Thursday, December 08, 2011

Classical Global Skepticism and the EAAN

Update (Sep. 12, 2015): I'm temporarily taking this post offline -- like for a year or so -- because it inspired me to write a more detailed article that is being published in an academic journal. Even though a blogpost doesn't (or at least shouldn't) count as a prior publication of something, and even though the article and blogpost are only similar in very broad strokes, I'd like to avoid any appearance of impropriety.

Discuss this post at the Quodlibeta Forum

17 comments:

Andrew Brew said...

Jim, I don't think your third paragraph holds up, does it? The only conclusion from Descarte's cogito is the sum - I exist. If you introduce an evil genie to manipulate my thoughts, whose thoughts is he manipulating again? Mine, so I must exist. QED.

Or am I overlooking something obvious?

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.

Andrew Brew said...

So, if you accept the possibility of the evil genie, you must suspect the validity of all arguments. Even then, though, you must accept the conclusion of this particular argument, even if not for the argument's sake.

Crude said...

G. Kyle,

Any ideas? The article is in the latest Analysis and available at his blog.

I've seen that sort of move before. There's a lot of problems with it, but here's one criticism I make that I rarely see made - and I personally think it's the most devastating.

The EAAN isn't an argument against evolution, but an argument against the conjunction of evolution and naturalism. And insofar as someone makes a move like Law's and insists that evolution is very likely to produce beliefs that not only benefit survival but are also true, he's loading up evolution with enough teleology and direction to make it increasingly inhospitable to the naturalist, and increasingly hospitable to the theist.

Crude said...

G,

Adding "conceptual constraints" would further strengthen the case for teleology in evolution (at least from an A-T perspective...but not necessarily a mechanical one).

Actually, I think it would go through even in a mechanical one. I think the A-T and non-mechanistic perspectives are powerful, but I also think the arguments that run with a mechanistic assumption do a lot better than many people think.

People often forget that evolution isn't just some thing that happens out there in nature. It's also a process used by minds for various ends. If evolution is directed towards the production of true beliefs (such that if there are minds then there are very good odds that most of their beliefs are true), you have an evolutionary process that is pretty painful to absorb into what passes for most naturalisms.

Jim S. said...

I haven't read Law's article but Plantinga makes a distinction between indicative representations and depictive representations. Indicative representations indicate something about the world: a thermometer indicates the temperature for instance. These are directly connected with behavior so evolution would certainly "see" them and select them according to whether they promote survival or not.

Depictive representations depict the world as being a certain way, they have content, and can be true or false. Indicative representations do not have content and are not true or false. Beliefs are inherently depictive, and it is difficult to see how evolution can see them. In fact, the problem here is, as Dretske points out, essentially the mind-body problem. If a philosopher wants to define beliefs in such a way that they include indicative representations Plantinga simply grants them the right and says that what he's talking about are those representations that have content and can be true or false.

Of course the common-sense, folk-psychology view is that the content of belief does influence behavior. If it is difficult to accommodate this within naturalism, then we should reject naturalism.

Staircaseghost said...

Yep, there's the problem.

Well, one of them anyway.

"My philosopher's story based on armchair speculation and folk psychology appears to conflict with this well-established scientific theory. Better throw out the science!"

Alvin question-beggingly insists on a peculiar notion of true beliefs in which there is an unbridgeable dichotomy between indicative and depictive representations. But ask yourself, if a belief consistently, over time, predicts the contents of future experience and consistently plays an inferential role alongside other beliefs which successfully predict the content of future experience, in what meaningful sense could that belief be said to be false?

Staircaseghost said...

The Law article is disappointing.

He is of course right that there are conceptual constraints on what constitute beliefs that make Plantinga's suggestion that they might be utterly divorced from behavior risible.

But he's been snookered by Plantinga's way of talking as though evolution can "select for" beliefs -- as though there is a gene for each belief!

