Qualia seem to be real and widespread among humans, but I sure as hell don't have them. I could be mistaken about anything, but not about this. I know it... directly.
I'm sick if people just presuming I'm conscious.
You say I'm not a zombie, per definition, since I claim to be one? Well zombies get psychiatric illnesses too. Perhaps my condition is something like Cotard syndrome. Only it happens to allow me to make true statements about my lack of consciousness. The fact that I would make the same statements claiming to be a zombie if I were an ordinary conscious human with Cotard syndrome doesn't change that one bit.
Yeah I remember our conversation in which you claimed not to be conscious despite all manner of behavioral evidence you were. Absent a settled theory of consciousness, I can't prove you're sentient but, as we agreed before parting, it would be a big mistake to treat you as if you weren't.
I have a hard time believing I ever denied without qualification being conscious. It’s frustrating when people who know that the meaning of term “consciousness” is under dispute pretend that it isn’t in order to portray their opponents as idiotic or psychopathic. At least you didn’t do so in the New York Review of Books!
Oops, my comment and link to my "conversation with a zombie" was meant to be a light-hearted response to Mark Slight who was (in jest, we assume) claiming to be a zombie. I guess I have to figure out how the comment/reply structure works in Substack, my apologies. In any case, no, you've never denied being conscious tout court that I know of. I know Galen Strawson in the NYRB accused Dennett of committing the silliest mistake ever in philosophy by denying what Strawson sees as the obvious fact of phenomenal consciousness.
Ah, no apology needed. I replied without checking the larger context, and now look like someone who can’t process humor. Allow me to apologize instead! My bad.
Your claiming that I might be conscious is the silliest clame ever made. It's just obvious to me that I'm not. How would you like if I claim you may not be conscious?
Forget about theories about consciousness. I know this, directly. Introspection, right?
Respect me like anyone who's conscious, but that entails not implying that I'm conscious. Thank you.
Asking for a neutral definition of *phenomenal* consciousness is asking for trouble given that "phenomenal" is philosophically freighted, no? Rather, one might ask what's an innocent, example-based definition of what being conscious consists in, but leave out such terms as phenomenal, qualitative, etc. One pretty innocent folk-psychological definition is that sensory experiences like seeing red have a distinctive character that distinguishes them from their sensory cousins like blue. Illusionists and qualia quietists should agree that such characters exist, but will dispute there's anything necessarily phenomenal or qualitative about them. Experiential characters, they might say, can be cashed out functionally or behaviorally or physically. I'm not sure if appeal to experiential character is non-circular, but if someone says no such characters exist, then they have to say what does present itself in perceptual experience.
It's cool GPT identified “recursive access modeling with internal access by downstream interpretive modules” as the common feature of its vision analog. Let’s say an AI representational architecture is designed so that such access can’t decompose what’s being accessed into sub-components. This would parallel one aspect of a basic experiential character like red: it’s monadic, non-decomposable into other color components, cognitively impenetrable, so its color character is only reportable as being “like this”.
Well, just giving you guys the benefit of the doubt. I remain a phenomenal realist since I'd say phenomenality *just is* to have in one's experience basic sensory characters that can't be decomposed, so aren't effable, only nameable. I expect as a quietist (in your illusionist moments) you might say that experiential characters *seem* to exist, but remain agnostic about their actual existence since it isn't obvious how they comport with physicalism. I agree it isn't obvious, but I'm not wedded to physicalism and keep my options open as the investigation proceeds.
Yeah but I was talking what's less doubtful about *experience* such that you can claim it really doesn't seem to me that red has a distinctive character. What backs up that claim in terms of what's evident about experience? Or is nothing evident on your view?
it’s evident that people make judgments about the world, and often are reliably caused to do this by the parts of the world those judgements are about. Any supposed intermediaries between the judgments and everything else either have their third-personal physical bonafides or else are a warped fantasy of magical pseudojudgements like qualia, sensations, experiences, sensory intuitions, sense data, figments, mental paint, qualities, the mythical “given,”raw feels, phenomenal appearances, yadda yadda, blah blah blah, no thanks.
