= 27d567bc2d48d9f40e9ccc20ea370b15)
Jonathan_Lee comments on Consciousness - 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 (221)
Projecting the ontology of your (flawed) internal representations onto reality is a bad idea. "Doing a Dennet" is also not dealt with, except by incredulity.
This is simply not the case. The fact that we can compare two stimuli more accurately than we can identify a stimuli merely means that internally we represent reality with lesser fidelity than our senses theoretically can achieve. On a reductionist view at most you've established "greater than" and "round to nearest" are implemented in neurons. You do not need to have colour.
Let's unpack "blueness". It's a property we ascribe to objects, yet it's trivial to "concieve of" blueness independent of an object. Neurologically, we process colour, motion, edge finding and so in in parallel; the linking of them together occurs at a higher level. Furthermore the brain fakes much of the data, giving the perception of colour vision, for example, in regions of the visual field where no ability to discriminate colour exists, and cases of blindness with continued concious perception of colour.
Brains compress input extensively; it would be crass to worry about the motion of every spot on a leopard separately - block them up as a single leopard. Asserting that the world must fit with our hallucination of reality lets you see things that are marginally visible, and get by with far worse sensory apparatus than needed. Cue optical illusions: this, this and this, for example. Individual shades don't exist as you want them to.
It is absurdly clear that the map your brain makes does not correspond to either the territory of your direct sense perception (at the retina) or reality. On precisely what basis do you assert to project from the ontology of a bad map to the territory?
"Blue" is a referent to properties of internal representations, which is translatable across multiple instances of primate brains. You say "X is blue", and I can check my internal representation of X to see whether I would categorise it as "blue". This does not require "blue" to be fundamental in ontology. There isn't a "blue thing" in physics, nor should there be. "Blue" existing means simply that there are things which this block of wetware puts in some equivalence class.
Lets move on to computation:
Again, you seem project from an internal map of your own brain to the territory. Simply because I can look at a computer at multiple levels, say: Starting Excel, API calls, Machine instructions, microcode, functional units on the CPU, adders/multipliers/whatever on the CPU, logic gates, transistors, current flows or probability masses in the field of electrons, does not in principle invalidate any of the above views as correct views of an operating computer. The observer dependence isn't an issue if (modulo translation/equivalence classes for abstraction between languages) they all give the same function or behaviour. You can block things up as many low level behaviours or a smaller number of high level ones; this doesn't invalidate a computational view. What is the computation implemented by starting Excel? What details do you care about? It doesn't matter to a functionalist, as the computations are equivalent, albeit in different languages or formalisms.
The critique of aboutness is similar to your issues over colour. You percieve "X is about Y" and thus assume it to be ontologically fundamental. Semantic content is a compressed and inaccurate rendition of low level states: Useful for communicating and processing if you don't care about the details. Indeed the only reason we care about this kind of semantics is that our own wetware implements theory-of-mind directly. Good idea for predicting cognitive agents; not neccessarily a true statement about the world. The "Y" that "X" is "about" is another contraction - an infered property of a model.
"Time" is as flexible as your neural architecture wants it to be. Causality is a good idea, for Darwinian reasons, but people's perception of the flow of time is adjustable. I will point out that your senses imply strongly that the world is a 2D surface. Have you ever been able to see behind an object without moving your head? I haven't either, therefor clearly this 3D stuff is bunkum - the world is a flat plane and I directly percieve part of one side of it. Ditto time. Causality limits the state of a cognitive thing to be dependent on its previous states and its light cone at this point in space-time, and you percieve time to flow because you can remember previous brain states, and depending on them (compressed somewhat) is good for survival.
And now for unity of conciousness. It isn't unitary. Multiple personality, dissociative disorders, blindsight, sleepwalking, alien hand, need I go on? I percieve my own representation of reality to be unitary; I know for a fact that it's half made up. You claim that the individual issues "just can't" be the whole story. Why? Personal incredulity isn't an argument. The brain in the skull you call yours isn't just running a single cognitive entity. You move before even realising "you" were going to; you are unconcious of breathing until you decide to be. Why is a unitary conciousness fundamental? Why isn't it just a shortcut to approximate "you" and others in planning the future and figuring out the present?
