Yep, I didn’t see the edits. No problem.

I don’t think your edits change anything in my comment. I’m comparing class (1) (or class (1) plus class (3)) to class (2), and calling the Christian restriction of God to class (3) worlds an unjustified fine-tuning.

(3b) may indeed be an empty set. That’s fine by me. Makes the argument simpler.

I apologize for making changes after the first posting.

I’m basically agreeing with you that class (3) probably looks identical to class (1). My point is that

A) the number of worlds in class (1) is greater than or equal to the number of worlds in class (3), and

B) the number of worlds in class (2) is many orders of magnitude greater than the number of worlds in classes (1) and (3) put together.

So the argument is that, all things being equal, if God designed us, it’s massively more likely we would find ourselves in a world in class (2) than class (3).

Now, it’s perfectly easy to imagine a God who wants to put us in a class (3) world. In that case, all things would not be equal. However, you don’t have a good justification for believing this is the case. The interpretation that we’re in a class (3) world is a new invention on the part of theologians to explain the inconvenient fact that it looks like we’re in a class (1) world.

Indeed, before Darwin, theologians weren’t saying that the world wasn’t designed. They didn’t react to Darwin as if Darwin was confirming their expectations about God. They were horrified at the prospect that their world, which looked so obviously designed, wasn’t designed after all.

If there was any suggestion before Darwin that God was hiding, it was probably as a response to the problem of evil, which dates back to the Ancient Greeks or before. What we see again and again is theology reacting in hindsight to reality, not reality matching theology’s predictions.

Again, reacting to reality by fine-tuning your theory is okay as long as it gets you some predictions. Scientists do this all the time. They look at a system, observe something about it, and then devise a model of that system. The thing that proves they’re not deluding themselves is that fact that the model makes predictions. If the model is wrong, it won’t stand. If it does stand, it tells us things we would not otherwise know.

But theology is always looking backward. You cannot tell me anything about future discoveries, yet you use scientific discoveries to tune your theory of God. Nothing is inconsistent with your God theory because your theory takes no risks with truth. It takes plenty of risks with consequence, but that’s something different!

If you’re so confident that God wants us to live in a class (3) world, why wouldn’t you predict that evolutionary biology would be highly successful? If we get proof of design, that moves us into class (2), and God doesn’t want that. Of course, if that happened, theologians would invent (in hindsight as always) some rationale that explains why God changed his mind or why today’s Christians were so wrong in the past.

The preceding comment may be taken as an argument that
(A) If this is a class (3) world, then
(B) There are no incompatibilities between evolutionary theory and what God has done.

If evolutionary theory were perfectly demonstrated, then that would be true, I think (I’m still exploring here), and we would conclude that something like theistic evolution must be true (or we live in a class (1) possible world). But it is not just God’s activities that are ambiguous in our world. The success of evolutionary theory is still in question, too.

The question is best summarized in this, which I’m quoting at length.

Evolutionary biology and neuroscience make specific predictions about what we have to see. We have to see common descent, we have to see a progression of species, we have to see central nervous systems with causal links, we have to see brains with computing ability, etc. Dualism and ID do not predict these things. They are perfectly happy with other things. God could have created a world in which we think but don’t have brains, or in which species aren’t connected by common descent or common composition (e.g., rabbits could be robots).

But ID advocates and dualists simply pretend that they can fine-tune their theories (eliminating the robot rabbits and brainless thinkers) without paying a penalty.

One more thing I want to add: an explanation for why supernatural claims will never gain traction. What does theory A have to do to catch up with theory B?

To catch up, theory A has to be specific. Theory A has to be consistent with something narrow where theory B is consistent with something broad. Then you do the experiment, and if theory A’s possibilities are vindicated, theory A gains traction over theory A. Of course, the strength of theory A’s prediction has to be weighed against the strength of B’s, so A can win a round without winning the match, but at least A can come back.

The problem with supernatural claims is that they generally do not make predictions. Suppose God designed the species. In order to convince us of this, you need to create a predictive model of God that says what he would design and why. You have to set things up so that evolution could do anything, but God would do something specific. You’ll never get there unless you’re willing to talk about the designer.

doctor(logic) is proposing something that can be handled through the logic of possible worlds. (For those who are unfamiliar with this approach to logic, please see the Wikipedia introduction. A “possible world” here means something different than in ordinary language.)

Let us consider three classes of possible worlds containing complex, intelligent life:
(1) one in which there is no God, or no God active in nature;
(2) one in which God exists, acts, and makes his authorship of his actions unequivocally knowable; and
(3) one class in which God acts in nature, but his authorship of such activities is ambiguous, such that a human observer could have freedom with respect to his or her views toward God: the person could conclude either that God existed or that God did not exist.

