Williamson on Naturalism

I’d be interested in hearing what our readers think of Timothy Williamson’s case against naturalism in this post at the NYT’s Stone blog (H/T NewApps). As I read it, his argument has the form of a dilemma. There are, he supposes, two ways that the thesis of naturalism can be defined:

  1. Naturalism might be the ontological thesis that the only things that exist are things of some sort.
  2. Naturalism might be the methodological thesis that the scientific method is the only legitimate method of finding out what there is.

If the naturalist adopts characterization (1), then Williamson’s response is that there is no adequate way to characterize the relevant “sort” of things: if the naturalist simply says that she means that there are only natural (or non-super-natural) things, then her definition is circular, and if she tries to give a more substantial characterization, then she is likely to be out of step with present or future science. There is then a natural temptation to move in the direction of characterization (2), perhaps defining the relevant sort of entities as the sort that science postulates. But this characterization is problematic unless there is an answer to the demarcation problem (i.e. the problem of drawing the line between science and non-science), and finding such an answer has proven to be rather difficult, to say the very least.

It should be noted that there is a large and ever-growing philosophical literature on these issues, and many believe that challenges like this can be resisted. It is nice, though, to have the matter summed up so succinctly. So what do you think? Are you a naturalist? If so, how do you respond to Williamson’s challenge?

P.S. One other thing to note here is that there are some forms of philosophical naturalism that aren’t harmed by this criticism, but only because they’re views of how to do philosophy, rather than views about what there is. For example, I am quite sympathetic with the idea that the methods of philosophy aren’t entirely discontinuous from those of the natural sciences (I’d call this “Quinean naturalism”, except that Quine was also a naturalist in the senses of (1) and (2)), and I have argued in my own work that the conclusions of apparently sound a priori arguments can be challenged on empirical grounds. At the same time, though, I think that philosophical theorizing has a place of its own, especially in places where experimental evidence isn’t (or, perhaps, can’t be) available. That is, I agree with Williamson when he says this:

The scientific spirit is as relevant in mathematics, history, philosophy
and elsewhere as in natural science. Where experimentation is the
likeliest way to answer a question correctly, the scientific spirit
calls for the experiments to be done; where other methods — mathematical
proof, archival research, philosophical reasoning — are more relevant
it calls for them instead. Although the methods of natural science could
beneficially be applied more widely than they have been so far, the
default assumption must be that the practitioners of a well-established
discipline know what they are doing, and use the available methods most
appropriate for answering its questions. Exceptions may result from a
conservative tradition, or one that does not value the scientific
spirit. Still, impatience with all methods except those of natural
science is a poor basis on which to identify those exceptions.

In this sense, however, the scientific spirit is not restricted to (what is usually called) “science”, nor is it prejudicial against the existence of entities of any sort, no matter how “unnatural”.


  1. I don’t see how it’s circular (to the point of intellectual failure) to define naturalism in terms of there not being a supernatural realm or any supernatural entities inhabiting that realm. For it seems like the supernatural realm has been sufficiently and concretely described throughout religious history so as to make a rejection of such realms & entities to be a respectable metaphysical worldview that is intellectual superior to and not simply “equally dogmatic” as supernaturalism itself. Thus, it strikes me as absurd to put the claims of the naturalist (“there is no supernatural realm; heaven and hell are fictive”) and the supernaturalist (“there is a heaven and hell along with demons, angels, and gods”) on the same epistemic footing. For whatever the naturalist posits in his worldview, it is at least discoverable by means of our experiences as fleshy bodies, whereas the supernaturalists claims are not even in principle verifiable because heaven and hell are described as being outside of time and space. To put a bronze-age theologian talking about supernatural realms on the same epistemic footing as a modern-day naturalist is absurd, particularly when the naturalist can, in principle, give a concrete account of the evolutionary and cultural factors involved in the very development of religious and supernatural ideas. Naturalism has recourse to the possibility that supernaturalists have simply been hallucinating because of inherited neurological systems, whereas the supernaturalists simply explains his/her experience in mysterious and unprincipled ways not amenable to the scientific spirit. For this reason, I vehemently disagree with Williamson that naturalism is somehow “antiscientific”. Quite the contrary.

