Slightly known unknowns

post by NancyLebovitz · 2011-11-10T10:29:37.863Z · LW · GW · Legacy · 25 comments

Contents

25 comments

Inspired by What visionary project would you fund?, I'm wondering about whether there are known blank areas in our knowledge which might turn up surprising knowledge.

Once upon a time, the sun was a mystery. People had a pretty good idea of its mass, and how much chemical energy would be needed to keep it shining. I don't remember how long the sun could last running on chemical energy, but it didn't seem plausible that it could be so new.

It turned out that chemical energy wasn't the only possibility.

I don't get the impression that there was wide appreciation that most of the world was being ignored by scientists because the calculations were too difficult-- until the calculations were made easier.

Do people here expect to be surprised by whatever it takes to understand qualia?

25 comments

Comments sorted by top scores.

comment by Richard_Kennaway · 2011-11-10T13:49:43.573Z · LW(p) · GW(p)

Do people here expect to be surprised by whatever it takes to understand qualia?

Yes.

Note that "It's neurons!" is not an explanation, any more than "It's atoms!" is an explanation of turbulence, or an empty outline of Africa is a map of its interior.

Replies from: Logos01, lessdazed
comment by Logos01 · 2011-11-14T11:23:16.977Z · LW(p) · GW(p)

Note that "It's neurons!" is not an explanation, any more than "It's atoms!" is an explanation of turbulence, or an empty outline of Africa is a map of its interior.

But it is true that you cannot have physical turbulence without atoms -- not that anyone would discuss turbulence in terms of atoms -- no more than one would discuss genetics in terms of subatomic reactions. It's nonsensical to do so; the variances of scale are such that entire categories of manifest phenomena in one scale are entirely transparent to the lower.

Qualia -- consciousness in general -- seems most likely to me to be something akin to the quality of "wetness". There is no specific atom or collection of hydrogen & oxygen atoms which are "wet". Not even an individual H20 molecule at room temperature is "wet" -- nor even, say, a dozen of them. But if you get a sufficient quantity of them together, the relationship between them manifests the quality of "wetness".

Given our ability to decode and encode visual and motor cortex activity, and the vast swathes of observations of how aberrant neuroanatomy, it is fairly clear that consciousness is dependent upon the brain in the same way that rats are "dependent" upon atoms.

That doesn't make discussing consciousness in terms of specific neurons any more intelligible than discussing rats in terms of their atoms. It just means we need to understand the intermediate 'scales of behavior' between the atom and the rat, with the opposing "ends" of the 'spectrum' being the individual neuron and the fully 'emergent' consciousness.

I do strongly believe that Minsky is onto something with the Society of Mind hypothesis; I suspect that many of the individual substructures of consciousness may very well be able to be broken down and isolated for investigation. As to where individual's "qualia" come from in all this -- well, subjective experience, I again suspect -- but of course cannot prove -- may very well be a result of individual agents within the 'society' whose role is to arbitrate/play intermediary between the various secondary agents; something akin to low-level metacognition -- observations recorded in real-time and recalled nearly instantaneously. We "experience" events and those "experiences" are real phenomena, but they of course are not recorded in any specific individual neuron in the exact same way that the structure of a rat's motion through a maze is not "recorded" in any specific atom within its body.

Replies from: Richard_Kennaway
comment by Richard_Kennaway · 2011-11-14T13:11:02.177Z · LW(p) · GW(p)

Qualia -- consciousness in general -- seems most likely to me to be something akin to the quality of "wetness". There is no specific atom or collection of hydrogen & oxygen atoms which are "wet". Not even an individual H20 molecule at room temperature is "wet" -- nor even, say, a dozen of them. But if you get a sufficient quantity of them together, the relationship between them manifests the quality of "wetness".

This is exactly the sort of non-explanation I was warning against. None of that goes any way towards explaining wetness. In fact, wetness is explained, nowadays, in terms of atoms. We know how atoms combine into molecules; how forces exist between molecules which, under certain temperatures and pressures, result in the liquid state; how interatomic forces between a liquid and a solid in contact with it result in the liquid spreading over the solid surface or forming a bead on top of it. It takes a lot of atoms together to show these phenomena, but "emergent" (a word you used later) wetness is not something separate from the story I outlined, and without that story you do not have an explanation.

But it is true that you cannot have physical turbulence without atoms -- not that anyone would discuss turbulence in terms of atoms -- no more than one would discuss genetics in terms of subatomic reactions. It's nonsensical to do so; the variances of scale are such that entire categories of manifest phenomena in one scale are entirely transparent to the lower.

