Thank you for laying out your initial position. Let me immediately address the major points as they currently stand.

No, I don’t think that consciousness is a sui generis problem. That wold be rather unparsimonious, a move to be invoked only in desperate situations. As you point out, the fact that consciousness is an unobservable is not — per se — an issue for science, which is used to plenty of unobservables (e.g., electrons). Contra your statement, we do postulate consciousness in order to explain certain behaviors — that, for instance, is the whole point of Turing-type tests, and the problem comes up when biologists investigate the degree of consciousness in other animals.

The difference with other phenomena is that we also experience it. But this is an advantage, not a limitation. Consider an analogous situation: life. It used to be that people would make the kind of argument you are putting forth to the effect that there was something special, irreducible to materialism, about life. They called it élan vital, vital essence. And yet, biophysicists and molecular biologists have done away with that notion as superfluous more than a century ago: life is a particular type of physical phenomenon, best explained not at the level of fundamental physics — though of course it is compatible with everything fundamental physics holds — but at the levels of the disciplines I just mentioned.

The analogy between consciousness and life is particularly apt because we experience being alive, just like we experience consciousness. And we face similar scientific problems: to determine whether other things (advanced AI, viruses, alien forms) are alive or not. Again, just as in the case of consciousness. So the way I see it you are assuming that the problem of consciousness is special while it isn’t (though it is “hard” to crack, just like the problem of life), and you are postulating the consciousness equivalent of an élan vital, for which there is no need. Incidentally, I’m not alone in making this parallel, you will find it also in Sabine Hossenfelder’s “Electrons don’t think.” She is a physicist, providing an additional science-based point of view to my own as a biologist. I will get back to her take on the issue below.

Now to the Russellian view, which you summarize as “the idea is that physics only tells us about the causal structure of the physical world – what things do – and leaves us completely in the dark about the intrinsic nature of matter that realizes that structure.”

As you yourself note, there is no agreement in philosophy about the necessity to invoke any intrinsic nature. Indeed, even after reading your letter I don’t have a good grasp of what you mean. For instance, do you think physicists are missing something crucial because they don’t give us an account of the intrinsic nature of electrons? What, exactly, would such an account look like? Electrons are theoretical constructs invoked to explain the empirical data, and if we take their interpretation as elementary particles, then there is no intrinsic further to talk about. How is this different from consciousness, or any other property of matter, at any level of complexity?

Hossenfelder has argued that the notion of elemental consciousness is incompatible with modern physics, on the ground that it should make a difference in our experiments. It doesn’t seem to, since particles’ behaviors appears to be entirely accounted for by their known properties according to the Standard Model. You may argue that this assumes physicalism, materialism, or some other kind of unpalatable “ism,” but my position is that if something is not compatible with scientific findings, or does not stem out of established scientific theories, or is based entirely on non-empirical arguments, then it is a non-starter. In the specific sense that it may be logically coherent, even true, but if we don’t have an empirical handle on it, then it is useless as an explanation of anything.

This isn’t “scientism” on my part (I’m actually a critic of that intellectual attitude), but rather a form of reasonable skepticism of what used to be called “first philosophy,” the notion that somehow we can think hard about stuff and discover new things about the cosmos. That project died with Descartes, who carried out the last valiant attempt in that direction.

You invoke coherence as a criterion, but there are two problems with that: first, coherence with what? The Russellian account may be internally coherent, but it doesn’t join up with anything from physics or biology, and indeed — if Sabine is right — it is incoherent with fundamental physics. Second, coherence is far too low a bar. Lots of notions are coherent but not true. Which is why we need empirical evidence, the only criterion that is potentially capable of reducing the essentially infinite number of logically coherent notions about the world to the one notion that best corresponds with the world as it actually appears to be.