  2. John Schwenkler

    Most of this is irrelevant to Williamson’s point. He agrees with you that heaven, hell, demons, angels, and gods do not exist. What he denies is that you can put (1) all these things together with (2) the entities postulated by non-Western religions plus (3) things like moral facts, beauty, and immaterial mental states within a single, well-defined category of “the natural”. To say that the problem is that such things are “outside of time and space” is clearly problematic, as it relies on applying these concepts in ways that our best science does not.

  3. John Schwenkler

    Yes, van Fraassen’s The Empirical Stance contains a much more worked-out version of this sort of criticism, together with a deeper and more controversial diagnosis of what has gone wrong and what an alternative position would look like. Of course one significant thing differentiating van Fraassen from Williamson is that he does believe in God.

  4. Eric Thomson

    His comments seem quite relevant to “if the naturalist simply says that she means that there are only natural (or non-super-natural) things, then her definition is circular.”

  5. John Schwenkler

    Sorry if I was unclear. What I meant is that Williamson isn’t disputing the claim that there isn’t good reason to believe in the truth-claims of various religions. It doesn’t follow from this, though, that the category of the supernatural can be given an adequate definition: for then we’d have to be able to define what a religion is, as well as defining which claims are proprietary to religions and which ones are not (i.e., can also be among the claims of “science”).

  6. When pushed by philosophers on these types of issues, I am usually able to avoid getting bogged down by excessive talking by clarifying what I am a (metaphysical) naturalist about. E.g., consciousness, I think neuronal processes are sufficient for conscious experience. By pretty much everyone’s definition, that is naturalistic.

    On the other hand, if my hand is forced, I don’t find the concerns in the post very persuasive. Even if ‘naturalism’ isn’t crisply defined, or delineated by a set of non-vague necessary and sufficient conditions, that’s OK. By analogy, that doesn’t dissuade me from saying something is/isn’t ‘science’ or that something is/isn’t ‘alive.’ 

    Like with both of these concepts (life, science), for ‘naturalism’ there seems to be polydimensional semantic space, most dimensions with many values (not just two). Certain corners are canonical instances of the terms, and there are fuzzy cases. Demarcation is possible, it is just nontrivial, and not everything falls clearly on one side of the fuzzy demarcation boundary. Some likely fall right into the fuzz.

    I consider myself a naturalist because I end up in an extremely deserted corner of this space where there are no abstract objects, no law-avoiding gods, and everything ultimately depends on configurations of atoms, energy, mass, and various fields in spacetime.

    At this point people bring up QM, and things do get weird down there, but within this space of QM interpretations there are more and less “naturalistic” views (from explicitly dualistic views of wavefunction collapse to those for which measurement is just another law-governed interaction between masses and fields in spacetime).

  7. John Schwenkler

    I consider myself a naturalist because I end up in an extremely deserted
    corner of this space where there are no abstract objects, no
    law-avoiding gods, and everything ultimately depends on configurations
    of atoms, energy, mass, and various fields in spacetime.

    So if an empirically-minded philosopher believes, say, that consciousness is an irreducible feature of material reality (e.g. Russellian monism), or an atheist metaphysician believes in possible worlds on the basis of philosophical arguments (perhaps including ones involving the proper interpretation of quantum mechanics), this counts as rejecting naturalism? And if the more “naturalistic” interpretations of QM turn out to be false (i.e., are disproven on empirical grounds), then is naturalism false, too?

  8. Eric Thomson

    My belief that consciousness is a brain process makes me naturalistic about consciousness. Whether a panpsychist is also a naturalist is a different question. I gave what I consider roughly sufficient conditions, not necessary. I
    can know that  my dog is alive even if I can’t tell you if a virus is

    More generally, telling someone you are a metaphysical naturalist is informative about your overall beliefs. It leaves open some questions, yes, but that isn’t to say it isn’t informative. It’s like saying that you rolled an even number greater than 3 on a 6-sided die. That’s pretty useful to know, even if it leaves some questions open.

