DATING THE END-DATE OF THE QUMRAN CAVE TEXTS
Excerpts from recent publications

(Supplementary handout, Nov. 12, 2006, Gig Harbor, Washington)

The following is excerpted from K.L. Rasmussen, J. Gunneweg, G. Doudan, J. Taylor, M. Belis, J. van der Plicht, J.B. Humbert and H. Egsgaard, "Cleaning and Radiocarbon Dating of Material from Khirbet Qumran", in J. Gunneweg, C. Greenblatt and A. Adraens (eds), Bio- and Material Cultures at Qumran (Brussels: Fraunhofer IRB Verlag, 2006): 139-163 at 156-160.

Linen datings in the caves associated with scroll deposits

Of particular interest are the datings of linen from caves near Qumran which are likely associated with scrolls deposited in those caves. Taylor et al. (2005) argued that all of the large quantity of linen found in Cave 1 was associated with deposits of scrolls to that cave. While hypothetically scenarios could be imagined by which linen could end up in scroll-bearing caves near Qumran not associated with scroll deposits, it is highly plausible that all of the linen items dated in the present series from Qumran caves 4, 8, and 11 were associated with scrolls in each of those caves.
    As further discussed in Taylor et al. (2005), the linen could be associated with scrolls in more than one way. Each scroll likely had its own linen wrapper. But also, in some caves the scrolls were placed in jars and the jars then sealed with linen wrapping and extra linen packed in the jars as stuffing (Crowfoot 1955, 19). Therefore the linen found in scroll-bearing caves near Qumran may have been associated with individual scrolls either prior to or at the time of their deposit in the caves.
    As pointed out by Taylor et al. (2005) with reference to the early work of Grace Crowfoot on linen found in Cave 1: "Crowfoot notes how 'many of the [linen] cloths show signs of wear and tear and have several repairs' and 'there are only one or two instances of cloths whose fringe ends seem to show no signs of fraying caused by use'. The fact that so many of the cloths are cut down from larger pieces also means there may have been a separate period of use for the larger pieces ... which may then mean that from the earliest probable radiocarbon date we should add a period of use before the employment of the linen for storing the scrolls in Cave 1".
    The first Qumran linen radiocarbon date was Willard F. Libby's famous 1950 dating of Cave 1 linen - the second radiocarbon dating of any item in history. Libby reported this linen had an age of 1917 +- 200 "Before Present" (BP) (Libby 1951: Sellers 1951). Because the need for calibration of radiocarbon age measurements was not known and implemented until over a decade later, this was straightforwardly, though slightly erroneously, reported at the time as an age estimate for the Qumran Cave 1 linen of 33 AD +- 200 years, at one standard deviation. After calibration by today's standards, Libby's Cave 1 linen of 1950 dated to between 170 calBC and 340 calAD at one standard deviation (68% confidence), or an even wider 9-1/2 century range, between 400 calBC and 550 calAD, at two standard deviations (95% confidence). It can be seen that although the Libby radiocarbon date was compatible with human activity in Cave 1 contemporary with Qumran habitation of the Second Temple era, in itself this radiocarbon date did not prove a Second Temple era dating for the scroll deposits (as opposed to somewhat earlier or later by 1-2 centuries). And the Libby dating was worthless in weighing for or against individual decades or centuries within the time Qumran was inhabited - since the one standard deviation range alone extends to much greater than the entire range of Second Temple period habitation at Qumran. The only practical outcome of the Libby Cave 1 linen date was a rather compelling argument against a medieval dating of the scroll deposits, which effectively ceased to be an issue thereafter.
    Only in 1994, when a series of radiocarbon datings were carried out on Dead Sea items at the University of Arizona in Tucson, was more useful radiocarbon information obtained on Qumran cave linen. A linen scroll wrapper from Qumran Cave 4 was dated at this time by Accelerator Mass Spectrometry (AMS). For the first time a high-precision radiocarbon date was obtained on such linen. This particular Cave 4 linen wrapper was attached to a leather thong of the type used to wrap scrolls (Jull et al. 1995, 16). This linen measured a radiocarbon age of BP 2069 +- 40, which after calibration by today's standards means between 170 and 40 calBC at one standard deviation, or between 200 calBC and 20 calAD at two standard deviations.
