Sean M. Carroll shows that panpsychism is unlikely and unnecessary

July 26, 2021 • 11:00 am

I’m heartened to see that other scientists and philosophers of mind I respect, like Sean Carroll and Patricia Churchland, have analyzed the idea of “panpsychism” and found it wanting. As I noted yesterday, adding some of my own criticisms, panpsychism is somewhat of a philosophical fad (or even a religion). It claims that we’ll never understand consciousness through a combination of neuroscience and philosophy, but instead must posit that every bit of matter in the universe has its own form of consciousness. And if you put enough of those conscious atoms and molecules together, you get “higher” consciousness: the feelings of subjectivity, pain, pleasure, and the perception of colors known as “qualia”.

The problems with panpsychism are at least fourfold: the theory is untestable, there’s no evidence for consciousness of inanimate matter, there’s no explanation how the “rudimentary” consciousness of molecules and atoms can combine to produce to the complex consciousness of humans and (surely) other mammals, and we have made no progress in understanding consciousness by considering or adhering to panpsychism. It seems to be a view that, ultimately, will not help us understand consciousness.

The physicist Sean Carrol takes another angle in a new (and yet unpublished) paper that was cited by reader Vampyricon and is online. Click on the screenshot to read it. I’ve included the abstract and the place where it will be published (the book referenced, Galileo’s Error, is advocate Philip Goff’s big defense of panpsychism, but I wasn’t impressed).

As he has in previous books and papers, Carroll demonstrates that our present theory of physics is perfectly adequate to explain the physics of everyday life—unless we go sticking our heads in a black hole or something. Further, adding “mental” properties to our known core theory of physics not only changes that core theory, but is unlikely to explain consciousness, which, though we don’t yet understand it, is in principle perfectly consistent with the laws of physics, with consciousness being an epiphenomenon of physical processes. Yes, we don’t understand it, but that doesn’t mean that we must go tinkering with the laws of physics to explain consciousness or positing untestable mental properties of inanimate matter.

Carroll’s is a long paper, and has some equations that I don’t understand, but his conclusions are clear, and demands that panpsychism clarify its propositions in explicit physical terms beyond merely saying “all matter has consciousness”.  Here’s his conclusions:

Any discussion of mental aspects of ontology must specify one of two alternatives: changing the known laws of physics, or positing that these aspects exert no causal influence over physical behavior. We cannot rule out the first option either through pure thought or by appeal to existing experimental data, but we can ask that any modification of the Core Theory be held to the same standards of rigor and specificity that physics itself is held to. The point of expressions like (1) and (3) is not that mentally-induced modifications of physical parameters are impossible, but that a promising theory of consciousness should be specific about how they are to be implemented.

The passive mentalism option, where mental aspects have no impact on physical behavior, seems even less promising. “Behavior” should not be underrated; the behavior of physical matter is literally “what happens in the universe.” Crying at a funeral is behavior, as is asking someone to marry you, as is arguing about consciousness. No compelling account of consciousness can attribute a central explanatory role to metaphysical ingredients that have no influence on these kinds of behaviors.

We don’t know everything there is to know about the laws of physics, and there is always the possibility of a surprise. But the solidity of our confidence in the Core Theory within its domain of applicability stands in stark contrast with our fuzzy grasp of the nature of consciousness. The most promising route to understanding consciousness is likely to involve further neuroscientific insights and a more refined philosophical understanding of weak emergence, rather than rethinking the fundamental nature of reality.

I have a feeling that in one or two decades panpsychism (which has been around in one form or another for centuries) will no longer be regarded as a fruitful way to understand consciousness.

71 thoughts on “Sean M. Carroll shows that panpsychism is unlikely and unnecessary

  1. The “money shot”, in my opinion, would be the following:

    The idea of new independent mental degrees of freedom runs into serious trouble with the framework of quantum field theory. In conventional field theory, the existence of new degrees of freedom quantitatively affects processes that rely on quantum fluctuations, in which each property value represents a separate contribution that should be added together (e.g. we “sum over spins” in a scattering calculation). But we know empirically how many degrees of freedom actual electrons have – two spin states for the electron, and another two for its antiparticle, the positron. If electrons could also be found in both “happy” and “sad” states, it would have an unmistakable impact on their scattering rates, in flagrant contradiction with experiment.

