Direct Realism and the Brain-in-a-Vat Argument
Here, I explain how direct realists can avoid the brain-in-a-vat argument for skepticism.*
1. The BIV Argument
Imagine that there’s a brain being kept alive in a vat of nutrients, connected to a powerful computer capable of giving the brain any desired pattern of stimulation. The computer is programmed to give the brain a simulation of life in the 21st century. The computer can read the outputs from the brain so as to figure out what the brain is trying to do in the virtual world and modify the pattern of stimulation accordingly, so everything seems to the brain just the way it does to normal people living in 21st century society. Q: How do you know that you’re not a brain in a vat (BIV) right now?
Philosophical skeptics claim that this shows that we don’t know anything about the world around us. Note: The interesting version of this view is not that we lack absolute certainty in our external-world beliefs. That’s uninteresting, in my view, because that’s basically trivially true. The interesting version of the view is that we lack justification for our external world beliefs – i.e., we don’t even have good reason to think that the world is the way it appears.
The skeptic’s argument:
1. If you have justification for believing any (contingent) claims about the external world, then you must have justification for thinking that you’re not a BIV.
2. You don’t have justification for thinking that you’re not a BIV.
3. Therefore, you don’t have justification for any (contingent) claims about the external world.
2. The Anti-Closure Objection
Premise (1) is an application of the Closure Principle for justification, the principle that if P is justified, and P entails Q, then Q is justified. (I.e., the set of justified propositions is closed under entailment.)
Some philosophers reject the closure principle. Note: The issue is not whether, if P is true and P entails Q, then Q must be true. Obviously, truth satisfies closure – that’s trivial. But “justification” is different from “truth”, so it’s conceivable that justification might not satisfy closure even though truth obviously does.
An example from Fred Dretske: You see some animals at the zoo, which look like small horses with black-and-white stripes. They are in a pen labelled “Zebras”. Stipulate that everything else about this situation is normal. Do you know what animals are in that pen?
Yes, you know (and are justified in believing) that they are Zebras. But now consider: in this scenario, could you know that the animals were not cleverly disguised mules that were painted black and white by the zoo authorities to trick visitors into thinking that they were zebras?
Dretske thinks the answer is no, because the evidence you had for thinking them zebras doesn’t count against their being cleverly disguised mules. E.g., the fact that they are horse-shaped, black-and-white striped, and in the zebra pen surely counts as strong evidence for their being zebras. Yet none of that is evidence against their being cleverly disguised mules, since all of that would be true of cleverly disguised mules too!
Indeed, as you can see from the photos I’ve included here, painted mules are indistinguishable from zebras to all but the most seasoned zoologists. Believe it or not, one of the pictured animals is actually not a zebra at all!
This is a counter-example to the closure principle: you have justification to think that the animals are zebras, being zebras entails not being mules (cleverly disguised or otherwise), yet you don’t have justification to think that they aren’t cleverly disguised mules, since you have no evidence that they’re not cleverly disguised mules.
So, according to Dretske, the skeptic’s premise (1) is false. So we can know stuff about the external world without knowing that we’re not BIV’s.
3. A Defense that Leads to Question-Begging
Dretske’s example seems to show that justification isn’t closed under entailment, since you can have justification for P, where P entails Q, yet lack justification for Q.
But wait, that’s not exactly what the example shows. What Dretske actually says about the example (and what is true) is that the evidence for their being zebras is not itself evidence against their being disguised mules. But we can’t conclude that there isn’t any evidence against their being disguised mules, since something else might be evidence against their being disguised mules.
So here’s a way of defending Closure for Justification: one could say that the evidence for P need not directly justify Q. Rather, when P is justified and P entails Q, P itself justifies Q. This is the most natural explanation of why Closure would be true, particularly after we’ve ruled out the idea (per Dretske) that the evidence for P has to be evidence for Q.
So that’s what the skeptic could say to defend premise (1).
But now here’s the problem, pointed out by Peter Klein (who was my PhD advisor). If the above is how we defend the Closure Principle, then in order to defend the second premise of the skeptic’s argument, the skeptic is going to have to beg the question. The second premise claims that we lack justification for thinking that we’re not BIV’s. But why think that? Why can’t I just look down at my body, see that I have arms and legs, which a mere BIV doesn’t have, and deduce that I’m not a BIV?
The skeptic just told us that, if I had justification for thinking (e.g.) that I had two hands, the belief that I had two hands could be used to justify whatever it entails, which would include “I’m not a BIV.” That’s why the Closure Principle is true. So the only thing the skeptic could say at this point is that I can’t justify the belief that I'm not a BIV in this way because I'm not justified in thinking I have two hands. But that italicized claim is just (an instance of) the skeptic’s conclusion. So they’d have to appeal to their conclusion to justify their second premise. So the skeptic’s argument essentially begs the question.
4. A Better Defense for the Skeptic
When you think about it, it’s counterintuitive that either Dretske’s or Klein’s response to the skeptic should work.
