Plantinga’s EAAN: Criticism and Discussion

Alvin Plantinga’s Evolutionary Argument Against Naturalism has attracted a great deal of serious critical discussion (e.g. Naturalism Defeated?) and has had a substantial impact on ‘popular’ appraisals of naturalism.  (For example, William Lane Craig frequently uses it, and it also appears in the dismissal of naturalism in The Experience of God.)  Many philosophers have pointed out various problems with the EAAN, and in my judgment the EAAN is not only flawed but fatally flawed.  Nevertheless, it’s a really interesting argument and it could be worth exploring a bit.  I’ll present the argument here and then we can get into it in comments if you’d like — though I won’t be offended if you’d rather spend your time doing other things!

The EAAN has gone through various iterations, but here’s the latest version, from Where the Conflict Really Lies: Science, Religion, and Naturalism (2011).  Intuitively, we regard our cognitive capacities — sense-perception, introspection, memory, reasoning — as reliable, where “reliable” means “capable of giving us true beliefs most of the time” (subject to the usual caveats).  Call this claim R (for ‘reliable’).   But how probable is R?

Suppose that one accepts evolution (E) but also affirms naturalism, defined here as the belief that there is no God or anything like God (N).  What is the probability of R, given N & E?    One might think it’s quite high.  But Plantinga thinks that, however high the probability of R, nevertheless the probability of R given N&E is low or inscrutable.  Why’s that?

Now, here’s the key move (and in my estimation, the fatal flaw): beliefs are invisible to selection.  Why?  Because selection only works on behavior.  If an unreliable cognitive capacity is causally linked to adaptive behavior, then the unreliable capacity will be selected for (i.e. not selected against).  Even a radically unreliable capacity — that one never or almost never yields true beliefs — can be selected for.  Selection only “cares” about adaptive behaviors, not about true beliefs.  (More precisely, we have no reason to believe that the semantic content is not epiphenomenal.)

So, Plantinga thinks, given N&E, the probability of R is very low. But, if the probability of R is low, given N&E, then that should ‘infect’ the likelihood of all of the beliefs produced by those capacities — including N&E themselves.  So, given N&E, we should it think it extremely unlikely that N&E is true.  And so the initial assumption of N&E defeats itself.  (Here I’m being much too quick with the argument, but we can get into the details in the comments if you’d like.)

Anyway, it’s a really cool little argument, and it’s not immediately clear what’s wrong with it — and I thought it might be worth discussing, given how influential it is.

 

 

500 thoughts on “Plantinga’s EAAN: Criticism and Discussion

  1. Kantian Naturalist: But dreams don’t have success conditions. The dream-imagery isn’t anything that one can get into a better position from which to view it, or ask another person if one is perceiving it correctly. It’s an activity of sensory consciousness, yes, of course. But the idea that all activity of sensory consciousness thereby counts as perceiving is precisely the idea that I mean to reject

    There are lots of conditions and situations in which perception is subjective, and I would suggest that there are “objective” means available for outsiders to confirm that dream imagery is occurring. And those means are improving through technology.

    I am not persuaded there is any difference between dream imagery, hallucination, and “correct” perception, whatever that might mean.

    Perception is something we do. There is no correct or incorrect.

  2. Kantian Naturalist:
    Erik would probably want to say that the normativity of perception is due entirely to the intellect. As Descartes says, “the senses do not judge”. And that is, in some sense, correct (or: acceptable by my lights). I say that because judging takes place within the space of reasons, which is a distinct kind of normativity.

    I wonder if this might require a bit of qualification. Where, exactly, do the senses give way to reason within the brain? Experiments show that by the time the frontal lobes receive a visual signal, about 98% of it has already been discarded, and you have no way of knowing this. There is simply too much visual input to process, so the brain ignores everything reaching the eyes that is not deemed “interesting” by the visual center, and even within that small remainder, either discards or INVENTS material when there are unexpected anomalies. That way, we can generally see what we expect to see without excessive workload.

    I don’t know if neuroscientists have determined exactly where which kind of editing occurs in the visual chain, but definitely judgment is being exercised at multiple levels well before the intellect gets a peek at any of it.

  3. Kantian Naturalist: I don’t think we perceive in dreams, actually.

    I agree, which means that I disagree with petrushka on this.

    I think of perceiving as a normative concept.

