How was the Old Testament put together?

Jews call the Hebrew Bible the TaNaKh, a name built from its three divisions: Torah (Instruction, or Law), Nevi'im (Prophets), and Ketuvim (Writings). Counted by the scrolls they were traditionally written on, these come to 24 books (Britannica) ↗. The parts did not close together. The majority view in reference scholarship is that the Torah most likely reached its final form during the Babylonian exile, in the 6th to 5th century BCE; the Prophets closed by the late 4th century BCE; and the Writings stayed open longest, lacking even a fixed name. Britannica states the limit of this reconstruction in its own words: “no more than a plausible reconstruction of the successive stages involved can be provided.”

The threefold grouping itself is ancient. Around 132 BCE, the grandson of Ben Sira translated his grandfather's wisdom book into Greek and wrote in his prologue of “the law and the prophets and the others that followed them” (quoted by Britannica). Three parts, already visible, the third still without a settled name.

The rabbis' own debates survive in the Mishnah, the first great written collection of rabbinic law. Its phrase “defiles the hands” is its technical way of marking a book as sacred scripture, and two books were still being argued over:

Rabbi Judah and Rabbi Yose each accept one of the two books and report a dispute over the other. Rabbi Akiba insists no one in Israel ever disputed the Song of Songs; the dispute, he says, was only over Ecclesiastes. A canon still being argued about is a canon not yet closed. Esther's debate is recorded in the Talmud:

Not rendering the hands ritually impure is the same technical vocabulary: Shmuel's recorded position is that Esther was not accorded the sanctity of the sacred scrolls. The Gemara itself answers him on the same page: Esther was stated with the Divine Spirit to be read in public, though not to be written. What survives either reading is that the book's standing was still being argued. Britannica adds a quiet corroboration: no fragment of Esther has turned up among the biblical scrolls from Qumran.

What about a council that settled all of this? The idea that a council at Yavneh (Jamnia in older spelling) closed the Jewish canon around 70 to 90 CE was proposed by Heinrich Graetz in 1871. It has not held. Jack P. Lewis examined the evidence twice, in 1964 and again in the Anchor Bible Dictionary in 1992, and found the hypothesis resting on thin evidence. Albert C. Sundberg Jr. found canonical discussion at Yavneh attested only for specific books. F. F. Bruce judged it unwise to speak of a council that laid down the limits of the canon. Britannica keeps only a softened version: a synod at Jabneh “seems to have ruled on the matter,” with its decisions taking a generation or two to win general acceptance. The record attests discussions of specific books; it stops short of a council that closed the canon.

For the whole collection, Britannica's dating is plain: “The Hebrew Bible probably reached its current form about the 2nd century CE.” That is roughly a century after Jesus. The count of 24 books appears first in a work called 2 Esdras, in a passage usually dated about 100 CE (Britannica), inside a legend this page returns to below.

And the list still differs by community. The Jewish canon counts 24 books, most Protestant Old Testaments 39, and the Catholic Old Testament 46 (Britannica). The difference is a set of books Protestants call the Apocrypha and Catholics call deuterocanonical. Jerome, the translator of the church's standard Latin Bible, doubted some of them and was overruled; most were included in the Vulgate. On April 8, 1546, the Council of Trent declared nearly the entire Vulgate canonical for Catholics. Luther kept the extra books in his German Bible but set them apart in a separate section, as books that deserved to be read but did not stand on the same level as scripture. The Orthodox Old Testament is larger still: the Orthodox Church in America describes its additional books as the “longer canon,” “accepted by Orthodox Christianity as canonical scripture” (Orthodox Church in America ↗), and printed Orthodox lists vary. The Jewish 24 and the Protestant 39 are largely the same books, grouped and counted differently; what separates the communities is which added books each receives. Four communities, and no single list shared by all. No council, in any century, settled the difference for everyone.

