Keith Frankish on Illusionism, Consciousness, and the Hard Problem

Keith Frankish

I am conscious. But of course we need to sort out what that means, and that's really the whole focus of my work — trying to get a grip on what we're talking about when we talk about consciousness. You have experiences: you have pains, you smell things, you see things, you hear things. These are what we call conscious experiences. They're perfectly real things that happen to us. The question is what is happening.

Keith Frankish

Thank you for the opportunity. What Galen imagined I was saying is that no one ever feels pain, no one ever feels pleasure, no one ever has conscious sensations or perceptions. I'm not saying that. What I am saying is that people — particularly some philosophers, and people who have thought about this a lot — have a mistaken view of what is happening when we have conscious experiences. I think that mistaken view is due in part to an illusion created by our introspective systems, the systems that monitor and model our own mental activity. An illusion isn't nothing happening — take a stage illusion: when you see someone appear to levitate, something is really happening on the stage. It's just not what you're invited to think it is. That's what I think about consciousness: something real and significant is happening, but you're invited, by the nature of your introspective systems, to misconceptualise it. It's that misconceptualisation that lies at the basis of the philosophical debates about consciousness.

Keith Frankish

There's a lot to unpack in that, because it sounds completely innocent, as if no theoretical framework were being assumed — but there is one. Start with a term that gets smuggled in without much reflection: 'presented to me'. What is this 'I'? I mean this particular animal, this biological organism — a product of a certain genetic makeup and developmental history, capable of what it does in virtue of that history. It's entirely part of the natural world; the 'I' isn't something distinct from the body, somehow associated with the animal — it just is this animal. So when I say something is 'presented to me', I mean that systems inside this animal are doing things that acquaint me with things around me or inside me, in virtue of evolved mechanisms — fallible, capable of misrepresentation, tuned by evolution and developmental experience.

Now, what's presented to me? At a pre-theoretical level, it doesn't seem that an inner world is presented to me at all — an outer world is presented to me. In front of me there are bookshelves, all different colours: blue, red, brown, purple. That's what's presented to me — vivid, evocative, and all apparently out there, none of it in my mind. Consciousness, at this starting point, is being aware of the world around me, and the world inside me too — the feeling in my stomach, my heart. All of that is just everyday description. We haven't yet arrived at a private world of 'mental qualities'. The question is how we get from that pre-theoretical position to a world of inner mental qualities — and to whom that mental world is supposed to be presented.

Keith Frankish

That's a nice way of putting it. There is a contribution from the perceiver. Ancient philosophers tended to think the qualities really were out there, and that copies of them somehow came off the surface and hit the eye — or, on a projectivist view, that the eye went out and touched them. We're already in theory here. A more modern view says: the colour of the book turns out not to be colour as we ordinarily think of it, but a reflectance profile of the surface, characterisable in terms of which wavelengths of light are reflected and absorbed. And then there's supposed to be another kind of colour — the blue that I thought was out there but is 'really' in my mind. That's what people call qualia. But that relocation is a theoretical move, not something given at the start.

Your point about the blind person is important, because perception isn't a matter of passively opening your eyes and receiving stuff — it has to be learned, built up through association. Let's talk about it as an interaction between me and the thing out there, determined partly by the thing's nature and partly by mine. I'm thinking of a patient studied by the psychologist Nicholas Humphrey in the early 1970s, called HD. She had gone blind early in life from an infection that damaged her corneas, and later — I think in her twenties, living in Iran — she was given a corneal graft that restored her sight. She had been looking forward to it hugely, expecting a revelation. But the operation, though technically a success, was not that. She could detect and respond to things visually, but none of it had the significance or richness we associate with experience. Classifying things visually was an effort; nothing delighted her. It distressed her so much that she decided it was not worth it, put her dark glasses back on, and went back to living as a blind person.

My explanation is that knowing what it's like to see is something you have to learn — a matter of building up associations and patterns of reaction to stimuli, from birth. Our sense that the world is infused with richness, with 'what-it's-likeness', comes from us: we are projecting the psychological significance those things have for us. HD's world did not mean anything to her because she had not had the chance to build up the neural circuits and associations that would give it meaning. The idea that all that significance is simply packed into a pure quality — and, if it isn't out there, must be in here — is a very simplistic theory. There's another way to think about it: as an interaction, where redness is something like vibrating psychologically with a particular stimulation.

