On Philosophy

June 13, 2006

Hilary Putnam on Why We Aren’t Brains in a Vat

Filed under: Intentionality,Language — Peter @ 1:41 am

Another long title, I know. This post will be the first in which the arguments I present aren’t my own, but rather a slight improvement on Putnam’s argument in Reason, Truth, and History (amazon). The goal of the argument is to show that we cannot meaningfully talk about, or even speculate about, being brains in a vat, or more generally that reality is somehow an illusion. And if we can’t speak meaningfully about it why bother considering it at all?

The crucial premise in this argument is that to be able to refer meaningfully to something we must have had some perception of the thing we are referring to. Somehow (depending on your theory of intentionality) it is the perception that has generated the meaning. For example since I have seen the tree in front of my house I can meaningfully refer to it. Now let us say that you have a tree in front of your house, which I have never seen, but that you have told me about. If I talk about the tree in front of your house the meaning of my words does not come from the tree but from your description of the tree. It is even possible that there is no tree at all in front of your house; even so my talk about it has meaning, because the meaning was derived from the description I had heard and not the tree itself; in actuality the reference of my words was not the real tree but the “tree” that was created by your description. I know this sounds vague, but since I am trying to be theory-of-intentionality-neutral you will have to fill in the details based on whatever particular theory you subscribe to.

Now suppose that there are people who are physically brains in a vat but who are living in a simulated world. All they can refer to, and think about, meaningfully are entities within their simulated world. (And possibly abstract entities as well, depending on the theory of intentionality again.) When we quote what people in this situation might say let us then follow their words with *s to show that meaning behind their words is different from ours. For example if we speak of cats we mean collections of atoms that tend to lay in the sun; when people in the simulated world think of cats* they are referring to some aspect of the program that provides them with certain visual stimuli, and not a collection of atoms. Thus if they speak about brains* and vats* they aren’t speaking of physical brains and vats but objects, possibly imagined, within the simulation itself. They cannot speak of brains or vats (without *s), but it would be these words that could give their hypothesis meaning, because clearly they are not brains* in vats*, after all they have arms* and legs*. Thus neither we nor our hypothetical people in a simulated world can talk meaningfully about ourselves possibly being brains in a vat.

Even if they cannot meaningfully talk about the external reality, which they have no access to, me might still think that it could be meaningful for someone in their position to deny the reality of the world. Upon reflection on what such a denial entails however it becomes apparent that this claim too is meaningless. Consider for example the denial of the reality of a hallucination. The person suffering from the hallucination claims that it is “not real” because no one else is able to perceive it (although it was a real hallucination). However in the case of a simulated world this is not a possible use of the claim of unreality, since other people share the same “unreal” world. Perhaps then they mean that it is like a hologram. However when we deny the reality of a hologram we do not deny that it is a real hologram, only that it is not what it appears to be. To make a denial of reality in this fashion however requires that the speaker be able to say what it really is (it is really a hologram), and once again the people in the simulated world cannot meaningfully refer to their experience of a physical universe as being something else, since they have no experience of what that something else could be.

What is left that a person in such a simulation could meaningfully claim? We might think that even if they couldn’t deny their reality they might be able to meaningfully insist that there were some other aspects of reality, inaccessible to them, that were really the cause of their experiences. In this case the person is not denying the reality of the simulation, simply saying that there may be more to reality than is perceived. However once again we run up against the problem that if this extra reality can’t be observed to have a casual effect then there is no way that person could meaningfully talk about it, and in fact could have no reason to believe that it even exists.

Thus the only meaningful claim we can make along these lines is something to the effect that “there may be more to reality than what we have observed so far, although I can say not what”, and this claim is so empty that we might as well simply not make it at all.

