Tuesday, September 29, 2009

A Robin Collins paper on design.


Clayton Littlejohn said...

Correct me if I'm wrong, but isn't the conclusion of this argument incredibly weak?

Suppose I find that there's 23 pieces of fruit in my fridge and 23 pieces of change in my change bowl. One hypothesis is that a ninja obsessed with the number 23 sneaked into my apartment to make it so. The conditional probability of the fruit and change being this way on the hypothesis that there is such a ninja is quite high. My observations, however, seem to give me little reason to believe that there is such a ninja (as we ordinarily think of reasons for belief) because it seems that the prior probability that there is such a character is quite low. Can't the atheist say something similar in response to Collins' argument? The prior probability that there is a god is quite low so the appearance of design (even if we grant that the conditional probability of the appearance of design on the hypothesis that there is a fine tuning deity is high) does not make it unreasonable to continue to be an atheist.

Mark Frank said...

I agree with Clayton. He bases the whole thing on likelihoodism which he calls prime principle of confirmation. Comparative likelihoods is a good way of comparing the support that evidence gives to hypotheses but he is quote wrong when he writes:

"there does not appear to be any case of recognizably good reasoning that violate this principle"

whether it is Clayton's example or gremlins the attic there appear to be plenty of cases. I think the key is that before you can use comparative likelihoods you have to have some reason for supposing the hypothesis to be plausible other than the evidence you are taking into account. Otherwise you can always offer the hypothesis that "an omnipotent being wanted it to happen" for any observation and it will have the greatest likelihood of all hypotheses.

Gordon Knight said...

How does one evaluate the prior probability of theism? Any probabilistic argument would have to take in all the evidence. Considerations re "fine tuning" would only be part of the mix.

The argument from evil may lower the probability of theism. But this argument only purports to show that that the traditional theistic God is improbable, it does not weigh against any sort of designer.

I guess I agree that a full defense of any probabilistic argument for theism does require considering all the relevant evidence--as Swinburne tried to do in The Existence of God. But I don't think you can just assume theism is improbable prior to adducing evidence for its improbability.

Re: Gremlins

Suppose you run into the number 23 all over the place. not just once or twice, but every time tyou could your change you get 23, everytime you look at your fruit bowl you get 23 etc.. Further, suppose you know that there are people who claim to have talked to a dude who claims to be a gremlin with powers to make the number 23 pop up all over the place. The dude calls himself "The Gremlin"

Doctor Logic said...

In South America, there's a very rare orchid. There's also an equally rare moth with a proboscis that is extremely long. Only this one moth can fertilize the orchid.

What are the odds that there are two extremely rare entities finely tuned to work together in this way?

If we assume they are independent, then the probability looks terribly small. Of course, they are not independent. The thing about biospheres on Mars is that we know enough about the natural world to know about the independence of the variables. In the case of the parameters universe, not so. But unified field theories hold out that promise.

Moreover, design doesn't even solve the problem. Why does God WANT to create a universe like ours? Given God's powers, universes aren't even required for agents such as ourselves to exist. Nor would agents such as ourselves be interesting to an omniscient God. So, a generic God has to be fine-tuned at least as much as our universe in order to "explain" it. If our universe is a one in a trillion, why propose in a one-in-a-trillion God to explain it? You're not making any headway. Just making things more complex.

And if your theory of God is good enough to predict the constants of our universe, you ought to be able to predict something else. The fact that theism predicts nothing at all proves that it has no explanatory power.

God is no more explanatory than a Theory of Everything that I haven't formulated. Both will explain the universe by definition, once we can formulate them in a predictive manner. But just as I don't know the formula for the Theory of Everything, I don't know the mind of God. As it stands, they are equally predictive (i.e., not at all). But you don't see naturalists walking around saying everything has been explained by the Theory of Everything.

Clayton Littlejohn said...

"How does one evaluate the prior probability of theism?"

That's a good question. It's a question that someone offering an argument should work out.

I think that you might be running two things together in your comment. The first is the requirement of total evidence. The second has to do with assigning some prior probability to a hypothesis. My point had to do with prior probabilities and I think that's an issue that arises even when we make sure to look at the conditional probability of a hypothesis on all the evidence rather than just some.

The reason that the observations of 23 objects seems to give no reason to believe in my ninja (in the colloquial sense of 'reason to believe') is that the prior probability of a character with these powers and these motives seems really, really low. So, the discovery of just the evidence that you'd expect to find seems to give little reason to believe this character is real.

Anonymous said...

