From what I have been able to gather about this fraught field, Goff’s argument is both clear and persuasive, though I would add one caveat. Its force seems greatest against functionalist or “neo-behaviorist” forms of physicalism, rather than every possible physicalist account. Even so, his central point is compelling: experience is known from within, not merely inferred from behavior. Any theory that explains consciousness by explaining that away risks mistaking its abstraction for reality.
Every form of "physicalism" that holds that consciousness can be fully explained in the terms of physical science falls prey to these worries.
We can define "physicalism" more broadly, as Galen Strawson does, so that even panpsychists are "physicalists", but I don't think that's a good use of words.
The boundaries of reality themselves may need to be rethought. Consciousness does not have to occupy a reserved gap in an otherwise complete physical picture; it may instead belong within a more general framework of reality.
A broader framework is exactly what’s needed. Consciousness isn’t a gap in physics — it’s what a pattern looks like from the inside when its internal measurement loop closes. Physics gives the external readout; experience gives the internal one. Same system, different scales of coherence.
The introspection test is doing real work here, and I think it succeeds against illusionism and against the ostensive/functionalist move both — pain isn't a placeholder for "something, I know not what" and it isn't exhausted by what it does. But I'd push back gently on where that leaves us. The argument, as run, is really Russell's: physics gives us the relational skeleton of things and leaves their intrinsic nature open, and introspection is our one glimpse of that intrinsic nature from the inside.
Panpsychism is one way of populating that gap — put some flicker of the inside at every level, down to the electron.
But it isn't the only way, and it isn't obviously the most economical one. You could just as well say the inside appears only where structure is organized reflexively — where a system represents itself to itself — rather than smearing it across everything physics quantifies over. That keeps your entire argument intact (physicalism still can't be the whole story, introspection still reveals something behaviour can't capture) without asking anyone to grant electrons a point of view. The homework rules out physicalism. It doesn't, on its own, rule in panpsychism over a more restrictive Russellian monism.
Reflexive organization is the right pivot. A system becomes conscious when its internal measurement loop closes — when the pattern can read its own state from the inside. Behavior is the extrinsic readout of that pattern; experience is the intrinsic readout. That scale distinction does more work than any metaphysical commitment.
Here’s a thought experiment (in the vein of Frankish and Dennett’s drone): you have a robot with an LLM and a camera. The camera feed goes to a pattern recognizing unit and the output of that goes to the LLM as special tokens. The LLM is trained as usual but with the extra tokens from the camera unit. Say the pattern recognizing unit can only recognize 5 things: “red”, “green”, “ball”, “chair”, movement left to right. The system also recognizes combinations of these, so potentially “green + ball + moving left to right”.
So what can the LLM say about its experience? When it sees a red chair, it knows for sure it sees a red chair. We can call that introspection. Even if it’s an illusion (“red + chair” token sent directly to LLM without the camera) it still knows for absolutely sure that it’s experiencing a red chair. Does it know the nature of the experience? It knows that it is different from other possible experiences, so it knows something about the experience. But does it know the *nature* of the experience?
We may do well to remember "introspection" is a metaphor. We can't actually turn the eyes around and look within. A computer with data on its own data, although the metaphor of "introspection" is often suggestively used to describe that, is not the same "introspection" we do. It's important not to conflate the two.
What’s the difference between the introspection that we do and the introspection that the LLM does? Preferably, the difference that makes a difference.
Well, if we're going with Bateson ("a difference which makes a difference"), let's start from mind being both ecologically embedded, and ecological within (thus "Steps to an Ecology of Mind").
What we metaphorically call "introspection" that we do is not literally turning our eyes inward. It's obviously quite something else, despite the metaphor being adequate to point to it. When a computer recycles its some of its own data as input, it's literally no different than if it had printed the data out on a tape, and had the tape fed back into it, through the same tape reader as accepts any other data as input to it. If reading tape is how a computer "spects" (see: Turing machine), then computers can literally introspect; we cannot. We're doing something other than the literal meaning of the term here.
Saying more precisely what we're doing may be precisely the same problem as consciousness itself. This may be where there's no "difference which makes a difference" -- regarding our own introspection and consciousness. Our "introspection" and computers' though -- a difference!
Let’s get back to the LLM in the thought experiment. Note: I’m not talking about “the computer”, I’m talking about the system that is taking tokens as inputs (representing both words and the patterns recognizing things from the camera). This system cannot tell you what tokens were input. All it can tell you is what it is disposed to tell you when those tokens get input. The question is, is that different from you “introspecting” and saying what you are disposed to say when you have your inputs?
James, An LLM is still a Turing machine, so a subset of "the computer." I can "introspect" far more than I can possibly say. Even to approximate much of it would require far better poetic skills than I possess, despite a summer at the Jack Kerouac School of Disembodied Poetics. Ever done any chakra meditation? That sort of bodily awareness is a constant in my "introspection." Now, even the Rivian I have on order, claiming some onboard AI capability to go with a good set of sensors ... nah, I can't see that having anything remotely like human bodily awareness. Can you? Was My Mother the Car prophetic TV?
[I was gonna let this go, but then Philip showed up, so …]
An LLM is not a Turing machine. A Turing machine can instantiate an LLM, but it’s not identical. An LLM has a specified structure and identifiable capabilities. Yes, you can introspect far more than the LLM in my example, and far more than any LLM currently implemented. So what? That doesn’t change the nature of “introspection”, which is what I’m trying to get at. Yes, the self awareness you have of your body is significantly greater than the self awareness of the Rivian. But that’s a difference that doesn’t make a difference to this discussion. The Rivian still has some self awareness. Yes, I can see that.
I disagree that there is no first-person subject. The LLM looks like a subject to me. What’s the difference that makes a difference?
Not sure I want to start this, but … just what is “knowledge in the ordinary epistemic sense”? Why is an LLM’s knowledge that Paris is in France different from your knowledge? Again, what’s the difference that makes a difference?
In the ordinary epistemic sense, an LLM does not “know” anything, because there is no first-person subject there to know it. There is no inner perspective, no awareness, and certainly no “what it is like” to see red or anything else.
If knowledge is at least justified true belief (or, better, properly warranted true belief) then we first need an account of belief. By what metric does an LLM “believe” a proposition in a way that a calculator, thermostat, or simpler computer does not? Without a subject of experience, “knowledge” here seems to be reduced to a metaphor for reliable or useful output, and not knowledge in the ordinary epistemic sense at all.
In the ordinary epistemic sense, an LLM does not “know” anything, because there is no first-person subject there to know it. There is no inner perspective, no awareness, and certainly no “what it is like” to see red or anything else.
If knowledge is at least justified true belief (or, better, properly warranted true belief) then we first need an account of belief. By what metric does an LLM “believe” a proposition in a way that a calculator, thermostat, or simpler computer does not? Without a subject of experience, “knowledge” here seems to be reduced to a metaphor for reliable or useful output, and not knowledge in the ordinary epistemic sense at all.