A canon fixes which books belong. The words inside them are a separate question, and the caves near the Dead Sea show what those words looked like before one text became standard. Britannica ↗ states the starting point directly: there is much evidence for a period when more than one Hebrew text of a given book was current, and the older the surviving copies are, the more they differ from one another.

The main witnesses are the Dead Sea Scrolls. Among the roughly 230 biblical manuscripts found at Qumran are “partial or complete copies of every book in the Hebrew Bible (except the book of Esther),” dating from the 3rd century BCE to the 1st century CE (Israel Antiquities Authority ↗, the scrolls' official custodian). Emanuel Tov, the scrolls' longtime chief editor, distinguishes several Hebrew text traditions among them. The custodian institution's own sentence about the copyists deserves to be read whole: “Additions and deletions in certain texts imply that the writers felt free to modify texts they were copying.”

The book of Jeremiah shows what that freedom could produce. The Septuagint, the ancient Greek translation of the Hebrew Bible, presents a Jeremiah about one-seventh shorter than the received Hebrew text (the standard text known as the Masoretic Text), and it arranges the material differently. Tov argues that the Greek translator worked relatively literally where the two texts overlap, and judges it unlikely that the translator shortened his Hebrew source on his own: the short Greek points to a short Hebrew. Most scholars hold that the short edition is the earlier one. And Qumran has produced a short Hebrew text of Jeremiah that probably underlies the Greek (Britannica). Two editions of a prophetic book, circulating at once.

The Samaritan community kept its own Torah, the Samaritan Pentateuch; Pentateuch is another name for the Torah's five books. It differs from the Masoretic Text in about 6,000 readings, nearly a third of which agree with the Septuagint; Britannica attaches the proportion in the same breath: “Only a minority, however, are genuine variants, most being dogmatic, exegetical, grammatical, or merely orthographic in character.” In plain terms: Britannica classes most of the 6,000 as differences of doctrine, interpretation, grammar, or spelling, and only a minority as genuine variants.

The largest systematic difference sits in Genesis 5 and 11, the chapters of lifespans and begettings. The three traditions transmit three different editions of those chronologies (the standard scholarly treatment is Ronald Hendel's Oxford study of the text of Genesis 1-11). The sources establish the divergence; they do not establish which community changed its numbers, and this page does not say. What they do establish is the scale: over a thousand years of world chronology rides on which manuscript tradition you open.

The other half of the record belongs to a later period. The text type behind the Masoretic Text is already well represented at Qumran before the Christian era. Scrolls from other desert sites, dated to the 2nd century CE, are “practically identical with the received text” (Britannica). Of the centuries after, Britannica says the single standardized text was “transmitted by the scribes with amazing fidelity”: not one medieval Hebrew manuscript, and none of the thousands of fragments from the Cairo Geniza, a storeroom of worn-out Jewish manuscripts found in Cairo, departs from it in any real significance. The Antiquities Authority calls the similarity between the Qumran scrolls and the medieval manuscripts remarkable, given that the scrolls are over a thousand years older. Both facts are the tradition's own record, and they sit at different points in time: many texts before, one steady text after.

Even the standardized text arrived in layers. A printed Hebrew Bible carries three things: consonants, vowel signs, and musical marks. The history of the bare consonants, in Britannica's phrase, “stretches back into hoary antiquity.” The vowel signs and accents were added by the Masoretes, the scribal scholars who fixed the reading of the text, working roughly from the 6th to the 10th century CE; the system developed in the city of Tiberias prevailed over the others.

Their care is part of the record too. The Masoretes “counted and noted down the total number of verses, words, and letters in the text” (Britannica), so that any future change could be detected. Britannica credits the consistency of the Hebrew text since their work to exactly that rigor.

