Saturday, February 25, 2012

Too Many Explanations

Yeni Cami Cat
This cat is conscious. Discuss.

My previous post sparked an unexpected and irritating discussion regarding physicalist versus nonphysicalist explanations of consciousness. I vowed I wouldn't go there, but there you are, I can't help myself. Because I'm really unhappy with all attempts to explain it that I've come across.

A bit of background.

In my early teens, I did a massive amount of reading into paranormal phenomena. I believed every word of it, too. Then I started reading up on the skeptical literature related to them, and pretty soon all of that belief was gone. For a few years, I hung out on alt.atheism, sci.skeptic, and a few other forums dedicated to debunking flim-flammery of all sorts. Then that got old too.

Then at some point I realized that actually those purely materialist, neurophysicalist explanations of consciousness—emergent materialism and all that—don't really explain anything either. They all end up as making perfect sense up to a certain point and then, poof! consciousness. That poof! is never even addressed. Instead, skeptics like to substitute some nonmaterialist explanation for it, and then tear that to pieces. It was good fun, too, but doesn't really help.

This is an attempt at writing up the reasons for my dissatisfaction with the various attempts at explaining consciousness—what it is, where it comes from, what it's for—that I've come across. Only at a very crude level, since there are so damn many when you drill down. So sit back and grab a beverage of your choice, 'cuz this is going to get long.

Ultimately, you can sort all of these explanations into two piles: the physicalist ones here, the nonphysicalist ones there, and I don't like any of them.

Definitions. I'll avoid using the word 'materialist' or 'materialism' here, instead using 'physicalist' or 'physicalism' as appropriate. The reason is that 'materialist' tends to create an association that excludes 'energy,' whereas I—and modern physics, for that matter—see matter and energy as different types of manifestations of the same thing, namely, the physical universe. Einstein even showed exactly what their relationship is. In the context of what I'm trying to say, the distinction between matter and energy doesn't matter at all.
—Albert Einstein
So, 'physical,' as opposed to 'nonphysical.' As a matter of fact, I have a real problem with the whole notion of 'nonphysical' in this context, which I'm going to explore further down. There are sensible uses of that word too, but I'm not treating them here. Thoughts, values, and information are nonphysical, for example, although as far as I know they're always associated with a physical substrate. More on that later too.

Most physicalist explanations of consciousness see things more or less like this:

Humans are bags of salty water programmed to reproduce themselves. Over hundreds of millions of years, due to evolutionary pressure, these bags have developed a very sophisticated internal organization that includes a hugely intricate analog computer we call 'the brain.' This analog computer continuously runs a bunch of as yet incompletely understood algorithms, crunching inputs into outputs. The inputs are provided by the senses. The outputs are externally manifested as action, and internally manifested as mental states. There's also all kinds of interal recursion, where outputs can also provide inputs to other algorithms. It's fiendishly complicated, but fundamentally no different from what a silicon-based computer does, other than in the degree of complexity and in that the computer is primarily analog rather than digital. This is a minor matter, because as per the Church-Turing thesis, a sufficiently complex digital computer can simulate any analog system to any desired degree of precision.

There are a couple of unspoken assumptions about this story.

One: it leaves consicousness pretty much undefined. Questions like "is a perfectly accurate simulation of consciousness actually conscious?" are usually dismissed as so much philosophical wankery.
It's impossible to tell, so who the fuck cares? Next question.
Two: it doesn't actually explain anything about consciousness at all! If we really are only analog computers crunching through algorithms, then why bother with consciousness at all? We would behave exactly as efficiently if we didn't have any consciousness, but merely the algorithms. What's the advantage of developing consciousness? Where did it come from? Poof!

Three: it takes a hugely reductionist view of what it means to be human. I may be made of a bag of salty water preprogrammed to replicate itself, but that's not what I am. In fact, I do not feel any particularly pressing need to self-replicate; I have no children, and am not terribly depressed about the fact. The things I value in life do rely on the reliable functioning of that mechanism (and are sorely disrupted if that functioning goes even a bit haywire, even by something as trivial as congested sinuses), but they go a lot beyond that.

