Thursday, October 19, 2017

The Hunt for Genius, Part 2: Crackpots, athletes, 4 kinds of judgment, training, and Cultural Context

I continue reposting my series on the MacArthur Fellowship Program. This time I take up the problem of identifying "genius"-class creativity by running through a variety of examples and ending on a brief discussion of the importance of cultural context. How do we bias our selection process toward the future, not the past? I've collected these posts into a working paper, The Genius Chronicles: Going Boldly Where None Have Gone Before?, which you may download at this link:

* * * * *

Part 1 is my post on the misguided MacArthur Fellows Program. And I thought that would be the end of it. I was wrong.

Now that I’ve gotten my brain revved up thinking about “genius”, whatever that is, I’ve got to think a bit more. The foundation is making judgments about people, judgments about the originality of their work, their ability to cross traditional disciplinary and institutional boundaries, their need for support, and their potential for future contributions of an extraordinary kind. The program has been consistently criticized for picking too many fellows who don’t meet those criteria.

In that post I argued there is in fact a simple way to improve those judgments relative to those criteria: don’t give fellowships to people with stable jobs at elite institutions. The purpose of this post is to clarify my reasoning on that point.

I begin by pointing out that it’s possible for one person to be both a genius and a crackpot. Then I have a brief note on the Nobel Prize, where the point is that even giving awards for accomplishment is difficult. In the following two sections I step through athletic and musical performance as a way of outlining different kinds of judgments, which I’ve called objective, complex, incommensurable, and predictive. I return to the MacArthur Fellowship Program in the final section where I once again talk about the importance of cultural context.

Two for One: Genius and Crackpot in a Single Package

First, let’s think about, say, Isaac Newton, a prototypical scientific genius. We remember him for his work in physics (optics, mechanics, and gravity) and mathematics. No one cares about his work in theology and alchemy except historians, yet it meant a great deal to Newton himself. In the last century Albert Einstein was quickly recognized as a genius, mostly for his work on relativity and photons. He spent the last part of his career looking for a unified field theory. For a long time that work was considered to be a waste of time. Now that unified theory has made a comeback in physics I don’t know whether that work has been re-evaluated or not.

Were these guys working on half a brain when they did that misbegotten work? Were they drunk? I mean, what happened to the supernal abilities that allowed them to make profound and permanent contributions to science?

Nothing happened to those abilities. There’s no reason to think that they weren’t firing on all cylinders when they did that work. The work just doesn’t fit very well with other knowledge of the world. Think of ideas as keys. What do we use keys for? To unlock doors. Some of the keys these geniuses crafted unlocked real doors. Other keys don’t unlock real doors. Whether or not a key unlocks a door is not a matter of how well the key is crafted. The most exquisitely crafted square peg is not going to fit into a round hole.

Well, it turns out that some of the locks these guys had in mind when crafting keys weren’t real. They were figments of their imagination. Just because the lock was imagined by a genius doesn’t mean it is real.

And so forth.

The point is that ability is not enough. That ability has to be fitted to context.

The search for genius, however, is always conceptualized as a search for ability. This is most obviously the case when genius is defined in terms of a score on some standardized test, an IQ test. If you score high enough on the test you’re a genius – as defined by the test. Otherwise, no.

Now, I rather doubt that anyone involved in the MacArthur Fellows program cares about scores on IQ tests. Whatever it is they’re looking for, it can’t be identified by an IQ test. If it could, then running the Fellows Program would be trivially easy. If tests did the trick there’d be no need for the program. The geniuses would be identified by the standard testing programs undertaken in schools. They aren’t.

So, how do you find a genius?

Nobel Prizes: Even Post Facto Judgments are Difficult

What about Nobel Prizes? They, of course, are awarded for accomplishment, not for promise. And so the prize is not conceived of as one given for ability, though we all assume that Nobel Laureates must have extraordinary ability in order to do whatever it is that got them the award.

And yet the fact that these prizes are awarded for accomplishments visible to all doesn’t insulate them from criticism. I’m sure if I were to did around in what’s written about Nobels I’d find lists of people who got them, but shouldn’t (e.g. Obama or Kissinger for the Peace Prize) and other lists of people who should have gotten them but didn’t. Judging the value of accomplishments such as these is not easy.

So, let’s start by thinking about some kind of ability where the tests are straightforward.

Looking for the Best Athletes

Let’s say we’re looking for the fastest sprinters in the world. One’s ability at sprinting can judged in a straightforward way. Get your candidates together and have them run a race. Determining who crosses the finish line first is easy in principle, though it can be tricky in fact. Still, if we have a photo finish, then run them again a day later. Award the prize to the test two out of three.