Sounds like you're a skeptic about experience, period, so no wonder you're quiet about qualia, except to say they are magical pseudojudgments. Dreams on your view are judgments about the world - wrongly applied concepts - with no experiential intermediary. And concepts have third-person physical bonafides - nothing abstract about them. Fascinating, captain!
Yeah, I might grant that you believe that (thus making me lean illusionist), but more likely I’d raise doubts about whether things actually seem that way to you or to anyone else (meta-illusionism&qualia quietism)—like how it neither seems that the sun goes around the Earth nor the other way around.
Ok, I figured you'd come back with that. Then what's less subject to doubt than my assertion that characters exist in sensory experience such as the character of how red looks to me? Well, it's the claim that they only *seem* to exist. But you say there's something even less doubtful than that by which the (illusionist) seeming claim is cast into doubt. What is that on your view?
I don't think Eric's search for an innocent definition can succeed in a framework where zombies are considered genuinely possible. If we had it on God's authority that there was a zombie world next door, then, as a direct consequence of the fact that zombies have been defined as our cognitive isomorphs, we could not draft a definition that was rational, defensible, included us in the list of conscious beings, and excluded them.
They will remain in definitional lockstep with us at every turn, and their defence of their definition will always match ours, leaving us totally unable to say anything sensible about where their definition falls down. This quandary is intrinsic to the idea of zombies, and to any definition of consciousness that leans on the idea of a human-zombie difference as the core feature of phenomenal consciousness.
I think Eric's search can be a *starting* point for a successful definition, provided that it takes care to put aside the possibility of zombies. By then, of course, it is no longer theoretically neutral. But there is no consistent framework available that keeps zombies.
I think standard uses of "phenomenal consciousness" suffer from fatal internal contradictions, including the ones you highlight, and the term has been so abused it is beyond rescue. There are, however, sub-concepts within "phenomenal consciousness" that can be named and considered separately. The component of the concept of "phenomenal consciousness" that separates us from zombies needs to be named and shamed, and then put aside.
As to your titular question, I would say that I would be embarrassed if me and my zombie twin disagree. I must be ready for my zombie twin to agree with my personal definition, and if I want my opinions to be defensible, I would need to keep revising my opinions until I am not embarrassed on its behalf when it discusses consciousness. I will be left with definitions that deny a human-zombie difference, and my zombie twin would also deny such a difference. It could not rationally believe in its own zombiehood or non-zombiehood; that dimension of separation is spurious.
This definitional issue can be *raised* in the presence of zombie agnosticism, but it cannot be resolved in the presence of zombie agnosticism.
if god is around and assuring us that there are zombies a world away, maybe he’ll also assure us that dualism is true. It would be irrational at that point to complain about the implied definition of consciousness.
You do understand the practice of considering hypotheticals, though, even ones we disagree with, because your own post is based on that practice.
It almost seems as though it is okay to raise the spectre of zombies to make some points, but once they enter the chat in someone else's line of thinking, it's irrational to consider the consequences of their existence. That would be somewhat unfair.
The point is that even if God gave us an explicit endorsement of zombies, and hence dualism, it still would not help us define the element that distinguishes us zombies. such a difference is fundamentally beyond our grasp even if we try to accommodate it by assuming that it exists. That makes the posited difference pointless to think about. It also means that it is *not* what we're actually puzzled about.
And, of course, I don't think it makes sense to posit that a human-zombie difference does exist, for that very reason, and I would politely suggest to God that whatever difference he is talking about it when he discusses our zombie neighbours, it cannot be a key part of any consciousness that we should care about.
the element would be something nonphysical. we would then know that no physical-compatible definition of phenomenal consciousness can be correct. I may still be missing something, but your conjectures that I don’t know how hypotheticals work or that I lack some capacity for fair treatment seem like jumping the gun. I have tried to indicate why I’m not understanding the point you’re making, and I continue to fail.
