Abstract
The Hebrew Bible or Old Testament did not drop down from heaven, as is sometimes believed. It is widely agreed that the canon is the construct of faith communities. But what was the process that led to the formation of the canon? This article discusses what we know about the formation of the canon, addressing assumptions that different readers might have about what constitutes ‘the Bible’. It will show that the Jewish Tanak and Protestant Old Testament have the same collection of books and it is the canon of Pharisaic-Rabbinic Judaism. It was not, and is not, the canon of other faith communities that considered different lists of books as authoritative.
According to the Mishnah, Moses received the Torah from Mount Sinai and transmitted it in an unbroken chain to Joshua, the Elders, the Prophets and the Men of the Great Synagogue (m.Avot 1:1), but the Bible did not drop down from heaven as implied in this Rabbinic dictum (cf. m.Sanh 10:1). There is broad scholarly agreement that the canon of the Hebrew Bible was a human product, constructed by religious communities that believed in the divine inspiration of their holy writings. 1 Exactly how this process was carried out and when the canon was formed, however, are much-debated issues. 2
Different Conceptions of the Canon
The formation of the canon is a complex and multifaceted subject, because there are different conceptions of the Bible. The most widely held view is that the canon comprises twenty-four books that are divided into three sections, the traditional Jewish Bible or Tanak. 3 This canon is the canon of Rabbinic Judaism and of the Pharisees, and the Masoretic Text is the textual version of the books of the canon. This same collection is designated by scholars as “the Hebrew Bible”. Christians call the collection “the Old Testament”, but they differ from Jews and from each other in the way that they categorize the books and include other books. Protestants enumerate the same books differently from Jewish tradition, counting 39 rather than 24 books, and divide the books into four rather than three categories. Catholics use the Latin Vulgate and Orthodox Christians the Greek Septuagint and have their own arrangements that include other books in their canons, which they call “the Apocrypha” and “Deutero-canonical books”. The Samaritans or Northern Israelites have fewer books in their canon, restricting their books to the Torah or the first five books, the Pentateuch. 4 Despite the existence of competing conceptions, each community calls its own collection “the Bible” and not “a Bible”, in effect recognizing only its own canon as authoritative.
Addressing Assumptions about the Canon
Scholarly disagreements are often based on assumptions about the significance of features and factors thought to be associated with the canon. Is it legitimate to speak about the closing of the canon only with the emergence of the codex or book-format? 5 Are ancient discussions of the criteria of inclusion in the canon ex post facto rationalizations? 6 Is the textual version of biblical books a factor in how a particular canon was formed? 7 Can holy scriptures be written in a language other than Hebrew? 8 How does one understand textual authority and is it restricted to the books that are included in the canonical lists? 9 Do the different ways of counting the same books and dividing them into categories mean that the canon was still in a flux? 10 Scholars offer a range of views on each of these and other questions.
The essential feature of the canon is that it is a list or collection of books that is considered authoritative by a religious community. This is evident in the earliest catalogue, known as the Bryennios list (ca. 150), which counted twenty-seven books in one undifferentiated, canonical register. 11 Other features include: how the books are counted (22, 24 or 27), the sub-categorization of the record, the format of presentation of the collection as separate rolls or bound in a codex, and the original language of the books. 12
Etymology and Concept of Canon
Jews in antiquity did not have a term for “canon”, but they did have the concept. The English word “canon” derives from the Greek kavōn (κανών), a loanword from a semitic root (in Hebrew qāneh [קָנֶה]) meaning “bulrush, calamus or stalk”, and by extension a measuring reed, rod or staff. Its figurative sense as a list of books is not found in ancient, Semitic usage but in the Greek meaning of a literary kavōn, derived from Hellenistic understanding of the importance of the list of the works of Homer and other classical authors.
The absence of the term, however, is not the same as the absence of the concept. In the ancient sources, scripture is denoted in Hebrew, Aramaic and Greek by various phrases and titles. In Ezra-Nehemiah, several designations are deployed with the terms “(ha-)torah”, referring to laws and narratives, and/or “sepher”, a book or scroll (Neh 8:1, 3, 8, 13, 18; 9:3; Ezra 6:18; 7:6), which imply large and small literary collections. Likewise, Aramaic expressions embedded in the documents and narratives (“the book of the law of the God of Heavens”, Ezra 7:12 and “the law of your God”, Ezra 7:26), suggest that there is a concept of canon.