Here is what the layers mean. The vowels, and with them many readings that settle what a word means, were fixed centuries after Hebrew had ceased to be anyone's mother tongue. Britannica dates the beginning of Aramaic replacing Hebrew in everyday speech to about the 3rd century BCE; Hebrew continued as a language of worship and study, and was revived as a spoken language only in the 19th and 20th centuries. When the Masoretes pointed the text, that is, wrote in the vowel signs, they were writing down a received reading tradition, choosing among readings the consonants alone leave open. Britannica's own example shows what was left open: the consonant string DBR can be read as dibber, “he spoke,” or as devar, “the word of.” One such choice can move the meaning of a verse. The pointing fixed one reading for every verse in the Bible.

Those words, the priestly blessing, survive in the earliest known copy of a passage also found in the Hebrew Bible. Two small rolled silver amulets from a burial cave at Ketef Hinnom in Jerusalem, excavated in 1979 by Gabriel Barkay and dated to about 600 BCE, carry an early form of the blessing along with the divine name. They are amulets, small inscribed charms worn for blessing; the earliest Bible manuscripts proper come centuries later.

For a complete Hebrew Bible, the wait is far longer. The Aleppo Codex, corrected and pointed by the Masorete Aaron ben Asher about 930 CE, is now incomplete: of what appear to have been 491 pages, 295 survive, and the lost pages include the Torah up to Deuteronomy 28:17 (the manuscript's official project site ↗). The Leningrad Codex, whose own closing note dates it to 1008 or 1009 CE, is the oldest complete manuscript of the Hebrew Bible, and the basis of the standard scholarly edition used today. Between the latest biblical books and the first complete surviving copy of the whole Hebrew Bible stand more than a thousand years. The Dead Sea Scrolls close part of that gap for portions of the text, and the amulets carry one blessing older still.

The tradition also discusses a famous list of reverent rewordings. Rabbinic literature calls certain passages tiqqune soferim, corrections of the scribes, classically counted at eighteen and tied to reverence for God; what that label records is itself argued inside the tradition, as the next passage shows. On Genesis 18:22, where the received text says Abraham stood before the Lord:

The Sefaria translation prints Rabbi Simon's label as “a scribal emendation,” and that label, tikkun soferim, is exactly what the tradition argues over. Ehud Ben Zvi's note in The Jewish Study Bible, on another of the eighteen, reads such comments as recording an original text changed by the scribes to avoid describing God too humanly, and the medieval commentator Rashi, following Rabbi Simon on this verse, wrote that the rabbis reversed the phrase. An early rabbinic commentary, the Sifrei, reads the label the other way: the original authors themselves chose the softer wording, and no scribe changed a letter. Marty Lockshin, whose treatment this paragraph follows (TheTorah.com ↗), notes that the majority of classical rabbinic passages on these corrections present them that second way. So whether these passages record actual alteration or original reverence is disputed within the tradition, and the community kept the whole argument in its own commentaries.

The traditional Jewish and Christian position is that Moses wrote the five books of the Torah, the books traditionally ascribed to him as the recipient of the revelation at Sinai (Britannica). That is the confessionally affirmed position, held by Orthodox Jews and many Christians today. The Talmud's own discussion is more layered than a slogan:

Moses wrote his book, the Talmud says, and Joshua wrote the last eight verses of the Torah, the verses that describe Moses' death. Rabbi Shimon, as the discussion continues onto the next page, holds that Moses wrote even those: God dictated, and Moses wrote in tears. The tradition itself, in its own central text, records at least one passage of the Torah as possibly not from Moses' hand. The same Talmud records a second layered fact, about the alphabet itself:

On this account, the Torah was first given in one script and given again in the days of Ezra in the script used in Torah scrolls today. That is one opinion among several: the same discussion records Rabbi Yehuda HaNasi's opinion that the Torah was given in today's script from the beginning, and Rabbi Elazar HaModa'i's view that the script never changed at all. What the record shows either way is that the tradition's central text preserves a live debate about the very alphabet of its scrolls.