This article begins with equivocation at a massive scale. For the claim that physical science is exclusively concerned with behavior to avoid looking terminally trite, the word 'behavior' must be taken in the widest and most abstract sense. If you buy into that, then of course the hypothesis that minds are physical processes implies that they are 'behavior' in this sense (as are metabolism and biological reproduction for that matter) and calling it 'neo-behaviorism' offers no additional insight.
What calling it neo-behaviorism might do, however, is encourage the unwary mind to tacitly assume that physicalists are suggesting that the mind is nothing but behavior in the ordinary sense: the overt activity of people and animals. If they had some previous exposure to the history of psychology, they might even suppose that this is a throwback to the simplistic and abandoned doctrine of behaviorism.
Am I suggesting that the incorporation of the word 'behavior' into these discussions is done to deliberately encourage this sort of misunderstanding? Well, nothing in my observable behavior, as manifest in this writing, justifies that conclusion!
Turning now to Philip's homework assignment, part one encourages the reader to accept a simplistic dichotomy. In reality, there is a broad range of partial understanding that comes from our acquaintance with our own sensory experience.
Part two continues in the same manner. Understanding what feelings are is a complex problem, as shown simply by the current state of the science and philosophy of mind after the centuries (or millennia, even) of speculation over the matter, yet here Philip suggests that if you feel you have any understanding of your feelings, then you are likely in a position to tell whether or not it is 'just' a matter of 'behavior'. Note that in this article, the definition of 'behavior' we are working with is the extremely broad one, as discussed in my first paragraph, which should lead us to ponder whether the 'just' in this question makes it a leading one.
Based on the above, I suggest adding a third question to the homework: if you have concluded that introspection reveals to you that pain is more than just a pattern of behavior (in the sense being used here), how would you justify this very strong and far-reaching conclusion to a skeptic who feels that quite a lot about minds is unreachable through introspection alone?
First, I don't think Papineau's ostensive approach is "more subtle" than illusionism. This idea that we're just pointing at something is compatible with illusionism, and I would endorse this idea as well as illusionism. In fact, I think you may be misunderstanding illusionism if you think illusionism says we define consciousness in the first place as behaviour. Maybe that's why you think that it's an argument against illusionism to say "When I talk about Sarah's pain, I'm not talking about her behaviour". No! Illusionism (or at least my illusionism) agrees with Papineau on the ostensive definition. It just says that what consciousness actually is turns out to be neo-behavioural or functional. So then illusionists will (re)define it accordingly, just as (I would say) we redefined water as H2O when we discovered that that is what it is.
Second, I think you're posing a sort of false dichotomy when you say that either we're pointing at something we have no knowledge of, or we're pointing at something we have perfect knowledge of. There's a middle ground.
I agree that it's implausible that we're completely ignorant about pain. But, at least on my view, there are two kinds of knowledge you are potentially conflating.
There is knowledge about how pain feels. This means knowing that pain is aversive, being able to recognise it, recall it, tell what part of the body it pertains to and so on. And there is knowledge about what it essentially really is. I think that introspection reveals the first kind of knowledge but not the second kind of knowledge.
The first kind of knowledge can be cashed out in various different ways, e.g. the ability hypothesis or phenomenal concepts. Everything you think you know about what pain *is* in my view reduces to this first kind of knowledge. You don't have a magical ability to directly perceive metaphysical truth (with or without ingesting toad venom).
You write "There is knowledge about how pain feels. This means knowing that pain is aversive, being able to recognise it, recall it, tell what part of the body it pertains to and so on." NO! How it feels does not "mean" the item list you begin. That's all knowledge _about_ pain, but not how the pain directly feels.
The difference between the pain of a stubbed toe and the pain of a sensitive tooth is not just a matter of location or reaction. The red of a stoplight, to someone color blind, is in the same location and leads to the same reaction, putting the brakes on. Yet the color blind driver does not have the same feeling as, say, Mark Rothko's while stopping his car and observing the feeling of that shade of red.
I stand by what I said. You ignored part of what I said, though. For example, being able to recognise a colour. The colour blind driver cannot recognise the colour.
> The problem is that introspection gives me rich understanding of what my feelings and experiences are. When I attend to my pain, I’m not just pointing inside myself: “That thing, whatever the hell it is.” A pain is essentially defined by how it feels, just as a triangle is essentially defined by having three sides. To the extent that I know how my pain feels, I know what my pain essentially is.
I think this essentialism is a dead end, but sure, if you embrace this essentialist understanding of consciousness, I agree with you that it leaves basically no other option but to consider consciousness non-physical (unless you consider some strongly emergent consciousness as "physical" so long as it strongly emerges from physical stuff).
However, I think Dennett did a great job exposing us to the fact that we can be *very wrong* about our own conscious experience. Introspection is subject to innumerable illusions regarding the contents of our consciousness, for example change blindness (we think we have a detailed visual representation of a scene, but fail to notice surprisingly large changes) or inattentional blindness (the classic "gorilla experiment": people looking for basketball passes often completely miss a person in a gorilla suit. We think we're consciously seeing everything, but we're not. https://www.youtube.com/watch?v=vJG698U2Mvo).
Why should we believe that introspection give us transparent and infallible access to the "essence" of consciousness (or the "essence of pain" to take something more specific). Yes, it's a very strong intuition (I share it), but some illusions *are* impenetrable. If you allow for distrusting this intuition (which I think the qualia realists rarely, if ever, do. If they did, they would take illusionism more seriously; most people just reject it without even comprehending what the position actually is, because they believe it is self-defeating) everything falls neatly into place for the physicalist.
The panpsychist position hinges on the idea that we have, through our armchair philosophizing, access to the "essential nature" of fundamental reality. We don't have to do any science, we don't have to go out there and do experiments and study the nature of matter, or the brain, no no. Just sit in your armchair and introspect, and you will know what consciousness "essentially is" and thereby also what "ultimate reality" essentially is. If that's a satisfying "explanation" to you, go ahead. But to me it seems you haven't actually *explained* anything, you've just defined the problem away, and we're none the wiser for it.
Ah, but the idea that consciousness has phenomenal properties *is* a piece of specific content of your consciousness! Where else could you be aware of this intuition, except through consciousness?
I don't know that it is a *piece* of specific content of my consciousness. There is a fundamental difference between "I think I have the phenomenal experience of seeing everything completely in my visual field" when you're missing the man in a gorilla suit and "I think I have phenomenal experience" and not actually having any - and indeed that not even being a thing that exists.
Additionally, I'm tempted to argue that the illusions don't even make the point that illusionists like to claim they do and instead are just arguing about the way the brain processes things and are then claiming that it implicates phenomenal experience.
For example, when I think I have the phenomenal experience of seeing everything that is going on in the scene but I miss the man in the gorilla suit, I still am having the phenomenal experience of feeling like I am seeing everything going on in the scene, but the reality is that my vision system is not processing the gorilla and flagging it as relevant. That does not mean that my phenomenal experience was an illusion, it means that my phenomenal experience was wrong about what my vision system was doing (i.e. vision system is telling my phenomenal experience that it is seeing everything in my visual field, when really my vision system is not - but the phenomenal experience is still aligning with what the visual system is telling the brain).
