Three weeks before the Gulf War began, Eliyahu Rips, an Israeli mathematician, decoded from Genesis the date on which Iraq would launch its first Scud missile against Israel. More than a year before Rabin's assassination, his method found the event predicted in Deuteronomy. Journalist Michael Drosnin describes Rips' method and some of its yet-unfulfilled prophecies informally and nonmathematically and relates his own struggle as an agnostic with the implications of twentieth-century details embedded in a book more than three millennia old.

If Rips' method is sound, evangelicals should exploit it for exegesis. In fact, several books by Christian authors use the method to find hidden messianic prophecies. While Rips' credentials and technical claims and the predictions themselves are impressive, his approach is lacking philologically, mathematically, and theologically.

The "Bible Code method" is built on two principles: skipping letters, and proximity.

The first principle is that letters separated by equal numbers of intervening letters may be read consecutively to yield a work or phrase. The computer searches the consonantal MT for a given word by first skipping no letters, then one, then two, and on up to several thousand if necessary. The successive characters of "Yitzhak Rabin" in Deuteronomy 2:33-24:16 are separated by 4,771 letters.

Most of the expressions discovered by skipping are single words or short phrases. Forming a coherent prediction requires combining several encoded expressions (perhaps using different skips) that are close to one another. The prophecy of Rabin's assassination consists of three such expressions: "Yitzhak Rabin" with a skip of 4,771 letters, "assassin will assassinate" with zero letters skipped (the plain text of Deuteronomy 4:42), and the name of the assassin (discovered after the event), "Amir," in reverse order with a skip of 8 letters (Numbers 33:14-15).

The tradition of patterns among nonconsecutive letters in the Old Testaments includes acknowledged acrostics in Psalms and Lamentations, the hidden Tetragrammaton in Esther, and the cabala. In spite of its long pedigree, however, Drosnin's treatment is not persuasive.

Concerning philological issues, Drosnin reveals only a superficial knowledge of nature and history of the Biblical text. Most of his errors do not directly affect the book's argument, but one is fatal. He claims: "All Bibles in the original Hebrew language that now exist are the same letter for letter" (page 194). A glance at the apparatus of BHS shows numerous ms variations in the consonantal text. Adding or removing a single character in the midst of a skip sequence will throw off the sequence and destroy the encoding. Given this sensitivity to textual variation, even if someone did encode messages 3,000 years ago it is unlikely that they would be recoverable from the MSS that exist today.

Concerning mathematical issues, biblicists may be intimidated by the scholarly article by Rips and his colleagues in the book's appendix. However, a more careful look at the article, a simple example, and some unexpected messages provide some perspective.

Rips and his colleagues give a rigorous definition of the method in Statistical Science 9 (1994, pages 429-438), a refereed journal for professional statisticians. The study searches Genesis for the names and birth or death dates of famous rabbis as recorded in Margalioth's Encyclopedia of Great Men in Israel (1961). It finds collocations of a rabbi's name and date more frequently in Genesis than in other texts, including Isaiah and the Hebrew translation of War and Peace, and it argues that this difference cannot be explained by chance. Drosnin repeatedly emphasizes the lack of any refereed challenge to this scholarly study.

Two cautions are appropriate here, First, as any active researcher knows, peer review is no guarantee that a study is correct, and much valuable work circulates informally for many years before reviewed publication. Brendan McKay, an Australian mathematician, has attempted without success to replicate Rips' effect (http://cs.anu.edu.au/~bdm/dilugim). The selection of a test sample (such as the names and dates used by Rips) is extremely sensitive to researcher bias, and such an effect may be responsible for Rips' results.

Second, even if Rips' paper were substantiated, it does not validate Drosnin's more popular predictions. These predictions have not been subjected to the elaborate statistical tests defined in the paper. Furthermore the paper shows not only the presence of coded material in Genesis but also the absence of such material in Isaiah, invalidating Drosnin's search of the entire Biblical text for hidden prophecies.

Drosnin's prophecies seem too striking to result from chance. This intuition is not valid, because the analyst controls many options (statistically, "degrees of freedom") in searching for messages. (1) A sequence of unpointed consonants generated by a skip can often be read in multiple ways. (2) Words can be spelled either forward or backwards. (3) When using a zero skip, word boundaries are ignored. (4) The searcher chooses the skip length and where to begin skipping. (5) Dates can be written in several different ways. (6) The words sought are not specified in advance but chosen as analysis proceeds. These variables create a rich palate from which a diligent analyst can construct almost any message.

