CodeTrawler Scientific Notes


scideas@iol.ie

Introduction
Is it just chance ?
Designing an Experiment

Introduction
There have been many serious claims to have found messages encoded into various texts, often by distinguished academics. One such recent claim was made by a mathematician called Eliahu Rips and concerned the dates of birth/death of a list of rabbis encoded into the Torah (the Jewish bible, first five books of the Old Testament).

You can read the original paper here and a scientific refutation here.

Following Rips' work a journalist claimed to have discovered the message assassin that will assassinate near to the name Yitzak Rabin in the Old Testament. These examples serve to illustrate a few points about the scientific use of skip codes.

Is it just chance ?

This is the nub of the whole thing. It has been suggested for instance, that the chance of Rips' results occurring randomly were in the region of 50% - 60 %; in other words they were indistinguishable from chance occurrences. Of course this does not mean that an intended code cannot exist - just that we know whether it was intended or not. In order to reduce the possibility of a chance result an experiment must be properly constructed - something Rips, surprisingly, did not do.
A mathematical treatment can be found here.

Designing an Experiment

Use a control
Test data that you expect to show a negative result to verify that it does. For example, if you are trying to find evidence of predictions for the future in an old text, you should make the same test on texts written after the events occurred. If you are claiming that Tolstoy encoded a message in one of his novels, you must show that the same message cannot be found in books not written by him.

Repeat your experiment
Results of one experiment on one set of data (should) carry no weight. If you reckon on finding messages in the Torah, you must also reckon, logically, on finding them in the rest of the Old Testament and in the New Testament.

Define Acceptable Positive Results
The phrase assassin that will assassinate has two obvious problems. Firstly it is not a correct English phrase; that should be who. Secondly, it is ambiguous; it could refer to Rabin as the assassinated or as the assassin. The decision about whether to accept only correct grammar, what to do about ambiguities and any other such decisions must be declared before the experiment is conducted.

Make your data un-ambiguous
One of the problems with analysing text is the difficulty with defining words. For instance, prior to the late seventeenth century, western scripts had no consistent spelling; you will often see the same word spelt differently by the same person in the same document. In the case of the Torah, written in Hebrew, not only is the language itself notoriously diverse in the spelling of names and dates but there is the additional problem of translating results into another language. Many controversies relating to the Bible have been resolved by more accurate translation.

Are you sure that you are spelling a name the way it was spelt at the time the text was written ? Have you tried all the alternative spellings ? Have you tried all the person's alternative names ?

back to top