# 28 philosophical zombies, by chaospet
Denyse O’Leary writes (ironically in a UD post that has attracted 500+ comments, in none of which is the word mentioned, or the topic touched):
Materialist neuroscience has a hard time with qualia because they are not easily reducible to a simple, nonconscious explanation.
And she is not alone in this view. I would agree that “materialist neuroscience” does have a Hard time explaining qualia, but that’s not, in my view, because qualia are some mysterious extra thing that neuroscience has to explain but because they are not – they are intrinsic, I would argue, to the ability to act as an intelligent autonomous agent capable of navigating the world, making use of its resources and avoiding its hazards. Which is another way of saying they evolve, but that’s not what I’m interested in arguing right now (although I do think it lies at the heart of the ID case).
The point of the cartoon (or at least the reason I posted it) , is that when we postulate a “philosophical zombie” we postulate a physical simulacrum of a person but with something missing. However, a philosophical zombie, being, by definition, undifferentiable from a normal person, cannot be any scarier than a normal person. The difference between a person and a p-zombie, is not detectable by any independent observer. Hence the Hardness of the Problem for science, which relies on independent observation in order to allow conclusions to be drawn.
If a person really does consist of their material contituents arranged into a person-capable system plus Something Else (S+) where that S+ is not necessary for its person-capability, then it is certainly not possible for science to detect, or detect the absence of, the putative S+. And yet, the argument goes, that doesn’t mean the S+ doesn’t exist, and moreoever, we have direct subjective evidence of it in the form of qualia.
My p-zombie may recoil from a hot plate just as I do, but it doesn’t actually feel the pain, it just acts as though it does. It may weep at the death of its p-zombie mother, but it doesn’t feel unhappy, it’s just that its serotonin levels adjust in response, and produce behaviour that we (and it) call “grief-stricken”. It also tells us how it feels, but it doesn’t actually feel anything, it just behaves as though it does, including the part that involves telling us how it feels. So we aren’t going to solve the S+ problem by trying to find it. We won’t. The issue is not how to solve such a Hard Problem, but whether there is a problem at all. But, contrary to some assertions, saying there is no problem doesn’t mean that there are no qualia. It would just mean that qualia are the direct result of the physical existence of a being made of the same stuff, in the same configuration, as we are.
So, on the hypothesis that an S+ is not required to account for qualia, how do we account for them?
Qualia are sometime described as “raw feels” – how something feels without any reference to any link between the thing-felt and anything else. The paradigm case is colour sense – we can know what blue is, in terms of its wave-length, in terms of what things are coloured blue, how our colour vision system works, what different blues are called – but there seems something missing from all that – just the experience of blueness, independently of any knowledge about blue.
But that doesn’t work, as can be readily shown by the evidence that the perceptual experience of a colour (the “raw feel”) is hugely influenced by context. Check out the brown square at the centre of the upper surface of this cube, and the yellow square at the centre of the surface facing you, then play the video:
And then there’s this one of course:
So, I think the idea of representation-free “raw” feels is somewhat bogus.
However, it remains reasonable to ask: how do I know that when you look at the coloured cube and see yellow and brown, which turn out to be the same colour, what you see is the same yellow and brown as I see? Well, I don’t, of course. So does that mean that there is an S+ that I experience in addition to simply “knowing” that the top square will be perceived by everyone as the colour they call “brown” and the side square will be perceived by everyone as the colour they call “yellow”? The brownness of that brown percept and the yellowness of that yellow percept?
Well, there may well be something extra. But while I think it’s additional to the knowledge that what I’m seeing will generally be agreed to be yellow/brown, I don’t think it’s additional to what is explicable in material terms. I think a p-zombie of me would also see “my” yellow/brown, in other words would have my qualia, and therefore wouldn’t in fact be a p-zombie.
