What Ancient Scrolls Teach Us about the Torah’s Formation
When trying to understand how biblical texts were produced, Bible scholars have long been influenced by the publication media familiar to them. For instance, the first generations of source-critical scholars in the 19th century focused on the idea that the Pentateuch may have begun as independent texts which were then combined by an editor acting as a modern publisher to produce a unified document.
In the mid-twentieth century, the possibility that the Pentateuchal stories have oral traditions at their core took strong hold among Bible scholars. This approach was inspired by how historians understood the transmission-history of Icelandic and other European cycles of oral traditions.
Over the past few decades, an older idea that Pentateuchal texts were formed through the accretion of layer upon layer of scribal supplements (now often termed “redactions”), has become extremely popular. The increased popularity of this model in recent times may result from the fact that this model of multi-layer revision of an existing text is reminiscent of the mechanics of contemporary computer word-processing, where an author, or multiple authors, can add new materials to an otherwise unchanged earlier computer file.
Each model illuminates important aspects of the literary development of the Pentateuch and other biblical books. Nevertheless, I would suggest that we need to look at a more ancient set of data, specifically the material dimension of ancient reading and writing.
Ancient Israelites and their neighbors utilized multiple media for writing, depending on the type of composition. Pottery sherds were used for letters and other sorts of compact records and exercises, and writing boards were used to record somewhat larger texts, at least temporarily. In contrast, scrolls, whether made of papyrus reeds or of treated leather skins, provided a larger writing surface on which longer texts could be written.
Some previous scholars, such as Lajos (Ludwig) Blau (1861–1936) and Menahem Haran (1924–2015), tried to think through the implications of writing specifically on scrolls. These early studies, however, focused primarily on limited data found in the Bible itself and rabbinic literature.
Archaeological finds, however, offer another important entry point to this discussion. We now have a wealth of information from scholarship on existing, preserved scrolls from Egypt (Hieratic, Demotic, Jewish-Aramaic, early Greco-Roman literary scrolls), Levantine sites like Deir ʿAlla, and the scrolls found near the Dead Sea. These documents bear directly on questions regarding the composition history of biblical texts. We will look at two of these implications below.
Revision by Extension
Adding new material to the end of a composition on a scroll—a phenomenon I call “revision by extension”—is easy to execute, and is well-attested in ancient examples. In Egypt, we see the example of Papyrus Chester Beatty I, where a scribe (or scribes) successively added unrelated texts to the same scroll over time; the scribe probably didn’t own these scrolls, so he copied them from an archive onto the blank part of the papyrus he owned. We must remember that both papyrus and parchment were expensive and required extensive treatment before they could be written on, and thus scribes preferred to use whatever blank space they had before attempting to acquire new scrolls. Revision by extension was also used for adding addenda to the main text.
The inscription of this and other papyrus scrolls started on the “recto” side of the scroll, where the fibers run parallel to the text. If someone wanted to add text to such a scroll, they would continue until they filled the recto, and then turn it over and write on the “verso” side.
We can see the same process in Qumran scrolls, although writing on the verso was more difficult for scrolls made from processed animal skins (an early form of parchment). These were typically inscribed on the formerly hairy side of the skin, while the flesh side was rarely written on, as the writing would not come out well. With papyrus, however, the option to write on the other side was utilized when necessary.
Biblical Examples: The Addendum to Leviticus
Leviticus 27 provides a good illustration of revision by extension in a biblical text that originally ended with Leviticus 26, a chapter consisting of blessings and punishments (26:1–38) followed by a final set of promises regarding YHWH’s ongoing commitment to those that remain in exile (26:39–45). This makes sense as the ending to the book, and the chapter’s final verse then affirms that all of the recorded laws were given by YHWH to Israel through Moses at Mount Sinai:
ויקרא כו:מו אֵלֶּה הַחֻקִּים וְהַמִּשְׁפָּטִים וְהַתּוֹרֹת אֲשֶׁר נָתַן יְ־הוָה בֵּינוֹ וּבֵין בְּנֵי יִשְׂרָאֵל בְּהַר סִינַי בְּיַד מֹשֶׁה.
Lev 26:46 These are the decrees, laws and teaching which YHWH set between him and the Israelites at Mount Sinai by the hand of Moses.