    However, I am dubious of this. Our verbal describing of what we perceive is normative, but I don’t think we can say that about perceiving itself.

    Yes, I expect that we have internal criteria used for perceiving. But these are private criteria that are not publicly visible, and not accountable to anyone else. I don’t think I would want to use “normative” for that case. However, our describing of what we perceive is potentially public and subject to public norms.

  4. 404 thoughts on “Plantinga’s EAAN: Criticism and Discussion”

    I thought that was amusing.

  5. petrushka: I am not persuaded there is any difference between dream imagery, hallucination, and “correct” perception, whatever that might mean.

    I strike a key on the piano. I hear a note.

    I drop a heavy book on the top of the piano. I hear the same not, a resonance by the same string. I do think there’s a difference between playing a note on the piano, and on hearing the string resonate for some other cause.

    I see the relation between dreaming and perceiving as analogous. They share some of the apparatus, but the causal chain is different.

  6. Neil Rickert: I agree, which means that I disagree with petrushka on this.

    How would you label what we see and hear in dreams?

    Hallucinations?

    Where would you draw the line, and how?

    What about misperception?

  7. petrushka: 404 thoughts on “Plantinga’s EAAN: Criticism and Discussion”

    It’s now at 406. And, after I post this, it will be a 407 (or higher).

    Were you thinking of the “404 page not found”? If not, then I don’t know what you were hinting at.

  8. Neil Rickert: How about “dream”?

    I don’t know why you think another term is needed.

    Why do you need the term perception, when you could use seeing or hearing?

  9. Neil Rickert: It’s now at 406.And, after I post this, it will be a 407 (or higher).
    Were you thinking of the “404 page not found”?If not, then I don’t know what you were hinting at.

    I was amused by the proximity of 404 and Plantinga. Perhaps I’m too easily amused.

  10. Neil Rickert: I strike a key on the piano.I hear a note.

    I drop a heavy book on the top of the piano.I hear the same not, a resonance by the same string.I do think there’s a difference between playing a note on the piano, and on hearing the string resonate for some other cause.

    I see the relation between dreaming and perceiving as analogous.They share some of the apparatus, but the causal chain is different.

    As a musician, I would disagree. Even though the same string can produce the same overtone series, there is still the matter of ASDR (attack, sustain, decay, release), also called the “envelope” among synthesizer people. There are a lot of very different tones you can get out of a given piano string, and your ear is perceptive enough so that it would be difficult to use two methods of causing the string to vibrate where you could NOT hear the difference.

  11. dazz: Do you have a point mungy?

    Do you ever step back and look at what you’ve written to:

    a.) see if it makes sense
    b.) see if it’s relevant

  12. keiths: If Mung weren’t quoting someone else, he’d have to make his own argument…

    It’s a pragmatic shortcut to avoid all the “you must support that or retract it” nonsense we see so often here at TSZ.

  13. Mung: It’s a pragmatic shortcut to avoid all the “you must support that or retract it” nonsense we see so often here at TSZ.

    So all you have to do to be comfortable is:
    1) find a place where you don’t have to support what everyone agrees with;
    2) where everyone agrees with you.

  14. I think of thinking as a normative concept.

    Either we can think correctly (we call this “thinking”) , or we can think incorrectly (we call this “not thinking”).

    I think the concept of saying perception is a normative concept as falling into the second category.

    Every time you perceive something incorrectly, you are not actually perceiving. You know like dreaming according to your waking state, which tells you your dreaming state was wrong (your dream state is not allowed to tell your wake state that it is wrong, because..well, that would be wrong). So if someone tells you their perception is different from yours, then it must be that neither of you is perceiving, because you have both been decided to be wrong, and wrong perception are the same as not perceiving. Got it?

    But actually neither of you is wrong, because you didn’t perceive at all, because perceiving incorrectly is the same as not perceiving at all, so since neither of you perceived neither of you is wrong.

    So since neither of you is wrong, you both must have perceived! And since you perceived it must be correct. Its exactly like the concept of fitness in evolution! It can’t be wrong!

    Pass the doobie, isn’t philosophy great!

  15. Flint: So all you have to do to be comfortable is:

    Why would I need to be comfortable?

    A number of TSZ regulars seem to think it’s their duty to make theists and IDists as uncomfortable as possible here at TSZ, and that is apparently what Elizabeth had in mind for this site.