Keith Frankish

I'm thinking of it in psychological terms — in terms of effects. Crudely: it's detecting something and being able to respond differentially to it — carving out certain stimuli that have significance and responding selectively to them, in an open-ended range of ways. If I detect something as red, that opens up different responses depending on context: a stoplight means one thing, a book another. It's an invitation to react in a certain way that has been taken up by my perceptual systems.

Keith Frankish

That needs heavily qualifying. It's behavioural in the sense that it concentrates on the effects of stimulation — but the observable, overt behaviour is just the tip of the iceberg. Every bit of stimulation causes a massive amount of micro-adjustment: to perception, to attention, to memory, to belief. If 'behaviourism' is pejorative because it suggests only overt behaviour, call it functionalism instead — a matter of what functions the system performs, characterised at a very fine level. Lots of stimuli have no effect on me at all — magnetic radiation, say, which I have no way of detecting. A bicycle rusting is quite different from a creature reacting: the rust is a change happening to the bicycle, but the bicycle can't do anything with that information — it isn't using the change to modulate its response to the world. If it had a detector for oncoming vehicles and could act on it, that would be a tiny sliver of what our eighty-six billion interconnected neurons are capable of. I don't think there's a bright line here — there are very few bright lines in nature.

Keith Frankish

We need to think about why the word 'consciousness' matters to us in the first place. One reason is that it marks out the boundaries of ethical concern — some people used to think only humans were properly conscious, conveniently, if you didn't care much about non-human animals. So the word comes loaded with ethical import, and I don't think that import can just be read off the psychology. There isn't some inner light that's on in some creatures and off in others, licensing us to do what we like to the ones where it's off — that's a simplistic, anthropocentric view. We need to look carefully at what's actually happening in these different kinds of systems and decide, as a community, how much we care.

There is one big difference between the self-driving car and us: we don't just respond to the world, we respond to our responding to the world. We have introspective systems — we can monitor not just that the world affects us, but how it affects us. That gives us another layer of functional organisation, self-awareness, which the self-driving car doesn't have. But whether that layer matters ethically doesn't just fall out of the description either — it's something we decide, as a community, putting pressure on each other to draw the boundary where we think it belongs.

Keith Frankish

For exactly the reason I gave for perception: perceiving the world doesn't involve creating a mental simulacrum that an inner eye then observes — it involves being sensitive to stimuli and generating a distinctive pattern of internal reactivity that reflects their significance. Introspection is the same story, except the stimuli are internal — patterns of neural activity — and the introspective system is selectively sensitive to the global aspects of that activity that are useful for me to know, such as whether something bad is happening to me. That itself triggers a further, higher-level pattern of reactivity — most basically, the disposition to say 'that's bitter, don't try it.' If I had made the mistake people accuse me of, it would be a real howler: denying a Cartesian theatre at the level of perception while smuggling one back in at the level of introspection. But I'm saying the same thing about introspection as about perception — sensitivities and reactions, first order, higher order, reactions to reactions to reactions. Perception tunes me into the world; introspection is sensitive to patterns of first-order reactivity in the brain and triggers appropriate further reactions, which is what lets me tell you how things affect me, remember how they affected me, and reason about reproducing or avoiding them.

Keith Frankish

Let's unpack the zombie properly, because it isn't just acting like a human — a clever outward facsimile. It's stipulated to be an atom-for-atom duplicate. Everything that happens physically inside a human happens inside it: eighty-six billion interconnected neurons, glial cells, all wired and working exactly as mine do. It isn't an actor playing a part; it's doing the whole thing.

Take what a neuroscientist would tell you happens when someone smells coffee or sees a rich, chocolatey brown. Even at the retina, the system is already filtering for significance. That continues through the brain: modulating attention, triggering categorisation — 'it's a cat, no, it's a tree' — some of it boosted into linguistic report, some into decision-making, some triggering emotion or memory. The brain, on the predictive-processing picture, is constantly trying to anticipate what happens next. All of that — every bit of it, by definition — happens in the zombie too. A psychologist could run the whole battery of tests for awareness of stimulation on the zombie, and it would pass every one, just as I would. It would tell you it can see the light has changed, that a smell reminds it of its grandmother's garden, that it is phenomenally conscious, that it can't imagine how consciousness could be physical — Chalmers calls this the paradox of phenomenal judgment: the zombie is just as convinced it's conscious, and that consciousness is a mystery, as any of us. But something is supposedly missing. That thing is what I say is illusory.