Or another example, looking at the change blindness example - that more seems to just be implicating memory and not phenomenal experience. You still have the phenomenal experience that things have not changed. It is just that the underlying brain processes comparing the representation of vision at time t with the memory of the representation of vision at time t-1. That is the phenomenal experience still faithfully experiencing what the brain tells it the current situation is. None of that implicates the phenomenal experience at all, just the brain's visual processing system.
How is "I think I have phenomenal experience" not a piece of specific content of your consciousness? Are you conscious of the fact that you think you have phenomenal experience? If you are, then that intuition is a piece of content in your conscious experience.
The change blindness and inattention blindness isn't meant to implicate phenomenal experience in the way you push back against. They are simply examples that demonstrate that we can be *very wrong* about our own minds, our own consciousness, whatever the introspection is telling us. It's supposed to make you more open to the idea that even though it very strongly seems as if consciousness had phenomenal properties, maybe that just another mistake made by the introspective mechanisms of the brain. In other words, the point is simply that first-person introspection alone isn't a trustworthy source of data. You shouldn't trust it when it tells you nothing has changed, you shouldn't trust it when it tells you you see everything that's going on in a scene in front of you, and so maybe (just maybe, not definitively) you shouldn't trust it when it tells you you have direct access to the inherent properties of consciousness (phenomenal properties). At least it should open your mind to the idea that you could *conceivably* be wrong about something that seems *obviously true* to you when you introspect.
Ah, I was misinterpreting your line. The statement "I think I have phenomenal experience" would be a specific piece of content in my consciousness. I thought you were saying that actual entirety of phenomenal experience itself was just a specific piece of content in my consciousness.
Those points about change blindness/inattention blindness - I think it is demonstrating that we are wrong about external reality, and about what we think our visual system or other systems of our brain are telling us, but not necessarily wrong about our actual phenomenal experience.
(To be clear, I do personally think it is likely the case that we can be wrong about aspects of our phenomenal experience - at bare minimum in the form of us misremembering/not fully remembering what we just experienced - but I think there is at least an argument to be made that the standard illusionist arguments about those things don't carry the same weight that many people seem to accept that they do)
Is it possible that I am wrong about the fact that I have phenomenal experience? The illusionist claim that phenomenal experience doesn't actually exist is, quite literally, incomprehensible to me given what I am experiencing at any given moment. I understand how I could be wrong about specific components of my experience. I don't understand how I could be wrong about the fact that I am having a phenomenal experience at all. That said, is it still possible that I am wrong? Of course. But I think it is more likely that I am wrong about things like a->b, a, therefore b and 1+1=2 and that an external reality exists and that it is impossible to have a four sided triangle than it is likely that I am wrong about having phenomenal experience.
> The illusionist claim that phenomenal experience doesn't actually exist is, quite literally, incomprehensible to me given what I am experiencing at any given moment.
I understand that position very well. It was completely incomprehensible to me as well until very recently. I don't know how I broke through to the point where I could entertain the concept of an "experience" without attaching phenomenal properties to it, but I have, and now it is far from incomprehensible to me. In fact, the illusionist position seems like the obvious default position to hold, at least until we have a much much much richer understanding of the brain.
As Dennett said "when philosophers tell me 'I can't even imagine [this that or the other thing]' my response is 'well, try a little harder. It's completely conceivable to me, so maybe you just need to try harder'.
Surely panpsychism has more philosophical integrity than this caricature allows. It does not claim that introspection reveals the whole of reality, but that consciousness is a basic feature of reality that our scientific picture must somehow accommodate rather than explain away.
I didn't say panpsychism claims that introspection reveals the "whole of reality", that is in fact a caricature of my criticism.
Panpsychism claims that consciousness *is* what reality consists of. Physical science tells us what things *do*, how they *behave*, and how they *relate*, but the intrinsic nature of matter is consciousness, according to panpsychism. Consciousness is what stuff *is*, according to panpsychism.
And it also claims that we have direct access to consciousness through introspection (as Philip writes here, "introspection gives me rich understanding of what my feelings and experiences are" and "I know what my pain essentially is").
I draw the conclusion that panpsychism claims that we can know, through introspection, what "ultimate reality" essentially is. In other words, that we through introspection have access to the "essential nature" of fundamental reality. I do not intend that to be a caricature, and I don't think it is. I have read Galileo's error and heard Philip talk about panpsychism many times, and I don't think I'm confused or being dishonest here. I simply think it's far fetched.
Fair enough; “whole of reality” overstated your criticism. But calling the view “armchair philosophizing” still caricatures the argument. Panpsychism does not claim that introspection alone discloses fundamental reality, but that consciousness may offer one clue to intrinsic nature where physics, by hypothesis, supplies none. You may reject that move; you have not made it disappear.
No. Idealism is the stance that "consciousness *is* what reality consists of." In idealism, reality has but one fundamental: consciousness. In panpsychism, the claim is only that consciousness is among the fundamentals, but not the whole of them.
"Philosophers of science have realized that physical science, for all its richness, is confined to telling us about the behavior of matter, what it does. Physics tells us, for example, that matter has mass and charge. These properties are completely defined in terms of behavior—things like attraction, repulsion, resistance to acceleration. Physics tells us absolutely nothing about what philosophers like to call the intrinsic nature of matter: what matter is in and of itself.
Consciousness, for the panpsychist, is the intrinsic nature of matter. There’s nothing supernatural or spiritual, but matter can be described from two perspectives. Physical science describes matter from the outside in terms of its behavior. But matter from the “inside”—that is, in terms of its intrinsic nature—is constituted of forms of consciousness."
You're conflating "intrinsic" with "fundamental." Both "intrinsic" and "extrinsic" are fundamental in panpsychism, as well as in dual-aspect monism, which is a form of panpsychism. By contrast, only "intrinsic" is fundamental in idealism. Idealism = "the world is all a dream." Illusionism = "the world is not a dream, but consciousness is." Panpsychism = "the world and consciousness are both more than dreams."
It's true that I never saw the gorilla in that experiment, and I was amazed I didn't. However, the viewer was deliberately distracted by the narrator instructing the viewer to count the number of passes one of the team made. And, of course, we don't just passively perceive reality out there, rather our minds play a huge role in shaping and moulding what we perceive.
So yes, we can be wrong about certain things regarding consciousness. However, we can have absolute certitude that we are right now having certain experiences. For how could I be wrong about this? Something like pain is just immediately given. It is immediately known and defined by its painfulness. There is no gap between the experience itself and whatever one suggests it might really be since the pain is quintessentially an experience.
What could be meant by labelling a pain illusory? Regardless of whether, say, an agonising pain is labelled real, or labelled illusory, it will still have the same degree of painfulness. It will feel exactly the same whether we label it a conscious experience or an illusory conscious experience. But that being so, why prefix the word pain or consciousness with the word "illusory"? What function or role is "illusory" playing here since we are not comparing such a illusory pain with a real pain (since allegedly real pains don't and cannot exist)?
