort of competitive endeavor, then chess would be let in - but so would many other activities, like put-down contests and job interviews. Nor is it enough to add to the competitiveness condition the further requirement that it's an activity where physical prowess can make a substantial difference to one's potential success: one candidate for a job may succeed due to his enhanced fitness (his healthy appearance impressed the hiring committee, his superior conditioning enabled him to successfully work longer hours at his previous job, improving his qualifications, etc.), but that still wouldn't turn job interviewing into a sport.

In sum, while chess is in some significant ways sports-like, and physical and mental training are of great value to ambitious tournament chess players, chess is not a sport - at least if an activity only counts as a sport if it includes some intrinsically physical component.

A few months ago, the ever-active Boylston Chess Club blog has taken on Howard Goldowsky as a guest blogger, and in his opening post he challenges my earlier argument that chess is not a sport, based, appropriately enough, on his rejection of my definition of sport. On my view, it's a necessary condition of some activity's being a sport that it has an irreducibly physical component, while Goldowsky thinks it's a sufficient condition that its practitioner relies on timing and pattern recognition.

It's clear that the pattern recognition condition fits chess, but it seems to fit universally acceptted sports as well: in football offensive players need to recognize defensive alignments and defensive players various offensive schemes; in basketball, there's the pick-and-roll and the zone defense; in baseball, the batter looks for characteristic arm movements, leg kick and ball rotation patterns from the pitcher, and so on. And certainly timing is important in those sports as well, as the reader can readily confirm for him or herself.

[A brief aside: the meaning of "timing" is pretty clear when it comes to sports, but it may be equivocal when applied to chess. There isn't some physical movement requiring excellence in timing; rather, timing in chess has to do, broadly, with the way in which one attempts to execute some idea - with the order of moves. Thus it's a conceptual sense of timing rather than a physical sense, and one might think that it's the latter sense of timing rather than the former that's appropriate to an activity's being a sport. I'm congenial to this objection, but will let it pass for the remainder of this post.]

That's a bit of the positive case, but now let's turn to critique. Goldowsky says that timing and pattern recognition are sufficient - presumably he means jointly sufficient - for some activity's being a sport, which means that any activity requiring those two conditions will automatically be a sport.

So here's a very partial list of new sports:

(1) Driving. Of course auto racing is a sport, but Goldowsky's definition makes all ordinary driving a sport as well. Clearly there's pattern recognition involved - one learns how to negotiate the roadways without getting into accidents, and preferably without getting into situations in which accidents are reasonably likely to occur. And certainly driving involves timing, too; ergo, driving is a sport.

(2) Poetry. Language use involves tons of pattern recognition - indeed, words are patterns of a certain sort - and timing (including but not limited to meter) is involved too. So, poetry is a sport.

(3) Making music. Recognizing key structures, chord progressions and so on are all clear cases of pattern recognition, and the role of timing in music is obvious. Music is a sport!

Without elaborating the details, we can also include (4) walking, (5) cooking and (6) brushing one's teeth as sports, too. But clearly, I think, a definition of sport that includes (1)-(6) as instances is an overly liberal definition.

Perhaps Goldowsky's definition can be improved by adding some further conditions - a competition condition, for example. That would plausibly render (1), (4) and even (6) as sports, though even then I remain skeptical about (2), (3) and (5).

Even if this is waived for the sake of argument, I think there is another problem. Even if we suppose that it's sufficient for something's being a sport that it involve pattern recognition and timing, that won't show that chess is a sport. The reason is that it's possible for someone or some thing to play chess without recognizing any pattern at all; say, by using a purely brute force approach. God, for example, or some sort of idealized computer [assuming, as I don't, that a computer plays chess at all] is the sort of being who could figure chess out from move 1 through the end without recognizing any patterns at all (beyond those required to involve and interpret the rules of the game).

It might be objected that even if God or some super-powerful intellect could play without relying on pattern recognition, we humans can't. True enough, but I claimed that chess isn't a sport, because the game doesn't necessarily have the conditions required of a sport. Likewise, even if chess sometimes fits Goldowsky's definition, it doesn't always - doesn't necessarily.

