If the Septuagint and the Masoretic Text stem from different Hebrew parent-texts, how did these different Hebrew traditions develop? It might help to picture the development of the Bible by examining Jeremiah as a hypothetical test case. Jeremiah dictated his prophecies to the scribe Baruch who wrote them down (Jer 36:4-6; 45:1). Each of these oracles would have been recorded separately at the various times Jeremiah prophesied. In other words, Jeremiah did not sit down and write out the book that we hold in our hands today. The book of Jeremiah is a collection of shorter texts. Possibly, Israelite practice was similar to neo-Assyrian wherein oracles were first written individually on a single document before being collected together later on an additional document (for neo-Assyrians, cuneiform tablets; for Israelites, papyri or parchment). When Baruch took down Jeremiah’s words, he was likely working with a fresh document. Eventually various separate scrolls containing Jeremiah’s prophecies were collected together onto one scroll. At the time scribes collected these texts, they might also have added narrative or made editorial emendations.
Significantly, Jeremiah began to prophesy around 626 B.C.E. The destruction of Jerusalem took place in 587/86 B.C.E. At this time the Babylonians took many Judahites into exile. Others, including Jeremiah, fled to Egypt (Jer 43). What happened to the recorded oracles at that time? Were some destroyed? Were some taken to Babylon or Egypt? Did some remain behind in Judea? Perhaps, all of the above? How many copies of Jeremiah’s oracles had been made during that time? As the Judahites were dispersed, some of their scrolls were dispersed with them. In this case, different textual traditions could develop in various locations in diaspora. Repetitive copying over many years and distinct collation of oral tradition could easily produce unique textual forms. Similarly, dispersed Jewish communities might only have had access to particular scrolls. None of the various texts that make up the Hebrew Bible were bound together at this point.
There are various theories as to how different textual traditions developed. Prior to the Dead Sea scrolls discoveries these theories were based primarily on the Masoretic Text (MT), Septuagint (LXX), and Samaritan Pentateuch (SP). These known traditions provided some information on how divergent texts arose. The two most prominent theories were proffered by Frank Moore Cross (local text theory) and Shemaryahu Talmon (group text theory). Cross surmised that the MT, LXX, and SP developed from one shared original source (“Urtext”) in three different geographical locations that had large Jewish communities: Judea (SP), Egypt (LXX), and Babylon (MT).As a result of Jews being scattered into diaspora various textual traditions arose at distinct sites. Later, these three textual traditions came together in Judea and influenced each other. However, some scholars reject this view as too simplistic to explain the diverse textual traditions that seem to have developed in a single place, such as at Qumran. As Eugene Ulrich notes, Cross is probably correct that copying and transmission of scriptural texts in diaspora played a role in producing variants, but the evidence suggests greater plurality beyond three traditions.Talmon, on the other hand, does not see streams of tradition flowing out from one Urtext, but rather textual plurality that gradually became unified when specific traditions were selected over others.He believes three socio-religious groups chose texts to further their theological objectives: Samaritans—the SP, Jews—the MT, and Christians—the LXX. Similar theories were proffered specifically for the Septuagint as well. Paul de Lagarde believed all LXX manuscripts could be traced back to one original Urtext. On the other hand, Paul Kahle held to a plurality of texts that eventually narrowed to one text.
The discovery of the Dead Sea Scrolls changed the conversation beyond the three previously known traditions (MT, LXX, SP). The Dead Sea Scrolls provide remarkable insight into the copying and transmission history of the biblical texts. From 1946-1956, 800-900 texts were found in eleven caves at Qumran near the Dead Sea. These texts are dated to approximately the third century B.C.E. to the 1st century C.E. and include biblical and non-biblical religious texts, as well as historical documents such as Bar Kokhba’s letters. Among the 230 biblical texts, all the books of the Hebrew Bible are represented except for Esther (and possibly Nehemiah). The most popular books based on the highest number of manuscripts (keeping in mind there might be others that were lost) are Deuteronomy, Isaiah, and Psalms, with Psalms leading the way. These three are the most frequently cited by the New Testament authors as well.
The pioneering voices for new theories based on the Dead Sea Scrolls are Eugene Ulrich and Emmanuel Tov. Eugene Ulrich maintains that the biblical books cannot be traced back to one original text (Urtext). Rather the textual diversity at Qumran indicates different traditions developed side by side as a result of scribal revision. Older traditions were reformulated for new audiences. These recontextualizations retained the intent of the older version while including the scribe’s creative adaptation.Sometimes these new editions replaced the old, but at other times they did not. Sometimes there were two or more editions of the same biblical book in circulation. The redaction history came to an abrupt end as a result of the Roman persecution of the Jews and challenges from the new Christians on interpretation of sacred texts.
