Monday, September 14, 2009

The Bayesian Argument from Desire

A redated post.

A good place to start in making sense of the argument from desire is Peter Kreeft’s formulation.

1) Every natural, innate desire in us corresponds to some real object that can satisfy that desire.

2) But there exists in us a desire which nothing in time, nothing on earth, no creature can satisfy.

3) Therefore there must exist something more than time, earth and creatures, which can satisfy this desire.

4) This something is what people call "God" and "life with God forever."

A good deal of poetry and literature seem to support the second premise. Human beings are deeply dissatisfied even when all of their earthly needs are satisfied. On the face of things, the most difficult premise to defend is 1). How could we, without first knowing that Joy can be satisfied?

Kreeft responds: This is very easy to refute. We can and do come to a knowledge of universal truths, like "all humans are mortal," not by sense experience alone (for we can never sense all humans) but through abstracting the common universal essence or nature of humanity from the few specimens we do experience by our senses. We know that all humans are mortal because humanity, as such, involves mortality, it is the nature of a human being to be mortal; mortality follows necessarily from its having an animal body. We can understand that. We have the power of understanding, or intellectual intuition, or insight, in addition to the mental powers of sensation and calculation, which are the only two the nominalist and empiricist give us. (We share sensation with animals and calculation with computers; where is the distinctively human way of knowing for the empiricist and nominalist?)

But doesn’t this just mean “We just know?” Why shouldn’t natural unsatisfiable desires arise? This difficulty is especially acute when you look at the naturalistic world-view to which theism is opposed. The naturalistic atheist is prepared to accept a substantial amount of absurdity in human existence, at least if it is measured by the standards of expectations conditioned by theism. Consider the following comments by Keith Parsons here.

Bertrand Russell said that a soul’s habitation must be built on a firm foundation of unyielding despair. However, Parsons, in his debate with William Lane Craig, maintained that the despair is a despair from a theistically conditioned set of expectations concerning life’s meaningfulness.

The question I am asking is this: Is there any reason why a nontheist should be surprised that human beings have desires that are doomed to permanent frustration?

John Beversluis, in his critique of Lewis’s apologetics, suggests that in preface to The Pilgrim’s Regress Lewis offers a justification for 1 in the principle that nature does nothing in vain, but then offers no good reason why we should believe that nature does nothing in vain. In fact, it might be argued that the very principle itself presupposes a teleological understanding of the universe that presupposes theism, thus reducing the argument from desire to begging the question.

But I wonder if some version of the principle that nature does nothing in vain might be accepted by both parties in the debate. Beversluis says that the fact that we are hungry is no evidence that food exists, the actual discovery of food is the only thing that would suffice. But if we were to find creatures with, say, sexual desires, but no way of having sex, and which reproduced asexually, wouldn’t that conflict with out expectations? Wouldn’t biologists he shocked to find such a creature? Wouldn’t the existence of sexual desires be evidence that sex was at least possible or surely of some biological use, even if we did not see any actual mates for these creatures? To argue thus we would not need creationism; even evolutionary biologists would have to agree.

In an earlier post I wrote:

Why should we think that a natural desire within us would not exist unless it was satisfiable? Well, let us suppose that God and evolution are the main two explanations for why we have the desires that we have. We can understand easily why we have those desires if God has outfitted us with the desires that we have. These desires are God’s “calling card” whereby He draws us to Himself. But suppose evolution were the explanation, as it would have to be on naturalistic assumptions. It is possible, of course, that these desires should evolve, but should we expect this? Should we not expect that desires that don’t directly promote survival would be shoved out of the way by desires for food, clothing, and shelter, power, and strength, which do us so much more good from an immediate survival standpoint. If we didn’t know better, we should expect this meme to become extinct. On the face of things, we have something that obviously provides Bayesian confirmation for theism. We have something that is very likely on the theistic hypothesis, and perhaps compatible with atheism, but not very likely given atheism.

At least that’s what I’d like to think. But I do know that evolution is not perfectly efficient. If nature does nothing in vain, how do we account for the human appendix, an organ which now has no use other than to get infected and make money for doctors? In one sense, it is something nature did in vain, in that it doesn’t do anything for us now. On the other hand, I am told that it was used by our ancestors to digest raw meat, back before we learned how to cook. Should we expect human creatures in an atheistic world to desire an object that nothing on earth can satisfy?

