Quiddity has a qualia all its own

I changed my mind about a significant philosophical issue today, and in the process parted ways with a thinker I’ve been a serious fan of for a couple of decades now. The issue is raised by a thought experiment, of which I was previously unaware, called Mary’s Room. The simplest way of getting involved the dispute is to ask “How can I know that my experience of (say) the color ‘red’ is the same as yours? Is it even possible to have such knowledge?”

The deeper question this gets at is whether physicalist theories of mind are sufficient; read the Wikipedia article linked above for the argument, I won’t rehash it here. Until a few hours ago I would have, somewhat reluctantly, agreed with Daniel Dennett’s position that Mary doesn’t learn anything when she steps into the world of color. I’ve long been a fan of Dennett’s bracing, unapologetic physicalism; I’ve especially enjoyed his witty takedowns of mysterian positions in the philosophy of mind.

My reluctance would have stemmed only from this: whether or not Mary actually learned anything when she first saw the color red, it seems certain to me that Mary would feel she had learned something. Dennett argues that with complete knowledge of the physics of light, the range of red wavelengths, the history and significance of the color “red” to other human beings, Mary wouldn’t know anything about the world that she hadn’t known before.

Under an operationalist, fallibilist account of “truth” – which I think is the only sane one – it’s at first hard to see how to argue with Dennet’s position. The Mary’s Room experiment conveys the assumption that Mary knows so much about “red” by indirection that she can make predictions about events involving “red” as accurately as anyone else. So if Mary says “Wow!” upon first actually seeing the color red, what does the “Wow!” mean?

Dennet insists that Mary’s “Wow!” is meaningless – he seems to think he has to maintain that in order to defend physicalism against the mysterians. This is where I now part company with him. It certainly means something to Mary. Any theory of mind that can’t support questions about that meaning to Mary is dangerously impoverished – Mary won’t buy it, for starters, and why should she? Dennett thinks he’s robustly defending physicalism, but I think he’s surrendering the high ground to the mysterians.

The Wikipedia article offers an alternative answer based on some experimental work by two cognitive scientists working with a color-blind synesthete. I think the Ramachandran-Hubbard answer (Mary will have blindsight about color distinctions) is as mistaken as Dennett’s, because Mary (by hypothesis) is not color-blind. But they did provide me with a vital clue.

The human brain actually includes two semi-separated signal-processing pathways for vision. One is luminance-oriented and good at picking up fine details: it “sees” in black and white, but very sharply. The other is good at color distinctions but poor at processing shape details. The take from these two pathways is integrated at a late stage in processing, a fact on which depend several classic “optical illusions” and the continuing niche appeal of black-and-white photography in a world of cheap color film.

Ramachandran-Hubbard’s color-blind synesthete is in the extremely odd position that he can have activations of his color-vision pathway from two different sources (his retina and the abnormal synesthesia pathways) which give him disjoint sets of color qualia. Ack! There! I’ve said it: the dread word “qualia”, which in Dennett’s universe no self-respecting physicalist is supposed to utter without firmly insisting that it is meaningless!

Have I, too, surrendered to the mysterians? No. In fact, not.

Here is my physicalist account of Mary’s “Wow!” What she learns is what it feels like to have the color-processing pathways of her brain light up. This is an objective fact about her subjectivity; with a sufficiently good MRI we could actually see the difference in patterns of occipital-lobe activity. And that will probably be a world-changing experience for Mary, fully worthy of a “Wow!”, even if we concede the Mary’s-Room premise that she has not learned anything about the world outside her own skull.

To see this, imagine being a precocious, prepubescent scientist who knows (objectively) everything all other human have reported about sex. Now the hormonal switch flips, and you feel it…and nothing is the same, is it?

What I’m really arguing here is that Dennett, and thinkers like him, are stuck hard enough in a theoretical set of distinctions about “objective” vs. “subjective” to have ignored an important part of the phenomenology. One’s own mental life – or, to put it physicalist terms, one’s perception of one’s own brain states – is part of the phenomenal field just as genuinely as Husserl’s copper ashtray is.

I’m proposing that, contra Dennett, there is a sense of the word “qualia” that is meaningful in physicalist terms. A “quale” (singular form) is a brain state with the following properties: (a) like the abnormal activation of a colorblind synesthete’s color pathways in the occipital lobe, or like the first-ever feeling of sexual desire, it is in principle an objectively measurable event with detectable correlates in brain and body, and (b) it’s incommunicable.

That is, I can learn to anticipate the phenomenal experience I will have when I look at something you have previously told me is “red”, but I can’t tell you what that experience is. Arguably, all brain states are incommunicable in that sense – but that’s actually part of my point; language and art and mathematics and music and so forth are all, in important ways, too narrow to shove our phenomenology through.

The last laugh goes to Alfred Korzybski: the map is not the territory, the word is not the thing defined. We can communicate linguistic and para-linguistic maps, but not the phenomenological territory – the qualia – from which we abstracted them.

138 thoughts on “Quiddity has a qualia all its own

  1. To erode a few important details in the service of sentiment:

    Mary has learned which color is her favorite!

  2. > I’ve long been a fan of Dennet’s bracing, unapologetic physicalism;

    But how could you be a physicalist and a neopagan at the same time? How could you reject all qualia and at the same time say: “The mode of the mystic is experience.”. If I understand the term correctly, your mystical experiences seems to be the examples of qualia.

  3. Joshua: Then it would depend on how the android is programmed. A human-intelligent android may very well have an exactly analogous experience: Despite knowing everything about color, previously-inactive circuits built to process color may activate, and presumably cause human-complex secondary reactions.

    That really cuts to the crux of the matter; are we in fact 1950s-style androids, in the style of Data (if not even less sophisticated), or is the situation more complex than our initial simplistic and unnuanced models of cognitive function would lead us to believe? Well… so far, nobody who has bet on “it’s more complicated than we thought” has lost yet. Presumably at some point that will occur. But it won’t be before we’ve long since given up on a cognitive model that is built on the rather silly idea that understanding how an explosion works is exactly the same in every way as experiencing getting your arm blown off. That’s what the broken model implies. Any android AI that actually tried to function that way would be insane and shut down.

    The only other real alternative is that the android is not programmed with custom-purposed color circuits, in which case the android would still have to contend with new sensations and new reactions to those sensations being built, even if it was just a matter of connecting them to previous knowledge. There’s got to be some fundamental separation between conceptual and actual in practice, or your AI is going to truly suck.

  4. >But how could you be a physicalist and a neopagan at the same time?

    Reread my essay on the topic.

    >How could you reject all qualia and at the same time say: “The mode of the mystic is experience.”. If I understand the term correctly, your mystical experiences seems to be the examples of qualia.

    No. It is not easy to communicate about them, but it is possible. Even supposing they were incommunicable, I just got done describing “qualia” in a physicalist way – any contradiction you think you’re seeing is a result of your own preconceptions about what a “mystic” must believe.

  5. Eric,

    FWIW, I had not hear of dear sweet Mary either. I read the article in Wikipedia before I read yours, and drew pretty much the same conclusion you did. However, it reminded me of my broader thoughts on nearly all these sorts of philosophical arguments, namely that they are largely meaningless because they all largely come down to finding a shared meaning of various words, and one spends all one’s time arguing about arbitrary definitions. For example, consider the philosophical question: “Do I have free will”. This question can easily be answered if we will all agree on what the words “I”, “have”, “free” and “will” mean.

    So too, in this example, what it basically comes down to is this: what does one mean when we say Mary has “knowledge”. If you include only intellectual knowledge then Mary learns nothing, however, if you include knowledge in a broader sense, including other parts of your central nervous system beyond our cerebral cortex, then she clearly does.

    To reiterate, what it, and frankly most philosophical questions, comes down to is the meaning of words. Thankfully for all University philosophy departments and writers of high brow books (including words like “qualia”) one can debate and discuss this with no hope of agreement, since the meanings of words are largely arbitrary. However, for the wise young things in university coffee bars (wise in the sense we use it in “sophomoric”) one can discuss such matters, look really smart, and hold firm passionate positions without fear of irrefutable contradiction, until our eyes turn red as Mary’s tomatoes.

  6. And just since I am on a rant about philosophy, can someone explain why it is called philosophy and not sophology, or at the least sophophilia? Same with philology, why not logology or logophilia? And these damned mathematicians are a bunch of etymological narcissists.

    OK, I got that off my chest.

    I’m off now. I’m going to dye my hair purple, tattoo “David Hume” on my butt, then head off to the local coffee shop to smoke some weed, drink quadruple expressos and say really deep, dark things. “The moving finger writes and having writ moves on; nor all thy piety nor wit shall lure it back to cancel half a line; nor all your tears wash out a word of it.” Yeah, man, deep, really deep.

  7. I disagree that qualia are inherently incommunicable, just that we aren’t equipped (yet?) to communicate them. (I’m new to the topic, maybe ways off in a direction proven false somewhere?)

    You said that a sufficiently good MRI would give us a view on qualia – on the brain’s pathways lighting up. But we (apparently) lack brain-internal sensors with good resolution – we can’t see by ourselves what the MRI would show. Imagine we could…

    There was an experiment out there with someone wearing a compass belt that constantly vibrated in the direction of north – they reported developing a sixth sense out of this, and improving their orientation abilities in the process – they were no longer getting lost in strange cities.

    So imagine we could see, all the time, a real-time 3D MRI image of our own brains. Seeing such an image would enable us to experiment with influencing the image: what do I have to do to get this or that pattern? Based on how we seem to handle other senses, I believe we’d learn how to manipulate our brain to create an MRI-identifiable state. Some states would be unattainable (I can’t lick my elbow), and some would be relatively easy (happiness? Where would that lead?).

    Now, the 3D MRI image can easily be communicated, therefore a recipient (Mary) would be able to attempt a previously unknown brain pattern (what others experience when seeing red) and get the feeling. If the MRI would show different people reacting to red differently, we’d be forming theories about that, and Mary in her B&W room could try out various existing and predicted reaction patterns.

    Then she’d come out and instead of a wow, she’d say “been there, done that”.

    There’s also a possibility in this line of argument that qualia are hiding under the resolution of MRI – we can sense them but not observe them; or that only really-well-trained persons (reminds me of meditation and yoga) would be able to influence their MRI patterns finely enough to simulate seeing red. Presumably, Mary in the thought experiment would be so trained, just like she’s trained in neurobiology.

  8. This essay meshes nicely with my biggest bone of contention against the Dennett/Bostrom/Yudkowsky/Drescher cabal: the substrate-independence premise of the simulation argument is nonsense. A defender of the school of thought that you’re critiquing in this essay would likely argue that complete knowledge of the cognition of color entails the ability to produce a perfect, complete mental simulation of it, and that a perfect simulation of any phenomenon is indistinguishable from the phenomenon itself, and therefore Mary can produce visual qualia at will. Now, if the Dennettites are willing to concede that a mental simulation of this nature entails Mary somehow using her frontal lobe to artificially stimulate her own visual cortex, then I’ll agree that nothing is learned upon stepping outside, but that this conclusion is then obvious and uninteresting. However, I wouldn’t anticipate concession of this nature; instead I’d expect to hear that the frontal lobe can carry out all of the actions of the visual cortex within itself (assuming, for the sake of argument, sufficient relative information capacity to carry this out); that is, execute the function of the visual cortex on a different substrate; and that this change of substrate is of no particular consequence.

    Here’s the trouble with substrate independence. When we say that system A is a simulation of system B, what I think we mean is that some homomorphism exists mapping the states of system A onto system B. But that’s actually an extremely weak statement! Provided that the isomorphism function can be exponential in complexity with respect to the simulated system, then every system can be a simulation of every other system of equal or lesser complexity. Are you prepared to bite the Tegmarckian bullet and assert that if you have a vat of boiling water whose mass and kinetic energy equals or exceeds that of my body, then every other possible person whom the state of that vat might be mapped onto is equally as real as I am?

    You can retort that some homomorphism is encoded in the mind of an observer of the simulation, and that this particular homomorphism is therefore real. But then, without trapping yourself in infinite regress, how do you define what you mean when you say “encoded”? I think the following is the only way out: given observer O, system A is a simulation of system B if the execution of A causes O to reconstruct B. In other words, a DVD of The Matrix is a perfect simulation of a television playing The Matrix, if and only if it the DVD has a television observing it. So once again, you haven’t brought yourself anywhere terribly exciting.

  9. Nice, you hit the nail on the head my man. But what is that subjective feeling? I think a full answer to Dennett and co needs to be based on a failure of reductionism and an information view. I’m been trying for years to rebut the reductionist/bayesian crowd and think I’ve finally done it. So here’s my take on consciousness after 7 years of thought on the topic, think if as my way of giving the Dennett/Bostrom/Yudkowsky/Drescher cabal the fingers on the way out:

    “The truth is most likely a combination of Dennett’s ‘Narrative Center of Gravity’, Baar’s ‘Global Workspace’, Tononi’s ‘Information Integration’ and Hofstadler’s ‘Strange-Loop Analogies’. Combine these four ideas and you get an explanation something along the lines of this (my own ideas):

    ‘Consciousness is concerned with integrating different high-level representations of goals into a single coherent high-level view-point which serves to coordinate all the seperate components of the brain. This integration is done via categorization, or equivalently, analogy formation. It takes the form of story plots, or narratives which are coherent plans for future actions’.

    I’m a ‘non-reductive’ materialist, I agree consciousness is physical, but think that there are different irreducible levels of explanation, and concepts at higher levels of explanation are not neccesserily reducible to concepts at lower levels of explanation. Our actions may be pre-determined at the physics levels, but all that exists at that level are mere particles, forces and fields. What hasn’t been pre-determined is the *meaning* of all that…. and that’s what consciousness can change.
    See:

    http://en.wikipedia.org/wiki/Physicalism

    “Non-reductive physicalism is the idea that while mental states are physical they are not reducible to physical properties.”

    Essentially, Quine critiqued reductionalism on the grounds that you can’t even talk about low-level concepts without referencing high-level ones. See:

    http://en.wikipedia.org/wiki/Two_Dogmas_of_Empiricism

    From the paragraph on ‘Reductionism’ :

    “The difficulty that Carnap encountered shows that reductionism is, at best, unproven and very difficult to prove. Until a reductionist can produce an acceptable proof, Quine maintains that reductionism is another “metaphysical article of faith”.

    The problem with reductionism is that the semantic meanings we attach to things are neccessary for explanations, but these semantic meanings are only determined at a high-level! Bayesian reasoning relies on fixed semantic meanings… the programmers have to precisely fix the initial meanings of everything at *some* level of description… but humans can use conscious reflection to *change* the meanings of everything (our brains can always *recode* how meanings map to raw bits of information) so this shows that: (a) Reductionism fails, and (b) Bayes cannot fully capture rationality.
    Consciousness could be both composed of physical processes AND have some non-physical properties. This sounds highly peculiar simply because science has never yet encountered a failure of reductionism, but consciousness could be the first instance of it. It would mean that when you beak the brain down into its individual parts it’s all physics, but when the parts are put together new properties emerge which can’t be explained as merely a combination of the parts.

    Physics itself relies on our understanding of abstract mathematical categories such as ‘integration’, vector’, ‘equation’ etc, and categories depend on the language or code the mind uses to define the meanings of information content.