Re. blueness: Mitchell is talking about qualia. Google the hard problem of consciousness.
Just a note - I don't disagree with your point; but the claim that we can't discriminate color in our peripheral vision is simply false. I've done some informal experiments with this, because I was puzzled that textbooks say that our peripheral vision is primarily due to rods, which can't detect color; yet I see color in my peripheral vision.
If I stand with my nose and forehead pressed against the wall, holding a stack of shuffled yellow and red sheets of origami paper behind my back, close my eyes, and then hold one sheet up in each of my outstretched arms, and open my eyes, so that the sheets are each 90 degrees out from my central vision and I see them both at the same time, I can distinguish the two colors 100% of the time.
There's a serious problem with resolution; but color doesn't seem to be affected at all in any way that I can detect by central vs. peripheral vision.
Of the same apparent intensity to a rod? If they're not, you'll guess correctly based on apparent brightness, and your brain fills in the colour based on memory of which colours of paper are around.
There are low levels of cones out to the periphery, but of such level as to be unreliable sources. For example, this notes that some monochromatic light is misidentified peripherally but not foevally, and that frequency discrimination drops by a factor of 50 or so.
Would be interesting to see you do this on video with a second person shuffling and displaying the cards.
Noting Jonathan_Lee's remarks, a suggestion for an experiment: place a monitor in the peripheral vision of the experimental which, at regular intervals, shows a random RGB color. The subject is to press a key indicating perceived color (e.g. [R]ed, [Y]ellow, [B]lue, [O]range, [G]reen, [P]urple, [W]hite, [B]lack) each time the color changes (perhaps an audio cue?). Compare results to same experiment with monitor directly in front.
It seems you agree that colors, the flow of time, meanings, and the unity of experience all appear to be there. The general import of your remarks is that reality isn't actually like that, it only appears to be like that. You state how things are, and then you state what's happening in the brain in order to create certain appearances. Color and time are external appearances, meaning and unity are internal appearances.
Some of what you say about the imperfections of conscious representations is not an issue for me. The fidelity of the mapping between external states and conscious states only has an incidental bearing on the nature of the conscious states themselves. Whether color is sometimes hallucinated is not the issue. Whether a color is nothing but an equivalence class is the issue.
In this regard I have observed a number of positions taken. Some people are at the stage of saying, color is a neural classification and I don't see any further problem. Some people say that color is how it "feels" to make such a classification. Since Dennett takes care to deny that there is anything that is actually colored, even in the mind, and that there are only words, dispositions to classify, and so forth, he arguably wishes to deny even that there is a "color feeling", though it's hard to be sure.
My position is very simple. All these things (color, time, meaning, unity) exist in consciousness; which means that they exist in at least one part of reality. The elements and the modes of combination offered by today's scientific ontology do not suffice to generate them. Therefore, today's scientific ontology is wrong, incomplete, however you want to put it.
So if we are to have a discussion, you need to say less about the imperfections of consciousness as a medium of representation, and more about the medium itself. Do you agree that color, time, meaning, unity exist in consciousness? If so, can you identify the physical or computational property which supposedly corresponds to, or is identical to, "appearing to experience" each of these various phenomena?
External only in that wetware is modelling something outside of the skull, rather than it's internal state. The intent was to state that merely because you perceive reality along certain ontological lines does not imply that reality has the same ontology.
This should be particularly obvious when your internal sense fails to correspond to reality; if conscious states are an imperfect guide to external states then why should the apparent ontology of consciousness be accurate?
None of which you refute here or in the OP, especially those who deny that "blueness" is a veridical property of reality.
No; it means that something referencing them exists in some part of reality (your skull). An equivalence relation; an internal tag that this object is blue.
To counter the realism, consider mathematicians, who consciously deal in infinite sets, or all theorems provable under some axioms (model theory). Just because something appears plainly to you does not mean it exists. Kant says it better than I can.
