David Papineau interviewed by Richard Marshall.
[Photo; Steve Pyke]
David Papineau is still roving in the deep philosophical waters even though he knows that he’ll never know everything. He keeps writing hard core books about his philosophical thoughts covering things such as physicalism and how come everyone isn’t a physicalist, substance and property dualism and Kripke’s worry that the mind brain identity is just contingent. He wonders why philosophers think there’s something wrong with just knowing the facts. He thinks about the nature of colour experiences, representation, and avoids mixing up methodological issues with metaphysical ones. He thinks about the significance of Schrodinger’s cat,about whether there are any special laws that are not reducible to physics and about the usefulness of ‘historical kinds.’ This is a deep water big beast from the philosophical depths: bangin’.
3:AM: What made you become a philosopher? Has it been worth it?
David Papineau: My first degree was in mathematics. That was great, but it didn’t help with many of the things that puzzled me. I became a philosopher because I wanted to understand everything, especially those things that didn’t make sense. And that has continued to be my philosophical motivation. That’s one reason I have such a roving philosophical eye— once I have figured out a philosophical topic to my satisfaction, I find myself moving on to new problems.
Has it been worth it? Absolutely. (I now realize that I won’t have quite enough time to understand everything—but that hasn’t stopped me wanting to understand as much as I can.)
3:AM: You are an ontological naturalist. You think that modern science makes some species of physicalism an irresistible position don’t you? Can you explain what your arguments are?
DP: It’s simple enough. Nearly everybody nowadays accepts the ‘causal completeness of physics’—every physical event (or at least its probability) has a full physical cause. This leaves no room for non-physical things to make a causal difference to physical effects. But it would be absurd to deny that thoughts and feelings (and population movements and economic depressions . . .) cause physical effects. So they must be physical things.
Note how this argument only bites for those things that do have physical effects. If numbers say, or moral properties, have no physical effects, then this argument gives us no immediate reason to say that they too must be physical.
You might want to ask—if there is such a simple argument for physicalism, how come everybody hasn’t always been a physicalist? That’s a good question, and there is a good answer. The ‘causal completeness of physics’ wasn’t widely accepted until recently. A century ago mainstream science was still quite happy to countenance vital and mental powers which had a ‘downwards’ causal influence on the physical realm in a straightforwardly interactionist way. It was only in the middle of the last century that science finally concluded that there are no such non-physical forces. At which point a whole pile of smart philosophers (Feigl, Smart, Putnam, Davidson, Lewis) quickly pointed out that mental, biological and social phenomena must themselves be physical, in order to produce the physical effects that they do.
3:AM: This is not an eliminativist position regarding the mind but is reductionist isn’t it?
DP: Yes—at least in the sense in which ‘reductionist’ simply means neither eliminativist nor dualism. Philosophers sometimes also use ‘reductionist’ more strictly, to mean ‘type-identities’ between mental and physical categories, and to exclude ‘non-reductive physicalisms’ like metaphysical functionalism. I’m not so sure that I am a reductionist in the strict type-identity sense. The issues here are messy. But I certainly a reductionist in the more general sense which is opposed to eliminativism and dualism.
3:AM: Substance dualism is a target of this approach isn’t it?
DP: Yes. But so is property dualism.
3:AM: Tim Crane, for example, might happily concede the arguments about substance dualism but not concede that this means no species of dualism can’t be sustained. How do you respond to that sort of challenge?
DP: Well, the causal argument I gave above doesn’t just imply that there can’t be a non-physical mental substance, but also that there can’t be non-physical mental properties. (Tim is always a bit cagey about exactly what he thinks at this particular point. I’m having dinner with him on Saturday and will press him about it.)
3:AM: Kripke has anti-materialist arguments at the end of his Naming and Necessity and you think he’s wondering how mind brain identity seems false even to people like yourself doesn’t he? How do you handle his challenge?
DP: Kripke says that physicalists like me can’t explain the ‘apparent contingency’ of mind-brain identities. He maintains that, if I really believed that pains are C-fibres, then I ought no longer to have any room for the thought that ‘they’ might come apart. His argument is that, since pains aren’t identified via some contingent description, but in terms of how they feel, I have no good way of constructing a possible world, so to speak, where C-fibres are present yet pains absent.
(For the experts, note that I’m here reading Kripke quite differently from the widespread ‘two-dimensionalist’ reading which takes him to be saying that the problem for physicalists is simply that mind-brain identities are a posteriori. This seems to me an absurd misreading of Kripke.)
