Alternative Psalms Manuscripts: 4QPs-q

For previous posts in this series, see here.

4QPsq (4Q98)

• Date: Turn of the Era
• Preserved contents: Ps 31:24–25; 33:1–18; 35:4–20
• Alternatie sequences: 31→33 (cf. 4QPsa)

I cannot find a picture of this scroll online, so I've had to resort to scanning an image from DJD 16 (plate XIX). So the picture doesn't look great partly because it's just a scan from a printed book, but also "Towards the right side of the scroll the surface has flaked off, evidently because of stitching on the next revolution, and worm-holes are clearly visible" (DJD 16, p. 145).

A single fragment of this scroll has survived.

Scan of 4QPsq from DJD 16 (plate XIX).
Red line = Ps 31:25; Purple line = vacat; Yellow line = superscription;
Orange line = Ps 33:2–3; Green line = vacat after Ps 33:12

The red arrow in the above scan shows the end of Psalm 31, followed by a vacat at the beginning of the next line (indicated by the purple arrow), and then a superscription at the end of that same line (indicated by the yellow arrow). Psalm 32 has a superscription in the MT (לדוד משכיל), but it does not match the superscription in this scroll (לדויד שיר מזמור).

In the third line (see orange arrow), we can see the last words of Ps 33:2 (בכנור בנבל עשור זמרו לו) and the first word of v. 3 (שירו). So, our scroll does not contain Ps 32 in the spot where it appears in MT, but rather has put Ps 33 immediately after Ps 31. Moreover, whereas Ps 33 has no superscription in MT, our scroll contains a superscription for this psalm. The same sequence (31→33) is also found in 4QPsa, but this latter scroll does not contain a superscription for Ps 33.

Another interesting feature of this fragment is that there is a vacat after Ps 33:12 (green arrow in the above scan), apparently indicating that Ps 33:13–18 forms a separate poetic composition (DJD 16, p. 148; Yarchin 779–80).

Col. ii contains material from Ps 35.

Alternative Psalms Manuscripts: 4QPs-n

This is the fifth post in my series. See previously here, and follow the links back.

4QPsn (4Q95)

• Date: late-I BCE
• Preserved contents: Ps 135:6–9, 11–12; 136:23–24
• Alternative sequences: Ps 135:11–12→136:23–24

There are three fragments of this scroll.

4QPsn frg. 1

This fragment preserves part of Psalm 135:6–8, with a longer (vis-a-vis MT) reading of v. 6 mentioned earlier in relation to 4QPsk. The first line shows לעשות יעש[ה, i.e., "to do he does" at the very beginning of the plus. The second line shows the last word of v. 6, תהו]מות, and the first words of v. 7, מעלה נשיאים, "raising clouds."

4QPsn frg. 2

Fragment two has a little bit of text from Psalm 135:11–12. The first line has מלך הב[שן from Ps 135:11 and the second line has נחל[ה ]ל[נו from v. 12.

4QPsn frg. 3. 

This third fragment contains part of Psalm 135:12 and part of Psalm 136:23–24. At the very top, we can see ונתן, which is the very first word of Psalm 135:12. The DJD editors also think they can see on the same line the next word of the verse: ארצם, which they can put together with the נחל[ה ]ל[נו of frg. 2. The next line of this fragment preserves לישראל עמו כ[י, which are the last words of Psalm 135:12 in the MT, except for the כי, which is an extra element. DJD proposes that this is the first word of the phrase כי לעולם חסדו ("for his lovingkindness is everlasting"), taken over from the refrain of Psalm 136.

The third visible line on frg. 3 shows the words זכ[ר] לנו כי לעו[לם, corresponding to Psalm 136:23 (I myself can't see any of the זכר.) That lamed at the bottom of the fragment plausibly corresponds to the לעולם in the refrain of Ps 136:24.

According to DJD 16 (p. 136), the combination of Ps 135 and 136 happened in two ways:

(a) by the introduction of the refrain that is characteristic of Psalm 136 (כי לעולם חסדו) after 135:12a and again after 135:12b; and (b) by the colon 135:12b (נחלה ]לישראל עמו) in line 3, which is very similar to the colon found in 136:22a (נחלה לישראל עבדו). 

And on p. 137:

The preserved text represents a new Psalm, which forms a coherent whole and presumably comprised 135:1–12 + 136:23–26. But since the refrain suddenly appears from v 12 onwards, it seems that the compiler fashioned the new Psalm by combining material from Psalms 135 and 136. Strictly speaking, the transition is from 135:12, which ends a pericope, to 136:23, which begins a new one. However, by introducing the refrain twice into 135:12 and in view of the close similarity between 135:12ab and 136:21–22, the compiler has succeeded in blending material from Psalm 135 with that of 136 at points where the separate Psalms contain very similar readings. 