I suggest that the additional “feel” of yellow is something to do with the vast cocktail of hard-wired and learned associations between yellow and the world – brightness, lightness (yellow activates a double set of receptors), sunlit scenes, bananas, egg-yolks, dandylions, Simpsons, parking restrictions; each of which induces a tiny ghost of a motor response related to those associations, and which we reify as the “feel” of “yellow”.
It’s probably easier with red, possibly the most canonical colour, used for fire engines, to , to correct errors, to note that you are overdrawn, warn you of danger. Why red? It’s the color of blood, and fire – both exciting and dangerous, warmth, food, hunting, violence, catastrophe. It’s also the colour of a lot of fruit. Why? Possibly because fruit needs to be eaten if it’s going to be propagated, and evolved to be red because humans notice red (because of fire, blood, etc). And why are we good at seeing red? I suggest it’s no coincidence that our most sensitive receptors are for red and green – green the colour of “nothing happening” (more leaves, meh), and red the colour of something requiring attention – useful to distinguish between the two, and thus also useful for fruit to turn red when ripe, remaining green while it doesn’t “want” to be eaten. And green is supposed to be a soothing colour (hence “green rooms”). Blue receptors are maybe a bit of a luxury, but perhaps worth having anyway, if only to help distinguish between, say, purple and red things, or cyan and green things, and perhaps they co-evolved with the colours of fruit and flowers. Maybe. Maybe it was just a freebie.
My point being, that what seems like a “raw feel” may in fact be a highly processed nutritional supplement of motor response-sets, a kind of “stance” we take towards a colour, superimposed on the specific associations induced by that colour in that context (a red rose vs red sunburn, for instance) – Extra, yes, but not anything beyond the reach of a neuroscientific account.
But that still doesn’t address the issue of why that motor set should be experienced, you (well some) cry.
Well, I think it does. I think it’s key. The point about a motor program that is activated (even if not executed) in response to a stimulus is that it can both bring in new sensory data (an eye or head movement for instance, or bring something in to touching range), as well as actually alter the world, and the organism’s relationships (e.g. spatial relationship) within it. That additional data, and the changes in the world brought about by the motor action then require that the organism updates its world-model, including its own parameters within that model. In other words, an organism that can move (and I think it is no coincidence that organisms with brains are all organisms that can move, i.e. not plants, or even sponges) must possess some kind of world-model that incorporates the organism itself, as well as some kind of updating system. This is likely to involve both old maps and potential new maps (so that the organism can make decisions depending on where it will end up, given the next action). So it has to model itself as a persisting entity with a dynamic relationship with the world. And we have good neuroscience models for this – we even have robotic models for this.
Now I’m not suggesting that Asimo is any more conscious than the most primitive of aware organisms. But I am suggesting that inasmuch as Asimo resembles a human being (not much, but a little), Asimo is not a zombie. Asimo’s “experience” may be so simple as to be unimaginable by we whose experience is so rich. The nearest I can manage is that that dream state where you are looking for something, but you don’t know what it is, nor why you want it, but it’s important, and stuff keeps getting in the way.
So I suggest that the reason p-zombies are not possible, even in theory, is that anything capable of responding as richly and as flexibly as we can to the world (and which, by definition, it can) must necessarily have a rich and complex map of the past and future world on which the p-zombie itself is represented, and which can be predictively updated in order to make sensible (and even foolish) decisions (interesting that the robotics team can set Asimo’s risk threshold. It must be able to simulate, extremely rapidly, many potential motor responses to a stimulus, and get them on standby in order to react appropriately as more data comes online. Indeed those motor responses must include responses that will bring appropriate data – sensory input – into the system, and ready potentially appropriate motor programs. In other words, it must have qualia.
I rest my case.
I am what is sometimes called a “qualiaphobe”. I see “qualia” as a useless word that does not refer to anything.
Roughly speaking, I see it as a word introduced to the language to allow us to pretend to talk about what cannot be talked about. It was introduced as a first step in providing a reductionist account where no reductionist account is possible and no reductionist account is needed.