In Chronicles, the Passover tradition is retold in a way that claims dependence on the earlier prescriptions of the laws of Moses, legal ordinance, authority of the ruling king and the prescriptions of David and Solomon (2 Chr 35:4-12). Significant is the mention that the Passover tradition is to be found “in the writing of David”, “in the document of Solomon” and “in the book of Moses”.
Among the sectarian texts of the Dead Sea Scrolls, designations of scriptures include: “the torah” (CD 4:8; 5:7; 6:4, 7; 1QS 8:15; 4Q159 fr. 5, l. 6 etc); “the scroll of the sealed torah” (CD 5:2); “the torah of Moses” (CD 15:12; 16:2; 1QS 5:8); “the scroll of Moses” (4Q397 fr. 14, l. 10; 2Q25 fr. 1, l. 1; verso of 4QpapCrypta); “the scrolls of the prophets” (CD 7:17//4Q266 fr. 3 col. 3, l. 18; 4Q177 fr. 1, col. 4, l. 14); “the songs of David” (11Q13); and “the book of psalms” (4Q491 fr. 1, l. 4).
In Jewish texts of the Hellenistic and Roman periods and in the writings of the early church, scripture is most commonly called “the writing” (e.g.,
In Rabbinic literature, scripture is designated commonly by “what is read” (המקרא), “what is written” (הכתוב), “the writings” (הכתובים), “the holy writings” (כתבי הקודש), “the book or scroll” (הספר), “the books or scrolls” (הספרים), “the law” (תורה), and “the law and prophets” (תורה והנבאים).
These sources do not have the same books in view, but they had incipient understandings of collections of textual authority that are consistent with the concept of canon.
Brief History of the Study of Canonization
There is no ancient source that describes how the canon was formed. The canonical process has to be pieced together by a reading of the sources, which are incomplete and biased in some respects and open to interpretation. A historical reconstruction is based on a plausible interpretation of a multitude of sources.
Between the end of the 19th century and first half of the 20th century, there was a broad scholarly consensus that the canon developed in three-stages, the Torah or Pentateuch was closed around 500 BCE, the books of the Prophets (nevi’im) in the third or fourth century, and the Writings (kethuvim) at the council of Javneh in 90 CE. This general agreement no longer holds sway. 13 Questions have been raised about the strictly sequential manner in which the canon was thought to develop, and the christianization of Javneh as a “council”. 14 More recently, the schism between the Judaeans and Samaritans, thought to be decisive for the formation of the Pentateuch in the three-stage theory, has also been shown to be untenable. 15
In the past generation, several theories have emerged. Roger Beckwith argued that the canon was not closed in three but two stages: the Torah or Pentateuch first, followed by the subdivision of non-Mosaic material in the second century BCE. 16 Andrew Steinman likewise believed that the whole canon was closed by the end of the fifth century, before splitting into two strands that subdivided the books into the “the Law and the Prophets” or “the Law, the Prophets, and the Writings/Psalms”. 17 John Barton proposed that throughout the post-exilic period to the time of the New Testament period and beyond, scripture was bipartite rather than tripartite. He believed that the canon, remained open and could have included books in addition to those books that were eventually canonized. 18 More recently, Barton thinks that the canon was closed in the first century CE. 19
These scholars do not understand the closing of the canon in the same way. Does it mean the sequential closing of each subsection of the Torah, the Prophets, and the Writings? Is the canon closed only when there is a list of biblical books? The closing of the canon was a process that cannot be defined by a schema and a point in time. It was an undertaking that occurred in various religious communities over time. Moreover, disputes about the canon continue to the present day. The canon has never been closed, if by this is meant the absence of any doubt by anyone about the identity and collection of books of the Bible. The closing of the Rabbinic canon is akin to the reaching of a scholarly consensus. It does not imply the absence of dissenting voices. It means that most accepted the 22/24 books of the canon. 20
Reconstructing the Canonical Process: The Starting Point
How does one begin to discuss the process that led to the canonization of the Rabbinic canon? One way is to take as starting point the earliest evidence in the biblical texts themselves. 21 This approach investigates concepts and features that are believed to be relevant for the canon. 22 Discussions invariably begin with the reform of Josiah where a book of the law, believed to be an Urdeuteronomium, is mentioned (2Kgs 22). But the evidence is too meagre and ambiguous, allowing for theoretical possibilities rather than plausible reconstructions. In what sense is Josiah’s one book of reform a canon?
