Thursday, December 8, 2011

Classical Global Skepticism and the EAAN

Global skeptical claims are those that call virtually everything we think we know into question. Two of the most common are the evil genie and the brains-in-vats scenarios. The idea behind such claims is that they introduce situations where our experiences would be exactly the same as they are even though virtually everything we experience is a massive illusion. And if our experiences would be identical, then there's no way to test whether our experiences are veridical or whether the skeptical scenario is true.

Descartes, in establishing his philosophy, proposed a hyperbolic doubt for methodological purposes -- that is, he said, "Let's doubt everything we can and see if there's some level we can't doubt". So he doubted that his senses give him reliable information about the physical world, he doubted that he was awake since it was possible that he was dreaming, etc. When he got to a priori knowledge, such as mathematics, he argued that it's possible that there be an evil genie who manipulates his thought processes whenever he tries to add two numbers together so that he comes up with the wrong answer every time. Of course Descartes wasn't suggesting this as a real possibility, or even that we could doubt such things on a practical level, he was just saying that it's logically possible to doubt it. He ultimately got to the ground level with his cogito: I can't doubt that I'm doubting. Doubting is a form of thought and thought requires a subject who is doing the thinking. As such, I must exist. I think therefore I am (cogito ergo sum).

Now this fails for a very obvious reason. If we're willing to consider the possibility of the evil genie, why can't we just say the genie is also manipulating our thought processes when we try to derive the conclusion of the cogito? Yes it seems as if I can't doubt that I'm doubting, but maybe that's just the evil genie having his way with me. Once the genie is proposed, it applies to everything, including Descartes's argument.

The brains-in-vats scenario is sometimes used in science-fiction. The idea is that we are all disembodied brains being manipulated by scientists or aliens or something to think we are interacting with objects and other people in a physical world. This was used in the Matrix movies, although there they weren't disembodied brains. John Pollock begins his Contemporary Theories of Knowledge with a cute little short story illustrating the problem here. A man finds scientists taking people's brains out of their heads and hooking them up to electrodes. The man is discovered and is told that the people don't know the difference because they're programming them to think that their lives are continuing on without interruption as they had before; they can't tell the difference. The upshot is that, just as the man thinks they're going to do the same thing to him, the scientist laughs and says, "Oh no. We did it to you three months ago." Then they let him go.

Now global skeptical claims are fascinating and they play a huge role in the theory of knowledge. Michael Williams uses them as the main method in establishing his epistemology in his brilliant Problems of Knowledge. But no one takes them seriously as actual possibilities. For whatever reasons, we just don't feel threatened by them. Yes, it's possible that my experiences of the physical world and other people are all illusory, but why should I think so? Simply pointing to the possibility doesn't really make them realistic options. They're just bizarre stories that someone made up. We can certainly use them to further our concept of knowledge, but that's all.

The reason I'm bringing this up is that one of the criticisms given to Alvin Plantinga's Evolutionary Argument against Naturalism (EAAN) is that it is just one more form of global skepticism, and so should be treated the same way. Philosophers already have ways of dealing with such claims, and even if they didn't they wouldn't need them to reject the claim as nothing more than an interesting puzzle.

There's some truth to this: Plantinga's argument is that if naturalism is true, the likelihood that our beliefs would be mostly true is low or inscrutable. Therefore, for any particular belief, regardless of how reasonable it seems, the probability that it be true is low or inscrutable. And of course, naturalism is itself a belief. Therefore, if naturalism is true, belief in naturalism would be irrational. I'll go into more detail about that in future posts, but for now I'll just point out that the unreliability of our cognitive faculties would amount to a form of global skepticism.

There's more to the story though. The problem with traditional global skepticism is that it calls everything into question a priori. Before we have the right (deontologically) to accept any particular belief, we have to show that the belief in question is not subject to the skeptical claim. To do this, however, would require some form of argument -- and that argument would be under the same cloud as the belief it seeks to defend since the skeptical claim would equally apply to it. There's no way out.

