Welcome to The Skeptics Digest, wherein we discuss scientific skepticism, its uses and misuses, and where we apply the methods of skepticism to controversial subjects.
The previous diaries can be found here:
The Skeptics Digest: What is Skepticism? by rfall
This diary set the basis for the diary series, and defines the meaning of "skepticism" for purposes of the discussions therein.
The Skeptics Digest: Superstition and Fallacies by XNeeOhCon
This diary deals with defining and providing examples of a number of common logical fallacies committed in the process of researching controversial issues.
While we do not have a set schedule for diaries in this series, it appears to be working out nicely that we will post a weekly diary, on Sundays, for the foreseeable future.
For our first application of the methods of skepticism I've chosen a topic which is somewhat dated, and which will likely not have many adherents here among DKos members, but which is a "soft start" for the process of applying skeptical methods: Bible Codes.
Bible Codes, for those not familiar with the subject, are:
The Bible code, also known as the Torah code, is a series of messages alleged to exist within the Bible text, that when decoded form words and phrases supposedly demonstrating foreknowledge and prophecy. The study and results from this cipher have been popularized by the book The Bible Code. Wikipedia
Before delving into the subject of Bible Codes, we need to define a term or two:
The root of all Bible Code theories is the concept of Gematria:
Gematria or gimatria (Hebrew: גימטריה, gēmaṭriyā) is a system of assigning numerical value to a word or phrase, in the belief that words or phrases with identical numerical values bear some relation to each other, or bear some relation to the number itself as it may apply to a person's age, the calendar year, or the like. Wikipedia
Gematria itself is a subset of the practice of numerology, in which a connection is made between numbers and physical objects or persons.
Numerology and gematria go far back in human history, and appear to have emerged as a mechanism for drawing a connection between the easily understood reality and rules surrounding mathematical concepts and the natural world, which often appeared unruly and difficult to predict. Perhaps, the thinking went, the Creator hid the secrets to a better understanding of the Universe in mathematics for Man to find.
(The ancient roots of numerology and gematria are neither proof nor disproof of their correctness or "truth value", but do indicate that we humans may have a deep-seated drive to map "numbers" to the "real world".)
The most recent upswelling of interest in Bible Codes was driven by the publication in 1997 of The Bible Code by Michael Drosnin, and a sequel, The Bible Code II in 2002. In both of these books, Drosnin, an American journalist and author, claimed that applying the methods of gematria and utilizing computer programs which made searching for patterns much faster he had found "messages" in the Bible which accurately foretold events—some of which had already occurred and some of which are yet to occur. The most famous of these, according to Drosnin—and the one which turned him from a "skeptic" to a believer, was the "prediction" of the 1995 assassination of Israeli Prime Minister Yitzhak Rabin.
Drosnin also believes that the Bible Codes were placed there by extraterrestrial intelligences—a claim which we will not examine here in conjunction with the Bible Code claims, but will leave for a more general examination in a later Skeptic's Digest.
So what of the claim that there are "messages" in the Bible which can be decoded and which foretell future events?
Here's how the Bible Code works:
The code is revealed by searching for equidistant letter sequences (ELS). The code is called the Bible Code or the Torah Code. For example, start with any letter ("L") and read every nth letter ("N") thereafter in the book, not counting spaces. If an entire book such as Genesis is searched, the result is a long string of letters. Using different values for L and N, one can generate many strings of letters. Imagine wrapping the string of letters around a cylinder in such a way that all the letters can be displayed. Flatten the cylinder to reveal several rows with columns of equal length, except perhaps the last column, which might be shorter than all the rest. Now search for meaningful names in proximity to dates. Search horizontally, vertically, diagonally, any which way. Skepdic.com
In the case of the Bible Code, the "letters" referenced here are in Hebrew, though some Bible Codes and other gematria-based letter codes use English or some other language.
From The Straight Dope: What's up with the "Bible Code"? comes this description of a simple version of the Bible Code methodology:
Ancient Hebrew has no vowels and no punctuation. Furthermore, the letters can also represent numbers. Rather than try to teach you ancient Hebrew, let's use English as an example. First, we need to set up a number equivalency, similar to ancient Hebrew, as follows:
A = 1, B = 2, C = 3, D = 4, . . . , H = 8, I = 9, J = 10, K = 20, L = 30, M = 40, . . . , S = 100, T = 200, . . .
Got it?
Now, let's take some text, let's take our title: The Straight Dope by Cecil Adams. First, we take out all the vowels, and get THSTRGHTDPBYCCLDMS.
Next, we put these 18 letters into a 2 x 9 matrix array, as follows:
T H S T R G H T D
P B Y C C L D M S
With me so far? Now, we play "word find." We look vertically, horizontally, diagonally, slantwise, forward, or backwards to see what we find in the way of words. Notice, right away, the letters "HST" and almost adjacent the letters HM (on the diagonal). Hmmm. Now, numerically by our system, HM would be 48, right? And HST is clearly Harry S Truman, right? And so we have a prediction that Harry Truman won the 1948 election! Astounding! Remarkable! Incredible!
There are several things to note about this process, relevant to a skeptical analysis:
- The mapping of letters to numbers in the analysis process is up to the Bible Code "researcher".
- Even within one application of the Bible Code method, changes can be made to the starting point, the letter skip number and these methods can be reapplied to the resulting letters to achieve "significance".
