Probability and NHM

[Bill Hamblin]

Don insistently speculates that there should be some relationship between the modern Arabic pronunciation of Nihm and the ancient Sabean pronunciation of NHM. Here are some examples of known shifts in toponym pronunciation between Hebrew and Arabic that occurred after the Arab conquest of Palestine in the 7th century. The Hebrew is given first, the Arabic equivalent second, followed, in parentheses, with a possible parallel of what a similar shift would create with Nahom. (These examples come from Y. Aharoni, The Land of the Bible, 2nd ed. 1979, p. 119-23)

Shaalbon > Selbit (Nahom > Nahmit)

Shiloh > Seilun

Betonim > Batneh (Nahom > Nahmeh)

Beth-lehem > Beit Lehm (Nahom > Nehm)

Beth-Shemesh >

[DonBradley]

Dan,

Thanks for the information. That explains why NHM shows up as "Nehem" in the early map image posted to the thread. I concede that this seems close enough to "Nahom" to me to make the "difference" between "Nahom" and "Nih[e]m" inconsequential.

As I've mentioned both on this thread and to you personally, I believe the NHM correlation does serve as evidence favoring the BoM narrative. The questions for me are how much evidentiary value it has and in what total evidentiary context.

I'll post tomorrow on why I believe the improbability of the NHM find tends to be overstated by believers and offer my thoughts on what evidentiary force it should be accorded.

Don

[DonBradley]

Bill,

I don't find the argument I employed correctly characterized in your post. The argument was that a place named X now is more likely to have been named X in the past than is a place now named Y. This is far different from asserting that the name should have remained essentially constant over millennia.

In any case, your arguments of earlier this afternoon, now strengthened by Dan Peterson's explanation of the voiced "H," were a sufficient refutation as far as I'm concerned.

Don Bradley

[DonBradley]

We also have:

Kefar Nahum (Hebrew:

[Alf O'Mega]

Unfortunately, it is attempting to answer the wrong question. What Alf is describing is the probabilities of finding ANY match whatsoever. If I'm not mistaken this is quite different from the probabilities of finding a very specific single match.

Exactly so. I don't see how this is the wrong question. If any of the other candidates had matches, we'd be discussing them, not NHM. There is no a priori reason to single out Nahom from the list and therefore be amazed that this one neonym matches a target in Arabia.

As I have said before, I believe there are too many unknowns and variables to come up with precise numbers. Rather it should give us a sense of relative probability and improbability.

We agree. As I said, "Obviously subjective factors swamp the objective factors, so let's not be too easily impressed by intuitively improbable hits."

You go on to argue that the number of candidate neonyms and target toponyms are small enough for NHM to be an impressive hit probabilistically. I think that you are vastly underestimating the number of targets. The Bible includes about 1200 toponyms. If only half of these are places in and around Palestine (a much smaller area than South Arabia), we still have more than the 500 target toponyms I used for my illustration. I don't see why we should assume that South Arabia is a more toponymically impoverished area than Palestine.

So if we make 5 guesses against a set of 500 targets, we would have about a 1 in 4 chance of a hit. For 200, 1 in 10; for 100, 1 in 19; and for your absurdly low 50, 1 in 37. I hope that it's clear that we are orders of magnitude away from the 1 in 1000 chance you ventured earlier in the thread. (Charity disqualifies discussion of any earlier figures you may have ventured.)

Also, although on-the-ground Arabian hits are apparently only possible with the half-dozen candidates you cite (I haven't double-checked you, but I have no reason to doubt that the list is comprehensive), let's not forget that BoM names from the New World portion of the narrative also appear in archaeologically informed apologetics (e.g., Alma). Who doesn't think that finding a Zarahemla in Arabia would launch an entirely new round of BoM triumphalism?

[New Kid OTB]

Alf has it right. Post hoc restrictions on the definition of a hit artificially drive down the probability of achieving one.

[Dan Vogel]

4. Nahom (16.34)

[Bill Hamblin]

I think that you are vastly underestimating the number of targets. The Bible includes about 1200 toponyms. If only half of these are places in and around Palestine (a much smaller area than South Arabia), we still have more than the 500 target toponyms I used for my illustration. I don't see why we should assume that South Arabia is a more toponymically impoverished area than Palestine.

You'll have to provide some evidence instead of mere assertion. I'm done proving the obvious to people who refuse to do any homework for themselves.