Put another way, at their best, physicalist explanations of consciousness can be useful at explaining things about how consciousness works, but they have fuck all to say about what it is, or even where it comes from, or what it's for. It just... appears. Somehow. From somewhere.

This is in fact a fundamental limitation of rationalism. Rationalism relies on the intersubjective, whereas there is no way that we know of to directly share subjective experiences. The Vulcan mind-meld would sort that out nicely, but we don't know how that would work in real life. This is the end of the road for it. Wovon man nicht reden kann and so on. There is simply no way to create a test that would distinguish a perfect simulation of consciousness from actual consciousness, since we're dealing with internal experience. We cannot know if such a perfect simulation really is conscious or not. Yet to claim that that makes the question meaningless is a bit of a leap.

Here we run into one fundamental problem: that of definitions. So let's get some of those clear here. What I mean by 'consciousness' is my subjective experience of it. I believe that you, dear reader, also have it. I believe my cat, currently napping on my lap and purring, has it, albeit with important differences. I think a honeybee might have some sparks of it. I think it's unlikely that a flatworm or paramecium shares this experience, without stretching the definitions beyond breaking point, but I'm open to the possibility. And I don't think a thermostat has it, even though it responds to external stimuli. And that's about the best I can do—a Popperian nominalist attempt at defining it by pointing at things that I think do and don't have it.

So. Other than the Vulcan mind-meld, there is no conceivable way to create a test that would be able to distinguish between a convincing simulation of consciousness with no actual internal experience associated with it, and the real thing, using this definition. Physicalists would therefore discard this definition as useless.

Yet I can't think of anything quite as important to anyone as consciousness (by this definition). Without consciousness, there is no subject to value anything to start with! If it were possible to use a MegaDeConscionizer to excise the internal experience of consciousness from an individual but leave him otherwise entirely intact, in my mind, that would be a close ethical equivalent of murdering him. If you dismiss this problem—as the physicalist must—you will thereby dismiss everything that makes it meaningful to be human.

Note that this definition is certainly not the only one, nor even the most reasonable one under all circumstances. Emergency first response personnel, for example, have a different definition of consciousness, which is extremely useful when determining how to respond to an emergency. It's based on how a (human) individual responds to common external stimuli—alert and normal, disoriented, responsive to pain, unresponsive. Thing is, they don't use this definition to distinguish between hypothetical non-conscious androids and actual conscious humans, or conscious alpha-level simulations and (possibly?) non-conscious beta-level ones, to borrow some words from Alastair Reynolds.

A slight tangent. I write software as my day job. There are a couple of hundred thousand lines with my name on them in our source repository. Some of that software I'm even quite proud of. I cannot conceive how such code, written by me or any being remotely like me, could become conscious, in any reasonable sense of the word. If true artificial intelligence is possible—and I see no reason it shouldn't be—I'm inclined to agree with Hubert Dreyfus (HT: NellaLou): the work done on it so far has been more like alchemists trying to transmute lead into gold with the methods and theoretical framework they used to successfully extract quicksilver from cinnabar. Or, to use his other metaphor, like someone who wants to go to the moon and considers learning to climb a tree—as opposed to, say, inventing rocket propulsion—as tangible progress towards that goal.

So, I'm unhappy with the purely materialist/physicalist/neurophysicalist view of consciousness. It's neither ethically tenable, since it makes no difference between an entity with no internal experience of consciousness but only the appearance thereof and one with an internal experience of consciousness, nor possessed of explanatory power, since it's based on a poof! between biochemical activity and the 'emergence' of 'mind.' Which doesn't mean it's worthless, of course—I'm quite convinced that physicalist approaches will be able to tell us a great deal more about the how of the mind. But it's a big, big mistake to assume that means they have any ability to tell us anything about the what or the wherefore.

Taking a little breather here, then going on to my beefs with nonphysicalist explanations of the mind.

Okay, breather over. Nonphysicalism.

For starters, lets get rid of what I'll dub 'weak' nonphysicalism. This is pretty much your standard physicalist explanation with the 'poof!' designated as some weakly defined nonmaterial 'property' that then may influence the the physical activity of the brain (as demonstrated e.g. in studies about neuroplasticity and meditation) in some vaguely defined way and undefined mechanism. I have no particular beef with this story; it just doesn't add anything beyond the label to the physicalist one, so I have nothing particular to say about it either.