Why then does the claim “we are all really brains in vats” seem meaningful? It is because we are confusing objects that we can talk meaningfully about with objects in some hyper-reality (more real than what we are perceiving), objects about which we have no information whatsoever. Really saying “we are all brains in vats” is just as meaningful as saying “we are all akhaf in uyawer”; since we have no way of knowing what the hyper-reality is like why use the same words for it? Yes we could imagine real brains in vats, but then those people, living in a simulated world would have no way of knowing what our reality was like, and thus could not meaningfully form a hypothesis about it. Of course all bets are off if you let information from one reality leak into the other, but since there is no evidence that happens in our “real” world we should be satisfied that we are not akhaf in uyawer.

About these ads

21 Comments

  1. Here’s the trouble:

    If *people can only talk about being *brains in *vats, not about being brains in vats … then how is it that it seems we can sensibly imagine being brains´ in vats´?

    We are quite capable of entertaining the notion that our consensus reality is an illusion. Even though we are unlikely to agree with them, we can comprehend the propositions posed by, e.g. “The Matrix”. They are not absurd or incomprehensible, in the sense that a colorless green idea is, or a round square.

    We are familiar with notions of unreality: optical or auditory illusions, in which the senses can be reliably fooled; psychedelic hallucinations, in which we perceive patterns in qualia which are not there; suspension of disbelief, in which we become both perceptually and emotionally involved in a story or drama we know is fictional; delusions, in which we form and defend utterly erroneous and unfounded beliefs; mistaken impressions, as where we think we see our friend’s face in a stranger’s. We know (or can know) what it means to have momentary or lasting erroneous beliefs about the nature or contents of reality, and to subsequently discover that these beliefs are erroneous.

    It is true that our present perceptions of reality are all that we have to go on. It is likewise true that we are justified in acting as if we aren’t brains in vats. (But it is likewise true that our simulated *people are justified in acting as if they are real too, according to the *physical laws of their *reality, the simulation.)

    However, not only is it possible that we are brains in vats, but we are capable of comprehending the proposition. Just as I can know what it feels like to put in contact lenses every morning and go from a world of fuzzy visual images to a world of clear visual images, I can likewise reason about what it would be like if suddenly the vat technician hooked up a camera to my brain and I could see the vat I’m in.

    Comment by Frater Plotter — June 13, 2006 @ 2:15 am

  2. “The crucial premise in this argument is that to be able to refer meaningfully to something we must have had some perception of the thing we are referring to.”

    This is simply incorrect. Almost mindbogglingly so. You may as well say “we can’t meaningfully refer to something without a signed note from our Doctor” or “we can’t meaningfully refer to something unless we know how to spell its true name.”

    “Really saying “we are all brains in vats” is just as meaningful as saying “we are all akhaf in uyawer”; since we have no way of knowing what the hyper-reality is like why use the same words for it?”

    Because that’s a distinct claim. We could well be real, human brains in real glass vats, in this world. We might also be these other things you refer to, ether as an alternative or in addition. While the ideas may be related, there is no reason to suppose they are identical.

    –MarkusQ

    Comment by MarkusQ — June 13, 2006 @ 6:11 am

  3. If *people can only talk about being *brains in *vats, not about being brains in vats … then how is it that it seems we can sensibly imagine being brains´ in vats´?

    That’s just the crux of the argument, we can’t speak meaningfully of brains+ in vats+ where those are hyper-reality terms because we have no idea what hyper-reality is like or have any basis for speculation into it. Yes we can imagine that we are normal brains in normal vats, but that is an impossible situation since clearly we are not. Yes we could make an assumption like “what if hyper reality is just like normal reality”, but we have no reason for making that assumption, and thus any talk based on it is quite meaningless.

    Comment by Peter — June 13, 2006 @ 12:00 pm

  4. This is simply incorrect. Almost mindbogglingly so. You may as well say “we can’t meaningfully refer to something without a signed note from our Doctor” or “we can’t meaningfully refer to something unless we know how to spell its true name.”