Alternative Psalms Manuscripts: 4QPs-k

This is the fourth post in my series on Qumran Psalms manuscripts that feature psalms known from the MT in sequences out-of-step with the MT. Previous posts covered 4QPse, 4QPsb and 4QPsd, and 4QPsa. This post covers 4QPsk.

4QPsk (4Q92)

• Date: first half of I BCE
• Preserved contents: 135:6–16; 99:1–5(?)
• Reconstructed sequence: 135; 99 (with, perhaps, another psalm in between)
• Alternative sequences: 135; 99

4QPsk (entire preserved contents) 

On that piece of twine connecting the two fragments, DJD 16 (p. 123) says:

...the two fragments are connected by a coarse piece of cord which extends upward through a hole just below the he of העמים in frg. 1 ii 4, enters frg. 2 through a hole from behind, and emerges just below the waw of לדו[ד in line 3. On PAM 43.030, frg. 2 is at a perpendicular angle above fr. 1, while PAM 42.029 documents a partially successful attempt to align the pieces correctly. The artificial join seems to be the remains of a repair undertaken in antiquity. 

The first column (on the right) contains material from Psalm 135:6–16. There's a little bit at the very top right of the fragment, more easily seen in this infrared photo: 

4QPsk col. i (top)

Those little lines at the point are plausibly reconstructed as the yod and final nun in the word ואין, which comes in the middle of a reconstructed form of Psa 135:6 that matches what we find for this verse in 11QPsa and 4QPsn (and does not match the MT reading of this verse). In Sanders's Cornell edition of 11QPsa, this verse reads in English:

What the LORD pleases he does, in heaven and on earth, to do he does; there is none like the LORD, there is none like the LORD, and there is none who does as the King of gods, in the seas and in all deeps. (p. 61)

(The italics mark divergences from MT. I have underlined the portion that corresponds to the proposed reconstruction of 4QPsk.)

The first words you can really see on the fragment (in what is actually the second line of preserved writing) constitute the very end of Psa 135:7 (מאצרתיו) and beginning of v. 8 (שהכה בכורי). The last visible words in that same column are the final words of v. 15 (וזה]ב מעשי ידי אדם) and the first word of v. 16 (פה), followed by the bottom margin of the scroll.

As for the second column, here's the picture that shows the "partially successful attempt to align the pieces correctly," mentioned in the DJD quotation above.

4QPsk  col. ii

Apparently those first visible letters above the coarse cord in the second column are the remnants of a superscription attached to Psalm 99 (the name David, לדוד), though MT contains no superscription for this psalm. On the second line we have the word העמים (= the last word of MT Psa 99:2). Remember, the fragments are not properly aligned; you can see the he on frg. 1 and the mem on frg. 2, preceded by the ayin (partially visible) and part of the top stroke of the he. After the mem on frg. 2 you can see the yod, but the cord is covering the final mem which is visible (upside down) on this picture:

4QPsk frg 2 (upside down)

On the third line of col. ii we have רממו, followed by the first two letters of the Tetragrammaton; these words are the beginning of Psa 99:5.

The identification of the words in this column with Psa 99 depends on the rarity of the term רממו, which in the MT appears only in Psa 99:5, 9 (spelled both times רוממו). The word also appears in Psa 135:2 in 11QPsa, but this would leave the העמים in the preceding line unexplained.

Based on this evidence, Lange (p. 387) can say: "There is no question that Ps 99 in 4QPsk followed Ps 135. The material reconstruction of 4QPsk makes it probable that between Ps 135 and Ps 99 in this manuscript there stood another psalm: Ps 135→Ps ?→Ps 99."

Alternative Psalms Manuscripts: 4QPs-e (4Q87)

This is the third in a series of posts on the Psalms manuscripts from Qumran that contain psalms known from the MT Psalter but in a sequence that does not correspond to the MT. Today's post covers the fourth of these manuscripts, 4QPse.

4QPse  (4Q87)

• Date: mid-I CE
• Preserved contents: 76:10–12; 77:1; 78:6–7, 31–33; 81:2–3; 86:10–11; 88:1–5; 89:44–48, 50–53; 103:22(?); 109:1(?), 8(?), 13; 114:5; 115:15–18; 116:1–4; 118:29; 104:1–3, 20–22; 105:1–3, 23–25, 36–45; 146:1(?); 120:6–7; 125:2–5; 126:1–5; 129:8; 130:1–3, 6. 
• Reconstructed sequence: Ps 76→77; 78; 81; 86; 88; 89; [103?]→109(?); 114; 115→116; [118(?)]→104[→147(?)→]105→[146(?)] or [106(?)]; 120; 125→126[→127→128→]129→130
• Alternative sequences: [103?]→109(?); [118(?)]→104[→147(?)→]105→[146(?)] or [106(?)]