  9. John Schwenkler

    I agree that describing oneself as a naturalist is informative, and I think that Williamson would agree. The question is whether the class of “natural” things is possible of being given even a minimally adequate definition. Put another way, the question is whether irreducible mental properties, angels, moral facts, numbers, properties, and possible worlds have something in common that they do not have in common with black holes, curved spacetime, and the occupants of the quantum realm. The only distinction I can think of is that things of the former sort are not posited by our best current scientific theories, while they latter are posited by them. (For note that on certain understandings irreducible mental properties and moral facts *are* spatiotemporally located.) But then there is the apparent impossibility of demarcating science from non-science, plus the fact that it is possible in principle that future science *could* postulate things in the first category, in which case the existence of those things wouldn’t disprove naturalism at all …

  10. Eric Thomson

    Let’s say that the class of “natural” things, and the set of disciplines that count as natural science, are hard to characterize precisely because both make up an evolving, somewhat vague, set (and even worse, different people use different senses of the term). What is supposed to follow? We shouldn’t call ourselves ontological naturalists? I don’t see how this follows. The analogy with ‘life’ is particularly apt. I am happy to say my daughter is alive, even though my concept of life is plastic and somewhat vague, and other people may understand the term differently. That’s the key analogy that I’d like to see addressed.

  11. John Schwenkler

    I think the problem is that if by “the entities postulated by natural science” you mean the entities postulated by some future natural science, then you essentially mean nothing at all, as there’s no way to predict how future science will go (e.g. perhaps it will happen upon God). The disanalogy with the case of life is that we have a rough-and-ready concept of what it is; by contrast the concept of “future science” amounts to nothing more than “science at some point in the very distant future”, and we have no idea (1) what that science will look like or (2) what it will take to exist.

  12. Eric Thomson

    But we have a “rough and ready” depiction of many entities in science, at least as rough and ready as life: atoms, neurons, lungs, etc.. Because there is some uncertainty at some level, why should this percolate up to all the other levels (and why shouldn’t it percolate up to ‘life’?).

    Note I never put faith in “future” science. I think much of what science provides now is pretty damned good. E.g., how do neurons work?

  13. John Schwenkler

    we have a “rough and ready” depiction of many entities in science

    That’s not the issue. The problem is that future science may show that such things don’t exist, or – more likely – exist only in a very different form than we take them to. Future science may also show that there are things very different from the things postulated by present-day science – as different from them as e.g. the postulates of quantum theory are from those of Newtonian mechanics. So if naturalism is the claim that all there is is the kind of stuff postulated by present-day science, it’s bound to be false; and if it’s the claim that all there is is the kind of stuff that will be postulated by some future science, then it’s simply empty, and doesn’t rule out the existence of supernatural things at all.

  14. Eric Thomson

    If that isn’t the issue, then your response to the life analogy breaks down, because you said the disanalogy with life was that we have a “rough and ready” depiction of what life is.

    These discussions almost always end up with people talking about quantum mechanics, as if concerns at that level will infect all of science and crack naturalistic thinking tout court. I think revolutions at that level will have very little impact on our understanding of how action potentials are generated, for instance, or how birds fly, or whether my daughter is alive or not (again getting to the ‘life’ analogy).

    Note I agree that ‘naturalism’, used in an unqualified sense, isn’t a precise term.However, it is far from empty, far from uninformative, far from useless. Just like the term ‘science.’

  15. John Schwenkler

    Sorry if I was unclear. But the key point is that what we need to define naturalism in terms of the existence only of entities postulated by science is not a rough and ready description of the postulates of present science (which of course we have), but rather of the postulates of some future science (whose nature we have no idea of).

    And the problem with QM isn’t that it’s postulates are non-naturalistic. Rather the problem is that they’d have looked that way from the perspective of some prior scientific theory, say Newtonian mechanics once again. By the same token, to rule out the existence of some sort of thing on the basis of our present scientific understanding would be just foolish. This is what leads the definition of naturalism to the reliance on the concept of future science, and thus the above problem arises.

  16. Eric Thomson

    I understand the point about QM: my concern is that these discussions end up focusing on theories of subatomic particles, as if the conclusion reached there should decide whether we call ourselves naturalists or not.

    I’m pretty sure that present science is actually quite good. I call myself a naturalist based on that.