    Really? Then propose to me an alternate way our word have meaning without our being exposed to their reference in some way. I have admitted earlier that being told about it would count, as would infering something about the word from deduction, but the meaning cannot come from nothing. To deny that we must have in someway perceived that which we wish to refer to is to imply that somehow magically meaning is attached to our words. For example if we wish to discuss Xyltrl you might say “what is Xyltrl?”; if I have never seen or heard about Xyltrl how could I know about it, and how could I tell you about it? Thus without some pervious perception that informs be about Xyltrl my use of it is meaningless.

    We could well be real, human brains in real glass vats, in this world.

    No, we couldn’t, since we have arms and legs. You are saying that what we really think is a physical reality, P1, is really a simulation P2. However to make a hypothesis about what is going on in P1 we must have some way of knowing about it (some channel of information), but if we are really in P2 we have no way of knowing about P1, and thus can’t say meaningfully that we are brains and vats in P1. Yes we can construct a meaningful “what-if” situation for hypothetical people, but we can’t assert that is the situation we are in; to do so would be to accept a hypothesis that contains statements about objects which we can have no knowledge of. If I told you that in the Andromeda galaxy there were Xythyl wearing githli we would both agree that is nonsense. If I tried to define those terms then we would both agree that I was talking about a fiction and not a real state of affairs. Likewise we can accept the brains in a vat theory as a fiction but not as a claim about the world.

    Comment by Peter — June 13, 2006 @ 12:12 pm

  5. Yes we could make an assumption like “what if hyper reality is just like normal reality”, but we have no reason for making that assumption, and thus any talk based on it is quite meaningless.

    You’re making this needlessly complicated! It’s not meaningless to say, “what if we were brains in vats,” because based on what we know about brains* and vats*, there’s no reason to suppose that there isn’t a hyper-reality with brains′ that are almost identical to brains*, except they aren’t virtual. These brains′ aren’t unintelligible to us, for the very reason that I’m defining the brains′ to be nearly identical to brains*. On the other hand, there’s no reason that there couldn’t be a hyper-reality with brains′ that are unintelligible to us. Neither is more likely than the other, and neither is very helpful for living our lives.

    So while we could say it, it isn’t worth saying, “What if we’re brains in vats,” because there are a lot of possible What Ifs, but those without actionable implications and a high probablity of being true aren’t worth our time to talk about.

    Comment by Carl — June 13, 2006 @ 6:07 pm

  6. I guess that is another way to look at it. (“What if there are invisible and insubstantial unicorns everywhere?”) I will have more on this tomorrow, so save the really insightful comments until then, when you can respond to a more fully developed account of some of these issues (dealing with when it makes sense to say we know something).

    Comment by Peter — June 13, 2006 @ 6:27 pm

  7. A point of fact: Throughout human history, a number of people have proposed seriously the notion that we are brains in vats, or rather, that we are abstract minds (or spirits) subjected to delusions of living in a material world.

    The Hindu/Buddhist notion of maya comes to mind. So do certain Gnostic notions that our bodies and the phenomenal world are merely delusions spawned by an evil god. I do not assert here that these are true or even self-consistent; I do assert that they are understandable and not incoherent in the sense of “we are all akhaf in uyawer.”

    Descartes also specifically describes this idea in his “evil demon” argument. The fact that he is able to describe it shows that it is quite comprehensible and not incoherent. Descartes’ refutation of the evil demon is not convincing, since it relies on an argument for the existence of a benign God scarcely discernible from Anselm’s, and equally subject to the “most perfect tropical island” refutation: the fact that we are able to conceive of a perfect God does not imply that one exists.

    The notion that the “brains in vats” claim is incoherent seems to me to be an attempted end-run around refuting it. Instead of actually refuting the possibility that we are brains in vats, the author simply declares it incoherent. But history shows that it crops up with some frequency in religion, mysticism, and fiction as well as in philosophy. Quite an accomplishment for an incoherent notion!

    (For one recent example of fiction that deals with this idea in one of its more mystical incarnations, see Grant Morrison’s “The Invisibles”.)