It is widely agreed that the earliest list of the canon in Rabbinic literature is to be found in b.Baba Bathra 14a-15b. The context of the mishnah is the division of property and belongings in the event that partners break up, a dispute that digresses into a discussion of holy scriptures and how they should be divided. The halakhic concern is for the ordering of the scriptures when copied on a scroll. The Pentateuch is not mentioned, but assumed, since its order is not in dispute. The number of books enumerated is 24, five of the Pentateuch, eight of the Prophets and eleven of the Writings.
The dating of this passage is not straightforward. There is a reference to R. Johanan (250-290), a Palestinian amora, who was responding to some unidentified rabbis (“our rabbis taught”). But R. Johanan need not have lived at the same time as these unnamed rabbis who could be contemporary with or earlier than Johanan. Philip Alexander argued that the final closing of the Rabbinic canon took place around 200 CE in reaction to the “growing power of Christianity”, which, at the end of the second century, moved to define its canon, and to add a Second Testament to the First. 23 He used the canonical notice of Melito, the Bishop of Sardis, as corroborative evidence and dated the baraita of Baba Bathra to before 200 CE. 24
The canon had already taken shape a hundred years before. By 100 CE, there was already a canon of books. Josephus polemicized against his Greek detractors by stating that Jews do not have thousands of books that disagreed and are in conflict with each other, but only twenty-two books, consisting of the five books of Moses, the thirteen books of the prophets, and four books of hymns and instructions (Ag. Ap. 1.38-41). 4Ezra recounts the miraculous restoration of the Jewish law, destroyed during the destruction of the Temple in 70 CE, with a reference to a 94-book canon, consisting of 24 public books (widely agreed to be the biblical books) and a larger corpus of 70 books for the wise (2 Esdras 14:45-48). The Bryennios list (ca. 150 CE) provides a record of 27 biblical books identical to the content, but not the count, of the Rabbinic canon. The Mishnah states that only “holy scriptures” defile the hands, assuming that what constitutes kitvey ha-qodesh is known and not disputed, except for Qohelet and the Song of Songs (m.Yad 3.5, 5; 4:6). And Rabbi Aqiba declares that the one who reads “outside or heretical books” (sepharim ha-hitsonim) does not have a place in the world to come (m.Sanh. 10:1), implying that at least for him the books of the canon were agreed. 25
Starting from the end of the canonical process rather than its beginning has its potential pitfalls that must be avoided. Philip Davies warned against a teleological fallacy in constructing “a single line of evolution” that results in a single canon by reading the final shape of the Hebrew Bible back into pre-Rabbinic Judaism. This warning is well worth repeating, but it should be recognized that Baba Bathra is only one of several canonical lists in the first centuries of the common era (see Josephus, CA 1.38-41; 4 Ezra 14:45-48; mYad 3.5; Bryennios list [ca. 150 CE; folio 76a of Ms 54 of the Greek Patriarchate Library of Jerusalem]; Melito of Sardis [ca. 190; apud Eusebius, EH 4.26], Baba Bathra 14-15, Origen, [ca. before 232; Commentary on the Psalms apud Eusebius, EH 6.25, Jerome [ca. 390s, Prologus Galeatus]). The books mentioned on these canonical lists overlap to a large extent with each other and to earlier collections of authoritative scriptures, and these lists did not materialize ex nihilo. No single line of evolution is being retrojected back into the pre-70 period. There was not just one canon among ancient Jews. The Rabbinic canon was the canon of the Pharisees, and not of all Jews. The existence of the lists and notices suggests that the Hebrew Bible was more or less formed by the turn of the era, and that the canonical process preceded this time. 26