My response to Kripke is simply to point out that mind-brain identity claims are very counter-intuitive. They continue to seem incredible even to committed physicalists like myself. And that is why I go on half-thinking at an intuitive level that there is a possible world with C-fibres and no pains. I simply haven’t fully freed myself from the dualist intuition that even in the actual world pains involve something more than from C-fibres. So of course I intuitively think that they might come apart in other possible worlds, even if they contingently co-occur in the actual world. (If pains are extra dualist states ‘generated’ by brain states, courtesy of the contingent laws of nature operating in this world, then it immediately follows that those brain states could occur without the conscious states, in a world with different laws of nature.)
In truth, as Kripke points out, a clear-headed physicalist shouldn’t be thinking any of these dualist thoughts. If pains are one and the same as C-fibres firing, then there really isn’t any possibility of having ‘one’ without the ‘other’. Once you properly appreciates physicalism, this dissociation should cease to appear possible—C-fibres with pains should strike you as no more possible than squares without rectangles.
From my perspective, then, Kripke’s ‘intuition of contingency’ isn’t a thought that physicalists are somehow required to continue respecting even after they have embraced their physicalism. Rather it is simply a manifestation of the psychological difficulty of fully embracing physicalism in the first place.
This is a very straightforward response to Kripke, one that cuts through the huge literature on the ‘explanatory gap’ and two-dimensional semantics. This whole literature is motivated by the idea that there is something deficient about our current theoretical understanding of the mind-brain relation, and that therefore we need some different and deeper perspective that will somehow render mind-brain identities transparently true. I say that there is nothing deficient about our current theoretical grasp of mind-brain identities. The problem is only that they are counter-intuitive. This doesn’t show that there is anything wrong with our theoretical understanding, any more than the intuition that the Earth is at rest shows that there must be something theoretically wrong with Copernicanism, or the intuition that time is moving shows that there is something theoretically wrong with the block universe ‘B series’ view of change. (A hankering for ‘transparent understanding’, ‘grasp of natures’, ‘having things revealed as they are’, and so on, seems to run through a lot of current philosophical debate. I don’t get it. What’s wrong with just knowing the facts?)
Of course, there remains the question of why we should find mind-brain identities so persistently counter-intuitive, if they are true. But this is a simple psychological question, and there are a number of plausible explanations. Indeed this is a topic that is quite extensively discussed outside philosophy, by developmental psychologists and theorists of religion among others, under the heading of ‘intuitive dualism’. It is rather shocking that so few of the many philosophers working on ‘the explanatory gap’ are familiar with this empirical literature.
3:AM: While we are on conscious experience, you deny that we can really see a million colours. This is territory that Pete Mandik is also looking into isn’t it? You offer an alternative to the orthodox view, and then argue that phenomenological scrutiny isn’t going to help decide which is right. So first could you set out the two views?
DP: Colour experience is a new topic for me. I’m not sure how closely it relates to my previous work.
The orthodox view of colour experience assumes that, when we see a colour difference between two surfaces viewed side-by-side, this is because we have different responses to each of the two surfaces viewed singly. Since we can detect colour differences between something like ten million different surfaces, this implies that we are capable of ten million colour responses to surfaces viewed singly.
I don’t think that we are capable of anything like this many possible colour responses. Instead I argue that the perception of colour differences between two surfaces viewed side-by-side is a gestalt phenomenon. There is a brain mechanism that works to identify colour differences directly, without first identifying the absolute colour of each surface. So on my view there is no reason to suppose anything like ten million colour responses to surface viewed singly.
I think my view is rather more radical than Pete Mandik’s. Both of us want to show that colour perception doesn’t transcend what can be conceptualized, but I don’t think he goes so far as to deny that it doesn’t involve different responses to all the discriminable surfaces.
On the methodological issue, I think that would be hopeless to try to adjudicate between my view and orthodoxy by appeal to phenomenological introspection. We need to know about brain mechanisms.
3:AM: So why does phenomenological scrutiny not help? Does this relate your enthusiasm for phenomenal concepts? You’ve written about Wittgenstein’s Private Language Argument and argued that phenomenal concepts are inconsistent with Wittgenstein. So who wins—Wittgenstein’s argument or phenomenal concepts? Or both?
DP: The ‘phenomenal concept’ issue is rather different, I think. Here the question is whether there are concepts of experiences that are made available to subjects solely in virtue of their having had those experiences themselves. Is there a way of thinking about seeing something red, say, that you get from having had those experiences, and so isn’t available to a blind person? Many contemporary philosophers would say ‘yes’, despite the fact that such concepts seem to conflict with Wittgenstein’s private language argument. I have written a paper arguing that phenomenal concepts do indeed conflict with the private language argument, and that this is bad for Wittgenstein.
