The discipline of textual criticism as developed over the last two centuries has become one of the pillars of modern Bible research and interpretation. In the field of scientific Bible study it is commonly accepted that one of the first questions to be addressed before real interpretation can be undertaken is the nature of the text itself and what changes it has undergone during the long course of its transmission. The assumption underlying this approach is, on the face of it, seemingly simple: the Scriptural text is an entity that has been handed down over the centuries and is therefore subject to the same sort of errors as any other transmitted text. To what extent the received text has been preserved in its original form is a question that can be examined by critical-philological means, as developed in general textual criticism.

But, due to the sanctity of the Holy Scriptures, the text-critical approach was shunned by religious students of the Bible, and its use as a tool of interpretation summarily dismissed. Even those scholars who were willing to adopt some aspects and conclusions of scientific Bible study stopped short of textual clarification in the scientific manner.[1]

The strong opposition to textual criticism stems from the feeling that such practice contradicts the accepted religious view of the sanctity of the text, with no possibility of reconciliation. This ideal, as rooted in popular perception, is generally given an historical interpretation: the Bible text, down to the last of its letters, reached us unchanged from the time of its authorship. This idea gained currency through the generations, thanks to Halakhic and Aggadic statements and writings in the area of Jewish thought. Therefore, any method that casts doubt on the absolute reliability of the transmitted text arouses instinctive rejection on the part of believing Jews.

Even so, in an age where scientific awareness has become second nature to many religious Jews and so many subjects are treated in light of science and religion together, there is reason to bring up the issue of textual criticism of the Bible for renewed discussion. From both practical and educational standpoints, it is unhealthy for Judaism, which has long recognized the value of scientific method and its ability to provide answers in the empirical field, to place any empirical topic beyond the pale for fear of confrontation with prevailing religious views. In the natural sciences intelligent religious Jewry has long ago overcome the barrier of confrontation between scientific conclusions and accepted religious beliefs. Orthodox scientists work with the assumption that binding religious authority cannot be granted to traditions and statements in areas which are subject to empirical scientific study. Does it make sense to be exclude any empirical field from this rule?

The history of the transmission of the Biblical text is definitely an empirical subject, for it has left in its trail textual evidence, much of it available for research and evaluation. Lending religious authority to the conventional understanding of the ideal about the sanctity of the consonantal text means an ongoing confrontation with the facts. Is continued adherence to the popular historical interpretation of the sanctity of the text, coupled with disregard for the available textual evidence, the only path contemporary Judaism may pursue?

1. The standard approach

Before we begin to analyze the textual evidence for the Scriptures and its implications, we will present the conventional understanding as outlined above by means of a typical formulation, this one given by Don Yitzchak Abarbanel at the end of the Middle Ages. In his introduction to Jeremiah he disagrees with the RaDaK and R. Yitzchak Ben Moshe Duran HaLevi (The Efodi) on the nature of the Ketib and the Qeri, the written vs. read text:

...I think it appropriate here to discuss the written vs. read text, [Ketib/Qeri] and the reason why, in the Torah, Prophets, and Writings, there are words which in the main text are written one way, while in the outer margins they appear in a different form, though there is no doubt that the prophet or Divinely-inspired speaker spoke with only one version, and not two.
The RaDaK commented about this and wrote: "It appears that these words are here because during the first Exile, books were misplaced and lost and scholars died; when the Great Assembly restored the Torah they found conflicting information in manuscripts and went according to the majority. When that was unsatisfactory they wrote the word without vocalization or wrote in the margins and not in the text, and thus they wrote one way inside the text and one in the margins..." The author of the Efod agrees....
The opinion that these scholars agreed on is incomprehensible to me, for how can I believe and how can I state that Ezra the Scribe found the true G-d-given Torah and the books of His prophets and Divine Hagiographa in a state of inaccuracy and confusion? A Torah scroll missing one letter is ritually invalid; how much more so a Qeri differing from the Written text (which would imply that sundry letters are missing). Our one consolation is that the Torah is with us in our exile, and if we agree with these scholars that the Torah has undergone a process of textual damage and confusion, we will have nothing left on which to rely...The eighth pillar of faith, as laid out by the great master Maimonides in his Mishnah commentary, requires every believer to accept that the Torah we have today is the same as that given to Moshe on Mount Sinai, with no changes whatsoever...

Abarbanel's words illustrate how the halakhic ramifications of the notion of the sanctity of the Bible text brought about an historical interpretation of the ideal. Since Jewish law states that a single incorrect letter invalidates a Torah scroll, we cannot, according to the Abarbanel, entertain the notion that any error whatsoever crept into the received text. That is why Abarbanel specifically expresses concern over the loss of faith in the authority of the Torah were we to suppose the text might not have reached us exactly as first written.

There are many other no less valid arguments that could be added to the Abarbanel's, including the halakhic and aggadic Midrash which is often based on the exact letter-text and which serves as the foundation for both the spiritual ideas and the practice of Judaism. Can we assume that these Midrashic comments on the text spring from an "incorrectly transmitted" Torah?

On the other hand we have the words of the RaDak and the Efodi to prove that the Abarbanel's arguments were neither the only possible explanation nor were his objections unanswerable. It is a fact that a leading medieval traditional commentator, R. David Kimhi, was able to entertain the possibility of an incorrect transmission of the text without undermining the basis of his faith.[2] But neither the approach of RaDak nor the Efodi won out in mainstream Judaism; the dominant approach was and remains that expressed by Abarbanel.

2. Testimony from the Second Temple period

Now let us look at the historical facts about the text. The relevant questions are:

What is the textual state of affairs that can be deduced from the evidence available for the various stages of textual transmission?
Was the notion of the sanctity of the text as now commonly held, prevalent always?
What was the relation between the idea of the text's immutable sanctity and the textual reality throughout history, and how did the great Torah authorities of each generation relate to it?

Even before the discovery of the Dead Sea Scrolls, however, theories about the Biblical text in the Second Temple period abounded. These theories drew inspiration from two texts whose roots lie in the same period: the Septuagint and the Samaritan version of the Bible. The Septuagint Vorlage (the presumed underlying Hebrew text) differs from the Masoretic Text [the received Hebrew text, the Authorized Text, the Jewish Bible, abbreviated MT] in many aspects, several of them of great significance. We cannot determine the exact Vorlage of the Septuagint, but it appears to have contained thousands of differences from the received text, some minor (conjunctive vav, prepositions, etc.) and some quite significant, including words, sentences, and even whole sections (an outstanding example is The Book of Jeremiah, which in the Septuagint is almost one eighth shorter than in the Masoretic Text).

The Samaritan text shows similar differences; in addition, since the text is in Hebrew, several thousand differences in spelling are apparent to the eye in the Five Books of the Pentateuch. The Samaritan text has distinctive features, and even though it holds almost two thousand differences in common with the Septuagint, it is in no way identical to the Septuagint; many of its changes are unique and in many places it differs from the Septuagint and agrees with the Masoretic Text.

Since the Septuagint is not a Hebrew text and the Samaritan version reached us through a breakaway sect, their value to reflect the early stages of the Biblical text was debatable. This battle was first waged between Catholic and Protestant scholars on theological grounds, and in the 19th and 20th centuries amongst Bible scholars against a scientific background. Some maintained that extreme caution must be exercised when using the Septuagint as a text-witness for an ancient Hebrew text-type fundamentally different from the Masoretic version; one must take into account the changes that were made in the course of translation for linguistic, exegetical, and interpretative reasons. There were also similar claims that the Samaritan text could not be taken as representing the general transmission of the Biblical text outside of the specific Samaritan recension; its variants do not reflect the earlier text-form which the community had adopted, but are changes that were made within the closed frame of the Samaritan community.

On the other side, insofar as the Septuagint is concerned, many signs point in the direction of a different Hebrew Vorlage, at least for several of the Biblical books. A parallel claim has been made for the Samaritan version-- that it mirrors a vulgar text-type common in the Second Temple period in the general Jewish community which was then adopted by the Samaritans, who merely added a few "ideological" changes (such as the "Tenth Commandment").