    For instance consider a radio wave digitally modulated which is transmitting the voice of a professor giving a physics lecture. It’s just a string of bits;
    e.g. 00001011101111000001101011001 etc

    But what does it mean? That depends on the coding system used to assign meaning to the information:
    e.g.
    0000=’Dirac Equation’,
    1011=’Feynman’ etc

    But the coding choice is not itself reducible to low-level physics, because only a mind can understand it. So we appear trapped in a circular explanatory loop. In fact you could argue that Bayesian Reductionism is just the modern vastly more sophisticated version of the old discredited fallacy of Hume, who saw everything as mere associations between sensory data.”

    Suck on that reductionists!

  10. Jacek – whilst I can see, somewhat, your premise, there’s a complicating factor that I believe people brains can differ substantially in the detail of their layout whilst having the same broad layout.

    I only get this from tv – but brain surgeons always have to probe specific areas to determine their affect, they don’t appear to have a good map of the brain. Also, in the case of brain damage, other areas of the physical brain can take over activities.
    So I’m not sure that it’s a given that communication on that level of detail would be meaningful

  11. Interestingly, I also gave Korzybski the last word (actually first) on this affair in my analysis: http://fare.livejournal.com/121167.html

    From the common-sense agreement of everyone on the situation of Mary, the main argument of the mysterians is an instance of “insanity” according to Korzybski: confusion between the thing and its representation. They shift the meaning of knowing “everything” about red from reality to representation so as to fallaciously reach their nonsensical conclusions.

    Dennett’s mistake may be to accept the ambiguous hypothesis of knowing “everything” about red. If meant in the wider meaning from which the mystics later deduce their conclusion, that she actually has a deep knowledge about the red and experiencing it, then Dennett is correct that her experience of real red is nothing new as compared to her previous experience of simulated or imagined red (possibly achieved through a intercessive reflective interface to her brain) as implied by the hypothesis. She has already experienced the “qualia” to reuse the mystics’ vocabulary, though never by natural means. If the hypothesis was meant in the narrower meaning of knowing everything about a *model* of red, then all the mystics are pointing at is the difference between this narrower and wider meaning — a difference they are trying to argue away in their fallacies.

  12. Interesting. I had not heard of this “Mary’s room” thought-experiment. Coincidentally, I have previously approached this topic somewhat tangentially – from the angle of “communication” in general.

    I wondered “how can I describe color to another?”. Mary has all the ‘physical’ knowledge of color, in terms of wavelengths and retinas etc, yet until she steps out into the world and sees a red object, she is ignorant of what gives red its ‘redness’ to humans.

    We can communicate color by mutual reference – I point at the grass and say “that’s green” – but I still have no idea that what I see as green is not, in fact, what you see as blue. If I could ‘wear’ your eyes, what would my colorscape look like? Although I hadn’t heard the word, these are the ‘qualia’ that are so hard to communicate.

    Mary has ‘learned’ what the _experience_ of color is.

    FWIW, I see the world of art as an extension of language, striving to communicate these qualia. A painting that captures ‘love’. A poem that captures the innocent bliss of a child.

  13. @esr: So basically you’re a neopagan atheist? IOW, you see the gods as historical or Jungian archetypes, but only in the psychological and evolutionary senses, ignoring the the concepts of ‘soul’ and ‘psyche’ pervasive in New Age Jungian thought, hence the ‘collective unconsciousness’ being an illusion that simply arose as a matter of evolutionary pressures?

    @Ivan: It’s certainly possible to be a neopagan physicalist. Basically, it means that while ESR might have some sort of altar or tools, and though he might practice some sort of ritual, the tools and ritual and communication with gods or other mythical entities are simply a means of reprogramming the deep mind. Think of it has a form a hypnosis. Which isn’t too far off the mark, either. The use of color or scent associations, for example, green for money or growth; the scent of (very small amounts of) patchouli oil is said to have effects similar to male pheromones (I can’t come up with a reliable source on the patchouli/pheremones thing at the moment, so I’m using weasel words :-P) and all this certainly has its basis in human psychology because we has humans, being emotional animals, tend to associate certain emotions/feelings with certain scents, colors, textures, etc. It’s a product of how our memories and imagination work.

    Something to think about, however, that, at least from my perspective. tends to poke holes in physicality is this: studies have shown that the mind literally does not know the difference between what it sees and what it remembers or imagines. For instance, they’ve looked at the pattern of synapses firing in the brain, and they’ve found that when subjects were shown a picture of an apple, and were also told to imagine an apple, the synapses within the brain fired in exactly the same pattern.

    Thinking of that apple might you make hungry; it might make your stomach growl and it might leave with you a hint of the taste of that juicy apple, which might start to make your mouth water, yet nothing outside of you caused that — it was that which was within you that caused that.

  14. >Are you prepared to bite the Tegmarckian bullet and assert that if you have a vat of boiling water whose mass and kinetic energy equals or exceeds that of my body, then every other possible person whom the state of that vat might be mapped onto is equally as real as I am?

    Um, yes actually, sort of? At least within the conscious observer’s own local context. You seem to be using something along the lines of Searle-esque “ping-pong balls and beer cups analogy”, a sort of fallacy through unrelated absurdity. I can’t seem to really grok what you are even trying to get at beyond this.

    On the post at large, I must say this reminds me of Hofstadter’s attempt to reclaim the “soul”: not as a magic observer, but as a mental pattern.

  15. A quick note on “Do we all see the same red?”

    My answer for many years has been “No, because if everyone saw the same colors, we’d all like the same colors. Instead, we all have different tastes.”

    Similarly, no one could enjoy what happens in my mouth when I eat an olive. Ergo, people who like olives taste something different.

    The color sense is not spectroscopy; taste is not chromatography.

  16. @DJMoore – It seems to me that your answer is confusing two separate things – our perception of color, and our subjective tastes for color. It also does not seem to follow that if we all saw the same color, we’d all like the same color.

    Flavor is another interesting topic. I have a friend that is literally as repulsed by the flavor of coffee as he is dog shit….I love a rich espresso….go figure.

    If the entire universe were lifeless, would there be color?

  17. >However, it reminded me of my broader thoughts on nearly all these sorts of philosophical arguments, namely that they are largely meaningless because they all largely come down to finding a shared meaning of various words, and one spends all one’s time arguing about arbitrary definitions.

    All correct, except that you appear to think “arbitrary” implies “meaningless”. It doesn’t. The reason philosophers spend so much time arguing about definitions is that it’s the only way to arrive at a shared vocabulary so that you can tackle the really hard questions – the ones that are still left when you actually have a shared vocabulary. Of course, there are many traps on the way – for the free-will case, see this as an example.

  18. >Now, the 3D MRI image can easily be communicated, therefore a recipient (Mary) would be able to attempt a previously unknown brain pattern (what others experience when seeing red) and get the feeling.

    See dfranke’s criticism of substrate-independence in the comment following yours. Your “therefore” is highly questionable in this case.

  19. >I’m been trying for years to rebut the reductionist/bayesian crowd

    I’m not interested in ‘rebutting’ them: I largely share their goals. I’d rather fix the limitations in their theory. I know Eliezer Yudkowsky and do not believe he would interpret my analysis of the Mary’s Room experiment as hostile criticism.

  20. >the main argument of the mysterians is an instance of “insanity” according to Korzybski: confusion between the thing and its representation.

    Good point! I think your analysis of Dennett’s confusion is also largely correct.

  21. >@esr: So basically you’re a neopagan atheist? IOW, you see the gods as historical or Jungian archetypes, but only in the psychological and evolutionary senses, ignoring the the concepts of ’soul’ and ‘psyche’ pervasive in New Age Jungian thought, hence the ‘collective unconsciousness’ being an illusion that simply arose as a matter of evolutionary pressures?

    All correct, except that I’m not sure the “collective unconscious” is an illusion. It may be an information network in which humans are nodes and the links are low-bandwidth and high latency. Isaac Bonewits has written some interesting speculation about this.

  22. I’d like to quibble on the ‘incommunicable’, by extending it to “incommunicable with the present technology”. It does not seem impossible in principle to construct a brain-state-copier. If I had such a machine, I could look at something red, copy my brain state (or just the relevant parts of it that correspond to “RED! RED!”; how much brain states factorise is a difficult question but presumably answerable in theory) and transfer it to Mary while she is still in the room. She would then have the experience of having her brain light up in the “RED!” pattern without having her optical nerves stimulated.

    It might even be possible in principle to do the same thing by words alone, if you knew exactly how a given word (or sound, more generally) affects the brain state, and exactly the brain state you want. (I say “might”, because it’s not clear whether the coupling from the sound- and language-processing circuits to the visual-processing circuits is strong enough.) This would be equivalent to programming by butterfly using Brainf*ck as your language, but what of it? I wave my hands and say “in principle”, and lo, many things become possible! :D

  23. >It does not seem impossible in principle to construct a brain-state-copier.

    Again, this depends on an assumption about substrate independence that is shaky at best.

  24. >>Now, the 3D MRI image can easily be communicated, therefore a recipient (Mary) would be able to attempt a previously unknown brain pattern (what others experience when seeing red) and get the feeling.

    >See dfranke’s criticism of substrate-independence in the comment following yours. Your “therefore” is highly questionable in this case.

    I’m afraid I didn’t quite get dfranke – usually the point is that a simulation is not perfect – it’s intentionally simpler so it can run faster (and with real-world systems, often it is used to predict *future* behavior of a system which wouldn’t be possible if it couldn’t run faster). And what’s the point with people being equally real as him? Maybe I’m just not well-enough versed in philosophy (which I wouldn’t claim to be).

    If we had the MRI sense, we’d have an easier job of trying to create maps of the brain (easier, not necessarily easy), and then Mary would have an easier job of *attempting* to reach the same state of mind.

    I don’t believe in incommunicable. That seems to be at the heart of the issue.

  25. >All correct, except that you appear to think “arbitrary” implies “meaningless”.

    No, I don’t think that at all. My point was not the philosophy was worthless, (or meaningless) it was that most discussions of philosophy are destined to be fruitless. Let me put it a different way: philosophical systems are chaotic. That is to say, a minuscule tweak in the meaning of a word causes massive, highly unpredictable consequences throughout the whole system based on that meaning.

    This situation is excellent if you are a professional philosopher (or dope smoking, expresso drinking college kid trying to impress and bag dizzy blondes with how deep you are.) Simply speaking, there is never a right answer. There is always more talking and book writing to do. There are always more dizzy blondes. There is always more dope. There is rarely any way to pin down any philosophical discussion because one can always wriggle free with a tiny unstated tweak. It makes for great book sales, accelerates the sale of dope and coffee, and allows for increased recreational sexual activity. So all in all, the fact that most philosophy is pointless and meaningless is compensated for by its auxiliary benefits.

    Perhaps defining a well accepted set of jargon is helpful in this regards, but I doubt it. Philosophers have such divergent viewpoints in general, and philosophy largely comes down to the meaning of words. So the likelihood of success in that endeavor is not unlike the likelihood of fundamentalist Christians from Georgia agreeing with NeoPagans from Seattle on the nature of the anthropogenic global warming problem. (Or worse, where to get the best cup of coffee.) If there is no shared common ground, there is little shared common vocabulary. Especially so with the specificity necessary in philosophical arguments.

  26. >This situation is excellent if you are a professional philosopher (or dope smoking, expresso drinking college kid trying to impress and bag dizzy blondes with how deep you are.)

    Personally, I was always more interested in bagging non-dizzy redheads. More years ago than I care to remember, I discovered that one of the principal advantages of actually being a deep thinker is that it substantially reduces the difficulty of convincing such women that you are one :-)

    All you’re demonstrating here is that philosophy reduces to a pointless definitional tail-chase when it’ s done poorly. This was not in dispute. I try to do it well.

  27. “There are always more dizzy blondes. There is always more dope”…”It makes for great book sales, accelerates the sale of dope and coffee, and allows for increased recreational sexual activity”

    Come on now, ‘fess up….you’re a plant from the Philosophers’ Guild!

  28. Main article sounds right to me.

    Gary Drescher in *Good and Real* offers an idea that might support this; the idea is that when you encounter a new sensory experience, it gets an anonymous GENSYM that operates as its internal name.

    Ergo, when Mary sees red, she “finds out” what the red-GENSYM is – but it’s an anonymous GENSYM, so she can’t describe it to anyone else.

    Marc J. Geddes (mjgeddes) is a lunatic/crackpot type who goes about the Net pronouncing himself triumphant over me in a rather extremely one-sided rivalry; don’t worry about unraveling his meaning.

  29. >Gary Drescher in *Good and Real* offers an idea that might support this; the idea is that when you encounter a new sensory experience, it gets an anonymous GENSYM that operates as its internal name.

    Um, Eliezer, did you know that I’m a LISP-head of ancient vintage? I find this analogy hilarious. I don’t mean that in a hostile way; I see what Drescher is driving at, I think. It’s just a hellaciously frickin’ funny way to present the idea.

    I don’t think an anonymous GENSYM is quite the right analogy here. though, because the internal “symbol” for Mary’s red is probably an activation of a network of neurons that is causally tied to the V4 area. Thus, it’s not really arbitrary, not in the way that the internal GENSYM for (say) the word “qualia” would be. I recognize that this is mainly a technical quibble.

  30. > Again, this depends on an assumption about substrate independence that is shaky at best.

    Both substrates in this case are carbon-based brains of the same species. They can’t be *that* different. Still, I do take your point; the question will have to be referred to empirical testing over the next fifty years or so, as we learn more about how brains work.

  31. PS. Do I detect an oblique allusion to “Johnny Got His Gun” in the title?

    ESR says: The title alludes to “Sometimes, quantity has a quality all its own”, which is variously attributed – most often to Joseph Stalin.

  32. All correct, except that you appear to think “arbitrary” implies “meaningless”. It doesn’t. The reason philosophers spend so much time arguing about definitions is that it’s the only way to arrive at a shared vocabulary so that you can tackle the really hard questions – the ones that are still left when you actually have a shared vocabulary. Of course, there are many traps on the way – for the free-will case, see this as an example.

    So why isn’t there a shared vocabulary, like a “standard”? And I don’t mean all those trendy terms that people are always misusing like ‘zeitgeist.’ I mean a real, agreed-upon and universal set of symbols (as in programming) that represent various agreed upon concepts. The problem with English and most other spoken languages is that they’re so … imprecise.

    Take, for example, this dictionary. It’s filled with loaded terms like “communism,” “conflict,” and “dogma.” If one were to comment on someone’s stated doctrine or tenet of belief, and then refer to that as “dogma,” instantly one is viewed as demonizing that person’s dogma, due to the negative stigma often attached to that word, whether the use of the term is meant pejoratively or not.

    Or perhaps I’m seeing a problem where none exists? I mean, language is just that, isn’t? A universal and agreed upon set of symbols, right?

    Um, Eliezer, did you know that I’m a LISP-head of ancient vintage?

    In fact, rumor has it that ESR’s ‘(‘ and ‘)’ keys have the labels worn right off of them. :-P

  33. Morgan Greywolf writes:
    >So why isn’t there a shared vocabulary, like a “standard”?

    Because to have a shared vocabulary you need to live on the same planet. The diverse philosophy of, for example Muslims or Buddhists compared to utilitarians or objectivists, liberals or conservatives, is so different having a shared vocabulary is essentially impossible. What tends to happen is that philosophers fraction into little cliques, and form their own vocabulary that is often entirely non understandable to the uninitiated.