    If I wanted to be comfortable I’d go elsewhere.

  16. Kantian Naturalist: I don’t see what the point is of posting that long quote from Rosen.

    I guess I could have just shouted insults at Neil.

    : I can post long quotes from fictionalists about mathematics (Hartry Field, Hilary Putnam) but why bother?

    Are you a fictionalist like Neil? Was Quine a fictionalist? Did Neil say why he’s a fictionalist, because if he did, I missed it.

  17. Flint: Where, exactly, do the senses give way to reason within the brain?

    Neither “the senses” nor “reason” fit easily within a neurcomputational model of how brains contribute to successful coping of organisms within an environment. I myself wouldn’t even try to use “the senses” or “reason” when talking about what brains actually do, according to our best models of what brains do.

    Experiments show that by the time the frontal lobes receive a visual signal, about 98% of it has already been discarded, and you have no way of knowing this. There is simply too much visual input to process, so the brain ignores everything reaching the eyes that is not deemed “interesting” by the visual center, and even within that small remainder, either discards or INVENTS material when there are unexpected anomalies. That way, we can generally see what we expect to see without excessive workload.

    Right, I’ve read Andy Clark’s nice book on this, and also the critical discussion in Brain and Behavioral Sciences about a year before. He has a nice way of describing how brains are basically in the job of predicting what signals are going to detected, and the bottom-up information propagating from sensory receptors is basically a question of detecting the degree of mismatch between what was predicted and what actually transpired. But Clark’s model of brains as executing Bayes’ Theorem has been severely criticized as well by other neuroscientists and philosophers.

    I don’t know if neuroscientists have determined exactly where which kind of editing occurs in the visual chain, but definitely judgment is being exercised at multiple levels well before the intellect gets a peek at any of it.

    There are certainly multiple branches in the visual information pathway, and many of them are reentrant as well, meaning that latter information processing modifies earlier information processing.

    I’m be extremely reluctant to use the term “judgment” for any of this, since I think that judgment is an activity of rational agents. But there is no rational agent inside the skull!

  18. Mung: Are you a fictionalist like Neil? Was Quine a fictionalist? Did Neil say why he’s a fictionalist, because if he did, I missed it.

    I don’t have any views on the ontology of mathematics.

    Quine wasn’t a fictionalist; he thought that there’s no distinction in kind between believing in numbers, quarks, tables, or gods. As I said before, on his view, once you know what the sentences of a language look like in first-order predicate logic, that’s all there is to ontology. Though I should add that the ontology is constantly being revised, too — but the only criterion, ultimately, is the pragmatic criterion of allowing us to make better predictions about future sensations.

  19. Kantian Naturalist: I’m be extremely reluctant to use the term “judgment” for any of this, since I think that judgment is an activity of rational agents. But there is no rational agent inside the skull!

    Well, we were talking about judgment and reason. I prefer the term “editing” for what goes on in the processing of sensory data. And in my semantics, editing implies judgment.

    Though I should add that the ontology is constantly being revised, too — but the only criterion, ultimately, is the pragmatic criterion of allowing us to make better predictions about future sensations.

    I suspect one need only watch a child learn to experience its world, to see the AHA! moment when a child’s eyes tracks a moving object, which has for the first time been characterized AS an object in the child’s brain. And one implication of this is that as our models are refined through life and our expectations become increasingly accurate, unexpected inputs are increasingly likely to be edited out altogether. A form of Morton’s Demon, perhaps, but as we age we trade more of our ability to notice the unusual, in exchange for more of our ability to make good predictions.

    Maybe all this is wandering too far off topic…

  20. Kantian Naturalist: Quine wasn’t a fictionalist; he thought that there’s no distinction in kind between believing in numbers, quarks, tables, or gods.

    No kidding. By the way, who brought up Quine?

    Was there something you wrote then that indicated Quine also took a position contrary to Neil’s position? You mentioned Quine to indicate that my position had merit? Perhaps I misunderstood your post.

  21. Mung: I guess I could have just shouted insults at Neil.

    That’s pretty much what you did.

    You posted ridicule, which is a way of insulting.

    The funny thing is, that you did not understand what you were ridiculing, so in the end it was you who looked ridiculous.