However, everyone (except me) accepts that Plantinga is an externalist, and part of the strength of externalist epistemologies is their ability to avoid global skepticism. A belief constitutes knowledge if it is connected in the right way to its object. Thus, if you believe that there is a tree in front of you because there is a tree in front of you, you know it. The global skeptical questions only come into play when we ask second order questions (do you know that you know there is a tree in front of you?). But you do not have to solve that problem before you can know that there is a tree in front of you. Thus externalist epistemologies don't really solve global skepticism so much as they bypass it.

Moreover, Plantinga denies that all beliefs have to be believed for a reason before they constitute knowledge (he calls his epistemology Reidian foundationalism after Thomas Reid); rather, he maintains that some beliefs are properly basic, i.e. they are simply given and we are justified or warranted in accepting them (thus they constitute knowledge) until we have a reason to doubt them. As such, they are not beyond doubt, they can potentially be shown to be false, they are just innocent until proven guilty.

With these two points in mind, Plantinga's EAAN is significantly different from classical global skepticism. First, we do not have to have a reason for a belief if it is properly basic, and such a belief can constitute knowledge even if we don't know that we know it. We are justified, or our beliefs are warranted, up until the point where we have a reason for thinking them to be false. The EAAN provides just such a reason: if naturalism is true, then it is improbable or inscrutable that any given belief would be true. After this, the EAAN has the same effect as the more traditional global skeptical arguments: any reason you can give for a particular belief is itself subject to the EAAN and is therefore not trustworthy. There is no stopping the rot once it's started. Indeed, part of the genius of Plantinga's argument is that it amounts to a global skeptical argument that arises from within externalism. Not to mention the fact that by saying that belief in our cognitive faculties' reliability is warranted until we have some reason to deny it Plantinga is also able to ward off a tu quoque argument being constructed against theism.

Another significant difference is that the other global skeptical claims involve scenarios that are logically possible but that we don't take seriously. Plantinga's, however, involves a scenario that is actually believed by many people, namely naturalism, and even those of us who don't believe it tend to take it seriously (that is, we don't consider it as crazy as the evil genie or brains-in-vats scenarios). It's like if someone came up with an argument that if theism is true, it leads irrevocably to the evil genie scenario. If the argument were sound it would be much more than just an interesting puzzle.

A third difference cuts the other way: traditional global skepticisms posit situations where it is extremely probable, almost certain, that our cognitive faculties are unreliable. Plantinga's EAAN merely argues that it is either less than 50% probable that they are reliable, or if we feel we cannot ascribe any probability, inscrutable. We can certainly modify the traditional scenarios to make them more parallel, but the point is that in their traditional formulation they are stronger than Plantinga's argument.

A fourth difference is that the traditional global skepticisms do not allow for any way out. The brains-in-vats suggestion applies to everybody. But the EAAN allows for a way out, since it only holds if naturalism is true. We can avoid it by simply rejecting naturalism. Since naturalism entails the non-existence of God or any supernatural agency, it follows that in order for us to have knowledge of anything there must be a God or some sort of supernatural agency (although "supernatural" comes with a lot of baggage, so perhaps we could come up with another term that doesn't have as many connotations).

Two final points: first, Plantinga's argument only applies to those who have heard it; the naturalist who hasn't heard the EAAN or a similar one (such as J. R. Lucas's Gödelian Argument against physical determinism) does not have a reason to reject any particular belief. So it's not the case that one has to affirm the existence of God in order to have knowledge. The claim is that there must be some supernatural agency in order for us to have knowledge, not that we have to recognize that there is a supernatural agency. The problem here is very similar to axiological (moral) arguments for the existence of God. The point of these arguments is not that one has to consciously believe in God in order to be a moral person or believe in objective moral truths. Rather the claim is that one is being inconsistent in believing that there are objective moral truths without an objective anchor for them which transcends individual people and cultures. Of course atheists do not deny that murder is immoral, the argument just seeks to show that this is inconsistent with atheism.

