Many studies on the formation of the collection of the Twelve Minor Prophets take for granted that the 4QXIIa manuscript provides evidence that the Book of Jonah stood at the end of a scroll of the Twelve and that this sequence could have been the original one. Examination of the reconstruction of the scroll published in the DJD XV volume reveals that the Malachi-Jonah sequence is highly hypothetical and should not be considered as firm evidence.
In 1988, a Harvard thesis by Russell Earl Fuller presented the first critical edition of six manuscripts of the Minor Prophets from Cave 4 near Qumran, concluding confidently that these remains of scrolls originally contained all twelve Minor Prophets and that 4QXIIa preserves a unique sequence with Jonah following Malachi.1 At the time, there was no way to evaluate Fuller’s conclusions since the section of his thesis dealing with 4QXIIa nowhere discusses how the Malachi and Jonah texts join. The relevant fragments are reproduced on two different plates (Pl. A.2 and A.3) with no attempt to physically join Jonah after Malachi.
A thesis by Barry Jones concluded that the Malachi-Jonah sequence reflects the original position of the Book of Jonah in the Minor Prophets (XII). The argument is based on the manuscript published by Fuller which Jones describes as apparently exhibiting a previously unattested sequence of books where Jonah most likely followed Malachi as the final book of the scroll.2
A year later, Odil-Hannes Steck integrated the Malachi-Jonah sequence preserved by 4QXIIa into a discussion on the origin of the Book of the Twelve. Steck checked the photographs in March 1995 with Qumran specialists in Göttingen and refers to a letter by A. Maurer confirming that the end of Malachi on col. IV was followed by the beginning of Jonah on column V.3 Since 4QXIIa is one of the oldest witnesses of the Minor Prophets, dated to the second century BCE, it was legitimate to suggest that Jonah, a most unusual narrative within the anthology, was originally appended to the collection of Minor Prophets, after Malachi.
The same year, Fuller published an article based on a paper presented at the SBL Consultation on the Formation of the Book of the Twelve, offering a convenient overview of the evidence provided by the most ancient manuscripts.4 Fuller is more cautious than he was in his thesis, writing that around 150 BCE, 4QXIIa&b seem “to confirm that the collection of the XII is complete” and may “preserve the unique transition/order Malachi—Jonah”. The transition is deemed “uncertain” and a question mark is added to the Malachi—Jonah (?) transition.5
Yet, in the same volume, James Nogalski presses the case for a single Book of the Twelve. He affirms that the “ancient traditions irrefutably establish that the writings of the twelve prophets were copied onto a single scroll and counted as a single book from at least 200 BCE. Naturally, one presumes that someone intended these twelve writings to be read together.6” Within a few pages, the existence of the XII becomes irrefutable and moves back 50 years. The following challenges the validity of the reconstruction in order to temper the excessive confidence with which it is widely claimed that originally, Jonah was set at the end of the XII.
The evidence for the connection between Malachi and Jonah was finally provided when Fuller’s work was integrated into the Discoveries of the Judean Desert series.7 Among the 4QXIIa fragments, the end of the Book of Malachi is well preserved with a column bearing 40% of the text of Mal. 3.14–24 on four large fragments with top, left and bottom margins. Thanks to these margins and to the fact that only one line is totally missing, the reconstruction of the end of Malachi is undisputable. This column was clearly not the final column of the scroll since the left side of frg. 9 (Mal. 3.19–21) transmits three letters of the beginning of two lines of the next column (see below Fig. 1).8 Thanks to the preservation of column IV, there is no doubt that these lines correspond roughly to lines 10–11 of column V. The question is what came after Malachi in 4QXIIa ?
It was reasonable for Fuller to postulate that Jonah and Malachi were closely connected since all but one of the 23 fragments attributed to 4QXIIa transmit parts of Jonah and Malachi. Hence, Fuller managed to establish a physical join between frgs. 9 and 15, linking the end of Malachi and the beginning of Jonah (Fig. 1). The physical joint is very narrow (5 mm) and there are no letters or marks in the leather running directly from one fragment to the other to confirm the validity of the joint. Moreover, Plate XLI shows a marked difference in the colouration of the leather. The left hand side of frg. 15 comes out as dark as frg. 9 but the right hand side of frg. 15 that joins with frg. 9 is almost transparent. I am not in a position to evaluate the significance of such difference in colour. It may be due to the way the two fragments were photographed and thus may not be significant. It is important to remember that the image is computer generated which is not exactly the same as a physical joint. In fact, the joint may be less perfect than it looks since the scale indicated on Plate XLI only applies to frgs 11–18. Since no scale is provided for frgs 7–10, it is possible that the fragments of the last column of Malachi were reproduced on the same plate at a different scale than the fragments of the first column of Jonah.