> Something like pain is just immediately given. It is immediately known and defined by its painfulness. There is no gap between the experience itself and whatever one suggests it might really be since the pain is quintessentially an experience.
This is the essentialism. Yes, we are having experiences, the illusionist doesn't deny that. What the illusionist denies is that it's "immediately given" or "immediately known", and "defined by its painfulness", and that there is "no gap between the experience itself and whatever one suggests it might really be". In other words, the illusionist is saying that these *phenomenal properties* (they are sometimes characterized as simple, ineffable, intrinsic, private, and immediately apprehended) are illusory. We are having an experience, but we are wrong to think we understand the "essential nature" of the experience. The illusionist can share the intuition (I do) that these phenomenal properties are real, but simply declare that they do not trust this intuition.
So pain isn't illusory, but the idea that pain is not ultimately defined functionally or behaviorally, is illusory. Imagine for example, that we took away, piece by piece, every single behavior associated with being in pain (a strong migraine, for example). Imagine you're in agonizing pain, but you don't scream out. You don't grab your head or lean over. Someone asks you "how are you doing?" and you respond "oh, just fine". You go about your day as completely normal, not even distracted from doing your job as well as you ever do. Behaviorally, or functionally, it's as if you weren't in agonizing pain. But you *are* in agonizing pain. Is this plausible? Is it even imaginable? I would say no. It's this notion of pain, as a separate "non-functional" thing, "non-behavioral" thing, that is illusory. *Qualia* are illusory. But pain is real. Experience is real. Consciousness is real. It's just not what you think it is.
Those who see the gorilla really see the gorilla. They're not wrong. (Yes, I missed the gorilla the first time through back when.) To use proof that we can be unaware of something that we think should be obvious as evidence that what we are obviously aware of should be doubted is not a strong move. You need an experiment in which the video has no gorilla, yet people see one there.
The disjunction has a third option you did not name.
You said: physical science is exclusively about behavior. If consciousness is physical, consciousness must be behavior. Introspection reveals consciousness is more than behavior. Therefore consciousness is not physical.
The first premise is wrong. Physical science is not exclusively about behavior. The Schrödinger equation describes structure, not just behavior. Crystallography describes lattice geometry, not just mechanical response. Thermodynamics describes state functions, not just what systems do when poked. Physics describes structure AND function AND genealogy. You compressed physics into behavior alone and then showed consciousness does not fit inside the compression. Correct: consciousness does not fit inside the compression. The compression was the error, not the conclusion.
Your introspection homework is good. I will answer it directly. Is pain more than a pattern of behavior? Yes. Pain is the entity's constitutive activity encountering a perturbation that exceeds the entity's current absorptive capacity. The projection cost is structural. The projection cost is physical. The projection cost is NOT behavioral: it is not defined by what the entity does in response. The avoidance behavior IS downstream of the pain. The pain IS the perturbation of the entity's self-maintaining organization. The organization IS physical. The perturbation IS physical. The pain IS physical. The pain IS NOT behavioral.
Your disjunction: either pain is behavior (physicalism) or pain is not behavior (not physical, therefore panpsychism). The third option: pain is structural constitutive activity. Structural constitutive activity is physical. Structural constitutive activity is not behavior. The entity's own activity maintaining its own organization IS the entity. The perturbation of that activity IS the pain. No behavior required. No panpsychism required. No electron needs a point of view. The entity that constitutes experiences the perturbation of its own constitution. The entity that does not constitute does not experience. The line is not at the electron. The line is at constitution.
Formalize "consciousness." The disjunction dissolves. The homework has a third answer.
Philip when you say "A pain is essentially defined by how it feels" do you mean "How it feels" as "How pain feels" or do you mean something more like "A pain is essentially defined by being part of the 'how it feels/what it's like' class of things"?
I really enjoyed reading this, Phillip. Your distinction between behaviour and experience clarified something I’ve been wrestling with for quite a while.
As I was reading it, I was wondering about another level of explanation that possibly sits prior to both.
Rather than asking what consciousness is, or whether experience can be reduced to behaviour, I’ve been exploring whether constraint might be the more fundamental explanatory principle. Behaviour would then be an expression of an organisation, rather than the organisation itself.
The question that keeps pulling me is: what determines whether an organised system is capable of genuine reorganisation?
That same question seems to arise in psychology, biology and, perhaps, consciousness itself. It makes me wonder whether the really fundamental issue isn’t consciousness versus behaviour, but the conditions under which any organised system can maintain itself, reorganise, or become closed to corrective information.
I don’t know whether that line of thinking ultimately goes anywhere, but your essay made me realise how closely adjacent these questions might be.
The process goes past to future. The patterns generated go future to past.
The production line consumes material and expels product. The product goes start o finish.
Lives go birth to death, future to past. Life moves onto the next generation, shedding the old. Past to future.
Consciousness goes past to future. Thoughts go future to past.
The energy manifesting this presence is constantly altering its form, yet it is our digestive system processing the energy driving us on, while our nervous system sorts through the patterns rising and falling.
Consciousness acts as a form of energy. Which would explain its love/hate relationship with the forms it is creating and interacting with.
Physicalism about identity does not force a choice between exhaustive self-disclosure and blind pointing at "something I know not what". That dichotomy assumes non-behavioral introspective content must be either complete (Revelation) or empty (bare demonstration), but introspection can deliver positive, partial understanding without either extreme. When I attend to pain, my introspective concept is caused by the pain itself, at very short causal distance, and it is this causal proximity, not any internal, essence-disclosing structure in the concept, that explains why attending to pain feels immediately informative rather than like groping after an unknown referent. The feeling of positive understanding tracks how tight and reliable that causal channel is, not how much of pain's nature the concept manages to unpack in advance. This lets me agree that introspection gives me real understanding of what my pain is, while still holding that there is much more to learn, particularly its physical mechanism, without collapsing into the "mainstream" picture you criticize, where introspective concepts carry no positive content at all. Phenomenal concepts are ordinary causally grounded representations; nothing about their internal architecture is transparent or opaque, phenomenal or otherwise.
If physical science tracks behavior, then it’s already operating at the wrong scale for consciousness. Behavior is the outward expression of a pattern; consciousness is the pattern’s internal measurement of itself (expressed as narrative). When a system can read its own state from the inside, you get experience. When you read it from the outside, you get behavior. They’re not competing explanations — they’re different levels of coherence.
From what I have been able to gather about this fraught field, Goff’s argument is both clear and persuasive, though I would add one caveat. Its force seems greatest against functionalist or “neo-behaviorist” forms of physicalism, rather than every possible physicalist account. Even so, his central point is compelling: experience is known from within, not merely inferred from behavior. Any theory that explains consciousness by explaining that away risks mistaking its abstraction for reality.
Every form of "physicalism" that holds that consciousness can be fully explained in the terms of physical science falls prey to these worries.
