There are probably some kinds of intellectual problems to which only geniuses can make any contribution. That is, for some problems there may just be no way to break off a nugget of the problem that a non-genius could handle. But this is not true of most kinds of problems. Most kinds of problems can be broken up so that the non-geniuses handle the small nuggets and the geniuses handle the big nuggets and/or figure out how to put the nuggets together. If success and failure could be determined unambiguously, no non-genius would have an incentive to tackle a genius-sized nugget because there would be no (or an unacceptably small) chance of success. But if failure can be obscured, then it might be worth the while of a non-genius to tackle a genius sized nugget, and just obscure the fact that they didn’t get it right. This is worse than nothing because it not only makes no contribution but also makes it hard to figure out who really knows what they are talking about. It seems to me that a lot of this goes on in empirical economics. It is clear that there are a lot of small nuggets that non-geniuses could work on, like carefully checking the work of genius researchers for robustness, or redoing the genius analysis on lots of other data sets. etc. But this won’t get you any love, and a bad effort at tackling a big nugget will likely not be easily discovered, because much of the work is very complicated (either of necessity or by design) and no one will ever take the trouble to unpack it, and even if someone did, there would be some remaining ambiguity about who was right and who was wrong. I know this is a problem for me, I often have a lot of doubt about who to listen to. I’m not sure whether it introduces any systematic bias.
Geniuses don't do Economics.
Following more directly on Stuart, sometimes the big nuggets are serendipitous, a result of a sideshow to a main research effort, although the most dramatic cases tend to be more purely empirical, such as the discovery of penicillin.
Stuart, I guess I am implicitly assuming that anyone who has gotten through the necessary training can actually do the non-genius sized nuggets, so it doesn't matter whether or not failure there could be easily obscured. If you accept that failure can be obscured in the genius-sized nuggets, and you are prepared to accept as given that prestige will still attach to attempts of unknown success to do them (which does seem to be the case), then I think my point holds. This creates a bias in the allocation of talent, and it muddies the waters. What I'm not sure about is whether it introduces any systematic bias in economic results.
Stuart and Meredith,You are right that my proposed division into geniuses and non-geniuses is way too pat and misses a lot. But I think it does serve to make my point.
Alan: There's a big payoff in academic legal circles for inventing whole systems from scratch.
I certainly agree that there's a big payoff, but I think it's a misnomer to say that law professors "invent" whole systems from scratch. Rather, we seem to crib whole systems from existing liberal arts and science fields. (See, e.g., Seidman, Tushnet, Posner, Sunstein.)
The problem -- at least in America -- ultimately seems to stem from the admissions, grading, and academic hiring criteria. But that's a different discussion for a different time.
If you think it's bad in empirical economics, you ought to look at law. There's a big payoff in academic legal circles for inventing whole systems from scratch. Actually mastering a tough field tends to get you labeled as "nuts and bolts." Some (certainly not all, but quite a few) of the "big names" among academic lawyers are people I wouldn't trust to draft a simple will. I know one lawyer whose law firm fired him for incompetence and who then went into academe and wrote ambitious but nonsensical papers. His career aver since has been very successful.
Genius appears to be more about creativity than the "size of the nugget." Like the rest of us, most of them start off with one question: it just turns out that the question's solution is far more extensive than they had originally anticipated.
It would tend to be a bias towards studies where failure can be obscured more easily. Whether those will be simpler or harder models... well we had inconclusive fun with that here.
Also, figuring out if something is a genius size nugget can only be done in retrospect - once Einstein has announced General Relativity and we look at all the other endevours in the field and realise how far off they were,
There is error in the evaluation of success. Failure can look like success and success can look like failure. I don't yet see that this produces an overall bias to try problems that are too hard.
The other problem is figuring out if you actually are a genius or not. ;)