This is a faulty conception of physical law. Physical laws do not constrain — not in that way. Sure, a physical law constrains what’s possible — *if the law is valid*. But a physical law is merely an assertion that something is universally true. If some phenomenon P violates “physical law” L, that doesn’t make P “supernatural”, it merely makes the assertion that L is a physical law erroneous. And being unconstrained by any physical law does not make something “supernatural” — whether Schrödinger’s cat is found to be alive or dead when we open the box is not governed by any physical law; both outcomes are equally likely (ex hypothesi; see http://en.wikipedia.org/wiki/Schr%C3%B6dinger%27s_cat#The_thought_experiment).

“The falsification of naturalism on this view is, as you say, harder, since it’s always an open question as to whether something will eventually fall to a law-like explanation.”

That doesn’t just make it harder, it makes it impossible. But that’s no problem — the applicability of Popper’s falsification criterion distinguishes between a scientific (empirical) theory and pseudo-science, but naturalism is neither of those, it’s an ontological claim, which is allowed to be analytically true (I believe it is).

“I’m not sure whether this necessarily makes my proposed definition of the supernatural untenable.”

Not at all. Your definition makes the supernatural something that cannot, analytically, exist, which is no more untenable than “a mechanical procedure to prove all true statements of arithmetic” — people once thought that such a thing was possible, but they were wrong, and we now understand the insights they lacked; this is by no means “trivial” or “without much utility”.

Of course I very much doubt “psychons” or anything panpsychist or proto-phenomenal will be discovered, but I don’t think we can rule out the possibility that nature might contain more than what we might want to classify as matter and energy. After all, our current conceptual grasp of reality as delivered by science may expand in ways we can’t predict. So to say that something we find that doesn’t fit into a matter/energy paradigm must count as supernatural seems to me unwarranted. After all, it will be incorporated into our best, most reliable understanding of reality, which will make it inherently unmysterious. To me that strongly counts in favor of naturalization, but maybe you’re right that the supernatural needn’t be mysterious to count as such.

I agree that the idea that supernatural beings can interfere with and control events in the natural world, but themselves are not constrained by natural laws, is probably central to what most folks have in mind by the supernatural. But the idea that supernatural beings are themselves constrained by laws that science could discover doesn’t seem to me what most folks have in mind by the supernatural. But maybe I’m wrong. It would be interesting to find out what people think about this.

“…can you describe a situation in which a being that is not constrained by laws can be determined by science to exist?” If we all reliably observed something that didn’t seem constrained by natural laws, that would certify its existence according to science. It would then be a further open question as to whether it was constrained by any laws or not. Until such laws were discovered, it would be a live hypothesis that the phenomenon was unconstrained by any laws. But as I suggested previously, it’s an inherently *defeasible* hypothesis given progress in science, since we might discover laws which constrain the phenomenon. The falsification of naturalism on this view is, as you say, harder, since it’s always an open question as to whether something will eventually fall to a law-like explanation. I’m not sure whether this necessarily makes my proposed definition of the supernatural untenable. There might be tension between the conception of the supernatural (assuming there is a single unambiguous correct conception) and what makes naturalism easily falsifiable, or maybe not. Anyway, I’m keeping an open mind about all this and very much appreciate our exchange.

Keep in mind that consciousness is not a singular construct; drugs, level of sleep, perceived threat, and other conditions will change how someone experiences any given phenomena; in this sense, there is no “pure qualia”. There is no one irreducible thing called “redness” for example; the experience is a product of visual sensation, perception, access to memory, and then perhaps higher level functions, such as awareness (“I’m seeing red”), judgment (“I like red”), and intention (“I want to paint my house red”). This is what Minsky was getting at and I think he was right. That we are aware of our perceptions and thoughts is a product of a prefrontal cortex, and the reason it is limited to my own head is that brains aren’t connected. That doesn’t *explain* subjective experience, of course, but it does suggest that non-brain explanations are not necessary.

But even if we did discover “psychons”, I think that they would very likely fit somehow within what we know to be the universe of matter/energy within time/space. If they did not, then they would be supernatural, and naturalism would be rejected.

One (let’s call it the Bowie Model) says that both are (in principle) knowable, understandable, and yet fundamentally different. The difference is embodied in physical laws, what we now call the Standard Model (or any other better theory that comes along in the future). The natural is that which mindlessly conforms to or arises from the laws that underlie the universe of matter/energy existing in space/time (which I guess falls under the header of physicalism). The supernatural is that which can intentionally interfere with the natural world but is not constrained by its laws. I get the mindless/intentional part from Carrier—I would call any mindless process that seems to violate natural laws (e.g. astrology) to be paranormal/magical rather than supernatural. Under the Bowie Model, naturalism would be falsified if the supernatural is definitively proven to exist, regardless of how little or how much we can scientifically understand about it.

In the other model (the Clark Model), the natural is any thing that can be brought under a law-like explanation, regardless of its relation with other known laws—if it obeys laws that are explainable (in principle) by science, then it is natural. The supernatural is that which exists and yet inherently evades reliable scientific comprehension, due perhaps, to a form of being that does not exist in a schema of law-like relations. Under the Clark Model, naturalism would be falsified if any thing can be proven to exist that cannot be understood via a law-like description and explanation.