I think the issue here us that you are linking God's endorsement of a particular subspecies of dualism with the end of any need to explain the many versions of consciousness that are not covered by the idea of a human-zombie difference. Knowing zombies were in the world next door would tell us we had some extra of no particular importance.
We would still have the original true target of our puzzlement waiting to be explained.
The idea of the world next door is merely raised to distinguish two very different meanings of phenomenal consciousness. The one separating us from notional zombies cannot be the important thing that stands in need of definition, in part because the human-zombie difference evades definition completely, even if we go out on a limb and grant its existence.
This is your own point being expressed from a different view.
My question for any particular definition of "phenomenal consciousness" is, does it still lead to an explanatory gap / hard problem? I generally don't see it with the most innocent pre-theoretical versions. And explaining why there's still a hard problem seems to inevitably bring back in things like ineffability, irreducibility, privacy, direct acquaintance, etc, all of which I think amounts to a theoretical commitment to fundamental consciousness. But that theoretical commitment seems optional for the innocent versions of phenomenal consciousness.
depends on what “lead to” means. If it means “entail,” that would be a poor definition, and would make arguments for gaps and hardness plainly question-begging. If it doesn’t mean “entail”, it’s hard to see how anyone could credibly claim to have a priori arguments against physicalism.
I meant "entail". And I agree on a priori arguments against physicalism. I actually don't think there is a hard problem without assuming non-physicalism. If so, trying to solve it with non-physicalism seems like another case of circularity.
Terrific post! Thanks for the thoughtful engagement. I'm inclined to respond to your dilemma by requiring first-personal obviousness and defining "us" not phenomenologically (which I agree risks circularity) but in biological and social terms -- that is, ordinary adult biological human beings in our 21st century mainstream Anglophone context.
This definition by example does require that this "we" all have introspective access to their experiences, but I think that's okay. I don't think it rules out marginal cases in any problematic sense of "ruling out". We focus on a certain group for the sake of picking out the relevant property, but of course the property can extend beyond the group used to pick it out. Compare: We focus on obvious cases of "pink" to define that color property by example. Non-obvious cases, such as Lucia's well-hidden socks, we can't use as examples for the sake of the definition, but that doesn't prevent them from being straightforward bearers of the property once it is defined.
Thanks, Eric! I do suspect that there are residual worries about sufficient neutrality here. In this response, you're not just limiting who (humans), but also limiting how (introspection). That might tilt things in ways some parties might object to (e.g. arguably privileges certain claims about relations of phenomenal consciousness to access consciousness). Also, I suspect that your response here does a better job at screening off robot worries than zombie worries. It might be that our world has one or more zombies in it. "Ordinary adult biological human..." might nonetheless let in a lot of zombies. Maybe even *only* zombies, depending on who you ask.
Fair enough. The definition isn't wholly innocent, only as innocent as I can manage. It does assume that phenomenal consciousness is often accessible and that the seemingly obvious, "accessed" cases are indeed cases of phenomenal consciousness, and that this "we" doesn't include any significant number of zombies -- and several other things besides. Undercut any of these assumptions and the definition falls apart.
From my own perspective, I think the biggest problem with your innocent definition is that it requires on an act of ostension within a representational system.
Even thinking of introspections as "ostension" implies that the act of entertaining a concept is akin to pointing; it implies that the target is there, and we just point to it, and it ignores the act of creative representation that is always involved, which can introduce fictions.
A more substantial issue is that pseudo-pointing within a representational system is inherently ambiguous. Under what ontological assumptions is the pointing being done? Are we pointing at the substrate we just activated, or its represented content? Can we continue to consider your definition innocent if we exclude the folk (like me) who don't believe the represented content is straightforwardly real?
What do we do about the many folk who think they are pointing at something that exists above and beyond the representational substrate, such that a zombie performing the very same ostensive act is deemed not to be pointing at the same thing? When they notionally subtract the zombie-contaminated aspects of the pointing, they are potentially left with a phantom: content that puzzles them, but others might think has no attachment to any ontology.