Fig. 1: Joint between frgs 9 and 15 with first letters of lines 10–11
(from Plate XLI)
Fuller logically seeks confirmation of the joint from the reconstruction of the column bearing Jonah 1. The task is, however, daunting as Jonah 1 is in a much more fragmentary state than Malachi. The text of Jon. 1.2–5,7–8 is preserved on eight small fragments which transmit no more than 30% of the text (Fig. 2 below). There is no doubt that frg. 15 belongs to the initial column of Jonah but it is necessary to verify that its position at the level of lines 9–11 of the previous column bearing the end of Malachi is acceptable in order to confirm the Malachi-Jonah joint. At this point, Fuller faces insuperable difficulties entailed by the poor state of preservation of the Jonah 1 column. The length of the lines of this column is indeed attested by frgs 17 and 18 which preserve portions of the bottom and left margins of the column. Unfortunately, there is no physical contact between the end of the column and the fragments above. The crucial frg. 15 is separated from the fragments transmitting the bottom of the column by a vertical gap of at least four lines which are entirely missing. Fragments 11–14 bearing elements of Jon. 1.2 are also “floating” and bear no trace of margins that would have helped positioning them. Hence, rather than confirming the validity of the Malachi-Jonah joint, Fuller’s placement of frg. 15 presupposes its validity. Methodologically, there is nothing wrong with the procedure as long as one remembers that the joint is a working hypothesis requiring confirmation. The confirmation is, however, impossible since, in addition to the vertical gap of four lines, there is also a horizontal gap of 4–5 letters that separates the beginning of the two lines on frg. 9 and their purported continuation on frg. 15. This gap prevents confirming the alignment of the lines suggested by the physical join and cannot dispel doubts over the validity of such a narrow point of contact between the end of Malachi and the beginning of Jonah.
The difficulty of establishing the validity of the “physical” joint between Malachi 4 and Jonah 1 is illustrated by the evolution of Fuller’s reconstructions. At first, in his thesis, Fuller read a ל at the beginning of what he then called line 6 on frg. 9ii and a clear ה followed by a possible ט at the beginning of the following line. Then, in the DJD edition, Fuller modified the reading of the initial letters, the line numbering and the verse distribution because he realized that the text of Jonah would better fit column V if Jonah did not start immediately at line 1 of the column but at line 3, the blank lines indicating the beginning of a new book.9 In spite of the modification, the photograph published as Plate XLI does not entirely support the reconstruction:
|לׄ[בוא||Fuller’s thesis, Fig. A.6 line 6 (Jon. 1.3)|
|ו[ייר]א֯ו֯[ המ]לחים ויזעקו[||DJD XV, p. 229 line 10 (Jon. 1.5)|
|וי֯[ ]לח֯ים ויזעקו [||My reconstruction on the basis of Plate XLI|
To shore up the connection between Malachi 3 and Jonah 1, DJD XV suggests the presence of aleph and waw (]א֯ו֯[) on line 10 as part of the first words of Jon. 1.5 (המלחים וייראו). However, the photograph (redrawn as Fig. 1 above) shows the trace of a second letter after the initial waw on frg. 9 line 10 but then the lacuna extends all the way to the lamed on frg. 15 with no material to bear traces of the purported א֯ו֯. There seems to be the shade of the top of two letters on the right side of frg. 15 just before the joint with frg. 9 but they are too small and too faint to support the reconstruction of א֯ו֯, and they are too high to belong to line 10 as it is reconstructed. I have thus not drawn them on Fig. 1. Hence we are left with the warning in the introduction of the DJD XV volume which suggests that parts of letters, especially around the manuscripts, may have since broken off.10 The photograph on PAM 43.09911 does not support the purported א֯ו֯ which would have partly filled the crucial lacuna between frgs. 9 and 15 and would have considerably consolidated the Malachi-Jonah sequence. At line 11, Fuller reconstructs הכ֯[ל]י֯ם אשר באניה[ but Plate XLI only shows the top half of the final mem of הכלים while there is no trace of the previous yod. No firm conclusions can be drawn on the sole basis of the photograph, but the DJD edition shows that the Malachi-Jonah sequence is based on a maximalist reconstruction of the evidence.