    This should be obvious to anyone who has ever engaged in political debate. If you are a conservative, and you talk to a democrat you can disagree about some policy issue, and as you begin to examine why the difference you find that the problem is not whether to increase taxes on the rich to help the suffering single Mom get health insurance, the problem is that at a most fundamental level you think in completely different ways about the human condition. You don’t share enough beliefs to discuss anything and come to any agreement at all.

    Furthermore it is a pretty difficult thing to define the precise meaning of a word that describes an abstract anyway. You description immediately gets subjected to an analysis of what the words in the description mean. Then, the words in the description of the description and so on, down the spiral with no closure in sight. However, the dope and coffee flow freely as we all enjoy discussing these things, feeling a great sense of self importance, and superiority to the plebeians who only discuss the perfunctory rather that the profound (like what we do.)

    I don’t mean to be contemptuous, however, most of these discussions serve very little benefit unless the discussion focuses consciously on practical realities. The handwaviness of philosophy needs to be grounded in some practical reality to save you ending up on a trip to Jupiter. If you don’t then you end up in a pointless spiral of nonsense. Mary’s room describes something quite unrealistic. I understand that the purpose of the non-realism is to control for variables that might muddy the answer. However, what is the practical benefit of Mary’s little story? Simply to reveal what seems rather obvious to me: namely that “knowledge” is an extremely vague term, that the human brain is multi-layered, and that there is more to learnin’ than book learnin’. This doesn’t seem all that useful or illuminating a discovery.

    But perhaps I missed something.

  34. >Marc J. Geddes (mjgeddes) is a lunatic/crackpot type who goes about the Net pronouncing himself triumphant over me in a rather extremely one-sided rivalry

    I think I’m entitled to rely to this don’t you? I have traced the whole disagreement between me and the Yudkowskyian fan-boy crowd back to a single issue: reductionism versus non-reductionism; these terms are both coherent well-defined positions in professional philosophy, the non-reductionist postion is hardly ‘crackpot’. Virtually every point of annoyance I have with Yudkowsky’s crowd traces back to this single issue. That it, I were to convert to reductionism, I think I would support everything they said. Yudkowsky is a reductionist, others like me beg to differ. Simple. The issue of reductionism versus non-reductionism is still open, and has not been resolved to my satisfication.

    If readers want something which is concrete and clear, where I disagree with Yudkowsky here it is:

    “I think the strict distinction between values (preferences) and decision options in standard decision theory will fail. A mind needs a way to set the priors to act in real-time, but this relies on our intuition (values, preferences). Specifically, I believe setting the priors relies on our aesthetic preferences.

    Uniform priors might work for simple problems, but increase the complexity and they’ll fail. Formal solutions for complexity priors that actually work (i.e. formalizations of Occam’s razor) are all uncomputable or intractable – in other words, they’re no solutions at all ;)

    I predict the problem of priors will only be solved with a sentient system, which can use conscious intuition to determine the correct approximations (i.e. aesthetic preferences).

    There’s a close connection between Occam’s razor, subjective beauty and priors. See:
    http://www.idsia.ch/~juergen/beauty.html

    This is a strong clue suggesting aesthetic preferences as the means of setting priors.

    I suspect that as you grow more intelligent, ordinary Bayesian updating matters less and less and setting priors matters more and more; that’s because the more intelligent you get, the more information is coming in, leading to a computational explosion in possible correlations, you need more and more to be able to cut down fruitless reasoning avenues, you have to rely on priors more and more.

    The irony is that by the time you get to super-intelligence, Bayesian updating is probably largely irrelevant, and setting the priors is almost everything aka non-Bayesian conscious intuition”

  35. > ESR says: The title alludes to “Sometimes, quantity has a quality all its own”, which is variously attributed – most often to Joseph Stalin.

    AFAICT this quote (as “quantity has a quality of its own”) is by David Glantz on Soviet army during WWII. It’s mis-attributed to Stalin and Lenin all around the English-speaking web though.

  36. Elementary particles and their interactions, mediated by other elementary particles, are all that exist. Non-reductionists confuse their impressions of reality with the reality. Rather like vitalists considered life as obviously different from everything else.

  37. William,

    This issue is debateable, and I’d to like a world class non-reductionist philosopher debate Yudkowsky on the topic – who knows, if I thought Yudkowsky did better in the debate, I may even start issuing public apologies and sending money to SIAI!

    Meanwhile, I leave you to think about what you said in your post, you actually separated reality into TWO levels. I quote you; ‘Elementary particles AND their interactions’. This was the same amusing slip Yudkwosky made on his OB post on reductionism where he actually talked about ‘fields AND forces’ ;)

  38. >This was the same amusing slip Yudkwosky made on his OB post on reductionism where he actually talked about ‘fields AND forces’ ;)

    This was just a temporary concession to ordinary language; Eliezer knows better. He put it better in my favorite quote from him *ever*:

    “You have the absolutely bizarre idea that reality ought to consist of little billiard balls bopping around, when in fact reality is a perfectly normal cloud of complex amplitude in configuration space. This is your problem, not reality’s, and you are the one who needs to change.”

    Priceless…

  39. >So why isn’t there a shared vocabulary, like a “standard”?

    I think I can explain that, but be warned: most of the philosophy professors in the world will disagree violently with this account. That’s their problem.

    There is no standard vocabulary because until very recently (a little over a century ago) all of philosophy was still in a confused prehistoric sort of state, treated as a form of historical hermeneutics or a branch of literature. Many thinkers, especially outside the U.S. and England, still treat it that way. In that state, individual philosophers could have insights, but nothing could be settled because there was no theory of confirmation that didn’t disappear into circularity, religious dogma, or sheer handwaving.

    Then there was C.S. Peirce, and everything changed. His breakthrough – an enormous and still underappreciated one – was to identify the “truth” of hypotheses with their predictive power (“How To Make Our Ideas Clear”, 1878) By doing this he unified philosophy with experimental science, gave philosophers a way out of endless semantic traps, and reframed the mission of philosophy as the theory of theory-building. And he created a potential for a “standard vocabulary” where it had not previously existed.

    There has been no more important insight in the history of the field, ever. But for various historical, political and esthetic reasons, many philosophers have not yet caught up to where Peirce was in 1878. In fact, some of his immediate successors misunderstood and vulgarized his central breakthrough so badly that it was nearly buried. It was later rediscovered in various partial forms by by Hans Reichenbach, Karl Popper and other philosophers of science – but their formulations lacked the generality of Peirce’s and its ability to cut through questions in classical philosophy as well as the philosophy of science.

    So the answer is, it’s too soon. Despite having a couple millenia of history, philosophy today is about where chemistry was in, say, 1850 – the principles needed to make it rigorous and to ground that shared vocabulary (like Dalton’s atomism) have been discovered but not yet generally accepted. Someday – probably fairly soon in historical time – all philosophers will begin their theory-building with phenomenology and Peircean operationalism as automatically as chemists begin with atomism and Dalton’s law. But that time is not yet.

  40. “Dennett argues that with complete knowledge of the physics of light, the range of red wavelengths, the history and significance of the color “red” to other human beings, Mary wouldn’t know anything about the world that she hadn’t known before.”

    This seems like a pretty crazy position to commit one’s self to. It only takes a slight extension to get to ‘Dennett argues that with a complete knowledge of physics and a minimum set of observations, there is nothing that Mary wouldn’t know’.

    Obviously, humans (and everything else) are limited by complexity; the view you propose seems like a straight forward acknowledgment of this.

  41. ESR – I haven’t been able to go directly to your homepage for 2 days now, I keep getting a 404. I haven’t had any trouble following a link to a specific page from another site however.

    ESR says: I’m seeing it too. I’ve sent a trouble report to the ibibilo admins.

  42. I’m deeply pessimistic about academic philosophy’s ability to advance much of anywhere. Philosophy has produced lots of good ideas, but they’re smothered in a far vaster sea of stupid ones, because philosophers are chronically incapable of telling the difference. We still argue about Zeno’s paradoxes two-and-a-half millennia after Zeno. Can you name any philosopher who was ever widely-known, but is now generally agreed to be discredited?

  43. ESR,

    You are of course aware that even in fundamental physics a single unitary level of description has not been achieved? (i.e quantum mechanics contradicts the field physics of general relatively) And QM requires an unexplained bridging law (Born probabilities) derivation to convert math into human observables? If even in physics there is doubt about reductionism, imagine how much more doubt there is for something like the mind!

    The reductionist/non-reductionist issue is related to virtually all the other key issues of cognitive science, for instance the idea that all reasoning can be reduced to Bayes. (Can priors be derived from Bayesian inference or is something non-Bayesian going on?)

    On the OB blog, Hanson estimated the chances of a failure of reductionism as less than 1%, but in light of the above I think the odds are much much higher , >%50


    As an alternative to reductionist philosophers, I recommend the work of Bohm, as Bohm seems to have been a non-reductionist, and his pilot-wave interpretation of QM is a strong alternative to Many Worlds (yes, even the interpretation of QM depends on whether reductionism or non-reductionism is true!). If Gary Drescher is the champion of reductionism, then Bohm is the champion of non-reductionism.

  44. >If even in physics there is doubt about reductionism, imagine how much more doubt there is for something like the mind!

    I think you would confuse yourself (and others) less if you stopped using the word “reductionism”. You don’t appear to know what it actually means, or you would not take the well-known problems with relativity and QM as evidence that it’s untenable. For that to follow, physicists would have to believe at a minimum that QM and Relativity *necessarily* cannot be unified, rather than simply observing as they do now that they are not yet unified.

    I don’t think “reductionist” is a very interesting word, anyway. It seems to function mainly as a term of abuse directed at people the speaker thinks are too physicalist, or not willing to bow before the numinous quality of epiphenomena, or something like that – I stopped caring years ago. If you think Bayesianism is an insufficient account of reasoning, I suggest you focus on that criticism instead of haring off into metaphysics.

  45. I remember reading about Mary’s Room in the (one) philosophy course I took in University. You don’t need to come up with something as elaborate as “Mary’s Room” to have this sort of debate, though. Any sort of “experience” versus “description of the experience” contrast will provoke the same debate.

    On the other hand, you can also think of other, equally elabourate thought experiments that drive the point home. My personal favourite was to take Mary’s room and apply it to *us*. Imagine Fred who claims that, while we can only see blue, *he* sees two colours, which he calls “zapo” and “lupo”. Imagine, furthermore, that a group of neuroscientists figure out what’s happening on a physical level and invent a set of glasses that would enable us to see the new colours. Would you learn something new if you tried them on? What if you were like Mary and you knew *everything* there was to know about coloured light and its effect on the human body?

    I think still side with Dennet: We tend to think it’s “just obvious” that Mary would learn something because we don’t imagine the thought experiment in enough detail. Mary is supposed to know *everything* about colour and its effect on her – absolutely *everything*. This includes a complete breakdown, in stunning, page by page (or book by book even, since there’s just so *much* of it) detail, of how she would react to the new colour, what mood she would be in, what she would say (to the very word) upon her introduction to a tomato, etc.

    This task is, to put it very mildly, mind-numbingly immense (or Immense, with a capital I, to steal a stylistic tactic from Dennet’s writings). It is, of course, impossible for us to really imagine what it would be like to be in this god-like position – what it would be like to know *everything* – absolutely, totally, EVERYTHING – about colour and its effect on Mary (or ourselves). So we compromise with ourselves, and simply imagine that she knows “alot” about the topic. Of course “alot” isn’t the same as “absolutely everything” and if Mary merely knew “alot” about colour and its effect on herself well, then, of course she learns something upon seeing a tomato. But then we aren’t fulfilling the premises of the thought experiment.

    Of course, one could argue that our brains are simply not equipped to construct, out of nothing but physical data, an experience like seeing a tomato to such a detailed degree that it would be exactly like experiencing the vision of the tomato directly. Chalk that up to lack of imagination; perhaps our brains are limited in this regard.

    But to say that this proves physicalism is incomplete is wrong. At best it shows that our brains are limited in their ability to process large amounts of data.

  46. I hope my musings won’t be an annoyance (been there-done that)…
    I am guessing that ‘qualia’ is used by some to indicate some other ‘thing’ that is not strictly physical -in that case i would like to see the word removed from existance, but what it represents is something we don’t have a handle on. We know what it is- we just can’t write up a formula yet.(?)
    Isn’t qualia actuality only states of the brain which we are yet unable to fully map out?
    It seems that if we had the ability, we could take an image of a brain having a thought, and then, if we could replay that image back into he same brain before it had a chance to change itself (or BE changed), that same thought (or experience) would recur.
    While that would only work for the SAME brain in the same state as it was when the image was created, it seems that in doing that we would prove any experience or thought to be physical.
    While it may be impossible to do that, to say that there is no state of the brain that IS a thought is to have to say “qualia”.
    It is impossible to map out all the universe at one instant – therefore the universe is an experience. (?)
    Is that wrong?
    All of it?

  47. Daniel Franke wrote:
    >I’m deeply pessimistic about academic philosophy’s
    > ability to advance much of anywhere.

    Here is the question that one really needs to ask about philosophy: does it predict anything useful? I think the answer is a heavily qualified yes. In some areas, such as the philosophy of science, and the invention of the self correction of the feedback loop called the scientific method, most of the modern world owes its existence to this. However, the philosophy of science is a very narrow branch of science, some might even deride it as an ugly stepchild. If you broaden philosophy to include logic and mathematics then obviously there is great value there too. However, when one considers philosophy in a fairly strict sense that is to say questions of metaphysics and epistemology, and perhaps include ethics or morality I’d argue that not much of use has come out of that at all.

    (That, BTW is why I suggest it be rebranded sophiology, so that we get the darned emotion out of it, and treat it like a science.)

    Let me give you an example of what I mean. Consider the subject of artificial intelligence. Debate rages in that community as to what “intelligence” means, and this is much like a lot of philosophy — arguing over the meaning of a word who’s meaning is entirely arbitrary anyway.

    It leads to something like the Turing test that is considered by some as a gold standard of intelligence. But is in fact nothing of the sort, as is evidenced by the many entries offered up for the test, which are a tangled web of heuristics.. The Turing test is really more a test of knowledge of the experience of the human condition, and the ability to understand human language rather than any measure of intelligence. For example, ask the Turing test candidates the following question: “Laundry detergent costs three ninety nine, a laundry machine costs two ninety nine, which is more expensive?” A very sophisticated candidate might be able to answer this question by knowing a lot about ambiguity of human language, and typical prices of consumer goods. However, the actual intelligence involved (namely that $3.99 < $299) is almost zero.

    An intelligent machine would take a design for the Xeon microprocessor, and data on the performance and physical properties of various configurations of materials, and would produce a better design. Or it would take data from a variety of streams, such as bank reports, newspaper stories, government reports and sales performance data, and would predict the stock price of IBM over the next two months.

    However, you could disagree and argue that intelligence encompasses the understanding of the human condition too. And you would have no more claim on defining intelligence that I do. So we might as well fire up the espresso machine. We have a long night of pointless arguing ahead of us.

