How To Connect Dead Sea Scroll Fragments

024

025

When the first Dead Sea Scrolls came to light in 1947, putting their pieces together wasn’t really a problem. Indeed, one scroll, from what is now known as Qumran Cave 1, is almost complete. There was nothing to put together. That was the famous Isaiah Scroll, known to scholars as 1QIs^a. The siglum stands for Qumran Cave 1, Isaiah Scroll; the superscript “a” distinguishes this Isaiah scroll from another one found in the same cave, known as 1QIs^b.

Some other scrolls from Cave 1—as much as survived—are mainly in one large piece, so there is little to put together. That is true of such scrolls as the Habakkuk commentary, the scroll of the War of the Sons of light and Sons of Darkness and the Manual of Discipline. Only two of the larger Dead Sea Scrolls from Cave 1 were partly broken into pieces and had to be restored, although, in addition to these larger scrolls, many fragments were also found in Qumran’s Cave 1. Approximately 75 fragmentary scrolls have now been identified from Cave 1.

Gradually, as a result of searches by both professional archaeologists and Bedouin tribesmen, scrolls and fragments of scrolls have also been found in ten other Qumran caves, designated 2Q through 11Q. In addition, scroll materials were found in excavations and in caves in other wadis^a leading down to the Dead Sea— Masada, Nahal Hever, Wadi Murabba‘at, Nahal Se’elim and Wadi DaÆliyeh.^b

In all, fragments from about 800 different scrolls have been recovered from the 11 Qumran caves alone. Their fragments vary in size from large panels containing several columns to thumbnail pieces containing only a single letter.

One composition, found in 3Q and known appropriately enough as the Copper Scroll, is written on copper sheets. In addition, a few ostraca (inscriptions on pieces of pottery) have been found. But the balance are all written on leather (parchment) or papyrus.

The quantity of fragments varies, of course, from cave to cave. The biggest cache came from 4Q, which contained more different documents than all the other caves combined—about 580 different manuscripts. The number of fragments into which these 580 manuscripts were broken has never been determined with any accuracy. Estimates vary between tens of thousands to hundreds of thousands. About 75 percent of these manuscripts from 4Q still remain unpublished and are therefore not yet available to most scholars.

The first task in reconstituting any fragmentary Dead Sea document is to isolate and collect the pieces that come from the same scroll. Sometimes it’s easy to identify such fragments because they are stuck together. But more often they are scattered all over the place.

Three basic clues enable the scholar to gather together the fragments from a single scroll. The first clue comes from the material on which it is written. Obviously, a fragment on parchment is not part of the same scroll as a fragment on papyrus. But beyond this, parchment itself varies and so does papyrus. Some scrolls are written on thick material, others on thin. The color may also vary from brown to yellow or reddish and it may be bright, or dark or in-between.

Another clue comes from the fact that ancient 026scribes prepared their scrolls for inscribing by scoring the scroll at regular intervals with fine lines to guide their hand. The space between these lines varies and, after inscribing the text, the space between the lines of text varies. The text is sometimes written hanging on the lines, sometimes between them; this, too, helps identify the fragments from the same sample. The size of the spaces between the lines is another important clue as to which fragments come from the same document. Finally, the number of lines in each column of a given scroll is somewhat regular, as is the width of the columns. Observation of these similarities also helps in correlating fragments coming from the same scroll.

Third, the handwriting of the different scribes varies. Scholars who work with these scrolls regularly are able to distinguish between these different hands. Usually, although not always, a single scribe worked on each scroll. By identifying his handwriting, a modern scholar can tell whether or not a particular fragment belongs to a scroll written by that scribe.

In this way, the pieces of a single scroll are assembled. The result is often a box of scraps that resembles a jigsaw puzzle. There is a difference, however. Many, often most, of the pieces of this jigsaw puzzle remain missing. How then do we reconstruct the scroll itself? How do we tell the way in which the pieces related to one another in the original manuscript?

In the case of biblical manuscripts, this is not too difficult. The biblical text itself, like the picture on the box of a jigsaw puzzle, provides the grid, or pattern, on which each of the surviving pieces of the ancient manuscript can be located. The complete biblical text, as it has come down to us, serves as a kind of mirror of the ancient text, onto which surviving pieces of the ancient text can be placed. We can place these pieces on the mirror even if we have very few pieces. This process becomes somewhat, but not much, more complicated because there are variations, usually minor, even in biblical texts.

This same method of reconstruction is used for non-biblical manuscripts of which we have other, complete copies or ancient versions. Fragments of apocryphal books such as Tobit and Ben Sira are among the Dead Sea fragments that can be reconstructed by using a modern edition of these well-known books. Other Dead Sea fragments come from still other manuscripts of which we have later documents—such as the so-called Damascus Documents^c or fragments of the Books of Enoch.

