Three consistent positions for computationalists
Yesterday, as a followup to We are not living in a simulation, I posted Eight questions for computationalists in order to obtain a better idea of what exactly my computationalist critics were arguing. These were the questions I asked:
As it is used in the sentence “consciousness is really just computation”, is computation:
a) Something that an abstract machine does, as in “No oracle Turing machine can compute a decision to its own halting problem”?
b) Something that a concrete machine does, as in “My calculator computed 2+2”?
c) Or, is this distinction nonsensical or irrelevant?If you answered “a” or “c” to question 1: is there any particular model, or particular class of models, of computation, such as Turing machines, register machines, lambda calculus, etc., that needs to be used in order to explain what makes us conscious? Or, is any Turing-equivalent model equally valid?
If you answered “b” or “c” to question 1: unpack what “the machine computed 2+2” means. What is that saying about the physical state of the machine before, during, and after the computation?
Are you able to make any sense of the concept of “computing red”? If so, what does this mean?
As far as consciousness goes, what matters in a computation: functions, or algorithms? That is, does any computation that give the same outputs for the same inputs feel the same from the inside (this is the “functions” answer), or do the intermediate steps matter (this is the “algorithms” answer)?
Would an axiomatization (as opposed to a complete exposition of the implications of these axioms) of a Theory of Everything that can explain consciousness include definitions of any computational devices, such as “and gate”?
Would an axiomatization of a Theory of Everything that can explain consciousness mention qualia?
Are all computations in some sense conscious, or only certain kinds?
I got some interesting answers to these questions, and from them I can extract three distinct positions that seem consistent to me.
Consistent Position #1: Qualia skepticism
Perplexed asserted this position in no uncertain terms. Here’s my unpacking of it:
“Qualia do not exist. The things that you’re confused about and are mistaking for qualia can be made clear to you using an argument phrased in terms of computation. When you talk about consciousness, I think I can understand your meaning, but you aren’t referring to anything fundamental or particularly well defined: it’s an unnatural category.”
The internal logic of the qualia skeptic’s position makes sense to me, and I can’t really respond to it other than by expressing personal incredulity. To me, the empirical evidence in support of the existence of qualia is so clear and so immediate that I can’t figure out what you’re not seeing so that I can point to it. However, I shouldn’t need to bring you to your senses (literally!) on this in order to convince you to reject Bostrom’s simulation argument, albeit on grounds completely different than any I’ve argued so far. If you don’t buy that there’s anything fundamental behind consciousness, then you also shouldn’t buy Bostrom’s anthropic reasoning in which he conjures up the reference class of “observers with human-type experiences”; elsewhere he refers to “conscious experience” and “subjective experience” without implication that he means anything more specific. That’s taking an unnatural category and invoking it magically. In the statement that we are something selected with uniform probability from that group, how do you make sense of “are”?
Consistent Position #2: Computation is implicit in physics
This position is my best attempt at a synthesis of what TheOtherDave, lessdazed, and prase are getting at. It’s compatible with position #1, but neither one entails the other.
To understand this position, it is helpful, but not necessary, to define the laws of physics in terms of something like a cellular automaton. Each application of the automaton’s update rule can be understood as a primitive operation in a computation. When you apply the update rule repeatedly on cells nearby each other, you’re building up a more complex computation. So, “consciousness is just computation” is equivalent in meaning, essentially, to “consciousness is just physics”.
This position more-or-less necessitates answering “algorithms” to question #5, or if not that then at least something similar to RobinZ’s answer. If you say “functions” then you at least need to explain how to reify the concepts of “input” and “output”. You can pull this off by saying that the update rules are the functions, the inputs are the state before the rule application, and the outputs are the state afterward. Any other answer probably means you’re taking something closer or identical to position #3 which I’ll address next. This comment by peterdjones and his followups to it provide a (Searlesque) intuition pump showing other reasons why a “functions” reply is problematic.
I have no objection to this position. However, it does not imply substrate independence, and strongly suggests its negation. If your algorithmic primitives are defined at the level of individual update-rule applications, then any change whatsoever to an object’s physical structure is a change to the algorithm that it embodies. If you accept position #2 while rejecting position #1, then you may actually be making the same argument that I am, merely in different vocabulary.
Consistent Position #3: Computation is reified by physics
I was both shocked and pleased to see zaph’s answer to question #6, because it bites a bullet that I never believed anyone would bite: that there is actually something fundamental in the laws of physics which defines and reifies the concept of computation in a substrate-independent fashion. I can’t find any inconsistency in this, but I think we have good reason to consider it extremely implausible. In the language of physics which is familiar to us and has served us well — the language whose vocabulary consists of things like “particle” and “force” and “Hilbert space” — the Kolmogorov complexity of a definition of an equivalence relation which tells us that an AND gate implemented in a MOSFET is equivalent to an AND gate implemented in a neuron is equivalent to an AND gate implemented in desert rocks, but is not equivalent to an OR gate implemented in any of those media — is enormous. Therefore, Solomonoff induction tells us that we should assign vanishingly low probability to such a hypothesis.
I hope that I’ve fairly represented the views of at least a majority of computationalists on LW. If you think there’s another position available, or if you’re one of the people I’ve called out by name and you think I’ve pigeonholed you incorrectly, please explain yourself.
- 's comment on Confused as to usefulness of ‘consciousness’ as a concept by (16 Jul 2014 1:13 UTC; 5 points)
- 's comment on How sure are you that brain emulations would be conscious? by (24 Aug 2013 11:31 UTC; 2 points)
- 's comment on Not all theories of consciousness are created equal: a reply to Robert Lawrence Kuhn’s recent article in Skeptic Magazine [Link] by (8 Sep 2016 2:21 UTC; 1 point)
I believe I have found the perfect, modern theory of consciousness, completely supported by every study yet done on the matter!
“We really don’t know what’s going on. More research is needed. But there’s probably no magic involved.”
Hmm. My comment is the most highly upvoted response to your survey at the moment, and the second highest upvoted one is by XiXiDu expressing basically the same position as mine, but I don’t see it on your list. Here’s a summary: we don’t yet have enough insight to choose any specific answer or even to know if we’re asking the right questions. We’re facing an unsolved scientific problem. The wisdom of crowds doesn’t apply here. If no one has yet discovered Maxwell’s equations or Watson and Crick’s double helix, no amount of surveying can lead you to the right answer. You have to do, like, actual math and physics and biology and stuff.
I agree with this, but would like to add that it’s actually one step worse: most of the interesting experiments one can do with regard to consciousness have results that, for various reasons, cannot be transferred between observers. The quantum immortality hypothesis is one example—if someone else does an experiment, you don’t get to see the result. But the problem is more general; you also don’t get to see the results of experiments that other entities perform to test observer symmetry, or the subjective results of self-copying and merging. So the only information we can have is prior probabilities, which are not very informative, with no experimental data. Perhaps after a dozen one-way trips through cryopreservation, nested simulations, afterlives, etc., I’ll have an answer to how subjective experience works; but no one will ever find an answer in this universe, or convey an answer back to it, so the question has little point here.
I don’t like statements like “we can never know” this or that. For example, you can convince everyone that quantum immortality works by killing them along with yourself. (This shouldn’t pose any risk if you’ve already convinced yourself :-) Paul Almond has proposed that this can solve the Fermi paradox: we don’t see alien civilizations because they have learned to solve complex computational problems by civilization-level quantum suicide, and thus disappeared from our view.
It seems probable to me that if we think a little harder, we can figure out a way to investigate observer-dependent statements scientifically.
Interesting, particularly in light of the recent “What is analytic philosophy, that we should be mindful of it?” discussions. It almost seems that dfranke, taking the philosopher’s role, should respond: “We are facing an unsolved philosophical problem. You can’t contribute to the solution without taking a position.”
I agree, and furthermore this is a true statement regardless of whether you classify the problem as philosophical or scientific. You can’t do science without picking some hypotheses to test.
That’s not strictly speaking true. First of all, this doesn’t quite match what Perplexed said since Perplexed was talking about taking a position. I can decide to test a hypothesis without taking a position on it. Second of all, a lot of good science is just “let’s see what happens if I do this.” A lot of early chemistry was just sticking together various substances and seeing what happened. Similarly, a lot of the early work with electricity was just systematically seeing what could and could not conduct. It was only later that patterns any more complicated than “metals conduct” developed. (Priestly’s The History and Present State of Electricity gives a detailed account of the early research into electricity by someone who was deeply involved in it. The archaic language is sometimes difficult to read but overall the book is surprisingly readable and interesting for something that he wrote in the mid 1700s.)
Those early experimenters with electricity were still taking a position whether they knew it or not: namely, that “will this conduct?” is a productive question to ask—that if p is the subjective probability that it will, then p\(1-p)* is a sufficiently large value that the experiment is worth their time.
Ok. Yes, this connects to the theory-laden nature of observation and experimentation. But that’s distinct from having any substantial hypotheses about the nature of electricity which would be closer to the sort of thing that would be analogous to what Perplexed was talking about. (It is possible that I’m misinterpreting the statement’s intention.)
Perplexed intended to contrast science—where it is not respectable to take a position in advance of evidence (pace Karl P.) - with philosophy—where it is the taking and defending of positions which drives the whole process. Last philosopher left standing wins. You can’t win if you don’t take a stand.
Thanks for clarifying. Is that true though? If so, I’d suggest that that might be a problem about how we do philosophy more than anything else. If I don’t have evidence or good arguments either way on a philosophical question I shouldn’t take a stand on it. I should just acknowledge the weak arguments for or against the relevant positions.
There are no specifically philosophical truths, only specifically philosophical questions. Philosophy is the precursor to science; its job is to help us state our hypotheses clearly enough that we can test them scientifically. ETA: For example, if you want to determine how many angels can dance on the head of a pin, it’s philosophy’s job to either clarify or reject as nonsensical the concept of an angel, and then in the former case to hand off to science the problem of tracking down some angels to participate in a pin-dancing study.
I didn’t list this position because it’s out of scope for the topic I’m addressing. I’m not trying to address every position on the simulation hypothesis; I’m trying to address computationalist positions. If you think we are completely in the dark on the matter, you can’t be endorsing computationalists, who claim to know something.
I … don’t think there’s much empirical support for the actual existence of the painfulness of pain. Sure, humans experience pain in very similar ways, and you can lump all those experiences into the category pain, and talk about what characteristics are present in all the category members, but those common characteristics aren’t a physical object somewhere called painfulness.
As for how this bears on Bostrom’s simulation argument: I’m not familiarized with it properly, but how much of its force does it lose by not being able to appeal to consciousness-based reference classes and the like? I can’t see how that would make simulations impossible; nearest I can guess is that it harms his conclusion that we are probably in a simulation?
