Skip to: Site Menu | Main content

Redating the Dead Sea Scroll Deposits at Qumran: the Legacy of an error in Archaeological Interpretation

(page 3)
The creation of 1st-century CE palaeographic dates for Qumran texts

But what about palaeography? Is it not the case that many Qumran texts have been dated to as late as the 1st century CE, Qumran’s Period II, independently by palaeography, such that discussion of scroll jar uncertainties is rendered almost irrelevant? That is a complete misconception. In fact, those who created the currently published datings of the scribal hands of the Qumran texts—Avigad, Cross, and currently Yardeni—did so based on the assumption that the Qumran cave texts ended in 70 CE. They assumed this date was an external checkpoint, an independent archaeological fact. That is how the scribal hands of Qumran texts came to be thought of as extending into the 1st century CE. It is completely circular reasoning.

Remember that before the excavation of Qumran, all of the published Qumran texts were palaeographically dated no later than mid-1st century BCE. But then in 1951, de Vaux and Harding excavated Qumran, found the scroll jar in locus 2, and announced to the world that that the locus 2 jar proved that the scroll deposits were 1st century CE. This created an expectation that there should be Qumran texts as late as the 1st century CE in terms of palaeographic dating, even though none was yet known. The gap was soon filled. In 1953, Frank Cross, one of the original team of scholars working on the scrolls, reported that he had identified sequences of both cursive and formal hands in the unpublished scrolls from Cave 4, which, he said, went to the time of the First Revolt in the 1st century CE.

And soon after that, the supposed First Revolt dating of the scroll deposits caused Cross and other scholars to shift the dates of scribal hands of many more Qumran texts forward into the Herodian period and 1st century CE, including texts from Cave 1 which formerly had been palaeographically dated earlier. The First Revolt date for the scroll deposits "pulled" those palaeographic dates forward.

But the actual data has not cooperated with this now-standard scholarly construction of palaeographic datings of the Qumran texts. In 1961, Cross reported that one entire class of scribal hand—the semicursives—is completely missing from Qumran texts for the first century CE.

Cross (1961): "A gap of considerable length must be posited between the latest of the semicursives of Qumran and the extant Herodian cursives and post-Herodian semicursives."

Over three decades later, in 1998—in DJD 27—Ada Yardeni showed that the small number of 4Q texts identified as having 1st century CE cursive writing appear to be misidentifications of provenance. That is, those 4Q texts indeed are from the 1st century CE, but they are not from Qumran. And there went the cursives supposed to be from the 1st century CE at Qumran.

And so as it stands today, no semicursive or cursive scribal hand among the texts found in the caves at Qumran is even claimed by most Qumran scholars to postdate the 1st century BCE. Today the chronologically floating formal hands are all that remain claimed to be paleographically as late as the 1st century CE among the texts found in the caves at Qumran.

All palaeographical dates for Qumran texts in formal hands published since 1961 have been assigned on the basis of a single chart made by Cross showing the evolution of the formal hands of the Hasmonean and Herodian periods (Cross 1961: Figure 2). In this chart, lines of letters from ten texts are listed in typological sequence. Dates are given for each line of script, running from c. 175 BCE to 133 CE. The scribal hands of Qumran texts are compared to the lines of letters on this script chart, and the dates of the corresponding lines on the chart become the dates for the Qumran texts.

However, of the ten dated lines for the formal hands of the Hasmonean and Herodian periods in Cross’s chart, only the final one, line 10, a contract in Hebrew, is actually dated. It is dated 133 CE, from a date formula. The other nine are all undated (floating). They were given their dates by Cross in relationship to each other in a manner that seemed plausible, in terms of the assumption that the Qumran texts ran continuously to the time of the First Revolt. To cite the resulting palaeographical dates as showing that Qumran texts existed in the 1st century CE is completely circular.