'Strong' nonphysicalism, however, makes more ambitious claims. The specifics of these claims can vary wildly. Some people posit personal, immortal 'souls;' others talk about âtman and brahman, purusa and isvara; yet others about universal consciousness. What's common to these claims is that they all posit some nonmaterial substance or entity which interacts with the material in some at least somewhat specifically defined ways.

I have a few fundamental objections to these ideas, and some specific and particular ones to commonly-expressed specific and particular expressions of them.

First, one that I've asked here—and elsewhere—repeatedly, which I'll elaborate on a bit more this time.

Isn't a nonphysical substance or entity that interacts with the physical a contradiction in terms?

Consider the brain. Let's assume it demonstrated that mind-states are closely bound to brain-states. A fever will affect your mind-state; electrical stimulus of parts of your brain will affect it; drugs affecting brain chemistry will affect it; brain damage will most dramatically affect it. We also know a fair bit about the specifics of this brain activity. We know that it's biochemistry involving clusters of ions passing through partially permeable membranes; proteins knitting together and transforming as axons grow to touch each other, and so on and so forth. We know that messing with these processes will affect mind states. We also know—e.g. from those studies of neuroplasticity and meditation—that sustained application of mental effort will noticeably show up as different brain states. So far so good.

How does this happen?

If the mind is a nonphysical property or entity, how, exactly, does it cause those ions to jump over axons—or not—or those axons to knit—or not? These are physical objects and structures. Unless just about everything we think we know about the way the physical world works is wrong, you need to apply some kind of force to effect those changes. What is that force? What mediates it? If it acts on these physical objects, how does it make any sense to say that it itself is nonphysical?

Now, some genuinely weird shit does go down at the quantum level. The uncertainty principle and all that. Quantum computing, where you have an algorithm that depends on qubits in a state of superposition, but you read the result by collapsing them into some particular state. That's plenty weird, and it does not involve Newtonian forces pushing things around. That's a tempting thought, and charlatans like Deepak Chopra have made major hay from it. I have been seriously tempted by the idea myself.

Trouble is, the brain's too damn big. At the scale of neural structures, quantum effects cancel out. I mean sure, there's still the butterfly-flapping-its-wings-causing-a-tornado kind of thing, but that's so many steps removed from ions actually jumping or an axon actually knitting that it doesn't work as an explanation. Not unless you can somehow demonstrate that the mind actually operates on a far smaller scale than the brain—and if that's the case, then why is the brain so damn important to mental states? So as far as I'm concerned, the quantum consciousness dog don't hunt either.

The second fundamental objection I have to the nonphysical mind concerns information theory. A nonphysical mind implies disembodied information: information that is not encoded in anything. I can only think of one example of such information, namely, the stuff we call laws of nature. E=mc2 even if you never try to convert matter to energy, or vice versa. Perhaps consciousness, in some sense, is coded into the universe in the same way.

However, 'laws of nature' are qualitatively different from 'information' in the usual sense of the word. As far as we know, there's no way to alter the message written into them, and we can only read them indirectly, by observing regularities in the universe. While this could work for a general concept like 'consciousness,' I don't see what bearing that could have on specific features of individual consciousnesses. This would require reading and writing, not just reading through indirect means. You can't write if you only have a pen and nothing to write on. You can, and many have, glibly dismissed this objection by positing a "nonphysical medium" of some sort, but from where I'm at, a nonphysical medium that's able to carry information and act on the physical world is starting to look pretty damn physical to me. So if it can do this, then where is it? How can we detect it? What creates it? And, most importantly, where's my fucking lightsaber?
Well, the Force is what gives a Jedi his power. It's an energy field created by all living things. It surrounds us and penetrates us; it binds the galaxy together.
As good an explanation as any, as far as I'm concerned.
This is a pretty good bridge to the specific objections, which have to do with the specific claims made about this nonphysical entity or property called 'the mind.' Things like paranormal phenomena (ESP, out of body experiences, psychokinesis, etc.), miracles (miracle healing, walking on water, statues weeping blood or water, etc.), past lives (remembering them in a trance-like or even normal, waking state), and so on.

