An Intro to the Philosophy of Religion, Ch. 4

The nature of faith
Murray and Rea offer their account of faith as follows:

“…to say that a person S has faith in proposition p is to say that S believes p despite the fact that (a) there are alternatives to p that are compatible with whatever evidence supports S’s belief that p, and b) there is genuine and somewhat weighty evidence in favor of one or more of those alternatives.”

Since they classify faith as a kind of belief, there can be degrees of faith just as there can be degrees of belief (credences).  In addition, their definition of faith implies that faith can be judged as rational or irrational depending on the evidential context.  But what counts as evidence?

Reliabilism and Evidentialism
The evidentialist maintains that the rationality of a religious belief (merely) depends on the degree to which the belief is supported by arguments–in other words, only propositional evidence counts.  On the other hand, the reliabilist maintains that a belief is rationally held when (a) produced by reliable, properly functioning cognitive faculties, and (b) unaccompanied by overriding reasons to deem the belief irrational.  This implies that experiential evidence (e.g., religious experience) can justify a belief under the right conditions.  Properly basic beliefs are ones which can be justified or held without relying on other beliefs; for instance, perceptual belief is held on the basis of perceptual experience.  Likewise, Rea and Murray think that religious experience might justify a religious belief so long as it satisfies the criterion mentioned above.  What kinds of evidence might defeat our confidence in religious experiences?

Religious disagreement
Just as scientific theories explain sensory experience, Rea and Murray construe religious theories as collective attempts to explain religious experience.  A problem comes to light here: while scientific theories enjoy a fair level agreement, most religious theories are quite the opposite.  In light of this disagreement, one might ask some questions:

1. Do religious theories track truth about reality?

2. Are the arguments for any religious positions successful?

3. If not, should religious belief and practice be abandoned?

The religious skeptic generally replies in the negative to the first two questions, but affirms the last one.  Religious pluralism holds that no religion has a weighter share of (religious) truth than the others.  Pluralists maintain that many religions can capture isolated truths about reality, but none of the religions are superior in this sense.  The third view is religious exclusivism.  The exclusivist affirms what the skeptic denies: that religious theories track truth about reality.  He also affirms what the pluralist denies: that there is one objectively true religious narrative.  This doesn’t entail that all religious truths are known with certainty, but merely that we can have substantial knowledge of the truth about spiritual reality, and we can use this knowledge to judge disparate religious views.

Religious diversity poses a problem for exclusivists, and Rea and Murray undertake to show that this problem doesn’t render exclusivism unreasonable.  Putting aside pragmatic reasons for embracing pluralism, they examine how widespread religious disagreement might constitute evidence against exclusivity.  They conclude (as reliabilists) that it depends on whether or not disagreement should cast doubt on the believer’s reliability to judge the facts at hand.[1]  The mere fact of disagreement is insufficient.  Disagreement plus backgrounds beliefs (beliefs about the supernatural, morality, your own memory) work together to determine how heavily the evidence weighs.  Unfortunately, we don’t have adequate background beliefs about “a typical person’s ability to judge religious matters.”  More importantly, Christian background beliefs predict religious diversity.  So, the evidential force of religious diversity might be weakened.  Rea and Murray acknowledge that this only scratches the surface of the debate over exclusivity.

Is atheism irrational?
Alvin Plantinga has fashioned an argument which, if successful, renders atheism irrational.  Imagine that you’ve acquired evidence that your cognitive faculties are completely unreliable.  What should you make of the following proposition?

(R) Your cognitive faculties are reliable.

Well, firstly, you couldn’t have evidence for (R) that hasn’t already assumed (R), since reliable cognitive faculties are a prerequisite for assessing propositions about reliable cognitive faculties.  More importantly, if you have reason to reject (R), on the basis of the evidence you’ve acquired, then now you have a reason to doubt all of your beliefs, including the belief that (R) is true!  Defeating (R) defeats everything … resulting in global skepticism.