An example of "Herodian formal" writing from, as Cross put it in 1958 and 1995, "the first century A.D.," turned up on a bowl found by de Vaux in locus 86 in a Period Ib context. Not a single archaeologist today disputes that that locus 86 bowl is from Period Ib or that the writing on that bowl was inscribed in Period Ib. (The bowl was inscribed before it was fired.) The Ib dating of this locus 86 bowl is corroborated by the finds at Jericho, where that kind of bowl is found only in the 1st century BCE and not at all in the 1st century CE. Yet this Period Ib dating is before the writing found on the locus 86 bowl is supposed to exist, according to the system defined by Cross. In 1961, Cross, puzzled but honest, wrote that if that locus 86 bowl really was from Period Ib, then dates of development he had defined for the formal hands might be earlier than he had published for them. Cross’s reasoning on this was completely correct, and there is no reason whatever not to take Cross quite seriously on this. In fact, there is no evidence that any of the scribal hands of the texts found in the caves of Qumran is later than the end of Qumran’s Period Ib.

(See further discussion in section 13 of Greg Doudna, "Ostraca KhQ1 and KhQ2 from the Cemetery of Qumran: a New Edition", in Journal of the Hebrew Scriptures, Vol. 5, No. 5.)

(The small number of Greek texts among the Qumran finds have also received palaeographic date estimates, generally centering on or compatible with 1st century BCE. The expert brought in to do the palaeography for the Qumran Greek texts in DJD 9 sought to caution Qumran scholars concerning uses of these dates. Parsons [1992]: "dated parallels are rare, and all are from other parts of the Graeco-Roman world—Egypt … and Herculaneum … The notes which follow … need a double pinch of salt: the process is unreliable even with substantial manuscripts, but here we are dealing with dim and scrappy fragments which allow no more than a partial impression of the graphic ensemble …")



It is also sometimes claimed that radiocarbon dating has proven that dates of Qumran texts are as late as the 1st century CE.

Magness (2002): "[R]adiocarbon dating confirmed the 2nd-century B.C.E. to 1st-century C.E. date that paleographers (specialists in ancient handwriting styles) had already suggested for the scrolls (a date consistent with the pottery types found with the scrolls in the caves)."

Dimant (1995): "In the case of the Qumran manuscripts a few facts are firmly established: the manuscripts extend over a period of more than three centuries, from the second century B.C.E. to the eve of the fall of Jerusalem in 70 C.E. … [a]s was recently reestablished by the new, more advanced carbon-14 tests [at Zurich] … These tests have independently confirmed the soundness and precision of the dating achieved by paleographical typological chronology. For the palaeographical charts of the scrolls see [Cross 1961]…"

This too is incorrect and a misunderstanding. Such claims have arisen from interpreting ambiguous data through the filtering effect of the First Revolt construction and the palaeographic "dates" themselves, rather than evaluation of the radiocarbon data viewed independently of such presuppositions. Unfortunately, time does not permit a full discussion here, but the following is a good summary: the existing radiocarbon data, while confirming 2nd and 1st century BCE dates of Qumran texts, do not confirm that Qumran texts are from as late as the 1st century CE. A sound interpretation of the existing data is that true dates of texts from the caves at Qumran as late as the 1st century CE are neither confirmed nor refuted on grounds of radiocarbon data alone. This ambiguity will not always be the case. The picture will become clearer with further radiocarbon data.

In radiocarbon datings of Dead Sea texts at Zurich in 1991 and Tucson in 1994, nineteen Qumran texts were dated. Here a methodological point must be emphasized. In dating items from a similar time-frame, one cannot simply radiocarbon date a number of items and then focus on the latest radiocarbon date, whichever it is, as if that proves the latest date for the entire group. To illustrate, below are radiocarbon dates for five contemporary Dead Sea documentary texts from the time of Bar-Kochba, which were done in the same Zurich and Tucson series which dated the Qumran texts.


Radiocarbon dates for five documentary texts from the time of Bar-Kochba of identical true dates (128-135 CE)

Item Lab measurement 95% confidence 68% confidence True date (internal date)
Wadi Seyal 1917 +/- 42 2—220 CE 32—129 CE 130-31 CE
Mur 30 1892 +/- 32 32—224 CE 77—132 CE 134 CE
5/6 Hev 19 1827 +/- 36 84—322 CE 131—240 CE 128 CE
5/6 Hev 21 1799 +/- 57 80—329 CE 132—324 CE 130 CE
XHev/Se 8a 1758 +/- 36 140—390 CE 237—340 CE 135 CE


The latest of these five radiocarbon dates—the last one above (XHev/Se 8a)—is 140-390 CE, at 95% confidence. But the true dates of all five of these texts are known: they are all written 128-135 CE (from date formula). Imagine that the true dates of these texts were not known. How would that last radiocarbon date be interpreted? Focusing on the latest radiocarbon date of a group can be misleading in determining the latest true date among a group.

