A question of philosophy?

demagogue on 5/2/2011 at 03:06

Quote Posted by Subjective Effect

I totally reject functionalism because I think it's part of the organic nature of the brain that lends that particular substrate it's particular properties.

This is a respectable position. This is John Searle's position in the book The Rediscovery of the Mind, though I wish he went into more detail what was actually special about the carbon-based organic part. I never was clear about that part.

I'm sympathetic with its vibe of respect for nature and humility, in contrast to some functionalist arguments with a bit of hubris in thinking it must be so easy to make a brain and consciousness if only we wired it right. (Though you could argue it's another kind of hubris to think human brains are so special they just must have a special domain of physics just for them beyond our engineering capacity.)

Anyway, three things come to my mind... (I'll try to keep them short!)
(1) Re: "Unless they are made of the same material I don't think they can function the same." To be clear, are you thinking about this as a fundamental law or just as practical engineering?

Are you saying functionalism is technically right. It's the functional organization doing the actual work (so any substrate organized the same would theoretically hold consciousness) but it'd be practically impossible to engineer it carefully enough outside of actual brains. Or that functionalism is fundamentally wrong? It's not the functional organization at all, but some irreducible physical property of the stuff after all that's doing all the work, as a fundamental law (possibly with new physical properties we have yet to discover)? I don't know if we even know enough yet to answer that, but do you have a gut feeling about it?

(2) Now they've got models of non-carbon-based organic chemistry, and even real-world examples like around ocean vents. If you want to say only earth-type carbon-chemistry brains can have consciousness, (or maybe including a few other "natural" organic chemistries?), then first, doesn't it seems like a huge coincidence that the chemical composition of earth and then X billion years of evolution just happened to stumble in *exactly* the right combination to instantiate consciousness, but other types of chemistry in other evolutionary-trees, or human-made engineering, can't?

(3) And then, getting down to brass tacks -- a normal neuron signal is made up of three parts: (1) In-coming neutrotransmitters trigger a build-up of an electric potential in the nucleus, (2) When that reaches a threshold, gates open and choride ions stream inside, changing the local electric potential which open gates farther down the chain, cascading all the way to the ends. (3) Then at the axiom ends, it releases pods of neurochemicals into the synapse, starting the cycle over.

It's all efficient at what it does, but as a matter of physics there's nothing particularly special about it that Maxwell couldn't have described 150 years ago. It doesn't seem like there's anything more special going on that electrons streaming across gold atoms couldn't do just as well, or a billion other ways to carry a signal from eyes, to decisions, to muscles. This is similar to Phatose's point too.

There are people that like to say there's other kinds of signalling going on, like quantum effects or some mysterious other physical force involved in the signaling, but I don't know why you'd need all that extra physics when plain old vanilla electromagnetism seems to do all the work just fine. (Not saying we might not be surprised, but they've been studying the physical mechanism of neurons for a long time now and AFAIK haven't really needed extra physics yet to get the job done.)

Good fodder for debate, thanks.

Phatose on 5/2/2011 at 05:20

Did Searle ever refine his position since "Minds, Brains, Programs"? I thought it was terribly unconvincing there, and the causal powers he was constantly going on about practically translated to "A wizard did it."

demagogue on 5/2/2011 at 06:40

I guess not since I had about the same reaction as you. Searle could have decent critical arguments against functional or AI approaches to the mind. The (http://en.wikipedia.org/wiki/Chinese_room) Chinese Room argument is still a very vivid and influential argument (if mostly in how well it's generated counter-arguments. And like all good "love-to-hate" arguments, even if masses of people want to agree it's "wrong", there is still great disagreement in why they think it's wrong, and that in itself is very interesting).

But when he got into positive arguments about what's really special about grey matter, in causal terms... I didn't buy it either. Why should plain old bio-chemistry be special here? We're still just talking about things like electron sharing and charge & potentials. It struck me as an argument coming entirely from being uncomfortable with the claims AI people were making, but not too concerned with how neurons & their nets actually work.

PS, For the record, there are parts about classic AI approaches to minds I don't like either, even in similar ways as Searle. I agree with him that classic AI is completely botching what philosophers call the "intentionality" of sentences (the part about a sentence that makes it actually about what it's talking about, as opposed to just scribbles). I just don't think it's the physical wetware of neural nets that's adding the missing pieces. I'll have to save that argument for another post though (short answer: I was impressed by the model of Radu Bogdan in his book Predicative Minds).

