How Close is the Great Isaiah Scroll to the Medieval Masoretic Manuscripts?

I recently heard an author talking about the Great Isaiah Scroll (1QIsa-a) from Qumran, who said that "today's Hebrew version of the book of Isaiah is the exact same as that ancient copy, to the very letter." This reminded me of Wes Huff's overly optimistic appraisal on Joe Rogan's interview, where he said that it was "word for word identical to the Masoretic text." After some pushback, Wes, to his credit, did an interview with John Meade (where he generously cites me several times), clarifying the situation. But my recent experience urged me to write up this quick note to help set the record straight.

I want to be clear that I agree with the general consensus that the scroll is indeed very close in wording and meaning to the medieval Masoretic manuscripts. It has the same passages in the same order with the same message and mostly the same wording. This is indeed a remarkable observation, and I agree that it is legitimately impressive. I would go so far as to say that the Isaiah scroll is "almost word-for-word identical." But there are three nuances I would add:

First, speaking of letter-by-letter agreement should be avoided, because there are very many spelling differences. In fact, the scribe of the Great Isaiah Scroll fairly systematically updated the spelling of the scroll to follow his contemporary spelling conventions, whereas the Masoretic text more typically preserved the archaic, classical spellings. The scribe of 1QIsa-a also made frequent spelling mistakes, especially leaving out silent letters and such. These differences are usually not meaningful, but they are so frequent that we probably should not speak of particularly close letter-by-letter agreement.

Second, there are meaningful textual variants in the Great Isaiah Scroll. In most cases, they appear to be simply scribal errors. But in some cases modern scholars and translators have actually preferred readings of the Great Isaiah Scroll over the medieval manuscripts. Two examples may help illustrate:

Isa 19:18

On that day there will be five cities in the land of Egypt that speak the language of Canaan and swear allegiance to the LORD of hosts. One of these will be called the City of the Sun. (NRSV)

MT עִ֣יר הַהֶ֔רֶס "the city of destruction" (NKJV, ESV, NASB)

1QIsa-a עיר החרס "the city of the sun (= Heliopolis)" (supported by some medieval manuscripts and ancient translations) (CSB, NIV, NLT, NRSV)

The difference is only one graphically similar letter, but many modern scholars conclude that the reading of 1QIsa-a makes the most sense in context.

Isa 53:11

Out of his anguish he shall see light; he shall find satisfaction through his knowledge. (NRSV)

MT lacks אור "light" (ESV, NASB, NKJV, NLT)

1QIsa-a (plus additional DSS and the Septuagint) have the word אור "light" (CSB, NIV, NRSV)

Here the word makes good sense in the context and makes for a clearer reading than the obscure MT. Some have argued that this well-attested word was added to clarify the difficult text, but it may also have been omitted in the medieval tradition. Interestingly, the word in question is found in the middle of a cluster of similar letters (i.e., ', w/y, r) that could have prompted its omission: יראה אור ישבע yr'h 'wr ysb'.

Third, I would point out a series of difficult passages in the second half of the scroll that differ considerably from the Masoretic text, often with 1QIsa-a missing entire verses of the MT. Eugene Ulrich has argued that 1QIsa-a reflects an earlier version of Isaiah in these passages and that the MT has grown with additional verses being added. I have argued in an article that most of these passages were actually missing in the exemplar from which 1QIsa-a was copied, because it was damaged along the bottom edge. Thus, every column was physically damaged and missing one or more lines. You can see this, e.g., in how the scribe originally left blank lines for the missing verses. If correct, and I'm quite convinced it is, it means that most of the biggest differences between the Great Isaiah Scroll and the MT are easily explained as copying problems in the former.

So the big takeaway, I think it is fair to say that the Great Isaiah Scroll (1QIsa-a) is "almost word-for-word identical" to the medieval Masoretic text. But that doesn't mean it doesn't sometimes have its own independent text-critical contributions and problems, and the identity certainly doesn't extend to the letter level once we account for different spelling conventions. I hope that helps clear up some confusion and give popular writers some guidance for talking about the significance of this fascinating scroll.

