The Narrowness of Greatness

A few months ago, I published a book on the history and influence of chess. One of that book's consistent themes is that even young, not-very-accomplished chess players enjoy a nice fringe benefit of improved general thinking skills that they can apply in a wide variety of non-chess tasks. Studies show that chess "can enhance concentration, patience, and perseverance," according to the University of Sydney's Peter Dauvergne. Memphis State University's Dianne Horgan has shown that chess can sharpen evaluation skills, enhance "process feedback," and improve "calibration" -- a person's self-perception of his or her own ability.

But there's a surprising corollary to this point. When researchers look closely at what makes truly great chess players great, they don't find a similar carry-over into other tasks. Chess masters can remember thousands of positions and hundreds of games perfectly, and can even follow intricate games blindfolded, but they don't have an extraordinary memory for anything else. On all other subjects, their recall is rather ordinary.

According to the cognitive scientists who have studied this matter intensively over the last several decades, the same goes with chemists, political scientists, surgeons, athletes, and musicians.

This has proven to be one of the most enduring findings in the study of expertise...There is little transfer from high-level proficiency in one domain to proficiency in other domains - even when the domains seem, intuitively, very similar. [Paul Feltovich et al., 2006].

In short, it seems you can't be truly great at something without developing a phenomenal memory, but that phenomenal memory is always limited in context and content.

Talent researchers have a number of names for this phenomenon. One is "specificity in expertise." The phrase may come off as clinical and unexciting, but for cognitive scientists discovering it in the 1970s and early '80s, it was both revolutionary and profoundly instructive. Before, they had assumed that skills were skills, and that specific knowledge was just the stuff you plugged into a working brain. If you had a great memory, you had a great memory. If you were highly verbal, you were highly verbal. Seen in that generic light, one could plausibly assume that abilities were natural-born. "He's got a head for figures."

But the discovery of intensive domain specificity blew that idea away, and helped clear the road for a more sophisticated understanding of where extraordinary abilities -- and ordinary abilities, and everything in between -- come from.

Much more on this to come . . .

Comments

[Ray Cheng]

This talk about "specificity" is kinda funny, because I was just at the university coffee shop and overheard a couple psych professors talking. One of them apparently believes in innate talent, while the other takes the opposing view.

Anti-Talent: "Research shows that acquired skills are highly non-transferable to other disciplines. We call this 'specificity.' This pretty much sews up my case, since it proves that their advantages in their chosen discipline could not have been innate."

Pro-Talent: "Nonsense. It proves nothing of the sort. It only proves that their innate talent is highly specific. We both agree it takes tremendous work to express to reach expert level performance. Besides, why do you think they chose their particular disciplines?"

AT: "Sounds circular to me -- they're talented because they can perform, and they perform because they're talented."

PT: "It's not circular, it's a tautology. Look, if a person weren't talented, she wouldn't reach those levels of ability no matter how hard she practiced."

AT: "Come on, that would be a function of motivation, not talent. Some people just don't have the motivation to put in all those hours of grueling work."

PT: "Well, I'm glad to hear that you think *something* might be innate, and I do agree that some people are naturally more driven than others. But look, motivation also derives from talent. If you practice for two hours and see great improvement, then you're more likely to practice another two hours. If you practice for two hours and get nowhere, then you might just give up that day, and perhaps eventually abandon the discipline altogether."

AT: "You suck!"

PT: "No, YOU suck!!"

Thereafter, they threw their lattes at each other and began pulling on each others' beards. I got out of there as quickly as I could....

[Ray Cheng]

Please, I mean no disrespect to Mr. Gevers, and forgive me for any unintended offense. But Stephen Wolfram is a crackpot. Sure, we're all grateful that he invented Mathematica, but his book is, well, let me refer you to this review, which expresses my view nicely:
http://www.cscs.umich.edu/~crshalizi/reviews/wolfram/

Peter, I seem to recall that C.F. Chabris did something in his Ph.D. thesis at Harvard that might be the basis for a better model for expertise (and he takes the chess angle for his experiments). It goes beyond pattern recognition, adds another layer to the "chunking" theory. Something similar to your idea of analytical ability, but with a visual element (here, encoding and manipulating a chess position in the mind's eye).

