Thursday, January 08, 2015

The no evidence charge revisited

A redated post.

Loftus asks if there is any evidence for Christianity. Of course, he's not the only one, lots of people say that about religious beliefs, not that there is poor evidence, or that the evidence is outweighed, but that there is no evidence. 

Wouldn't it be an idea to come up with a concept of what we mean by evidence before we ask whether we have any? X is evidence for Y just in case Z? 

To me, X is evidence for Y just in case X is more likely to exist if Y than if not-Y. But now, if we go with that definition, then the existence of reports that Jesus was resurrected from the dead is unlikely given the claim that Christianity is false. After all, most people do not have people claiming they were resurrected after they died. (Not even Elvis Presley, though there are people who claim he never actually died). But we should expect it to be reported if Christianity is true, so, in and of itself, the existence of resurrection claims on behalf of Jesus are evidence that Christianity is true. Plug it into Bayes' theorem and it ups the probability. 

Now, you might say that that's crummy evidence, and in and of itself it surely wouldn't persuade much of anyone. But if you want to deny that it is evidence at all, you need to supplant my definition with one of your own. 

I am willing to embrace the logical consequence that the testimony to the Golden Plates is evidence for Mormonism. But my view would be that the weight of the evidence is against Mormonism, not that there is absolutely no evidence at all for it. I've, for a long time, been asking for a definition of evidence that allows us to draw the conclusion that there is no evidence for Christianity, a claim I would NOT make even about such patently false claims as Mormonism, or even Scientology. 

29 comments:

grodrigues said...

"Not even Elvis Presley, though there are people who claim he never actually died"

Of course he didn't. Who do you think is winning all those Elvis impersonation contests?

Anonymous said...

"if we go with that definition, then the existence of reports that Jesus was resurrected from the dead is unlikely given the claim that Christianity is false."

Doesn't that ignore the idea that people might have a reason to invent the story? The Golden Plates is the same thing. Someone was motivated to create a religious cult with a following, so they invented a story, came up with some "evidence", and the gullible followers, desperate for good news in their miserable lives, swallowed it. They then repeated the stories about the evidence.

This explanation fits with our understanding of physics and human psychology, and is FAR more likely than the miraculous one.

VinnyJH57 said...

Evidence is an effect from which we infer a cause. If we find a body in the hall with a knife sticking out of the back and on the handle of the knife there appear little swirly patterns that match the swirly patterns on a particular human being's fingers, we can say that we have evidence of who did the stabbing. We can say that because we understand what causes those little swirly patterns to appear on objects other than human fingers. If we didn't understand the natural process of cause and effect, those little swirly patterns on the knife handle wouldn't be evidence of anything.

Miracles don't follow the normally observed patterns of cause and effect that we use to draw inferences from evidence so I am not sure how we could ever claim to have evidence of a miracle.

Bradley C. said...

I agree. It is one thing to say "The evidence as a whole does not make X more probable than not." It is a completely different thing to say "There is zero evidence for X."

I don't see a problem saying there is evidence for X, but I still don't believe X. That can be rational provided there is greater evidence for not-X. It seems that to say there is NO evidence for X means something like: there is no reason that can be given in support of the claim that X is true. That seems absurd when it comes to most beliefs we hold.

William said...

Vinny:
"
Miracles don't follow the normally observed patterns of cause and effect that we use to draw inferences from evidence so I am not sure how we could ever claim to have evidence of a miracle.
"

Miracles are unlikely events which are anticipated and interpreted in a non-naturalistic context.

From a naturalistic context, you can see the unlikely event, but, unless you succeed in seeing that event from someone of a differing belief's perspective, you will never see it as a miracle.

Mark Frank said...

"To me, X is evidence for Y just in case X is more likely to exist if Y than if not-Y."

Uhm - suppose X is "the ground is wet" and Y is "a giant watering can recently appeared in the sky and sprinkled water". X is more likely given Y but you could hardly say X is evidence for Y. It is necessary to take into account the plausibility of Y and compare it to other possible explanations for X.

William said...

Mark:

Yes, there you have a "head-to-head collider" type Bayes network, where rain and the UFO sprinkler both would make the ground wet, and the ground is wet. In such a case, the more likely that it could have rained, the less likely it was the UFO sprinkled water.

VinnyJH57 said...