Now the question is whether or not our world is a member of class (1) or class (3) among possible worlds. The atheist or agnostic might phrase it this way: we want to know whether we live in a possible world of class (1) or class (3). In order to have some confidence regarding the answer, the atheist (and the theist) wants to set up some distinguishing characteristics that would differentiate (1) from (3).

doctor(logic) finds it rather over-convenient (for the theist) that the world we live in is one in which certain features of life are under-predicted by theism. That is, given the full range of class (3) possible worlds, a large proportion of them would likely have characteristics that are more obviously incompatible with evolutionary explanations than those of the actual world in which we live. Some unknown and inscrutable proportion of class (3) worlds would be completely incompatible with evolutionary explanations, whereas presumably all class (1) worlds would be compatible with evolution. (Throughout this comment, by “evolution” I am referring specifically to naturalistic, undirected evolution.)

The implication is that there are at least two subclasses in class (3):
(3a) Possible worlds in which life is generally compatible with what is taken to be true so far concerning evolutionary theory, and
(3b) Possible worlds in which life is generally not compatible with what is taken to be true so far concerning evolutionary theory.

doctor(logic) says that the existence of life that is compatible with what we know so far about evolution supports class (1) possible worlds. I think it’s a fair extension of what he has said to add that it supports class (3a) possible worlds but not (3b) possible worlds. The ratio of (3b) to (3a) possible worlds is unknown, inscrutable, but likely to be so large that the actual existence of a (3a) world is unlikely with respect to a class (1) world. If we are postulating a class (3) world, we should be surprised that what we see is not a (3b) class world.

But it seems to me that the class of (3b) worlds may just be an empty set.

If God had set out to create a class (3b) possible world in which life was clearly incompatible with evolution (as doctor(logic) has suggested he might have done), then God would have created a class (2) world instead. The existence of natural features that are so incompatible with natural processes would require or demand assent to the existence and action of God.

So the probability calculus doctor(logic) has built on robot rabbits and brainless thinkers (examples by which, using my new terminology, he illustrated a class (3b) possible world), is rendered invalid.

P.S. In discussing God and possible worlds, some have argued that if God exists in any possible world (as philosophers use the term) then in that world he exists necessarily, from the definition of God. And if in that possible world he exists necessarily, then necessarily he exists in all possible worlds. I think this argument carries considerable force. If any reader wonders how I can speak of a “possible world with no God,” I would say that I have chosen to speak in such a way not because I think it possible for God to exist in some “worlds” and not in others. I have chosen to speak this way only because it has been helpful for me in setting forth this particular argument.

Have you ever had a gut feeling that X was true, only later to discover that it wasn’t true? If so, how is this different? It seems to me that the only difference is that no evidence can challenge your gut belief.

doctor(logic), yes, I have had that experience. What’s different this time is that after hour upon hour upon hour of study and debate, I have encountered nothing to show me that it isn’t true. The Christian message “holds together,” it is rationally coherent, it’s connected to real history.

Why are you reacting so negatively to the word ‘magic’? How does God’s power differ from magic?

Magic is often defined as man’s attempt to manipulate the spiritual world for man’s own ends. Alternatively, it is sham or trickery. How do you define magic?

So learning English, and associating words with familiar experiences had nothing to do with it?

I’ll yield to your point there. The perception of God a direct sensory perception; it too requires some context for correct interpretation. Hence Scripture, theology, etc. But that does not eliminate the direct perception of God, through the Holy Spirit, as part of the chain of knowledge.

Two things. First, prior probabilities generally refers to our initial confidence in the various theories. If the theories are supernatural design versus evolution, the priors would be our initial confidence in each theory. Let’s just say it’s 50/50.

Wrong if. The theories in question, as I read it, are various theories about how God would do things. Why wouldn’t he make a robot rabbit, for example? You say as much in your paragraph following the one I’ve just quoted, and when you say, “Suppose your model of the designer had some strong constraints.”

Alas, ID folk insist on not constraining the designer, probably because they think he’s God, and God has no constraints. Also, any constraints on God narrow enough to force him to make our world look evolved are going to result in a lot of other predictions that might prove inconvenient.

It’s not so clear-cut as that. When ID scientists do ID as science, they do not identify the designer. When ID theorists do ID as philosophy, they often identify the designer. I’m sure you’ve seen this before.