So would I reject the possibility of a zombie? If you build something atom-for-atom identical to a human, it just is a human, on my view. The 'interior private experience' that's supposed to be missing isn't interior to me spatially — I am this animal, not something hiding inside it. The point to hold onto is: if there really were something extra, it wouldn't be playing any role — it wouldn't be necessary for any of that other behaviour and self-report to happen. It would be a fifth wheel. Our talk of qualities is a pre-theoretical, clumsy way of describing reactivity that's too complex and multi-dimensional to pull apart — so we say 'what's it like to see that', and it's tempting to think there's a pure quality we're directly detecting. But instead of direct, revelatory access to pure qualities, we have indirect, selective, distorted access to immensely complex patterns of reactivity, and we mistake one for the other. That's the illusion.

Keith Frankish

A nice question — though who really knows their own motivations? I don't think we have sound introspective access to all of them; to a large extent we fabricate a narrative from things that have actually happened to us. But here's one story I could tell. I don't think humans are as special as some of my opponents think. I don't think sitting in an armchair reflecting on your own experience gives direct insight into the fundamental nature of reality — I don't think evolution needed to equip us with that, and I don't think the universe has been kind enough to make it pop out for free. It seems too arrogant to think otherwise.

A second reason: I'm not a religious person, though I understand the attraction of these views — I've dwelt on the mysteries of consciousness and tried sincerely to think myself into the mindset of people like Philip Goff and David Chalmers, and I think I understand it. That's one reason I don't want to just call them wrong — illusionism says the experience is real, compelling, vivid, but an illusion. Part of why: I'm an anxious person, anxious about self-deception, about believing things because I'd like to believe them. I'd dearly love to believe I'll see lost loved ones again, but I don't want false comfort; I'd rather live with my eyes open, because self-deception lets you down eventually.

A third reason is that I love scientific explanation — reading about how we've unpicked the mechanisms of cellular metabolism, developmental biology. It's more wonderful to me than anything religion offers, and I think there's a fascinating story still to tell about unpicking consciousness. I don't want philosophers discouraging people from trying, by insisting it's a hard problem best left to armchair metaphysics.

People sometimes suggest I can't have had very rich conscious experience. I haven't taken psychedelics, but I do get vivid migraines — arguably more unpleasant, and just as strange: feeling that you're part of the wall, that you've split in two, that your senses have merged in inarticulable ways. So anyone hoping to convince me I'm wrong with a punch on the nose won't get far — though it might make me pretend to agree.

Keith Frankish

I said I'm a particular animal, a token animal, and we can debate the identity conditions for that animal separately from the self. This animal could lose a hand and remain the same animal; there's no hard-and-fast line for when it ceases to exist — dementia, ageing, the point of death — but broadly, as long as there's a pulse, this animal persists in some sense. There are no bright lines; looking for essences and definitions is a fairly pointless philosophical preoccupation. There are just words, and how we choose to use them.

On the self specifically, I'm heavily influenced by Daniel Dennett: there's a sense in which we have a self distinct from the body, but it's the sense of a fiction — a narrative that talkative animals like us construct. We don't just have experiences, we recognise having them, report them, tell other people about them, and from that we build a story: I am the one to whom this happened, who likes this, wants that. It's selective, and you can tell yourself bad stories and land in a bad state; some people construct parallel stories to cope with trauma. This is Dennett's approach to folk concepts generally — free will, consciousness, intentionality, the self — as things that can't bear heavy metaphysical or scientific weight as ordinarily conceived, but that we can re-engineer to keep their pragmatic value, rather than making them hostage to physics turning out one way or another. If you think freedom and responsibility depend on physics being indeterministic, you're at risk of losing them if physics turns out otherwise; better to re-engineer the concept so it doesn't collapse under that strain.

Keith Frankish

Dennett's line was that we're puppets, but puppets who can control their own strings — puppets with the ability to control the first set of strings using further strings, at different levels of organisation, with no ultimate puppet-master. That ability to control your own strings, at different levels, is what gives you freedom — it's a mechanism of autonomy, and you should love these deterministic mechanisms for allowing you to exercise it. Metaphysically, Harris and Dennett were pretty much on the same page; it was the moral each drew from it that differed.