That can be repaired in other ways; given that time travels in one direction for us, our experiences have one chance to be in the real universe, and n chances to be in simulated universes—where n is the total computational power that ever will be directed at simulating historical moments, over the computational cost of simulating a historical moment multiplied by the number of moments at least as interesting as this one. Even if you assign a low probability to the future containing computational power (ie we nuke ourselves before Matroishka shells or Jupiter brains are completed or something), that low chance times n is still large relative to 1. So our prior for being in a simulation should still be high.
I can’t see how it is remotely relevant that painfulness isn’t a physical object.Electron spin isn’t either.
Right. All the probabilistic reasoning breaks down, and if your re-explanation patches things at all I don’t understand how. Without reference to consciousness I don’t know how to make sense of the “our” in “our experiences”. Who is the observer who is sampling himself out of a pool of identical copies?
Anthropics is confusing enough to me that it’s possible that I’m making an argument whose conclusion doesn’t depend on its hypothesis, and that the argument I should actually be making is that this part of Bostrom’s reasoning is nonsense regardless of whether you believe in qualia or not.
You’ve missed a major position: that the entire idea of “substrate independence” is a red herring. Detecting the similarity of two patterns is something that happens in your brain, not something that’s part of reality.
This whole thing, AFAICT, is an attempt to have an argument war, rather than an attempt to understand/find truth. It is possible that no position on this subject makes any sense whatsoever, for example.
Or, to put it another way, failure to offer a coherent refutation of an incoherent hypothesis doesn’t represent evidence for the incoherent hypothesis.
If I’m correctly understanding what you mean by “part of reality” here, then I agree. This kind of “similarity” is another unnatural category. When I made reference in my original post to the level of granularity “sufficient in order model all the essential features of human consciousness”, I didn’t mean this as a binary proposition; just for it to be sufficient that if while you slept somebody made changes to your brain at any smaller level, you wouldn’t wake up thinking “I feel weird”.
I have no reason to assume that you couldn’t replace me entirely, piece by piece. After all, I have different cells now than I did previously, and will have different cells later, and all the while still perceive myself the same.
The only thing weird here, is the idea that I would somehow notice. I mean, if I could notice, it wouldn’t be a very good replacement, would it?
(Actually, given my experience with mind hacking, my observation is that it’s very difficult to notice certain background characteristics of one’s thought processes, such that even if a machine translation did introduce a systematic distortion, it seems unlikely to me that anyone would notice it in themselves, at least easily or at first!)
“Or, to put it another way, failure to offer a coherent refutation of an incoherent hypothesis doesn’t represent evidence for the incoherent hypothesis.”
Although perhaps a tangent, the point is important: the above is quite wrong. If A believes h is incoherent, but B is unable to demonstrate h’s incoherence, A should regard B’s inability to coherently explain h’s incoherence as evidence that h is not , in fact, incoherent. (I think that’s what you mean to deny.). This is because (at least in ordinary circumstances) A should regard h as more probably true if B can explain why; less probable if B can’t.
The denial that B’s failure to coherently explain h’s incoherence increases the probability that h is coherent expresses the common failure to regard others’ beliefs as evidence of what’s true. This fallacy is why Aumann’s agreement theorem seems so counter-intuitive to many people. (See my fishbowl analogy at http://tinyurl.com/3lxp2eh)
Er, ‘A’ believes ‘h’ is coherent in this case.
I have realized, though that my statement was profoundly unclear, even after the edit.
Let me attempt to rephrase yet again, more precisely:
“If a bunch of people on LW tell you your hypothesis is incoherent and you need to dissolve your question, this should not be considered evidence that your hypothesis is sound, merely because nobody directly refuted your incoherence, in terms currently comprehensible by you.”
Or, by analogy, if you go to a biology forum and ask about missing links or why there are still apes, and then when you get explanations that dissolve the wrong questions involved, you say, “aha, but you still haven’t answered my [wrong] question, so therefore I’m right”, this is not sound argument.
In this case, though, the incoherence has actually been quite clearly counterargued by many, and is already thoroughly refuted by the sequences.
Could you edit this? I can’t decipher it.
[eta: Cyan and Pavitra have come up with nice obviously-true statements that are textually similar to the original bungled sentence and similar in meaning, but I can’t be sure of what you meant.]
Sorry, that was a messed up edit—I was at first writing “doesn’t represent evidence for incoherence” and then messed up the edit to “doesn’t represent evidence for the incoherent hypothesis”.
More colloquially, if somebody can’t coherently answer your incoherent question, it doesn’t mean that the viewpoint which created the question is therefore sensible or true.
How about, “If I offer a not-even-wrong refutation of your not-even-wrong hypothesis, you can’t take the not-even-wrongness of the refutation as evidence for the hypothesis.”
I read it to mean that once one has demonstrated a hypothesis to be incoherent, one does not then also need to demonstrate it to be false.
One can only detect, as opposed to invent, what is already there. Being a NAND gate is not a physical property that is already there, nonetheless not everything is a NAND gate. There are constraints on what substrate can do what, but they are not fully determinate facts, for all that they are not imaginary.
Actually, “NAND Gate” is a term that we use to label something that is there—a tag we assign to patterns in the physical world that follow similar patterns of behavior to a representation we hold in our minds.
This is a bit like trees falling in the forest. If there is nobody there to label it a NAND gate, then it will still do the exact same thing… but there’s no “NAND gate” there.
And, when the person does show up and label it, there’s still no “NAND gate” there… there’s just a label in that person’s mind, saying, “that thing there is a NAND gate”.
Not understanding this basic concept (that reality does not contain any labels, and has no “is-ness”) leads to all sorts of confusion.
(Sadly, this kind of confusion is also the natural human state.)
If it’s doing what a NAND gate does, it’s a NAND gate. Reality does not come pre-labelled,but things also do not spring into existence just because someone has labelled them.
Only if you think that “X is Y” means something other than, “My brain has associated the label Y with the cluster of sensory experiences denoted by X”.
I do: I think it means “X is a mind-independent object that would and should be labelled Y by an onlooker speaking my language”. I believe there are stars and planets no one has ever seen, or had a chance to label as such, Don’t you?
I think you’ve missed the part where that is still a label in your mind, being attached to a cluster of sensory experiences.
In such cases, the sensory experience clusters you’re labeling are memories associated with the labels “star” and “planet”.
However, this has little to do with an X-is-Y identity. In order to say “X is Y”, there has to be an X and a Y, and you are speaking only here of the hypothesized existence of various X’s that you would then label Y.
In any event, this and this are relevant here, in case you’ve missed them.
Maybe I should clarify a bit. I have two intuitions about the relation of consciousness and calculation. The first is that abstract existence of a computation, in the mathematical sense (where “X exists” basically means that the definition of X is free of contradictions), doesn’t guarantee consciousness. The computations should be physically implemented somewhere, by which I mean there should be a physical structure isomorphic to the abstract process of computation.
The second intuition is that the specific qualities of consciousness should be invariant with respect to some transformations of the physical implementation. One can get a quale of hearing a high-pitched sound by actually hearing it, or because of dozens of physically different causes, many of which lie inside the brain. And because not all details of the physical structure are important, there must be some property which the systems with indistinguishable qualia share, and there is a non-negligible chance that this property is computational isomorphism between these systems. So, I don’t express the same objection as you in other language, since I think there is a non-negligible probability that the simulation could be isomorphic to the real world to degree which enables the same qualia.
(Edit: even if the qualia of the simulated agents are different from qualia of the real agents, how does this constitute an argument against us being in a simulation? If so, we know our simulated qualia and not the real ones and can’t compare.)
The most confusing question to me is how the boundaries between different conscious systems are set, i.e. why aren’t there two or more consciousnesses in one brain or one consciousness in more brains. The question is not only confusing, but probably confused, but I don’t see a resolution. But this is off topic here anyway.
I would not bet much money on any of the above positions.
I disagree, and think that in any case substrate independence is of two types. The directions are: replacing basic units with complex units and replacing complex units with other complex units. Replacing basic units with complex units that do the same thing the basic unit did preserves equations that treated the basic unit as basic. I will attempt to explain.
Consciousness is presumably not a unique property of one specific system. If you’ve been conscious over the course of reading this sentence, multiple physical patterns have been conscious. I am quite different than I was ten years ago and am also quite different than my grandmother and someone living in an uncontacted tribe, also conscious beings. If all humans are conscious, no line between consciousness and non-consciousness will be found within the range of human brain variation.
Whole brains, complex things, can be replaced with giant lookup tables, different complex things, and not have consciousness. The output of “Yes” as an answer to a specific question may be identical between the systems, but the internal computations are different, so it is logically possible that the new computations are not within the wide realm of computations that produce consciousness.
Above I was referring to replacing complex biological units with complex mechanical units, in which “substrate independence” will depend on the specifics of the replacement done. However, all replacement of a unit that is basic with a more complicated unit that will give the same output for each input will leave the conscious system intact as the old equations will not be altered.
For example: suppose that a mechanical system of gears and pulleys produces knives (or consciousness) and clanks. It is possible to replace a gear with a sub-system consisting of: a set of range finders, a computer, mechanical hands, and speakers. The sub-system can measure what surrounding gears are doing and use the hands to spin gears as if the missing gear were in place, and use the speakers to make noises as if the old gear was in place.
Everything produced by the old system will also be produced my the new system, though the new system may also produce something else, such as GTA on the computer. This is because we replaced a basic unit with a more complicated system that produces additional things.
Similarly, replacing biological cells with analogously functional mechanical cells should certainly preserve consciousness. Probably, but not by logical necessity, cells are not needed to produce consciousness as a computed output.
tl;dr: computationalism implies substrate independence insofar as anything upon which computations act may be replaced by anything of any form, with the only requirement being to give the same outputs as the old unit would have. Anything a computation uses by mapping it first may be replaced by anything that would be identically mapped.
Agreed that “replacing biological cells with analogously functional mechanical cells should certainly preserve consciousness,” but this is a very limited sort of substrate “independence”. This approach makes the difficulty of producing an AI with consciousness-as-we-know-it much more severe. Evolution finds local optima, while intelligent design is more flexible, so I expect AI to take off much faster and more successfully, at some point, in a different direction, rather than brain emulation.
Like dfranke, I favor option #2, but like peterdjones, I don’t think it fits under “computationalism”.
This sounds an awful lot like “making the same argument that I am, merely in different vocabulary”. You say po-tay-to, I say po-tah-to, you say “computations”, I say “physical phenomena”. Take the example of the spark-plug brain from my earlier post. If the computer-with-spark-plugs-attached is conscious but the computer alone is not, do you still consider this confirmation of substrate independence? If so, then I think you’re using an even weaker definition of the term than I am. How about xkcd’s desert? If you replace the guy moving the rocks around with a crudely-built robot moving the rocks in the same pattern, do you think it’s plausible that anything in that system experiences human-like consciousness? If you say “no”, then I don’t know whether we’re disagreeing on anything.
I don’t necessarily understand your argument. Recall I don’t understand one of your questions. I think you disagree with some of my answers to your questions, but you hinted that you don’t think my answers are inconsistent. So I’m really not sure what’s going on.