The rest of the science world wants ID to name the designer for reasons such as you have just mentioned. The rest of the science world wants ID never to name the designer because if it’s God, then “that’s not science!” Whichever way we approach it you’re not going to like it. Even though there is a nuanced and thoughtful third option, we seem frequently to get gored with whichever horn of the supposed dilemma is most convenient at the time.

Paul, I’m going to answer your question in a main blog post and see if others will add to it.

Bayesian probabilities include a term for prior probabilities based on background knowledge. Your equation, as I read it, calls for a background knowledge estimate of how probable it is that God would create a world like this one as opposed to another kind of world. That’s the number you need to insert into the equation to make it valid and complete, I believe, unless I’m misreading you completely. So I’m saying you don’t have that prior probability estimate, and you have no place applying Bayesian analysis in this case.

Two things. First, prior probabilities generally refers to our initial confidence in the various theories. If the theories are supernatural design versus evolution, the priors would be our initial confidence in each theory. Let’s just say it’s 50/50.

The second thing you are referring to is the distribution of probabilities for observations given each theory. Basically, what we want to know is the probability of finding the world to look evolved when it was actually supernaturally designed. You are saying, and I agree, that we cannot determine this probability with much precision. We probably cannot even get an accurate assessment of its order of magnitude! However, it is patently obvious that the probability of finding the world looking evolved after it has been supernaturally designed is many orders of magnitude smaller than finding a world not evolved. Why? Because making a world look evolved places a huge constraint on design. Assuming supernatural design, there are many more worlds with N species in them that don’t look evolved, than there are worlds with N species in them that do look evolved.

In modal reasoning (which I think is heavily abused, especially when it refuses to say what necessity means), there are many more supernaturally designed worlds that don’t look like evolution than there are supernaturally designed worlds that do look like evolution. Just look at this one little example. Suppose there are N species. In an evolved world, all N are related by common descent, and the differences between the N species are limited by time, energy, and resource constraints. However, the supernaturally designed worlds have no such constraint. For every world that looks evolved, there is another world in which species Ni is not related to Nj, or in which species Ni and more distant than Nj. That means that there are at least N^2 more worlds in the designed space, and that’s only considering 2 individual species being different. Given that there are millions of species, we’re already looking at millions more designed worlds that look designed than look evolved. The actual number is probably something closer to N! more unconstrained worlds than evolutionarily constrained worlds. (A million factorial would give evolution a 5 million order-of-magitude advantage.)

Can theism improve on such dismal performance? Yep. Suppose that your model of the designer has some strong constraints. In that case, the number of worlds this designer makes that look evolved might be relatively higher than the number of worlds that don’t look evolved. Such a theory of the designer could be far more competitive with evolution, and lag by less than an order of magnitude. Alas, ID folk insist on not constraining the designer, probably because they think he’s God, and God has no constraints. Also, any constraints on God narrow enough to force him to make our world look evolved are going to result in a lot of other predictions that might prove inconvenient.

Where does that leave us? If you’re not going to put in any constraints on the designer, then I’m justified in assigning each possible designed world equal weight. All things are equal. Consequently, belief in unconstrained design is ruled out by many many orders of magnitude.

Why does Paul often get straight to the point in one line while it takes me 15 paragraphs?

Like Paul, I want to know what this direct knowledge is like. It sounds to me exactly like personal bias. That is, you have a gut feeling that you’re right.

Have you ever had a gut feeling that X was true, only later to discover that it wasn’t true? If so, how is this different? It seems to me that the only difference is that no evidence can challenge your gut belief.

Why are you reacting so negatively to the word ‘magic’? How does God’s power differ from magic?

I just had a direct experience of reading the sentences, “The only knowledge we get (without inference) from direct experience is knowledge of the direct experience. It’s trivial.” I gained knowledge from that reading, directly.

So learning English, and associating words with familiar experiences had nothing to do with it?

Doesn’t language learning require Bayesian analysis, albeit informal analysis? Suppose you can’t speak French, and you’re dropped on an alien world where only French is spoken. You see a man look up at the sky at night, point and say étoile. You develop theories about the meaning of this word. It might mean dark, night, cloud, star, moon, moonlight, Milky Way, planet, Mars, space, God, etc. How will you distinguish between the theories? You’ll perform tests. If people say étoile when it’s cloudy, it’s less likely they are referring to astronomical objects. If the night is moonless, you’ll think it more likely that étoile does not mean Moon. Anyway, you know how the story goes.

The same applies when we learn our native tongue. That means that interpreting a sentence is an inference from past experiences, essentially an approximately Bayesian inference.