There was I time when I really wanted to believe in any or all of those. I did, too, in a lot of it. Yet all that delving into skeptical literature about it wrecked it. Simply put, I discovered a terribly, terribly strong pattern: when looked at closely enough, miracles tend to disappear.

Believers in paranormal phenomena like to point out that lots of experiements demonstrating them have been performed under conditions at least as well controlled as those that apply to standard scientific procedure. That is quite true. It is also insufficient. Thing is, the universe doesn't lie, but people do. A scientist with honest intent to test a hypothesis with an experiment will not expect the Universe to make those neutrinos scoot a few nanoseconds faster just to mess with his head. So he controls for all possible sources of error arising from any mistake he might make.

But people do lie. A lot. Convincingly. Even to themselves. And scientists are demonstrably lousy at catching them at it. So whenever someone who is genuinely good at tricking people—a stage magician, for example—is brought in, those phenomena somehow mysteriously evaporate. The usual defense for this is to claim that the skeptic's skepticism inhibits the phenomena; that they only manifest if everybody present really really believes in them. Not good enough for me, unfortunately.
There was just recently a story about an ascetic who never eats. He was verified medically not to eat, poop, or piss, but to be in perfect health. Miracle! Only on close inspection it turned out that he wasn't actually under medical supervision all the time—he wandered off on his own for a bit every day, you know, to gargle, sunbathe, meet his students, that sort of thing. And the medical records actually collected turned up some pretty weird stuff when looked at closely. Like, at one point there was poop in his colon and urine in his bladder, then later, there wasn't. Which is more likely—that his body magically transformed the poop and urine back into living tissue, blood, and lymph, or that he took a poop and a pee on the sly? Naturally, a sadhu like him would never, ever take advantage of that kind of opportunity to cheat. Right? Right.
Thing is, most of the claims about paranormal phenomena should be pretty easy to test. You say you can do ESP? Fine: let's have you guessing which cards I'm holding... in conditions specified by a close-up magician that carefully eliminate any sensory way of getting that information, such as through 'tells,' marks in the cards, or something as simple as seeing them reflected—even inaccurately—in the holder's glasses. Psychokinesis? Fine: here's a feather inside a sealed glass tube with near-vacuum in it. Move it.

This sort of thing has been done lots of times too, and it never. fucking. works. Nobody has claimed Randi's famous million-dollar prize, not because of the evil shenaningans believers attribute to Randi, but simply because all those famous paranormal phenomena fail to materialize—even the least bit—in conditions controlled not only for inadvertent experimental error, but for intentional human cheating.

Another tangent on one particularly Buddhist claim related to the nonphysical nature of the mind: past lives. I struggle with swallowing this concept whole on several levels.

First, there's the information theory thing. If karmic seeds and memories are transmitted from one life to another, what is the medium that carries them? Again, disembodied information makes no more sense to me than the idea of writing that isn't written on anything.

And second, there's a subtler thing. Even if things somehow happened exactly as the Tipitaka puts it—and I by no means completely reject the prospect!—then in what sense can my future life be regarded as 'me' or 'mine' any more so than, say, my life can be regarded as my father's, or his life his father's? My father's and mother's karma have conditioned a great deal of mine (something I come to realize more and more as I get older—the vast majority of that is very good karma, thank goodness—) and theirs has certainly been conditioned by that of their parents (a much more mixed bag, especially, I think, in the case of my father). I'm a lot like him. He's a lot like his father. He acts out behavior patterns passed to him by his parents. I act out behavior patterns passed to me by mine.

How is this any different from my karma—after my death—conditioning another birth in the future? Yet I don't regard my life as the same as my father's, nor he mine. So what's so special about rebirth anyway? From where I'm at, it boils down to the general principle of trying not to make a mess of things. My actions are conditioning other people's actions and existences all the time. I remember very little about my early childhood. What difference does it make if the karmic chain hops from one—and only one—individual to another when interrupted by death-and-birth, one karmic generation to another? How is it any different? How does it even make sense?