After this apparent conclusion, Leviticus 27 introduces a new speech report of YHWH to Moses (Lev 27:1) וַיְדַבֵּר יְ־הוָה אֶל מֹשֶׁה לֵּאמֹר “And YHWH spoke to Moses,” with a command for him to speak to the Israelites (Lev 27:2a) דַּבֵּר אֶל בְּנֵי יִשְׂרָאֵל וְאָמַרְתָּ אֲלֵהֶם “Speak to the Israelites and say to them,” and deliver a detailed set of instructions on how several specific sorts of vows can be discharged (Lev 27:2b–33).
As has long been recognized by scholars, this set of laws represents a new, distinct body of material added to the rest of the book. Apparently, the author of these materials wanted to make clear that these instructions, like those in Leviticus 1–26, belonged to the legislation given by Moses to the Israelites at Mount Sinai. So, he concluded this chapter with a resumptive repetition (Wiederaufnahme), restating the conclusion to Leviticus 26:
ויקרא כז:לד אֵלֶּה הַמִּצְוֹת אֲשֶׁר צִוָּה יְ־הוָה אֶת מֹשֶׁה אֶל בְּנֵי יִשְׂרָאֵל בְּהַר סִינָי.
Lev 27:34 These are the commandments which YHWH commanded Moses at Mount Sinai for the Israelites.
Several other biblical texts have additions or appendices with are examples of revision by extension:
Judges 17–21; Following the stories of Israel’s judges is an appendix with two stories framed by the phrase “in those days there was no king in Israel.”
Numbers 33–36; Following the narrative about Gad, Reuben and part of Manasseh settling the Transjordan, Numbers has four chapters of addenda, including an itinerary list, and various laws dealing with geography and inheritance.
2 Samuel 21–24; Following the story of Sheba’s rebellion and how it was quelled, Samuel ends with several stories, poems, and lists relevant to an earlier period in David’s career.
Psalm 151 (Greek tradition)/Psalms 151–155 (Syriac Psalters); Psalm 150 is a clear ending psalm, listing the musical instruments that are to accompany the recitation of hymns. Some traditions add extra psalms after this.
Jeremiah 52 (MT); at the end of MT Jeremiah, a section is copied from the book of Kings (2 Kgs 24:18–25:30) describing the destruction of Jerusalem, thereby showing the fulfillment of his prophecy of doom.
Isaiah 36–39; At the end of the early section of Isaiah, attributed to the 8th century prophet of that name, three chapters from the book of Kings appear, which narrate this period of time, including some of Isaiah’s actions (2 Kgs 18–20). This is now followed by the work of a different prophet, Deutero-Isaiah (beginning in chapter 40), which describes the early Persian period, but the original scroll first ended with some form of ch. 35, to which this addendum was added.
Non-biblical scrolls feature similar examples, such as
Papyrus Amherst 63 includes the Egyptian Tale of Two Brothers on the last few columns (columns xviii–xxiii).
The Qumran Community Rule; different editions of this text offer different endings.
Scholars have long agreed on the secondary character of these materials. What many have not recognized is the way this sort of revision by extension is particularly characteristic of scroll media. In other words, these evident additions to the ends of Leviticus and multiple other biblical and other ancient books are reflections of the fact that these compositions were inscribed on a certain sort of ancient media: scrolls, which allowed for the easy use of revision through extension.
Scroll Length and Capacity
Theoretically scrolls could be made to virtually any length, and we have at least one example from ancient Egypt, the Harris scroll, that is over forty meters long. This huge scroll contains a copy of the “Book of the Dead,” a collection of mortuary spells that was intended to be buried and read by the gods, not for living human beings who would actually open the text and use it.
Many wide and long scrolls in Egypt contain administrative records; their large format allowed for the recording of a maximum number of transactions. Such administrative scrolls were too unwieldy for any kind of regular reading, but they provided a written record that could be consulted later in the case of a dispute.
Scrolls used for reading texts, however, were a more modest size. Indeed, many literary scrolls in Egypt are repurposed administrative scrolls that were cut to half-width, erased with a wet rag, and reinscribed, thus creating a “palimpsest.” The narrower scroll format provided a reader with a scroll of more manageable width scribes could hold in their laps to read and study. As the French historian Jacques Monfrin (1924–1998) wrote regarding page format in Egypt and the Levant (translation mine),
From the beginning, the limits of the human visual field imposed a distribution of writing in columns. The height of the sheet defined the vertical dimension while the width [of the column], for its part, was determined by the surface that a reader’s two hands could easily unroll before their eyes and by the length of a series of signs that an eye can cover without mixing up the parallel running lines above and below.