There’s little or no knowledge that doesn’t arrive by this approximate Bayesian method. Direct sensations are assumed to be knowledge by virtue of an assumption that they are knowledge. However, direct sensations have no implications. Seeing red by itself doesn’t imply anything except that I am seeing red. Implications of a sensation are always derived by Bayesian reasoning.

Except, we are supposed to believe, for your “knowledge” of God. You somehow gain some direct knowledge of God (which has, we presume, lots of consequences and implications) without that knowledge being an inference. How does that work? It would be the only knowledge of its kind.

But it is also more than an impression. It is the Spirit bearing witness with our spirit (Romans 8:16) to the fact that we are children of God. God is not some deaf dumb blind mute ghost. What he wants to communicate, he can communicate. Therefore this is not applicable to the situation.

I guess this brings up the same question Paul asked, because I don’t know what this means. It sounds circular to me.

No, you can’t know. Beliefs are not validated or justified by your own bias. Belief in X does not validate belief in X, and how you feel about belief in X does not validate belief in X.

I didn’t say it was my own bias that validated my belief. Please re-read.

You seem to contradict yourself in this paragraph. He most certainly is claiming magical knowledge. God uses magic, not technology.

Okay, class, listen well as I explain knowledge. There are two routes to knowledge:
1. Technology
2. Magic

Because these are the only two routes to knowledge, what we do not know by technology we know by magic. Except that there really isn’t such a thing as magic leading to knowledge, so that what we do not know by technology we do not know.

Okay, doctor(logic), I know that’s not how you view epistemology. But it sure is a glaring false dichotomy if I ever saw one!

The only knowledge we get (without inference) from direct experience is knowledge of the direct experience. It’s trivial.

Wrong. I just had a direct experience of reading the sentences, “The only knowledge we get (without inference) from direct experience is knowledge of the direct experience. It’s trivial.” I gained knowledge from that reading, directly.

I’m not saying the experience of God is a bare uninterpreted quale. It is an experience that enters in with other experiences and knowledge, and part of a network of inferences.

But it is also more than an impression. It is the Spirit bearing witness with our spirit (Romans 8:16) to the fact that we are children of God. God is not some deaf dumb blind mute ghost. What he wants to communicate, he can communicate. Therefore this is not applicable to the situation.

Perhaps after handling red balloons enough, I might know them more directly when I see them. This doesn’t apply to belief in an invisible God.

God is invisible but not incapable.

As for the fine-tuning discussion you brought up in your most recent comment, I need more time to develop an answer in modal logic, which I think is the right approach to this issue. Since it’s a holiday, I may just decide to let that slide until later. But I will answer this:

Are you saying that because I don’t know exactly how to spread the probabilities, I can’t infer anything at all?

Bayesian probabilities include a term for prior probabilities based on background knowledge. Your equation, as I read it, calls for a background knowledge estimate of how probable it is that God would create a world like this one as opposed to another kind of world. That’s the number you need to insert into the equation to make it valid and complete, I believe, unless I’m misreading you completely. So I’m saying you don’t have that prior probability estimate, and you have no place applying Bayesian analysis in this case.

But I could be wrong about that, no doubt. My full answer will be modal rather than Bayesian.

The discussions about evidences for evolution is an old one and often-repeated, so again, since it’s a family day and the kids are waking up at last, I’m going to let it pass.

Why would it surprise you that the outcomes are fine-tuned for a universe created by God, if indeed the universe were created by God?

Fine-tuned to be the way we find it, no matter how we find it? Everything is fine-tuned in this way, even physical theories.

Fine-tuning doesn’t mean that someone is fine-tuning a knob on a radio set. It refers to the exclusion of other alternatives without accounting for why they were excluded. Fine-tuning is not a sin in itself. However, when weighing two theories, the more fine-tuned theory is at a disadvantage unless the fine-tuning enables it to predict its way out of debt. If supersymmetry is discovered at LHC, it will result in a theory with many more fine-tuned parameters. The reason it will beat the Standard Model is that it makes strong predictions in exchange for the fine-tuning. God theory isn’t making strong predictions in exchange for its fine-tuning.

What background knowledge of God are you calling on, to show what the probabilities really are of God creating robot-rabbits? What makes you think God would have wanted to create a world in which we could think but not have brains? Can you run a Bayesian analysis without some basis for that opinion? Of course not.

God is compatible with many possible worlds. There’s nothing he could not arrange. He can blend intervention and non-physicality in with physicality in arbitrary proportion. We could have been 99% non-physical and 1% physical, or the reverse. Even if we were 99% physical, God did not need us to look evolved. He could intervene continuously to provide us with food, etc.

Are you saying that because I don’t know exactly how to spread the probabilities, I can’t infer anything at all?