Not every substance can perform every sub-part role in a consciousness producing computation, so there’s a limit to “independence”. Insofar as it means an entity comprised entirely of non-biological parts can be conscious, which is the usual point of contention, a conscious system made up of a normal computer plus mechanical parts obviously shows that, so I’m not sure what you mean.
To me, what is important is to establish that there’s nothing magical about bio-goo needed for consciousness, and as far as exactly which possible computers are conscious, I don’t know.
Plausible? What does that mean, exactly?
What subjective probability would you assign to it?
I don’t know what the “usual” point of contention is, but this isn’t the one I’m taking a position in opposition to Bostrom on. Look again at my original post and how Bostrom defined substrate-independence and how I paraphrased it. Both Bostrom’s definition and mine mean that xkcd’s desert and certain Giant Look-Up Tables are conscious.
The substrate independence of computation (without regard to consciousness) is well known, and just means that more than one material system can implement a programme, not that any system can. If consciousness is more “fussy” about its substrate than a typical programme, then in a strict sense, computationalism is false. (Although AI, which is a broader claim, could still be true).
I’ll accept option #2 as close enough to my view.
Wrt necessitating an “algorithms” view for q5… maybe. My initial answer there was to observe confusion, either in myself or the question, precisely in the area you point out here: it’s unclear how the labels “input” and “output” map to anything we’re talking about. I don’t reject your proposed mapping, but I don’t find it especially compelling either. I’m not sure that those labels necessarily do mean anything, actually.
Wrt not implying substrate independence: sure, I agree in principle; it’s not impossible that only protoplasmic substrates can implement consciousness. All I’m saying is that if that turns out to be true, it will be because certain kinds of computations can only be performed on protoplasmic machines.
Similarly, to say that heavier-than-air flight is a property of certain mechanical operations doesn’t imply substrate-independence for flight; it might be true that those mechanical operations can only be performed by protoplasmic machines.
That said, that would be a surprising result in both cases. Once we built/discovered a heavier-than-air nonprotoplasmic flying machine, the idea that doing so was impossible was rightly discarded; I expect something similar to happen with nonprotoplasmic consciousnesses.
As for strongly implying the absence of substrate-independence: sure, in the strict sense you mean it here, that’s true. Change the substrate and there will always be some difference, even if it turns out to be a difference you-the-observer could not conceivably care less about.
I suppose I could say my understanding of substrate-independence is implicitly a 2-place predicate: system S is subtrate-independent with respect to observer O iff O considers some system S2 identical to S, where S is implemented on a different substrate than S2.
A 1-place version, I agree, is unlikely on my view (its negation is, as you say, strongly suggested). I would also say that time-independence (that is, the idea that my consciousness is precisely the same from one moment to the next) is equally unlikely. I would also say that neither of these things matters very much.
Physicalists can reject substrate independence and accept the Church-Turing thesis, while still taking consciousness seriously. One can argue that consciousness in the physical world is implemented on protoplasm, and that this is the only kind of consciousness which is directly experienced. The fact that conscious beings can be simulated on a computer would be true but irrelevant.
Physcalists can’t reject substrae independence and accept the Computational Theory of Mind, however.
That is false, since we can build Universal Turing Machines (up to a certain finite memory) out of non-protoplasm, and a UTM can compute anything.
An observer-relative notion of computation is problematic for a computational theory of consc, since an observer-relative notion of consciousness is problematics. Surely the point is that I know i am conscious, not that he thinks I am.
You have a proof of the Church-Turing thesis? You should write it up and become famous in the CS community!
The other guy needs a disproof of the CTT...an effective procedure that can only be computed in protoplasm.
I guess the only quibble I would have, and I don’t know that it really changes your critique much, is that I wrote that neurons would be some sort of gate equivalent. I wouldn’t say that neurons have a simple gate model (that they’re simply an AND or an XOR, for instance). But I do see them as being in some sense Boolean. Anyway, I would just try to clarify my fairly short answer to say that I believe that computation can always be broken down into smaller Boolean steps, and these steps could be rendered in many different media.
Computationality in any fashion needs to be reified by physics doesn’t it? Otherwise it wouldn’t exist. Now, I would say it’s an emergent feature; physics doesn’t need to provide anything beyond what is provided for anything else to explain it. Maybe that’s the point of contention?
I’m not trying to hold you to any Platonic claim that there’s any unique set of computational primitives that are more ontologically privileged than others. It’s of course perfectly equivalent to say that it’s NOR gates that are primitive, or that you should be using gates with three-state rather than two state inputs, or whatever. But whatever set of primitives you settle on, you need to settle on something, and I don’t think there’s any such something which invalidates my claim about K-complexity when expressed in formal language familiar to physics.
Isn’t that only a problem for those who answer “functions” to question 5? Desert-rocks-AND-gate and MOSFET-AND-gate are functionally-equivalent by definition, but if you’re not excluding side-effects it’s obvious that they’re not computationally equivalent.
Edit: zaph answered algorithms, so your counter-argument doesn’t really target him well.
They’re computationally equivalent by hypothesis. The thesis of substrate independence is that as far as consciousness is concerned the side effects don’t matter and that capturing the essential sameness of the “AND” computation is all that does. If you’re having trouble understanding this, I can’t blame you in the slightest, because it’s that bizarre.
(Didn’t realize this site doesn’t email reply notifications, thus the delayed response.)
What I’m saying is that someone who answers “algorithms” is clearly not taking that view of substrate-independence, but they could hypothesize that only some side-effects matter. A MOSFET-brain-simulation and a desert-rocks-brain-simulation could share computational properties beyond input-output, even though the side-effects are clearly not identical.
(Not saying that I endorse that hypothesis, just that it’s not the same as the “side effects don’t matter” version.)
I wonder if I’m a qualia skeptic. I think that qualia are Humean impressions, the most “forceful and vivacious” contents of the mind. Dan Dennett has recently revived this view (without sufficiently crediting Hume, sadly); at one point he calls it the fame model of consciousness. What makes a thought conscious is that it does a lot; it has a very rich variety of interactions with other things going on in the mind.
This explains why there can be perception without consciousness; the much discussed (by philosophers) case of blindsight is an example where visual perception has a much more limited impact than usual, and so doesn’t have enough force and vivacity (or fame or clout if you prefer Dennett’s terminology, or whatever you want to call it) to feel conscious. And that’s why something like the David Lewis “mad pain” case is probably possible; the range of different interactions a conscious experience has is sufficiently great that even something lacking one of the core functions of experiences of a certain type could still probably feel pretty much like that experience if it had enough of the right secondary connections.
I think I’m talking about qualia when I talk about these Hume/Dennett items. But I’m talking about things with certain kinds of functionally defined inputs and outputs, a certain kind of computations, in fact. Does this mean I am not talking about qualia as you mean them? If not, then I stand with perplexed; references to qualia should be committed to the flames, for they can contain nothing but sophistry and illusion.
Commit the references to the flames, but not the referees? You are no fun! :) Though since you have invited Hume to join us, I suppose I am satisfied.
Your mention of qualia and functionality in the same paragraph caught my attention. Yes, indeed. If qualia were not functional, then they could hardly be intersubjective. And if they are functional, why the instinctive appeal of the idea that the inimitable ‘essence’ of qualia can not be generated by a simulation?
I don’t understand you comment about intersubjectivity. Qualia surely are not intersubjective in the sense of being publically accessible. If you just mean that qualia are broadly the same between people under she same circumstances, then that is given by supervenience, which AFAICS has nothing to do with functionalism.
I am philosophically unschooled, so I may misunderstand “supervenience”. I will take it to mean, roughly, that distinct instances of the same phenomenon will have features in common. Yes, but how do we know we are talking about different instances of the same phenomenon unless they have the same function. Cartoon dialog:
Joe: I feel something.
Mary: I feel something too.
Joe and Mary: We both feel the same way.
One doesn’t have to be a very strong skeptic to suspect that that third step was something of a leap. But perhaps less of a leap if what they feel is nausea after eating at the same restaurant.
We can say that the qualia will be the same if their supervenience bases are the same, and we can say that if they have the same properties. Non functional things like blobs of chewing gum still have properties.
Yes, and we determine those properties using senses that exist because, in other contexts, their use is functional. Do we have a ‘sense’ that detects the presence of qualia and apprehends their properties? If we do have such a sense organ, would you care to speculate on its function or lack of function?
I’m using functional to mean “something that has inputs, outputs, and internal workings”, not to mean “something that does something somehow”.
I don’t think we have such a sense. More importantly, nothing I have said implies it.
Ah! I was using it in the biological sense. As roughly the same as “purpose”. (You are, of course, welcome to add as many additional scare quotes as you think necessary to immunize us from the taint of teleology.)
It appears we have been talking past each other. This may be a good place to stop.
As I have already argued, it is not the case that everything is functional or has a functional analysis off the bat: that cannot be assumed apriori. I cannot see the functiona analysis of a blob of chewing gum or a magnetic field. Funcitonal things need well defined inputs, well defined outputs, and a well-defined separation between them and its inner workings,
Since funtionalism is not a universal apriori truth, I see no reason to “codemn to the flames” any non-functional notion of qualia.
I think we know what qualia are because we have them, But that is knowledge-by-acquaintance. It is again question-begging to say that the very idea of qualia has to be rejected unless they can be described. The indescribability of qualia is the essence of the Hard Problem. But we cannot say that we know apriori that only describable things exist.
Can you solve the “explaining colour to a blind man” problem with this proposal? I think not: vivid blue is just as famous and vivacious as vivid green, but that does not tell us what blue and green are...what their phenomenal feels are.
It is a little rude of you not to wait for me to answer before insisting that I can’t. And it wouldn’t hurt to be clear about what the problem is anyway. Color is extremely complicated, and most of the associations that make color perception conscious are not themselves conscious, so I personally certainly couldn’t explain color to a blind man. But there’s no problem there. And if someone did know enough about color to explain all the associations that it has, well, having associations explained to you isn’t normally enough for you to make the same associations in the same way yourself, so perhaps it couldn’t enable the blind man to imagine the color. But it’s hard to say, and anyway, I don’t know why he’d need to be able to imagine it to know what it was. I can say that when I’ve read articles about how echolocation works, and what sorts of things it reveals or conceals, I’ve felt like I know a tiny bit more about what it’s like to be a bat than I did before reading the articles.
I think you are interpreting Peter’s comment in an overly negative fashion. I believe he simply means “it seems that this proposal won’t solve the problem of explaining color to a blind person” or something close to that.
I suppose you’re right that I was a little snappy, but his response did seem to indicate that he wasn’t really paying attention. Indeed, my response to him was too charitable; on rereading Peterdjones comment, he seems to have been responding to some straw-man view on which I claimed it was vividness that made green green, while I responded as if he’d tried to address my actual view (that vividness makes green conscious, and other functional characteristics make it green).
The claim that consciousness is fame in the brain, and the claim that qualia are incommunicable because of complexity are somewhat contradictory, because what is made famous in the brain can be subjectively quite simple, but remains incommunicable.
A visual field of pure blue, or a sustained note of C#, is not fundamentally easier to convey than some complex sensation. Whilst there maybe complex subconscious processing and webs of association involved in the production of qualia, qualia can be simple as presented to consciousness. The way qualia seems is the way they are, since they are defined as seemings. And these apparently simple qualia are still incommunicable, so the problem of communicating qualia is not the problem of communicating complexity.
Something that is famous in the brain needs to have a compelling quality, and some qualia, such as pains have that in abundance. However, others do not. The opposite of blindsight — access consciousness without phenomenal consciousness — is phenomenal consciousness without access consciousness, for instance “seeing something out of the corner of ones eye. Not only are qualia not uniformally compelling, but one can have mental content that is compelling, but cognitive rather than phenomenal, for instance an obsession or idée fixe; .
“And if someone did know enough about color to explain all the associations that it has, well, having associations explained to you isn’t normally enough for you to make the same associations in the same way yourself, ”
To some physicalists, it seems obvious that a physcial description of brain state won’t convey what that state is like, because it doesn’t put you into that state. Of course, a description of a brain state won’t put you into a brain state, any more than a description of photosynthesis will make you photosynthesise. But we do expect that the description of photosynthesis is complete, and actually being able to photosynthesise would not add anything to our knowledge. We don’t expect that about experience. We expect that to grasp what the experience is like, you have to have it. If the 3rd-person description told you what the experience was like, explained it experientially, the question of instantiating the brain-state would be redundant. The fact that these physicalists feel it would be in some way necessary means they subscribe to some special, indescribable aspect of experience even in contradiction to the version of physicalism that states that everything can be explained in physicalese.Everything means everything — include some process whereby things seem differnt from the inside than they look from the outide. They still subscribe to the idea that there is a difference between knowledge-by-aquaintance and knowledge-by-description, and that is the distinction that causes the trouble for all-embracing explanatory physicalism.
Weaker forms of physicalism are still posible, however.
“can say that when I’ve read articles about how echolocation works, and what sorts of things it reveals or conceals, I’ve felt like I know a tiny bit more about what it’s like to be a bat than I did before reading the articles.”
But everyone has the experience of suddenly finding out a lot more about something when they experience it themselves. That is what underpins the knowledge-by-acquaintance versus knowledge-by-description distinction.
I think that the “Mary’s Room” thought experiment leads our intuitions astray in a direction completely orthogonal to any remotely interesting question. The confusion can be clarified by taking a biological view of what “knowledge” means. When we talk about our “knowledge” of red, what we’re talking about is what experiencing the sensation of red did to our hippocampus. In principle, you could perform surgery on Mary’s brain that would give her the same kind of memory of red that anyone else has, and given the appropriate technology she could perform the same surgery on herself. However, in the absence of any source of red light, the surgery is required. No amount of simple book study is ever going to impact her brain the same way the surgery would, and this distinction is what leads our intuitions astray. Clarifying this, however, does not bring us any closer to solving the central mystery, which is just what the heck is going on in our brain during the sensation of red.
To say that the surgery is required is a to say that there is knowledge not conveyed by third persons descriptions, and that is a problem for sweeping claims of physicalism. That is the philosophical problem, it is a problem about how successful science could be.
The other problem, of figuring out what brains do, is a hard problem, but it is not the same, because it is a problem within science.
No it isn’t. All it says is that the parts of our brain that interpret written language are hooked up to different parts of our hippocampus than our visual cortex is, and that no set of signals on one input port will ever cause the hippocampus to react in the same way that signals on the other port will.
But if physicalism is correct, one could understand all that in its entirety from a third person POV, just as one can understand photosynthesis without photosynthesising. And of course, Mary is supposed to have that kind of knowledge. But you think that knowledge of how her brain works from the outside is inadequate, and she has to make changes to her brain so she can view them from the inside.
The very premise of “Mary is supposed to have that kind of knowledge” implies that her brain is already in the requisite configuration that the surgery would produce. But if it’s not already in that configuration, she’s not going to be able to get it into that configuration just by looking at the right sequence of squiggles on paper. All knowledge can be represented by a bunch of 1′s and 0′s, and Mary can interpret those 1′s and 0′s as a HOWTO for a surgical procedure. But the knowledge itself consists of a certain configuration of neurons, not 1′s and 0′s.
No, the premise of the Mary argument is that Mary has all possible book-larnin’ or third person knowledge. She is specifically not supposed to be pre-equipped with experiential knowledge, which means her brain is in one of the physical states of a brain that has never seen colour.
No, she is not going to be able to instantiate a red quale through her book learning: that is not what is at issue. What is at issue is why she would need to.
Third person knowledge does not essentially change on translation from book to paper to CD, and for that matter it should not essentially change when loaded into a brain. And in most cases, we think it doesn’t. We don’t think that the knowledge of photosyhtesis means photsynthesising in your head. You share that the qualiaphobes assumption that there is something special about knowledge of qualia that requires instantiation.
Well, then when she steps outside, her brain will be put into a physical state that it’s never been in before, and as a result she will feel enlightened. This conclusion gives us no insight whatsoever into what exactly goes on during that state-change or why it’s so special, which is why I think it’s a stupid thought-experiment.
It isn’t intended to answer your question about neuroscience.It is intended to pose the philosopher’s question about the limitations of physicalism. If physicalism is limited, that eventually folds back to your question, since one way of explaining the limitation of physicalism is that there are non-physical things going on.
When she steps outside, something physical happens in her brain that has never happened before. Maybe something “non-physical” (huh?) also happens, maybe it doesn’t. We have gained no insight.
If we agree that she learns something on stepping outside we have learnt that a version of physicalism is false.
Can you state what that version is? Whatever it is, it’s nothing I subscribe to, and I call myself a physicalist.
There are broadly speaking two versions of physicalism: ontological physicalism, according to which everything that exists is material, spatio-temporal, etc; and epistemological physicalism, according to which everything can be explained in physical terms. Physicalism can be challenged by the inexplicability of qualia in two ways. Firstly, qualia might be physically inexplicable because they are not physical things, which contradicts ontological physicalim. Secondly, the phsyical inexplicability of qualia might be down to their having a first-person epistemology, which contradicts epistemological physicalism. Epistemological physicalism requires that eveything be explicable in physical terms, which implies that everything is explicable in objective, descriptive, public, third-person terms. If there are some things which can only be known by acquantance, subjectively, in first person terms, then it is not the case that everything can be explained in physicalese. However, ontological physicalism could still hold.
Or, you could notice that the apparent inexplicability of qualia is a sign that you are confused. ;-)
OK. You understand qualia, Please de-confuse me on the subject.
Have you read the LessWrong Sequences yet?
Some of them. I have read Block, Chalmers, Dennet, Flanagan, Jackson, Levin, Nagel, Searle, etc, etc as well. Which sequence did you have in mind?
For this discussion, the one of principal relevance is the one on the use of words, especially the mind projection fallacy (including the non-existence of mental or supernatural entities). Reductionism would be useful as well, and quantum physics.
The quantum physics part is particularly helpful for disabusing one’s self of many naive intuitions about object identity, that otherwise lead to belief in things like souls, or consciousness as something separate from bodies, or the idea that an exact duplicate of you wouldn’t actually be you. If you don’t get at least that much about the basics of physics, then it’s way too easy to believe in fairy tales when they have words like “consciousness” and “qualia” attached.
In other words, human beings are born with various intuitions (hardwired into the brain, as has been shown by experiments on babies who can’t even talk yet) that, without sufficient education, we use as the basis for reasoning about minds and reality. Huge amounts of philosophy and “common sense” reasoning are then based on these false premises.
Of course, this makes most philosophical discussions equivalent to nothing but hot air: reasoning based on false premises. Attempting to refute the conclusions without first refuting the premises is pointless, which is why I keep pointing to the Sequences. They contain the necessary information to refute the premises that support the vast majority of philosophical and supernatural nonsense. (Such as some of Chalmers’s and Searle’s, for example.)
Of course, if you don’t agree that physics, cognitive science, and Bayesian updates based on them are the basis for reaching an objective conclusion, then this discussion is entirely moot. That’s why I asked whether you’ve read the sequences—and implicitly, accept their premises about the nature of reasoning, as well as the specific facts of physics and cognitive science—so I can determine whether there’s anything worth talking about.
Why did you link to Zombies: The Movie, which is really fun for people who already understand that the concept of irreducible qualia is nonsense, rather than Zombies! Zombies?, which explains why it is nonsense?
Because “Zombies: The Movie” demonstrates why it is nonsense, and is thus harder to tune out and argue with.
Also, I personally found it much more convincing on the first read than the complex arguments of the other article. ;-)
Zombies?! I never said a word about zombies...
It would have been helpful to say how it is relevant.
That mental entities don’t exist at all is a very bold claim: much bolder than the claim about the supernatural that you bracket it with, and one that many physicalists would disagree with. Moreover, neither claim follows from the very general consideration (which in itself I do not contend) that there is a “mind projection fallacy”.
What do you mean by “reductionism would be useful”? If there were a generally accepted reduction of qualia, there would be no problem of qualia. There isn’t such a reduction. So are you talking about promissory reduction (we have to believe it will arrive one day)....or what?
I have studied quantum physics, and I don’t think my ideas about qualia are based on naive ideas about identity. I think they are based on what I have said they are based on. If you have a criticism that is relevant to something I have said, I will be glad to hear it. I would rather you did not guess at my motivations.
Calling something a “fairy tale” is not an argument. I am still waiting for an argument relevant to something I said.
That argument is a non sequitur. The fact that an intuition is hardwired does not make it false.
That’s another non-sequitur based on the on the previous one: you haven’t shown that philosophical arguments are mostly based on intuitions. Moreover, you are in danger of throwing out the arguments of your fellow qualiaphobes, such as Dennett.
They contain a bunch of stuff about logic and language that most philosophers (in the anglosphere at least) are vey familiar with. I have read arguments for and against qualia, and found them both to be based on reason. I think it is possible for reasonable people to disagree deeply.
And the equation between philosophy and the supernatural remains uninformed, to say the least.
OK. Someone doesn’t like Chalmers’s Zombie argument. Guess what? I don’t like it either. I know it is possible to have qualiaphilia without p-zombies because my own qualiaphilc arguments work that way. I never mentioned zombies in the first place...
The mind projection fallacy (or more specifically, the Less Wrong sequence on it) is more than sufficient as an explanation for how mental and supernatural entities are perceived; what “many physicalists” may or may not believe is not really relevant here.
I’m saying the Less Wrong sequences on reductionism and quantum physics will be useful in dissolving your confusion about qualia.
