Wednesday, March 05, 2014

A model for defining evidence

How about the following as a conception of evidence, based on Bayesianism? X is evidence for Y if X is more likely to exist given Y than given not-Y. 

Now if we accept this, it looks like there are lots of things that qualify on behalf of theism and on behalf of atheis. Beginning of the universe? Maybe it can be reconciled with atheism, but it's not what an atheist would expect. Ditto for the fine tuning of the universe? 

Evil and suffering? Sure it's possible given theism, but is it more likely given atheism? 

With this model, we might say of our opponents that there isn't enough evidence, or that the evidence is outweighed by the other side, but can we really make the "no evidence" charge? 

Is there something wrong with this definition of evidence? 

21 comments:

Jeffery Jay Lowder said...

Victor -- You've got my vote on this.

Jeffery Jay Lowder said...

In fact, on a very related note, see my criticism of Victor Stenger's appeal to the "Lack of Evidence" Argument (LEA) for atheism.

http://www.patheos.com/blogs/secularoutpost/2014/03/05/the-argument-from-silence-part-7-victor-stenger-on-the-absence-of-scientific-evidence-for-god/

unkleE said...

For what it's worth, I agree with this and have used this line of thinking for several years now. It is my conclusion that there are only a few things that are more probable, given atheism (suffering god's hiddenness), but many things that are more probable given theism (the creation of the universe, the design of the universe, human rationality, objective ethics, consciousness, human freewill, religious experience including miracles and the life of Jesus.

Marcus said...

Serious question, who are these people who literally mean there is no evidence for theism or atheism when they say there is "no evidence" for a position?

Naturally, there will always be outliers but it seems the vast majority of people mean something more like "no good evidence" or "nowhere near enough evidence to believe X is true" when they literally say there is "no evidence for X". I'd agree such people should probably be more clear with their use of language but, at least personally, I've rarely seen this be a major sticking point in discussions of theism v atheism.

im-skeptical said...

The reply I gave in the other thread:

Bayseanism sounds quite reasonable until you get down to the details. How do you calculate the P(X|Y) and P(X|not Y)? You might be able to do it for some things, but when it comes to the existence of God, you're probably just pulling numbers out of a hat. Is there any way to agree what these values should be?

Marcus said...

"How do you calculate the P(X|Y) and P(X|not Y)? You might be able to do it for some things, but when it comes to the existence of God, you're probably just pulling numbers out of a hat. Is there any way to agree what these values should be?"

In theory, yes but given the vague nature of theism, and the way people tend to apply Bayesianism selectively, even the examples Victor gave in the OP cast doubt on how this would work in practice.

For example, take the fine tuning of the universe. How is this supposed to be more likely given theism? If the constants of the universe were such a way that they could vary wildly and life would still exist I imagine theists would say "Look how robust the constants are, they would support life even if they were wildly different. Clearly, god made the laws with the goal of creating life." Indeed, I find this argument more convincing but even if the universe were like that I don't think it would be a huge boost relative to atheism.

The problem with this application of Bayesianism is people are likely to just belittle the probability of seeing the evidence we do if their favorite theory was false particularly when that theory allows everything like theism. To paraphrase John Hawthorne, if you discover the constants are fine tuned with regards to the creation of diamonds does that boost the hypothesis that there's a being with a diamond fetish? If your answer is no but you think fine tuning for life does give credence to a being who intended to create life, then you aren't applying Bayesianism universally.

Papalinton said...

The 'fine tuning of the universe' is the most recent addition of the Apologetical harmonizing strategies in reconciling science with religion. In essence the historical timeline illustrates it is most clearly a product of a worldview, the Genesis theological theory of origins, in search of evidence to fit the model, a thoroughly post hoc rationalisation. For twenty centuries there was no such conception of a fine-tuned universe. Now that science has gleaned the conditions that support life as we know it, the lazy theist latches onto it as 'evidence' for a god. It would be inconceivable for those who are god-bound in their thinking that there is more probability of life being fine-tune to exist in the environment it finds itself and there is a far greater probability that it also adapts and changes as that environment continually changes. Either that or it goes extinct. Period. The evidence for this epistemologically robust narrative is simply overwhelming. No religion can hold a candle to the insights that science has been able to achieve.

