= 27d567bc2d48d9f40e9ccc20ea370b15)
Mitchell_Porter comments on It's not like anything to be a bat - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (189)
"Qualia" is effectively a name for all those properties which constitute your experience of the world, but which do not exist in the current ontology of natural science (thus we have the spectacle of people on this site needing to talk about "how it feels" to be a brain or a computer program, an additional property instinctively tacked on to the physical description precisely to make up for this lack).
This is a problem that has been building in scientific culture for centuries, ever since a distinction between primary and secondary properties was introduced. Mathematical physics raised the description and analysis of the "primary" properties - space, quantity, causality - to a high art, while the "secondary" properties - all of sensation, to begin with, apart from the bare geometric form of things - were put to one side. And there have always been a few people so enraptured by the power of physics and related disciplines that they were prepared to simply deny the existence of the ontological remainder (just as there have been "irrationalists" who were really engaged in affirming the reality of what was being denied).
We are now at the stage of figuring out rather detailed correlations between parts and states of the brain, described in material terms, and aspects of conscious experience, as experienced and reported "subjectively" or "in the first person". But a correlation is not yet an identity (and the verifiable correlations are still mostly of the form "X has something to do with Y"). Mostly people are being property dualists without realizing it: they believe their experiences are real, they believe those experiences are identical with brain states, but out of sheer habit they haven't noticed that the two sides of the identity are actually quite different ontologically.
Dennett belongs to that minority of materialists, more logically consistent but also more in denial of reality, who really are trying to deny the existence of the secondary properties, now known as qualia. It's possible to read him otherwise, because he does talk about his own experience; but if you look at his checklist of properties to deny, you can see he's a sort of neo-behaviorist, focused on verbal behavior. Indeed, the only thing neo about his behaviorism is that he has a physical model of how this behavior is caused (connectionist neural networks). But he is careful to say quite explicitly that there is no "Cartesian theater", no phenomenal color, no inner life, just people talking about these things.
I cannot tell if you are truly in Dennett's camp, or if you're just rejecting the view that there's something especially problematic about explaining sensations. A lot of people who talk about qualia are trying to emphasize that a description of human beings in terms of causal interactions between pieces of matter is leaving something out. But the things being left out are not in any way elusive or ineffable.
Science seems to be telling us that your whole life, everything you have ever experienced, is nothing but changes of state occurring in a few trillion neurons which have been sitting inside the same small dark space (your skull) for a few decades. Now if that's the case, I may not be able to write an equation describing the dynamics, but I do know what that is physically. It's a large number of electrons and quarks suspended in space by electromagnetic fields. If we are to unconditionally accept this as a description of what our lives and experiences really are, then we have to be able to identify everything - everything - we have ever thought, known, or done, the whole of our subjective realities, as a process composed of nothing but changes of states of particles all occurring within a few cubic centimeters of space. And I have no hesitation at all in saying that this is impossible, at least if the basic ingredients, those particles and fields, are understood as we currently conceive them to be.
Quite apart from the peculiar difficulty involved in identifying complex subjective states like "going diving in the Caribbean on your 25th birthday" with the electrical state of a captive neuronal porridge, the basic raw ingredients of subjective experience, like color qualia, simply aren't there in an arrangement of pointlike objects in space. This is why materialists who aren't eliminativists like Dennett are instead dualists, whether they realize it or not - because they simultaneously assert the existence of both the world of atoms in space and the world of subjective experience. These two worlds may be correlated, that is being demonstrated every day by neuroscience, but they simply cannot be identified under the physical ontology we have.
In my opinion the future lies with a new monism. But "physics" will have to be reconceptualized, if that world of subjective experience really is going to be found somewhere inside the skull, because as things stand there is nothing like it in there. I would also say that doing this is going to require a leap as big as anything in human intellectual and cultural history. It won't just be a matter of identifying the "neural correlate of consciousness". Someone is going to have to go right back to the epistemic beginning, before the distinction between primary and secondary properties, and rethink the whole of natural science from Galileo through to molecular neuroscience, while keeping the secondary properties in view. You can always reduce science to subjectivity, if you're prepared to let go of your models and remember that everything that has ever happened to you has occurred within your own subjective experience, so that's the easy part. What we're aiming for is far more difficult, namely, an objective world-picture which really does contain subjective experience and is true to its nature while also encompassing everything else. Of course, all those people who are out there trying to "naturalize subjective experience" or "naturalize phenomenology" are trying to do this, but without exception they presuppose the current "naturalistic" ontology, and yet somehow that is where the change and the progress has to occur.
Suppose I write a computer program (such as Second Life or World of Warcraft) that simulates the properties of an imaginary reality. Have I now created new "subjective secondary properties"? After all, in the real world, objects do not have owners and copyability, nor levels of mana or hit points. Is this "duality", then?
What about a book that describes an imaginary world? Is it duality because there are only words on the page, and these have no physical correlate to the things described?
The reasoning that you're using is an application of the mind projection fallacy. Human brains have built-in pattern recognition for seeing things as "minds", and having volition -- and this notion is itself an example of an imaginary property projected onto reality. The projection doesn't make the projected quality exist in outside reality, it merely exists in the computational model physically represented in the mind that makes the projection
tl;dr version: imaginary attributions in a model do not create dualtiy, or else computer programs have qualia equal to those of humans. Since no mysterious duality is required to create computer programs, we need not hypothesize that such is to create human subjective experience.
(My emphases.)
You seem to be contradicting yourself there. The mind only exists in the mind?
The intuitive notion of "mind" exists only in the physical manifestation of the mind.
Or to put it (perhaps) more clearly: the only reason we think dualism exists is because our (non-dual) brains tell us so. Like beauty, it's in the eye of the beholder.
Our judgment of whether something is intelligent or sentient is based on an opaque weighing of various sensory criteria, that tell us whether something is likely to have intentions of its own. We start out as children thinking that almost everything has this intentional quality, and gradually learn the things that don't.
