This article examines the categories of “holy” and “unholy” texts within the Hebrew Scriptures through the lens of the Judean Desert scrolls. Despite later assumptions about special scribal procedures for sacred writings, the evidence from Qumran and related sites demonstrates that Scripture and non-Scripture scrolls were produced using identical scribal practices. Authority and sanctity were therefore attributed primarily to content, not to the physical scrolls themselves. A significant exception emerges in the proto-Masoretic (proto-MT) tradition, which introduced a theologically motivated no-change approach to the biblical text from at least the first century BCE. This innovation resulted in an unusually stable textual transmission that continued into the medieval Masoretic Text. The study analyzes the origins and implications of this scribal ideology, contrasting it with other textual traditions such as the LXX and Samaritan Pentateuch. Ultimately, the proto-MT tradition transformed certain Scripture scrolls into holy scribal artifacts.
Keywords: Proto-Masoretic Text, Qumran Scrolls, Scribal Practices, Textual Transmission, Hebrew Bible Canonization
Este artículo examina las categorías de “textos santos” y “no santos” dentro de las Escrituras hebreas a partir del análisis de los manuscritos del Desierto de Judea. A pesar de las suposiciones posteriores acerca de procedimientos escriturarios especiales para los escritos sagrados, la evidencia de Qumrán y de sitios relacionados demuestra que los rollos bíblicos y no bíblicos fueron producidos mediante prácticas escriturarias idénticas. La autoridad y la sacralidad se atribuían, por tanto, principalmente al contenido y no al objeto físico. Una excepción significativa emerge en la tradición proto-masorética (proto-MT), que introdujo un enfoque teológicamente motivado de no modificación del texto bíblico desde, al menos, el siglo I a. C. Esta innovación dio lugar a una transmisión textual inusualmente estable, que continuó hasta el texto masorético medieval. El estudio analiza los orígenes e implicaciones de esta ideología escrituraria, contrastándola con otras tradiciones textuales como la LXX y el Pentateuco samaritano. En última instancia, la tradición proto-MT transformó ciertos rollos en artefactos sagrados.
Palabras Clave: Texto proto-masorético, Rollos de Qumrán, Prácticas escriturarias, Transmisión textual, Canonización de la Biblia hebrea
When talking about1 the categories ‘holy,’ ‘proto-holy,’ and ‘authoritative,’ it would be best to refer to the period or periods for which we have written evidence, preferably in connection with a group that used the literature that has been preserved, such as a Judean Desert community. Any reference to such a group is a luxury, because usually we have to limit ourselves to literatures without a link to a community, but the availability of a group enables us to examine the community’s use of that literature. At Qumran, we can relate to the views on Scripture held by the members of the Qumran community, their scribal practices, and to the evidence of the Scripture scrolls themselves.
Having said that, I don’t know whether I can commit myself to any definition of a holy text in the period covered by the Qumran scrolls, from the third century BCE to the first century CE, or to the second century CE if we include the other Judean Desert scrolls. I, myself, never use the term ‘holy text’, although I do speak about sacred texts, which may be the same. The use of different terminologies is probably due to the variety of disciplines in which we work.
The Judean Desert corpora are usually subdivided into Scripture and non-Scripture or biblical and nonbiblical texts. This distinction is significant because the ancients may have related differently to these two types of literature in several areas. The distinction itself is anachronistic because the depositing of the Qumran scrolls in the caves preceded a clear conception of a biblical canon. All the same, the ancients may nevertheless have had different views about or approaches towards most of the books that we name ‘biblical’ and ‘nonbiblical’, even if the confines of the biblical canon had not yet been fixed. Therefore, continuing our investigation into holy and unholy scrolls, we first focus on the approach towards the Scripture scrolls found in the Judean Desert. That area comprises the early site of Masada (texts copied between 50 BCE and 30 CE) and the later sites of Wadi Murabba‘at, Wadi Sdeir, Naḥal Ḥever, Naḥal Ṣe’elim, and Naḥal ‘Arugot, dating to the period of the Bar Kokhba revolt in 132–135 CE. It also includes the site of Qumran (texts copied between 250 BCE and 70 CE), but I often mention Qumran separately from the Judean Desert sites even though it is technically found in the Judean Desert.
