“Primacy of Theory” – Hmm. Dan Klein Chimes In

Dan Klein sent me his thoughts on the discussion that has been taking place concerning Human Action. With his permission, I am posting them here:

"Reading Emily’s post, I find myself unsympathetic with her distinction between “economic theory” and “empirical economics” (or “applied economic theory”).

Continued below the fold.
+/-

Where philosophers including Hayek and Polanyi—and Berkeley(?), Hamilton(?), Kant(?), Spencer(?), James (?), etc., etc. – say that all observation or perception emerge from higher-level cognitive goings-on (associations, mechanisms, categorizings, etc.), no one nowadays would disagree. I take this as the nub of Hayek’s “Primacy of the Abstract,” (in New Studies), a title that is regrettable in the “Primacy.”

Hayek is against “the assertion that the abstract presupposes the concrete” (p. 37). Fine: It is wrong to say “first comes the concrete and then comes the abstract”. But I think it is a mistake to take Hayek to be saying “first comes the abstract and then comes the concrete.” In speaking of babies and lower animals, Hayek makes it pretty clear that the abstracts evolve in relation to a species/life experience.

So when Hayek says “primacy of the abstract” what he means is something like the following: Abstracts are as primary as concretes. He is not saying that all abstracts come before all “concretes”.

Now, in Hayek’s paper, I don’t notice a single occurrence of the word “theory.” Hayek is essentially talking about things like a squirrel’s perception of a nut being emergent from higher cognitive goings-on.

Emily writes:
“The observation of all facts and phenomena in the world take place through a lens of theory. For Mises, there is no such thing as observation without prior theory.”

I object to this. We should avoid saying that squirrels theorize. Likewise, when I merely observe a baseball game, I am not theorizing. Theorizing is a form of human discourse (which could of course be discourse with oneself in one’s own head). Theory is an artifact of human discourse.

A useful way to think of “theory” is as explanation. Explanation is a matter of discourse – the squirrel does not explain.

Explanation implies an explanandum. An explanandum implies some facts.

Another useful way to think of “theory” is as articulate interpretation. That too implies facts to be interpreted.

“Factual” statements are presumed acceptable to all parties of the communication.

Sure, in any discourse situation, factual statements reside within what Hayek calls “abstractions”, but if the set of statements are regarded as “factual,” then those “abstractions” are not, within the discourse situation, theory.

(My take of facts and interpretation is given on pp. 3-5 of the following paper)

Emily expounds what she identifies as the “Misesian position in acknowledging the primacy of theory.” She champions “the primacy of theory.”

But, irrespective of whether we view “theory” as explanation or as interpretation, we must have facts to do theory. I don’t see a contest of “primacy” between facts and theory. If we must speak of “primacy” – the value of which I doubt – then, in theorizing, facts and theory are dually “primary.”

Any kind of theorizing must presuppose some facts. It is by recourse to the factual that we answer the questions: Theory of what? Explanation of what? Interpretation of what?

When Emily speaks of “the primacy of theory,” I am tempted to ask her, “The theory that you say is primary is a theory of what, exactly?” Notice how Emily shifts to: “a method for developing an [unlimited number of relevant refutable hypotheses].” A “method,” hmm. Has it turned into the “primacy of method”?

Within a discourse situation, neither of the following is true:
1. The facts determine the theorizing.
2. The theorizing determines the facts.

To say that you can always open up what are treated in one situation as “factual” to dispute, that you can change the discourse situation, is true enough. But such recursion will never bring you to a “primacy of theory.”