Both are viable models, I think. The Clark Model would make naturalism harder to falsify, I think, because I’m not sure how science could reliably determine that something exists that is not itself constrained by any laws. This is a bit beyond my pay grade, but I’m not sure how it is possible to say that something exists if it is unconstrained by any fundamental laws of existence, no matter how bizarre. For example, if a ghost is unconstrained by any laws, then it is free to be and do anything within our physical world…and yet, ghosts are often “trapped” at locations and have common behavioral traits; this would not be the case if they were free from laws.

And so, outside of unusual hypotheticals, can you describe a situation in which a being that is not constrained by laws can be determined by science to exist? If not, I don’t see how naturalism can be falsified.

Re consciousness, I agree that all the evidence indicates that physical correlates have to be present for consciousness (phenomenal experience, e.g., pain) to exist. But it’s difficult to draw an identity between those correlates and a phenomenal experience like pain for the simple reason that pain is categorically private. My pain is available only to me, the person who consists of the correlates, whereas the correlates themselves (my brain states) are in principle publicly available for observation. How does something categorically private and qualitative like pain come to exist by virtue of certain physical, objective states of affairs? As far as I know, there’s no accepted solution to this problem (the “hard problem” of consciousness) on offer. So I see Minsky as patently evading the difficulty. My latest attempts to tame the hard problem using representationalism are at http://www.naturalism.org/appearance.htm

Re defining the supernatural: I agree we want to make naturalism falsifiable, or more or less probable in response to evidence, since otherwise it’s a dogmatic assertion. This means we have to say in advance what the supernatural is, so that if something fitting that definition comes to light, or is probably the case, then naturalism will have been falsified. According to Carrier’s definition, should we discover mental phenomena that don’t depend on non-mental phenomena, that would do the trick. This equates naturalism with physicalism, which I think doesn’t work since many (perhaps most) philosophers would say the discovery of independently mental phenomena (e.g., “psychons”) would simply add to our knowledge of what exists *within* nature.

Perhaps another way of defining the supernatural that fits the facts, avoids equating naturalism with physicalism, and still makes it falsifiable is to say that what’s supernatural is just what can’t be brought under a law-like explanation. I think this definition captures common intuitions about the supernatural.

As I’ve suggested before, I think naturalists are those that stick with our most reliable modes of justifying beliefs about reality, science and other empirical disciplines, and they stick with empiricism when deciding *all* questions of fact. If something independently mental shows up, so be it. So the way I see it naturalists are simply thorough-going empiricists who are ontologically non-dogmatic about what’s real, and they call what’s been demonstrated thus far to be real, “nature.” Thus far, what’s been shown to empirically real seems to exist *lawfully* in relation to everything else that exists. These natural laws include psychological laws that apply to agents with mental states, states which seem to supervene (lawfully) on physical states (how and why is the unsolved hard problem).

Naturalist philosophers generally suppose that *if* any categorically independent mental phenomena show up (a big if, I agree) they will likely manifest some sort of regular relations to physical phenomena, relations that David Chalmers calls psycho-physical laws. After all, as far as we know mental states don’t seem to be haphazardly connected to physical states. So perhaps the naturalist’s falsifiable claim is most generally that what’s certifiably real is an organized state of affairs potentially describable by law-like generalizations. Supernatural entities would be just those which transcend natural laws in their characteristics and capacities; they can’t be described or explained in terms of science or other empirical disciplines. This seems to me to capture common intuitions about the supernatural. The supernaturalist agrees with the naturalist that science-certified, law-governed nature exists, but claims that there’s more, the supernatural, which isn’t constrained by observable laws. Typically the supernatural phenomenon is claimed to be an agent of some sort, possessing intentional states like beliefs and desires.

Let’s say that intersubjective evidence comes along that seems to confirm the existence of a phenomenon supernaturalists believe in, such as a god, demon, soul, ghost or spirit. We all observe it and agree that it has these amazing powers, so it’s unequivocally empirically real. For this being to be supernatural, it has to be the case that its characteristics and capacities *cannot* be shown to depend on what the supernaturalist has agreed are natural phenomena, namely phenomena which science has certified to exist and which conform to empirically determined regularities. If they did it would be a natural being by virtue of that conformity.

You say “the supernatural object would still likely be ruled by some set of laws, but they would be different from natural laws.” What I’m suggesting is that *if* such laws exist and were empirically observed and incorporated into science, then we should say that the entity in question isn’t supernatural. Despite it’s *apparently* miraculous powers, the being would have been revealed as just another part of law-governed reality, however bizarre (and physics already presents us with some pretty bizarre realities!) Even if, let’s say, its laws were completely proprietary and sui generis, so that they bore no causal, emergent or supervenient relation to anything else known to science, it would still have been incorporated into our understanding of reality as law-governed. This is very much like the idea contemplated by some physicists that other parts of the universe or multiverse might conform to completely different sets of laws: for the scientist they still count as parts of nature.

