Chess and Robots and Humanity and the Nature of Artistic Expression
Updated: Feb 11, 2020
I learned to play chess at the age of six. I don't play much anymore, but the game has influenced me greatly. For instance, the part of the brain that recognizes faces, the occipital lobe I believe, is responsible for all complex pattern recognition, and it turns out that once you've played enough chess, you stop seeing the board and the pieces, and you stop thinking about this or that move. Rather, you just see the whole thing at once, as a unit, much like you don't see someone you recognize and think Oh there's Steve's nose, or That's Betsy's earlobe. You just see the face, take in all of the minutia at once (the shape, the coloring, the distance between features, etc.), and immediately know who it is. Similarly, I see a chess board and just kind of know what it is and what move I should make. This is probably pretty common in chess players. The reason that it's noteworthy is that, having exercised my occipital lobe so often at an early developmental stage, I really kind of never forget faces. Like ever. If I have seen someone, even only briefly, and then I see them again, I recognize them.
For example, I submit Liza Proctor and Cyrus Karni. I met the both of them at philosophy club on the last day of high school my freshman year. I guess I was fourteen-years old. I had never been to philosophy club before. I sort of stumbled upon a flier, and I never really wanted to go home, so I decided to see what it was about. I walked in, and there were some kids on a couch, all listening to a boy read what seemed to be a dissertation on Love. It was actually really good, or at least I remember thinking so, and the boy was quite smart. I talked with him later, and found out his name was Cyrus. I asked him what he had been reading, thinking that he, like I, was a hopeless romantic, but it turned out that he had just lost his father to cancer and was reading about and experiencing real love (as opposed to my childish obsession with courtship). We talked, he ended up giving me a ride home that afternoon, and later I read an article in the Chronicle (Houston's local paper) about him making a perfect score on the SAT shortly after his father passed. He was a real cool guy. But he was a senior that year, and I never saw him again.
Liza Proctor was also there. I had a much smaller interaction with her. She asked my name, I told her, and she told me about an ex-boyfriend of hers that had my name. I could describe exactly how she looked, with the bobbed brown hair and plain features: small brown eyes, small, slightly upturned nose, roundish face, defined chin, and whatnot. Her left ear pushed out from under her hair, a bit like an elf's might.
Anyway, I met her that once, and then didn't see her again for three years. I was working at Target on South Main Street, and she walked into the health and beauty section that I was busy tidying. I immediately said, "Hello, Liza," and of course she didn't have any idea who I was, and I spent a good five minutes explaining what I've just begun to tell you, that I can't seem to forget faces, no matter how insignificant they might be. Eventually she believed me because she was in philosophy club in high-school and how could I know about Derek (even the British spelling)? A couple years later, I saw her in Austin at a gas station and decided, like I usually do, to ignore the fact that I could tell her about the time we met in Target two years before and the time we met at Lamar High School three years before that.
It's really useless and awkward information. I often recognize people that I've never spoken to at all. I just saw them some time, maybe a couple of years ago at a convenience store or maybe in a friend's yearbook or maybe just riding their bike past my car at a stoplight. I don't need that memory. It doesn't help to see someone randomly for a second time and immediately remember the first.
However, when walking across the street in DC just a month ago, I saw Cyrus coming towards me. He was much older, bearded, wearing some sort of baggy Indian shirt-thing, now ten years removed from his preppy, high-school attire. I felt like I should say hello. So I yelled out "Cyrus," and he stopped for a second before we realized we were both in the middle of a street and should retreat to one side or the other. I explained things to him and what I was doing and why I was in DC and how I was actually going to meet some friends from UT, and it turns out that he actually knew the people I was going to meet, so we decided to meet up at a bar later that evening, and we did and got along swimmingly. So now I have Cyrus Karni phone number and a standing invitation to his poker games in DC every Thursday.
All this because of chess. I guess I also got a strong understanding of the difference between tactics and strategy (tactics being discrete maneuvers to gain advantages, strategy being long-term plans for victory), a bit of ridicule as a chess-nerd and later the president of my high school chess club, and finally, a silly yet strong obsessive compulsive disorder when it comes to circles resting within or atop squares. For instance, if I have a glass of water at the dinner table, and the table cloth has a large checker pattern, I must set the glass in the exact center of the squares. And you have to, too. No really you have to. It's this thing. I mean... look just do it, alright. I need this.
Chess and Robots.
So I mentioned the other day that Gary Kasparov played a super computer at chess. The computer was called Deep Blue. In the first match of six games, they split, three and three. I've always loved that. I mean, you have to think that you could program any computer to make any number of calculations based on some criteria and find the optimal choice. That's what computers do. And they do it really well. Much better than you or I could.
But if you asked a computer to paint a picture or sculpt something, the story would change drastically. You could teach any computer the technical end of things. You could train a robot to make skilled and precise movements. But art is about emotional expression and short of programming pre-ordained emotional qualities, or rather specific ways to express the emotions we have, into the instructional code, computers are essentially sociopaths. Similarly, though we could ask a computer to observe and record all of the characteristics of some set of artistic examples, we would never expect the computer to create an accurate formula for artistic expression of any sort. Nearly all attempts of the like are, despite superficial or coincidental successes, ultimately shoddy and inauthentic.
So when you think of chess, you probably think that chess is just a game of glorified number crunching. At least I did. There are sixty-four squares, and twenty possible opening moves for white. Then you have twenty moves black could make in response, so four-hundred combinations within the first move, and it grows more than exponentially from there, since some moves open up more possibilities than others.
As a human, this makes for a rather confusing game, and it's a good thing you don't have to memorize all of the possible moves and their outcomes to become a decent player. Luckily, because of the magic of my occipital lobe, I just see the whole board at once and the right moves kind of glows a bit brighter than all of the others. So I make it and move on from there. Occasionally I calculate into the future of the game, making assumptions and educated guesses about my opponent and what will or will not seem like a good move to him or her, depending on the position of their pieces and their personal disposition. Perhaps I can accurately see four or five, sometimes eight, moves into the future, using filters to disqualify unlikely moves and their progeny from consideration. And Bobby Fischer or Gary Kasparov can see much further than I. But at some point the raw computing power of the human mind gives way, and a super computer is just much stronger, much faster than we can hope to be.
Of course, if that were the case, wouldn't Deep Blue have won all six games? Do we believe that Kasparov is just smarter than a computer? There are autistic savants that can do arithmetic faster than calculators can. But I don't really think that's it.
I think asking a computer to play chess is a bit like asking a robot to choreograph an interpretive dance. Perhaps it can learn to plié and all that. It can analyze other dances and emulate certain bits. But it can't create. It can't be original. At least I don't think it can. Computers are confined to the code that bore them.
I can tell you from personal experience that creativity is essential to becoming a good chess player. I once played this player who was much much better than I. (He was a drug dealer, and I was hanging out with him because he was friends with my girlfriend at the time. She was, unbeknownst to me, a bit of an addict. A story for another time.) We played speed chess, one-minute games, where we each had one minute on our clocks before the game ended. My time ran when it was my turn. Immediately after my move, I pushed the clock, and it was his turn, and his time began running. This type of chess is exciting fun, but it requires an intimate knowledge of the game. You have very little time to process or look ahead. You play almost entirely on instinct and creativity. This guy was demolishing me. He won nearly every game. (He was on crystal meth at the time, which in hindsight seems a little unfair, doesn't it?)
Still, I kept pace and played well enough to impress him, so he decided he would teach me some things. He taught me tactics and strategy and mentioned that I would be much better if I stopped trying to make all of the right moves all of the time. He told me I needed to be more creative, and he figured that, having been told to get creative, I would be very creative in a week or so. He was right, and the next time we played, I made moves that were not only good, but surprising, which made them harder to anticipate and harder to defend against.
