Eric Funkhouser --- A Call for Modesty: A Priori Philosophy and the Mind-Body Problem

[New Waves in Philosophy of Mind]

A Call for Modesty: A Priori Philosophy and the Mind-Body Problem

Eric Funkhouser

While philosophy has conceded much ground to the sciences when it comes to investigating the natural world, many have held onto the mind-body problem as still presenting a distinctively philosophical problem. In particular, mind-body relations remain a subject for armchair speculations. The two most prominent arguments on the mind-body problem from the last two decades – Chalmers’ zombie argument and Kim’s exclusion argument – are primarily driven by a priori premises. I advocate specific objections to the a priori premises of each argument. More generally, I urge that such a priori approaches overreach the territory for which philosophy is properly suited. I conclude by suggesting a more modest role for a priori (e.g., metaphysical) reasoning

A PDF of the paper is ready to view and download in the attachment below. 

A direct link to the PDF: http://goo.gl/bzivU

[Graeme A Forbes]

I am largely sympathetic to your paper.

I was thinking about the distinction between a priori and a posteriori you rely on. C.I. Lewis has an interesting view of the a priori where for something to be a priori is for us to reach a decision about how we will interpret future experience. For example, to say it is a priori that 2+2=4 is to say that any time we have 2 things and 2 more things and get 5 things we shall reject that as a miscount or defective in some other way. Lewis then goes on to advise that the decisions we make about future experience get updated in light of the experiences we have, such that if making sense of experience would be easier with different a priori principles, we should adopt those.

This sort of view means that our view of the relation between empirical and a priori work should be much more integrated than is often accepted. (Grist for your mill, I hope). Thought experiments then taken on the role of exploring consequences (as you point out)  and exposing the presuppositions that we already had, but were unaware of. 

You say "To be sure, a priori speculation has something to contribute to the investigation of scientific kinds. But let’s be more modest here – it never settles the issue." With Lewis view in mind, we might describe the situation as follows: Decisions about how to interpret future experience settles how we will view natural kinds, but then evidence comes in saying that different decisions would make more sense of the issue. So on Lewis view of the a priori, the a priori is only meant to settle the issue in so far no experience will disprove it. The changing of our concepts to make something else an a priori truth is done on pragmatic grounds, rather than evidential grounds, precisely because the a priori consists, on his view, in decisions about what counts as evidence. So there is a doxastic sense in which the a priori always settles the issue, but a pragmatic sense in which the a priori never settles the issue.

I don't take myself to have come up with an objection, more to have put a (plausible?) view of the role of the a priori that might be useful to consider in thinking about the relation of the a priori to the empirical.

[Philip Goff]

On the big picture:

I agree that, in general, it is difficult to see how philosophy can make substantive claims about the nature of reality without some empirical input. But in the case of consciousness, it seems to me we have a non-observational source of data that must be respected by any metaphysical theory: the certain knowledge each of us has of the existence of phenomenal consciousness. Science can tell me my pre-theoretical concepts get things badly wrong: maybe our folk concepts of ‘time’ and ‘freedom’ are not strictly satisfied. But when I’m in agony, and I form a phenomenal concept ‘feeling like that’, I know for certain that that very concept is satisfied, and hence any metaphysical theory that cannot account for the satisfaction of this concept is certainly false.

We can then have a debate about whether the physicalist can account for the satisfaction of this concept. The type-A physicalist tries to do so by giving a causal analysis of phenomenal concepts; the type-B physicalist tries to do so by giving a semantic externalist account of phenomenal concepts. I reject physicalism as I think these accounts of phenomenal concepts are incredibly implausible. But even if you don’t agree with me on this, you must surely concede that there is a source of data which is insulated from empirical refutation: my concept ‘feeling like that’ is certainly satisfied (in contrast to my pre-theoretical concepts of ‘time’ and ‘freedom’).

On the specific details:

Do you mean to suggest that Chalmers may have an impoverished notion of the physical in taking the physical world to be pure structure/dynamics? If so, I’m not sure how this criticism is supposed to work, give that – as you acknowledge – he is open to Russellian monism. 