Finally, sometimes more than one copy of a text has been found at Qumran itself. Fragments from eight or nine copies of the Songs of the Sabbath Sacrifice were recovered there. In addition to the large Temple Scroll—which is 027nearly 26.5 feet long—fragments from another copy survived. Occasionally, as in these cases, one copy can serve as a grid for the reconstruction of another copy.

Nearly 40 percent of the Dead Sea Scrolls can be reconstructed totally, or at least partly, in this way. About 20 percent of the documents are biblical documents—approximately 170 manuscripts. Fragments from every book of the Hebrew Bible, except Esther, have shown up at Qumran. Another 20 percent of the documents are from texts otherwise known, either from modern copies of ancient versions or from ancient copies. In addition, some fragments, while not strictly biblical, quote biblical passages or paraphrase biblical texts, so the Bible provides help in the reconstruction of the original order of the surviving fragments.

But this leaves more than 50 percent of the Dead Sea Scroll manuscripts—formerly unknown texts—with no grid to use for reconstructing the fragments, for placing the pieces in relationship to one another. Most of these previously unknown texts found in the Dead Sea caves have survived in very fragmentary pieces that are in a poor state of preservation.

The first Dead Sea Scroll publication of previously unknown texts that survived partly in fragments made no effort to place them in any order, let alone to reconstruct the text. In 1954 (two years after the death of the senior editor, E. L. Sukenik), Nahman Avigad published the fragments from the famous Thanksgiving Hymns scroll from Cave 1. He divided the dislocated pieces according to the different scribal hands that had copied the hymns, presented the large fragments first, and concluded with the smaller ones, a total of 66 dislocated fragments from this scroll. This pattern has generally been followed by later editors: Usually the editors of such manuscripts arrange the fragments of their scrolls according to size, from the larger ones to the smaller ones, without any attempt at reconstructing the original state of the scroll. Indeed, very few attempts have been made in the Qumran editions—most of which have been published by J. T. Milik—toward any kind of reconstruction of the original scrolls. The main problem is the total lack of any well-established method for reconstructing fragmentary scrolls without the help of parallel texts.

The principal means for reconstructing such texts is that employed in jigsaw puzzles—finding pieces that join. They have complementary borders. The join has to be sufficient and sufficiently distinct to assure that in fact the two pieces do connect. In addition to the physical shape of the pieces, joins are often indicated because the two pieces divide a word or even a letter. These kinds of connecting joins are called “material joins” because they are based on the physical characteristics, rather than on the thematic content, of the pieces.

028

Incidentally, material joins are also useful in connection with manuscripts for which we have modern texts to use as a grid. The material join is often even more reliable than the grid of a parallel text, for the particular text may vary from the text in the grid. Even in the case of biblical manuscripts, few of them are word-for-word identical.

In addition to material joins, some reconstruction can be accomplished on the basis of thematic context. Even though the fragments themselves do not connect with one another, they can be arranged in relation to one another.^d These are called “distant joins.” Obviously, such joins are often rather speculative.

Recently, however, I have developed a method of identifying distant joins—placing non-connecting fragments in relation to one another—that has produced some remarkable results. The key element in this new method is the relationship of fragments containing similarly shaped, damaged areas.

By definition, scroll fragments are damaged. That is why they are fragments instead of complete scrolls. By studying how they were damaged, we can learn how—at least in part—they can be put together again.

Almost all of these Qumranic fragments come 029from scrolls (as opposed to separate sheets^e) that had been rolled up and then damaged in the Dead Sea caves while still rolled up.

Those who are familiar with modern Torah scrolls used in synagogues know that these scrolls, containing the five books of Moses, have posts or rolling sticks at both ends of the scroll to facilitate rolling from one end to the other. But that was not true of ancient scrolls such as the Qumran scrolls.^f They were simply rolled up, forming a hollow tube or shaft in the center of the rolled-up scroll. At both the beginning and end of the ancient scroll were blank sheets that the reader could hold to roll the scroll backward or forward without having to place his hands on the inscribed parchment or papyrus. These blank sheets at the beginning and end of the scroll are called “handle sheets.”

Some of the scrolls found in the Dead Sea caves were stored there in jars, but most were simply laid, stacked or leaned against one another, otherwise unprotected.

Over the millennia, many of the scrolls were very extensively damaged. The two principal agents of damage were humidity and animals (rodents and insects). The damage thus produced, however, was patterned and repeated. When the bottom part of the scroll touched a wet place, the last lines of all the columns often disappeared. When the scroll stood on its head, it was the tops of all the columns that vanished in the course of the centuries. The edges thus damaged follow a repeated pattern. Other than at the edges, humidity might attack a scroll from the outer layer or from the inner layer (via the hollow shaft at the center of the rolled-up scroll). Sometimes the damage would eat all the way through, but occasionally the innermost layers would remain undamaged, protected by the sheer bulk of the scroll. Damage by rodents and insects, who enjoyed eating the scrolls, occurred in much the same patterned way.

The result is that the holes and breaks in a scroll have similar or even identical shapes through the several damaged layers. This pattern is also found in the fragments that originated from these holes and breaks.