It's as if brains have a built-in (at or near birth) "mind detector" circuit that triggers for some things, and not others, and which can be trained to cease seeing certain things as minds.
What it doesn't do, is ever fire for something whose motions and innards are fully understood as mechanical - so it doesn't matter how sophisticated AI ever gets, there will still be people who will insist it's neither conscious nor intelligent, simply because their built-in "mind detector" doesn't fire when they look at it.
And that's what people are doing when they claim special status for consciousness and qualia: elevating their genetically-biased intuition into the realm of physical law, not unlike people who insist there must be a soul that lives after death... because their "mind detector" refuses to cease firing when someone dies.
In short, this intuitive notion of mind gets in the way of developing actual artificial intelligence, and it leads to enormous wastes of time in discussions of dualism. Without the mind detector -- or if the operation of our mind detectors were fully transparent to the rest of our mental processes -- nobody would waste much time on the idea that there's anything non-physical. We'd only get as far as realizing that if there were non-physical things, we'd have no way to know about them.
However, since we do have an opaque mind-detector, that's capable of firing for the wind and the rain and for memories of dead people as easily as it does for live animals and people in front of us, we can get the feeling that we are having physical experiences of the non-physical... when that's a blatantly obvious contradiction in terms.
It's only by elevating your feelings and intuitions to the level of fact (i.e. abandoning science), that you can continue to insist that non-physical things exist in the physical world. It's pointing to reality and saying, I feel X when I look at it, therefore it is X.
(A bit like the religious fundamentalists who say that they feel icky when they see gays, therefore homosexuality is disgusting.)
I would have said, "A bit like philosophers of free will who say that they feel like they could have done something else, and therefore determinism must be false". (:
I upvoted you back to 0 because your comment was thoughtful and well-written, even though I disagree.
Yes, I'm in Dennett's camp. Aside from what other commenters have said, think about it like this:
I have a novel here. It's made of the letters A-Z as well as punctuation, arranged in a complicated pattern. But, somehow, the novel also talks about a plot and characters and a setting and so forth, even though all there is to the novel is letters and punctuation. The plot and characters don't have some magical separate state of existence: they exist because they're built out of the letters.
Same with conscious experience. Right now I'm eating goat cheese and crackers. This experience arises out of the neurons in my brain, and it's intimately tied up with them and the patterns they make. You can't separate it from my past experience and associations and memories (which is Dennett's point about qualia). Of course the experience exists: it's just built out of and associated with a complex pattern of neuron firings in my brain. The experience is not the same as the series of neurons: that would be a category error, just like a character in a book is not the same as the series of letters that make up his description. No property dualism needed. Of course it's difficult to explain this association, because we don't know enough about brain chemistry.
Me too.
Me too.
I think it's a good illustration, but I can give you 'the standard reply' from the anti-materialist: As a physical object, the novel is just a hunk of matter with funny shaped ink blotches on it. The 'plot' and 'characters' you speak of have a mental character to them: they don't exist outside of some mind apprehending the novel, a mind which actively 'constructs' these things rather than passively 'finding them' somewhere in the matter of the book.
So book --> plot is not after all an analogy that helps us understand how a mind can reduce to a pattern of physical matter, because "plot" already presupposes the mind, so any "reduction" would presuppose that the mind is itself reducible.
Yeah, I know this is all wrong - but I've learned to make myself "flip" between a materialist and anti-materialist view.
Hmm. Maybe a better analogy is three stones in a field making a triangle. The triangle exists and is formed by the stones, but this doesn't require dualism, just an understanding that relationships and structures exist and are built out of smaller parts. (I know, that's not exact either.)
Earlier you wrote
The ontological ingredients, and the ways of combining them, which physics gives you are quite limited. You can make shapes (like your triangle), you can count objects, you can consider their motions and other changes of state, you can average quantitative properties, you can consider causal dependence and counterfactual situations. There might be a few other things you can do. But if you are going to have a mind-brain identity theory, and not property dualism, then something built solely using methods like the ones I just listed has to be the experience. It can't just be "associated with" the experience - that would be dualism.
Color is usually mentioned at this point, because it is pretty obvious that no amount of piling up particles, averaging their properties, and engaging in causal and counterfactual analysis, is going to give you redness where there was none, in the simple way that putting three stones in a field really does give you a triangle. If someone proposes that the experience of a certain shade of red is some complicated but purely physical predicate, object, or condition, then from the perspective of orthodox physical ontology, they are proposing a form of strong emergence. (Weak emergence is like the triangle.) And strong emergence is property dualism - it introduces new ontological ingredients.
But although color is the standard counterargument - because of its vividness - any sensation, any thought, anything involving a self, anything like the "experience of an object", is just as much unlike anything that can be made from physics in a weakly emergent way. I challenge you to find a single aspect of your experience which you can unproblematically identify with (and not just associate with) some imagined neurochemical correlate. In every case, you will be taking some subjectively manifest reality, and then saying to yourself, "that is really just neurons doing something"; and in every case, physics alone gives you absolutely no reason to think that neurons doing that has any subjective side to it.
If you don't want to be a dualist, you are going to have to take that subjectively manifest reality, admit that it exists somewhere in exactly that form, and somehow rebuild physics around it. But that is really hard to do.
Three rocks in a field aren't a triangle until there's a brain with a concept of 'triangle' that identifies them as such. Photons of a particular wavelength aren't red until there's a brain with a concept of 'red' that identifies them as such. A creature isn't conscious until there's a brain with a concept of 'consciousness' that identifies it as such.
Third one's tricky because of the self-reference, but that doesn't make it an exception to the general rule. Concepts are predictive models, a model can't make predictions unless it's running on a computer, brains are the one kind of computer that can be mass produced by unskilled labor. Qualia, to the extent that they can be coherently defined at all, are a matter of software. Software can be translated between hardware platforms, but cannot exist in any useful form in the absence of hardware.
