Friday, March 14, 2003

Bible Code Madness From The New York Times comes this item about the lengths to which intelligence officials are willing to go in their search for Osama Bin Laden. It seems they granted an audience to Michael Drosnin, author of the bestsellers The Bible Code and The Bible Code II, who claims that hidden codes in the Bible can be helpful in locating bin Laden.

The Bible Codes first became big news in 1994 with the publication of the paper "Equidistant Letter Sequences in the Book of Genesis" by Israeli mathematicians Doron Witztum, Eliyahu Rips and Yoav Rosenberg. The paper appeared in the highly regarded journal Statistical Science. Equidistant Letter Sequences, or ELS's, refer to sequences of letters taken from the text of Genesis in which the letters are spaced at equal intervals. Thus, we might start at some arbitrary letter and look at every third letter thereafter, or every seventh letter, or every 24th letter. When we do this we sometimes find that a coherent word or sentence is spelled out. Witztum, Rips and Rosenberg performed an experiment in which they compiled a list of famous rabbis together with their birth or death dates, and looked for ELS's corresponding to their list. For most of their rabbis they were able to find an ELS spelling the rabbi's name, with a second ELS corresponding to the birth or death date located nearby.

Of course, a moment's thought reveals that in any sufficiently long text there will be so many possible ELS's that simply by chance some of them will form coherent words. The significant part of the original paper was the claim that its findings were far beyond anything that could plausibly be attributed to chance. If this claim was correct the implication was obvious. If the sequences were not attributable to chance and if no human author could have been clever enough to plant them himself, we are led to the conclusion that some supernatural intelligence planted them there.

Drosnin exploited this idea in his books, arguing that these codes could be used to predict future events, an idea that Witztum, Rips and Rosenberg wanted no part of. Drosnin's claim would have had more merit had he actually predicted a surprising event on the basis of his analysis of the codes. Alas, all he actually did was take events that had already occurred, and retroactively find ELS's in the Bible that corresponded, often roughly, to them. In fairness, Drosnin sent this letter in reply to the Times' piece. Suffice it to say that his claim that "the Bible Code keeps coming true" is open to debate.

The Codes Explained Due to the extravagant nature of the claims, the paper went through an unusually thorough process of peer review. All the referrees came to the same conclusion: They were pretty sure there was some flaw in the experimental design, but there was no obvious flaw and therefore the paper merited publication. Subsequent research by Brendan McKay, Dror Bar-Natan, Maya Bar-Hillel, and Gil Kalai found the elusive flaw. Briefly, there were some idiosyncrasies in the way the list of famous rabbis was compiled. There is a certain amount of freedom in how a name or a date is spelled out in Hebrew characters, and it seems that this flexibility was exploited in ways that made the results seem more statistically significant than they were. It must be said that there is no allegation of fraud here, only a certain amount of carelessness. When the experiment was repeated with this flexibility taken into consideration, researchers were able to reproduce the original results in non-religious books like War and Peace. These results were subsequently published in Statistical Science as well. Which leads us to conclude either that the codes are not good evidence for the divine origin of the Bible, or that War and Peace was authored by God as well.

This site contains the original paper, the published refutation, and several other resources as well.

Religious Bias?Incidentally, note that the referrees recommended publishing this paper despite their reservations and despite its religious overtones. This shows that the creationist complaint that scientific journals are too closed-minded to accept papers with religious themes is not corrrect.


Post a Comment

<< Home