    We have a good understanding of action potentials. Things called neurons exists. Natural selection happens. DNA explains inheritance. These things won’t likely change, even if things are tweaked in quantum mechanics. I think the basic infrastructure is in place, and won’t radically change, for these types of things. Even if some theory of everything ends up showing that schrodinger’s equation is only an approximation to some other more interesting underlying reality. That will actually have little effect on the types of things I’m talking about, just as special relativity had little effect on how we think about our solar system, and quantum mechanics hasn’t changed how we think about natural selection (except perhaps suggesting hypotheses about how mutations are generated–I think this is actually a good model of the type of minor tweaking and parameter-explaining, at higher levels, we might get with some future physics).

  17. Eric Thomson

    And note I don’t mean to say that even if the only science was quantum mechanics, that naturalism would be a vacuous term. I would still bet that I would call myself a naturalist if that was the only subject under discussion. But that is too far outside my expertise to say anything strongly. For instance, I don’t think QM looks ‘supernatural’ compared to classical mechanics. We still have rule-governed patterns described in quantitative terms, of mass/energy/fields or whatever. Yes, the particular mechanics changed, but it is still mechanics.

  18. John Schwenkler

    I’m pretty sure that present science is actually quite good. I call myself a naturalist based on that.

    That doesn’t capture the sense in which the term ‘naturalist’ is ordinarily used. For one thing, belief in the goodness of present science is compatible with belief in e.g. God. For another, in the ordinary sense it’s entirely possible to be a naturalist while believing on inductive grounds that present science is likely to be wildly inadequate. Lastly, this description of what naturalism is makes the naturalisms of all previous ages (e.g. the naturalism of the ancient atomists, or of Hume or La Mettrie) either false (b/c their science turned out not to be very good) or not naturalism in the same sense as you’re committed to it.

    I don’t think QM looks ‘supernatural’ compared to classical mechanics.
    We still have rule-governed patterns described in quantitative terms, of
    mass/energy/fields or whatever. Yes, the particular mechanics changed,
    but it is still mechanics.

    IMO the only reason you think this is that QM is a respectable scientific theory. From the perspective of classical mechanics QM is not mechanics, and the idea of indeterministic processes not describable in terms of particles bumping against one another is entirely incompatible with materialism. Put differently: if it had turned out that the right description of the quantum level didn’t follow rule-governed patterns at all, I think you’d be quite willing to adopt this view (as you should, if the evidence required it), and that you’d view it as compatible with naturalism.

    None of this is to say that calling oneself a naturalist doesn’t mean anything. My point, and I think Williamson’s, is that what it means can’t be coherently captured in terms of statements about what there is or isn’t. This is why van Fraassen describes it instead as a “stance”, though without his book in front of me I can’t recall exactly what he means by that.

  19. Popped out response because of thin thread above…

    I just don’t buy into the worries here.

    I’m fine talking about things like action potentials, and if someone pushes down to some basic physical level (e.g., of what does the electron in the sodium ion consist in this action potential?) I am happy to say “whatever physics says, and it may not be worked out yet.” I’m fine with such mushiness, and we know that whatever theory replaces QM (assuming one does) it won’t make wildly different predictions than QM in the domains in which it has been tested. So, even if QM ends up being an approximation in the ‘quantum limits’ of some super-theory, I’m OK with that. It won’t change my understanding of action potentials in a serious way.

    This doesn’t make talk of ‘naturalism’ empty, any more than talk of ‘science’ itself is empty. We are dealing with fuzzy open-ended things here. But to say I think the world consists of these types of things discussed in the natural sciences, and nothing more, is what makes me a naturalist (so you are right, simply my overall optimism is not sufficient).

    I think you are not addressing the ‘life’ analogy enough, incidentally. It captures quite well my problem with your view toward ‘naturalism.’ We have a good overall idea of the shape of things, there are many details to be worked out, but calling something alive is informative, even if we sometimes have to clarify it to someone (and some people think the Earth is alive, so we have to make caveats and such).

    Stepping back, ‘naturalists’ tend to be people that think the world described by science is all that exists. Sure, it is open ended, just as the term ‘science’ is open ended. Whether that fuzziness concerns you enough to stop using the term ‘naturalist’ is largely a matter of convention. That, I think, is the crux.

    Because philosophy rewards contrarians, and ‘naturalism’ has been something of a fad, I won’t be surprised if there is some kind of backlash against using the term. But it is such a trivial issue, in the big picture, why do we care, why would someone use precious NYT space to talk about such a semantic quibble? As I said above, when this issue comes up, ” I am usually able to avoid getting bogged down by excessive talking” by simply focusing on a specific topic and trying to discern what is true about it. Leave it up to the dictionaries to decide if it counts as naturalistic.

