Join 3,572 readers in helping fund MetaFilter (Hide)


Help me find and explore the methodological pigeon-hole of my department
December 1, 2012 10:05 AM   Subscribe

Help me to philosophically pigeon-hole what I take to be the methodological views of many professors in my economics department. Then help me to explore that and other pigeon-holes.

My university department does a lot of work in macroeconomics. If I were to ask many of the professors of my department to succinctly summarize their methodological approach to macroeconomics, I believe that they would say something like this:

1. “We build models which are highly simplified representations of national economies, which may contain some false premises, but contains enough essential elements of truth that (we hope) will allow us to fit the data.”

2. “We reject models which are flagrantly at odds with the data.”

3. “We “rank” the models according to how they fit the data, although this is not always straightforward, since the models can fit the data on many different dimensions, and we are not always interested in all of the dimensions on which the model can fit. ”

4. “We can learn things about the phenomena we are interested in from this approach.”

My questions are as follows. (a). If you were to hear these sorts of statements, would you agree that that person in question is coming from a methodologically anti-realist and falsificationist perspective? (b.) If not then what is it philosophically/methodologically- if anything? (c.) Can you give me examples of other areas of science/social science where this sort of methodology is used? (d.) Your thoughts on/recommendations for sources of information on methodological realism/anti-realism in social science in particular.

Please note: I am fully aware of the tough old econ blogosphere out there, so no need to refer me to it. I am also aware that economics has gone through reputational wringer of late, and that deep divisions exist within the profession. I don’t really want to discuss economics per-se here – it’s the methodology above I’m trying to zero in-on for the purposes of this question. If you know something about epistemology in general, or in hard or social science I would particularly like to hear from you.

About me: A grad student in economics. Have read one book on philosophy of science recently, after a lot of implicit absorption of economists’ views on methodology, it was a real eye-opener. From it I think I got the basic basics of some of the major philosophy of science ideas, which is what has prompted this question.
posted by mister_kaupungister to Religion & Philosophy (10 answers total) 4 users marked this as a favorite
 
Have read one book on philosophy of science recently

Which one? Multi-model selection is pretty standard for most sciences for which falsifiability is not possible. It's one of the best approaches people have agreed upon as a method for making inference about complicated hypotheses. Are you looking for sources that better explain the machinery of model selection, or are you looking for sources that critique science as an oracle for truth?
posted by one_bean at 10:50 AM on December 1, 2012


These are very complicated questions. I have a decent idea of what "realistic" means for me as a philosophy/theology student. What you describe doesn't strike me as "anti-realist" from that perspective. But looking briefly at some economics references online, it seems they use the terms in a different way and from that perspective it may be that your professors are "anti-realistic" in an economics sense, but I'm not sure there's enough information here to tell.

This is heavily trod ground in economics. That it's heavily trod doesn't make it easier, and perhaps harder to sort out.

Christopher A. Sims, "Macroeconomics and Methodology" (1996) seems like a decent place to start... then pulling his references and looking for places that cite him.
posted by Jahaza at 10:50 AM on December 1, 2012


Why do you say anti-realist, rather than pragmatist?

Would your profs deny that the phenomena they posit (the theoretical entities named by their models) really exist? (That is what I think of, when I think of anti-realist.)

Or would they say "the picture of the phenomena that's given by our models is 'true-enough' because it fits the data, or it works, and that's the best we can say about anything in this domain; there's no such thing as an absolute fact about whether these phenomena exist outside of those considerations"? (What I think of as pragmatist, or maybe instrumentalist.)
posted by LobsterMitten at 11:13 AM on December 1, 2012 [2 favorites]


(Should clarify, I'm coming from analytic philosophy perspective, not from any knowledge about economics - so if the term "realist" has a special meaning in econ, I don't know about it.)
posted by LobsterMitten at 11:14 AM on December 1, 2012


That is the same approach used in ecological modeling, especially ecosystem modeling. Ecology and economics have a surprising amount of mutual sharing of modeling ideas, so reading in the ecological modeling literature might also be interesting to you.
posted by hydropsyche at 11:21 AM on December 1, 2012 [2 favorites]


To give some clarification about what I've read and how I understand realism/antirealism:

I'm not looking for critiques of scienctific knowledge as an oracle for truth. Model selection is closer to it, but with more of a philosophical flavour.