What is visible to selection is not individual beliefs, but belief-forming mechanisms. Brains, especially ones like in higher primates, are domain-general induction machines, and as such the accuracy of beliefs about water location and tiger footraces are not statistically independent of one another.

Thus the "problem" is not one of billions and billions of individual beliefs being linked one at a time to individual adaptive behaviors. It is the creation of a domain-general induction machine creating beliefs which in totality, on average, across time have adaptive utility. Natural selection acts to make these better, and there is just no non-miraculous way a lineage could stumble and slip its way across millions of years with blind luck after blind luck after blind luck after blind luck forming false-but-adaptive belief/behavior pairings.

Crude said...

But he's been snookered by Plantinga's way of talking as though evolution can "select for" beliefs -- as though there is a gene for each belief!

From what I've read of Plantinga, he largely claims the exact opposite: that the contents of beliefs are invisible to selection. A "belief" never gets selected for or against directly. What does get selected for or against are behaviors. Maybe those behaviors are linked to a belief, and maybe they're not.

It is the creation of a domain-general induction machine creating beliefs which in totality, on average, across time have adaptive utility.

But that seems wrong as well: what would have "adaptive utility" are the behaviors and actions. The beliefs, whether individual or in total, aren't what selection acts on. This is drawn out more with this sort of example...

Natural selection acts to make these better,

All natural selection could act on directly is the behavior in this case. Whether or not thirsty Rick believes there's water over the hill for him doesn't matter. What matters for selection is if Rick goes over the hill and drinks the water. He can even have no beliefs at all. (More below.)

there is just no non-miraculous way a lineage could stumble and slip its way across millions of years with blind luck after blind luck after blind luck after blind luck forming false-but-adaptive belief/behavior pairings.

This seems out and out false, and there are quick examples on-hand: do bacteria have beliefs? Does a bacteria moving towards food "believe there is food over there"? I think most people would agree no. But if you have an utter lack of belief + behavior pairing that's adaptive, then it seems far more plausible to have an extremely wrong set of beliefs + behavior pairing that's adaptive as well.

Another way of putting it is Plantinga's argument doesn't require "blind luck after blind luck" in an evolutionary sense, since he's not denying that natural selection can result in adaptive behavior. The issue is that you can seemingly get adaptive behavior without true beliefs (and again, without any beliefs at all). And fiddling with evolution to guarantee that where there are beliefs, true beliefs propound, embarks on a project of retrofitting evolutionary theory that a theist could admire.

Crude said...

Minor edit: Prevail, not propound.

jamiesrobertson said...

there is just no non-miraculous way a lineage could stumble and slip its way across millions of years with blind luck after blind luck after blind luck after blind luck forming false-but-adaptive belief/behavior pairings.
What about theism? Wouldn't many scientists say this was a false-but-adaptive belief which has been with us for thousands of years anyway?

Staircaseghost said...

From what I've read of Plantinga, he largely claims the exact opposite: that the contents of beliefs are invisible to selection. A "belief" never gets selected for or against directly. What does get selected for or against are behaviors. Maybe those behaviors are linked to a belief, and maybe they're not.

You've missed the target of the objection.

Beliefs are not heritable. Therefore, beliefs are not subject to selection.

Law uncritically picked this infelicity up when responding to Plantinga and it needs to be called out because calling it out invalidates the entire enterprise. Anyone with a high school understanding of biology (a category which apparently excludes Plantinga) knows that there can be no such thing as a "gene for believing in tiger footraces".

All natural selection could act on directly is the behavior in this case. Whether or not thirsty Rick believes there's water over the hill for him doesn't matter. What matters for selection is if Rick goes over the hill and drinks the water.

Thirst is a general desire that, when it goes right, tells an organism to go where it believes water is. The content of a belief is exhausted by the functional role it plays in guiding behavior antecedently conditioned by desires. This latter point is now screamingly obvious to me, but the fact that it used to not be screamingly obvious to me gives me a moment's pause in evaluating the philosophical competence of others who base their arguments on it.