[TParent]

Dear Professor Goff,

On the big picture, I think much of what you say is interesting and important. However, it seems Funkhouser was allowing a role for the apriori--though perhaps the role you mentioned was not sufficiently emphasized (depending on how important you take that role to be.)

I wanted also to ask about your statement that, in a case of a phenomenal concept, "I know for certain that that very concept is satisfied, and hence any metaphysical theory that cannot account for the satisfaction of this concept is certainly false." Did you mean to say 'false' or just 'inadequate'? (I assume that a theory of mind could be incomplete in its explanations, but still contain only true sentences.)

Perhaps your point was directed at a physicalist theory of mind--and more specifically, a physicalist theory of mind which purports to explain *all* there is to explain about the mind. Then, I may well sympathize with your claim that it is certainly false...if only because it claims to explain everything. Nonetheless, in conceding an "explanatory gap," I'm not sure why an ontological, anti-physicalist thesis follows. Couldn't the gap just reflect a limit on what human beings can know? (Granted, I'm not prepared to defend such a suggestion...but is there any special reason why we should rule that out?)

[Justin Fisher]

Hi Eric,

Interesting paper. I definitely share your horror at the immodesty of much a priori philosophy. Here's a couple of quick questions.

(1) I was surprised that you didn't focus more on the question of how confident we could be that there isn't an appearance/reality distinction for conscious-experience. Such supreme confidence has always struck me as the height of philosophical immodesty! Worse, it seems to me that there are also plenty of uncontroversial examples: aspects of my conscious experience that, after having my attention drawn to them, I'm confident were there all along, even though they weren't apparent to me before. E.g., someone might point out that my experience of a light that had recently appeared steady to me was actually a gently pulsating experience, or that my experience of snow-shadows that had appeared pure grey to me before, actually was a bluish experience all along (this example is from John Pollock). Simple examples like these show that not all aspects of a conscious experience are immediately apparent, which invites the question of how many other not-so-apparent aspects of conscious experience might be lurking there, including perhaps even some underlying physical/functional essence...

(2) In note 11 (page 14) you suggest that in cases of apparent causal overdetermination by the mental, there might actually be distinct *effects*, so that these might not actually be causal overdetermination afterall. And you say this can be supported by the same sorts of multiple-realizability arguments that show that mental causes can be multiply realizable. I'd like to hear more about how this would work.

E.g., take the effect of there being a japanese maple tree at that particular spot. I grant that there's a purely physical explanation of this, and I'd also like to say that there is a psychological explanation for this (involving my desire to have such a tree there). I also grant that there are multiple ways that the psychological states in question could have been realized. But I don't see any corresponding ways that the tree's being there could have been realized. I do see that it could have been there and a slightly different shape, color, etc, but none of those seem to correspond to the ways in which my mental state was multiply realizable.

Perhaps you mean that there are a multitude of different fine-grained states that would have equally satisfied my desire, and that the psychological explanation really can explain only the general effect of having one-or-other of these configurations obtain, whereas the purely physical explanation can explain only some much more specific effect? What would you say about a case in which I have a maximally specific desire (or rather precisely as specific as physical explanation is, which might still allow Heisenberg uncertainty, etc...)? Would this be a case of overdetermination, or do you still think there are different effects to be explained?

Anyway, I'm curious to hear more about this (and am sorry for pouncing upon a tangential footnote).

Thanks!

-Justin

[Eric Funkhouser]

Thank you for reading the paper and for all the comments -- much appreciated! Sorry I’m a bit slow in responding.

 

Graeme: Yes, my understanding of the a priori here is rather naïve. I certainly think that the a priori has something to contribute to “empirical” questions. I don’t have a position on the C.I. Lewis proposal, though. (It reminds me of Grice/Strawson versus Quine on the dogma.)

 

Philip: I am willing to grant that when it comes to certain phenomenal mental states non-observational data must be respected, as you state. But I wouldn’t go so far as to claim that these data provide us with the essences of those mental states. The data could be respected as telling us that we are in those phenomenal states without informing us (decisively, at least) of their essences.