But not Bayesian evidence, which is what’s relevant on LessWrong.com. This is a community devoted to furthering the practice of Bayesian rationalism, not the discussion of philosophy in general, or what philosophers consider to be reasonable or not reasonable. This is a community that considers dissolution of the confusion about “free will” to be a basic exercise in rationality, rather than an abstruse philosophical question requiring years of argument, or something that’s still considered an unsettled open question, subject to disagreement.
...and on LessWrong, we agree that’s true… IF and only if one or more of these conditions apply:
The reasonable people have different information,
The reasonable people are using different methods of evaluating the same information (due to e..g different values/desires), or
One or both of the “reasonable” people aren’t really reasonable at all
If you believe that there is some other way for reasonable people to disagree, then it’s a good indication that we’re not on the same page enough to bother talking about this at all.
If you think that’s just an opinion, you don’t get Bayesianism yet; that’s why I suggested the Sequences to you, in case you’re genuinely interested in being able to settle philosophical arguments once and for all, instead of just having philosophical arguments. ;-)
You didn’t need to. Any argument for epiphenomenalism reduces in roughly the same way: if it has an effect, then the effect is phenomenal and reducible. If it doesn’t have an effect (i.e. produces no difference in our predicted observations), why do we care?
Ontologically fundamental mental entities of any sort require one to think of the mind as a supernatural entity, rather than a physical one. But it’s very hard to notice that you’re doing this, because it’s implicit in how we think about thinking by default.
What’s a “supernatural entity”? The word ‘supernatural’ is ill-defined: if something exists in the real world, then it is natural by definition.
For the record, I don’t think minds are ontologically fundamental per se, because minds are far too complex and they’re explained already by physical brains. But it may be that some precursor of subjective experience is fundamental.
Yeah, just like the word ‘metaphysics’ is ill-defined. If something exists in the real world, then it is physical by definition.
Or to be even more snarky but at least more explanatory: I doubt that ‘exists’, ‘physical’, ‘meta-’, ‘super-’ or ‘natural’ are sufficiently well-defined in these contexts for your accusation of ill-definition to hold any weight. If I try to interpret what you’re saying in roughly the same manner in which it seems to me that you’re interpreting what most folk mean by ‘supernatural’, except instead of being uncharitable in the direction of being snobbishly literal I reverse it and be uncharitable in the direction of not paying attention to your explicit message, it looks something like this: “People who use the word ‘supernatural’ tend to be wrong in obvious ways and I like to point this out in a mildly intellectual fashion so that I can feel superior to them; also since I just denounced the enemy tribe you should like me more”. But that would be no more accurate a characterization of what you meant, than your characterization of what is typically meant by ‘supernatural’, and nobody on either side would learn anything from such analysis.
(This comment is not really a reply to User:bogus so much as an expression of annoyance at certain traditional rationalist memes. Sorry you got caught in the crossfire, User:bogus.)
The naive impression of “mind” in general philosophical discussion is a good example of a supernatural entity—the concept of mind separated from a specific human brain, some almost spirit-like entity.
In order to commit the mind-projection fallacy, you have to forget (really: not notice) that your brain actually exists and is not an objective observer of fact, but only an opinion-generating machine. Thus, discussions of consciousness and “qualia” are hugely hampered by forgetting that the mind is not an abstraction, it’s a specific physical thing, and that the various properties being attributed to it in these discussions exist only in the brain of the beholder, rather than in the thing being discussed. (As a natural consequence of physics not having layers or levels.)
Exactly.
Well, I don’t have a naive conception of the mind, and I do remember my brain exists, so I am not committing the MPF. Hurrrah!
I disagree.I don’t see the specific application at all.
OK.I’m saying I already know quite a lot about both subjects, and I don’t see the application. You need to stop assuming that I am ignorant, and start putting forward relevant arguments. Repetition of “you are confused” won’t cut it.
I don’t see the relevance of Bayes. The topic is at the level of of clarifying concepts, not of making computations on datasets.
To say that qualia don’t exist, as you have been, is philosophy in general. To say that philsophy as a whole is wrong-headed, as you have been, is metaphilosophy. Your position is inconsistent. You say both that philosophy is wrong headed and that a certain philosophical problem is (dis)solved in the Sequences (in a typically philosophical way, dismissed as a verbal/conceptual confusion).
If it is a community based on reason, it will be open to reasoned objections.
That seems laughably naive to me. You don;t have an algortihm for settling phil. arguments, because they do depend on evaluations, and other stumbling blocks you haven’t thought of. You think it is just obvious that we should ditch the idea of qualia to retain physicalism and avoid epiphenomenalism. But that isn’t an obvious objective fact which other people are toostupid to understand: that is you de-valuing qualia and subjective experience.
I didn’t mention epiphenomenalism either, and I don’t believe in it..or, rather, I value theories that avoid it.
I haven’t said qualia are fundamental. and they are not defined that way.
If you don’t think that these arguments can be settled, there is no point in continuing this discussion.
And if you don’t think that Bayes matters to updating your beliefs, then you are not a Bayesian rationalist.
The reason I asked about the sequences was to find out whether you were someone trying to learn an application of Bayesian rationalism, or someone who’s just trying to have a philosophical argument.
Apparently, you fall in the latter category, which means I have no interest in continuing the discussion.
What is considered “reasoning” by philosophy doesn’t reach the level of rigor that is required here… as was amply demonstrated by statements of yours such as:
They only depend on evaluations if you’re interested in having an argument, as opposed to finding the truth (with or without a capital T) of a situation. Here, we expect arguments to be supported (or at least not opposed) by physics and cognitive science, in order to be considered “reasonable”, and we expect that hypotheses not be privileged.
I don’t think they have been settled. And I think there is value in reversing the Dunning Kruger Effect: getting someone to realise how difficult something really is.
I didn’t claim to be a Bayesian or not. I am comparing Bayes to Popper and various other things at the moment. What I did say, and stand by, is that the formal part of Bayes is only applicable to problem areas that have already been marshalled into a less ambigous and non-linear form than typical phil, problems.
You can say you have some wonderfully high level of reasoning, but I don’t have to believe you. I can judge from the examples supplied. You have not applied Bayesian reasoning as a formalism to any problem. and the material you directed me to in the sequences didn’t either. It is all typical philosophical reasoning, neither particularly good not particularly bad.
Ie...you value science.
But the idea that just by basing your philosophical arguments on science, you can Avoid Arguments and Find Truth is very naive. Most English-speaking philosophy is science based, and is full of plenty of disagreements. Why don’t you know that? Oh yeah: the Dunning-Kruger effect means that the less someone knows about a subject, the more they over-estimate their own abilities at it...
I have. I don’t understand qualia either. Do you have a particular relevant link you were thinking of?
ps: “You should really read the sequences” is telling people to read 1,000,000 words or go away, and as such is functionally equivalent to an extremely rude dismissal. Please don’t do that. Link to a particular post if you actually think the pointer is helpful, i.e. make it an actually helpful pointer rather than a functionally rude one.
There is a difference between telling someone to go read the sequences, and asking if someone has read the sequences. If you ask, the other person is allowed to say “no, I haven’t”, and that is useful to know what sort of inferential distances would be appropiate in your explanation.
I think that’s splitting hairs. It is still a functionally rude response in the guise of one that isn’t, and it’s still not actually a helpful response.
Particularly when I know the sequences don’t actually explain what qualia are such that we should care about them rather than dimissing them as a circular argument for magic. This post calls them things that “cannot arise from computable processes”, which doesn’t actually answer the question. Are there any I’ve missed that do?
It seems like you think PJEby was implying with his question that anyone who has read the sequences would understand qualia. I don’t think that is what he actually meant.
I would agree that it would have been better if he had been more explicit about why he was asking.
Yeah, it seems to me that it’s usually intended to ask something more like “Can I use terminology and concepts found in the Sequences without explaining them further or explicitly linking to them?”, though I’ll agree that we should replace “Have you read the Sequences?” with something more specific in any case. (Of course, if someone wants to claim that a particular set of posts directly answers someone’s question, it’s always best to point to the specific posts. Unless they depend on so much before them that it makes sense to link to an entire sequence, which is probably unavoidable sometimes.)
http://wiki.lesswrong.com/wiki/Jargon needs expansion (and isn’t formatted much like a jargon file for newbies).
Pjeby’s use of the word ‘yet’ leads me to agree with David on this one.
I used “yet” because the specific person I was replying to appears to be a new user who has joined specifically to discuss this topic, while being unaware of the relevant basics on the topic.
Based on my own recent past discussions with pjeby, I think that’s what he meant.
No, what I meant is that if you’ve understood the sequences, then you would be able to dissolve questions regarding qualia, in the same way that you’d be able to dissolve questions regarding “free will”, and other confusing ideas.
Pjeby, the specifics of dissolving each dissolvable question are different.The word “sound” may have two definitions, once we tell them apart the question of falling trees producing sound gets dissolved. The word “free will” has a fuzzy definition, once we define it precisely, the question gets dissolved. The word “qualia” on the other hand is about categories of subjective experience which we cannot define or communicate adequately, and we aren’t sure if they’re even communicable, because they’re too much tied in with the concept of subjective experience, anticipation, and these in turn seem tied in with other unsolved questions like existence, causality and the fundamental relationship between mathematics and physics.
So, yes, perhaps the issue of qualia will one day be dissolved, but we’ve not managed to do so yet. If you can tell us the exact way to dissolve it, please oblige us. “Read and understand the sequences” is obviously not sufficient for us, so either you’re a few levels above us on this matter (by being able to dissolve this confusion), or we’re a few levels above you (in actually noticing ours).
Either way, telling us “just understand the sequences” isn’t sufficient or helpful, for us atleast.
On the matter of qualia. A truly naive person, untutored in philosophy, would not understand claims about qualia. He would say something like, “I see an apple. The apple is red.” He would assign redness to the apple itself—and so on like that. You on the other hand, having been through the process of tutoring in which your teachers directed your attention at the qualia themselves, as distinct from physical objects which may by impinging on our senses give rise to them, would know better than the naive person. But importantly, the difference between you is not that you have very different evidence in front of you, but that you have very different conceptual tools with which to think about that evidence. The possibility arises that your concepts are deeply flawed, and have led you to grossly misidentify the evidence in front of you.
One of the methods by which we come to recognize the qualia themselves, as distinct from the objects that give rise to them through their effects on our senses, is to imagine that there is no apple (say) in front of us, but something else indistinguishable from it (to us). So two very different things can be indistinguishable to us. Nevertheless (so the reasoning can go on) plainly something has not changed: namely, our own subjective experience remains the same, despite variation in its cause. This realization directs our attention at that which remains the same despite the physical changes. Voila, we are now attending to our qualia.
But the reasoning is flawed. The fact that we cannot distinguish between two things does not mean that when one replaces the other, something remains the same. Our senses of course do insist to us that something remains the same (which we will eventually identify as the quale). By by assumption, they are not to be trusted!
This is one of the flawed paths of reasoning by which we may come to believe in the existence of qualia.
The flawed path you’re attacking is not one I followed, and so it just seems a strawman to me.