Keith Frankish

Suppose we have a science of fundamental reality — basic physics, the universal wave function, superstrings, whatever it turns out to be. I'm not a physicist, and I've got no dog in that particular fight. What I do think is that just sitting and reflecting on my own mind doesn't give me access to it. It would be arrogant to think that physicists spend their careers building enormous particle accelerators spanning national borders to crack this problem, while I can crack it from my armchair by thinking about my own mind. Evolution has plausibly given us a powerful sense of our own self-conception, a way of modelling ourselves that feels vivid and potent — Nicholas Humphrey has written a wonderful book, Soul Dust, about why the idea, not the reality, of a magical inner realm is adaptive. Plenty of people have a sense, as powerful as the sense of acquaintance with qualia, that they have direct experience of God. That's an interesting psychological phenomenon worth studying, but I don't think we should trust it just because it's powerful — and the thought that I might discover the fundamental nature of reality just by introspecting seems to me arrogant and anthropocentric. Imagine a race of aliens who are zombies, or were simply never subject to this illusion, studying the fundamental nature of reality, and learning that some species on a remote planet thinks it found the answer by reflecting on what it's like to look at a brown surface.

Keith Frankish

There's a historical case that Descartes's move was tactically aimed at licensing him to overturn the whole Aristotelian-scholastic framework, to make room for the new science — but set that aside. It's often been pointed out that he wasn't entitled to 'I exist' — at best, he was entitled to 'thinking exists'. What is thinking? Does it require words? Could he have doubted without language? If not, then at best he can know that language exists — which is already a long way from our ordinary, basic cases of knowledge: I open my eyes, I look around, I know stuff, even though I could be in error. We are biological organisms, and all our knowledge depends on some mechanism for tracking things, noticing differences, and responding — including knowledge of our own minds. If you think you're an immaterial soul, transparent to itself, needing no mechanism to know its own states, you could say otherwise — but that's the conception you'd need to make Descartes's argument work.

There are no foundations here, in the sense of a mathematical system where you fix axioms and build upward. We have a whole web of beliefs — common sense, experience of the world, reflection on ourselves, science — and we try to find the best overall balance. Descartes's thought experiment is captivating, and that's psychologically very significant — it tells us something about how the brain generates a compelling conception of itself — but you can't build a metaphysics on it. I don't think we need a different metaphysics for ourselves than for any other part of the natural world; we're not that special.

Keith Frankish

Good question. If I had a copy of a Sherlock Holmes story, I could open it, read about Holmes smoking his pipe, and ask: where is Sherlock Holmes's pipe? I can say true things about Sherlock Holmes — that he was a detective, that he took cocaine — and false things — that he was a two-hundred-pound wrestler from Montana. Where is Sherlock Holmes? He's fictional; you can make true statements within the scope of a fiction operator. That's what I want to say about the triangle: it's not anywhere, but you can say true things about it because it's a construction, with the same status as your thoughts about the Sherlock Holmes you imagine.

Keith Frankish

Yes — though not necessarily a representation made of little real triangular lines. Exactly how this works is a fascinating question: your brain uses the resources it has for representing the world to construct representations of non-existent things, and you react to them much as you would to the real thing. The force of the imagining comes from it triggering similar patterns of reaction — imagining the scent of my grandmother's room triggers a wave of emotional response, not because it carries some intangible smell-quale, but because it has effects on me, which is also why we enjoy fiction: not just because we like entertaining non-existent things, but because they move us, puzzle us, spark trains of thought.

Keith Frankish

I was with you until the last step. Yes, it's a mental construction, realised in a pattern of neural activity — but its significance lies in what it means for us. Tell someone the bare facts about Sherlock Holmes and they form a thin image; embed him in the stories, and it becomes rich, triggering thoughts of Victorian London, an escapist world you can fill out in your mind — all of that is a rich body of representations producing effects on you that you find attractive and that hold your attention. Where is Sherlock Holmes? There's a representation created by interaction with Conan Doyle's words; there's no deeper mystery — asking 'where is Sherlock Holmes himself' is one question too many. The right question is how the impression is created, and that's how I want to think about consciousness too.

Suppose light hits the retina, processing happens, and at some point that processing produces an inner image that is simply there. What happens to that image? Is it just aware of itself? I don't see how. For the image to mean anything, some system would have to scan it, notice its features, and produce all the downstream effects the image was supposed to produce. Now suppose we cut out the middleman: the processing produces all those downstream effects without ever creating the image. It's just as good — you don't need the image, you just need the effects an image would have had. You'd have a vivid image of Sherlock Holmes only if it's hooked up to react to and mean something; an inner cinema with no one watching, reacting, is no good at all. If you can get all the effects without the picture on the screen, how would you tell the difference between undergoing an illusion of qualia and undergoing actual qualia? My answer: if you're getting all the effects qualia would produce, you don't need the qualia. Occam's razor: suppose the brain is tricking itself into thinking it has qualia. It's the simplest hypothesis.