Not if you mean more than perception by consciousness. Even in perception, they're just the ontology imposed by our neurology, and have neural correlates that suffice.
Consciousness isn't prior to perception or action; it's after it. There isn't a homunculus in there for experience to "appear to". If anything, there's a compressed model of your own behaviour to which experience is fed into; that's the "you" in the primate - a model of that same primate for planning and conterfactual reasoning.
Let's suppose I have a hallucinatory perception of a banana. So, there's no yellow object outside my skull - we can both agree on that. It seems we also agree that I'm having a yellow perception.
But we part ways on the meaning of that. Apparently you think that even my hallucination isn't really yellow. Instead, there's some neural thing happening which has been tagged as yellow - whatever that means.
I really wonder about how you interpret your own experience. I suppose you experience colors just like I do, but (when you think about it) you tell yourself that what naively seems to be a matter of seeing a yellow object is actually experiencing what it's like to have a perception tagged as yellow. But how does that translate, subjectively? When you see yellow, do you tell yourself you're seeing the tag? Do you just semi-visualize a bunch of neurons firing in a certain way?
We went over this issue a bit in the previous discussion. My response (following Drescher) was: "To experience [yellow] is to feel your cognitive architecture assigning a label to sensory data."
As I elaborated:
The point being: I can't give a complete answer now, but I can tell you what the solution will look like. It will involve describing how a cognitive architecture works, then looking at the distinctions it has to make, then looking at what constraints these distinctions operate under (e.g. color being orthogonal to sound [unless you have synaesthesia], etc.), then identifying what parts of the process can access each other.
Out of all of that, only certain data representations are possible, and one of these (perhaps, hopefully, the only one) is the one with the same qualities as our perception of color. You know you're at the solution, when you say, Aha! If I had to express what information I receive, under all those constraints, that is what qualities it would need to have.
To that, you replied:
Though you object to the comparison, this is the same kind of error as demanding that there be a fundamental "chess thing" in Deep Blue. There is no fundamental color, just as there is no fundamental chess. There is only a regularity the system follows, compressible by reference to the concept of color or chess.
I am intrigued by your wording, here. I suppose I experience colors just like you do, but - when I think about it - I tell myself that what is, in fact seeing a yellow object is, in fact the same thing as experiencing what it's like to have a perception tagged as yellow. I believe these descriptions to be equivalent in the same sense that "breaking of hydrogen bonds between dihydrogen monoxide molecules, leading to those molecules traveling in near-independent trajectories outside the crystalline structure" is equivalent to "ice sublimating".
The relevant part of the optical cortex which fires on yellow objects has fired; the rest of your brain behaves as if there were a yellow banana out in front of it. "Tagging" seemed like the best high level term for it. A collection of stimuli are being collected together as an atomic thing. There's a neural thing happening, and part of that neural thing is normally caused by yellow things in the visual field.
The most obvious point where it has subjective import is when things change[1]. I probably experience colours as you do; when I introspect on colour, or time, I cannot find good cause to distinguish it from "visualising" an infinite set or a function. The only apparent different is that reality isn't under concious control. I don't assume that the naive ontology that is presented to me is a true ontology.
[1] There are a pair of coloured mugs (blue and purple) that I can't distinguish in my peripheral vision, for example. When I see one in my peripheral vision, it is coloured (blue, say); when I look at it directly, there is a period in which it is both blue and purple, as best I can describe, before definitively becoming purple. Head MRI's do this too.
Edit: The problem is that there isn't an easy way to introspect on the processes leading to perceptions; they are presented ex nihilo. As best I can tell, there's no good distinguisher of my senses from "experiencing what it's like to have a perception tagged as yellow"
Citation or he didn't say it. Daniel Dennett coined the phrase "greedy reductionism" - partially to emphasize that he does not deny the existence of color, consciousness, etc. Unless you know of some place where he reversed his position, I would argue that you have misinterpreted his remarks. My understanding is that his position is that color is an idiosyncratic property of the human visual perception system with no simple referent in physics, not no referent at all.