Such folk cannot point innocently, as far as I am concerned, and my idea of what they are picking out as the most obvious thing on the list and their idea of what they are picking out cannot be reconciled.
Compare pointing at chairs to define "chair" by example. Some pointers might think chairs aren't robustly real; there are only particles arranged chairwise. Others might think that in pointing at chairs they are pointing at instances of some Platonic ideal. This introduces problems, yes. But still, assuming those people are wrong, we are in fact all pointing at the same type of thing, and the definition by example still works. Even if one or the other group with a nonstandard view is right, the definition by example still might work, depending on how much such detailed metaphysics really matters to property identification.
I'm really enjoying this discussion! Let me say first to Eric Schwitzgebel: I'm a fan of your work, and I hope you weren't bothered by my comparing you with Andy Partridge in my silly series of philosopher of mind / rock star posts on Bluesky... FWIW I'm a *huge* XTC fan - in the same ballpark for me as Tom Petty (and Pete Mandik :-) - so that was definitely meant as a compliment.
My concern though with the approach you suggest picks up on the comment above about ostension. You've rejected comments (e.g. from Keith Frankish) suggesting that the obvious feature the positive examples have in common (and that distinguish them from the negative examples) is that we're aware of them, noting a difference between awareness as an epistemic relationship and awareness as experience.
There are two possible concerns here. One is that it may imply that experiential awareness is a special form of awareness (perhaps implying direct acquaintance?) raising questions about neutrality. But the concern I'm more interested in highlighting is about the notion of ostension in this context. If the obvious common property is a special form of awareness, how can we point to it? That is, if experience is a form of mental pointing (which I think would apply here but maybe not?), wouldn't that be like pointing to an act of pointing? To relate it to the chair analogy, the concern is that it's not like pointing to chairs to identify the shared property, but more like saying "point to the obvious property shared among various acts of pointing to chairs."
I'm allergic to the word "awareness" in this context: It just seems too ambiguously confusing between the epistemic and phenomenal senses. If "experiential awareness" just means "phenomenal consciousness that you correctly judge to be present", then I don't see relying on it as a problem any more than building accurate judgment into an ostensive definition of chairs is a problem. (It is a problem if one is too inaccurate, of course.)
On mental pointing: I'm not inclined toward a theory on which phenomenal consciousness is itself a type of pointing, so it's more like pointing to a chair. But let's allow that phenomenal consciousness is a type of pointing -- maybe it has inherent intentionality or self-reference that becomes a kind of pointing in a way. I don't see a problem here. As an analogy, consider examples of physical pointing with fingers. I could say, "by 'pointing' I mean *that* kind of thing, and then gesture to several examples of pointing". I might even point at those examples and use my own pointing as a further example. So far, no self-referential paradox or other problem that I can see.
Given how an LLM is trained — by massive amounts of human-created text — it delivers a distillation of what humans have written. So, it may well be offering "folk-psychologically obvious" content. In fact, that may be the only thing it can deliver.
I've come to believe that some things are both too complicated and too fundamental to be defined. They can only be described, which often involves pointing to enough positive and negative examples to create a convergent picture of what's meant. I very much suspect human consciousness is in that category.
You saying I'm not a zombie?
Qualia seem to be real and widespread among humans, but I sure as hell don't have them. I could be mistaken about anything, but not about this. I know it... directly.
I'm sick if people just presuming I'm conscious.
You say I'm not a zombie, per definition, since I claim to be one? Well zombies get psychiatric illnesses too. Perhaps my condition is something like Cotard syndrome. Only it happens to allow me to make true statements about my lack of consciousness. The fact that I would make the same statements claiming to be a zombie if I were an ordinary conscious human with Cotard syndrome doesn't change that one bit.
Anyway, great piece!
THIS is the zombie uprising we need.
