Biblical Criticism & History Forum - earlywritings.com

Bernard Muller wrote:to Ben,

But what you are describing is a set of interdependent events. For example, if you were able to find out for certain that, when Paul wrote "descendant of Jesse", he meant it literally, would that not change your perception of the odds that, when he wrote "descendant of David", he also meant it literally?

I do not give a hoot if Jesus was a descendant of Jesse or not. What matters is that phrase implies literally Paul was taking Jesus as an earthly human (in the past).

That is exactly what I am talking about, too: what is the probability that calling Jesus the descendant of Jesse means that the person so calling him thought of Jesus as a literal human being?

I rated the probabilities for the two phrases at 80% & 80% separately, regardless of the other one (and other phrases in Romans implying historicity through a literal reading).

Yes, and that was a mistake.

BTW, what would be your arguments on one or two of these phrases about not implying the historicity of Jesus?

I have none. I think that they definitely imply belief in an earthly, human Jesus. That is not our difference here. This is all about the math, which you are misusing because you cannot, for some reason completely opaque to me, see that your factors are interdependent, that one of them being reckoned either true or false necessarily must change our estimate of the odds of (at least some of) the others.

For whatever it may be worth, I personally think that 50% is, if anything, too low as an initial estimate for any one of those statements. But Peter's 1% for the remaining statements, contingent upon that initial statement being false, sounds pretty close to correct to me. After all, if we could be 100% certain that Paul meant one of those statements in a nonliteral sense, the rest of the statements of a similar nature would immediately become very suspect.

So, to recap and hopefully save us some confusion, if those statements are legitimately part of Paul's letters (and not interpolations), then I think the chances are very high that Paul thought of Jesus as an earthly human. I am not at all a fan of any of Doherty's or Carrier's business about a superterrestrial, sublunar crucifixion. You and I agree (at least mostly) on that, I believe.

If you do not agree with my 80%, please tell me why and change my rating.[/quote

I am not entirely uncomfortable with 80% as a rating for your first instance; I am okay with that. But it is impossibly high for the remaining instances which are interdependent with the first. (For the record, I do not even remember which one you listed first; all that matters is that the initial probability can be quite high, while any interdependent events described thereafter must necessarily be lower, precisely because they are not independent.)

Please come up with arguments defending that Paul could have meant "descendant of Jesse" mythically.

Why on earth are you asking me to defend arguments that I myself find insupportable??

He did! He went through the whole process, assigning values he deemed fit for interdependent events. You objected to the low values of those events after the first one, but you did so because you still did not understand why it makes a difference whether the events are independent or not. (Even once you understand the principle, you are free to object, but hopefully on the basis of true knowledge, not of ignorance.)

No he did not. His values were not mathematically defined, and he suggested they may not be realistic.

ALL such values are going to be subjective. We are assigning probability to human behavior here, not to coin flips. His illustration showed you exactly how the odds are changed for interdependent events: that is what you are not getting. You think his initial 50% was too low or too high? Very well, then change it. Make it 80%. No problem. You think his successive estimates of 1% (contingent, as interdependent events have to be, upon the falsity of the first one) are too low? Very well, then change them. Make them 5% or something. But to make them 80%, same as the first, demonstrates that you are thinking of these events as independent, like coin flips; it demonstrates that you are treating Paul like a random writing machine, sometimes spouting mythicist rhetoric, sometimes historicist, all willy-nilly, as random as a coin flip.

This is my last memorandum on this topic, unless specific critiques of the mathematical principles I am discussing can be leveled at me. If you still do not "get it", despite Peter's best efforts to educate you on the matter, then I leave you to it.

We can revisit some of the statements from a couple pages ago where it's demonstrated that the "multiplication rule" and "conditional probability" were not well-understood, at least by that time.

Bernard Muller wrote:I always acknowledged what you call conditional probability and that's why I lowered my probabilities to an average of 50%, which is very generous,

This doesn't really make sense.

Suppose we return to one of the things that are easier to think about ("arguments" about history are relatively hard to think well about, in probabilistic terms, IMO, especially if an understanding of the basics of probability have not been built up from easier-to-understand situations).

Suppose a blood test measures for the presence of something in the blood, to detect a disease, but also that some people have a disease but don't have that thing in their blood yet. So of the people that have the disease, the vast majority of them will get a negative reading on a second try, if they had a negative reading on the first try.

Your (bad) doctor says "Look, if you don't trust two or three readings to provide great certainty, I tell you what. Let's assume the test is 50% accurate and do ten readings. That way, it's basically certain!" (The readings are negative, doctor is arguing patient doesn't have the disease.)

You mention the problem that you might not have the thing in your blood, and something about conditional probability, but the doctor says, "look, that's why I reduced the presumed accuracy to 50%, to account for what you call conditional probability! That's very generous!"

You see the problem, I hope.