Isaiah 9:6 and the Antiquity of the Masoretic Tradition

Today I was reading Isaiah 9:6 (Eng. v. 7) with some colleagues and came across a textual variant I had never noticed before.

5 וַיִּקְרָ֨א שְׁמ֜וֹ פֶּ֠לֶא 

יוֹעֵץ֙ אֵ֣ל גִּבּ֔וֹר אֲבִיעַ֖ד שַׂר־

שָׁלֽוֹם׃   6 לַם רַבָּה הַמִּשְׂרָ֜ה 

וּלְשָׁל֣וֹם אֵֽין־קֵ֗ץ 

and he is named Wonderful 

Counselor, Mighty God, Everlasting Father, Prince of 

Peace.  ⁷ His authority shall grow continually, 
and there shall be endless peace (NRSV)

In the Aleppo and Leningrad codices, the second letter of v. 6 is a final mem yielding לם רבה, whereas a Qere note in the margin (marked with קׄ) says to read למרבה as one word with a non-final mem.

Detail of Leningrad Codex. Photographer: Bruce E. Zuckerman.

My first instinct was to explain this away as a scribal error in the Aleppo codex I was reading from. This kind of confusion is common in the Dead Sea Scrolls, but I noted that it most commonly occurs on the archaic ־מו prefix (meaning "them") rather than in the middle of a word. A little more digging showed that the Leningrad codex has the same exact peculiarity, so this is certainly no accidental innovation in these manuscripts. 

In 1QIsa-a, the scribe does not use a distinct form of final mem, but does leave a space between למ and רבה. It is worth noting that the scribe does not always write a distinct final form, as is evident in the second detail below from just a few lines above our verse. Thus, it appears that our scribe here likely read למ רבה as two words rather than one.

Details of 1QIsa-a. © The Israel Museum.

Looking at the Septuagint, Ziegler reconstructs the Greek text as:

καὶ καλεῖται τὸ ὄνομα αὐτοῦ Μεγάλης βουλῆς ἄγγελος·

ἐγὼ γὰρ ἄξω εἰρήνην ἐπὶ τοὺς ἄρχοντας, 

εἰρήνην καὶ ὑγίειαν αὐτῷ (some mss αυτων). 

^{(6) 7} μεγάλη ἡ ἀρχὴ αὐτοῦ, 

καὶ τῆς εἰρήνης αὐτοῦ οὐκ ἔστιν ὅριον

and he is named Messenger of Great Counsel, 

for I will bring peace upon the rulers, 

peace and health to him (some mss them). 

⁷⁽⁶⁾  His sovereignty is great, 

and his peace has no boundary (NETS)

For our verse, the Greek translator--who often translates rather freely--is clearly struggling to make sense of a difficult text, reading something like:

אבי(א) עד שר(ים) שלום (שלום ובורי/בריאות?) למ(ו) רבה המשרה

The extra "peace and health to him" may have been added by the translator for some reason, or it could have been a Hebrew reading known to him or at least supposed by him. Note the similarity between a conjenctured retroversion like שלום שלום ובורי and שלום למו רבה. If so, the translation could perhaps reflect a double translation of the Hebrew representing two possible readings of the difficult Hebrew text before him. Either way, the translation αὐτῷ μεγάλη seems to imply a source text that read לם רבה as two separate words.

Thus, the final mem in the medieval Masoretic codices is no accident, but an accurate record of a textual variant that can be traced back to the 2nd century BCE in 1QIsa-a and the LXX! The inherited reading tradition preferred to read למרבה as one word (cf. the similar noun form in Isa 33:23), a tradition which can be traced at least as far back as the 2nd century CE (see Theodotion's translation τω πληθυνειν την παιδειαν; so also το εβρʹ λεμαρβη αμμισρα according to Chrysostum). But for over 1000 years, some (not all) Jewish scribes faithfully copied the alternative text with two words לם רבה inherited from their exemplars. This variant then gives us a remarkable window into the deep roots of the (sometimes multiple and conflicting) readings embedded in the traditional Masoretic text in one of the most famous passages in all of the Hebrew Bible.