- "Meaningful" is in the eye of the beholder—making an objective analysis somewhat difficult.
- Often the gematriatic analysis is done on the letters in one language, resulting in "words" in that same language, which are then "translated" to the speaker's native language—English, in most cases.
- Adding punctuation and spaces between letters, and capitalization (all of which are not a part of the gemtria method itself) often changes the "meaning" of the "sentence", adding a level of flexibility to the outcome. "Newt" versus "newt"—Former Speaker of the House or a lizard?
The Straight Dope columnist has this to say, as a result:
My point is that finding initials and names and dates in an arbitrary "word find" array is not hard. Especially if you're dealing with no vowels, and with numbers equivalent to letters. In a text that is 300,000 letters long, there should be hundreds of such coincidences. And if you can re-matrix as you please, the number of coincidences increases enormously.
So the methodology is suspect from the start. Skeptics have found incredible word combinations ("predictions") in other works (even in English) such as Moby Dick and War and Peace.
As a particularly devastating example of this ability to find messages in works other than the Bible, Drosnin originally said that applying his Bible Code methods to other works didn't produce similarly meaningful results. He so believed this that he challenged anyone to find some of the same messages in other texts.
When my critics find a message about the assassination of a prime minister encrypted in Moby Dick, I'll believe them.
Rising to this challenge, at least one researcher—Mathematician Brendan McKay of Australian National University—found:
...assassination "predictions" in Moby Dick for Indira Gandhi, Rene Moawad, Leon Trotsky, Rev. M. L. King, and Robert F. Kennedy (see http://cs.anu.edu.au/...). CSICOP
Unsurprisingly, Drosnin disputed the meaningfulness of these results, but McKay used the same methods as Drosnin applied to the Bible, so it's difficult to know why he feels this way.
In the case of The Bible Code, an even more stunning "miss":
...the code contains predictions of disasters and an apocalypse to occur between 1998 and 2006. Wikipedia
Even allowing for some leeway in the definition of "apocalypse" to include a worldwide financial meltdown, the prediction seems not to have been well borne out.
For those interested in a more in-depth and mathematically oriented analysis of the Bible Code methodology, there are many sources of information. Among the best:
- The CSICOP article linked above: Hidden Messages and The Bible Code has some easy-to-understand mathematics in this article, as well as examples of number of "hidden messages" which can be found in common works, makes this a good starting point.
- A follow-up article: Bible-Code Developments is also worth a read.
For our purposes, let's just look at a few of the logical fallacies that are part of the Bible Code phenomenon which should cause a true skeptic to pause in wholesale acceptance of the technique:
Selective Observation, or "See the Hits and Ignore the Misses".
The promoters of hidden-message claims say, "How could such amazing coincidences be the product of random chance?" I think the real question should be, "How could such coincidences not be the inevitable product of a huge sequence of trials on a large, essentially random database?"
An important contributor to this pattern of seeming coincidences is the fact that "misses"—which in this case would mean combinations of derived "messages" which are dismissed as "not meaningful" out of hand—are ignored in determining whether the pattern of "hits" has any statistical meaning.
As elegantly described in the CSICOP article:
Let’s look at Drosnin’s approach with a lottery analogy. The probability of winning a lottery with a single ticket is very small, and Drosnin says the probability of getting an improbable match (such as "Clinton" and "president") is also very small. But what happens if you buy more than one ticket?
In the New Mexico "Daily Millions" lottery, the odds of winning the $1 million jackpot with just one ticket are about ten million to one against. With two tickets, the odds plummet, to about five million to one. If you buy one million tickets, your odds drop to only about ten to one against. And if you invest $10 million in tickets, the odds become approximately two to one in your favor! Most people can't afford to buy millions of tickets. Those who do have that kind of money usually don't dump it on the lottery, because you almost always end up losing.
But in Drosnin’s game, you don't have to win more than you lose. You don't even have to break even. All you need for success is to win every once in a while. And, you can have what amounts to millions of "free lottery tickets" simply by running a computer program, or poring over crossword-puzzle printouts.
Hindsight bias.
Hindsight bias is the tendency to construct one's memory after the fact (or interpret the meaning of something said in the past) according to currently known facts and one's current beliefs. In this way, one appears to make the past consistent with the present and more predictive or predictable than it actually was.
Hindsight bias is the equivalent of shooting an arrow at the side of a barn, then painting circles around it and declaring it a bulls-eye. With a nearly infinite number of past events, any combination of "meaningful" words found using the Bible Code methods can be matched to something, and then declared a "hit".
The only meaningful test, if there one, is to use the Bible Code to predict a reasonable number of future events—ones not likely to be predictable any other way (for instance, predicting that an elderly public figure will die within a large span of time would not be useful), and in significant detail, would provide some evidence in support of this method.
What does all of this mean?
None of this proves, beyond a shadow of a doubt, that the Bible Code methods and results are false. However, based on what one can see from a "skeptical approach" to the Bible Code, a true skeptic would likely place this topic in the "unlikely to be true until more and better evidence is provided" category. The same would apply, of course, to any similar methods which find "hidden messages" in everyday objects.
=============================
Interestingly, the Doonesbury cartoon from last Sunday dealt with the subject of skepticism. Enjoy.