I gave a deck of cards example. I did not specify how the cards were shuffled. Maybe there are 50 people who might have shuffled the cards, and they have varying levels of shuffling ability. Card shuffling never is totally random, anyway. Are you saying that if I deal an ascending straight flush of the deck, I can’t say anything about whether it was shuffled?

I don’t have to put a firm number on the weighting of final states in order to make a rational judgment, and I can still be pretty sure that the deck is sorted. Maybe the odds of drawing a low-ball straight flush from a human-shuffled deck aren’t exactly 311 million to 1 as with perfect shuffling, but would you think I was unreasonable if I said that it was 10 thousand to 1 against the deck having been shuffled?

like the missing-fossils-of-the-gaps theory, that says “oops, we were wrong about our predictions of innumerable transitional forms

You are attacking naive assumptions about the mechanisms of fossilization and evolution, not the theory itself.

If evolutionary biology is correct, species evolved from a common ancestor (or a small ancient number of such). We can make simplifying assumptions, such as the assumption that evolution proceeds at a constant rate, and that instances of every species get fossilized. These assumptions are not central to the theory, but the simplest possible case.

It turns out that such simple assumptions are wrong. Evolution does not proceed at a constant speed, and only a tiny minority of species get fossilized and survive to the present. But constant speed and universal fossilization are not implied by the theory. They were just simplifying assumptions about the mechanism. That these assumptions were wrong has no impact on the central claims of evolution.

Your argument is like me saying that divine creation is wrong because it took millions of years more than 6 days. In response, you can claim that God created the universe, and that shoving that creation all into 6 small, equally-long periods is a simplifying assumption that’s not central to the theory. (Actually your criticism is far weaker than that because Genesis actually says it was 6 days.) I expect you would argue this, and it’s one of the reasons I don’t bother criticizing 6-day creation arguments. I think there are much better arguments.

In the case of evolution, we find common descent, we find transitional fossils (if not all of them equally spaced as if in a constant speed evolution), and a mountain of genetic evidence confirming evolution.

There are several overarching predictions in the ID program: one is that undirected evolution will be found to be inadequate to have done what has been claimed for it.

So the theory that the deck is shuffled predicts that the next card will be… well, it doesn’t say what the next card will be. It just says that it might not be what you expect if it is sorted.

Another is that the information that exists in life cannot be shown to be an exception to the general rule that information comes from minds.

There is no such rule. This has already been proven false by genetic algorithms.

Another is that no viable theory for the formation of the first life by unguided means will be forthcoming. (In the meantime, ID calls naturalist OOL’s bluff: “future science of the gaps” answers get low marks for credibility.) There are many sub-predictions under these.

The low-ball straight flush has just been dealt off the top of the deck. I’m saying the deck is almost certainly sorted. Are you going to accuse me of issuing the “promissory note” of sorting theorists? Or maybe a future sorting theory of the gaps? Or bias against shuffling theorists?

Craig distinguishes, rightly, between knowing and showing. We can know that God exists, and more, by the working of the Holy Spirit within us

No, you can’t know. Beliefs are not validated or justified by your own bias. Belief in X does not validate belief in X, and how you feel about belief in X does not validate belief in X.

You know about cognitive bias and psychology. You know that people often believe what they want to believe, and that people claim direct knowledge (especially psychics). You know that such claims are false whenever they are testable. So you have positive evidence that the kinds of justification you are using are generally useless when we have the ability to test them. Why should such forms of justification be reliable when the claims cannot be tested?

So Craig is not 100% wrong in light of that. He is not claiming “magical knowledge,” for one thing. He claims there is knowledge that comes by direct experience of God. Knowledge by direct experience is certainly not outside of human expectations!

You seem to contradict yourself in this paragraph. He most certainly is claiming magical knowledge. God uses magic, not technology.

The only knowledge we get (without inference) from direct experience is knowledge of the direct experience. It’s trivial. If I see something that looks red, then I get knowledge that I see something red. If I see something balloon-shaped, I get direct knowledge that I see something balloon-shaped. However, I do not get direct knowledge that I see a red balloon. I have to infer that. Perhaps after handling red balloons enough, I might know them more directly when I see them. This doesn’t apply to belief in an invisible God.

It just occurred to me–and it’s too late in the evening to write out the full development of it–that your card-shuffling analogy has the same fatal flaw that I pointed out earlier in Tom Clark’s epistemology. It creates a situation, a set of rules, under which there is no possibility of discovering evidence for God in any possible world, including worlds where God actually exists and has revealed himself. Therefore it rules God out illegitimately.

I’ll have to explain that further tomorrow. As I said, it’s late now.

