Analogies work really well with anything mathematical (including pure math). There, you're explicit about what the stuff is, and which parts are similar. Like modeling the spread of Islam in Spain using logistic growth, on analogy with adaptive radiation or an epidemic disease.

Then we fit the data to such a model and see how close the analogy is. In that case, I think the only tweak needed was adding a constant migration rate to the source of new Muslims.

"The same equations have the same solutions."

Agnostic said it, but model-building is a kind of analogy-making. Of course, unrealistic models are the bugaboo of social science.

I remember when I saw that information theory used formulas derived from the thermodynamic descriptions of volumes of gas, I thought that that was really far-fetched, but actually it was a deep principle.

The analogies that really get my goat are the ones used by emergentists/supervenientists. Ie, "the mind emerges from the coordination of neurons not conscious in themselves, like assembled water molecules have a boiling point but a single water molecule does not." I have always been 99% sure this analogy has got to be complete garbage. The phase transitions of water can be explained reductionistically, no sweat. If there were only one water molecule in the universe, it would still be possible to figure out its temperatures for phase transitions.

(I've never been able to find any ideas of non-reductive physicalism that make any sense to me, admitted rank amateur in philosophy though I am. I really don't think identity, awareness, or freedom can be reconsiled with physicalism at all, and I am not only agnostic rather than atheistic, but in fact have no real ontological convictions about anything.)

A bad ontology is worse than none. One guy has the idea that everything real is an arrangement of quarks. It's probably true, but it doesn't get you anywhere.

Ontologically, computer programs are just enormous strings of + or - bits, but when you're trying to understand what's going on on a computer, you're not going to go down to the bit level during your investigation. (If you have to go that far, presumably the computer is trashed). Thinking that Windows is a "thing" is pretty helpful if you're working with Windows.

I think there are two different common usages notions of reductionism. One is soft/metaphorical/pragamtic and the other hard/ontologic. The hard usage applies to, say, the belief that everything comes down to quarks and could in principle be derived from the theory of quarks. I don't deny that the boiling point of water is a "thing" in its way, but new theory is not needed in principle for understanding the boiling point; nothing new is needed at all. The theory of quarks would in principle have it fully covered, and I suppose the distribution of electron density maps for the (solitary) water molecule would by itself come damn close to determining the boiling point.

In the study of really complex phenomena that are hard to experiment with (say, history), people level the charge of reductionism in a slightly different way. They basically mean someone is trying to model something in terms of some sort of fundamentals, but that it is just inadequate - in your words, "unrealistic models are the bugaboo of social science." (All models are bad but some are useful, as the saying goes, but of course some are so bad they aren't useful.) This usage of the term reductionism doesn't really pertain to ontology - when a historian is accused of being a reductionist, whether he believes world history is at bottom simply the history of quarks, isn't really at issue. If I inveigh against some reductionist historian, I'm not complaining that he has a physicalist ontology or physicalist philosophy of mind.

Perhaps these multiple uses of "reductionism" contribute to people being able to swallow the analogical idea that the boiling point of water is a "supervenient property" existing only when multiple water molecules are assembled, and that the advent of awareness or freedom when neurons are assembled is barely any more mysterious than the advent of boiling point.

And yet I still I don't think I can really eviscerate the analogy to boiling point - I feel like I just know it's wrong and that this whole branch of philosophy of mind is no good, but I can't quite put my finger on it.

The phase transitions of water can be explained reductionistically, no sweat. If there were only one water molecule in the universe, it would still be possible to figure out its temperatures for phase transitions.

Maybe you should try it sometime and see how far you get. Not disagreeing with your argument in principle, just saying... water is actually quite a complex beast and you would be surprised at just how *much* sweat is involved in trying to model it from first principles.

Maybe you should try it sometime and see how far you get. Not disagreeing with your argument in principle, just saying... water is actually quite a complex beast and you would be surprised at just how *much* sweat is involved in trying to model it from first principles.

I'm certainly not capable of it. I definitely didn't mean to dis physics; I was just casually using a figure of speech.

Now that you mention it, maybe what I said could be wrong. First, I don't really have a clue whether you can really get a worthwhile prediction of the boiling point if you ignore the nuclear forces and gravity. Second, even if you can figure out the distribution of electric fields of the solitary molecule (and I'm not sure how that is done), I'm not sure that gets us to the boiling point. Humble biologist though I am, I do know that quantum effects apply on this scale, and I don't know any quantum mechanics.

Maybe I should just report back in ten years if I learn some physics and read more anglophone philosophy! But suffice to say, I agree with Razib that analogies are sketchy. I wonder if there is much more going for supervenience and nonreductive physicalism than analogies of this sort.

Proposing an analogy is proposing an isomorphism.

In math, the goal is to reduce to a previously solved problem.

In moral argument, the goal is to reduce to an accepted tradition, taboo, or "obvious" reaction.

heisenberg once said that, compared to turbulence in fluids, quantum theory is easy.

I just ran into a distinction between "constructionist" vs. "reductionist" theories. Reductionist theories say that ontologically "emergent" phenomena are the outcome of the interactions of lower-level phenomena, but deny that there's any real way to start with the lower level phenomena and construct the higher level phenomena. It's something of an informational way of looking at it.

It's in an article "More is Different", by the physics Nobelist P W Anderson. He did "condensed matter" physics and was annoyed by the claims that condensed matter physics is really just a consequence of fundamental physics on a large scale.

The article is in an otherwise rather disappointing book "Emergence", ed. Bedau and Humphries.

I think that the mystique of fundamental physics, the so-called "theory of everything", and the most sophisticated math have been an impediment to the serious study of anything more complex. Considering that Anderson feels that way even about the way plasma physicists think about the physics of solids, liquids and gases, this really isn't a reactionary humanist thing to think.

I should have said:

Reductionist theories say that ontologically "emergent" phenomena are the outcome of the interactions of lower-level phenomena, but do not necessarily claim that there's any real way to start with the lower level phenomena and construct the higher level phenomena.