Skip to: Site Menu | Main content

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

Abstract: There was no actual basis for de Vaux’s confidence in 1952 (when he announced the first excavation findings from Qumran) that the scrolls of Cave 1 had been deposited as late as the first century CE, since the dating of a "scroll jar" found in locus 2 was uncertain. A distinct, earlier first-century BCE occupation at Qumran was discovered by de Vaux in the second excavation season in 1953. Yet the perception of certainty surrounding the First Revolt deposit date for the scroll deposits have remained to the present day. 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 in the first century BCE. The dating of the Qumran text deposits is a classic example of an unfounded scholarly paradigm filtering subsequent perception of data (archaeological, palaeographic, and radiocarbon), creating illusions of independent corroboration.


By Greg Doudna
Bellingham, WA 98227
June 2004

It is difficult to name a statement concerning the texts at Qumran which has had a more complete scholarly consensus during the past five decades than whether the Qumran text deposits either occurred or ended at the time of the First Jewish Revolt against Rome in 70 CE. In this talk, I will trace how this scholarly consensus came to be, why this consensus was never soundly established, and why it is very likely mistaken.

In 1947, the first Dead Sea Scrolls from the caves around Qumran became known. In 1949, the cave where the first scrolls were found—Cave 1—was excavated, and in 1951, Qumran was excavated. It is little known today (because it is not mentioned in modern accounts of the Qumran finds), but before the excavation of Qumran, all of the relevant archaeologists—de Vaux, Harding, Albright, and so on—were in unanimous agreement that the scrolls found near Qumran were dated no later than the 1st century BCE. This was based on the archaeologists’ dating of wide-mouthed, cylindrical jars found in Cave 1 in association with the scrolls—the so-called "scroll jars." Note the language of certainty.

Albright (1949): "It cannot be too strongly emphasized that the bulk of the pottery [in Cave 1] (all but those few Roman pieces) consists of absolutely homogenous jars, bowls (made specifically to cover the jars), and lamps, whose pre-Herodian date in the last two centuries B.C., is beyond dispute."

Sellers (1949): "[T]he archaeological evidence [from Cave 1] confirms the views of those who had pronounced the manuscripts pre-Christian from the epigraphic and literary evidence. They were deposited in the cave not later than the first century B.C."

Sellers (1951): "Mr. G. Lankester Harding and Father R. de Vaux, who conducted the excavation [of Cave 1, in 1949], are two of the most competent archaeologists who have worked in Palestine. No one has seriously questioned their dating of the pottery as late Hellenistic."

At the time, this verdict of the archaeologists—"beyond dispute," according to Albright—claimed independent confirmation from palaeography. The leading authority on the palaeographic dating of Jewish scripts at that time was Solomon Birnbaum. In a series of studies in 1948 and 1949, Birnbaum had painstakingly dated all of the famous published Cave 1 scrolls on palaeographic grounds to no later than the middle of the first century BCE. This was before any of the caves, or Qumran itself, had been excavated.

Birnbaum (1949): "It was at all times beyond question that the handwriting of none of the Scrolls is later than about the middle of the first pre-Christian century."

Birnbaum (1951): "It may be mentioned that the dates of the Scrolls [mid-1st century BCE and earlier] had been fixed by palaeography, months before the pottery had even been taken out of the Cave, and still longer before the findings of the archaeologists had been made known."

So this was the original consensus, which is very different from today’s consensus on the dating of the scrolls at Qumran. What changed?

What changed was the first excavation of Qumran in 1951, conducted by Roland de Vaux of the École Archéologique Française de Jérusalem, and G. Lankester Harding, director of the Department of Antiquities of Jordan. What de Vaux and Harding found startled them: they found Qumran was inhabited in the 1st century CE, and they found a jar which was the same kind as the jars which had been found with the scrolls in Cave 1. At this time—and this is an important point—this 1st-century CE habitation was the only habitation known to de Vaux and Harding for Qumran that could be relevant to the scrolls. (They also found an earlier Iron Age settlement, but that was centuries earlier.)

This 1st-century CE habitation at Qumran—according to what de Vaux and Harding learned from this first excavation—had ended at the First Jewish Revolt of 70 CE. De Vaux found the identical jar as the kind found with the scrolls—a "scroll jar" (de Vaux never personally used the term "scroll jar," but others did, and the name stuck)—buried in a floor of a room at locus 2. There were 1st-century CE coins and pottery on this floor: all of which were covered over by a destruction layer from a fire at the time of the First Revolt. Two coins were also found underneath the broken floor next to the jar, both also tentatively identified as 1st-century CE, one possibly from the First Revolt.

De Vaux concluded from this that this buried jar in locus 2 dated to the 1st century CE. De Vaux reasoned that the same jars in Cave 1, with the scrolls in them, therefore also dated to the 1st century CE. On the basis of this new information, de Vaux courageously said to the world concerning the earlier dating of the scroll deposits that he and the other archaeologists had held: "je me suis trompé," "I was wrong," in a famous announcement in Paris in 1952.

