Thinking en Passant

I read an interview of a chess hustler in New York expounding on his personal philosophy. It was certainly an interesting interview with an unusual character. His theories struck me with a few thoughts. But perhaps a disclaimer and aside first.

I sucked at chess for a child of my intelligence. I used to play against the computer in primary school. I knew the rules, sure, and the goal of the game as well, but I could never figure out how to string them together. My eyes saw material and nothing else, even though the software had fancy features to help you track legal moves, fields of influence, and so on. I saw only the surface level, playing it like a Democritean atomist (I nearly said reductionist, but that would be false because I saw nothing to reduce). Barren of abstractions, the chessboard is a mere particle accelerator, one governed by a physics which permits only annihilation, not transmutation (except the occasional promoted pawn).

No one, not even the most advanced supercomputers, can play chess by calculating all the possibilities that lie ahead - there are simply too many possibilities. So the only way to play is by applying sets of heuristic rules, of which mimicking successful moves from a game database is one. Chess then becomes a clash not between two players, but between the rule systems which they choose to enact. Of course, players choose when to switch between rules and between rule systems even, but these choices too are governed by metarules, and the rules and metarules can be subsumed under greater rule systems. The rules also dictate a player's gaze, because observation is theory-laden. It is tempting to think of these rules as being the evolving products of an incomplete science, renewed every so often by paradigm shifts. But why not view the rule systems as expressions of aesthetic sensibilities, with players trying to sculpt the board into forms of their desire (as Marcel Duchamp (he of the urinal) said, not all artists are chessplayers, but all chessplayers are artists)? Or as religions, where the right articles of faith are rewarded after the jihad by some Divine Chessplayer? (This may be, according to personal preference, a deity of war, or love, or politics, or wisdom, or whatever.)

How might I play the game now, having been exposed to strategy in other domains, chiefly that of football? Perhaps I would overload the middle of the pitch and form two banks of four to prevent direct assault on my keeper. Soak up pressure, then play a quick diagonal and counter at pace along the flanks. Play around the defenders with neat little triangles, or simply go through them with individual brilliance. Move into space and pass through gaps in the defensive lines. Strike swiftly and surely from distance, giving no chance to react. As a South American commentator would shout, MAAAAA~hold for 20 seconds~AAAATE!!!

Or I could go back to first principles, building a system from ground-up like what Isaiah Berlin called a Hedgehog. I could side with Heraclitus, who declared that everything is Change. Following George S. Patton, I then claim that the object of chess is not to change for your king but to make the other bastard change for his. Bruce Lee's voice floats into my head, saying "be like water, my friend", guiding me to go with the change. Ross Ashby would remind me that to make a system change as I want it, I have to be capable of an even greater degree of change. So I could begin by pitching loosely connected groups of pieces against my opponent's pieces, as long as my squadron has more degrees of freedom than that of the opponent.

But we quickly run into problems. There is only so much space to manouevre on the chessboard before my pieces run into other pieces and start interacting, sorta like quantum decoherence. The cybernetic way of playing is only practicable at the limit of low piece counts.

What alternatives remain? Do we view the chessboard using actor-network theory, where each piece is an actor, and the active pieces are mediators passing on quasi-objects that enable the continuation of a piece's function? Do we see the game as a Markov process, where the current game state is all that is required to determine future game states? Or take the parallax view with structural realism, and consider the moves to be more real than the pieces?

All these analyses ignore something crucial. Graham Harman argued that Latour's actor-network theory fails to capture the totality of objects, missing out for instance the micro-organisms on the chess piece, but the most important organisms are surely the macro- ones sitting either side of the board. To adapt Sartre, everything in chess is complicated by the presence of the other player.

Daniel Dennett posited that we must adopt the intentional stance towards our opponent when playing chess, even against computers. This brings us to the realm of game theory. Chess is a zero-sum game, so whatever weakens the opponent also strengthens your position. There is only one goal, which is checkmate. Zermelo's theorem states that since both players have perfect information about the game and chance plays no part, either the game ends in a draw unless someone makes a mistake (think Tic-Tac-Toe), or one of the sides has a winning strategy and can only lose or draw by deviating from it.

The computational intractability of chess prevents us from knowing what type of game chess is, as well as what these unstoppable strategies are if they exist. Your rule system helps restrict your search space to a few moves, and attributing certain beliefs to your opponent allows you to predict their possible responses and hence evaluate the efficacy of each move you might make. Iterating this a few times may improve your choice, but even grandmasters don't look ahead by more than a few moves. So there is nothing to do but to make your move and see what happens.