People who favor the use of the word “qualia” are sometimes called “qualiaphiles.” As I see it, if somebody is a qualiaphile, we can be reasonably sure that person is on a wrong track in his inquiries into questions related to human cognition.
Blue seems unlikely to be a luxury, since it predates red in the primate repertoire of color perception. Indeed, many new world monkeys still don’t have red cones. And while fruit is often thought to be the reason for the evolution of primate red cones, others have suggested that young red vegetation might be more important for old world primates evolving red receptors.
It’s backward to suppose that red fruits evolved for us to eat those fruits (if indeed that is why we perceive red), as birds with their 4-color vision systems are almost certainly more important for dispersing seeds. We may have evolved to join in the feast of red fruits, however, or perhaps for eating the young reddish vegetation–or both.
The sense that qualia simply are part of existence, or whatever, seems not to take account of the fact that they can be quite different, and indeed, basic qualia can be gone from human experience in extraordinary cases. One may see only black/white/gray at high G forces. There is also “blindsight,” wherein people are unaware of having vision, yet respond to questions as if they are seeing (probably in a very simple manner) at least shapes and objects. No, not a miracle, there apparently are live pathways for visual information in these people, it’s just not reaching a level of conscious awareness–at least until questions are being answered (but that’s not visual consciousness per se). However, that seems unsurprising as well, since it’s hardly new that we can pick up visual information without any apparent consciousness at the time, and much of the brain evidently operates without consciousness (or at least it’s a quite separate consciousness).
My sense is that consciousness is the one thing we know that is experienced as what it is “in itself.” Everything else we “know” is something translated into nerve impulses, some of which do not become conscious, some of which do. The interactions of electric fields being the difference. If it’s a jumble of chaotic electric field interactions, then it’s just noise and not experienced as any kind of consciousness, unless it’s felt as a sort of a sense space at the edge of coherent consciousness. “Normal consciousness” is where there are coherent interactions of the electric fields, that is, forces above the level of noise, the register of forces in the electromagnetic realm. Anyway, it’s a model of actual observable causes, not some sort of woo having nothing to do with the fact that we simply experience data encoded within electric impulses.
Above all, I have to note that qualia do not obviously pertain to robots, or even to our unconscious minds.
Glen Davidson
Interesting about blue preceding red.
Still, it doesn’t spoil my argument about why red should be such a salient colour!
I’ve seen drawn a distinction between logical modality (necessity and possibility) and metaphysical modality (necessity and possibility).
Something is logically possible if it can be conceived of without contradiction, and logically impossible if it cannot be conceived of without contradiction (e.g. square circles). Whereas metaphysical possibility (and necessity) take into account space, time, causation, mathematics, and a very large set of possible laws of physics.
In that sense, when Lizzie says that
I take her to be saying that zombies are not metaphysically possible, even if they are logically possible. And I think that’s basically correct. However, I think that Lizzie’s way of putting it isn’t quite right.
On her view, it is metaphysically necessary that anything that displays a complex repertoire of adaptive behaviors must be able to not only map its environment, but also map its own location in that environment and map its relations to that environment. (In Sellars’ terms, it must be able to “picture” its environment.) And that’s right.
But usually “qualia” are understood as an ineffable Something Else — not just a third-person accessible cognitive mapping from inputs to outputs, but the very pith of the first-person, subjective stance on the world. Now, I do not find “qualia” a useful term for elucidating what is available from the first-person, subjective stance. But I think it would be a mistake to identify qualia with picturing, because it collapses the distinction between the first-person and third-person points of view. The most we could hope for, I think, is correlating those standpoints as precisely as we can — a project that is sometimes called “neurophenomenology” — rather than collapsing one into the other.
I didn’t mean to give the impression that I thought qualia were to do with “picturing” – although visual sensory experience tends to figure largely in discussions of qualia. What I was trying to express was the idea of qualia as a preparatory stance with regard to potential motor action. Which would also include a preparatory autonomic “stance”.