    Did Neil say why he’s a fictionalist, because if he did, I missed it.

    You could have just asked.

    I’m a fictionalist, because that is the best account of how I do and understand mathematics. If you want more detail than that, I suppose I could start a new thread. But I’m not sure it’s of great interest to others at TSZ.

  22. Mung,

    I brought up Quine as an illustration of what happens if we adopt existential quantification in first-order predicate logic as our sole criterion of ontological commitment, as you appeared to do when you suggested that one could not use existential quantification without using it as implying ontological commitment, since your complaint against Neil was that, as a functionalist about mathematical entities, he could not intelligibly use that symbol in his work.

  23. KN,

    …since your complaint against Neil was that, as a functionalist about mathematical entities, he could not intelligibly use that symbol in his work.

    I think you meant to write “fictionalist”, not “functionalist”.

  24. Kantian Naturalist:
    Mung,

    I brought up Quine as an illustration of what happens if we adopt existential quantification in first-order predicate logic as our sole criterion of ontological commitment, as you appeared to do when you suggested that one could not use existential quantification without using it as implying ontological commitment, since your complaint against Neil was that, as a functionalist about mathematical entities, he could not intelligibly use that symbol in his work.

    FWIW, Quine dropped that as a “sole criterion” sometime between the publication of “On What There Is” in 1948 and that of “Existence and Quantification” in 1966.

  25. Kantian Naturalist:
    This distinction between top-down and bottom-up information processing is a neurocomputationally-informed descendant of Erik’s distinction between “the senses” and “the intellect” . But whereas Erik has only those two “levels”, so to speak, my model has two different directions in a multi-level hierarchy. This hierarchy allows the organism to extract information about the distal causal structures of the world from the proximal energetic flux active at its sensory receptors.

    It would be nice if you used the word “hierarchy” in its correct meaning. In a hierarchy, there is a top and a bottom, up and down, and up is up and down is down. In your case, whenever you use “higher”, “lower”, “top-down” and “bottom-up”, you put them in quotes as if indicating you are using them metaphorically.

    More crucially, from your description it is never clear whether the priority is, up or down, higher or lower. You are talking about different directions in a multi-level structure, not hierarchy. The difference is that in a hierarchy there is a distinct level that rules, guides, or organizes the rest. In your structure there is no such thing.

    The closest to the actual point you come when talking about affordances, but what you say about them is an anarchic mess.

    Kantian Naturalist:
    Hence, on the naturalistic theory of cognition I’m lightly sketching here, cognition is reliable if it allows the animal to successfully cope with the affordances and solicitations that it perceives in its environment. This is strikingly at odds with Plantinga’s characterization of reliable cognition. But my conception of reliable cognition is one that naturalists themselves would find congenial, and Plantinga’s is not.

    It’s hard to make out what you mean here. “Cognition is reliable if it allows the animal to successfully cope with the affordances.” How do you determine success? That the animal got a meal/mate for the time being? If yes, then this is indeed at odds with Plantinga’s definition of reliable cognition, but it’s also at odds with any sensible definition of reliable cognition, because you lack the definition of success. For example, all animals die eventually, so none of them are successful. Everybody gets at least one meal during their life, so everybody is successful. Which way is it?

    Lacking the definition of reliable cognition which would be at least as complete as Plantinga’s, you are not really addressing EAAN. When you have any lesser definition of reliable cognition, you are not really talking about cognition. Animals don’t know what cognition is and they don’t care if it’s reliable or not. They just eat and mate. If you are happy with that, then you are not contributing much to human dialogue.

    But thanks for providing the references. I realize you need authorities to bolster your legitimacy. That’s how the aircastle of postmodernism also works.

    Earlier I made the point that perception, misperception and non-perception also need clear definitions. People here seem to think dream is non-perception. If so, then dream should be a non-event with no effects. But for example fear in dream has the exact same effects as fear on any other occasion, so obviously you are missing something. Carry on.

  26. Erik: The closest to the actual point you come when talking about affordances, but what you say about them is an anarchic mess.

    (Erik is responding to KN).

    Affordances are an anarchic mess. Cognitive systems are an anarchic mess. Culture is an anarchic mess. Evolution produces an anarchic mess.

  27. Erik does make one good point: what is the criterion of “successful coping with the environment” going to be? It can’t been too narrow (e.g. permitting the organisms to obtain just one meal) or too broad (e.g. preventing the organism from ever becoming extinct).