Of course, knowing that there is another human across the board, you could try to get into his head and interfere with his playing. The most obvious way to do this would be to knock out his prefrontal cortex, as players may attempt to do in chess boxing, but would clearly violate the spirit of the game in regular chess. It would also lose you all willing opponents in the future and be very detrimental to your long-term chessplaying career.

So, back to the article. Harry's worldview divides life into three stages like chess, the opening, the midgame, and the endgame, each requiring philosophy, mathematics, and science respectively. He then says that each piece represents something. Rooks are security, knights mobility, pawns wealth, queens power, bishops heart, and kings the self.

Is not the choice of applying mathematics in the midgame and science in the endgame decided by the philosophy chosen at the beginning? Are there not other valid pathways, such as starting with science to learn about the world (he is vague about what he means by science but I shall interpret it as empiricism), following with a guiding philosophy, then doing the math? What of other possibilities, like philosophy, math, then architecture (building something), or science, philosophy, then medicine (healing self and others)? Philosophy seems to be the common denominator in our pathways here, but it needn't be so. As Bertrand Russell noted, most people would rather die than think, and most do. Any philosophy that they might follow is merely inherited from society, a philosophical baggage that weighs them down without them even realising what they are carrying.

There is something intuitively satisfying about his use of chess pieces to read into a person's psychology. After all, these representations cohere with our cultural connotations about mediaeval archetypes, and also with our folk notion that people's choices reflect their personalities. But don't other interpretations of these pieces exist? For example, pawns could be endeavour because they advance step by step, knights creativity because they don't go straight, bishops dissent because they only go diagonally, rooks authority because they govern everything in line with them, and queens freedom because they can go almost anywhere. It might be quite fun to come up with some of your own. In any case, is there anything to make this more of a science than graphology?

The point is not to argue against his views, but to think about how he arrived at them. Our activities and environments provide us with metaphors for thinking about things. Whatever the metaphysics you wish to adopt, there are perceived isomorphisms between things, which lead us to use something as a metaphor for another. Being immersed in certain environments leads one to perceive certain things and certain isomorphisms, and hence apply certain metaphors. The more complex a system is, the more features it has and hence the more possible isomorphisms it can have with another system. It should come as no surprise that Harry chose chess as a metaphor for life.

Of course, we may be getting things in the wrong order here. Chess was probably designed in the first place as a model of war, which also happens to be one of the most common metaphors that people use for life, along with journeys. By the approximate transitivity of metaphors, chess could also serve as a metaphor for life. The burden of acting as a metaphor for life falls on chess because of the anthropomorphism of its pieces. Nonetheless, chess is a useful toy ontology. It models many scenarios and instantiates many dynamics, and its elements themselves provide us with metaphors.

Ian Bogost's idea of the procedural rhetoric is that the rules of a game construct arguments. The constitutional rules of chess were indeed used to promote a feudal social hierarchy, with the commoners and laity beneath the aristocracy and bishopry, and thus the first to die in battle, with everything expendable for the sake of the king.

What might chess persuade us of today? Should we replace the king with the president or CEO? Certainly Socialist chess, in which all pieces move equally, would not be very fun, unless some pieces are more equal than others. Perhaps that explains part of the appeal of chess in the Soviet Union, as it portrayed the injustice in society. Furthermore there is no money in the game, with pieces moving at the whim of a tyrant. But a more Hobbesian equality prevails, where even the lowliest pawn can defeat a knight or royalty. This makes teamwork indispensable to victory.

Feminists might be pleased that the queen is the most valuable piece in the game, although she is subservient to the king, like lionesses in a pride. Monogamy is ostensibly the message, until more pawns get promoted to queens. Imagine a game where pawns were promoted to kings instead, and you had to checkmate them all! But the gender of a pawn remains mysterious; the same pawn that can become a queen can also become a bishop. Perhaps the king doesn't mind male queens.

The game also stops at the moment of checkmate, before any scene of regicide can be enacted. Is this a model of conquest, where all the king's subjects become peaceful citizens thereafter? Or is it hiding the cruel aftermath of war, which may involve acts of retribution and retaliation?

As James Carse said, finite players play within the rules, infinite players play with the rules. Perhaps the time has come to conceive a new form of chess which is capable of capturing the complexities of our world, yet simple enough to have universal appeal. It must be full of uncertainty, where a winning position can become a losing one in a single move. It must be contextual, where the behaviour of each piece depends on its context and networks of relations. And it must be dynamic, where goals may shift and payoffs may change. And who knows, one day someone might use it to come up with a deep theory about life.