 

My point against Chalmers isn’t just the straightforward point that he might have an impoverished understanding of the physical as structural/dynamical. I think I had two main points against the zombie argument. First, we can have confidence in a metaphysical necessity on empirical grounds even if we don’t yet know how/why it obtains. Realization, I think, is best understood as a kind of metaphysical necessitation. And we can know (or be justified in believing) that a realization relation holds without seeing how/why it holds – like knowing that water is chemically realized even before discovering how this is so. Second, if we have some reason to think that phenomenal consciousness is physically realized, then the assumed lack of an appearance/reality distinction can then explain why either a) sameness of physical facts but without phenomenal consciousness is not ideally imaginable or b) it is ideally imaginable but not metaphysically possible. a) is even more palatable given that ideal conceivability is a pipe dream. And b) would not be an unacceptable exception to the conceivability-possibility claim, contrary to what Chalmers claims. Rather, the empirical grounds for believing it is a case of metaphysical necessitation (realization) combined with, critically, the lack of an appearance-reality distinction explains why this is a unique exception.

 

Justin: Yes, I am also skeptical of the claim that there is no appearance/reality distinction for phenomenal consciousness. Given the theme of my paper (immodesty), I should probably add this point. Thanks for bringing it up. I just wanted to grant as much as possible to the argument. But I also wanted to show that that very premise can also be used to help undermine the zombie argument (see my response to Philip above).

 

About the multiple realizability of effects. Take your example, where the effect is supposed to be a Japanese maple being at that particular spot. The cause involves, say, a desire that is multiply realizable – a desire that the tree be right there. What’s the psychological effect here? Well, the desire causes a series of actions – going to the nursery, digging a hole, planting the tree, etc. These actions are multiply realizable just as much as is the desire itself. I’m not inclined to say that microphysical “causes” caused these actions, because the actions were not sensitive to those precise details. (Here we might disagree, especially if you are more Davidsonian about the causal relata.) But, those actions are sensitive to his beliefs and desires – e.g., if he no longer wanted the tree there, it wouldn’t be planted there.

 

Now, you might want to grant me all this and still say that the tree’s being there is not multiply realized, even if the actions are. Well, his desires aren’t maximally determinate here. His desires caused the tree to be, say, in that general vicinity (but not in that maximally determinate spot). The microphysical facts caused the tree to be in that maximally determinate spot. I think that’s what I want to say. And if the desire were maximally determinate (very unlikely), I might accept this as overdetermination. But, I’d still point out that the intermediate actions in the causal chain that produced that effect (going to the nursery, etc.) were probably multiply realizable. Regardless, as you point out, this doesn’t bother me too much. Since I am not bothered by this kind of overdetermination, I wouldn’t have a problem if the true metaphysics of causation had (contrary to what I think) that consequence.

[Justin Fisher]

Hi Eric,

Just one other thought on that footnote about multiple realizability of effects.  Your view feels like it's in the same intuitive ballpark as contrastivist accounts of mental causation, which would say things like the following:  

1. The exact physical state of my brain caused the tree to end up where it is, rather than micrometers away.  
2. My psychological state caused the tree to end up where it is, rather than on the other side of the yard.  
3. But my psychological state did not cause the tree to end up where it is, rather than micrometers away.  

Contrastivists don't claim that there are different *effects* -- just different contrast classes for a given effect.  But if you can help yourself to some sort of vague, non-specific, or multiply realizable "effects", I can see how it might deliver roughly the same perks as a contrastivist account would.  (One potential difference will involve cases like the ones I asked about above, where someone has a maximally specific desire.)  Anyway, contrastivism seems very similar in spirit to what you were suggesting, so it's probably worth looking into if this is a project you're pursuing.

Cheers, 

-Justin

[Eric Funkhouser]

Justin,

Yes, I actually have a whole account of the causal relata, developed in a book manuscript on properties. I take properties to be the primary causal relata, so I think that the effects (e.g., psychological and microphysical) actually are different. But as you say, it is similar to the contrastivist's results -- I just build it into the ontology.

Eric

[Philip Goff]

Ted: I’m not just suggested an epistemic gap. The claim is that we have a concept that we know for certain is satisfied (phenomenal consciousness). This has ontological implications, namely that whatever demands the concept of consciousness makes on reality, those demands are satisfied. My overall view – although I have not argued for that here – is that there is no plausible account of the concept of consciousness such that its satisfaction is consistent with physcialism; therefore physicalism is false.

