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

  • Rav Amnon Bazak

E.        Development of the Masoretic text


Even after the Masora text of Ben Asher became the accepted version of the Tanakh, there remained many small discrepancies between the various manuscripts. For instance, from Rashi's commentary we see that the Tanakh text that he used, based on Ashkenazi manuscripts, differed in dozens of minute instances from the text of Ben Asher.[1] Let us consider one interesting example. In our Masoretic text, one of the prophecies of Yishayahu reads:


"On that day the Lord with his fierce and great and mighty sword will punish Leviatan, the flying serpent, and Leviatan, the crooked serpent, and will slay the crocodile that is in the sea… I, the Lord, guard it; I will water it every moment lest any punish it (pen yifkod), I will keep it night and day." (Yishayahu 27:1-3)


Rashi comments:


"'I will water it every moment' – little by little I will water it with the cup of punishment that will come upon it, lest I punish it (pen efkod) in a moment and consume it.”


His explanation indicates clearly that in the version he used the verse read, "pen efkod" (“lest I punish it”) and not "pen yifkod” (“lest any punish it”) Rashi's disciple and colleague, R. Yosef Kara, notes these two variants (in his commentary on verse 3):


"'Pen yifkod aleiha' – thus it is written in all the texts in Spain. Accordingly, the meaning of the verse is: 'I am the Lord Who guards it (Israel) from the crocodile, lest the crocodile punish them and make them like thorns and weeds.' In our texts, where the verse reads, 'pen efkod aleiha,’ the commentators have explained the verse as meaning, 'Lest I punish it in a moment – for if I were to punish their transgression every time they sin, I would consume them in a moment…' And only God knows which is the proper version."


The same phenomenon is to be found in the works of other commentators.[2] The manuscripts commonly used in the Middle Ages did not take pains to exactly mirror the Masora version, and overlooked thousands of tiny discrepancies.[3] A clear example of a version that has been preserved even though it contradicts the Masoretic version is to be found in the inclusion or omission of two verses in Sefer Yehoshua. Chapter 21 of Sefer Yehoshua deals with the cities of the Levites, and at the beginning of the chapter we find that the children of Merari received twelve cities from three tribes:


"The children of Merari by their families [received] from the tribe of Reuven and from the tribe of Gad and from the tribe of Zevulun – twelve cities." (verse 7)


Afterwards, however, the Masoretic text lists eight cities given to the children of Merari: four from the tribe of Zevulun (verses 34-35) and four from the tribe of Gad (36-37), followed by a concluding verse:


"All the cities for the children of Merari by their families, which remained of the families of the Levites, were by their lot twelve cities." (verse 38)


Clearly, there are verses missing here which should mention the four cities given by the tribe of Reuven. Indeed, in the parallel chapter in Divrei Ha-yamim, the four cities given from the tribe of Reuven do appear:


"And on the other side of the Yarden, by Yericho, in the east side of the Yarden, [they were given] from the tribes of Reuven – Betzer in the wilderness with its pasture lands, and Yahatz and its pasture lands. And Kedemot with its pasture lands and Mefa'at with its pasture lands." (Divrei Ha-yamim I 6:63-64)[4]


In his commentary on Yehoshua (21:7), Radak writes:


And there are versions (of Sefer Yehoshua) which have been redacted to include ‘And from the tribe of Reuven, (the city of) Betzer and her pasture lands, Yahatz and her pasture lands, Kedemot and her pasture kinds, and Mefa’at and her pasture lands – four cities.’ Yet I have not seen these two verses included in any ancient and authentic manuscript, rather they have been added to a small number of texts. And I saw that Rabbeinu Hai Gaon z”l had been asked regarding this, and responded that even though here (in Sefer Yehoshua) they (the cities of Reuven) are not enumerated, nevertheless in Divrei HaYamim they are enumerated. Thus it would appear from his response that the verses are not written in the authentic versions of Sefer Yehoshua.


Radak is aware of the existence of other manuscripts in which these verses appear, but believes that the textual version that includes them is not reliable and he refuses to accept it – even though it solves the difficulty posed by the Masoretic version. The author of Minchat Shai brings a different version of the first verse, in the name of "ancient manuscripts from Spain":


"And from the tribe of Reuven – the city of refuge for murderers, Betzer in the wilderness and its pasture lands and Yahatz and its pasture lands."[5]


It would therefore seem that the Masoretic version was not universally relied upon for all manuscripts, and in this instance the variant in the other manuscripts reflects a more logical version of the text.[6] In the various contemporary printed editions, these verses are sometimes incorporated in the text, or appear in the margins, while in other versions they are omitted.


Withthe advent of the printing press, the problem of discrepancies among manuscripts arising from errors ceased to exist, and most editions rely on the Mikraot Gedolot edition of the Tanakh, which was published in 1524-1525, based on the textual version established by the editor Yaakov ben Chaim Adonia. This edition, although much closer to the Masoretic version than the manuscripts that had been used in Germany, was still far from perfect, and sages such as R. Menachem de Lonzano, R. Eliyahu Bachur,[7] and R. Yedidia Norzi criticized its defects.[8]


Today there are different editions of the Tanakh available, based on manuscripts that have many slight differences.[9] There are nine known discrepancies between the text of Yemenite Torah scrolls (which follow the version of the Aleppo Codex) and that of Ashkenazi scrolls, including one difference that is expressed in the manner of reading: while the version in the Ashkenazi manuscripts starts the verse detailing the age of Noach when he died (Bereishit 9:29) with the word "va-yehi,” the Yemenite manuscripts read "va-yihyu.”[10]


These very slight discrepancies demonstrate the very careful transmission of the Masoretic text, but at the same time they also show that even with regard to the text of the Five Books of the Torah, there are some slight differences even today. Among the different printed editions that exist today there are a total of about 100 variants.[11] Obviously, the number of discrepancies in the Tanakh as a whole is greater than the number found in the Torah.


We may therefore summarize by saying that the Masoretic version is indeed accepted as authoritative, but since the Tanakh is such a remarkably complex work, including tens of thousands of details (letters, vowels, cantillation marks, etc.), in many instancesthe general acceptance of an authoritative version was not sufficient for it to be implemented with perfect accuracy throughout the Jewish Diaspora.


Translated by Kaeren Fish



Appendix – The Texts of the Medieval Commentators – Rashi, Rashbam, and Meiri


J.S. Penkower (see fn.1 above) examined manuscripts of Rashi's commentary and enumerates no less than 63 instances – in the Books of Bereishit and Shemot alone – in which the Torah text that Rashi used was different from the one familiar to us. To rule out the possibility that these discrepancies arose from errors in copying the commentary itself, Penkower took the trouble to investigate textual witnesses of the Bible from the Middle Ages, in which the verses that Rashi cites appear in the identical form to his quotes. Most of the variants involve the inclusion or omission of the letter 'vav' at the beginning of a word, but Penkower also notes four instances in Sefer Shemot in which Rashi's commentary contains a different word:

1.    Shemot 20:5 – The Masoretic version reads, "and performing (ve-oseh) kindness to the thousandth [generation]," while Rashi's version reads, "and preserving (ve-notzer) kindness to the thousandth";

2.    Shemot 23:18 – The Masoretic version reads, "You shall not offer the blood of My sacrifice with leavened bread," while Rashi's version reads, "You shall not slaughter [it] with leavened bread";

3.    Shemot 24:17 – The Masoretic version reads, "And the appearance of God's glory was like a consuming fire atop the mountain, in the sight of Bnei Yisrael," while Rashi's version (as he quotes it in his commentary on Shemot 34:30) reads, "in the sight of all of (kol) Yisrael";

4.    Shemot 26:24 – the Masoretic version reads, "And they shall be coupled together above the head of it (al rosho)," whereas Rashi's version reads, "el rosho."


The differences between the Biblical text used by Rashbam (Rashi’s grandson) and the Masoretic version has been researched by S. Japhet, (see fn.2 above). Sheproposes 23 instances where Rashbam's commentary suggests that his textual version differed from that of the Masora. In several places in his Commentary on the Torah, the Rashbam makes slight changes to the version that is his reference, and notes that he finds support for his commentary in the manuscripts of Spain (e.g. Shemot 23:24). Commenting on the verse that prohibits various types of witchcraft, including a "chover chever" (charmer) (Devarim 18:11), Rashbam argues that the emphasis on the second word should be on the first syllable of the word and not the second, with the cantillation mark on the letter 'chet' and not on the 'bet'. He adds, "And I found the cantilation sign on the 'chet' - as I maintain it should be - in Spanish manuscripts." It should be noted that in both instances the version that Rashbam proposes as the correct one, having arrived at it through his own logic, is the Masoretic version as we know it. See further in his commentary on Devarim 7:14.


Another interesting example is to be found in the commentary of R. Menachem ben Shelomo ha-Meiri on the first few mishnayot of the sixth chapter of Massekhet Shevu'ot. The Meiri discusses the halakhic concept of the "shevu'at heset" – an oath uttered by a defendant who rejects outright the claim made against him, where neither he nor the claimant has any evidence to show in support of their position. One of the possibilities that the Meiri raises is as follows: "Some explain that the oath is so named as a derivation of the term 'hasara' (to remove), as in, 'David removed them (va-yesitem) from upon him' – in other words, the later sages instituted this oath to be uttered by the defendant [in those cases where] he is not so required by Torah law, in order to put the mind of the claimant at rest." However, in our Masoretic text the verse cited reads, "va-yesirem" (Shemuel I 17:39).


Translated by Kaeren Fish


[1]See J.S. Penkower, "Nussach ha-Mikra she-Amad Lifnei Rashi," in A. Grossman and S. Japhet (eds.), Rashi – Demuto vi-Yetzirato, Jerusalem 5769, pp. 99-122, and appendix below.

[2]For instance, see S. Japhet, Dor Dor u-Parshanav, Jerusalem 5768, pp. 189-206 and appendix below).

[3]See M. Cohen, "Mavo le-Mahadurat ha-Keter," in Mikraot Gedolot – Ha-Keter: Yehoshua-Shoftim, Jerusalem 5752, p. 4.

[4]In the Septuagint, which we will discuss further on, the verses in Divrei Ha-yamim appear also in Sefer Yehoshua. However, it must be remembered that the Septuagint displays a tendency towards harmonization of the Tanakh, and we therefore cannot rule out the possibility that these verses were simply copied from Divrei Ha-yamim, rather than being an integral part of the textual version upon which the translation was based. See below in our discussion of the attitude towards translations of Tanakh.

[5] In his commentary on 13:26, Minchat Shai points out – correctly – that based on the number of verses comprising Sefer Yehoshua and the mid-point of the Sefer, according to the Masora, "the two verses that are to be found in some of the manuscripts in chapter 21 should be removed, since in any case the calculation is inaccurate."

[6] Concerning these manuscripts, too, it is of course possible that the additional verses were an addendum introduced on the basis of the version in Divrei Ha-yamim, rather than an ancient variant that had been preserved.

[7]R. Eliyahu Bachur, one of the most important grammarians and Masoretic scholars in the early 16th century, expressed praise for the Mikraot Gedolot edition, in the second introduction to his book Masoret ha-Masoret, Sulzbach 5531, declaring it the most orderly, accurate, and beautiful set of Tanakh volumes that he had ever seen, but he also expresses reservations concerning errors (as well as concerning the editor, Yaakov ben Chaim, who converted to Christianity at the end of his life).

[8]See Cohen (n. 3 above), pp. 10-11.

[9]The five best-known Hebrew editions are as follows: 1. The Koren Tanakh was first published in 1962 (5722), and its text was based on the editorial comments of E. D. Goldschmidt, A.M. Haberman, and M. Medan, "on the basis of the opinions of the Masoretes and of the grammarians and the commentators, and in accordance with the majority of the manuscripts and printed editions generally regarded as authoritative" – i.e., without relying on any specific manuscript as the textual basis. 2. The Dotan Tanakh was first published in 1973(5733), and is based on the Leningrad Codex. This important manuscript was written in Cairo in the year 1008, and served as the basis for many Tanakh editions. In the colophon the scribe notes that it was copied "from the checked manuscripts produced by the teacher Aharon ben Moshe Ben-Asher," but its text contains hundreds of discrepancies in relation to the Aleppo Codex. 3. The Breuer Tanakh is based on the Aleppo Codex and other very similar manuscripts, although Rabbi Breuer occasionally deviates from the Aleppo Codex version with regard to vowels. 4. Mikraot Gedolot Ha-Keter, which started to appear in 1992 (5752), edited by M. Cohen. This edition, too, is based on the Aleppo Codex, and remains more closely dependent on it without the changes introduced by Rabbi Breuer ("except for a handful of instances which are unquestionably corruptions"; M. Cohen, pp. 44). 5. The Simanim Tanakh, Jerusalem 2010 (5770), based for the most part on the Aleppo Codex, and where this version is lacking, on the Leningrad Codex.

[10] Concerning this difference and the different textual versions, see Minchat Shai,  ad loc. Aside from these differences there are also a few discrepancies in the division of the parashiot, as well as in the Song of the Sea and the Song of Haazinu, and in the division of letters, as for example the name "Poti Fera" (Bereishit 41:45, 50).

[11]In most cases, the differences arise from discrepancies between the Leningrad Codex and other manuscripts.