The Torah in the Persian Period
In the post-exilic period, there emerged an authority that was based on the written text, which the returnees from Babylon called “the torah”, a biblical term that variously means direction, instruction, and law. 27 Scholars often assume that “the book of the law of Moses (ספר תורת משה; Neh 8:1)” is equivalent to the Pentateuch as we know it, but this assumption is open to question, since there are differences between the two. In the account of the feasts and celebrations, there is no reference to Yom Kippur alongside Sukkot (cf. Lev 23:27-36, 39-43; 25:9). Conversely, there are references to regulations not found in the Pentateuch in the offering of wood (Neh 10:34; cf. Temple Scroll 11Q19 11.3) and in the annual one-third shekel charge for the Temple (Neh 10:32). Some of the regulations on the same practice also differ (e.g., rules governing the feast of booths, Neh 8:13-18 and Lev 23:33-36). Ezra-Nehemiah cites or alludes to the first six books of the canon, from Genesis to Joshua. 28 It evidences knowledge of legal content similar to, but not identical with, those found in the Pentateuch, and of the tradition of the conquest in the book of Joshua. It is a plausible inference that the book of the law of Moses refers to the Pentateuch plus Joshua or a collection of six books from Genesis to Joshua, a hexateuch. 29
From Authoritative Scriptures to Canon
The next important phase in the canonical process is the period between the Maccabean revolt and the destruction of the Temple of Jerusalem by the Romans, ca. 200 BCE to 100 CE. During this transitional time, other compositions became authoritative and canonical for different Jewish communities. How one describes these literary works is a matter of debate? One option is to call them “biblical”, since most of these compositions were eventually included in the canon. It seems contrived, for example, to eschew calling the book of Genesis “biblical”, when it is the first book of the canon. Whether “biblical” is anachronistic depends on when one believes the canon was closed. The use of this term is problematic in another more significant way. By definition, it excludes compositions that functioned in the same way as biblical books, but were not included in the Rabbinic canon. For instance, the communities reflected in the sectarian Dead Sea Scrolls also considered the book of Jubilees, the Temple Scroll, the book of Enoch, as well as their own writings (Pesharim, Rules, Hodayot, Rules) in the same way that they regarded the biblical books. But these books were not included in the canon.
A second option is to use the terminology of “authoritative scriptures” that has the advantage of avoiding unwarranted assumptions of canonical inclusion and closing while underscoring the essence of what is connoted by “biblical”, namely its authoritative nature. At the beginning of this transitional period, we see the emergence of “authoritative scriptures” among Jewish communities, and by the end of the first century CE, the Pharisaic/Rabbinic canon was more or less formed. 30
Translation of Jewish Laws to Greek
It was during this transitional period that the Jewish authoritative scriptures were translated into Aramaic and Greek. The Dead Sea Scrolls attest to the earliest Aramaic translations or targumim of Leviticus (4QTgLev [4Q156], ca. 150-50 BCE) and Job (4QTgJob [4Q157], ca. 20-50 CE and 11QTgJob [11Q10], ca. 30-68 CE). 31 But it was in the origin story of the translation of Jewish laws into Greek that we find a conception of an authoritative list.