Why should we think that atheism gives us evidence of unreliable faculties?  According to Plantinga, evolutionary theory is concerned with survival and reproduction, and doesn’t give a fig for true beliefs.  Thus, conjoining naturalism with evolutionary theory leads us to believe that our cognitive faculties are unreliable.  Why?  The ability to arrive at truth about the world might be (and Plantinga would say is) irrelevant to the function of enabling us to survive and reproduce.  For example, consider a man who wants to die and he believes that the best way to secure death is to run away from tigers and mate frequently with women.  He’ll do just as well as the man who doesn’t want to die, and believes the best way to secure life is the same.  Thus, Plantinga argues that the probability of our faculties being reliable–given evolution and naturalism–is low.  But if so, then the atheist who accepts evolutionary theory and naturalism has a defeater for his belief that naturalism is true … and a defeater for all his other beliefs too.  Rea and Murray suggest that a more modest version of this argument might work: arguing that natural selection might produce some true beliefs, but why should we think our philosophical and religious beliefs are among these? [2]

1 There are “undercutting” and “opposing” defeaters.  It sounds like Rea and Murray are arguing that an undercutting defeater is needed, since presumably there is no opposing defeater for basic beliefs such as my belief that I was at work yesterday morning.  See this article for more information on defeaters.
2 See Plantinga’s paper “Naturalism Defeated.” Stephen Law has been working up a response.  Also, see Fitelson and Sober’s response.

Alvin Plantinga – Science and Religion: Are they Compatible? Part 3

What follows is part 3/3 of my transcription of Plantinga’s opening statement during his exchange with Daniel Dennett in 2009. The mp3 is available here. Also, Plantinga and Dennett co-authored a book shortly after this debate.

“Science and Religion: Where the Conflict Really Lies”

Speech delivered by Alvin Plantinga at the 2009 American Philosophical Association Central Division Conference.

---

Section 3: Naturalism versus Evolution

Naturalism comes in more than one variety.  Here, as I said, I take it to be the view that there is no such person as the God of the theistic religions or anything at all like God.  So taken, naturalism is not a religion.  Nevertheless, it is a crucial part of the naturalistic worldview which in turn plays at least one of the most important roles of a religion.  This worldview functions as a sort of myth in the technical sense of that term.  It offers a way of interpreting ourselves to ourselves; a way of understanding our origin and significance at the deep level of religion.  It tells us where we come from, what our prospects are, what our place in the universe is, whether there is life after death, and the like.  We could therefore say that it is a quasi-religion.  What I propose to argue next is that naturalism and current science are incompatible, so that there is a religion (or quasi-religion) science conflict sure enough, but it is between science and natural not science and theistic religion.  What I’ll argue is that naturalism is incompatible with evolution in the sense that one can’t rationally accept them both.  Since I’ve given this argument elsewhere I can be brief.

First, note that naturalists are all (or nearly all) materialists about human persons.  A human person is a material object through and through with no immaterial self or soul or subject.  For present purposes therefore, I’ll assimilate materialism to naturalism.  The central premises of the argument are as follows: where ‘N’ is naturalism, ‘E’ is current evolutionary theory, and ‘R’ is the proposition that our cognitive faculties are reliable.  The argument goes like this:

(1) The probability of R/N&E is low

(2) One who accept N&E concedes that 1 is true as a defeater for R

(3) This defeater can’t be itself defeated

(4) One who has a defeater for R, has a defeater for any belief he takes to be produced by her cognitive faculties including N&E itself

(5) Therefore, N&E is self-defeating and hence can’t be rationally accepted

These premises need defense, perhaps the first in particular; so, suppose there are beliefs.  This isn’t essential to the argument for (1), but it will facilitate a brief statement of it.    From the point of view of materialism, a brief will presumably be an event or structure in the nervous system—perhaps in the brain.   It will be a structure involving many neurons connected in various ways.  This structure will respond to input from other such structures: from sense organs and so on.  It may also send signals along a structure of nerves to muscles and glands thereby causing behavior.  Such a structure will have at least two kinds of properties.  On the one hand, it will have neurophysiological properties—called “NP properties”—specifying, for example: the number of neurons involved in the structure; the rate of fire in various parts of it; the change in rate of fire in one part in response to the change in rate of fire in another; the way in which it is connected with other structures, and muscles; and the like.  But if it is a belief, it will also have a property of quite a different sort: a mental property.  It will have a content.  It will be the belief that p for some proposition p.  NP properties are physical properties; having such and such a content is a mental property.  There are three ways in which, given materialism, mental and physical properties can be related.  First, non-reductive materialism: while mental properties can’t be reduced to physical properties, they supervene on them.  And take supervenience to be like this: properties of sort A supervene on properties of sort B just if necessarily, if entities x and y differ with respect to their A properties then they differ with respect to their B properties.  A necessity involved could be either broadly logical metaphysical necessity or nomological necessity—­giving us two varieties of supervenience: logical and nomological—and hence two possibilities as to the relation of mental properties to physical properties.  The third possibility for that relation is reductive materialism: according to which ever mental property is identical with some physical property.