This is Bayes’ theorem. H is the hypothesis, K is background knowledge and e is the evidence.

p(h/e & k) = p(h/k) x p(e/h & k)/p(e/k)

Suppose we haven’t considered the evidence concerning the human desire for the infinite, and so we consider this data as e. In using Bayes’ theorem, we begin by considering the probability of the hypothesis, in this case theism, on background knowledge alone. To see if the Bayesian argument from desire has any weight at all, let’s assume that we have a person who thinks that theism and atheism are equally likely. Bayesian theorists have tried in vain to find a method of determining objective antecedent probabilities. So let’s assume that p (h/k) = .5. The next question is how probability is the desire evidence to arise if theism is true. It seems that theism gives us a reason to suppose that these desires would be likely to arise in a theistic universe, especially if that universe were a Christian universe. On Christian theism God’s intention in creating humans is to fit them for eternity in God’s presence. As such, it stands to reason that we should find ourselves dissatisfied with worldly satisfactions. Let’s put the likelihood that we should long for the infinite given theism at .9. Now, what is the likelihood that infinite longings should arise on background knowledge alone. This is the hard part. If we don’t know whether theism is true or not, how likely are we to have desires like Lewis is talking about? I wouldn’t say that such desires couldn’t possibly arise in an atheistic world. Even though such desires seem to have limited evolutionary use, they could well be byproducts of features of human existence that do. But how likely would they arise in such a world? So long as the answer is “less likely than in a theistic world,” the presence of these desires confirms theism. Let’s say that, if we don’t know whether theism is true or not, the likelihood that these desires should arise is .7. Plugging these values into Bayes’ theorem, we go from .5 likelihood that theism is true to a .643 likelihood that theism is true. Thus, if these figures are correct, the argument from desire confirms theism.

I can see Bart Ehrman throwing a fit already. And I will admit that I don’t have supreme confidence in this argument. But I can’t help thinking that there must be something to the argument from desire, especially if the argument is presented in Bayesian terms.


Anonymous said...

For me, the problem is in premise 2. I don't have these desires which nothing can satisfy. I long for love, understanding, creativity, peace. These are earthly desires which can be satisfied but which often are imperfectly satisfied. So are people's desires for the physical - sex, shelter, food, water. People use technology to fulfill other desires which nothing can naturally satisfy: superhuman strength (weapons), the ability to fly (airplanes, etc), the ability to see through matter (x-rays), the ability to perfectly recall the past (photographs, writing).
Those desires which the writer on prosblogion calls "transcendent" seem to me to be simply natural human desires which he has stated in such a way as to make them seem that nothing on earth can satisfy them. This is misleading.

Those things he says we wish to transcend were: Self, Space, Time, Death, Finitude

Self: I assume he is speaking here of companionship, love and understanding from someone outside ourselves. Because this desire is not perfectly satisfied does not make it transcendent.

Space and time: Does this mean time travel? Being able to fly or to travel instantaneously? I don't think fantasies about being invisible or traveling through time in a nonlinear fashion indicate anything but the ability to imagine.

Death: We wish to live. We have a strong biological drive to live and we preserve our lives as long as we are able. This does not grant us eternal life. In fact, many old people are quite ready to die as they are worn out with living.

Finitude: Seems to me this falls under the above category. As long as we are enjoying life, we wish to prolong it. Living for eternity is something that most of us cannot accurately imagine, and if we did, we would not particularly like it.

So these desires are hardly transcendent, unless you or someone else can explain them in a meaningful way that shows they are not extensions of other earthly desires. Obviously the desires for love/companionship (in its many forms) and life (avoiding death) would be directly linked to survival.

Regarding the desire for an "infinite" amount of love, life and so on, humans are quite greedy about many things. We often eat more than we need, have more sex than is biologically necessary, develop addictions to various substances, buy more crap than we could conceivably use in one lifetime, live in bigger houses than we require from a functional standpoint and on and on.

I find this an unconvincing argument. The one feeling which I might conceivably call "transcendent" in some sense is the sense of awe I have when contemplating the majesty and beauty of the natural universe. But I don't think this "proves" the existence of God. It is merely recognition of um, the beauty and majesty of the universe.

Steven Carr said...

Kreeft says he knows that all humans are mortal. Did Elijah die? Did Enoch?

1 Thess. 4:15 'According to the Lord's own word, we tell you that we who are still alive, who are left till the coming of the Lord...'

If Kreeft knows that all humans will die, where does that leave Christianity, which teaches that not all humans die? Refuted, once again, by what everybody knows to be true?

And if Christianity is true, how did Kreeft come to the conclusion that all men are mortal?

BTW, does the Philosopher's Stone exist - seeing that so many people desired it for so long?

Steven Carr said...

What does it mean to say that people desire a god to exist?

Victor was complaining recently that there were people who desired god not to exist. Don't their desires get factored into the argument from desire?

Anonymous said...

