There are 255,168 possible games of tic-tac-toe and they’re all boring.
If both sides play perfectly, tic-tac-toe is a draw. And playing perfectly is relatively straightforward because you aren’t five anymore. So you’ve probably stopped playing altogether.
There are over 10^120 possible games of chess, which is over ten thousand trillion trillion trillion games for every atom in the observable universe.
A lot of them are pretty interesting.
If both sides play perfectly, chess is probably a draw too. But playing perfectly is beyond the ability of even our best computer engines. Beyond that, chess is a lot more subtle and demanding than tic-tac-toe. A great sequence of moves has the inevitable elegance and beauty of a mathematical proof. So chess retains significant interest for us.
But hold on—what would it mean to play chess perfectly?
That's a tricky question, and I'll try not to get unbearably technical. But notice that tic-tac-toe and chess are more alike than they might seem:
Both games are fully deterministic—there’s no randomness about whether my X can block your winning O, or my queen can capture your pesky knight.
Both games give players full information—there are no hidden trap cards. Everything's out in the open.
Most importantly, both games have fully sorted outcomes—there are only wins, losses, and draws. Even if we might disagree about which of several available checkmates is most beautiful, they all count the same, and they all beat losing or drawing.
As a result, there’s always (at least) one best next move. If your opponent has played perfectly, you can preserve a draw; if they haven't, you can punish them. Both tic-tac-toe and chess are fully solvable if you just have enough computational power to figure out what to do next.
There are so many possibilities to consider in chess that we happen to be practically unable to solve it, but it’s still theoretically possible. A superintelligent alien race armed with quantum computers might find chess just as trivial as we find tic-tac-toe.
Even if we did solve chess, it would still retain most of its interest for us. That’s because there’s a tremendous gap between playing perfectly in theory and playing as best as we can in practice. So if the aliens offer us their supercomputers, we should probably agree not to rely on them when we play. (It will be more fun that way.)
But I’m really interested in what playing perfectly looks like beyond games like chess. Very few practices are fully deterministic, supply full information, and fully sort their outcomes. For example, what does playing perfectly look like in the context of jazz or basketball?
There are uncountably many ways to play a jazz standard. While Miles Davis and John Coltrane reached electric heights, surely even their performances could be improved upon.
But I don’t think God Himself could play jazz perfectly, because jazz lacks internal standards by which perfection might be measured. Even if I'm clearly worse than Herbie Hancock, the practice of jazz doesn't give us fully sorted outcomes. So playing jazz perfectly isn’t just practically impossible; it seems theoretically impossible, too.
What about basketball? My coach once told me not to play afraid unless Jesus Christ himself laced em up for the other team. Talk about a perfect player.
But what would Jesus do?
Let's assume Jesus’s team would always score, and his opponents never would. But would Jesus dunk the ball every time, or sink full-court bombs? Run up the score as quickly as possible, or set up his teammates and rebound whenever they missed? (He does seem like a nice guy.)
On defense, would he bother to block my shot? Or just take the ball away before I made it across half-court?
And what would happen if everyone on both teams played basketball perfectly? Would they score every time, or never? I have no idea how to tell. The received wisdom is that “Good offense beats good defense,” but we don’t have any data on perfection.
So even though basketball supplies us with a perfectly clear notion of what it means to win, it still doesn't seem to give us a clear notion of what it means to play perfectly.
This may seem a trifling matter, but I think it has tremendous implications for thinking about morality.
Some moral frameworks imply that perfect play is possible, at least in theory. If you’re a utilitarian, and think morality is just about maximizing pleasure and minimizing pain, then (setting aside difficulties comparing and adding up pleasures and pains) there is some way of playing perfectly and achieving theoretical optimization, even if we can't practically figure out what that would be.
But that’s a surprising result. Is morality really more like chess than jazz?
When you play chess, there’s basically just one value you’re trying to maximize: how close you are to winning compared to your opponent. If you could win in three moves but I can win in two, I gotcha.
But in morality, it sure seems like we’re weighing competing values against one another: harm versus consent; liberty versus security; stability versus progress; and so on.
Is there really a single all-encompassing moral value to theoretically maximize?
I have my doubts.
Moral outcomes don't look fully sortable. There seem to be genuine dilemmas that pit our values against each other, not just tricky optimization problems on some underlying supervalue.
So what should we do?
Virtue theorists suggest that we should consider what the virtuous person would do, and how they’d balance these competing values. ‘The virtuous person’ is usually idealized but need not be perfect. Still, they’re a damn good exemplar, like Magnus Carlsen is at chess, or Duke Ellington was at jazz, or Nikola Jokic is at basketball.
I think we are all virtue theorists about chess and jazz and basketball because playing perfectly isn’t practically possible for us. So instead, we identify virtues (and vices) based on exemplary (and poor) play. Then we think about how to cultivate the virtues, avoid the vices, and put it all together. In this way, we can start to reason about the kind of person we should be and how we should live without losing track of what’s practically possible for us.
So here are the beginnings of a case that we should be virtue ethicists too.
Someday I’ll write it up real precise for the pros to tear apart. Together, we might not get it perfect, but I bet we can do pretty damn well. Because in the end, I’m a virtue theorist about philosophy too.