Dualism and ID do not predict these things. They are perfectly happy with other things. God could have created a world in which we think but don’t have brains, or in which species aren’t connected by common descent or common composition (e.g., rabbits could be robots).

why would it surprise you that the outcomes are fine-tuned for a universe created by God, if indeed the universe were created by God?

Or otherwise: how are you going to run a Bayesian analysis of what God would do? What are your prior probabilities based on? What background knowledge of God are you calling on, to show what the probabilities really are of God creating robot-rabbits? What makes you think God would have wanted to create a world in which we could think but not have brains? Can you run a Bayesian analysis without some basis for that opinion? Of course not.

And have you noticed the evidences that don’t run the direction the atheist predicts, like the missing-fossils-of-the-gaps theory, that says “oops, we were wrong about our predictions of innumerable transitional forms, but hey, we can wave that off!” Or the (well-debunked) theory that says the earth is just an ordinary planet, with nothing special about its position relative to its sun, its moon, its other solar system members, the galaxy in which it sits? You have a lot of unfulfilled predictions to deal with. If theory B predicts E2, and E3 is found, then your analogy is considerably weakened.

The problem with supernatural claims is that they generally do not make predictions. Suppose God designed the species. In order to convince us of this, you need to create a predictive model of God that says what he would design and why. You have to set things up so that evolution could do anything, but God would do something specific. You’ll never get there unless you’re willing to talk about the designer.

In order to conduct your Bayesian analysis you need your own predictive analysis of what God would design and why. I invite you to develop it and apply it. Have fun.

There are several overarching predictions in the ID program: one is that undirected evolution will be found to be inadequate to have done what has been claimed for it. Another is that the information that exists in life cannot be shown to be an exception to the general rule that information comes from minds. Another is that no viable theory for the formation of the first life by unguided means will be forthcoming. (In the meantime, ID calls naturalist OOL’s bluff: “future science of the gaps” answers get low marks for credibility.) There are many sub-predictions under these.

The difference isn’t in HOW you know versus how the jury knows. You and the jury are essentially following the same rules. The difference is in the information available to you versus the jury.

Craig distinguishes, rightly, between knowing and showing. We can know that God exists, and more, by the working of the Holy Spirit within us. Plantinga, in Warranted Christian Belief lays out an extended explanation of how this is a rational position to take for those who take it, given certain other conditions of their being epistemically responsible about how they evaluate evidences.

Simply put, given the evidences, it is possible that God exists. Some people (myself included) look at the evidences and say that God’s existence is considerably more plausible than his non-existence. Others disagree. A believer such as myself (or Craig or Plantinga or the apostle Paul in the Bible) would say that we have more working for us than just the evidences: we have the witness of the Holy Spirit. By his working we have strong assurance of the knowledge of God. We can know that God is real; and this knowledge can be as assured to us as the knowledge of any other phenomenal experience or inner state. It can be as assured to me as the knowledge that right now my toe doesn’t hurt but my shoulder is a bit tender. God can do that, and I am quite convinced that he does do that.

But none of us would put that forward as an apologetic argument to convince someone who does not have the same experience, that same sensus divinitatis. That’s not a valid expectation.

So Craig is not 100% wrong in light of that. He is not claiming “magical knowledge,” for one thing. He claims there is knowledge that comes by direct experience of God. Knowledge by direct experience is certainly not outside of human expectations!

The reason you know you are innocent and the jury doesn’t is that you have information they don’t have (i.e., your memories).

The reason Steve knows there is a God and you don’t is that he has information you do not have (i.e. his experience of God).

This leads to other questions, obviously, including the one Tom Clark raises with respect to fallibility of belief, and also questions relating to who is granted that knowledge and who is not. I don’t mind taking those issues up later, but the point I want to make for now is this: there is nothing irrational or incoherent in the proposition that God can provide assurance of knowledge of himself to persons as he chooses, by means other than publicly available evidence. This allows one to know truths about God. It is not something that anyone would (or at least should) say is capable of providing evidence to show another person any truths about God.

There are, as you know, a whole host of publicly available evidences for God, open to public discussion. That’s a different topic than the one Steve has raised just now.

Unfortunately for you, “Theory B” wins out and you are declared guilty because the Bayesian analysis showed it was the better theory to believe in. But you’re not guilty!!

I think you (or, rather, Craig) have this analogy backwards. You and I are the jury, and God is the suspect. God knows he exists, and our sound reasoning says he probably doesn’t.

Of course, in a typical court case the suspect may be unable to cough up the evidence that would clear him. Not so with God.