In contrast, Emmanuel Tov believes each biblical book reached a final composition.These completed books were then copied over and over again by scribes to replace tattered ones or to accommodate community needs. Tov’s work focuses primarily on classifying documents at Qumran. He believes there are more than three traditions for the biblical texts (MT, LXX, SP). Specifically, he notes the significant diversity among manuscripts at Qumran and concludes there are four primary streams represented:
1. Proto-Masoretic, AKA proto-rabbinic (52% of Torah texts; 44% of the Prophets and Writings)
2. Non-aligned texts (no coherent pattern that suggests connection with the other witnesses; 37% of the Torah texts; 53% of the Prophets and Writings)
3. Pre-Samaritan (6.5%)
4. Texts similar to the Hebrew Vorlage of the LXX (4.5% of Torah texts; 3% of the Prophets and Writings)
Others scholars have objected to Tov’s high percentage in the proto-Masoretic category. Some of the texts also equally meet the criteria of the pre-Samaritan group given the similarities. This results in a bloated percentage for proto-Masoretic texts. In response to this criticism, Tov has delineated between an “inner circle” and “outer circle” of Masoretic texts. He now places the more ambiguous texts in what he classifies as the “outer circle” of the proto-Masoretic category. Tov tends to favor the Masoretic tradition and sees textual diversity at Qumran primarily stemming from early official copies of biblical texts kept in the Jerusalem temple. In his view, this temple tradition was proto-Masoretic and the precursor to our modern-day Masoretic Text. However, Tov does acknowledge the possibility of Hebrew traditions that preceded the proto-Masoretic in some cases.
In contrast to Tov, Armin Lange proposes the following categories and percentages:
1. Non-aligned (52% of Torah texts; 51% of Prophets and Writings
2. Pre-Samaritan (5%)
3. Vorlage of the LXX (5% of Torah texts; 4% of Prophets and Writings)
4. Equally close to MT and SP (27%)
5. Semi-Masoretic (5% of Torah texts; 35% of Prophets and Writings)
6. Proto-Masoretic (5% of Torah texts; 10% of Prophets and Writings
Lange only includes in the proto-Masoretic category what constitutes Tov’s “inner circle.” As a result, his percentages suggest the proto-Masoretic tradition is not nearly as dominant at Qumran as Tov suggests. Lange’s numbers indicate non-aligned texts—those that did not fit any particular tradition’s pattern—were dominant until mid-first century C.E. He also analyzed dating of the texts and found that in general a large increase in scribal activity occurred during the Roman conquest of Judea (63 B.C.E) and reign of Herod the Great. Most of the proto-Masoretic texts at Qumran as well as those found at other sites such as Masada date to this time period. At the turn of the eras, the proto-Masoretic texts began to gain dominance as non-aligned texts decreased. After 68 C.E. only proto-Masoretic manuscripts are found. Lange concludes that the proto-Masoretic Text was created in the 2nd half of the 1st century B.C.E. by priests during the scribal boom. This became the standardized text.
According to Lange’s theory the standardization effort that resulted in the proto-Masoretic Text was built upon majority readings. The priests selected some books “as-is” from the textual traditions available. But other texts, such as Jeremiah, they edited. Thus the proto-Masoretic Text was not one “original” book. Different biblical books had various scribal histories and were given the “finishing touch” by the priests. As a result of this standardization of the proto-Masoretic Text, we see evidence already during the same time period to bring the Septuagint translations into closer alignment with the proto-Masoretic version (e.g. Greek Twelve Minor Prophets scroll, 8HEVXII).
Lange proposes that the LXX and proto-Masoretic Text have the same Hebrew Vorlage for Jeremiah. The LXX remained closer to the Vorlage, while the proto-Masoretic employed more editing. He also believes the proto-Masoretic Jeremiah finds its seeds in the Egyptian Jewish community. He concludes this based on the allusions to: 1) the sack of Gaza by Ptolemy I Soter; 2) the proto-Masoretic rebuke of the Judean Jewry; and 3) proto-Masoretic universalization of diaspora beyond Babylon. On the other hand he believes the apocryphal Letter of Jeremiah was written in Babylon and not Egypt.
Other scholars have built upon or modified these proposals. Sidnie Crawford agrees with Tov that each book reached a “recognizable shape” before being copied. However, the text within that shape was not fixed. The two editions of Exodus (proto-MT and proto-Samaritan) are both recognizable as the book of Exodus, but there are also distinctions between them. Like Tov and Ulrich, Crawford believes the disparity is the result of two forms of copying—scribes who were conservative and precise in their copying, and those who were “creative” or “free.” However, Crawford objects to Tov’s privileging of the Masoretic Text as the “original” text. In some cases the LXX preserves the more conservative tradition (Jeremiah), while at other times the LXX contains a revisionist edition (Esther). The opposite is also true. Sometimes the Masoretic Text has the more conservative text (Esther) or a revisionist one (Jeremiah). However, ultimately the Masoretic texts, comprising books of different scribal histories, were accepted as the standard.