    If panpsychism is true, each type particle can only hold one mental state, which means we are back to physicalism with causally inert “mental properties” tacked on, and these mental properties cannot communicate with each other to form a more complex consciousness except via regular physical interactions. At this point, panpsychism would be strictly more complex than physicalism, and Occam’s razor takes care of the rest.

    1. Panpsychists would immediately reply that at the level of elementary particles, consciousness would also be at an elementary level, flitting randomly from one blip of awareness to another.

      1. That’s exactly what the quote addresses. Its consciousness can’t flit from one blip to another because that implies extra degrees of freedom, and extra degrees of freedom affect scattering experiments.

      2. Being able to flit from one blip of awareness to another implies more internal degrees of freedom than there exist, which affects scattering amplitudes.

    2. Yeah, I’m surprised Carroll claims alternatives 1 and 2 are separate. As your money quote points out, in QM you can’t really have a “passive” thing the way this is described. Even the mere act of physical states being detected by mental ones would have an impact on the physical. So really, alternative 2 (purely passive) is just a specific example of alternative 1 (laws of physics must be changed).

      I’m guessing maybe he treated them separately because alternative 2 is a somewhat common defense? So worthy of a separate discussion even if technically it falls under #1.

  2. I’ve already seen a rush of woo ads on the internet purposefully misrepresenting this paper. “not that mentally-induced modifications of physical parameters are impossible” has become a rallying cry not unlike the creationists use.

    1. The evidence for consciousness seems to boil down to ‘I experience this sensation and you have it too, right?’ This is anecdotal. Studies of neural systems will probably be best served by ignoring non-physical phenomena.

  3. I’m a fan of the idea that our consciousness is a kind of illusion. I don’t think qualia in the traditional sense really exist. The whole idea of some intrinsically first-person private qualia is just physically and probably even logically impossible. How can something only exist for one person? Existence is third-person. If something exists then it exists for everyone.

    Our sense of experiencing some private qualia is probably a kind of useful fiction our brains use. Philosophers used to worry that mind is fundamental and reality may be a figment of our imaginations, but it looks like it’s the other way round. There are parts of objective physical reality that believe they have some intrinsically first-person qualia. Reality is not a figment of the mind, the mind is a figment of reality.

    1. “I’m a fan of the idea that our consciousness is a kind of illusion.”

      Pity that poor rock, sitting there thinking it is conscious, but in fact it is just an illusion.

      1. I think Bernd’s point is that the rock, very likely without anything like consciousness, is the one that’s got it right. It’s US who are under the illusion of some ontologically real set of states corresponding to identity. I don’t think it’s true, though, that `our sense of experiencing some private qualia is a fiction’: if it’s not real—doesn’t exist—then we shouldn’t be experiencing it, right? What’s fictional would rather have to be the ontological conclusion that there really is an ‘I’ which is having these illusory experiences. This was a theme that Sartre used to bang on about, as I recall. But is there any actual data that might have an empirical bearing on this point?

        1. Hard to guess sometimes what may have been meant, but illusion doesn’t necessarily mean it doesn’t exist. It can also mean that “it” isn’t what it seems to be.

    2. Our sense of experiencing some private qualia is probably a kind of useful fiction our brains use.

      Could well be. But for the time being, I’m gonna go with anesthesia. 🙂

    3. “Illusion” seems a bit inapt.
      My opinion is that is that the ‘hard problem of how neural signals produce qualia’ is only hard because we don’t yet grok that neural signaling is qualia. The latter is just our way of interpreting the former. No extra, inexplicable step needed. Thinking there is some additional step we need to explicate is the illusion….not the sensations themselves. I hazard a guess I’m not the only one with that opinion.

      1. Are you saying that just as our sensation of pain is nothing more than registration of activity in pain receptor neural complexes, the sense of experiencing object shapes is nothing more than the registration corresponding to neural activity in the retinal cell complexes that detect edges in space and in the downstream parts of the visual cortex that compute out spatial relationships based on these? And so on?

        Could well be… but there’s still the question, where does the `I’ that is felt to be the center of all these qualia experiences come from?