The Bible has hundreds of toponyms because it is a large book with numerous geographically based narratives. First millennium BC South Arabia has no comparable narrative textual history. All known first millennium BC South Arabian toponyms come from inscriptions, most of them dedicatory. It is not the number of names that once existed that is the question. It is the number of names from that period attested in surviving texts.

[Beowulf]

Just a note:

1 Nephi 16:34: Ishmael died, and was buried in the place which was called Nahom.

Later,

1 Nephi 17:5: And we did come to the land which we called Bountiful...

It seems to me there is purpose in the different use of grammar here. The first implies that the name was already there, given by someone else. The second is plainly the name that Lehi's family gave it.

The second would not last beyond the next visitor to the area. The first, on the other hand, could last, and might even merit an inscription sometime along the way.

Dan, surely you can read the BofM as well as I can...

Beowulf

[Bill Hamblin]

So if we make 5 guesses against a set of 500 targets, we would have about a 1 in 4 chance of a hit. For 200, 1 in 10; for 100, 1 in 19; and for your absurdly low 50, 1 in 37. I hope that it's clear that we are orders of magnitude away from the 1 in 1000 chance you ventured earlier in the thread. (Charity disqualifies discussion of any earlier figures you may have ventured.)

Once again, your figures are responding to a different question, not the question I asked.

[Alf O'Mega]

The question you asked:

I

[Bill Hamblin]

The question you asked:

I

[Alf O'Mega]

Bill, it may be time to stop digging. I've tried my best not to be condescending, but your innocence of probability is making it extremely difficult.

By your count, Joseph Smith placed six wagers in South Arabia. Yes, the chances of the NHM wager paying off are small (1 in 370, by your most miserly set of assumptions: 1 guess against 25 targets; better than that with more realistic parameters), but he would have collected the same payoff if any of the wagers had won. I gave you the odds of one of them hitting, and I used your own assumptions (while pointing out that an additional assumption about how many hittable targets exist must also be made). Each of these assumptions is contestable on subjective grounds, particularly the idea that JS made random guesses. (He was familiar with Semitic names: he had one, and the Yankee frontier was crawling with them.) But working within the parameters of the problem as you defined it, I have provided an answer to your original question.

The chances of Nahom being a coincidence? Not too bad.

[New Kid OTB]

The question is the probability of finding NHM, not any match of any random three consonant.

I ask what is the probability of getting 7 red on a roulette wheel, and you tell me the probability of hitting any odd red number. It seems clear to me that these are two different probability questions.

Let's look at three different questions:

1. What are the odds of the NHM match occurring by coincidence?

2. What are the odds of any toponym match occurring by coincidence?

3. What are the odds of any match occurring by coincidence?

Alf is cutting you some slack by allowing you to stop at question 2. Technically, the first two questions, by themselves, cannot be shown to be appreciably relevant to the BoM authenticity question because they're tailored post facto to a cherry-picked data point. In order to show any appreciable relevance, you must come up with an estimate for the third number and compare it to the number of apparent matches. This you haven't done.

(It's true that, by themselves, the first two questions are negligibly relevant. It's also true that the fact that Dr. Hamblin has dark hair negligibly increases the odds that we're long-lost brothers.)

[Beowulf]

(It's true that, by themselves, the first two questions are negligibly relevant. It's also true that the fact that Dr. Hamblin has dark hair negligibly increases the odds that we're long-lost brothers.)

Not for long, though. Since he verging on gray quickly... (Ducking fast)

[asbestosman]

Alf is right is right in his analysis except for one thing:

The chances of Nahom being a coincidence?

[Analytics]

In my opinion, the chances of Joseph Smith

[Dan Vogel]

Beowulf,

It seems to me there is purpose in the different use of grammar here. The first implies that the name was already there, given by someone else. The second is plainly the name that Lehi's family gave it.

You are making unwarranted assumptions about the text, especially since

[Daniel Peterson]

I agree with Bill that NHM weighs to some extent on the

[asbestosman]

Maximum Likelihood is interesting, but it isn't always the best solution. It seems to rely on how you split up categories. We could ask which particular theory is most likely, in which case Maximum Liklihood works quite well. On the other hand, we can ask ourselves whether we should expect this particular event to happen or not. We would have to say that we expect no particular event to happen, but we still expect at least one of them to happen.