William,

The thing is that I don't see the event. What I see are the effects and I infer the event that caused them. In the case of an alleged miracle, the effect that I see is usually the fantastic story that someone tells, but knowledge and experience lead me to believe that the usual cause of such fantastic stories is some combination of human foibles like ignorance, superstition, prevarication, gullibility, and wishful thinking.

Anonymous said...

I agree with Victor. X is evidence for Y iff Pr(Y|X) > Pr(Y).

Or, X is evidence for Y iff Pr(X|Y) > Pr(X|~Y).

Which is true even if Pr(Y|X) turns out to be low, as is the case of giant in the sky watering the ground.

I think the confusion happens because when most people think "X is evidence for Y" means "X makes Y probably true," (Pr(Y|X) > 0.5) when it actually means "X incrementally increases the probability of Y."

I suppose someone will try to deny this anyway.

unkleE said...

"This explanation fits with our understanding of physics and human psychology, and is FAR more likely than the miraculous one."

But nevertheless, if we use Bayes Theorem, the question isn't which explanation is more likely, but how much the event is probable given each explanation - which is a different thing. So if the God of Jesus exists, how likely are the stories of resurrection, and if not, how likely are the stories of the resurrection? I think Vic is correct.

"Miracles don't follow the normally observed patterns of cause and effect that we use to draw inferences from evidence so I am not sure how we could ever claim to have evidence of a miracle."

If you study enough alleged miracles, this is not necessarily true. Apparent healing miracles seem to cluster around certain people and certain practices (christians praying for healing). Of course many will be bogus or have natural explanations, but it appears that about 300 million christians claim to have observed or experienced a healing miracle, so it only requires a small percentage to be plausible (i.e. a recovery did occur after prayer and there is no obvious medical explanation) and the probability of them all being bogus or natural becomes quite small.

"where rain and the UFO sprinkler both would make the ground wet, and the ground is wet. In such a case, the more likely that it could have rained, the less likely it was the UFO sprinkled water."

The Bayes question still is: what is the relative probability of the ground being wet if it rained compared to the probability if there was a UFO sprinkler? Since the wet ground follows from both hypotheses, that particular fact is of no value in assessing the probabilities of the two hypotheses by Bayes - they will have to assessed on other grounds - which would surely lead to the rain hypothesis being favoured.

Anonymous said...

unkleE,

I agree with what you are saying, but if i read Victor's remarks correctly, he hasn't considered the possibility of an alternate explanation that might be more likely.

Mark Frank said...

Chris W.

I agree with Victor. X is evidence for Y iff Pr(Y|X) > Pr(Y).
Or, X is evidence for Y iff Pr(X|Y) > Pr(X|~Y).

Which is true even if Pr(Y|X) turns out to be low, as is the case of giant in the sky watering the ground.
I think the confusion happens because when most people think "X is evidence for Y" means "X makes Y probably true," (Pr(Y|X) > 0.5) when it actually means "X incrementally increases the probability of Y."
I suppose someone will try to deny this anyway.


Well of course you can define "evidence" any way you want. The fact remains that

"X incrementally increases the probability of Y."

is compatible with

"X incrementally increases the probability of Y to a value which is still so low that we would be foolish to believe Y is true."

"X incrementally increases the probability of Z (which is incompatible with Y) even more than it does Y."

Although you can define "evidence" how you like, your definition does seem to be different from common practice. Normally if new evidence makes explanation A vastly more probable than explanation B, then we would count that as evidence against B - even if the evidence makes B more probable than it was before.

Take a legal example. Suppose there are 100 suspects who are equally likely to have committed a crime. Then DNA is found which is known to come from the perpetrator. It is incompatible with 98 of them - there is 1.1% chance that it could come from suspect Bob and a 99% chance that it could have from suspect Charlie. If you do the sums, you will find that in the light of the new evidence the probability of Bob doing it has increased from 0.01 to 0.011 while the probability of Charlie doing it has shot up from 0.01 to 0.989. You would say the DNA is evidence that Bob did it – but I would say the massive evidence that Charlie did it is evidence against Bob doing it. I would be less inclined to believe Bob did it after finding the DNA – you presumably would be slightly more inclined. I don’t think there is an objective way of proving who is right.

Crude said...

Although you can define "evidence" how you like, your definition does seem to be different from common practice. Normally if new evidence makes explanation A vastly more probable than explanation B, then we would count that as evidence against B - even if the evidence makes B more probable than it was before.

Take Claim X. Claim X is judged, for whatever reason, to be 1% likely. New evidence A shows up: Claim X is now 1.5% likely.