    It took another decade for a second Qumran cave linen wrapper to be dated by high precision AMS. A sample from a linen wrapping from Qumran Cave 1 was obtained in 2001 by a coauthor of the present article (Taylor) from the Palestine Exploration Fund, London (Taylor et al. 2005, 6). This Cave 1 linen gave a radiocarbon age of BP 1984 +- 28, which after calibration is between 40 calBC and 55 calAD at one standard deviation, or between 50 calBC and 80 calAD at two standard deviations.
    To these two high-precision radiocarbon datings of Qumran cave linen done up to 2005 the present series adds either six or eight more, depending on whether the two "possibly trustworthy" AMS datings of cave linen in the present study are included. However note that by the criteria used in the present study, the 1994 Tucson AMS linen dating would also be classified as "possibly trustworthy", as well as most of the radiocarbon datings of texts from the Qumran caves that have been done so far. The 2005 AMS linen dating obtained by Taylor did involve decontamination prior to the AAA pretreatment, however, and is "highly trustworthy". The total database therefore now consists of seven "highly trustworthy", or ten if the "possibly trustworthy" dates are included, high-precision AMS dates on linen from the Qumran caves, with each of these linen items apparently reflecting an association with scrolls deposited in those caves.
    Even though still further radiocarbon datings of linen from the caves would give better information, some preliminary observations can be made. A first observation concerns the difference between the dating of the charred linen from Qumran's locus 96, and the datings of the linen from the caves surrounding Qumran.
    Unlike the linen in the caves, there is no known basis for identifying the locus 96 linen as associated with scrolls. Nevertheless, the dating of the locus 96 linen when compared with that of the linen of the scroll-bearing caves is of chronological interest.
    As noted earlier, the radiocarbon dating of the Qumran locus 96 charred linen was BP 1965 +- 40, which after calibration is between 20 calBC and 80 calAD at one standard deviation. The cause of the charring of the linen is not known. It could be from the fire at the site at the time of the First Revolt, or it could be from an earlier fire, or it could simply have blown into the open area of locus 96 from a cooking fire.
    Seven or ten (depending on how counted) high-precision AMS datings of Qumran cave linen in every case gave radiocarbon dates that appear older than the age of the charred linen of locus 96. To see how striking this is, consider the data in Table 5, with particular focus on the pattern of the lab BP age measurements.
    Ever since de Vaux's first excavation of Qumran in 1951, it has been assumed in Qumran scholarly discussions that the text deposits in the caves at Qumran either occurred or ended at the time of the First Jewish Revolt, immediately prior to a fire at Qumran of c. 68 AD. In his introduction to the first volume of the Oxford University Press flagship series Discoveries in the Judean Desert published in 1955, the director of antiquities of Jordan, G. Lankester Harding, reported without equivocation, "Excavation of the settlement at Khirbet Qumran has established beyond doubt that all of the material was deposited in these caves in the late first century A.D." (Harding 1955, 4). Some Qumran scholars later allowed that some texts could have been deposited earlier, but without questioning that the text deposits continued through the 1st century AD and ended at the time of the First Revolt. This conventional schematic dating of the Qumran text deposits are occurring as late as the 1st century AD was challenged by Doudna in a series of studies as lacking secure evidence on archaeological, paleographic, or radiocarbon grounds (Doudna 1998, 1999, 2001, 2004, 2006). For the first time since the excavation of Qumran the question was raised: whether any scroll deposits in the caves of Qumran can be known to have occurred as late as the 1st century AD, as distinguished from all being 1st century BC.

Table 5. High-precision AMS dates of linen from Qumran locus 96 and from caves near Qumran presumably associated with scroll deposits.