[Johann Gevers]

David,

What a fascinating blog! Scientific American carried a series of articles on "The Expert Mind" (http://tinyurl.com/nw6q6) which supports several of your key points. You may also find it interesting to research Stephen Wolfram, one of the greatest living scientific geniuses. This is a good start: http://www.stephenwolfram.com/publications/talks/98-WWMCKeynote.html

I look forward to your future blog posts and your book.

Johann

[Kirk McElhearn]

Chess is an interesting case. While expertise in chess does represent some level of cognitive skill, it is often "compensated" by poor social skills. The same is the case, it seems, for other skills that revolve around math, calculating, etc.

Chess is, in my opinion (you wanted skeptics...) a two-dimensional skill. It's so easy that computers can now beat grand masters, and all serious players use computers to help them analyze matches. Perhaps it's this two-dimensionality that makes it accessible to people lacking in other skills, such as true reasoning, empathy, etc.

Now, to really look at a brain activity, check out the game of go. There is little brute-force memory involved, and it's more of a gestalt thing, clearly using different parts of the brain. It may be true that go players are asocial - I've never lived anywhere where there were more than a handful - but it's a real three-dimensional expertise.

Kirk

[Peter]

Okay this is interesting. It has me wondering whether the folk psychological distinction between memory and analytic ability, which I think I have been taking for granted all my life, is actually legitimate. At the neural level, can memory and analytic ability be disaggregated? You have a neural net, and when activated in a certain way a memory or image appears in the mind, but probably other things happen simultaneously, for example the biasing of attention. Like when we go out for pizza, we remember all the pizza we've ever had or at least easiily retrieve the memorable ones, and simultaneously become attentive to certain key features that we think are key to a good pizza, which helps us evaluate what we're served - charcoal, crust thickness, mozz quality, amount of sauce, etc. I guess the more we've experienced the more we remember and the more adept we are at recognizing whether the pizza before us will be good or not. Those two things covary, memory and analytic capacity, but that doesn't mean they are separate. So I think this is leading me to have the same kind of criticism in a different way, which is that i don't think it is the memory part of expertise that is what makes someone an expert, it's the pattern recognition. I also think motor memory is different than cognitive memory, because the former is in basal ganglia/cerebellum premotor and motor cortices and really immune from verbal analysis or representation, while the latter is temporal pole, hippocampus, parietal association cortices, frontal cortex (OFC, PFC, DLPFC, mPFC) and very very amenable to verbal analysis and representation. I am obsessed with your book. I predict a huge success. Everyone I talk to about it gets very into it and we have long discussions.

[David Shenk]

Peter -- Henceforth, you are forever enshrined as GeniusBlog's First Commentor, and I'm submitting an early nomination for you as its Skeptic-in-Chief.

Based on what I've written so far, you are dead right to call me on the distinction between memory and skill. Common sense says that memory -- as important as it is -- is just memory, just one component of skill/ability. But my sense so far from reading a lot of the expertise research is that the expansion of storage and quick-recall turns out to be *the* pivotal feature of any advanced ability. The upper range of experts are actually able to convert space in the brain normally reserved for long-term memory and make it more like instant-recall RAM. The real point here is that memory is much more than lists or recitable information -- it is recognition and recall of movements, thoughts, patterns, etc, all prior relevent experience that enables an expert to make sophisticated judgements in a split second. An actor's or baseball player's suite of thoughts and movements are made possible by their easy ability to draw on so much experience.

This is central and I'm not doing it justice -- partly because I'm not up to speed on it. But rest assured, much more to come on this.

[Peter]

David - hey great website! Amazing really. It's hard to believe someone who is a great writer can also have talent for website design. I would have thought the two abilities would be independently sorted. My critical comment on this entry: are you conflating memory with ability? I know San Francisco very well but have no skill in city planning. The same is true of most 'city planners'. Many actors can recite whole Shakespeare plays but couldn't get paid to act in them. Most baseball statisticians can't play ball. Etc. So tracking the memory abilities of experts probably doesn't tell you that much about their expertise because it just lacks specificity. Does this make any sense? I type well but am not reasonable.