Books in Review

Copyright (c) 1998 First Things 85 (August/September 1998): 61-64.

The "Bible Code" is the name for computer-generated sequences of letters taken from the Hebrew Bible. Researchers string all the letters together, deleting spaces between words. Then, instead of running through the text letter by adjacent letter (as we do in ordinary reading), they run through the text by skipping a fixed number of letters. The resulting sequences, known as equidistant letter sequences (or ELSs), are then inspected for patterns that cannot reasonably be attributed to chance. The Bible Code comprises such equidistant letter sequences.

The Bible Code is controversial because some have presented it as a preprogrammed time capsule set to go off once humans invent computers. The human authors of the Bible, writing well before the advent of computers, would have been incapable of consciously introducing into the Bible the patterns that Bible Code researchers are finding by means of computers. Hence these patterns, if not attributable to chance, must stem from a non-human intelligence. Moreover, if the patterns contain information about subsequent events in world history, this nonhuman intelligence would also have to possess preternatural foreknowledge. And since the Bible claims to be inspired by precisely such a being, the most obvious solution to the identity of this nonhuman intelligence is the God of the Hebrew Scriptures, to wit, YHWH. Here, in broad strokes, is the logic underlying the Bible Code.

Jeffrey Satinover’s Cracking the Bible Code is the place to begin for anyone interested in the subject. It is engagingly written, well-informed, and generally sober. In particular, it avoids the statistical fallacies too commonly associated with coverage of the Bible Code. Even so, it is an uneven text. It is superb in describing the history leading up to today’s investigations. Indeed, the window it provides on Jewish intellectual life from the Middle Ages through the present is itself worth the price of the book. It is also very good at making intelligible the mathematics needed for deciding whether the Code is genuine. On the other hand, its speculative portions about the significance of the Bible Code are often diffuse and controversial.

The best known example of the Bible Code, and one Satinover treats in detail, is the Great Sages experiment. Several Israeli mathematicians took thirty or so prominent rabbis from the Encyclopedia of Great Men of Israel, and looked among the equidistant letter sequences of Genesis for a juxtaposition of rabbi names with their dates of birth or death. They found some. On calculating the improbability of so close a match between names and corresponding dates, the mathematicians concluded that the match couldn’t have happened by chance (the probability was less than 1 in 60,000). The journal Statistical Science tacitly agreed. After an arduous review, its editors finally decided to publish the results of this experiment. Suddenly the Bible Code had a measure of academic respectability.

Nevertheless, one paper in a prestigious journal is hardly enough to settle so controversial a topic. The Bible Code raises a number of sticking points. For instance, even if a nonhuman intelligence can convincingly be shown to have introduced information into the Bible, the identity of that intelligence remains controversial. That’s not to say that the Orthodox Jewish preference for identifying this intelligence with YHWH wouldn’t be the most plausible one. But New Age adherents will be sure to have their own preferences. Moreover, if other sacred books exhibit codes, we may enter a new form of theological disputation: My code is better than yours.

Another sticking point concerns how to use the Bible Code should it prove genuine. There are roughly two approaches, one modest, the other extravagant. The modest approach uses the Bible Code simply to authenticate the Bible, showing that there is a nonhuman intelligence behind the text. For example, Satinover describes how Bible Code researchers looked among the equidistant letter sequences of Leviticus 1:1-13 and found an overwhelming number of references to the high priest Aaron. What led the Bible Code researchers to look for Aaron’s name were the persistent references throughout this passage to the Aaronic priesthood, yet no references to Aaron himself. In this passage of thirteen verses one would expect on average to find eight references to Aaron among the equidistant letter sequences. There are actually twenty-five. The probability of this happening by chance is less than 1 in 2,000,000.

This approach to the Bible Code as a scheme for authenticating the Bible is modest and seems to me the most promising course for deciding the genuineness of the Code. There is, however, also an extravagant approach. It uses the Code for divination, retrodicting the past as well as predicting the future. Satinover explicitly adopts the modest approach, stressing that the Bible Code’s only legitimate function is to authenticate the Bible. Nevertheless, Satinover and most of the Bible Code advocates find the lure of the extravagant approach hard to resist. Thus Satinover describes, without criticism, how Israeli intelligence during the Gulf War used the Bible Code to predict Iraq’s first Scud attack on Israel.