When analyzing the Judean Desert scrolls with regard to their special nature, we need to focus on two different aspects that are sometimes interlinked. The first pertains to all the scrolls as scribal products. The second pertains to a small group of sacred texts, the MT group, that introduced a no-change approach to the Scripture text. These are two different issues: the first has to do with the technicality of the book-making procedure; the second has theological implications in the case of the Scripture text (see p. 6). The issue of the authoritative status of the scrolls is only partially related to this discussion. In my view, almost all Scripture texts were considered authoritative but at the same time they allowed for changes, as did the non-Scripture authoritative writings.
(1) The scrolls as scribal products. For the communities that left the Scripture scrolls behind, the scrolls were relevant at several levels. They were valued because of their content. They were part of Israel’s sacred literature; for their owners, the content of many scrolls had both sacred and authoritative aspects. For us, these scrolls should be examined also as scribal and textual products, although the ancients themselves did not pay attention to these aspects. Since the scribes did not distinguish between the preparation of the two types of texts, I refer to the totality of the Qumran corpus, sacred and nonsacred, or using a different term, ‘holy’ and ‘unholy,’ in the same way.
(2) The no-change approach of the MT group. Non-Scripture and Scripture texts were prepared in the same way, including the Torah (see next section). In one aspect, the MT group went its own way as it innovated a no-change policy (see p. 6), motivated by theological principles.
At the scribal level, all Qumran scrolls were equal while they possessed different levels of authority. For most scribes it made no difference what kind of scroll they were preparing. Therefore, some known Qumran scribes copied both Scripture and non-Scripture scrolls. The most well-known example is the scribe who apparently copied the nonbiblical texts 1QS, 1QSa, 1QSb, and the biblical 4QSamc (4Q53), and his hand is also visible in several corrections in the biblical 1QIsaa.2 This is not what we would have expected from the scribal procedures known from the later history of Judaism. We have become accustomed to the descriptions in rabbinic literature that make us believe that special procedures were instituted for the copying of sacred books. However, the description of the book-making procedures in the rabbinic literature, presented as prescriptions, were not novel since they reflect the realia of all the Judean Desert scrolls, sacred and nonsacred.3 I refer to such matters as the leaving of open and closed paragraphs and the sizes of margins. We have also become used to the personal purity of scribes copying Torah scrolls in the Middle Ages and modern times. However, in the reality of the Judean Desert texts, things were different.
According to the research performed in my monograph Scribal Practices2,4 scrolls of all types were prepared and copied according to identical scribal practices, with very few exceptions. They share all the main scribal features relating to the material of the scrolls and its preparation, handwriting, writing, horizontal and vertical ruling, stitching of sheets, size and shape of columns, correction systems, scribal marks, length of scrolls, number of columns per sheet, height of columns, margins, paragraphing, repair-stitching, patching, initial and final handle sheets, use of guide dots/strokes, scribal approaches, etc. Although further research is required, seemingly the material used for biblical texts was not of superior quality to that used for nonbiblical compositions. In other words, as a rule, scribes did not make a distinction between the types of text they were copying. These descriptions include the Torah and the MT group of all the books.
Similar to the lack of difference between Scripture and non-Scripture scrolls in the book-making process, there was no difference between these two categories in the approach towards the copying process. That is, there was no distinction between the groups, but there was between the individual scribes. In both groups we find scribes who adhere precisely to the right and left columns, write neatly under the lines, are distinguished by a fine handwriting, and rarely err in copying their source. On the other hand, in both groups we also find scribes who are imprecise and sloppy in all these areas. These parameters thus depend on the inclination of the scribe and not on the nature of the text he is copying.
As a result, we find negligent Scripture scrolls and this is the more remarkable if this is in the Torah. Several Torah scrolls are written carelessly in the super-plene Qumran scribal practice, in which famously 1QIsaa was also written. These Torah texts involve many corrections and changes. This pertains to: 1QDeuta (1Q4), 4Q[Gen-]Exodb (4Q13), 4QDeutk2 (4Q38a), 4QDeutm (4Q40), as well as several tefillin such as 4QPhyl A, B (4Q128, 4Q129).
