I hate to say "So now let's accept this and move on," over such a philosophically controversial question, but it seems like a considerable majority of Overcoming Bias commenters do accept this.

Yep, I accept this, let's move on!

[Warning: Here be sarcasm] No! Please let's spend more time discussing dubious non-disprovable hypotheses! There's only a gazillion more to go, then we'll have convinced everyone!

Consciousness, whatever it may be - a substance, a process, a name for a confusion - is not epiphenomenal; your mind can catch the inner listener in the act of listening, and say so out loud. The fact that I have typed this paragraph would at least seem to refute the idea that consciousness has no experimentally detectable consequences.

Eliezer, I'm shocked to see you write such nonsense. This only shows that you don't understand the zombie hypothesis at all. Or, you suppose that intelligence requires consciousness. This is the spiritualist, Searlian stuff you usually oppose.

The zombie hypothesis begins by asserting that I have no way of knowing whether you are conscious, no matter what you write. You of all people I expect to accept this, since you believe that you are Turing-computable. You haven't made an argument against the zombie hypothesis; you've merely asserted that it is false and called that assertion an argument.

The only thing I can imagine is that you have flipped the spiritualist argument around to its mirror image. Instead of saying that "I am conscious; Turing machines may not be conscious; therefore I am not just a Turing machine", you may be saying, "I am conscious; I am a Turing machine; therefore, all Turing machines that emit this sequence of symbols are conscious."

Eliezer, I'm shocked to see you write such nonsense. This only shows that you don't understand the zombie hypothesis at all. Or, you suppose that intelligence requires consciousness. This is the spiritualist, Searlian stuff you usually oppose.

Um, no. What it IS is a radically different meaning of the word than what the p-zombie nonsense uses. Chalmers' view requires stripping 'consciousness' of any consequence, while Eliezer's involves leaving the standard usage intact.

'Consciousness' in that sense refers to self-awareness or self-modeling, the attempt of a complex computational system to represent some aspects of itself, in itself. It has causal implications for the behavior of the system, can potentially be detected by an outside observer who has access to the mechanisms underlying that system, and is fully part of reality.

Do not confuse two totally different concepts just because the same word is used for both.

This isn't your clearest essay, and I'm not completely sure of the point you're making, but I think you make quite a leap at the end. It seems like you want to equate my awareness of changes to myself with my identity; so any change that's imperceptible to my awareness of myself would not change my identity. This seems odd. For one thing, if you tell me you're going to change my neurons to robotic equivalents (or if I study the procedure), aren't I then aware of the change? I think you'd have a hard time defining consciousness as something I could detect a change in.

"Zombies" are putatively beings that are atom-by-atom identical to us, governed by all the same third-party-visible physical laws, except that they are not conscious.

That seems to me be a bit beyond current technical ability (whether or not 2 things on the scale of a human being are atom-by-atom identical).

I'm not sure there's huge value in spending a lot of time on that "problem", except a very small fraction of our energy as a persistence-maximing hedge, sort of like spending a very small amount of time (if any) on planning how to beat proton decay trillions of years from now. http://www.pbs.org/wgbh/nova/universe/historysans.html

I've seen the term zombie used however, in ways other than your definition in this piece. For example, mental "uploads" that profess to be the person uploaded. That could be a little trickier, because just because something fools an observer into thinking that it a particular subjective conscious entity (for example, that it's me, HA) doesn't mean that it is. And since our technology can't current do atom-by-atom comparisons of humans, it requires less than that to fool almost any current observer. That to me is the more relevant problem currently. In attempting to maximize my persistence odds, I don't want to minimize my chances of being replaced by a "zombie" in that sense: something that meets current discernment technology but doesn't actually preserve my subjective conscious experience. Practically, it seems to me this results in giving somewhat greater weight to persistence strategies that are more conservative in keeping my subjective consciousness in something closer to it's current wet brain in a human body experience (as opposed to 'uploading', etc.)

I always took it to be that a zombie could catch itself thinking (as a listener) in the same way a non zombie could, the zombie doesn't lack inner speech in that sense, the whole causal chain remains and doesn't create an issue well at least not for a certain set of definitions and understanding of logic.

which really just reflects my agreement with Frank's "Please let's spend more time discussing dubious non-disprovable hypotheses!"

however I like "Thou shalt provide operational definitions for your terms!"

Phil Goetz: Have you been reading the recent posts and comments? I'm very surprised to see you being surprised if you have.

Caledonian: Has anyone ever suggested to you that you look into early-mid 20th century refutations to "positivism"? Operational definitions etc are good heuristics, not divine edicts.

Eliezer, does this whole theory cause us to anticipate something different after thinking about it? For example, after I upload, will I (personally) feel anything or only the death-like dark nothingness comes?

I think I did find such a thing, involving copying yourself in parts varying in size. (Well, it's leading to a contradiction, by the way, but maybe that's why it's even more worthwhile to talk about.)

Caledonian: Has anyone ever suggested to you that you look into early-mid 20th century refutations to "positivism"? Operational definitions etc are good heuristics, not divine edicts.

They are neither heuristics nor edicts. They're what's necessary for a definition to be functional and make sense - if you cannot divide the world into A and ~A based on a provided definition, it is invalid.

As for positivism, the 'refutations' made certain assumptions critical to their validity that I assert do not hold. With a whole field dominated by Richards, why would you assume that long-standing consensuses are valid?

The concept of logical positivism is certainly wrong... but it's the 'logical' part that's the problem.

Will Pearson [about tiny robots replacing neurons]: "I find this physically implausible."

Um, well, I can see it would be quite hard. But that doesn't really matter for a thought experiment. To ask "What it would be like to ride on a light beam?" is quite as physically implausible as it gets, but seems to have produced a few rather interesting insights.

How can we possibly move on when there are still people who are wrong on the Internet?

One of the very many problems with today's world is that, instead of confronting the root issues that underlie disagreement, people simply split into groups and sustain themselves on intragroup consensus.

If we do this every time we run up against a persistent disagreement, we will never actually resolve any issue; we'll just winnow down the number of people we're willing to listen to until we're secure in a safe and comfortable echo chamber, with our own opinions bouncing back at us forever.

That is an extraordinarily bad way to overcome bias.

Caledonian: It doesn't look to me like Philosophy has always been dominated by people with such a weak grip on the opposing positions, at least with respect to reduction. Russell, for instance, is a clear counter example, though he was weak in his understanding of the economic mind set. Definitions are a useful tool for thought, not the whole thing. The classical disproof of positivism is that it is self-contradictory. "Only the empirical can be true", but that statement is not empirical.

Will Pearson: I share some of your suspicion that replacing neurons may not be possible. Reactions to mind altering substances should be fairly easy. Ditto reaction to acceleration etc. I don't think that any of those, nor the diseases nor the EM radiation really need to be copied according to the "generalized anti-zombie principle" in any event. Would you really worry about a pill that eliminated all of them possibly also eliminating phenomenal consciousness? Seems far more likely that it would preserve it from those things. OTOH, mimicking the construction of new neurons and connections sounds very tough. It doesn't seem very likely to me that after replacing my brain with these robots I wouldn't still be "conscious" and "me" but it seems not unlikely that I would fairly soon be a brain damaged version of "me", possibly in a manner that was opaque from outside, possibly in a manner that was at first opaque from outside and later not.

Caledonian: Good point about echo chambers, but its far from clear to me how to fix it. It's a fairly clear empirical fact that most people are not receptive to arguments on many topics. Since they persist in disagreement at some point we have to stop listening to some of them if we are ever to get on to doing anything else.

but its far from clear to me how to fix it.

Well, the first thing we have to do is stop talking about the argument as though it were a matter of possibilities and probabilities: "it seems unlikely that", "I suspect that", "this is unlike what we would probably expect", et cetera, need to be abolished. The argument must be logically resolved, not merely trail off with stated positions that one side feels are "reasonable".

The p-zombie advocates are confusing physics-as-it-is and physics-as-we-understand. It is entirely possible that there are phenomena that our current understanding of physics and limited powers of observation might not include. But those hypothetical new things would be detected IF AND ONLY IF we noticed that the world did not act as our model said it should, given the available conditions. That would be the evidence we'd need to conclude that our model was missing some parts - perhaps our representation of the conditions were wrong and our rules were right, or perhaps our rules were inadequate.

If we had 'souls', 'consciousness', 'experiences', 'qualia', whatever we wish to call the hypothesized "new things", they would bring about changes in the world that the models that did not include them could not account for. The p-zombie advocates explicitly rule out this possibility: the p-zombie world acts precisely as ours does in all respects, not just the ways we can currently see.