    Comment by Frater Plotter — June 13, 2006 @ 8:26 pm

  8. Ok, I’ll hold off after one last note: Remember that even as implausible and unhelpful as it can be, things can be both unproveable and true. This is the conclusion of Gödel’s theorem.

    OK, that’s all.

    Comment by Carl — June 13, 2006 @ 8:35 pm

  9. Actually that’s not quite what the Godel’s proof showed; it showed that given any axiom system there are some sentences that cannot be proved nor disproved. Effectively they are neither true nor false, since one can either consistently arrest or deny them.

    Comment by Peter — June 13, 2006 @ 9:03 pm

  10. Oh religions are frequently incoherent. In fact if someone tells me that a belief is religious that is all the more reason to question it. Tomorrow I will show why the kind of refutation you want is impossible, and yet we would still have good grounds for rejecting it and claims like it.

    Comment by Peter — June 13, 2006 @ 9:12 pm

  11. Wikipedia on Gödel’s Incompleteness Theorem,

    For any consistent formal theory that proves basic arithmetical truths, it is possible to construct an arithmetical statement that is true 1 but not provable in the theory. That is, any consistent theory of a certain expressive strength is incomplete.

    1 The word “true” here is being used disquotationally; that is, the statement “GT is true” means the same thing as GT itself. Thus a formalist might reinterpret the claim

    for every theory T satisfying the hypotheses, if T is consistent, then GT is true

    to mean

    for every theory T satisfying the hypotheses, it is a theorem of Peano Arithmetic that Con(T)→GT

    where Con(T) is the natural formalization of the claim “T is consistent”, and GT is the Gödel sentence for T.

    Comment by Carl — June 13, 2006 @ 10:59 pm

  12. Yes it is possible that the statement is true, but it is also possible that the statement is false. The conclusion of the proof is that it is undecidable, not one or the other. Perhaps Wikipedia should be edited to make this more clear. I would refer you to the book Gödel’s Proof by Ernest Nagel and James R Newman, it is very brief; or if you hate math Gödel Escher Bach by Douglas Hofstander. Often we say that they are true statements simply for convenience, the reality is, as proven by Gödel, that there is no way of knowing. If we can conclude that the statement in question is true, we do so with the aid of axioms not in the system, but then according to the proof those axioms too will yield undecidable statements. You can assume that the sentence is either true or false and build a consistent mathematical theory, both the statement and its negation are equally “true”. (There is a demonstration of this on page 222 of GEB.)

    Comment by Peter — June 13, 2006 @ 11:18 pm

  13. Note the reason your quote says that the statement GT is true is because when applied to the formalization of logic commonly used the Gödel sentence can be “proved” with another intuitive axiom. This is not true for all Gödel sentences, nor is it true for the Gödel sentence generated by applying the theorem to formal logic + the axiom used to prove GT.

    Comment by Peter — June 13, 2006 @ 11:22 pm

  14. This is simply incorrect. Almost mindbogglingly so. You may as well say “we can’t meaningfully refer to something without a signed note from our Doctor” or “we can’t meaningfully refer to something unless we know how to spell its true name.”

    Really? Then propose to me an alternate way our word have meaning without our being exposed to their reference in some way. I have admitted earlier that being told about it would count, as would infering something about the word from deduction, but the meaning cannot come from nothing.

    Your position is self contradictory. You claim that meaning “cannot come from nothing”–how could that possibly be a meaningful statement? You may have been told about it, but the person who told you had to get it from somewhere. And eventually, you have to have had someone who “directly perceived” that all meaningful statements must be rooted in perception. This requires the ability to directly perceive a universal truth, which is a rather dubious basis for a logical argument. But even if we were to accept that someone once had this ability, noticed the fact that all meaning had to be rooted in direct perception, and told his friend about it, and we accept that the observation is true, how on Earth can you expect the friend to understand what he’s being told?

    By your reasoning, the big-bang couldn’t have happened because (if it had) meaning, like everything else, would have had to “come from nothing.”