While any individual radiocarbon dating in Table 5 would be indecisive, considered in aggregate the pattern is interesting: seven out of seven (if the three "possibly trustworthy" AMS datings are excluded), or ten out of ten (if the three "possibly trustworthy" AMS datings are included) Qumran cave linen radiocarbon datings measured older than the date of the charred linen of Qumran locus 96. This is what would be anticipated on the hypothesis that the cave scroll deposits occurred earlier than the time of the First Revolt. But it is also what we would expect if the linen in the caves had a long use prior to its employment in the caves. There is no claim here that these radiocarbon dates prove the hypothesis correct that the Qumran cave text deposits date to the 1st century BC, earlier than commonly supposed. But we note that the data above are compatible with such an hypothesis. What we do not see is the radiocarbon date of the locus 96 linen at Qumran measuring more or less equivalent in date with many of the linen items found in the caves.
    Of course there are many explanations for why cave linen samples could have radiocarbon dates c. 1st century BC without meaning scrolls were deposited in the caves 1st century BC. The radiocarbon dating of the linen establishes only terminus a quo information for the dating of scroll deposits. Nevertheless the present AMS datings on the cave linen leave the question of the dating of the scroll deposits open. Many, perhaps all, of the linen items associated with scroll deposits in the caves which have been radiocarbon dated appear to reflect a period earlier than that of the locus 96 linen. In light of these data, we believe both the "late dating" of scroll deposits to the 1st century AD time ending at the time of the First Revolt traditionally held by Qumran archaeologist and text scholars, and the alternatively proposed "early dating" of the Qumran cave scroll deposits as an exclusively 1st century BC phenomenon, merit closer study, to see whether either of these hypotheses can be falsified or shown correct.

References in order cited:

J. Taylor, K. Rasmussen, G. Doudna, j. van der Plicht, and H. Egsgaard, "Qumran Textiles in the Palestine Exploration Fund, London: Radiocarbon Dating Results", Palestine Exploration Quarterly 137 (2005), 159-167.

W. Libby, "Radiocarbon dates, II", Science 114 (1951), 291-96.

O. Sellers, "Date of cloth from the 'Ayn Feshka Cave", Biblical Archaeologist (1951), 14, 29,

G. Harding, "Introductory, the Discovery, the Excavation, Minor Finds", in D. Barthelemy and J. Milik, Qumran Cave I (DJD 1; Oxford: Clarendon Press, 1955).

G. Doudna, "Dating the Scrolls on the Basis of Radiocarbon Analysis", in P. Flint and J. VanderKam (eds), The Dead Sea Scrolls after Fifty Years (Vol. I; Leiden: Brill, 1998), 430-71.

G. Doudna, "Redating the Dead Sea Scrolls found at Qumran" (Qumran Chronicle special edition 8/4; Kracow: Enigma Press, 1999).

G. Doudna, "Appendix A. Palaeography and the Dating of Individual Qumran Manuscripts", in G. Doudna, 4Q Pesher Nahum: A Critical Edition (Sheffield: Sheffield Academic Press, 2001), 675-682.

G. Doudna, "Redating the Dead Sea Scroll Deposits at Qumran: the Legacy of an Error in Archaeological Intepretation", June 2004, Bible and Interpretation website (http://www.bibleinterp.com/articles/Doudna_Scroll_Deposits_1.htm).

G. Doudna, "The Legacy of an Error in Archaeological Interpretation: the Dating of the Qumran Cave Scroll Deposits", in K. Galor, J.-B. Humbert, and J. Zangenberg (eds), The Site of the Dead Sea Scrolls: Archaeological Interpretations and Debates. Proceedings of a Conference held at Brown University, November 17-19, 2002 (Leiden: Brill, 2006), 147-57.

The following sections are excerpted from G. Doudna, "The Legacy of an Error in Archaeological Interpretation", in K. Galor, J.-B. Humbert and J. Zangenberg (eds), The Site of the Dead Sea Scrolls: Archaeological Interpretations and Debates (Leiden: Brill, 2006), 147-157 at 152-153.