[Philip Goff]

But if Russellian monism is true, then the mental is realised/metaphysically necessitated by the physical (in terms of its intrinsic nature rather than its structural/dynamic features). So I’m not clear why, or even how, you are disagreeing with Chalmers. Do you have an argument to the conclusion that we ought to think that consciousness is realised by the structural/dynamic features of the physical?

[Ted Parent]

I apologize, Eric, if my discussion with Philip is more relevant to Philip's paper! Fwiw, I found your paper entirely convincing!

Thanks very much Philip for your reply. (I'm not entirely sure why Russellian monism got introduced into our discussion, so I might ask for further clarification before I try to address your latest post.) I am sympathetic with your view that it is possible for me to know for certain that a pain-sensation falls under my pain-concept. And I agree that a complete theory of the mind would need to explain all that. I realize that your complete argument may be beyond the scope of the present discussion, though I am hoping to ask one further question. But since I have a similar question vis-a-vis your paper, which I quite enjoyed reading, I'll go post my question over there!

[Eric Funkhouser]

Philip: I don't take a position on the mind-body problem, at least not on a priori grounds. One of my stated convictions is that it's odd to give an a priori (or even armchair) argument for such a conclusion. But, I have clearly stated disagreements with Chalmers' zombie argument.

I object to the argument by denying either the ideal conceivability premise or the connection between ideal conceivability and 1-possibility (in this special case). I support this by also noting that we could have empirical reasons for thinking that phenomenal consciousness is physically realized without knowing how it is realized -- e.g., without knowing whether it is realized in virtue of the structural/dynamical features of the physical or in virtue of Russellian monism (I'll let those chips fall where they may). Compare: we could have empirical reasons for thinking that water is chemically realized without knowing how it is realized. (Realization is a relation of metaphysical necessitation, as I understand it, and we can have knowledge of such necessities through means besides conceivability.) These empirical reasons could be our reasons for holding that zombies are not 2-possible. But, I am granting the lack of an appearance/reality distinction for phenomenal consciousness. Then, given that we can have reason for holding that zombies are not 2-possible, we can have reason for denying either that the zombie world is ideally conceivable or that, if it is ideally conceivable, ideal conceivability tracks 1-possibility in this case. Chalmers has claimed that it would be bizarre, though, to find a break between ideal conceivability and 1-possibilty, as we do not find this elsewhere. I counter his point here by noting that this would be a perfectly explicable exception -- it is explained by the lack of an appearance/reality distinction in the manner I just offered.

[Mark Sprevak]

Hi Eric! I really liked the paper! I’m really sympathetic with your big picture. I’m was just thinking about being devil’s advocate here trying to come up with a few random worries. Please ignore this if it is not useful!

My main worry was about the evidence that physicalism is the default reasonable position: i.e. we should accept that P metaphysically necessitates Q pending good reasons otherwise (reasons that Chalmers fails to provide on your view).

You give three reasons why we should have some prior commitment to the belief that P metaphysically necessitates Q. 

I was wondering if you could explain them a bit more?

1. ‘we should note that the relationship between P and Q does not appear to be like the causal connections that are prime examples of nomological necessities.’
   
   I wasn’t sure that causal connections really are our prime examples of nomological necessities. Consider laws in physics (e.g. F=ma). Aren’t these prime examples of nomological necessities? But those laws are hard to interpret as causal relationships (equality is symmetrical but acceleration does not ‘cause’ mass and force, the laws describe synchronic not diachronic relationships, etc.). These physical laws appear to be brute correlations between properties that hold, not accidentally, but in some modally robust way. Why can’t the relationship between Q and P be a modally-robust form of correlation like this?
   
   
2. ‘Second, realization is very plausibly a relation of metaphysical necessity. Given the possibility that phenomenal consciousness is multiply realizable in the physical, we might not want to identify Q with P (and so ground the metaphysical necessity in a claim about identities between rigid designators).’
   