In fact, out of the nineteen Qumran texts dated by Zurich and Tucson, only two gave radiocarbon dates with 95% confidence ranges entirely later than the time of Qumran Period Ib. The first was 4QSd, which gave a radiocarbon date of 129-318 CE, at 95% confidence. The second was 4QpPsa, which gave a radiocarbon date of 3-126 CE, at 95% confidence. 4QSd was rechecked because a 2nd to 3rd-century CE date for a Qumran text was considered impossible. A second sample cut from a different area of 4QSd gave a significantly earlier dating, indicating the radiocarbon date on the original sample from 4QSd had been affected by some modern contaminant. 4QpPsa was not rechecked since it was consistent with the First Revolt construction. The First Revolt construction—an idea in the minds of scholars having nothing to do with radiocarbon data—had determined which text radiocarbon date was further checked and which was not. 1QpHab, which is almost certainly contemporary to 4QpPsa, radiocarbon dated earlier, in the 1st century BCE (160 BCE-2 CE at 95% confidence [at 68% confidence, 88-2 BCE]). In light of the "outlier" status of the radiocarbon date of 4QpPsa—4QpPsa gave the latest, unrechecked date of all 19 dates (the actual latest turning out to be contaminated, when rechecked)—it is simply wrong to claim that the radiocarbon date for 4QpPsa proves true dates of Qumran cave texts as late as the 1st century CE. (Remember the Bar Kochba text example above.)



No actual evidence on archaeological, palaeographic, or radiocarbon grounds justifies the existing scholarly certainty concerning 1st century CE/ First Revolt Qumran text deposits. In fact, there have been signals all along suggesting that the Qumran scroll deposits are earlier than commonly supposed. But these signals have not been appreciated.

For example, there are numerous allusions in the Qumran texts to figures, circumstances, and events in the 1st century BCE. But after the end of Qumran’s Period Ib, historical allusions in the Qumran texts cease completely and permanently, without any exception. This is well known to Qumran scholars. But why? This pattern might suggest—if one did not know otherwise—that perhaps the text deposits are a phenomenon of Qumran’s Period Ib. If one did not know better, this might be the simplest, first, and default interpretation that would come to mind. But here is how this gets interpreted in mainstream Qumran scholarship today:

Stegemann (1998): "[I]t is still surprising that, among all the rich Qumran finds, there seems to be not a single Essene work that we can prove to have been composed only after the middle of the first century B.C. … From that time forward, they concentrated essentially, perhaps even entirely, on the biblical writings, on other works of pre-Essene tradition, and on writings of their own that they had already produced, studying and copying these again and again [until 68 CE], but neither revising their contents nor expanding or abridging them …"

Stegemann gives no reason for this odd change. This doubtful story is simply ad hoc, made necessary by the unquestioned starting assumption of First Revolt text deposits.

A more interesting response to the data is proposed by Michael Wise. In his 1999 book, The First Messiah, and more recently in an article in Journal of Biblical Literature in 2003, Wise notes flourishing allusions in Qumran texts to figures and events of the 1st century BCE down to about 37 BCE, followed by zero after that. Wise concludes from this that the group which produced the Qumran sectarian texts died out late in the 1st century BCE—because if the sect had continued to exist, Wise argues, there would be new text compositions with 1st century CE allusions among the text deposits at Qumran. Wise also proposes a single-decade theory for production of the majority of scribal copies of the sectarian texts.

