In Defense of Being a Specialist

In Defense of Being a Specialist

The most interesting and timeless writers in the social sciences are the ones who concern themselves with the foundations of civilization; who dare to ask big questions.

The craziest heterodox crackpots in the social sciences are also the ones who concern themselves with the foundations of civilization; who dare to ask big questions.

What’s the difference between the two?

The danger of asking big questions is that, at that level, there’s very little feedback from the world to a mental model. Once a significant investment has been made in constructing a framework, it’s almost always easier to adjust it to accommodate new observations than to start over from scratch. Lakatos called this the “protective belt” of assumptions. The more rarified your mental model, the more inclined you’ll be to tweak the assumptions in the protective belt instead of looking for a new framework.

The problem for someone who wants to tackle big questions, then, is: what’s the best way to make observational feedback more salient? Given some problematic fact, how does one know whether to ditch the mental model and start over, or to tweak the assumptions and keep going?

A similar, and nearly as important, problem is: for someone asking big questions, is there a way to distinguish credible and good-faith authors from dilettantes and free-thinkers?

One distinctive feature of timeless writers is that they never dive straight into the big questions. They always start from a launchpad of a much more mundane research program.

There are two reasons why specialists would be better equipped to answer big questions. First, in light of the feedback problem, a deep familiarity with the minutiæ of a research program provides the opportunity for inconvenient facts to put pressure on the researcher’s worldview – especially if these minutiæ are about as salient to him as the specifics of his grand unifying theory. The conscious mind is fundamentally inductive: it generalizes from specific instances, rather than deducing from general rules.1 A research program concerned with the minute details of some particular problem or field is the anchor that prevents a theory of civilization from floating off into irrelevance.

Second, quite apart from the fact that specialization helps one approach big questions, it also indicates that one is competent to do so as well. Specialization requires discipline, and it is fair to assume of a non-specialist asking big questions that they are not sufficiently disciplined to produce answers of much value. Obviously there will be exceptions – undisciplined specialists and disciplined generalists – but given the sheer number of practitioners vying for attention, specialization is an important signal to nonspecialists who otherwise would have no way to separate the theorists with enough conscientiousness and good faith to be insightful, from the dilettantes and freethinkers. In this case, being a specialist is credible precisely because it is costly!

It should be emphasized in closing that specialists approaching big questions doesn’t necessarily imply a balkanization of frameworks. If a grand unifying theory is indeed grand and unifying, it should be able to be approached from a great number of starting points. I myself have very little interest in public finance or public administration, but this starting point is precisely what makes Buchanan’s more interesting moral philosophy credible. It might seem wasteful to occupy the best social theorists with minutiæ of interest only to a limited set of specialists. In practice, however, I am even tempted to say that no exciting political or social theory is worth the time of day unless its author has first approached it from the angle of a mundane research program.

Next: Monetary theory as a Prototype for a Theory of Civilization


  1. This is not to deny that the mind can make sense of reality only via deductive cognitive first principles. Learned rules, however, at least in order to be amenable to articulation, must be generalized from specific instances.




Facebook Twitter Reddit Threads

Leave a Reply

More Content