But if you’re going to argue that qualia don’t exist, then I’ll have to believe that you’re the one confused, not me—the existence of qualia, the existence of my current subjective experience is the only thing I know for certain, more certain by far than I can of the existence of the physical world. Whether I’m in the Matrix, or in a true flesh body, or if perhaps my consciousness is simulated by hordes of monks using abacci to simulate my neurons… whether reality itself is a lie… I can never place absolute certainty on those things.
But one thing I know for certain: I experience qualia, more certainly than 2+2=4.
No, I am not saying that qualia are something well-construed which does not exist, like a tea pot orbiting Mars, or like an aquatic dinosaur living today in a lake in Scotland. I am saying that the concept of qualia critically misconstrues the evidence. In fact your own statement here nicely illustrates what I take to be the reason the concept of qualia exists and persists, which is that by its very conception it provides the philosopher with something that he is supremely confident in. To employ my example: the philosopher does not know whether he is looking at an apple, or a plastic fake, or whether he is a brain in a vat hooked up to a computer simulation of an apple. What can he do to escape his uncertainty which grows with each new possible scenario that he imagines? Why, in the end it turns out to be simple: he declares that all these possibilities which he is unable to tell apart share a common element, and that element is a common subjective experience. Thus he turns his own failure (to tell things apart) into a supposed success (the supposed discovery of a common element). Voila: qualia. The concept is in effect defined to minimize uncertainty. So for you to write:
strongly confirms my own view about the nature of the philosopher’s concept of subjective experience, of qualia.
I always enjoy reading your contributions to these philosophy of mind debates, even when I disagree on some fairly minor points.
This is a rather Wittgensteinian diagnosis—it reminds me of Philosophical Investigations §308, for instance.
To play devil’s advocate I shall temporarily pretend to believe in qualia.
If I am unable to distinguish between visual stimuli A and B then some properties must be remaining the same each time, namely whatever it is that predisposes me to judge that present stimuli stand in such-and-such relations to other stimuli. (E.g. “this green thing, whatever it is, has a lighter, more garish shade than the leaves of that tree over there.”) Qualia are precisely those properties of my subjective experience which enable me to make these judgements. To be conscious of X at all implies the ability to synthesize X with other mental contents Y and make judgements like the comparison above. So since I am conscious, my mental contents must have qualia.
Therefore, I can no more doubt the existence of qualia than I can doubt the fact that I am conscious.
Not necessarily. Here’s what could be happening in your brain and which could underlie your power of discrimination: there is some operator which takes two inputs and produces an output “same” or “different”. Since this operator could, in principle, be anything at all, then it is in principle possible for any two arbitrary inputs to be assessed as “the same” by the comparison operator. They don’t have to have anything in common.
That’s in principle. In reality, we don’t expect brains to be so badly designed. We expect that under normal conditions (though not necessarily in highly artificial laboratory setups which test the limits of perception, such as for example “change blindness” experiments) the comparison operators which operate inside the brain output “the same” only when the two inputs are from objects in the real world which are really pretty similar in some important way. And in the case of an apple versus a fake plastic apple, there is something that does remain the same: the pattern of light traveling in the space between the object and your eyes.
But this common pattern of light is surely not a quale. It’s not even in the brain. In fact, the sameness of the two patterns can be demonstrated by taking a digital photo of each scene and then having software compare the two photos. Can you take a photo of a quale?
We have jumped from “some properties must be remaining the same”—which I’ve acknowledged is probably typically the case in normal circumstances—to “properties of my subjective experience”. So we’ve slipped in the term “subjective experience”.
Fine. I have evidence for qualia, but it is not certain. Does that change anything?
Yep.
Fundamentally, though, all of these sort of mistakes arise from assuming that conceptual entities have some sort of existence outside of the mind of the conceiver. “Qualia” are just one example of such conceptual entities.
Some concepts refer to entities outside the mind, some to mental entities, and some don’t refer. So the observation that something is “conceptual” tells us nothing, basically. The phrase “conceptual entities” seems empty to me. Did you mean something like “only and purely conceptual entities”.
And all of them are physically represented in the brain. And even the ones that refer to outside reality, are an arbitrary division. In other words, physics doesn’t have layers—layers exist only in brains.
That’s why, when you make claims about qualia or consciousness as if they were something that existed outside of some particular observing brain (not the one within which they are deemed to exist), you’re making a mistake about physics, as well as philosophy, and committing the mind projection fallacy at the same time.
I don’t understand that. Please give an example.
How true. Oh, and there’s no guarantee that any particular question is disolvable ahead of disolving it....
In these discussions I’ve concerned myself only with improper use of the word “qualia” to support mystical arguments that attempt to place human consciousness into a special category exempt from simulation or duplication. That is, arguments that attempt to use “qualia” to justify naive human intuitions about consciousness.
The rest of what I’ve seen has been questions or arguments roughly equating “zombie worlds”… that is, ones where the presence or absence of the thing described yield no difference in predictions for the currently-observed world, and thus (AFAICT) meaningless, and therefore pointless to talk about.
There may exist non-confused, non-meaningless topics that somehow involve qualia, and even open questions for research. So far, however, I have not seen them in any recent LW discussion. (With the caveat that I have not been reading much outside the replies to my own comments, and the random bits that catch my eye on the main comments page.)
I don’t think anyone here has ever argued that qualia can’t be duplicated—what we argue is that perhaps they can’t be duplicated in a qualia simulator: same way that a gravity simulator can’t actually duplicate gravity. You need mass to duplicate gravity, a Turing machine doesn’t suffice.
And that argument is a basic MPF error, that you should be able to see through if you really understood “how an algorithm feels from the inside”, or the rest of the mind projection fallacy sequence.
In order to claim that qualia can’t be duplicated in a qualia simulator, you are claiming that a purely mental property exists, outside physical reality.
After all, if a simulated person behaves in exactly the same way as a non-simulated one, we have no evidence regarding the state of these hypothesized qualia, one way or another.
The untrained mind takes this to mean that there must be no qualia in the machine… and only the machine, instead of realizing that this just means there’s no such thing in the first place. That we only think they exist because that’s how the algorithm feels from the inside—that is, the algorithm that our brain has for labeling things in the world as minds.
Our brains are built to suppose that things which move by themselves have minds and intentions. We can learn that things do not have minds, or that they do, but this labeling carries with it a host of specialized biases in our thinking.
And when you point this assembly of biases at something that is otherwise very simple, it’s trivial to see that the question isn’t, “can machines duplicate qualia?”, because clearly, we are machines, so the question is silly.
It only seems like a question, because our brains have prebuilt categories for “animate” and “inanimate”, and so the question feels like a big mystery to us… “how could something inanimate be made animate?”
So, the whole concept of qualia (as applied to this topic) is basically the human brain grasping at straws to preserve its inbuilt intuition that these are separate categories, instead of simply dropping them to realize that we are all made of the same stuff as machines are, and there’s absolutely no evidence for the animate-inanimate distinction being anything other than an evolutionary convenience. It’s not a “natural” category or distinction, outside the human mind.
Um, no. The whole point of a qualia simulator is that it’s running on a physical substrate which may or may not instantiate qualia itself. This is a property of physical reality, not one of mind or algorithms.
Not quite. I may have no direct evidence that the machine has any subjective experience or qualia, but then again, the only subjective experience I have any evidence of is my own. Nevertheless, I can reasonably suppose that subjective experience also applies to other people, because I share a common body structure and biological ancestry with them, and as a physicalist I believe that subjective experience depends on some kind of physical substrate. But as far as behavior goes, I would regard the output of a qualia simulator (say, a whole-brain emulation) as being indistinguishable from any other mind.[1]
It’s not obvious that an algorithm should feel like anything, per se. Will the algorithm feel the same when run on an Intel Core computer, or a SPARC workstation?
[1]And yes, the simulated minds could talk about their own subjective experiences. If these subjective experiences are instantiated somehow as part of their algorithm, we could then take that as their equivalent to our qualia. But these minds would not share our physical substrate, so the status of these “qualia” would be radically different from our own.
I’m using the phrase in this sense… that is, the distinctions that are available for an algorithm to make—the states that are reachable, in some sense.
Human brains have special states to represent conscious or “intentional” animate entities, so to us it “feels” as though this category is special.
And what’s your extraordinary evidence for that extraordinary claim? You’re basically claiming that there is something special about human brains that makes them different. Why is that?
How did you arrive at this hypothesis, as opposed to any other?
Obvious answer: you privileged this hypothesis, out of any number of equally complex hypotheses, because it has inutitive appeal. Your brain has a special category for this, so it feels sensible.
But it isn’t sensible as a hypothesis, because if you didn’t have that special category already built into your brain, you would have no reason to single out human brains as being, not only the only substance in all the world that currently has these miraculous qualia, but to also be the only substance that ever will have them.
However, if you set your intuition aside, then there is no other reason whatsoever to single out such an extraordinary hypothesis for special consideration.
Obviously, regarding a system as complex as the human brain as “the only substance in all the world that currently has these miraculous qualia” is an unlikely hypothesis. Nevertheless, subjective experience could be instantiated in a simpler physical system as a result of brain activity. There is plenty of precident for biological lifeforms tapping into “exotic” physics for some of their adaptive functions, and subjective experience might be no different.
It’s not an extraordinary claim. As a physicalist, if I’m going to take subjective experience seriously as anything other than what some algorithms (or minds) like to talk about, then it’s reasonable to suppose that the physical substrate matters.
In a way that produces no distinguishable physical effect?
You seem to be hypothesizing the existence of an invisible dragon. Why? Explain to me how you came to select this hypothesis, out of all similarly complex possible hypotheses.
Like, for example, let’s say you grab a philosopher out of the past who insists that women don’t have men’s reasoning power because, you know, they’re not men, and that surely must be some physical reason why this is so!
Wouldn’t you want to know why he hypothesizes this? What his evidence is? Why he insists that, even if a woman were—hypothetically speaking, in his view—to make the same statements or draw the same conclusions as a man, from the same inputs as a man… then somehow, she still wouldn’t “really” be reasoning like a man, because she has female “qualia” instead of male “qualia”?
Pretty soon, you’d have to come to the conclusion that he’s arguing from the bottom line: trying to provide argumentative support for a conclusion he already had before he started, rather than simply investigating what truth there was to be found.
(Especially if he has nothing in the way of physical evidence for the existence of these “qualia” things… but appears to have just seized on them as a way to justify the already-existing intuitions.)
Whatever. How about the arguments of qualiaphiles who don’t think simulations will necessarily lack qualia?
I treated qualia the same way I treated gravity—does that mean I must also believe gravity to be a purely mental property outside physical reality?
The question wasn’t if machines can duplicate qualia, but if Turing Machines can duplicate qualia.
Turing machines can’t duplicate gravity. Turing machines can’t duplicate sound. Turing machines can’t duplicate lots and lots of physical phenomena. Machines can duplicate these things, Turing machines can’t.