And, for the record, the math necessary to fully define a rock is a hell of a lot more complicated than "1+1." Don't dismiss it until you've properly studied it.
It's not just tricky, it's self-contradictory. The mind exists only in the mind, you say?
If you really want to try reducing all of this to physics, I'd recommend that you first deliberately try to dispense with terms which have a technological or user-semantic connotation, because no such thing exists in physical ontology. "Computer" and "software" are being used as metaphors here, and a "model" is an intentional concept. Computer science has the concept of a "state machine", which is a little better from a physical standpoint, because it doesn't attach any semantics to the "states".
OK, fine, you can do such a translation, and you get e.g. qualia are equivalence classes of state machines. At least your claim has now truly been expressed in terms that do not implicitly exceed physical ontology. But it's still a wrong claim, because it says nothing about the properties that really define qualia, like the "<red>" that we've been talking about in another thread.
I don't study rocks, but I study physics every day. I know the mathematics is complicated. What I'm saying is that physics is not mathematics.
So we can set up state machines that behave like people talking about qualia the way you do, and which do so because they have the same internal causal structure as people. Yet that causal structure doesn't have anything to do with the referent of 'redness'. It looks like your obvious premise that redness isn't reducible implies epiphenomenalism. Which is absurd, obviously.
Edit: Wow, you (nearly) bite the bullet in this comment! You say:
I claim that mental states can be regarded as causes, that they are indeed a shorthand for immensely complicated physical details (and significantly less but still quite a lot complicated computational details), and claim further that they cause a lot of things. For instance, they're a cause of this comment. I claim that the word 'cause' can apply to more than relationships between fundamental particles: for instance, an increase in the central bank interest rate causes a fall in inflation.
So, which do you disagree with: that interest rates are causal influences on inflation, or that interest rates and inflation are shorthand for complicated physical details?
No, it just means that <red> plays a causal role in us, which would be played by something else in a simulation of us.
There's nothing paradoxical about the idea of an unconscious simulation of consciousness. It might be an ominous or a disconcerting idea, but there's no contradiction.
See what I just said to William Sawin about fundamental versus derived causality. These are derived causal relations; really, they are regularities which follow indirectly from large numbers of genuine causal relations. My eccentricity lies in proposing a model where mental states can be fundamental causes and not just derived causes, because the conscious mind is a single fundamental entity - a complex one, that in current language we might call an entangled quantum system in an algebraically very distinctive state, but still a single entity, in a way that a pile of unentangled atoms would not be.
Being a single entity means that it can enter directly into whatever fundamental causal relations are responsible for physical dynamics. Being that entity, from the inside, means having the sensations, thoughts, and desires that you do have; described mathematically, that will mean that you are an entity in a particular complicated, formally specified state; and physically, the immediate interactions of that entity would be with neighboring parts of the brain. These interactions cause the qualia, and they convey the "will".
That may sound strange, but even if you believe in a mind that is material but non-fundamental, it still has to work like that or else it is causally irrelevant. So when you judge the idea, remember to check whether you're rejecting it for weirdness that your own beliefs already implicitly carry.
So you're taking the existing causal graph, drawing a box around all the interactions that happen inside a brain, and saying that everything inside the box counts as one thing.
That's not simplification, it's just bad accountancy.
Where else would it be?
I'm saying that a brain is an environment where ideas can do interesting things (like reproducing themselves, mutating, splitting and recombining) comparable to the interesting things that started happening a very long time ago between amino acids and phospholipid membranes and assorted other organic chemicals which eventually resulted in the formation of brains. Any Turing-complete computer is also a sort of environment for ideas.
An idea outside an environment capable of supporting it does not do interesting things. It might be dormant, like a virus or bacterial spore, and colonize any less-hostile environment to which it's introduced. It might not. As yet, the only reliable way to distinguish between a dormant idea and a different arrangement of the same parts which does not constitute a dormant idea is to find an environment in which it will do interesting things.
For example, if you find a piece of baked clay with some scratch-marks in it, and want to know if they're cuneiform or just random scratches, you could show it to an archaeologist. The archaeologist looks at the tablet and compares it to prior knowledge about cuneiform - that is to say, transfers information about shape and coloration into her brain via the optic nerve and, once inside, drops them into the informational equivalent of a dish of agar. If anything interesting pops up, it's an idea. If not, either it's just noise, or it's an idea that the archaeologist can't figure out. There's no way to definitively prove the absence of potential ideas in a given information-bearing substrate.
If these disembodied qualia-properties don't help you make any actionable predictions beyond what physicalism could do, and their presence is unfalsifiable, I can't see any point to this debate. Is it a social-signaling contest of some sort?
Let's go back to your original statement:
OK, so according to you, we have concepts existing before and independently of consciousness, and we also have that consciousness is not a property that is objectively present (or else there'd be no need to appeal to the conceptual judgement of a brain, as a necessary cause of consciousness's existence). Both of these have to be true if you are to avoid circularity.
The second one already falsifies your account of consciousness. The difference between being conscious and not being conscious is not a matter of convention. It's an internal fact about you which is not affected by whether I am around to express opinions.
It sounds like you want the consciousness of a brain to depend on the conceptual judgements of that same brain, which is at least less abjectly dependent on the epistemology of outsiders. But it's still false. If you are conscious, you are conscious regardless of whatever opinions or concepts you have. Your conceptual capacities limit your possible conscious experience, in the sense that you can't consciously identify something as an X if you don't have the concept X, but whether or not you're conscious doesn't depend on how you are using (or misusing) your conceptual faculties at any time.
Just to clarify, by consciousness I mean awareness in all forms, not just self-awareness. What I said still applies to self-awareness as well as to awareness in general, but I thought I would make explicit that I'm not just talking about the sense of being a self. Even raw, self-oblivious sensory experience is a form of consciousness.
Maybe my very latest comments will clear things up a little. The immediate problem with physicalism is that reality contains qualia and physicalism doesn't. In a reformed physicalism that does contain qualia, they would have causal power.