De Vaux (1952): "Je me suis trompé en attribuant les jarres des manuscrits à l’époque préromaine; elles sont d’un bon siécle plus tardives" ("I was wrong in dating the jars with the scrolls to before the Roman period. They are a good century later.")

De Vaux’s announcement was reported and accepted throughout the scholarly world. Every archaeologist accepted this correction, for—based on what de Vaux had found—the argument seemed, at the time, virtually unassailable. This is pivotal to understand. This announcement of de Vaux based on the find of the locus 2 "scroll jar" was the basis for the First Revolt date construction of the scroll deposits which is believed to the present day. Note carefully the reasoning explaining the change to the First Revolt deposit date.

Wright (1953), Biblical Archaeologist: "[A]n excavation in December of 1951 at Khirbet Qumran … has had the effect of correcting the dating originally assigned to the cave pottery from the 1st Century B.C. to the 1st Century A.D. Coins were found dating as late as the 1st revolt (67 A.D.), in connection with a jar identical with those found in the cave. This also implies that those who lived at Qumran deposited the scrolls there."

Kelso (1955), Journal of Biblical Literature: "These jars [of Cave 1] were of a type never before known and the excavators at first dated them to the 1st century B.C. … Later, however, when they excavated Khirbet Qumran itself, the site of the Essene Community, they found the very same type of jars. Here they were definitely dated by coins to the 1st century A.D. but before the destruction of Jerusalem."

De Vaux (1955), Discoveries in the Judean Desert, Volume I (DJD I): "des dépôts homogènes et contemporains de la première grotte. La date est fixée par les monnaies recueillies dans le bâtiment du Khirbet et par les parallèles qui ont été indiquès à propos de chaque forme. C’est le 1er siècle après J.C., plus précisèment avant 70 de notre ère." ("The homogenous deposits are contemporary in Cave 1. The date is fixed by the coins recovered from the buildings of Qumran and by the parallels which are indicated with respect to each form. This is the 1st century AD, more precisely before 70 AD.")

Harding (1955), DJD I, Introduction: "Excavation of the settlement at Kh. Qumrân has established beyond doubt that all of the material was deposited in these caves in the late first century A.D."

But here is where it gets interesting, for things are not always as they at first seem. There was a problem with de Vaux’s reasoning: it was wrong.

In the second season of excavations at Qumran, in 1953, de Vaux discovered a distinct, substantial, earlier habitation period at Qumran in the 1st century BCE. This earlier habitation had also ended by fire. De Vaux called this newly discovered, earlier Qumran habitation period, "Period I," which he later divided into Ia and Ib. The habitation at Qumran in the 1st century CE that ended at the First Revolt became called Period II. De Vaux belatedly realized—after this second season at Qumran—that the floor of locus 2 had actually been built in the earlier period, Period I. Locus 2 had been cleared out and the same floor had been reused in Period II in the first century CE. That is why coins and pottery from Period II were found on the floor of locus 2—they were from the people who used the room in Period II. The coins do not date the installation of the floor. Nor do the coins date the installation of the "scroll jar" in the floor. The paved floor was broken around the top of the buried jar, meaning coins could have fallen through from above after the jar was already in the floor.

De Vaux reported the new discovery of the earlier habitation period. This should have raised at that time the question of the certainty of the dating of the "scroll jar" buried in the locus 2 floor. For how could it any longer be considered certain that the jar buried in that floor was installed in the later period of use of the floor ending in the First Revolt? How could it be excluded that that jar had been installed in Period Ib, in the 1st century BCE, in the earlier period of use of the floor, and then inherited, along with the floor, by the later people who re-used that room in Period II? But there is no sign in the published record that this question was ever raised. The certainty attached to de Vaux’s 1st-century CE dating of the jar buried in the locus 2 floor remained unchallenged and unquestioned, even after de Vaux moved the date of the floor in which the jar was buried a century earlier.

Could de Vaux have been mistaken on his dating of the locus 2 jar in 1951? Yes, because it is not certain that the locus 2 jar was installed in Period II. But de Vaux thought it was. That is how the scroll deposits were dated to the First Revolt to begin with—based on an illusory claim of certainty when there was no actual basis for that certainty.


A coin correction and the locus 2 scroll jar

De Vaux originally reported in 1953 that one of the two coins found under the floor next to the scroll jar in locus 2 was "peut-être Première Révolte," "possibly First Revolt." That was mistaken. That coin was subsequently identified as from Antigonus Mattathias, 40-37 BCE. The correct identification of this coin first became known to the world only in the publication of Humbert and Chambon’s Fouilles de Khirbet Qumrân et Ain Feshkha in 1994. But de Vaux knew this identification, although he never disclosed it, at the time of his Schweich lectures in London in 1959. This is deduced from the following. De Vaux mentions four Antigonus Mattathias coins at Qumran in his Schweich lectures in 1959, compared to his mention of two Antigonus coins in 1954, which becomes three in 1957. The four Antigonus coins in his Schweich lectures of 1959 equal the number of secure Antigonus coin identifications given in Humbert and Chambon 1994. The four in Humbert and Chambon are the four known to de Vaux at the time of his Schweich lectures.