Around 100 CE, a Jewish writer told the story of a deeper rupture. In the book called 2 Esdras, the scribe Ezra prays:

In this legend the Law has been burned, and Ezra re-dictates ninety-four books in forty days: twenty-four to be published openly, seventy to be kept for the wise. It is a legend, told in a book Britannica classes as apocryphal, one the Vulgate prints only as an appendix, and it is also the first place the 24-book count of the Hebrew Bible appears (Britannica). Its value here is what it shows: at least one ancient Jewish author imagined the transmission of his scripture as a story of loss and restoration.

The biblical text also points to books it used and we no longer have:

The book of Jashar is cited again in 2 Samuel 1:18. These are lost sources that the text quotes; nothing in the text calls Jashar inspired scripture. But a text that cites its sources is a text that had sources.

Critical scholarship draws the conclusion the majority of its practitioners hold: the Pentateuch is composite, edited together from several sources over centuries. The classical version of that view named four source documents, dated between perhaps the 10th and the 5th century BCE. What happened to that classical version is described by Gordon Wenham ↗, an evangelical Pentateuch scholar surveying his own field in 1996: by the mid-20th century the four-document theory “was almost universally accepted”; from the 1970s that consensus was disturbed, on dating, on history, and on the principles of dividing the sources; and “no new paradigm or scholarly consensus has emerged to displace the old theories.” So the positions stand thus: composite authorship over centuries remains the majority view; the classical four-document map, on Wenham's account, lost its consensus with nothing yet replacing it; and the traditional position remains Mosaic authorship.

Even where the Hebrew consonants are not in dispute, moving them into another language involves choices. Two famous verses show how much a single choice can carry. Isaiah 7:14 is the classic case. Here is the verse in a Jewish translation:

The Hebrew word rendered “young woman” is almah. The translators' note on this verse in the NET Bible ↗, an evangelical translation, states the linguistic core: the word “does not carry this meaning inherently,” and “The word seems to pertain to age, not sexual experience.” The Septuagint rendered it with “the more specific Greek word παρθένος (parthenos),” the Greek word for virgin. Matthew's Gospel quotes the verse in its Greek form:

The Christian defense stands at full strength in the same evangelical note. The Greek rendering predates Christianity: Jewish translators made that choice long before there were Christians. And whatever the Hebrew word means in Isaiah's context, “Matthew's usage of the Greek term παρθένος clearly indicates that from his perspective a virgin birth has taken place.” The note adds that one's view of the doctrine is ultimately unaffected by the translation question. What this page marks is the category: this is a divergence of translation and interpretation, not a Hebrew textual variant. The consonants are not in dispute here.

Psalm 22:16 is different in kind, because there the Hebrew manuscripts themselves diverge:

The World English Bible quoted above attaches its own footnote to that last line, printed here exactly as it stands: So Dead Sea Scrolls. Masoretic Text reads, “Like a lion, they pin my hands and feet.” The two readings differ by one Hebrew letter. The JPS translation renders the line “like lions [they maul] my hands and feet,” following the medieval commentator Rashi; the ancient Greek and Latin versions read a verb, they have pierced. A psalms fragment from Naḥal Ḥever, from the Dead Sea era, is read by its editors (Abegg, Flint, and Ulrich) as a verb form meaning pierced or dug. Jewish and Christian translations still divide on the verse today, and a verse Christians read as a crucifixion prophecy hangs partly on that one letter.

This section stands apart from the history above. It reports what the Quran states about earlier scripture, offered as the Quran's own claim; the manuscript record above stands on its own evidence. The Quran affirms the Torah as real revelation:

The Torah held guidance and light. The prophets judged by it, and the rabbis and scholars were entrusted with keeping God's Scripture and stood witnesses over it. The Quran also condemns a specific act:

Writing scripture by hand and attributing it to God; words moved from their places; a portion of the reminder forgotten. And the structural contrast is the Quran's own. In 5:44, the earlier scripture was entrusted to its human custodians. Of the Quran, God says:

There, the guardian is God Himself. How the Quran itself was compiled and transmitted, set beside the Bible's history and judged by the same standard of evidence, is its own study: how the Quran and the Bible reached us.