Still, as I said, this issue about phenomenal concepts is different from your previous question of whether we can decide the structure of colour perception by phenomenological introspection. After all, in supporting phenomenal concepts I am in a sense siding with introspection against the more behaviourist Wittgensteinians. But even so I don’t think that introspection is powerful enough to resolve the specific issue about how many colours you can see.
3:AM: So what approach do you recommend?
DP: As I said, I don’t think that we can figure out what is going on in conscious colour perception just by phenomenological introspection. We need to know about brain mechanisms as well. We need to figure out what information is present in the mechanisms that constitute conscious colour perception. If neuroscientific research shows that those mechanisms only contain comparative information about colour differences, and have ‘thrown away’ more fine-grained information about the absolute colours of single surfaces, then that would support my position, in a way that just introspecting our colour experiences can’t.
The use of neuroscientific data to help resolve phenomenological questions is proving a common theme in much contemporary thinking about the mind. How rich are the contents of visual perception? Does vision only tell us about shapes and colours, or does it also represent higher categories like lemon or umbrella? Again, when we view a scene fleetingly, do we consciously see all the details even though we don’t retain them, or do we not see them in the first place? Neurological information is crucial to deciding these questions. After all, they are so interesting precisely because unaided introspection cannot resolve them. Rather we need to know what is going on in the brain activities that constitute visual awareness.
Of course, without any appeal to introspective phenomenology at all, we couldn’t get started on this kind of analysis in the first place, since the initial identification of ‘the brain activities that constitute visual awareness’ must depend on correlating brain processes with phenomenological reports. But we can engage in a kind of useful bootstrapping here. First we use uncontroversial aspects of introspective phenomenology to figure out which brain activities are in general responsible for visual phenomenology and other features of consciousness. And then we use the neuroscience to tell us what information is present in those brain activities, and so to decide the trickier questions about the structure of consciousness. We start and end with phenomenological data, but we couldn’t have completed our inferential journey without the detour through brain science.
3:AM: Does this relate to your anti-conceptualism about psychological representation? Don’t you want to defend your views about psychological representation as scientific reductions, rather than as results of conceptual analysis?
DP: I think it helps to distinguish the local semantic question about the specific representational contents of perception—what things do perceptual states represent?—from the more general meta-semantic question of the nature of representation—what it is for psychological states to have representational contents at all?
On the former question, I rather incline towards ‘conceptualism’, in line with my view of colour perception—I don’t think that we can represent objects and properties for which we have no concepts, not even in perceptual experience. In this sense I differ from those who defend ‘non-conceptual content’ like Michael Tye and Chris Peacocke.
But this local semantic question isn’t something that I have written about much, apart from my recent interest in colour vision.
On the general meta-semantic question, by contrast, I have written a lot, mostly under the heading of ‘teleosemantics’. And here—though this is an entirely distinct issue—I am very much inclined to be anti-conceptualist, in the sense that I think that the philosophical task (as always) is to come up with a synthetic theory that fits the empirical evidence, and not to analyse our a priori concept of representation or anything like that.
3:AM: You also look forward to reducing causality to probabilities as part of this same approach don’t you? And an issue here is avoiding mixing metaphysics with methodology. Can you explain this, and also why this is not a methodology issue but a metaphysical one?
DP: A certain kind of methodologically-minded philosopher of science is quick to read off metaphysical conclusions from features of scientific practice. Chemists don’t derive their laws from fundamental physics, so reductive physicalism must be false. Biologists refer to natural numbers in some of their explanations, so numbers must exist. I think that this kind of thing makes for bad philosophy. The relevant features of scientific practice often have mundane explanations which don’t point to any deep metaphysical moral. (Thus it would simply be messy and pointless for the chemists to essay physical reductions, or for the biologists to offer number-free explanations. It’s a weird kind of science-worship that views these practical considerations as clues to the nature of reality.)
Recent work on causation is a case in point. The metaphysical question is whether causal relations can be reduced to non-causal general regularities in some Humean style (though the modern Humean will work with probabilistic generalizations rather than deterministic ones). Now, methodological philosophers working on causal inference in practical areas of science (epidemiology, economics, agriculture, . . .) have observed that in practice causes are never inferred from probabilistic patterns alone. When scientists do infer new causal conclusions from probabilistic information, it is always against a rich background of prior causal assumptions. (‘No causes in, no causes out.’) And many philosophers of science then move quickly from this practical methodological observation to the metaphysical conclusion that causation must somehow transcend any Humean pattern of probabilistic generalizations. But this is not a good inference. Even if causation is at bottom constituted by patterns of probabilistic generalizations, there are obvious practical reasons for using prior causal knowledge to help identify new causes, rather than trying to work everything out from first principles every time.