    Unlike most naturalists, I think that we really can’t do biology without teleology, which means that if can’t naturalize teleology, then we don’t have a science of life. The stakes are high!

    However, I think that we can naturalize teleology, and that in fact the great achievement of Darwin was that he showed us exactly how to do that. (Hence I stand opposed to the “Epicureanized Darwinism” of Monod and Dawkins.) Okrent has a really elegant way of putting the complexity of the point:

    Darwin is often described as anti-Aristotelian. And, indeed, there are many respects in which he undercuts Aristotle’s legacy. Nevertheless, there is a central respect in which Darwin was the greatest Aristotelian of the nineteenth century. Darwin agrees with Aristotle — and disagrees with Christianity — on the central issue of whether individuals are evaluable in a nonarbitrary fashion even if they were not made by some rational creator. Darwin even agrees with Aristotle in his judgment concerning which things are so evaluable: living things. For Darwin and Darwinians, living organisms are those individuals that carry the principle of nonarbitrary normative evaluability in themselves. (p. 68)

    That is, what we should be looking for here is an immanent normativity, or a normativity immanent to life itself. This is an absolutely crucial move: normativity is immanent to life itself. This means that (a) any theory of life which fails to describe and explain this immanent normativity is going to be fundamentally inadequate and (b) the rejection of transcendent conceptions of normativity does not entail the rejection or unreality of normativity as such.

    (I do acknowledge that the immanent normativity of life itself does require its own explanation, which is to say that abiogenesis is in many respects the stumbling block of contemporary naturalism. Nagel recognizes this in Mind and Cosmos, though he vastly underestimates how close our current understanding of complex self-organizing systems has come to a solution.)

    Okrent emphasizes that organisms have not just functions but also goals. So what is the goal of a living thing, given the principle of immanent normativity? It is life itself:

    A living thing succeeds just in case it continues to go on living. Otherwise it fails. Since to be alive is to do something that allows it to continue to live, it is life itself that provides the standard against which a particular living thing is to be evaluated. The end or goal of the process is the very process itself, for the end of the process is that the process should continue.

    (This is quite clearly a deliberate echo, on Okrent’s part, of Aristotle’s conception of entelechia, or being-at-work, which Aristotle thinks is essential to all beings as such.)

    I mention Okrent’s work here because it now allows us to understand what “reliable cognition” means: an organism’s cognitive abilities are reliable just in case the functioning of those abilities tends to play a causal role in realizing the goals of the individual organism (to stave off death long enough to reproduce) and of the lineage (to avoid extinction) by allowing it cope with the affordances and solicitations in its environment.

  28. Neil Rickert: I’m a fictionalist, because that is the best account of how I do and understand mathematics.

    How would mathematics be any different if you weren’t a fictionalist? How on earth can others who are not fictionalists “do math” ?

  29. Kantian Naturalist: That is, what we should be looking for here is an immanent normativity, or a normativity immanent to life itself. This is an absolutely crucial move: normativity is immanent to life itself.

    Normativity is immanent everywhere. Life would not be possible without it. So I think I am rightly confused by your insistence on restricting normativity to living organisms.

    If you say that you do not restrict normativity to living organisms, then why would you feel compelled to restrict teleology to living organisms?

  30. Mung,

    I don’t think there is abiotic normativity. Organisms have purposes and goals, but galaxies and molecules do not.

  31. Kantian Naturalist: I don’t think there is abiotic normativity. Organisms have purposes and goals, but galaxies and molecules do not.

    I understand that you do not think there is abiotic normativity. However, if there were not abiotic normativity there would be no life. No galaxies, no life. No molecules, no life.

    You need to ask yourself how life is possible, absent norms.

  32. Mung: How would mathematics be any different if you weren’t a fictionalist?

    Many (perhaps most) mathematicians are platonists (or “mathematical platonists” in full).

    One example of where I disagree with platonists, in the continuum hypothesis. Platonists tend to assert that there is an actual fact of the matter, as to whether the continuum hypothesis is true. By contrast, I see its truth as purely a matter of the choice of axioms.