The Dead Sea scrolls decided these issues, by showing that there was indeed a Hebrew text-type on which the Septuagint-translation was based and which differed substantially from the received MT. These findings also confirmed that most of the textual phenomena in the Samaritan version (aside from ideological changes) were part of a Hebrew text-type in common use outside of the Samaritan community as well, during the Second Temple period in the Land of Israel.[3]

The following conclusions about the state of the text in the Second Temple period can be drawn from the evidence of the scrolls:

The people of Qumran (apparently the Essene sect) had access to several text-types of the Scriptural text: In essence, all Hebrew text-types represented by the MT as well as those behind the Septuagint and the Samaritan text, served the sect in its study and worship.
Beside the multiplicity of text-types described above, the Qumran scrolls show a variety of sub-types: within each text-type variations appeared, such as orthographic variants, conjunctive vav [present/absent], variant words, and grammatical forms. These variants did not remove the specific text from its general text-type, but they created sub-groups within the text-type.
Even the Masoretic text-type found at Qumran is not identical, letter for letter, with the Masoretic text we know today. Only in very rare instances can it be said that the Qumran scribes copied from a consonantal base close to the MT in use today. In most cases, the similarity between the Qumran texts and the MT are in the general lines, not in the orthographic identity.[4] This shows that the Qumran scribes did not have the notion of an "ideal" text in the exact image of the current MT.
The lack of an authorized MT at Qumran can also be proven from the correction-methods used by the scribes. Corrections and additions can be found in many of the scrolls and it is obvious that they are not based on a single ideal text which was letter-perfect. Rather, they are based on the scribe's personal taste and his discretion to choose between the variant readings.[5] Often, the later hand in the manuscript did not mean to correct the original version, but to mark the existence of an alternative reading and to preserve it. In this fashion, many doublets were created.[6]
Illuminating in this regard is the writing of tefillin and mezuzot at Qumran. The exacting halakhic requirements that govern the writing of Torah scrolls, tefillin, and mezuzot, so familiar from later periods, is not present at Qumran; a missing or superfluous letter-or even word-did not invalidate tefillin or mezuzot. In the tefillin and mezuzot of the Qumran dwellers we find many kinds of orthography, and some variants are actually evidence for different text-types.[7]

We can conclude that the textual situation at Qumran was substantially different from the one we know today. There were many different consonantal texts, both variants within a particular text-type group and also multiple text-types, and there are no signs that the people of Qumran entertained the idea of one single version that was a fixed, sanctified text. The clear-cut belief of the Qumran people in the Scriptural message as the words of a living G-d, eternally vital and always meaningful, was not dependent on any notion of the sanctity of one particular consonantal text.

Can we draw a comparison between the situation in the Qumran community and that in the entire Land of Israel during that era, or was the Qumran reality unique to one separatist sect?

It seems that there is no reason to differentiate between the situation in Qumran and that of other places. The Qumran sect was composed of members who came from throughout the land, and therefore it must be supposed that its salient phenomena reflected the situation of Jewry as a whole. No one would suggest that the wide variety of text-types was created within the Qumran community. The Hebrew vorlage of the Septuagint text-type was undoubtedly used by the Jews of Alexandria in the late centuries BCE, as this was the version chosen for the Greek translation. The above-mentioned "Samaritan text-type" found at Qumran was also common in the Land of Israel, adopted by the Samaritans who added their ideological changes to that version. It can also now be proven beyond doubt that the author of Chronicles used a version of Samuel different from the MT and closer to the Lucianic version of the Septuagint, whose Hebrew prototype was found at Qumran.[8]

All the evidence we possess points to textual pluralism in the Second Temple era, as opposed to the notion of a single sacred consonantal text as later conceived. Even so, it might still be possible to distinguish between different groups within Jewry, not as concerns the very existence of the above- mentioned pluralism, but regarding the length of its existence and the efforts to change the situation. There are several signs that Pharisaic circles attempted to reject the multiple text-types long before the destruction of the Temple, while at Qumran there are no such signs until close to the destruction of the Temple, when the sect ceased to exist.

3. The transition from multiple text-types to a single text-type

Contra Apion

specific

We can now adduce textual proofs to these indirect proofs, and they testify that the Pharisees and Zealots possessed, for some of the books of the Bible, a version that can be identified with the current MT. In the ruins of Masada were found remnants of Biblical texts, brought there by the final defenders, which show clear ties with the present-day MT. We can therefore safely state that the move towards textual unification and the establishment of a definitive text preceded the destruction of the Temple, though the exact dates, as well as the measure of success in that early period, can not be determined.

In any case, it seems that after the destruction the array of text-types disappeared from normative Judaism, and the Masoretic type alone remained. This conclusion is bolstered by all textual evidence of the era, whether direct or indirect, both original and in translation. In the fifties, remnants of Scriptural scrolls used by Bar Kochba's soldiers were found in the Judean desert (Wadi Murabba'at and Nahal Hever). They all show that Bar Kochba's people used the same text which we call the MT, with only the slightest of differences. During the same period, new Greek translations were being prepared in place of the Septuagint, which, by virtue of its becoming an official Christian text, was rejected by the Jews. These translations, especially that of Aqilas which was praised by the Sages, reflected the Masoretic text-type. Likewise, the Aramaic translations such as Onqelos, Targum Jonathan, and the Palestinian Targum, whose roots date back to that same period, reflect a common text-type.

The most impressive testimony of the period for an authoritative consonantal text and for the idea of a sanctified consonantal text is the monumental work of the Sages: the Mishna, Baraita, Tosefta, Midrash, and Talmuds. All are based on the notion of a consistent text with an exact set of letters. This can be learned from direct statements (such as, "Scribal tradition [Masorah] is a fence around the Torah"), the detailed rules for the precise writing of Torah-scrolls, tefillin, and mezuzot, and the derashot themselves which are based on a precise orthography (in particular, the derashot of Rabbi Akiva).

It is important to try and visualize the swift process of transition from the multiple-text situation which characterized the Second Temple period to the single text-type which characterizes the period after the destruction of the Temple. This will make it easier to understand the dynamics of the transmission of the text from the destruction of the Temple onward.

It can be said that the unification of the text was hastened by two parallel processes: (1) rejection and removal of "deviant" text-types like the Septuagint and the Samaritan texts, which left the MT as the single legitimate text-type; (2) the formulation of one particular consonantal text and its prevalence in as wide a circle of transmission as possible.

A realistic examination of matters shows that the first process was the main cause for the relatively sudden and swift changeover to the single-text-type reality; the second process rapidly spread the notion of a sacred consonantal text, but it did not succeed in uprooting the variety of sub-types which existed within the MT framework even before an official text was fixed. The battle between the Authorized Text or the textus receptus and other shadings of this text-type continued another 1,500 years, until the era of print.

Specifically: the speed and ease with which the battle against the Septuagint and the Samaritan text was waged is to be explained as follows: (1) the motives behind the struggle were, apparently, ideological and polemical, aimed at deviant groups (Christians, Samaritans) who had adopted these text-types; (2) these texts differed from the MT substantially, sometimes even in whole sentences and sections, and therefore no great expertise in the consonantal text was required to identify one of them and to remove it from use; (3) there was an available alternative to the rejected texts: the MT. The Qumran discoveries proved, as previously mentioned, that this text-type was popular, available, and within reach of the Jewish communities. Rejection of the other text-types thus did not leave a vacuum; there were now more copies of MT available or easily obtainable to fill the vacuum.

It was a much more difficult task to establish a single unique consonantal text, for the following reasons:

Establishing this sort of version involves a battle not only against the rejected text-types, but also against many texts of the MT group itself -- those which have alternate spellings, changes in prepositional prefixes, differing grammatical forms, etc. Since we are speaking of a precise letter-sequence, propagating the Authorized Text was dependent upon the actual availability of the one precise text and expertise in identifying this one proper version. As will be seen, during the lengthy period of manual text-transmission until the first printed editions were produced, these conditions were never more than partially fulfilled.
Add to this the fact that the struggle against texts of the MT type was not an ideological/polemical fight against deviant groups, but served only to unify the MT from within. To convince Jewish communities to abandon their Torah scrolls or other Biblical books in favor of what is proclaimed to be the sole approved version, requires that the approved version be physically present, and that the local scribes be aware of its preferred and uniquely authorized status. This was not self-evident, as the history of textual transmission shows.

