Monday, June 20, 2011

Daniel Kuehn: Subverter of Positivism!

I suppose...

Ryan Murphy posts this: "That seems like the important question in the most recent spat between Daniel Kuehn and Robert Murphy. Kuehn’s Keynesianism and RPM’s [no relation] Austrianism sound an awful lot like what Karl Popper would call a metaphysical doctrine. That is, if your theory doesn’t tell us that it is impossible to observe something in the empirical world, the “theory” is nothing more than an arbitrary way of categorizing whatever you see. RPM’s acceptance of such a label is defensible in the sense that many Austrians have acknowledged this and hand-wave the issue away because economics allows us “understanding.” Kuehn’s statement is quite a bit more subversive since it is far afield from the positivism of mainstream macro."

I think it's very important to distinguish between the way scientists talk about what they do and some buttoned-up epistemology you learned somewhere. Yes, positivism talk is all over the place. I talk like this too on occasion. Economists who talk in terms of "identification" implicitly have a sort of falsification in mind, after all, and all empirical economists (which is what I do for a living) are constantly obsessed with identification problems.

That's all well and good. But let's not confuse that positivist sentimentality and lingo with the epistemological logic that also bears the same name.

To quote Rorty, "We can tell you about justification but we can't tell you about truth - there's nothing to be said about it. We know how we justify beliefs, we know that the adjective "true" is the word that we apply to the beliefs that we've justified. We know that a belief can be true without being justified. That's about all we know about truth. Justification is relative to an audience and to a range of truth candidates. Truth isn't relative to anything. Just because it isn't relative to anything there's not much to say about it. Truth with a capital "t" is sort of like God - there's not much you can say about God - that's why theologians talk about "ineffability". Contemporary pragmatists tend to say the word true is indefinable, but none the worse for that. We know how to use it. We don't have to define it."

I ultimately think talk about proof is just a shorthand way of saying we're impressed by the evidence in its favor and we can't think of an obvious rival candidate to explain our observations. If we can formulate our theories in a way that is more precise and puts our theories to a more apparently stringent test, that's a good thing. We ought to keep on using the language of falsifiability because it's awfully useful language to use and awfully profitable to approximate. But as Marshall said, nature's action is complex. The hope for genuine falsifiability is a pipe dream. I think it's far better to have a clear, formal, meaningful theory that is specific enough to be interesting and to be persuasive when confronted with the evidence. Testing those sorts of theories rigorously enough is going to get us incredible useful understandings of the world. We can talk like positivists - that's fine. We can even operate on the mistaken impression that we are positivists. You might even catch me saying I'm a positivist as shorthand for the scientific method that I ascribe to. But ultimately, I think a lot of this epistemology is just word-games and not all that useful. What do we "know"? What an absurd question! Forget what we "know" and spend time constructing persuasive explanations for what we experience.

4 comments:

  1. In truth, I prefer to talk as a Bayesian. Falsifiability is just very useful way of framing issues since it is more powerful than verifiability. Are these word games? I don't know- Bayesian updating is a pretty straightforward way of thinking about things. I didn't look at the graph RPM alluded to, but I can tell you that if someone found in the future that from 2012-2017, there was a negative correlation between free trade and growth, even without controlling for anything else, it would force me to Bayesian update against it. It would make me think that the underlying fundamentals of the economy that make Krugman's international economics theoretically interesting but practically irrelevant may have been much more important, for instance.

    ReplyDelete
  2. Popper's main methodological insight is that one should avoid internal stuff and describe the way people are actually found in their surroundings.

    ReplyDelete
  3. "What do we "know"? What an absurd question!"

    This was the question that drove Descartes, Hume, and Kant to greatness and immortality. Epistemology is frightfully important for the intellectual (read: anyone that doesn't repair cars for a living) and it doesn't do you any good to treat questions of knowledge and knowledge acquisition with such flippancy.

    ReplyDelete

All anonymous comments will be deleted. Consistent pseudonyms are fine.