Likewise, the witnesses of the Torah (MT, SP, Qumran scrolls, and the Vorlage of the LXX) reflect the same degree of editorial and scribal intervention as the other Scripture books.5
The identity of the scribal practices across the board can also be examined when a scribe is identified who copied multiple copies as in the case of the scribe of 1QS, 1QSa, 1QSb, etc. (see n. 2). Further, with the aid of advanced paleographical methods, Gemma Hayes6 identified one scribe who wrote eight scrolls and corrected another one.7
Despite my statement that there is no major difference between the preparation of Scripture and non-Scripture scrolls, I noted a very few distinctive features of the Scripture scrolls. However, these features do not change the overall picture:
• Scripture texts from the Judean Desert were written almost exclusively on skin-based material (thus also the rabbinic prescriptions for the writing of Scripture texts).8 The relatively small number of papyrus fragments of Scripture texts9 possibly served as personal copies.
• Scripture texts (except for tefillin) were inscribed on only one side of the material unlike several nonbiblical opisthographs from the Judean Desert.10
*The Innovation of the MT Group: The No-Change Procedure
I now turn to MT, a relatively small group in antiquity. I will first explain the terminology:
• MT, Masoretic Text: the full-fledged medieval text of the traditional Bible
• Proto-MT: the consonantal predecessor of MT from Judean Desert sites, but not from Qumran
• MT-like texts: texts that are a little removed from MT, found at Qumran
While all other texts allowed for a constant flow of changes during their textual transmission, the proto-MT texts found in the Judean Desert initiated a no-change approach for the text, possibly already in the fourth–third century BCE, but certainly by the first century BCE. This approach brought about an innovation in this area in which all copies of any composition, Scripture or non-Scripture, differed from one another. This innovation was based on a concept of the holiness of the Scripture text in a certain milieu.
The scribes of the proto-MT scrolls shared the book-making procedure with all other scrolls, but in one aspect these scrolls went their own way as they innovated a no-change policy motivated by theological principles.
| Source | Word Count | Content Variants |
Spelling Variants Total |
Variants | Percentage of Variation |
| XḤevSeNumb (50–68 CE) | 95 | 0 | 0 |
0 |
0 |
| 34SePhyl (50–100 CE) | 122 | 0 | 0 | 0 | 0 |
| SdeirGen (50–100 CE) | 139 | 0 | 0 | 0 | 0 |
| MurPhyl (100–125 CE) | 393 | 0 | 0 | 0 | 0 |
| En-Gedi Lev (1st or 2nd cent. CE) | 185 | 0 | 0 | 0 | 0 |
| MasLevb (30 BCE–30 CE) | 456 | 0 | 1 | 1 | 0.22 |
| 4QGenb (50–68 CE) | 358 | 0 | 1 | 1 | 0.28 |
| MurXII (ca. 115 CE) | 3,605 | 18 | 23 | 41 | 1.13 |
| 5/6ḤevPs (50–68 CE) | 509 | 2 | 6 | 8 | 1.6 |
| MasEzek (50–1 BCE) | 382 | 3 | 8 | 11 | 2.87 |
| MasPsa (end 1st century BCE) | 284 | 5 | 4 | 9 | 3.16 |
Table 1: Well-Preserved Proto-Masoretic Texts Compared
with MT (Codex L), no/minute variation
As always, it is hard to determine a scribe’s approach to his source when that source has not been preserved, but in this case the decision is relatively easy since the main feature of the MT group is its closeness to and often identity with the medieval MT tradition of 1500 years later. This fact enables the identification of the ancient proto-MT texts. A glance at the details of some of the typical proto-MT texts will establish this statement (see Table 1).
