Plantinga’s EAAN revisited

Blogs are places where one tosses out a hastily constructed piece of argument, or commentary, and not where one slowly and thoughtfully writes something that one will eventually earn an income from (unless you are PZ Myers). So when I responded to Plantinga’s Evolutionary Argument Against Naturalism, I did so based on what I could find online (not having access to a library, my own or anyone else’s right now). I could not find a clear formulation, so I reconstructed it, based on these sources:

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

http://books.google.com.au/books?id=p40tc_T7-rMC

http://books.google.com.au/books?id=XQqBP5trqCIC

http://books.google.com.au/books?id=AEnhhfR7N70C&pg=PA307

Since I could only see some of the published argument, I was sloppy.

Now Professor Plantinga has replied, after my post was brought to his attention by blogger Maryann Spikes, and he makes some interesting comments.

First, he disputes that I have understood his argument properly and that none of the premises I adduced in the last post were in fact his. Of course, he is entitled to state his premises as he sees them; how else could he do it? But as I was reconstructing the argument without a clear statement in front of me, I did my best. If my premises are not his, then I cannot claim to have argued against his conclusion [I will say, though, that it doesn’t seem to me to have been all that clearly stated in the past in what I could locate; and although I won’t say there are several versions of the EAAN, at the least I have seen others interpret it this way on the web, which is enough to justify a webby argument against it]. I apologise to Professor Plantinga for ascribing to him views he says he does not hold.

Second, he objects to my definition of “naturalism”. This is crucial because the entire argument, and often indeed the recent debate against simple evolution by Christian theologians and philosophers, depends on a definition of “naturalism”:

the second main problem is that in attempting to state my argument, he uses the term ‘naturalism’ in a way completely different from the way I use it. In my argument I take naturalism to be the claim that there is no such person as God or anything like God—no angels, demons, or anything else supernatural.  Naturalism is therefore stronger than atheism; you can’t be a naturalist without being an atheist, but you can be an atheist without rising to the full heights (or descending to the murky depths) of naturalism.  Wilkins, however, takes naturalism to be something entirely different: “the view that everything about humans, including their cognitive capacities, evolved”. I reject and argue against naturalism taken my way; it is naturalism taken my way that I argue against.  On the other hand, I have no objection to naturalism taken Wilkins’ way: since I have no objection to the view that we have come to be by an evolutionary process (one guided or orchestrated by God), I also have no objection to the view that everything about us, including our cognitive capacities, has evolved.

This is clear and unobjectionable. We are both in agreement then. Human cognitive faculties are evolved. However this is not a definition of “naturalism” that I recognise from the philosophical literature, although it is often asserted as a truism by the apologetics literature, basically since Philip Johnson defined “naturalism” in his anti-evolution book Darwin on Trial (1992):

Naturalism assumes the entire realm of nature to be a closed system of material causes and effects which cannot be influenced from anything “outside”. Naturalism does not explicitly deny the existence of God, but it does deny that a supernatural being could in any way influence natural events, such as evolution… Scientific naturalism makes the same point by starting with the assumption that science, which studies only the natural, is our only reliable path to knowledge. [p145 of the 2010 edition]

Now this is clearly a source of Plantinga’s definition. But what about the prior philosophical tradition (which is surely relevant in a philosophical discussion)? What is naturalism in that context? Well it is a lot less dramatic. When G. E. Moore in 1903 named the famous “naturalistic fallacy” in ethics, naturalism was held to be the claim that a moral property was a natural property (like the presence of pleasure or the absence of pain). This was later extended to include semantic properties like the intentionality of words and names (their “meaning”), or the notion that mind could be reduced to physical states of nervous systems, and the like. Each naturalism employed a limited and specific form of naturalism. Ethical naturalism was the claim that ethics was natural; mental naturalism the view that the mind was natural, and so forth. So how does this become a claim about God? Even Johnson did not assert it was outright, but had to extrapolate from his overstated formulation. Plantinga’s version, however, is gerrymandered to the point that one has to ask why it is even there. If the argument is devised to show that a view defined as opposition to God is, well, opposed to God it is not interesting until that is motivated, and since philosophical arguments must appeal to what Peter van Inwagen once called “ideal agnostics” who have not yet made up their minds, this is questionable if not question begging (but see this).