Renzatic on 5/2/2011 at 07:44

Quote Posted by Subjective Effect

And reductionism scares the crap out of me because, well wtf?

See, I don't believe reductionism works with the brain. If you break it down to its base mechanical actions, then all you have are unaware but somehow intelligent meat puppets, or Demagogue's aforementioned philosophical zombies. From my personal subjective experiences with reality here, it seems to me that isn't the case. I can look on my desk and see my can of soft drink. I can identify it as a can of soft drink. I can look at and break it down to it's individual colors and shapes. I can imagine a myriad uses for this can of soft drink, from the logical, to the creative, to the what the hell am I thinking. It's not just an object, but a concept, and a concept upon a concept. I can play with it in my mind. And whatever parts of my brain I'm flexing coming up with these uses, these concepts, there seems to be a me in there doing the flexing. Or at least there is from my perspective.

I'm probably exposing my ignorance here, but I believe the behaviorists believe there isn't such a thing as you or I. That all thought processes are completely separate things, that only appear to combine as a whole because they have to coordinate themselves to function. To break that down, citing poetry is a meaningless thing, because there is no conscious drive behind it. It's a function of one particular part of the brain of one particular entity. Much like passing gas, it makes a noise, and is about as ultimately empty of deeper meaning. And those that appreciate it? Same thing. A bunch of other unconscious entities babbling about in response, in much the same way that a kicked can tumbles in response to your foot. All of civilization, everything we know, is nothing more than an evergrowing chain of complex responses driven by empty headed bags of bipedal beef. If an actual conscious third party were to observe us, it'd look like we all know what we're talking about...but we don't. There's nothing behind this facade of humanity except a blank void.

So why am I here pondering this? Why do we have a definition of reason, if reason itself doesn't exist? If I'm an illusion, I must be an illusion so complete as to be indistinguishable from the real thing...whatever that is.

Why do I hate mayonnaise so much?

Someone educate me here.

Renzatic on 5/2/2011 at 09:55

Quote Posted by Wormrat

Because it's useful. We have definitions for all sorts of things that don't exist, strictly speaking. For example, we can define geometric shapes in terms of points and lines, even though points and lines aren't real things that can be found in the universe (nor is geometry, for that matter). They're concepts build on concepts, symbol-games built on symbol-games that, when scrutinized, boil down to a kind of "useful nonsense." Likewise, it doesn't really matter that no one has a rigorous definition of what "pain" means; we've managed to make the term work for us anyway. When we toss it around in conversation, it seems to trigger similar patterns of thoughts in people, enough so that we can communicate and get something done.

I could've worded that a little better. What I should say is that if we're incapable of reason per the behaviorist definition of the mind, then why do we have a concept of it?

Quote:

The problem, as I see it, comes when people start treating these concepts as real things and then go looking for them in our physical models of the world. It seems to me a pointless exercise from the outset. You're not going to be able to discover "consciousness" or "qualia" or "intentionality" amidst particles and energy any more than you are going to find truth, beauty, or an embodiment of the number 4.

Yup. I see things the same way. Consciousness, what we experience within our own heads, can't be quantified. We can talk about them. But we can't offer up any concrete proof that we all actually feel and think the way we do.

I'll try to explain something I have no words for to prove my point. When I think back to certain points in my life, good or bad, I associate them with this, I'll call it a sub feeling, not really an emotion, but a kind of mental texture of that that encompasses the emotions I felt during those times. I can think back to a party I went to when I was 18. I had fun, I was happy, fell in love, ect. I can explain the emotions, because we all have words for them. But I can't quite explain that "texture" underneath it that ties the emotions, the location, and the time together into something very specific to me.

Do we all feel this? Am I unique in this regard? I never hear anyone else talk about them, maybe because if others do, it's such a difficult thing to define. But how do I prove to you that I feel these, for lack of a better word, "textures". I can't. It's literally impossible.

Cognitive science is interesting because it's one of the few places where the scientific method fails us. You can study what the brain does when someone gets angry, but you can't quantify why they got angry in the first place. We can talk about the concept of anger. We all feel it, after all. But does anger, despite having similar physical results, feel different for each of us?