  48. It seems to me that there’s a computer-science angle to at least the Mary’s Room problem: Mary’s brain is a computer of limited capabilities, and thus it cannot “know” everything about all its possible internal states.

  49. speaking of dalton, he’s credited as the discoverer of color blindness, and there’s an interesting quote at http://en.wikipedia.org/wiki/John_Dalton#Colour_blindness where he does his best to communicate what “red” is (or more to the point isn’t) like for him: “that part of the image which others call red appears to me little more than a shade or defect of light. After that the orange, yellow and green seem one colour which descends pretty uniformly from an intense to a rare yellow, making what I should call different shades of yellow.”

  50. JessicaBoxer Says:

    > It leads to something like the Turing test that is considered by some as a gold standard of intelligence.
    > But is in fact nothing of the sort, as is evidenced by the many entries offered up for the test, which are
    > a tangled web of heuristics.. The Turing test is really more a test of knowledge of the experience of the
    > human condition, and the ability to understand human language rather than any measure of intelligence.

    > However, you could disagree and argue that intelligence encompasses the understanding of the human
    > condition too. And you would have no more claim on defining intelligence that I do. So we might as well
    > fire up the espresso machine. We have a long night of pointless arguing ahead of us.

    It’s interesting that the Turing test was designed precisely to avoid this. Turing avoided defining terms like “think” and “intelligence” in his paper because of all the “pointless arguing” that would result (I’m basing this on hearsay; I haven’t read the paper). He asked a much more answerable question: “Can a machine act intelligently enough to fool someone into thinking it was human?”

    He thought, I suppose, that any machine capable of fooling a judge under a no-holds-barred barrage of any question under the sun would just *have* to be intelligent in any reasonable sense of the word.

    Personally, I think this is reasonable, as long as you realize that the judge is supposed to be able to ask the computer *anything*. S/he can ask the computer about current events, what the computer is feeling, what jokes the computer has heard lately. S/he can tell the computer a joke and then ask the computer to explain it back, in as much detail as you please. It’s hard to imagine the computer being able to do all that and not be “intelligent”.

  51. >Here is the question that one really needs to ask about philosophy: does it predict anything useful?

    No. It’s not a philosopher’s job to make predictions. That job belongs to scientists.

    The philosopher’s proper work is the theory of theory-building. Like mathematicians, philosophers deal not in what is contingently true but what is necessarily true.

    “Does Mary learn anything when she first sees something red?” is a philosophical question because it inquires into what we mean by “learning” and the relationship of experience to learning. The way the question is posed leads to questions about the observed capabilities of mind can be accounted by a purely physicalist model.

    There is a great deal of bad, shallow philosophizing out there. That does not mean all philosophizing is bad and shallow.

    >However, when one considers philosophy in a fairly strict sense that is to say questions of metaphysics and epistemology, and perhaps include ethics or morality I’d argue that not much of use has come out of that at all.

    One very important thing we have learned (though many philosophers still resist the knowledge) is that once you’ve understood how to do sound epistemological analysis, metaphysics disappears. Or to put it another way, one of the wonderful things about the predictive definition of truth value is that it doesn’t care what metaphysical presuppositions you walked in with – the predictive criterion will steamroll right the fuck through ‘em and destroy any metaphysical premises or categories that don’t actually support a theory with an operational payoff.

  52. Desmond: Go ahead and read the original paper. It is very readable; it is a philosophical paper written in English, not a math paper. While it is not perfect, it is certainly one of those papers that has more misinformation about it than information being spread, and you should read it for yourself without somebody intermediating the interpretation for you.

  53. ESR wrote:
    > No. It’s not a philosopher’s job to make
    > predictions. That job belongs to scientists.

    Thanks for pointing out my error, you are completely right. Perhaps you would allow me to restate my original point better: The question to ask about philosophy is this: does it *produce* anything useful?

    > The philosopher’s proper work is the theory of
    > theory-building.

    That is an interesting point of view. I hate to contradict someone so evidently smart and well read as you on something like this, however let me offer my humble point of view. Philosophy, from what I understand anyway is largely about explaining meaning, purpose, value, rightness and so forth rather than meta-theory (or perhaps meta-meta-theory) as you seem to suggest. That is surely in the domain of mathematics. I will grant you that philosophy does tend to have fairly soft edges, often everything is thrown in and called philosophy. so in a sense some parts of mathematics do end up in there. BTW, I find it ironic that philosophy, which I have accused of being often futile due to the inability to define words, has the problem that the discipline itself is rather hard to define.

    > Like mathematicians, philosophers deal not in
    > what is contingently true but what is necessarily true.

    And they differ how? Remember, they both use a lot of funny Greek letters. (Well Socrates and Aristotle did anyway.)

    > The way [Mary's room] is posed leads to questions about
    > the observed capabilities of mind can be accounted by
    > a purely physicalist model.

    As I said, in this particular case I don’t think it does lead to that at all. All it leads to is a conclusion that words are imprecise, and nobody can really agree on what “learn” means. Regardless of what we agree about the word, still learning, or something readily confused with learning, of some nature takes place. I find the spooky interpretation to be more one of deliberate obfuscation by confusing semantics, much like a side show card trick. “Where is the lady, where did she go?” “Did she learn? Yes, well she knew it already! No, well she thinks she learned.” It is all a meaningless dance around what the words mean, and brings no real light on the the physical structure of the mind. All it seems to me to reveal is the weakness of the discipline trying to explain it. It calls for no such doubts about physiology or pedagogy. The spooky interpretation guys need to come up with something a lot more convincing and less tricky than that.

    > One … thing we have learned … once you’ve
    > understood how to do sound epistemological
    > analysis, metaphysics disappears.

    If that is the best we have learned it doesn’t seem much to me. In fact it sounds a whole lot like inside baseball.

  54. >The question to ask about philosophy is this: does it *produce* anything useful?

    I think it can. If nothing else, there are certain kinds of errors and inconsistencies in theory construction that you can avoid with good philosophical analysis.

    >Philosophy, from what I understand anyway is largely about explaining meaning, purpose, value, rightness and so forth rather than meta-theory (or perhaps meta-meta-theory) as you seem to suggest.

    The difference between these descriptions is, I think, one of emphasis rather than actual content. My formulation reflects an emphasis on epistemology, confirmation theory, and metamathematics. Yours centers more on moral philosophy and esthetics – and I’m dubious whether esthetics belongs in the “philosophy” category at all, I suspect it’s there only by historical accident and should really be a branch of psychology.

    The abstractions you’re talking about are all emergent from attempts to unify theories formed in very different domains of experience; the questions they raise require a meta-theory that relates (for example) “correctness” in a predictivist sense with “correct” used of ethical behavior. When you say “nobody can really agree on what ‘learn’ means”, what you are actually pointing out is that we have a meta-theoretical problem disguised as a simple language issue. The difficulty in agreement comes exactly from the fact that people carry around different theories of what “learn” means, each theory with different metaphysical and ontological premises hidden in it. The philosopher’s job is to expose and (if possible) reconcile those premises.

    >And they differ how?

    Mainly in that mathematicians have a much, much easier job. The theories they analyze are formalized and relatively tractable.

  55. I think the construction of the experiment is flawed.

    When Mary walks out of her room, regardless of the response she makes, if she’d never seen “color” (never mind red) before, then neural pathways that were never active before will suddenly be activated. Whatever mental or emotional event occurs after that is still caused by the physical operation of the brain.

    So unless she found a way to stimulate the optic nerve while in her room studying, she never fully grokked “red”. And if she DID stimulate the optic nerve by other means, is that really functionally different from experiencing the reflection of the red wavelengths of the EM spectrum?

  56. I find Mary’s Room to be another example of “philosophers” setting up straw men so that they may enjoy knocking them down. They conveniently ignore Mary’s knowledge of her own body and its functioning (has she never cut herself and bled red blood?), including the receipt of color information from the eye. That no other brain is connected to her optic nerves does not exclude either her brain, the nerves, or the eyes from reality. It merely means that there can be no independent confirmation of what her brain receives from those nerves.

  57. > When Mary walks out of her room, regardless of the response she makes, if she’d never seen
    > “color” (never mind red) before, then neural pathways that were never active before will
    > suddenly be activated. Whatever mental or emotional event occurs after that is still caused by
    > the physical operation of the brain.

    Actually, this is a bit of an interesting question in brain development. If the colour-receiving parts of the eye are never stimulated, does Mary in fact *have* a colour-responsive pathway in the brain? We do know that human brains are quite plastic and develop in response to the stimuli they get. Perhaps, in a black-and-white world, you would get a black-and-white vision processor. Against this, colour response evolved long before anything approaching human intelligence, and so is presumably hardwired into the primate rather than human parts of the brain. I’m not sure how plastic that part is. It would be a fascinating experiment, at any rate.

  58. >Actually, this is a bit of an interesting question in brain development. If the colour-receiving parts of the eye are never stimulated, does Mary in fact *have* a colour-responsive pathway in the brain?

    This is one of the reasons the experimental evidence from the color-blind synesthete is so interesting. It confirms what we’d guess from neurology, which is “yes”. His color pathways are activated by colors he has never seen with his retinas.

  59. That’s interesting, but not necessarily conclusive. Our synesthete is, if I understand correctly, not fully colourblind; he can see some colours but not others. It follows that this part of the brain would develop in response to the colours he can see, and thus be present for activation in unfamiliar ways by the synesthetic miswiring. And at the present level of technology and ignorance about substrates, there’s no way for us to tell whether his “Martian colours” are anything like what the rest of us perceive that he cannot get through his retinas. I wonder if the experiment can be repeated with a fully colourblind, synesthetic subject? But the difficulty then is, how would he know that he’s seeing “colour”?

  60. One of the other sources we can turn to for information about this phenomenon is the experiments that people do with adding senses to their body, such as the magnet in the finger or the feelSpace belt. It appears that the brain, even as an adult, is extremely plastic, and even things that we do not have any argument for having gene-based brain pathways for, such as the magnetic sense, rapidly get integrated into the body’s self-image and become full senses, complete with the brain’s ability to extract remarkably amounts of signal from relatively shoddy or fuzzy data.

    In fact, reformulate the Mary’s Room problem as “If we give Mary a magnetic-field sense that she didn’t have before, even though she understood everything about magnetic fields before, and is there a qualitative difference?” and you no longer have to theorize. Just ask the guys that lived that experiment. It’s equivalent in every essential way I can think of, except that it can be actually done, and has been actually done.

    ESR says: Excellent point. Cuts right to the core of the matter.

  61. Jaron Lanier liked to tell a story about giving people lobster or other multi-legged avatars to control in his VR experiments, breaking down input from the body parts they had in order to control their virtual arthropod legs. The participants became surprisingly dextrous with their new legs.

  62. One of the other sources we can turn to for information about this phenomenon is the experiments that people do with adding senses to their body, such as the magnet in the finger or the feelSpace belt. It appears that the brain, even as an adult, is extremely plastic, and even things that we do not have any argument for having gene-based brain pathways for, such as the magnetic sense, rapidly get integrated into the body’s self-image and become full senses, complete with the brain’s ability to extract remarkably amounts of signal from relatively shoddy or fuzzy data.

    and

    Jaron Lanier liked to tell a story about giving people lobster or other multi-legged avatars to control in his VR experiments, breaking down input from the body parts they had in order to control their virtual arthropod legs. The participants became surprisingly dextrous with their new legs.

    I actually think this says a lot about ESR’s main point, mainly that Mary’s “Wow” is a result of the qualia of having those pathways in the brain light up.

    There’s a lot that’s not understood about the brain, but there is one phenomenon that is well-understood: the brain is capable of making up “neural nets.” This is literally what happens in the process called “learning.”

    Let’s take learning billiards (“pool”) as an example, only because it’s an easy example to work with as the mechanics are very straightfoward and I assume most of you have at least a passing familiarity with the game. You take your first shot, and the typical n00b jabs the cue ball without proper follow-through and has way too much chalk on the tip of his cue stick, and the ball takes a seemingly random direction and the n00b looks on in disbelief as the ball seems to defy all physics.

    Then an experienced player might typically show him the right way to make bridge with his hand, how to chalk his cue properly, etc., and then explain that shooting (like so many other games involving balls), is all about follow-through and aim. The n00b, after a few tries on a practice table, starts to “get” it — his strokes become more smooth, he’s able to shoot some easy more-or-less straight in shots. As he does this more and more, his brain builds neuro-pathways that give him his “pool legs” — an additional “sense” about billiards, if you will.

    As this happens, and the n00b becomes better and better through much practice and experimentation, more and more neural nets are built in his brain for playing billiards.

    Do these neural nets necessarily have to be about learning “software applications” for the existing “hardware”? No, not necessarily. It’s obvious that if you provide the brain with new hardware, the brain can form neural nets to drive that new hardware. (Look into recent research developments related to aiding people with disabilities as an example.)

    The brain is very versatile and can learn many, many things. Adding additional senses or limbs is possible because of the brain’s ability to essentially rewire itself.

  63. Jeremy Bowers Says:
    > the magnet in the finger or the feelSpace belt.

    What a very interesting contribution, thanks Jeremy. I have often wondered what it would be like to have other senses, and this goes some way to helping understand. However, I did want to comment on how this applies to Mary’s room. If your read the article, both of which are quite fascinating, you will recognize that these explorers did not add new senses at all, they hijacked an existing sense, and added a peripheral. In both cases the data was communicated to the brain using the existing touch sense. It is very interesting to see how their brain rolled up these new sensations into higher level abstractions: the feelspace guy did not just feel a buzz on his waist, his brain rolled it up into a higher order understanding of space, so too with the magnet guys, though perhaps less so.

    This is a fascinating experiment from the point of view of cognitive science, and, to a lesser extent neuroscience, however, I would argue that it really has little to do with philosophy at all. Philosophy searches for meaning and purpose and this doesn’t seem to me to have much to do with meaning at all. The question that Eric originally raised was over the hoary old question of whether your mind can be explained and described in purely physical terms, or is some sort of spooky non corporeal thing in play. These experiments don’t seem to add anything in that regard, since the physiological process by which they work is pretty straightforward. It is fascinating to see how that great abstraction machine that we have between the ears can pattern out physical sensations into higher order concepts, but I see no arguments for or against spookiness in these experiments.

    The only place where philosophy seems to play a role here is to recognize that “understanding” is largely a filtration and synthesis process converting sensory inputs into recognizable, repeatable patterns (often times by hiding details, or ignoring minor variances), and that these patterns offer a predictive power. And I suppose there is some value in that idea.

  64. Humans have introspective capabilities, making Dennet’s position meaningless.

  65. The only place where philosophy seems to play a role here is to recognize that “understanding” is largely a filtration and synthesis process converting sensory inputs into recognizable, repeatable patterns (often times by hiding details, or ignoring minor variances), and that these patterns offer a predictive power. And I suppose there is some value in that idea.

    I disagree with this statement. Philosophy != mysticism, and I think you are making the common mistake of conflating the two. Philosophy is all about building theories. It’s a meta-theory of theories. Philosophy attempts to solve fundamental problems of things like existence, truth, validity, beauty, etc. It is contrasted with other methodologies of solving these problems such as mysticism because it uses (or at least attempts to use) reasoned arguments.