Michael Weitzman, reader in Hebrew at University College, London, published a helpful review of The Bible Code in The Jewish Chronicle, July 25, 1997. He offers the helpful example of searching for a given six-character phrase (longer than many of Drosnin's items) in the Pentateuch. Assuming equal frequencies, the chance of randomly picking a specified Hebrew letter is one out of 22 (1/22). The chance of randomly picking two specified letters in a given order is (1/22)*(1/22). Thus the chance that a specified six-letter phrase would randomly occur is 1/(22*22*22*22*22*22), or about 1/10,000,000. These odds seem incredible, but the analyst gets to choose the letter at which to begin (300,000 options), the interval to skip (on average 30,000 options), and whether to spell the phrase forward or backwards (two options). Thus the Pentateuch yields about 300,000*30,000*2, or about 18,000,000,000 six-letter phrases. The probability that one of these will be the desired one is thus 180/1.1, or about 163-that is, the required phrase should occur by chance not just once but more than 160 times. The occurrence of multiple words near each other seems less likely, but the large number of potentially relevant words and the fact that only one letter in each of two "nearby" words actually has to be nearby once again make interesting patterns inevitable.
 
 

To show that the method can yield any desired message, McKay presents a detailed argument from the book of Revelation that Bill Gates is the antichrist, ten arrays predicting six different twentieth-century assassinations from Moby Dick, and a single region of the Hebrew translation of War and Peace (Rips' control text) that contains at least 59 words about Hanukkah as well as the names of the analysts. Weitzman offers another instance from Jeremiah 8:8, whose last two words are "deceit" and "scribes." The roots of "scribes" means "to count," so "scribes" are really "counters," people who tally Biblical letters to find hidden messages, and the first three letters of the Hebrew word spell Rips' name backwards. Unfortunately, the immediately previous word declares all this effort to be "deceit."

In addition to the philological and mathematical shortcomings of the method, it presents at least two theological problems to those who accept the Bible as the Word of God.

First, its predictions sometimes fail. Drosnin reports an elaborate prophecy of world war to have started on July 25, 1996, an event that did not transpire, and much of the book speculates about the deeper philosophical meaning of failed prophecies. For the believer, speculation is unnecessary. Failure means that the prophet is not from the Lord (Deuternomy 18:21-22) and therefore not to be trusted.

Second, the book exemplifies the human lust to seek out and believe a hidden meaning in the Scriptures while rejecting their open teaching. The Bible itself claims to be clear and patent to the believer. We do not need computers or statistical analysis to understand its message. It is God's communication to an unsophisticated people, a plain and simple message to confound the wise and mighty (1 Corinthians 1:26-29), hidden from the wise and prudent and revealed unto babes (Matthew 11:25). These are indeed "secret things" that "belong unto the Lord our God," but we are to be preoccupied with "those things which are revealed," and that not for abstract theological speculation but "that we may do" all that God commands us (Deuternomy 29:29). One who believes these testimonies will realize that even if the Bible contained hidden messages they would be less important than understanding and obeying the plain teachings of the text. Drosnin's attitude is just the opposite. He enthusiastically promotes the importance of hidden messages while rejecting the most explicit teachings of the Scriptures concerning the existence of God and the nature of his revelation.

This inconsistency between the "Bible Code method" and the plain sense of Scripture holds a warning for biblical studies in general. Some scholars feel that their particular specialty, arduously cultivated through years of graduate study and professional research, is essential to understanding the true message of the Bible and that those who ignore those arcane investigations will forever be ignorant. Biblical scholarship sounder than The Bible Code can indeed illuminate, clarify, and illustrate the plain teaching of the Bible, and it is a blessed privilege to have the time and training for formal exegesis. But such research is the icing on the cake, not the main course. It is not the key without which the Scriptures remain closed. Rather, it is supplementary to the personal reading and meditation that are the privilege of every believer.

In sum, the prophecies of The Bible Code result from the interplay between coincidental distributions of letters and investigative creativity. There is no scientific reason to believe them to be intentional messages encoded millennia ago. The method is fundamentally flawed philologically, mathematically, and theologically, and it is useless for serious exegesis. The book's popularity warns of the human weakness to prefer the hidden and sophisticated over the plain and simple. It exhorts us as believing scholars to pursue and package our work in a way that encourages lay people to diligence in their own interaction with the text rather than persuading them to surrender their individual theological discretion either to biblical scholars or to Bible decoders. 

Back to Contents Does God Exist?, NovDec99.