Deprived of a firm basis, the rest of the reconstruction of Jonah 1 is necessarily hypothetical. There is no physical evidence of the top margin of column V that would have indicated at which line Jon. 1.1 started and would have greatly helped the reconstruction of the beginning of Jonah. Since frg. 8 does attest the top margin of column IV (see Fig. 2 below), Fuller postulates it for the next column thanks to the joint at lines 10–11. According to Fuller’s reconstruction, the first line of Jonah 1 (line 3 of the column) is attested by the bottom parts of four letters (two on the photograph) at the top of frg. 11.
Fig. 2: Reconstruction of the top of Col. V with purported blank lines
(from Plate XLI)
It is impossible to identify these letters and thus equally impossible to be sure that this fragment transmits Jon. 1.1. The same fragment (frg. 11) bears three readable letters belonging to the following line. Therefore Fuller places frg. 11 next to frgs 12, 13 and 14. However, they do not fit well and Fuller has to postulate a hypothetic flaw on the leather to explain the presence of an unusually large space on line 4 between העיר on frg. 12 and הגד[ולה on frg. 11 to reconstruct the words “the great town” from Jon. 1.2.12 There is no doubt that frgs. 12, 13 and 14 transmit letters from Jon 1.2–3, but the small size of frg. 11 (1.5 x 1.5 cm), its lack of fit with the adjacent fragments and the necessity to postulate a flaw in the leather suggest that the three legible letters (הגד) on frg. 11 could just as well belong to other occurrences of the adjective הגדול. Hence frg. 11 could be part of Jon 1.12 and would easily fit in the 7–9 missing lines of Fuller’s reconstruction of the next column. Or, frg. 11 could transmit a piece of Jon 3.2 and would then fit next on the right of frg. 22. This means that there is no evidence of the first line of Jonah and no way to know on which line of the column it started. It is reasonable to postulate that it started at the top of a column but whether or not it was preceded by a blank space is beyond the evidence. Considering that the top margin on frg. 8 of column IV provides an indication of a top margin for column V is entirely dependant on the precarious connection of Jonah 1 with Malachi 3 at lines 10–11. The question of the top margin of col. V provides no confirmation for the Malachi-Jonah sequence. Another problem with Fuller’s reconstitution is that it requires verse 3 to be shorter than MT’s.13 For this reason, Heinz-Josef Fabry considers Fuller’s reconstruction most unlikely.14 In fact, 4QXIIa does transmit variants which produce a shorter text in Jonah.15 In the text below, I added in shadow letters the words Fuller omitted from his reconstruction to show that in fact only the words תרשישה עמהם are clearly in excess of the average length of the bottom lines of the column:
|קום2 ויהי דבר יהוה אל יונ]ה֯ ב֯ן֯[ א]מ֯[תי לאמר1||11?|
|הגד[ו]ל̇[ה וקרא עליה flaw לך אל נינ]ו֯א העיר||11?,12|
|ויקם̇ י֯ו֯נ֯]ה לברח3 כי עלתה ר]ע֯ת֯ם לפני[||13,14|
|תרשי]ש מלפני יהוה [ויר]ד̇ יפא וי[מצא||13,14|
|אניה באה תרשיש ]וֹיתן שכ̇ר֯[ה וירד בא לבוא||14|
|וי]ה֯ו֯ה֯ ה֯[טיל רוח גדולה אל הים4 עמהם תרשישה מלפני יהוה||15|
|וי]הי סער גדול בים[ והאניה חשבה להשבר||15|
|ו[ייר]א֯ו֯[ המ]לחים ויזעקו[ איש אל אלהיו ויטלו את5||15|
|הכ֯[ל]י̇ם אשר באניה[ לים להקל מעליהם ויונה||15|
The words תרשישה עמהם can be removed and verse 3b still makes sense: “He went down (to) Jaffa. He found a boat going to Tarshish. He gave its fare. He went down in it to go away from in front of Yhwh.” Therefore, the necessity for verse 3 to be shorter than MT’s is not enough to invalidate Fuller’s reconstruction. However, postulating a variant is no evidence, only a possibility.
Hence, the only certainty about col. V is that frgs 17–18 (Jon 1.7–8) belong to the bottom of a column since the bottom and left margins are preserved.16 The gap of at least four lines between line 11 and line 17 precludes using the firm base of the bottom of the column to support the reconstruction of the joint at lines 10–11 (Fig. 3).