    Using only empircal science to answer these questions is a philosophy — it is the philosophy of empiricism, perhaps with elements of rationalism. When ESR states that he is a ‘physicalist’, that’s just a type of empiricism, essentially, that generally focuses on the mind or mind/body connection, and says that no matter what, for every characteristic of x, there is an associated physical property y.

    When ESR attacks the “mysterians,” he makes it seem like there are only two choices: physicalists and mysterians. This is a false dichotomy. There are, in fact, a wide spectrum of various philosophies that would generally stand in contrast to physicalism, such as idealism or existentialism. To say or imply that these branches of philosophy are not based on rational thought because they stand in contrast to physicalism is perhaps a bit intellectually dishonest, but such debates are common among philosophers.

    What, for example, makes something “beautiful”? The physicalist might say “beautiful” is just another way of saying “well-ordered” or “appealing to instinct derived from hundreds of thousands of millennia of evolution” or whatever, but then I can think of things that I would count as “beautiful” that are neither of those. *shrug*

    I’m not saying I entirely disagree with physicalism, nor am I saying that I agree with existentialism or idealism. I just think it is false to think physicalism is not a philosophy and it is false to think that science stands in contrast to philosophy. Quite the contrary, the two often work hand-in-hand.

  66. “you will recognize that these explorers did not add new senses at all, they hijacked an existing sense, and added a peripheral. In both cases the data was communicated to the brain using the existing touch sense.”

    I acknowledge that, but let me at least argue the contrary even as I don’t know if I entirely believe it. Arguably, it really is a separate sense in every way that matters. All that comes in to the brain is a mass of nerves and their firings. There’s no fundamental separation between vision and hearing, or taste and vision (forgot that one went by a couple of days ago!). The brain has to learn that separation, just as it has to learn which nerves go where. We perceive a rigid separation because there is little correlation between the different senses and so the brain naturally partitions them quite strongly in most cases, but that’s the data, not the brain or the senses.

    (Synesthetes do not fully learn this distinction; it would be interesting to know whether that’s because the brain is somehow unable to do so, if the signals are physically crosstalking such that they are in fact too correlated for the brain to pull apart, or, most likely, some other more complicated mechanism neither of these two English phrases really correctly describes. But this is an interesting counterpoint and part of why I’m not sure I completely buy the argument I’m making in this comment. IIRC nobody quite knows the answer to this but if I’m wrong please let me know.)

    Extracting from the “touch” channel a new, uncorrelated signal is indeed different than burning in truly new pathways (patience! we’ll conduct that experiment soon enough), but I don’t know that you can get a rigid yes/no distinction out of it; it may be more similar than different.

    I look forward to someday owning some convenient, high-res EEG gear and perhaps a rig I can set up to send signals to my fingertips from a computer program. I want a rig that reads my brainwaves and does something to give me tactile experience in a four-dimensional space (small shocks correlated to touching a hyperplane or something); I want to see if I could develop an intuition for four spatial dimensions. If I could use such a rig to train myself to actually think in four dimensions, I think it would offer further evidence about this conundrum, because goodness knows that for all my abstract, pre-frontal-cortex knowledge about how four dimensions work, I’m hardly any better than anybody else at it.

  67. morgan greywolf Says:
    > Philosophy != mysticism, and I think you are
    > making the common mistake of conflating the two.

    I make plenty of mistakes but not this one. I never suggested that philosophy = mysticism; in fact in some respects philosophy is man clawing his way out of mysticism.

    > When ESR attacks the “mysterians,” he makes it seem like
    > there are only two choices: physicalists and mysterians.
    > This is a false dichotomy.

    I don’t agree. A philosophy based on the idea of simple physical cause and effect is something that can be demonstrated with the physical sciences, that is to say, the very same physical sciences that have produced all the richness of life we enjoy today. As soon as you get beyond what is clearly evident to the senses by science and introduce some additional, non-measurable, non corporeal force then the burden falls heavily on you to provide some reason to believe it. So to put it another way the “dichotomy” you talk about is between reality and spooky nonsense. Nonetheless, you are correct that nonsense comes in many flavors.

    In regards to existentialism, all I can say is that both Nietzche and Dostoyevsky make me want to blow my brains out. So it isn’t a particularly productive philosophy. (Also, can someone explain why every existentialist has a name that is REALLY difficult to spell?)

    > What, for example, makes something “beautiful”?

    I tend to agree with Eric on this one: aesthetics really has very little relationship to other branches of philosophy. Except insofar as it is a great thing to discuss when high, or caffeinated and the dizzy blonde chicks really dig it. Nonetheless, as I argued earlier, philosophy has such fluffy boundaries, who knows what can fairly be included in that grab bag.

    How the heck did I end up writing this comment? I started out saying that I thought philosophy discussions were largely pointless trips to Jupiter, and now I am not just discussing philosophy but meta-philosophy? “If you look to the left ladies and gentlemen you will get a perfect view of the dwarf planet Eris.”

  68. >When ESR states that he is a ‘physicalist’, that’s just a type of empiricism, essentially, that generally focuses on the mind or mind/body connection, and says that no matter what, for every characteristic of x, there is an associated physical property y.

    True, though not quite how I would put it. I would say instead that all the information in a mind is implied by the arrangement of matter and energy in the brain. The difference from your formulation is that it avoids a long excursion into what constiutes a ‘property’.

    >This is a false dichotomy.

    Nope, I’m with Jessica on this one. You can only be a physicalist, a mysterian, or in denial – what I mainly admire Dennett for is bulldozing through a great deal of bullshit and equivocation on this issue from people with a fervent desire to fool themselves that there’s a “third way”. Metaphysical idealism is nonsense instantly falsified by experience (there are some subtler kinds of idealism that aren’t obvious nonsense, but they all turn out to be mere word games that reduce to conventional realism once you’ve untangled the terminology). And existentialism has nothing to say about the mind-body problem.

    Mind you. physicalism doesn’t enable the claim that we understand the physical embodiment of mind, or even that we necessarily someday will – could be the problems are computationally intractable. Nor does it rule out the possibility of spooky quantum-entanglement effects – it’s more a form of sytematic skepticism against easy, stupid vitalistic handwaving.

    >Using only empircal science to answer these questions is a philosophy

    That’s trivially true, but no other “philosophy” is compatible with sanity. What else are you going to put in your ontology – alien space bats? All explanations other than empirical-scientific ones reduce at some point to “and then, a miracle occurs”. And then you’re done – you can’t reason past that point, because you just threw causality out the window,

  69. >I don’t think “reductionist” is a very interesting word, anyway. It seems to function mainly as a term of abuse directed at people the speaker thinks are too physicalist, or not willing to bow before the numinous quality of epiphenomena, or something like that

    You make a good point about trying to cut down on philosophy, a lot of it is bullshit, so yes, I agree, we need to try to cut out ‘isms’ and get precise definitions.

    I won’ t use the vague term ‘non-reductionism’ then, let me give the precise term ‘ontological object hierarchies’ and a precise definition: ‘reality has different hierarchial levels of description, where properties of objects on higher levels cannot be derived (even in principle) from those on lower levels’ ;

    We were discussing Yudkowsky’s claim to have eliminated object hierarchies in physics and reduced everything to configuration space; I pointed out he couldn’t possibly have done this, since we know that relatively contradicts quantum mechanics, which in turn contradicts classical physics. If Yudkowsky could solve these puzzles, he’d win the Nobel prize (along with all the funding and status he would ever need for SIAI research).

    In practice three levels are used in pure physics – this is not some airy-fairy philosophy, we are talking clear mathematical/physical principles here, with real experimental and theoretical consequences;

    (1) The field level (general relatively) needs an as yet unknown ‘bridging law’ to move to –
    (2) The mechanics level (quantum mechanics) needs another ‘bridging law’ (derivation of Born probabilities) to move to-
    (3) Classical level (particles/objects in space)

    So why can’t Yudkowsky (or anyone else) reduce these three levels to a single one, and eliminate the bridging laws? I put it to you that it can’t be done, because even pure physics *really is* divided into three irreducible levels, and we can expect that will be many more irreducible levels on the way up to a high-level thing like the mind.


    Yes I do claim to have demolished the reductionist world-view, further I claim to have succeeded in grasping the deep structure of reality. Not all philosophy is nonsense (although yes a lot of it is). My position is closest to that of Kant, who argued for categories built-in to the structure of minds in general, categories that underpin and are logically necessary for all thought. I claim to have discovered exactly what these categories really are, and even to have achieved direct conscious awareness of them! (I bullshit you not, they blaze like the sun – I’m NOT bluffing)

    My theory has 27 fundamental categories of thought (though the real categories are not the same as Kant’s), and these exactly correspond to the aforementioned irreducible levels of organization – I believe that reality is separated into 27 fundamental levels. These big claims would require books and papers to try to justify (someday perhaps!)

    But meantime, if anyone wants to look into these wild ideas, I recommend the works of David Bohm; many of his ideas appear similar to my own; Bohm was a top-rate physicist by the way, not an airy-fairy philosopher, there’s a marvelous volume of papers you can read called ‘Quantum Implications’ (Hiley, Peat, Routledge Books) , should bring readers up to speed on his ideas, and yes I do think that an extended version of his pilot-wave theory is the correct interpretation of QM.

  70. esr, I like your treatment of the issues in this article. My only question is to what extent Dennett would actually disagree with you. I’ve read his detailed discussion of this thought experiment, and I think he would zero in on the following statement you make:

    [quote]Here is my physicalist account of Mary’s “Wow!” What she learns is *what it feels like to have the color-processing pathways of her brain light up*.[/quote]

    I think Dennett would ask how you can be sure that Mary doesn’t already know what it feels like to have the color-processing pathways of her brain light up. After all, by hypothesis, she knows *everything* about how color processing works in the brain, including what effects lighting up her color-processing pathways will have on other areas of her brain, such as the ones in which “what it feels like” is instantiated. And of course she also knows what further effects that particular excitation of her “what it feels like” brain regions will have, such as causing further activity in her speech area that will cause her to say, “That apple is red” or something similar.

    You appear to be saying that no amount of that kind of knowledge can possibly be the same as the experience of actually having her color-processing pathways light up–but that’s precisely the point at issue. I don’t think we’ll know for sure until we can actually run experiments that test how much leverage, if any, detailed knowledge of how the brain works can give people in imagining (simulating) what various experiences will be like. (I pass over the question of whether “simulating” an experience is really the same as having it in the “normal” way, because the final test will be perfectly objective: either people in situations like Mary’s *will* actually say “Wow!” and admit that they didn’t really know what it was like in advance, or they won’t.)

    But I think Dennett would agree with your more fundamental point, expressed here:

    [quote]One’s own mental life – or, to put it physicalist terms, one’s perception of one’s own brain states – is part of the phenomenal field just as genuinely as Husserl’s copper ashtray is.[/quote]

    I’m not sure what in Dennett’s writings has led you to think he doesn’t believe this. In _Consciousness Explained_ he explicitly includes one’s inner life as part of the “phenom”–the whole menagerie of first-person experience that any theory of mind must explain. (I don’t have my copy of the book handy so I can’t give a direct quote.)

  71. After all, by hypothesis, she knows *everything* about how color processing works in the brain, including what effects lighting up her color-processing pathways will have on other areas of her brain, such as the ones in which “what it feels like” is instantiated.

    The hypothesis asserts that she has “all the physical information” about how vision works, but she knows these things as learned facts, but not as experienced sensation. Epistemologically, there is a difference between facts known by direct observation and those learned from others. As an obvious example, I know certain things about The Bride of Monster via direct observation. ESR may tell us things about his wife, but I don’t know those things; I know that he said them. I accept their truth conditionally, because I’ve never seen any reason to disbelieve what he says, but I cannot say I “know” those things in the same sense as I know things from personal experience.

  72. >I think Dennett would ask how you can be sure that Mary doesn’t already know what it feels like to have the color-processing pathways of her brain light up.

    That depends. Are we in fact supposed to assume that Mary says “Wow!” or not?

    I think Dennett can reasonably maintain that Mary already knows that feeling only if she doesn’t indicate that she’s having a novel experience. So he can propose that she won’t say “Wow!” on first seeing red, but not that if she says “Wow!” it is meaningless – which is what the article describes as his position.

  73. > My theory has 27 fundamental categories of thought (though the real categories are not the same as Kant’s), and these exactly correspond to the aforementioned irreducible levels of organization – I believe that reality is separated into 27 fundamental levels. These big claims would require books and papers to try to justify (someday perhaps!)

    Could you list these “27 fundamental categories”?

  74. I am not a philosopher. Nor do I play on on television or the Internet.

    I am, however, someone who is both mildly synesthetic, and who was blind for the first eight years of my life from congenital cataracts. The best way to describe my vision

    I can tell you from comparisons with other people, and experiments conducted in neurology labs, that my visual perceptions are a bit different from the middle of the bell curve of the sample sets.

    1) I did not learn to ‘see’ (build the neural pathways) at the same age everyone else did. In particular, even with attempted correction on my left eye, it’s unlikely that I will ever develop true binocular vision. My ability to spot objects against visual clutter is bad enough that it’s casually noticeable.

    2) I do not see IR reflectance as well as most people do. (Most people, if you point a Wii-mote at the right surface, can catch the spot of IR light backscatter in a dark room. I can’t.)

    3) I am MUCH more dependent upon my inner ear for kinesthetic cues than most people are; I literally cannot watch first person shooter games without getting motion sick within a few minutes, because the difference in signals between visual and inner ear sensations.

    4) I very vividly remember the fact that my parents faces were about 1/4 the size I expected them to be when the bandages were taken off; my prior frame of reference had been touching them.

    5) I a appear to be mildly synesthetic – I see an overlay of information on objects in motion that’s best described as blurry arrows showing where they might be and what their probable range of motion is. I still lack binocular vision – which means that judging time and distance is tricky.

  75. esr wrote:

    Metaphysical idealism is nonsense instantly falsified by experience (there are some subtler kinds of idealism that aren’t obvious nonsense, but they all turn out to be mere word games that reduce to conventional realism once you’ve untangled the terminology).

    (Off-topic note: I don’t understand the strange aversion around these parts to using <blockquote> tags. The days of Usenet are long over. :-P)

    I guess it depends on what you mean by ‘metaphysical’. It seems to me that you have trouble with anything that can’t be described in pure, empiricist terms. The big one most empiricists have trouble with is “love.” Most empiricists, upon encountering this concept, will simply say it either doesn’t exist or that it is a myth (or some other sort of delusion). Which, incidentally, is why Hallmark doesn’t hire empiricists. :) Anyway, all of the arguments most empiricists will use when discussing “love” all boil down to one of these two statements, both of which can easily be proven to be false.

    As for quantum entanglement: it’s been shown that the brain uses quantum effects. Quantum physics opens up some interesting possibilities. After all, once we decide that everything is mostly empty space with a few zero mass particles blinking in and out of existence in a probability cloud, then we have to accept that there is nothing concrete — all things are only tendencies.

  76. The Monster wrote:

    The hypothesis asserts that she has “all the physical information” about how vision works, but she knows these things as learned facts, but not as experienced sensation. Epistemologically, there is a difference between facts known by direct observation and those learned from others.