The fact that the proto-MT has remained virtually unchanged for 1500 years is a man-made marvel that was probably enabled by centralized copying procedures. The nineteen proto-Masoretic Judean Desert scrolls large enough for analysis were found at both the earlier site of Masada (texts copied between 50 BCE and 30 CE) and the later sites of Wadi Murabba‘at, Wadi Sdeir, Naḥal Ḥever, Naḥal Ṣe’elim, and Naḥal ‘Arugot, dating to the period of the Bar Kokhba revolt in 132–135 CE.11 The latter texts were copied between 20 and 115 CE. 4QGenb, mislabeled as a Qumran text, but probably deriving from one of the Judean Desert sites, needs to be included in this group.12 This group also includes the En-Gedi Leviticus scroll that agrees with the medieval Codex L in all its details. It is fair to say that we now have access to only a small amount of proto-MT text, possibly no more than five percent of the Bible text. Nevertheless, since the surviving texts are virtually identical to the medieval MT, it stands to reason that the nonextant portions of proto-MT also would have been identical to the consonants of the medieval text. It should be emphasized that all the scroll fragments from these sites reflect the proto-MT and no other text patterns have been found there.13 To the best of our knowledge, this was the standard text of these Judean Desert sites, which remained virtually without change until the Middle Ages.14 The Qumran evidence is different since copies of proto-MT were not found there. Slightly more distant members of that family (MT-like texts) are found among the Qumran scrolls. These are ancient texts belonging to the masoretic family that can be identified by statistical analysis as they differ by up to 10 percent from the medieval masoretic texts, while the proto-Masoretic texts differ from MT by up to 2 percent.15
The special nature of MT has been described often, but we now turn to the ideology behind that unique approach. When explaining the decision to single out the MT group as the only no-change text tradition, we base ourselves not only the precision in the transmission but also on the fact that the text of the MT group has remained stable for 1500 years, at least from the first century BCE until the Middle Ages.
Is the MT the only no-change text? Like the MT group, the Hebrew source of the LXX also presents a precise text, except for the Torah, but there is a basic difference between the LXX and the MT group. The translation of the LXX represents a one-time translation from a Hebrew source, while the MT group preserved a continuous almost-identical transmission over more than 1,500 years. We don’t know how the Hebrew source of the LXX would have developed, with or without internal changes, since in our mind that source has remained frozen since the act of the translation. By the same token, the SP group does not provide a good parallel to the no-change procedure of the MT group: there is an element of relative stability between the textual condition of the pre-Samaritan scrolls and the medieval Samaritan texts, but in all these texts spelling differences abound causing many internal differences. Furthermore, SP differs relatively much from the pre-Samaritan texts in each column of these texts. Finally, SP includes some Samaritan sectarian readings16 as well as phonetic variants not found in the earlier pre-Samaritan texts.
The proto-MT texts innovated a no-change approach to the Scripture text, thus differing from the other sacred texts. From the proto-MT onwards, both the content and the writing surface of this text were considered sacred. In due course, this text, possibly once a minority text, was to become the majority text of ancient Israel.
Proto-MT texts, in addition to Table 1 (earliest texts)
Masada17 MurIsa (Mur 3) 20–84 CE
MasPsb (Mas 1f) 50–25 BCE Naḥal Ḥever
Wadi Murabba‘at XḤev/SeNumb (XḤev 2) 50–68 CE
MurDeut (Mur 2) 20–84 CE
MT-like scrolls (Qumran)
1QIsab 50–25 BCE 4QJerc 30–1 BCE
4QJera 225–175 BCE18 4QPsc 50–68 CE
Table 2: Well-Preserved Proto-Masoretic and MT-like Texts Compared with MT
(Codex L), little variation
The suggestion that the MT group initiated a no-change approach does not depend on chronological presuppositions. We have good evidence for the proto-MT for the first century BCE from sites in the Judean Desert, especially Masada (Table 1 above), and for the MT-like texts from Qumran at least from 150 BCE. The texts present in Table 2 (above) are large enough for textual analysis.
4QJera (225–175 BCE) may point to an earlier date for the proto-MT, but the evidence is problematic. This scroll, copied in the MT-like tradition, displays proximity to the medieval text, even in matters of spelling.19 Furthermore, the text clearly displays the content of the edition of MT as opposed to that of the LXX (the same pertains to the content of 4QJerc). At the same time, the scribal character of this text is a far cry from precision. It contains more mistakes and corrections than any other scroll found at Qumran. As a result, when considering the no-change approach of the proto-MT text a novelty, this approach is reflected in this text. However, the scribe was sloppy at the same time, making multiple mistakes and correcting them towards MT.
