TedM

Ok, let me just ask you this Amaleq13:

Which of the following is more likely to point to an author making up a name for the messiah he depicts, as opposed to accurately reporting the name of the messiah he depicts:

Scenario 1:
1. the name is George, which has a non-messianic meaning
2. the name is Jesus, which for the sake of argument, has a messianic meaning

Scenario 2:
the name is Jesus, which for the sake of argument, has a messianic meaning and
1. 90% of the people of his time had the name Jesus
2. 5% of the people of his time had the name Jesus


IF you can answer and explain your answers for each of these, maybe I can finally understand what you are saying.

thanks,
ted

Amaleq13

The latter is more consistent with the notion of fabrication and the former more consistent with historical constraints but neither can actually tell us anything meaningful with regard to whether the character is more likely fiction or historical. The author could very well have had some non-apparent reason to choose "George" (eg the earlier 'Alice' example) and the historical man could very well have simply been given an appropriate name for the beliefs that developed about him.

You are placing far too much weight on the mere fact of the name than it can reasonably support.

Neither piece of information can tell us anything meaningful with regard to whether the character is more likely fiction or historical. All they tell us is whether the name was popular or not.

TedM

I really don't understand this answer. How can something be more consistent with hypothesis A over B yet not also be evidence that hypothesis A is more likely than B?

You appear to think that a result is of no value for determining a cause, when the cause is not random. Would you agree with that?

ted

Amaleq13

You have to keep in mind that you are working with a single piece of evidence and, unless that piece of evidence somehow necessarily requires the conclusion, it can only be considered consistent or inconsistent with it. You can start talking about probability, I think, only if you gather multiple pieces of consistent evidence and, even then, only in general terms rather than mathematic specificity.

The fact that I own a gun is consistent with the possibility that I shot the victim but it clearly does not make it more likely. You don't have to own a gun to obtain one, for example.

The fact that I tested positive for gunpowder residue is consistent with the conclusion and indicative of the conclusion that I recently fired a gun (or was really damn close to one as it fired). It is only consistent with the first conclusion because it does not require it but it is indicative of the second because there really is no other explanation for the evidence. The latter can even be directly expressed as a precise mathematical probability given whatever the parameters are of the test (ie it is 90% probable that the results indicate I fired a gun within X number of hours).

The fact that I initially denied firing a gun recently is consistent with the conclusion but I may have lied because I didn't want to get blamed.

The fact that I recently had an argument with the victim is consistent with the conclusion. It could just be a coincidence, though.

The fact that I cannot account for my whereabouts at the time the shooting is believed to have taken place is consistent with the conclusion. There are numerous times when people have no witnesses to their whereabouts.

None of the above, on its own, requires the conclusion and none, on its own, actually makes it more likely that I committed the murder.

Taken together, however, a case can certainly be made that the confluence of all these consistent facts is best explained by the conclusion (ie it is more likely than not that I did it).

IIUC, the best way to express this mathematically is by applying Bayesian theory to it but that causes numbness in my tongue and bleeding from my ears.

I have no idea how you obtained that from anything I've written but I also have no idea what it is supposed to mean.

At the very least, what you should have obtained from what I have written is that statistics based on non-random factors are not applicable to random factors. You have to calculate separate statistics for random factors.

TedM

Not sure I agree. Multiple pieces will increase the accuracy of the probability, but I think you can talk about probability as soon as you have data which distinguishes a random act from a non-random act.

If you have data that says that victims are shot by people that own guns 60% of the time, then it is more likely that you shot the victim than someone that doesn't own a gun.

ted

Amaleq13

Multiple pieces will increase the probability that the conclusion is true but the accuracy of that prediction is another question entirely. A single piece of merely consistent evidence simply provides too little information to make any conclusion more probable. This is not from an understanding of statistics so much as it is from an understanding of logic.

Your assumed statistic cannot make that differentiation because it only describes the results of non-random choices. The frequency distribution of non-random choices tells us absolutely nothing about random choices. I'm getting tired of repeating what should be a very obvious fact. You have presented no data that is capable of differentiating a random from a non-random act and you have completely ignored the fact that such a differentiation is entirely irrelevant to the use of the name "Jesus" in Christian texts since nobody thinks it was a random choice.

This is simply wrong. You are, again, misusing the available information. All that statistic tells you is that the victim was most likely shot by someone who owned the gun. That I own a gun is merely consistent with that prediction but there is nothing about that mathematical fact that makes it any more likely that I'm the gun owner who did it. That 60% of gunshot victims are shot by gun owners is merely consistent with the notion that I did the shooting. On its own, it does nothing to suggest it is more likely. Likewise, the fact that the vast majority of gun owners do not shoot people is consistent with my innocence but it does absolutely nothing to suggest it is more likely.

You are trying to base more on a single piece of information than is rational and misinterpreting the available information to suggest something it does not. Both have been true throughout this discussion.

TedM

You are right, and I am wrong. The example doesn't have enough information to conclude what I did. For that we would have to know what pct of the population owns guns.

I can't say anything about the odds of whether a particular author intentionally chose a messianic name for a FICTIONAL messianic character without knowing what the pct of all such authors use. Also, since I can't say that if a person with a messianic name is more or less likely than a person in the general population to behave as a messiah, I can't make any reliable predictions as to whether a person with a messianic name was likely to have been fictional or not either.

If we assume that a name doesn't affect his behavior, then I still think we can say something about that.

So, on the whole I did make a number of false assumptions and conclusions and many of your objections were on point.

As for my original attempt to use the frequency of messiac names to determine if a person described to be a messiah was likely fictional or not, I think there is validity to that ONLY if one can conclude that a person's name doesn't strongly affect his behavior. With such an assumption, if only 5% of the population has a messianic name, I would conclude that a person described as messianic who has a messianic name is most likely a fictional character, because with the assumption given, we would expect 95% of actual messiahs to have a non-messianic name. I thought that you were primarily objecting to this kind of thinking.

If you still do, then I guess I still don't fully understand your objections, though I do admit error with a number of the statements and examples I made after first raising the idea.

Thanks for your patience. And, I apologize for being so thickheaded.

ted

Amaleq13

Ted,


I sincerely appreciate your honesty and integrity as well as your persistence in continuing to try to understand what was wrong. Apology accepted.

I'm not sure I'm following how the numbers lead to the conclusion but it doesn't seem right.

I still think you are much better off considering the actual names of actual messianic claimants than trying to extract probabilities from statistics. If the men Josephus identifies can be considered "messianic claimants", I don't think his list supports any connection between name and role in the real world.

In the recent edition of the local newspaper, I read a story about a man with the last name "Cook" who is a cook and a story about a man named "Jack Frost" who is running for mayor of Anchorage. I suspect that, if you tried to generate a similar probability statement using name frequencies that you would conclude that both are more likely fictional creations than historical figures but you would be wrong. Actually, that might be a good way to test any probability theory you might develop. Simply apply it to circumstances where you already know the correct answer and see if the equation agrees.

TedM

I would agree as long as the list of messiah claimants is long enough. IF there were only 3 or 4, it wouldn't be a big enough sample. Also, you get into the problem of what constitutes a messiah claimant. But, in general I agree that the closer we can come to a representative sample, the more reliable the statistics will be.


Good points. Of course, you would also have to look at what all the other "Cook's" and "Jack Frost's" did too.

Amaleq, thanks for your patience, and acceptance of the apology. I shouldn't have just disappeared like I did.

ted