    That said, I broke my usual rule, and do describe myself as a naturalist without apology for all the reasons I’ve stated above and in this thread.

  20. John Schwenkler

    … to say I think the world consists of these types of things discussed in
    the natural sciences, and nothing more, is what makes me a naturalist
    (so you are right, simply my overall optimism is not sufficient).

    Again, which natural sciences? Present ones? Or future ones? If the latter, then you don’t mean anything in particular. If the former, then what gives you reason to believe that future science won’t show that our present conception of the world is woefully inadequate, as it has shown of every conception before ours? And if it does show this, then will it follow that naturalism is false?

  21. Eric Thomson

    As I said, I think the broad outlines of most of the natural sciences (bio, chem, physics) as they stand now are roughly right. We have lungs, for instance, and DNA, &etc.. Sure, in some future subset of physics we will have to modify or supplement our view, and some of this might percolate up to influence our view of DNA. I’ll deal with that when we get to it.

    You seem to be insisting on naturalists fixing the term more finely than is justified, rather than just accept the acceptable vagueness of the term. You are right, it is not super precise. But it is useful as a rough flag-planter in (for me) present science. I don’t think my conception of the world as containing lungs, DNA, neurons, ion channels, etc will be shown “woefully inadequate”. It seems to be a really good conception, basically right.

    If that is what it comes down to, I am happy to leave it as an unargued premise that DNA, lungs, ion channels and such are here to stay (even if Schrodinger’s equation is not).

  22. John Schwenkler

    The question isn’t whether we have things like lungs and DNA. Rather the question is whether there’s anything to the claim that the only things there are are things like lungs, DNA, black holes, quantum particles, etc. – for that’s what the thesis of naturalism is supposed to come to. And the concern is that the concept of “likeness” that’s operative in the phrase “things like x, y, z” can’t be clarified in a non-trivial way without either prejudicing the development of future science (as a Newtonian would have, by ruling out QM or curved space-time as contrary to the nature of matter) and so making naturalism turn on something it shouldn’t, or making the empty claim that all there is is the stuff that the natural sciences ever will posit, which likely includes stuff wildly unlike lungs and DNA (though of course lungs and DNA will be posits of those theories, too).

    Ahh, I don’t think either of us is going to make any progress with the other.

  23. Eric Thomson

    But I do think that’s all there is, with some fuzziness in darker corners of physics that I’m happy to accept, and that’s why I call myself a naturalist. 

    I don’t share the worry that it is damning that this fuzziness includes things that might happen in the future. Hell it includes things I don’t understand now very well, like quantum field theory or cutting-edge stem cell research. I see the dumping off to the future as the same in kind as dumping off to present expert knowledge I don’t understand. If there is some radical break, some crazy-ass revolution, then I’ll concern myself with whether to continue calling myself a naturalist.

    Incidentally, how many decades before the pessimistic induction over the history of science becomes the optimistic induction over the recent successes of science? 🙂

  24. John Schwenkler

    Sorry for the slow reply – I’ve been traveling. I’ll just say that while the position you’ve been staking out does indeed constitute a distinctive philosophical attitude (this is what van Fraassen means, I think, in calling naturalism a ‘stance’), it’s still not well enough defined to mark off a metaphysical thesis concerning what there is and is not. I take it that the latter thing is what Williamson – who is of course a stickler for precision – thinks we should be after, and it’s that concern, together with a frustration with the idea that the scientific spirit is limited to experimentally-driven disciplines (indeed I find Williamson’s work a perfect illustration of the idea that philosophy can make meaningful discoveries of its own), that keeps him from calling himself a naturalist.

    Re: optimism and pessimism, every other scientific worldview had many highly successful decades (centuries, even) of its own. It’s clear enough from the history that they had the very same perspective on their science as many of us (myself included) have on ours. I just find it too unlikely to believe that I’d have been born at the end of scientific history rather than somewhere in the mushy middle.