The key problem confronting the Bible Code, at least for now, however, is whether the patterns found in the Bible’s equidistant letter sequences decisively preclude chance. Throughout Cracking the Bible Code one finds references to "p-values," some of which are quite small. "P-value," short for "probability value," refers to the probability that a given pattern might be present in the equidistant letter sequences by chance. The smaller this probability, the more compelling the inference that a nonhuman intelligence influenced the writing of the Bible. There is, however, a catch.

A small p-value by itself isn’t enough to preclude chance. In addition, the pattern for which one calculates a p-value must be independently given. The Bible Code researcher can’t just willy-nilly search through equidistant letter sequences until something interesting or unusual turns up. Statisticians call this cherry picking or data snooping, and it vitiates any statistical analysis. Patterns must be given independently of the search, and only then can small p-values indicate the finger of God.

To Satinover’s credit, he is fully aware of the dual requirement that p-values be small and that patterns be given independently. Moreover, despite his enthusiasm for the Bible Code, he never loses sight of this requirement. By keeping to a sound statistical methodology throughout, Satinover earns the credibility that others writing on the Bible Code fail to achieve.

If all that is required to demonstrate divine agency is to find an independently given pattern that also has a small p-value, why does the Bible Code remain controversial? One problem is that Bible Code research is stretching what statisticians mean by "independently given patterns." Moreover, the p-values are not always easy to calculate. The Bible Code is forcing statistics to extend its research frontiers, an unprecedented case of biblical studies advancing statistical research.

Although it’s too early to decide whether the Bible Code is genuine, Satinover makes a good argument for taking it seriously. The credibility of the key players is significant. The main figures involved in this debate, whether critics or advocates of the Bible Code, are top rung, well-credentialed mathematicians. All the fundamental research on the Bible Code has been done openly, and without a hint of deception. It is notable that advocates of the Bible Code have jeopardized their careers, converted to Orthodox Judaism, refocused their research, and in some cases changed disciplines.

Statistical analyses, though yet to decide this debate, have at least been suggestive. Highly improbable, independently given patterns seem to occur in the Hebrew Bible with an obstinacy displayed in no other text. What’s more, if this phenomenon had to occur reliably in some ancient text, there would be no better candidate than the Hebrew Bible, especially its first five books, known as the Torah.

The textual transmission of the Hebrew Bible, and especially of the Torah, is far more accurate than anything else in antiquity. Because the Bible Code is sensitive to copying errors, the Hebrew Bible satisfies a crucial necessary condition for possessing such a code. Indeed, if the Bible Code ultimately proves genuine, we can credit its discovery to generations of obsessive-compulsive scribes who meticulously transcribed the Hebrew Bible. We might then even regard their obsessive-compulsiveness as a divine grace.

The philosopher Bertrand Russell was once asked why he didn’t believe in God. He replied, "Not enough evidence." Satinover’s fascination with the Bible Code is that it may provide evidence for God’s existence that would have convinced even a Bertrand Russell. There has been a long dismal trend of science eroding faith in God. In the Bible Code Satinover sees a unique wedding of science and theology, wherein each reinforces the other.

But perhaps that wedding is not unique. At the same time that research in the Bible Code has taken off, research in a seemingly unrelated field has taken off as well, namely, biological design. These two fields are in fact closely related. Indeed, the same highly improbable, independently given patterns that appear as the equidistant letter sequences in the Bible Code appear in biology as functionally integrated ("irreducibly complex") biological systems, of the sort Michael Behe discussed in Darwin’s Black Box.

The relevant statistical methodology is identical for both fields. As a result, the two fields stand to profit from each other. For instance, my forthcoming book, The Design Inference, gives a thorough account of universal probability bounds, i.e., how small a p-value one needs to eliminate chance decisively. (Although the literature on universal probability bounds dates back to the French probabilist Emile Borel, it seems not to have been engaged by the Bible Code researchers.)

This convergence of the Bible Code and biological design should not seem surprising. There is a tradition within both Judaism and Christianity of speaking of two "books" where God reveals himself—the Book of Scripture, which is the Bible, and the Book of Nature, which is the world. I commend Jeffrey Satinover for his efforts to read both books.

William A. Dembski, a mathematician and philosopher, is a fellow of the Center for Renewal of Science and Culture at the Seattle-based Discovery Institute. His new book, The Design Inference, will be published by Cambridge University Press this August.