Of course, we CAN make things more complicated if we wish. Maybe you don’t think one distance (100 yards or 100 meters) is enough. You want to test them on two distances (add in 200 yards, or 200 meters). What if one is best at one distance while another is best at the other distance? Maybe we simply add the times for the two distances and award the prize for the best combined distance. And if you don’t like that, well, propose another formula. But once we’ve picked a formula, the winner’s going to fall out pretty directly.

Maybe you want to add another distance. OK. Do it. Now we’ve got to ponder another set of formulas.

And so forth.

There are practical problems as well. How do we get all the sprinters together for a race? We don’t have to do that. We can run a tournament with preliminary rounds here and there and first round winners advancing to later rounds, and so forth. This will, of course, introduce difficulties. Maybe one of the best sprinters gets knocked out in an early round for any of a number of reasons. We’ll just have to deal with it.

For that matter, we don’t even have to run the sprinters in head-to-head competition. We can just take their X best times over the course of some period of time – a year, three years, whatever – average them, and use that result.

We can play these conceptual games for quite awhile. But the end result isn’t going to change. Making a determination is relatively straightforward, but only relatively so. It’s not going to be flawless. And the range of possibilities we play with is relatively constrained.

Things change when we consider disciplines such as gymnastics, diving, or figure skating. In these disciplines there are no sharp criteria for judgment. Performances are judged on the difficulty of the skills involved and how well those skills are executed. Multiple judges are used and the final score is some function of those individual scores, but not necessarily a simple average (high and low scores may be tossed out).

This is a bit more like the situation facing the MacArthur selection committee. But only a bit. Gymnasts are judged only against gymnasts, not against divers, skaters, or for that matter, sprinters and point guards. And the range of variation available to a gymnast is relatively narrow compared to that available to a novelist, a particle physicist, or a community activist.

Now, let’s open things up: Who’s the best athlete in the world? That’s very different from looking for the fastest sprinter or the best figure skater.

What counts as athletic activity? Track and field events, sure. But which ones? Swimming events. Sure. Which ones? Basketball, football, soccer, cricket, rubgy, vollyball, ice hocky? What about mountain climbing? Horse racing? Ping pong? Golf? Curling? Skate boarding? Dance sport (aka competitive dancing)? Arm wrestling? And so forth.

And whatever list of events you include there’s the problem of comparing athletes in different disciplines. How do you compare an expert in ping-pong with a 10-meter diver, or a golfer, or a kick boxer?

Let’s forget about picking the best athlete and instead go for the 10, 20, 30, 40, 50, 100 best. That allows us to pick more winners, so it takes some of the pressure off of comparing athletes in different disciplines. But it doesn’t eliminate that problem entirely, unless we pick one best in each discipline, but there are ways to mess that up too (e.g. do we allows performance enhancing drugs, physical prosthesis?). Now we’re going for a suite of winners, which means we’re likely to want to present an interesting array of athletes. We’ll have a swimmer, a runner, a basketball or volleyball player, and so forth.

Once again we’re moving toward the problem faced by the MacArthur Fellows selection committee. You have a wide range of people, all of them excellent at something, but these different kinds of excellence aren’t directly comparable. So how do you judge?

Still, I’d say this is easier than the MacArthur problem in two respects. In the first place, excellence in individual disciplines is easier to judge. To be sure, excellence in gymnastics or basketball cannot be judged as directly as excellence in weight lifting or pole vaulting, but it’s not as open-ended as excellence in physics, or merely particle physics, or excellence in poetry.

In the second place, we’re talking about judging athletes at the top of their ability. But the MacArthur Fellows Program, at least in theory, is looking to identify geniuses before their genius has fully flowered.

So, let’s return to sprinters. Now our task is those sprinters between, say, nine and eleven, who will be the best when they are at their prime, whenever that is (late teens to mid-20s I’d guess). How do we do that?

Finding the best child sprinters is no more difficult than finding the best adult sprinters, nor is it any less difficult. But we cannot assume that the best child sprinters will be the best adult sprinters. Some first-rate child sprinters will taper off as they get older and some second-rate child sprinters will blossom and become first-rate. We cannot predict which is which.

What do we do? Well, if the problem is to identify those 10 child sprinters who will be the best adult sprinters 10 years later, then we’re stuck. We might get one or two of them if we’re lucky, but we certainly won’t get a half-dozen much less all of them. And we might not find any at all.

Our odds get considerably better if our task is to identify that group of 1000 child sprinters from which the ten best adult sprinters will emerge a decade later. The odds get even better if we start a training program to help those 1000. Maybe we work with all 1000 for two years, then cut the number in half and continue working with those. Then make another cut three years later or so. And so forth.

And that, more or less, is how organized athletic programs operate. A lot of young children go into the large end of the funnel so that 10 or 15 years later we have a small trickle of superb adult athletes coming out the small end.

