Trees: Eden, Cross, New Jerusalem

A long text is written as one stream of letters with no spaces or vowel points. A reader picks a starting letter and a fixed skip, takes that letter and then every letter that many places on, and if the letters spell a word, the word is an equidistant letter sequence. Laid out in rows as wide as the skip, the sequence falls into a straight column, which is why the results are drawn as grids.

In any long text short words appear at some skip by chance, so a control text is required, such as the same search run on Moby Dick or on a Hebrew War and Peace. At a fixed skip the letters tav, vav, resh and he spell torah.

Doron Witztum, Eliyahu Rips and Yoav Rosenberg published their study in Statistical Science 9 (1994), pages 429 to 438. They reported that the names of 32 famous rabbis lie closer to the Hebrew dates of their birth or death, as letter sequences in Genesis, than chance allows. The figure usually quoted is 1 in 60,000. The editor, Robert Kass, offered the paper as a challenging puzzle, since his referees were baffled.

Brendan McKay, Dror Bar-Natan, Maya Bar-Hillel and Gil Kalai replied in Statistical Science 14, number 2 (1999), pages 150 to 173. They showed that the rabbis' names left wide room for choice and that the choices made leaned toward the result. With a small part of the same freedom they built a list of names that scored as well in a Hebrew translation of War and Peace. Every test they ran with data prepared by an independent expert found nothing above chance.

A committee formed at the Hebrew University in 1997 ran two fresh replications of a related experiment on the rabbis' cities, with data compiled by experts it chose. Yisrael Aumann, Hillel Furstenberg, Eliyahu Rips and Dror Bar-Natan sat on it. The experiments were completed in 2002, and neither found any sign of codes. A claim of this kind is weighed by its statistics.

Every fiftieth letter from the first tav of Genesis spells torah, and a like sequence stands near the start of Exodus. The observation was published after Weissmandl's death, with earlier forms claimed in Rabbeinu Bachya and in Chabad's The Purim Code.

Finding a four-letter word at some skip is trivial in any long text, as McKay and his colleagues (1999) showed, so the observation is real history and proves nothing. The sequence stands in the public domain Leningrad Codex as well.

Yacov Rambsel, Yeshua: The Hebrew Factor (1996), and Grant R. Jeffrey, The Signature of God (1996), claimed that Yeshua and related words cluster as letter sequences through Isaiah 53 and other messianic passages.

The four letters of Yeshua stand at some skip throughout any long Hebrew text, and contrary and blasphemous strings appear by the same method. David E. Thomas answered the claim in Skeptical Inquirer 21, number 6 (1997). Randall Ingermanson, a Christian physicist, tested it, found no signal and published the result in 1999 as Who Wrote the Bible Code?

Michael Drosnin claimed in The Bible Code (1997) that the Torah encoded the assassination of Yitzhak Rabin, reading the name across a phrase of Deuteronomy 4:42. Jeffrey H. Tigay, The Bible Codes: A Textual Perspective (1999), records that Drosnin read an atomic war after Rabin's death into the same grid and later settled for the bus bombings of 1996 as its fulfilment. McKay reports on his Torah Codes page that the leading codes researchers have disowned him.

Drosnin challenged his critics in Newsweek on 9 June 1997 to find such a message in Moby Dick. McKay answered with Assassinations Foretold in Moby Dick, grids of the same kind for Indira Gandhi, Leon Trotsky, Martin Luther King, John F. Kennedy, Abraham Lincoln and Yitzhak Rabin, and on further pages for Princess Diana and for Drosnin himself.

Chuck Missler, Cosmic Codes (1999), read the meanings of the names from Adam to Noah in order as a sentence of redemption. It is an interpretive reading, so it is weighed by the tests of an echo in Richard B. Hays, Echoes of Scripture in the Letters of Paul (1989), which do not apply to a statistical claim. Three of those tests are availability, volume and the history of interpretation. The names are available, the volume is low because one meaning must be chosen for each disputed name, and the history of the reading is modern, from the nineteenth and twentieth centuries.

The disputed meanings weaken it, since Methuselah is not securely his death shall bring, and Mahalalel and Jared are disputed in the standard Hebrew lexicons. It stands as a devotional observation and falls short of a proof of design.

The 1994 study used the Koren edition, published in Jerusalem in 1962. Jeffrey H. Tigay counts 304,805 letters in the Torah of the Koren edition and some 45 more in the Leningrad Codex, finished in 1009, and most such differences are spellings that leave the sense unchanged.

Yemenite Torah scrolls differ from Koren in the spelling of nine words. Tigay reckons that each Dead Sea manuscript of the Torah, if it were complete, would differ from the Masoretic text by hundreds of letters. One letter added or lost shifts every later position and breaks every sequence that spans it, and the Talmud grants in Kiddushin 30a that the sages were not expert in full and short spellings. Witztum settled the question of the text by asking Rabbi Shlomo Fisher, who answered that they could fully rely on their text. The reply of McKay and his colleagues (1999) puts the main fault in the loose definition of the data.

Koren's letter stream differs from the Leningrad text, so the 1994 searches cannot be repeated on the open text. Menachem Cohen of Bar-Ilan University set out the differences in Genesis between Koren and the great manuscripts in a letter published on Brendan McKay's Torah Codes pages. The Leningrad Codex is published at tanach.us as a Unicode text that may be copied without restriction.