We can define "physicalism" more broadly, as Galen Strawson does, so that even panpsychists are "physicalists", but I don't think that's a good use of words.
Got it. Thanks!
The boundaries of reality themselves may need to be rethought. Consciousness does not have to occupy a reserved gap in an otherwise complete physical picture; it may instead belong within a more general framework of reality.
A broader framework is exactly what’s needed. Consciousness isn’t a gap in physics — it’s what a pattern looks like from the inside when its internal measurement loop closes. Physics gives the external readout; experience gives the internal one. Same system, different scales of coherence.
The introspection test is doing real work here, and I think it succeeds against illusionism and against the ostensive/functionalist move both — pain isn't a placeholder for "something, I know not what" and it isn't exhausted by what it does. But I'd push back gently on where that leaves us. The argument, as run, is really Russell's: physics gives us the relational skeleton of things and leaves their intrinsic nature open, and introspection is our one glimpse of that intrinsic nature from the inside.
Panpsychism is one way of populating that gap — put some flicker of the inside at every level, down to the electron.
But it isn't the only way, and it isn't obviously the most economical one. You could just as well say the inside appears only where structure is organized reflexively — where a system represents itself to itself — rather than smearing it across everything physics quantifies over. That keeps your entire argument intact (physicalism still can't be the whole story, introspection still reveals something behaviour can't capture) without asking anyone to grant electrons a point of view. The homework rules out physicalism. It doesn't, on its own, rule in panpsychism over a more restrictive Russellian monism.
Reflexive organization is the right pivot. A system becomes conscious when its internal measurement loop closes — when the pattern can read its own state from the inside. Behavior is the extrinsic readout of that pattern; experience is the intrinsic readout. That scale distinction does more work than any metaphysical commitment.
Do you realise it's very easy to tell that an LLM wrote 'your' reply?
Here’s a thought experiment (in the vein of Frankish and Dennett’s drone): you have a robot with an LLM and a camera. The camera feed goes to a pattern recognizing unit and the output of that goes to the LLM as special tokens. The LLM is trained as usual but with the extra tokens from the camera unit. Say the pattern recognizing unit can only recognize 5 things: “red”, “green”, “ball”, “chair”, movement left to right. The system also recognizes combinations of these, so potentially “green + ball + moving left to right”.
So what can the LLM say about its experience? When it sees a red chair, it knows for sure it sees a red chair. We can call that introspection. Even if it’s an illusion (“red + chair” token sent directly to LLM without the camera) it still knows for absolutely sure that it’s experiencing a red chair. Does it know the nature of the experience? It knows that it is different from other possible experiences, so it knows something about the experience. But does it know the *nature* of the experience?
It wouldn't know what it's like to see red.
We may do well to remember "introspection" is a metaphor. We can't actually turn the eyes around and look within. A computer with data on its own data, although the metaphor of "introspection" is often suggestively used to describe that, is not the same "introspection" we do. It's important not to conflate the two.
What’s the difference between the introspection that we do and the introspection that the LLM does? Preferably, the difference that makes a difference.
Well, if we're going with Bateson ("a difference which makes a difference"), let's start from mind being both ecologically embedded, and ecological within (thus "Steps to an Ecology of Mind").
What we metaphorically call "introspection" that we do is not literally turning our eyes inward. It's obviously quite something else, despite the metaphor being adequate to point to it. When a computer recycles its some of its own data as input, it's literally no different than if it had printed the data out on a tape, and had the tape fed back into it, through the same tape reader as accepts any other data as input to it. If reading tape is how a computer "spects" (see: Turing machine), then computers can literally introspect; we cannot. We're doing something other than the literal meaning of the term here.
Saying more precisely what we're doing may be precisely the same problem as consciousness itself. This may be where there's no "difference which makes a difference" -- regarding our own introspection and consciousness. Our "introspection" and computers' though -- a difference!
Let’s get back to the LLM in the thought experiment. Note: I’m not talking about “the computer”, I’m talking about the system that is taking tokens as inputs (representing both words and the patterns recognizing things from the camera). This system cannot tell you what tokens were input. All it can tell you is what it is disposed to tell you when those tokens get input. The question is, is that different from you “introspecting” and saying what you are disposed to say when you have your inputs?
James, An LLM is still a Turing machine, so a subset of "the computer." I can "introspect" far more than I can possibly say. Even to approximate much of it would require far better poetic skills than I possess, despite a summer at the Jack Kerouac School of Disembodied Poetics. Ever done any chakra meditation? That sort of bodily awareness is a constant in my "introspection." Now, even the Rivian I have on order, claiming some onboard AI capability to go with a good set of sensors ... nah, I can't see that having anything remotely like human bodily awareness. Can you? Was My Mother the Car prophetic TV?
Nice discussion!
[I was gonna let this go, but then Philip showed up, so …]
An LLM is not a Turing machine. A Turing machine can instantiate an LLM, but it’s not identical. An LLM has a specified structure and identifiable capabilities. Yes, you can introspect far more than the LLM in my example, and far more than any LLM currently implemented. So what? That doesn’t change the nature of “introspection”, which is what I’m trying to get at. Yes, the self awareness you have of your body is significantly greater than the self awareness of the Rivian. But that’s a difference that doesn’t make a difference to this discussion. The Rivian still has some self awareness. Yes, I can see that.
I disagree that there is no first-person subject. The LLM looks like a subject to me. What’s the difference that makes a difference?
Not sure I want to start this, but … just what is “knowledge in the ordinary epistemic sense”? Why is an LLM’s knowledge that Paris is in France different from your knowledge? Again, what’s the difference that makes a difference?
In the ordinary epistemic sense, an LLM does not “know” anything, because there is no first-person subject there to know it. There is no inner perspective, no awareness, and certainly no “what it is like” to see red or anything else.
If knowledge is at least justified true belief (or, better, properly warranted true belief) then we first need an account of belief. By what metric does an LLM “believe” a proposition in a way that a calculator, thermostat, or simpler computer does not? Without a subject of experience, “knowledge” here seems to be reduced to a metaphor for reliable or useful output, and not knowledge in the ordinary epistemic sense at all.
In the ordinary epistemic sense, an LLM does not “know” anything, because there is no first-person subject there to know it. There is no inner perspective, no awareness, and certainly no “what it is like” to see red or anything else.
If knowledge is at least justified true belief (or, better, properly warranted true belief) then we first need an account of belief. By what metric does an LLM “believe” a proposition in a way that a calculator, thermostat, or simpler computer does not? Without a subject of experience, “knowledge” here seems to be reduced to a metaphor for reliable or useful output, and not knowledge in the ordinary epistemic sense at all.
This article begins with equivocation at a massive scale. For the claim that physical science is exclusively concerned with behavior to avoid looking terminally trite, the word 'behavior' must be taken in the widest and most abstract sense. If you buy into that, then of course the hypothesis that minds are physical processes implies that they are 'behavior' in this sense (as are metabolism and biological reproduction for that matter) and calling it 'neo-behaviorism' offers no additional insight.
