The Bible Code Debunkers
I posted the following article here on my web site in September, 1999, following an article in Statistical Science debunking the Bible Code. I believe this article is still useful today. The predictions I made in this article have been largely fulfilled over the years.
--------------------
In August of 1994, Doron Witztum, Eliyahu Rips, and Yoav Rosenberg published an article in the American journal Statistical Science, presenting evidence that the book of Genesis contained encoded information about some dozens of rabbis who lived and died many centuries after Genesis was written. (This is the famous Great Rabbis Experiment.) These gentlemen didn't say outright that "therefore God wrote the Bible," but many of their readers have reached this conclusion and it's clear that Witztum, Rips, and Rosenberg also take this position.
The article was reprinted in Michael Drosnin's best-seller The Bible Code, and might therefore be the most-published scientific article in history! In the five years since the "StatSci" article appeared, plenty of critics have appeared. By far the most productive in word count (and in intellectual fire-power) have been the articles on the Web site of Brendan McKay, an Australian mathematician. McKay has a number of coauthors, including Dror Bar-Natan, Maya Bar-Hillel, Gil Kalai, Mark Perakh, and others, and he includes links to many other important articles on other sites.
In September of 1999, McKay, Bar-Natan, Bar-Hillel, and Kalai published an article in Statistical Science presenting a synthesis of several years of hard work. They've done an excellent job of presenting the case against the codes. Here I'll discuss their main points and suggest areas where they'll be attacked by the codes proponents.
First, though, let's note that some of the news releases got things garbled. One article claimed that the StatSci article predicted the assassination of Yitzhak Rabin and an earthquake in the year 2010. This is totally wrong. These were claims made in Michael Drosnin's sensationalist book. The claims of the StatSci article dealt mainly with the Great Rabbis Experiment, the alleged encoding of the birth and death dates of a set of rabbis in the book of Genesis.
OK, let's look at the new "debunking" article. The abstract reads as follows:
A paper of Witztum, Rips and Rosenberg in this journal in 1994 made the extraordinary claim that the Hebrew text of the Book of Genesis encodes events which did not occur until millennia after the text was written. In reply, we argue that Witztum, Rips and Rosenberg's case is fatally defective, indeed that their result merely reflects on the choices made in designing their experiment and collecting the data for it. We present extensive evidence in support of that conclusion. We also report on many new experiments of our own, all of which failed to detect the alleged phenomenon.
To paraphrase, they're saying that the Great Rabbis Experiment is bogus because Witztum, Rips, and Rosenberg (hereafter called "WRR") designed a fuzzy experiment that allowed them to find an "amazing result" that isn't really there. This fuzziness is generally called "wiggle room." I believe Caltech Professor Barry Simon coined this phrase.
McKay, Bar-Natan, Bar-Hillel, and Kalai (hereafter called "MBBK") ask two questions in their article, which I paraphrase here:
- Did WRR have enough wiggle room to get remarkable results, even if there is no Bible code?
- Is there any evidence that WRR actually exploited that wiggle room?
MBBK answer both questions with a resounding yes. Their 45-page article explains why. Here, I'll summarize their arguments (and I hope I'll not distort their case).
They begin by pointing out some technical statistical flaws, and refer the reader to a detailed 1998 critique by A.M. Hasofer. These are interesting, but not a showstopper, and MBBK conclude by saying,
"Serious as these problems might be, we cannot establish that they constitute an adequate explanation of WRR's results."
Expect WRR to argue that MBBK are nitpicking and have failed to make their case.
MBBK's next point is much easier to understand for the layman. The "amazing results" found by WRR in their second list of rabbis would have been far less amazing if they'd left out only 4 of the 32 rabbis they studied. And these 4 rabbis weren't the most historically important of the bunch. (Some details: WRR claimed odds of 1 in 60000. But without those 4 special rabbis, the odds would have nosedived to 1 in 30. What's so great about that?) MBBK note that,
". . . the result of the experiment is extraordinarily sensitive to many apparently minor aspects of the experiment design . . ."
This is bad news for those who believe in the Great Rabbis Experiment. It means that it's easy to "cheat" by finding wiggle room in only a few places. Expect WRR to argue vehemently that they didn't cheat, they conducted the experiment with extreme objectivity, and those 4 rabbis are certainly important folks.
The point remains that MBBK have shown that a little wiggle could in principle go a long way. But in order to wiggle the experiment, you've got to have something to wiggle. Is there actually such wiggle room?
Yes! In their next section, MBBK consider the many sources of wiggle room.
Some are minor: the method of choosing the rabbis and the method of choosing the birth and death dates. Expect WRR to argue that these just don't make that big a difference.
But some are major. MBBK list eight different ways to write dates in Hebrew. WRR used only three of these. Oddly enough, these three tend to give excellent results in the Great Rabbis Experiment; the other five ways of writing dates would have yielded very ordinary results for the experiment. This raises a major question: Why didn't WRR use all eight methods? Expect WRR to argue that those three methods of writing dates are the most natural.
But the biggest source of wiggle room in the Great Rabbis Experiment is the choice of "appellations" for the rabbis. Many of the rabbis were known by several names, abbreviations, acronyms, etc. Many (but not all) of these "appellations" were used in the Great Rabbis Experiment. And these appellations have a number of variant spellings. From these variants, WRR chose a particular set. Again, it's very odd that the choices WRR made yielded "amazing results," while most of the other available choices would have yielded much poorer results.
