What is "reductionism"?

[Robin]

As the term "reductionism" is being bandied about quite a bit, I thought I would put my understanding of it and invite others to do the same.

AFAIK there are two ways in which the term "reductionism" is used in philosophy:

reductionism: the belief that each meaningful statement is equivalent to some logical construct upon terms which refer to immediate experience

WVO Quine, "Two Dogmas of Empiricism".​

This concept is not a Materialistic one - in fact it was introduced by the Idealist Berkeley and promoted by the radical Empiricists in the Vienna Circle. (It is also common on this forum to confuse Empiricism with Materialism, in fact most Empiricist philosophers tend to reject Materialism).

The concept was more or less killed off by Quine in the early 1950's.

The other meaning of reductionism is a way of describing the way in which the various branches of science relate to each other. When the term Reductionist Materialist is used it is mostly in this sense although not all reductionists in this sense are Materialists - Stephen Hawking for example, states that he is a Reductionist, but not a Materialist.

A common way to state this is to say that one branch of science is reducible to another, or that one kind of theory is reducible to another, often by a series of trans-theoretic identities.

Not all reductionists agree about how reductionism will proceed and even if there can be a complete reduction to a single theory.

The most controversial issue is whether psychology is reducible to neuroscience (and presumably therefore to physics), and this obviously has relevance to the philosophy of the mind.

Even if a reductionist believes that psychology is reducible to neuroscience, this does not imply that they want psychology to be abolished and replaced by neuroscience (that would be eliminativism).

Reductionism, in this sense, is in contrast to Physicalism (in it's original meaning at least) which was an attempt to unify science by devising a common language in which statements about a range of science could be placed, without making any assumption about how the hypotheses might relate to each other.

 

[Madalch]

I thought it was the opposite of oxidationism.

 

[Rairun]

When I say I'm a reductionist, I mean that things emerge from the interactions that take place among their parts. I don't think a house is more than bricks and mortar arranged in a very specific way. I also don't think 'personhood' exists as a thing in itself; it's just a word we use to label a complex system composed of smaller parts.

But strictly speaking, I'm not a reductionist. I don't think that current biology can necessarily be reduced to current physics. The main reason for that is that science doesn't describe the world as it really is; even when theories work, even when they have predictive power, it doesn't necessarily mean that they are compatible with other theories that are predictively successful. So, while I believe that sociology, psychology, biology, chemistry and physics can be unified, I don't think the higher level theories can necessarily be reduced to the lower level ones. We might need to tweak all of them in order to accomplish that.

 

[Robin]

When I say I'm a reductionist, I mean that things emerge from the interactions that take place among their parts. I don't think a house is more than bricks and mortar arranged in a very specific way. I also don't think 'personhood' exists as a thing in itself; it's just a word we use to label a complex system composed of smaller parts.

Yes that is another common use of the word that I missed.

It is more or less my position too.

But strictly speaking, I'm not a reductionist. I don't think that current biology can necessarily be reduced to current physics. The main reason for that is that science doesn't describe the world as it really is; even when theories work, even when they have predictive power, it doesn't necessarily mean that they are compatible with other theories that are predictively successful. So, while I believe that sociology, psychology, biology, chemistry and physics can be unified, I don't think the higher level theories can necessarily be reduced to the lower level ones. We might need to tweak all of them in order to accomplish that.

I think the idea of science-theoretic reductionism is that a completed biology could be reduced to a completed physics.

I wouldn't pretend to be an expert on it.

 

[Wowbagger]

Even if a reductionist believes that psychology is reducible to neuroscience, this does not imply that they want psychology to be abolished and replaced by neuroscience (that would be eliminativism).

It would also be called "Greedy Reductionism".

There is a saying in acadamia: All models are inaccurate. Some models are merely more useful than others.

In science, we use whatever models are appropriate for the job.

That does not mean we can't think about how those models can be reduced to more fundamental ones. Sometimes, some teams might even specialize in the link. If it adds value, then it is a positive application of reductionism.

But, for the most part: The findings at one level work (or not) independently of anything going on at other levels.

 

[Hux]

I sometimes have the feeling that reductionism is wrong. What is wrong with being a reductionist?

 

[Dancing David]

You take a liquid , like wine or orange juice and you heat it to a slow boil, and that is how you engage in reductionism. You make the base of a sauce called a reduction.



It is also a pejorative term for the scientific method.

 

[schrodingasdawg]

I tend to think of reduction in terms of what the philosophy rules out rather than what it asserts. Rairun mentioned a house being made of bricks and mortar. I would state my understanding of reduction as: the fact that bricks and mortar are arranged into a house don't cause them to behave in a fundamentally different way than they would if they were not arranged into a house.

The reason I try to work with this understanding of reduction is that a "whole" may be understood in a way that precludes defining the whole in terms of its parts. e.g. there are a large number of arrangements of bricks and mortar that could be called a house, and it's not really clear how to rigorously define a house as any member of a class of brick and mortar arrangements.

Applied to the mind and brain, this form of reduction would assert that an individual neuron (or an individual atom, perhaps) does not behave in a fundamentally different way than it would if it was not part of a brain. (Assuming it is subjected to the same external forces, of course.)

I've heard that something called supervenience principle has been suggested for similar reasons, but I'm not savvy on what the principle actually says.