Join us! We're hanging out next Friday. Not that it's gonna be like anything, but still.
Yeah I remember our conversation in which you claimed not to be conscious despite all manner of behavioral evidence you were. Absent a settled theory of consciousness, I can't prove you're sentient but, as we agreed before parting, it would be a big mistake to treat you as if you weren't.
https://naturalism.org/philosophy/consciousness/pain-vs-propensities-conversation-with-a-zombie
I have a hard time believing I ever denied without qualification being conscious. It’s frustrating when people who know that the meaning of term “consciousness” is under dispute pretend that it isn’t in order to portray their opponents as idiotic or psychopathic. At least you didn’t do so in the New York Review of Books!
Oops, my comment and link to my "conversation with a zombie" was meant to be a light-hearted response to Mark Slight who was (in jest, we assume) claiming to be a zombie. I guess I have to figure out how the comment/reply structure works in Substack, my apologies. In any case, no, you've never denied being conscious tout court that I know of. I know Galen Strawson in the NYRB accused Dennett of committing the silliest mistake ever in philosophy by denying what Strawson sees as the obvious fact of phenomenal consciousness.
Ah, no apology needed. I replied without checking the larger context, and now look like someone who can’t process humor. Allow me to apologize instead! My bad.
Remember to always check the larger context of Mark Slight
my context checkers have checked out for spring break and i’m missing a lot of context in the meantime
You not process humor? That's your damn day job bro! ;-)
time to send me to the glue factory
Your claiming that I might be conscious is the silliest clame ever made. It's just obvious to me that I'm not. How would you like if I claim you may not be conscious?
Forget about theories about consciousness. I know this, directly. Introspection, right?
Respect me like anyone who's conscious, but that entails not implying that I'm conscious. Thank you.
Asking for a neutral definition of *phenomenal* consciousness is asking for trouble given that "phenomenal" is philosophically freighted, no? Rather, one might ask what's an innocent, example-based definition of what being conscious consists in, but leave out such terms as phenomenal, qualitative, etc. One pretty innocent folk-psychological definition is that sensory experiences like seeing red have a distinctive character that distinguishes them from their sensory cousins like blue. Illusionists and qualia quietists should agree that such characters exist, but will dispute there's anything necessarily phenomenal or qualitative about them. Experiential characters, they might say, can be cashed out functionally or behaviorally or physically. I'm not sure if appeal to experiential character is non-circular, but if someone says no such characters exist, then they have to say what does present itself in perceptual experience.
It's cool GPT identified “recursive access modeling with internal access by downstream interpretive modules” as the common feature of its vision analog. Let’s say an AI representational architecture is designed so that such access can’t decompose what’s being accessed into sub-components. This would parallel one aspect of a basic experiential character like red: it’s monadic, non-decomposable into other color components, cognitively impenetrable, so its color character is only reportable as being “like this”.
Wow, you sound like a Qualia Quietist! What happened to you?
Well, just giving you guys the benefit of the doubt. I remain a phenomenal realist since I'd say phenomenality *just is* to have in one's experience basic sensory characters that can't be decomposed, so aren't effable, only nameable. I expect as a quietist (in your illusionist moments) you might say that experiential characters *seem* to exist, but remain agnostic about their actual existence since it isn't obvious how they comport with physicalism. I agree it isn't obvious, but I'm not wedded to physicalism and keep my options open as the investigation proceeds.
i’m not ruling out the abstract
less doubtful: “Some people have written or uttered ‘there is something it’s like to see red’.”
Yeah but I was talking what's less doubtful about *experience* such that you can claim it really doesn't seem to me that red has a distinctive character. What backs up that claim in terms of what's evident about experience? Or is nothing evident on your view?
it’s evident that people make judgments about the world, and often are reliably caused to do this by the parts of the world those judgements are about. Any supposed intermediaries between the judgments and everything else either have their third-personal physical bonafides or else are a warped fantasy of magical pseudojudgements like qualia, sensations, experiences, sensory intuitions, sense data, figments, mental paint, qualities, the mythical “given,”raw feels, phenomenal appearances, yadda yadda, blah blah blah, no thanks.