This was brought home in 1998, when McKay and Bar-Natan published an article on McKay's Web site with the results of their own "Great Rabbis Experiment." In their version, they used a set of appellations similar to the one used by WRR, but differing in a fraction of the appellations. And they found "amazing results" -- in the Hebrew translation of Tolstoy's book War and Peace. Unfortunately, they found nothing special at all in Genesis.
Doron Witztum has long ago sharply critiqued the McKay/Bar-Natan paper, claiming that they made many mistakes (whereas he claims that his original version in the WRR paper was correct). Recently, an anonymous person has posted an exhaustive examination of this question on McKay's Web site. The conclusion: McKay and Bar-Natan's experiment is at least as valid as the original WRR paper. (Though one expert has called both sets of appellations "equally appalling!") The argument is intensely technical, and you won't follow it unless you're very knowledgable about rabbinic literature. Expect Witztum to counter this latest paper with his own detailed rebuttal.
But there's a much simpler problem with the Great Rabbis Experiment. This was pointed out by A.M. Hasofer in 1998 in an article posted on McKay's Web site. The problem is that Genesis is not considered "more inspired" than Exodus, Leviticus, Numbers, or Deuteronomy. Therefore, if God encoded information about the great rabbis in Genesis, He presumably also included similar information in the other four books of the Torah. (Otherwise, how would we know where to look?)
Now here's the clincher: If you perform the Great Rabbis Experiment, exactly as WRR specify, on Exodus, Leviticus, Numbers, or Deuteronomy, rather than Genesis, you get . . . nothing special at all.
So if Witztum, Rips, and Rosenberg had gone looking in one of the other four books, they'd have found nothing! That's a serious problem. The Bible code proponents justify their search for codes by saying that the Torah encodes vast quantities of information -- about every person, animal, plant in the world. The clear implication is that you need only to go looking anywhere in the Torah and you'll find codes. Expect WRR to find a natural reason why rabbis should be found in Genesis.
The above arguments make it abundantly clear that the Great Rabbis Experiment has enough wiggle room, in principle, to entirely explain its "amazing results."
Next, MBBK ask whether there is evidence that the data was actually wiggled. They conclude that it was:
". . . there is significant circumstantial evidence that WRR's data is indeed selectively biased towards a postive result. We will present this evidence without speculating here about the nature of the process which lead to this biasing. Since we have to call this unknown process something, we will call it tuning."
One piece of circumstantial evidence is that the great majority of minor modifications that you might make to the Great Rabbis Experiment result in much less "amazing results," and sometimes such minor changes lead immediately to "no results worth talking about." Since many of these minor changes in the experiment seem perfectly reasonable, this suggests that the experiment has been somehow optimized (that is, "tuned.") This evidence is probably the strongest part of MBBK's case.
Note: As I made clear in my own book on the subject, I don't believe that this tuning was done intentionally. But I do believe it happened -- not only because of the data MBBK present, but also because my own analysis shows that no substantial information is encoded in the Torah. However, WRR won't see it that way. Expect them to argue that the Great Rabbis Experiment, as performed, was more natural than any other variation suggested by MBBK. I have no prediction for how they'll respond to my own work.
MBBK present other circumstatial evidence of tuning, namely the surprisingly close agreement between the two sets of "amazing results" found in the Great Rabbis Experiment. You can make a statistical argument that this close agreement is "too close to be true," but the argument is not watertight. Expect WRR to argue that this evidence is pretty weak.
Stronger evidence (at least to MBBK) is the fact that they've created their own "unbiased" attempts to find Bible codes and haven't found any. It's clear that they believe they've avoided bias quite well, just as WRR believe they have. Yet the layman might be hard pressed to guess which group is really unbiased and which one has bungled the experiment. It seems unlikely that both groups could be right. Whom should the poor reader trust? Unless you're an expert, you may find it hard to decide. Expect WRR to make a spirited attack on MBBK's experiments.
Fortunately, there is plenty of other evidence for deciding the case in favor of MBBK.
MBBK present a final argument against the codes based on the history of transmission of the text of the Torah. In a nutshell, most Biblical experts agree that a great many letter-variations have crept into the text through the last three thousand years. A few dozen such variations in any book would spoil the codes in that book. While we don't have access to the original manuscripts, the surviving manuscripts do in fact show dozens of variations, and we have reason to believe there are hundreds of them. This is also an extremely strong case. Expect WRR to argue that their experiments demonstrate the purity of the Koren text; therefore, any arguments about the transmission history must be invalid.
In summary, McKay, Bar-Natan, Bar-Hillel, and Kalai have presented a vast amount of data that compellingly undermines the credibility of the Great Rabbis Experiment. If you'd like to read the full document, it's available at this page on McKay's Web site. The other articles by McKay and Bar-Natan, and by Hasofer are also available at the same place.
It's very likely that Doron Witztum and Eliyahu Rips will soon respond with a rebuttal to the debunkers. They've done so repeatedly over the last few years. They'll try again now. Just remember, you heard their arguments here first!