The funny thing is that recent methodological work on causation itself opens the way to a successful metaphysical reduction of causes to probabilistic generalizations. I am thinking of ‘Bayesian Nets’. The Bayesian Nets literature shows that, for any arrangement of causes, there is a possible set of probabilistic relationships that entails that arrangement. (‘No causes in, no causes out’ is a practical precept, not a principled constraint.) Dan Hausman has written a terrific book—Causal Asymmetry—building a reductive account of causation on this basis, and I have written a couple of articles in the same strain. But as far as I know we are the only two people who read the Bayesian Nets stuff in this way. The reason, I suspect, is that nearly everybody else who works on Bayesian nets is a methodologist rather than a metaphysician, much more interested in the way science proceeds than in the nature of reality. And so they think that if scientific practice treats causes as irreducible, then that’s good enough for them. Still, as I said, it is a bad idea to run metaphysics together with methodology in this way. I’ve nothing against philosophers who are interested in the practicalities of science per se. It’s their metaphysical aspirations that irk me.
3:AM: Why don’t you want to get into the box with Schrodinger’s cat and what is the significance of this thought experiment?
DP: Schrödinger’s cat has a 50% quantum chance of coming out of the box alive and a 50% quantum chance of coming out dead. If you got in the box with it, the same would apply to you. So you really don’t want to do that.
I favour an interpretation of quantum mechanics (the ‘Everett interpretation’) according to which reality branches in any chancy quantum situation. On this view, Schrödinger’s set-up will give rise to in two future branches of reality, one with a live cat, and one with a dead cat—and the talk of ‘50% chances’ just indicates that the two branches are both equally real futures of the cat that originally entered the box.
Now, some philosophers have tried to make trouble for this interpretation by arguing that, if it were true, then you would have no reason not to get in the box with the cat. For on the Everett interpretation you would be sure to come out of the box alive. True there’d also be a future in which you come out dead. But what’s so bad about that, given that you won’t be there to experience it, and that you survive happily in the only future that you will experience?
This is a terrible argument (and not made any better by David Lewis defending it at length in his last published paper—see my ‘David Lewis and Schrödinger’s Cat’.) There may be good objections to the Everett interpretation, but this isn’t one.
Everybody agrees that a future in which you are dead is a very bad thing, and that it isn’t made any better by your not being around to notice how bad it is.
Everettians will simply agree with this, and observe that it follows that it’s a bad idea to get in the box with the cat. Doing so will cause the universe to contain a future where you are dead, alongside the one where you are alive, rather than leaving it as a universe where you are alive in all futures. Since a future where you are dead is a very bad thing, you really don’t want to do that.
3:AM: It seems to many people thinking about such matters that many of our complex cognitive capacities are innate, but that raises the issue about how they could be? How could we have evolved an innate capacity to recognize doorknobs, say, given that presumably when we were evolving our minds we didn’t have doorknobs?
DP: I don’t have much use for the concept of innateness. The everyday concept incorporates a number of different notions that can come apart in in many ways, and as a result encourages a range of dangerously fallacious inferences.
Nor is it easy to tidy up the concept. I guess the best move is to try to equate innate with ‘not learned’. But this really only works as a necessary condition. It looks a bit dotty as a sufficient condition. (Is my newfound ability to sing innate, just because it wasn’t learned, but caused by that bang on the head last week?)
Even if we go with the idea of innate as ‘not learned’, I doubt that anything worth calling a cognitive capacity will come out as innate. This is because it seems unlikely that evolution would ever bother to write the whole of any cognitive capacity into the genes, so to speak, instead of allowing information from the environment to play at least some part in shaping it. Of course our genes will make some capacities very much easier to learn than others, and of course our genes themselves are not learned. But the point remains that genes themselves are not cognitive capacities, and that anything worth calling a cognitive capacity will depend to some degree on learning and so not be innate.
Having said that, I do have quite a lot of sympathy for Fodor’s picture of concepts as information-free atomic entities which get locked onto their referents causally, and to that extent they needn’t involve anything much in the way of learning. But even so it seems perverse to call them ‘innate’. Here we see again the oddity of treating ‘not learned’ as sufficient for innate. Even if no learning to speak of was involved in locking my mental term onto doorknobs, it is odd to say that therefore my possession of a doorknob concept is innate, just as it is odd to say that my head-injury-caused singing is innate.