I think we all need to get together and write some new age self help book.

Sg3 on 5/2/2011 at 14:46

Quote Posted by Renzatic

I'm probably exposing my ignorance here, but I believe the behaviorists believe there isn't such a thing as you or I. That all thought processes are completely separate things, that only appear to combine as a whole because they have to coordinate themselves to function. To break that down, citing poetry is a meaningless thing, because there is no conscious drive behind it. It's a function of one particular part of the brain of one particular entity. Much like passing gas, it makes a noise, and is about as ultimately empty of deeper meaning. And those that appreciate it? Same thing. A bunch of other unconscious entities babbling about in response, in much the same way that a kicked can tumbles in response to your foot. All of civilization, everything we know, is nothing more than an evergrowing chain of complex responses driven by empty headed bags of bipedal beef. If an actual conscious third party were to observe us, it'd look like we all know what we're talking about...but we don't. There's nothing behind this facade of humanity except a blank void.

So why am I here pondering this? Why do we have a definition of reason, if reason itself doesn't exist? If I'm an illusion, I must be an illusion so complete as to be indistinguishable from the real thing...whatever that is.

If my reality or experiences are an illusion, then what "real" sapient being is experiencing that illusion, if not me? I just can't get my head around the "I am an illusion" thing—I don't mean because of fear, although I do find it horrifying, but because I can't understand it. An illusion requires a real being to experience it, doesn't it?

Quote Posted by Renzatic

You can study what the brain does when someone gets angry, but you can't quantify why they got angry in the first place. We can talk about the concept of anger. We all feel it, after all. But does anger, despite having similar physical results, feel different for each of us?

I think it can feel different. It seems to me that many people like to feel angry; I don't. And so on. In a literature class a few days ago, we read (http://www.jacklondons.net/lawoflife.html) Jack London's "The Law of Life". I thought it was heavily depressing. The literature professor came to the same basic conclusions that I did, but called the "law of life" beautiful and good instead of dreadful. I guess I'm citing sadomasochism.

SubJeff on 5/2/2011 at 23:42

Quote Posted by demagogue

(1) Re: "Unless they are made of the same material I don't think they can function the same." To be clear, are you thinking about this as a fundamental law or just as practical engineering?

A combination of the two actually. It's my belief that the brain only works in just the way it does because of it's organic physical and logical nature. Neural connections and networks are not simply logic gates because there is an inherent amount of instability in them by dint of the fact that they rely on semi-quantitative processes like neurotransmitter release.

I'll give a simple example. At the neuromuscular junction about, not exactly, about 125 vesicles of acetylcholine are released with every impulse. And in each vesicle there is different amount of acetylcholine released. It's a similar amount each time (until the stores are exhausted) but it's never identical amounts. And not all of the transmitter will bind to receptors, and not all receptors will be bound to. Temperature alters the kinetics of the binding, as does acidity, other chemicals and tens if not hundreds of other factors. The reason it doesn't just go haywire is that there are feedback pathways that enhance or inhibit neural activity in such a way as to keep everything "stable".

This variability occurs in every single synapse in the brain. The networks may be governed by logical structure but it is nothing like digital data, despite our ability to describe these processes in a digital manner.

So it's both an engineering problem - this will never be replicated, and a fundamental law - it is a property of organic brains to be able generate this particular set of almost chaotic patterns.

Quote:

(2)If you want to say only earth-type carbon-chemistry brains can have consciousness, (or maybe including a few other "natural" organic chemistries?), then first, doesn't it seems like a huge coincidence that the chemical composition of earth and then X billion years of evolution just happened to stumble in *exactly* the right combination to instantiate consciousness, but other types of chemistry in other evolutionary-trees, or human-made engineering, can't?

No, it's not a coincidence. It was inevitable. This is veering onto a different subject but I believe it's basic probability in action. You think it's a huge coincidence because you're biologically programmed to have that type of mystical thinking. It's a function of the human mind and it may have been seeded there by a Creator/Manipulator or it may just be part of the "will to live" paradigm that I think is a product of selection pressure. (Selection pressure that may be there by coincidence or design, but that's another story).

Quote:

(3) And then, getting down to brass tacks -- a normal neuron signal is made up of three parts:

It's actually made up of a lot more than this as I've alluded to earlier. And just to correct your description (because that's how I roll) 1. neurotransmitters open ion channels which can be for a range of different ion types (although in the brain the main excitatory transmitter opens channels for positively charged ions and vice versa - how coincidental is that?), 2. these ions move down an electrochemical gradient because of an already built up potential and thus alter that potential (depolarising the cell locally) opening voltage gated channels, 3. the nucleus isn't involved in this, 4. the electrical cascade (depolarisation) then travels down the axon.

Quote:

It's all efficient at what it does, but as a matter of physics there's nothing particularly special about it that Maxwell couldn't have described 150 years ago. It doesn't seem like there's anything more special going on that electrons streaming across gold atoms couldn't do just as well, or a billion other ways to carry a signal from eyes, to decisions, to muscles.

... I don't know why you'd need all that extra physics when plain old vanilla electromagnetism seems to do all the work just fine. (Not saying we might not be surprised, but they've been studying the physical mechanism of neurons for a long time now and AFAIK haven't really needed extra physics yet to get the job done.)

With respect I disagree. This is why I reject functionalism - it's not a simple as that and there is something more special going on than electron flow. It's incredibly complex, so much so that I could go on about the factors that alter a single nerve cell activation for 100s if not 1000s of words. Include all the bits I'd miss, all the bits we don't know yet, and multiply it by the number of connections we have in our brains, further complicate it by mixing in the functional organisation of different areas and projections and all the different types of transmitters, receptors and reciprocal, enhancing and antagonistic relationships and you realise that thinking you can replicate that, ever, is Goddamn dreamland.

Going back to the Chinese Room - it's the perfect compliment to my argument. I've no doubt that we can build something that looks like it's conscious but it won't be. Not in the same was as we are. I'm not saying no AI can ever have any kind of consciousness, on the contrary I think it perfectly possible. But it won't be human consciousness.

Quote Posted by Renzatic

This is what I hate about it so much, why it scares me so much. Perhaps I'd be happier about it if I was a pure pragmatist and just tried to accept it. It's an incredibly dehumansing theory though and if it were true it would be a fantastic argument for utter selfishness.

Quote Posted by Wormrat

Likewise, it doesn't really matter that no one has a rigorous definition of what "pain" means; we've managed to make the term work for us anyway.

Definition of pain: perception of noxious stimulus.

Divinum sedare dolorem.

Word.

demagogue on 6/2/2011 at 03:03

@SE, I think, as far as which technical position you're actually arguing for, it's still a functionalist in its technical philosophical sense, and you're actually arguing against specific functionalist theories. (Worth pointing out that most biology is functionalist: digestion, respiration, circulation, all these are functionalist systems; the key isn't what they're made out of but how they work. And IIRC it was actually the biologists pushing functionalism to begin with, because they wanted consciousness studies to fit in with the rest of biology. I think you're the biologist's brand of functionalism, but not the AI brand.)

Physicalism is that consciousness is a physical primitive, like "magnetic field". Functionalism is that consciousness is a system like a clock, or other biological systems like digestion, which can use different physical primitives to do the job.

E.g., you can use gravity or EM or whatever to run a clock, but you can only use EM to "run" a magnet. With consciousness, at some point do you need to appeal to "consciousness" forces, or can you use any physical force that does the job? That's the technical debate between physicalism and functionalism. (You can challenge the distinction too, of course; that's just how the tradition has framed the debate.)

You're using terms like "chaos", "probability", "AI consciousness is still possible", "organic", which are quintessential functionalist arguments.

Quote:

Since I'm a functionalist, I agree it's inevitable because it's basic probability in action. But that's a classic functionalist argument (you're taking my position I meant to use against you). It doesn't matter what chemistry you have to work with (to a degree of course; apparently without liquid water your chances go way down, etc); but after a threshold, the basic probability of natural selection will take whatever chemistry is on hand and you'll get consciousness out of it. That's philosophical functionalism. It's basically saying consciousness is a functional product of evolution, with whatever chemistry is on hand, not a physical primitive like a magnetic field. (Most functionalists don't say AI makes consciousness; they say evolution makes consciousness. But some functionalists have more confidence in AI, and I think it's them you're taking issue with, not functionalism generally.)