Now, in order to avoid inter-specific chauvinism, suppose we think not about ourselves but about a population of creatures (perhaps in on of those other cosmoi proposed by inflationary scenarios) who resemble us in holding beliefs, changing beliefs, making inferences and so on.  Suppose also that naturalism holds for these creatures, and that they have come to be by the processes specified in contemporary evolutionary theory.  Now ask about P(R/N&E)—specified not to us but to them.  And consider that probability with respect to each of the three suggestions about the relation of mental and physical properties.

Consider first logical non-reductive materialism: call it ‘LNM.’  Mental properties are distinct from physical properties but supervene upon them where the necessity involved is broadly logical.  What is P(R/N&E&LNM)?  Well these creatures have evolved; we may therefore assume that their behavior is adaptive in their circumstances and that accordingly the neurophysiology producing that behavior is also adaptive.  But natural selection doesn’t give a fig about true belief just as such.  It rewards adaptive behavior and punishes maladaptive behavior, but doesn’t care about the truth of a belief.  As Patricia Churchman says, “Truth, whatever that is, definitely takes the hindmost.” [viii]

So truth in a particular belief B held by one of these creatures.  We may assume that B is adaptive and that its NP properties are adaptive.  But of course nothing so far follows about the truth of falsehood of the content that supervenes on these properties.  If the supervening content is true, excellent.  But if it is false that’s just as good.  Its falsehood in no way interferes with the adaptively of the NP properties.  We should assume therefore that the probability of that belief being true (given N&E and LNM) is about a half; but then the probability of their faculties being reliable would be low.  If you have a hundred independent beliefs and the probability of each is a half: the probability that three-fourths of them are true (which is a modest enough requirement for reliability) will be less than one out of a million.  So P(R/N&E&LNM) therefore is low.  But the same thing holds (and for the same reasons) for P(R/N&E&NMN)—where ‘NMN’ is the version of non-reductive materialism where mental properties supervene upon physical properties with nomological necessity.

That leaves reductive materialism, which we’ll call ‘RM’.  What is P(R/N&E&RM)?  Here the property of having such and such a content is identical with some physical property—­presumably a neurological property.  Again consider any particular belief B held by one of these creatures.  We may suppose that having this belief B is adaptive, and adaptive by virtue of its content as well as its other physical properties.  But once again, it doesn’t matter whether the content associated with B is true or false.  We may assume that the physical property identical with the property of having B’s content is adaptive.  The content associated with B is, of course, true or false.  If it happens to be false, this in no way compromises the adaptivity of B.  Once more then, we must suppose that the probability of that belief’s being true is about a half.  But then it would be unlikely that the cognitive faculties of these creatures are reliable.  It follows therefore that P(R/N&E) with respect to these hypothetical creatures is low.  But then of course the same goes for us: P(R/N&E) is low specified to us as well.

The next step to note is that anyone who sees that P(R/N&E) is low and also accept N&E has a defeater for R in her own case—a reason for rejecting R, for giving it up, for no longer believing it.  This defeater cannot itself be defeated.  That is because the defeater for this defeater would have to take the form of an argument; but, of course, one who accepts N&E will also have a defeater for the premises of this argument as well as for the proposition that if the premises are true so is the conclusion.  Another way to put it: any argument for R will be epistemically circular in that reliance on the argument presupposes that the conclusion of the argument is true.  But anyone who has a defeater for R has a defeater for any belief that has been produced by her cognitive faculties, including of course N&E itself.  Hence, one who accepts N&E and sees the truth of that first premise has a defeater for N&E.  N&E, therefore, is self-defeating and cannot rationally be accepted.  If so, however, there is a conflict between naturalism and evolution.  There conjunction cannot rationally be accepted.  Evolution, however, is one of the pillars of contemporary science.  Hence, there is a science/religion or perhaps science/quasi-religion conflict in the neighborhood of evolution alright.  But not between evolution and theistic religion.  The real conflict is between evolution (that pillar of contemporary science) and naturalism.  Thank you [audience applause].

---

viii Churchland, Patricia. “Epistemology in the Age of Neuroscience.” Journal of Philosophy. 84. (1987): 548.