The supposed join between Ps 103 and 109 is preserved on frg. 9. I cannot find a picture of this fragment online. The photo (plate XI) in DJD 16 shows that it is exceedingly small, and the reconstruction on p. 79 shows that it preserves parts of only three words.

Here's a (not great) scan of the fragment.

scan of DJD 16, plate XI, frg. 9

The second line contains the words לדו]יד מזמו[ר, clearly a superscription, attributed by DJD to Ps 109, because "[o]nly three Psalms that fall within the preserved range of this scroll (76–130) have a superscription containing the two words preserved in this fragment," those being Psalms 101, 109, and 110, but on the other fragments of this scroll we have portions of Psalm 109 but not the other two. Obviously, this reconstruction is very speculative, especially in light of the fact that the DSS sometimes have superscriptions where the MT does not (as proposed later for this very scroll--see on Ps 104--and, e.g., in 4QPsk, to be discussed in a later post).

In the first line of frg. 9, only a few ink traces are visible.

While these traces are somewhat consistent with the bottom tip of nun and the waw in צרינו at the end of Psalm 108, they are better read as the two legs of the he in [the Tetragrammaton] which ends Psalm 103. The combination of Psalm 103→109 is also the most likely sequence in 11QPsa. (DJD 16, p. 79) 

The connection to 11QPsa is pretty important for the DJD reconstruction of the entire scroll of 4QPse. Here's a passage from DJD 16, pp. 74–76: 

Although very fragmentarily and badly preserved, 4QPse appears to be textually affiliated with the Psalter represented by 11QPsa (and 11QPsb) since the preserved pieces share some features with the large Cave 11 scroll. 4QPse contains several differences in arrangement from the received text; most notably, Psalm 104 in frgs. 14–16 cannot be followed by Psalm 105 in frgs. 17–24 given the location of the two column tops on frgs. 15–16 and frgs. 18 ii, 20–24.

And here are the relevant fragments.

4QPse frgs 15–16 (Ps 104:20–22) 

4QPse frg 18; col. i = Ps 105:3; col. ii = Ps 105:36–45 + 146:1(?)

4QPse frg 20; Psa 105:36–40

It's fragment 20 here that actually preserves the column top, along with frgs. 15–16. According to DJD 16 (p. 81): "A text of approximately 15 lines would be required between the two compositions [Ps 104 and 105]; the only ancient source incorporating such a text is 11QPsa, where Psalm 147 is placed between Psalms 104 and 105." 

As for Ps 146 rather than Ps 106 following Ps 105 on frg 18--the final word visible on frg 18 is a hallelujah, immediately preceded in the previous line by the end of Ps 105 (v. 45). While DJD 16 (p. 82) admits that this hallelujah could be the one found at the beginning of Ps 106 in MT, "it more likely denotes the beginning of Psalm 146 (as originally in 11QPsa)."

Now, back to the beginning of this sequence. The DJD reconstruction identifies Ps 118 as preceding Ps 104. This idea is based on the evidence of frg. 14.

4QPse frg. 14. 

This fragment contains 4 lines of text. Lines 2–4 correspond to Ps 104:1–3. In line 2, the first letter that is visible (at least to my eye and based on the above photo) is the yod in ברכי, which is the first word of MT Ps 104. (The DJD reconstruction proposes a superscription לדויד based on spacing considerations in the fragment. MT does not have this superscription.) The fragment continues on line 2 with נפשי את, then the Tetragrammaton followed by the yod of another Tetragrammaton (and DJD thinks it can also see the first he). So, anyway, we have Psalm 104 here. Line 3 contains almost the end of v. 2, and line 4 shows the very end of v. 3. The top line contains some writing that DJD identifies as טוב כי לע[ולם, and reconstructs the line as the end of Ps 118, "since in the received text this ending for a Psalm occurs only at 118:29, which also precedes Psalm 104 in 11QPsa" (p. 81).

As a general conclusion to the discussion on this scroll, Willgren (p. 100) says: "Although intriguing, the quite extensive reconstructions have been questioned, and other scholars have chosen a more cautious approach, stating that the argument can be neither confirmed, nor refuted" (citing Jain, p. 104).

Alternative Psalms Manuscripts: 4QPs-b (4Q84) and 4QPs-d (4Q86)

This post picks up the theme introduced previously concerning the Dead Sea Psalms Scrolls that preserve alternative sequences (alternative to the MT sequence) of psalms in the MT. There are nine such scrolls, and I've already introduced the first, 4QPsa. Here are the second and third. 