Ergo, the properties that they postulate, that make p-zombies different from non-p-zombies, do not exist. Imagining a p-zombie as distinct from a 'conscious entity' is not possible, because the two things are the same. They have precisely the same properties, it's just that the labels that point to them are different.

Eliezer doesn't go far enough. Chalmers' idea of consciousness isn't just unnecessary, it's incoherent. It's not merely improbable, it is wrong. The people postulating effective epiphenomena aren't fiddling with trivia, they are logically contradictory.

If we cannot perceive a logical contradiction of this simplicity and directness, how do we expect to resolve subtler questions?

The only way I can see p-zombieness affecting our world is if

a) we decide we are ethically bound to make epiphenomenal consciousnesses happier, better, whatever; b) our amazing grasp of physics and how the universe exists leads our priors to indicate that even though it's impossible to ever detect them, epiphenomenal consciousnesses are likely to exist; and c) it turns out doing this rather than that gives the epiphenomenal consciousnesses enough utility that it is ethical to help them out.

Caledonian:

One of the very many problems with today's world is that, instead of confronting the root issues that underlie disagreement, people simply split into groups and sustain themselves on intragroup consensus. [...] That is an extraordinarily bad way to overcome bias.

I disagree. What do we have to gain from bringing all-and-everyone in line with our own beliefs? While it is arguably a good thing to exchange our points of view, and how we are rationalising them, there will always be issues where the agreed evidence is just not strong enough to refute all but one way to look at things. I believe that sometimes you really do have to agree to disagree (unless all participants espouse bayesianism, that is), and move on to more fertile pastures. And even if all participants in a discussion claim to be rationalists, sometimes you'll either have to agree that someone is wrong (without agreeing on who it is, naturally) or waste time you could have spent on more promising endeavours.

there will always be issues where the agreed evidence is just not strong enough

Then no one arguing can justify their positions, and everyone is incorrect in their assertions.

In any argument there can be at most one correct side. There's no principle saying that any of the sides involved must be right - only that only one can be.

There's also no principle mandating that any of the sides must be wrong. Incoherent arguments aren't wrong. They would have to go up in ontological status to be wrong. It would take a great deal of work and some serious improvement for them to be wrong.

P-zombies aren't right. They aren't wrong. They are merely nonsense.

Caledonian writes:

Um, no. What it IS is a radically different meaning of the word than what the p-zombie nonsense uses. Chalmers' view requires stripping 'consciousness' of any consequence, while Eliezer's involves leaving the standard usage intact.

'Consciousness' in that sense refers to self-awareness or self-modeling, the attempt of a complex computational system to represent some aspects of itself, in itself. It has causal implications for the behavior of the system, can potentially be detected by an outside observer who has access to the mechanisms underlying that system, and is fully part of reality. What Eliezer wrote is consistent with that definition of consciousness. But that is not "the standard usage". It's a useless usage. Self-representation is trivial and of no philosophical interest. The interesting philosophical question is why I have what the 99% of the world who doesn't use your "standard usage" means by "consciousness". Why do I have self-awareness? - and by self-awareness, I don't mean anything I can currently describe computationally, or know how to detect the consequences of.

This is the key unsolved mystery of the universe, the only one that we have really no insight into yet. You can't call it "nonsense" when it clearly exists and clearly has no explanation or model. Unless you are a zombie, in which case what I interpret as your stance is reasonable.

There is a time to be a behaviorist, and it may be reasonable to say that we shouldn't waste our time pursuing arguments about internal states that we can't detect behaviorially, but it is Silly to claim to have dispelled the mystery merely by defining it away.

There have been too many attempts by scientists to make claims about consciousness that sound astonishing, but turn out to be merely redefinitions of "consciousness" to something trivial. Like this, for instance. Or Crick's "The Astonishing Hypothesis", or other works by neuroscientists on "consciousness" when they are actually talking about focus of attention. I have developed an intellectual allergy to such things. Going on about zombies and consciousness as if you were addressing philosophical issues, when you have redefined consciousness to mean a particular easily-comprehended computational or graph-theoretic property, falls squarely into the category of ideas that I consider Silly.