  33. Neil Rickert: Many (perhaps most) mathematicians are platonists (or “mathematical platonists” in full).

    One example of where I disagree with platonists, in the continuum hypothesis. Platonists tend to assert that there is an actual fact of the matter, as to whether the continuum hypothesis is true. By contrast, I see its truth as purely a matter of the choice of axioms.

    How would mathematics be any different if you weren’t a fictionalist?

    How on earth can others who are not fictionalists “do math” ?

  34. Mung: You need to ask yourself how life is possible, absent norms.

    I don’t see how there need to be norms in order for there to be life. If you have an argument, please go ahead and make it.

  35. As I’ve indicated elsewhere on numerous occasions, I think that Varela and Maturana’s concept of autopoeisis is a very promising way of understanding how to naturalize teleology.

    From what I can tell, an autopoeitic system requires nothing more than a network of autocatalytic molecules (a network in which some chemicals synthesize their own precursors) and a semi-permeable membrane (to ensure that the network is thermodynamically open but organizationally closed to its exterior).

    When I talk about norms, I mean either (1) the kind of first-grade, one-dimensional normativity of organisms whose responsiveness to and engagement with their environments can be described in terms of the satisfaction or non-satisfaction of the organism’s goals and (2) the kind of second-grade, two-dimensional normativity of agents who are responsive to and engage with patterns of commitments and entitlements of propositional contents in a space of reasons.

    I don’t see any room for (1) or (2) in taking about how autocatalytic molecular systems are dynamically self-organized or in how those systems become enclosed in semi-permeable membranes so as to become autopoietic systems.

    From where I sit, anyone who wants to defend abiotic normativity as a requirement for life needs to specify a conception of normativity that is distinct from (1) and (2) and explain how that conception allows us to understand how life came into being.

  36. Kantian Naturalist: I don’t see how there need to be norms in order for there to be life. If you have an argument, please go ahead and make it.

    If you can’t see how norms are needed in order for there to be life, I see no point in trying to argue with you about it. But from where I sit your position demonstrates the utter arbitrariness of your stance on teleology.

  37. KN, to Mung:

    I don’t see how there need to be norms in order for there to be life. If you have an argument, please go ahead and make it.

    Mung:

    If you can’t see how norms are needed in order for there to be life, I see no point in trying to argue with you about it.

    Translation of Mung’s comment:

    I don’t have an argument. This is embarrassing. I was hoping you wouldn’t ask for one.

  38. keiths,

    I doubt that Mung’s refusal to provide us with an argument — involving, at a minimum, a conception of normativity distinct from my (1) or (2) above — causes him the slightest embarassment.

  39. I think Mung is embarrassed. Otherwise he’d admit that he has no argument instead of pretending to condescend to you.

    Mung, tell us why norms are needed for life, and identify some of those norms for us. Perhaps KN and I are too dense to understand, but everyone else will nod sagely and admire your erudition.

    Or not.

  40. keiths:
    I think Mung is embarrassed.Otherwise he’d admit that he has no argument instead of pretending to condescend to you.

    Mung, tell us why norms are needed for life, and identify some of those norms for us. Perhaps KN and I are too dense to understand, but everyone else will nod sagely and admire your erudition.

    Or not.

    I’d certainly be interested in see the argument as to why norms are necessary for life, and if there is a conception of normativity that doesn’t fall into either biological normativity or discursive normativity (as I outlined above).

  41. keiths: Perhaps KN and I are too dense to understand, but everyone else will nod sagely and admire your erudition.

    I don’t think either you or KN are dense. otoh, you regularly act as if you think I am dense. Think, keiths!

    This is just one of those things that too me seems blatantly obvious, if not self-evidently so. I don’t see why I even ought to try to convince you if you disagree.

    But it seems as if KN is willing to acknowledge the existence of different “kinds” of norms. Wikipedia defines Norms as follows:

    Norms are cultural products (including values, customs, and traditions) which represent individuals’ basic knowledge of what others do and think that they should do.

    It might help you to think of how life might be possible if everything life depends on were abnormal.

  42. Kantian Naturalist: I’d certainly be interested in see the argument as to why norms are necessary for life, and if there is a conception of normativity that doesn’t fall into either biological normativity or discursive normativity (as I outlined above).

    Wikipedia defines normative as follows:

    Normative means relating to an ideal standard or model, or being based on what is considered to be the normal or correct way of doing something.

    I’m following the normal use of the term.

Leave a Reply