That’s how the MacArthur Fellows program is supposed to work. Nominees are already pretty deep into the funnel, but they’ve not yet made it to the output tube. The program’s objective is to pick the best among the nominees and give them the support they need to make it to the output tube. No one expects them all to make it.

But none of this involves the kind of contextual problem that we faced with Newton and Einstein spending a lot of time on fruitless research. Oh I suppose we might compare that to Michael Jordan’s golf game, but the comparison is not a particularly good one. After all, there really are good golfers in the world, touring pros are better than Jordan; but no one’s yet made a go of alchemy or, so far, of a unified theory.

Scorecard: Where Are We?

Let’s recap.

1. We started with the problem of identifying the best sprinter. There the criterion is an objective one. That’s the easiest case.

It takes relatively little skill to make these judgments. Close finishes may require careful analysis of photographs, but other than that, there’s little to judge.

2. Then we considered gymnastics, which is more difficult. Let’s call that judgment complex without, however, worrying too much about just what complex means. This is just a rough sort we’re doing.

These judgments are trickier. But there’s not reason to think that one well-trained panel would be different from another. If is, of course, possible for judges to cheat exercising biased judgments for favorites. But then any system can be gamed somehow.

3. Then we looked at the problem of identifying the best athletes regardless of discipline. There we’ve disciplines where both objective and complex criteria are in use and this is compounded by the fact that skills in widely different athletic disciplines are incommensurable with one another.

Here the judgments are much more difficult. Given the great variety of athletic disciplines, a panel of judges might well be systematically biased. For example, a panel from the United States is going to be unfamiliar with athletic disciplines not practiced here, or only practiced in a minor way. Sumo wrestlers and cricket players are not likely to be given sufficient consideration. That’s comparable to the kind of bias operating in the MacArthur Fellowship Program.

4. Next up, identifying elite sprinters when they’re young. Call it predictive judgment.

As long as we’re dealing with judgments that are made within specific disciplines, we have nothing more than objective judgments and complex judgments. But they’re applied to immature athletes rather than mature ones.

5. Finally we talked about instituting a training program, where we work with young sprinters to develop their abilities. We’re now deliberately intervening in the development process. But there’s always a chance that we’ll misallocate our training resources. Some kids who would in fact blossom will be dropped, or never even enter the program, while others in it will not pan out.

This is the same as with the previous case, predictive judgment, except that we’re now making judgments about who will be supported in the next tier of the program. Some judgments need to be made about cut-off points, and those judgments will be a function of resource availability. Once the cut off points have been set, determining who qualifies and who doesn’t is fairly straightforward.

Let’s quickly run through the same series of judgment disciplines in a different arena, one I know fairly well: musical performance.

1. Objective judgment: What trumpet player can play the highest note? Believe me, lots of trumpet players really care about this one, though it’s mostly irrelevant for all but a very small number of contexts and players. But the call is an easy one to make. Either note X is higher than note Y or it isn’t. Things get trickier if we worry about tone quality. Does it count if X is higher than Y, but the tone is thin?

2. Complex judgment: Who’s the most technically skilled trumpet player? I specify technical skill so as to eliminate the whole range of judgment issues that come into play when we consider matters of artistic interpretation, a range that gets even larger when we think about improvisation (in a variety of styles, jazz, Latin, contemporary classical, whatever).

Technical skill on the trumpet involves many things in addition to range. Tone quality counts, and we need to think about tone quality in all registers, lower, middle, and upper. What about tone quality for pedal tones? There’s the cleanness and speed of single, double, and triple tonguing. The ability to play wide intervals cleanly and accurately. Speed in all kinds of playing. What about dynamic range? Can a trumpeter keep the tone spinning while playing very softly in the lower or upper registers?

Trumpet players will vary in their mastery of different technical elements. Jazz players, for example, are much less likely to have mastered double and triple tonguing than classical players because those skills aren’t really required in jazz. So how do we assess the technical skills of, say, a Dizzy Gillespie as against those of, say, a Maurice André?

3. Incommensurate Judgment: Who’s got the best technical skills of any living musician? How do you compare the skills of trumpeters, tabla players, beatboxers, guzheng virtuosi, and pianists?

4. Predictive judgment: Which ten-year old trumpeters will mature into stars? While I am by no means a star, I’m better than any of the kids I started out with back in the fourth grade. In fact, I fell behind the others in my group after several months of lessons. It took me two years to catch up and them move ahead of them.

5. Training programs: It’s going to work out the same as it does for athletes. If you train them, they get better.

But this is all simple stuff compared to the situation the MacArthur Fellowship attempts to address. All of these judgments take place in cultural contexts that we may consider to be “closed” in some sense. The skill being judged is well defined in existing terms. That is not the case with the MacArthurs.

The Context of Genius

The happiest situation is when the supernal capacities of an exceptional individual emerges to evolve in time and in step with the larger movements of his or her society.