Sounds like you're a skeptic about experience, period, so no wonder you're quiet about qualia, except to say they are magical pseudojudgments. Dreams on your view are judgments about the world - wrongly applied concepts - with no experiential intermediary. And concepts have third-person physical bonafides - nothing abstract about them. Fascinating, captain!
Yeah, I might grant that you believe that (thus making me lean illusionist), but more likely I’d raise doubts about whether things actually seem that way to you or to anyone else (meta-illusionism&qualia quietism)—like how it neither seems that the sun goes around the Earth nor the other way around.
Such doubts have to be placed in the context of what can't be doubted about our experience. What is that on your view?
everything can be doubted
Ok, I figured you'd come back with that. Then what's less subject to doubt than my assertion that characters exist in sensory experience such as the character of how red looks to me? Well, it's the claim that they only *seem* to exist. But you say there's something even less doubtful than that by which the (illusionist) seeming claim is cast into doubt. What is that on your view?
Interesting piece.
I don't think Eric's search for an innocent definition can succeed in a framework where zombies are considered genuinely possible. If we had it on God's authority that there was a zombie world next door, then, as a direct consequence of the fact that zombies have been defined as our cognitive isomorphs, we could not draft a definition that was rational, defensible, included us in the list of conscious beings, and excluded them.
They will remain in definitional lockstep with us at every turn, and their defence of their definition will always match ours, leaving us totally unable to say anything sensible about where their definition falls down. This quandary is intrinsic to the idea of zombies, and to any definition of consciousness that leans on the idea of a human-zombie difference as the core feature of phenomenal consciousness.
I think Eric's search can be a *starting* point for a successful definition, provided that it takes care to put aside the possibility of zombies. By then, of course, it is no longer theoretically neutral. But there is no consistent framework available that keeps zombies.
I think standard uses of "phenomenal consciousness" suffer from fatal internal contradictions, including the ones you highlight, and the term has been so abused it is beyond rescue. There are, however, sub-concepts within "phenomenal consciousness" that can be named and considered separately. The component of the concept of "phenomenal consciousness" that separates us from zombies needs to be named and shamed, and then put aside.
As to your titular question, I would say that I would be embarrassed if me and my zombie twin disagree. I must be ready for my zombie twin to agree with my personal definition, and if I want my opinions to be defensible, I would need to keep revising my opinions until I am not embarrassed on its behalf when it discusses consciousness. I will be left with definitions that deny a human-zombie difference, and my zombie twin would also deny such a difference. It could not rationally believe in its own zombiehood or non-zombiehood; that dimension of separation is spurious.
This definitional issue can be *raised* in the presence of zombie agnosticism, but it cannot be resolved in the presence of zombie agnosticism.
i’m puzzled by your first paragraph. “by definition” of what?
I will edit it to make it clearer. I mean, "as a direct logical consequence of what we accept when we agree to the suggested definition of zombies".
if god is around and assuring us that there are zombies a world away, maybe he’ll also assure us that dualism is true. It would be irrational at that point to complain about the implied definition of consciousness.
You do understand the practice of considering hypotheticals, though, even ones we disagree with, because your own post is based on that practice.
It almost seems as though it is okay to raise the spectre of zombies to make some points, but once they enter the chat in someone else's line of thinking, it's irrational to consider the consequences of their existence. That would be somewhat unfair.
The point is that even if God gave us an explicit endorsement of zombies, and hence dualism, it still would not help us define the element that distinguishes us zombies. such a difference is fundamentally beyond our grasp even if we try to accommodate it by assuming that it exists. That makes the posited difference pointless to think about. It also means that it is *not* what we're actually puzzled about.