Cheers,

Massimo Pigliucci

Massimo Pigliucci

Starting with the problem of consciousness, I can’t see why we’d ever need to postulate consciousness to explain behaviour. Why wouldn’t a mechanism do? It’s not true that the Turing test is about consciousness; rather it concerns thought (indeed, in his classic paper, Turing redefined ‘thought’ to mean passing his test, thus making it completely about observable behaviour).

Yes, the difference with consciousness is that we experience it, but the point is that our most fundamental way of knowing about experience is by attending directly to experience itself. When I attend to my feelings and experiences, I am immediately aware of their qualitative character: the redness of a red experience, the smell of coffee, the distinctive feel of pain. It is those qualities we know about through attention to our experience that are the basic phenomena a theory of consciousness needs to explain, rather than anything known on the basis of third-person observation and experiment. This is what makes the problem unique: in any other scientific problem, we are trying to explain what we know on the basis of observation and experiment, in this case we are trying to explain what we know in a quite different way, i.e. through direct attention to our experience.

I reject materialism because that I don’t think these qualities – the qualities we know about through our immediate awareness of our own experiences – can be captured in the purely quantitative vocabulary of physical science. That is not to say that those qualities do not have quantitative structure: colour experiences can be mapped out in terms of the similarity space of (the conscious analogues of) hue, saturation and brightness. But this kind of abstract mapping doesn’t fully capture, e.g., the redness of a red experience. And this expressive limitation entails an explanatory limitation. For if we wanted to explain those qualities in the purely quantitative language of neuroscience, we’d have to be able first to convey them in that vocabulary before explaining them in terms of more fundamental physical process. If neuroscience can’t even express those qualities, then it certainly can’t explain them. That’s why I think I physical science alone cannot explain consciousness and hence we must turn to alternative ways of accounting for it.

This should not surprise us, as physical science is designed to account for what we know on the basis of observation and experiment, something it does that very well. But the basic datum of consciousness is known in a quite different way.

You say this is like the problem of life because we experience life. But the difference is that we don’t have any special way of knowing about the phenomenon of life. We know about life through observation and experiment, and science is very good at explaining what we know through observation and experiment. I guess in a sense we also know about life through experiencing what it’s like to be alive, but this just gives us access to forms of consciousness associated with being alive. We are not directly aware of the essence of life in the way we are directly aware of the essence of consciousness.

Turning to panpsychism, I think it’s really important to carefully distinguish three questions:

1. Is panpsychism coherent?

2. Is there empirical reason to doubt panpsychism?

3. Is there reason to believe panpsychism?

Let’s begin by assuming (for the sake of discussion) that panpsychism is coherent and addressing (2). I think Hossenfelder misunderstands the view she’s attacking (and for what it’s worth, I think you’ll be hard pressed to find a philosopher of mind on either side of this debate that would disagree with this). When one first hears about panpsychism, one thinks it’s the view that in addition to its physical properties – mass, charge, spin – a particle also has non-physical consciousness properties. That kind of panpsychism would lead to the kind of problems Hossenfelder points to, because we’d want to know what the consciousness properties of particles are doing over and above their physical properties. But Russellian panpsychism is very different: the view is that mass, charge and spin are forms of consciousness. If that makes sense (which we’re currently assuming), i.e. if micro-level forms of consciousness are identical with the properties invoked in the standard model, then clearly it’s mistaken to wonder what these forms of consciousness do over and above the properties of the standard model (because this implies that they’re distinct, when ex hypothesi they are identical).

But how on earth could mass, charge and spin be identical with forms of consciousness? This brings us to (1). You seem to suggest that the postulation of intrinsic natures is incoherent if particles are elementary. I’m happy to accept that quarks and electrons are fundamental, but we still need to ask about the nature of their properties. In my view, physics tells us what mass, charge and spin do (or more precisely the behavioural dispositions they endow to their bearers) but does not tell us what they are. Hence, it is coherent for the panpsychist to suppose that they are forms of consciousness.