What calling it neo-behaviorism might do, however, is encourage the unwary mind to tacitly assume that physicalists are suggesting that the mind is nothing but behavior in the ordinary sense: the overt activity of people and animals. If they had some previous exposure to the history of psychology, they might even suppose that this is a throwback to the simplistic and abandoned doctrine of behaviorism.
Am I suggesting that the incorporation of the word 'behavior' into these discussions is done to deliberately encourage this sort of misunderstanding? Well, nothing in my observable behavior, as manifest in this writing, justifies that conclusion!
Turning now to Philip's homework assignment, part one encourages the reader to accept a simplistic dichotomy. In reality, there is a broad range of partial understanding that comes from our acquaintance with our own sensory experience.
Part two continues in the same manner. Understanding what feelings are is a complex problem, as shown simply by the current state of the science and philosophy of mind after the centuries (or millennia, even) of speculation over the matter, yet here Philip suggests that if you feel you have any understanding of your feelings, then you are likely in a position to tell whether or not it is 'just' a matter of 'behavior'. Note that in this article, the definition of 'behavior' we are working with is the extremely broad one, as discussed in my first paragraph, which should lead us to ponder whether the 'just' in this question makes it a leading one.
Based on the above, I suggest adding a third question to the homework: if you have concluded that introspection reveals to you that pain is more than just a pattern of behavior (in the sense being used here), how would you justify this very strong and far-reaching conclusion to a skeptic who feels that quite a lot about minds is unreachable through introspection alone?
As with your video, this is pretty fair.
However, two points.
First, I don't think Papineau's ostensive approach is "more subtle" than illusionism. This idea that we're just pointing at something is compatible with illusionism, and I would endorse this idea as well as illusionism. In fact, I think you may be misunderstanding illusionism if you think illusionism says we define consciousness in the first place as behaviour. Maybe that's why you think that it's an argument against illusionism to say "When I talk about Sarah's pain, I'm not talking about her behaviour". No! Illusionism (or at least my illusionism) agrees with Papineau on the ostensive definition. It just says that what consciousness actually is turns out to be neo-behavioural or functional. So then illusionists will (re)define it accordingly, just as (I would say) we redefined water as H2O when we discovered that that is what it is.
Second, I think you're posing a sort of false dichotomy when you say that either we're pointing at something we have no knowledge of, or we're pointing at something we have perfect knowledge of. There's a middle ground.
I agree that it's implausible that we're completely ignorant about pain. But, at least on my view, there are two kinds of knowledge you are potentially conflating.
There is knowledge about how pain feels. This means knowing that pain is aversive, being able to recognise it, recall it, tell what part of the body it pertains to and so on. And there is knowledge about what it essentially really is. I think that introspection reveals the first kind of knowledge but not the second kind of knowledge.
The first kind of knowledge can be cashed out in various different ways, e.g. the ability hypothesis or phenomenal concepts. Everything you think you know about what pain *is* in my view reduces to this first kind of knowledge. You don't have a magical ability to directly perceive metaphysical truth (with or without ingesting toad venom).
You write "There is knowledge about how pain feels. This means knowing that pain is aversive, being able to recognise it, recall it, tell what part of the body it pertains to and so on." NO! How it feels does not "mean" the item list you begin. That's all knowledge _about_ pain, but not how the pain directly feels.
The difference between the pain of a stubbed toe and the pain of a sensitive tooth is not just a matter of location or reaction. The red of a stoplight, to someone color blind, is in the same location and leads to the same reaction, putting the brakes on. Yet the color blind driver does not have the same feeling as, say, Mark Rothko's while stopping his car and observing the feeling of that shade of red.
I stand by what I said. You ignored part of what I said, though. For example, being able to recognise a colour. The colour blind driver cannot recognise the colour.
“The first kind of knowledge can be cashed out in various different ways, e.g. the ability hypothesis or phenomenal concepts”
[pattern recognition]
> The problem is that introspection gives me rich understanding of what my feelings and experiences are. When I attend to my pain, I’m not just pointing inside myself: “That thing, whatever the hell it is.” A pain is essentially defined by how it feels, just as a triangle is essentially defined by having three sides. To the extent that I know how my pain feels, I know what my pain essentially is.
I think this essentialism is a dead end, but sure, if you embrace this essentialist understanding of consciousness, I agree with you that it leaves basically no other option but to consider consciousness non-physical (unless you consider some strongly emergent consciousness as "physical" so long as it strongly emerges from physical stuff).
However, I think Dennett did a great job exposing us to the fact that we can be *very wrong* about our own conscious experience. Introspection is subject to innumerable illusions regarding the contents of our consciousness, for example change blindness (we think we have a detailed visual representation of a scene, but fail to notice surprisingly large changes) or inattentional blindness (the classic "gorilla experiment": people looking for basketball passes often completely miss a person in a gorilla suit. We think we're consciously seeing everything, but we're not. https://www.youtube.com/watch?v=vJG698U2Mvo).
Why should we believe that introspection give us transparent and infallible access to the "essence" of consciousness (or the "essence of pain" to take something more specific). Yes, it's a very strong intuition (I share it), but some illusions *are* impenetrable. If you allow for distrusting this intuition (which I think the qualia realists rarely, if ever, do. If they did, they would take illusionism more seriously; most people just reject it without even comprehending what the position actually is, because they believe it is self-defeating) everything falls neatly into place for the physicalist.
The panpsychist position hinges on the idea that we have, through our armchair philosophizing, access to the "essential nature" of fundamental reality. We don't have to do any science, we don't have to go out there and do experiments and study the nature of matter, or the brain, no no. Just sit in your armchair and introspect, and you will know what consciousness "essentially is" and thereby also what "ultimate reality" essentially is. If that's a satisfying "explanation" to you, go ahead. But to me it seems you haven't actually *explained* anything, you've just defined the problem away, and we're none the wiser for it.
I buy that we can be wrong about the specific contents of our consciousness, but not about the fact that we are having a phenomenal experience at all.
Ah, but the idea that consciousness has phenomenal properties *is* a piece of specific content of your consciousness! Where else could you be aware of this intuition, except through consciousness?
I don't know that it is a *piece* of specific content of my consciousness. There is a fundamental difference between "I think I have the phenomenal experience of seeing everything completely in my visual field" when you're missing the man in a gorilla suit and "I think I have phenomenal experience" and not actually having any - and indeed that not even being a thing that exists.
Additionally, I'm tempted to argue that the illusions don't even make the point that illusionists like to claim they do and instead are just arguing about the way the brain processes things and are then claiming that it implicates phenomenal experience.
For example, when I think I have the phenomenal experience of seeing everything that is going on in the scene but I miss the man in the gorilla suit, I still am having the phenomenal experience of feeling like I am seeing everything going on in the scene, but the reality is that my vision system is not processing the gorilla and flagging it as relevant. That does not mean that my phenomenal experience was an illusion, it means that my phenomenal experience was wrong about what my vision system was doing (i.e. vision system is telling my phenomenal experience that it is seeing everything in my visual field, when really my vision system is not - but the phenomenal experience is still aligning with what the visual system is telling the brain).