Actually, you didn’t.
You have exactly the same evidence for other people having qualia, as you do for a simulation of a person having qualia: namely, that the person or simulation says they do.
Then the question is whether reality is Turing computable, and we currently have no reason to believe it isn’t.
(Or, more precisely, we have no evidence that we would experience anything differently if our universe were itself a simulation running on a Turing machine.)
Not exactly the same. My own internal experience is applicable with more certainty to other flesh brains that use the same physical processes as mine than it is to electronic simulations thereof.
And my whole point is that reality being Turing computable doesn’t mean that reality is Turing-producable. Same way that a black hole can be Turing computable but not Turing-producable.
We have no evidence to believe that anyone would experience anything on a Turing machine algorithm. Turing machines produces an output whose meaning is deciphered by external observers according to their own definitions. There’s no magic in the Turing Machine that creates an independent universe with its own internal meaning and internal experience—the Turing Machine returns some computations to us, to our universe, not to some different internal one.
It’s the belief the Turing Machine creates an internal universe with its own life that is a confusion.
Unless Tegmark IV—in which case we don’t even need the Turing machine, the math alone would sustain said internal universe.
Why, if it’s a complete simulation of those “same physical processes”?
You’re missing the privileged hypothesis here. The hypothesis you’re privileging is, “consciousness or qualia are physical side-effects of the brain’s operation”… a proposition for which no evidence exists, first of all.
You’re then using this hypothesis as a basis for an extended argument for the possibility that, if this hypothesis were true, then simulation of consciousness wouldn’t really be the same as regular consciousness.
But, you’re not only privileging that hypothesis (i.e., putting it forward without any evidence), you’re not even thinking through the ramifications of it being true!
Because if it’s true, then either those physical effects feed back into the operation of your brain, or they don’t.
And if they don’t, then, what do they matter? Why should we care about them, any more than we care about how much heat your brain radiates as a side-effect of its operation? (And how could this side-effect be perceived by you as “qualia” in that case?)
And conversely, if they do feed back into your brain’s operation (as they would have to in order for you to perceive them!), well, then a simulation which lacks this element will diverge in behavior from a real human, and thereby make apparent what is missing from the simulation.
In other words, you’re simply making another zombie-world argument. (That is, you’re making another “epiphenomena” argument—that some mysterious thing exists which has no effects and yet somehow matters, thereby providing a rationale for the bottom-line intuitions you started out with.)
For the same reason that a complete simulation of a plane, down to the quark level, still won’t fly.
That’s vaguely phrased. There may be no direct evidence that qualia are epiphenomenal side effects, but there is plenty of evidence that they covary with brain operation.
It is perfectly possible for qualia to be causally effective in brains, and to be missing from simulations, and for the simulation to be behaviourally identical. It For instance, transistors are causally effective in electronic computers. But any electronic computer can be rebuilt as a behaviourally identical optical, mechanical, hydrualic (etc) computer.
Because I understand the word “simulation” to mean an algorithm which can answer questions about the states of the thing it simulates; not something which actually performs those same processes. The simulation of an Boeing can’t actually fly actual people to actual America. Why should the simulation of a brain actually experience actual qualia?
The confusion doesn’t lie with me about this—nobody, not even you, has ANY problem understanding why the simulation of a physical phenomenon (e.g. gravity, flight, fire, weather patterns) is different to the physical phenomenon itself. And yet many people have trouble understanding why the simulation of a mental phenomenon may be different to the mental phenomenon itself.
Not so. The zombie-world argument is about an atom-for-atom identical to our world reality whose people nonetheless don’t experience qualia, even though they talk about them.
A simulation of a thing isn’t atom-for-atom identical to the thing it simulates. It might compute every single atom, so that it produces answers which an outside observer can somehow use their own cognitive processes to map its responses onto the relevant thing. But the map isn’t the territory.
Unless, again, Tegmark IV—keep in mind that my own position is NOT actually “No qualia in simulation” but rather “Tegmark IV or no qualia in a simulation”.
I certainly believe qualia have an effect, including the fact I’m talking about them right now. Same as mass, gravity, sound waves, radiation, etc… I treat them as a physical phenomenon, which can be reproduced by machines, but only simulated and not reproduced in a Turing machine.
Yes, generally speaking. But certain kinds of simulation necessarily perform that which they simulate. For example, a perfect simulation of a scientific calculator necessarily is a scientific calculator, because in reporting what a scientific calculator would show as the answer, it necessarily itself shows that same answer.
Let’s re-word that last bit:
See above.
Certainly a Turing machine can be anything which is strictly defined by the return of an abstract output after the manipulation/transformation of a symbolic map it receives as input/initial state—so, a Turing machine can be a simulator, a calculator, a chess playing program, etc, etc.
Are mental phenomena are like calculations, or like physical event and processes? That is the question. You don’t seem to have answered it.
But first, you have to demonstrate that this “qualia” thing is physical, which you haven’t done. When you can show qualia are physical in this same way, you might have a case.
But at this point, you’re still arguing the purely-theoretical possibility of a very-privileged hypothesis.
And of those alternatives, Tegmark is by far the simpler hypothesis: it doesn’t require privileging a complex carving out of categories to support it.
Meanwhile, the only thing “no qualia in simulation” offers as a selling point is that it appeals to human intuition… which should make it a suspicious alternative, indeed, where science is concerned.
(After all, the history of science seems to be a never-ending march of finding out ways the world doesn’t really work like our intuitions and biases say they do.)
Actually, it might appeal to the intuition in the abstract as we talk about it now, but if someone were really watching a simulation of a person—and better, interacting with it—then, given a perfect simulation it would very difficult for them to believe anything but that the simulated person was real, conscious, and had genuine emotions.
“no qualia in simulations” requires qualia to be physical. “Tegmark IV” requires mathematical constructs to be physical.
I’m not at all sure which is the simplest hypothesis. Tegmark IV does have the benefit of explaining a larger chunk of reality though.
The problem with this argument is that an external observer cannot tell apart the output of a simulated mind from the output of a real actual brain, other than by looking at its substrate. Thus, the simulated mind is sort of like a zombie: it will talk about its own subjective experience, but the only real basis to that claim will be a portion of its algorithm.
I think it can still be argued that subjective experience is physically “real”, on Occam’s razor grounds: but it’s not an easy claim to make.
No,. there is extra evidence that other people have qualia: you know they are made out of stuff that can generate qualia, because it is the stuff you are made of.
No, that isn’t the question: it is naturalistically possible for qualia to only exist “on the metal”. And the computability of physics is a much more complex subject than you are assuming.
“A decisive refutation of any claim that our reality is computer-simulated would be the discovery of some uncomputable physics, because if reality is doing something that no computer can do, it cannot be a computer simulation. (Computability generally means computability by a Turing machine. Hypercomputation (super-Turing computation) introduces other possibilities which will be dealt with separately). In fact, known physics is held to be (Turing) computable,[11] but the statement “physics is computable” needs to be qualified in various ways. Before symbolic computation, a number, thinking particularly of a real number, one with an infinite number of digits, was said to be computable if a Turing machine will continue to spit out digits endlessly, never reaching a “final digit”.[12] This runs counter, however, to the idea of simulating physics in real time (or any plausible kind of time). Known physical laws (including those of quantum mechanics) are very much infused with real numbers and continua, and the universe seems to be able to decide their values on a moment-by-moment basis. As Richard Feynman put it:[13]
The objection could be made that the simulation does not have to run in “real time”.[14] It misses an important point, though: the shortfall is not linear; rather it is a matter of performing an infinite number of computational steps in a finite time.[15]”
Upvoted for this in particular. There is one possible interpretation that is something of a status hit, but on LessWrong, this should be considered as a statement of the form “you making this statement is strong evidence you do not understand something. If you believe you do understand that something, now is the time to notice you are confused!”
It would be nice if there were a way to tell somebody that they don’t understand something, without it being a status hit!
Of course, if such a thing were possible, human history and civilization would look a LOT different than they currently do. ;-)
Yeah, precisely. It’s doable if difficult in person, but I haven’t the slightest idea how to phrase it in text.
The category of things whose simulations are not duplications isn’t special or exceptional.It includes most things. Simulated planes don’t fly, simulated gravity doesn’t attract, etc, etc. There is a smaller category of things whose simulations are duplications. It is not an extraordinary claim to say consc. belongs in the first category. You need to examine the intutions that make you think it belongs in the second.
That doesn’t follow, as explained above. That simulated gravity does not atrract, does not imply gravity is non-physical.
There clearly is such a things since I experience qualia every time I suck a lemon or sit on a brass tack. The “untrained mind” should have stuck with “we don’t know one way or the other”.
If the algorithm feels like anything from the inside, there are qualia, because qualia are what something feels like. This is the error Searle is always pointing out:: you can’t say conscious experience is just an illusion, because to be able to have illusions, you must be able to have experiences in the first place...
That has almost nothing to do with qualia.
In some senses of “machine” (eg artificial construct), we are clearly not machines. Absent a definition of “machine”, that comment is almost meaningless.
You mean quarks and electrons? Good luck building an electromagnet out of soap, then. We do have qualia,and they could depend on some vary specific physical and chemical properties, as does being a ferromagnet or a liquid crystal. That being the case, qualiaphilic arguments should not be lumped in with the supernatural, at least not without consideration of the specific argument.
You responded to my post, when I think you meant to respond to pjeby’s.
Did you try section 2.5 of “Good and Real”?
Haven’t read it, evidently it might be an idea ( /me adds it to the extensive queue). The description looks like large chunks of the sequences already written up as a book.
I mean, I can come up with my own idea of what “qualia” means (starting from “a word that means whatever wins my argument that consciousness is irreducible”, mentioning the narrative fallacy and getting more acerb from there), but have trouble coming up with what it could mean as part of such an argument without being really obviously silly …
edit: Found and glanced at section 2.5 in a PDF. Yeah, “a word that means whatever wins my argument that consciousness is irreducible” looks like the actual substance of the term “qualia” as an argument for irreducible consciousness, i.e. none to speak of. “I feel something! That counts as actual magic, doesn’t it?” “Er, no.” (The term “qualia” may have uses in a reductionist’s conception of consciousness—I might have use for it in thinking about aesthetics—but those uses aren’t these ones.)
Neither do I. However, having understood (a certain subset of) the sequences, I am capable of dissolving nonsensical questions about qualia… which is what most discussion of qualia consists of. (I.e., nonsense questions and confusion.)
The uses of words, the mind projection fallacy, reductionism, and part of QM are probably the sequences with the most important tools for dissolving that sort of confusion.
My conclusion in the Mary’s room thought experiment doesn’t challenge either of these versions: something new happens when she steps outside, and there’s a perfectly good purely physical explanation of what and why. It is nothing more than an artifact of how human brains are built that Mary is unable to make the same physical thing happen, with the same result, without the assistance of either red light or appropriate surgical tools. A slightly more evolved Mary with a few extra neurons leading into her hippocampus would have no such difficulty.