As a result, the date for the evidence of the proto-MT approach may be firmly set at the mid-first century BCE, most likely at the beginning of that century. The evidence of 4QJera (225–175 BCE) may allow us to push this date back to the third century BCE even if the scribe erred much.
The evidence of the LXX compared with that of MT may provide another anchor for pushing the date of MT back to the fourth–third centuries BCE. When compared with the MT, the LXX and SP reflect much harmonization (harmonizing pluses). By implication, these pluses were based on an unharmonized text like MT, so that the Greek translation of the Torah performed in approximately 285 BCE must have been preceded by the earlier text of MT or a similar one. I admit that this argument is based on circular reasoning.
Further, some evidence is available for the assumed presence of the proto-MT in the fourth century BCE, the presumed time of composition of Chronicles.20 The Chronicler quoted from genealogies in the Torah, from a text close to MT-Torah.21 MT-1 Chr quoted extensively from MT-Torah in the first seven chapters, staying close to MT.22 However, the comparison of these texts is complicated by the fact that all known texts of Chronicles were at least slightly harmonized to those of the Torah. Nevertheless, other lists in Chronicles were not harmonized, making me believe that MT-Chr indeed knew MT-Torah in the approximate form known to us.
Finally, probably due to coincidence, none of the proto-MT texts from the fourth–second centuries have been preserved. They were kept in communities whose scrolls and libraries have not been preserved. We meet these scrolls for the first time when fugitives from Jerusalem and Qumran brought them to an arid area in the first century BCE (Masada) and to the Judean Desert caves in the second century CE. These scrolls were preserved due solely to the dry climate of these sites.
We don’t know who initiated the novel approach of the proto-MT movement, or which events may have brought about that change. Could it have been the introduction of Scripture scrolls to the temple library? Following this act, the books would have been considered sacred objects requiring a special approach. All this is pure speculation, but I’ve reached the conclusion that in the Torah, and in the Torah only, MT reflects the purest text form, that is, the earliest obtainable text form based on the available evidence. All other known Torah texts (especially the LXX, SP, many tefillin, and liturgical texts have been harmonized and altered by secondary changes. I believe that this assumption accords with another assumption that explains the background of the lack of changes in MT over the course of 1500 years since that text was locked up unchanged in the temple as a master copy.
Scribal precision as an abstract principle is an insufficient explanation for the virtual lack of differences between the copies of the proto-MT in early times and through the centuries until the medieval text. There are two main possible explanations for this near identity. (1) One explanation is that all manuscripts in ancient Israel were constantly corrected towards the proto-MT. According to this scenario, in the beginning there were many differences between the scrolls, but these differences were slowly obliterated by the systematic change of the texts towards an abstract entity that we now call the proto-MT. However, there is no evidence for such a scenario, at least not in the scrolls that have been preserved. This process has been named the stabilization of the text.23 (2) The other option is that the proto-MT scrolls were copied from or revised according to a master copy located in a central place, probably the Jerusalem temple, as suggested above. Likewise, this hypothesis is not easy as it would involve massive logistic problems, necessitating centralized activity in Jerusalem.
However, there is some later evidence for this theory in the Talmud.24 On the basis of this option, I postulate that some of the scrolls found in the Judean Desert (recorded in Table 1) served as such “corrected copies,” the term used in the Talmud, that is, copies corrected according to a master copy in a central place. It seems that the assumption of correction according to a master copy provides the only possible explanation for the abstract identity.25
Ultimately, the motive behind this no-change approach of the proto-MT was theological. The content of the authoritative text was considered so binding that not even the smallest change to the text and the writing surface was allowed after a certain period. We notice the lack of change in the transmission of spelling and content readings and even in the reproduction of the scribal signs in the text. As for the signs, ancient scribes left marks in the text for the correction of errors that were not supposed to be passed on to the next generation of scribes or readers. However, within the masoretic tradition even these scribal signs were considered part of the Scripture text. As a result, cancellation dots and parenthesis signs in the proto-MT were perpetuated for eternity. A parallel from modern times would be if we were to leave in the printed text the printers’ remarks in the proofs. As a result of these developments, the Masoretic copies became ‘holy’ as scribal products; in later times, the MT of the Torah became in use in the synagogue service.