This is not to say that there are not those who take a line like this overall. In my experience, though, these people are typically called “materialists” or “physicalists”, not “naturalists”. Professor Plantinga may redefine or define terms as he likes, but it is unsurprising that someone like me (who is not American and therefore not likely to understand the history behind his use of the term) might become confused and impute a definition that is not his own. I am a naturalist, for example, but I do not rule God out as impossible, instead holding that natural methods of knowledge gathering give me no confidence such a being exists. My naturalism is about cognition: we can only know what is acquired by natural means. Whether that implies there are not such beings as supernatural deities depends upon a further premise that only science is the way to know anything, a view I reject as circular. It happens in my view that only science delivers reliable knowledge, but I’m a pragmatist and so that is to be expected. When or if theology starts to deliver reliable knowledge I shall include the ontology that warrants such beliefs in my own.

Anyway, let us look at Plantinga’s stated argument in detail:

The argument goes as follows. First, I’ll use ‘N’ to abbreviate ‘naturalism’, ‘R’ to abbreviate ‘our cognitive faculties are reliable with respect to metaphysical beliefs’ and ‘E’ to abbreviate ‘we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory’).  Then we can state the argument as follows:

P1 P(R/N&E) is low i.e., the probability that our cognitive faculties are reliable when it comes to metaphysical beliefs given the conjunction of naturalism with evolution is low.

P2 One who sees that P1 is true and accepts N&E has an undefeated defeater for R.

P3 One who has an undefeated defeater for R has an undefeated defeater for any of her metaphysical beliefs.

P4 N&E is a metaphysical belief.

Therefore

C One who sees that P1 is true and accepts N&E has an undefeated defeater for N&E and hence can’t rationally accept N&E.

There is a kind of over-formalism that philosophers use based on the days in which typesetting cost money, and which often . I will make the substitutions so that we have the argument in a more ordinary language format:

P1 The probability (our cognitive faculties are reliable with respect to metaphysical beliefs/[both naturalism [is true] and we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory]) is low.

P2 One who sees that P1 is true and accepts [both naturalism [is true] and we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory] has an undefeated defeater for our cognitive faculties being reliable with respect to metaphysical beliefs.

P3 One who has an undefeated defeater for our cognitive faculties being reliable with respect to metaphysical beliefs has an undefeated defeater for any of her metaphysical beliefs.

P4 That [both naturalism and we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory] is a metaphysical belief.

Therefore

C One who sees that P1 is true and accepts naturalism&we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory has an undefeated defeater for [both naturalism and we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory] and hence can’t rationally accept [both naturalism and we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory].

This is still clumsy, but it is no less exact than Plantinga’s formalist version, by definition. So what I want to do now is do what is called an argument mapping of the argument as presented here in “standard form”. The reason I will do this is that it might make the argument clearer and also make clear some hidden premises and weak spots. I will replace the complex form “[both naturalism and we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory] with a name: the conjoint thesis for simplicity.

Argument maps are read “upwards” so the top claim or contention is the final conclusion. I give a rough draft first, which employs Plantinga’s premises and structure only.

First draft

The numbering is relative, so it will not survive into the next version. Also, I added in 3A-d a name – the conjoint thesis – to make the statements cleaner. Now as it stands, this is clearly not a valid argument yet. It is what philosophers call an enthymeme, an argument which has a lot of unstated reasons, and even conclusions. So the next version will include all the additional premises needed to make it work as it stands, and I have removed personality claims like “A person who…” and changed tenses to make it read more naturally. If you and I were in the same room, I could take you through each step to get here. The major point, however, is that one should make sure that terms and phrases used in shared reasons are similar or identical, and that nothing should pop out in the next level of the argument that wasn’t there beforehand. These shared term rules are “holding hands” for shared reasons, and “the rabbit rule” (you can’t pull out of the hat a rabbit that wasn’t put there in the first place), and they match traditional logical rules pretty closely.