    I had a hard time understanding this one. My very hazy interpretation of it was:
   
   
1. It is plausible that Q is multiple realisable (MR)—i.e. Q can be realised in various media, including P.
2. Our best account of how any X is MR is that the realisation relation obtains between X and the relevant media.
3. The realisation relation is a relation of metaphysical necessitation
4. Hence, we should think that P metaphysically necessitates Q
   
   Is that right?? If so, I guess I was worried that someone hostile to your position would disagree with it. For example, couldn’t they just deny that the relevant realisation relation is a relation of metaphysical necessitation, rather than, say, nomological necessitation?
   
   
3. ‘the simple fact that the physical sciences have had widespread success at eventually providing similar metaphysical necessities for a host of other contingent kinds — many of which were mysteriously realized, at one time — gives some reason to think that the physical-mental correlation is best explained as a metaphysical necessity.’
   
   This seemed to me the most likely one to motivate all sides in the dispute.
   
   But it uses an induction on past successes. And this induction only carries warrant if the current case (Q) is relevantly similar to past cases. The worry is that there seem to be plenty of disanalogies between Q and past successes of science for an anti-physicalist to mine: Chalmers’ point about structure and dynamics seems relevant here, also Phil’s point about the privileged role of introspection for Q.
   
   The worry is that this consideration just isn’t going to be motivation that will move all sides. It will only look as having probative value to someone who is already a physicalist. Anyone with non-physicalist tendencies will see the disanalogies as relevant reasons to refrain from performing your induction. Unless you can show that these disanalogies are relevant by their lights the reason you give will only be convincing for those who already accept your conclusion…?

But maybe I misunderstood your position, and you are not shooting to provide reasons that would convince a non-physicalist?

[Philip Goff]

Erik

ON THE BIG PICTURE (I forgot to post this part of my last comment (which was in response to your claim about the starting data being that we are in conscious states, but not what their essences are):

My starting point is not that we have data about the essence of mental states. My starting point is that we have concepts that we know for certain are satisfied. We can then have an argument about what those concepts are like, for example, whether or not they reveal the essence of the states they denote. But if you agree with me that we have concepts that we know for certain are satisfied, then we are in a very special epistemic situation relative to other metaphysical debates. Do you agree with this?

ON THE DETAILS

Denying (i) the ideal conceivability of zombies, or (ii) the move from their ideal conceivability to their 1-possibility is disagreement with Chalmers *only* if this is not done via the adoption of Russellian monism. So when you suggest in your paper that the failure of (i) and (ii) may be due to the fact that ‘Like Leibniz we have an impoverished conception of the physical’ (p. 7), you are in no way expressing disagreement with Chalmers (probably this isn’t important, I just found it slightly confusing).

On the more important point, it seems to me that at best we just have some considerations to balance (which I don’t think Chalmers would deny). We have good reason to accept (i) and (ii), but we may also have reasons to deny (i) and (ii). What is the argument that the reasons to deny (i) and (ii) are stronger that the reasons to accept them (Or do you think we have no reason to respect a priori judgements about what is ideally conceivable, in which case what is the argument for that?)?

Moreover, I don’t see how we have *any* reason to deny (i) and (ii) given that Russellian monism allows us to make sense of the realisation of the mental in the physical without having to do so. Therefore, the balance of considerations clearly leads us to Russellian monism rather than conventional physicalism.

[Bence]

Eric, great paper and it's good to see that there is a nice synergy between a handful of papers in the volume that all urge for a less a priori way of looking at the mind. I was especially pleased that your general recipe for how the a priori and the a posteriori could/should interact is very similar to the account of naturalism I give in my chapter. 

Two comments:

1. In Section 1, you suggest that the interaction between a priori and a posteriori should be a two way street. A posteriori should serve as the basis for a priori theorizing, but also vice versa. But in the conclusion (at the very end) you seem to go for a much weaker claim, according to which a posteriori is the input for a priori theorizing and that's the end of the story. I'm an enthusiastic advocate of the bidirectional interaction between the a priori and the a posteriori, so I was hoping that the official version is the former one...

2. This is about your footnote 15, where you are restricting the scope of your claim. I agree that we should exclude necessary kinds. But why normative kinds? I have heard this argument about how empirical stuff is completely irrelevant when it comes to normative phenomena, but, frankly, I never understood why this would be so. And this is a dangerous potential exclusion to make - with all these people arguing that pretty much everything about the mind (consciousness, belief, even perception, for crying out loud) is normative. 