Second, when we ask what the supernatural agency in question is, Plantinga immediately points to God. I think instead that the agency is the individual human being. We are supernatural agents, and rational thought is a supernatural process. I would argue further that this ultimately requires God's existence via a less direct route, but that's a post for another day.

(cross-posted at Quodlibeta)


Ron said...

Excellent post. You have obviously read a lot more on epistemology than I have. One thing that I don't understand is your refutation of Descartes. The evil genie can't trick you into thinking that you are thinking when your really not because both your thinking and your existence are prerequisites for being tricked. Thus the cognito.

I agree with you that the EAAN can't be treated like global skepticism because it is specifically targeted at the combination of naturalism and evolution. Take one of those away and you've resolved the problem. Also, naturalism is really a metaphysical philosophical theory so it would be more threatened by the EAAN than any other more basic beliefs or really most beliefs about science or how the world works.

Jim S. said...

The problem with this is that my existence is presented as the conclusion of an argument. If the evil genie is tricking me, then I must exist (there must be someone he's tricking). But the whole point of the evil genie is that we can't trust conclusions of arguments, even deductive (a priori) arguments. All "if ... then" arguments are called into question.

One partial solution is to reject treating one's own existence as the conclusion of an argument. Descartes sets the whole thing up so we can only affirm our own existence via argument and that's simply not why we believe that we exist. Rather, we should treat it as something simply given. We don't have to have a reason for believing that we exist. I think that part of Plantinga's epistemology is right on the money. This doesn't entirely avoid the evil genie scenario (although a lot of epistemologists think the genie can't apply to a priori knowledge at all), but it certainly avoids it at one level.

Ron said...

Ah, I think I'm tracking with you.

Btw, I'm reading Dallas Willard now and am enjoying it. I checked out his book, Knowing Christ Today" from a recommendation on your blog. I love how fluidly he writes. Whether the book is truly good or not however will depend upon whether he can deliver on his claim that Christianity is knowledge rather than just faith or commitment. I think that is a tough bar to meet for anyone so I'll see how that goes.

Timothy Mills said...

The best modern psychological research basically confirms the important parts of the EAAN: our reasoning is fallible, our perceptions are subject to physiological, neurological, and psychological errors. In many ways, our brains work just the same way as our evolutionary cousins - other apes, other mammals, etc.

But just because a machine is fallible does not make it useless. A Bayesian approach to reasoning and belief is still valid, even once we acknowledge that all of the bases of our beliefs are imperfect.

Actually, it seems to me that something very like Plantinga's concept of "warranted belief" can be used to rescue naturalism from the problem he thinks he has discovered. If it is the case that evolution would favour organisms whose beliefs are true over those whose beliefs are false, then we are warranted in accepting our beliefs even knowing that they are constrained by our evolutionary history.

Jim S. said...

Well, Timothy, people have always known that our cognitive abilities are fallible and can be impacted by physiological issues. The first person who ever got drunk figured that out. The debate over fallibilism in epistemology is not about whether we are fallible but whether knowledge is fallible. If it is, then it would be possible to know something that is false, which seems to be a contradiction. But if we hold it up to that standard then we can't be certain of anything.

At any rate, you're missing the challenge of the EAAN. Classical global skepticism played on the fact that our knowledge is not 100%. The EAAN claims that if naturalism is true any particular belief one selects has a 50% chance or less of being correct. In other words, you have a defeater for that belief, a reason for withholding belief in that proposition. You can't defend that belief by giving an argument for it, because the belief that the argument is successful itself has only a 50% chance or less of being true. In other other words, it is irrational to believe anything. It's not merely a matter of being fallible but of being unreliable. Of course the argument may fail, but that's what it claims to show. In order to avoid it one has to reject naturalism.

Timothy Mills said...

Jim, thanks for the clarifications.

I suppose this is why I'm not a philosopher. My main interest is in whether my beliefs (or people's beliefs in general) are fallible - and to what extent, and in what circumstances. The question of whether "knowledge" is fallible seems to balance rather precariously on the definition of the term "knowledge". I recognize this is not a trivial question, but I feel that some intuitive correspondence theory version is probably good enough for most discussions.