The Scribe(s?) of the Great Isaiah Scroll

Mladen Popović, Maruf Dhali, and Lambert Schomaker have now published their groundbreaking PLOS ONE paper arguing that there are strong paleographic indications that the two halves of the Great Isaiah Scroll were written by two different writers.

Artificial intelligence based writer identification generates new evidence for the unknown scribes of the Dead Sea Scrolls exemplified by the Great Isaiah Scroll (1QIsaa)

More background on the project can be found in their recent presentations here, here, and here.

Peter Flint and the Dead Sea Scroll

Peter Flint gives a nice preview of his work on the Hebrew Bible Critical Edition of the Psalms and his Eerdmans' edition of the Qumran cave 1 scrolls here.

Developmental Stage, Scribal Lapse, or Physical Defect? 1QIsa-a’s Damaged Exemplar for Isaiah Chapters 34–66

My article on the Isaiah scroll is now out in the latest volume of Dead Sea Discoveries, for those who might be interested.

Longacre, Drew. "Developmental Stage, Scribal Lapse, or Physical Defect? 1QIsaa’s Damaged Exemplar for Isaiah Chapters 34–66." Dead Sea Discoveries 20, no. 1 (2013): 17-50.

Abstract

The Great Isaiah Scroll from Qumran (1QIsa^a) does not generally reflect a text form earlier than the Masoretic text. Instead, the convergence in 1QIsa^a of patterns of spacing irregularities, literary and textual problems, and secondary supplementations, as well as a consistent pattern of distribution, are best explained on the basis of the hypothesis of an exemplar for chapters 34-66 with a damaged bottom edge. Upon reaching the defective edge in each column of his exemplar, the scribe dealt with any lacunose or illegible text in one of two ways before continuing with the unaffected text at the top of the subsequent exemplar column. Sometimes he left blank spaces in his new copy to be filled in with the correct text from other manuscripts at a later time. At other times he attempted full or partial reconstructions of the text based on whatever text remained legible in the damaged exemplar, memory, and contextual clues.

I previously blogged about this here.

The Great Isaiah Scroll and Its Damaged Exemplar

In DJD 32, Ulrich and Flint make the case that the Great Isaiah Scroll does not reflect a later stage of development than the MT (contra Kutscher), but instead often reflects an earlier stage of development to which the MT adds significant insertions. They cite seven examples of "Insertions in M highlighted by 1QIsa(a)": 2:9b-10; 34:17-35:2; 37:5-7; 38:20b-22; 40:7; 40:14b-16; 63aβ-bα. In reading the Isaiah scroll recently, I believe I have noticed a pattern which would undermine the major pillars of their thesis.

Preliminarily, it should be admitted that it is quite possible that the Isaiah scroll preserves earlier, pre-expanded texts. 40:7 may be an example, as the original scribe omitted the verse entirely. The Old Greek also omits this verse. The verse is redundant with v. 8. It also disrupts the connection between vv. 6 and 8. Furthermore, the way v. 7 explicitly identifies the referents in the metaphor is characteristic of explanatory glosses. So v. 7 may be a late expansion in the MT. On the other hand, it may be a simple case of haplography in the Greek and Isaiah scroll.

That said, I have noticed a pattern, which may alternatively explain four (34:17-35:2; 37:5-7; 38:20b-22; and 40:14b-16) of Ulrich's and Flint's seven examples. First, none of these can be explained by homoioteleuton. Second, each of these verses are secondarily inserted into the manuscript, bringing the text into alignment with the MT. Third, the original scribe intentionally left one or more blank lines at each of these points, possibly indicating awareness of textual problems. And fourth, each of these four minuses occur at approximately the same location in their respective columns. Noting these patterns, I would like to propose an alternative explanation for these examples.

The four first-hand omissions in 1QIsa(a) at 34:17-35:2; 37:5-7; 38:20b-22; and 40:14b-16 do not reflect a pre-MT stage in the development of the text of Isaiah, but rather a damaged exemplar used by the original scribe of 1QIsa(a). Upon reaching the damaged edges, the scribe left blank spaces in his new copy to be filled in with the correct text from another manuscript at a later time and then continued with the first subsequent legible texts. It is unlikely that the original scribe would have known the expanded text of Isaiah well enough to note the absence of such innocuous additional texts and to note them while copying a non-expanded text (contra Ulrich and Flint). Rather, the consistent pattern of blank spaces and their consistent locations in the text indicate a defective exemplar, which the original scribe copied as best as he could, leaving blank lines for later insertion of the lacunae.