At the same time, there is no evidence regarding the persons who created or shaped the proto-MT. It is enticing to assume that certain theological circles were involved in the rewriting of at least a minute layer of the proto-MT before it became sacrosanct, but such evidence is lacking. There is no proof that details in the proto-MT were changed in line with the views of protorabbinic circles despite the attempt by Abraham Geiger26 and others to find a few Pharisaic or anti-Sadducean changes in MT. Such theological changes as are found in that text were probably inserted by individual scribes without connection to overall tendencies.27 The proto-MT influenced the rabbis and not the other way around, because the text could no longer be changed when these circles were operating.
Summarizing my study, I started off saying that there are two levels of viewing the Judean Desert scrolls. As scribal products, all scrolls were identical; there was no additional value placed on Scripture scrolls. In both Scripture and non-Scripture scrolls we find scribes who are precise and imprecise in their copying. These parameters depend on the scribe’s inclination and not on the nature of the text he is copying.
However, a small group of texts, the proto-MT texts, innovated a no-change approach to the Scripture text, thus differing from the other sacred texts. The content as well as the writing surface of this text were considered sacred. In due course, this text, possibly once a minority text, became the majority text of ancient Israel due to historical developments among the Jewish people.
Because of these two different developments, I named my paper “‘Holy’ and ‘Unholy’ Texts of Hebrew Scripture.” The content was always ‘holy’ or sacred, but as scribal products these scrolls were not. On the other hand, in due course, one group of texts, the proto-Masoretic copies became ‘holy’ also scribal products. The persons that embraced that text innovated a no-change approach to the content and writing surface of the text. Because of historical developments within the Jewish people, this stream and this text became the majority text in Judaism.
1 This study was presented to a workshop of the Bonn-Jerusalem project “The Rise and Function of Textualism: Concepts, Histories and Traditions of Holy Texts in the Ancient East Mediterranean,” September 8–11, 2025 at the University of Bonn.
2 See Eugene Ulrich, “4QSamc: A Fragmentary Manuscript of 2 Samuel 14–15 from the Scribe of the Serek Hayyaḥad (1QS),” BASOR 235 (1979): 1–25; Eibert J.C. Tigchelaar, “In Search of the Scribe of 1QS,” in S.M. Paul, R.A. Kraft, L.H. Schiffman, and W.F. Fields, with the assistance of E. Ben-David, eds. Emanuel: Studies in Hebrew Bible, Septuagint, and Dead Sea Scrolls in Honor of Emanuel Tov, VTSup 94 (Leiden: Brill, 2003), 439–52. Likewise, Eibert J.C. Tigchelaar, “4Q26b (4QLeviticusg),” Textus 29 (2020): 53–56 suggested that 1QGen (1Q1), 1QMyst (1Q27), 4QLevg (4Q26b) were copied by the same scribe.
3 If no innovations were made to these procedures, why were they put into writing in rabbinic literature? The writing of Torah scrolls had become a ceremonial act within Judaism since the scrolls were used in the religious ceremony. The Torah scroll had become a religious object like tefillin and mezuzot and the writing of the three texts is often mentioned together in rabbinic literature. Therefore, all the details relating to the preparation and writing of Torah scrolls were of importance within the religious experience. The details had to be performed correctly since they were now discussed among the rabbis and were codified. The details of the book-making process were presented as prescriptions since nothing could be left to coincidence. In that way the rabbis ascertained that religious objects were prepared according to the customs that had become part of tradition. See Emanuel Tov, “Scribal Practices Mentioned in Rabbinic Sources Compared with the Judean Desert Scrolls,” in Missing Pieces: Essays in Honour of Eibert J.C. Tigchelaar, STDJ 152 (Leiden: Brill, 2025), 560–78.
4 Emanuel Tov, Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert: Material Culture and Writing Practices in Modern Research, Revised Second Edition, STDJ 153 (Leiden: Brill, 2025).
5 Tov, Scribal Practices2, Appendix 1b, col. 7; ibid., Appendix 8.
6 Gemma Hayes, “Searching for Dead Sea Scribes: A Study on Using Artificial Intelligence and Palaeography for Writer Identification in Correlation with Spelling and Scribal Practices, Codicology, Handwriting Quality, and Literary Classification Systems for Dead Sea Scrolls” (Ph.D diss., Groningen University, 2023).