And, of course, I don't think it makes sense to posit that a human-zombie difference does exist, for that very reason, and I would politely suggest to God that whatever difference he is talking about it when he discusses our zombie neighbours, it cannot be a key part of any consciousness that we should care about.
the element would be something nonphysical. we would then know that no physical-compatible definition of phenomenal consciousness can be correct. I may still be missing something, but your conjectures that I don’t know how hypotheticals work or that I lack some capacity for fair treatment seem like jumping the gun. I have tried to indicate why I’m not understanding the point you’re making, and I continue to fail.
I think the issue here us that you are linking God's endorsement of a particular subspecies of dualism with the end of any need to explain the many versions of consciousness that are not covered by the idea of a human-zombie difference. Knowing zombies were in the world next door would tell us we had some extra of no particular importance.
We would still have the original true target of our puzzlement waiting to be explained.
The idea of the world next door is merely raised to distinguish two very different meanings of phenomenal consciousness. The one separating us from notional zombies cannot be the important thing that stands in need of definition, in part because the human-zombie difference evades definition completely, even if we go out on a limb and grant its existence.
This is your own point being expressed from a different view.
My question for any particular definition of "phenomenal consciousness" is, does it still lead to an explanatory gap / hard problem? I generally don't see it with the most innocent pre-theoretical versions. And explaining why there's still a hard problem seems to inevitably bring back in things like ineffability, irreducibility, privacy, direct acquaintance, etc, all of which I think amounts to a theoretical commitment to fundamental consciousness. But that theoretical commitment seems optional for the innocent versions of phenomenal consciousness.
depends on what “lead to” means. If it means “entail,” that would be a poor definition, and would make arguments for gaps and hardness plainly question-begging. If it doesn’t mean “entail”, it’s hard to see how anyone could credibly claim to have a priori arguments against physicalism.
I meant "entail". And I agree on a priori arguments against physicalism. I actually don't think there is a hard problem without assuming non-physicalism. If so, trying to solve it with non-physicalism seems like another case of circularity.
Terrific post! Thanks for the thoughtful engagement. I'm inclined to respond to your dilemma by requiring first-personal obviousness and defining "us" not phenomenologically (which I agree risks circularity) but in biological and social terms -- that is, ordinary adult biological human beings in our 21st century mainstream Anglophone context.
This definition by example does require that this "we" all have introspective access to their experiences, but I think that's okay. I don't think it rules out marginal cases in any problematic sense of "ruling out". We focus on a certain group for the sake of picking out the relevant property, but of course the property can extend beyond the group used to pick it out. Compare: We focus on obvious cases of "pink" to define that color property by example. Non-obvious cases, such as Lucia's well-hidden socks, we can't use as examples for the sake of the definition, but that doesn't prevent them from being straightforward bearers of the property once it is defined.
Thanks, Eric! I do suspect that there are residual worries about sufficient neutrality here. In this response, you're not just limiting who (humans), but also limiting how (introspection). That might tilt things in ways some parties might object to (e.g. arguably privileges certain claims about relations of phenomenal consciousness to access consciousness). Also, I suspect that your response here does a better job at screening off robot worries than zombie worries. It might be that our world has one or more zombies in it. "Ordinary adult biological human..." might nonetheless let in a lot of zombies. Maybe even *only* zombies, depending on who you ask.
Fair enough. The definition isn't wholly innocent, only as innocent as I can manage. It does assume that phenomenal consciousness is often accessible and that the seemingly obvious, "accessed" cases are indeed cases of phenomenal consciousness, and that this "we" doesn't include any significant number of zombies -- and several other things besides. Undercut any of these assumptions and the definition falls apart.
From my own perspective, I think the biggest problem with your innocent definition is that it requires on an act of ostension within a representational system.
Even thinking of introspections as "ostension" implies that the act of entertaining a concept is akin to pointing; it implies that the target is there, and we just point to it, and it ignores the act of creative representation that is always involved, which can introduce fictions.
A more substantial issue is that pseudo-pointing within a representational system is inherently ambiguous. Under what ontological assumptions is the pointing being done? Are we pointing at the substrate we just activated, or its represented content? Can we continue to consider your definition innocent if we exclude the folk (like me) who don't believe the represented content is straightforwardly real?