You are essentially arguing for a biased jury. Suppose you are on the jury, and before you even hear the charges, you look at the defendant and decide he seems like a bad guy. The trial proceeds and you find that the prosecution’s case is empty and pathetic. Should you think it likely that the defendant is guilty? Should you discount your bias? Or should you cling to your bias and assume it very likely that the defendant is guilty?

I can make this scenario even worse. Suppose again that your intuitive preconception is that the defendant is guilty. Consequently, discount all evidence in the defendant’s favor as some sort of elaborate deception, and accept only damning evidence from the prosecution. Will you reach a rational conclusion about the guilt of the defendant? No. We would say that you were being a bad juror, and that your conclusion was set before you ever heard the evidence.

Dr. Craig’s point is there are ways of knowing without resorting to evidence, argument or statistical analysis. God can be known this way.

Craig is just making excuses for believing despite the rational case against belief.

Let’s go back to the trial story, but use it in a different way. Suppose that you are the defendant. Why do you believe you are innocent? Do you have reasons for believing so?

There are at least two major reasons you believe you are innocent. First, you don’t remember committing the crime. Second, you remember doing something else instead (i.e., you remember experiencing your alibi).

Why would you think that these are valid reasons for believing you are innocent?

You assume two principles implicitly. Statistically, you remember significant actions you have taken, and, statistically, you don’t remember significant actions you have not taken. Roughly speaking, you have taken a significant action if and only if you remember taking it. There might be rare exceptions to the rule, but that doesn’t make the rule thoroughly unreliable, it just makes it slightly less than airtight.

Back to the trial. The reason you know you are innocent and the jury doesn’t is that you have information they don’t have (i.e., your memories). They would agree with your reasoning and you would agree with theirs. Both are rational. If there were ways to read a suspect’s memory, the jury would agree with you and clear you. (Well, up to a point. If you remember walking on the Moon in nothing but your pajamas, the jury would think your memory faulty, and you probably ought to side with the jury, too!)

So Craig is 100% wrong. You don’t just magically have knowledge. That would be superstitious nonsense. Rather, you have statistical, Bayesian-approved reasons for knowing what you know. The difference isn’t in HOW you know versus how the jury knows. You and the jury are essentially following the same rules. The difference is in the information available to you versus the jury.

Answer: we should think theory B is more likely to be true than theory A!!

I’m reminded of a short example Dr. Craig gave in one of the podcasts I linked to above.

You know you’re innocent of a particular crime that you have been charged with – you’ve been framed – and so you put together evidence hoping to convince a jury that you are innocent. The prosecuting attorney does the same to try and prove you are guilty. Unfortunately for you, “Theory B” wins out and you are declared guilty because the Bayesian analysis showed it was the better theory to believe in. But you’re not guilty!!

Dr. Craig’s point is there are ways of knowing that don’t require that we rely on evidence, argument or, in this case, statistical analysis. God can be known this way.

(NOTE: If you’re talking about science and ID, not God, then my comment is off the mark)

Back to the card analogy. I’ll try to state it as a Bayesian probability problem.

I have two theories, theory A and theory B. Each predicts a different likelihood for the states of evidence we might find. Suppose theory A predicts equal likelihood for finding evidence in state E1 and evidence in state E2. Theory B predicts that we won’t find evidence in state E1, but only in state E2. We look at the evidence and find it to be in state E2. Which theory is more likely to be true?

Well, theory A predicts one of the following:

A with E1
A with E2

and theory B predicts

B with E2

Finding E2 helps theory B, and hurts theory A because it rules out half the possibilities that A was advancing.

Suppose the theories also predict results in another context/experiment. Theory A predicts equal weight for finding evidence in states K1 and K2. Theory B predicts we will only find evidence in state K2. We look at the evidence in this new context and find the evidence to be in state K2. Which theory is helped by finding this result?

As before, theory B is advanced.

However, the problem for theory A is cumulative. Theory A predicted one of

A with E1 and K1
A with E1 and K2
A with E2 and K1
A with E2 and K2

while B predicted

B with E2 and K2

I’m sure you can see that the problem has become progressively worse for theory B. 75% of the possibilities advanced by A have been ruled out, while B is undamaged.

Now suppose we move into another testing context in which we have two possible results, L1 and L2. Again, A is fine with L1 and L2, and B is fine only with L2. Due to the data from the past experiments, A advocates are advancing the following possibilities:

A with E2, K2, and L1
A with E2, K2, and L2

while B is advancing its original proposition:

B with E2, K2, and L2.

Here, theory A has been fine-tuned. Possibilities of A that were ruled out by past experience are being excluded. We’re excluding possibilities like:

A with E1, K1 and L2.

