I recently began reading Strobels "the case for faith" and in chapter two WLC repeat his refutation that "extraordinary claims require extraordinary evidence". His argument is as follows:
This standard would prevent you from believing in all sorts of events that we do rationally embrace. For example, you would not believe the report on the evening news that the numbers chosen in last night's lottery were 4, 2, 9, 7, 8 and 3, because that would be an event of extraordinary improbability. The odds against that are millions and millions to one, and therefore you should not believe it when the news reports it (p. 65).
Most refutations I have seen in debates and elsewhere are not very clear. Paul Dolan, on Internet infidels (when discussing the book) argue: http://www.infidels.org/library/modern/paul_doland/strobel.html#obj2
Does Craig honestly believe that the results of a lottery are equivalent to the Resurrection? Even if any specific set of numbers is improbable, we know that some such set will be picked in a lottery. For resurrection to be analogous, being raised from the dead would have to be known to happen, even if very rarely; then it might be reasonable to suggest that Jesus is one of these rare resurrected persons, even with only a modest amount of evidence. But since it is clearly not known that some people are raised from the dead, Craig's analogy fails.
The problem is that while i think the argument touch some of the problematic areas in WLC argument, it dosnt really explain why WLC is wrong. For instance, suppose we just invented lottery and played it for the first time; in that case we wouldnt "know" people won all the time and WLCs argument could be reframed to use that example. Would that not circumvent this argument?
The second argument is more confused:
Using Craig's logic, if I were to say either that "I walked to the store" or "I flapped my arms and flew to the store," it would be unreasonable to demand more evidence for the latter than for the former! Frankly, I'd expect a little better understanding of logic from somebody with Craig's credentials. It appears that the discussion of "extraordinary events" and "extraordinary evidence" is placed here so that the rest of the chapter can appeal to merely "ordinary" evidence. And while Strobel apparently argues that he doesn't needstrong evidence because he doesn't want to have to provide it, I doubt he would accept merely ordinary evidence for the miraculous claims of any other religion.
The problems with these argument is that the author is simply claiming that miracles must be rarer than walking into a store and therefore WLC must be wrong. Thats certainly true, but its not a refutation of WLCs argument. The final argument is even worse:
According to the concept of "initial probability," the amount of evidence that one would need in order to reasonably believe a proposition is inversely proportional to its probability given our overall background knowledge. The initial probability of someone walking to the store is high, as it happens all the time; therefore, the amount of evidence required in order to reasonably believe the proposition is low. But the initial probability of someone flapping their arms and flying to the store, or of someone rising from the dead, is very low, and therefore the amount of evidence required to affirm such propositions is very high. Nevertheless, Craig tries to bypass the problem of the Resurrection having a low initial probability:
...which is saying nothing beyond reframing the original claim "extraordinary claims require extraordinary evidence" as a demonstrating the claim itself is true; the argument is entirely circular.
I think this is unfortunate because WLCs argument can be shown to be in error very easily by just using WLCs example.
Lets assume that winning in the lottery and someone (Jesus) raising from the death are equally extraordinary, ei. impropable, to begin with. WLC ask what evidence should convince us either of these things has happened; in the case of the lottery, the evidence we have is (for instance) what is being shown to us in the TV. Lets suppose the probability the announcer will get the sequence right is only 9/10 -- in other words, in one of 10 cases the TV will report a sequence different from what was actually drawn. Lets suppose that in the case the announcer get the sequence wrong, the announcer will mention another sequence different from the true winning sequence (what else should he say?).
The question WLC pose is now: Given the announcer mention the sequence S, what is the probability the sequence that was actually drawn was S?
But this can very easily be computed! To my surprice, it come out as 9/10 = 90% -- exactly the chance the announcer get the answer correct. So even very modest amount of evidence, in this case an announcer who make many errors each year, will provide sufficient evidence to make us highly suspect the true sequence is what he claim it is.
What about Jesus? The evidence we now have are the gospels. We are going to be generous to WLC and assume that if jesus rose from the dead, we would get exactly the gospels we have with their many contradictions, etc. In other words, if jesus rose from the dead, the bible would look the way it look today with chance 1. This is equivalent to assuming the announcer gets the lottery sequence correct every time.
But we also need some other input, namely that if jesus didnt rise from the dead, how likely is it we would have the evidence we have today? This is a slightly subjective judgement, but i think its fair to assume based on (1) not requiring breaking known laws of nature (2) reports of people rising from the death in other cultures (3) humans are known to make errors that there is some chance that even if jesus didnt rise from the dead, we would still end up with books like the gospels. For the sake of argument:
- (degree-of-extraordinary) The chance jesus rose from the death without any evidence is 1 : 1'000'000 (about the same as winning in the lottery)
- The chance the gospels would be the way they are if jesus rose from the death is 1 : 1 (perfect announcer)
- The chance the gospels would be the way they are if jesus didnt rise from the death is 1 : 1000 (false positives)
In this case it is easy to work out the chance jesus rose from the death given the evidence is slightly less than 0.1% - compared to 90% before.
Conclusion:
When scientists say extraordinary claims require extraordinary evidence they are proposing a general rule of thumb. In a real setting one would ofcourse try to quantify what is meant by extraordinary (or sufficient) evidence, and this is determined by the actual conditions.
When WLC attempt to equate the lottery to jesus rising, he is doing no more than using their superficial resemblance to argue they are the same, while any actual treatment quickly reveal they are simply different situations. An example i dont think is unfair is to say water ice does not melt at 0C because carbon dioxide ice (which presumable look roughly similar to water ice) does not melt at 0C.
WLC can't possibly argue as a philosopher that one shouldnt work through ones thought-examples. Another thing he cant suggest is that we shouldnt compute probabilities the way i have attempted, as he did exactly this in his debate with Bart Ehrman on the resurection of Christ. What could save WLC was to show i have made unfair assumptions, but i have a hard time seing what they could be.
I suppose my question is if WLC know his argument is flawed and still use it, or he just havent worked it through, or something else.
Another thing i was surpriced about was how clearly wrong the lottery example is. I was surpriced the numbers came out the way they do, 90% to < 0.1%, even though the two situations seem to make the same type of assumptions; it also quite clearly show that arguing the bible is consistent with jesus rising from the death (something most appolegists consider a victory in itself) is a far cry from getting them anywhere near demonstrating jesus rose from the dead. I also think internet infidels treatment leave a lot to be desired; a smart christian can clearly see two of the arguments used are bad and the first dont really refute WLC.