    (Your next paragraph likewise demonstrates that language could never have developed).

    You ask for an “alternate way our word[sic] have meaning,” The problem is, you are expecting to find meaning in the words, which is clearly counterfactual (think of foreign languages). Meaning isn’t a property words intrinsically “have”, it’s something we ascribe to them. Like Humpty Dumpty, we can give them any meaning we care to. Like beauty, it is in the eye of the beholder.

    We could well be real, human brains in real glass vats, in this world.

    No, we couldn’t, since we have arms and legs. You are saying that what we really think is a physical reality, P1, is really a simulation P2. However to make a hypothesis about what is going on in P1 we must have some way of knowing about it (some channel of information), but if we are really in P2 we have no way of knowing about P1, and thus can’t say meaningfully that we are brains and vats in P1. Yes we can construct a meaningful “what-if” situation for hypothetical people, but we can’t assert that is the situation we are in; to do so would be to accept a hypothesis that contains statements about objects which we can have no knowledge of.

    Oh, nonsense. If I’m going to doubt the reality of the entire universe, why should I balk at doubting my arms and legs? For all I know, I “died” last week and Alcor froze my head. A hundred years later they thawed me out and put me in this vat with a simulator that matches my memories (doesn’t have to match reality, only what I remember) and they’re running tests on me now (including leading me to this page) to see how I’ll handle the shock of learning what happened to me.

    I just asserted a meaningful “what-if” and asserted that that was the situation I am in, without even breaking a sweat (or making bubbles, as the case may be). To address a point made by another poster, this is even an actionable claim, in that there are certain things I ought to do / ought not do, based on my belief in it. It is, in fact, testable, though probably not falsifiable.

    — MarkusQ

    Comment by MarkusQ — June 13, 2006 @ 11:37 pm

  15. 1: The statement meaning cannot come from nothing is based on reasoning from: physical laws and experience. It does not come from nowhere. Where do physical laws come from: observation. Where do observations come from: they simply exist, but observations do not themselves have meaning, they require interpretation. Thus meaning has been derived from that without meaning, but not from nothing.
    2: the big bang doesn’t have meaning, it simply exists. Meaning is not the same as existence.
    3: Read the second quote more carefully. It doesn’t involve claims about what you can doubt, merely what you can know. You can doubt anything you wish, but it is only reasonable to so when you can say what can exist instead. Since you have know knowledge of hyper-reality you cannot say what might or might not exist in it.

    Comment by Peter — June 13, 2006 @ 11:50 pm

  16. Carl- finally, even if you don’t believe me about what the Gödel theorem means it is relatively unimportant to our discussion (although interesting in its own right), since knowledge about the physical world and physical laws / reality isn’t axiom based, and thus the Gödel theorem can’t be applied to it. (it is falsification based, you might even say that the word true when applied in physics and math has two different meanings).
    Secondly the Gödel theorem is one of the three “forbidden” theories. It is considered unwise to use them in a philosophical paper unless you can get someone with a PhD in math or physics to validate your claims. (The three forbidden theorems are Gödel’s theorem, special relativity, quantum mechanics).

    Comment by Peter — June 13, 2006 @ 11:56 pm

  17. Yeah, I’m well aware that Gödel is the second refuge of crackpots, right behind quantum mechanics. I using that to say that there exist true things which are not proveable, not to give a particular example of such a thing.

    In any event, if you’re asserting that it’s impossible for something to be “true” without it also being proveable, then you and I must have different definitions of true.

    In fact, I find that assertation quite ridiculous and suspect that to the contrary, the vast majority of true things are either impossible to prove or so impractical to prove that they may as well be impossible (for strict values of “prove”). This may just be a symptom of my latent platonism.

    Comment by Carl — June 14, 2006 @ 1:14 am

  18. Well for example I think it is true that I have at least impressions of having arms and legs, but I cannot prove it from an axioms system, nor do I feel that I need to. This is what I meant by saying that truth in a practical context is probably different than truth in a mathematical context; there are different standards for when we know something to be true.