It is the paradox of the prefix. In a book there are many things written. For any particular thing written in it, you are 99.9% certain that it is correct. However, in the prefix you state that you are also 90% certain that the book contains a mistake. It is because the independent probabilities multiply. Since those numbers have a magnitude less than 1, multiplying them decreases their magnitude until you can be reasonably sure that there is at least 1 error among them.

Maximum likelihood is the strategy I'd use if I only get rewarded for being right, not for being close. Is that really the best way to handle this situation? I think a better option is to withold judgment.

In fact, with the previos coin example, a vital piece of information has been assumed instead of stated: the a priori chance that a fair coin will be used in .5 and likewise for the weighted coin. This a priori assumption is not always warranted. If the a priori chances were different, then the a posteriori chances would also be different.

Look at it this way. Let's say I have a fair die, a fair coin, and a weighted coin. If I roll a 1 on the die I flip the fair coin 10 times. If I roll 2-6, I flip the weighted coin 10 times. Now my chances become 1/(6*1024) = 1.6276e-4 for a fair coin vs 5/(6*1316) = 6.33232e-4 for a weighted coin.

Indeed, instead of maximum likelihood, what seems to be argued in the previous post is Bayesian conditional probability. Bayes is good, but I don't think we can use it without making lots of assumptions (just as we had an unstated assumption in the previous coin example).

And yes, Analytics, there is little doubt in my mind that Joseph Smith made use of the KJV to aid in the translation of the Book of Mormon. There is plenty of evidence for this in my mind. Everything from translating the word "Christ" instead of "Messiah" and "Jesus" instead of "Joshua" to the switching back and forth between "has" and "hath" to passages copied with little or no alteration from the KJV. But there seems to be many possible reasons this could be. Saying that the entire book screams 19th century origin seems a bit too hasty a judgment in my opinion.

[Dan Vogel]

Dr. Peterson,

I agree with Bill that NHM weighs to some extent on the

[DonBradley]

Asbestosman,

It seemed clear to me that Analytics used the KJV as an example, rather than claiming that every aspect of the BoM screams 19th Century for the sole reason that it uses the KJV.

BTW, you gave a nice, if brief, list of reasons to believe the BoM relies directly on the KJV.

I believe the use of the KJV in the BoM causes considerable problems for the latter's claimed antiquity. Because this is a new issue, I'll address it in a new thread.

Don

[asbestosman]

It seemed clear to me that Analytics used the KJV as an example, rather than claiming that every aspect of the BoM screams 19th Century for the sole reason that it uses the KJV.

I could have worded it better. I meant that just because there is evidence that the translator put his own spin on things, it does not seem like enough evidence to me to judge the book as 19th century fiction. There are so many other factors.

But keeping in the topic of this thread, I do concede that such things may be seen as evidence in favor of a 19th century origin. The question is whether it means the entire book is of 19th century origin, or if it merely means that there is a 19th century "accent" to the translation because of the translator. In other words how much weight, if any, does this evidence provide in favor of any theory?

Catch ya in the new thread sometime, hopefully.

[Daniel Peterson]

Inductive probability being the subjective enterprise that it is, especially as it is trying to find expression here, various value judgments should be no surprise.

What surprises me is not the fact that there are various estimations of NHM's evidentiary weight, but that you are apparently unwilling to grant it even the slightest smidgin of evidentiary value in favor of the Book of Mormon's antiquity.

Or have I misperceived your position?

I understand weighting it fairly highly. I also understand granting it relatively little importance as evidence. But I confess that an assignment to NHM of absolutely zero evidentiary weight strikes me as remarkably rigid and dogmatic.

[Analytics]

Maximum Likelihood is interesting, but it isn't always the best solution. It seems to rely on how you split up categories. We could ask which particular theory is most likely, in which case Maximum Liklihood works quite well. On the other hand, we can ask ourselves whether we should expect this particular event to happen or not. We would have to say that we expect no particular event to happen, but we still expect at least one of them to happen.

That is precisely the value I see in an MLE approach and why I offered that example; in MLE calculations, the specific events that are observed tend to be enormously unlikely, regardless of the actual values of the parameters being estimated. But despite the minuscule scale of the likelihood function, some estimates are a lot less likely than others.

Maximum likelihood is the strategy I'd use if I only get rewarded for being right, not for being close. Is that really the best way to handle this situation? I think a better option is to withold judgment.

But the whole object of the exercise is to come to a judgment; that is the question! What