Ah, so we're arguing over definitions.
Let's say you take an organism capable of receiving and interpreting information in the form of light, such as e.g. a ferret with working eyes and a visual cortex. Duplicate it with arbitrary precision, keep one of the copies in a totally lightless box for a few minutes and shine a dazzling but nondamaging spotlight on the other for the same period of time. Then open the box, shut off the spotlight, and show them both a picture.
The ferret from the box would see blindingly intense light, gradually fading in to the picture, which would seem bright and vivid. The ferret from the spotlight would see near-total darkness, gradually fading in to the picture, which would seem dull and blurry. Same picture, very different subjective experience, but it's all the result of physiological (mostly neurological) processes that can be adequately explained by physicalism.
Does the theory of qualia make independently-verifiable predictions that physicalism cannot? Or, if the predictions are the same, is it somehow simpler to describe mathematically? In the absence of either of those conditions, I am forced to consider the theory of qualia needlessly complex.
Or three quarks making a proton.
Suppose on Tuesday I perceive object O as red.
For labeling convenience, I'm going to start referring to my subjective experience of that perception as <red>. In other words, on Tuesday I experience O as <red>.
If I've understood you, you claim the <red> is due in part to color qualia in some way associated with O, which are distinct from the set of things happening inside my skull.
So, OK, assuming that, some questions.
I assume we agree that if I suddenly become color-blind, I might suddenly stop experiencing <red>. Do you assert that in that case the <red>-causing qualia continue to exist, I just stop experiencing them? (I would say something analogous about photons and perception, for example, if I suddenly lose my eyes.) Or do you assert that they stop existing? Or something else?
Either way: is that assertion something someone has confirmed in some way, or is it a purely theoretical prediction?
I assume we agree that if I suddenly manifest synesthesia -- say, due to a stroke -- I might also start experiencing a honking car horn as <red>. I assume you would therefore say that there must be <red>-causing qualia present, since my brain is unable to construct <red> on its own. Do you assert that the <red>-causing qualia were always present, and I've only just become able to perceive them? Or that they became present when I had the stroke, but not previously? Or something else?
Again: is that assertion something confirmed or theoretical?
No. I think that in reality, <red> is in the head. But our current physical ontology contains no such entity. That is why I say that if you accept our current physical ontology, you're either an eliminativist or a dualist.
I'm not in the least bit interested in the labels. But yes, if we're agreed that <red> is constructed by my brain, rather than being a property of my environment, then I don't understand what grounds you have for believing that <red> isn't explicable by entities in our current physical ontology.
Just imagine if you were having a discussion with someone who said that the world is made of numbers. And you picked up a rock and said, so, this rock is made of numbers? And they said, sure. And you said, that's absurd. How could a rock be equal to 1+1, for example? They're completely different kinds of things. And they went off on a riff about how science has shown that all is number, and whenever you tried to point out the non-numerical aspects of reality, they'd just subsume that back into the all-is-number reductionism, and they'd stubbornly insist that, even if the rock was not equal to 1+1, it might be equal to some other numbers, and besides, what other sort of things could there be, besides numbers?
For me, the idea that <red> is identical to some arrangement of particles in space is just like saying that 1+1 is a rock. The gulf between the nature of the allegedly identical entities is so great that the problem with the assertion ought to be obvious. In a sprinkling of point objects throughout space, where is the color? It's really that simple. It's just not there. It's not intrinsically there, anyway. You might propose that redness is a property of certain special configurations, but when you say that, you've embarked upon a form of dualism, property dualism. It's a dualism because on the one side, you have properties which are intrinsic to a geometrically defined situation, like distances and angles and shapes; and on the other side, you have properties which are logically independent of the geometric facts and have to be posited separately. For example, the existence of color experiences, or indeed any kind of experiences, in a brain.
In other words, the onus is on you to explain just what you think the connection is between arrangements of particles in space (e.g. a brain), and experiences of color. I have my own answer, but I want to hear yours first.
You won't find my answer interesting, but since you asked: I think experiences of color are among the states that particles in space can get into, just as the impulse to blink is a state particles in space can get into, just as a predisposition to generate meaningful English but not German sentences is a state that particles in space can get into, just as an appreciation for 17th-century Romanian literature is a state that particles in space can get into, just as a contagious head cold is a state that particles in space can get into. (Which is not to say that all of those are the same kinds of states.)
We can certainly populate our ontologies with additional entities related to those various things if we wish... color qualia and motor-impulse qualia and English qualia and German qualia and 17th-century Romanian literary qualia and contagious head cold qualia and so forth. I have no problem with that in and of itself, if positing these entities is useful for something.
But before I choose to do so, I want to understand what use those entities have to offer me. Populating my ontology with useless entities is silly.
I understand that this hesitation seems to you absurd, because you believe it ought to seem obvious to me that arrangements of matter simply aren't the kind of thing that can be an experience of color, just like it should seem obvious that numbers aren't the kind of thing that can be a rock, just as it seems obvious to Searle that formal rules aren't the kind of thing that can be an understanding of Chinese, just as it seemed obvious to generations of thinkers that arrangements of matter aren't the kind of thing that can be an infectious living cell.
These things aren't, in fact, obvious to me. If you have reasons for believing any of them other than their obviousness, I might find those reasons compelling, but repeated assertions of their obviousness are not.
An arrangement of particles in space can embody a blink reflex with no problems, because blinking is motion, and so it just means they're changing position in space.
Generating meaningful sentences - here we begin to run into problems, though not so severe as the problem with color. If the sentences are understood to be physical objects, such as sequences of sound waves or sequences of letter-shapes, then they can fit into physical ontology. We might even be able to specify a formal grammar of allowed sentences, and a combinatorial process which only produces physical sentences from that grammar. But meaning per se, like color, is not a physical property as ordinarily understood. (I know I'll get into extra trouble here, because some people are with me on the color qualia being a problem, but believe that causal theories of reference can reduce meaning to a conjunction of known physical properties. However, so far as I can see, intrinsic meaning is a property only of certain constituents of mental states - the meaning of sentences and all other intersubjective signs is not intrinsic and derives from a shared interpretive code - and the correct ontology of meaning is going to be bound up with the correct ontology of consciousness in general.)