4. Ideal and reality in the era of the Sages

We have posited that the process of arriving at a single version of the Biblical text which took place around the time of the Destruction (70 CE) was primarily a transition from textual multiplicity to a single text-type and, only in a much more limited fashion, to a specific consonantal base. But the idea that there existed a text sanctified in its consonantal base, a notion which developed following the establishment of an authorized version (MT) and the efforts to propagate it, quickly spread throughout the Jewish communities via supportive statements in Halakha, Aggada, and Jewish thought, quite independent of the text which produced it. The ideal of "the perfect text" caught on and spread much more rapidly than the Authorized Text itself, and the lack of synchronization between the ideal and the real left its imprint on the entire subsequent history of textual transmission. What ensued was the following odd situation: a single recension, MT, which actually included diverse texts, side by side with the ideal of a single sanctified text. Within this framework the struggle continued between the authoritative text, which reached more and more circles, and the other texts within the MT type, a struggle not resolved until the era of print.

This reading of the situation is affirmed mainly by the textual witnesses mentioned above. All texts from the period immediately following the adoption of the MT show the predominance of this type, but not all mirror the single textus receptus. A glance at contemporary Greek translations, written to replace the Septuagint among Jews, shows that though they reflect the consonantal base of the Masoretic type, they can in no way be identified with what is currently known as "the" Masoretic Text. Though it is impossible to know what the spellings were in their Hebrew vorlage (and the spellings are an important component in the Masoretic recension), the aggregate of known differences in the Greek translations is enough to rule out the possibility that we have before us today's Masoretic Text. The same can be said of the various Aramaic translations; the differences they reflect are too numerous for us to class their vorlage as our Masoretic Text.

In addition, even Rabbinic literature, whose techniques cannot be understood without presupposing a unique consonantal text, paradoxically point to a variety of sub-types within the framework of the Masoretic type. This conclusion can be drawn both from descriptions in this literature of variant texts existing side by side, as well as from an examination of Biblical spellings on which the various Sages based their inferences of Aggada and Halakha.

Important information about a Torah scroll from the end of the Second Temple- period which contained several variants from the MT has been preserved in Midrash Bereshit Rabbati (Albeck edition, page 209 ff). The Midrash details a substantial list of differences, all of which fit into the framework of the Masoretic text-type, but not the MT. According to the story, this was an important Torah scroll which had been in Jerusalem during the Destruction (perhaps in the Temple?), and which the Romans took back with them to Rome. At a later period, some differences in this Torah scroll were attributed to the Torah scroll of Rabbi Meir (Midrash Bereshit Rabbah, Parsha IX, sec. 5 [Theodor-Albeck ed. 70,16]; Parsha XX, sec. 12 [ib. 196₄]; Parsha XCIV, sec. 6 [ib. 1181₇]). It seems that some of these variant readings survived into the Mishnaic period.

At a later period, in Babylonia, we find an explicit admission that the Talmudic Sages were not expert in the orthography of the established text. Bavli Kiddushin (30a) states: "The first scholars were called "soferim" (counters) for they would count the letters in the Torah, saying: "the vav of gahon (Lev.11:42) is the middle letter of the Torah, darosh darash the middle word(s), and wehitgallah (Lev.13:33) the middle verse." Rav Yosef asked, "The vav of gahon, on which side is it?" He was told, "Let a Torah scroll be brought and a count made." Rabba Bar Bar Hanna said, "They did not move until a Torah scroll was brought and a count undertaken." He said to them, "They [=the Tannaim of Eretz Israel] were expert in plene/defective spellings, we [=the Babylonians] are not, etc."

Rav Yosef recognized that the Sages of Babylonia were not experts on plene and defective spellings, and their scrolls were not identical letter-for-letter with the Jerusalem texts. Perhaps the most interesting part of this story is that the letter count, attributed to the original Jerusalem scribes, does not reflect the situation in the MT today, in which the middle letter is not the vav of gahon but the aleph of hu (Lev. 8:28), far from the word gahon.

A check of midrashim based on wordplays shows that in many cases the Biblical texts which the Sages used for homiletical purposes had a different consonantal base than our MT. There are hundreds of instances in which a midrash depends on the appearance of one letter more or less, and sometimes even on entire words whose spelling differs from the Masoretic text. This holds true not only for Aggadic Midrash, but also for the Midrash Halakha (such as the midrash based on the vav of [Dt. 11:18] adduced in Bavli Sanhedrin [4b], which is spelled defectively in our text).[9] These facts illustrate the difficulties in establishing the MT recension even within the circle of Sages.

Yet there is no doubt that all the Sages, translators, and scribes of this period labored by virtue of the ideal of a sacred consonantal text. The very existence of midrashim based on the exact spelling proves this, as do the direct statements on the topic. So, for example, the words of Rabbi Yishmael to Rabbi Meir: "Be careful in your work [R. Meir was a Torah scribe], since it is the work of Heaven; for if you delete a single letter or add one, you destroy the entire world" (Bavli Eruvin 13a). The very real gap between the ideal of one sacred text and the textual reality was not perceptible in the Rabbinic statements or in the subjective feelings of people. So long as one does not actually compare Biblical texts and is not aware of the existence of texts different from his own, he has no sense of the contradiction between the ideal and the real.

Though the decision to choose one consonantal text and promulgate it was an activity carried out by the circle of the Sages, there is no reason to suppose that all the Sages in the Mishnaic and Talmudic periods were experts in textual studies to the extent that they could decide if the text which they were using was "letter perfect". By way of analogy, we have much information on this subject from the medieval period, and it appears that very few Torah scholars of those days were proficient in issues of text and transmission. There is no reason to suppose that this situation was any different in the Mishnaic and Talmudic periods. Therefore, as long as the Sage making the derasha was unaware that the Bible text in his hands was not identical with the authorized version, his assumed that his exegesis was in accord with the ideal of one sacred text.

Unfortunately, we have only indirect witnesses to the textual situation in the Mishnaic and Talmudic period in the Land of Israel and Babylonia. Aside from those few remnants of manuscripts from Wadi Murabba'at and Nahal Hever which allow us but a glimpse into the milieu of the Bible manuscripts of that period, we can neither profile the textus receptus of the period nor sketch the outlines of the Masoretic text-type. This lack of direct textual evidence continues right up to the ninth century (perhaps a bit earlier, if we accept the earliest dating postulated for Bible fragments in the Cairo Genizah). From then on we possess a few manuscripts through the eleventh century and many more from the eleventh century until the era of printing.

These manuscripts show that during the preceding centuries, intensive efforts were made in the Land of Israel and Babylonia to preserve the precise form of the Received Text and to reject other versions of the MT text-type. These efforts were made over the course of generations by people who devoted their time to developing a variety of mechanisms which would preserve all the details of the Masoretic text. These people came to be known as the Ba'ale Ha-Masorah, "Masters of the Transmission", and their complicated devices to preserve the textus receptus are known as the Masorah, after which the Authorized Text has come to be known as the "Masoretic Text."

5. Ideal, masorah, and textual reality in the Middle Ages

The Masorah is a system of annotations and instructions whose purpose was to preserve the consonantal text, and, at a later period, also its pronunciation and cantillation. The Masoretes counted the letters, plene and defective spellings, as well as other phenomena, and transmitted the information from generation to generation. The transmission was first in the form of an oral tradition, then written, and finally they entered the essentials on the Codex-page itself to serve as a guide to copyists. Slowly a new form of transmission crystallized which featured [A] vocalization and accents, [B] the inclusion of annotations comprising the Masorah parva ("small Masorah") and the Masorah magna.[10] Manuscripts of this sort are termed "Masorah codices"; they are fully-developed from the ninth century onwards.