  25. Eric Thomson

    Re: your last paragraph, it depends on the science, I think, as you seem to agree as you don’t wish to take issue with my use of terms like ‘life’ or argue that lungs, DNA, evolution, etc will be jettisoned.  Also, science by its nature is open-ended, and “success” is typically gradual and rarely if ever final. I don’t really mind if you call what I’m arguing for a worldview, orientation, attitude,  thesis, or hypothesis. I find it precise enough to be informative, but not precise beyond its means. I tend to think of it as a well-confirmed hypothesis, but don’t really care about the words.

    Overall, it really is easiest to avoid all this discussion and just be specific about a particular topic (e.g., consciousness, Godel’s theorem, whatever). Since naturalistic folk as a rule tend to avoid these extremely general discussions because of impatience and concerns about fruitfulness, there might be a tendency for folks like Williamson to think they are evoking acquiescence rather than just polite avoidance. Sort of like a pushy religious person that we eventually just avoid arguing with.

    [Edit:  final sentences are a bit too strong. I should be clear I am not saying these worries about naturalism are as silly as beliefs in angels and such; I just think the worries seem a bit anemic with questionable payoff for the effort. It is more fruitful, if someone has cool results suggesting that more than the entities in natural sciences exist, then they should argue for those specific cases rather than these generalities.]

  26. Richard Marshall

    Holes, silence, shadows, the elephant not in my room all exist but physics, chemistry, biology can’t access them. Ever. Roy Sorenson argued this to point out a miguided bias against non-material existence in various philosophical positions. Seems relevant.

  27. Eric Thomson

    Natural science can’t access shadows? Really? I think you don’t want to say that.

    Elephant not in your room? How about that there isn’t a planet between Mars and Earth? Science talks about nonexistent things all the time. It doesn’t mean they actually exist…Meinong.

    Holes? Study semiconductors

    Propensity to eschew nonmaterial objects doesn’t seem a bias as much as a hard-won conclusion, or at least a plausible hypothesis. Perhaps the set of even numbers would be the best type of example for the antimaterialist.

    I’m to some degree being glib here, just my first reaction to a set of things the materialist probably shouldn’t worry about.

  28. Richard Marshall


    Seems like science is just redefining things – as when they start to make empty space do the work of ether or vacuums are no longer empty but have energy that can be converted into mass. But I take your general point – scientists can access these things eg biologists point to use of perceiving absences in animals eg mosquito larvae.

  29. Josh Weisberg

    All right–very late, sorry. But thanks to John and Eric for an excellent debate (if it did get remarkably skinny there for a while!).

    It seems to me that accusing naturalism of “not being a (precise enough) metaphysical thesis” misses the point of naturalism. The idea (attitude, stance, etc.) is that there are no useful ways to establish such metaphysical theses beyond what science licenses us to do. When one goes beyond such (admittedly rough-and-ready) boundaries, there seems to be no good way to settle disputes. And so one spins one’s wheels without getting anywhere.

    So, does that mean that all there is is what science tells us? Perhaps no, but if you go beyond the science, how do we tell who is right and who is wrong, even in principle? The alternatives out there are formal proof methods and… what? Formal methods aren’t going to do it, IMHO–that is the message of the last 60 years of analytic philosophy, perhaps. So what’s the alternative?

    And if Williamson accepts all this (does he?), then I think he’s attacking a straw man.

    Anyway, no doubt I am posting to a dead thread, but what the hey. Thanks again for the stimulating debate!

  30. Eric Thomson

    The skinniness of the debate was fairly comical.

    It seems you are advocating a kind of methodological naturalism (focusing on the paucity of alternatives to determine who is right is a question about methods, no?). I am very sympathetic to this concern, but think philosophers tend to be pretty good at stretching our imaginations in new and interesting directions, opening up conceptual possibilities that people hadn’t considered before (e.g., twin earth, eliminative materialism). This is less a matter of settling who is right about something than opening up new questions that hadn’t really even been considered before.

    Note scientists do this as well, I’m just saying philosophers can be pretty good at doing it in ways that an entrenched scientist might not.

  31. Josh Weisberg

    No doubt that’s correct, Eric. In its “proposing” mode philosophy (and science, as you note) needn’t be bound by worries of how to settle who’s right. But once we turn to that question, methodological naturalism seems like a good guiding principle. My “naturalist hackles” rise when people take such thought-expanding scenarios like twin Earth as more than that.

    Thanks again for the illuminating debate!

Comments are closed.

Back to Top