Keith Frankish

I'm not suggesting only an external stimulus — with the imagined triangle, the stimulus may be internal, arising through self-monitoring of the system's own activity. Let's try the migraine case, because I like testing it on people. Everyone agrees a migraine involves lots of reactions. First: I'm in a state I don't like — that's a reaction. Imagine removing it: someone asks how I'm feeling, and I say something strange is happening, but I don't mind it. Another reaction: I believe I'm having a migraine — imagine removing that too: I say no, I don't think so, though something weird is happening. Another: it seizes my attention — imagine that released: I'm happy to chat about something else. Another: waves of nausea, which I'll treat the same way — imagine I don't feel like being sick. Strange lights: imagine I'm not seeing any. A desire to lie down: imagine that's gone too. Keep stripping out every psychological reaction, everything characterisable in terms of what the brain is doing — no effect on belief, emotion, desire, attention, memory, behavioural disposition; I'm free to walk around chatting about anything. And yet the horrible thing that supposedly makes a migraine morally significant is meant to still be there. I find that very hard to imagine.

Keith Frankish

So the objection is that we've taken one big pain-quale and broken it into sub-qualia — a nausea-quale, a seeing-lights-quale. I do the same thing to those: strip out the reactive effects of the nausea one by one, and eventually you bottom out in things that don't even look like qualia — beliefs, dispositions. The point of qualia talk is that it bundles together complexes of reactive disposition we're aware of in ourselves — and here's the kicker: it has to be something like that for qualia talk to track anything that matters, because if it isn't tracking something with effects on us, it can't matter at all. Say an animal is 'in horrible pain' but it has no effects on it whatsoever — it's playing happily, nothing bad is happening — then it doesn't matter. It's the people who reify pain as something wholly nonfunctional who aren't taking consciousness seriously; I take it seriously because it's part of this messy, embodied biological world.

Keith Frankish

Then we can make a terminological adjustment: let's define qualia as whatever it is that causes that whole range of effects — because it's the effects that matter, whatever the underlying cause turns out to be. The phenomenal realist says what causes them is some inexplicable, purely qualitative thing outside science; I say it's a pattern of brain activity. If we define qualia by their effects, it's far more likely to be something natural, part of the world we already understand physically, than something requiring an extension of that world.

Keith Frankish

Now we're back to where we started. The world I know most directly is just the world around me when I open my eyes — there's stuff there. This inner world of qualia is a theoretical invention, dating from the late sixteenth and early seventeenth centuries, when Galileo and others realised colours couldn't really be 'out there' as they seem, and relocated them into the soul. Dennett once said the problem of consciousness is the problem of what to do with colours once you realise they aren't really in objects. That inner world could be true, but being an invention doesn't make it accurate, and you can't just assume its accuracy as an objection to rival theories. Before the scientific revolution, people didn't need this move — everyone believed in souls, so recalcitrant qualities could simply live there. The problem became philosophically acute once people stopped believing in souls and had to relocate those qualities into the brain, or into fundamental reality. The alternative — which I take, in a sense, from Philip Goff, though I draw the opposite conclusion — is not to take the qualities out of the world in the first place: there's a contribution from us, consciousness is an interaction with the world, which is exactly why it's not a good candidate for a fundamental building block of the universe. It's not a thing at all — it's a process, an interaction between complex systems and a complex world, and it requires an immensely complex system to support it.

Keith Frankish

No — they're competing theoretical worldviews, and the question is which is more economical and more productive. The Cartesian picture doesn't earn its keep: it gives no real predictive or explanatory leverage, creates a host of problems, and its only payoff is vindicating an intuition we already have about ourselves. If that intuition can be explained another way — as the product of a compelling act of self-modelling, which of course feels compelling to the system doing the modelling, since it isn't observing some independent piece of the world the way it observes other things — then what do we gain from staying Cartesian? Abandoning it removes an epistemological obstacle to understanding other minds: they stop being secret, impenetrable subjectivities and become complex, mappable patterns of sensitivity and reactivity, elsewhere in the world, that we can come to fully understand — which lets us extend or adjust our ethical concern appropriately, and gives us a handle on artificial intelligence. It integrates us fully into the natural world — the final step in de-anthropomorphising our conception of it. The alternative view gives you only the vindication of a private world no one else can ever know about.