    What facts, exactly, do you know by direct observation when you experience something–say, for concreteness, seeing red? Do you know by direct observation that your brain is in a particular state? Surely not–it’s only a huge body of scientific knowledge (knowledge we don’t fully have yet) that would let you deduce that your brain must be in some particular state (or one of some set of states) from the fact that you are seeing red. That being the case, the *connection* between your seeing red and your brain being in a particular state is also something you can’t know by direct observation. So you have no advantage in that respect over Mary.

    esr wrote:

    I think Dennett can reasonably maintain that Mary already knows that feeling only if she doesn’t indicate that she’s having a novel experience.

    I agree, and I think Dennett would agree too. The Wikipedia article may not be getting across his position very well; from reading Dennett’s own discussion of this topic, I think his position is not that it would be meaningless if Mary did say “Wow!”, but only that the thought experiment (at least in the version he discusses) is begging the question by asserting without argument that Mary *would* say “Wow!”, when in fact it’s perfectly possible, on our present knowledge, that she wouldn’t.

    Btw, an online draft version of one of Dennett’s papers on this topic (not the exact one I have in book form at home) is at
    http://ase.tufts.edu/cogstud/papers/RoboMaryfinal.htm. (I see on looking again that it’s the second general reference linked to in the Wikipedia article.)

  77. # Morgan Greywolf Says:
    > The big one most empiricists have trouble with is “love.”

    Really? I don’t see why it is a problem at all, except insofar as it is the usual obfuscation and confusion raised to the power ten (as emotions always do.) The word love means dozens of different things in English, Consider the following statements:
    He loves ice cream,
    She loves the guy I met last night,
    They love their children,
    She loves her husband of thirty years,
    We made love to the sound of the rain,
    I love God,
    God loves you.
    I just wuuvvv you to deaffff my wittle snookie wookie wuvvv bunny.

    In each of these the word love means something quite different. So unless you can explain which kind of love you are talking about you certainly can’t explain its physical manifestation. Nonetheless, all of these can readily be described in terms of developed habits, preferences, behavior patterns and hormonal states of your body. Perhaps you can explain why you disagree with that point of view? What spooky component do you need to add?

    > Most empiricists, .. say it either doesn’t exist or that it is a myth

    Aren’t these two dismissals the same? Nonetheless, I can’t speak for most, but I certainly think it exists, just that it doesn’t exist as some force outside of the normal cause and effect forces of the physical universe. Cupid’s arrow is made of atoms not scotch mist.

    > As for quantum entanglement: it’s been shown that the brain uses quantum effects.

    I don’t think “it has been shown” at all. (Though it depends on what you mean, in a sense a covalent chemical bond is a quantum effect.) However, I imagine you are referring to Roger Penrose’s work here. I am not an expert on it, (I can barely grok Newton, never mind QM) but what I always found interesting about Penrose is more how he defended his work rather than what his work said. This is especially evident in his debate with Ray Kurweil. Penrose seemed to be very uncomfortable with the whole consequences of an understanding of the mind as a physical artifact (being uncomfortable, for example with the idea of transferring the mind to an electronic substrate), and his theory is almost like a physicalist excuse to have spooky effects. That is to say, he wants to be a physicalist without accepting the very real consequences of physicalism, and deals with the problem by introducing spooky quantum effects that Kurzweil can’t implement on his PC. It feels like religion with scientific words.

    Perhaps I do the guy an injustice, and I certainly can’t say there is no quantum effects in the brain, but from what I read, most of the way the body works, including the brain, is explicable at an molecular level without the need to resort to probabilistic explanations. Dopamine and RNA molecules are, after all, pretty big.

  78. >Most empiricists, upon encountering this concept, will simply say it either doesn’t exist or that it is a myth (or some other sort of delusion).

    *What* *the* *fuck*?

    What sort of vulgar parody of empiricism have you been rolling in, man? Empiricism is just a theory of knowledge which asserts that knowledge arises from experience and sensory evidence; an empiricist is required to treat the word “love” as he would any other linguistic abstraction tied to observable behaviors. Nothing in the theory requires an empiricist to consider “love” any more mythical than (say) electrons.

    >It seems to me that you have trouble with anything that can’t be described in pure, empiricist terms

    Nope. I have trouble with any theory that can’t be tested by observable consequences.

    Please think about the difference between your formulation and mine. Carefully. Notice that mine doesn’t require any metaphysical posits about what’s “real” or not. It doesn’t require that I know a difference between “mind” and “matter”; an observable can be anywhere in my phenomenal field. It doesn’t require a prior theory of knowledge such as “empiricism” or “rationalism” (Let these two asses be set to grind corn!)

  79. Nothing in the theory requires an empiricist to consider “love” any more mythical than (say) electrons.

    True, but can “love” be described in purely physicalist terms? Not as a linguistic abstraction, but as a concept? If “love” is just some observable behavior, then how can it be recognized? Why is that no one can seem to describe love to someone who has never experienced it? I posit that it cannot be described in purely physicalist terms. I’m not saying that makes it mythical.

    Please think about the difference between your formulation and mine. Carefully. Notice that mine doesn’t require any metaphysical posits about what’s “real” or not. It doesn’t require that I know a difference between “mind” and “matter”; an observable can be anywhere in my phenomenal field. It doesn’t require a prior theory of knowledge such as “empiricism” or “rationalism” (Let these two asses be set to grind corn!)

    Okay, I see what you’re saying. Yeah, that is actually a pretty big gap, and I’m making the mistake of conflating physicalism with empiricism in general.

  80. >True, but can “love” be described in purely physicalist terms?

    This is exactly equivalent to the physicalism question about minds, because love is a behavior of minds. It’s just a more emotionally loaded way of posing the question.

  81. esr Says:
    > This is exactly equivalent to the physicalism question about
    > minds, because love is a behavior of minds.

    Just to nit pick, love, or some meanings of the word love pertain to more than the mind. Emotions in general often have a significant hormonal component, manifesting in various physical feelings and reactions. (I am not referring to sexual arousal here, though that is an example, but other reactions, like feeling like throwing up, butterflies, increased pupil size, release of pheromones and so forth. In the case of other emotions, such as fear or anger, release of hormones cause the heart to beat faster, the blood vessels to dialate, the release of energy from the muscles and so forth.)

    All are, of course, physical manifestations resulting from the stimulation of neural patterns, which cause various bodily organs to release the aforementioned chemicals. These chemicals then react with the cellular mechanisms in various parts of your body (carried in the bloodstream usually), which cause the physical symptoms. Nothing spooky here.

  82. >Just to nit pick, love, or some meanings of the word love pertain to more than the mind.

    I did not intend “mind” as a synonym for “intellect”.

  83. Hmmm…didn’t see this earlier:

    I don’t think “it has been shown” at all. (Though it depends on what you mean, in a sense a covalent chemical bond is a quantum effect.) However, I imagine you are referring to Roger Penrose’s work here.

    Actually, I’m NOT talking about Roger Penrose, whose claims are … extraordinary. (As in extraordinary claims require extraordinary proofs). I’m talking about Efstratios Manousakis, a physics professor at Florida State University, who talks about actual testable quantum effects in the brain (and not just covalent bonds) in this paper written about 2 years ago. Here’s the abstract:

    A theory of the process of perception is presented based on von Neumann’s quantum theory
    of measurement and conscious observation. Conscious events that occur are identified with the
    quantum mechanical “collapses” of the wave function, as specified by the orthodox quantum theory.
    The wave function, between such perceptual events, describes a state of potential consciousness
    which evolves via the Schr¨odinger equation. When a perceptual event is observed, where a wavefunction
    “collapse” occurs through the process of measurement, it actualizes the corresponding
    neural correlate of consciousness (NCC) brain state. Using this theory and parameters based
    on known values of neuronal oscillation frequencies and firing rates, the calculated probability
    distribution of dominance duration of rival states in binocular rivalry under various conditions
    is in good agreement with available experimental data. This theory naturally explains recently
    observed marked increase in dominance duration in binocular rivalry upon periodic interruption of
    stimulus and yields testable predictions for the distribution of perceptual alteration in time.

    So, IOW, yeah, the quantum effects have been shown in the brain.

  84. What facts, exactly, do you know by direct observation when you experience something–say, for concreteness, seeing red? Do you know by direct observation that your brain is in a particular state? Surely not–it’s only a huge body of scientific knowledge (knowledge we don’t fully have yet) that would let you deduce that your brain must be in some particular state (or one of some set of states) from the fact that you are seeing red.

    I directly experience my brain state. Discussing Mary’s brain state as we are doing, and her own “knowledge” of what messages would be sent to her brain when red light impinged upon her retina, are an entirely different sort of “knowledge” from the direct experience of what that red light does to the brain that receives those messages.

    That being the case, the *connection* between your seeing red and your brain being in a particular state is also something you can’t know by direct observation.

    I can read the numbers from a meter measuring the wavelength of the light, and I can read the outputs of my optic nerves. When I do so at the same time, and the numbers match the definition of “red”, I am indeed able to make that connection. I can correlate the perceptions of different colors with the numbers on the meter, and in doing so I have two independent means of understanding light wavelengths (which can serve as checks upon each other should one become unreliable).

    So you have no advantage in that respect over Mary.

    Until Mary is allowed to directly observe red light, I have the distinct advantage of being able to compare the effect on my brain of getting those signals from the optic nerves at the same time I can confirm by means of the meter what kind of light it is. I really don’t understand how this internal knowledge is waved away as if it doesn’t add anything to Mary’s knowledge.

  85. esr says:

    > I think Dennett can reasonably maintain that Mary already knows that feeling only if she doesn’t indicate that she’s having a novel experience. So he can propose that she won’t say “Wow!” on first seeing red, but not that if she says “Wow!” it is meaningless – which is what the article describes as his position.

    Where did you get this from? My understanding of the Wikipedia article points to Dennet saying that she won’t say “Wow!”, not that if she did it would be meaningless. I might have missed something, but I think this is keeping in line with his position in Consciousness Explained as well.

    Mary knows precisely what it’s like to see red because she knows everything (physical) there is to know about seeing red, including her behavioural dispositions towards the colour. So she wouldn’t say “Wow!”.

  86. The Monster wrote:

    Until Mary is allowed to directly observe red light, I have the distinct advantage of being able to compare the effect on my brain of getting those signals from the optic nerves at the same time I can confirm by means of the meter what kind of light it is. I really don’t understand how this internal knowledge is waved away as if it doesn’t add anything to Mary’s knowledge.

    And what is the effect on your brain of getting those signals? To put your brain into another state, the state of “knowing what it is like to see red” (for example). Call that state S, and the state of “seeing red” itself (the state you’re in while you’re “reading your optic nerve” as it’s stimulated by red light) state R. Then your claim is basically that having been in state R is necessary in order to get into state S.

    Now if Mary were a normal person like you and me, I would agree with you. For normal people like us, it’s simply not possible for us to get into state S, or anything close to it, without having first been in state R, and it’s not possible for us to get into state R except by the “normal” route of having red light shining into our eyes. But that may be just a contingent fact, a consequence of our ignorance of any of the physical details of states R and S and how they sit in our brains. Mary, by hypothesis, has *complete* knowledge of those physical details.

    Consider what that means: it means that, if she wants to describe state R, she isn’t limited (as we are) to just “pointing at it”, as it were–she isn’t limited to just saying that state R is “whatever state my brain is in when I see red”, which is pretty much all that you or I can say about it. Yes, we could give *some* details, such as picking out particular areas of the visual cortex that would light up in an fMRI scan, but that’s pretty paltry compared with what Mary can do. She can give a *complete* physical description of state R with as much micro detail as you like, along with all the functional specifications that explain why it is that state R subserves seeing red, as opposed to state G, which subserves seeing green, etc., etc. And the same goes for state S. Given that Mary has all that knowledge that we don’t have, it’s not obvious to me that she can’t get into state S until she is first put into state R. Or, alternatively, it’s not obvious to me that she couldn’t contrive to put herself into state R by some unusual means, *without* having any actual red light shining on her optic nerves.

  87. >Where did you get this from? My understanding of the Wikipedia article points to Dennet saying that she won’t say “Wow!”,

    Hm…on second reading, I see the report of Dennett’s position is phrased rather ambiguously. I interpreted it as “Mary will learn nothing, therefore the ‘Wow!’ Frank Jackson expects is meaningless.”

    Searching for “Dennett Mary’s Room” does not seem to turn up Dennett’s own analysis.

  88. > Searching for “Dennett Mary’s Room” does not seem to turn up Dennett’s own analysis.

    He wrote about it in Consciousness Explained (which I’ve read, albeit a long time ago).

    Frank Jackson has said that it is “just obvious that Mary will learn something about the world and our visual experience of it” when she sees a red tomato. Dennett retorts that the only reason this seems “just obvious” is because we haven’t imagined the thought experiment in enough detail. The idea of Mary knowing *everything* physical there is to know about seeing red is so Immense, so Huge, that we don’t even try; we just imagine that she knows lots and lots. But if Mary just knew lots and lots – if she just knew everything that today’s neurophysiologists know, for example – then *of course* she learns something new upon seeing a rose.

    Dennett then goes on to imagine Mary’s captors releasing her from her room and, for fun, showing her a banana painted bright blue. Mary isn’t fooled; she knew, in advance, how she would react to both yellow and blue, and hence she knew that she was being had.

    Dennett uses the same tack with Searle’s Chinese Room.

  89. >Dennett uses the same tack with Searle’s Chinese Room.

    I think his attack is very well-founded in that case.

    I’m going to need to read your cite a couple times and think about it for a while. Dennett does appears to be maintaining that that Mary will not necessarily say “Wow!” but might contingently say “Wow!” – I think I see why the Wikipedia summary is ambiguous. Unfortunately, Dennett is so set on carpet-bombing the territory that it’s not easy to construe his central point(s).

  90. What Mary learns on seeing red is what the experience of seeing red feels like, unless learning everything there is to know about red somehow includes the experience of seeing red, maybe by direct stimulation of the optic nerve, she will be learning something new. And if Mary says “Wow”, then she did not experience seeing red before.

  91. >Could you list these “27 fundamental categories”?

    I will post the basic list of categories, but not the definitions. Here then are my names for the 27 categories (drum-roll please):

    Symmetry, Transform, Field, Producer, Builder, Signaller, Substance, Structure, Design, Virtue, Freedom, Aesthetic, Actor, Decision-maker, Narrator, Meme, Project, Story, Ordering, Relation, Category, Deducer, Predictor, Categorizer, Operation, Program, Ontology

    Note that ‘Category’ is itself a category – here Category is referring to a *mathematical* category (in the sense of category theory). Believe it or not I think that these 27 categories, once properly defined, are sufficient to ‘capture’ ALL of reality (in the sense of acting as prototypes for ANYTHING that needs to be modelled). They are the ‘ontological alphabet’ of the universe.

  92. Note that ‘Category’ is itself a category – here Category is referring to a *mathematical* category (in the sense of category theory). Believe it or not I think that these 27 categories, once properly defined, are sufficient to ‘capture’ ALL of reality (in the sense of acting as prototypes for ANYTHING that needs to be modelled). They are the ‘ontological alphabet’ of the universe.

    That’s a pretty bold statement. How about aspects of reality that are not yet discovered? The universe is a very big place and humans have thus far experienced only a tiny fraction of it. Humans are not unlike Mary, sitting in her little black and white room, only, unlike Mary, we only know what we see in the little black and white room.