Incidentally, while agreeing with your main point, I feel I ought to challenge the implications of “more evolved.” This has nothing to do with Mary’s position on some scale of evolution; she could be “less evolved” and have those neurons, or “more evolved” and lack them.
I should have predicted that somebody here was going to call me on that. I accept the correction.
Mary still doesn’t have to make anything special happen to her brain have knowledge of anything else. She can still understand photosynthesis without photosynthesising.
She can understand the sequence of chemical reactions that comprises the Calvin cycle just as she can understand what neural impulses occur when red light strikes retinal rods, but she can’t form the memory of either one occurring within her body.
Which, yet again, only matters if there is something special about qualia that requires memory or instantiation in the body to be understood. She can understand the Calvin Cycle full stop.
You’re converting “physicalism” from a metaphysical thesis to an epistemological one, or at least adding an epistemological one. That’s not the usual usage of the term.
Since qualia are widely supposed to impact physicalism, and since they don’t impact ontological theses such as “everthing is material”, then it is likely that people who suppose that way have the descriptive/explanatory/epistemological version in mind, however implictly.
I don’t understand how Mary’s room is supposed to be epistemologically relevant. Supposing that physicalism is true (and that physics is computable, for simplicity) Mary can run a simulation of herself seeing red and know everything that there is to know about her reaction to seeing red, including a comprehensive description of its phenomenology. Yet, she will still lack the subjective experience of seeing red. But this lack has nothing to do with epistemology in the first place.
It does have something to do with epistemology, because the experience delivers knowledge-by-acquaintance, which is a form of knowledge.
Yes, clearly an experience can deliver knowledge. But does experience yield any additional knowledge over a simulation of same? One could plausibly argue that it does not.
I think this does get at one of the key issues (and one of the places where Hume was probably wrong, and Dennett constitutes genuine progress). On my theory, qualia are not simple. If qualia are by definition simple (perhaps for your reason that they seem that way, and by definition are how they seem), then I am a qualia skeptic. Simple qualia can’t exist. But there is independent reason for being skeptical of the idea that phenomenal conscious experiences are as simple as they appear to be. Indeed, Hume gave an example of how problematic it is to trust our intuitions about the simplicity of qualia in his discussion of missing blue, though of course he didn’t recognize what the problem really was, and so was unable to solve it.
Given that qualia ere what they appear to be., are you denying that qualia can appear simple, or that they are just appearances?
I suppose I am denying that they are just appearances.
Which is a strange thing to say, since qualia re widely defined as appearances.
It certainly becomes stranger when you drop a word. But either way, strangeness is rarely evidence of very much.
Having to use a strange definition of qualia to explain your views may be evidence that you actually a qualia sceptic, a possibility which you seem open to.
As I have already argued, it is not the case that everything is functional or has a functional analysis off the bat: that cannot be assumed apriori. I cannot see the functiona analysis of a blob of chewing gum or a magnetic field. Funcitonal things need well defined inputs, well defined outputs, and a well-defined separation between them and its inner workings,
Since funtionalism is not a universal apriori truth, I see no reason to “codemn to the flames” any non-functional notion of qualia.
I think we know what qualia are because we have them, But that is knowledge-by-acquaintance. It is again question-begging to say that the very idea of qualia has to be rejected unless they can be described. The indescribability of qualia is the essence of the Hard Problem. But we cannot say that we know apriori that only describable things exist.
Unpack this. You know what your qualia are because you have them. I know what my qualia are because I have them. We come to use the same word for these impressions … why, exactly?
What was it Wittgenstein said about remaining silent?
We also both call our kidneys kidneys. I don’t see the big deal.
I didn’t realise Witt was 100% correct about everything.
Only because we are able to describe our kidneys.
I can describe qualia in general as the way thing seem to us. I can’t describe them much more specifically than that.
I don’t believe so. I’ll accept that you can describe them as the way things seem to you. Or define them as the way things seem to us. What I am saying is that you cannot convince me that the definition has a definiendum unless you get more specific. Certainly, your intuitions on the significance of that ‘seeming’ have no argumentative force on anyone else until you offer some explanation why they should know what you are talking about.
Maybe this analogy is helpful: saying “qualia” isn’t giving us insight into consciousness any more than saying “phlogiston” is giving us insight into combustion. However, that doesn’t mean that qualia don’t exist or that any reference to them is nonsensical. Phlogiston exists. However, in our better state of knowledge, we’ve discarded the term and now we call it “hydrocarbons”.
Not really helpful (though I don’t see why it deserved a downvote). It is not that I object to the term ‘qualia’ because I think it is a residue of discredited worldviews. I object to the term because I have never seen a clear enough exposition of the term so that I could understand/appreciate the concept pulling any weight in an argument.
And, as I stated earlier, I particularly object when philosophers offer color qualia as paradigmatic examples of atomic, primitive qualia. Haven’t philosophers ever read a science book? Color vision has been well understood for some time. Cones and rods, rods of three kinds, and all that. So color sensation is not primitive.
And moving up a level from neurons to mind, I cannot imagine how anyone might suggest that there is a higher-level “experience” of the color green which is so similar to an experience of smell-of-mothballs or an experience of A-major-chord so that all three are instances of the same thing—qualia.
Yes, I agree that this kind of atomism is silly, and by implication that things like Drescher’s gensym analogy are even sillier. Nonetheless, the black box needs a label if we want to do something besides point at it and grunt.
I’m saying that there wasn’t a box until someone felt the need to label something. The various phenomena which are being grouped together as qualia are not (or rather are not automatically) a natural kind.
Nothing much hinges on the claim that colour qualia are not really primitive. If we could use their non-primitivity to communicate them, you would be on to something, but the scientific understanding you mention isn’t subjectively accessible.
You seem happy with the idea that all those disparate experiences are experiences. Why not be happy with the idea that they are all qualia?
We do use their non-primitivity to communicate them. I learned the names of the colors in kindergarten as a result of a demonstration involving three kinds of rods in my eyes and photons of various wavelengths. Of course, I didn’t realize all that at the time, but the learning would have been less successful if one of my three types of photoreceptors had not been generally functional.
I don’t object to the word—I object to the exclusionary way it is used—the discriminatory weight it is called upon to bear. “Yeah, buddy, you and me, we have qualia. Unlike those nasty robots over there.” How did my humanity become so inextricably tied to my possession of certain kinds of sense organs? Or rather, my use of those organs.
You were not using their non- primitivity to communicate them: you were not offering or being offered a descrption that sed a decomposition of a quale into primitive terms. You were using a knowledge-by-acquaintance mechanism that happened to be complex, although you did not know it was complex, or needed to know it was complex. In fact it could have been primitive for all you knew.
The issue about discrimination you bring up is not, I think, central. Moral relevance is tied to the ability to suffer and enjoy. These are qualia. No one feels guilty about kicking rocks because we believe rocks don’t have pain qualia. But even if we did believe rocks have pain qualia, the epistemological and metaphysical issues remain.
Does your objection to “discrimination” extend to treating rocks as sensitive beings?
I certainly would treat a rock as sensitive if I had reason to believe that it would be willing to treat me as sensitive. (Maybe the only thing Kant got right!). Certainly my decision regarding how to treat rocks would have absolutely nothing to do with my guesses as to whether the way they experience the world was ontologically similar to the way I experience the world.
You’re sensitive. If they were, that would be a broad similarity
Even if their sensitivity were perfectly well understood in terms of geochemical cause and effect? Understood well enough to simulate? A simulation that could be connected to actuators that would act in my interests (assuming I reciprocated)? Great. Then we are in agreement. There is nothing mysterious or unsimulatable about qualia.
The word “qualia” doesn’t have to justify its existence by providing a solution. It can justify its use by outlining a problem.
Sit on a brass tack. If you feel nothing, I will accept that the deifintion has no definiendum for you, even though it does for me and everyone else.
That reply would be cogent if I claimed not to feel pain. It not useful in this context, though, since I claim not to understand exactly what you do and don’t mean by “qualia”.
It does serve as a single example though. Provide a few dozen more, plus another few dozen examples of mental events that are not qualia, and a brief explanation of what it is that separates the negative and positive examples—do that and you will have communicated a concept. That is, you will have done so if no one disagrees with your lists and explanations and has a different understanding of the word ‘qualia’.
Sometimes I wonder whether this is the reason that philosophers never produce such lists. When you read a text on thermodynamics, say, you will spend many pages going over the meanings of important technical terms like system, open, and closed. But, it is worth it because when you are done, everyone is on the same page. :)
An example would be the way lemons taste to you, as opposed to their chemical composition. Other examples: the way a sunset looks, the way sandpaper feels, the smell of coffee, a stomach-ache a sharp pain, such as sitting on a thumbtack. There are therefore qualia corresponding to all the traditional sensory modalities, although nothing need be sensed to have a quale — they occur in dreams, for instance. Some people include the “phenomenal feels” of emotional states along with qualia, although these are something of an edge case. Phenomenal states or qualia form one of three large and overlapping categories of mental states, the others being cognitive/intentional states and states of volition, will and decision-making.
I decided to look back to see how I got into this fruitless conversation. I think it started with this question from dfranke:
I responded flippantly that I was unable to make any sense of the concept of experiencing red. I didn’t make clear that what I was really objecting to was the reification of “red”. I certainly can imagine any person with the same kindergarten education as myself, recognizing a red apple, or a red light, or a red object of indeterminate shape. I can imagine someone classifying the sensation of a tack in the bottom as painful, and someone identifying the taste of lemon juice as sour.
I also have no trouble at all imagining that a robot trained by a naive Bayes classifier (rather than by a kindergarden teacher named Mrs. Weiskopf), could easily recognize those red things as ‘red’, the piercing thing as ‘painful’, and the lemon thing as ‘sour’. Yet people keep suggesting that robots can’t have qualia.
So, I respectfully suggest that your examples are not successfully communicating to me what it is about qualia that you know you have, you suspect that I have, and you (or maybe it is just dfranke) claim a robot cannot have.
Agreed. As you illustrate here, a lot of talk about “qualia” and “subjective experiences” can easily be interpreted as everyday talk about perception and discrimination between different real things out there in the physical world (e.g. there really is something about the chemistry which we are discriminating when we notice how the lemon tastes, which a robot could, if he has the right sensors, just as easily discriminate), and furthermore I think that this everyday interpretation is the main part of what makes such talk seem obviously true. That is, the philosophical talk about qualia is powered by ambiguity about what is being said, by equivocation between the everyday meaning and the “philosophical” meaning.
Equivocation—saying one thing which is accepted as true, then silently shifting its meaning in order to draw a false conclusion—is a huge problem in philosophical discussion, and it’s hard to deal with precisely because the meaning shifts are not easy to notice, since the words remain the same.
Most qualiaohiic philosophers are explicit that qualia are not just discriminative behaviours or abilities.