I have rearranged the columns in 1QIsa(a) from photographs on the Israel Museum website to approximate what this reconstructed exemplar must have looked like. The red base line indicates the bottom margin of the exemplar, and the blue boxes mark off the damaged portions of the text. This reconstruction can only be approximate, since 1QIsa(a)--and probably its exemplar--has uneven column widths, but it is sufficient to show that the physical reconstruction of the manuscript provides an obvious explanation for the minuses.

As we can see, all four of the major minuses occur at the bottom margin (potentially you could make the same case for the top margin) of the exemplar in close proximity. If the scroll had been damaged at that point, we would have a very straightforward explanation for how the verses could have been omitted in 1QIsa(a).

This proposal has several advantages of Ulrich's and Flint's proposal.

1) It does not require complex hypotheses about the scribe copying the non-expanded text while having in the back of his mind the expanded text, leaving room for (but not including) this alternative text. I would be interested to see if they can cite any other examples where this kind of procedure was followed. But I consider it quite unlikely that the scribe would have known both texts that well and followed such a scribal practice. My solution is much simpler and much more natural.

2) It explains why these "MT insertions" concentrate on such a small section of text, rather than being equally spread out across the entire book.

3) It explains why all of these were later inserted according to the MT, including one claimed by Ulrich and Flint to be in the original hand (37:5-7). Also interesting is the fact that (if their analysis is correct), the original scribe in 37:5-7 stopped mid-word, and then secondarily inserted the missing text! This would seem to be clear evidence for an exemplar whose bottom edge was defective.

For all of these reasons, my theory of a damaged exemplar for 1QIsa(a) seems to me to account for the evidence in a much simpler and more realistic manner than Ulrich's and Flint's proposal that it reflects a pre-MT stage of development. Careful attention to the physical characteristics of the manuscript and the practical mechanics of copying manuscripts has helped resolve several major textual problems. Without these central pillars, Ulrich's and Flint's theory is based on slender evidence indeed.

Update (10 April 2013): I have posted about the publication of my full DSD article on the Isaiah scroll here.

Paleography and Second Hand Corrections

At the end of chapter 34 in the Great Isaiah Scroll, I came across a significant correction. The first hand omits the end of 34:17 - 35:2. A second corrector has gone back and corrected this omission by adding in the missing text.

A similar situation occurs in 37:5-7.

Interestingly, in both places, the original scribe left an unusual blank line. Perhaps he was aware of his omissions? Perhaps his exemplar was damaged at these points, given the physical proximity of the corrections?

These corrections offer good examples for paleographic study of the different hands in the manuscript. When I first began reading Hebrew manuscripts, I was often bewildered by paleographers' claims to be able to distinguish multiple hands within manuscripts and their corrections. The differences often seemed so small and the letter forms so inconsistent, that I found identifying different hands very difficult. I must admit, I am still far from expert in paleography, but I can pick up on much more now than when I first began. Reading through the Great Isaiah Scroll in particular, I have become intimately familiar with the original hand of the manuscript by sheer repetition. With this familiarity in the background, many of the differences in the corrections practically jump off the page now. And I'm not just talking about the size and color of the scripts. I think this would be a good point to show some of the distinguishing characteristics of the hands to show how such work is done.

* - Original Hand
1 - First Correction (34:17 - 35:2)
2 - Second Correction (37:5-7)

Aleph

Both corrections show significant differences in the general form of the aleph. The left downstroke meets the diagonal almost at the far upper left corner, instead of the more normal location closer to the middle. The original hand also has more curve to it.

Heh

The second correction is quite close to the original hand, but the first correction is clearly distinct. It is more boxy, the horizontal stroke meets close to the top of the rightmost vertical stroke, and the leftmost vertical stroke barely touches the horizontal.

Waw

Once again, the second correction is much closer to the original hand than the first correction. The first has sharper angles and bends to the right (instead of the left) at the bottom.