    To understand what I’m saying, you need to go someplace with a beautiful vista, like a mountainscape, on top of a mountain, or even a deserted beach (No, not a picture on the Internet, go there and experience it, and then observe and realize fully that you and your fellow man are naught but a tiny spec of nothing in the grand scheme that is the universe.

    There’s a difference between knowing something on an intellectual level and actually having the real experience of it. I think that’s the point of Mary’s Room.

  93. There’s a difference between knowing something on an intellectual level and actually having the real experience of it.

    But he does have experience of the categories which capture all reality — NO JOKE, THEY BLAZE LIKE THE SUN.

    All I can say is that must have been some mighty powerful LSD.

  94. > Hm…on second reading, I see the report of Dennett’s position is phrased rather ambiguously. I interpreted it as “Mary will learn nothing, therefore the ‘Wow!’ Frank Jackson expects is meaningless.

    Not to beat a dead horse, but…the Wikipedia article says this about Dennett’s position:

    “Daniel Dennett argues that Mary would not, in fact, learn something new if she stepped out of her black and white room to see the colour red”

    The paper cited by Peter Donis says this:

    “I [Dennett] was saying that Mary had figured out, using her vast knowledge of color science, exactly what it would be like for her to see something red, something yellow, something blue in advance of having those experiences”

    Unless there’s another, contradictory statement somewhere else that I missed, I don’t read this as ambiguous. In other words…no “wow!”, meaningless or not.

    OTOH, Dennett *has* said that under any sort of even semi-realistic circumstances, Mary would, in fact, learn something new upon seeing a rose because under any sort of semi-realistic circumstances, she wouldn’t even be close to knowing everything physical there is to know about colour vision. The thought experiment demands a form of omniscience as one of its basic premises, which Dennett sees as pure fantasy.

  95. Although the main topic is certainly interesting too, I’d rather like to take a look at the operationist, fallibilist definition of truth. I’m not sure what it actually means, but I think it means what you wrote a while ago: we are prediction-seeking animals because we seek to control our environment because we are goal-seeking animals.

    But if it is so, why define truth on the third level and not on the first, original level: truth is a statement that satisfies a goal? Which would make all lies, all propaganda truths if they help satisfying a goal, so it would be a very cynical position to take, but if the only reason to form true statements is to make predictions that help in controling our environment in order to satisfy our goals, then what is exactly the value-free rational argument against skipping the intermediate level and making statements for satisfying goals directly, thus, lying whenever it is useful to do so? Of course I’m not making an argument for lying, but rather against a definition of truth derived from goal-seeking.

    One possible way out is to go one step futher: we are goal-seeking beings because we are happiness-seeking beings. In this sense a truth would be something that makes generates the largest possible sum of happiness, which would, at least, not be so cynical. Good predictions would probably still be important in everything that’s falsifiable, because feeling let down after we believed something and worked hard for something for 50 years clearly doesn’t help human happiness, so actually most of it would be predictionalist truth, however, it would liberate us to make up pretty much anything we think believing will help human happiness regarding unfalsifiable statements, because in unfalsifiable statements there is no chance for being let down later on.

  96. “Someday – probably fairly soon in historical time – all philosophers will begin their theory-building with phenomenology and Peircean operationalism as automatically as chemists begin with atomism and Dalton’s law. But that time is not yet.”

    Would it mean everything from Aristotle to Michael Oakeshott being thrown overboard and forgotten like phlogiston or alchemy?

    I don’t think I would like it. Every system needs a “rubber” part to it, that stretches to connect everything else, pretty much the same way that every small company has a “rubber” employee that does all those things nobody cares about doing.

    Philosophy in the pre-modern sense is the “rubber” part of human intellectual activity: it’s not quite science and not quite poetry and not quite theology or atheology and not quite politics and not quite literature, but it stretches to connect them all.

    Philosophy, the very term, rests on two very human and very unrigorous terms: wisdom, which nobody knows what it is, and the love of it, the most ultimately subjective and unrigorous experience of them all.

    Thus, I can love the wisdom of Chesterton because it’s basically nothing but very good writing, a good use of the paradox as a literary device, I can love the wisdom of Michael Oakeshott because it’s basically an extremely ingenious use of the English language, I can love the wisdom Platon for the sheer architectural beauty of his text, I can love the wisdom of Aristotle for the large amounts of common sense embedded in it and so on.

    I have no objection against a “Husserlian”, rigorous study of the logic of the human language and the human experience, but please call it something else than philosophy or it will outcompete this sort of stuff completely and that would be a big loss.

  97. >One possible way out is to go one step futher: we are goal-seeking beings because we are happiness-seeking beings. In this sense a truth would be something that makes generates the largest possible sum of happiness,

    That is the move the early post-Peircean pragmatists made – in fact, Peirce had a falling-out with William James over this issue. Peirce was right and James was wrong; if you can’t distinguish truth from expediency you end up doing politics, not science.

    >Would it mean everything from Aristotle to Michael Oakeshott being thrown overboard and forgotten like phlogiston or alchemy?

    Almost everything, I think. Some of the better-quality skeptical arguments will survive. Some relatively peripheral areas of philosophy won’t fit in the synthesis; esthetics is the obvious one, and moral philosophy is a question – it depends partly on how much of that field survives the evolutionary psychologists.

    >I have no objection against a “Husserlian”, rigorous study of the logic of the human language and the human experience, but please call it something else than philosophy or it will outcompete this sort of stuff completely and that would be a big loss.

    Sorry, but this seems to me to be like wishing alchemy had survived because the illuminated manuscripts were so pretty. Silly, and ultimately futile.

  98. Desmond wrote:

    OTOH, Dennett *has* said that under any sort of even semi-realistic circumstances, Mary would, in fact, learn something new upon seeing a rose because under any sort of semi-realistic circumstances, she wouldn’t even be close to knowing everything physical there is to know about colour vision. The thought experiment demands a form of omniscience as one of its basic premises, which Dennett sees as pure fantasy.

    I think there’s another aspect to this as well. The omniscience of Mary in the thought experiment is not gratuitous; it’s necessary to get to the conclusion that Jackson, the originator of the thought experiment, originally intended: that physicalism is false. But Dennett’s point is that the thought experiment pulls a kind of bait and switch: it assumes that Mary knows *everything* physical (because it has to to be a valid test of physicalism), but then encourages you to accept the conclusion that physicalism is false based on intuitions that only make sense if Mary *doesn’t* know everything physical or anything close to it.

  99. Mary, by hypothesis, has *complete* knowledge of those physical details.

    Hogwash. Mary, like a person with red-green color-blindness, has never had her brain in “state R” until she leaves the room and is allowed to actually use her ability to perceive color. The very phrase “complete knowledge” is ridiculous. No one ever has complete knowledge. Rumsfeld’s infamous “known knowns, known unknowns, and unknown unknowns” may have been mocked by his political opponents, but no military leader worth a damn will challenge that those three categories are real, and that battles often turn on them (especially on that last category).

    Nor does the hypothesis stipulate this “complete knowledge”; it says she’s an expert on neurophysiology. She “knows” abstract facts about how brains function, but she has never personally experienced her own brain functioning in “state R”. When her own brain enters this “state”, it is a new experience. I don’t understand the idea that nothing is added to her knowledge thereby.

  100. The Monster wrote:

    Nor does the hypothesis stipulate this “complete knowledge”; it says she’s an expert on neurophysiology.

    From the Wikipedia article, the hypothesis is stated as follows:

    “Mary is a brilliant scientist who is, for whatever reason, forced to investigate the world from a black and white room via a black and white television monitor. She specializes in the neurophysiology of vision and acquires, let us suppose, all the physical information there is to obtain about what goes on when we see ripe tomatoes, or the sky, and use terms like ‘red’, ‘blue’, and so on. She discovers, for example, just which wavelength combinations from the sky stimulate the retina, and exactly how this produces via the central nervous system the contraction of the vocal cords and expulsion of air from the lungs that results in the uttering of the sentence ‘The sky is blue’. [...]”

    So she has, by hypothesis, “all the physical information” — not just what an expert in neurophysiology would have today, but *all* of it.

    That’s a big difference, enough of a difference to make me think it’s possible that Mary, with *all* of the physical information, might be able to get her brain into state S without having been in state R, or, alternatively, to get her brain into state R without having the normal means of doing so available to her.

    Of course no real person would ever have such complete knowledge, but Mary does, as the thought experiment is stated. If you want to say that’s impossible, then you’re saying that the situation described in the thought experiment is impossible because it depends on an impossible premise. That’s fine by me, since it’s not my thought experiment. :-)

  101. When her own brain enters this “state”, it is a new experience. I don’t understand the idea that nothing is added to her knowledge thereby.

    Agreed, which means despite what Peter Donis is saying, she doesn’t have complete physical knowledge because she is missing this experience.

    I said it before, I’ll say it again: there is a difference between knowing something on an intellectual level and experiencing that same thing.

    A better example that many here would understand: you can read all the damned books you want about programming, but until you do it, you won’t be able to code your way out of a wet paper bag.

  102. morgan greywolf wrote:

    Agreed, which means despite what Peter Donis is saying, she doesn’t have complete physical knowledge because she is missing this experience.

    As I noted in my previous comment, by the hypothesis of the thought experiment, Mary *does* have complete physical knowledge. If you’re saying that, nevertheless, she doesn’t know what it’s like to experience seeing red, you’re assuming that physicalism is false (complete physical knowledge complete knowledge). This was, of course, the point of the original thought experiment. But I believe that physicalism is true, so I reject the claim that Mary, with complete physical knowledge, doesn’t know what it’s like.

    A better example that many here would understand: you can read all the damned books you want about programming, but until you do it, you won’t be able to code your way out of a wet paper bag.

    You’re assuming that Mary’s possession of all physical knowledge doesn’t include *using* that knowledge. Why wouldn’t it? For example, why wouldn’t her knowledge of all the physical details of how her brain would react if red light impinged on her retinas include *imagining*, in detail, what it would be like for her brain to be put in that state? After all, she could have already done the same thing for many other states of her brain, including “calibrating” her imagination by comparing her experience of imagining her brain being in a certain state (say, the state of seeing a black and white image of a tomato) with her experience of her brain being put into that state by the appropriate external stimulus.

    In other words, Mary’s possession of complete physical knowledge is *not* analogous to just “reading books about programming” without ever writing a program. A better analogy would be: suppose there were a programmer who never had access to an actual computer to test his programs, but whose knowledge of the physical workings of computers was so complete that he could *imagine*, in detail, the effects that any piece of code would have when run on a computer. Such a programmer could, in effect, teach himself to program by reading books about programming and then writing programs (in his head, on paper, whatever) and imagining their effects to see if they did what he wanted.

  103. Oops, I forgot to format something correctly in the comment I just entered; I meant to write (in my first paragraph of reply): complete physical knowledge < > complete knowledge.

  104. Peter Donis writes:
    >I think there’s another aspect to this as well. The omniscience of Mary in the thought experiment is not gratuitous; it’s necessary to get to the conclusion that Jackson, the originator of the thought experiment, originally intended: that physicalism is false. But Dennett’s point is that the thought experiment pulls a kind of bait and switch: it assumes that Mary knows *everything* physical (because it has to to be a valid test of physicalism), but then encourages you to accept the conclusion that physicalism is false based on intuitions that only make sense if Mary *doesn’t* know everything physical or anything close to it.

    Exactly. Dennett’s response to the Chinese room argument is very similar. Searle asks you to imagine that a non-Chinese speaker is in the room cranking through an algorithm as complex as you please to produce appropriate Chinese answers to whatever questions the judge comes up with. The algorithm might be good enough to pass the Turing test – it certainly seems like someone in the room understands Chinese – but, according to Searle, it’s “just obvious” that no real understanding is going on.

    Dennett’s retort to the Chinese room (in Consciousness Explained) is similar to his retort to Mary’s room – the conclusion only seems obvious because we’re not imagining the thought experiment in enough detail. He invents an imaginary transcript between the judge and the room where the judge asks the room to explain an ethic joke about Irishmen liking their Guinness a bit too much, to which the room a) asserts that this is boring and one should never explain jokes b) says that he finds ethic jokes in extremely poor taste c) admits that he nonetheless laughed in spite of himself and c) then proceeds to explain why the joke is supposed to be funny.

    We have a hard time imagining a computer program so supple, and so sophisticated, that it can do that sort of thing. We have a hard time imagining that it could talk about love, red, hate, blue, beer, cheese and Angelina Jolie so convincingly that we wouldn’t know there wasn’t a human behind the keyword. We have a hard time imagining it – so we don’t try. We just imagine that the program is “fairly sophisticated” or that it would fool a naive judge, or that it looks like a typical modern AI program (none of which stand a snowball’s chance in hell of fooling anyone under pointed and prolonged questioning). We imagine that there would have to be *some* way of detecting the artificiality of the responses.

    But…what if there wasn’t? Would Searle’s conclusion still seem obvious? What if you could look at the program as it was doing it’s work? People underestimate the power of behaviourism.

    Sorry, that was long and off topic :) I’m just saying that “not imagining the thought experiment in enough detail” is a standard Dennett response to these sorts of intuitive thought experiments.

  105. morgan greywolf wrote:
    > A better example that many here would understand: you can read all the damned books you want about programming, but until you do it, you won’t be able to code your way out of a wet paper bag.

    Well, sure. Except that Mary isn’t reading “all the damned books you want” about neurophysiology. She’s gaining complete physical knowledge of colour. Of course, I agree that reading books is probably not the path to complete physical knowledge of anything, but we’re not really concerned with how she became omniscient, are we?

    If I really had “complete physical knowledge of programming”, then I could program my way out of a wet paper bag, even if I had never seen a computer.

  106. @Peter Donis
    You’ve moved the goalposts:

    So she has, by hypothesis, “all the physical information”

    That’s not “complete knowledge”. No knowledge is absolutely complete. Heisenberg proved it (absolutely ).

  107. >That’s a pretty bold statement. How about aspects of reality that are not yet discovered? The universe is a very big place and humans have thus far experienced only a tiny fraction of it. Humans are not unlike Mary, sitting in her little black and white room, only, unlike Mary, we only know what we see in the little black and white room.

    Morgan, it follows from Kant. Kant (I think rightly) said that there are a priori categories built into the mind that are *logicallly neccessary* to all understanding/experience ; so if we knew what these categories were, we would in essence have the tools to understand anything in reality that could possibly affect us. Mary does NOT just have to rely on sensory data, she could (in theory) peer into her own mind…. even in a dark room… and *directly experience* Kant’s categories of thought. The problem is that the human brain has poor reflective capabilities, so this is very difficult to do, as Jeff suggested, some serious LSD might help ;) But you are right that any sort of ‘theory of everything’ must remain open to new knowledge… I view the categories only as a sort of ‘reality kernel’ (prototypes) around which an infinite amount of new knowledge can coalesce – I’m not suggesting that they are some sort of closed system.

  108. The Monster wrote:

    That’s not “complete knowledge”. No knowledge is absolutely complete. Heisenberg proved it (absolutely ).