    Comment by Peter — June 14, 2006 @ 7:01 am

  19. 2: the big bang doesn’t have meaning, it simply exists. Meaning is not the same as existence.

    You are being obtuse. I brought in the big bang to forestall a “turtles all the way down” epistemology; if the universe has only been around a finite amount of time and, as we agree, there was no “meaning” at its inception, and some things have “meaning” now, then there must be some way of creating “meaning” out of nothing–albeit possibly with an intermediate larval stage of “observation” as you claim, though possibly not.

    1. The statement meaning cannot come from nothing is based on reasoning from: physical laws and experience. It does not come from nowhere. Where do physical laws come from: observation. Where do observations come from: they simply exist, but observations do not themselves have meaning, they require interpretation. Thus meaning has been derived from that without meaning, but not from nothing.

    So now instead of a “Universal Principle Detector” you need a “Valid Interpretation Sniffer”–otherwise I can look at a flower, say, “Ah, I must be a xgfdy floating in a fdess!” and you will be forced to conclude that it is meaningful. (I’ll let you ask me what in the hell I’m talking about, but I’ll just say “Look at this flower!” and expect you to see the light. Or the fdess, as the case may be.)

    If you want to throw this out, please tell me how you can infallibly distinguish “valid” interpretations of observations that produce new meaning from invalid ones. If you can’t, then requiring observation for the production of meaning is pointless. If you can, than we have found the royal road to Truth; we can set up a production line to generate and test new meanings wholesale without all the pitfalls of philosophy, math, and home economics.

    3: Read the second quote more carefully. It doesn’t involve claims about what you can doubt, merely what you can know. You can doubt anything you wish, but it is only reasonable to so when you can say what can exist instead. Since you have know knowledge of hyper-reality you cannot say what might or might not exist in it.

    This point makes no sense to me.

    * What second quote?
    * Why is it only reasonable to doubt something if you can say what can exist instead? This seems like an arbitrary (and pointless) restriction–and one, I might add, that you consistently violate. You not only doubt things (e.g. specific claims of hyper reality) you assert that your doubt is justified by the very fact that we can’t “say what can exist instead.”

    I note that you seem to have totally ignored my counter example to (and thus disproof of) your claim that I could not be a real brain floating in a real vat at this very moment, based on your (unfounded) assumptions about my limb count. Remember, I don’t have to prove that it is true (which would be hard) or false (which would be harder), but merely that it is meaningful, which I have done. Your claim that “being a brain in a vat” perforce involves some hyper-reality can be broken by a single explication of how one might become a brain in a vat within this reality. If technology could (even in principle) accomplish something without resorting to an alternative plane of existence, then philosophical theories of why the alternate plane is a logical necessity must be flawed.

    –MarkusQ

    Comment by MarkusQ — June 14, 2006 @ 7:29 am

  20. I was kind of hoping that the post would deal more with the “brain in the vat” thought experiment in Dennett’s “Where am I?” paper, where the brain, which is in a vat, controls the body by a kind of remote control. I think that an argument against this version of the brain in a vat would be pretty interesting, since it also work against a “soul in a vat” version of religion.

    Comment by Jeff G — July 2, 2006 @ 10:53 am

  21. Dennett’s piece is not really designed to raise the serious possibility that you are in remote control of your body. That could be easily defeated on simple epistemological grounds and the best explanation. The doubt raised as to whether there may be an identical consciousness to yours is the more interesting question, but it wouldn’t cause us to doubt that we are conscious or physically part of the world. (See the commentary in “The Mind’s I”)

    Comment by Peter — July 2, 2006 @ 2:57 pm


RSS feed for comments on this post.

The Rubric Theme. Create a free website or blog at WordPress.com.

Follow

Get every new post delivered to your Inbox.

Join 162 other followers

%d bloggers like this: