Peter Godfrey-Smith interviewed by Richard Marshall.

[Photo: Stephanie Mitchell/Harvard News Service]

Peter Godfrey-Smithis the go-to guy in the philosophy of biology. He is forever evolving his thoughts on externalism, complexity and why we shouldn't expect a settled outcome, the contribution of pragmatists to philosophy of biology, why Fodor gets it wrong, on how best to understand what science is, on Darwinian theory, Darwinian populations, on why Richard Dawkins and David Hull are wrong and on the contribution of philosophy to biology. Like Cool Hand Luke, this one bites like a 'gator!

3:AM:What made you become a philosopher?

PGS:In my last few years of high school I had a very good English teacher with strong philosophical interests. His name was Alan Goldsby-Smith (no relation). He did not have philosophical training, and didn't talk about official philosophers like Locke and Kant, but he was passionate about the philosophical side of novels ¬– Solzhenitsyn and Conrad, for example. That gave me an initial interest. Then in the months between leaving high school and starting at university I read Douglas Hofstadter's Gödel, Escher, Bach. What an amazing book. I came across it through a newspaper review of the Dennettand Hofstadter collection The Mind's I, which had just come out (1982). I read GEB, very slowly, and then The Mind's I. So I'd read quite a lot of that sort of material before starting university.

The first philosopher I met at Sydney University was Kim Sterelny, then just a tutor, who was manning the department's "orientation" desk on the first day, answering student questions. We've been talking ever since. The Sydney Department at that time (or more precisely, one of the two Sydney University philosophy departments, this one known as "Traditional & Modern") was a fantastic place. David Armstrong was the chair, and he regularly gave first-year lectures. David Stove, Michael Devitt, Keith Campbell, and (as I mentioned) Sterelny were all there. Huw Pricecame in during my last year. One set of discussions taking place was in naturalistic philosophy of mind. Dennett's Brainstormsand Fodor'sRepresentationshad come out, and there was a lot of talk about how that cognitive-science style of work related to home-grown Australian materialism. There was also a lot of metaphysics; Armstrong was working on his book on possibility, while Devitt advocated a more Quinean approach to that whole area. The student society (the "Russellian Society") had evening talks followed by long parties, which most of the faculty attended. The students were very good – Fiona Cowie was in my year, Rae Langton a couple of years ahead, and Danieland Natalie Stoljar came through then, too. A lot of other smart and very philosophical people were around, from computer engineers to taxi drivers. Visitors from the US – Bill Lycan, Steve Stich – would come in for a term or two and shake things up. David Lewis visited regularly. It was a truly exceptional environment. After all that, staying with philosophy was an easy decision.

3:AM:In your first book, written nearly twenty years ago now, you raised the question as to whether it was possible to develop a informative philosophical theory of mind by linking it to properties of environmental complexity and went on to wonder about what externalist explanations in general were like and what issues do they generally raise. You think there’s lots of attractive things about externalism but wrote the book via an exchange with someone who was more suspicious about it. So first of all, could you explain what you mean by an externalist explanation of mind and what problems it hopes to solve that the internalist, mentalist explanations of the likes of Putnam, Armstrong, Lewis and Fodor can’t?

PGS:Externalism is the attempt to explain properties of systems like us – organisms, minds, living populations – in terms of features of their environments. It's an "outside-in" style of explanation. The discussions of externalism in my 1996 Complexitybook had both a diagnostic side and a more positive side. I think that lots of debates in biology, philosophy, psychology, and other areas are debates between people who look for externalist explanations and people who think there is generally something deeply wrong with those explanations. Rationalism versus empiricism; self-organization versus Darwinism about evolution. In part through the discussions you mention with Richard Francis, who was then working in fish biology, I became interested in the similarities seen in debates between externalist and more internalist views across very different fields.

That's the diagnostic side. On the more positive side, I defended a version of a general view about the mind: that the function of cognition is to enable agents to deal with environmental complexity. I saw that as one externalist view, which applies to a particular phenomenon at a particular time-scale, that is worth defending. Looking at the book now, I agree with its diagnostic side, and I also think there is something in the environmental complexity thesis (ECT), but I think I was wrong to see the ECT as worth defending in an externalist form. I'll say more about this later. What I want to emphasize here, though, is that I think it's an error to choose sides on the basic divide between externalist and internalist mindsets, and the 1996 book did not try to do that. You mention people like Putnam and Fodor as internalists. In some ways, in the context of some debates, that's right (though, of course, Putnam is in another sense an externalist – "meanings just ain't in the head"). But think of standard "evolutionary psychology," of the kind associated with people like Leda Cosmides, John Tooby, and Steve Pinker: they're Darwinian about evolution and anti-associationist about individual learning. They're "mentalists" in any reasonable sense.