Anyway, you say it's not obvious to you that "arrangements of matter simply aren't the kind of thing that can be an experience of color". Okay. Let's suppose there is an arrangement of matter in space which is an experience of color. Maybe it's a trillion particles in a certain arrangement executing a certain type of motion. Now, we can think about progressively simpler arrangements and motions of particles - subtracting one particle at a time from the scenario, if necessary... progressively simpler until we get all the way back to empty space. Somewhere in that conceptual progression we stopped having an experience of color there. Can you give me the faintest, slightest hint of where the magic transition occurs - where we go from "arrangement of particles that's an experience of color" to "arrangement of particles that's not an experience of color"?
I could also simply ask for you to indicate where in the magic arrangement of particle the color is. That is, assuming that you agree that one aspect of the existence of an experience of color is that something somewhere actually is that color. If it turns out that, according to you, brain state X is an experience of <red> only because the brain in question outputs the word "red" when queried, or only because a neural network somewhere is making the categorization "red" - then that is eliminativism. There's no actual <red>, no actual color, just color words or color categories.
The reason it is obvious that there is no color inherently inhabiting an arrangement of particles in space is because it's easy to see what the available ontological ingredients are, and it's easy to see what you can and cannot make by combining them. If we include dynamics and a notion of causality, then the ingredients are position, time, and causal dependence. What can you construct from such ingredients? You can make complicated structures; you can make complicated motions; you can make complicated causal dependencies among structures and motions. As you can see, it's no mystery that such an ontological scheme can encompass something like a blink reflex, which is a type of motion with a specified causal dependency.
With respect to the historical case of vitalism, it's interesting that what the vitalists posited was a "vital force". That's not an objection to the logical possibility of reducing life, and especially replication, to matter in motion. They just didn't believe that the known forces were capable of producing the right sort of motion, so they felt the need to postulate a new, complicated form of causal interaction, capable of producing the complexly orchestrated motion which must be occurring for living things to take shape. As it turned out, there was no need to postulate a special vital force to do that; the orchestration can be produced by the same forces which are at work in nonliving matter.
I'm emphasizing the way in which the case of vitalism differs from the case of qualia, because it is so often cited as a historical precedent. The vitalists - at least, the ones who talked about vital forces - were not saying that life is not material. They just postulated an extra force; in that respect, they were proposing only a conservative extension to the physical ontology of their time. But the observation that consciousness presents a basic ontological problem, in a universe consisting of nothing but matter in motion through space, has been around for a very long time. Democritus took note of this objection. I think Leibniz stated it in a recognizably modern form. It is an old insight, and it has not gone away just because the physical sciences have been so successful. Celia Green writes that this success actually sharpens the problem: the clearer our conception of material ontology and our causal account of the world becomes, the more obvious it becomes that this concept and this account do not contain the "secondary qualities" like your <red>.
Even at the dawn of modern physical science, in the time of Galileo, there was some discussion as to how these qualities were being put aside, in favor of an exclusive focus on space, time, motion, extension. It's quite amazing that from humble beginnings like Kepler's laws, we've come as far as quantum mechanics, string theory, molecular biology, all the time maintaining that exclusion. Some new ontological factors did enter the set of ingredients that physical ontology can draw upon, especially probability, but those elementary sensory qualities remain absent from the physical conception of reality. The 20th-century revolution in thought regarding information, communication, and computation goes just a little way towards bringing them back, but in the end it's nowhere near enough, because when you ask, what are these information states really, you end up having to reduce them to statistical properties of particles in space, because that's still all that the physical ontology gives you to work with.
I'm probably an idiot for responding at such length on this topic, because all my experience to date suggests that doing so changes nothing fundamentally. Some people get that there's a problem, but don't know how to solve it and can only hope that the future does so, or they embrace a fuzzy idea like emergence dualism or panpsychism out of intellectual desperation. Some people don't get that there's a problem - don't perceive, for example, that "what it feels like to be a bat" is an extra new property on top of all the ordinary physical properties that make up a bat - and are happy with a philosophical formula like "thought is computation".
I believe there is a problem to be solved, a severe problem, a problem of the first order, whose solution will require a change of perspective as big as the one which introduced us to the problem. Once, we had naive realism. The full set of objects and properties which experience reveals to us were considered equally real. They all played a part in the makeup of reality, to which the human mind had a partial but mysteriously direct access. Now, we have physics; ontological atomism, plus calculus. Amazingly, it predicts the behavior of matter with incredible precision, so it's getting something right. But mind, and everything that is directly experienced, has vanished from the model of reality. It hasn't vanished in reality; everything we know still comes to us through our minds, and through that same multi-sensory experience which was once naively identified with the world itself, and which we now call conscious experience. The closest approximation within the physical ontology to all of that is computation within the nervous system. But when you ask what neural computation are, physically, it once again reduces to matter in motion through space, and the same mismatch between the apparent character of experience, and the physical character of the brain, recurs. Since denying that experience does have this distinct character is false and therefore hopeless, the only way out must be to somehow reconceive physical ontology so that it contains, by construction, consciousness as it actually is, and so that it preserves the causal structural relations (between fundamental entities whose inner nature is opaque and therefore undetermined by the theory) responsible for the success of quantitative predictions.
I imagine my manifesto there is itself opaque, if you're one of those people who don't get the problem to begin with. Nonetheless, I believe that is the principle which has to be followed in order to solve the problem of consciousness. It's still only the barest of beginnings, you still have to step into darkness and guess which way to turn, many times over, in order to get anywhere, and if my private ideas about how to proceed are right, then you have to take some really big leaps in the darkness. But that's the kernel of my answer.