Given only the information stated, can we say that Claim X has had evidence presented in its favor? The answer seems to clearly be 'yes' by a normal view, come what may. I think arguing that Claim Y is vastly better supported by Evidence A doesn't really argue against this - it just confuses things. 'Better evidence for Y' doesn't change the increase in likelihood for X context.

William said...

unkleE:


Look up Berkson's paradox.

Mark Frank said...

Take Claim X. Claim X is judged, for whatever reason, to be 1% likely. New evidence A shows up: Claim X is now 1.5% likely.

Given only the information stated, can we say that Claim X has had evidence presented in its favor? The answer seems to clearly be 'yes' by a normal view, come what may.


I don't find it all clear - but in end the decision is subjective. Take my example above. I would argue that before the DNA was discovered there was no reason to dismiss Bob as a suspect. After the DNA was discovered we have overwhelming evidence that Charlie did it and therefore good reason to dismiss Bob and all others - despite the fact that the probability that Bob did has increased marginally.

Crude said...

Take my example above. I would argue that before the DNA was discovered there was no reason to dismiss Bob as a suspect. After the DNA was discovered we have overwhelming evidence that Charlie did it and therefore good reason to dismiss Bob and all others - despite the fact that the probability that Bob did has increased marginally.

I know you're saying it's subjective, and really, that's reasonable and normally good enough.

But here's my question: in the example, the probability that Bob is guilty has increased marginally. This increase is due to (call it) data X. That much doesn't seem to be in dispute by anyone, regarding the example.

Then what do you call Data X when Data X leads to an increased likelihood of Possibility A of being true? Apparently 'evidence for possibility A being true' is out. So what's left? Meaningless data?

Unknown said...

I noticed that no-one cared to answer my previous challenges on the previous thread. But what do you expect? when their theo=philosophy is challenged by one of a sceptical disposition, what can they do but flee to the high mountains of self-deceitical predispositional persiflage?

The subject of this latest post concerning no evidence is clearly approach from an apologetical and theological in content, and thus does not deserve to be discussed in any depth by those freed from the mammoth memeplex of such alchemical endeavours. [We haven’t the heart for such endeavours, oh.] To arbitrarily introduce such ideas as minds, souls, essences, forms, ghosts and other such is supernaturalistic, at best. Fortunately for the enlightened, religio-philsophico-superstitio-supernaturalo-theological predilection of those spiritworld-ers which has been doggedly [dogmatically] supernaturally superstitiously unencumbered by a great progress in the knowledge of humanity. Consider the recent conference “Moving Naturalism Forward”, for example:

http://preposterousuniverse.com/naturalism2012/video.html

People are gradually realising that a supernaturalistic predilection abrogationally/appositionally shackles anti-rationalism to a scientific conclusion of pabloistic sophotrometric atronomico-estertetico philosophical underpinnings of metaphysical naturalism as the best explanation for what we understand about ourselves, the world and the universe. As Alfred E. Newman mischievously said:

it’s crackers to slip a rozzer the dropsy in snide.”

[woo-meisters all subconsciously au fait with this , but if you want to learn a little about what has been discovered in this area by science see:

http://en.wikipedia.org/wiki/the_scientific_underpinnings_of_ pabloistic_sophotremetric_atronomico-estertetico _investigation

Perhaps those of a more sceptical and less barbaric predilection may gain some non-supernaturalistical benefit from this, roundly rejecting any recherché recourse to anti-amaranthinate apologetical aspersions? Indeed, slipping a rozzer the dropsy in snide is somewhat reminiscent of the unfortunate human dallying and fascination with astrology, alchemy, superstition, spiritualism etc. of which supernaturalistic woo is but a subcategory]

Unknown said...

To continue, an apologetical effort congruently touches on a raft of options, supervening on the decaying and increasingly callow memeplex that is the superstitious supernaturalistic mythos. All that remains is for scientifically-informed philosophy (SIP) [here see Dennett, Eller, Loftus, Harris, Edis et al.] to overtake scientifically uninformed philosophy (SUP) [see everyone who subscribes to apologetical efforts doggedly [dog-matically, dog-wood may be considered anti-theologically through the lens of the history and progress of science of a canine variety, or , at a stretch dendrology [http://en.wikipedia.org/wiki/dendrology] given that dogwood is a tree] to instantiate superstitious supernaturalism in the 21st century without harmonizing onto-theo-semantico-bobishafto-empirical based efforts which produce knowledge.]