Final version

Inserting all the missing reasons and conclusions, we get a valid argument for the claim that naturalism sense Plantinga is unacceptable (rationally speaking). I have added what I take, from the very name of the argument, to be the implicit conclusion that naturalism is unacceptable (I have not used the terms “true” or “false” because Professor Plantinga doesn’t. One can do the substitutions if one wishes).

But is the argument sound? If the argument is valid but unsound, then one or more of the reasons adduced are false (or, to be consistent with the epistemic nature of this argument, unacceptable). Let us look at the claims made outright and not further argued for in this formulation:

1ac

Professor Plantinga makes this claim in his rebuttal to me. The evolutionary claim, remember, is this:

5ab

I concur with Professor Plantinga here, and anyone who think that science provides reliable knowledge must think this, for if anything is reliable knowledge it is the evolutionary claim. So no trouble at this point. How about these two:

3aa3ac?

Both of these may be challenged. For a start, if one does not accept the definition of “naturalism”, one need not accept the definition of the conjoint thesis as a metaphysical belief. Suppose one adopts “naturalism sensu Wilkins”: that “we can only know what is acquired by natural means”, or that “naturalism is a claim that some property or object is natural” (epistemic and substantive naturalisms respectively). In this case the claim that naturalism is acceptable is a claim either that we do know some things or that some things are not supernatural. So since the conjoint thesis involves naturalism (of either kind) conjoined with the evolutionary claim that our (cognitive) capacities evolved, what we might have is an epistemic thesis of a limited kind:

Some of our knowledge is expected to be reliable + our knowledge capacities evolved.

This doesn’t seem to me to be a metaphysical claim, not yet (you have to attach all kinds of theses about knowledge being metaphysically constrained, and so forth). In fact it is an empirical claim: both claims can be observed and tested. Unless you think everything is metaphysical (and there are such folk; I do not know if Professor Plantinga is one), empirical claims are the antithesis of metaphysical claims. So far, then, the argument rests squarely on the definition of naturalism:

5aa

3A-a can be challenged on the grounds of scope. I can think some of my metaphysical beliefs are unfounded (because they rely on evolved capacities, for example) but not that all of them are, unless I think the only way to acquire metaphysical beliefs is via evolved capacities. So having an undefeated defeater for the evolutionary claim doesn’t translate to an undefeated defeater for all my metaphysical claims (I might, for example, be a kind of evolutionary Platonist). However, I am inclined to think that only evolved capacities are feasible ways to acquire metaphysical beliefs, because all beliefs are states of the brain, so I won’t challenge it. All you dualists might, though.

So we are back at the definition of naturalism given in 5A-a. Without this, the argument has no purchase. I will now explain why I think it is a bad claim to assert (apart from questions of history of philosophy).

First, terms that are ambiguous cause confusion. If “naturalism” means “atheism”, then call it “atheism”. Professor Plantinga says that

In my argument I take naturalism to be the claim that there is no such person as God or anything like God—no angels, demons, or anything else supernatural.  Naturalism is therefore stronger than atheism; you can’t be a naturalist without being an atheist, but you can be an atheist without rising to the full heights (or descending to the murky depths) of naturalism.

This is entirely unclear to me. I’d like to see the argument for it. I have previously discussed what is meant by “atheism” and came to the view that there are three kinds: positive atheism (the denial there are gods and the supernatural), negative atheism (the denial that gods are to be believed), and privative atheism (the complement to a theist claim that God exists). What Plantinga is saying here is that privative atheism does not imply positive atheism, which I agree with. But privative atheism is not closely related to naturalism as I understand it.

So what does the argument show? Does it show that one cannot both believe in positive atheism (as I define his “naturalism”) and the conjoin thesis? Not if the conjoint thesis is an empirical claim, and not a metaphysical claim, which is what I, and a good many others, think is the case. One may think that empirical evidence shows there is no god or supernatural, or one might, as I do (and apparently Dawkins does too) think that it merely makes it an unwarranted empirical belief. Either way, this is about the formation of empirical, not metaphysical beliefs.

I’d like to thank Professor Plantinga for clarifying this argument. I still think Darwin ran the EAAS though, and that I haven’t much misinterpreted his argument, and that was the focus of my last post.

