The Myth of Non-reductive Materialism
Labels: Jaegwon Kim, mental causation
posted by Victor Reppert @ 5:17 PM
17 Comments
Dangerous Idea 2
A blog to discuss the argument from reason.
Monday, October 01, 2007
About Me
- Name: Victor Reppert
- Location: Glendale, Arizona, United States
I am the author of C. S. Lewis's Dangerous Idea: In Defense of the Argument from Reason, published by Inter-Varsity Press. I received a Ph.D in philosophy from the University of Illinois at Urbana-Champaign in 1989.
Previous Posts
- Is non-reductive materialism an oxymoron
- Why only mentalistic monism or dualism will do
- Leopold Steubenberg's Stanford Encyclopedia entry ...
- An Obsevation I made on Rational Perspectives
- Hasker on how not to be a reductivist
- Some further responses to exapologist
- What makes arguments from reason argument from reason
- Reply to exapologist on neutral monism
- The argument from objective cognitive values
- My new way of putting the argument from reason tog...
Subscribe to
Posts [Atom]
17 Comments:
I am sympathetic to Kim's arguments, as I have always tended toward reductionism. Calling the position an oxymoron (and somehow getting that from the points about epiphenomenalism) was what I was arguing against in the previous post.
Say I am an externalist about propositional content. To what does saying 'Brain state X refers to Y' reduce? It can't reduce to neural events, as the content fixation necessarily involves a description of the world. There are an infinite number of different possible interactions with Y (direct, indirect via pictures, other people, sounds) that suffice to make X refer to Y. To many, an infinite disjunction is not a proper reduction base (it doesn't look anything like a standard lawlike generalization, for instance, but an ad-hoc ever-expanding list of possibilities with no real explanatory bite).
So you either end up with nonreductive physicalism, or perhaps try to save reductionism by describing the relevant interactions at a higher-level (e.g., in terms of information channels, a story that generalizes over the particular content-giving interactions). I am sympathetic to the latter, but the former strikes me as sane.
Note that even the nonreductionists think that token mental events reduce to physical events. But they are after a description at a higher level of generality, one that will generalize and extend in lawful ways to other and future tokenings.
All that said, I am sympathetic to the anti-anti-reductionist arguments, and think we can give a general description of content fixation in Dretskian fashion.
I, unfortunately, don't have access to that particular Kim article, but I am assuming he is making arguments about causality and supervenience, and how supervenience implies reduction. I am not sure, though, how he handles the type-token issues, and concerns about generalizability of stories about reference fixation. I seem to remember some good criticisms (that he agreed were good) that he needed to be more clear about whether he meant his criticisms to apply to mental state types or tokens.
Perhaps if you explicitly put your argument in terms of mental state tokens (as opposed to leaving it ambiguous whether you are referring to types or tokens) it would all flow more gracefully. Killing the claim that token mental events reduce to physical events would kill all but the most ridiculous naturalist (e.g., the 'common sense' naturalists may remain unscathed, but they are wrong anyway).
I think the reason Victor called non-reductive materialism an oxymoron (which I agree doesn't seem strictly right as a description), is because the result looks incoherent with the claims of atheistic materialism. An atheistic materialist who avers that active intentionality can't be reductively explained as merely reactive non-intentionality, is basically saying that an atheistic reality in principle doesn't and cannot explain our active intentionality. And yet... he's still an atheist. It's like admitting he can't get there from here, whlie still insisting it must be gotten there from here anyway.
{{Perhaps if you explicitly put your argument in terms of mental state tokens (as opposed to leaving it ambiguous whether you are referring to types or tokens) it would all flow more gracefully.}}
I'd be intrigued to see that attempt, too; but I have a minor suspicion that 'token' salts the scales too much in favor of physical events. Or isn't it standard operating procedure right now to treat the physical realities as tokens of the mental states? (This may be why Victor isn't being more specific about whether his argument refers to tokens or to types--the latter would similarly seem to be salted a little too much in favor of non-physicalism, via a connection with archetypes and ectypes to Platonic form theory.)
JRP
JP: I don't understand most of what you said, but I gave an example of what I consider a sane argument for nonreductive physicalism in terms of the infinite disjunction argument.
Functionalists also have a noneliminativist argument against reduction. IF a mental state type is fixed by its causal role, and different types of systems can have states witht the same causal role (e.g., silicon, carbon, etc), then the mental state is not reducible to the physical (unless you, again, allow a theoretically awkward disjunction as a reduction base).
This argument doesn't work against mental state tokens, as each token mental state is implemented in carbon or something. My attitude is more biological in inspiration: we can have limited type-reductions. E.g., mammalian minds work this way, but I can't say if this will apply to all minds. So I restrict my attention to thought in terrestrial organisms rather than 'all thinking things.' If we succeed with rats and monkeys, we'll have a better idea how powerful these multiple realizability arguments are.