4QPsb (4Q84)

• Date: mid-I CE
• Preserved contents: Ps 91:5–8, 12–15; 92:4–8, 13–15; 93:5; 94:1–4, 7–9, 10–14, 17–18, 21–22; 96:2; 98:4–5; 99:5–6; 100:1–2; 102:5?, 10–29; 103::1–6, 9–14, 20–21; 112:4–5; 113:1; 115:2–3; 116:17–19; 118:1–3, 5–10, 12, 18–20, 23–26, 29.
• Reconstructed sequence of preserved contents: Ps 91[→]92[→]93[→]94, 96, 98, 99[→]100, 102→103[→]112[→]113, 115, 116[→117→]118
• Alternative sequences: 103→112

To focus on that alternative sequence, here's a picture. 

Frag. 25, cols. ii–iv

In this picture, the column on the right (col. ii of the fragment = col. XXV of the scroll) is the end of Ps 103; you can see all of v. 20 and the beginning of v. 21. The next column (col. iii = col. XXVI) contains part of Ps 112; you can see most of v. 4 and the beginning of v. 5 (in the third line). In the next column you can sort of see a hallelujah, presumably the one that begins Ps 113.

4QPsd (4Q86)

• Date: mid-I BCE
• Preserved contents: Ps 106:48?; 147:1–4, 13–17, 20; 104:1–5, 8–11, 14–15, 22–25, 33–35.
• Reconstructed sequence: 106:48?→147→104
• Alternative sequences: 106:48?→147→104

First alternative sequence: 106:48?→147

Frag. 1

This fragment shows a hallelujah at the top, preceded by something. Here's the comment in DJD 16, p. 66: "Halleluyah is preceded by a final letter that extends below the line which cannot be the reš of ודר in Ps 146:10. The only Psalm with such an ending is 106." Of course, this reasoning assumes that we are looking at a psalm that we know from the traditional Psalter. At any rate, what comes next, in the second visible line, is the opening of Ps 147: [הללו]יה כי טוב זמרה אלהינו נא[וה זמרה].

Second alternative sequence: 147→104

Frag. 6

The top of this fragment shows the very end of Ps 147. The second visible line has the hallelujah at the end of the psalm, followed by a vacat on the same line. The next line (third visible line) contains the opening of Ps 104.

Alternative Psalms Manuscripts: 4QPs-a (4Q83)

I thought I would take a moment to lay out some evidence for alternative arrangements of the psalms in the Dead Sea Scrolls. These scrolls present some traditional psalms (traditionally found in the Tanak) in sequences that are non-traditional (i.e., the Tanak has a different sequence). Of the 36 Psalms-bearing manuscripts from Qumran (Yarchin's terminology, 777n4), 9 of them have these alternative sequences. This post covers only the first of these. Subsequent posts will treat the others. 

4QPsa (4Q83)

• Date: mid-II BCE
• Preserved contents: Ps 5:9–13; 6:1–4; 25:8, 10, 12, 15; 31:23–24; 33:2, 4–6, 8, 10, 12; 34:21–22; 35:2, 13–18, 20, 26–27; 36:1, 3, 5–7, 9; 38:2, 4, 6, 8–10, 12, 16–23; 71:1–14; 47:2; 53:2, 4–5, 27; 54:2–3, 5–6; 56:4; 62:13; 63:2, 4; 66:16, 18–20; 67:1–2, 4–8; 69:1–19.
• Alternative sequences: 31→33 (cf. 4QPsq), 38→71 (considered one psalm)

Ps. 31→33. 4QPsa frg 4.   Yellow line: middle of Ps 31:24 (אמנם[ נ]וצר) Green line: vacat after Ps 31:25 Pink line: end of Ps 33:2 (עשור זמרו לו) Purple line: end of Ps 35
Red line: start of Ps 36

4QPsa frg 9 ii. The arrow shows the beginning of Ps 71, following immediately Ps 38.

A Temple Collection? (part 3)

I'm finally back at this subject, which has, I admit, ended up being more complicated than I anticipated (what else is new?). These posts are my attempt to get a handle on the evidence and arguments for a temple collection of scripture, especially as that relates to the development of the biblical canon. The first post surveyed some of the ancient testimonia for such a collection, and the second post presented the argument put forward by Emanuel Tov in regard to the textual data arising from the Judean Desert scrolls.

I did promise some evaluation of Tov's arguments for our present purposes (biblical canon), so here goes.