What do we do about the many folk who think they are pointing at something that exists above and beyond the representational substrate, such that a zombie performing the very same ostensive act is deemed not to be pointing at the same thing? When they notionally subtract the zombie-contaminated aspects of the pointing, they are potentially left with a phantom: content that puzzles them, but others might think has no attachment to any ontology.
Such folk cannot point innocently, as far as I am concerned, and my idea of what they are picking out as the most obvious thing on the list and their idea of what they are picking out cannot be reconciled.
Legitimate concerns, of course!
Compare pointing at chairs to define "chair" by example. Some pointers might think chairs aren't robustly real; there are only particles arranged chairwise. Others might think that in pointing at chairs they are pointing at instances of some Platonic ideal. This introduces problems, yes. But still, assuming those people are wrong, we are in fact all pointing at the same type of thing, and the definition by example still works. Even if one or the other group with a nonstandard view is right, the definition by example still might work, depending on how much such detailed metaphysics really matters to property identification.
I'm really enjoying this discussion! Let me say first to Eric Schwitzgebel: I'm a fan of your work, and I hope you weren't bothered by my comparing you with Andy Partridge in my silly series of philosopher of mind / rock star posts on Bluesky... FWIW I'm a *huge* XTC fan - in the same ballpark for me as Tom Petty (and Pete Mandik :-) - so that was definitely meant as a compliment.
My concern though with the approach you suggest picks up on the comment above about ostension. You've rejected comments (e.g. from Keith Frankish) suggesting that the obvious feature the positive examples have in common (and that distinguish them from the negative examples) is that we're aware of them, noting a difference between awareness as an epistemic relationship and awareness as experience.
There are two possible concerns here. One is that it may imply that experiential awareness is a special form of awareness (perhaps implying direct acquaintance?) raising questions about neutrality. But the concern I'm more interested in highlighting is about the notion of ostension in this context. If the obvious common property is a special form of awareness, how can we point to it? That is, if experience is a form of mental pointing (which I think would apply here but maybe not?), wouldn't that be like pointing to an act of pointing? To relate it to the chair analogy, the concern is that it's not like pointing to chairs to identify the shared property, but more like saying "point to the obvious property shared among various acts of pointing to chairs."
No problem about the comparison, of course!
I'm allergic to the word "awareness" in this context: It just seems too ambiguously confusing between the epistemic and phenomenal senses. If "experiential awareness" just means "phenomenal consciousness that you correctly judge to be present", then I don't see relying on it as a problem any more than building accurate judgment into an ostensive definition of chairs is a problem. (It is a problem if one is too inaccurate, of course.)
On mental pointing: I'm not inclined toward a theory on which phenomenal consciousness is itself a type of pointing, so it's more like pointing to a chair. But let's allow that phenomenal consciousness is a type of pointing -- maybe it has inherent intentionality or self-reference that becomes a kind of pointing in a way. I don't see a problem here. As an analogy, consider examples of physical pointing with fingers. I could say, "by 'pointing' I mean *that* kind of thing, and then gesture to several examples of pointing". I might even point at those examples and use my own pointing as a further example. So far, no self-referential paradox or other problem that I can see.
I think it is more akin to a character in a complicated simulation pointing to a chair, convinced within their software brain that it is made of wood.
What are they actually pointing at? In this case, we know the underlying ontology, and it is still ambiguous.
Given how an LLM is trained — by massive amounts of human-created text — it delivers a distillation of what humans have written. So, it may well be offering "folk-psychologically obvious" content. In fact, that may be the only thing it can deliver.
I've come to believe that some things are both too complicated and too fundamental to be defined. They can only be described, which often involves pointing to enough positive and negative examples to create a convergent picture of what's meant. I very much suspect human consciousness is in that category.
Zombies and robots agree with you and they are also pointing at and calling something "consciousness".