Reading over the EAAN link you gave, it seems to me that part of the disagreement here may be a matter of definition. Plantinga makes much of the fact that beliefs may be causally unconnected to behaviour, and thus invisible to natural selection. Fine, but then they are less interesting. I know that sounds like a dodge. On the other hand, it seems strange that Plantinga is talking about "reliable" and "unreliable" beliefs, while explicitly disavowing the necessity of a causal connection between beliefs and behaviours. Surely a "reliable" belief would be one that, if it were used to drive behaviour, could be generally relied on to produce a desired result.

I think I'd want to see a good solid definition of "knowledge" or "belief" before accepting his criticism that beliefs may correspond only syntactically rather than semantically to the truth. What is the important distinction between these two? What makes a proposition syntactically "true" but semantically "false"? (I'm not trying to be cheeky here - as a linguist, I recognize that there may be an important distinction here, but I don't know what the philosophical definitions are that Plantinga is using.)

I appreciate his calculation of probabilities, but I just don't find his estimates of the probabilities to be particularly compelling (as he seems to anticipate).

Finally, setting aside the (to me uninteresting) causally unconnected beliefs, I think my comment about evolution producing "warranted beliefs" may still apply. Of course, knowing how evolution works, we would only be warranted to provisionally accept any particular belief. But I would submit that, knowing human fallibility, the same could be said of warranted Christian belief.

Ian said...

This is where I don't get philosophers. The whole argument is based on an "if x, then y" construction (as best I understand it). But it only takes a cursory examination of the argument to realise Plantinga's "x" hinges on a failure to understand basic biology.

What I fail to grasp is why anyone cares? Plantinga erects and defeats a strawman. His argument may be interesting, but since the caricature he defeats is based only on his misunderstanding or misrepresentation of basic biology.

While I understand that the structure of the argument may be interesting to philosophers, the conclusions lack any validity. So it baffles me that people continue to discuss the implications of the conclusions.

Jim S. said...

Ian, Plantinga's claim is philosophical, not biological. Even the most naturalized of naturalized philosophers will still insist that what they're doing is philosophy, not science. You can certainly point out that, insofar as they are relying on science, that the science has to be valid. But then I don't see what basic biological facts Plantinga's philosophy contradicts.

At any rate, my point in this post is not to defend his argument but simply to show the similarities and differences between it and traditional global skepticism.

dara ghaznavi said...

Hi, very nice analysis.
the problem I find with Plantinga's argument is its kind of circularity.

suppose that I am a naturalist. Now why I should believe that I have a defeater for R? Isn't it because of accepting the first premise of EAAN, i.e. "P(R/N+E) is low"? Of course it is. So in order to accept the second premise, I, the naturalist, must BELIEVE that the first premise (I will call it Q, hereafter) is true. Q is one of my beliefs which is produced by my cognitive faculties, and if I have a defeater for R, then I will have a defeater not only for N+E, but also for Q; and this makes the argument problematic. Plantinga supervenes the truth of Q over N+E, though he does not have any good reason to do so while the cognitive reliability is the subject of discussion. We can call this error "non-reasonable truth supervention".
If I have enough evidence for natural selection, then I can arbitrarily comprise "evolutionary argument against professor Plantinga":

1) N&E
2) Anyone who accepts (believes) "P(R/ N&E) is low" and sees that N&E are true has a defeater for R.
3) Anyone who has a defeater for R has a defeater for any other belief she thinks she has, including "P(R/ N&E) is low" itself.
4) If one who accepts "P(R/ N&E) is low", thereby acquires a defeater for "P(R/ N&E) is low" , "P(R/ N&E) is low" is self-defeating and can’t rationally be accepted. Conclusion: "P(R/ N&E) is low" can’t rationally be accepted.
What if naturalism will be proved by evidence? Then how would we act? deny it, for the sake of an analytic argument?
the problem is with the form of argument which ultimately leads to self referring propositions like
I am truthful or I am liar.