Turbulence is less well understood. As far as I know, it isn't even known if the Navier-Stokes equation has turbulent solutions. Either way, a real explanation of turbulence, if we ever find it, will provide a detailed connection between the macroscopic phenomenon and the atomic level.

That doesn't make discussing consciousness in terms of specific neurons any more intelligible than discussing rats in terms of their atoms. It just means we need to understand the intermediate 'scales of behavior' between the atom and the rat, with the opposing "ends" of the 'spectrum' being the individual neuron and the fully 'emergent' consciousness.

A real explanation of consciousness as a physical process of the brain must go all the way down to atoms, as surely as an explanation of how rats work must go all the way down to atoms -- and does, for all the explanations that we have of how rats work.

Replies from: Logos01
comment by Logos01 · 2011-11-14T13:54:21.157Z · LW(p) · GW(p)

This is exactly the sort of non-explanation I was warning against. None of that goes any way towards explaining wetness. In fact, wetness is explained, nowadays, in terms of atoms.

The fact that you don't like the explanation doesn't make it a "non-explanation".

In fact, wetness is explained, nowadays, in terms of atoms.

Reductively, yes. But I defy you to find me an explanation of wetness written in terms of atoms. Which atom in H20 contains the "wetness" property, exactly? That's exactly my point: "wetness" itself is not a property of atoms. Nor even of individual molecules, nor even of small quantities of molecules -- but rather of the interaction of a large quantity of molecules.

We know how atoms combine into molecules; how forces exist between molecules which, under certain temperatures and pressures, result in the liquid state;

That is more a "description" than an "explanation".

Turbulence is less well understood.

Indeed. But what explanations we have for it do not depend upon the behaviors of individual quarks, and we never discuss it in those terms. That's exactly the point: even attempting to discuss turbulence in terms of quarks is failing to properly address the right scale.

So, too, with consciousness and neurons -- so goes the assertion which you reject out of hand as a "non-explanation". It states that there are intermediary levels of explanation which must be engaged between consciousness and neurons. Minsky's "Society of Mind" qualifies as this; neurons comprise individual components of select "agents", and those agents together manifest "consciousness" in the same way that atoms comprise molecules which in turn manifest "wetness" when present in the correct quantity.

A real explanation of consciousness as a physical process of the brain must go all the way down to atoms

Ahem. This is absurd. Nowhere in scientific endeavors is this done. Emergence (non-epiphenomenalistic emergence that is) is not a rejection of reductivism, but an assertion of it.

as surely as an explanation of how rats work must go all the way down to atoms

Find me an explanation of rats written in terms of atoms -- and I'll accept what you say. You must, for your rejection of emergent-phenomena-as-consciousness-explanation to be accepted as valid by me -- should this be your goal -- find an explanation of rats that has no shift in scale other than atoms. It must explain everything about rats in terms of their atoms.

Replies from: Richard_Kennaway
comment by Richard_Kennaway · 2011-11-15T17:01:43.975Z · LW(p) · GW(p)

The fact that you don't like the explanation doesn't make it a "non-explanation".

The causality is in the opposite direction.

neurons comprise individual components of select "agents", and those agents together manifest "consciousness" in the same way that atoms comprise molecules which in turn manifest "wetness" when present in the correct quantity.

"those agents together manifest "consciousness"" is not an explanation: it is a hole where an explanation should go. There is nothing in common with the way that atoms together produce wetness beyond that general outline. It is like the "map" of Africa consisting of the outline only. With wetness, we do know (pretty much) the actual explanation that fills that outline. With consciousness we do not.

Find me an explanation of rats written in terms of atoms -- and I'll accept what you say.

Pick up any general textbook on physiology.

Replies from: Logos01
comment by Logos01 · 2011-11-15T19:55:42.541Z · LW(p) · GW(p)

The causality is in the opposite direction.

Explain.

"those agents together manifest "consciousness"" is not an explanation: it is a hole where an explanation should go.

No, it's an assertion that you can't use the wrong scale to discuss a given type of phenomenon.

There are not atoms that carry the property of "wetness". They don't exist. Pointing this out doesn't mean that I'm not a Physicalist; nor does it mean that I have rejected reductivism. It's frankly juvenile for it to be insisted that I do.

There is nothing in common with the way that atoms together produce wetness beyond that general outline.

And yet there is no such thing as the quality of "wetness" in an atom.

It is like the "map" of Africa consisting of the outline only. With wetness, we do know (pretty much) the actual explanation that fills that outline. With consciousness we do not.

