When the editors of the scholarly journal Statistical Science decided, in 1994, to publish the paper “Equidistant Letter Sequences in the Book of Genesis,” they did not realize they were going to kick up a controversy that would last for more than 10 years. The question that the authors—Doron Witztum, Eliyahu Rips, and Yoav Rosenberg—investigated was whether secret messages were hidden in the Book of Genesis. The alleged messages pointed to events that took place only millennia after the writing of the Bible.

According to Jewish law, the Hebrew text of the Bible must not be changed even by a single letter in the course of myriad transcriptions. This is why, even today, the content of the Bible is believed by many to be identical to the text originally dictated to Moses by God on Mount Sinai.

The three authors believed they had found statistical proof for the existence of Bible codes. If the text of the Book of Genesis is strung along a line, without spaces, and letters are picked out at regular intervals, word combinations allegedly arise that make some sense. These words are called ELSs—equidistant letter sequences. (The intervals can be of any length, sometimes thousands of letters.) The Proceedings of the National Academy of Sciences had rejected the paper, but since the mathematical tools that were used seemed sound, Statistical Science agreed to publish it. The editorial board did not take the paper’s claim all too seriously, however, and in an introduction to the piece questioned its scientific validity. The purported discovery of Bible codes was not referred to as a scientific achievement but as a puzzle.

Witztum, Rips, and Rosenberg claimed that the ELSs

of pairs of words that bear a relationship lie closer together in the Book of Genesis than would be expected by pure chance. To prove their thesis, they inspected the names and dates of birth and death of 66 famous rabbis. (In Hebrew, numbers are represented by combinations of letters.) To the authors’ satisfaction, pairs of ELSs that referred to the same rabbi seemed to lie significantly closer than in randomized texts or when incorrect dates were assigned to the rabbis. This, they claimed, proved with high probability that the Bible predicted the appearance of Jewish scholars many centuries before they were born. A cryptographer at the National Security Agency, Harold Gans, followed up with an analysis of his own in which dates were replaced by the names of cities in which the scholars had been active. His study also seemed to show that the textual proximity of the ELS pairs was not solely due to chance.

The news that messages of the Almighty had been deciphered created a furor. In 1997 the first best-seller about the Bible codes appeared, and skeptics started to take notice. Brendan McKay, a mathematics professor from Australia, and Maya Bar-Hillel, Dror Bar-Natan, and Gil Kalai from Israel set out to debunk what they were convinced was pseudoscientific humbug. As was to be expected, the skeptics did not find any statistical evidence of hidden codes. Worse still, they charged that the data in the original paper had been “optimized,” a euphemism for the charge that Witztum, Rips, and Rosenberg had adjusted the raw material to suit their investigation. After appraisal by several statisticians, their assessment was published in Statistical Science in 1999.

If the editors had thought the controversy would now be laid to rest, they were utterly mistaken. The debate was only heated up by the second paper. It did not dampen the enthusiasm of the proponents of Bible codes that “secret” messages were soon also identified in Moby Dick and War and Peace. In this charged atmosphere, scientists at the appropriately named Center for Rationality at the Hebrew University in Jerusalem decided it was time to subject the question of Bible codes to a sober, scien-

tific analysis. A five-member commission was created and charged with getting to the bottom of the matter. It consisted of proponents of such codes and objectors and skeptics. Among the commission members were mathematicians of the highest caliber, like Robert Aumann (2005 Nobel Prize in economics), a leading mathematical expert in game theory, and Hillel Furstenberg, a world expert on an esoteric part of mathematics called ergodic theory.

Why does the test of the well-documented thesis turn out to be so difficult? One of the problems is that there are no vowels in Hebrew, and words appear much more often than in other languages, just by randomly juxtaposing letters. A randomly chosen collection of letters would spell out the name of a city, say Basle, with a probability of 1 to 12 million, while the same word in Hebrew, Bsl, would appear with a much higher probability of 1 to about 10,000. (There are only 22 letters in the Hebrew alphabet.) A more important reason for the controversy is that names can be written in different manners in Hebrew, especially when transcribed from Russian, Polish, or German. How, for example, should the German town where Rabbi Yehuda Ha-Hasid was active during the 12th century be written: Regensburg, Regenspurg, or Regenspurk? The resulting flexibility leaves researchers many degrees of freedom when preparing the database.

To eliminate any doubts about the data-gathering process, the commission charged independent experts with compilation of the place names. As a measure of precaution their identities were to remain anonymous and instructions were to be given in writing. After everything had been decided on and painstakingly recorded, the commission started its work—by throwing all instructions overboard. Experts were instructed partly in writing, partly orally, and part incorrectly. Some misunderstood the explanations; others made orthographic mistakes like confusing the Spanish towns Toledo and Tudela, Rabbis Sharabi and Shabazi, and places of death and of burial.

Then things started unraveling. Two members left the committee while it was still in its initial stages, and of the three remaining professors, one refused to sign the

final protocol. Consequently, the majority report that was published in July 2004 was issued by just two members (Aumann and Furstenberg). Two others wrote minority reports, while the fifth had lost all interest in Bible codes and no longer wanted to be disturbed. That two out of five committee members do not form a majority is only one of the inconsistencies in the committee’s work.

The “majority report” suggested that there was no statistical evidence of codes in the Book of Genesis—which is, of course, not equivalent to the statement that no Bible codes exist. The minority reports charged that the experiment was fraught with errors and therefore void of any meaning. In rejoinders, Aumann and Furstenberg rebutted the charges, and new statements followed. In the meantime the criticisms, refutations, responses, and rejoinders to the responses—all prepared with forensic meticulousness—fill binders. Words like liar, fake, and fraud that are rarely found in academic disputes were used by all sides. The three original authors publicly bet a million dollars that more ELS word pairs exist in the Book of Genesis than in Tolstoy’s War and Peace. There were no takers, but Furstenberg challenged the proponents of the codes to design a test that would be more meaningful. Probably the most poignant statement was made by Aumann: No matter what the evidence, everybody would stick to his or her preconceived ideas.

This page intentionally left blank.