First of all, Tov's views are shared by other scholars. For example, David Carr:

The proto-Masoretic texts that start to appear in the late Hasmonean period probably linked back to reference exemplars stored there [= the Jerusalem temple]. (Formation, p. 163)

Armin Lange also believes that a standard text was created in the second half of the first century BCE by priests in the Jerusalem temple ("They Confirmed the Reading," p. 79). Lange (Flores Florentino): accepts much of Tov's presentation, but adds: (1) diversity of biblical quotations (see now his book and the review of same at JHS); (2) Greek recensions; (3) Greek textual scholarship (but precisely this third argument was refuted in 1978 by Bertil Albrektson, pp. 48-51, citing Lieberman before him). Lange thinks Jewish standardization began in Egypt (evidence: Alexandrian pagan scholarship; LXX mss earliest examples of Jewish standardization; Aristeas) and thence influenced Palestine after 63 BCE. Originally it was not aimed at MT exclusively--"The proto-Lucianic recension as well as the supralinear corrections of 5QDeut demonstrate that originally other texttypes were also favored as standard texts" (p. 126)--but eventually the MT was housed in the temple, perhaps by the late 1st cent. BCE.

But Tov's hypothesis is speculative, as he admits. This is how I summarized his position in my previous post. 

If the non-Qumran Judean Desert texts reflect exclusively MT and if those who used these scrolls can be reasonably considered to have recognized the authority of the Jerusalem temple and its leaders--and to have obtained their biblical texts in reliance on this Jerusalem leadership--and if rabbinic writings testify to the existence within the temple of a scroll of the Torah and possibly other biblical books used for corrections, then perhaps the manuscript evidence confirms these rabbinic statements and we should connect the non-Qumran textual uniformity to the temple.

Tov here says "This is a mere hypothesis" (p. 9) and he goes on in the same passage to admit that the rabbinic evidence really concerns only a Torah scroll in the temple, "but it stands to reason that other Scripture books were also found in the temple." There's the rub for our question (= canon): exactly which scripture books does it stand to reason were to be found in the temple?

So, it's a 'mere hypothesis'. On the other hand: 

The textual unity described above has to start somewhere and the assumption of master copies is therefore necessary. (same article, p. 8; a nearly identical sentence at TCHB p. 31)

So, the assumption of master copies is an assumption, but a necessary one. The location of these master copies in the temple is more of a guess, but a guess based on good sense and on the ancient testimonia that we surveyed in the first post.

That doesn't mean everyone has to like the guess. Eugene Ulrich, for instance:

Despite suggestions to the contrary, the future still awaits demonstration that the texts preserved in the medieval MT transmit the texts guarded by the priests in the Jerusalem temple as opposed to other popular or "vulgar" texts that were less well preserved by less well qualified people. Nor has a line of succession--from temple priests to Pharisees to rabbis--been convincingly shown. (here, p. 155)

In the footnote on the same page, he writes:

Moreover, to my knowledge, no one has demonstrated how we could know either the textual nature of the priests' manuscripts in the Jerusalem temple, or how the Pharisees/rabbis, usually considered a lay group, would have received them in contrast to the (probably priestly) LXX translators and the Qumran leaders who were presumably very strict priests. (p. 155 n. 18)

What does that leave us with?

First, Ulrich seems to accept the idea that the temple in Jerusalem housed some scrolls of scriptural books. The ancient statements affirming this point--and the comparative evidence from other cultures--put the matter beyond dispute to my mind. That is to say, I think it would be very hard to argue that the temple contained no scriptural scrolls, and I'm not sure why someone would want to deny it. The exact content of this temple library can certainly be disputed: besides the Pentateuch, which other scriptures were included? The entire Tanak? More than the Tanak? And which textual form were these scrolls? Just one form or multiple?

Second, Tov seems to have established that there were master copies of some scriptural books that led to the production of scrolls textually equivalent to the later MT. As he says, the idea of master copies is an assumption but a necessary one. While Ulrich does in another essay question to what extent the Masada scrolls actually reflect later MT, he must also admit in regard to his principal example (MasGen, a fragment with 8 complete and 3 broken words, containing 3 variants vis-a-vis the MT): "such small variants are to be expected even within the Masoretic group" (p. 456). Lange also says that MasGen cannot be classified, and he says the same about seven Bar Kokhba era mss (Handbuch, p. 24). Nevertheless, Lange basically agrees that all of these non-Qumran texts are protomasoretic and that they thus stand apart from the texts at Qumran (ibid.).

Granted the existence of master copies (though I'm not positive that Ulrich would grant this point), where would these master copies be housed? The Rabbis possessed MT-like scrolls, and their predecessors are usually thought to have been the Pharisees. Ulrich seems to accept this in the quotation above, so let's go with it. I guess it's possible that the Pharisees were the ones at Masada leaving behind the biblical scrolls, and at Wadi Murabba'at and elsewhere (especially if a wing of Pharisees really were into violent revolution, as Wright has argued here, pp. 185-86, and here, pp. 190-95).