Here’s an analogy. Let’s say we have some weird phenomena in quantum mechanics, and there are two interpretations of the evidence: the Copenhagen Interpretation (which has some crazy stuff about observers creating reality, reality being indeterminate, etc.) and Bohm’s Interpretation (which is completely sensible). Suppose physicist A asserts the Copenhagen Interp based on the quantum mechanical phenomena that have been observed. Physicist B says, “Not so fast. All that evidence could instead by explained by Bohm’s Interp. If you’re going to accept the Copenhagen theory, you have to explain what's wrong with Bohm’s theory.”
Sounds reasonable. But imagine that Physicist A has been reading Dretske, so he knows that justification isn’t closed under entailment. So he replies: “You’re right, I can’t rule out Bohm’s theory. But that doesn’t matter because justification isn’t closed under entailment. For all I know, Bohm is right; nevertheless, I still know that the Copenhagen Interpretation is right.”
Or suppose that A has been reading Klein, so he knows that if the closure principle is correct, it’s because when P is justified, you can use it to justify anything that P entails. So he announces: “Okay, maybe I have to be able to rule out Bohm’s theory. But if that’s true, then I can rule it out straightforwardly: the Copenhagen theory is correct; therefore, Bohm is wrong. And if you reject this argument, you’re begging the question.”
Intuitively, neither of these replies should work. The Bohmian “skeptic” cannot be dismissed so easily. And this seems, at least arguably, analogous to the BIV skeptic – the BIV skeptic would want to portray himself as doing something like what the Bohmian physicist is doing in the above example.
So let’s try to reformulate the skeptic’s argument with that in mind. I suggest that the skeptic thinks something like this:
1’. Given two competing explanations for some evidence, the evidence justifies one explanation only if you have some strong, independent reason for preferring that explanation over the other.
Note: An “independent” reason must not be based on the hypothesis that's in question.
2’. The BIV Hypothesis and the Real World Hypothesis (that we are perceiving the real world normally) are two competing explanations for our sensory experiences.
3’. We have no strong, independent reason for preferring the Real World Hypothesis over the BIV Hypothesis.
Comment: If you think you have such a reason, try saying what it is. It’s really hard to do.
4’. Therefore, our sensory experiences do not justify the Real World Hypothesis.
From there, they’ll conclude that you lack justification for the RWH, etc. This is the best formulation of the BIV argument, which is true to the skeptic’s intentions and lets the skeptic avoid Dretske’s and Klein’s critiques.
5. The Direct Realist’s Escape
Background definitions:
Epistemological Direct Realism (“DR”): The view that perception gives us foundational (non-inferential) justification for certain beliefs about the external world.
Epistemological Indirect Realism (“IR”): The view that perception gives us only inferential justification for certain beliefs about the external world, based upon our justified beliefs about our own sensory experiences (or ‘ideas’, ‘mental images’, ‘sense data’, etc.)
Historically, IR is the more common view among philosophers, many of whom just assume it. See the likes of Descartes, Locke, Hume, & Russell. However, DR is the common sense view (as is also generally acknowledged).
Now notice that the skeptic’s argument, as we’ve formulated it above, is really only directed against indirect realists. The Indirect Realist thinks that, during normal perception, the basic evidence we acquire is that we have such-and-such sensory experiences. The BIV scenario is a competing explanation for that evidence.
The Direct Realist, however, thinks that, during normal perception, the evidence we acquire is that the external world is such-and-such way. You don’t have to make an inference from propositions about mental states to propositions about physical objects. You’re allowed to start, epistemologically, from external world propositions when you (seemingly) perceive them to be true. And of course, on that account, the BIV hypothesis does not even attempt to account for the evidence at all.
The BIV argument, as formulated in section 4, doesn’t answer the Direct Realist; it just assumes that the realist the skeptic is talking to is an Indirect Realist.
6. Objections
Is this begging the question?
I've appealed to Direct Realism, which directly contradicts skepticism. Is this question-begging?
No, provided that you understand what this response is supposed to do. The above remarks do not, and are not supposed to, positively show to a skeptic that we have knowledge of the external world (which would likely be impossible). Rather, they simply rebut the skeptic’s argument to the contrary. I.e., they show that the skeptic’s argument does not succeed – the skeptic has not shown that we lack knowledge of the external world. If you started from the common sense view, the skeptic hasn’t given you any reason to change your mind.
Is this dogmatic?
Suppose you have a sensory experience that really would be best explained as a hallucination. E.g., after taking a lot of LSD, you seem to see a ghost. Does the Direct Realist view imply that you should believe there really is a ghost? After all, you’re foundationally justified in believing what you seemingly perceive, so you don’t have to refute the hypothesis of a drug-induced hallucination.
In reply, no. The reasonable, moderate version of Direct Realism holds that perceptual experiences give one foundational but defeasible justification for believing external-world propositions. This is to say that you’re justified in believing what seems to you to be the case, without the need for any argument for this, provided that you have no specific grounds for doubting the appearances.
In the LSD/ghost case, you need to consider the hallucination hypothesis because you have specific grounds for suspecting that you’re hallucinating. By contrast, when the skeptic just mentions the BIV possibility, you have no specific grounds for suspecting that you’re a BIV, so you’re allowed to ignore that hypothesis.
[*Based on: “Direct Realism and the Brain-in-a-Vat Argument,” Philosophy & Phenomenological Research 61 (2000): 397-413.]