So my beef with the nonphysicalist explanations of consciousness is that either (a) they don't explain that poof! either, but only label it (weak nonphysicalism), or (b) they make testable claims which don't pan out when tested under conditions designed to exclude lying and cheating. So on the one hand, we have a vague mushy something that looks a lot like the Deists' inert creator God, only immanent, or some more specifically defined active principle that's so terribly shy that it hides whenever somebody looks very closely.

Unsatisfactory. So very unsatisfactory. Yet besides these two approaches, what's left? I'm damned if I know. There's a huge-ass hole right in the middle of everything. If you work up from physicalist principles—the algorithmic view of the mind—the inescapable conclusion is that consciousness is totally unnecessary. Things would be so much simpler if we could dispense with the concept altogether. Some people do, in fact; I've heard a few hardcore physicalists angrily dismissing the idea of freedom of will, for example. One of 'em might even be reading this.

Yet there it is. Bam. Every waking moment, permeating everything. Yet attempt to explain it through the nonphysicalist approach, and you invariably end up making it up as you go, perhaps convincing yourself into believing (usually relatively harmless) stuff on remarkably flimsy evidence in the process.

So here I am, acting out my old karmic patterns from my alt.atheism days, all derisive on the "I want to believe" set, yet deeply, profoundly unsatisfied by the lack of explanatory power of the physicalists—not to mention irritated at the lightness with which they shrug off that huge poof! in the middle.

What's left is just that: a mystery. Something that appears to be nothing at all—unnecessary, unlikely, inexplicable, even impossible—when examined closely, yet that's incontrovertibly and inexplicably there, all the fucking time.

There are ways to explore it, but they have fuck all to do with labs and science and experiments and chemistry and information theory and programming. I even believe there is an answer to be had, even if that answer isn't communicable to anyone who doesn't already know it.

That's what makes it so exciting. And frustrating at the same time.

No, I don't have any references. I lost track of my Martin Gardners and Carl Sagans, and I haven't stayed up to date with my James Randis and what have you. And I really have lost my stomach for one-on-one debate with paranormalists on this, especially as I no longer think I have anything substantially better to offer. So comments are welcome, but my reactions may be even more unpredictable than usual. This is a can of worns I haven't opened in a while, and I'm not quite sure what might crawl out. Be warned.


  1. Interesting view, but I do think the first problem we'd have in discussing this is exactly your definition of consciousness.

    It will be very hard for me to say anything conclusive from that kind of definition.
    Do you mean consciousness as the ability to think ? The ability to think about more than just survival ? The ability to think about complex problems ? the ability to see consequences to actions ? The ability to .... ?

    Also, one very specific point I would like to make is that a lot of current research is being done about the brain not just being an analog device, but partly digital too. The digital signals being the Action Potentials.

  2. That, David, is the point. In order to be able to say anything meaningful about it in the physicalist framework, you end up with such a reductionist definition of consciousness that it excludes the thing itself—the internal experience. But just because you can't define it doesn't mean it's not real.

    The answers to your questions in your second paragraph are (1) no, (2) no, (3) no, (4) no, and (5) ... .

  3. Lol :D

    I get that it can be real without a good definition.
    Even the definition of life is very hard. I'd say that defining the things we take for granted are the hardest. One of my teachers once challenged the class to define a chair.

    It really is hard.

    I do think that it's useful to keep researching all the avenues and the current project I'm working on in university is exactly part of that process.

    The project is trying to find a set of metrics to better understand how the brain encodes information using action potentials. While I also think it simplifies too many processes, it still is a very interesting avenue of thinking of the ionic currents in our brain as electrical signals, which encode things in binary exactly like we use computers :)

  4. Just following reading this I happened to listen to an episode of Quirks and Quarks about dolphin self-awareness (the "Dolphin in the mirror" section of

    It doesn't really go into explaining consciousness, but the interesting bit is that the frontal cortex where (the assumption apparently was) this self-awareness bit happens (barring non-physical explanations) for humans and other primates is sufficiently different in a dolphin to make that explanation a bit suspect.

    In fact, what if anything, are your thoughts of consciousness being defined as being self-aware?