The book I read before was "Philosophy of Science, a Very Short Introduction" by Samir Okasha.

Okasha defines anti-realism as the stance which says we can only obtain knowledge of the observable world, and not about things which are unobservable.

E.g. philosophical debates over the particles predicted by physics. Okasha says that those that think they are real are realists and those that argue that we have no reason to believe that they are real (since they are not observable) are anti-realists.

My profs like models which are premised on (at least some) unrealistic assumptions. The "representative agent", for example, in which one optimizing individual representing the whole of the "household" sector of the economy, which in reality is made up of many individuals.

However, the representative agent is an input, it is a premise of the model. The "output" of the model is just a set of predicted time-series, based on these assumptions. So I think perhaps the place of the unreal "entity" is not quite the same as for the case with particles in the example in Okasha's book.

Perhaps this is a form of LobersterMitten's distinction between pragmatism and anti-realism.

Not sure though.
posted by mister_kaupungister at 1:11 PM on December 1, 2012


Actually, it might be useful to look at the Stanford Encyclopedia of Philosophy entry on Models in Science. That will give some good pointers for followup reading.

The Stanford Encyclopedia of Philosophy is generally a pretty reliable guide to present-day academic philosophy in the Anglo-American world, so you can click around through the related articles and feel reasonably confident about what you find. (Note that other disciplines sometimes use outdated versions of philosophical notions, so be a bit careful when carrying over terminology from philosophy to your field! Better to stick to the way these notions are used in your field, if you need to choose.)

It sounds like you are interested in "de-idealizing" the models your profs use, in the terminology of the SEP article?
posted by LobsterMitten at 1:28 PM on December 1, 2012 [2 favorites]


Ha, SEP has an entry on philosophy of economics which may help with these questions too!

Pointing to a few subsections that might be especially useful:
Contemporary schools of economic thought
Methodological questions: Abstraction, idealization, and ceteris paribus clauses in economics
Inexactness, ceteris paribus clauses, and “unrealistic assumptions”
Popperianism/Falsificationism
"Realism"
posted by LobsterMitten at 1:41 PM on December 1, 2012 [1 favorite]


Bas van Fraassen's Constructive Empiricism may be another place to look.
posted by jade east at 1:49 PM on December 1, 2012


Looks like constructive empiricism with a fictionalist attitude toward entities postulated by economic models.

The question, I think, is whether your profs would say that good fit of one of their models is evidence that the entities postulated by the model really exist. That is, suppose a "representative agent" model fit the data really, really well. Would that mean that there really was a representative agent? Answering "no" indicates a kind of anti-realism; answering "yes" indicates a kind of realism.

Constructive empiricism is realist with respect to the observable world, which it takes at face value. Constructive empiricists are realists with respect to tables and chairs, trees, people, and whatever else we can see, hear, taste, etc. with unaided human senses. The goal for constructive empiricists is to produce theories that are empirically adequate in the sense that they fit the phenomena, where phenomena are just those things we observe with unaided senses.

The relationship between constructive empiricism and pragmatism is not straightforward. On the one hand, there is a strong instrumentalist strain in pragmatism (following Dewey). And instrumentalism looks a lot like fictionalism in that it says roughly that a theory or postulate is only as good as its usefulness. On the other hand, there is a strong realist strain in pragmatism (following Peirce). For example, abductive inference, which Peirce called the logical basis of pragmatism, is generally rejected by constructive empiricists and accepted by realists.
posted by Jonathan Livengood at 4:28 PM on December 1, 2012 [3 favorites]


« Older UKTenantFilter: My heating bro...   |  Using a tablet in China. Buy h... Newer »
This thread is closed to new comments.