"B.C.E." (Before Common Era) = BC
"C.E." (Common Era) = AD

The Creation of First Century C.E. Paleographic Dates for Qumran Texts

Many today think that some Qumran texts have been independently dated to as late as the first century C.E. by paleography. But, in fact, those who defined the absolute datings of the formal scribal hands of the Qumran texts - Nahman Avigad, Frank Moore Cross and, currently, Ada Yardeni - did so based on the assumption that the Qumran cave texts ended c. 70 C.E., which was assumed to be an external checkpoint, an independent archaeological fact. That is how Qumran texts came to be as late as the first century C.E. paleographically. It is one hundred percent circular reasoning, as the following brief historical sketch will make clear. Recall that prior to the first excavation season at Qumran, all published Qumran texts were paleographically dated pre-Herodian. Then, in 1951, de Vaux excavated Qumran, found the scroll jar in locus 2, and announced to the world that this proved the cave scroll deposits had taken place in the first century C.E. This created an expectation that there should be first century C.E. texts, although at that point none were known. The gap was soon remedied: in 1953 Cross reported the discovery of the first scribal hands in Qumran texts from as late as the Herodian period among the unpublished Cave 4 fragments.

[T]hanks to the enormous quantity of material in the fourth cave, examples of every stage in the evolution of the 'Aramaic' character, from cir. 200 B.C. to 70 A.D., are in hand. [They] continue into the script of the Herodian period, known especially from funerary inscriptions (and thus later than the latest of the 1947 finds published to date).1

Soon the First Revolt construct retroactively pulled the paleographic datings for a large number of additional Qumran texts, including most of the previously published Cave 1 texts, later into the Herodian period.

The organization of a typological series with scores of exemplars of the formal script, both from MSS and inscriptions, is now in progress ... Similarly, the cursive series can be set up, though with fewer specimens. From Qumran, MSS exhibiting both hands stand side by side from the second century BC until the First Revolt.2

However, external data has not cooperated with this construction, which is derived from the First Revolt premise. In 1961, Cross reported that semi cursive hands from as late as the first century CE are entirely missing from the Qumran texts (they never existed in the Qumran texts to begin with).3 And in 1998, Yardeni argued that a tiny handful of first century CE true cursive hands among the 4Q texts have nothing to do with Qumran (that is, those 4Q texts indeed are first century CE, but they are not from Qumran).4 Today the chronologically floating formal hands are all that remain among the Qumran finds which are still generally believed to run as late as the first century CE.
    But an example of "first century AD" writing (according to Cross) turned up on a locus 89 bowl from Period Ib - decades before this was supposed to exist in Cross's system. This caused a perplexed but honest Cross to say that if the locus 89 bowls really were from Period Ib, then the actual dates for developments in the formal hands might be systematically earlier than he had published for them.5 Few today question that the locus 89 bowls are Period Ib, and the first century BCE dating of the locus 89 bowls has been corroborated by finds at Jericho.6 Yet the implications of Cross's reasonable statement have never been taken seriously. There is no non-circular argument for dating Qumran texts found in the caves on paleographic grounds later than the end of Period Ib.

Conclusion

It is a curious paradox that scholarly constructions often retain momentum after the original reasons which created them are acknowledged to be mistaken. There was no actual basis in the data for de Vaux's confidence (when in 1952 he announced the first findings at Qumran and declared) that the scrolls of Cave 1 were deposited as late as the first century CE, since the dating of the locus 2 scroll jar was uncertain. But, de Vaux did not know this, because at the time he found the locus 2 jar he knew of only one occupation period for Qumran, in the first century CE. The discovery of the distinct, earlier first century BCE occupation at Qumran, including locus 2, was reported by de Vaux after the next excavation season, in 1953. Yet, the perception of certainty surrounding the First Revolt date for the scroll deposits remained uncorrected down to the present day. The first century CE dating of the Qumran text deposits is a classic example of a mistaken scholarly paradigm filtering subsequent perception of data (archaeological, paleographic and radiocarbon), creating illusions of independent corroboration. In fact, it has never been soundly established that texts found in the Qumran caves were composed, copied or deposited in the caves later than the time of Qumran's Period Ib. Once this is acknowledged, the question is raised whether there is a sound basis to suppose first century CE Qumran text deposits in the absence of evidence.
    A significant difference between Qumran Periods Ib and II with respect to the texts is already accepted: the texts in the caves reflect flourishing authorial activity during the time of Qumran's Period Ib, but, strangely, none at all in Period II. According to the prevailing scholarly construction, the inhabitants at Qumran switched over to copying old texts, without authoring a single new one, through the entirety of the first century CE until the First Revolt.7 No reason is given. However odd this may seem, it has been regarded as a necessary interpretation in light of what has been assumed to be "archaeological fact" (the First Revolt deposit date).
    In light of the foregoing analysis, a different possibility suggests itself. The complete absence of even one allusion to a figure, circumstance, or event in the first century CE in a corpus of texts on the scale of the finds at Qumran - compared to dozens of such allusions from the first century BCE - is well explained if the text deposits themselves ended in the first century BCE.
    In the same way, the fluid, pre-stabilization character of the biblical texts found at Qumran, compared to the post-stabilization character of biblical texts found at Masada, also is well explained if the Qumran text deposits ended earlier than commonly supposed.8