Poke: . It seems like you want to equate my awareness of changes to myself with my identity; so any change that's imperceptible to my awareness of myself would not change my identity.

As Charles says, many real cognitive processes are not accessible to introspection, or are real and substantial without rising to the point of producing an immediate effect on the internal narrative. Maybe, as you were reading just now, your brain forgot some small fact about your thirteenth birthday. That would be a real change to your personal identity; if enough such changes accumulated, you would cease to exist; and you didn't notice it as it happened.

Albert's reply is that we aren't talking about a small change here; Charles has just postulated that your personal continuity was absolutely interrupted - one person died, another person was born, and neither of them noticed. This is what Albert thinks the GAZP prohibits, on account of it strongly resembling the notion that consciousness can be eliminated entirely without your internal narrative noticing; in other words, if you don't notice yourself dying and being born, you probably didn't - that what makes Albert think that something epiphenomenal is being postulated.

As a side note, Bernice might say that you don't notice the change because the area responsible for noticing it has been damaged (destroyed, actually) like an anosognosic patient who can't believe that their left arm is paralyzed. But Albert and probably even Charles would agree that this kind of specific functional brain damage to awareness-areas, is not occurring here.

Frank Hirsch: Riding a light beam is well specified. Silicon neurons having the same behaviour as protein ones is not well specified and not likely to be fruitful until it is so.

michael vassar : My statement was not about eliminating phenomenal conciousness, it was changing personal identity. I don't personally but other people strongly associate with altering their conciousness through drugs. Someone who couldn't get stoned with their mates or have a night out on the town drinking while enjoying the altered mind state might think that something was wrong with them, and that they weren't the same person as they were before.

And even if personal identity is not affected, social identity might be. That is other people would see you as a different person even if you didn't. If your personal identity is founded upon your relationships with others, this may be a problem.

I must admit I found the previous articles on Zombies somewhat tedious as I find the entire concept of philosophical Zombies to be specious. Still, now I'm glad I read through it all as I can see why you were so careful to lay down the foundations you did.

The question of what changes one can make to the brain while maintaining 'identity' has been been discussed many times on the Extropians list, and seldom with any sort of constructive results.

Today's article has already far exceeded the signal to noise ratio of any other discussion on the same topic that I've ever seen, so I am really looking forward to seeing where you go from here.

Michael Vassar:

The classical disproof of positivism is that it is self-contradictory. "Only the empirical can be true", but that statement is not empirical.

I have always been mystified at how this glib dismissal has been taken as some kind of definitive refutation. To the contrary, it should be perfectly obvious that a meta-statement like () a statement is nonsense unless it describes an empirically observable phenomenon is not meant to be self-referential. What () does is to lay down a rule of discourse (not meta-discourse). Its purpose is to banish invisible dragons from the discussion.

You cannot appeal to the "legitimacy" of sentences like (*) in order to argue on behalf of your favorite invisible dragon. But this is exactly what is going on in exchanges like the following: A: "The concept of consciousness is meaningless because it has no empirical consequences." B" "Silly amateur! Don't you know that logical positivism has been refuted?"

Expecting an argument to be able to justify itself is unreasonable, to my mind. Nothing can justify itself; everything must be justified by referring to something else, and the references cannot be circular.

Sure, you can always reference a deeper, more fundamental set of assertions to justify any particular claim, but what justifies those? You could construct an infinite chain that way, and still not explain how "assertions you make" can be justified, because you must always assume that the latest claims are justified themselves in order for them to support everything you claimed before them.

The key, I think, is to recognize that you can justify your claims only by pointing to something outside yourself. This applies as much to the totality of humanity as it does to an individual. You can construct an argument, but what validates your ability to construct is not yourself, but something greater; no argument can be constructed that validates that thing.

"It" is empirical reality, and the justification for human claims is observation of that reality. It does not need our support to function, and its functioning is unaffected by the arguments we make. The truth points to itself.

But this is another argument and will be argued another time.

For this discussion I use "consciousness" to refer to the mind's internal awareness of qualia. Consciousness may be an inherent property of whatever makes up the universe, i.e., even individual photons may have some essence of consciousness. Human type consciousness might then arise whenever sufficient elements group together in the right pattern. Other groupings into other patterns might generate other types of consciousness. Consciousness may have no purpose. Or perhaps certain types of consciousness somehow enhance intelligence and provide an evolutionary advantage.