Wise (1999): "nearly 90 percent of all the Society’s books were copied in the first century B.C.E. … 52 percent of them originate in a single script period that likely represents approximately the years 45-35 B.C.E. … But then production falls off a cliff … the failure to write new books; and the failure to copy old ones; taken together, these argue that the Society died in the first century B.C.E. At least, we have no textual evidence that the movement survived and strong reason to think it did not … The final extinction must have occurred around the turn of the eras, if not a decade earlier … The relatively few copies of the society’s writings that were produced in the first century C.E. can be explained as the work of carriers, people who for one reason or another found the works valuable but who were not related to the original movement."

Wise (2003): "… six references [in Qumran texts] to second-century B.C.E. people, processes and events as against twenty-six to those of the first century B.C.E. … This considerable overbalance toward the first century [BCE] must be accorded due weight … no references to the first century C.E. appear … Surely this puzzle requires explanation …"

Wise does not seem to question the notion of 1st century CE/ First Revolt Qumran text deposits. But that premise is the problem. Once that premise is removed, Wise’s interpretation remains a possible explanation, but is not the only one. But whether one agrees with Wise or not about the extinction, Wise has with devastating force correctly focused on the 1st century BCE as the time of flourishing authorial and scribal copying activity reflected in the Qumran texts, followed by a sharp dropoff—which seems to correspond to Qumran’s Period Ib.

Second, it is recognized that all of the biblical texts at Masada are of the Masoretic text type. The Qumran texts, on the other hand, as is well known, have an astonishingly wide variety of versions and editions of biblical texts. Emanuel Tov has noted further that the biblical texts at Masada are closer to the medieval Masoretic Text in terms of minor variants and letter-perfect copying than even those biblical texts at Qumran which are proto-Masoretic Text in type.

Tov (2000): "Les textes de ces trois sites [Nahal Hever, Murabba‘at, Masada] sont ainsi presque identiques au texte consonantique médiéval du TM, encore plus que ceux ‘protomassorétiques’ de Qoumrân." ("The texts of the three sites [Nahal Hever, Murabba‘at, Masada] are practically identical with the medieval consonantal Masoretic text, even more so than the ‘protomasoretic’ texts of Qumran.")


How is this to be interpreted?

In a detailed study in Dead Sea Discoveries in 2002 on this point ("The Stabilization of the Biblical Text in the Light of Qumran and Masada: A Challenge for Conventional Qumran Chronology?" DSD 9: 364-90), Ian Young of the University of Sydney argued convincingly that the stabilization of the biblical text seen at Masada—whereas this is not the case in the Qumran texts—suggests quite simply that the Qumran text deposits are from an earlier period than those at Masada. Young’s important study has received no published response or rebuttal to date.

(It is not clear how many Qumran scholars are even aware of Young’s 2002 article. For example, see the usually well-informed James Davila in his weblog "Paleojudaica" on Oct. 11, 2003. Davila overlooks not only Ian Young’s study but also the note in Young’s article citing an endorsement of the early dating theory by Alan Crown, the authority in Samaritan studies.)

And third, palaeographically (although this has not yet been noticed by Qumran scholars), texts at Masada have formal scribal hands which are more developed typologically than the latest formal scribal hands at Qumran—specifically, a Masada text called "Masada Shir Shabbat" or "Songs of the Sabbath Sacrifice."

Each of these three observations just mentioned are substantial. Each suggests a reality concerning the Qumran text deposits that differs from standard, conventional assumptions.

And here the issue must be correctly framed. The question is not whether there is proof the Qumran scroll deposits are earlier than commonly thought. That presupposes the existing construction has some basis for being correct by default. But that is exactly what is at issue. The correct question asks whether there ever was evidence in the first place that scrolls found in the caves at Qumran are later than Qumran’s Period Ib? If so, what—exactly—is this evidence? If reasons historically claimed for a scholarly construction are invalid, the burden of proof is upon those upholding a scholarly construction to explain exactly the nature of their claim, the degree of certainty being claimed for it, and on what grounds.

There are other cases in which de Vaux reported phenomena at Qumran as being Period II or common to both periods Ib and II, but which later have turned out to be Ib only. This has happened with the dating of the "hellenistic" lamps, the dating of the animal bone deposits, and the dating of the most common type of bowl found at Qumran. It should hardly be unthinkable that the dating of the scroll deposits may be simply one more such case.

| Page 1 | Page 2 | Page 4 |