But what about all those 'second-circle' texts (as Tov calls them) at Qumran? These 'second-circle' texts are copies of the proto-MT (I know Ulrich rolls his eyes when he reads that--because I'm sure he reads my blog!--sorry, it's a useful though anachronistic term) that are not quite as precise representations of the MT as are the non-Qumran Judean Desert scrolls, but they seem to have been copied from one of these latter 'inner-circle' texts. That is, they were not copied from the master scrolls themselves but from copies of the master scrolls. While the actual scrolls are earlier than the 'inner-circle' scrolls, they are typologically later. (See the summary of Tov's argument in my previous post.)

Does not the presence of the 'second-circle' texts at Qumran indicate that the master scrolls, or at least the text represented by them, must have been important, or, let's say, available, to groups beyond the Pharisees? What's more, I think the numbers in which they are represented at Qumran would indicate not just their availability but truly their importance. While Qumran undoubtedly exhibits textual pluriformity, it is equally undoubted that the textual form that became the MT was important there.

I don't know that the master scrolls were housed in the temple, but that is a pretty good guess. I'm basing this on the previous two points: (1) the 'necessary assumption' of master scrolls that provided the model for (a) the non-Qumran Judean Desert scrolls, (b) the biblical text adopted by the Rabbis and apparently earlier used by the Pharisees, and (c) many of the Qumran scriptural scrolls at a secondary level (second-circle texts); (2) the presence in the temple of some scriptural scrolls, at least the Pentateuch if not more. Moreover, the Rabbis thought that the temple Torah scroll was used for correcting other scrolls that were more-or-less proto-MT (since that is the only text form the Rabbis seem to attest), so we apparently have some ancient evidence (granted, it's later rabbinic testimony and so open to doubt) connecting the temple Torah scroll to the proto-MT. But, given the presence of proto-MT texts at Qumran and the apparent animosity of the Qumran group to the Pharisees, they probably weren't adopting this text from each other. It seems necessary to imagine that the text was disseminated from another authoritative source.

In light of these 'second-circle' texts at Qumran, I would also want to call into question Ulrich's statement quoted above that the Pharisees somehow ended up with these standard scrolls and the group at Qumran did not. Or, again, from Ulrich:

If any group had temple texts that they preserved and copied, the Qumran group would seem to be the most likely candidate. Their early members are widely believed to have been priests in the temple who separated themselves because they believed the temple had been defiled. (here, pp. 155-56)

It's true that scrolls of the inner-circle were not found at Qumran (except for 4QGen-b; Tov, TCHB, p. 31), but plenty of 'second-circle' texts were. I don't know why the 'inner-circle' is almost completely absent, but the presence of the 'second-circle' is still significant.

[By the way, Armin Lange has his own classification of the Qumran scrolls (Handbuch, 16-17): he counts 20 semimasoretische Handschriften and 7 protomasoretische Handschriften, along with the other text types: pre-SP (2 scrolls), scrolls equally near MT and SP (11), LXX Vorlage (4), independent (47), non-classifiable (83), and some which might not actually attest biblical books (35).]

What biblical books are represented by the 'inner-circle' and 'second-circle'?

'Inner-Circle'
(See my previous post, especially the 'update' at the bottom. The following list of biblical books represents the books attested by the non-Qumran Judean Desert scrolls as give in Tov's Revised Lists, pp. 126-29. )

• Genesis
• Exodus
• Leviticus
• Numbers
• Deuteronomy
• Joshua
• Judges
• Isaiah
• Jeremiah
• Ezekiel
• Twelve Prophets
• Psalms
• Daniel
• Ezra-Nehemiah

Some of these are of course more impressive specimens than others. Not all of them have enough text to say for sure that they absolutely represent proto-MT and not other textual editions (see the doubts about MasGen expressed by Ulrich and Lange, noted earlier).

'Second-Circle'
(The following list names the biblical books of the Qumran scrolls with exclusive closeness to MT as given by Tov here, pp. 154-57.)

• Genesis
• Exodus
• Joshua
• Samuel
• Kings
• Isaiah
• Jeremiah
• Ezekiel
• Twelve Prophets
• Psalms
• Job
• Proverbs
• Daniel
• Ezra

(Once again, I have a bit of a problem with Tov's statistics. Here, p. 22, he mentions "the exclusive closeness of fifty-seven Qumran texts to the medieval texts," but he doesn't list them. In the article cited above, he does list them but they only total 24 texts. He goes on to list other scrolls that are equally close to MT and either SP or LXX, which account for the remaining 33 texts.)

(Another caveat: some of these Qumran scrolls still present problems, even if the extant portions exclusively reflect MT. For example, the Book of Psalms is listed above in the 'second-circle,' but anyone familiar with the attestation of the Psalter at Qumran will know that it's more complicated than that. See Flint.)