Lizzie,
Why must a “rich and complex map”, complete with self-representation, necessarily be accompanied by subjective experience? Why can’t all that representation and information processing take place “in the dark”, so to speak?
Well, I’d say that because by representing the
Lizzie,
But you can represent something without consciously experiencing it, as the famous phenomenon of ‘blindsight’ demonstrates.
People with blindsight experience knowing what they know as a result of their blindsight, just as a bat experiences knowing what it knows as a result of its sonar. It’s almost certainly different, but it is still experienced, and would almost be more useful if it were more accessible to a forward model. Indeed that’s what therapy for blindsight focusses on. But blindsight is a lovely example as it is lays out in explicit form the idea that perception consists essentially of a preparatory motor response.
You probably know of the blindsight patient that can play ping pong – visual knowledge seems to be accessible as an retrospective analysis of the motor response that was made.
Lizzie,
Yes, visual knowledge is (indirectly) accessible, but the visual qualia are not. The experience of seeing is missing, as the patient himself will tell you.
Well, of course. There is an entire pathway missing – and one that was previously present. So the visual qualia of colour, shape, edge etc, will all be missing,because those are all properties only handled by the lesioned pathway.
What is preserved is knowledge of movement and direction, specifically in relation to what requires a motor response.
I guess what I’m saying is that qualia is knowledge. But you acquire different knowledge via other pathways than via primary visual pathway.
Checking the wikipedia page, I do think that the sentence “Blindsight challenges the common belief that perceptions must enter consciousness to affect our behavior” is misleading. I don’t think “consciousness” is a box that can either be “entered” or bypassed. I think we can become conscious, or not, of things that go via V1, or solely via the other pathways, for example something in peripheral vision. It’s just that one of the functions of extrastriate cortex is to trigger eye movements to a moving object where it is foveated and would therefore enter V1. If V1 is missing, that “something went past me” information won’t be updated with more data acquired via fovea. But one can still be aware of it.
Lizzie,
But not the qualia of movement and direction. That’s my point — their brains represent movement and direction, but they don’t have the conscious experience of those things.
There must be more to qualia than mere representation.
In what sense don’t they have conscious experience of them?
They know they knew. They just didn’t know they knew until they knew.
Also, I don’t think qualia are “representations” – who would they be representing to?
I think they are preparatory states – at least that is my proposal – and that we reify those preparatory states as “qualia”. Although normally we don’t – we rapidly drill down from our generic response to “yellow” to “yellow banana” or “yellow dog poo”, which can be very different. Specifically we bind the “yellow” percept to the “banana” percept.
It’s only when we sit back and try to think of the abstraction “yellow” that we get all excited about qualia. And I would speculate that after a while, exactly the same thing could happen with blindsight – that people would begin to reify the experience of being stimulated to reach for a ping pong ball, just as we can reify – or at least recall – the “qualia” of being surprised, or needing to pee, or thinking we’ve forgotten something but we can’t remember what.
Lizzie,
Qualia are more than mere knowledge.
To have qualia is to experience them as they happen. If qualia were nothing more than ‘knowledge’, Lewis wouldn’t have bothered to coin the term.
I may be able to infer that the cat knocked over the vase while I was asleep. That is knowledge without qualia.
Lizzie,
I don’t think they are identical to representations, which is why I argued above that we can have representations without qualia. However, you seem to be saying that knowledge = qualia:
If you don’t know that you know something at the time, then you aren’t experiencing it. Qualia are absent.
Lizzie:
Those preparatory states can be unconscious, as you seem to acknowledge in writing this:
The experience isn’t available at the time. The knowledge can only be inferred via a “retrospective analysis”. The patient has the experience of inferring the knowledge, but not the experience of actually seeing the ping-pong ball.
I think all experience is retrospective. And we all use “blindsight” to play ping pong,
In fact there’s a different lesion (damage to MT).where you retain the experience of “seeing” statically but cannot see motion
I think “actually seeing” isn’t the “simple” experience that qualiophiles suggest. I suggest we construct experience as we need it, like the light in the fridge. It’s only “on” when we open the door, but because it’s always there when we need it, we don’t think of it as not always there.