Your remove-an-atom argument also disproves the existence of many other things, such as heaps of sand.
Let's try to communicate through intuition pumps:
Suppose I built a machine that could perceive the world, and make inferences about the world, and talk. Then of course (or with some significant probability), the things it directly perceives about the world would seem fundamentally, inextricably different from the things it infers about the world. It would insist that the colors of pixels could not consist solely of electrical impulses - they had to be, in addition, the colors of pixels.
Stolen from Dennet: You are not aware of your qualia, only of relationships between your qualia. I could swap <red> and <blue> in your conscious experience, and I could swap them in your memories of conscious experience, and you wouldn't be able to tell the difference - your behavior would be the same either way.
Two meditations on an optical illusion: I heard, possibly on lesswrong, that in illusions like this one: http://www.2dorks.com/gallery/2007/1011-illusions/12-kanizsatriangle.jpg your edge-detecting neurons fire at both the real and the fake edges.
Doesn't that image look exactly like neurons detecting edges between neurons detecting white and neurons detecting like should look like?
Doesn't the conflict between a physical universe and conscious experience feel sort of like the conflict between uniform whiteness and edgeness?
My latest comment might clarify a few things. Meanwhile,
No-one's telling me that a heap of sand has an "inside". It's a fuzzy concept and the fuzziness doesn't cause any problems because it's just a loose categorization. But the individual consciousness actually exists and is actually distinct from things that aren't it, so in a physical ontology it has to correspond to a hard-edged concept.
Consider Cyc. Isn't one of the problems of Cyc that it can't distinguish itself from the world? It can distinguish the Cyc-symbol from other symbols, but only in the same way that it distinguishes any symbol from any other symbol. Any attempt to make it treat the Cyc-symbol really differently requires that the Cyc-symbol gets special treatment on the algorithmic level.
In other words, so long as we talk simply about computation, there is nothing at all to inherently make an AI insist that its "experience" can't be made of physical entities. It's just a matter of ontological presuppositions.
As I've attempted to clarify in the new comment, my problem is not with subsuming consciousness into physics per se, it is specifically with subsuming consciousness into a particular physical ontology, because that ontology does not contain something as basic as perceived color, either fundamentally or combinatorially. To consider that judgement credible, you must believe that there is an epistemic faculty whereby you can tell that color is actually there. Which leads me to your next remark--
--and so obviously I'm going to object to the assumption that I'm not aware of my qualia. If you performed the swap as described, I wouldn't know that it had occurred, but I'd still know that <red> and <blue> are there and are real; and I would be able to tell the difference between an ontology in which they exist, and an ontology in which they don't.
A neuron is a glob of trillions of atoms doing inconceivably many things at once. You're focusing on a few of the simple differential sub-perceptions which make up the experience of looking at that image, associating them in your mind with certain gross changes of state in that glob of atoms, and proposing that the experience is identical to a set of several such simultaneous changes occurring in a few neurons. In doing so, you're neglecting both the bulk of the physical events occurring elsewhere in the neurons, and the fundamental dissimilarity between "staring at a few homogeneous patches of color" and "billions of ions cascading through a membrane".
It's more like the difference between night and day. It is possible to attain a higher perspective which unifies them, but you don't get there by saying that day is just night by another name.
Degree-of-existence seems likely to be well-defined and useful, and may play a part in, for example, quantum mechanics.
However, my new response to your argument is that, if you're not denying current physics, but just ontologically reorganizing it., then you're vulnerable to the same objection. You can declare something to be Ontologically Fundamental, but it will still mathematically be a heap of sand, and you can still physically remove a grain. We're all in the same boat.
Do you think Cyc could not be programmed to treat itself different from others without use of a quantum computer? If not, how can you make inferences about quantum entanglement from facts about our programming.
Does Cyc have sensors or something? If it does/did, it seems like it would algorithmically treat raw sensory data as separate from symbols and world-models.
Is there anything to inherently prevent it from insisting that? Should we accept our ontological presuppositions at face value?
No you wouldn't. People can't tell the difference between ontologies any more then math changes if you print its theorems in a different color. People can tell the difference between different mathematical laws of physics, or different arrangements of stuff within those laws. What you notice is that you have a specific class of gensyms that can't have relations of reduction for other symbols, or something else computational. Facts about ontology are totally orthogonal to facts about things that influence what words you type.
My consciousness is a computation based mainly or entirely on regularities the size of a single neuron or bigger, much like the browser I'm typing in is based on regularities the size of a transistor. I wouldn't expect to notice if my images were, really, fundamentally, completely different. I wouldn't expect to notice if something physical happened - the number of ions was cut by a factor of a million and made the opposite charge, but it the functions from impulses to impulses computed by neurons were the same.
Uniform color and edgeness are as different as night and day.
A few thoughts in response:
I agree with you that if my experience of red can't be constructed of matter, then my understanding of a sentence also can't be. And I agree with you that we don't have a reliable account of how to construct such things out of matter, and without such an account we can't rule out the possibility that, as you suggest, such an account is simply not possible. I agree with you that this objection to physicalism has been around for a long time.
I agree with you that insofar as we understand vitalism to be an account of how particular arrangements of matter move around, it is a different sort of thing from the kind of "sentientism" you are talking about. That said, I think that's a misrepresentation of historical vitalism; I think when the vitalists talked about elan vital being the difference between living and unliving matter, they were also attributing sentience (though not sapience) to elan vital, as well as simple animation.
I don't equate the experience of red with the tendency to output the word "red" when queried, both in the sense that it's easy for me to imagine being unable to generate that output while continuing to experience red, and in the sense that it's easy for me to imagine a system that outputs the word "red" when queried without having an experience of red. Lexicalization is neither necessary nor sufficient for experience.
I don't equate the experience of red with categorization... it is easy to imagine categorization without experience. It's harder to imagine experience without categorization, though. Categorization might be necessary, but it certainly isn't sufficient, for experience.
