The Brain in a Vat Argument | Hacker News Reader

The Brain in a Vat Argument

iep.utm.edu

28 pointsinfinity53 comments

Wow, what a tangle of verbosity. Yes, you could be a brain in a vat. Or you could be a brain on a biped. Obviously you cannot distinguish between these two possibilities merely by thinking about them, because your thought processes are equally compatible with either scenario (by construct, by the premise of the question). A simulation good enough to be indistinguishable from reality, cannot be distinguished from reality -- that's the definition.

The fun thing about this quack philosophy is that's it's falsifiable, that in a couple decades, we'll actually have brains in vats, and kids will play games where they shuffle a kid's brain around and have him guess if he's in a vat or not. And with the higher-end iVats, they'll never be able to tell. And they'll totally ridicule the tenured philosophy faculty, they'll be so mean, what the hell have you been doing with your grant money?

The brain in a vat that you predict in a couple of decades could be a brain in a vat that thinks they have put a brain in a vat. This type of assumption is _not_ falsifiable (like $deity existence and so on) therefore open to speculation, based on each one their own heuristics and belief, but not rock-solid science.

I personally believe that the brain in a vat doesn't pass Occam's razor... but I have no proof.

Occams Razor is a good general rule, but you're right, not proof.

The related "This is all a simulation" argument is interesting, especially as computer power increases. If we ever get to the stage where we can simulate even a simpler sentience in a far smaller universe at a fraction of real speed then how do we know that's not exactly what's happening with us... ?

if you believe that we can reach this stage, and that we will create simulations when we do, it seems wildly implausible that we are _not_ in a simulation - because that would imply we'd be the first people anywhere, in any universe, to do such a thing.

Well yes, as a thought exercise I started idly wondering the other day how, if/when we reached the stage of development ourselves, we might go about making contact with the next layer up...

Probably make a decent sci-fi short story. Now if only I had the talent or patience for writing...

http://qntm.org/responsibility is an interesting go at it, and where I first learned of the concept. His other fiction is pretty fun, as well.

That's really quite cool. Thanks for that :)

--edit-- My own thoughts were along the lines of - humans decide we're probably in a simulation, and try to figure out how we could attempt communication with the simulators. To do so we decide to build our own simulation, then run it at double speed and watch how they try to communicate with us. Then recurse...

If you ever read a story like that, you'll know someone with the same idea but more motivation than me wrote it :)

Very nice story indeed :D In the story, the characters talk about nested levels of simulation and how slightly disheartening is to know that they are not in the top layer, therefore not "real".

Afai(!)k, there might as well not be a top layer but infinite nesting -- or, even more charmingly, endless recursion, each simulation containing not another layer, but another iteration of itself. This at least would eliminate the embarrassment of being "less real" than the above layer.

Already exists, well at least the movie does - http://www.imdb.com/title/tt0139809/

edited: initially linked to the wrong movie

I am not sure about the implausibility you mention.

"There are worlds who are simulations" !== "Most / all worlds are simulations". The second assumption is possible but by no means implied by the first. If we could create such complex simulations, it would prove only that it's possible that we live in one.

On the bottom end of the stack, for such thought exercises ordered by descending charm, it could be that we live in a simulation created by someone / something that exists in the real reality, and we aren't even in a sophisticated pinnacle of technology achievement like a supercomputer, but in the real reality equivalent of a cheap videogame given away as a real reality equivalent of swag: "Universe 2000, Acme Inc special edition".

Ouch. What a depressing thought.

The problem with Occam's Razor is that the simplest explanation for anything is:

      p AND ~p --> TRUE

The evil demon of skepticism is both evil and a demon.

That's modern philosophy for you. It's the guys who believe they're excluded from science and regurgitate unscientific discussions ad nauseam.

These complaints are ridiculous. Of course philosophy is going to deal primarily with philosophical discussions, and not with scientific ones, that's the job of science!

If you're trying to claim that only scientific ideas are worth entertaining, then the problem is that that claim itself is not scientific.

The commenter you're responding to does not even understand the problem if he thinks science can solve it, as others have pointed out.

Have a look at some articles on philosophy of science. You might not like them, but contemporary philosophy is very much engaged with scientific practice: http://plato.stanford.edu/contents.html.

Or, for that matter, look at Hillary Putnam. Notice that he was a rather accomplished mathematician: https://en.wikipedia.org/wiki/Hillary_Putnam#Mathematics_and.... I believe he also had an extensive knowledge of physics, though I don't think he had any discoveries in the field.

(Edited to be a little nicer).

Without "modern philosophy" the idea of falsifiability would not be a major component of the modern science. Much of what is now considered to be standard scientific practice is the product of Karl Popper's philosophical work.

Far from being excluded, philosophers have defined a large chunk of what modern science is.

I think this vastly overstates Popper's influence. First, if you're trying to say that he's responsible for how science is done, I think that's flatly untrue. Popper's role is to describe and explain scientific practice, and understand how it is effective in producing understanding (assuming it does so--and most philosophers of science tend to agree).

