Footnotes on Epicycles

Suits not of cards, but of chess

Mon 30 Aug 2010 01:35 PM

Here is more about Bernard Suits' Grasshopper. It picks up where the post on RPGs and the post on Suit's definition of 'game' left off.

Recall that Suits defines a game as "the voluntary attempt to overcome unnecessary obstacles." As such, playing a game involves (1) pursuing a goal (2) while accepting constraints on what means can be used to attain it (3) such that the constraints are partly constitutive of the activity.

The first part of the this formula is what he calls the prelusory goal. It is essential that the goal can be characterized independently of the game itself, since the game is constituted in part by constraints on how players may achieve the goal. If the goal itself depends on the game, then the definition ends up being circular.

Last time I discussed the example of poker. Although Greg points out (in the comments) that my attempt to specify the prelusory goal of poker leaves out a lot, I think the details could be filled in without reference to the constitutive rules of poker.

In this post, I want to consider the game of chess. The goal is to put your opponent's king in checkmate. The way that the pieces move, the fact that players alternate moves, and so on are the constraints on how that goal may be achieved.

One might worry: Even if you get out a chess set and try to checkmate, you are helping yourself to the fact that these are chess pieces. A little horsie is only a knight because it makes the L-shaped knight move. The piece qua chess piece presumes the rules.

Suits considers such a worry and has the grasshopper say this:

Although it is necessary to refer to the rules of chess in describing checkmate, and also necessary, when playing chess, to obey the rules in seeking to achieve that state of affairs, the involvement of chess rules in the two cases should not obscure the fact that the uses to which the rules are put are quite different.

So his reply is that rules may be used as mere descriptions, and descriptive checkmate does not necessarily involve anyone's victory.

I think this misses the point. Of course accepting checkmate as a goal is different than describing it. For checkmate to be the prelusory goal, however, the game must involve limitations on how the goal can be achieved. If the way that the pieces move is presumed in specifying what checkmate is, then there are no further constraints that can constitute the game.

Perhaps Suits' notion of descriptive checkmate might still do the job, but it takes a bit more work. Separate the abstract structure of chess into instantaneous board descriptions and permissible transitions between positions. If we catalog all of the possible board positions, giving symbolic labels to the various pieces, some count as checkmate (for one side or the other) and the rest do not. The permissible transitions are constraints on how to get to checkmate board positions. By separating the state space from the kinematics of chess, we can specify a prelusory goal.

Of course, this way of thinking about chess is counterintuitive. When learning chess, one first learns how the pieces move and capture. Then one learns that checkmate is when a king is under attack and can't get out of it.

We end up saying that what makes chess a game is just the opposite. The goal is to have your opponent's king in checkmate, and the ways that the pieces move (as restrictions on achieving that goal) enter the picture so as to make it a game.

This entry has already gotten long, so I'll hold my one more thing to say about Suits for another day.

Comments

from: Ryan Hickerson

Tue 31 Aug 2010 03:28 AM

What's your one more thing to say about Suits? Is it an objection to the definition? ('cause I think I might've missed your actual objection somewhere in there...) Didn't you say, a post or two back, that his definition was too narrow? It seems to me that we can specify a prelusory goal for any game, if we are given the latitude of including lively diachronic processes and not merely boring states at particular temporal intervals. If I were Aristotle I would say that goals were what everything was about.

But the part of his definition that bothers me is the inefficiency part. It was fun for me to read your observation that the book doesn't really have anything to do with the philosophy of science, because my sense when reading it was just the opposite: this definition smells *too much* of the philosophy of science! It's like the old adage: "If your only tool is science, then everything will look to you like an efficiency problem!" Games become mysterious, until the philosopher of science lights upon the idea of defining them as the *opposite* of science! Because science is the discovery of the most efficient means to a possible end, games must be the opposite: an intentionally *inefficient* means to achieving the end!

Isn't that the backdrop for Suits' definition? Isn't science the discovery of the most expedient means to a possible end? (I could be behind the times, I suppose.)

p.s. I don't think games are, by definition, an inefficient means. Quite the contrary, game playing often leads to the discovery of the *most* efficient means! (I'm thinking of Ender's Game, at the moment... or beta testing...)

from: Greg

Tue 31 Aug 2010 05:18 AM

Why can't the prelusory goal of chess just be taking your opponent's king, and the word 'checkmate' not appear until we get to the constitutive rules?

(p.s. In my previous comment about poker, I just wanted to say that the prelusory goal of a single round should be winning the money in the pot, not having the highest hand at showdown. The other stuff was mostly thinking-out-loud fluff.)