Lizzie,
Only in the trivial sense that processing takes a finite amount of time. There is still a big difference between real-time analysis and retrospective analysis.
If you mean that subconscious visual processing is necessary to play ping-pong, then sure, I agree. But we don’t only blind-see while playing ping-pong; we also see.
The very fact that we use the term ‘blindsight’ indicates that it is qualitatively different from normal sight. In characterizing visual qualia as nothing more or less than visual knowledge that we act upon, you are failing to acknowledge that important difference.
I am open to being persuaded, a la Dennett, that qualia are incoherent in some important way, but no satisfactory account can neglect the difference between knowledge and qualia, and more specifically between blindsight and conscious sight.
What do you think the difference is?
I’d say we movement-see and colour-see and shape-see and location-see and thing-see. But with blindsite some of those are missing – with other lesions, different ones are missing.
I’m fully acknowledging they are different. Patients with V1 lesions are really quite disabled. I just don’t think there is “real” seeing and “blindsight” seeing. I think they are all real, and all contribute to the experience of seeing – and if we lose any one pathway, seeing is different.
I’ll keep trying!
I think the difficulty is that you are thinking of “blindsight” as not “conscious” sight because for a “blindsight” patient, there are aspects of the visual world they are no longer concious of – because they no longer have the means to construct the full model. But that doesn’t mean that don’t have a model that they can consciously access, and and therapy focuses on helping patients do just that, just as aural training can help us become aware of aspects of sound that we were hitherto unconscious of.
Actually, the story is a little bit more complicated than that. Lewis (Mind and the World Order, 1929) posited qualia as nonconceptual episodes of mere awareness. But Lewis also held that all knowledge requires the use of concepts. Accordingly, he concluded that qualia have no epistemological function at all. Their function rather is semantic (on my interpretation of Lewis, which is a bit contentious) — he thinks that a concept must be correlated with some range of qualia in order for that concept to have any empirical meaning at all.
Dennett’s attack on qualia follows through on (and is largely dependent on) Sellars’ criticisms of Lewis: there is no “sky-hook of given meaning” (Sellars, “Physical Realism”, 1954). All conceptual meaning is inferential — the meaning of a concept lies in the role it plays in reasoning. The ‘rubber hits the road’ — conceptual frameworks hook up with reality — in terms of how well they allow for successful navigation of the environment, not in terms of their correlations with states of consciousness.
(And if one thinks that Sellars’ criticisms of Lewis parallel Hegel’s criticisms of Lewis, congratulations — you’ve just hit upon the basic idea of the book I’m working on now. 🙂 )
Lizzie,
We can obviously have knowledge without qualia, as I pointed out earlier:
Given that, what else do you think is required for knowledge to qualify as qualia? Does it have to be direct sensory knowledge? Is indirect knowledge good enough?
You also wrote:
Do you think that a retrospective analysis, unaccompanied by the sensation of seeing, qualifies as qualia? I don’t. The very definition of ‘quale’ requires an accompanying experience.
Qualia are experiences, not merely retrospective analyses.
I’m not sure about your argument. Let’s say someone made a robot that was very human looking, in fact, indistinguishable from a human unless you cut it open. Let’s say this was a German robot – that makes it more believable, because Germans are very good at engineering. Anyway, Germans are also (in my experience) awful at programming, so this programmer, rather than constructing a whole cunning mapping routine, controlled the robot via a massive long string of ‘if’ ‘then’ statements – i.e. it was entirely responsive to the environment via simple logical patterns. It’s a really really long string that could account for all things that could ever happen to the robot, but all it ever did was scroll through the list until an if statement read “true” and then responded accordingly. can anyone really argue that a robot mind, scrolling through an if-else table and then responding according to a particular preprogrammed instruction is consciousness?
the motion one is an odd one…. I find that incredibly difficult to comprehend – compare that with colour blindness and such which are easy to imagine.
still, the post hoc rationalization by the guy walking down the hall is amazing:
http://en.wikipedia.org/wiki/Blindsight
Then what about the guy in the BlindSight case study? What was his experience?
