Gematria

A genuinely ancient technique with a genuine foundation — and a combinatorial problem so severe that the same hidden codes turn up in War and Peace.

What it is

The interpretation of words by their numerical value, based on the fact that in Hebrew — and in Greek, and in Arabic — the letters were the numerals.

Before the adoption of separate digit symbols, these languages wrote numbers with letters. Hebrew aleph is 1, bet is 2, through yod at 10, then kaf 20, lamed 30, up to tav at 400. Greek did the same with alpha, beta, gamma. Arabic abjad numerals follow the same principle.

This matters, and it is what separates gematria from numerology. The values are not assigned by a mystic. They are how the culture wrote arithmetic. Any Hebrew or Greek word already had a number, unavoidably, before anyone thought to make anything of it.

Gematria is the claim that words with equal values are meaningfully related.

Where it came from

Greek practitioners called it isopsephy, and it was not confined to mystics. A piece of graffiti at Pompeii reads, in effect, "I love her whose number is 545" — a lover encoding a name in a public wall. The technique was common enough to be a game.

In Judaism, gematria appears as an exegetical method in rabbinic literature from the early centuries CE, and is listed among the interpretive rules attributed to Rabbi Eliezer. It sits alongside two sister techniques: notarikon, treating a word as an acronym, and temurah, systematic letter substitution.

It became far more prominent in Kabbalah, where the letters themselves are creative agents, and where the numerical relation between words is evidence of a relation in the structure of reality.

The most compelling single application is a scholarly one. The number 666 in Revelation is introduced with the instruction to calculate it, which is an explicit invitation to gematria. The Hebrew letters of Neron Kesar — Nero Caesar — sum to 666. And the Latin form Nero Caesar, one letter shorter, sums to 616 — which is precisely the variant reading preserved in some early manuscripts. Two different spellings of the same emperor's name producing exactly the two numbers attested in the textual tradition is about as strong as this kind of argument gets, and it is widely accepted by New Testament scholars.

How it is practised

A word or phrase is summed and compared with others of the same value, and the equivalence is read as significant. Several counting methods exist — standard values, ordinal position, reduced values, and schemes assigning the final forms of certain letters higher values — and practitioners may add one for the word as a whole, a convention called the kolel.

What we can and cannot say

We can say the technique is ancient, internal to the traditions that use it, and legitimate as exegesis. Within a religious framework that holds scripture to be divinely composed, exploring numerical relations is a coherent activity, and it has produced two millennia of genuine interpretive literature. It is not the same activity as claiming to have found a code.

We can say the combinatorial problem is severe, and it has been tested rather than merely asserted.

Consider the freedoms available: multiple counting methods, optional kolel, alternative spellings, the option to compare a word with a phrase, and no constraint on which pairs you may look for. The number of possible connections grows so fast that finding matches is guaranteed. The question is never whether a match exists but whether its existence is surprising, and with this much freedom it never is.

The Bible Code episode settled this experimentally. In 1994 Statistical Science published a paper by Witztum, Rips and Rosenberg claiming that equidistant letter sequences in Genesis encoded the names and dates of medieval rabbis at rates far beyond chance. It was presented as rigorous and reached a mass audience through Michael Drosnin's The Bible Code (1997).

In 1999 the same journal published a response by Brendan McKay, Dror Bar-Natan, Maya Bar-Hillel and Gil Kalai. They showed that the original result depended on choices made in preparing the data — which spellings of rabbis' names and which date formats were used — and demonstrated the point decisively by finding equivalent "codes" in the Hebrew translation of War and Peace. Drosnin had publicly challenged critics to find the assassination of a prime minister predicted in Moby Dick; McKay found several.

That is the cleanest available demonstration of the whole problem: a method that finds meaning in any sufficiently long text will find it in Tolstoy.

We cannot say the traditional practice is refuted by this, because the traditional practice was never making the Bible Code's claim. A rabbi noting that two words share a value is drawing an interpretive connection within a text he already regards as designed. Drosnin was claiming a falsifiable statistical result about hidden information, and it failed.

What it is

Where it came from

How it is practised

What we can and cannot say

Further reading