(I normally wouldn't make such a big deal of it, but Dennett is one of the major figures on the physicalist side of this debate, and a mischaracterization of his views impedes the ability of bystanders to perform a fair comparison.)
In Explaining Consciousness, chapter 2, p. 28, Dennett says there is no purple in the brain when we see purple. That may be what he means.
I also heard Dennett quoted as saying there is no such thing as qualia, allegedly in "The taboo of subjectivity", p. 139, which I don't have.
Here is a full quote that makes clear in exactly what sense he doesn't believe in qualia:
Originally in Quining Qualia, 1988, by Dennett, and quoted on Multiple-Drafts Model.
The Taboo of Subjectivity is a book by B. Alan Wallace. It appears that Dennett wrote a review for that work, but I couldn't find it online. Are you referring to that review, or to something else?
I see what you meant now. Dennett was quoted in Wallace's book, on p.139. Sorry for the misunderstanding.
The quote, with some context, is:
18 . Paul M. Churchland, 1990, Matter and Consciousness: A Contemporary Introduction to the Philosophy of Mind, p. 41 & 48.
19 . Daniel C. Dennett, 1991, Consciousness Explained, p. 74.
Belatedly:
I think the reference on p. 28 is pointing out that the brain doesn't turn purple (and a purple brain wouldn't help anyway, as there are no eyes in the brain to see the purple). The remainder of the page is extending the example to further elaborate the problem of subjective experience.
I cannot find the reference to qualia quoted in The Taboo of Subjectivity at all - p. 74 is before Dennett even defines qualia, and p. 374 does not have those exact words - only the conclusion of a thought experiment illuminating his rejection of the concept.
Thanks for the page number - I'll see if I can find it in my copy when I get home.
Upvoted. I searched Google for about 15 seconds looking for the quote and didn't find it, but I remember seeing or hearing Dennett say once how flabbergasted he is about being "oh yeah, that guy who thinks we don't see color."
He does not use the expression "color feeling", but here's a direct quote from Consciousness Explained, chapter 12, part 4:
He explicitly denies that there is any such thing as a "private shade of homogeneous pink" - which I would consider a reasonably apt description of the phenomenological reality. He also says there is something real, a "complex of dispositions". And, he also says that when we refer to color, we think we're referring to the former, but we're really referring to the latter. So, subjective color does not exist, but references to color do exist.
That still leaves room for there to be "appearances of pink". No actual pink, but also more than a mere belief in pink; some actual phenomenon, appearance, component of experience, which people mistakenly think is pink. But I see no trace of this. The thing which he is prepared to call real, the "complex of dispositions", is entirely cognitive (in the previous paragraph he refers to "innate and learned associations and reactive dispositions"). There is no reference to appearance, experience, or any other aspect of subjectivity.
Therefore, I conclude that not only does Dennett deny the existence of color (yes, I know he still uses the word, but he explicitly defines it to refer to something else), he denies that there is even an appearance of color, a "color feeling". In his account of color phenomenology, there are just beliefs about nonexistent things, and that's it.
The references to red together definitely form a physical network in my brain, right? I have a list of 10,000 things in my memory that are vividly red, some more vivid than others, and they're all potentially connected under this label 'red'. When that entire network is stimulated (say, by my seeing something red or imagining what "red" is), might I not also give that a label? I could call the stimulation of the entire network the "essence of red" or "redness" and have a subjective feeling about it.
I'm certain this particular theory about what "redness" occurs frequently. My question is, what's missing in this explanation from the dualist point of view? Why can't the subjective experience of red just be the whole network of red associations being simultaneously excited as an entity?
Above you wrote
So I guess I'm just asking, what's the further problem? (If you've already answered, would you please link to it?)
What are in those ellipses? In what you quote, I see that he's denying that it's "a private, ineffable something-or-other in your mind's eye". From what else I've read of Dennett, I'm sure that he has a problem with the "private" and "ineffable" part. Is it so clear that he has a problem with the "component of experience" part?
In the book, a character called Otto advocates the position that qualia exists. The full passage is Dennett making his case to Otto once again:
And how would you answer that passage of Dennett's ?