If we put these two lists together, we find that all of the Torah is attested in this "standard" MT-like form, all of the Nevi'im, and five of the eleven Ketuvim. The books missing are the Five Megilloth (Ruth, Lamentations, Esther, Qoheleth, Song of Songs) and Chronicles.

If the books represented here--18 of the 24 books of the Hebrew Bible--were copied from master scrolls ('inner-circle') or from copies of the master scrolls ('second-circle'), then we can say that the text form represented by the master scrolls was available and important to the Qumran group and the Pharisees.

It seems to me that the most likely location for such master scrolls to be housed is the Jerusalem temple, especially given the ancient testimonia on precisely that point.

Remaining Problems

The textual diversity presents some difficulties, especially as Ulrich argues it. Not only does he point to Qumran, but he also argues that the SP, the LXX, the NT quotations, and Josephus "all resoundingly confirm this widely accepted state of pluriformity" (p. 155). That means that the evidence from Qumran is "representative of the Jewish scriptures generally in that period." I don't see that this overturns the possibility of a standard text, though it would limit the reception enjoyed by this standard text. Whereas we don't necessarily know what the Qumran group thought about the available textual options--we don't know, in other words, that they valued all these texts the same, as Ulrich wants us to think (p. 155)--the textual diversity exhibited by the LXX vis-a-vis the MT presumably meant that the translator valued his Vorlage over other options (though maybe even here other scenarios could be imagined; and for which books is there a real difference between LXX Vorlage and MT?). But there might still be a standard text promoted by an influential group, even if not all Jewish groups or individuals accepted or had access to this standard text.

Another problem is that MT is textually inferior in some books. This is especially true for Samuel, which is the typical example given in these discussions.

Is it at all plausible to picture the MT of the Books of Samuel as the outcome of a careful comparison of manuscripts and textual traditions, when it is obvious that it is an inferior text, full of errors and lacunae which could easily have been remedied with the aid of contemporary manuscript material? (Albrektson, p. 57)

[...] it is at the same time a good text--as a whole it is clearly superior to other textual traditions like the LXX or the Samaritan--and an uneven text with obvious and in places rather embarrassing defects. (Albrektson, p. 60)

The specific texts for each book in the rabbinic collection as reflected in the MT are, as far as we can tell, not selected or chosen but chance or coincidental. The poor state of the text, for example, of Samuel and Hosea, would seem to preclude conscious textual preference and selection; and the criteria for the choices of the MT versus the Hebrew Vorlage of the LXX could not have been the same for all books (cf., e.g., Jeremiah and Daniel). (Ulrich, p. 156)

I am not sure what to do with this fact. There were scrolls of Samuel circulating in antiquity with a higher quality of text than that of the MT. Why did not one of these higher quality scrolls become the standard text of later generations (i.e., the MT)? It is tempting to say, with other scholars (such as Albrektson and Ulrich) that this simply resulted from chance, that following the desolation of Judah enacted by the Romans, only this defective text of Samuel survived among the Pharisees/Rabbis. Does this undermine the entire idea of a standard text prior to the destruction? I don't think so, because Tov's assumption of master scrolls would still be a necessary assumption. How the text of Samuel works within this assumption requires further thought.

Conclusions

How does this relate to the canon of Hebrew scripture? To quote myself from the first post in this series:

the evidence makes it probable that the temple housed a collection of scripture, but it is hard (impossible) to know the exact contents of this collection, and it seems unlikely that it contained precisely the Jewish canonical books, no more and no less.

A collection of scripture in the Jerusalem temple that may have served the function of housing master scrolls for the dissemination of a particular text form of various scriptural books does give us a possible way of thinking about how the canon (to use again an anachronistic term) might have functioned in antiquity and how the ancient Jews might have conceptualized the canon. Since we are not in a position to know the precise contents of the temple collection, we cannot determine from this line of argument exactly which books counted as canonical or, indeed, whether anyone had given any thought to delimiting the scriptural collection in such a way.

Did the ancients notice that there were different editions of biblical books?

As I have been trying to press on with my thoughts on the ancient library housed in the temple (I really have, I promise), I've read some brief pieces by Eugene Ulrich. He likes to emphasize the textual plurality of the pre-rabbinic biblical text; in other words, there were multiple editions of various books of the Bible.

The Qumran biblical MSS show that at least six books (or possibly ten) of the twenty-four in the Masoretic canon circulated in variant literary editions in the closing centuries of the Second Temple period: Exodus, Numbers, Joshua, Judges(?), Samuel(?), Jeremiah, the Twelve Prophets(?), Psalms, Song, and Lamentations(?). When the study is widened to include the witness of the LXX and SP, seven (or eight) more can be added or become clear: Genesis (chapters 5 and 11), Samuel (at least 1 Samuel 16-17), Kings, Ezekiel, the Twelve Prophets, Job(?), Proverbs, and Daniel. Thus, we have MS evidence that Judaism during the last two and a half centuries while the Second Temple stood knew variant literary editions for half or more of the books that would become Tanakh: thirteen (or up to sixteen) of the traditional twenty-four books. ("Two Perspectives," p. 460). 