7 Hayes, “Searching” § 2.4.1.1 went far beyond the analyses recorded in the previous paragraph when identifying with the aid of advanced paleographical methods one scribe who wrote eight scrolls and corrected another one: the biblical 4QSamc as well as 1QS, 1QSa (1Q28a), 1QSb (1Q28b), 4QTest (4Q175), 4QParaGen-Exod (4Q422), 4QPersonal Prayer (4Q443), 4QEschatological Hymn (4Q457b). His hand is also visible in two corrections in 1QIsaa.
8 Hayes, “Searching”, Table 8.9 (3).
9 Hayes, “Searching”, Table 3.9.
10 Hayes, “Searching”, ch. 4b and Appendix 3.
11 This number is composed of all the texts large enough for analysis from Masada (MasLeva,b, MasDeut, MasPsa, b), Murabba‛at (MurGen, MurExod, MurNum [possibly all three represent one scroll], MurIsa, MurXII; MurPhyl), Naḥal Ḥever (5/6ḤevPs, XḤevSeNumb, XḤevSeDeut), Naḥal ‘Arugot (ArugLev), Naḥal Ṣe’elim (34SePhyl), Sdeir (SdeirGen), En-Gedi (En-Gedi Lev); 4QGenb (recognized as deriving from the Judean Desert; see James Davila, DJD XII, 31). The link with the medieval text of MasPsa, especially regarding its stichometric layout, was shown in detail by Peter J. Gentry and John D. Meade, “MasPsa and the Early History of the Hebrew Psalter,” in From Scribal Error to Rewriting: How Ancient Texts Could and Could Not Be Changed, ed. A. Aejmelaeus et al., De Septuaginta Investigationes 12 (Vandenhoeck & Ruprecht: Göttingen, 2020), 113–45. The contrast between the Masada and Qumran texts was stressed by Ian Young, “The Contrast between the Qumran and Masada Biblical Scrolls in the Light of New Data: A Note in Light of the Alan Crown Festschrift,” in Keter Shem Tov: Collected Essays on the Dead Sea Scrolls in Memory of Alan Crown, ed. S. Tzoref and I. Young (Piscataway: Gorgias Press, 2013), 113–19. For reservations regarding the proto-MT character of MasGen and MasPsb, see Ulrich, The DSS and the Developmental Composition, 251–63 (MasGen) and Eibert J. C. Tigchelaar, “Notes on Two Masada Scripture Manuscripts: Mas1 (MasGen) and Mas1f (MasPsb),” in Hokhmat Sopher, Mélanges offerts au Professeur Émile Puech en l’honneur de son quatre-vingtième anniversaire, ed. Jean-Sébastien Rey and Martin Staszak (Leuven: Peeters, 2021), 335–46. Both scrolls are not included in the statistics as they are too small for analysis.
12 See James R. Davila in Eugene Ulrich and Frank M. Cross, Qumran Cave 4.VII: Genesis to Numbers, DJD XII (Oxford: Clarendon, 1994), 31.
13 It has been claimed that in the late Second Temple period there was insufficient cohesion in Judaism and the lack of a clearly acknowledged leadership guiding the Jews to recognize a single authoritative text (Ulrich, The DSS and Developmental Composition, 18–19).
14 See Tov, “Approaches of Scribes,” Table 4.
15 In TCHB3 I introduced the notion of MT-like texts found among the Qumran scrolls.
16 Even if is now suggested that the amount of the Samaritan sectarian readings is small, they are there. See Tov, TCHB4, 190–92.
17 The dates are based on those suggested by the publication of Shemaryahu Talmon in idem and Yigael Yadin, Masada VI: The Yigael Yadin Excavations 1963–1965 Final Reports: Hebrew Fragments from Masada (Jerusalem: Israel Exploration Society and the Hebrew University of Jerusalem, 1999), 92.