That is, de Vaux knew, already in 1959, that one of those four Antigonus coins was found underneath the floor of locus 2 next to the very scroll jar which he had earlier told the world was evidence of First Revolt scroll deposits. But de Vaux did not disclose this. Publicly, de Vaux referred to these four Antigonus coins in his Schweich lectures as indicating Period Ib activity at Qumran. He never disclosed that one of those four Antigonus coins, which he interpreted as from Period Ib, was found underneath the floor of locus 2 next to the buried scroll jar—or how this affected his thinking.

The other of the two coins found underneath the floor next to the locus 2 jar is identified in Humbert and Chambon 1994 as an Augustus procurators coin, c. 6 CE. As noted, it is difficult to be certain that the event of installation of the jar can be associated with these coins, since it cannot be excluded that one or both of the coins could have fallen through after the jar was in the floor. Nevertheless, the Antigonus coin found under the floor with the locus 2 scroll jar and de Vaux’s reticence on this point are curious.


Dating of scroll jars

The locus 2 "scroll jar" was only the first of a number of this kind of jar found at Qumran in following excavation seasons. This raises the larger question: What are the correct datings of these so-called "scroll jars" found at the site of Qumran? De Vaux’s excavation records from Qumran remain not yet fully published, and this has left many questions and areas of dispute concerning the archaeology of Qumran. However, two years ago a significant breakthrough occurred with the publication of the pottery of the Netzer excavations at Jericho by Rachel Bar-Nathan, in 2002 (Hasmonean and Herodian Palaces at Jericho. Final Reports of the 1973-1987 Excavations. Volume III: The Pottery [Jerusalem: Israel Exploration Society, 2002]). This publication is extremely important for interpreting the finds at Qumran since Jericho is so close to Qumran, the same pottery was found at both Jericho and Qumran, and the procedures followed by the Jericho excavators were more professional and accurate than de Vaux’s excavations of the 1950’s, particularly concerning stratigraphy. In fact—this is a method statement—these reasons mean that in most cases Bar-Nathan’s Jericho volume is more accurate and reliable for dating Qumran pottery than de Vaux’s reports of the excavations of Qumran.

In Bar-Nathan’s volume, the Qumran-type "scroll jar" is Jericho Storage Jar Type 2B. Bar-Nathan reports only one of this type of jar found at Jericho: it was found in a context dated by the excavators 31-15 BCE. This date of this only example of a Qumran-type "scroll jar" reported from the Netzer excavations at Jericho is contemporary with the end of Qumran’s Period Ib.

Date of the end of the Qumran Period Ib.

Traditional (early de Vaux; Milik; Charlesworth): 40-37 BCE

Later de Vaux: 31 BCE

Bar-Nathan: c. 15 BCE

Magness: c. 8-4 BCE

The exact date of the end of Qumran’s Period Ib is disputed. There are currently four competing leading theories in the scholarly literature. Each of these dates has arguments in its favor, but none seems to be decisive or certain on the basis of present information. In any case, nothing in Bar-Nathan’s Jericho data suggests that Qumran-type scroll jars existed at Jericho as late as Qumran’s Period II in the 1st century CE.


Other Dead Sea sites

In agreement with the picture at Jericho, no Qumran-type scroll jars are known as late as the first century CE at Ein Feshkha, Ein Gedi, Ein el-Ghuweir, or indeed any other Dead Sea location, from published reports of excavators. However, there is an unverified rumor concerning scroll jars at Masada that requires comment.

No published report from the Masada excavations has reported scroll jars at Masada. However Jodi Magness’s 2002 book, The Archaeology of Qumran and the Dead Sea Scrolls, refers to Qumran-type "scroll jars" found at Masada in First Revolt contexts. Magness derives this from an unpublished master’s thesis of Bar-Nathan in Hebrew done in 1988. Magness does not quote an exact citation, and the source referred to, since it is unpublished, is unavailable for verification in libraries. Magness notes that "none [of these Masada jars] is illustrated and they are otherwise unpublished." This kind of argument from hearsay is impossible to evaluate; if such jars are published, scholars can evaluate them at that time. In fact, this claim is missing in Bar-Nathan’s published 2002 volume on the Jericho excavations, and this is notable since Bar-Nathan’s 2002 volume is a more developed version of the M.A. thesis. Bar-Nathan does, however, refer in the 2002 volume to pottery at Masada from the time of Herod the Great, secondarily used by Zealots at the time of the First Revolt. (Is that what is behind the rumor of First Revolt "scroll jars" at Masada?) In any case, Qumran-type "scroll jars" at First Revolt Masada are not supported by published reports of the Masada excavations or by anything in Bar-Nathan’s published Jericho volume.

(For the sake of completeness, a cylindrical jar like the Qumran "scroll jars" was also reported to have been found in a 2nd-century CE context in a cemetery way up at Quailibah [Abila] in Transjordan, far to the north of the Dead Sea. This jar is not drawn or illustrated, and its present whereabouts are unknown. It is not clear what conclusions can be drawn relevant to Qumran from this. Compare potsherds from the Iron Age found in a Roman period context in the Qumran cemetery.)

| Page 2 | Page 3 | Page 4 |