If you mean that he's responsible for our understanding of science, that's debatable. Outside of philosophy, Popper is extraordinarily well known, and you can't avoid people speaking as if falsification is all there is to philosophy of science. Within philosophy, he is respected, but I don't think his particular views are particularly popular.

It's going to suck for all those scientists when they realise their "evidence" is just false data fed to their brains in vats, though.

(Or, more generally, this "quack philosophy" is trying to get us to think about what knowledge is, and you can't understand how science works without thinking about the concept of knowledge).

Yes, you could be a brain in a vat.

That depends on whether by "could be" you mean "I can imagine it", or "I have evidence-based reasons to suppose it is possible."

    Yes, you could be a brain in a vat.

This sentence is false, since you do not have any basis to refer to a vat. This is simply the same as talking about pointers in java script. Either you are not talking about pointers, or you are not talking about java script.

    The fun thing about this quack philosophy is that's it's falsifiable [...]

It is actually not, the moment you set up any experiment, I just assert that the experiment is a elaborate simulation, which does not tell us anything about the "true universe."

>> This sentence is false, since you do not have any basis to refer to a vat.

I know what a vat is here in subjective world, I am able to imagine a situation in which such a thing exists somewhere with something like my perceived centre of consciousness within it being fed false data. Should that (or some other solipsistic scenario as yet unimagined) actually be the case I am fully aware I do not have the terms of reference to accurately describe the scenario as I have been relying on false data for my entire life. However an accurate description of the scenario is entirely unnecessary to the thought experiment itself and dismissing the entire argument based on the statement that we cannot properly describe a vat is to fail to address the root of the argument and attempt escape on a technicality.

Or, in other words, you could be a brain in a vat.

I think:

     Yes, I could be a brain in a vat.

Is a better formulation, epistemologically speaking. Since I could be in meat space servicing your brain in a vat and denying that the world that each of us experiences has much overlap. In general, propositions that lend themselves to solipsism tend to form a better basis for discussions around philosophical skepticism.

If Hillary Putnam practices 'quack philosophy' then it is only because philosophy admits duck typing. And alas, it's duck typing all the way down to the turtle.

If we are brains in vats, then the hypothesis that we are brains in vats is only brains-in-vats-falsifiable, unless one includes the very premise of metaphysical reality that Putnam is suggesting is in need of justification.

Keep in mind that Putnam's paper substantially predates The Matrix and that the search for the premises of scientific rationality goes back at least to Descartes, and that so far the two best solutions are still basically Kant's model of mediated access to reality and claims relying upon a priori that we have direct access to a reality as it exists outside human experience [see Nagal's What is it like to be a Bat?]

http://en.wikipedia.org/wiki/What_Is_it_Like_to_Be_a_Bat%3F

This is the kind of philosophical argument I cannot understand. The question is essentially: "Given that it is impossible to know X, how could you know if X"?

There is nothing to think about. The scenario answers its own question.

Your boiled-down version of the question leads on to the following questions:

* Why is impossible to know X?

* How can we prove that X impossible to know?

* How close to knowing X can we get?

* What else can't we know? Is there a model for these kinds of unknowable?

* Given that we can't individually perceive X, can we infer X by other means? Is it valuable to do so?

I think the even more interesting question would be:

* Might any of these questions be answered in such a way that the answer ought to influence our future behaviour?

If the answer to that particular question is No, then why bother with the others? If the answer seems to be No, it makes sense to delegate bothering about it to a few philosophers in their ivory towers ;-)

The question is more like "We usually think we know X, but it looks like it's impossible to know X; what's going on?" That is, the real question is, what is it about our understanding of "knowledge" that leads to this apparent contradiction?

Ah, ok. That makes a bit more sense.

Although if it is truly impossible to know X, but everything works as if X holds, we might just take the easy way out and assume that X until proven otherwise. But that may be my personal philosophy.

Apologies to David Byrne.

  You may find yourself living as a brain in a vat
  You may find yourself in another biped body
  You may find yourself behind the wheel of a large automobile
  You may find yourself in a beautiful house with a beautiful wife
  You may ask yourself, well, how did I get here?

The question of whether you are living in a simulation is not merely up for grabs. There is a fascinating argument here[0] that approaches the problem probabilistically; the conclusion is that you probably are.

[0]http://people.uncw.edu/guinnc/courses/Spring11/517/Simulatio...

Bostrom's simulation "argument" is really embarrassingly stupid. "The real world might be a simulation" is not only not probably true; it's false by definition, because the real world is definitionally not a simulation.

According to Paul Nahin[1], Richard Feynman had a probabilistic argument that Fermat's Last Theorem was true, which can be dated to before 1963 because Feynman later wrote that Morgan Ward (who died in 1963) told him that "the same argument would show that an equation like x^7 + y^13 = z^11 would be unlikely to have integer solutions..."[2]

[1] Number Crunching, 2011.

[2] The remainder of the quote is, "...but they do, an infinite number of them!"