Please don't say "but." Please don't interpret this particular example in order to illustrate a contrast, hoping to prove some point, based on an incomplete understanding of the topic. At this point I just want to hope we can reach any level of understanding at all...

Bernard Muller wrote:Furthermore, in the cases addressed by Carrier in OHJ (4, 5, 6, 8, 9) the worst probabilities (from the historicity view point) are respectively 50%, 33%, 50%, 50% & 50%. http://historical-jesus.info/110.html

I was looking for a serious discussion of conditional probability in Carrier's pair of books. I didn't find it. I was disappointed.

Bernard Muller wrote:And I still do not understand why these probabilities should be lowered so much because of (assumed) conditioning on the (assumed) negation of previous points.

The "assumed" bit needs to be taken as 100%, absolute, undeniable, indisputable, unalterable, factual, completely certain certainty.

This is because the "conditioning" is part of the formula that you are using, called the multiplication rule:

http://stattrek.com/statistics/dictiona ... ation_rule

Rule of Multiplication

If events A and B come from the same sample space, the probability that both A and B occur is equal to the probability the event A occurs times the probability that B occurs, given that A has occurred.

P(A ∩ B) = P(A) P(B|A)

When talking about the second number, you must take it as an absolute fact that the other event A occurred, on top of whatever background knowledge there is. You must imagine what the universe is like when you know that the event A had occurred; in the probability that remains, then... think about that. Then think about the probability of the event B, knowing A.

If the event A (for example) is equivalent to "Paul is a card-carrying 'mythicist' and believes that Jesus existed in the sky only, which is why he said X," then. Think about the chances that saying "Y" should be interpreted as a 'mythicist' comment or a historicist comment. (This is not my view. It doesn't have to be your view either. But it doesn't matter at this point. We're talking about the far side, where we have to take a different view on event A than what we believe is the most likely. We're talking about 'the rest' of the probability space, which is already considered to be unlikely in general, and divying that up.)

Or if the event A (for example) is equivalent to "Someone went through the letters of Paul and interpolated stuff about HJ, including X" then, think about the chances that "Y" statement about a HJ was interpolated or not... when we already know about this interpolator, who definitely existed, who definitely interpolated X, and who definitely was doing this sort of thing.

If you can't think this way, don't use the multiplication rule. Don't talk about probability... that's fine too.

(This is a somewhat rough and ready explanation. At this point I'm just trying to be as loud and clear as possible, in the hopes that you might hear.)

You're not taking it as an "absolute fact" in the general sense. But since you're talking about the probability that both events happened, you need to consider first the probability of A in the general sense. Then, in the probability space corresponding to where A happened, you need to think about B's proportion there. Not about B's proportion of the total space.

I made some thinking, and I came to that:

Let's say a man said something which may imply Objectification of Women (OW).
Then there is a debate and discussion, with exchange of arguments, among a group of 100 persons. Then they vote: 50 are in favor of the man implied OW, the other 50 not ... So we have a probability of 50% about positive for OW.

Then days later, the same man said something else which may imply Objectification of Women (OW).
There is a debate and discussion, with exchange of arguments, among a different group of 100 persons, about what the man said the second time. Then they vote: 50 are in favor of the man implied OW, the other 50 not ... So we have a probability of 50% about positive for OW on the second event.

Then again days later, the same man said something different which may imply Objectification of Women (OW).
There is a debate and discussion, with exchange of arguments, among a another different group of 100 persons, about what the man said the third time. Then they vote: 50 are in favor of the man implied OW, the other 50 not ... So we have a probability of 50% about positive for OW on the third event.

However some people (the YES persons) who think the man implied OW three times say, for the second event: "we knew it, yes, because this man implied the first time OW, so we are more certain than ever he did the same the second time". Therefore, for them, that would raise the probability by, let's say 10% (arbitrary value). So now we have 55% fot the second event.

That's what I would call taking care of some dependence or condition, with the first event affecting the second one.

About the third event, same process. But this time, because the two previous two events reinforce the conviction of the Yes persons about the man implied OW, then the probability would be raised to 50% + (2*10%) = 70% for the third event.
That's what I would call taking care of some dependence or condition, with the first & second events affecting the third one.

And if the third event is rated at 70%, then there is no reason, with hindsight, why the two first events should be rated at 70% also. But that would be only for the (biased) people who think the man implied OW three times.

Now let's look at the other side of the 50% (also 50% probability), from the points of view of some people (the NO persons) who think the man did not imply OW for anyone of the three times.
Same process concluding for them that there is a 70% probability for each event about the man never implying OW.

So now we have:
70% <=> 70%
70% <=> 70%
70% <=> 70%

We cannot have probabilities of 70% on each side for a total of 140% and that has to be reduced by a factor of 0.7142857 (100/140):
50% <=> 50%
50% <=> 50%
50% <=> 50%

Back to where we started, meaning the so-called dependencies or conditional relationship are cancelling each other.