If it's admitted that the prior probability of theism is not low (or at the very least, cannot be ascertained) that wreaks havoc on the atheist response to design arguments across the board. Mind is the one thing we're absolutely certain of existing, and (our own) design nearly occupies a similar level of certainty. The existence of 'totally unguided things that just have lucky and utterly unforeseen combinations and results' is utterly unverifiable and vastly more foreign to the mind than, simply, some other mind as an ultimate source or even ground of being.

It gets worse when one starts thinking about ideas like Nick Bostrom's simulation hypothesis, etc. Minds are just too darn familiar, have too much potential even for we limited beings, etc. Showing how unlikely minds seem to be given 'chance', and realizing what minds of all kinds are fundamentally capable of, pushes the 'reasonable view' chip into the hands of the theist/proponent of design.

Gordon Knight said...

One should add that the unlikeliness of mind is given naturalistism. There are atheistic philosophical views that hold mind to be central to the universe (Mctaggart, Bradley, Hegel probably)

If one is serious about consciousness, one should be a theist, an idealist, or both.

I am open to being persuaded, but I find the claim that theism is improbable independent of any relevant data hard to understand.

Doctor Logic said...


The existence of 'totally unguided things that just have lucky and utterly unforeseen combinations and results' is utterly unverifiable and vastly more foreign to the mind than, simply, some other mind as an ultimate source or even ground of being.

Incorrect. Evolution is verifiable. Indeed, it has been verified.

Design doesn't predict evolution. It doesn't predict descent. It doesn't predict common descent. It doesn't predict common biology or architecture. It doesn't predict extinctions. It doesn't predict survival of the fittest. Evolution does predict these things. Design predicts things like manufacturing, and a vastly richer set of possibilities for life than evolution does.

This is why a designer God is effectively ruled out by evolution. Of the trillions of ways of designing life, God chose the one way compatible with naturalistic evolution?

We can agree on one thing: set the priors at 50/50. The likelihood of design is minuscule once the evidence for evolution is factored in.

Doctor Logic said...

Theist: "Suppose there is a God who (for unknown reasons) wills to create X. This God explains X."

Materialist: "Suppose there is a theory that (if we could write down its formula and calculate) would deterministically predict X. This theory explains X."

That's what comparable claims would look like. In reality, neither of these vague theories explains X.

If you've got no predictions, you don't even have a theory. Instead, you've got a name for theory you would like to have.

God can't explain the universe because there's no reason for God to make a universe like ours. Sure, a God COULD make the universe as we see it, but I can just as easily say that an ultimate physical theory COULD necessitate the universe as we see it.

The issue isn't possibility - they both have equal spheres of possibility. The issue is probability, and I have never seen a theist justify God's choice of this universe over, say, no universe or a radically different universe.

Anonymous said...


What a bunch of crap. Evolution is not some magical atheist non-design fairy. It's yet one more design method, and is entirely compatible with design.

Look at the way you're formulating your question. "Design doesn't predict evolution. But evolution predicts evolution." Hello, what? You claimed design doesn't predict anything, period. I may as well say "Theistic evolution predicts evolution. Many forms of intelligent design predicts evolution. But atheism doesn't predict evolution, therefore the likelihood of atheism/a designerless universe is vanishingly small once evolution is figured in."

I know evolution is the closest thing to a fetish a lot of atheists have, but surprise: It's amenable to teleology, and always has been. Right back to Asa Gray and Alfred Russell Wallace, and many since then. And considering evolution has been and is used by programmers and engineers, the proof of "evolution as a design tool" is available in abundance. Evolution as un-designed? There's no evidence for such - it's always been either axiomatic, or an assumption.

So, sorry. Even evolution fits better in the design category than the naturalism category. It just happens to be a naturalist favorite due to its opposition to YEC.

(PS: Who agreed to 'set the priors' to 50/50? Stop making up nonsense out of desperation.)

Gordon Knight said...

If my choice to write this down explains why I am writing this, then there are cases of explanation that have nothing to do with predictability.

Of course you can ask why I decide to write this, and there are various motives. But, arguably, the motives do not determine. So there is no predictability in free actions.

So perhaps the question of whether God can be an explanation relies on the prior question of whether some libertarian account of action (for human beings, which we have greater access to than God), is true.

Doctor Logic said...


You would see my point if your knickers weren't twisted so tightly.

Some questions while you try to get them untwisted.

1) Why don't we evolve cars? Is there any reason why we might not want, say, cars hunt animals for fuel?

2) Suppose I ask you to design a refrigerator. However, I tell you that you must use descent, and you can't use any method other than evolution. Does this constraint increase or decrease your design choices? Are there more ways of designing fridges with evolution alone?