Fig. 3: Fuller’s positioning of the fragments of Columns IV and V
(from Plate XLI)
Finally, Fuller claims that the greater fragmentation of Jonah compared to Malachi supports the Malachi-Jonah sequence.17 The idea is that the beginning of the 4QXIIa scroll which found itself on the most external parts of the rolled-up scroll would have deteriorated first and only the end of the scroll remains with its final columns. Statistically, however, about half of the fragments of the 800 scrolls from the caves come from the middle of the scrolls because the ends, close to the inner hole of the scrolls, were also subject to more deterioration than the scrolls’ middle parts.18 Hence Jonah’s greater fragmentation can either indicate that it belonged to the end of the scroll, or that it came before Malachi on a scroll recording only a few prophetic books. Fuller provides no diagram of the position of fragments for 4QXIIa, presumably because, apart from the stitches holding col. II and col. III together which were separated in the course of photographing, the scroll was already made up of loose fragments when it was bought.19 This adds an extra layer of uncertainty and provides no confirmation that Jonah was the last book of the 4QXIIa scroll.
In a recent reconsideration of the textual evidence from the Dead Sea Scrolls, Georges J. Brooke warns that the Malachi-Jonah sequence has been too easily accepted as factual. Brooke claims that the reconstruction in light of the damage pattern strongly suggests that the end of the scroll was on the outside when the scroll was put in the cave and that it is far from certain that there would be enough room on this scroll to preserve all twelve Minor Prophets. Brooke also claims that Malachi and Jonah, in that order, belonged to the middle of the collection, or that the manuscript contained some rather than all of the Twelve.20
The possibility that the scroll was rolled with the end out rather than inside invalidates any argument based on the comparative states of preservation of Malachi and Jonah. In fact, it is not even sure that the Jonah and the Malachi fragments attributed to 4QXIIa belonged to the same scroll. Cave 4 is the most problematic of all the Qumran caves since it is in fact constituted of two caves, 4a and 4b. There is no way to know from which side fragments came from and whether a given set of fragments classified as belonging to the same scroll were actually collected from the same side of Cave IV.21 The number of actors involved in the recovery and the publication of the fragments and the long time it took to complete the entire process multiplies the uncertainty as to the accuracy of the allocation of fragments to individual scrolls. Moreover, 4Q76 has been renamed 4QXIIa on the presumption that the handful of letters on frg. 1 (Plate XL) are part of Zech 14.18. In his 1996 article, Fuller includes Zechariah in the contents of 4QXIIa but eight pages later he does not.22 The words לא and אשר לא on frg. 1 could come from Jon 4.10–11 since the fragment is unconnected and bears no evidence that it belongs to the postulated Column I. Attributing this fragment to Zechariah was dictated by the presupposition that a scroll containing two Minor Prophets is a scroll of the XII. Brooke mentions a fragment of Joel 4.1–4, supposedly coming from Cave 4, and 5QAmos, containing only Amos or Amos at the beginning of a collection.23 There is thus a considerable amount of unconnected fragments of the Minor Prophets that calls for a reconsideration of the 4QXII phenomenon.
As far as 4QXIIa is concerned, it is sure that Malachi 3 is not the end of the scroll. That the Book of Jonah was copied by the same hand on leaves of a similar size as the ones used for Malachi is also sure. But the point of anchorage of Jonah 1 after the last Malachi column cannot be confirmed. The column transmitting the beginning of Jonah bears no evidence of a top margin and the gap of at least four lines between the joint and verse 8 preclude all confirmation of the Malachi-Jonah sequence with the aid of the textual reconstruction. It is thus impossible to affirm that Jonah came after rather than before Malachi. If Malachi and Jonah were copied on the same scroll, a point over which a considerable amount of uncertainty remains, it is not sure that this scroll contained other books of the Minor Prophets. The renaming of 4Q76 as 4QXIIa is thus misleading. Something like 4QJon, Mal-? would be preferable if it was clear that the position of Jonah and Malachi relative to each other remains entirely open.
