Shiur #7f: Nusach Ha-mikra - Accuracy of the Biblical Text

  • Rav Amnon Bazak

F.         Textual witnesses


As the last few shiurim have argued, the Masoretic text is the "most complete and most accurate extant testimony" for the Tanakh.[1] Nevertheless, there are many other ancient textual witnesses, which contain numerous instances of different versions of words or verses. Let us first conduct a general review of these textual witnesses, and then address the question of how they should be regarded.


The textual witnesses may be divided into two main groups:


1.            Hebrew manuscripts


The first group consists of the ancient Hebrew textual witnesses, first and foremost among them the Dead Sea Scrolls.[2] In the caves on the western side of the Dead Sea, especially around the Khirbet Qumran area, hundreds of manuscripts were discovered, mostly during the years 1947–1956. These manuscripts were written between the 3rd century B.C.E. and the 1st century C.E. – i.e. more than a thousand years before the earliest Hebrew manuscripts that were known before their discovery. The majority of the scrolls appear to belong to a cult that split away from the central stream of Judaism, and much of the material is devoted to their cultic ideology, including the "War of the Children of Light against the Children of Darkness" and the "Temple Scroll.” However, this collection also included some two hundred biblical scrolls, with the remnants of all the books of Tanakh (with the exception of the Scroll of Esther). Most of these were written in square script used today; only a few copies of the Books of Torah and of the Book of Iyov are written in the ancient Hebrew script (ketav Ivri).


The biblical scrolls discovered at Qumran may be further divided into two main groups. One group includes those that reflect the normative Masoretic tradition, and here the letter-text is adhered to very strictly. Within this category we find "the Second Scroll of Isaiah" (including parts of chapters 53-60). Although a thousand years separate this scroll from the earliest versions of the Masoretic text, the differences between them are very minor.


The second group includes scrolls whose letter-text differs from the Masoretic Tanakh, and in general they display a fully 'plene' spelling and elongated language forms (such as הואה instead of הוא; היאה instead of היא; and אתמה instead of אתם), and they are characterized by a freer version of the letter-text, sometimes to the point of carelessness.


Until the discovery of these scrolls, many biblical scholars had dismissed the significance of the Masoretic text owing to its relatively late appearance. The scrolls brought about a change in attitude, since they demonstrated that the Masora preserved a tradition that was older by a millennium or more, and that had been passed down meticulously. We may therefore state that the text upon which the Masoretic version is based was one of the textual versions that existed in the centuries prior to the Common Era. In addition, even those scrolls that cannot be easily categorized with the textual tradition of the Masoretic version do not display far-reaching deviations from it; the differences can be attributed to the normal processes of copying. Moreover, the instances of significant deviations from the Masoretic version do not necessarily reflect a version that is earlier than the Masora. There is a strong correlation between some of the scrolls and the version that appears in the Septuagint, which we will discuss below, and there are also links between some of the scrolls and the Samaritan version of the Tanakh.


The Samaritan Torah[3] is another Hebrew textual witness for the Chumash. The Samaritans – or, as Chazal refer to them, "Kuttim" (Cuthites)[4] – accepted only the Five Books of the Chumash and rejected the rest of the Tanakh. The existence of a sefer Torah in the possession of the Kuttim is mentioned already by Chazal, who note differences in spelling between this version and the Jewish Torah:


"Mar Zutra (or, some say – Mar Ukva) said: In the beginning, the Torah was given to Israel in Hebrew script and in the holy tongue. In the times of Ezra it was given to them over again in Ashurit script and in the Aramaic tongue. Ultimately, they settled on the Ashurit script and Hebrew tongue, leaving the (ancient) Hebrew script and the Aramaic tongue for the 'hedyotot.’ Who is referred to by the term 'hedyotot'? Rav Chisda said: the Kuttim." (Sanhedrin 21b)


The Samaritan Torah is indeed written in ancient Hebrew script, but also in the Hebrew language. There are no extant ancient manuscripts of the Samaritan Torah, and the oldest scroll that the Samaritan community has in its possession dates back to approximately the 12th century. It is generally agreed today that the Masoretic text predates the Samaritan one, especially owing to the addenda and corrections found in the latter. In any event, it is essentially based on an ancient text, which – as noted above – has parallels in some of the Dead Sea Scrolls as well.


There are some 6,000 differences between the Samaritan text and the Masoretic text, and in about a third of these instances the Samaritan version is identical to that of the Septuagint. A great many of the differences concern linguistic phenomena (such as plene spelling), exegetical addenda, etc.; and these seem to derive from a text of a late date. There are also differences that reflect the Samaritan ideology, especially in matters relating to the unique status that the Samaritans attributed to Mount Gerizim.[5] However, there are also many other differences that may reflect the existence of a version more ancient than that of the Masora – as indicated by the connection between the Samarian text and some of the Qumran scrolls.[6]


2.            Translations


The second group of textual witnesses consists of ancient translations of the Tanakh.[7] The most significant translation, for the purposes of our discussion, is the Septuagint. The Letter of Aristeas[8] speaks of a delegation of seventy-two elders that arrived in Alexandria, at the order of King Ptolemy, to translate the Torah. A similar account of the story appears in the Gemara:[9]


"‘It is related concerning King Ptolemy that he brought together seventy-two elders[10] and placed them in seventy-two [separate] rooms, without telling them why he had brought them together. He went in to each one of them and said to him, Write for me the Torah of Moses your teacher…" (Megilla 9a)[11]


Without getting into the details of the account and the many questions that surround it, it should be noted that while the story concerns only the Five Books of the Torah (Chumash), it has served as the source for applying the name "Septuagint" to the Greek translation of Tanakh as a whole – a project which was undertaken by different translators using different styles. There are various extant manuscripts of the Septuagint, the oldest of them dating to the 4th century C.E.


Notably, with regard to our discussion, the Septuagint deviates in many instances from the Masoretic text. In some places the differences may be attributed to the nature of the translation, exegetical problems, or different emphases, but there are places where this version may reveal something about the Hebrew manuscript that the translators worked with and the differences between this text and our Masoretic version. As noted, the Dead Sea Scrolls also include some manuscripts that reflect a Hebrew text similar to the one used by the creators of the Septuagint.


Aside from the Septuagint and other translations that were based upon or influenced by it, it is also important to note the Aramaic translations from the 1st-8th centuries C.E., including the translations of Onkelos and Yonatan ben Uzziel, and the Syriac translation – the Peshitta. These translations generally adhere closely to the Masoretic text, but there are occasional differences, as we shall see further on.


G.        Proposals for textual amendments


As we have seen, the issue of the accuracy of the Tanakh text raises many questions. First, we must consider instances where there is a discrepancy between the version found in the Dead Sea Scrolls and the Masoretic text with which we are familiar. Then there are the instances where Chazal based their teachings on a version that is different from the Masoretic text. In addition there are the disagreements between the various Masoretes: "easterners" vs. "westerners," and among the "westerners" themselves, between Ben Asher and Ben Naftali. To this we may add the range of textual versions that were used during the Middle Ages, and – finally – the discrepancies between different contemporary printed editions of the Tanakh. Thus, it is difficult to assert that the textual version that appears in any Tanakh today is in every detail a perfect copy of the "original" text. As Rabbi Mordekhai Breuer writes:


"Anyone today who seeks to prove that the biblical manuscripts of the Second Temple Period – or of the period following the Destruction – differed from one another in some lettering or some words, will find that this assertion is in no way revolutionary. I might apply to him the words of Iyov (12:9), 'Who knows not among all these…'! Admittedly, the Jewish sages of earlier generations could not have been aware of the full scope of the problem: perhaps they did not imagine the number of textual variants, or the extent of the discrepancies between different versions. However, this is of no fundamental importance, since the principle itself is what matters – this being that the early Sages were aware of the fact that the scribes were divided concerning the textual version, and the Sages decided among the alternative versions on the basis of the majority."[12]


Nevertheless, the version that is universally accepted today, with its very slight variations, is the version that was decided upon by the Masoretes and this, of course, is the version that is halakhically binding.


It must be asked - does all of the above open the door to proposing amendments to the text where our version gives rise to textual difficulties? This question is especially pertinent in those instances where the form of a verse in the Masoretic version is more problematic than its parallels in the other textual witnesses. Furthermore, what is the proper attitude towards a textual amendment that solves a textual difficulty, where the amendment has no basis in earlier textual witnesses?


The fundamental rule here would seem to be set forth by Rabbi David Zvi Hoffmann (Germany, 1843-1921), in the introduction to his commentary on Sefer Vayikra:


"Even if we acknowledge that certain places in the text are not free of error, we lack the necessary means to restore the version [originally] written with Divine inspiration. There is no textual amendment – even if it rests on arguments drawn from exegesis and from history – that can force us to believe that the prophet, or the author of the holy books, wrote the text in the exact form proposed by the amender."


This approach demands, first and foremost, that we try to understand the version that is before us. Even if we are aware that errors may have crept into the text, we have no way of proving this conclusively. For this reason, our responsibility is always to try to understand the version that we have, with its difficulties, even though we are aware of the theoretical possibility of textual errors.


In addition, it must be remembered that it often turns out that the seemingly less coherent or logical version is actually the accurate one. The reason for this is that the scribes would always aspire to produce clear and intelligible manuscripts, and where the ancient version was difficult to understand, they would tend to "amend" it to the best of their ability. In many cases we find that the amendment arose from a lack of comprehension on the part of the scribe, or from an attempt at independent, unfounded exegesis. It is for this reason that the general rule among scholars is, "lectio difficilior potior" – "the more difficult reading is the stronger [option].”


This is especially relevant in the realm of translation, where the content needs to be conveyed in a different language, and this transition is highly dependent on subjective interpretation, along with the objective challenges of translation, as noted in Massekhet Sofrim (1:7, Higger edition p. 101), concerning the Septuagint:


"That day was as vexing for Israel as the day when the golden calf was fashioned, for the full meaning of the Torah could not be translated."


In the past, since academic biblical scholars had little confidence in the accuracy of the Masoretic text, they would propose textual amendments on the slightest pretext. The work of some of these scholars gives the impression that there is not a single verse in Tanakh that is free of corruption. However, while textual amendments may be an easy and convenient solution for textual difficulties, they may turn out – and have often turned out – to be a superficial and unsatisfying solution.


In recent generations, the trend of exaggerated reliance on textual amendments has been somewhat put in check, and scholars are more cautious in proposing such solutions.[13] The proposed amendments often turn out to miss significant literary messages which the text manages to convey specifically by means of the seemingly more complicated wording. In-depth familiarity with the linguistic and literary character of the Tanakh shows that the text consistently and systematically uses various means to emphasize the messages conveyed by the content, and therefore the scholar needs to be sensitive to these means in order to attain a better understanding of the text. Today we also have a broader and deeper knowledge of the idiosyncrasies of ancient Hebrew, and forms of expression which in previous generations might have been dismissed as scribal errors, have been shown to preserve a certain tradition of writing or pronunciation. In the next shiur we will look at some examples of such instances, each representing a different type of proposed textual amendment to a textual difficulty which may turn out to reflect the deliberate literary expression of a concept or idea.


(To be continued)


Translated by Kaeren Fish


[1] Talshir, p. 52

[2] For more on the scrolls see Tov, pp. 80-93; Talshir pp. 56-67.

[3] For a discussion of the relationship between the Samarian Torah and the Masoretic text, see Tov, pp. 62-80; Talshir pp. 67-76; A. Tal and M. Florentin, Chamisha Chumshei Torah: Nussach Shomron ve-Nussach ha-Masora, Tel Aviv 5771, pp. 11-57.

[4] Regarding Chazal's attitude towards the Kuttim in general, see "kuttim,” Encyclopedia Talmudit 27, Jerusalem 5752, columns 649-730.

[5] For instance, in the Samaritan Torah the tenth Commandment concerns the building of an altar on Mount Gerizim. Similarly, the expression that appears repeatedly in Sefer Devarim, "the place which God will choose (yivchar)" (Devarim 12:5, 11, 21, and others) is replaced with the expression "the place which God has chosen (bachar)" – in other words, the place is already designated (this, of course, being understood as a reference to Mount Gerizim). Chazal note these ideological differences: In Devarim 11:30 there is a description of the exact location of Mount Gerizim and Mount Eval – "Are these not on the other side of the Jordan, by the way where the sun goes down, in the land of the Canaanites who dwell in the Arava over against Gilgal, beside Elonei Moreh.” The Samaritan version appends two words to the end of the verse: "mul Shekhem" (over against Shekhem) (Tal and Florentin edition, p. 559). Chazal heap scorn on this addendum: "R. Elazar son of R. Yossi said: In this matter I proved the Samarian scrolls to be false. I said to them, 'You have falsified your Torah but have gained nothing by doing so. For you say that 'Elonei Moreh' refers to Shekhem; we, too, agree that 'Elonei Moreh' is Shekhem. We infer this using the principle of 'gezerah shavah;' but [since you do not accept the sanctity and authority of the hermeneutical laws,] on what basis do you infer it?!" (Sifri Devarim, 56; see also Talmud Yerushalmi, Sota 7:3).

[6]Unlike the Masora, the Samaritan Torah has no uniform text, and the various manuscripts reflect different versions. These textual discrepancies between various Samaritan texts do not trouble the Samaritans, since their tradition places greater emphasis on an oral, rather than a written, tradition.

[7] On Tanakh translations, see, inter alia: Y. Komlos, Ha-Mikra be-Or ha-Targum, Tel Aviv 5744; C. Rabin, Targumei ha-Mikra, Jerusalem 5744.

[8] An extra-canonical work from the 2nd century B.C.E. and the earliest source relating to the translation of the Torah into Greek. The letter was written in Greek and several translations exist. Hebrew translations include that of A. Kahana, Ha-Sefarim ha-Chitzoniim, vol. 2, Tel Aviv 5697, pp. 1-71; A. S. Hartoum, Ha-Sefarim ha-Chitzoniim: Sippurim ve-Divrei Chokhma, Tel Aviv 5728. For more on the period of the letter and its author, as well as on the nature of the Septuagint, see the introduction to the above editions as well as Z. Doribel, "Al ha-Mekorot shel Targum ha-Shiv'im la-Torah,” Beit Mikra 50, 1 (180), 5765, pp. 3-19. An English translation of the letter can be read here:

[9] The Gemara makes no explicit mention of a translation into Greek, but it can be deduced both from the content of the narrative and from a comparison to the account in the Letter of Aristeas.

[10] Avot de-Rabbi Natan, version B, chapter 37, Schechter edition p. 94, and Massekhet Sofrim 1:7, Higger edition p. 101, mention "five elders.”

[11]Rabbinic sources recount how the king placed the elders in separate buildings and how each of them miraculously amended the biblical text in their translations in precisely the same way as one another, owing to different problems – mostly theological questions – that might have arisen from a literal translation. For instance, the verse "Let us (in the plural form) make man in our image and after our likeness" (Bereishit 1:26) might create the impression of a multiplicity of divinities, or of God's corporeality. For this reason each of the sages translated the verse "I shall make man in the form and in the likeness.” The various sources offer differing counts of the instances where the translators introduced amendments: Avot de-Rabbi Natan asserts "they changed ten things in the Torah"; Massekhet Sofrim puts the count at 13; and Massekhet Megilla 9a enumerates 15 amendments. The version of the Septuagint that we have today features only four of the changes mentioned (see Kahana p. 16, n. 6).

[12] Rabbi M. Breuer, "Emuna u-Mada' be-Nussach ha-Mikra,” reprinted in Y. Ofer (ed.), Shitat ha-Bechinot shel haRav Mordekhai Breuer, Alon Shvut 5765, pp. 71-72.

[13] As noted by Cohen, p. 69: "The demand to avoid imaginary speculation and to exercise caution in the examination and assessment of facts is heard increasingly today among scholars themselves as one of the fundamental conditions for sound exegetical conclusions."