... I figured you'd say this, actually. It's rather discouraging of any hope of getting heard by you.

The concept of consciousness as a recursive record of disparate agents is in no way shape or form "not an explanation", nor is it "a hole where an explanation should go". It is a solid, realistic, testable prediction of how consciousness operates. There are real things that must occur for it to be valid.

Tell me -- are you familiar at all with superatoms? Do you understand that when I say to you that the elemental mimickry superatoms produce is an emergent phenomenona resultant from the interaction of those atoms -- that is not an "outline with no details" but a precise statement?

Pick up any general textbook on physiology.

Done. Read 'em cover to cover. Several. Never found a single one that ever discussed physiology -- of any organism, let alone a mammal, in terms of atoms. As I said: find me an explanation of rats written in terms of atoms. Being as generous as I possibly can, 'pick up any general textbook on physiology' is just wrong. The closest I can conceive of to what you're saying is actually one that deeply contradicts your assertions: "it's all genes!" is precisely the same as "it's all neurons!".


At this point, this conversation has lost all interest to me. It's pretty clear that you are very skilled at rejecting this position regardless of how valid or useful the position might be. So let me instead ask you this:

What path would you recommend we follow towards gaining comprehension of consciousness? Note: I will not accept any answers that come in the form of a statement of 'hard problem of consciousness'. I want real, practical approaches. Steps researchers could take to make the problem "simpler", so we can worry at it like a dog with a bone.

I feel justified in asking you to provide this because that is exactly what you are rejecting when you reject my position as a "hole where an explanation ought to go". For example; consciousness as an emergent property resultant from recursive realtime interaction between various agents predicts amongst other things "bleedover" in sensory faculty for those with failing senses; it predicts the existence of the utility of mnemonic devices such as the Roman Room, it predicts the ability of small portions of the brain to have major/significant roles in consciousness -- as the hippocampus ; it also predicts that there should be no "central organ of consciousness" -- which again has been experimentally corroborated; or how it could even be possible that the removal of specific "sub-agents" might improve cognition. Such as induced "autism-like symptoms" through TMS shows to be the case -- or autistic savants in general. There are even edge-cases where we can see chances that it might be false, such as the case of the man with "no" brain -- if there weren't such a phenomenon as neuroplasticity, and if that process hadn't developed so slowly, we would have expected very different results if the Socially-emergent Consciousness Hypothesis weren't correct. We would also be left with absolutely no means of explaining how it could come to be that split brains should be able to produce two independent personalities which are not mirrors of one another.

It is a model which puts these pieces into place and creates of them a roadmap towards deeper understanding. It also makes predictions about whether or not "qualia" are explicable. To me, and anyone who accepts this position, there is nothing "mystical" or "inscrutable" about qualia. We can expect entirely to eventually uncover the existence of portions of the brain where specific categories of qualia occur when they are directly sensed. We can map where, when, and how -- and we further assert what from the body of evidence governing how qualia can be prevented from occurring (as with the split brain patients). In fact, because of the 'social'-emergent consciousness hypothesis, I am able to predict that split-brain patients are related to the 'problem' of the nature of qualia.

So what is this solution you offer? What is it about the 'social'-emergent consciousness hypothesis's deep explanatory power, exactly, that causes you to reject it as a "non-explanation"? What explanation can you provide that so deeply exceeds this that you are justified in your assertion?

Replies from: lessdazed
comment by lessdazed · 2011-11-16T00:23:46.460Z · LW(p) · GW(p)

... I figured you'd say this, actually. It's rather discouraging of any hope of getting heard by you.

That's downvote worthy.

Replies from: Logos01
comment by Logos01 · 2011-11-16T00:59:57.427Z · LW(p) · GW(p)

Why? It's a factual statement.

comment by lessdazed · 2011-11-14T14:59:58.626Z · LW(p) · GW(p)

Note that "It's neurons!" is not an explanation

That's been explicitly noted since at least 2008 on OB:

the promise, though not yet the realization, of a full scientific account of thought. The problem may now be declared, if not solved, then solvable.

(Note: quote is from a fictional press release).

comment by Richard_Kennaway · 2011-11-10T14:38:01.879Z · LW(p) · GW(p)

Excuse me while I geek off on a tangent about Bayesian inference and Kolmogorov complexity.

Can one ever expect to be surprised?

If I generate a 100-digit random number, whatever number I get will have had a prior probability of 10^-100, but that does not mean I will be surprised by it. If it turned up all zeros I would be surprised, but that is most unlikely to happen. In this case, my surprise on seeing the generated number might theoretically be measured as the amount by which it can be compressed, the expectation value of which is surely zero.

But what is the expectation of my surprise at the answer to a question of which I cannot think of anything that could even be a possible explanation (including explanations that deny the question)? I don't know how to measure that within the Bayes/Kolmogorov system of ideas.

Replies from: printing-spoon, magfrump, roystgnr
comment by printing-spoon · 2011-11-11T02:01:43.413Z · LW(p) · GW(p)

You implicitly have hypotheses "this random number generator is broken always returns 0" and "this random number generator works fine." You start off being pretty sure the latter is true. Your shift to the former upon seeing 0 is where the surprise comes from.

Replies from: JoshuaZ
comment by JoshuaZ · 2011-11-11T02:06:35.543Z · LW(p) · GW(p)

Right, and one thing to note is that given basic programming assumptions, a return of all zeros is a much more likely failure mode of a random number generator than say always returning 72 or something similar.

comment by magfrump · 2011-11-11T05:55:15.974Z · LW(p) · GW(p)

This seems like the same conversation being had here.

comment by roystgnr · 2011-11-10T18:30:05.715Z · LW(p) · GW(p)

You can expect surprise, because "surprise" isn't a linear function of the data.

Simplest example: assign a prior probability distribution to some uncertain real variable x. Your expected value E[x] is the average of this distribution. Your expected error E[E[x] - x] is indeed zero! But unless you already know the exact true value, your expected error magnitude E[abs(E[x] - x)] is always positive.

comment by Eugine_Nier · 2011-11-11T02:19:48.970Z · LW(p) · GW(p)

Do people here expect to be surprised by whatever it takes to understand qualia?

Going by the outside view, I expect it will in involve a lot of unknown unknowns.

I won't be surprised by this when it happens, but expect a lot of people in the field to be.

comment by wedrifid · 2011-11-10T12:19:08.973Z · LW(p) · GW(p)

Do people here expect to be surprised by whatever it takes to understand qualia?

No. Not even remotely. This struck me as a massive leap away from the build up with the sun, etc. I expect the most difficult part of 'explaining qualia' is going to be persuading philosophers that the neurobiology really is all that is needed.

Replies from: NancyLebovitz, ArisKatsaris
comment by NancyLebovitz · 2011-11-10T15:00:15.182Z · LW(p) · GW(p)

It seemed to me as though the article trailed off rather than building to a conclusion, but I couldn't think of anything else to say on the subject.

comment by ArisKatsaris · 2011-11-10T12:52:08.279Z · LW(p) · GW(p)

I don't think it's just philosophers who have trouble understanding how existence can arise from non-existence. And I consider that to be one and the same with the qualia problem -- or atleast highly similar to it.

Replies from: wedrifid
comment by wedrifid · 2011-11-10T13:02:25.674Z · LW(p) · GW(p)

:s/philophers/philosophers and amateurs who happen to be engaging in philosophy about qualia/

comment by Douglas_Knight · 2011-11-10T15:48:30.475Z · LW(p) · GW(p)

I don't get the impression that there was wide appreciation that most of the world was being ignored by scientists because the calculations were too difficult-- until the calculations were made easier.

What's this about?

Replies from: NancyLebovitz
comment by NancyLebovitz · 2011-11-10T15:53:05.428Z · LW(p) · GW(p)

Chaos theory.

comment by lessdazed · 2011-11-10T13:40:13.056Z · LW(p) · GW(p)

Do people here expect to be surprised by whatever it takes to understand qualia?

I expect to be unsurprised by whatever is sufficient to get people to talk about qualia. I know the talk is a real thing needing explanation.

comment by Cthulhoo · 2011-11-10T16:07:39.234Z · LW(p) · GW(p)

Do people here expect to be surprised by whatever it takes to understand qualia?

On one hand the answer is no: I don't expect a deviation from reductionism or even the birth of a completely new branch of science (like nuclear physics in the sun example).

On the other hand I expect that, on the path of properly understanding the arising of consciousness we will learn something surprising about complex systems.

comment by Logos01 · 2011-11-10T11:52:24.992Z · LW(p) · GW(p)

Do people here expect to be surprised by whatever it takes to understand qualia?

(In my case) only if it turns out they're not already understood.

comment by Manfred · 2011-11-10T11:47:38.089Z · LW(p) · GW(p)

Nope. Um, since "understanding" qualia in some sort of final sense means getting an evaluable definition out of human brains (a slightly impossible task, but good approximations are possible), I might have the opportunity to be surprised at some of the weird implicit definitions people have :P