3:AM: As a physicalist you’ll say that all laws are physical laws I guess. But a non-physicalist will say that there are non-physical laws – such as laws in economics, or biology and psychology. Fodor writes about these as ‘special sciences’. Do you think there can be special sciences?
DP: No, I think that there are non-physical laws all right: genuine (if not strict) laws written in the language of biology, economics, and so on. But I don’t regard that as a contentious issue. Even reductionists about chemistry will think that there are special chemical laws whose formulation makes essential use of chemical terminology.
The contentious issue is whether there are any special laws that aren’t reducible to physics. Fodor says yes, but I have always thought that there are issues here. It has always puzzled me, along with Jaegwon Kim and Ned Block, that there should be genuine lawlike patterns at the special level, if physicalism is true (as Fodor agrees) yet the special laws are ‘variably realized’ by different physical processes in different cases (as Fodor insists). Why should we always get the same results in the same circumstances, if what is going on at the physical level is so different in each case?
In a number of papers I have explored the idea that natural selection might fill the gap. Sometimes selection processes can ensure that there is always some mechanism to produce such-and-such an effect in such-and-such circumstances, even though that mechanism will be different in different cases. All territorial birds have some way of discouraging conspecific invaders, but the mechanisms vary (songs, displays, odours, . . .). Natural selection has ensured that each species achieves the requisite effect somehow, but it doesn’t care, so to speak, how the trick is done.
I still think that this story works in some cases, especially in the case of people learning skills and other social behaviours (individual learning is a kind of selection process). But more recently I have become interested in another possible source of variably realized special science laws. The idea is inspired by Ruth Millikan’s notion of a ‘historical kind’. Millikan observes that some categories—chemical compounds, clouds, stars–enter into a range of generalizations because their instances have a common physical essence. These are ‘eternal kinds’. But other categories enter into a range of generalizations because they are all copied from a common source. These are ‘historical kinds’. For example, all the many copies of the Bible have the same first word, the same second word, and so on. Each individual version of the Nuer belief system contains the same tenet about twins, about ancestral spirits, and so on.
I now think that many generalizations of interest in the special sciences have this kind of basis. Entities share properties not because of any common physical basis, but because of copying mechanisms. And, because of this, these generalizations may well be variably realized at the physical level (think of all the physically different versions of the Bible, or the physically different Nuer brains). It is an obvious idea, when you think of it, and I’m kicking myself for having missed it all these years. I haven’t written anything on this yet, but my ex-student Marion Godman, now a postdoc in Helsinki, has some very nice papers on the subject.
3:AM: And finally are there five books (other than your own) you could recommend to the readers here at 3ammagazine to help them further delve into your philosophical world?
DP: Saul Kripke. Naming and Necessity. Everybody knows that this is a terrific book, but it may not be obvious to younger philosophers how much it reshaped the philosophical landscape when it came out. It contains two ideas—externalism about representation, and the difference between necessity and a priority—that just weren’t there when I started doing philosophy. Kripke persuaded us of both of these ideas pretty much single-handedly, and they now inform all serious philosophical writing.
Jerry Fodor. The Modularity of Mind. Fodor has become increasingly opinionated and eccentric, culminating in the embarrassment of his recent What Darwin Got Wrong. But his earlier work set the agenda for many philosophers of mind working in the naturalist tradition. The subject of cognitive architecture as we know it scarcely existed before The Modularity of Mind, and subsequent views on this topic still define themselves against that book.
Hartry Field. Science without Numbers. I wish more philosophers worried about the topic of this book. I don’t mean the ‘fictionalism’ about numbers, but rather the general puzzle about the relationship between abstract objects and the concrete world. How can references to abstract objects be important for understanding the concrete world? Field explores this question in connection with the physical applicability of mathematics, but his book also holds many lessons for those contemporary philosophers of mind and language who bandy ‘propositions’ about so freely without stopping to wonder what difference they can possibly make to concrete reality.
Let me finish with more two recent books that I learnt much from.
Peter Godfrey-Smith. Darwinian Populations. This book focuses on concepts that are generally taken for granted in philosophic thinking about natural selection, such as ‘organism’, ‘heredity’ and ‘reproduction’. By challenging these notions Godfrey-Smith brings out what is and isn’t essential to natural selection and opens up a fascinating range of new issues in the philosophy of biology.
Richard Holton. Wanting Willing Waiting. Holton distinguishes two notions of ‘weakness of will’—acting against your better judgement, and failing to stick to your resolutions—and shows that they are quite different. The book explores the latter idea, and uses a wide range of empirical studies to cast new light on such topics as will-power, temptation, addiction and free will.
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
First published in 3:AM Magazine: Monday, April 8th, 2013.