As for your specific comment, though, actually I was thinking of something more pedestrian. What if we had evolved on a planet giving rise to a silicon organic chemistry? If consciousness is restricted only to carbon organic chemistry, then there's no possibility for consciousness on this planet. But of course the difference between a planet having carbon or silicon organic chemistry is a cosmic roll of the dice. That's why I wondered isn't it a huge coincidence our life went down the carbon-chemistry track and consciousness, rather than silicon chemistry and none?

But if you think consciousness *can* be in a silicon organic chemistry -- and you already admitted it's possible in AI, so silicon organic chemistry shouldn't be a problem -- then that's philosophical functionalism. The traditional debate isn't talking about the "quality" of consciousness; only what it is, or how it exists, as the most basic scientific domain.

But there is this interesting point you made, an AI could have consciousness, but it wouldn't be "human consciousness". That crosses the boundaries a bit. It's still a functionalist position I think, since I don't see how consciousness could ever exist in a totally different chemistry without it being a "functional" entity (as opposed to a physical primitive). But differences in the "quality" of different chemistries raises a second question: is there some physical primitive that's distinguishing human and AI consciousness, or is it go back to the functional organization is different, more complex non-linear dynamics are involved with human consciousness, etc, and functional differences like that explain the difference. (It's still consistent with functionalism to say it's the physical differences that are involved, the crazy nuances and dynamics of ions, gates, polarities, etc, as long as it's the functional operation of those nuances that are doing the work; it's just if you have to appeal to specific "consciousness forces" that differ where you cross the line. Your response to the "replace human neurons with functionally equivalent artificial neurons" where consciousness gradually shifts does suggest that, though.) Again, though, this is just how the classic literature has framed the debate.

Interesting stuff. Also great to hear the perspective from someone trained in biology.

.....

Edit: Maybe it'd be good to simplify it to a clear question.

Given the thesis that radically the different chemistries of different worlds can hold consciousness in radically different ways (so it can't be any known physics that's connecting them), so you have these two conscious beings, one in earth carbon chemistry, one in alien-planet chemistry: do you think we need new physics to explain their mutual consciousness and both brains (otherwise radically different) appeal to that same extra physics in some way, that's the connecting piece, OR is there no new physics involved and the uniting element is the similar functional arrangement of the mechanisms involved (connected by the same inevitable logic of natural selection, like the ways eyes keep being independently evolved, although in very different arrangements.) The first is a version of physicalism, the second of functionalism.

My feeling is the logic of evolution practically ensures functionalism (otherwise you get into the "huge coincidence" problem of physicalism, I think, where you even agreed with me consciousness is no coincidence, whatever habitable planet you're on). Consciousness inevitably evolves, even in very different arrangements, like eyes = functionalism.

SubJeff on 6/2/2011 at 06:47

Short answer: a lot of my reasoning may seem like functional reasoning but the undercurrent differs because I believe in a divine spark.

Long answer later when I'm not on my mobile.

SubJeff on 6/2/2011 at 22:45

Ok, long answer.

Quote Posted by demagogue

@SE, I think, as far as which technical position you're actually arguing for, it's still a functionalist in its technical philosophical sense, and you're actually arguing against specific functionalist theories.

Physicalism is that consciousness is a physical primitive, like "magnetic field". Functionalism is that consciousness is a system like a clock, or other biological systems like digestion, which can use different physical primitives to do the job.

So now you're going to shoot me down for trying to be logical but simultaneously believing in sky wizards.

I don't think consciousness is a physical primitive. I believe it's a property of the substrate, in this case the brain, but that there is a greater power that is responsible for the physical laws of this universe and that events have been set in motion in order for consciousness to evolve in the way it has.

Consciousness could exist in a system that consists of different components but which is functionally similar to our brains so silicon, or other, organic brains are a possibility. But again, I don't believe that the type of consciousness will be equivalent to ours. The sticking point here is going to be the term "consciousness" because I think that different forms of it can exists, and that on a superficial level those forms may look similar and behave in similar ways. Two separate forms may behave in an indistinguishable fashion until one day something testing happens that shows you that they are completely different.