[Eric Funkhouser]

Sorry for not replying sooner – busy grading final exams and papers, plus a job search. I’m glad the conference has been extended a week!

Mark:

Yes, I think you’ve put your finger on the critical issues for my position. What prior reasons do we have for thinking that P metaphysically necessitates Q? I also need to get clear as to who exactly is supposed to be my audience here – those who have previous commitments to physicalism, those who are genuinely undecided, or everyone? I really appreciate you pushing me on this – your comments are spot on. Let me at least say something about the very helpful points you make about my 3 reasons in favor of taking the relationship to be metaphysical necessitation rather than nomological necessitation.

1. You say that F=ma is a non-causal, nomological necessity, so why not think of the phenomenal consciousness case as being like this? I admit to not knowing exactly what position to take here. The obvious possibilities to consider are that F=ma is a causal claim or that it is a metaphysical necessity. While I don’t think most laws of nature are metaphysically necessary, this general law relating fundamental concepts might be such. I’m inclined to say that F=ma is metaphysically necessary (perhaps with the qualification that no non-alien forces are present). This is outside my expertise, though, so I don’t know.

2. What reason do we have for thinking this particular relation is one of metaphysical necessity rather than just nomological necessity? I really want to push this point more than the others. Isn’t it true that in other areas we have thought (or do think) that a relationship is metaphysically necessary rather than nomologically necessary, even though we don’t see why/how? We could have thought (or I do think) that the liquidity of water is metaphysically necessitated by its chemical properties without knowing at all how this is the case? I would challenge someone who claimed, instead, that there is merely a nomological connection between the chemical properties and liquidity. (To be clear, I am imagining this exchange taking place prior to the discovery of how liquidity is realized.) Why would we think this? Perhaps because we think there is constitutive dependence of liquidity on the chemical, and this suggests a stronger connection than just nomological connection. Similarly, few think that physical depends on the phenomenal (though, of course, there are some), but it is natural to suppose that there is a constitutive relationship in the other direction. Finally, the metaphysical necessitation picture is simpler than the nomological necessitation hypothesis, since it gets rid of nomological danglers – the phenomenal is there because it (really) must be there.

3. Why rely on the past successes of physicalism, given that consciousness is unique in not being a structural/dynamical property (supposedly) and also being open to introspection? I don’t have anything major or novel to say here. But gravity seemed unique and bizarre at one time. And so did life and a host of other phenomena that were eventually incorporated into the physical sciences.

 

Philip:

Yes, I grant the privileged access we have to knowing that phenomenal concepts are satisfied. (Does this rise to certainty? I don’t know – probably.) But, I don’t see how it at all follows from this that we are in an especially good position to know the essences of phenomenal kinds on a priori grounds.

“Denying (i) the ideal conceivability of zombies, or (ii) the move from their ideal conceivability to their 1-possibility is disagreement with Chalmers only if this is not done via the adoption of Russellian monism.” This is true. I take the zombie argument to be an argument primarily for dualism, though, with the Russellian monism thing added on as more of a caveat. Maybe I shouldn’t. Regardless, my point here is to defend physicalism (in whatever form) by running the argument backwards and showing that we have reason to deny either (i) or (ii) without this being an ad hoc strong necessity.

“What is the argument that the reasons to deny (i) and (ii) are stronger than the reasons to accept them?” Well, this hinges on whether we have independent reasons for taking the relationship between P and Q to be metaphysically necessary. This is critical, of course, and it’s addressed in the discussion with Mark, above. So, I think there are empirical reasons to think the relationship is one of realization (understand as metaphysical necessitation), and then I run the zombie argument backwards to explain why this isn’t an unexpected or ad hoc exception to the general connection between conceivability and possibility.

 

Bence:

Yeah, I’m not a fan of a priori input that is specific to the kind in question (as opposed to contributing general logical/metaphysical input that is domain-neutral), unless we are talking about math or logic (or maybe ethics). Also, I mainly had ethics in mind (in footnote 15), when I mentioned normative kinds that are within the domain of the a priori. I should maybe just avoid bringing this up. Ha. You very well might be right about the relevance of empirical input to epistemic or moral justification, say.

 

Thank you all for the very helpful comments!