Staircaseghost said...

(cont'd)
Consider that lying requires more cognitive overhead - and thus, by definition, more neural complexity - because you need to keep two sets of books, one the set of things that are true, and two the set of things you have to remember that you've convinced someone else is true. Do you see how the latter scenario is exactly equivalent to saying that perception might be consistently lying to us about its contents?

This seems out and out false, and there are quick examples on-hand: do bacteria have beliefs? Does a bacteria moving towards food "believe there is food over there"?

This is not a counterexample to my point because it fails to be an example of an organism which has beliefs!

What I am talking about is a race of people who think tigers are friendly sportsmen but who somehow manage to keep from getting eaten when they walk up to them after the race to shake hands and congratulate them on their performance.

But if you have an utter lack of belief + behavior pairing that's adaptive, then it seems far more plausible to have an extremely wrong set of beliefs + behavior pairing that's adaptive as well.

You haven't thought it through WRT any plausible implementation in real biology in the real world. How, specifically, would such a monstrosity evolve? What are the intermediate steps, the transitionals? Present a plausible evolutionary pathway for statistically dependent beliefs to be paired with statistically dependent behaviors where the former are random with respect to truth.

Consider a desire whose content is "get some water and drink it when you need it". Now consider a variety of stimuli: every door with a green light has water behind it, every door with a red light gives you an electric shock. Every time you smell pine trees there is water underneath, every time you smell gasoline there is poison. Every time you hear a bell there is water, every time you hear a buzzer there is another electric shock etc.

Because brains are domain-general induction machines it is very easy to see how ONE simple, parsimonious linking of stimulus and response will get you the right behavior in each of these cases. But according to Alvin, it is at least as likely that there are three separate mechanisms fortuitously in place, and each one not only has to link the response to the correct stimulus, it has to do it in a way utterly unrelated to every other stimulus known to be correlated with it (i.e. generate false beliefs) and never pay a price for it! Under what possible evolutionary scenario could such a mechanism evolve?

Another way of putting it is Plantinga's argument doesn't require "blind luck after blind luck" in an evolutionary sense, since he's not denying that natural selection can result in adaptive behavior. The issue is that you can seemingly get adaptive behavior without true beliefs.

"Seemingly" as in a bare logical possibility. Not seemingly as in any kind of remotely likely possibility. This is classic apologetic obfuscation.

It is "blind luck" if I am blindfolded with banana peels on my feet but am somehow the winning quarterback in the superbowl. Plantinga proposes that the probability that the winning quarterback in the superbowl knows what he is doing is "low or inscrutable".

You'll have to pause me as I deal with hernia caused by the ensuing laughter.

Crude said...

there can be no such thing as a "gene for believing in tiger footraces".

I'd like to see where Plantinga says that there is a "gene for believing in tiger footraces", or where his argument depends on that kind of claim. What I've seen Plantinga say is that the universe as construed by the naturalist along with evolution as construed by naturalists yields a process which can be expected to give a variety of results - and that "typically having true beliefs" is not among those results.

Thirst is a general desire that, when it goes right, tells an organism to go where it believes water is.

You don't need 'thirst' in that sense to 'go where the water is'. See the bacteria example.

Do you see how the latter scenario is exactly equivalent to saying that perception might be consistently lying to us about its contents?

Not at all, because presenting it as "perception might be consistently lying" is an anthropomorphism which confuses the subject. A closer analogy would be an instrument that gives a readout which doesn't do what we think it does.

This is not a counterexample to my point because it fails to be an example of an organism which has beliefs!

It's an example of how an organism can engage in behaviors that can be selected for or against regardless of what beliefs it holds - in this case, because it (presumably) has no beliefs whatsoever. Your beliefs don't need to map accurately to the world for all (to use another anthropomorphism) evolution 'cares'. All that matters are the behaviors.

You haven't thought it through WRT any plausible implementation in real biology in the real world.