    Hmm. If I understand you correctly, you’re saying something like the following: “what it’s like” to have your brain in a particular state is something that can’t be known in advance, or “imagined”, or anything like that, because there is an unavoidable element of quantum indeterminacy in your subjective experience. In other words, the state of the parts of your brain that are involved in seeing red, before you’ve actually seen red, is something like the state of a photon that’s gone through a half-silvered mirror; until you actually make some measurement that distinguishes the alternatives, you can’t predict what state the photon will end up in. In the case of your brain, the “measurement” would just be the physical act of impinging red light on your retinas, and the “alternatives” would presumably be some spectrum of “possible ways that red could look like to you”, and which alternative would actually be realized would not be predictable in advance (though you could, in principle, calculate probabilities for the various alternatives).

    That’s an interesting viewpoint, and as far as I can see, it’s completely consistent with physicalism; knowledge of “what it’s like to see red” could still be physical knowledge, it would just be physical knowledge that nobody can predict in advance, just as the state of the photon is before it’s measured. So as far as the Mary’s Room thought experiment is concerned, if your viewpoint is right, then the thought experiment does *not* prove that physicalism is false. (It doesn’t prove that physicalism is true, either, of course; it’s consistent with either viewpoint.)

  109. Hmm. If I understand you correctly,

    You don’t.

    What I am saying is that it is one thing to observe other eyes reacting to what Mary is told is “red” light, to use external instrumentation to determine the wavelength of that light, to perform experiments to isolate the portions of the brain stimulated by red light, and to form a mental model of what the experience of seeing red must be like, based on all of that indirect information.

    It is entirely a different thing to see red light with her own eyes, along with the independent confirmation from her instruments that the wavelength falls within the range her teachers and books have defined as “red”.

    Others have granted unto Mary the ability to have artificially created in her own brain the same state that would otherwise be the result of physically seeing red light. Apparently that comes from her assumed omniscience, which some have accepted can’t exist in reality, but still insist applies to this hypothetical.

    We might as well argue about how many angels can dance on the head of a pin.

    The zeroth thing we know about Mary is that she is a human. Anything else in the hypothetical that contradicts that invalidates the whole Gedanken thing. Humans do not have detailed knowledge of how our own brains function, especially to the point of being able to construct out of whole cloth the perception of red light. Mary’s Room, being based on false premises, proves nothing.

  110. The Monster wrote:

    We might as well argue about how many angels can dance on the head of a pin.

    The zeroth thing we know about Mary is that she is a human. Anything else in the hypothetical that contradicts that invalidates the whole Gedanken thing. Humans do not have detailed knowledge of how our own brains function, especially to the point of being able to construct out of whole cloth the perception of red light. Mary’s Room, being based on false premises, proves nothing.

    Well…yes?

    That’s Dennett’s whole point; the entire premise of the thought experiment is crazy. Much crazier than people generally give it credit for; it goes beyond just the logistics of how we would ever, practically speaking, keep Mary in the dark (so to speak) about colour.

    People imagine the “experiment” and think that the premise is something like “Mary is really, really smart with a lot of book-learnin’”. That’s not it; Mary, according to the premises of the thought experiment, is practically God.

    You can argue that this is silly, and you’d be right. One might as well start a thought experiment with “Imagine for a moment that humans beings are actually turtles”.

  111. The Monster wrote:

    Humans do not have detailed knowledge of how our own brains function, especially to the point of being able to construct out of whole cloth the perception of red light. Mary’s Room, being based on false premises, proves nothing.

    As I said in a previous comment, I’m fine with that, since it’s not my thought experiment. :-) It was your mention of Heisenberg that sent me off into quantum-land (since quantum indeterminacy was the only kind of incompleteness of knowledge that Heisenberg proved anything about). When I read the thought experiment I was willing to play along to the extent of granting that it would be possible in principle for humans to become advanced enough in knowledge and cognitive power to be capable of something like the feats of imagination I was attributing to Mary, but I admit that’s a lot to grant. :-)

  112. Given two ways of interpreting something, where one way is useless and pointless and the other is an interesting, pointful question, I without shame will choose the latter interpretation.

    It is true that if you start with a definition of “knowing” that encompasses a definition of knowledge that includes absolutely everything, where if you ask “Does it involve knowing X?” the answer is always “yes”, then yes, you’ve defined a trivial and worthless question. Yes, if you start with a definition such that the answer to the question is already, by definition, that Mary has experienced (in every way) the color red without ever actually seeing it, then by definition, the experience is nothing new.

    But that’s a terribly boring question. I would even go out on a limb and say it’s objectively boring. I also don’t think it’s the point. I think it’s more about trying to figure out why having knowledge, with a reasonably standard definition of “knowledge”, is different from having the experience, with a reasonably standard definition of “experience”. Now that’s an interesting question that might actually result in some insight, and also has interesting tie-ins to artificial intelligence as well as natural intelligence.

    So I favor the second interpretation, because, quite frankly, even if the original poser did “mean” the boring formulation, nothing prevents me from discussing and caring about the interesting formulation! To be honest, I don’t really care what the original question was.

  113. Well, sure. Except that Mary isn’t reading “all the damned books you want” about neurophysiology. She’s gaining complete physical knowledge of colour. Of course, I agree that reading books is probably not the path to complete physical knowledge of anything, but we’re not really concerned with how she became omniscient, are we?

    No. But you’re still not getting my point. Take as an example this quote from Aleister Crowley in Magick in Theory and Practice”(which I just happened to be reviewing the other day):

    A blind man might cram up astronomy for the purpose of passing examinations, but his knowledge would be almost entirely unrelated to his experience, and it would certainly not give him sight. A similar phenomenon is observed when a gentleman who has taken an “honours degree” in modern languages at Cambridge arrives in Paris, and is unable to order his dinner.

    The point is that Mary’s knowledge of science sitting in total isolation in a little black and white room and never leaving is just that knowledge. To develop an understanding, one must not just necessarily know about certain things, but they must experience those things themselves.

    Here is an exercise for you: read all about a stunningly beautiful place somewhere on the planet that you’ve never been to. Study it on Google Earth or in books or on Wikipedia or whatever you want. Get all the knowledge you can about it. Call the locals about and ask them about it. Whatever. And when you’re done, actually go there. Many people will say that their experience of a place doesn’t always match their knowledge of a place, no matter how much knowledge they accumlate about it beforehand. Do you see what I mean?

  114. Morgan Greywolf wrote:

    To develop an understanding, one must not just necessarily know about certain things, but they must experience those things themselves.

    That’s true of us normal humans, but the cognitive powers Mary must have in order to satisfy the assumptions of the thought experiment (she has to know *all* the physical information) make her so unlike us normal humans that your proposition may not apply to her.

    If you don’t believe it’s possible for her to have such cognitive powers, then you believe (as The Monster does) that the thought experiment’s premise is invalid. But if (like me) you’re willing to grant the premise as valid, at least for the sake of argument, then you can’t rely on your beliefs about what normal humans have to do to understand what something is like, to get a handle on what omniscient Mary would have to do to understand what something is like.

  115. Examples for communicating qualia: http://www.frenchscout.com/types-of-wines

    “Typical taste in varietal wine: often wider-bodied (and more velvety) than other types of dry whites, with rich citrus (lemon, grapefruit) flavours. Fermenting in new oak barrels adds a buttery tone (vanilla, toast, coconut, toffee). Tasting a USD 12 Californian Chardonnay should give citrus fruit flavours, hints of melon, vanilla, some toasty character and some creaminess. ”

    Is it a clumsy, awkvard kind of domain-specific language? It is. Is it useful for making predictions, thus, is it, according to your definition of truth, true ? It is – I can predict which ones would I probably like more than the other ones, thus helps if I’m facing a buying decision without an opportunity to taste it first.

  116. Is it a clumsy, awkvard kind of domain-specific language? It is. Is it useful for making predictions, thus, is it, according to your definition of truth, true ? It is – I can predict which ones would I probably like more than the other ones, thus helps if I’m facing a buying decision without an opportunity to taste it first.

    Are you sure you’re not falling for a sophisticated form of pareidolia, exploited by the wine makers for advertisement purposes?

    An example: When I chew on a strawberry-flavored Starburst® fruit chew, what I taste is only vaguely, if at all, like strawberries. Yet because the packaging tells me it’s strawberry-flavored, I’m more inclined to think “this tastes like strawberry” when I eat it.

    The utility of blurbs like this in predicting whether a given person will like a particular wine is, to me, far more dubious than its utility in enticing people to buy wine.

  117. For me “hints of melon, vanilla” is a turn-off: I’m quite literal-minded and think it tastes like ice cream, not like proper wine, so, no buy.

    Yes, I can imagine certain groups of buyers who can decode this mumbo-jumbo and for them it can a good enticer to buy. However, wouldn’t that mean it serves other purposes too? I suppose there must be a reason why buyers learn this language.

  118. A question to materialists/physicalists:

    Doesn’t it follow from the materialist/physicalist assumption that “good” or “bad” only make sense as pleasure and pain, and pleasure and pain only makes sense as certain brain functions that result in the feeling of pleasure and pain, i.e. it’s impossible to define “good” in any other way than “actions that result in the most pleasure and least pain generated in the brains of the most possible people” ? Meaning: we should consider the man who researches the stars or physics, the man who writes poems or composes music or directs movies, a man who desires to be an athlete etc. either irrational or immoral because everybody should focus his energies on researching the perfect recreational drug or brain stimulation device or Matrix-style VR (except that this VR should provide an extremely pleasureable paradise, not an ordinary world), because this would result in more pleasure induced in the brains of more people? Thus, the world of Huxley’s Brave New World is not horrible or inhuman, but actually a worthy goal to pursue?

    Because I think we all “instinctively”[1] see there is something wrong about it, but can you escape this logic and believe in materialism / physicalism while also agreeing that painting or composing or competing in athletics is a worthwhile human activity?

    Of course one must ask what is, then, outside brain-level pleasure and pain? And it’s a very hard question but I’ll try to answer it with a thought experiment. Suppose there is a very rich man who arranges the following for himself: he’ll spend the rest of his life in a hospital bed, fed artifically, continously asleep under the effect of a drug that induces perfectly happy dreams and does no harm to the brain or other parts of the body. Wouldn’t we “instinctively”[1] feel it’s kinda disgusting, worthless way to live? And not only because of its selfishness – we even find similarly selfish activities such as engaging in sports or painting for one’s own pleasure better ways to live. Why? We can find two reason why. One is that if we have Reason we have a duty to use it. Another reason is that such life is simply ugly to the external observer, that means, we are making a kinda aesthetical judgement.

    [1] “instict” = something we or at least the person who uses this word cannot (yet) explain. The sentence “migrating birds fly southwards by instinct” means “I don’t know how they do it.”

  119. >Doesn’t it follow from the materialist/physicalist assumption that “good” or “bad” only make sense as pleasure and pain, and pleasure and pain only makes sense as certain brain functions that result in the feeling of pleasure and pain, i.e. it’s impossible to define “good” in any other way than “actions that result in the most pleasure and least pain generated in the brains of the most possible people”

    That’s correct. But you can bootstrap your way from there up to something resembling a virtue ethic by asking what sorts of behaviors cause the best consequences over long time frames. This maneuver is not even very difficult.

  120. Yes, but there are many possible counter-examples leading to stable equilibriums in repulsive configurations. F.e. what about a global society where 90% of the have this bed-and-happy-drugged “life” I described above and 10% care for them – their payment is that they have pretty much all the wealth in the world to play around with because those in the beds use very little actual resources?

  121. >Yes, but there are many possible counter-examples leading to stable equilibriums in repulsive configurations

    Very true. And how would repudiating physicalism be supposed to fix this, exactly? You seem to have multiple different issues confused here.

    To avoid the “Joy Makers” scenario you’ve described, what we we need is a functioning virtue ethic. The only way physicalism could be relevant is if it somehow prevents people from developing or maintaining a vitrtue ethic. How would it do that, exactly?

  122. Are you hinting at an SF novel that could be a good idea to read? http://www.amazon.com/Makers-Classics-modern-science-fiction/dp/0517551845

    How would it do that? Isn’t it obvious? By the usual “deadly truths” stuff (Leo Strauss). Virtue ethics in this framework would be a useful illusion that works only as long as nobody figures it out (and manages to popularize the criticism) that it is only a useful illusion. Then it doesn’t work. Virtue ethics usually need to be cast into religious, aesthetical, emotional, nationalistic, poetical-heroic (Homer) or whatever other similarly non-rational frameworks, myths, in order to work i.e. to provide sufficient motivation. But mat./phys. , done right, deconstructs these frameworks and shows them as false and then they don’t work.

    Do you seriously mean you haven’t noticed that the “deadly truths” stuff accurately explains much of the history of the last 50-100-200-300 years? That for example mat./phys. successfully deconstructed the myths behind most kinds of patriotism or sheepdogism and what we got is what you described (accurately) as suicidalism?

  123. To put the above question in more accurate terms, how can one simultaneously maintain the “doublethink” of realizing that some myths sometimes are clearly useful and yet they are false, and even communicate this to each and every person even to those who aren’t able to maintain such “doublethinks”?

  124. >Are you hinting at an SF novel that could be a good idea to read?

    Why yes, in fact. That very one.

    I don’t buy your presumption that the truth about the universe is deadly to virtue ethics, It isn’t math and physics that gave us suicidalism, it was Soviet memetic warfare.

  125. Sorry, not math and physics, mat/phys I just used as an abbreviation for materialism-physicalism in the question of human nature/consciousness. (Soviet memetic warfare certainly has something to do with materialism too.)

  126. Don’t mean to light up an old thread, but I was reading through your back posts and am glad to see that someone with such a solid philosophical background came to the same conclusion I did. Perhaps it’s my background in physics that made my response fairly easy, but my immediate reaction to hearing the “paradox” was that of course Mary didn’t learn anything new about “red”; she learned something new about Mary (specifically, what Mary feels on seeing red).

  127. I think you’re right — quiddity has a qualia all its own.

    How about spliting the universe into 3 axies (objective/subjective/qualitative) instead of 2?

    You may be interested in this blog: coyotequiddity.blogspot.com/

  128. “What it *feels like* to have the” …etc etc.
    How is “feeling” as you use it here any different from the notion of qualia in the first place?
    Its not. You’ve played a game of three card Monty here.

    The answer to Dennet is simple: we experience qualia first and develop notions
    about science, frequencies, particles, wave lengths later. Any person who had
    no qualia operant in their beings would not possess any scientific or philosophical
    knowledge either, either factual or errant.

  129. “Arguably, all brain states are incommunicable in that sense ”

    so what are you communicating right noew?

  130. “Believe it or not I think that these 27 categories, once properly defined, are sufficient to ‘capture’ ALL of reality (in the sense of acting as prototypes for ANYTHING that needs to be modelled). They are the ‘ontological alphabet’ of the universe”.

    No biggie. “Sausage” and “not sausage” capture everything too.

  131. “Metaphysical idealism is nonsense instantly falsified by experience”

    You can call your experiences experiences of phsyical things. The idealist will just insist that apparent weight and solidity are qualia in your head.

  132. Mr. Raymond, as much I usually disagree with everything you have to say – and that disagreement remains even with this post – I must commend you on the wittiness of your choice of title.

  133. @Peterdjones: “You can call your experiences experiences of phsyical things. The idealist will just insist that apparent weight and solidity are qualia in your head.” Speaking as an idealist, qualia in your mind, thank you very much!

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>