Especially insofar as some ancient readers may have been of smaller stature than well-nourished contemporary Western males, large scrolls with column heights surpassing around 20 centimeters would have been difficult for the average ancient scribe to read and manipulate. This column size limit for texts meant for ongoing reading has to do with the materiality of the human body—the arms that hold a scroll and the eyes that read sections of it.
Body size, the limits of arm reach, and the mechanics of handling scrolls also impose some limits on the length of literary scrolls intended for ongoing reading. My survey of literary scrolls across diverse ancient cultures (e.g., Egypt, Greece) suggests that scrolls bearing such texts were usually less than 10 meters in length. These limits correspond roughly to the modest size of most scrolls found at the Dead Sea. The exceptions are a handful of large format, mostly Pentateuchal scrolls, that share features with Greek De Luxe scrolls designed more for display than for ongoing reading.
In sum, preparing inscribed scrolls for use by human readers compels some constraints on size and format. Mundane and technical investigations of scroll length have a direct impact on theories regarding the formation of important Hebrew texts like the Torah.
Implications of a Material-Historical Scroll Approach for Biblical Studies
The limits on (literary) scroll length in the ancient world likely led to the division of the Torah into five scrolls, referred to as חומשים “fifths” in many ancient rabbinic texts or “Pentateuch” (literarally “five scrolls”) in Latin. It also could explain the basic book divisions found in the continuous narrative of the biblical historical books, leading to what we know as the free-standing books of Joshua, Judges, Samuel, and Kings, etc., came to be transmitted on separate scrolls, many concluding with “revisions by extension” of the sort discussed above.
Considerations of typical literary scroll length could also inform more speculative theories surrounding the sources of ancient biblical books and the process by which they were collected. Though biblical scholars often postulate the existence of larger prebiblical sources, like a Priestly work or Deuteronomistic History, they have not considered the extent to which these hypothesized works actually could have fit on a typical-length literary scroll. My initial investigations, using scroll-density figures from well-preserved Dead Sea Scrolls, suggests they would not have.
Of course, there remains the possibility that such works may have been transmitted across multiple scrolls, much like the Homeric epic tradition (Iliad and Odyssey) was. I myself think this is likely for the Priestly tradition, with potential marks of early scroll divisions (in P) in an apparent scroll introductory section in Exod 1:1-5 (P) and possible marks of early endings of early Priestly scrolls, e.g., in the evident appendices to the Sinai pericope that now appear midway through the book of Numbers (Numbers 5-10). Nevertheless, as pointed out years ago by Menahem Haran, such multi-scroll compositions were the exception rather than the rule in the ancient world.
It would be helpful for biblical scholars to know whether the text of the sources they are postulating once existed could have been inscribed on one scroll. The issue is yet more acute when we realize that the Dead Sea Scrolls, written as they were in a much smaller script than that used in the Iron Age and Persian periods, likely could contain far more text per square centimeter than scrolls written in pre-Hellenistic periods.
Indeed, we need to investigate the possibility that the major shift in the density of writing, that occurred sometime in the late Persian or early Hellenistic periods, led to the gathering of separate written collections onto single scrolls. It may be that bodies of related traditions that previously had been written on separate scrolls up through the Persian period (e.g. psalm collections or prophecies associated with Isaiah or Jeremiah) could later be joined and written (in smaller script) on single, larger Hellenistic-period scrolls.
A Scrolls Approach
A wealth of ancient scroll data now make it possible to collect, measure and compare literary scrolls from diverse periods and loci, and this work is already underway by the present author and others. This work on what might be termed a “material historical, scroll approach” to the Bible has much potential to inform future research into the sources and processes that produced the Pentateuch and other important biblical compositions.
TheTorah.com is a 501(c)(3) nonprofit organization.
We rely on the support of readers like you. Please support us.
May 20, 2022
February 14, 2024
Previous in the Series
Next in the Series
Prof. David M. Carr is Professor of Hebrew Bible at Union Theological Seminary in New York. He holds a Ph.D. from Claremont Graduate School and an M.T.S. from Emory University's Candler School of Theology. Carr is the author of several books, including Genesis 1-11, International Exegetical Commentary on the Old Testament (Kohlhammer, 2021), The Formation of Genesis 1-11: Biblical and Other Precursors (Oxford University Press, 2020), The Hebrew Bible: A Contemporary Introduction to the Christian Old Testament and Jewish Tanakh (Wiley Blackwell, 2020), and Holy Resilience: The Bible's Traumatic Origins (Yale University Press, 2014).
Essays on Related Topics: