Chapter 21: The Masoretic Text: How Faithfully Has the Torah Been Copied?, section 2 of 3 Overview of chapter 21 Book contents Previous Next

How Great Are the Differences?

Differences between copies of the Torah do exist. Even where they are found, however, they are very minor. The typical case is an extra or missing vav or yod. These are the “helper letters,” which in Hebrew spelling serve to mark a vowel. They are not letters that carry the meaning of the word. The same word can be spelled with or without them, much as English writers might spell a word with or without a silent letter, and it remains the same word.

The best-known comparison is between the accepted text and the text of the Yemenite Torah scrolls. The Yemenite community lived for many centuries in relative isolation and kept its own scribal tradition. The two texts differ in nine letters in the entire Torah. In all but two of those places, the difference is a helper letter, present in one text and absent in the other.

Of the two remaining cases, one is the end of a word written with a final alef in one tradition and a final he in the other. At the end of a Hebrew word both letters are silent and both serve to mark the same vowel. The word is pronounced the same and means the same.

Only one of the nine affects the meaning, and even there only technically. In the verse that gives the age of Noah at his death (Gen 9:29), Ashkenazi manuscripts begin the verse with the word va-yehi, while the Yemenite manuscripts read va-yihyu. The extra final vav turns a singular verb (“and [all the days of Noah] was”) into a plural one (“and [all the days of Noah] were”). Strictly speaking, these are two different words. Yet both are correct Hebrew for the same idea, and both are used for it elsewhere. In the genealogies of Genesis 5, the same formula, “and all the days of [a man] were,” appears sometimes with the singular verb and sometimes with the plural. Whichever reading is original, the verse says the same thing.

Nine letters out of more than three hundred thousand, and not one of them changing what the Torah says, is a remarkable degree of agreement between two communities separated for many centuries.

A second line of evidence comes from the work of Rabbi Mordechai Breuer in the twentieth century. Breuer set out to test the accuracy of the traditional text against the oldest and best manuscripts, above all the Aleppo Codex and the Leningrad Codex, and against the Masoretes’ own notes. Those notes record, for example, how many times a word is spelled with a helper letter and how many times without. They therefore act as an independent check on the text they accompany. Breuer found that the notes and the best manuscripts agree with one another to a very high degree. According to Breuer, the Masoretic notes can be reconciled to point to a single text, and the text they point to is, with very few exceptions, the text in our scrolls.

Scholars may disagree about how far back this agreement reaches. Academic scholars of the biblical text use other witnesses as well, such as ancient translations, the Samaritan version of the Torah and the scrolls found at Qumran, and some prefer readings found in them. In traditionally religious Jewish communities, however, the Masoretic text remains the only authentic text of the Torah. It is the text every kosher scroll must match, and the text on which code research has almost always been carried out.

For the purposes of codes, this leaves a mixed picture. The differences between reliable copies of the Torah are few and slight in meaning, and the code researchers are right to work on the text that the tradition treats as authoritative. But a code does not care about meaning. A missing vav changes nothing in the sense of a verse, yet it shifts every letter after it by one. A code that crosses such a place exists in one text and not in the other. However faithful the tradition, any code must be stated together with the text in which it was found.

The Text Used in This Book

The text on which this book relies, for the codes of the previous chapter and for every count of words and letters in every chapter, is the traditional Masoretic text of the Torah, the text that a kosher scroll must match. The electronic edition of the Leningrad Codex most widely used in scholarship, the Westminster Leningrad Codex, differs from it in about 35 letters in the Torah. The codes and the counts in this book have all been checked against that edition as well. No difference between the two texts has affected any of them. Where a count did not match, the error turned out to be in the counting, not in the text.

Last revised 13 November 2026