Keith Frankish

I wish we could get rid of 'emergence' — it suggests one thing popping out of another, like a whale emerging from the ocean. Chalmers distinguishes weak from strong emergence. Weak emergence: you put components together in a certain way and you get the higher-level thing, and it's perfectly intelligible how that assembly gives rise to it — the property of being a television weakly emerges from its components; that's exactly how engineers design one. Strong emergence: something genuinely new appears over and above what the organisation of the parts would inevitably produce — many people think pain is like that, that the brain does all its processing and then a further thing, the quale, emerges on top. I don't think strong emergence occurs anywhere. Pain is weakly emergent from all those sensitivities and reactions hooked up in the right way — there's no more a hard problem of explaining it than there is a hard problem of explaining what makes something a television.

Keith Frankish

You probably couldn't, given how we currently conceptualise pain — but I'm asking you to reconceptualise it in a way that makes it more tractable, and you're trying to imagine the product of billions of years of evolution, which even an individual cell embodies at almost unimaginable complexity. If the best evolutionary biologists and neuroscientists in the world could walk you, in complete detail, from nociceptor stimulation through everything each system does and how it primes and prompts every other system, and how all of that eventually connects to a conceptual system trying to encapsulate it for communication — and we're nowhere near being able to do that — I think you'd come away saying, yes, that does make sense. That's the work we should be doing, instead of declaring in advance that it couldn't possibly work, or concluding that everything must be conscious because that's the only way anything could be.

Keith Frankish

I'm naturally cautious of soundbites, because they get quoted back at me — but here's an attempt. It's my vivid, vibrant, potent response to the world in all its dimensions and significance. It's being in the world and being aware of it as packed with significance for me, both as a creature and as an individual. Consciousness is what makes the world like something for me — and it's significant because it's constituted by a whole set of responses to it. The illusion is the idea that it's something separate from all those processes: an irreducibly private world, belonging to a different dimension of reality, not part of the shared public world, accessible only to an inner 'I' that's itself hard to make sense of — a view that ultimately builds in a Cartesian dualism. Consciousness is being richly alive and sensitive to the world in virtue of highly complex evolved processes, not some further, unexplainable feature of reality on top of them.

Keith Frankish

Absolutely not — because I'm describing a vibrant, multi-dimensional set of responses, and ChatGPT has one line of response, one dimension of sensitivity: text in, text out. It doesn't explore the world; it has no range of sensitivities beyond written input, and it doesn't even go looking for that input, it just waits for it. It's an amazingly complex text-response system, but still just one dimension, and it isn't tuned to an individual developmental history the way every red thing you've ever seen shapes your response to redness. It's never seen anything red — only text — and even the text has no long-term effect on it: the context window is deleted afterwards. If we built autonomous artificial agents with the range of sensors needed to survive and operate in the world, equipped with something like evolution's toolkit, plus self-modelling and self-monitoring, we might start to enter territory where the word 'consciousness' becomes appropriate to extend to them. But there's no moment where some inner light switches on — that's the wrong picture entirely. We look at what a system does, what its capacities are, and ask whether it's similar enough to us to extend the word — the way we already do with other animals.

Keith Frankish

Conscious like an ant, conscious like a fish — we need to drop the anthropocentrism. If all consciousness ultimately is this vibrant responsiveness, and vibrancy is relative and hard to define precisely, then in principle ChatGPT could be an extremely impoverished form of consciousness — a difference in degree, not simply in kind. But the word 'consciousness' isn't well enough defined for that framing to do much work. We should be thinking in terms of many different dimensions of consciousness, mapping out the different dimensions of reactivity and self-monitoring a given system has — for other animals, for us, for hypothetical aliens, for machines — rather than asking a single yes-or-no question about an inner light that we could never actually observe anyway. Once we've mapped the dimensions, the further question of how much ethical concern, autonomy, or responsibility to extend to a given system is a separate negotiation, to be settled in the light of facts that are, in the end, publicly available to everyone — there's nothing hidden. Study bats carefully enough and you'd know everything there is to know about bat consciousness. That's the constructive work; asking flatly 'is it conscious?' is the least useful question in the whole area, because it treats consciousness as hidden behind symptoms we can never be fully sure of. There are no bright lines anywhere — that's the bottom line.

Keith Frankish

Thank you, Alex — and if anyone wants to convince me by creating some vivid experience in me, I'd rather they send chocolate than a punch on the nose. I enjoyed the conversation; you asked genuinely helpful questions, which helped me articulate what I wanted to say. Not every interviewer manages that.