If I don't trust that other people have a self awareness much like mine, then I have no reason to trust any of my senses or memories or beliefs. So I trust the evidence that other humans look like me, act like me, have brains like mine, and express internal thoughts in language as I do. I am only slightly less certain that mammals, birds, reptiles, and fish are conscious as they share common ancestry, have similar brain structures, and exhibit similar behavior. I am less certain about insects or worms. As I don't know the physical correlates of consciousness, the further from myself an entity is in structure and behavior, the less certain I am that it has an internal awareness similar to my own.

Animal consciousness can be explored by experimentation on humans, primates, mice, and fruit flies. The boundaries of consciousness can be mapped in the neural tissue of the brain. Cognitive scientists can explore what stimuli provoke a conscious response, what provoke an unconscious response, and what don't provoke any response. Scientists can observe what brain tissue is active when we say we experience qualia and what is active when we say we don't experience qualia. Studying brain injury patients provides a wealth of information concerning the brain's generation of consciousness...split brain, phantom limbs, aphasias, personality changes, delusions, etc.

Such experimentation indicates that our internal concept of self is largely an illusion. The mind tries to make sense out of whatever is available. If both brain hemispheres are strongly connected then there is a strong illusion of one internal person. If the brain hemispheres are disconnected, then experiments show two different personalities inhabiting the same brain. Each personality has no awareness of the other personality. When the second personality acts independently, the first personality rationalizes why the first personality "chose" to perform the action. It is possible that many such self aware personalities co-exist in our brains, each with its own illusion of being in control and each with its own perception of qualia. (In some brain injuries, a person no longer believes that their own arm is part of self. Even though they can control the arm and feel what the arm touches, they think it is someone else's arm. The brain function that creates the illusion of self is broken.) These illusions of self may not be necessary to experience qualia but probably are necessary for a human to describe or relate the experience of qualia.

Speculation about zombies should take into account what science has already discovered. I.e., our internal concept of ourselves is only a blurred reflection of reality. "Self" is manufactured on the fly out of bits and pieces that change with every experience, with every hormonal change, with every drug we take, or with every injury we experience.

What would our internal "self" experience as each neuron were gradually replaced by a nano computer simulator? If the simulator generated a similar essence of qualia (i.e., simulating a brain pattern is sufficient to generate the experience of qualia) then the internal experience should be the same. If the simulator produced no such experience of qualia, then our internal self would be unable to recognize that our internal awareness was shrinking. We would not be able to remember that we could once hear more sounds or see more colors as memory itself depends on that internal awareness. Our internal self would fade away unnoticed by that internal self. (In some cases of dementia, the patient doesn't comprehend that his mind is failing. He don't understand why his family has brought him to see the doctor.) With nano-simulators mental function would continue, but internal awareness might disappear.

I envy that your blog has attracted a much richer discussion on this topic than has mine (see for instance Phil Goetz's & Fly's recent posts).

"But if you add up enough tiny little gravitational pulls, they will eventually yank you across the room and tear you apart by tidal forces, so clearly a small effect is not 'no effect at all'."

When I reread this passage, I can't help but think of the combined gravitational pull of 3^^^3 dust specks.

Such experimentation indicates that our internal concept of self is largely an illusion.

It is relatively easy to discover that the self / me is only a thought, not the reality it is assumed to be. Some basic inquiry into the nature of the assumed "self" will dissolve the illusion rather quickly.

It is often assumed that this is some kind of "religious" belief, but in fact it is also easily available to atheists as well, such as Susan Blackmore and Sam Harris. I suspect Nick Tarleton would also include himself in this category.

Wittgenstein's post-Tractatus work concentrated on the role played by language in the kind of talk we sometimes call "philosophical discussions." In this later portion of Wittgenstein's work, his central activity consisted of pointing out, by means of numerous distinct examples, that one cannot stretch the use of words and phrases into new, synthetic realms, without incurring a significant risk of ending up talking nonsense.

The trap against which Wittgenstein warns us is something like writing "Where has a wireless mouse's tail gone? Possibly it is still somehow attached, but invisible."

In this case it is obvious that the word "mouse" can be used in a number of different contexts, and that these various contexts are not actually linked together in any but the trivial fact of both containing the word, "mouse."