I guess what I am trying to say is that, within any system of rules that allows sufficient variation, there is room for human preference, for arbitrary aesthetic decisions to be made, and ultimately for artistic expression. And that, maybe three times out of six, artistic expression might trump the raw computation of even the strongest super computer.
Chess and Robots and Humanity.
Of course, chess isn't actually infinite. It is vast. But really it's finite. There are a certain number of moves that can be made at any point, which lead to another certain number of moves, and each of those separate trees grows and grows until the pieces begin to get taken, or the board gets cramped, and eventually, rather than more and more moves with each round, there are fewer and fewer and finally none.
Humans can't possibly consider each of those moves, so we simplify. We decide which moves seem better based on really ridiculously complex algorithms of position and past position and player style and preference and emotional audacity or whatever. We say that no one would ever make this move because it's clearly stupid. And that this move is likely and should be defended against based on the threat to this or that piece. We filter.
And the people who write the programs that run chess-playing computers do the exact same thing. They identify the limits of each filter and code it into the computer. When we think about Kasparov splitting with Deep Blue, we have to remember that at the time he was the best chess player on Earth. So maybe whoever was helping IBM define the limits for each filter (it was a brain trust of about 20 of the world's best chess minds) just didn't know enough about chess to best Kasparov's filters. At least not three times out of six.
And maybe my assumptions about computers being about to out-calculate humans is unfounded. Perhaps we are closer to organic computers than I've thought. You use a PC right? If you press Alt-Ctrl-Delete and the task manager comes up, you can look at a list of processes your computer is currently performing. There are something like fifteen or twenty or forty. Maybe ninety if you open a lot of programs. Whatever, there is some relatively small (two-digit) number of things that your computer is considering. It is not, along with those things, twiddling its thumbs or constructing a list for the grocery store or thinking about sex nearly six times a minute (as the average man supposedly does). Your computer, when compared to the average human, is quite focused.
Humans have to breathe and think and metabolize and do all of these biological things that you undoubtedly know much more about than I. We have five senses and input in four dimensions, along with virtual sensation coming from the mind nearly constantly. Still, we are able to think in straight lines, somewhat logically and with persistent focus, such that we can make great strides as humans. We can consider terribly complex things like Metaphysics and Epistemology or beautifully simple things like the wheel or urgently safe things like repression and even things in our own image like robots and the computers that act as their brains (but perhaps not their minds.) Perhaps the computational power of machines just seems to dwarf our own because it is so singular. Even the simplest computer can focus better than any human. And in that simplicity, computers have reached the point where they can consistently outperform their complex creators in matters of objective calculation. Even chess. Three times out of six.
Chess and Robots and Humanity and the Nature of Artistic Expression
But computers can't create. Computers can't surprise the person who programmed them, though they may be able to surpass them at a given task in a given amount of time. We can assume that computers can perform desired tasks faster than we think desired thoughts, but our variability is often really fantastic. Art might be the best thing about life, if only because each human gets to have their own version of it. You can make whatever art you want (skill permitting), and you can appreciate whatever art you want, based on whatever criteria you want, and who cares what anyone else thinks?
And maybe that bit, about beauty being in the eye of the beholder, makes it hard to claim that computers can't create art. I mean, art has parameters. Paintings use paint on some surface to create something. If a robot takes a brush in its robot-y hand and randomly dabs the canvas, is that art? If an artist programs the robot to draw landscapes like Bob Ross and his happy little trees, and they aren't mere replicas of previously painted pictures, isn't that art? How is the product appreciably different than any painting I create? And can we, as humans, really claim that our artistic forays are anything other than programmed by previous observation and experience. For instance, if I were to take up painting and try to draw trees, they would look a lot like Bob Ross's forests. Because that's all I know about painting trees.