Of course, just because something is coherent and not contradicted by physics, that doesn’t give us any reason to believe it. This brings us to (3). The case for panpsychism is built on the arguments against materialism. I believe that we cannot account for the qualitative reality of consciousness in the purely quantitative terms of physical science. But we know that qualitative consciousness exists, so we have to fit it into our overall theory of reality somehow, and Russellian panpsychism gives us a way of doing this. It’s justified by a kind of inference to the best explanation: once we’ve rejected materialism, it’s the simplest account of how consciousness fits into reality.

Look forward to hearing your thoughts!

Philip Goff

Philip Goff

Thanks for several clarifications you have provided. At least now I understand better your position.

Let’s start with the need for consciousness to explain certain behaviors. I didn’t mean to imply that the Turing test is a good way to infer consciousness, it most clearly isn’t, for the reasons you briefly explain. I meant to say that we do ask ourselves whether certain beings are conscious (other primates, dogs, cats, possible future AI), why other beings don’t seem to be (plants, bacteria), and why the question doesn’t even arise for still other things (all inanimate objects). Behavior isn’t enough, but a combination of behavior plus knowledge of physiology and functional anatomy are a good start. It is because other primates have a large and complex brain, while plants have no brain at all, that we think the first are conscious and the latter aren’t.

You reject my analogy between consciousness and life on the ground that we need consciousness to feel alive. Fair enough. But the fact remains that “life” seems to be a qualitatively distinct form of matter from non-life, just like consciousness is from non-consciousness. And yet no biophysicist thinks either that the problem of life is beyond the reach of science or that we need to postulate aliveness as a fundamental property of matter. That’s why I see panpsychism as analogous to the old-fashioned élan vital.

You keep contrasting “qualitative,” as in the experience of consciousness, with “quantitative,” as in what science is good at. But all sorts of sciences use qualitative methods to investigate certain phenomena, so I suspect there is a lurking ambiguity here in the way you use these terms. Could you explain?

When you say that neuroscience cannot “express” the quality of consciousness, what exactly do you mean? If neuroscience were one day capable of delivering a complete mechanistic account of how consciousness is made possible, what would be missing? Don’t say “the experience itself,” since that would be a category mistake: we are talking about explaining the experience, not having it.

When you talk about “alternative ways” to account for consciousness, are you invoking some kind of dualism? If so, what kind?

Now to your three questions about panpsychism, which — you are absolutely right — it is useful to distinguish.

(1) Is panpsychism coherent? I will return to this in a moment. For now I will simply note that coherence is an extremely low bar (though not as low as Chalmers’ infamous “conceivability”). All sorts of false notions are coherent.

(2) Is there empirical reason to doubt panpsychism? This strikes me as exactly the wrong question. The issue is whether there is empirical reason to consider panpsychism. It isn’t the default hypothesis, its current status is that of an extraordinary claim, so we need proportioned evidence in its favor before we can take it seriously.

Thanks for clarifying where you think Hossenfelder goes wrong. The problem is, she at least provides a coherent view of panpsychism, and one that would be empirically testable to boot. By contrast, when you say that mass, charge, and spin are forms of consciousness I think you crossed into incoherence, and so I withdraw my provisional assent to (1). What does that even mean? Is there, for instance, a different kind of consciousness that accounts for each of the fundamental properties of matter? How does consciousness account for such properties? Why do physicists think that it is other things, like the Higgs boson, that account for some of these properties, like mass? Are they wrong? What reason do you have to say so?

You are happy to agree that quarks and electrons are fundamental, but then you ask about the nature of their properties. Physics does provide such an account already. In terms of measurable things like spin, mass, charge, etc.. What else is needed, and why?

As for (3), reasons to believe panpsychism, your only reason seems to be your belief that materialism is incapable of accounting for consciousness. But that strikes me as a massive instance of begging the question: what is under debate is precisely whether the scientific view of the world will or will not be capable of accounting for consciousness. Your version of (3) simply assumes that it isn’t and declares victory.