Straight from the Creationist's Book of Evolutionary Skepticism!

That organisms can survive and thrive with presumably zero beliefs is accepted. That evolution selects for behaviors, not beliefs, is relatively uncontroversial. That the belief which provokes a given fitness-enhancing behavior has no necessity to be a 'true' belief is likewise uncontroversial. That evolution is rigged such that any population which has beliefs tends to have, more often than not, true beliefs is pretty damn controversial - not only does it have to be argued for, but arguing for it just impales the naturalist more deeply on the teleology horn.

But best of all is this: According to a number of naturalists, evolution has resulted in some dramatic deceptions at the most fundamental levels - doing everything from deluding 'us' into thinking 'selves' exist, to tricking us into thinking 'beliefs' exist, to otherwise. (Somehow, our "beliefs" about science are supposed to be immune to this.) To the degree those moves are made, Plantinga's point is bolstered - by naturalists themselves.

Plantinga proposes that the probability that the winning quarterback in the superbowl knows what he is doing is "low or inscrutable".

Wow, you don't get Plantinga at all - this is flatly false. Plantinga proposes no such thing, because Plantinga rejects naturalism and thinks our minds are oriented towards truth generally. What Plantinga argues is that, given the conjunction of naturalism and evolution, not only are the reasons to think the winning quarterback in the superbowl knew what he was doing (again, a claim that pops up from a number of naturalist philosophers themselves) low or inscrutable oddswise, but that *if you accept naturalism and evolution* you've got no good reason to accept what you think you know about quarterbacks.

I think biting the bullet and insisting that evolution is such that, when beings with minds evolve, it's in essence a law of nature that their beliefs and minds will be oriented towards true beliefs at the population level. It just happens that this wreaks havoc on naturalism, but hey - the alternative is worse.

Staircaseghost said...

I'd like to see where Plantinga says that there is a "gene for believing in tiger footraces", or where his argument depends on that kind of claim.

He doesn't SAY it (that would imply that he was actually talking in terms of actual biological systems, which would just be too much effort for him), but he (and Law) talk as though it is true, and their thought experiments depend on it as a suppressed premise.

No gene for tiger-footrace beliefs means natural selection is never in a position to select for it. It's that simple.

What I've seen Plantinga say is that the universe as construed by the naturalist along with evolution as construed by naturalists yields a process which can be expected to give a variety of results - and that "typically having true beliefs" is not among those results.

What Plantinga says is utterly divorced from what anyone can pick up from a high school biology textbook.

According to actual biology (biology, not the apologists' boogeymen "naturalism", "humanism", "communazism"), we would expect belief-forming systems to have outputs which accurately associate past experience with future expectation, on the whole, across time. But Alvin refuses to learn biology, so he spins creationist fantasies where tiger-footrace genes poof into existence fully formed, one for each belief!

You snipped away without addressing the multiple stimulus/response pairing examples I gave. Please address the fact that brains are domain-general induction machines linking perceptions to future expectations. Please address the fact that one domain-general induction machine is more parsimonious than three or some indefinite number of specialist induction machines which each do miraculous double-duty creating fortuitously adaptive links to behavior and somehow generating false beliefs.

Not at all, because presenting it as "perception might be consistently lying" is an anthropomorphism which confuses the subject. A closer analogy would be an instrument that gives a readout which doesn't do what we think it does.

The two (tiger-footraces and lying) are exactly equivalent in terms of parsimony cost and therefore plausibility.

And I ask again, in what sense does an instrument which consistently, across time and through robust counterfactual testing lights up in the presence of X not give us true beliefs about the presence of X?

It's an example of how an organism can engage in behaviors that can be selected for or against regardless of what beliefs it holds - in this case, because it (presumably) has no beliefs whatsoever. Your beliefs don't need to map accurately to the world for all (to use another anthropomorphism) evolution 'cares'. All that matters are the behaviors.