Like you, I can't come up with a physical account of sentience. I have little faith in the power of my imagination, though. Put another way: it isn't easy for me to see what one can and can't make out of particles. But I agree with you that any such account would be surprising, and that there is a phenomenon there to explain. So I think I fall somewhere in between your two classes of people who are a waste of time to talk to: I get that there's a problem, but it isn't obvious to me that the properties that comprise what it feels like to be a bat must be ontologically basic and nonphysical. Which I think still means I'm wasting your time. (I did warn you in the grandparent comment that you won't find my answer interesting.)
If it turns out that a particular sensation is perfectly correlated with the presence of a particular physical structure, and that disrupting that structure always triggers a disruption of the sensation, and that disrupting the sensation always triggers a disruption of the structure... well, at that point, I'm pretty reluctant to posit a nonphysical sensation. Sure, it might be there, but if I posit it I need to account for why the sensation is so tightly synchronized with the physical structure, and it's not at all clear that that task is any simpler than identifying one with the other, counterintuitive as that may be.
At the other extreme, if the nonphysical structure makes a difference, demonstrating that difference would make me inclined to posit a nonphysical sensation. For example, if we can transmit sensation without transmitting any physical signal, I'd be strongly inclined to posit a nonphysical structure underlying the sensation. Looking for such a demonstrable difference might be a useful way to start getting somewhere.
Perhaps we are closer to mutual understanding than might have been imagined, then. A crucial point: I wouldn't talk about the mind as something "nonphysical". That's why I said that the problem is with our current physical ontology. The problem is not that we have a model of the world in which events outside our heads are causally connected to events inside our heads via a chain of intermediate events. The problem is that when we try to interpret physics ontologically (and not just operationally), the available frameworks are too sparse and pallid (those are metaphors of course) to produce anything like actual moment-to-moment experience. The dance of particles can produce something isomorphic to sensation and thought, but not identical. Therefore, what we might think of as a dance of particles actually needs to be thought of in some other way.
So I'm actually very close in spirit to the reductionist who wants to think of their experience in terms of neurons firing and so forth, except I say it's got to be the other way around. Taken literally, that would mean that we need to learn to think of what we now call neurons firing, as being fundamentally - this - moment-to-moment experience, as is happening to you right now. Except, the physical nature of whole neurons I don't believe plausibly allows such an ontological reinterpretation. If consciousness really is based on mesoscopic-level informational states in neurons, then I'd favor property dualism rather than the reverse monism I just advocated. But I'm going for the existence of a Cartesian theater somewhere in the brain whose physical implementation is based on exact quantum states rather than collective coarse-grained classical ones, quantum states which in our current understanding would look more algebraic than geometric. And the succession of abstract algebraic state transitions in that Cartesian theater is the deracinated mathematical description of what, in reality, is the flow of conscious experience.
If that is the true interior reality of one quantum island in the causal network of the world, it might be anticipated that every little causal nexus has its own inside too - its own subjectivity. The non-geometric, localized, algebraic side of physics would turn out to actually be a description of the local succession of conscious states, and the spatial, geometric aspect of physics would in fact describe the external causal interactions between these islands of consciousness. Except I suspect that the term consciousness is best reserved for a very rare and highly involuted type of state, and that most things count as islands of "being" but not as islands of "experiencing" (at least, not as islands of reflective experiencing).
I should also distinguish this philosophy from the sort which sees mind wherever there is distributed computation - so that the hierarchical structure of classical interaction in the world gets interpreted as a set of minds made of minds made of minds. I would say that the ontological glue of individual consciousness is not causal interaction - it's something much tighter. The dependence of elements of a state of consciousness on the whole state of consciousness is more like the way that the face of a cube is part of the cube, though even that analogy is nowhere near strong enough, because the face of a cube is a square and a square can have independent existence, though when it's independent it's no longer a face. However we end up expressing it, the world is fundamentally made of these logical ontological unities, most of which are very simple and correspond to something like particles, and a few of which have become highly complex - with waking states of consciousness being extremely complex examples of these - and all of these entities interact causally and quasi-locally. These interactions bind them into systems and into systems of systems, but systems themselves are not conscious, because ontologically they are multiplicities, and consciousness is always a property of one of those fundamental physical unities whose binding principle is more than just causal association.
An ontology of physics like that is one where the problem of consciousness might be solved in a nondualistic way. But its viability does seem to require that something like quantum entanglement is found to be relevant to conscious cognition. As I said, if that isn't borne out, I'll probably fall back on some form of property dualism, in which there's a many-to-one mapping between big physical states (like ion concentrations on opposite sides of axonal membranes) and distinct possible states of consciousness. But physical neuroscience has quite a way to go yet, so I'm very far from giving up on the monistic quantum theory of mind.
So, getting back to my original question about what your alternate ontology has to offer...
If I'm understanding you (which is far from clear), while you are mostly concerned with being ontologically correct rather than operationally useful, you do make a falsifiable neurobiological prediction having something I didn't follow to do with quantum entanglement.
Cool. I approve of falsifiable predictions; they are a useful thing that a way of thinking about the world can offer.
Anything else?
I find this argument irresistably compelling, and would appreciate a post or a private message letting me know what your answer is. I don't have one; it's all I can do here to notice that I am confused.
I think you need to be taken outside and shot...
...
...j/k.
It's just that over recent years I've spent quite a long time arguing with people educated principally in philosophy, who hate Dennett and think his version of materialism is absurd (or at least that it's manifestly wrong), and think it's absolutely essential to go around saying things like 'all we know about are correlations between body and mind'.
It's sort-of interesting/refreshing for me to arrive here, with a bunch of people who are (I assume) educated principally in computer science (with perhaps a few mathies and physicists), who are almost unanimously Dennett fans, think that functionalism is just blindingly obvious, that 'zombies' are blindingly obviously impossible, that it's blindingly obvious that the 'Systems Reply' is correct, that anything we build capable of passing the (full) Turing Test would have to be conscious etc.