Sorry to burst your bubble but the god hypothesis simply does not earn a guernsey in the reality stakes.

im-skeptical said...

The fine tuning question is particularly interesting. First, we have absolutely no way of saying what the probability is of physical laws being what they are. It could be anywhere from practically zero to one. We just don't know. And even if we did know, you can spin it any way you like.

Say we know that the probability of such a world is 1/10000000000. The theist argues the world is so improbable, it must have been created for life. The atheist argues that we must necessarily find ourselves in a world that supports life, so given that we exist, the probability of our world having these characteristics is very high.

BeingItself said...

Who says there is no evidence for theism?

Even if one person said they had an experience of God, that would constitute evidence.

BeingItself said...

Does anyone else think that the fine tuning argument is in conflict with ID theory?

Fine tuning alleges that God designed the laws of physics or constants such that intelligent life could evolve, whereas ID argues that God had to intervene at various points along the way because his tuning just did quite cut it.

Many apologetic arguments seem to undercut other apologetic arguments. Not that it matters to an apologist.

im-skeptical said...

"Who says there is no evidence for theism?"

I think Marcus was right. I am guilty of that.

Papalinton said...

"Even if one person said they had an experience of God, that would constitute evidence."

That would constitute evidence for having the experience. I think it is a stretch to extrapolate that experience and claim it constitutes evidence for the existence of a god. After all professional channelers and mystics claim to have all sorts of experiences conversing with dead grannies and other people's deceased relatives. Does that constitute evidence for the postmortem existence of these people?

BenYachov said...

>Fine tuning alleges that God designed the laws of physics or constants such that intelligent life could evolve, whereas ID argues that God had to intervene at various points along the way because his tuning just did quite cut it.

Very good you just gave Thomists another reason why they don't accept ID & Paley's design mishigoss both of which are in conflict with the 5th Way.

William said...

Bayesian probabilities require priors, just as logic requires premises. Different people and groups of people choose different premises as prior probabilities.

Victor Reppert said...

But regardless of priors, the difference in likelihoods creates confirmation. Whether such confirmation is sufficient or not is a function of individual priors.

Jeffery Jay Lowder said...

For those of you who haven't deeply studied inductive logic, especially Bayesian confirmation theory, here are 2 rules for participating in this conversation.

1. Victor knows what he's talking about.
2. You don't.

William said...

Vic: yes, I agree, evidence for God exists, and our priors are sometimes used to determine how we respond.

David Brightly said...

I've been puzzling over this for a while. Victor uses the phrase 'more likely' a couple of times. Now I think I understand Bayes theorem in its application, say, to disease diagnosis via imperfect clinical tests. I can see what the underlying probability space looks like. In Victor's model of evidence are we talking about genuine probabilities, or are we talking about something like 'degree of belief' on which we can impose the Bayesian mathematical structure? If the former, then the conditional probabilities are indeed objective, but I'm really puzzled as to what the underlying probability space is. Are its points Lewisian possible worlds, for example? And how does this fit with a classical theistic conception of God as necessary? If the latter, then I fear we may be no further forward. For then the conditional probabilities we need correspond to cogency or persuasiveness of argument, and this is exactly the subjective issue over which there is dispute.

im-skeptical said...

David,

It appears you'll have to keep wondering, because Victor isn't saying. It appears he may have done some calculations, but he isn't sharing them with us.

Victor Reppert said...

I'm pretty much a subjectivist, or at least a pluralist, about prior probabilities. I don't think they can be gotten from frequencies. But my definition of evidence doesn't require any particular prior probabilities.

David Brightly said...

Thanks, Victor. I accept your point about prior probabilities. But if conditional probabilities here are not relative frequencies, what on earth can it mean to say that P(X|Y)=p, with X and Y propositions. And how do we go about calculating them? I'm left with the feeling that we are standing in mid-air.