Were the Masoretes successful? Did they achieve the end for which the Masorah was developed and did a single Authorized Text finally replace the variety of texts within the MT group? A surface scan would seem to point to an endorsement of such an assumption: textual evidence in the Middle Ages unequivocally shows the spread and adoption of the Masoretic tradition throughout the Jewish world, and Masorah-codices became the accepted transmission model everywhere. The Masorah's prestige as the decisive authority in textual matters was never questioned in any Jewish community.

A deeper examination of the facts shows, however, that even the enormous activity of the Masorah circles was not sufficient to significantly alter the textual situation which had developed throughout the Diaspora after the Destruction. Variant consonantal bases within the MT group which differed from the Received Text continued to flourish right up until the advent of printing. It is apparent that under certain conditions these inner variants could coexist not only with the idea of a single hallowed consonantal text, but even together with the actual textual preservation mechanism of the Masorah.

We will illustrate this state of affairs by briefly examining the textual situation in four centers of Scriptural transmission which were also important centers of Medieval Jewry: the Land of Israel, Babylonia, Spain, and Ashkenaz.

Surviving MSS from the early Middle Ages (the ninth through eleventh centuries C.E.) come mainly from the Land of Israel and Babylonia. Among the manuscripts, two are attributed to the last of the Masoretes, the Ben-Asher family: the Cairo manuscript of the Prophets, attributed to Moshe Ben-Asher (late ninth century) and the Keter Aram Sova [=Aleppo Codex] attributed to his son, the most famous member of the family, Aharon Ben-Asher. It was apparently upon this MS that Maimonides based his laws of Torah scroll writing (Mishne Torah, Hilkhot Sefer Torah, 8,4). We also have several MSS from the same period akin to these two typologically and it is likely that this group of ancient manuscripts represents the Tiberian school of Masorah in its prime. Some of the manuscripts contain or did contain the entire Scriptures (Keter Aram Sova, MS Leningrad, MS Sassoon 1053), while others contained only parts of Scripture (such as the MS Cairo of the Prophets).

Aside from manuscripts of the Tiberian school, we also have fragments of other Palestinian manuscripts as well as Babylonian ones, mainly from the Cairo Genizah. These manuscripts are important for their vocalization signs, which allowed us to rediscover non-Tiberian systems of pointing that had gone to oblivion following the "victory" of the Tiberian signs and system. No less important than the vocalization are their consonantal texts, which testified to the continuation of Masoretic sub-types in the Land of Israel and Babylonia during this period.

If we were to judge according to the statistical findings in the available manuscripts of that period we would see that the circle of transmission of the Received Text did succeed in gaining a certain foothold in both the Land of Israel and Babylonia during the first millennium CE, but it did not utterly uproot the other texts of the Masoretic group which lived on in differing Jewish communities.

Furthermore, from the Land of Israel and Babylonian findings we can infer that even in places where the Authorized Text held sway, the "defeated" texts left their traces in the "victorious" text. Thus, even within the transmission tradition of the Authorized Text there remained a slender swath of variants, as can be seen via a comparison between the best texts of MT. No one model copy is identical to any other, and the variants between one copy and the next amount to a few hundred over all of Scriptures.

The most well-known differences within the bounds of the textus receptus are to be found between the Palestinian and the Babylonian manuscripts. Some of these variants were listed by the Masoretes themselves, in their notations of the differences between ma'arva'e (Westerners, i.e. Palestine) and madinha'e (Easterners or Babylonians) and in the annotations of the Masorah parva, as found in various manuscripts. Additional differences can be found via collations of the superior Babylonian and Palestinian texts discovered in the Cairo genizah. But differences also exist in sundry texts of Babylonian provenance and of Palestinian origin as well.

So we find variants between the two schools of Nehardea and Sura in Babylonia, which were recorded in the small Masorah and likewise between different schools in Erez Israel. A check of surviving MSS bears out this conclusion.

Alongside the limited range of differences between manuscripts of the MT, we find a wider range of differences in Babylonian and Palestinian manuscripts which lie outside this tradition. We can prove that most of these changes are not the result of copyist's errors but rather the remnants of texts which were at one time part of the Masoretic text-type and which managed to survive in certain places, despite the effort to establish the Authorized Text exclusively. These manuscripts are the direct successors of the text-types in the Mishna-Talmud period, which we described above on the basis of indirect testimony of that period. These variant manuscripts reaffirm the great difficulties encountered by the MT to establish itself exclusively even after the ideal of a single sanctified consonantal text was adapted by all Jews.

We possess a large and ever-growing number of Bible manuscripts from centers of Diaspora Jewry from the twelfth century on. Some of them comprise the full text of Scriptures, while some contain sundry books of the Bible; their numbers reach into the thousands. Most of the manuscripts are from the main centers of Jewry in Ashkenaz and Spain; the minority from other locations (Italy, North Africa, the Land of Israel, Yemen, etc.). This wealth of findings allows us to test some aspects of the transmission-tradition which we could not check in earlier periods because of the paucity of the evidence. These include: the extent of MT's dominance in various Jewish communities, the various outcomes of the struggle between the MT textus receptus and other texts of the Masoretic text-type, the reciprocal relationship between the notion of a sanctified consonantal text and the actual situation of textual multiplicity, and the circulation of Masorah literature and the influence of these books to eliminate textual diversity.

It appears that the rate of success in eliminating diverse text-types varied from community to community. In Spain, for example, the situation was fundamentally similar to that of the Land of Israel and Babylonia, with the Authorized Text in ascendance, but without completely eliminating the various sub-texts. In Spain hakhme Masorah, "Masoretic experts" clarified various aspects of the text and directly influenced the local scribal traditions. It is not for nought that the phrase "The exacting books of Spain" (sifre Sefarad ha-meduyaqim) was coined in the late Middle Ages, reflecting the high esteem accorded to Bible texts which emanated from the Spanish scribes. Especially known for his expertise and authority in Masorah matters was Rabbi Meir Ben Todros HaLevi (The RaMaH) of the thirteenth century, whose book Masoret Seyag LaTorah, along with the Torah scroll he wrote, served as a revered model for many scribes in and outside of Spain.

Even so, other textual traditions whose consonantal bases were modeled on earlier versions of the Masoretic text-type lived on in Spain, and a number of scribes continued to write texts, particularly Prophets and Writings, on the basis of these traditions. But the indirect influence of the main trend towards the Authorized Text is seen in the gradual shift of divergent texts which developed into something resembling the Authorized Text more and more. It is clear that the scribes themselves did not know exactly what the Authorized Text was, as their work still shows a consonantal base modeled after other texts of the Masorah group, but the very fact that their writing takes place on the margins of a strong MT transmission tradition blurred many of the variant features in the text-traditions which lay at the base of their work.

One of the main reasons for this process of convergence toward the Authorized Text was the Masorah apparatus. The prestige of the Masorah material cut across all Jewish groups, and the text prototype contained in the Masorah codices, produced within the milieu of the MT, spread rapidly throughout the Diaspora, and was adopted even in those transmission circles which held on to a Bible text that differed from the Masoretic Text. Of course, in the model manuscript codices of the textus receptus, the Masorah and the Biblical text are two sides of the same coin, each feeding and being fed by the other, while in the codices of the Masorah-group written outside the sphere of the Authorized Text, there is an artificial conjoining of the Masorah apparatus copied onto the page together with a variant local text of the Bible. The results of this "meeting" are interesting and instructive.

A Masorah codex (mashaf) is generally the work of two people, the copyist-scribe and the masorete/vocalizer. Each brings his materials from a text accepted by him. In places where scribes were unfamiliar with the MT consonantal text, a scribe would copy the consonantal text from an available Codex or scroll and give it to the masorete so that he might add the pointing and the Masoretic annotations. The masorete had his own sources and might take his Masoretic material from a Masorah-Codex or from his own Masoretic compendiums. Entering the Masorah on the folio automatically creates an open confrontation between the annotations of the Masorah Parva, written adjacent to the words to which they refer, and the variant spellings in the consonantal text of the scribe. Since everyone gave preference to the accuracy of the Masorah, the masorete generally corrected the orthography of the Biblical text to reflect the Masorah.

These corrections, of course, bring the text of the Bible closer in line to the authorized version, but a check of dozens of such Masorah-codices proves that it is only a partial convergence. Many of the masoretes working in the Middle Ages were merely copying "technicians" and their expertise was not sufficient to bring the scribe's text fully in line with the MT. The corrections are generally limited to instances of immediate and obvious confrontation between a Masoretic annotation and the spelling of a word; the masorete did not bother to check out additional words in other parts of the Bible to which the masoretic note referred and certainly did not attempt to elucidate the full meaning of the Masorah Magna comments.

We thus conclude that the Masorah, which was created as a tool for preserving and establishing the Authorized Text, often could not fulfill its purpose due to a flaw in the human component-- the copyist's lack of expertise. Even so, the Masorah apparatus did lead to differing levels of convergence with the Authorized Text. In Spain, for example, the process of textual unification was more intensive and encompassing because the many scribes and experts in issues of Masorah and text created a strong basis for transmission of the Authorized Text. On the other hand, in Ashkenaz, which had few such experts, the process of unification was slow and limited and had relatively little influence on the state of textual multiplicity.

In effect it can be said that the textual situation in Medieval Ashkenaz was largely based on text-traditions extraneous to the MT, local traditions passed from one generation of scribes to the next. Most of the German manuscripts contain a large number of orthographic variants from the MT, and it can be shown that these variants are generally not the mistakes of one scribe or another, but rather a genuine "inheritance," stemming from earlier versions belonging to the Masoretic text-type. These text-traditions were not created in Ashkenaz but were brought from the Land of Israel and Babylonia, and there is a genetic relationship between them and the "unauthorized" Palestinian texts which were found in the Cairo genizah, as well as between them and Spanish manuscripts of the same type.[11]

The situation in Ashkenaz serves as a perfect example for the continued prevalence of various consonantal texts of the Masoretic text-type in many places throughout the Jewish world, even after the adoption of a standard text and the firm establishment of the notion of its sanctified status. There is no doubt that the scribes of Ashkenaz, just like their colleagues everywhere, on the basis of profound conviction in the sanctity of the consonantal base, copied from their model texts with the same diligence as their colleagues who worked from the MT. Paradoxically, it appears that under certain conditions the ideal of a sanctified text actually strengthened the variant text traditions, so long as the scribes were not aware that the ideal text was not identical with the one they were copying from. Indeed in the lands of Ashkenaz, most scribes were unaware of this.

This situation in Ashkenaz continued until the beginning of printing. Even some of the first printed editions of the Scriptures were printed according to Ashkenazic text-models, by German editors who had emigrated to Italy. A prime example is the Soncino version of the Prophets of 1485-6 and of the entire Bible by the same editors, dated 1488. The editor's approach to textual accuracy is reflected in his introduction to the edition of Prophets: "One needs to testify only to that which is unknown. But this book is open and accessible to all and therefore needs no testimony from me as to its accuracy. Nevertheless, for those who may not have time to look through this work properly, I certify that this Bible was proofed and checked by experts knowledgeable in the Biblical text and the Masorah. No one will be able to find mistakes, in content or spelling. The closest thing to errors may be exchanged letters ( for , for , etc.) which the proofreader might have missed since these letters resemble one another. Sometimes a single letter is skipped in a word, but even these will hardly be found here. We have done the utmost to make this work perfect....We are perfectly certain that there is none among the codices written with the pen as correct as these printed copies."

This paragraph is imbued with the conviction that it is necessary to give readers a faithful and accurate text of the Scriptures. The editors therefore employ proofreaders who are "knowledgeable and intelligent" (Dan. 1:4) and they are certain that, though there be some unavoidable errors in minutiae, they have issued a first-rate work which exceeds the accuracy of even the most scrupulously written codices available. This conviction results from the notion of a sanctified consonantal text, which motivated all those involved in transmitting the MT.

But what is the truth in those transmission circles in which these editions were created? The editor tells us that he had manuscripts which are "accurate copies which have been studied for days and years." He bases his printed edition on a tradition of manuscripts which in his circle had been considered reliable for quite some time. How accurate was this tradition? Comparing the two above-mentioned Soncino editions with model codices of the MT shows an immense number of variants-sometimes scores in a single chapter, and tens of thousands throughout the Scriptures. These variants are generally not the mistakes of habochur hazetzer, "the young and inexperienced typesetter," as the editor would have it in his introduction, but reflect the tradition of the Ashkenazi manuscripts. A comparison of the Soncino consonantal text with many manuscripts of German provenance proves the strong genetic ties between them.

The penetration of the Masorah apparatus into the Ashkenazi circle of transmission and the adoption of Masorah codices as models for copyists even in Ashkenaz could not substantially alter textual facts in this region. We have explained above that a complex mechanism of preservation such as the Masorah cannot fundamentally alter divergent traditions unless it is applied by Masoretic scholars who are expert in the consonantal base of the Authorized Text and able to use the Masorah to make decisions in cases of doubt. There were very few of those experts in Ashkenaz and they could not alter the basic situation. Even so, we do find a certain convergence of manuscripts towards the Authorized Text as a result of cumulative-albeit partial-revision by Masoretes.

The event which substantially changed the textual situation throughout the Diaspora (including Ashkenaz) and led to the final victory of the Authorized Text was the second edition of Miqraot Gedolot, printed by Ya'akov Ben Chaim in Venice, 1524/5. This edition served as the basis for subsequent printed editions. Though a narrow margin of variants has continued to appear in each edition from then until today, the number of changes does not exceed the amount found in the best codices of the Authorized Text throughout its transmission, and they do not impair the status of any printed edition as representative of the Authorized Text of MT.

The attitude of Medieval scholars to textual multiplicity

The existence of tens of thousands of variants in text-traditions of the Bible should be cause for wonder-why didn't scholars and sages of that period point to this reality as a fundamental religious problem, for it seems to contradict the accepted historic notion of a single sanctified text? How could textual multiplicity be compatible with faith in the accuracy of the Masoretic transmission? A short discussion of the issue is appropriate before we examine the stand Judaism takes today on the subject of textual corrections in the Bible.

First we will present three quotes from three leading medieval Torah authorities: Maimonides (in the east), R. Meir Ben Todros HaLevi (RaMaH, in Spain), and R. Yom-Tov Lippman Milhausen (in Ashkenaz). All three, by virtue of their involvement in halakhic questions about the writing of Torah scrolls, were well-acquainted with the situation as it was:

Maimonides (Rambam), Hilkhot Sefer Torah 8, 4:
Since I have seen great confusion in all the scrolls [of the Law] in these matters, and also the Masoretes who wrote [special works] to make known [which sections are] "open" and "closed" contradict each other, according to the books on which they based themselves, I took it upon myself to set down here all the sections of the Law, and the forms of the Songs [i.e. Ex.15, Deut.32], so as to correct the scrolls accordingly. The copy on which we based ourselves in these matters is the one known in Egypt, which contains the whole Bible, which was formerly in Jerusalem [serving to correct copies according to it]. Everybody accepted it as authoritative, for Ben Asher corrected it many times. And I used it as the basis for the copy of the Torah Scroll which I wrote according to the Halakha.
RaMaH (R. Meir Ben Todros HaLevi) in his introduction to Masoret Seyag LaTorah:
...All the more so now that due to our sins, the following verse has been fulfilled amongst us, "Therefore, behold, I will again do a marvelous work among this people, Even a marvelous work and a wonder; And the wisdom of their wise men shall perish, And the prudence of their prudent men shall be hid"(Is. 29:14). If we seek to rely on the proofread scrolls in our possession, they are also in great disaccord. Were it not for the Masorah which serves as a fence around the Torah, almost no one would find his way in the controversies between the scrolls. Even the Masorah is not free from dispute, and there are several instances disputed [among the Masorah manuscripts], but not as many as among the scrolls. If a man wishes to write a halakhically "kosher" scroll, he will stumble on the plene and defective spellings and grope like a blind man through a fog of controversy; he will not succeed. Even if he seeks the aid of someone knowledgeable, he will not find such a one. When I, R. Meir HaLevi Ben Todros of Spain, saw what had befallen the scrolls, the Masorah lists, and the plene and defective spelling traditions, due to the ravages of time, I felt the need to search after the most precise and proofread codices and the most reliable Masoretic traditions, to resolve the conflicts. The newly-produced scrolls should be abandoned in favor of older, more faithful ones and among these the majority of texts should be followed as commanded in the Torah to decide any controversy, as it is written: "After the multitude to do..."(Ex. 23:2).
R. Yom Tov Lippman Milhausen, in his work Tikkun Sefer Torah:
Because of our many sins, the Torah has been forgotten and we can not find a kosher Torah scroll; the scribes are ignoramuses and the scholars pay no attention in this matter. Therefore I have toiled to find a Torah scroll with the proper letters, open and closed passages, but I have found none, not to mention a scroll which is accurate as to the plene and defective spellings, a subject completely lost to our entire generation. In all these matters we have no choice [i.e. we are halakhically considered anusim]; but how to write the correct forms of the letters we do know and their laws are like that of tefillin. Thus if we allow the ignorant scribes to continue to follow their usual practices [in shaping the letters], here we sin on purpose [mezidin].

The problem these Sages faced was purely halakhic. They would never have bothered to compare manuscripts unless the issue under discussion halakhically invalidated the Torah scroll on account of a single missing or extra letter or an imprecision in marking open and closed passages [petuhot, setumot]. Indeed these Sages do not even discuss the other books of the Scriptures. Even after they compared manuscripts and discovered the dismal reality, the halakhic framework which motivated their inquiries did not force them to inquire beyond what lay at arm's reach. The textual facts definitely point to "mistakes," but these same texts included the "proper" reading. These sages never postulated that the precise consonantal base of the MT, as given to Moses on Sinai, was irretrievably lost during the history of transmission and could no longer be reconstructed. For them the Authorized Text was alive and well, but it had to be retrieved from the existing manuscripts which were flawed through the human errors of the copyists.

As to the way to reconstruct the text, each of the mentioned Sages differed, based on the scope and quality of the material at his disposal. Maimonides, who had at his disposal a famed codex proofread by the Masorete Aharon Ben Asher himself, relied on this text as an accurate representation of the Authorized Text. The RaMaH, who had no such book, did have ones he calls "proofread and precise," also "masoretically exact traditions." He used an eclectic method to decide each question, based upon the halakhic mandate to follow the majority (of texts). He was certain that this method did indeed lead to the true image of the Authorized Text.

The only one who at first glance seems to despair of ever reconstructing the authentic text is R. Yom Tov Lippman, who lived in the textual reality of Ashkenaz. But even his stance -- that his generation was to be considered "forced" (anusim) in the matter of orthography and open and closed passages-- does not suppose that the accurate text is irretrievably lost. His basic theory is similar to the opinion of Rav Yosef in the Talmud, mentioned above, that the Babylonians were not expert in plene and defective spelling, and their scrolls do not represent the precise Authorized Text. From a halakhic standpoint, Rav Yosef meant that even the Babylonians were "forced," as R.Yom Tov Lippman later expressed the matter, except that Rav Yosef simultaneously expresses his belief that the true text of the MT is preserved in the Land of Israel. This sort of supposition, it seems, is also present in Yom Tov Lippman's words: the "correct" Text undoubtedly exists, but within the Ashkenazic reality there is no way to find it.

We can sum up by saying that the motivation of Medieval scholars to clarify the text for halakhic ends, together with the data which they used for this purpose which was entirely the product of inner-Jewish transmission, and the nature of their decision-making mechanisms which were mandated by legitimate Jewish criteria (majority rule, or in accord with the Masorah), all prevented a head-on collision between the ideal of a single sacred consonantal text as a historical reality versus the textual multiplicity which was a fact of life. All the above processes of clarification still left room for the belief in a reliable transmission which was able to pass on the consonantal base of the Bible as written by the original Author.

6. The situation in our day

Why is the present situation different from the Middle Ages? Why does the assumption of textual multiplicity for the Bible affront the religious Jew and threaten the foundations of his faith, when the Medieval reality described above seems to prove that such is not the necessary conclusion?

It seems that the difference is to be found in the contemporary grounds for textual clarification, which place the believer on the horns of a dilemma between what appears to be a principle of faith and what the evidence seemingly shows. The most daunting factor, the factor from which all other factors are drawn, is both psychological and substantial in nature: the evidence for a situation of textual multiplicity no longer stems from the transmission within the Jewish community but is forced on us from outside the Jewish world. To understand that, we will briefly analyze the developments in the history of the Biblical text from the beginnings of the modern era until today and the Jewish and Christian reaction to these developments.

Within Judaism in modern times there has been a unprecedented unification of the Biblical text. We have already mentioned that the revolutionary new method of transmission, the printing press, enabled the Authorized Text of the MT to win a final victory over other traditions of the MT group. Even books which continued to be written by hand, namely the scrolls used in the synagogue, became more unified thanks to printing, as it gave scribes everywhere a large number of identical "writers' handbooks" (called tikkune soferim) from which they copied their Torah scrolls. Even though changes in the text did not completely disappear, especially not in the Prophets and Writings, their number does not exceed the amount of variants which have always been present even in the model copies of the MT and hence do not create a sense of deviation from the consonantal base of the MT. We refer especially to the Torah scrolls used in synagogues today, where a number of differences exist which can be counted on the fingers of both hands and are in the main differences between ethnic traditions (Ashkenazi/Yemenite).

The obvious unity among Torah texts provides no reason for textual clarifications motivated by those considerations which drove the RaMaH in the Middle Ages, namely his recognition that "if a man wants to write a Torah scroll he will stumble over plene and defective spellings, and will have to feel his way through the fog of controversy...." This description of the way things were does not reflect today's reality. From the standpoint of internal Jewish development, not only does the current textual reality present no threat to the belief in an historical, sanctified consonantal text; field conditions have never been so favorable as to actually justify such an interpretation.

Outside the Jewish world, developments took a completely different turn. In the Christian milieu, not only did the era of print not foster the unity of the text, it weakened the idea. In the first centuries of the modern age this was due to religious polemics between the Protestant Reformation movement and Catholicism, which included the question of the Biblical text. The Reformation reintroduced to the Christian world the Hebrew Bible and claimed it should be preferred over the accepted versions in Christianity (the Septuagint or the Vulgate). Luther's translation into German, which was accepted by Protestants as authorized, was translated from the text-type of one of the first printed editions of the Scriptures (Brescia, 1494) and immediately sparked polemics between Catholics and Protestants about the origin of the various versions.

Oil was added to the flames of controversy with the discovery and publication of the Samaritan Hebrew text in the beginning of the 17th century (1616). Here was a Hebrew language text which was preserved for many years outside the transmission traditions of the Masorah, and it contained many variations also found in the Septuagint (nearly two thousand). Thus, for theological reasons within the Christian community, different text-types were compared and compared again, while the developing science of philology served as a tool in this internal debate.

But in the nineteenth century, thanks to the great advances in philology in general, the discipline of Biblical philology was released from the bonds of religious polemics and became an important tool for straightforward literal interpretation, with Protestant researchers leading the way. All the evidence which had been gathered by comparing the three texts-- MT, Greek Septuagint, and Samaritan Hebrew-- was used for the history of the text and its interpretation.[12]

In our century textual research took a giant step forward, particularly with the startling discovery of the Judean Desert scrolls. These findings for the first time enabled the community of scholars and researchers to get to know, independently and from a first-hand source, Hebrew Scriptural manuscripts from the Second Temple period, as well as the working methods of the scribes of that period. As a result, some disputes between researchers about the history of transmission during that era were resolved and many valuable facts were added to the arsenal of philological interpretation. The importance of these findings and the main conclusions to be drawn from them were discussed at the beginning of this study.