Note I actually don't like multiple realizability arguments. Temperature is multiply realizable, but reducible. Carbon, silicon, oxygen gas, all have temperature.
So as I said I actually like the anti-anti-reductionist arguments, but they are not decisive, and nonreductive materialism seems a reasonable option. More reasonable, say, than theism. :)
Hi Jason,
Many non-reductivists follow Davidson in supposing that types are linguistic entities, not Platonic universals, so Physicalism could still stand.
Hi BDK,
The problem with token fucntionalism is that it leaves nothing in common between one mental instance and another. For example, when we ask 'what is pain', how would a token functionalist respond? He can't say it is the raw feel of pain that all instances of pain share, since he defines pain by its causal role. He can't say there is a common causal role in all organisms, since he has granted multiple realizability, and every organism is a different system which reacts differently to stimuli. Ie, the stoic may stand up strong in the face of pain, the child may wimper and cry, so causal input and output is not constant. Finally, he cannot say the physical realizer is constant, since he has granted multiple realizability. Ie, pain is C-Fibres firing in this organism, but green goop squishing in the alien. All that remains in common is the wrongheaded tradition of using a word in a certain way.
Rino: I said multiple realizability arguments against reduction don't work for tokens. I don't think a token functionalist exists. I have never heard of such a position and it may be inconsistent!
What I said was that even functionalist naturalists think that tokens are reducible. But what they are after is what all these tokens have in common, and once we think at that level reductionism (they argue) becomes unlikely because of the disjunction problem.
BDK: I have certainly heard of this position (functionalist token-state identity theory) . It's the one advocated by my old professors (they may have changed their minds since). The professors in questions are Peter Smith (of Peter Smith and O.R. Jones The Philosophy of Mind, CUP) and Peter Carruthers (author of Introducing Persons).
I think the position appears stronger when you concentrate on propositional attitude type mental states rather than raw feels. I've always taken this position as sitting nicely with the anomolous monism.
Rino: can you say a little more about why such a theory cannot give a functionalist answer to the "what is pain?" question? The functionalist wants to say "that mental state which causes avoidance behaviour". If the Stoic needs to "stand up strong" that shows his "other things being equal" reaction is avoidance.
Steve
Rino: Since no theory handles raw feels well, I agree with the latter of what you said.
Carruthers is a functionalist, that much I know, but a token identity theorist as well? I assume a type-token identity theorist, then? (e.g., in humans each mental state type is identical to some brain state token: pain types are identical to brain tokens). Token-token identity theory is just crazy, so he can't be that.
Well, Peter Carruthers was well known at Sheffield for endorsing positions that other people found counter-intuitive. However, due perhaps to my lack of study in Philosophy of Mind, I'm rather failing to see why token-token identity theory is "crazy".
I'd have thought that once you reject type identity theory, your "species types" could become as narrow as you like so that the only benefit of a type-token theory would be simplicity and fit with experimental data. This in turn suggests to me that you can't avoid any genuinely philosophical difficulties with token-token theory by moving to type-token theory ... and that if you want to defend the latter you need to be able to defend the former.
So ... why is token-token theory crazy?
Steve
Steve: for one it goes against multiple realizability: the same mental state (e.g. pain) can be realized in a ton of different brain state tokens...this seems crazy. Let's say you argue it isn't crazy: the same experience, you may argue, can be implemented by one and only one brain state.
Then we run into another horn on the mental side of the token. There is no species of mental state, pain, that we share. Your pain is my snarfudingle. Without mental state types we lose...mental state types! :) This seems awful.
I was able to find an article which I haven't read, but seems sympathetic to token-token functionalist identity theory by Jackson here. Ya' learn something new every day.
Nevermind: that paper is a functionalist type-type identity theory.
Hi Steve,
Typically a functionalist will define 'pain' in terms of both inputs and outputs, not output alone. So, pain is the state caused by tissue damage and causing winces and groans. The problem is that no system needs to have the same set of outputs to match the inputs. For example, the stoic has the input of tissue damage but the output of standing strong. In another system, the child has the input of tissue damage but the output of crying. So, the input-output is not the same in every system, so there is nothing to unify it. Kim talks about this in a paper from 1992 I believe.
@BDK: Oh. That isn't really what I had understood by token-token theory. I agree that would be mad. The position I had in view is roughly:
Functionalism: Types of state are defined by functional role. States playing the same functional role are the same type of state.
Multiple-Realization: That functional role can be played by stuff of many different kinds, and even by the same kind of thing in different ways (such as when, after a head injury, one part of the brain takes on the functions that are normally carried out elsewhere).