Perhaps a robot wouldn't be able to refine its own skill over time based on its own experience. At least not the robots we have now. But honestly that seems like a real small jump from what we have now. Asking a computer to analyze its own output is not exactly revolutionary. Raytracers (which render light in 3D graphics) do it all of the time. The artist tells the computer to place a light source somewhere, and then the computer uses the information about the light and how light behaves and whatever other objects there are in the scene to simulate the angle and intensity of some millions of individual rays of light to approximate the lighting for the picture.
So why do we, as humans feel that that process is different, somehow more robotic and less human than our own eyes looking at where the sun is on a horizon and deciding the angle of the highlights in the canopy of such and such tree? We make more choices, with more nuance, and less conscious definition to our algorithms, but one can't help but wonder if the seeming infinince (this is not a word, but it should be) of our choices, as well as the accompanying infinite criteria we use to pare down our lives into units, is really finite. If, perhaps, art is just a name that we give to our imperfect filters.
Case in point, in 1956, a 13 year-old Bobby Fischer played International Master and U.S. Open Chess Champion Donald Byrne who, at the time, was twice Fischer's age. The game began with a standard variation of the English Opening that any club chess player would know. But at move 17, the young boy created something amazing. He created a threat that looked like a blunder, sacrificing his queen and in the process setting up a 24-move chain of events that led to an unprecedented upset by the young man. I've tried several times to explain how beautiful this game is, how profound Fischer's genius is, but words always fail me. They will again here, but I will try just the same.
This chain of moves is not simply a feat of foresight (it would be closer to clairvoyance anyway). And it is not merely a well planned trap. Much more awe-inspiring than those moments when life just falls into place, it's one of those things that's both shatteringly surprising and immediately clear, as if common sense would have led you there had you only thought to think of it. This move that started it all, not just this win, but Bobby Fischer's rise to chess fame, this move that must've come directly from heaven, bishop to E6, it's as if it fell directly from the collective unconscious. This bit of human invention from a thirteen-old boy is as profound as any I can think of: Newton and Leibniz simultaneous conception (discovery?) of calculus, or Einstein's relativity, or T.S. Eliot's Prufrock, or Satie's Third Gymnopédie or any other bit of art or science that you could think of or appreciate or love. It's a prime display of human capacity for creative genius.
Only maybe it's not. Maybe Fischer just filters better than Byrne, even at half his age. Deep Blue would not have fallen for Fischer's trap. It would have had the processing power to evaluate each of the millions of potential situations twenty-five moves ahead of move sixteen. Maybe Fischer's mind is just a better chess computer. And maybe some computers in our head are so good that they loose the appearance of calculation. Since I can't possibly fathom the chess arithmetic that he would have had to do to choose that move, or since no one other than him can, we all call it this amorphous thing, Art.
There is some evidence for this argument. One chess player, a friend of Fischer's remembers first meeting him. They were at a tournament, and, on his way to the restroom, Fischer glanced at the game the player was engaged in. Fischer then went on his way to the bathroom and didn't see the player again for several years. When they saw each other next, Fischer asked how that particular game all those years ago went. The player responded that he had played it out to a draw, and Fischer said that he would've won from the position that he saw that day. (Fischer was a bit of a prick.) The player, surprised and a bit insulted though wary of Fischer's known brilliance, said that he didn't think there was any way to win outright. At that point, Fischer pulled out a chess board and set up the entire game from that brief memory and proceeded to show the player that he could have indeed won had he moved such and such piece in such and such a manner.
Bobby Fischer can reenact chess matches from memory on a single glance, some people have photographic memories or perfect pitch, some people are able to instantly recall what day of the week a distant date is, and I cannot possibly forget the faces of the strangers I see. Computers could easily replicate any of these feats. Yet, for some reason I am reluctant to call the human brain a computer or the human existence a computer program. I imagine that the distinction between human-kind and robot-kind is less hard and fast than most humans would like. I wonder if maybe it is a difference of degree rather than type. I also imagine that humans will fight any and all progress towards sentience in machines. Still, the more I think about it, the closer and more inevitable that idea becomes.
P.S. Can you do the robot?