3:AM:The issue has been hotly debated. Are there grounds for thinking that a settled view about the merits and demerits of externalism has been reached or is convergence of opinion still some way off? Where is the contemporary debate now?

PGS:I think we shouldn't expect a settled outcome about the merits and demerits of externalism in general. Sometimes explanations of this kind work, sometimes not, and an externalist view can apply at one time-scale but not at another. I'd again illustrate this with the evolutionary psychology examples I gave above. As I also said above, though, in one way my views are changing here.

I was recently reading Evan Thompson's book Mind in Life, which is very interesting and very ambitious, more ambitious than my book. Evan's book, which discusses life in general as well as mind and its evolution, is a whole-hearted endorsement of autonomy, self-development, and windowlessness. Living systems, for Thompson, "need to be seen as sources of their own activity, specifying their own domains of interaction, not as transducers or functions for converting input instructions into output products." This is a strongly internalist perspective. Reading his book, I found myself seeing a sort of mirror image of my own, or seeing my own in a new way. I think I was too inclined to think about the role of environmental complexity in a way consonant with externalist habits, just as Thompson is too inclined, I think, to approach the evolution of the mind in a way consonant with internalist habits.

Evan acknowledges, of course, that there is a huge amount of causal traffic across the boundaries of a system like a cell or a larger organism. But he wants to re-interpret the role of external factors so they are not really external; their "form" is imposed by the organism itself. I think that's a mistake, and I made mistakes of the same kind in my book. In my book I discussed ways in which the complexity of an organism's circumstances give rise to demands and opportunities that can be addressed by getting smarter – producing more complex behavior with the aid of more complex internal machinery. Of course, the relevance of this external complexity to the organism, and often the structure of the environment itself, will often be due to what the organism is like and what it did on earlier time-steps. I acknowledged that, but saw it as not really compromising the externalist character of the claim I was making about environmental complexity. Now I think that was an error, and the diagnostic part of the book is really at odds with some of the positive part of the book. I think the right way to think about the role of environmental complexity in the evolution of cognition is a way based explicitly on feedback and reciprocal influence. Once we do this, both externalist and internalist ways of casting the process look deficient.

3:AM:You’re a philosopher of biology so you look at specific biological models which treat properties such as adaptive plasticity and their variability in the environment. You intersect philosophy of mind with both philosophy of biology and a pragmatist epistemology (linked interestingly with Spencer and Dewey) that addresses the connection between thought and action. Can you say how these intersect in your approach?

PGS:Before pragmatism, philosophers in the western tradition had done very little with the obvious links that exist between belief and action – ordinary action, things we do to achieve everyday goals. It's surprising to me how wholesale this neglect was. Dewey eventually gave elaborate diagnoses of why philosophers had so much neglected these obvious links, diagnoses in terms of inheritances from the ancient Greek tradition, and so on. Whatever the reason, the neglect was real.

The pragmatists saw Alexander Bain, writing in the mid 19th century, as a precursor; he did treat belief in terms of its effects on action. Spencer, writing around the same time, also thought about these links, but in a more cosmic way, and a way the pragmatists came to react against. In any case, I see real progress in the pragmatist tradition on this point. Peirce, James, and Dewey developed theories of thought that did pay proper attention to the links between belief and action. They did so very imperfectly; Frank Ramsey, in the 1920s, developed those links in a more coherent form. Ramsey was influenced by Peirce much more than people realize (as Cheryl Misak is arguing in her current work).

Once you think about the links between thought and action in this way, it's natural to look to biology for models of the simplest, most basic forms of linkage between perception, inner states, and action. (Or at least, it seems natural to me.) In my 1996 book I discussed general models of adaptive plasticity in this light. At the moment I'm looking at some work on the early evolution of nervous systems, for the same sorts of reasons. This will involve a collaboration that is starting up with Gáspár Jékely (a biologist) and Fred Keijzer (another philosopher). I think that many of the philosophical themes standardly associated with "pragmatism" have quite close connections to cognitive science, especially to debates about mental representation and correspondence relations between thoughts and the world.