It is also a pejorative term for the scientific method.

I've noticed that, actually...

 

[HansMustermann]

But strictly speaking, I'm not a reductionist. I don't think that current biology can necessarily be reduced to current physics. The main reason for that is that science doesn't describe the world as it really is; even when theories work, even when they have predictive power, it doesn't necessarily mean that they are compatible with other theories that are predictively successful. So, while I believe that sociology, psychology, biology, chemistry and physics can be unified, I don't think the higher level theories can necessarily be reduced to the lower level ones. We might need to tweak all of them in order to accomplish that.

Sort of. It really helps if you look at the abstraction levels inside the same field, before looking at how different disciplines fit together.

It's really like in the joke about the guys hired to make sure the Don's favourite horse wins, and the Physicist starts his solution with "We assume the horse is a point in a uniform linear motion..." It's really how Physics works. You have to find the abstraction level with just the details absolutely needed for the problem at hand, but discard all the details that would be lost in the decimals or are irrelevant at all.

If you just have to calculate how long a train will take from New York to Washington DC at 100mph, you're not interested in what model of train, its weight, its length, its engine's power, etc. The train is just a point for the scope of the problem. If you want to calculate the maximum slope it can climb, then torque, weight and friction start being relevant, but the train is still basically a point. If you want to make a train that can go at 200mph, then shape starts to matter too, and maybe even such detail as what kind of paint it's painted with. (Coarser paint creates extra drag at high speed.) Etc.

For each problem you have to find the right abstraction level.

Or as an equally trivial example, newtonian mechanics are really a subset of GR, but you don't apply GR when calculating in how much time a penny dropped off the top of the school reaches ground. Heck, you probably don't even really apply the inverse square law for gravity, but assume a constant gravity. Because the difference is really lost in the decimals.

It's really like that between domains too. There is no contradiction between, say, chemistry and quantum mechanics. The latter explains perfectly well why those atoms do what they do. But if you just want to know why Na burns if you put a piece of it in water, you just really don't need that level of detail.

And then some stuff falls in between abstraction levels. For really complex mollecules, the usual approximations may no longer be enough. If you want to know how or why proteins fold like they do, you end up needing to figure out the minimum potential energy configuration, which is physics again. That's what Folding@Home does, btw.

 

[barrymore]

I think reductionism fits nicely with the sense that scientific disciplines are human constructs. We call physics, physics vs biology, biology because our understanding is so incomplete across the spectrum of "knowledge" that it makes sense to break these domains up and label them (for classification and specialization purposes).

Imagine the EM spectrum and say we could only detect visible light, gamma ray, and radio waves. At first, we might not understand how radio waves are related to visible light, if we are not able to detect microwave/infrared, and the same with gamma rays if we are unable to detect ultraviolet/x-ray. But as we become aware of these intervening types of light, it makes sense that we could use that knowledge to form a coherent theory on light that explains each type itself as well as the relation to each other (as we have). This is how I view the current discrete classification of scientific knowledge.

 

[HansMustermann]

Partially, but partially it's just different abstraction domains, really. Sort of how you might just call it "light" if you're just interested in what aperture to use to take a photo, or you might call it a wave if you have to explain the rainbow, or you might call it a particle (photon) if you have to explain something where that's apropriate, or you might actually have to treat it as both in a whole other situation.

 

[barrymore]

Partially, but partially it's just different abstraction domains, really. Sort of how you might just call it "light" if you're just interested in what aperture to use to take a photo, or you might call it a wave if you have to explain the rainbow, or you might call it a particle (photon) if you have to explain something where that's apropriate, or you might actually have to treat it as both in a whole other situation.

Yes, I agree that domains have usefulness for abstraction purposes. We use the terms light, wave, and particle in relation to EM because it allows us to direct and simplify our thinking, not because our knowledge is so limited that we do not understand how they relate. But, right now, we break down science into different disciplines due to the fact that we just do not have a great understanding of how they interrelate, and we are left to assume that they are "different" on some level.

Abstraction is more in line with what I think as classification. Even if we got to an understanding so great that we just "got" how the world worked, it would still not make sense to classify Science as one thing. Each human is limited to what he/she can study and think about, so there is still going to be classification merely as means to help organize our energies. But just because we study "biology" does not mean we would not understand that it is "biology" in the sense that we mean now.

 

[Beerina]

Knowing how 0-dimensional particles form composite subatomic particles isn't going to help explain how DNA works, much less why a chimp baby raised apart from a momma turns into a nerd.

 

[HansMustermann]

Actually, knowing how those atoms work is exactly key to understanding how DNA works or how the proteins encoded by it work, if we're to get past the current primitive stage of just copying existing genes or copying an antibiotic produced by a fungus. Or for example if you really want to understand why a prion can re-fold a healthy protein, getting into that level of detail is very much essential. Or if we're to ever take a less shotgun approach to cancer than the current destructive methods.

Because at that level, an undestanding at the level of "the carbon atom has 4 electrons to share" or even "UUU encodes Phenylalanine" just isn't enough any more.

Those chains of aminoacids work in a specific way precisely because they're folded, and the combination that breaks some other mollecule apart might be an atom from here and one from the other end of the chain, except they're folded so they're next to each other just right. And that folding has everything to do with, basically, physics.