In turn, Arie van der Kooij challenges Crawford’s view by suggesting the Temple preserved texts much more carefully than copies made outside the Temple, and this tradition was the Masoretic one. Thus textual fluidity is not as pronounced for all scriptural editions as some claim. He believes the freer copies outside the Temple were for reading and studying purposes as opposed to preservation. As for textual disparity between the LXX and MT in Jeremiah he suggests the possibility that the LXX is not more original but merely reflects a literary stage prior to the book’s final redaction.
And finally, to top off the discussion, David Carr suggests too much has been made of the textual diversity.Instead of 10-20 layers of revision history as proposed by some scholars, Carr sees only 2-3 major stages. He notes that while redaction activity is clearly evident, we simply do not know how texts grew to their final form, and should be more cautious about assertions. Scholars have been too presumptuous in making claims. For example, scribes in an oral culture would sometimes copy texts from memory leading to slight distinctions. Compare, for example, Proverbs 6:11-10 and Proverbs 24:33-34, as well as Proverbs 16:2a, 21:2a, 12:15a. Scholars tend to assume a textual transmission history governed by written text rather than orality and memory. But if scribes sometimes worked from memory, we cannot always trace back to a specific fixed text (albeit that does not mean there were no textual tradition as well). Similarly, we might understand textual variants, such as expansions, as related to memory (e.g. a scribe recalling a tradition from another biblical text and incorporating it). Carr believes the bulk of the biblical texts reached their general shape far earlier than many scholars now allow and that revisions over time were primarily for preservation, not creation of new texts. He suggests the standardization of the Hebrew Bible probably took place in the 2nd century under the Hasmonean/Maccabean dynasty. They were the first to assert the idea that prophecy had ended (1 Macc 9:27). Thus, their canon was limited based on this understanding. Only books in circulation prior to the Maccabean period are included in the Masoretic canon (with Daniel coming in at the tail end).
In essence, biblical texts underwent revision over time via copying. How that occurred is subject to debate. But, ultimately, standardization toward the proto-Masoretic tradition occurred around the 2nd or 1st century B.C.E. This standardized text paved the way for a closed canon. To that I turn in the next post.
 The Assyrian oracles were collated by date and included a mixture of prophets together. The biblical text, on the other hand, collates oracles by prophet (Karel van der Toorn, Scribal Culture and the Making of the Hebrew Bible [Cambridge, MA: Harvard University Press, 2007] 184).
 “The Evolution of a Theory of Local Texts” in Qumran and the History of the Biblical Text (eds. F. Cross and S. Talmon; Cambridge: Harvard University 1975) 306-20.
 Eugene Ulrich, The Dead Sea Scrolls and the Origins of the Bible (Grand Rapids: W.B. Eerdmans, 1999) 83.
 “The Textual Study of the Bible: A New Outlook” in Qumran and the History of the Biblical Text (eds. F. Cross and S. Talmon; Cambridge: Harvard University 1975) 321-400.
 Karen H. Jobes and Moises Silva, Invitation to the Septuagint (Grand Rapids: Baker Academic, 2000) 35-6.
 Ulrich, Dead Sea Scrolls.
 One example of adaptations is scribal harmonization of texts. One document at Qumran, 4QDeuteronomy, harmonizes Exodus 20:8-11 and Deuteronomy 5:12-15. These biblical texts contain the Ten Commandments but provide different reasons for why the Sabbath is honored. The scribe harmonized the texts by including both rationales into his copy of Deuteronomy 5. Similarly, at Qumran there are proto-Samaritan texts that harmonize. This type of expansion is common in the Samaritan Pentateuch because the Samaritans adopted and used this Hebrew tradition of the Pentateuch that was already in circulation. One example can be found in the proto-Samaritan text 4QpaleoExodm (Sidnie White Crawford, The Fluid Bible: The Blurry Line Between Biblical and Nonbiblical Texts, Bible Review XV  34-39, 50-51).
 The following discussion on Emmanuel Tov and Armin Lange is based on: Armin Lange, “The Textual Plurality of Jewish Scriptures in the Second Temple Period in the Light of the Dead Sea Scrolls” in Qumran and the Bible: Studying the Jewish Scriptures in Light of the Dead Sea Scrolls (ed. by Nóra Dávid and Armin Lange; Dudley, MA: Peeters, 2010) 43-96.
 Sidnie White Crawford, “Understanding the Textual History of the Hebrew Bible: A New Proposal” in The Hebrew Bible in Light of the Dead Sea Scrolls (ed. by Armin Lange, Kristin De Troyer, and Shani Tzoref; Oakville, CT: Vandenhoeck & Ruprecht, 2012) 60-69.
 Arie Van der Kooij, “Preservation and Promulgation: The Dead Sea Scrolls and the
Textual History of the Hebrew Bible” in The Hebrew Bible in Light of the Dead Sea Scrolls (Ed. by Armin Lange, Kristin De Troyer, and Shani Tzoref; Oakville, CT: Vandenhoeck & Ruprecht, 2012) 29-40.
 David M. Carr, The Formation of the Hebrew Bible: A New Reconstruction (Oxford,
Oxford University Press, 2011).