In principle, it could be that some functional role has never been played by the same underlying stuff twice. Indeed, depending on how strict you are as to what counts as "the same underlying stuff" (eg nanometres difference in the location of some neuron) it seems likely that the underlying state is never the same twice.
Does this count as what you call "type-token" theory? Perhaps so. Is that position also mad? I don't believe it myself, but I can see the attraction of it.
@Rino: I'm no defender of functionalism, and I don't know the literature very well, but your example doesn't do the work you are requiring of it. At least, not so far as I can see. Perhaps there are better examples? My problem with your example is simple. The stoic is impressive because he behaves differently, yes. But it isn't the actual behaviours which matter. The functionalist is surely interested in the behaviour which would result from a mental state "other things being equal". The stoic is impressive because he overcomes his natural instinct to wimper (or whatever) not because he has no such instinct. Or, if he has no such instinct, we might be tempted to say that the phenomenology of pain was indeed different for him. To be convinced I either need this example fleshing out and made more rigorous or a different example where your point is more immediately evident.
Steve
Hi Steve, Sorry that I have been unpersuasive to date. Again, I will refer you to Kim's 1992 article entitled 'Multiple Realizability and the Metaphysics of Reduction', especially Pg. 15-16 and 20-24. He explains it better than I.
He basically argues that pain as a functional property cannot be unified for two reasons: first, functional properties are physically realized, so they will obey what the physical objects are doing. Since the physical objects are multiply realizable, so are the functional aspects of pain. Second, pain as a causal property would be a disjunctive property, and disjunctive properties are not natural kinds (I believe BDK was making this point as well). For example, you seem to be saying that pain is avoidance behaviour, and avoidance behaviour is either running into the fray for the hero or avoidance behaviour is crying for the baby or avoidance behaviour is standing still for the stoic. What unifies 'avoidance behaviour' as a kind? All instances are different, there is no law upon which to form a generalization. Perhaps this gets to your 'other things being equal' issue. I guess the point is that all things are never equal in different systems. Agree/Disagree?
I don't know what to make of type-token theories. I haven't thought about it enough. I am getting way out of my league on the different versions of identity theory at this point!
@Rino: I agree, much depends on the "other things being equal" clause, at least when it comes to the behaviour.
However, as you point out, the causes are also relevant according to functionalism. Now the causes of pain are a little more unified and may fall more generally under the term "bodily damage". I also think you are probably reading "avoidance behaviour" rather too narrowly. Even the hero will avoid unnecessary pain during his rescue attempts, and the experience of pain tends to encourage certain courses of behaviour and discourage others. Of course, the functionalist also has a fair amount of evolutionary theory on his side when it comes to pain. If bodily damage didn't (via some intermediary mental state) generally cause creatures to avoid whatever it was that caused that damage, they probably wouldn't last that long.
No longer belonging to any academic institution, I have no way of getting hold of the Kim paper you refer to (unless I shell out for some appropriate essay collection or journal back issue), so anything you can say here would be appreciated.
Steve
Hi Steve,
Here is the clearest explanation that I have seen Kim give of the issue:
"...if the "multiplicity" or "diversity" of realizers means anything, it must mean that these realizers are causally and nomologically diverse. Unless two realizers of E[emergent property in the sense of higher-level, functional property] show significant causal/nomological diversity, there is no clear reason why we should count them as two, not one. It follows then that multiply realizable properties are ipso facto causally and nomologically heterogeneous. This is especially obvious when one reflects on the causal inheritance principle. All this points to the inescapable conclusion that E, because of its causal/nomic heterogeneity, is unfit to figure in laws, and is thereby disqualified as a useful scientific property. . . . The conclusion, therefore, has to be this: as a significant scientific property, E has been reduced – eliminatively"
- Kim, 1999, "Making Sense of Emergence", Philosophical Studies 95, Pg. 17-18.
I don't agree with Kim, but there is his argument.
Rino,
Hmm. Not sure what to make of this passage from Kim. The basic point he's making seems correct, but it doesn't seem sufficient to unseat functionalism.
I imagine that the functionalist will happily accept any wider sense in which the subvenient states are of the same type. More specifically, I would guess that functionalists imagine that all/most different subvenient states for a given supervenient state will have a similar structure even if that structure is composed from different stuff. I'm thinking along the "hardware/software" lines. The important causal features are at the software level, and many different hardware setups can produce the same software results.
As for the subvenient states not comprising scientifically acceptable "kinds" ... well no, I guess they wouldn't. This is why I see it as a form of anomolous monism and not a straightforward reductionism (Token-Token or Type-Token but not Type-Type identity theory).
To repeat myself: I don't buy the functionalist position. I'm just trying to give it a fair hearing.
Steve
Post a Comment
Subscribe to Post Comments [Atom]
<< Home