In the case of words such as "conscious," "aware," "thought" and the like, the lack of connection between contexts is less apparent. Likewise, the danger of conflating and distorting those contexts is significantly greater for such (actually quite narrow) words.

Another comment to add a few years later than the original post and hence be pretty useless:

My thoughts are that consciousness (as in the experience of it) is a kind of epiphenomenon:

The sensation is derived from cognitive processes that map isomorphically to an abstract model of consciousness in mindspace (and I do not make any distinction or heirarchy between realspace and mindspace in terms of privileged levels of existence).

It does this because the brain is doing exactly what it feels like consciousness does - integrating various inputs into a representation of self and environment, making plans and telling a consistent story about it all. And the mapping, by being possible, is also real.

have to say I agree with Charles' proposition. I mean, if one thinks "i am thinking" the neurons have to fire off in your head A) to think B) to say "I am thinking" C) to realize one is saying he or she is thinking D) determine the cause and thought process of all of the above and E) rationalize the behavior of our brains in a inductive reasoning-based processing sense of the word.

So, if all of the above are true, as are the aforementioned butterfly effect that causes a misplaced neuron to trigger a seizure, than if one's neurons were replaced by other completely identical neurons, then you would have consciousness, but not the same consciousness, and not necessarily a human consciousness. (Also the argument depends on whether one believes if random is really not random at all, and if it is random, than the robot neurons could not replicate that process in an algorithm, since there isn't one (in that case) than the randomness of human consciousness would constitute the definition of the difference between the robot consciousness and the human one, if the robot conscious is actually considered "conscious" at all, which would mean zombies COULD exist due to the lack of randomness in a robot-neuron-composed brain.) BUT, if randomness isn't actually random at all, and such variables as pi consist of a very complex pattern, then whose to say robots cannot replicate the pattern, in which case human existence would be replicable, and there would be no difference between conscious robots and conscious humans, but the unconscious would not be unconscious unless they were dead, thus proving the GAZP. Does anyone else agree or am I missing something?

I'm sorry to comment on such an old post, but I'm really new to rationality and especially bayesianism, and this discussion got me confused about something.

Non-reductionists such as Richard say there is a non-physical "thingy" called a consciousness, and that it is epiphenomenal. That means it has no consequences on the physical world.

Wouldn't this be a model that doesn't anticipate anything, as you described in your first posts? If one argues that conciousness has no effect on the observable world, isn't one arguing that there might not be any conciousness at all? That the whole argument is pointless?

The gravitational pull from a one-gram switch ten meters away is around 6 * 10-16 m/s2. That's around half a neutron diameter per second per second, far below thermal noise, but way above the Planck level.

[...]

The switch's flip does change the fundamental particles of your body and brain. It nudges them by whole neutron diameters away from where they would have otherwise been.

But when you flip the switch, it doesn't disappear altogether, it just gets displaced by a few millimetres, so the number you care about is this, which is hundreds of times smaller.

Albert's position is similar to how you know that two calculators will have the same output despite having different physical configurations. If you have an idealized abstract model of say addition, you can draw a boundary around different designs that perform addition despite being different. You will know that something like a unconnected switch won't be enough to make it stop matching the model of addition.

If we take the reason that Albert talks about being Albert and him talking about a person from five minutes ago as himself and build an abstract idealize model, we will see that a lot of physicals differences can take place without effecting the real reason for his report. There is a range of physical designs that match the models prediction and it includes ones were his brain is made of tiny robots. The cause of his report with remain the same regardless of the tiny robots replacing his neurons with themselves.

The implication I see is mind uploading. Where must we draw the boundary to capture the referent of "Albert". How that questions is answered may determine the future.

This is the most I've read on this Zombieism concept, and now I can see it may not be the first thing I read about it. There is a fantasy series called Skulduggery Pleasant by Derek Landy. In one of it's many side plots, two characters become zombies and even eventually get their brains fully replaced by genetically modified plant matter. They retain their consciousness and their personalities the entire time. They also continued functioning without a hitch after their Zombie Master died (the term was used in the books).

So I suppose the author would agree with this principle, and I find myself inclined to as well. It just makes so much sense, although I personally feel Eliezer could have been a bit more concise.

I will take this opportunity to recommend the series to all rationalists. It's the most rational piece of fiction I've read aside from HPMoR.