If one tries to reconstruct the text of a scroll (what is written on it), all the holes and breaks are annoying and only sources of trouble, repeatedly interrupting the text. But if one tries to reconstruct the scroll itself, the patterned shapes of these holes and breaks are a reliable aid in arriving at the original order of what remains of the scroll fragments.

What can we learn from these patterned shapes of damage? First, fragments showing corresponding shapes of damage must be positioned along the 043same horizontal axis as measured from the top or the bottom of the original scroll.

Second, the distance between the repeated patterns can often enable us to place the fragments in a particular order. This is because the distance between corresponding damage increases as one moves from the inner layers the scroll outward, and decreases as one moves inward from the exterior layers. The rate of this increase or decrease can actually be mathematically calculated, based on the thickness of the leather or papyrus, and the tightness with which the scroll was rolled or wrapped.

Other clues are also helpful. Was the beginning of the scroll in the outer layer or the inner layer? Generally, the larger scrolls—those with 50 or more columns, often exceeding 20 feet in length—were rolled with the end on the inside. These scrolls were also tightly rolled, so the increase in distance between corresponding points of damage is quite short—from about one twenty-fifth to one-fifth of an inch. Shorter scrolls—a class generally between 5 and 6 feet long and containing between 12 and 20 columns—were more loosely rolled. The increase in distance between corresponding points of damage in these scrolls is mostly about one-fifth of an inch. Finally, a third class of scrolls begins the text on the innermost layer, instead of on the outermost layer. Such scrolls are almost always loosely rolled, regardless of their length. In these, the increase in distance between corresponding points of damage is relatively large.

Accordingly, once we identify corresponding points of damage, we measure the thickness of the leather, or papyrus, and attempt to determine the tightness of the wrap, the original length of the scroll, and whether it began on the innermost or outermost layer. It is also helpful to know the width of the columns; strangely enough, there was no standard column width at Qumran. We also like to know the number of columns on a sheet; remember that all scrolls are ultimately made up of sheets that were sewn or pasted together. Lastly, it helps to know the number of lines in a column, which varies from about seven to about 50 in a given scroll, but is relatively constant through all its columns.

Unfortunately, there is no comprehensive, up-to-date survey of the physical characteristics of the Dead Sea Scrolls. Most of the published editions of the Dead Sea Scrolls present very little information of this kind and often totally lack it. An exception is Yigael Yadin’s edition of the Temple Scroll: What he presents is excellent, but unfortunately relates only to that scroll.

A survey of the scrolls’ physical characteristics would be of enormous assistance in reconstructing both published and unpublished scrolls. With this information, we could in many cases establish the position of scattered fragments in the original scrolls and thus make evident the structure and content of many ancient texts. This is true even with regard to scroll fragments that have long been published.

One example from my own work will suffice. In 1982, 215 fragments of a scroll from Cave 4 were published by Maurice Baillet.^g One fragment of 12 lines (4Q511, fragment no. 10) was clearly from the beginning of the text. Another fragment from another scroll (4Q510) was a parallel text; so on this basis, the first column of this scroll, with 18 lines of the text, could be reliably reconstructed without any technical support. The distance between corresponding points of damage on this first fragment is about 4.7 inches. Another long strip from the top of the scroll represents the last two columns of the text. On this strip the distance between points of damage is only about 1.9 inches at the left end. Other fragments reveal a distance between corresponding points damage of 2.75 inches and 3.5 inches. These fragments can thus be arranged according to their original positions in the scroll. The final result is a continuous sequence of all 12 columns of the original scroll, representing about 80 percent of the original text. This scroll is called Songs of the Sage. Instead of the 215 scattered fragments in the published edition,^h we now have a nice, well-established sequence of several songs. One can study their contents and style of composition, relate them to other poetical texts of their time, etc. True, it takes several weeks to complete such a reconstruction. But the reward is, instead of a list of hypothetical suggestions, an accurate, methodically well-established, new edition of a text formerly unknown to the scholarly world. We should hardly be surprised that the fragments as published give only a very poor notion of their true importance.

(For further details, see Hartmut Stegemann, “Methods for the Reconstruction of Scrolls from Scattered fragments,” in Archaeology and History in the Dead Sea Scrolls—The New York University Conference in Memory of Yigael Yadin, ed. L. H. Schiffman, forthcoming from the American Schools of Oriental Research.)

When the first Dead Sea Scrolls came to light in 1947, putting their pieces together wasn’t really a problem. Indeed, one scroll, from what is now known as Qumran Cave 1, is almost complete. There was nothing to put together. That was the famous Isaiah Scroll, known to scholars as 1QIsa. The siglum stands for Qumran Cave 1, Isaiah Scroll; the superscript “a” distinguishes this Isaiah scroll from another one found in the same cave, known as 1QIsb. Some other scrolls from Cave 1—as much as survived—are mainly in one large piece, so there is little to put […]