  4. I am not well read on any panpsychism work, however, I would say that Ockham might advise us to not multiply entities beyond necessity. We needn’t spend time trying to disproving every idea, only the ones which have reasonable evidence of being operative.

    Is there good reason to even interrogate panpsychism?

  5. Yes, we don’t understand it, but that doesn’t mean that we must go tinkering with [it].

    Or as I learned to say to myself “Ohh look at all the bits you managed to save!”, when considering taking something apart to try to work out how it works. And, in particular, the implicit “put it back together” stage.

  6. I have a feeling that in one or two decades, panpsychists will still be out there. The lack of evidence and progress is not discouraging to pushers of any pseudoscience.

    1. A move from the “god of the (vanishing) cracks” to the “supernatural of the virtual cracks”?

  7. “In our conventional understanding, consciousness exerts an important influence on behavior: I can have a conscious experience and talk about it.
    (Admittedly, highly trained philosophers are reported to be able to imagine removing consciousness from a being without affecting its behavior in any way.)” – Sean Carroll

    I’m not an epiphenomenalist myself, but epiphenomenalists do have two plausible replies to this:
    1. They will reply that experiences themselves don’t have to influence or cause anything in order to be reportable by their subjects, because cognitive access to an experience needn’t be caused by it. An experience’s cognitive accessibility (with reportability being a kind of cognitive accessibility) doesn’t require its non-epiphenomenality.
    2. They will also reply that epiphenomenalism doesn’t imply that “removing consciousness from a being” wouldn’t “affect[.] its behavior in any way.” For if a physical brain event P1 causes both another physical brain event P2 and an epiphenomenal mental brain event M, then if M hadn’t occurred, P1 wouldn’t have occurred and its physical effect P2 wouldn’t have occurred either; so the absence of M would make a difference to the (behavior-causing) physical processes in the brain. Epiphenomenalism is perfectly consistent with a counterfactual dependence of certain physical brain events on epiphenomenal mental brain events.

    1. Your point 2 is right, but point 1 seems to fly in the face of the definition of “epiphenomenal”. Anything that causes speech, or a particular content of that speech, is non-epiphenomenal.

      1. Yes, of course, but the very point made by epiphenomenalists is that verbal reports of experiences needn’t be and aren’t caused by them but by something else. Epiphenomenal experiences are reportable by virtue of certain nonepiphenomenal cognitive processes that enable subjects to access their experiences (cognitively/introspectively/reflectively).

    2. “Passive mentalism opens the door to contemplating the possibility of philosophical zombies: creatures that have exactly the same physical behavior as ordinary human beings, but lack any inner conscious experiences (…).” – Sean Carroll

      First of all, passive mentalism = (psychological) epiphenomenalism.

      As I already explained in my previous post, if there is a counterfactual dependence (* of certain physical brain events on epiphenomenal mental brain events, then a conscious agent cannot be turned into a physically (and thus behaviorally) identical zombie agent, because removing the epiphenomenal mind means changing the body.

      (* To say that event E1 is counterfactually dependent on event E2 is to say that if E2 hadn’t occurred, E1 wouldn’t have occurred either. Note that there are philosophers who regard counterfactual dependence as sufficient for causation; but epiphenomenalists certainly don’t, since according to them mental events are never causes of anything.)

    3. While point 2 is certainly valid, point 1 is not that easy to grasp, at least not intuitively.

      At any rate, it seems to me that a bigger hurdle for epiphenomenalists to overcome is that we’re conscious only of things that are useful for us to be conscious of—useful in that consciousness of such things gives us the possibility to act in a deliberate way in response to them (a way that on average should be better than a more automatic or reflexive response, i.e. better than a very simple algorithm). And things where deliberate action would serve no purpose, e.g. the dilation and constriction of the pupils or blood circulation, we’re unconscious of. This would be an astonishing coincidence if consciousness didn’t affect behavior. I wonder how epiphenomenalists account for it.

      It’s clear to me that consciousness has been carefully and precisely “designed” by natural selection for very specific and distinct purposes, pretty much like any bodily organ. From this understanding it follows that it can’t be a mere epiphenomenon without causal power.

      1. Like Jennifer Aniston in The Break Up, panpsychists should ask their (mental) estheticians for “the full Yul Brynner.”

  8. I think Sean really nails it, especially by constructing the dilemma and including his critique of the “passive mentalism” option. On the other hand, it seems a lot of work to refute a position whose only attraction is to people inclined to commit the fallacies of composition and division. (I swear I picked this top Google hit for “fallacies of composition and division” without knowing that they’d use panpsychism as an example!)

  9. I have a feeling that in one or two decades panpsychism (which has been around in one form or another for centuries) will no longer be regarded as a fruitful way to understand consciousness.

    Does anybody who is taken seriously in the field of researching consciousness regard it as fruitful now?

      1. There is both a neutralistic version and a panpsychistic version of Russellian monism: According to the former, the categorical properties in question are neutral (neither-mental-nor-physical) qualities; and according to the latter, they are mental (experiential/phenomenal) qualities.

      1. No it does not. Our “now” appears to be between a few tens of milliseconds and two three seconds in the past.

        But I would agree it does seem that consciousness has a role. I frequently talk without thinking … I know it shows.

        If we subscribe to the cause and effect paradigm, then either consciousness is brain chemistry (monism) or consciousness is floating about somewhere directing brain chemistry, possibly quantum chemistry (dualism). I subscribe to the former.

        What is clear, reality is illusory, or at least not as we perceive it. For me consciousness is part of reality. I really suggest you google Susan Blackmore’s “Am I conscious now?”

      2. People also talk about ‘the set of all sets which are not members of themselves’ (speaking of Russell!), the planet with the name ‘Vulcan’ between Mercury’s orbit and the Sun, the ‘ether’, etc.

        I don’t think causal efficacy has much to do with peoples’ nouns which happen to be in quite common use, either ‘then’ or now. What causes people to talk about non-existent objects is not those supposed objects.

        When serious, it’s about some theory which ultimately is abandoned via empirical evidence and alternative better theories, occasionally pure logic alone e.g. in the Russell paradox case above.

        Panpsychism would only become serious if its advocates began to behave dangerously like Qanon advocates. As it is their behaviour is much like that dickhead Bernardo Kastrup—one attempts to make a career out of it. Suckers: Please exit through the grift shop. I guess the man never contributes here, so I’m allowed to call him an impolite but appropriate name.

        Bothering about panpsychism seems to me a giant time waster, e.g. in comparison to discussing free will.

  10. I just want to know what species of ‘shrooms Goff has been eating, and can I try some. They must be far out, man, like cosmic.

    1. Those ‘shrooms are derivatives of indole ethylamines, as are compounds like serotonin, and melatonin. Tryptophan an essential amino acid is an indole ethylamine with a carboxylate added. Can’t help wondering if consciousness is a trip of a lifetime.

  11. Billionaires with expensive hobbies – evidence that they are not paying enough taxes.
    Discussion of panpsychism by professors of cognitive science – evidence that they are not teaching enough courses.

    1. I feel that’s quite harsh. Academia is the place for people with deep knowledge on a subject to explore and advance ideas that could potentially advance our understanding on a particular topic. Without a clear solution for all things in philosophy of mind, there’s a lot of room for different research paradigms to explore various solutions to see what comes out.

      Panpsychism may be nonsense (I think it is), but it may be a fertile research paradigm for the field. If it’s sterile, it’ll be abandoned for something potentially more fertile. And while consciousness has many mysterious elements to it, potential paths of inquiry may seem crazy or absurd yet might yield for us genuine insights by the exploration of them.

      1. “Academia is the place for people with..” supposedly “..deep knowledge on a subject to explore and advance..” what they call ” ..ideas that could potentially advance..” for some: their careers however evidently stupid the idea, and for others: “..our understanding on a particular topic.”

        I’d like to think I’ve read thought about enough philosophy, sociology, etc. for the above not to be an ill-considered opinion, esp. with respect to panpsychism.

      2. To add:

        “If it’s sterile, it’ll be abandoned…”

        It’s been around for centuries, with nothing approaching any delineation of what evidence would get them to abandon it, with no seriously considered meaning given to the words they regurgitate over and over.

        I’m afraid it’s more like those zombie rightwing Usian political horse manure ‘ideas’ (as Paul Krugman calls them), one of his examples being “trickle-down economics”, his recent one being “family values”. But the zombies, like panpsychism, keep rising from the graveyard tended by the grifters.

  12. Jerry: “Further, adding ‘mental’ properties to our known core theory of physics not only changes that core theory, but is unlikely to explain consciousness, which, though we don’t yet understand it, is in principle perfectly consistent with the laws of physics, with consciousness being an epiphenomenon of physical processes.”

    Sean: “The passive mentalism option, where mental aspects have no impact on physical behavior, seems even less promising…. No compelling account of consciousness can attribute a central explanatory role to metaphysical ingredients that have no influence on these kinds of behaviors.”

    Sean’s “poetic naturalism” has it that talk of consciousness is simply a useful “way of talking” about the underlying fundamental physical situation in the brain and body as described by the Core Theory. If so, then talk about consciousness is talk about something physical. It’s likely that whatever physical goings-on consciousness consists in (if indeed it does consist in them) are doing causal work, thus are not epiphenomenal. Of course, we ordinarily think of conscious experiences as picked out by their qualities, and it’s the *quality* of, say, pain that we commonsensically think of as playing a causal role in pain-related behavior. But if pain is physical, then the causal,, behavior-guiding work is already being done by its neural instantiation, so it isn’t clear what the quality per se adds to that work. As Dennett once pointed out, you won’t see pain in a flow chart of pain-related neuro-muscular activity.

    It isn’t obvious to me that poetic naturalism – the claim that consciousness is physical – is true, but even if it is, it doesn’t seem to allow for a distinct causal role for experiences conceived of as qualitative episodes, which is how we usually think of them. In any case, the claim that consciousness is physical has to be shown as the conclusion of a settled theory of consciousness, which we don’t have yet; we can’t assume it’s the case. To naturalize consciousness isn’t necessarily to physicalize it.

    1. What is it to naturalize consciousness without physicalizing it? To endorse mental/experiential fundamentalism or (strong) mental/experiential emergentism?

      1. My current speculation is that conscious experience might involve a species of representational content (qualitative content), and if we can naturalize content in a mature science of representation, that might help explain consciousness. I’m not sure if content can be aptly characterized as physical, but if not it still looks to be a natural phenomenon that arises in concert with physically instantiated mind systems like ourselves that represent reality. I’ve written this up in a 2019 Journal of Consciousness Studies paper, “Locating consciousness: why experience can’t be objectified” (Google will get it for you). I’m not sure if this proposal qualifies as strong emergentism but it’s definitely not mental fundamentalism, nor is it substance dualism.

        Btw, I appreciated your comments about epiphenomenalism and would be interested to see any work you’re doing on consciousness.

  13. It reminds me of Tim Manchin’s take down of homeopathy about water remembering all the poo that’s been in it. Wouldn’t these conscious molecules/atoms be conscious of all the poo they’ve been a part of?

    1. Indeed, the poo itself is conscious, but is it conscious of human disdain for it? Now that would be a Revenge of the Zombies to be scared of—cf. my #14 comments above.

  14. Galen Strawson summarizes panpsychism as being the “real physicalism”. That is, conscious experience is the key thing the physicist *has* when she sits writing QFT equations – it is the one thing you are really sure of. And if she is not a dualist like modern computationalists seem to be (eg if you were a simulation in a computer, they argue you would have the same phenomenal experiences as a meat person), then she wants something physical in the model of the world to explain why redness is like that and not like something else, and what one is “seeing” when asked to imagine a pink ice cube or a unicorn. We think, in this era of computer graphics, that we understand what it is to visualize a non-existent object, but many would see this as implying a regress – it is our minds interpreting the picture on the screen, and writing software that generates the analogous outside appearances (“yeah, that’s exactly what a unicorn looks like in my imagination”).

    1. I have some sympathy for Strawson’s position. If indeed, if everything has some quantum of consciousness, and that the ‘quality’ of consciousness depends on the arrangement of matter, then a brick may have an impoverished form of consciousness that might be active over years or centuries rather than the second or two our (human) consciousness operates.

      I am reminded of Sagan’s, “We are a way for the Cosmos to know itself”. Poetry? Perhaps.

      1. Can you tell me what the phrase “quantum of consciousness” means?

        If all you mean is ‘small portion of consciousness’, phrased that way would make it more worth considering, rather than hopping on the bandwagon of the ‘quantum’ grifters such as Deepak Chopra—“quantum healing” for chrisake.

        But perhaps that’s just extracted from some Strawson diarrhea of verbiage.

  15. Panpsychism is testable. If we can build a computer program that, when running on arbitrary hardware experiences qualia, panpsychism is refuted.

      1. You would know because you can only program something when you understand it. The program would possess qualia because you made it that way. This isn’t circular, but clearly a fair bit of progress needs to be made in philosophy before we can achieve the task. We might even ban the technology on ethical reasons.

        1. “…progress needs to be made in philosophy…”

          Is there any good example of such progress already having been made?

          These people are trying to claim panpsychism is, or will become, part of science, not just filosofical psychobabble.

      2. I disagree with tom here. How does a computer experiencing qualia disprove panpsychism? But how do I know you are experiencing qualia? You tell me you do? Would we believe a computer experiences qualia if it told us it is?

        I’m stuck in my own Cartesian theatre experiencing qualia, or so I believe.

        1. Because it is the computer program that is experiencing the qualia, not the hardware, which is arbitrary. You will know the program is experiencing qualia because you understand the phenomenon well enough to program it in the first place. As things stand, panpsychism is not falsified, of course.

  16. Thirty years ago I would have said that panpschycism would no longer be considered a fruitful way to describe consciousness within one or two decades. But here we are.

  17. It should be mentioned that most contemporary panpsychists don’t simply define panpsychism as the view that all physical things have mental/experiential properties; so they are not committed to asserting that e.g. rocks and books have a mind or consciousness. Instead, they define panpsychism as the view that all physical things have mental/experiential properties themselves or (basic) parts with mental/experiential properties. (Note that the “or” is not exclusive, because a physical thing can be minded itself and have minded parts.) Then they can consistently say that rocks and books don’t have minds but (at least some of) the elementary particles of which they are composed do.

    1. Even in this form the endeavor is still a crock and suffers from all the problems with “pure” panpsychism I already listed. Until panpsychism helps us make progress in understanding science, I am not going to take it seriously. See Hitchens’s Razor.

      1. But isn’t the point of panpsychism that it leaves science alone? It proposes a research program into how phenomenological properties exist in reality – properties that don’t appear in any scientific account of anything.

        1. I think that may indeed be the point, but to my mind that is a very good reason to not take it seriously. If a hypothesis is intentionally formulated so as to avoid testing it against reality while at the same time claiming to describe some aspect of reality, that is a very good reason to not take it seriously.

          1. To be fair to the panpsychics, they are doing metaphysics, not science. There may be no way science can address the problem they are trying to solve.

            1. More germane to human intellectual progress would be: “There may be no way..” they “.. can address the problem..” scientists “.. are trying to solve.”

              There certainly is no way.

              Metaphysics in the hands of this species of filosofer is hardly a step up from pataphysics (cf. Paul McCartney’s Joan and Maxwell).

        2. Sean’s point is that it *can’t* leave science alone. If you want mental properties you’ll have to affect scattering amplitudes, which are very well known and correspond to no extra degrees of freedom for new properties.

          1. Nowhere do the panpsychics claim that mental properties affect scattering amplitudes, or even that scattering amplitudes have phenomenological properties, for that matter.

      2. “I am not going to take it seriously”

        Good.
        But I must admit to being surprised that Sean Carroll does take panpsychism more seriously than I ever thought he would. Perhaps he felt the need to give general reasons why it would never meet your “helps us make progress in understanding science” criterion for paying more attention to it.

  18. I Am a Strange Loop by Douglas Hofstadter was a pretty fascinating crack at consciousness. The math overwhelmed me, but the description of how a consciousness emerges was interesting (and really disturbing). At one point I believe he described it as a “hallucination of a hallucination”.
    As for panpsychism, it is just some kind of spiritualism and certainly not scientific. I think a lot of people just like something they can be in awe of. Frankly, I would include myself in that group. I think Jonathon Haidt is very much on to something when he describes awe as a distinct emotion. A big problem with the world is what triggers it. I can have a sense of awe when Sean Carrol describes the cosmos, but probably have the same amount of awe when I watch Loki. Others get it from institutions, like Catholicism, that can all too quickly turn authoritarian.

Leave a Reply