These developments happened, as we have said, outside the Jewish world and were diametrically opposed to the direction the text was taking within Jewry. It is therefore no wonder that traditional Judaism shrank from any contact with Biblical text criticism and all which it offered. Even though it was accepted among Jewish exegetes of the peshat since the Rashbam and others of his day that one should pay heed to "the simple interpretations being newly revealed daily" and accordingly to write new commentaries, the mainstream of religious Jewry actually closed itself off from the "new peshat" which was born of modern philological science. The hasty decision to suspect heresy in every commentary which was prepared to admit the possibility of textual error or which considered variants found in textual witnesses outside the Masoretic text, testifies to the strong feeling that the new scientific methods directly threatened the accepted principle of a single, original, sanctified Biblical text.

If we break this feeling down to its elements and test them against the background of the Middle Ages, in which this sense was completely absent, we can point to three main reasons which now exacerbate the feeling of direct conflict with the principles of faith:

The nature and scope of the textual diversity: As opposed to the Middle Ages in which textual clarification was undertaken on the assumption that textual plurality came about within the circles of the MT and that the one "correct" text could also be found within it, our era (which no longer has inner-Jewish variation) presents us with multiple text-types that exceed the bounds of the MT. These other text-types (the Septuagint, the Samaritan text) were unknown in Jewish tradition for almost two millennia. Reckoning with this array of texts seems to contradict the previous assumption that the origins of the text can be clarified within the bounds of inner Jewish transmission.
Motivation for textual clarification: Interest in textual changes in the Middle Ages stemmed, as we have seen, mainly from halakhic motivations. That is not the case in our day: from a halakhic standpoint there is satisfaction with the state of Torah scrolls, and no one would think of checking the actual text on those grounds. But from the viewpoint of modern textual criticism there are different motivations for textual research, foreign to Judaism, such as interpretation and the history of the text.[13] Difficulties in understanding verses and the desire to understand the history of transmission throughout the generations are what presses modern researchers to find and examine all available textual witnesses. These motivations, by their very nature, presuppose the possibility that correct readings might also be found outside the MT version accepted by Judaism, and that seems to contradict principles of faith.
The Grounds for Deciding: Jewish Sages of the Middle Ages did not include considerations of reason in their decisions between conflicting variants. The halakhic framework for textual clarification also mandated the rules of decision, which were in accordance with halakhic principle (follow the majority of texts) and according to the Masorah, whose authority was undisputed throughout the Diaspora.[14] On the other hand, methods of decision for modern textual criticism are based on pure reason, and only logical considerations determine whether to prefer the Masoretic text over another. Is it possible to submit the belief in the accuracy of the Masoretic transmission to a test of reason?

7. A new religious approach

We have tried to briefly analyze the causes for the present situation and the question with which we began this article remains: Is it right that educated religious Jewry which is constantly engaged in furthering ties with the scientific world in all fields, should accept the present situation in which philological-textual research in the Bible is shunned on principle, and should this situation be seen as an educational desideratum? Can there be any moral-religious persuasion to the demand that we refrain from using the methods of textual criticism when dealing with the text of Scriptures, while simultaneously acknowledging that these methods are relevant to every other text in the world?

It seems that the path educated religious Jewry has chosen to tread upon, in taking a favorable stance toward science and its methods and in recognizing science's authority to decide in empirical matters, is a painful one which requires constant comparison with accepted Jewish stances about those same empirical matters, stances which are now coming under scrutiny by science. An acknowledgment of the authority of scientific methods is not conceivable without the constant willingness to reassess and change interpretations and stances which clearly conflict with the conclusions of scientific research. This refers not only to the natural sciences, where this approach is generally accepted, but also to the humanities and Jewish Studies.

Textual research or criticism, as outlined in this article, is a characteristic example of an empirical subject whose treatment via scientific methods leads to a clear conflict with accepted beliefs in Judaism. The solution forced upon us in this situation is to release the ideal of a sanctified consonantal text from its historical interpretation and limit it to the halakhic realm alone. The facts demand we acknowledge that there is no historical proof for the idea that the Biblical text has reached us as it left the hands of its original authors. Transmission of the text was a human activity and all the laws and processes which affect any long-term transmission of a text would affect the transmission of the Scriptural text as well. The frequently-heard claim that the Scriptural text was better preserved than other literary works because of its excellent preservation mechanisms, such as the Masorah, does sound logical, but this too is not a religious claim but an empirical one which must be examined.

It therefore appears to me that the notion of a sanctified text in our era must be based on an halakhic interpretation alone, i.e., it must derive its power not from a determination that people managed to preserve the text exactly as it was throughout the entire transmission, but from the faith that man was given authority to determine, using halakhic methods of decision, the image of the sanctified consonantal text. The model which was decided upon would then be obligatory from a halakhic standpoint, even if it is found not to be historically "correct" in every detail.

This in turn requires a new approach to the science of textual criticism on the part of religious scholars who engage in the simple meaning of the Bible, the peshat. The literal interpretation of the text, which aims to discover things as they are, must take into account all possible textual witnesses, judge each variant on its own merits, and decide on the most suitable by means of pure logic.[15] There are strong and compelling arguments against textual criticism and its accomplishments during the nineteenth and twentieth centuries, such as the exaggerated importance given to conjectural emendations in Biblical philology, which often bordered on interpretative recklessness. But these arguments should not figure in a religious argument against textual criticism, only in a methodological debate between researchers about the proper ways to conduct textual criticism. The call to forgo imaginative emendation and to exercise care in classifying variants and estimating their value is being heard more and more among the scholars themselves as a basic requirement to arrive at proper exegetical conclusions.

[1] Even in the Da'at Miqra series of commentaries on the Bible, which claims to be a blend of tradition and science, the commentators have consciously refrained from dealing with textual difficulties. Text analysis is limited to an introduction entitled "The Text and its Sources," which lists minor textual differences (mainly vocalization and cantillation) deriving from a comparison of definitive Medieval manuscripts (Keter Aram Sova, Leningrad B19a, and others) as well as the Venice edition of Miqra'ot Gedolot.

[2] Compare U. Simon, "R. Abraham Ibn Ezra and R. David Kimhi--Two Approaches to the Question of the Reliability of the Text [Hebrew]," Bar Ilan 6 (1968), 191-237.

[3] For example: In the fourth cave of Qumran, remnants of a scroll of the Book of Exodus written in the ancient Hebrew script were found, which includes almost all the characteristics of the Samaritan version for this book, e.g. harmonization between command and performance in the Ten Plagues, additions based on the parallels in Deuteronomy, minus the Samaritan ideological additions (such as "the tenth Commandment"). This is incontrovertible proof that the Samaritan version was a Jewish Bible prototype before it was adopted (and adapted) by the Samaritans.

[4]For example: All researchers agree that the second scroll of Isaiah (1QIs^b ) represents the Masoretic text-type; yet in the segment which has survived there are 248 divergences from MT as represented by the Leningrad manuscript with the following breakdown: spelling: 107, added waw conjunctive: 16, missing waw: 13, definite article: 4, switched letters: 10, missing letters: 5, additional letters: 9, gender switches: 5, number switches: 14, pronominal changes: 6, various grammatical forms: 24, changed prepositions: 9, changed words: 11, missing words: 5, added words: 6, switched word order: 4. Some scholars maintain that the first scroll of Isaiah (1QIs^a) with its thousands of divergences, should still be classified as part of the Masoretic text-type since most of the differences are not typical of a differing version but rather of changes within a particular text-type.

[5] Thus, a fragment of Deuteronomy from the second century B.C.E. which belongs to the Masoretic text-type contains corrections made approximately 100 years later which are based on a Septuagint-vorlage text-type. For example, Deuteronomy 7:15--

The first scribe wrote, as per the Masoretic text we have, while the later scribe added above the line as per the Septuagint. In that section there are also two additional corrections along the same lines. It would be difficult to explain these changes assuming that the MT was the model-text at Qumran.

[6] The first scroll of Isaiah shows many of these instances. There are many cases of a variant recorded above the text without erasing the original. For example, Isaiah 36:11, (which version differs even from the Masoretic text, which reads ); a later scribe added in the margin thereby noting that he found a variant in place of but he does not express a preference for one over the other. In later copies this phenomenon sometimes becomes one of doublets, in which both variants are written in the body of the text. 1QIs^ahas many examples, such as Isaiah 37:9, which is a conflate text: (which is the MT) and (which is the text of the parallel in II Kings 19:9).

[7] Some tefillin at Qumran are written in most plene spelling, as found also in the first Isaiah scroll, and some are written more like the MT. Occasionally, tefillin conform to the Septuagint or Samaritan text-type, as in Deuteronomy 10:13, ; MT has only the tetragrammaton, the Septuagint and Samaritan texts have both. (Deut.10:13 is found in the tefillin of the Qumran sect.)

[8] For example, in I Chronicles 21:15-17, several items are mentioned which are not in I Samuel 24:16-17, e.g. "And David raised his eyes and saw an angel of G-d standing between heaven and earth, his sword outstretched over Jerusalem, and David and the elders, covered in sack, kneeled." This whole story is not found in our text of Samuel, and one might say that the author of Chronicles completed the account imaginatively or drew upon some other source. But Qumran fragments of Samuel show that the author of Chronicles used a version of Samuel different from MT which includes this verse and some other readings found in Chronicles.

[9] The reader can find other examples in Gilyon HaShas by Rabbi Akiva Eiger, Shabbat 55b. Many more examples, systematically arranged, can be found in S. Rosenfeld, Mishpahat Soferim.

[10] The Masorah parva is a system of short comments, formulated as acronyms and abbreviations, which relate to various words found in the body of the text. The comments are meant, in general, to mark a way of writing or reading the word or combination of words and the number of times this phenomenon occurs in Scriptures or a portion thereof (for example: = three times [written with] defective [spelling] in the Torah; = fourteen times [with] plene [spelling] in Scriptures, etc.) and they are written in the margin alongside the line in which the word appears, to the right or left. The Masorah magna is more detailed than the Masorah parva and is written in the upper and lower margins of the page. Generally, the Masorah magna tends not only to give the number of times a phenomenon appears, but also to cite the other Scriptural references (using signs and keywords).

[11] The genetic relationship between manuscripts is affirmed by two strata of comparison: (1) a stretch of text (2) orthographic method. Through the first comparison we can see that a long line of identical variants appears in manuscripts from varying geographical regions (including the Land of Israel, Spain, Ashkenaz) and periods. The second method of comparison proves the antiquity and independence of these consonantal traditions. For if we see that the changes are not a collection of random scribal errors in copying the standard MT consonantal base but represent a systematic orthography contrary to the methods of the MT we can show that we have before us consonantal/orthographic traditions which preceded the Authorized Text and were created before the ideal of a sanctified consonantal text took hold. In this early period, different schools of spelling might find expression in Scriptural scrolls copied by scribes of each school.

[12] The glamour of the Samaritan text as an independent textual witness dulled in the nineteenth century following publication of a paper by Gesenius, in which he claimed that almost all changes in the Samaritan text were made by Samaritans and grafted onto a text base similar to the MT which the Samaritans adopted for their own. Nineteenth century researchers generally accepted Gesenius's opinion (except for A. Geiger), and commentators of that era usually weighed the Septuagint and the Masoretic texts against each other. After the research of P. Kahle at the beginning of the twentieth century, researchers again acknowledged the important of the Samaritan text as an early, independent textual witness from the Second Temple period. The Qumran findings, as mentioned at the beginning of this work, finally proved that Kahle was correct on this issue. (He was not correct on some others, such as the supposition that texts such as the Septuagint and Samaritan, which he calls "vulgar," were the sole texts in the Second Temple period and that the Masoretic text was a later revisionist attempt to give the text an archaic look. The Qumran findings proved beyond a doubt that this description of the situation is completely incorrect, and that the Masoretic text-type existed during the Second Temple period along with the other text-types.)

[13] It must be pointed out that precedents can be found in Medieval commentators for interpretive textual clarifications, but they are so rare that they must be seen as exceptions which prove the rule. One such example is the commentary of R. Yosef Qara on Joshua 9:4: "And they went in disguise" () "They sent delegates ()...and some scrolls have written "and they provisioned themselves," () they dried all their bread () to make it appear they had come from a distant land. Each brings proof for his words and neither convinces the other, and we can not clarify which is the correct reading, but I tend to agree with the scribes who write "provisioned themselves," because of a proof from context. Verse 14 reads, "and the people took from their provisions" () And in another place: "This bread of ours, which we took from our homes () [verse 13]. Each utterance uses the root . This is exactly the method of textual criticism for interpretive ends, but its use is completely by chance and was not developed as a method by R. Yosef Qara or others.

[14] Text-decisions based on these principles are found in Medieval commentaries. As opposed to R. Yosef Qara's approach, described in the last footnote, which preferred a variant reading not found in the model scrolls of the MT for contextual and logical reasons, we find in the RaDaK's commentary on Joshua 21:7 a clear example of a decision which goes against logic and context, based on the fact that the unlikely variant is found in the model books of the Masoretic text. "The verse states that twelve cities were owned by the sons of Merari of the tribe of Reuven, and the tribe of Gad and the tribe of Zevulun, and further on, in the count of cities, only eight are mentioned: the tribe of Zevulun has four and the tribe of Gad has four and none are specified for the tribe of Reuven. Some scrolls are corrected to read, "and the tribe of Reuven has Beser and its environs and Yahzah and its environs and Mefa'at and its environs, four cities," but I have not seen these passages in any old and exact scroll, only in those which are partially corrected. Rabbi Hai Gaon was asked this question and he answered, "Though four cities were not cited here, they were in Chronicles." We see from his response that they were not written in the books of Joshua which they had. From RaDaK's words we understand that he is not prepared to accept the variant reading which includes the two verses, even though logic dictates that these two verses are called for by the context, simply because he did not find them in books he considered "old and exact," and because of the support he finds in the words of Rav Hai Gaon.

[15] One example will suffice to illustrate how, in our opinion, a modern religious Bible commentary should deal with textual matters. Gen. 46:13 reads: "And the sons of Issacar are Tola and Puah and Yov and Shomron." The third name on the list, Yov () arouses some doubt as to its accuracy, in light of the following: In Num.26:24 the name is given as "Yashuv" (); In the Samaritan version of the verse in Genesis, as well as in the Septuagint, the name is Yashuv; I Chron.7:6 in the MT reads: "And the children of Issacar are Tola and Puah, Yashiv (Qere: Yashuv) and Shomron, four sons." These facts lead us to the conclusion that Issachar's son was indeed called Yashuv, and that Yov is an error which crept in to the transmission of the MT before sanctified status was granted to its consonantal base. A modern religious commentary will not have fulfilled its obligation to its readers if it did not bring these facts and their like to their attention. Nevertheless, the historical conclusion is to be kept separate from the question of writing a Sefer Torah. Since the spelling has been sanctified together with the sanctification of the Authorized Text of the MT, anyone who changes one letter of this word invalidates the Torah scroll, as only this spelling has validity and carries halakhic weight when writing a Sefer Torah.

Translated from HaMikrah V'anachnu, ed. Uriel Simon, HaMachon L'Yahadut U'Machshava Bat-Z'mananu
and Dvir, Tel-Aviv, 1979, with very minor changes. The English version was editted by Isaac B. Gottlieb. By permission.

English version first posted at The Idea of the Sanctity of the Biblical Text.

The idea of the sanctity of the Biblical text and the science of textual criticism

Menachem Cohen Professor of Bible, Bar-Ilan University Director, Miqraot Gedolot HaKeter Project

Menachem Cohen
Professor of Bible, Bar-Ilan University
Director, Miqraot Gedolot HaKeter Project