A designer has a choice at every step of the process to use evolution, traditional design, descent, manufacturing, for some utility to the designer. A design as complex as life on a planet could see each of these methods used countless times. There are countless permutations available to a generic intelligent designer.

However, there's only ONE way for life to arise given what we know of naturalistic theories, an, gosh darn it, wouldn't you know it, that's exactly what is observed. BTW, naturalistic theories lead to one utility in the evolutionary design: survival. Amazing that this is the only apparent utility discovered, huh?

Doctor Logic said...


If my choice to write this down explains why I am writing this, then there are cases of explanation that have nothing to do with predictability.

But that's not an explanation for why you wrote this instead of not writing it. You're saying that no outside force moved your hand against your will to type the characters (or held you down so you could not type the characters).

In other words, the will is not an explanation for any action in particular. The explanation for any particular action has to go beyond this.

So if God creates the universe, his will doesn't explain his particular act. He could just as well choose not to create a universe.

(Thinking about it, if no one can force God's hand, God's will seems even less explanatory than your will.)

Will in this context is like nomological necessity in a naturalistic theory. Suppose I said "The laws of physics necessarily made this ball fall up instead of down." Would that explain the falling up? Sure, there could be an unknown law of physics operating in this case, under these conditions, and that law would necessitate the ball falling up, but is generic "necessity of law" explanatory by itself? I don't think it is.

Anonymous said...


No twisted knickers here. I just don't have patience for inanity and crap. Not my fault you're dealing this out in spades here.

1) For the same reason we don't 'grow houses' by planting a seed in the ground. Not because there's no scenario in which it could be desired, whether for purposes of design or art (in this case, art comes to mind as far more likely), but we don't have that kind of technology onhand. We're designers with limitations.

Why not ask why we don't 'evolve transistors'? Oh, wait. That's because we're freaking doing it.

2) Depends what you mean by 'with descent', much less 'design choices', now doesn't it?

Good Darwin, man, look at your reasoning here! There are plenty of ways other than evolution for "life to arise" given atheism - ask any multiverse proponent and they'll admit everything from boltzmann brains to otherwise, and that's just one type of examples! All you're telling me here is 'well, naturalism is compatible with evolution, and evolution is a popular theory, ergo naturalism is preferred'. It compares to my 'theistic evolution' response.

As for 'survival', you apparently don't know how to divorce metaphysics from science. Go read what Conway Morris has to say about the outcomes of evolution. Or Margulis. Lovelock. Or, if you want to have some fun with common descent at the cellular level, Woese.

Good God, you'd think if someone were going to pop one off about evolution they'd at least know about it. This has to be the most inane argument I've heard since 'God doesn't because the universe is way too large'.

Doctor Logic said...


You're twisting your knickers the wrong way.

The neo-Darwinian synthesis makes predictions, and those predictions have been confirmed. There's one lineage of common descent, just as predicted. DNA verifies it.

That's what you're up against. Not naturalism as a whole, but a naturalistic theory.

The point is not whether a designer could use evolutionary algorithms to design something. That happens today. We know evolution works, and it's one technique in a large toolkit of design techniques.

If I'm designing something large, e.g., an airliner, i might use evolutionary algorithms to design the wing, but use traditional techniques to design the seats. But even when I use evolutionary techniques to design the wings, I don't make wings that actually breed. That would be counterproductive. I simulate evolution without breeding. I don't need to design a reproductive system into the wing to make a wing.

Now it should be obvious that, as a designer, I have far more flexibility in designing something when I'm not limiting myself to evolution with reproductive descent. It is extremely limiting to have to incorporate reproduction into a system when all I care about is the end product. There are far fewer ways to design a thing using purely evolution with reproductive descent than there are with the full design toolkit that includes traditional design techniques, simulation, etc.

And it's worse. At every step of my design, I can switch from one technique to another. I can switch from one material to another. I can mix and match power sources. I could make silicon bunnies and nuclear powered deer. And frogs can be manufactured by frog factories that are not frogs.

I present you with a deck of cards. I deal 2,3,4,5,6 of clubs off the top of the deck. Is the deck sorted ascending by suit and rank, or shuffled?

There are 310 million more ways the cards could have come out had it been shuffled, but only one way it could come out if sorted. That's why we would be foolish to think that the deck was shuffled. Sure it MIGHT have been shuffled, but it is irrational to think it actually was.

This is why design is ruled out. Of the trillions of ways of designing life on this planet, the designer picked the one peculiar way that looks identical to unguided evolution? It is irrational to believe in a designer.