ITCS Summary

[With Hebrew]

by Eliyahu Rips

The First Congress of the

International Torah Code Society

May 10-11th, 1999
Hilton Hotel
Jerusalem International Convention Center
Jerusalem, Israel

Following are my summary and impressions from the 1st Congrss of the ITCS held on May 10 and May 11 in Jerusalem.

The two days of our meetings were preceded by several months of intensive preparations. Three of us were particularly closely involved in these preparations - Professor Robert Haralick, Dr. Moshe Katz and myself.

Professor Haralick formulated the general outline of the goals of ITCS, as well as the practical documents such as the announcement of the Congress, the call for papers etc. He was singularly good at this due to his great power of expression and his experience as a leading scientist. No less important, every statement issued by him bears the seal of his great personal integrity. I am sure that such a seal is unmistakenly detected by anybody who has appreciation for integrity.

The very hard burden of organizational tasks was carried on by Dr. Moshe Katz. With endless patience and iron nerves he undertook numerous negotiations to make this Congress possible. Most of us would certainly give up, but he persevered. He explained, he convinced. More than once he was facing the fantastic unreliability of the bureaucracy. Yet Moshe persisted, and the fact that the Congress successfully took place is entirely due to his unflagging efforts.

On my side, I was able to contribute my very good familiarity with the work of almost all prospective participants. Roughly, the ongoing experimental research can be divided into two categories. One category is the completion and improvement of the past work, and the other is the exploration of new areas, glimpses into the unknown. We should seek the correct balance between the two domains. Our program was composed having this concern in mind. On the theoretical side, the dominant place justly belongs to the works of Professor Haralick who brought into the Code research the professionalism, competent knowledge and brilliant penetration of a leading scientist.

The morning session of the Congress started with the work of Professor Haralick entitled "ELS Combination Probabilities". Let me quote the abstract:

In other words, this work provides tools for the theoretical computation of probabilities for the code pictures, using the size as the basic parameter. It is especially important for tables with many ELS's, because the existing Monte Carlo methods allow only treating of the pair meetings. Introducing the new quantitative tool will undoubtedly broaden the scope of the experimental research, which hardly could be comducted in the areas where we are lacking quantitative estimates.

The second talk by Professor Haralick was given in the next morning session, on May 11. It was devoted to the Best Star Team methodology. Given a sample consisting of many individual experiments, one needs to assemble the outcomes of these experiments into a single number. Different statistics will accomplish the task with varying performance. The BST methodology is precisely sensitive to the kind of the data which is expected in the ELS experiments. Its intuitive behavior: it rewards strong successes and it is relatively tolerant to a moderate number of wrong guesses.

My own talk discussed application of BST in combination with the proximity measures used by WRR. In this way we obtain a measuring scheme with a wide range of applicability, flexible and simplicity of use.

The lecture of Harold Gans was a great event. Together with Nachum Bombach, he spent immense efforts redoing the Communities experiment. His goal was to verify with full confidence every aspect. Through numerous talks with competent Rabbinical personalities he again and again verified the Inbal rules, until he became confident that they were not artificially concocted, but describe the true usage of the Communities names. Then with Nachum Bombach he applied these rules from scratch, carefully documenting every detail. The success of the resultig sample serves as a new verification of the Torah Codes claims. The challenge that Gans poses before the critics is very powerful indeed: to show how reasonably manipulating the rules, and the rules only, one can construct a sample with this kind of success in War and Peace. After all, if they manage to do it, this will be the most effective way to impair the validity of Gans' results. So far, they seem not to be very enthusiastic to undertake the challenge.

He also discussed the particular shape of the histogram obtained for the Communities sample, which is very much like the shape of the histograms in WRR. He hypothesized on possible explanations, expressing a number of very interesting observations and insights.

Dr. Alex Rotenberg presented his beautiful "Bnei Haman" sample. It is based on the ELS expression YG B)DR PWRYM, which is remarkable in itself, it has a very simple logical structure, it uses only data from the Book of Esther, and the composition of the sample uses only ideas already introduced in the previous research. The computations of Art Levitt show that this sample has a very strong success.

Dr. Moshe Katz presented two samples. One is about the death of Diana. It is important to stress that he was specifically challenged to work on this topic. I find the composition of his sample very logical and uniform. He started with a single newspaper report about the event. The transliteration of proper foreign names is consistently done as in this newspaper report. The choices are few, natural and explicit. The results: out of 35000 randomized texts, the Torah text ranked 1, 0, and 2 for the additive, multiplicative and minimax versions of BST, respectively.

The second sample is very simple, basic and straightforward: the word MY$NYWT is coupled with the names of six parts of Mishnah: ZR(YM, MW(D, N$YM, NZYQYN, QD$YM, +HRWT, taken with |d|=1. For 1000 randomized texts, the results in Genesis and Numbers are 1:120 and 1:80 respectively, while in the other Chumashim there is no significance. This is a moderate, but nevertheless important success.