For those who wish to know, I used a program called Rationale, by Austhink Software. Unfortunately it is only for Windows (yuck), but those needing to run Windows on a Mac (yay) cannot do better than use Parallels Desktop for Mac.

See also Jason Streitfeld’s response, which is similar to mine in structure.

14 thoughts on “Plantinga’s EAAN revisited

  1. Shorter version: someone (you) finally called Plantinga on his BS quote-mine of Darwin, forcing Plantinga to rephrase his argument to say that if evolution and naturalism are true, we have a low probability of being able to ascertain correct *metaphysical* beliefs.

    But Plantinga’s original argument, advanced for a decade or more, was that our ability to get true beliefs *of any sort* was jeopardized if evolution+naturalism were true. This is why there were so many arguments about lions and whether or not it was really true that natural selection would favor brains that led to running from lions based on absurd beliefs, rather than on true beliefs like the dietary preferences of lions, their number, speed, and direction. Plantinga’s whole line of argument on *that* score was idiotic from the start.

    Plantinga’s new argument, though, doesn’t bother me so much. Basically it boils down to “if you believe what Darwin said about the dubiousness of human conclusions about grand cosmic metaphysical matters, then you should, like Darwin, be agnostic in your conclusions about the existence of God/the supernatural etc.; similarly, you should be agnostic about naturalism, since your conclusions about this grand cosmic metaphysical matter are dubious for the same reasons.” This is almost a truism, since agnosticism about e.g. “God exists” equals agnosticism about “God doesn’t exist.”

    There are still various ways to validly remain a naturalist in reasonable senses (outside of grand cosmic metaphysical senses of the word — I specify “grand metaphysical” statements to avoid confusion with more pedestrian types of metaphysics, where standard reasoning is reliable). One is to avoid grand cosmic metaphysical statements about ultimate reality and remain agnostic on those, while observe that naturalism taken as methodological principal has, as an empirical matter, worked incredibly well. Another version of the empirical reply is to note that a great many claims to supernatural explanations have been shown to be mistakes/frauds/wishful thinking/legends of various sorts, and others that are difficult to investigate are plausibly explained the same way.

    Yet another option would be to say “yes, Darwin was right that modified monkey brains are bad at reaching correct conclusions, but if those modified monkey brains work really hard and collaborate and review each other and pay careful attention to data and logic they can figure out the truth nonetheless.” This would be the argument I would use about why humans can succeed at math, despite the fact that our brains suck at math, probably because there wasn’t much selective point in being able to count beyond “one lion, two lions, a bunch of lion-holy crap, time to get the hell out of here!” I suspect grand metaphysical matters are different from math, mostly because the concepts are so malleable and the questions are so remote from the data and from everyday experience.

  2. I agree with what NickMatzke said in his last paragraph.

    Plantinga seems to ignore the difference between:
    a) ” the probability that a single person’s cognitive faculties are reliable” and
    b) “the probability that our cognitive faculties are reliable”

    His argument might work if all the phrases were recast into ‘a single person’ – yes, a single person might draw incorrect views about naturalism because of his or her inadequate cognitive faculties. But this is a trivial case (against philosophical naturalism) where any individual may believe in anything if there is no constraint.

    Cast the argument into what many people can do together and the case against (methodological) naturalism is considerably weakened.

  3. we and our faculties have come to be by way of the processes appealed to in contemporary evolutionary theory

    I suggest an improvement on this argument by replacing this with:

    we and our faculties have come to be by way of the processes appealed to in contemporary reproductive biology

    This is an improvement because it does not have the appearance of being a different argument, an argument against evolutionary biology. There is time wasted in a number of rejoinders by presenting arguments for evolutionary biology, when that is not relevant to EAAN. Also, the argument has been misapplied by some creationists to argue against evolution. With my suggested replacement, no one would be tempted to waste their time on biology.

    Moreover, there is the possibility that someone would waste time on arguing whether this involves the fallacy of composition or division, that is, whether the “we” is a reference to development of “human-kind” or to “each one of us humans”, and to the knowledge of “humans as a whole” or of “particular humans”.