This is all well-known. There would certainly be disagreement among scholars as to how much variation, or which types of variation, render a copy a "variant literary edition," and some would want to dispute whether certain biblical books are actually attested in multiple editions--and Ulrich's question marks show that he himself is not sure about some of these--but the fact that variant editions of some books did circulate in antiquity I will take as a matter beyond serious doubt.

But, what did ancient people think about these multiple editions? I think that is a question to be pondered (as I have done before). I am led to think about this issue again by another of Ulrich's essays, which contains this paragraph:

It is possible, although undocumented, that some individuals may have been conscious of differences between variant editions of particular books and may have chosen one deliberately instead of another. But scrolls not in use were usually rolled up; and if there was more than one scroll of a book, it seems more in line with the evidence that a reader would have picked up one of the available rolled-up scrolls marked "במדבר" without knowing, and apparently from the Qumran evidence, without caring which text form of Numbers was inscribed inside. If there were an awareness of different editions and a conscious choice between them, the articulation of the choice is less likely to have been in terms of "pre-Samaritan versus proto-MT" and more likely "the newer, fuller edition versus the earlier, shorter edition. ("Determining Scriptural Status," p. 156). 

I assume Ulrich had a wry smile on his face as he typed that last sentence. I too doubt that an ancient person would have thought about the textual diversity (if he thought about it at all) in terms of proto-MT and pre-Samaritan. Point well-taken. These are anachronistic names for the textual forms of antiquity, and if we want to get inside the heads of the ancients, we might want to think about how they labeled these variant editions (if they thought about the issue).

But did ancient persons think about the issue? Ulrich wonders whether people were even "conscious of differences between variant editions of particular books."

Surely some people were conscious of the differences. First of all, someone had to have created the "newer, fuller edition," so he must have known about it. I don't guess we can say confidently much more about this scribe, but we might speculate that he would have a desire to disseminate his "newer, fuller edition" and so would have wanted to tell others about it. But how that would have happened, and how it would have been received, we can only guess.

Second, the Samaritans apparently made a conscious choice as to their text form. I think Carr more-or-less represents the consensus when he dates this to no later than the mid- to late-second century BCE (Formation, p. 177), so right during the time of the textual plurality attested at Qumran. For their Pentateuch, the Samaritans chose basically the same type of text for each book (except for Leviticus, which only existed in one form; Ulrich, "Two Perspectives," p. 459). They chose expanded texts, but not the most expanded texts (Kartveit, Origin, 285-88; Eshel and Eshel, "Dating," 237). According to Kartveit: "This makes the SP a deliberately chosen text" (Origin, p. 299).

Third,

Those who studied scripture intensively, which of course includes Essenes and Pharisees in particular, knew the material inside out and could evoke a whole world of textual reference with a word or phrase. The rabbis continued this tradition. (Wright, Paul, 176-77)

This sounds about right to me, that is, Wright's description of ancient Jews knowing the text "inside out." At least the Rabbis surely had a lot of scripture memorized. I don't guess we really know this for the earlier period, but Carr has presented a thesis (developed in Writing and Formation) in which memorization was an important part of textual transmission. If that is the case--if some literate ancients can be assumed to have had a good amount of scripture stuck in their heads--then the chances increase that they would notice when they are reading a different edition of a biblical book. That is not to say that they would have cared about the textual variation, though I suspect that at least some ancients would have noticed and cared. Despite Ulrich's assertion (quoted above), it is not at all apparent from the Qumran evidence that the ancients would have picked up a scroll of Numbers "without caring" about the textual form. It is possible that they did not care, but it is also possible that they did care. (See again this earlier post.) The evidence of the Samaritan Pentateuch suggests that the Samaritans did care.

But if someone in antiquity did care about the textual form of Numbers, still, how could he have determined which text form he was picking up? The physical scrolls do not seem to have been distinguished in any way (at least not in a way that is still apparent to us). So I don't know how he could tell. But I do know that when someone at church asks me about a new translation of the Bible, which I've never before seen or heard of, there are certain verses I'll read to see how the new translation rendered them, usually some familiar verses (John 3:16, etc.), but also some that I know some translations have messed up in the past (e.g. Gen 2:8; Isa. 7:14; etc.). Okay, I know, if someone handed me a biblical scroll in Hebrew with no vowels or chapter divisions, it would be somewhat more difficult to look up these verses. But I'm just using this modern analogy to say that there are ways of figuring out what kind of Bible you're looking at, even when you can't tell from the cover.