I would say that the most valued experience of my life has been a longing for- and fleeting tastes of, an absolute perfection of beauty,and 'rightness' or 'goodness" .. I cannot find words to express it. It is simply always beyond expectation .. as though you have to be taken out of yourself and all things familiar.
which is misleading ... it is in the familiar as well .. they are transformed.
It seems to me that it is inpossible to hold in a mind or memory .. it must come so to speak, live into the perceptions and heart.
From where? i have no idea.
No doubt some would say it is a trick of the mind and it certainly involves a change in the body and brain to perceive it. Maybe it is just what some say they know in a drug experience. I don't know anything of that either.
I do not have belief in a personal god,or certainty about any sort of god, but this ... unspeakable overflowing bliss, has as its object of perception, a seemingly transcendent other, which may be simply the world when seen in its fullness. It may be total illusion.
but it is as much like 'godlikeness' as i would ever hope to know.

It does give a sense of worth to living which lingers. Is it just a biological necessity for creatures like us to survive?
I do not think so. I am inclined most to think that this universe, at the very least, is much more than materialsm gives it credit for.
And prepared to believe that it is an essential mystery and wonder beyond any easy explantion whatsoever.

sorry for that ouburst.

regards David.L

Jason Pratt said...

"es" has some pretty good refutals methinks. {shrug}

Steven's first remark is only clumsy: Kreeft's conclusion that men are naturally mortal has problems in regard to his intended application, I quite readily grant; but it's entirely specious to call in avowedly supernatural exception to natural human nature, as a defeater for a conclusion reached about natural human nature.

That being said, Steven does have a good point (maybe accidentally {wry g}) in his second comment: the argument from desire cuts both ways; as even C. S. Lewis knew to be true when he himself routinely (and properly) argued against the _anti_-theistic argument from theistic desire! (i.e. we only believe God exists because we want to believe it, per a routine anti-theistic argument from wish-fulfillment, but there are four varieties of wish-fulfillment and per Lewis they cancel one another out as explanatory factors.)

Going back to the report of Peter Kreeft's attempt at getting past a refutation of (1): the problem with this particular attempt, is _not_ that it only means "We just know" (though admittedly that would be a problem, too, if he attempted it as a parry--which may be why he tries something more impressive-looking instead {g}), but that he is appealing to a type of knowledge, by parallel, which was reached in a fashion that will _not_ work as a way of reaching the kind of knowledge he wants to defend for (1).

We can reach a conclusion that all humans will naturally die, by studying the natural structure of humanity and deducing how behaviors will naturally proceed (all other things being equal); thus reaching a reliable conclusion about what we can expect in natural process (other things being equal) for entities with similar biological processes. In other words, we can discover all humans are in fact (not just in typical human experience) naturally mortal.

But this is very different, not from the kind of knowledge PK needs to appeal to, in order to defend premise (1), but from any way we could reach a valid knowledge for reliably undergirding premise (1). What principle of natural operation (or whatever) have we discovered, that _confirms_ (a word I would not apply to likelihood estimates, by the way {wry g}) that every innate desire corresponds to some real object that can satisfy that desire? (If PK says, Victor didn't quote him saying so--what Victor does quote, is not at all a proper answer to the problem of grounding (1).)

Finishing up, Bart Ehrman isn't the only one who would be throwing a fit over the Bayesian attempt here--I have a lot of criticisms on it myself (though possibly not the same as his). Ignoring various technical crits, I will only say that the result is frankly nothing that needed a complex operation of logical algebra to reach: if we decide this kind of desire is likely given the truth of theism, then of course the existence of the desire is going to (or at least should) add to our estimate of theism's likely truth. This absolutely should not be called "confirming" theism, though--that term should be reserved for deductive results, not increases of perceived likelihood.

If (let's say) Keith Parson's initial likelihood of theism being true is very low (and, by the way, Bayesian likelihoods should not be quantified either, on pain of instantly introducing persistant category error); then he could easily introduce this into his likelihood estimate (assuming he agreed it was proper to do so in the first place), and still have a very low likelihood estimate of theism being true; especially since the result totally depends on what kind of intuitive weight he himself agrees to accept concerning the new evidence 'e' (another reason why numerical figures should not be introduced into the operation). No one would ever call _this_ result 'confirmation'; but in principle it's not a different kind of result than that which Victor reached.

Jason Pratt

Victor Reppert said...

The whole idea of a Bayesian argument is that of course, is that it allows for differences of antecedent probabilities. If Parsons starts off at a .001 for theism, and assesses the probability of the desire arising given theism at .9, and assesses the probability of the desire arising whether theism is true or not at .7, then he gets a final result of .00166 for theism. We have some confirmation, but far from enough to overthrow his atheism.

Of course, as the knight of Monty Python said about Camelot, it's only a model.

Anonymous said...

Dear friends,

I was interested by Victor's respondent who denied having any unsatisfiable desires. He admitted that certain desires had never been satisfied perfectly, but maintained that they could be in theory, or that the satisfactions he could find in this life were good enough. How does one respond to this line? It's rather like trying to convince the dwarfs in The Last Battle that they aren't in a stable!