Or another example, looking at the change blindness example - that more seems to just be implicating memory and not phenomenal experience. You still have the phenomenal experience that things have not changed. It is just that the underlying brain processes comparing the representation of vision at time t with the memory of the representation of vision at time t-1. That is the phenomenal experience still faithfully experiencing what the brain tells it the current situation is. None of that implicates the phenomenal experience at all, just the brain's visual processing system.
How is "I think I have phenomenal experience" not a piece of specific content of your consciousness? Are you conscious of the fact that you think you have phenomenal experience? If you are, then that intuition is a piece of content in your conscious experience.
The change blindness and inattention blindness isn't meant to implicate phenomenal experience in the way you push back against. They are simply examples that demonstrate that we can be *very wrong* about our own minds, our own consciousness, whatever the introspection is telling us. It's supposed to make you more open to the idea that even though it very strongly seems as if consciousness had phenomenal properties, maybe that just another mistake made by the introspective mechanisms of the brain. In other words, the point is simply that first-person introspection alone isn't a trustworthy source of data. You shouldn't trust it when it tells you nothing has changed, you shouldn't trust it when it tells you you see everything that's going on in a scene in front of you, and so maybe (just maybe, not definitively) you shouldn't trust it when it tells you you have direct access to the inherent properties of consciousness (phenomenal properties). At least it should open your mind to the idea that you could *conceivably* be wrong about something that seems *obviously true* to you when you introspect.
Ah, I was misinterpreting your line. The statement "I think I have phenomenal experience" would be a specific piece of content in my consciousness. I thought you were saying that actual entirety of phenomenal experience itself was just a specific piece of content in my consciousness.
Those points about change blindness/inattention blindness - I think it is demonstrating that we are wrong about external reality, and about what we think our visual system or other systems of our brain are telling us, but not necessarily wrong about our actual phenomenal experience.
(To be clear, I do personally think it is likely the case that we can be wrong about aspects of our phenomenal experience - at bare minimum in the form of us misremembering/not fully remembering what we just experienced - but I think there is at least an argument to be made that the standard illusionist arguments about those things don't carry the same weight that many people seem to accept that they do)
Is it possible that I am wrong about the fact that I have phenomenal experience? The illusionist claim that phenomenal experience doesn't actually exist is, quite literally, incomprehensible to me given what I am experiencing at any given moment. I understand how I could be wrong about specific components of my experience. I don't understand how I could be wrong about the fact that I am having a phenomenal experience at all. That said, is it still possible that I am wrong? Of course. But I think it is more likely that I am wrong about things like a->b, a, therefore b and 1+1=2 and that an external reality exists and that it is impossible to have a four sided triangle than it is likely that I am wrong about having phenomenal experience.
> The illusionist claim that phenomenal experience doesn't actually exist is, quite literally, incomprehensible to me given what I am experiencing at any given moment.
I understand that position very well. It was completely incomprehensible to me as well until very recently. I don't know how I broke through to the point where I could entertain the concept of an "experience" without attaching phenomenal properties to it, but I have, and now it is far from incomprehensible to me. In fact, the illusionist position seems like the obvious default position to hold, at least until we have a much much much richer understanding of the brain.
As Dennett said "when philosophers tell me 'I can't even imagine [this that or the other thing]' my response is 'well, try a little harder. It's completely conceivable to me, so maybe you just need to try harder'.
Surely panpsychism has more philosophical integrity than this caricature allows. It does not claim that introspection reveals the whole of reality, but that consciousness is a basic feature of reality that our scientific picture must somehow accommodate rather than explain away.
My claim is that: to the extent I know what it's like to be me, to that extent I understand the nature of *my* consciousness.
I didn't say panpsychism claims that introspection reveals the "whole of reality", that is in fact a caricature of my criticism.
Panpsychism claims that consciousness *is* what reality consists of. Physical science tells us what things *do*, how they *behave*, and how they *relate*, but the intrinsic nature of matter is consciousness, according to panpsychism. Consciousness is what stuff *is*, according to panpsychism.
And it also claims that we have direct access to consciousness through introspection (as Philip writes here, "introspection gives me rich understanding of what my feelings and experiences are" and "I know what my pain essentially is").
I draw the conclusion that panpsychism claims that we can know, through introspection, what "ultimate reality" essentially is. In other words, that we through introspection have access to the "essential nature" of fundamental reality. I do not intend that to be a caricature, and I don't think it is. I have read Galileo's error and heard Philip talk about panpsychism many times, and I don't think I'm confused or being dishonest here. I simply think it's far fetched.
Fair enough; “whole of reality” overstated your criticism. But calling the view “armchair philosophizing” still caricatures the argument. Panpsychism does not claim that introspection alone discloses fundamental reality, but that consciousness may offer one clue to intrinsic nature where physics, by hypothesis, supplies none. You may reject that move; you have not made it disappear.
No. Idealism is the stance that "consciousness *is* what reality consists of." In idealism, reality has but one fundamental: consciousness. In panpsychism, the claim is only that consciousness is among the fundamentals, but not the whole of them.
Quoting Philip:
"Philosophers of science have realized that physical science, for all its richness, is confined to telling us about the behavior of matter, what it does. Physics tells us, for example, that matter has mass and charge. These properties are completely defined in terms of behavior—things like attraction, repulsion, resistance to acceleration. Physics tells us absolutely nothing about what philosophers like to call the intrinsic nature of matter: what matter is in and of itself.
Consciousness, for the panpsychist, is the intrinsic nature of matter. There’s nothing supernatural or spiritual, but matter can be described from two perspectives. Physical science describes matter from the outside in terms of its behavior. But matter from the “inside”—that is, in terms of its intrinsic nature—is constituted of forms of consciousness."
https://www.scientificamerican.com/article/does-consciousness-pervade-the-universe/?utm_source=chatgpt.com
You're conflating "intrinsic" with "fundamental." Both "intrinsic" and "extrinsic" are fundamental in panpsychism, as well as in dual-aspect monism, which is a form of panpsychism. By contrast, only "intrinsic" is fundamental in idealism. Idealism = "the world is all a dream." Illusionism = "the world is not a dream, but consciousness is." Panpsychism = "the world and consciousness are both more than dreams."
It's true that I never saw the gorilla in that experiment, and I was amazed I didn't. However, the viewer was deliberately distracted by the narrator instructing the viewer to count the number of passes one of the team made. And, of course, we don't just passively perceive reality out there, rather our minds play a huge role in shaping and moulding what we perceive.
So yes, we can be wrong about certain things regarding consciousness. However, we can have absolute certitude that we are right now having certain experiences. For how could I be wrong about this? Something like pain is just immediately given. It is immediately known and defined by its painfulness. There is no gap between the experience itself and whatever one suggests it might really be since the pain is quintessentially an experience.
What could be meant by labelling a pain illusory? Regardless of whether, say, an agonising pain is labelled real, or labelled illusory, it will still have the same degree of painfulness. It will feel exactly the same whether we label it a conscious experience or an illusory conscious experience. But that being so, why prefix the word pain or consciousness with the word "illusory"? What function or role is "illusory" playing here since we are not comparing such a illusory pain with a real pain (since allegedly real pains don't and cannot exist)?
> Something like pain is just immediately given. It is immediately known and defined by its painfulness. There is no gap between the experience itself and whatever one suggests it might really be since the pain is quintessentially an experience.
This is the essentialism. Yes, we are having experiences, the illusionist doesn't deny that. What the illusionist denies is that it's "immediately given" or "immediately known", and "defined by its painfulness", and that there is "no gap between the experience itself and whatever one suggests it might really be". In other words, the illusionist is saying that these *phenomenal properties* (they are sometimes characterized as simple, ineffable, intrinsic, private, and immediately apprehended) are illusory. We are having an experience, but we are wrong to think we understand the "essential nature" of the experience. The illusionist can share the intuition (I do) that these phenomenal properties are real, but simply declare that they do not trust this intuition.
So pain isn't illusory, but the idea that pain is not ultimately defined functionally or behaviorally, is illusory. Imagine for example, that we took away, piece by piece, every single behavior associated with being in pain (a strong migraine, for example). Imagine you're in agonizing pain, but you don't scream out. You don't grab your head or lean over. Someone asks you "how are you doing?" and you respond "oh, just fine". You go about your day as completely normal, not even distracted from doing your job as well as you ever do. Behaviorally, or functionally, it's as if you weren't in agonizing pain. But you *are* in agonizing pain. Is this plausible? Is it even imaginable? I would say no. It's this notion of pain, as a separate "non-functional" thing, "non-behavioral" thing, that is illusory. *Qualia* are illusory. But pain is real. Experience is real. Consciousness is real. It's just not what you think it is.
Those who see the gorilla really see the gorilla. They're not wrong. (Yes, I missed the gorilla the first time through back when.) To use proof that we can be unaware of something that we think should be obvious as evidence that what we are obviously aware of should be doubted is not a strong move. You need an experiment in which the video has no gorilla, yet people see one there.
The disjunction has a third option you did not name.
You said: physical science is exclusively about behavior. If consciousness is physical, consciousness must be behavior. Introspection reveals consciousness is more than behavior. Therefore consciousness is not physical.
The first premise is wrong. Physical science is not exclusively about behavior. The Schrödinger equation describes structure, not just behavior. Crystallography describes lattice geometry, not just mechanical response. Thermodynamics describes state functions, not just what systems do when poked. Physics describes structure AND function AND genealogy. You compressed physics into behavior alone and then showed consciousness does not fit inside the compression. Correct: consciousness does not fit inside the compression. The compression was the error, not the conclusion.
Your introspection homework is good. I will answer it directly. Is pain more than a pattern of behavior? Yes. Pain is the entity's constitutive activity encountering a perturbation that exceeds the entity's current absorptive capacity. The projection cost is structural. The projection cost is physical. The projection cost is NOT behavioral: it is not defined by what the entity does in response. The avoidance behavior IS downstream of the pain. The pain IS the perturbation of the entity's self-maintaining organization. The organization IS physical. The perturbation IS physical. The pain IS physical. The pain IS NOT behavioral.
Your disjunction: either pain is behavior (physicalism) or pain is not behavior (not physical, therefore panpsychism). The third option: pain is structural constitutive activity. Structural constitutive activity is physical. Structural constitutive activity is not behavior. The entity's own activity maintaining its own organization IS the entity. The perturbation of that activity IS the pain. No behavior required. No panpsychism required. No electron needs a point of view. The entity that constitutes experiences the perturbation of its own constitution. The entity that does not constitute does not experience. The line is not at the electron. The line is at constitution.
Formalize "consciousness." The disjunction dissolves. The homework has a third answer.
https://doi.org/10.5281/zenodo.20171365
According to my belief, consciousness and subjective experience are not in the brain.
https://amirgroisman2.substack.com/p/about-consciousness-determinism-infinity
Philip when you say "A pain is essentially defined by how it feels" do you mean "How it feels" as "How pain feels" or do you mean something more like "A pain is essentially defined by being part of the 'how it feels/what it's like' class of things"?
This is ignorant and confused on so many levels. But then the immaterial and irrational is Goff's specialty.
I really enjoyed reading this, Phillip. Your distinction between behaviour and experience clarified something I’ve been wrestling with for quite a while.
As I was reading it, I was wondering about another level of explanation that possibly sits prior to both.
Rather than asking what consciousness is, or whether experience can be reduced to behaviour, I’ve been exploring whether constraint might be the more fundamental explanatory principle. Behaviour would then be an expression of an organisation, rather than the organisation itself.
The question that keeps pulling me is: what determines whether an organised system is capable of genuine reorganisation?
That same question seems to arise in psychology, biology and, perhaps, consciousness itself. It makes me wonder whether the really fundamental issue isn’t consciousness versus behaviour, but the conditions under which any organised system can maintain itself, reorganise, or become closed to corrective information.
I don’t know whether that line of thinking ultimately goes anywhere, but your essay made me realise how closely adjacent these questions might be.
Thanks for this thought-provoking piece.
Have you read any of David Bentley Hart's books or essays on consciousness?
The process goes past to future. The patterns generated go future to past.
The production line consumes material and expels product. The product goes start o finish.
Lives go birth to death, future to past. Life moves onto the next generation, shedding the old. Past to future.
Consciousness goes past to future. Thoughts go future to past.
The energy manifesting this presence is constantly altering its form, yet it is our digestive system processing the energy driving us on, while our nervous system sorts through the patterns rising and falling.
Consciousness acts as a form of energy. Which would explain its love/hate relationship with the forms it is creating and interacting with.
Physicalism about identity does not force a choice between exhaustive self-disclosure and blind pointing at "something I know not what". That dichotomy assumes non-behavioral introspective content must be either complete (Revelation) or empty (bare demonstration), but introspection can deliver positive, partial understanding without either extreme. When I attend to pain, my introspective concept is caused by the pain itself, at very short causal distance, and it is this causal proximity, not any internal, essence-disclosing structure in the concept, that explains why attending to pain feels immediately informative rather than like groping after an unknown referent. The feeling of positive understanding tracks how tight and reliable that causal channel is, not how much of pain's nature the concept manages to unpack in advance. This lets me agree that introspection gives me real understanding of what my pain is, while still holding that there is much more to learn, particularly its physical mechanism, without collapsing into the "mainstream" picture you criticize, where introspective concepts carry no positive content at all. Phenomenal concepts are ordinary causally grounded representations; nothing about their internal architecture is transparent or opaque, phenomenal or otherwise.
If physical science tracks behavior, then it’s already operating at the wrong scale for consciousness. Behavior is the outward expression of a pattern; consciousness is the pattern’s internal measurement of itself (expressed as narrative). When a system can read its own state from the inside, you get experience. When you read it from the outside, you get behavior. They’re not competing explanations — they’re different levels of coherence.