In 2003, a patient known as TN lost use of his primary visual cortex, area V1. He had two successive strokes, which knocked out the region in both his left and right hemisphere. After his strokes, ordinary tests of TN’s sight turned up nothing. He could not even detect large objects moving right in front of his eyes. Researchers eventually began to notice that TN exhibited signs of blindsight and in 2008 decided to test their theory. They took TN into a hallway and asked him to walk through it without using the cane he always carried after having the strokes. TN was not aware at the time, but the researchers had placed various obstacles in the hallway to test if he could avoid them without conscious use of his sight. To the researchers’ delight, he moved around every obstacle with ease, at one point even pressing himself up against the wall to squeeze past a trashcan placed in his way. After navigating through the hallway, TN reported that he was just walking the way he wanted to, not because he knew anything was there. (de Gelder, 2008)
his retrospective analysis was completely wrong when compared to the visual inputs that he was not consciously aware of. He was obviously seeing these objects and moving around them, but at no point did the existance of those objects enter his conscious thought processes.
The blindsight study reminds me of another one I heard of. In some blind patients who have completely lost their rods and cones and therefore have nothing that they visually process, they can still tell ‘light’ and ‘dark’, because of the response to melanopsin, although they can’t actually see anything. Very odd.
Experience and Consciousness
I’d say that experiences are analyses – not necessarily entirely retrospective.
I think we construct experiences from data – they are models both of what just happened and of what might happen next.
I don’t think there is any “merely” about this. I do think that models based on impoverished data will be less good, also less salient (less attention-demanding).
Sorry I’m being a bit cursory though – trying to juggle other stuff right now. It’s not meant to be a brush-off.
Yes, this is interesting! Like the split brain patient whose left hemisphere justified choosing an image of a bell for the word “music”, from out of a set of more obviously appropriate images, with some cock-and-bull story about liking bells, when the “real” reason was that his right hemisphere had seen the written word “bell”.
I don’t know about Keiths, but you are clarifying things for me )
I suspect qualia is related to language. Not sure how, but it seems that being aware is coincidental with being able to talk about experience.
I’m just suggesting something that many people have already discussed, and that is that consciousness may be related to the language faculty and may have co-evolved.
My observation of pets suggests to me that they are a bit like your unconscious ping pong player. They are aware of the immediate moment, but do not have the ability to reflect on it. Self awareness requires the ability to sequence memories and this ability is related to language.
JetBlack,
I’m confused. You addressed your comment to me, but you seem to be arguing against Lizzie’s position, not mine.
For example:
I don’t think he had the experience of seeing. That’s why I brought up blindsight in the first place.
As you point out, his visual knowledge was not conscious, which means there were no accompanying qualia. Lizzie, meanwhile, asserts that there were qualia, because for her visual qualia are merely visual knowledge:
I see a difference between concious and unconscious processing. I think qualia accompany the former, but not the latter. For Lizzie, there are qualia in both cases, because there is visual knowledge in both cases.
As far as I can tell, you agree with me, and disagree with Lizzie.
Lizzie,
Sure, but you seem to be arguing the converse: that analyses, whether conscious or unconscious, are experiences. I think TN’s trash can avoidance, and his subsequent verbal report, show that this cannot be true.
Well, if we are operating under the assumption that consciousness in living things is emergent and supervenes upon some more basic physical framework – neuroscience, biology, physics – which, I think, is implied in Lizzie’s position, then we ought to accept multiple realizability of consciousness as a logical possibility. Which is to say, you are right, the situation that you propose is possible under our assumptions. Possible, but unlikely – I think this is what Lizzie’s position amounts to. It would be hard to prove this stance, but I find it plausible.