Since I’ve already talked of trumpet players, let’s consider one of the most important trumpet players of the Twentieth Century, Louis Armstrong. While he had fine technical skills, they were not outstanding in comparison to the skills of the best classical and concert band trumpeters and cornetists. But technical skill is not what’s important about him. Armstrong is important for his musical inventiveness. He transformed jazz. He changed the direction of musical development.

But he wasn’t working alone. As I wrote in Beethoven’s Anvil (pp. 255-256):
The greatness of an individual musician such as Armstrong is a function, both of his power to forge compelling performances from the “raw” memes and of the existence of that meme pool. While Armstrong may have been ahead of his fellows, he couldn’t have been very far ahead of them, otherwise they could not have performed together. Beyond this, without a large population of music-lovers familiar with the same meme pool, Armstrong’s recordings would have had little effect. By the time he went to Chicago, a large population had been listening and dancing to rags and blues, show tunes, fox trots and Charlestons and marches, all with a hot pulse and raggy rhythms. Armstrong’s improvisations gave them a new wild pleasure, and their collective joy made him great.
Such good fortune is not always the case.

Consider the unhappy case of Ignaz Semmelweis, a Hungarian physician of the second half of the 19th Century. In 1847 he proposed to reduce the incidence of fever in hospitals by having physicians wash their hands. To us this idea seems like common sense. But it was considered heretical when Semmelweis proposed it. He was beaten to death in an asylum in 1865. It was to be a decade or more after his death that antisepsis began to gain acceptance in the medical community. That acceptance was contingent upon promulgation and acceptance of the germ theory of disease, which was unknown when Semmelweis first made his proposals.

Such stories are common enough in the history of creativity. It took decades before Wegner’s theory of continental drift was accepted in the geological community. Van Gogh wasn’t widely hailed until after his death. Ornette Coleman was regarded as something of a freak early in his career, though the MacArthur Foundation finally worked up the courage and imagination to gift him in 1994, long after his groundbreaking recordings of the late 1950s and early 1960s. Other accolades followed.

These people, and many others, were “ahead of their time”, as the saying goes. The ostensible purpose of the MacArthur Fellowships is to identify these “geniuses” before the times catch up with them and to help things along. The recognition conferred by the gift itself may well bring a fellow’s work to a wider ranger of people, some of whom may understand, appreciate, and help advance it. The money itself gives the follow a five-year financial cushion, allowing them to function creatively even without acceptance that’s wide enough to provide a livable income.

In practice, though, too many fellowships have been awarded to people who are either reasonably well established and don’t really need the money or recognition or whose accomplishments, though of a high order, are not particularly creative.

The point I was making in that earlier post is simply that, in effect, a whole new set of contexts has emerged in the last half-century, a new cultural regime, and the MacArthurs are simply ignoring them (and it). The world is changing on a massive scale; cultural regimes of all kinds – religious, artistic, scientific, engineering, organizational, lifeways in whole and in part – are drifting, colliding, merging, conflicting, and reconfiguring. The MacArthur Foundation’s commitment to gifting people at elite institutions is, by that fact, a commitment to the old world order. If the foundation were to no longer hand out awards to those institutions by gifting selected individuals within them, it would eliminate that bias.

That would be a turn toward the 21st Century. Perhaps not a strong turn, but a turn.

If the foundation had done that this year, then those 15 awards that went to people in elite institutions would be going to those outside the 19th Century institutional network.

Of course I don’t know who those outsiders might have been. I have no reason to think that they would have been people of greater personal abilities than the 15 insiders who were gifted. But there’s a better chance that their work be oriented toward the worlds that are struggling to emerge. To use trumpet playing as an analogy, there is no reason to believe that these 15 outsider trumpeters would have technical skills superior to those of the gifted insider trumpeters. But their artistic sensibilities are more likely to be attuned to the future.

And that’s what I’m aiming for, the future. I want to change the bias in the system. No more, no less.


  1. But what's a risk? Who defines it? To use a crude analogy, to one kind of person, taking a risk means getting in a boat and heading out into the ocean toward an island no one's ever visited before. But to someone who's been trained in and who works in the foundation world, taking a risk means giving a fellowship to someone who went to Penn State rather than U Penn.

  2. On Newton, consider the opening paragraph of a Jonathan Rée's review of a recent biography:

    The life of Isaac Newton falls into two halves, and the main problem for Newton studies is how to fit them together. In the first half he was a sulky Cambridge mathematician who, at the age of 44, astonished the world with a work of natural science that was soon recognised as one of the greatest books ever written. In the second he was a sleek London gentleman wallowing in power, wealth and prestige and devoting his intellectual energy to esoteric studies of the Bible. How could they be the same person?

    H/t Chris Campbell.