3:AM:A push-back against the position was recently given by Jerry Fodorwhich you vigorously disputed in a rather riveting exchange. As I saw it, Fodor’s challenge was this: if the associationist psychology embedded in behaviourist learning theory used a form of externalist explanations and if Chomsky and later Fodor showed it wouldn’t do to explain how we acquire language, then why do biologists continue to rely on such externalist explanations? Either Chomsky was wrong, or the biologists, or the parallel isn’t exact. No one thought it was a good argument, although it seems on the surface to have some merit. So why is Fodor wrong here? And if he is, does that mean Associationism as a way of explaining language acquisition is back on the table?

PGS:The short answer is that the parallel is not exact. Behaviorism of the sort Chomsky opposed is one externalist view, and an extremely strong one. Other views in the same family don't have the problems that Skinnerian behaviorism has.

Suppose you are someone like Steve Pinker. Then you think that Darwinism – which we can treat here as an externalist view – is true and those evolutionary processes built very complex machinery in our heads. That machinery enables us to acquire language, where this acquisition is not associationist, not a matter of operant conditioning, or any simple process like that. There is no internal tension in a view like Pinker's. Whether or not the view is right, it is coherent and it shows that an externalist program can be false on one time-scale (ontogenetic) but true on another (phylogenetic). Earlier I mentioned Cosmides and Tooby as further examples of people who are very adaptationist about evolution, but who think the day-to-day and year-to-year operation of our minds is quite different. A more extreme, and very interesting, case is seen in the psychologist Roger Shepard's views about psychology and the psychological basis of science. Shephard thinks that evolution has tuned the structure of our minds to fit deep patterns in our environments – patterns in the "external" world. But evolution has done this so well that pure self-propelled reasoning can achieve a surprising amount in science now.

Shepard is a kind of hyper-rationalist about intellectual progress, but he thinks the machine with which we do this thinking was built by Darwinian evolution. He's a Darwinian rationalist. I don't buy Shepard's view, but it illustrates some features of the landscape very well. Evolution was not bound to build brains which use the same mechanism of change that we see in evolution itself.

So Fodor is wrong about the consequences of the analogy between Darwinian evolution and operant conditioning. There is an analogy there – as was glimpsed by Spencer and James, and discussed by Piaget and Skinner, before Chomsky and Fodor – and the analogy is interesting. But it does not provide support to the anti-Darwinians.

Returning to your question, saying this does not put associationist views of language learning back on the table, or take them off the table. I don't know enough about psycholinguistics to have an opinion on that matter.

3:AM:What that debatedid was raise really interesting questions about the nature of science. There have been many attempts to answer this question. What’ do you think is the best way of understanding what science is, and which competing and initially plausible options does it preclude? Is the difference between ‘strategy’ contrasted with ‘method’ something you want to say is helpful, and is perhaps where critics like Fodor go wrong?

PGS:I do think that science follows a kind of empiricist strategy, even though there is not a recipe-like "scientific method" in the old sense. Science is a way of approaching questions about the world in which theories are exposed, as much as possible, to empirical testing. Questions in science are interpreted in ways that lend themselves to empirical testing. On this point I've been influenced by some classic work in the history and sociology of science – especially Steven Shapin and Simon Schaffer's Leviathan and the Air Pump. They see the message of their work as rather anti-empiricist, at least in the context of philosophical discussions, but I see it differently. One achievement of the scientific revolution was to establish new ways of adjudicating theoretical debates. These ways involve a role for social factors – such as the relations between competition and cooperation – and also involve habits and stategies with an empiricist character.

I don't see this as a central issue in debates between Fodor and myself, though. We do have different views about science in some ways – he is still wedded to a view of science as a search for laws. I don't agree with that. But even though Fodor makes a big deal of it, this isn't pivotal to the issues about Darwinism and learning.

3:AM:Darwinian evolutionary theory is a key to your work. You understand evolutionary theory using two central ideas: the ‘tree of life’ and a ‘Darwinian population.’ Can you say what you mean by these two terms?