Moreover, to claim that panpsychism is the result of an inference to the best explanation is more than a bit strange. The strongest inference, seems to me — based on the physics, chemistry, and biology we know — is by far that consciousness is an evolved phenomenon, firmly rooted in physicality (it requires complex brains, according to all the evidence we have), even though at the moment we don’t have a complete explanation of how it works.

Finally, you are not actually giving any evidence for panpsychism, you are simply arguing a priori that it must be the best explanation. But as I wrote last time, I believe first philosophy died with Descartes: we can’t arrive at firm conclusions about how the world works by simply thinking about it. We need evidence. Which means that we need science.

Bottom line: do you think one could ever have positive empirical reasons to accept panpsychism?

Massimo Pigliucci

Massimo Pigliucci

Thanks again for your response, really enjoying the discussion!

Let me start where you end, with ‘big picture’ considerations. How should we go about having our best guess at what reality is like? I suspect you are what I call a ‘conservative naturalist’, someone who thinks the following is the rational way to uncover the nature of reality:

Conservative naturalism: Work out the simplest theory of reality that can account for the data of observation and experiment.

A conservative naturalist holds that the job of science is to account for the data of observation; if you can find the simplest theory able to account for that data, your job is done.

In contrast, I’m ‘liberal naturalist’, because I think there is another datum to account for, over and above the data of observation and experiment: the reality of consciousness. This is something we know to exist not through observation but rather through our immediate awareness of our own experiences. I’m not sure what you mean by ‘first philosophy’, but here’s my philosophical methodology:

Liberal Naturalism: Work out the simplest theory of reality that can account for the (A) data of observation and experiment, and (B) the reality of consciousness.

Your perennial question is ‘What’s the evidence for panpsychism?’ I don’t think there is any observational evidence for panpsychism, just as there isn’t any observational evidence for materialism, or dualism, or any other theory of consciousness. For there to be observational evidence for a theory of consciousness, there’d have to be observational evidence that consciousness exists. But there isn’t because consciousness is unobservable. If one is a strict conservative naturalist, one wouldn’t postulate consciousness at all (I respect Dennett for his consistency). I think the case for panpsychism is that it’s the simplest theory that can account for (A) and (B) above. Dualism is less simple than panpsychism; materialism is incoherent (more soon); so panpsychism wins by a process of elimination.

I’d like to get specific about what goes on in the scientific study of consciousness. Given you agree Turing test has nothing to do with consciousness, I don’t know why you raised it. Here’s my view about how the scientific study of consciousness works. Experiences can’t be observed. But what we can do is ask a person what s/he is feeling and experiencing. And if we scan her/his brain at the same time, we can gather data about correlations between visible brain activity and invisible experiences.

Note that this is already transgressing conservative naturalism, as we’re trusting what people say in order to gather data about essentially unobservable phenomena. As Dennett realizes, the only data a strict conservative naturalist should feel obliged to explain are the behaviour associated with consciousness: what Dennett calls ‘heterophenomenology.’ Moreover, this isn’t a theory of consciousness, it’s just correlations. And neuroscience cannot go further than this because consciousness cannot be observed. So we need to turn to philosophy and try to work out which philosophical theory best accounts for these correlations. For reasons outlined above and below, I think that theory is panpsychism.

I think you mischaracterize me in the following:

‘You reject my analogy between consciousness and life on the ground that we need consciousness to feel alive.’

No, I didn’t. I rejected the analogy on the grounds that we have a special way of knowing about consciousness but we don’t have a special way of knowing about life. You haven’t responded to this point.