Then you're changing the subject from the EAAN, so this is an irrelevant tangent. The purpose of that argument isn't to litigate the HPOC, it's to argue that given a BFS which creates beliefs, these beliefs would (somehow) be utterly divorced from their causal backstory.

Moreover, you've not addressed the sense in which a belief which consistently makes accurate predictions could possibly be said to be false.

Straight from the Creationist's Book of Evolutionary Skepticism!

The irony is lost on one of us, and I'm quite sure it isn't me.

It was Plantinga who wrote a book about what we would or would not plausibly expect evolution to produce. How foolish, then, to expect his arguments to be at all tethered to actual knowledge about what actual biology is actually plausibly expected to produce!

(Cont'd)

Staircaseghost said...

It is plausible that domain-general BFSs would evolve from known precursors. It is wildly implausible that billions of individual belief-specific BFSs would evolve ready-made to link each stimulus with an adaptive behavior AND SIMULTANEOUSLY with a false belief that somehow never gets challenged.

Please address the fact that this point of elementary biology destroys Plantinga's argument.

That the belief which provokes a given fitness-enhancing behavior has no necessity to be a 'true' belief is likewise uncontroversial.

There you go again with the classic apologetic "Retreat to the Possible".

Whether it's papering over obvious Biblical contradictions or trying to explain away the existence of evil, don't worry about whether your scenario is forehead-slappingly absurd! Just get the evul skeptic to admit that it's logically possible.

(Note that you still haven't addressed the issue of what possible sense can be made of a belief which consistently makes accurate predictions being "false", much less one's entire set of beliefs doing so)

That evolution is rigged such that any population which has beliefs tends to have, more often than not, true beliefs is pretty damn controversial - not only does it have to be argued for, but arguing for it just impales the naturalist more deeply on the teleology horn.

It is not "rigged", it is a simple and straightforward consequence of known (i.e. actual) biological forces working on known biological mechanisms.

Please address the fact that a domain-general BFS (gosh, notice how often he bolds that term, maybe it's important?) is more parsimonious and has clearer biological precursors than billions and billions of belief-specific lying machines. Understand that natural selection is never in a position to select between domain-general BFSs and belief-specific lie-generators like tiger-footrace systems because there is no such thing as a gene for things like tiger-footrace beliefs. (notice how "there's no such thing as an X" is a pretty good argument against the claim that it is plausible that evolution could select for X?)

Wow, you don't get Plantinga at all - this is flatly false. Plantinga proposes no such thing, because Plantinga rejects naturalism and thinks our minds are oriented towards truth generally.

No, flatly true I'm afraid. You make my point for me. He thinks that it is so wildly improbable that someone could beat Kasparov in chess because they, I don't know, are good at thinking true things about chess instead of making hundreds of moves that just by blind luck map to him thinking he is scrubbing the breakfast dishes that the only explanation is a miraculous suspension of the laws of reality.

(cont'd)

Staircaseghost said...

What Plantinga argues is that, given the conjunction of naturalism and evolution, not only are the reasons to think the winning quarterback in the superbowl knew what he was doing (again, a claim that pops up from a number of naturalist philosophers themselves) low or inscrutable oddswise, but that *if you accept naturalism and evolution* you've got no good reason to accept what you think you know about quarterbacks.

And his argument is based crucially on the premise that because it is logically possible in an individual case that false beliefs might have lead to the behavior, that we should conclude that it is highly plausible that this is the case, and that this might have been the case for millions of generations of organisms, slipping on banana peels and running tiger footraces without ever once catching on or having to pay the price in neurological parsimony.

I think biting the bullet and insisting that evolution is such that, when beings with minds evolve, it's in essence a law of nature that their beliefs and minds will be oriented towards true beliefs at the population level.

It is no "bullet" because as I've explained at length, it is the straightforward consequence of any plausible scenario employing actual biology. Put more strongly: we would be more inclined to conclude that a miracle had occurred if we saw a race of tiger-footracers than if we saw a race of "run from things that look like things that were dangerous in the past".