PGS:These are two central ideas – for me, perhaps the central two theoretical ideas – in evolutionary theory. The "tree of life" is the structure of genealogical connections between all living organisms. If you look at an ant and an oak tree, for example, those two living things have a common ancestor if you go back far enough (in this case, over a billion years). The "tree" of life is only approximately tree-shaped, and perhaps soon we will be talking about the "net of life" instead, but this is an enormously powerful organizing principle in biology.

A "Darwinian population" is a collection of things which can undergo evolution of the kind Darwin described – evolution by natural selection. Roughly, it's a collection that exhibits variation, heredity, and differential reproduction (some things reproduce more than others). Why is this sort of collection so important? Because of what evolution by natural selection can do.

3:AM:How far does the addition of replicator approaches started by Richard Dawkins and David Hull to natural selection take us to developing a pure understanding of the evolutionary process itself?

PGS:My "Darwinian population" framework is intended as a replacement for the Dawkins-Hull framework. For Dawkins, evolution by natural selection is at bottom a battle between replicators, entities which are copied in high-fidelity ways and pass on their structure largely intact. For Dawkins, familiar living organisms like ourselves are vehicles which are built to serve the evolutionary interests of replicators. (Hull's version of the view is structurally fairly similar, but puts less emphasis on "interests" and the primacy of replicators as evolutionary agents.) I think the replicator-based view of evolution is wrong; replicators are not necessary to Darwin's process, and all that is needed is heritability, parent-offspring similarity to some degree. There are technical issues here, but the main point I want to make is not technical. The idea of evolution as a struggle between potentially immortal replicators is wrong; replication is one basis for heredity, but it's not necessary, and when we realize this, the status of objects like ourselves looks quite different. Humans are "Darwinian individuals" – members of Darwinian populations – as are cells, genes, and various other things. Other objects, like artifacts and ideas, can have at least a partial match to the requirements for being a Darwinian individual, and this is matter of degree.

3:AM:So what happens if we look at the world in terms of these?

PGS:These evolutionary concepts are only important to some parts of the world – I don't think they bear directly on fundamental metaphysics, for example. But I suspect they have more importance than is obvious at first. The "populational" mind-set has links to nominalism, as Ernst Mayr rightly saw back in the 1950s. And there are larger ideas, or larger pictures, that come out of the "tree of life" as well. There is a kind of unity that these tree-like connections give to the living world, for example. I've been thinking recently about life on earth as one large particular, one large connected physical object, linked by geneaology. I think this might be philosophically important in some way, though I'm not sure how. (Julian Huxley sketched a picture of this kind, in rather imperialistic terms, about 100 years ago; every organism "in spite of its separateness and individuality, is only a part of the single, continuous, advancing flow of protoplasm that is invading and subduing the passive but stubborn stuff of the inorganic.")

3:AM:I guess a question for philosophers like yourself is what’s the role for philosophy here? Why not let the biologists get on with their science and leave it at that? What can philosophers contribute?

PGS:It's fine with me if the biologists (and other scientists) get on with their scientific work without input from philosophers. Philosophy does often contribute ideas and theory-sketches to science, which then acquire a life of their own in the new setting, but this "incubator" role is a secondary role for philosophy. The same applies to the "clarification" of scientific concepts by philosophers. It happens, and sometimes it's helpful for scientists, and that's a good thing, but it's not central to philosophy. I don't think of philosophy as essentially a field that contributes to other fields. Philosophy is, roughly speaking, its own field, though it has a special status because it's so integrative – because the aim of philosophy is to get a coherent and defensible picture of everything going on. I very much like the one-line description of philosophy given by Sellars: philosophy is about "how things in the broadest possible sense of the term hang together in the broadest possible sense of the term." If we take this view on board, it implies that philosophy will always be interacting with the sciences and drawing on them, but it won't be swallowed up by them.

So I have no problem with scientists who do their scientific work while ignoring philosophy. It's a different matter when scientists start trying to answer philosophical questions, or trying to distill philosophical messages from their work. Sometimes they do this well, sometimes badly. Either way, then they are part of the philosophical conversation.

3:AM:You discuss the role of information and communication in living systems at all scales. Aren’t you stretching the terms communication and information somewhat if you claim even the smallest scale of biological system communicates and holds information? Are biologists being literal or are they using metaphors?