My own work "The Tribes of Israel" was presented on the second day of the Congress. The draft of the paper is electronically available here. The advantage of this sample is that it is completely explicit. All data is already given in Midrash Tadshe. However, we all know that no efforts will be spared by the critics to find any possible weak point. Therefore everything in this work should be scrutinized extremely carefully. And indeed, the audience (which included a number of brilliant minds) received every detail very critically. My impression is that the sample somehow survived the "friendly fire".

Art Levitt presented 11 very well defined samples. The best of them has significance 0.00075. He also presented a program for future work. This paper is presented here.

A talk by Doron Witztum was also scheduled, but did not take place. The subject of his talk: the recent work on the "patronymic sample".This is a very important work, because of the straightforward composition of the sample and because of being directly relevant to the controversy around the Rabbinic appellations sample. This work is detailed here.

In the exploratory part the speakers were Nachum Bombach, Dr. Moshe Katz, Dr. Alex Rotenberg and Dr. Leib Schwartzman.

Nachum Bombach discussed a single pair of ELS expressions )LHYMDYN and RXMYHWH with extremely close meeting having the significance level of about 1:50000. The further detection and exploration of such "heavyweight" pairs is an extremely interesting problem.

Dr. Alex Rotenberg defined certain quantitative tools for dealing with long expressions. It is a well known difficulty to estimate the significance of long expressions, because no one knows how to define the relevant probability space. Instead, Dr. Rotenberg introduced two "non-triviality measures". The first one, denoted Ra is defined as -log of the probability of the appearance of the given expression in the given text at least once. The second one, denoted Rb is defined as -log of the probability of the appearance of the given expression in the given text with skips less than or equal to the skip with which it actually appears. If the difference Rb-Ra is large, it gives an indication that the ELS expression appeared with "surprisingly small" skip. It should be stressed that both quantities Ra and Rb are scales rather than (-log of) significance levels. So, these scales Ra and Rb provide the necessary adjustments for the length of the text and for the appearance with short skips. Because of the elementary observation that for a given length there are more expressions formed from short words than from long words, a numerical characteristic called DC (=Difficulty Class) is introduced. It is heuristically defined as L-3N, where L is the length and N is the number of words in the expression. Now, within a given Difficulty Class, we are interested in finding ELS expressions with values of Ra and Rb as large as possible. A list of examples found by different researchers was provided.

Dr. Rotenberg's talk opened a new perspective and interesting field. Maybe, progress in this field will be achieved when we better know how to use the power of computational linguistics.

Now I came to the talk given by Dr. Leib Schwartzman. I have no doubt that Dr. Schwartzman is one of the most deep (maybe the deepest) researcher of the Torah Codes. His talk concentrated on two new phenomena discovered by him - the Dialog Mode and Multiclustering. The extremely beautiful example "MY YRM" - who will raise high - gives a certain idea of both effects.

1. The book of Numbers contains the passage WYRM M)GG MLKW. The straightforward application of the Dialog Mode transforms it into a question and answer: MYYRM -who will raise high - and MLKYRM- king will raise high. Now, it turns out that the minimal ELS's for this pair have a close meeting in Numbers with significance level of 1:200.

2. Rashi Commentary to the same verse mentions the first king of Israel who fought the Amalekites, namely Saul. The pair MYYRM $)WLYRM in Deuteronomy appears with minimal skips in the passage speaking about kings. Moreover, both minimal ELS's meet the passage $WM T$YM (LHTK MLK mentioned in Sanhedrin with the respect to the Laws of kings. The significance of the meetings of both ELS's with this passage is very high. (On 100000 texts, the rank for MYYRM with $WMT$YM(LKMKL is 4 and the rank for $)WLYRM with $WMT$Y(LKMLK is 2, though the measurements I did for these two pairs are not completely satisfactory.)

3. In Leviticus, the minimal ELS for MYYRM appears at the end of the weekly portion Behar (where the redemption of land etc is considered). Here MYYRM has the extention G)L the redeemer - MYYRMG)L. The verse from Isaiah: "HNH Y$KYL (BDY YRWM WN$) WGBH M)D" gives the Dialog Mode forms MYN$) and MYGBH. While MYN$) does not appear here, the expression MYGBH (with minimal skips) appears very close to MYYRM; at the level of 1:2000. (In addition, remembering the verse WHYTH LH' HMLWKH also mentioning the redemption, we find that the minimal ELS for HMLWKH appears at the same place, with skip 26, which is the numerical value of the Holy Name.)

4. We observe that the same expression MYYRM appears in three clusters developing different, but related aspects of the same topic. This effect is called Multiclustering.

What was presented here is virtually the tip of the iceberg concerning the work of Dr. Schwartzman.

The closing section was for the general public. Professor Haralick spoke about the Torah Codes controversy. His presentation was powerful, beautiful and crystal clear. I cannot pretend to summarize it with my own words. I think it is recorded. I highly recommend listening to it as it is.