    Reproductive biology is no less materialistic, naturalistic, or atheistic than is evolutionary biology, so, as far as I can tell, this replacement does not introduce any novelties or problems into the argument. The replacement thus seems to make the argument clearer with no loss of soundness.

    1. I do like your strategy here. In effect it is the argument about cognition in general – is cognition just biological? I don’t know many scientists working in the field, or too many philosophers of mind, who think otherwise. Not even Chalmers (his Hard problem derives from physicalist assumptions).

      Mutatis mutandis the argument goes through. One thing about a developmentalist account of cognition – it suggests that simple computational accounts of mind are simply wrong; because computers don’t develop (nor do they evolve in the same way).

  4. Excellent post.

    I’m not sure you’ve motivated the metaphysical/empirical distinction well enough, though. I don’t think everything is metaphysical, but I do think the claim that empirical and metaphysical claims are antithetical is next door to nonsense, and could only be maintained if we held some kind of strict sense-data epistemology. Perhaps this is just a problem with what is to be counted as metaphysical, though? (I also think, I should say, that if it could be motivated properly it would be a promising move against the argument, in that it would deal with it at a point that would not require the kind of extraordinarily complicated discussion people usually get mired in when discussing this argument, while at the same time not requiring any oversimplification of the argument.)

    1. See next post … but I don’t need to appeal to a sense-data language to motivate that distinction. Indeed the onus is on those who think metaphysics is everything to show that. I can live without metaphysics, but not without empirical observation.

  5. So Plantinga’s new climb-down is just that evolution means we should be highly doubtful that we can make accurate knowledge claims about unobservables?

    How then is the EAAN not simply a shaggy dog story, rephrasing a core classical Empiricist claim with the proviso “it’s not a feature, it’s a bug”?

  6. If Plantinga’s argument is effective, could we not also use an argument of identical form to show that no explanation for the origins of our cognitive abilities would give us warrant to believe in it?

    If God created us fully formed, 6000 years ago, is there a high probability that our judgment on questions within the scope of this argument (whether metaphysics or everything) is valid? If so, why? And if not, we have no reason to believe that our created minds are reliable on the subject of creation.

    Further, why should we believe our reasoning on that subject is valid? Once you doubt the efficacy of human reason and perception, it’s doubt all the way down. (See Dodgson’s essay on Achilles and the tortoise for something similar.)

    1. Yes of course, but only if we do not avail ourselves of some other avenue or warrant for reason. Plantinga has two outs: one is that God has made our reason and so it is reliable (Descartes’ argument) and the other is that God has revealed himself and the nature of the supernatural. Since both of these rely on there being a God to do these things, it is not a universal solution, but the Christian approach is to show that you really need belief in God to make everything work out.

      This goes to the heart of my criticism about the meaning of “naturalism”. I think it is a project to show that some capacities of humans are not based on anything transcendental. It isn’t incoherent to argue that God warrants our inferences; but it is at the very least circular.

      1. But why does if follow from “God made our reason” that “it is reliable”?Sounds like another circular claim on the order of the usual “refutations” of the Euthyphro argument, or the self-proof of biblical authority. We know God made our reason reliable because that’s the sort of thing he would do, and we know that’s the sort of thing he would do because our reason tells us so, and our reason is reliable. Or we might say we know so because of revelation, and revelation is reliable, because revelation tells us so. Can Plantinga really be so simpleminded as that?

        1. I think the point is that Plantinga gives a defaeter for Naturalism & Evolution.

          Only assuming naturalism and evolution, you have grounds to distrust you cognitive faculties.

          Assuming a benevolent creator god, you have no grounds to distrust you cognitive faculties.

          It’s an argument aimed against the naturalists worldview, showing it to be, if not self contradicting, then at least improbable.

          IANAP so your mileage may vary.

  7. Soren:
    Only assuming naturalism and evolution, you have grounds to distrust you cognitive faculties.
    Assuming a benevolent creator god, you have no grounds to distrust you cognitive faculties.

    These are both interesting claims. We have seen Plantinga’s argument, however poor, for why the first conclusion follows from the first premise. But what’s his argument for why the second conclusion follows from the second premise?

Leave a Reply