1. F.M Cross, Bulletin of the American Schools of Oriental Research 132 (1953), 16.

2. F. Cross, "The Oldest Manuscripts from Qumran", Journal of Biblical Literature 74 (1955): 147-8.

3. F. Cross, "Development of the Jewish Scrips" in G. Wright (ed.), The Bible and the Ancient Near East (Garden City: Doubleday, 1961), 188: "A gap of considerable length must be posited between the latest of the semicursives of Qumran and the extant [non-Qumran] Herodian cursives and post-Herodian semicursives."

4. Regarding economic texts 4Q342-348, 4Q351-354 and 4Q356-360b, Yardeni concluded: "Despite their designation, it is unlikely that they came from Qumran Cave 4" (H.M. Cotton and A. Yardeni, "General Introduction", in Cotton and Yardeni [eds], Aramaic, Hebrew and Greek Documentary Texts from Nahal Hever and Other Sites [DJD 27; Oxford: Clarendon, 1997], 6 and 283-4). Yardeni adds: "The cursive script utilized in these texts sets them apart from the other Qumran manuscripts" (ibid., 283).

5. Cross 1961, 190 note 9.

6. R. Bar-Nathan, Hasmonaean and Herodian Palaces at Jericho: Final Reports of the 1973-1987 Excavations. Vol. 3: The Pottery (Jerusalem: Israel Exploration Society, 2002), 89.

7. E.g. H. Stegemann, The Library of Qumran (Grand Rapids, Eerdmans, 1998), 137: "[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 BC... 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, but neither revising their contents nor expanding or abridging them".

8. I. Young, "The Stabilization of the Biblical Text in the Light of Qumran and Masada: A Challenge for Conventional Qumran Chronology?", Dead Sea Discoveries 9 (2002), 364-90. S. Talmon writes: "[T]he Masada biblical fragments witness to the existence of a stabilized proto-masoretic textual tradition. [...] In contrast, the textual fluidity, which can be observed in the Qumran scrolls and fragments of biblical books and bible-related works, which stem from the last centuries BCE, proves that these manuscripts were not subjected to such a stabilizing process" (in J. Aviram, G. Foerster, and E. Netzer [eds], Masada: The Yigael Yadin Excavations 1963-1965 [Jerusalem: Israel Exploration Society, 1999], 25). E. Tov states: "Les texts de ces trios sites [Nahal Hever, Murabba'at, Masada] sont ainsi Presque identiques au texte consonantique medieval du TM, encore plus que cuex 'protomassoretiques' de Qoumran" ["The texts of the three sites - Nahal Hever, Murabba'at, Masada - are nearly identical with the consonantal text of the medieval Masoretes, still more so than the 'protomasoretic' texts of Qumran"] (Tov, "L'importance des texts du desert de Juda pour l'histoire du texte de la Bible hebraque. Une nouvelle synthese", in E.-M. Laperrousaz [ed.], Qoumran et les manuscripts de la Mer Morte [Paris, 2000], 210).