Take reductionism as an example. Let's say we have a brain that is conforms to reductionist theory - Renz's "meat puppet". Let's say that the meat puppet knows it's a meat puppet, is acutely aware of it. It's behaviour may be very similar to yours or mine but when the day comes for it to choose between your life and its life its knowledge of its own meat puppetry will make it choose for very different reasons than you or I.

Quote:

E.g., you can use gravity or EM or whatever to run a clock, but you can only use EM to "run" a magnet. With consciousness, at some point do you need to appeal to "consciousness" forces, or can you use any physical force that does the job? That's the technical debate between physicalism and functionalism. (You can challenge the distinction too, of course; that's just how the tradition has framed the debate.)

Do you have to come down on one side of the fence? Can't it be neither?

Quote:

I don't think it's inevitable only because of probability, but because a greater power knew it would be inevitable if it set the rules of the universe in this way. So it may be an evolution of sorts, but it's a targeted evolution.

Quote:

As for your specific comment, though, actually I was thinking of something more pedestrian. What if we had evolved on a planet giving rise to a silicon organic chemistry? If consciousness is restricted only to carbon organic chemistry, then there's no possibility for consciousness on this planet. But of course the difference between a planet having carbon or silicon organic chemistry is a cosmic roll of the dice. That's why I wondered isn't it a huge coincidence our life went down the carbon-chemistry track and consciousness, rather than silicon chemistry and none?

As I've said - I don't think it's limited to carbon based chemistry, or even to the chemistry we know about. We've no idea what other elements will be out there. And as I understand it silicon chemistry as we know it isn't really suitable because at this temperature and pressure carbon is the only suitable element for the type of long chain organic molecules you need for this level of complexity.

Quote:

But if you think consciousness *can* be in a silicon organic chemistry -- and you already admitted it's possible in AI, so silicon organic chemistry shouldn't be a problem -- then that's philosophical functionalism. The traditional debate isn't talking about the "quality" of consciousness; only what it is, or how it exists, as the most basic scientific domain.

I'm reiterating a lot here but I'm trying to address all the points in your post. The traditional debate may not talk about quality of consciousness but I think that the quality of it is the most important flaw in functionalism. So I'll concede that functionalism is correct to say that a type of consciousness can exist in a silicon biology but it's not the same a human consciousness and that is an important distinction.

NB: I'm not saying that this a lesser type of consciousness, just that it has different qualities.

Quote:

Given the thesis that radically the different chemistries of different worlds can hold consciousness in radically different ways (so it can't be any known physics that's connecting them), so you have these two conscious beings, one in earth carbon chemistry, one in alien-planet chemistry: do you think we need new physics to explain their mutual consciousness and both brains (otherwise radically different) appeal to that same extra physics in some way, that's the connecting piece, OR is there no new physics involved and the uniting element is the similar functional arrangement of the mechanisms involved (connected by the same inevitable logic of natural selection, like the ways eyes keep being independently evolved, although in very different arrangements.) The first is a version of physicalism, the second of functionalism.

Ha ha. You know my answer to this. Neither. We don't need new physics but that doesn't mean that the alien-planet chemistry brain has a similar functional arrangement to the carbon chemistry brain.

The alien brain may well use regular physics but in a different way because the greater power has set the path of evolution on that alien world in that direction. Remember that our physical chemistry only works in the way it does because of the energy that is contained in this part of the universe. We experience water in it's three physical forms here but we never experience oxygen in it's solid form because it's melting point is -183 degrees C. If we take carbon, and silicon, as the template (because they have 4 potential bonds and so can chain) what about tin or lead based organisms where the temperature means those elements exist as liquids and gases?

Quote:

My feeling is the logic of evolution practically ensures functionalism (otherwise you get into the "huge coincidence" problem of physicalism, I think, where you even agreed with me consciousness is no coincidence, whatever habitable planet you're on). Consciousness inevitably evolves, even in very different arrangements, like eyes = functionalism.

Sky wizards say no to this.

But seriously, I don't want you thinking that this "greater power" is necessarily a God, or a consciousness in the way that we think of one, rather a guiding force that has informed the creation of the universe. Call it the first principle if you like, or the over-arching physical law. I can't define it, but I believe it's there.

Of course, this is probably the evolved meat puppet in me talking and now I'm going to kill you all.

:p