18 4QJera (4Q70) III 6 (Jer 7:30–8:3): The handwriting of the long addition in the interlinear space, in the intercolumnar margin, and below the text reveals a different writing style from that of the main scribe (note the different shapes of the ’aleph, bet, lamed, mem, final mem, ‛ayin, and shin). Likewise, in XI 6, the added לא was written with a different type of lamed than that used in the remainder of the scroll. This lamed has a streamlined shape and does not contain a horizontal line. Discussions of the long insertion in this scroll: Aviyah HaCohen, “4QJera: A Pre-Massoretic Text?” Textus 17 (1994): ח–א (Heb.); Joseph Riordan, “Sin of Omission or Commission: An Insertion in 4QJera?” in Gottes Wort im Menschenwort: Festschrift für Georg Fischer SJ zum 60. Geburtstag, ed. D. Markl, C. Paganini, and S. Paganini, ÖBS 43 (Frankfurt: Peter Lange, 2014), 99–112; Ulrich, The DSS and the Developmental Composition, 141–50; Kipp Davis, “Margins as Media: The Long Insertion in 4QJera (4Q70),” in Bible as Notepad: Tracing Annotations and Annotation Practices in Late Antique and Medieval Biblical Manuscripts, ed. L.I. Lied and M. Maniaci (Berlin: De Gruyter, 2018), 39–53; Falk, “Margins,” 10–38. Tov and Riordan consider this long insertion a scribal omission, while the other scholars regard it as a contextual addition. Davis, “Margins,” 47 assumed that 50–125 years passed between the original copying of the scroll and the time of the long insertion.
19 The scroll reflects the orthography of MT even in unusual spellings. For example, (ו)תהיינה (30x) and (ו)תהיין (14x) is the majority spelling in MT, in contrast to (ו)תהינה. This majority spelling occurs three times in MT-Jeremiah, whereas the minority spelling occurs only in 18:21. Remarkably, in this detail, 4QJera agrees with the medieval codices.
20 For the date, see Sara Japhet, I & II Chronicles: A Commentary, OTL (Louisville: Westminster John Knox, 1993), 24–28 (28); Gary N. Knoppers, I Chronicles 1-9, A New Translation with Introduction and Commentary, The Anchor Yale Bible Commentaries (New York, NY: Doubleday, 2003), 116–17; Ralph W. Klein, 1 Chronicles, A Commentary, Hermeneia (Minneapolis, Fortress, 2006), 16.
21 Emanuel Tov, “The Genealogical Lists in 1 Chronicles 1–7 and Their Textual Background,” forthcoming.
22 Genealogies are found in the first nine chapters of the book, but only in the first seven chapters can we trace the Torah sources.
23 Tov, TCHB4, 374–78.
24 Beyond the mentioned reference, see b. Pesaḥ. 112a, where one of the five instructions of R. Akiba to his student R. Simeon was: “and when you teach your son, teach him from a corrected scroll.” Another such precise copy was the “Scroll of the King,” which accompanied the king wherever he went. Y. Sanh. 2.20c tells us that this scroll was corrected to “the copy in the temple court in accordance with the court of seventy-one members.” This may well be an imaginary scroll and the description is equally imaginary, but the details of the description fit the practice of the “corrected scrolls.”
25 The discovery of nineteen Hebrew scrolls identical to the medieval Masoretic Text brought about an innovation in terminology: Prior to the discovery of the Dead Sea Scrolls, there was no special term for masoretic texts dating to the period before the Middle Ages. However, when texts close to MT were identified among the Dead Sea Scrolls, the term “proto-Masoretic” was coined by William F. Albright, “New Light on Early Recensions of the Hebrew Bible,” BASOR 140 (1955): 27–33 (30).
26 Abraham Geiger, Urschrift und Übersetzungen der Bibel in ihrer Abhängigkeit von der innern Entwicklung des Judentums, 2nd ed. (Frankfurt a. Main: Madda, 1928 [Breslau: Heinauer, 1857]); Alexander Rofé, “The Onset of Sects in Post-Exilic Judaism: Neglected Evidence from the Septuagint, Trito-Isaiah, Ben Sira, and Malachi,” in The Social World of Formative Christianity and Judaism, Essays in Tribute to Howard Clark Kee, ed. Jacob Neusner et al. (Philadelphia: Fortress, 1988), 39–49 (40–41).
27 See Tov, TCHB3, 254–56.