    And, again, thank you for a wonderfully thought-out and thought-provoking post.

  5. I don't see consciousness and self-awareness (in the sense it's used in this context) as the same thing. Self-awareness requires consciousness, but you can be conscious without being self-aware. I think the equation of the two is unnecessarily constricting.

    When I've been doing zazen, sometimes my state of consciousness alters. It can happen that my stream of verbalized thoughts gets gaps in it, or stops altogether. It can also happen that my sense of the border between self-and-other starts to fade.

    I have not experienced a complete extinction of the sense-of-self in meditation, but I have had that sense altered enough that I'm sure that it can happen, and an inkling of what it might be like, and I'm pretty confident that it is not unconsciousness. These meditational states are increased rather than decreased focus and clarity; the polar opposite of what happens when you're approaching unconsciousness.

    So if a definition of consciousness would exclude such meditative states, then in my opinion that definition is badly flawed, at least as a general definition.

    In fact, I'm starting to think that all we can work with are badly flawed definitions—they may still be useful in some specific domains (e.g. the algorithmic definition of consciousness may be highly useful when investigating the relationship between brain-states and mind-states), but you get into real trouble if you start applying them outside their pretty narrow domains.

  6. Petteri,

    I agree with all your main points. I find dualist arguments incomprehensible, and find physicalist arguments interpreting qualia as emergent properties to be no more than mere hand waving. And alas, there's no third alternative. Maybe when we have a better handle on what matter really is....

    I also tend to be skeptical of paranormal phenomena. Most of the time. On the other hand, I remember being impressed reading Jules Eisenbud's book on Ted Serios. Randi and Marvin Gardner have both debunked the Serios phenomenon claiming Serios inserted a marble with a photograph inside it into his "gizmo," but their claim that this explains Serios's "thought-o-graphs" is seriously flawed. Serios produced his thought-o-graphs even when his gizmo was many feet from the camera, and some of the "errors" in his thought-o-graphs (e.g., misspellings) seem unlikely to have been produced in that way as well. Philosopher Stephen Braude at the University of Maryland has maintained an archive of Serios's thought-o-graphs and has argued the case for why Randi and Gardener's fraud claims don't hold water. Serios could still have been a fraud -- but I haven't seen any convincing explanation as to how he pulled it off. I maintain an open mind in this case, as well as several other cases that are less public. But of course, keeping an open mind doesn't mean the paranormal folks have successfully plead their case.

  7. The standard for a "discovery" in physics is six sigma. That's high. If paranormal phenomena like Serios's "thoughtographs" are real, they point to hitherto unknown, tremendously important properties of the universe. I don't think six-sigma certainty is an unreasonable standard for such extraordinary claims. And I don't think Serios's case comes anywhere close to that.

    My default assumption for such claims is to assume error or deception until otherwise demonstrated. Not being able to show exactly how the deception worked isn't necessary. People invent new magic tricks all the time, and most magicians won't be able to guess the exact mechanism just by watching it performed. To my knowledge, Serios has not produced "thoughtographs" under conditions sufficiently well controlled for deception.

    So yeah, open mind. You just get more and more jaded when you see the same pattern repeat over and over again—of an extraordinary claim made for paranormal phenomena melting away when subjected to truly rigorous scrutiny. There are more quacks and marks than skeptics, which means there will always be instances of quacks successfully gulling marks. There just isn't time or resources to debunk them all, even if they were willing to subject themselves to the kinds of conditions that would exclude fraud, and most of 'em aren't. So if you really want to believe, you'll always have something to latch on with 'maybe this is really it, even if that, or that, or that wasn't.'

    But really, all it would take to convince me is one ESP or PK demonstration—a simple one, "seeing" cards, for example, or psychokinetically moving a light object—in conditions rigorously controlled for cheating. If someone really can do that, they ought to be able to pass any test James Randi cares to devise. Why haven't they? If they're anything other than a combination of cheating and self-deception, why are these phenomena so damn elusive?

    Quoting Tim Minchin:

    "You show me that it works,
    And how it works,
    and when I've recovered,
    from the shock,
    I will take a compass and carve
    'Fancy That',
    On the side of my cock."