PGS:Sometimes it's a metaphor, yes. Arnon Levy has written a good paper about the role of metaphor in this part of contemporary biology. In the case of the idea of "communication," though, it's unclear where the literal ends and the metaphorical begins. My approach to this area is to make use of recent evolutionary models of communication (due originally to David Lewis, updated by Brian Skyrms) and try to give an account of how clearer, more definite forms of communication emerge from marginal forms – from interactions that are quasi-communicative in various ways. This is the same methodology I applied to Darwinian evolution in my Darwinian Populations book. Sign-like things mediate between senders (producers, writers) and receivers (consumers, readers), but these three roles are often only present in a partial way. Cell-to-cell signaling is an example. Memory is a broad functional category that is important in cell biology as well as in psychology, and I treat memory itself as communicative – roughly, memory is communication over time.

3:AM:As a leading naturalist philosopher epistemic issues are clearly important, but what are the metaphysical issues that loom large for you? I’m thinking of how biology and physics, say, seem to describe different realities. Is physics the fundamental reality and biological systems their emergent properties, or is Ladymanright to argue that there is no fundamental reality or level?

PGS:I'd like to think more about reduction and those issues you mention concerning patterns that exist at different scales – physical, biological, social. Ladyman's claims are interesting; I don't know whether to agree or disagree there. As far as metaphysics goes, I'm doing some work at the moment on the mind/body problem in its most basic form – qualia, consciousness, and so on – looking at connections to some ideas in the philosophy of biology. What is the connection between theories of life and theories of mind? I am a bit deflationary about the concept of life, but there is one side of living activity – the metabolic side – that might have important connections to philosophy of mind. In systems like ourselves, and all other actual-world cases we know of, "cognitive" processes are mixed in with a complex underlying set of metabolic activities. Some of those metabolic activities have a quasi-cognitive side to them, too, as they involve signaling, memory, and so on. Those are some connections I am looking into at the moment (and this is why I was reading Evan Thompson's book, Mind in Life, discussed above).

I'm doing this work within a materialist framework, and I don't find recent anti-materialist arguments at all convincing, but I think the difficulties in this area make it worthwhile to look also at "neutral monist" views of the relation between mind and matter. These are sometimes now grouped with "panpsychism," but I see those options as very different. The panpsychist thinks we have a good idea of what the mental and physical are like, can see that the mental can't be produced by the physical, and should conclude that entities at the fundamental level (some or all of them) must have a combination of both. The neutral monist thinks that what we crudely call "mental" and "physical" are manifestations of something else which is more basic. I'm a materialist, but I think neutral monism is the next best option (and the border between these two might not be very clear). There's probably a connection to think about here between the Ladyman-type work about "levels" that you referred to, which is guided by physics itself, and neutral monism.

3:AM:And finally, for the curious readers here at 3:AM, are there five books you could recommend (other than your own which we’ll all be dashing away to read) that will help us go further into your philosophical world?

PGS:Two books I worked on during my undergraduate years did a lot to shape my thinking, and I still go back to them: Fred Dretske's Knowledge and the Flow of Information(1981) and Ruth Millikan's Language, Thought, and Other Biological Categories(1984).

The third book I'll mention is much older. John Dewey's Experience and Nature(1925) is, for me, the high point of the pragmatist philosophical tradition (so far). It's difficult and convoluted, but I think there's a lot to learn from it.

Moving ahead to recent books, I'd choose Kim Sterelny's The Evolved Apprenticeand Brian Skyrms's The Stag Hunt. These are both very interdisciplinary books; Sterelny's is full of arguments about human pre-history, and Skyrms's is full of evolutionary models. But they're full of philosophy too. Sterelny's book has an important treatment of the relations between evolution and learning; he really does manage to get beyond a "learned or evolved" dichotomy in a way that makes a difference to questions about human culture. The Stag Hunt is one of three short recent books Skyrms has written about evolution and social behavior. This is the second of the series, perhaps less well-known than the others, and it has a rich discussion of the links between cooperation and communication. I mentioned above my use of sender-receiver models of the kind pioneered by Lewis. I see Skyrms's naturalization and generalization of David Lewis's 1969 model of communication as one of the most important developments in recent philosophy.

Richard Marshallis still biding his time.