One is reminded of Jerry Coyne’s comment on archaeopteryx:

The birdlike traits number just two: large feathers and an opposable big toe.”

See how mischievously a supernaturalistally suffused netherworld of ghouls, ghosts, Plantinghasts, Persiflagellants, spirits, supernature, superstition, souls, and the whole Reppertoire of dualistic supernaturalism and Feserite scholasticism, etc. is being progressively somatificated into supernaturalism by the quantum increases of scientific knowledge supernaturalisticalifragilistic-expiali, as such. If you can stand the cognitive dissonance consult this brief overview of science:

http://en.wikipedia.org/wiki/Science

(And for a bonus, consult the following knockdown refutation of superstitious theories of mind:
http://en.wikipedia.org/wiki/Neuroscience)

Unknown said...

And now I tire because, as David Eller put it

“the verdant plenitude of myth makes it difficult to study”

[though really he should’ve written “mythos” rather than myth to describe such superstitious supernaturalistic trite tripe, because that sounds far more sophisticated. I prefer to use this term, dear reader, speaking with the authority and expertise of one whose prose was illumined [not in the supernaturalistic monastic sense, mind you!] by receiving Enlightenment from ineffable, unknowable and unseen supernatural spiritworldico-netherworldical belief in my thirties. Before that I was a Bible-crazy for thirty years. Having found the theological cupboard bare, save for tradition, I ditched the supernaturalistic mythos. Thought everyone might want to hear about that again. ;o)

However, the reader should bear in mind that I may have been engaging in persiflage at points.

Or As Dan Dennett put it “ mere quasi-meaning, or semi-semantics.”

William said...

"
But here's my question: in the example, the probability that Bob is guilty has increased marginally. This increase is due to (call it) data X.
"

Crude, Berkson's paradox applies here. Basically, the odds are distorted from what you would intuit because we are saying that of the following:

1. None did crime
2. bob did, charlie did not
3. charlie did, bob did not
4. both did

You are selecting ONLY possibilities #2 and #3. This causes at least some of the nonintuitive numbers you complain about.

ingx24 said...

Lapa Pinton,

I haven't laughed this hard in a long time. Kudos. :)

Crude said...

William,

You are selecting ONLY possibilities #2 and #3.

How? 'Bob did it' seems subsumed under #4 as well. He's just not exclusively guilty.

Either way, that seems to get into an issue that doesn't really touch on the example. If there's an increase in the likelihood of possibility X, there's an increase in the likelihood of possibility X, period. Sure, there may be all kinds of movement in possibilities Y, Z and otherwise - but I'm just concerned about X.

Papalinton said...

Lapa Pinton. ;o) Wonderful synopsis.
Many a true word said in jest. Your jesting is more immaculate than the immaculate conception , particularly in drawing attention to the nonsense of godworlds and other things that go bump in the night.

Awesome.

William said...

Crude:

Are you then including the possibility that all 100 suspects all did the crime? If so, then I agree with you. Otherwise, it's clear that Bob has indeed become more likely to have done it, compared to the 98 that are cleared by the DNA results.

Mark Frank said...

Crude

Then what do you call Data X when Data X leads to an increased likelihood of Possibility A of being true? Apparently 'evidence for possibility A being true' is out. So what's left? Meaningless data?

I would say that sometimes it is evidence for A and sometimes it is not depending on the circumstances – especially taking into account alternatives to A.

Unknown said...

Re: the OP

The trouble is, if you define 'evidence' in such a way that excludes Christianity from having any, you end up setting the bar too high for what counts as evidence. If someone can't have evidence, no one can, because nothing counts as evidence.

Unknown said...

The question "What counts as evidence" is a philosophical question. Yet another instance where philosophy underwrites science.

Also, it is not clear from the Bayesian philosophical formulation that evidence needs to be limited to empirical evidence.

Son of Ya'Kov said...

The true comedy behind Lapa Pinton's satire is that Paps takes it seriously as serious rational argument.


Nuff said.

Unknown said...

Samwell

Our concepts underwrite science, just like they do philosophy. What counts as evidence is indeed a philosophical question, and science definitely presupposes a definition of evidence, but the definition of evidence doesn't arise from the disputes of philosophers. That would put the question of what counts as evidence on the level of mere opinion. Our conceptions of what counts as evidence run much deeper than the level of opinion, to the level of what Wittgenstein calls a form of life.