In sum, Goldowsky's definition of a sport is too liberal, letting in many activities that clearly are not sports . The definition also fails to include chess per se. So I conclude that while chess has many characteristics of a sport - it's sports-like - it is nevertheless not a sport, strictly speaking.

Related Posts (on one page):

1. Why Chess Still Isn't a Sport
2. Chess Isn't a Sport, Revisited
3. Is Chess a Sport?

In a number of posts (here and here), I have both offered my own view as to why chess is not a sport (though I'm willing to acknowledge that it is like a sport) and have critiqued others' attempts to claim that it is a sport. Briefly, my reason for thinking it's not a sport is that, in my view, it's a necessary (but not sufficient) condition for an activity to count as a sport the activity must include some intrinsically physical component. Chess need not include such a component (if it's even conceptually possible for immaterial beings to play chess, then chess does not have an intrinsically physical component), however, and thus it's not a sport.

One can of course deny that a sport must include an intrinsically physical component, but it seems to me that that's how the term has traditionally been understood, and unless widespread usage shifts, I'd prefer to say that chess is sport-like, but not a sport.

Howard Goldowsky thinks otherwise, and in an initial post on the Boylston Chess Club's blog, he suggested that chess or any other activity is a sport if it meets two conditions, roughly:

G1: It involves pattern recognition.
G2: It involves timing. (Defined in a very broad sense.)

I suggested that if G1 & G2 are jointly sufficient, then virtually any activity counts as a sport - walking to the mailbox, reading, eating, writing poetry, etc. Surely those don't count as sports, but then if that's the basis for including chess as a sport, it's insufficient to the task.

Next, I proposed an improvement on Goldowsky's scheme that rules out some but not all of my counterexamples:

G3: It takes place in a competitive context.

That rules out walking to the mailbox and reading, but it's still insufficiently restrictive, to my mind: applying for a job or playing in a piano competition now count as sports, which seems clearly wrong.

Goldowsky has since replied, and he's not impressed. Why aren't these activities sports? I haven't really defended that claim, and so, if his definitions suggest that they're really sports, then by gum, they're sports, even if no one recognizes it!

Now, I'm not completely unsympathetic to this line of reasoning. Suppose, for example, that everyone agrees that it's wrong to kill innocent human beings and that a human being is defined a living organism which has the genetic code of a human being and is either a mature member of the species or will, ceteris paribus, develop into a mature member of the species. If everyone accepts such a definition, then if some large segment of the population also accepted the permissiblity of abortion, then it would be fair to criticize their beliefs: if the definition they accept implies that abortion is wrong, then their views are inconsistetnt, even if they don't recognize it.

Unfortunately, this defense doesn't help Goldowsky, because it presupposes agreement about the definition. If G1-G3 represented the mainstream understanding of the nature of sport, then that would be one thing, but his definition is contentious at best. [An aside: Goldowsky has included a fourth condition, that "[l]uck is not inherent in the rules of the competition", but this condition isn't relevant to the ensuing discussion.]

Let's be (very) generous and suppose that, a priori, his definition and the more mainstream definition are each exactly 50% likely to be true. What do we do to figure out which of the two better captures the concept of sport shared by English speakers? I think the answer is to look at examples. A putative definition of "sports" should do three things:

(1) Include in its extension all clear cases of sports.
(2) Exclude from its extension all clear cases of non-sports.
(3) Be such as to account for the vagueness of vague cases.

Does Goldowsky's definition achieve this? (1) isn't any problem at all, but I think it fails on (2). No one not in the grips of a theory - at least no one I'm aware of - considers applying for a job or composing poetry for a contest as sports. These activities are not sports by any common understanding, and since the meaning of ordinary words comes from usage, not fiat, that gives us reason to think Goldowsky's definition is flawed.

Thus Goldowsky has still not provided us with sufficient reason to label chess a sport. He can call it a sport, or he can stipulate that whenever he utters the word "sport" he means an activity featuring G1-G3 (which is to say, G3, as G1 & G2, as he defines them, seem to apply to any action).

But I will not follow suit.