On the other hand, for this being to remain supernatural it seems to me that some or all of its properties would have to remain utter mysteries to science, uncapturable by any schema of law-like relations. I think this is what your description of the supernatural ghost really suggests: it’s behaving in a way that transcends law-like explanation. To be supernatural, neither its psychological states (should they be recognizable as such) nor its powers would have observable correlates in any law-governed phenomena we know of. There’s no reason in principle such a phenomenon couldn’t exist, so it seems to me naturalism would be falsified by a failure to bring something real, in particular an agent with mental states and apparently miraculous powers, under a law-like description and explanation.

But note that on this definition of the supernatural naturalism’s falsifiability is conditional, not absolute, since future developments might show that the being *is* subject to laws. This is just to say that our understanding of reality as law-governed (nature), can grow by means of new cognitive and conceptual developments such that it encompasses heretofore apparently supernatural phenomena, as has happened many times (e.g., lightning). Thus far no phenomena have appeared that haven’t fallen to naturalistic explanation, and since consciousness seems firmly connected to and entailed (somehow) by naturally evolved brains, there’s no good reason to suppose it’s supernatural either. It’s just not yet understood.

Consciousness does seem to be the crux of the matter. The problem I keep running into is twofold: (a) we have a very incomplete understanding of consciousness and (b) what we do know does seem to suggest that mental activity does have physical correlates. While it’s good to ponder, I am not too hung up on hypothetical conjectures about independently-existing mental states…just because something is possible doesn’t make it likely or even feasible (I have a bit of bias here being graduate trained in psychology). So, when it comes to “qualia”, I agree with Marvin Minsky:

———
“Now, a philosophical dualist might then complain: “You’ve described how hurting affects your mind — but you still can’t express how hurting feels.” This, I maintain, is a huge mistake — that attempt to reify ‘feeling’ as an independent entity, with an essence that’s indescribable. As I see it, feelings are not strange alien things. It is precisely those cognitive changes themselves that constitute what ‘hurting’ is — and this also includes all those clumsy attempts to represent and summarize those changes. The big mistake comes from looking for some single, simple, ‘essence’ of hurting, rather than recognizing that this is the word we use for complex rearrangement of our disposition of resources.”
———

I guess what would differentiate them as natural, not supernatural, on my view (which is trying to capture the ordinary meaning of supernatural) is that they aren’t fundamentally mysterious, but rather part of the *known* world.

I still don’t see how this is different from saying: “that which is knowable is natural” or “the supernatural is that which is unknowable.” This is, I think, a trivial distinction, if I’m reading it right.

Although I like Carrier’s approach, I’m open to other ways of making a clear and operational distinction between natural and supernatural (or nonnatural). For example, I might propose that the supernatural is that which is able to affect changes in the macroscopic world of matter, even in violation of natural laws, but which itself is not subject to natural laws. So, for example, a ghost might be able to move physical objects or be seen by a human eye, but cannot itself be, say, constrained in a box if we try to trap it. So, a supernatural object can choose to affect us but we cannot affect it; manipulation is one-way only (unless the object chooses to be manipulated). Further, the supernatural object would still likely be ruled by some set of laws, but they would be different from natural laws.

As always, it’s possible that there exists a deeper level of fundamental laws that govern the totality of reality, natural and supernatural alike. But even if this were true and my hypothesis were also found to be true, then the supernatural would be both real and distinct from what we now think of the natural world.

Naturalism, of course, is the worldview that the supernatural does not exist. The problem with my model is that supernatural objects cannot be known to us unless they choose to be made known. We only have natural objects at our disposal, and my supernatural entities are not subject to them. That doesn’t mean they can’t be studied, at least indirectly, but in most cases they would have to volunteer. I could go on ad nauseum…

The big exception here is of course consciousness. There’s no consensus among philosophers or scientists that any clean reduction of it is in the offing. Some hard core physicalists like Jaegwon Kim have given up trying to reduce qualia, for instance in his book Physicalism, or Something Near Enough, a chapter of which is at http://press.princeton.edu/chapters/s7971.html (see the section “Can we reduce qualia?”).

“If there do exist natural mental events that are (at least in part) independent from the nonmental, what differentiates those events from supernatural events?”

I guess what would differentiate them as natural, not supernatural, on my view (which is trying to capture the ordinary meaning of supernatural) is that they aren’t fundamentally mysterious, but rather part of the *known* world. The supernatural, seems to me, has as an essential characteristic that it evades human understanding of the sort that makes something amenable to prediction or control. I agree that mental events that had no connection to physical events would seem supernatural if we didn’t understand them, but if we *did* come to understand them, then it would just turn out that that’s how the world works. So such mental events would be incorporated into our picture of the world, what we ordinarily call nature.

If we encounter a being whose powers and behavior are currently inexplicable according to science, then one hypothesis would be that it has supernatural powers, powers which transcend natural laws. Now let’s suppose we end up understanding that being, it’s provenance, capacities and behavior, so it’s no longer a mystery. Let’s also suppose that understanding has nothing to do with natural laws as science describes them. We might continue to say that the being is supernatural, by definition, but of course it wouldn’t *seem* very supernatural any longer, precisely because it’s understandable, or at least that’s my intuition.