Brooke’s 2006 article signals the end of the period of euphoria and the dawn of a more careful evaluation of the evidence concerning the collection of the Twelve Minor Prophets. Brooke does not challenge the Malachi-Jonah sequence but the above evaluation of the evidence indicates that the hypothesis is far too fragile to be used as evidence of a scroll of the XII ending with Jonah. The hypothesis of a Malachi-Jonah sequence was too quickly adopted as factual. Ben Sira 49.10 remains the earliest mention of a collection of Twelve Minor Prophets with no indication of the sequence of the individual books. The notion of a widely accepted literary unit already in the second century BCE is not supported by the evidence. That Ben Sira referred to the twelve prophets as the Twelve around the time of the production of 4QXIIa&b is insufficient to claim that other scholars considered the Twelve as a unit and that all twelve Minor Prophets were already copied together onto single scrolls in Palestine. The words of Ehud Ben Zvi should be heeded. The most complete manuscripts of the Minor Prophets dated before the turn of the era are 4QXIIg and 8HevXIIgr with eight and seven (six attested!) Minor Prophets respectively. Before the turn of the era, the Twelve constituted no more than an anthology gathered in a somewhat flexible order, which later on became fixed.24 A shift by a century or two of the period when the writings of the twelve prophets were copied onto a single scroll and counted as a single book is bound to have major consequences on the reconstruction of the formation of the collection since most studies of the subject give too much credence to hypothetical reconstructions of Dead Sea Scrolls and have embraced the 4QXII designation uncritically.25 I will deal with the larger 4QXII issue in a coming article.
 R. E. Fuller, The Minor Prophets Manuscripts from Qumran, Cave IV (PhD Diss., Harvard 1988; published Ann Arbor: UMI, 1995), pp. 151–2.
 B. A. Jones, The Formation of the Book of the Twelve (Atlanta: Scholars, 1995), p. 6.
 O. H. Steck, ‘Zur Abfolge Maleachi-Jona in 4Q76 (4QXIIa)’, ZAW 108 (1996), pp. 249–53 (249).
 R. E. Fuller, ‘The Form and Formation of the Book of the Twelve’, in J. W. Watts & P. R. House (eds), Forming Prophetic Literature (JSOTSup, 235; Sheffield: Academic Press, 1996), pp. 86–101.
 J. A. Nogalski, ‘Intertextuality and the Twelve’, in J. W. Watts & P. R. House (eds), Forming Prophetic Literature (JSOTSup, 235; Sheffield: Academic Press, 1996), p. 102.
 R. E. Fuller, ‘The Twelve’, in E. Ulrich, F. M. Cross, R. E. Fuller, J. E. Sanderson, P. W. Skehan & E. Tov (eds), Qumran Cave 4.X DJD XV (Oxford: Clarendon Press, 1997), pp. 221–318 and plate XL-XLIII.
 Jones, Formation, p. 6, quoting R. E. Fuller, The Minor Prophets Manuscripts from Qumran, Cave IV (PhD Diss., Harvard 1988; Ann Arbor: UMI, 1995), p. 5.
 Published in R. H. Eisenman & J. M. Robinson (eds), A Facsimile Edition of the Dead Sea Scrolls (Washington DC: Biblical Archaeology Society, 1991), plate 1183. PAM 42.629, apparently an earlier photograph of PAM 43.099 mentioned in the index volume DJD XXXIX, p. 46 is not published in Facsimile.
 Note that Col. V line 4 is based on frg. 11 as well as frg. 12 contrary to what is indicated in Fuller, DJD XV, p. 229.
 H.-J. Fabry, ‘Die Nahum- und Habakkuk-Rezeption in der LXX und in Qumran’, in E. Zenger (ed.), „Wort JHWHS, das geschah…“ (Hos 1,1) (Freiburg: Herder, 2002), pp. 159–190 = H.-J. Fabry, ‘The Reception of Nahum and Habakkuk in the Septuagint and Qumran’, in S. M. Paul, R. A. Kraft, L. H. Schiffman & W. W. Fields (eds), Emanuel. Studies in Hebrew Bible, Septuagint and Dead Sea Scrolls in Honor of Emanuel Tov (VTS, 94; Leiden: Brill, 2003), pp. 241–256 (245).
 S. A. Reed, ‘Find-Sites of the Dead Sea Scrolls’, DSD 14,2 (2007), pp. 199–221.
 E. Ben Zvi, ‘Twelve Prophetic Books or “The Twelve”: a Few Preliminary Considerations’, in J. W. Watts & P. R. House (ed.), Forming Prophetic Literature (JSOTSup, 235; Sheffield: Academic Press, 1996), pp. 125–156. M. Beck, ‘Das Dodekapropheton als Anthologie’, ZAW 118 (2006), pp. 558–583.