What if we have a probability for the 3 events of 70% about the the man implying OW?

Following the same process and calculation, the overall results would be for the three events:
84% <=> 36%
84% <=> 36%
84% <=> 36%

Which would have to be reduced by a factor of 0.8333333 (100/120):
70% <=> 30%
70% <=> 30%
70% <=> 30%

Again back to where we started, meaning the so-called dependencies or conditional relationship are cancelling each other.

Cordially, Bernard

Bernard Muller wrote:I made some thinking, and I came to that:

Let's say a man said something which may imply Objectification of Women (OW).
Then there is a debate and discussion, with exchange of arguments, among a group of 100 persons. Then they vote: 50 are in favor of the man implied OW, the other 50 not ... So we have a probability of 50% about positive for OW.

Then days later, the same man said something else which may imply Objectification of Women (OW).
There is a debate and discussion, with exchange of arguments, among a different group of 100 persons, about what the man said the second time. Then they vote: 50 are in favor of the man implied OW, the other 50 not ... So we have a probability of 50% about positive for OW on the second event.

Then again days later, the same man said something different which may imply Objectification of Women (OW).
There is a debate and discussion, with exchange of arguments, among a another different group of 100 persons, about what the man said the third time. Then they vote: 50 are in favor of the man implied OW, the other 50 not ... So we have a probability of 50% about positive for OW on the third event.

However some people (the YES persons) who think the man implied OW three times say, for the second event: "we knew it, yes, because this man implied the first time OW, so we are more certain than ever he did the same the second time". Therefore, for them, that would raise the probability by, let's say 10% (arbitrary value). So now we have 55% fot the second event.

That's what I would call taking care of some dependence or condition, with the first event affecting the second one.

About the third event, same process. But this time, because the two previous two events reinforce the conviction of the Yes persons about the man implied OW, then the probability would be raised to 50% + (2*10%) = 70% for the third event.
That's what I would call taking care of some dependence or condition, with the first & second events affecting the third one.

And if the third event is rated at 70%, then there is no reason, with hindsight, why the two first events should be rated at 70% also. But that would be only for the (biased) people who think the man implied OW three times.

Now let's look at the other side of the 50%, from the points of view of some people (the NO persons) who think the man did not imply OW for anyone of the three times (also 50% probability).
Same process concluding for them that there is a 70% probability for each event about the man never implying OW.

So now we have:
70% <=> 70%
70% <=> 70%
70% <=> 70%

We cannot have probabilities of 70% on each side for a total of 140% and that has to be reduced by a factor of 0.7142857 (100/140):
50% <=> 50%
50% <=> 50%
50% <=> 50%

Back to where we started, meaning the so-called dependencies or conditional relationship are cancelling each other.

What if we have a probability for the 3 events of 70% about the the man implying OW?

Following the same process and calculation, the overall results would be for the three events:
84% <=> 36%
84% <=> 36%
84% <=> 36%

Which would have to be reduced by a factor of 0.8333333 (100/120):
70% <=> 30%
70% <=> 30%
70% <=> 30%

Again back to where we started, meaning the so-called dependencies or conditional relationship are cancelling each other.

Cordially, Bernard

Citation, please. I don't think you're qualified to make up your own math.

to Peter,

Citation, please. I don't think you're qualified to make up your own math.

That's not math, that's theoretical thinking. Can you tell me what is wrong with that? does that not make any sense?
Some scientists (like Higgs or Einstein) would say: first comes theoretical thinking, then the math.

Cordially, Bernard

There is no placing mathematical interpretations to plausibility.

Carrier failed and that should have been a lesson to all. Sad really.

Bernard Muller wrote:to Peter,

Citation, please. I don't think you're qualified to make up your own math.

That's not math, that's theoretical thinking. Can you tell me what is wrong with that? does that not make any sense?
Some scientists (like Mr. Higgs or Einstein) would say: first comes theoretical thinking, then the math.

Cordially, Bernard

to Peter,

If events A and B come from the same sample space, the probability that both A and B occur is equal to the probability the event A occurs times the probability that B occurs, given that A has occurred.

P(A ∩ B) = P(A) P(B|A)

When talking about the second number, you must take it as an absolute fact that the other event A occurred, on top of whatever background knowledge there is. You must imagine what the universe is like when you know that the event A had occurred; in the probability that remains, then... think about that. Then think about the probability of the event B, knowing A.

For example, "descendant of Jesse" occurs towards the end of Romans. But I do not see any dependence with "descendant of David" at the beginning of the same epistle. That is, if "descendant of David" was a latter interpolation (made with the benefit of knowing about the gospels), that would not prevent "descendant of Jesse" to occur in the epistle as written by Paul.
And the opposite ("descendant of Jesse" an interpolation, ""descendant of David" not one), that would not prevent "descendant of David" to occur in the epistle as written by Paul.

Cordially, Bernard