The ones who don't 'get it' - that at the core of Dennett's view there's the difficult-to-swallow idea that there isn't a 'fact of the matter' as to whether a being is conscious and if so what it's conscious of - can at least fall back on a Greg Egan-style view of consciousness which is identical insofar as it agrees that the issues above are 'blindingly obvious'. (That's the other thing: the people here have actually read Greg Egan - woohoo.)
I can see you have a more in common with the philosopher-types than the locals. And actually, in your interpretation of Dennett I think there's a mistake - one I've seen elsewhere:
You think that in abolishing the 'Cartesian theater' he is ipso facto abolishing phenomenal awareness, but this simply doesn't follow. What he's abolishing is the idea that all of the 'bits' of a person's awareness are present 'together' in a single sharply-defined 'moment', such that there are well-defined answers to questions like "am I seeing a moving dot or a static one?" which would resolve the "Orwellian/Stalinesque" dilemma.
Even after the Cartesian theater is abolished, you can still be a dualist as long as you're prepared to give ground on things like 'the unity of consciousness', and admit that the various parts of the mindscape are slightly removed from each other - not as far removed as the mind of a different person altogether, or even as far as the two hemiminds of a split-brain patient, but certainly not bundled together in a brilliant 'point' of 'inner light'.
I'd just come back as a zombie.
That sums it up well. Next up, let's consider other startling possibilities, such as: there isn't a fact of the matter as to whether you're reading this sentence, there isn't a fact of the matter as to whether this planet exists, there isn't a fact of the matter as to whether there is a fact of the matter as to whether a being is conscious...
Yeah but come on... you always-a-fact-of-the-matter-ists have some startling things to think about too, like The Exact Moment When You First Became Conscious, and the Infinitely Precise Line one can draw across the phylogenetic tree demarcating species whose members are (or may be) conscious and those which never are.
(Afterthought: Or are you some kind of panpsychist? Then your startling possibilities incude the minds of rocks...)
See, it's not so hard! You just have to take the idea seriously, and stick with it. You might even have a talent for this. And here I was thinking that my labor here was in vain.
I believe Eliezer doesn't agree with that last one, and has talked about building an AI who isn't conscious.
Also, consider the following hypothetical: I get really drunk and/or take Ambien and black out at 2 am. I have no conscious experience or memory of the time between 2 am and 3 am, but during that time you have a (loud and drunken) conversation with me. Or maybe in my drunken state I sit at my computer and manage to instant message without being conscious of it, and the person at the other end is convinced I'm human and not a computer program. Counterexample?
Well, I think we can all agree that it's possible for a non-conscious person (or program or whatever) to be mistaken for a conscious being.
However, there are several objections I can make to this scenario being considered a counterexample:
(1) How do you know you're not conscious? Just because you don't remember it the next day doesn't mean you don't have any awareness at the time.
(2) In the Turing test the judge is supposed to be 'on the look-out' for which of its two subjects seems less able to respond adequately to their questions. And one of the subjects is presumed to be a healthy, sober human. So unless you think the judge would be unable to distinguish a drunken, unconscious conversation from a normal, sober one, you would presumably fail the Turing test.
What exactly do you take the purpose of an ontology to be? If you have a scientific theory whose predictions hit the limit of accuracy for predicted experience why do you need anything in your ontology beyond the bound variables of the theory?
An ontology is a theory about what's there. The attributes of experience itself, like color, meaning, and even time, have been swept under a carpet variously labeled "mind", "consciousness", or "appearance", while the interior decorators from Hard Science Inc. (formerly trading as the Natural Philosophy Company) did their work. We have lots of streamlined futuristic fittings now, some of them very elegant. But they didn't get rid of the big lump under the carpet. The most they can do is hide it from view.
We don't have access to "what is there". What we have are sensory experiences. Lots of them! Something is generating those experiences and we would like to know what we will experience in the future. So we guess at the interior structure of the experience generator and build models that predict for us what our future experiences will be. When our experiences differ from expected we revise the model (i.e. our ontology). This includes modeling the thing that we are which improves our predictions of our own experiences and our experiences of what other humans say are their experiences. One thing humans report is the experience is seeing color. So we need to explain that. One thing humans report is the experience is self-awareness so we have to explain that etc. You seem to want to reify the sensory experiences themselves just because they look different in our model than in our experience. But the model isn't supposed to look like our experience it is supposed to predict it. You're making a category error. Presumably you know this and think the problem is the categories. But you need to motivate your rejection of the categories. All I want are predictions and I've been getting them, so why should I reject this model?
But lots of scientists study these things! Last semester I learned all about auditory and visual perception. There is a lot we don't know which is why they're still working on it.
So we know that whatever is there must include those sensory experiences. They themselves are part of reality.
Most models of reality are partial models that implicitly presuppose some untheorized notion of experience in the model-user. Medicine and engineering aren't especially focused on the fact that doctors and engineers encounter the world, like everyone else, through the medium of conscious experience.
But there are two types of explanatory enterprise where conscious experience does become explicitly relevant. One is any theory of everything. The other is any science which does take experience as its subject matter. In the latter case, scientists will explicitly theorize about the nature of experience and its relationship to other things. In the former case, a theory of everything must take a stand on everything, including consciousness, even if only to say "it's made of atoms, like everything else".
So some part of these models is supposed to look like experience. However, as I have been saying elsewhere, nothing in physical ontology looks like an experience; and the sciences of consciousness so far just construct correlations between "physics" (i.e. matter) and experience. But they must eventually address the question of what an experience is.
Nice essay! I'm not yet won over by the suggestion in your final paragraph, but it's intriguing.
Phil writes, "Nice essay!"
Is there something in Mitchell's essay (comment) that Mitchell has not already said on this site 30 times or did you just like the way he phrased it this time?