In my experience (as a student of philosophy at a strong university program), Stanford Encyclopedia of Philosophy is the go-to reference resource for such articles. Here's their version:

http://plato.stanford.edu/entries/skepticism-content-externa...

I think it's a slightly better article in this case. But neither one is really suitable HN. They both presuppose too much background, I think.

In "Pandora's Hope", Bruno Latour goes to great length to attack the artificial subject-object and nature-culture separations that are the legacy of the Cartesian paradigm. It's a highly intriguing book. Here's the foreword, which deals directly with the brain-in-a-vat:

http://www.bruno-latour.fr/sites/default/files/70-DO-YOU-BEL...

Let me tell you why you're here. You're here because you know something. What you know you can't explain, but you feel it. You've felt it your entire life, that there's something wrong with the world. You don't know what it is, but it's there, like a splinter in your mind, driving you mad.

Stop trolling random kids on the internet, Morpheus. Come to bed and

ＹＯＵＮＥＥＤＴＯＷＡＫＥＵＰ

My initial impression of Putnam's argument is that it carefully constructed in order to avoid confronting the crux of the matter. It is reminiscent of how Zeno's 'Achilles and the tortoise' paradox fails to actually consider the time when Achilles catches up with the tortoise.

This is a modern updating of Decartes's Demon concept from 1641.

http://en.wikipedia.org/wiki/Evil_demon

The trope of a Brain In A Vat is an updating of it, yes. But the linked article is really about Putnam's argument that the thought experiment is somehow defective.

> For there is a good argument to the effect that if metaphysical realism is true, then global skepticism is also true, that is, it is possible that all of our referential beliefs about the world are false.

But they're not going to tell us what the argument is.

This extremely counter-intuitive claim seems to be the heart of their argument. I suspect they're using sloppy language to confuse the difference between knowledge and certainty, but since they never spell it out I don't know.

Defining "know" something to mean "know the probability of something is exactly 1", then yeah, you can't do that. No amount of Bayesian evidence can raise a probability from uncertain to infinitely certain.

So what you need to do, is tear up that useless definition of "know" and replace it with something sensible like "I know X if the evidence for it is very strong."

Then how do you know that the evidence is very strong? You'd have to have strong evidence that the evidence is strong, ad infinitum. That was essentially Plato's definition of knowledge, "justified true belief." Then a guy named Gettier came around and showed that you can have strong evidence supporting a belief, and still be wrong. Then a guy named Nozick[1] came along and provided his own definition of knowledge, and that brings us up to the early 1980's.

There's an entire branch of philosophy dedicated to figuring out exactly what "knowledge" is, Epistemology. The "brain in the vat" problem has be considered in context[2].

[1]http://www.iep.utm.edu/nozick/#H3 [2]http://plato.stanford.edu/entries/epistemology/

http://en.wikipedia.org/wiki/Solomonoff's_theory_of_inductiv...

Crudely, you evolve a simplicity prior over every possible universe-predicting theory according to Bayesian updates on input evidence.

If I were a brain in a vat, why would the truth or falsehood of my beliefs matter? I can't interact with the world from the vat.

My thoughts exactly.

If I am a brain in a vat, who cares whether I say "I am a brain in a vat", I'm a brain in a vat and my opinions on the matter are irrelevant.

What does it mean for something to "matter" anyway? Does interacting with the world mean that your beliefs now matter? What if your brain-in-a-vat was hooked up to a computer that would encode your synapses on paper on the other side of the world? Does that "matter"?

Not trolling, just trying to make sense of what you mean by your sentence.

In related news, the universe is a hologram - http://news.ycombinator.com/item?id=6883611

That's not what that means.

"The universe is a hologram" means that the universe's physics is implemented with a data store that has fewer dimensions than what the universe offers its contents. Which is interesting, but orthogonal to it being a "simulation".

Holographic-looking phenomena already recognized by science include black holes: The maximum entropy -- the same as the maximum information content -- of an area of space is achieved when that space is a black hole. And the entropy of a black hole is proportional to its surface area. Fascinating. Not really "related" news though. Unless you want to go for poetry, but poetry mixed with cosmology and quantum physics is a great recipe for misunderstanding physics.

On a somewhat similar note, there is the simulation argument:

http://www.simulation-argument.com/

In my case, the vat contains coffee. ;-)

Yes, this could be the case. Or maybe some vats contain the outside-of-the-vat-world analogue of LSD or other psychotropic substances. Which leads me to the following questions:

If I am indeed a brain in a vat, how far can I trust my own thought processes (like a logical argumentation)?

And taking this further, could it be that I am only a part of a brain in a vat where the other part has already been replaced by computer controlled cyborg implants changing the way information is processed?

The impression that an argument is logically correct and plausible may be a phantom signal resulting from input into a specialized area in my remaining brain.

Even the thought that I had a thought is then maybe wrong. How would it feel to have the convincing impression that I am having very deep thoughts right now without ever been able to perceive one single aspect of these thoughts?

It doesn't matter how the idea comes to us; we can still argue and come up with truth.

http://www.hist-analytic.com/MooreExternalWorld.pdf

Reply on news.ycombinator.com