I turn now to my objection to materialism: that you can’t explain the qualities of consciousness in the purely quantitative vocabulary of physical science. You ask what kind of explanation I’m looking for. If you know enough chemistry, you understand why water boils at 100 degrees. Analogously, I think if materialism were true, then sufficient knowledge of neuroscience would allow one to understand why a certain kind of brain activity yields experiences involving specific qualities, such as the redness of a red experience. But to do this, you’d need to be able to describe that red quality in the purely quantitative language of neuroscience (and then explain it in more fundamental terms). But this plainly can’t be done; if the predicates of neuroscience could convey what it’s like to see red, then a colourblind neuroscientist would be able to know what it’s like to see red by reading relevant neuroscience.

Later in the letter you accuse me of begging the question against materialism. The above is my argument against materialism. What is your response?

Turning to the question of the coherence of panpsychism, it’s broadly agreed by philosophers of science that physics characterizes physical properties in dispositional terms. And one popular view in philosophy (completely independent of panpsychism) is that dispositional properties are grounded in categorical properties. The panpsychist accepts this view and holds that the categorical properties that ground the dispositions of physics are experiential properties. If you think panpsychism is incoherent, I think the onus is on you to tell me either why the popular philosophical view that dispositions are grounded in categorical properties is incoherent, or why it’s incoherent to suppose that experiential properties could ground dispositions.

Cool point about the Higgs boson though! Although, as I understand it, it’s the Higgs field rather than the Higgs boson that grounds mass. Taking this into account, I think I’d be inclined to hold that mass is a non-fundamental disposition that’s grounded in the dispositions of electrons/quarks and of the Higgs field to interact with each other as specified by the standard model. This does nothing to undermine the point that physics tells us only about dispositions and remains silent on their categorical grounds, which allows the panpsychist to propose experiential properties as their categorical ground.

Over to you. I’ll be on vacation now until 6th Jan. Happy holidays!

Philip Goff

Philip Goff

I accept the label “conservative naturalist.” But I think that’s the only kind of naturalism. You call yourself a liberal naturalist, but you are arbitrarily making a single exception (consciousness) to naturalism, solely on the ground that you don’t see how else the problem can be solved.

By “first philosophy” I mean what Aristotle meant: the sort of speculative metaphysics you (and many others, to be fair!) are indulging in. You assume that one can discover new things about the world by just thinking about it. I should have thought that research program in metaphysics ended with Descartes, replaced by natural science.

More simply: if you think that your theory does not, and cannot, make contact with empirical reality, then you simply don’t have a theory. You have a speculation that can never be tested. It may be true, but there is no way to find out. Why pursue it, then?

You say that there is no (possible) observational evidence for panpsychism, just as there is no observational evidence for materialism, or dualism, or any other theory of consciousness. Setting aside that materialism is much more than just a theory of consciousness, what you are saying is that there is no evidence for philosophical accounts of consciousness. Agreed. That’s why I think those accounts are dead ends. All of them. If we ever get a decent theory of consciousness, it will come from empirical science. If it doesn’t, then we simply won’t have any such theory. I’m okay with that possibility. Human beings are not omniscient, after all.

You keep insisting that the problem is that consciousness is unobservable, but obviously unobservability cannot be the problem. As I already pointed out, science deals with plenty of unobservables, and you have given me no reason to think that consciousness is special in that respect. Sure, we experience consciousness and not the charge of an electron. So what? From the point of view of science, they are both unobservables. Indeed, I shall think that being able to experience consciousness is an advantage the neuroscientist has over the physicist, who can’t ask electrons what’s going on during her experiments.

You claim that neuroscience can go no further than correlation, but that’s wrong as well. When scientists manipulate systems experimentally, the data they get provide information about causality, not just correlations. And, again, I see no reason why neuroscience should be exceptional in this respect.

You say that you rejected my analogy between consciousness and life on the grounds that we have a special way of knowing about consciousness but we don’t have a special way of knowing about life. Of course we do: we are alive ourselves. But my point was different, so I’ll try to make it more clear: consciousness is a property that seems to emerge (no “heavy” meaning here, just descriptive) when certain kinds of matter are arranged in certain ways. The same goes for life. Now, biologists used to invoke a special principle, the elan vital to explain what they thought was unexplainable by physical-chemical means. Fortunately, they now know better. I think panpsychism is a throw back to the days of the elan vital, and it’s equally misguided.

You also insist in claiming that science can only deal with quantitative phenomena. I repeated several times that this is false. Plenty of good scientific research is done on qualitative properties, for instance in biology.

Helpfully, you state that if the predicates of neuroscience could convey what it’s like to see red, then a colorblind neuroscientist would be able to know what it’s like to see red by reading relevant neuroscience. This is a pseudo-problem rooted in an ambiguous use of the term “know.” Once a neuroscientist can provide a detailed causal account of how it is possible for a human being to see colors, her job is done. Nothing else can reasonably be asked above and beyond that. Obviously, causal knowledge of X will not thereby engender experience of X, and to set that bar is just bizarre. Here is an analogy: if I give you all the knowledge necessary to understand how bicycles are ridden, would you say that something is amiss because after reading my paper on bicycle riding you don’t feel the experience? Experiential knowledge is a different beast from theoretical knowledge. Science isn’t going to give you the experience. It is not in the business of giving you the experience. It’s in the business of giving you a causal account of how it is possible for you to have that experience.

Regarding the distinction between dispositional and categorical properties, and the notion of grounding, I’m extremely skeptical — as both a scientist and a philosopher of science — that they do any useful work at all. Even the Stanford Encyclopedia article drily observes that “with respect to this question [of what categorical properties are] we are in no better position than we are with respect to dispositional properties.” Oh boy. If you’d like to clarify why you feel this is a necessary sub-discussion, and provide me with concrete examples, I’d be happy to address it.

Finally, you still have not explained to me the bombshell claim you made last time: that mass, charge and spin are forms of consciousness. What does that mean, exactly?

Mass has a number of definition in physics, including inertial and gravitational. Charge corresponds to the time-invariant generator of a symmetry group. And spin is an intrinsic form of angular momentum carried by elementary and composite particles. Where does consciousness come in, in any of that? What work does it do?

One last question, if you don’t mind: the subtitle of your book is “Foundations for a New Science of Consciousness.” In what sense is what you are proposing a science, given your own admission that it makes no contact with the empirical?

Cheers,

Massimo Pigliucci

Massimo Pigliucci

My final letter sums up the central issues.

1st Core Issue: The Unobservability of Consciousness

A key point I have been emphasizing is that consciousness is unobservable. I’m surprised that you keep repeating that science often deals with unobservables, given that I’ve acknowledged this many times (indeed, did so when I first made the point!). What is unique about consciousness as a scientific phenomenon is that the basic data we are trying to explain are not known on the basis of observation. What we want a theory of consciousness to explain are the qualities of experience, e.g. the qualitative character of red experiences. These qualities can only be known about by attending to experience from the 1st person perspective; they are invisible to 3rd person observation. This makes the problem of consciousness utterly unique: in every other scientific problem, we are trying to explain the data of 3rd person observation.

In your last letter, you didn’t quite deny this but wondered why it mattered. This brings us to...

2nd Core Issue: The Case Against Materialism

If there were a neuroscientific theory that explained the subjective qualities of experience – e.g. what it’s like to see red – that theory would have to articulate those qualities before accounting for them in terms of physical processes in the brain. But the purely quantitative vocabulary of neuroscience cannot capture these qualities: you can’t express in these terms what it’s like to see red or to feel pain.

You claim this is a pseudo-problem. Either:

(A) I’m demanding the standard scientific approach,

…or:

(B) I’m demanding that reading a scientific theory of red experience will give you a red experience.

Of course, I’m not demanding (B). This is simply an incorrect interpretation of the knowledge argument. You and Patricia Churchland were pressing this interpretation on twitter, and Frank Jackson emailed me and others to say that of course that isn’t what the argument is. As the name of suggests, it's about knowledge. The claim is that there’s a kind of knowledge/information one gets from experiences that can’t be got from physical science, namely knowledge of the qualitative character of experiences.

You say that this relies on an ambiguity in the word ‘know’, but you don’t say what the ambiguous meanings are. Perhaps you’re reaching towards Conee’s ‘acquaintance hypothesis’ response or Lewis and Nemirov’s ‘ability hypothesis.’ These are not so popular these days, largely due to Loar’s point that sentences concerning what it’s like to have experiences can be put into valid arguments, which suggests they have propositional content. A much better response comes from the ‘phenomenal concept strategy,’ although this has problems too. In any case, I don’t think you can just dismiss the knowledge argument in a slapdash manner without taking the time to properly understand how it works and investigating whether the response you’re initially attracted to has well-documented problems that has led most materialists who work on this stuff to discard it decades ago.

You say:

"…if I give you all the knowledge necessary to understand how bicycles are ridden, would you say that something is amiss because after reading my paper on bicycle riding you don’t feel the experience?"

No, because in this case you’re trying to teach me how to ride a bike rather than trying to explain the subjective character of the experience of riding a bike.

So my demand is not (B) but nor is it (A). For what science standardly does is explain the data of observation. But what we want from a theory of consciousness is an explanation of why experience has the character it does, and the character of experience can only be known from 1st person perspective. The point is:

THERE IS SOMETHING THAT NEEDS EXPLAINING THAT CAN ONLY BE KNOWN ABOUT FROM THE FIRST-PERSON PERSPECTIVE
Your entire response to my case against materialism has involved dodging this point, so I’ve made it really big in my last attempt to render it un-dodgeable.

3rd Core Issue: Is Panpsychism Coherent?

Towards the end of your last letter you mock the view that mass, spin and charge are forms of consciousness but don't provide an argument against its coherence. You express skepticism that categorical properties do explanatory work, but this point is not relevant to whether the view is coherent. You wonder what other examples of categorical properties might be, without giving any grounds for rejecting the example I have already given, namely experiential properties. You then say some random stuff about the physical properties in question, but I couldn’t see how any of it disputes my (pretty uncontentious) claim that physics only trades in dispositional properties (spin isn’t quite angular momentum, but a dispositional property so named due to its resemblance to macro-level angular momentum). I recently discussed these topics on the physicist Sean Carroll’s podcast, and he didn’t dispute this specific claim. Overall, I can’t see any reason in what you’ve said to doubt the coherence of the claim that experiential properties are the categorical properties underlying those dispositions.

4th core issue: Why Should We Believe Panpsychism?

We both agree that if we’re just trying to find the simplest theory able to accommodate the data of 3rd person observation and experiment, we won’t have any justification for believing panpsychism. But I think there’s another datum that can’t be known on the basis of 3rd person observation: the subjective qualities known about by attending to experience. Panpsychism, I believe, is the simplest theory able to accommodate both 3rd person observation and experiment and the subjective qualities of experience (this is obviously dependent on the above rejection of materialist accounts of the latter). This, I believe, gives us strong reason to take the view seriously.

Thanks for a very enjoyable discussion!

Philip Goff

Philip Goff

Thank you for this dialogue. It has been illuminating for me, and hopefully it will be useful to our readers.

In your last contribution you repeated once more that first person experience is not accessible by third parties. Which is obviously true by definition. The contention is whether this presents a problem or not for scientific investigation. After some back and forth, we both agreed that unobservability per se isn’t the problem, since electrons too are not observable, for instance. Are you then discarding a lot of what psychology and cognitive science has done since the demise of behaviorism? Because part of the business of those sciences is to systematically study first person phenomena, including people’s intentions, motivations, emotions, and so forth. All of which are not directly observable, and become data only via self-reporting. That has not been an obstacle to the scientific investigation of those phenomena, which we can even study experimentally, for instance by inserting electrodes in the brain, or using localized magnetic stimulation and asking the subjects what they feel. Why do you think this is an issue at all is beyond my comprehension, frankly.