Going into this experiment, how confident should we be in theory A? Is A just as likely as B?

Answer: we should think theory B is more likely to be true than theory A!!

If you disagree, then I want to play poker with you!!

I’ll apply this back to the deck of cards. Theory A is that the cards are shuffled. Theory B is the theory they are sorted. E is the first card draw, K is the second card draw, L is the third. The sorted deck theory (B) says the cards have to draw 2, 3, 4 of clubs. The shuffled deck theory (A) says that the draws can be 2 of clubs or not, 3 of clubs or not, 4 of clubs or not. We draw 2, 3, 4 of clubs. Does it make sense to say that the fine-tuned theory A is competitive with theory B?

No. We cannot just throw out scenarios of theory A without penalty. We would be ignoring cases of A like

Theory A with King of hearts, Jack of diamonds, and 5 of spades.

If we want to ignore such cases, we have to factor that into our likelihood estimates.

Evolutionary biology and neuroscience make specific predictions about what we have to see. We have to see common descent, we have to see a progression of species, we have to see central nervous systems with causal links, we have to see brains with computing ability, etc. Dualism and ID do not predict these things. They are perfectly happy with other things. God could have created a world in which we think but don’t have brains, or in which species aren’t connected by common descent or common composition (e.g., rabbits could be robots).

But ID advocates and dualists simply pretend that they can fine-tune their theories (eliminating the robot rabbits and brainless thinkers) without paying a penalty.

One more thing I want to add: an explanation for why supernatural claims will never gain traction. What does theory A have to do to catch up with theory B?

To catch up, theory A has to be specific. Theory A has to be consistent with something narrow where theory B is consistent with something broad. Then you do the experiment, and if theory A’s possibilities are vindicated, theory A gains traction over theory A. Of course, the strength of theory A’s prediction has to be weighed against the strength of B’s, so A can win a round without winning the match, but at least A can come back.

The problem with supernatural claims is that they generally do not make predictions. Suppose God designed the species. In order to convince us of this, you need to create a predictive model of God that says what he would design and why. You have to set things up so that evolution could do anything, but God would do something specific. You’ll never get there unless you’re willing to talk about the designer.

As usual, my writing seems clear to me, but my message doesn’t seem to get across to anyone.

I’ll give it another shot.

When you rationally believe a proposition P, you rationally believe that you have justification for P that goes beyond personal bias (unless P is a proposition about personal bias).

Now, suppose that I do not believe P. Not believing P is not the same as believing not P (~P).

If I don’t believe P, then either I am confident that ~P, or I have no confidence in either P or ~P. (I can be atheist about P, or agnostic about P.)

Suppose P refers to belief in the existence of any god or gods. In that case, the P theory advances many possibilities:

1) P with god(s) obviously present.
2) P with god(s) impossible to see.
3) P with god(s) who have yet to be seen.

In case (2), the world would look no different if the gods existed or not, and so I am prevented from ever finding out whether P is true or false. Clearly, I ought to be agnostic towards the gods of case (2). If not agnostic, I probably should consider whether claim (2) is even meaningful because existence requires a possibility of detection.

The Christian god is presumably an instance of case (3). IMO, there is overwhelming evidence against such a god from evolutionary biology. That’s why I am atheist about Christianity’s god.

But in contrast to Christians, I am not pandering to my own bias. I can point to specific rational claims against the existence of the Christian god. Also, there are lots of experiments, lots of ways that God could prove himself to exist to me. He could show up and have a conversation with me, and then submit to testing so that we can see his magical powers. God seemed perfectly happy to show himself to the world in the old days (if the stories are to be believed), why not now? I’m open to the evidence. What I’m not open to is personal bias.

However, you reject the ability to test God in this way. I think the alleged reasons for this rejection are specious. Christianity didn’t predict God would hide from humanity. In my experience, Christians are all too happy to predict God will show up in various ways. No, the idea of God hiding is a rationalization for why God has never shown up and proven himself.

You admitted in your response to Tom C that you reject naturalism because it rules-out belief in a god who hides from us. Why is that a problem for you? Remember, it doesn’t rule out the existence of a god who hides from us. It rules out rational belief in the existence of a god who hides from us.

Christianity is consistently pro-bias because, without bias, there’s no way to believe in a hiding god. This is why churches say “give us a chance, believe in God, and then you’ll see prayers answered.” What they mean is “bias yourself, and walk into a trap in which everything looks like confirmation.” Personally, I find this to be offensive and dishonest behavior (not that I expect anyone here to care what I feel).

Have a good turkey day!