One conclusion might be that the argument from desire just doesn't work with a certain type of person. Some of us are just too emotionally undeveloped--or jaded--to be susceptible. But I would suggest that we make a mistake by taking such people's statements at face value. Solomon tells us that "God has set eternity in their heart." Either Scripture is wrong or the denial of transcendant desire is a smokescreen, a defense mechanism to protect dwarfish atheists from reality.

A person who is still human is not in fact satisfied by the temporal and physical, however hard he tries to convince himself that he is. But you probably can't argue him out of his position. You can only try to arouse the desire, to fan it to the point where he cannot ignore it any more. And the best way to do that might be to talk about the foretastes of fulfillment we have already been granted in Christ, or just to live a life of transcendant openness to Joy before him.

If you can get him to read Thomas Traherne's Five Centuries of Meditations, it wouldn't hurt. "Things unknown have a secret influence on the soul, and like the center of the earth unseen violently attract it. We love we know not what, and therefore everything allures us. . . . Do you not feel yourself drawn by the expectation of some Great Thing? . . . You never enjoy the world aright till you see how a [grain of] sand exhibiteth the wisdom and power of God. . . . You never enjoy the world aright till the sea itself floweth in your veins, till you are clothed with the heavens and crowned with the stars. . . . Infinite wants satisfied produce infinite joys. . . . You must want like a God that you may be satisfied like God. Were you not made in his image?"

Lewis learned the argument from desire from Augustine's Trinity-shaped vacuum and his heart that was "restless until it rest in Thee," as developed by Traherne and MacDonald. It will have a certain logical cogency--which Victor well analyzes--to those in whom Desire has been sufficiently aroused. The best service those earlier writers--and Lewis himself--may do us is to fan that flame. In it, let us burn.

From the Falls of Henneth Annun,


Donald T. Williams, PhD
Prof. of English, Toccoa Falls College
Poetry Editor, The Lamp-Posy

"To think well is to serve God in the interior court." -- Thomas Traherne

Anonymous said...

Can't help thinking that the desire for God not to exist is just as confirming of theism as the desire for him.
In a world without water, a fear of downing would be just as puzzling as a desire to swim.

Anonymous said...

Some people's comments suggest that they think the desire in the argument from desire is a desire that God exist. Well sort of.
When people are hungry they don't desire that food exist, they desire food! Of course they couldn't possibly get any food unless it exists so the existence of food is a prerequisit of the satisfaction of the desire.
Similarly, the desire for God is a desire for communion with Him, a desire which cannot be satisfied unless He exists.

Also, the argument doesn't require that everyone has the desire in question. It requires that the desire be natural. Humans naturally have five fingers on each hand, but some people have less than this (some have more). So 'es' not having the desire in question isn't an insuperable barrier to the argument. Even so, I do wonder whether he is really correct in his assessment of his own desires here. He certainly could be wrong.

Like many people, I think the difficult premise is "The world, including anything supernatural, is such that all natural desires can be satisfied" (or however we want to phrase it).

We could certainly substitute this with "Either life is absurd or the world is such that all natural desires can be satisfied".

Correctly read this is analytic. Of course the atheist may just shrug his shoulders and accept the absurdity ... but what about the agnostic?

I'd be interested in seeing anyone's justifications for the non-analytic versions of the premise.

Joshua Blanchard said...

I also like the argument from desire. Like a previous commenter, I think it counts as a confirmation argument in Swinburne's sense that, although not proving by itself, adds to the likelihood of theism.

That being said, this particular phenomenon is open to relatively plausible naturalistic explanations. An unsatisfied desire could arise as an evolutionary byproduct. Indeed those evolutionary theorists in the byproduct camp about religion would surely say this.

But even if it doesn't arise as an evolutionary byproduct, the naturalist might be tempted to say that this desire is satisfied, i.e. by religion. As a man-made phenomenon, religion is a plausible candidate for something that satisfies Lewis-type desire which is not necessarily supernatural. Even though religion does not satisfy desire for things like "eternal life" in the strict sense, religious belief and practice surely grants participants a deep sense of well-being and satisfaction, directly relevant to these desires.

Ken said...

Theism is about denial, not desire.

Unknown said...

It's perfectly consistent for someone to desire that God not exist, and to have a desire that nothing on earth can satisfy...

Megan said...

"Can't help thinking that the desire for God not to exist is just as confirming of theism as the desire for him.
In a world without water, a fear of downing would be just as puzzling as a desire to swim."

Perhaps the fear of drowning would be puzzling, but not the desire to not drown. In a world without water, if someone conjured up an idea of water, and an idea of how the human body would function when submerged for long periods of time, it seems like they would say that they desired to not drown. This seems perfectly acceptable to claim, even while not retaining a belief in water, and even if there was, in fact, no such thing as water. A la Quine, it seems as though we can still speak of things without then making an ontological commitment to them.