So let me try to be just as clear as you tried to be with me (but without using all-capital letters): a scientific theory of consciousness—if we will have one—will provide a detailed mechanistic understanding of how the human brain generates first person experience, using people’s self-reports as data. Once we have that, there is nothing above and beyond it that requires further explanation. We would be done.

What you call “knowledge of qualitative experience,” and allege to be beyond scientific reach, I call experience. You are using “knowledge” is a very loose fashion, which allows you to equivocate on the subject matter, as Patricia Churchland pointed out on Twitter. As I said, experience provides the raw data on which psychologists and neuroscientists work. But you seem to want it both ways: on the one hand, you are baffled by my distinction between scientific knowledge and first-person experience; on the other hand you keep telling me that the two are so different that there is no possible bridge between the two. Which one is it?

You are convinced that you or others have made a case against materialism. I’m a scientist, and I’m concerned with how we discriminate between competing accounts on the basis of facts. No metaphysical position can be rejected on that ground, because they are all compatible with the world as we observe it. But methodologically speaking, assuming materialism has led to the incredible successes of science over the past several centuries. What, exactly, have we discovered about the world by going with your ontology?

To put it otherwise, I like to keep my ontology close to my epistemology, meaning that if I can’t justify an ontological claim, I simply won’t make it, because I think there is insufficient ground to believe it. The problem with your approach is that you are far too comfortable with a vast gulf between your ontological speculations and your epistemic warrants for such speculations.

You insist that panpsychism is coherent, but there too you keep falling into an ambiguity. If by coherent you mean logically so, then sure, we agree. But literally an infinite number of models of the world are logically coherent. That doesn’t help at all. We want coherence with the kind of data and theories we get from science. And that’s where panpsychism spectacularly fails. There is absolutely nothing in modern physics or biology that hints at panpsychism, and you have acknowledged that no empirical evidence could possibly bear on the issue. That acknowledgement, for me, is the endpoint of our discussion: once data are ruled out as arbiters among theories, those theories become pointless, just another clever intellectual game.

Which leads me to my next issue. I note with a bit of surprise that you entirely ignored my more basic criticism of your way of doing metaphysics. You seem convinced that analytical metaphysics, the kind of approach developed in ancient Greece and that I would have thought died with Descartes, is still a valuable project. You are not the only one, of course, David Chalmers is another prominent advocate. But this is simply a rabbit hole that leads to an absurd proliferation of “coherent” or—worse yet—simply “conceivable” scenarios that tell us absolutely nothing about how the world actually works. What is the problem? Simple: as I mentioned above, there are infinitely many coherent accounts of the the way things are, but only one of them describes the actual world. Logic and argument, by themselves, are incapable of winnowing things down to a reasonable number of small alternatives. You need empirical data. If your account is impervious to empirical verification then it is dead in the water. Not false, necessarily. Just irrelevant.

What should metaphysicians do, then? What people like James Ladyman and Don Ross have been doing: re-conceive the field as being in the business of articulating a comprehensive view of our understanding of the world as it emerges from the special sciences. This is because the sciences themselves are too specialized and narrowly focused to attempt anything of the kind. Moreover, metaphysics in particular, and philosophy more generally, should be pressed into the task—outlined by Wilfrid Sellars—of reconciling the scientific and the manifest images of the world, so that we can make more precise sense and better use of the kind of normative concepts that do not enter scientific vocabulary and yet are indispensable for human life. That would make metaphysics and philosophy actually relevant, for a change. The path you, Chalmers and others are attempting to chart has already been tried, centuries ago, and has brought us—as David Hume put it—nothing but sophistry and illusion.

Cheers,

Massimo Pigliucci

Massimo Pigliucci