Yodh

As before, the second correction is much closer to the original hand than the first correction. The first correction does not have the distinct carrot shape of the other two, but is more of a closed wedge shape.

Final Peh

Another example of how the first correction is clearly different from the main hand is his use of a special final form for the letter peh in the word אף. The main hand uses the same form as in the middle of words.
We could line up many such examples to show the precise differences between the different handwritings. These are merely a few letters as examples, but we can draw a few conclusions from them. The first correction is clearly done by a later scribe. The letter forms are quite drastically different from the original hand and exhibit later influences. These differences are so many and so striking that it is impossible to miss them, after you have been reading the Isaiah scroll for a while.

On the other hand, the second correction is written in a hand much more similar to the original hand. From these initial investigations, I suspect that it is a different scribe, due to slightly different letter forms. Nevertheless, because the two hands are so similar in many ways, we would have to look much more carefully to confirm this.
While this is very basic, I hope it helps clarify how scholars really can tell the difference between the handwritings of different scribes. Sometimes it is easy to tell, and other times it is more difficult. But it is always essential to understanding the history of the manuscript and its text.

Dittography in the Great Isaiah Scroll

Generally speaking, I consider it much more common accidentally to omit one of two similar words or phrases in a text (haplography) than to accidentally repeat a phrase (dittography). This is due in part to the fact that mentally engaged scribes are likely to note the repetitive copying of dittography and correct their errors, whereas scribes commiting haplography are more likely to be left unaware of their errors. In reading ancient manuscripts, this seems to bear out statistically. I have been reminded recently, however, about the real possibility of dittography in my reading of the Great Isaiah Scroll. I have recently come across two clear instances of dittography in 1QIsa(a) which serve as good reminders.

Isaiah 31:6

MT: ...שובו לאשר העמיקו
"Return to him whom the sons of Israel have so greatly rebelled against."

1QIsa(a): שוביו לאשר לאשר העמיקו

The Isaiah Scroll here has accidentally repeated the word לאשר translated "to him whom," rendering nonsense.

Isaiah 34:15

MT: בצלה אך שם
"... in its shadow. Also there ..."

1QIsa(a): בצלה אך אך שמה

The Isaiah Scroll here has accidentally repeated the word אך translated "also," rendering nonsense.

Inconsistent Spelling in the Great Isaiah Scroll

It is important to realize that spelling was not standardized or consistent in many ancient manuscripts. I came across a good example recently from 1QIsaiah(a)--The Great Isaiah Scroll. The Hebrew adverb for "very" is spelled three different ways in the manuscript: מואד in 16:6; 56:12; מאדה in 31:1; and מואדה in 47:6, 9; 52:13; 64:8, 11. These were all apparently written by the same scribe (though there is some debate) within a relatively short period of time. There also do not appear to be any obvious patterns, since the different spellings are scattered about the manuscript, so this example does not support the assertion that the spelling drastically changes in the second part of the manuscript. There are many such cases, even some with more variants (Kutscher Language and Background of the Isaiah Scroll 166-7, lists, for instance, ראש-רואש-ראוש-רוש and זות-זואת-זאות-זאת, among others). We should not necessarily expect to find, therefore, consistent orthography in biblical manuscripts.

Demons, Tribulation, Commands, or Incomprehensible Utterances?

Despite the somewhat provocative title, this is not a post about modern Pentecostal practices! :) Rather, it is about a certain orthographic variation I noticed in reading Isaiah 28:10, 13 in the Great Isaiah Scroll. In these verses, the MT has a very difficult series of repetitive expressions. The first is written ץו tsav, and many associate it with the root meaning "command." But the Isaiah Scroll reads ץי tsi, which may mean either "demon" or "ship." The LXX has θλιψιν "tribulation," apparently reading ץר tsar. All of these readings look remarkably similar, as seen in the letter combinations taken from the Isaiah Scroll below:

What does all this mean? The MT reading is likely to have given rise to the other two, but this just goes to show that semantic confusion can often be reflected in textual confusion. There can be little doubt that in the Second Temple period this phrase was just as incomprehensible as it is to us today!