It is widely held that the diegesis or narrative, known as Aristeas, attests to the translation of the first five books, the Pentateuch. 32 This is a story about how Jewish laws, during the reign of Ptolemy Philadelphus (285-247 BCE), were rendered into Greek at the request of the chief of the great library in Alexandria, Demetrius of Phalerum, in order to fill a gap in the library collection. But there are inaccuracies and incorrect information that belie Aristeas as a propaganda written by an unknown Hellenistic Jew at around 100 BCE to show how successful Jews were in the Egyptian court and how they enjoyed the beneficence of the Ptolemaic king. 33 The initiative came from some unknown Jews who paid for the translation and ascribed the project to king Ptolemy whom they admired as a man of culture. 34
What was translated is described as “the divine law” (§3); “the laws of the Jews (§10); “the law” (§176, §313); “the books” (§176, 317); “the entire law” (§309); and “the rolls” (§179, 330). The etymology of penta-teuch is associated with one of these terms: ta teuche, “the rolls”. There is weak evidence that Aristeas knows all five books of the Pentateuch. 35 Whether it was the Pentateuch or a kind of tetrateuch or four-book corpus by 100 BCE Jews in Alexandria held that there was a collection of scriptures that was worthy of translation and inclusion in the great library of Alexandria. The authority of these scriptures is highlighted in the climax of the narrative, as “Demetrius” read out the translation to the gathered assembly of Jews, and the priests, translators and elders affirmed the complete accuracy of the translation and pronounced an imprecation against changing it. This was the giving of the law on Mount Sinai re-imagined. 36
The Curriculum of the Scribe
It was also in the second century BCE that we encounter evidence for a collection of authoritative scriptures in the subjects and books to be studied by the scribe. The Wisdom of Jesus Ben Sira prescribes in two passages the curriculum to be studied by the scribe. This book of Wisdom was originally composed in Hebrew by a Jerusalemite sage named Jesus ben Sira in the first quarter of the second century BCE, between 196 and 175 BCE (cf. Sir 51:23). 37 His grandson took this composition with him when he went presumably from Jerusalem to Egypt in the 38th year of the reign of Euergetes or 132 BCE. It is thought that he spent the next fifteen years translating his grandfather’s book of Wisdom into Greek and published it with a Prologue after 117 BCE.
There are two passages that describe the curriculum of the scribe, Sir 39:1-3 (alternatively enumerated as 38:24-39:1) and the Prologue. 38 The former is part of a larger literary context that compares the skilled worker with the scribe who devotes himself to study (Sir 38:24-39:11). The subjects to be studied by the scribe include several kinds of writings, law, wisdom, prophecies, discourse, parables, and proverbs, and it is unclear how many categories Ben Sira had in view. 39 Some of these descriptions are apt descriptions of the first division of the Law and Prophets of the traditional canon. The references to wisdom, discourse, parables and proverbs are suitable descriptors of the wide-ranging genres of the books included in the traditional Writings, but the third division of the canon remained open as Ben Sira recommends learning “the wisdom of all the ancients” (italics added; Sir 39:1), and not just that of the Israelites.
The Preface, written in Greek by the grandson, corroborates and diverges from this understanding of the curriculum of the scribe. It includes references to “the Law” and “the Prophets” or “the Prophecies”. There is also a reference to “the other books of our ancestors”, which is not a category as such, but a classification in the sense of “all the rest of it”. 40 The Prologue attests to two categories of Law and Prophets, plus a third open-ended one which includes Jesus ben Sira’s book of Wisdom. 41 In the Prologue, the grandson recommends his grandfather’s book of Wisdom as a book to be studied alongside books that were eventually included in the canon. 42
We can identity the authoritative scriptures of these open collections by attending to the latter half of Sirach and the praise of the fathers (chs. 44-50). 43 This literary parade of Israel’s great and good begins with Enoch and ends with Simon, son of Onias (219-196 BCE). It implies knowledge of all the books of the traditional Hebrew Bible, except for Ruth, the Song of Songs, Esther and Daniel. 44 The authority of these scriptures is assumed in the sources used and the account of the heroes of Israel. Sirach also self-attests to its own textual authority, including a literary memorial of praise to Simon (50:1-24). 45
Dual and Graded Authority of Scriptures
The Dead Sea Scrolls are undoubtedly the most important new evidence for understanding the canonical process. Discovered by Bedouin goat-herders in 1947, these “greatest manuscript discovery of modern times” 46 comprise some 900-1000 original scrolls that attest to all the books of the Hebrew Bible except Esther. In the past, it was thought that the Dead Sea Scrolls belonged to the library of the sectarian community of the Essene who lived in the Judaean Desert at Khirbet (“ruins of”) Qumran, from the middle of the second century BCE to 68 CE. 47 Identifying what the Essenes regarded as authoritative, so it was thought, was a relatively simple matter. 48 If such and such a book were found in the collection, then it would be authoritative.
But this misunderstands the collection and the concept of authority on several levels. An ancient library—it is debatable whether that is what it is—does not have to curate only books that agreed with the community that uses it. 49 The scrolls corpus is not a library but a heterogenous collection, consisting of books of one or more sects to be identified with the Essenes, the biblical books that belong to all Jews, works that later became authoritative for Christian communities, and other compositions (previously known and unknown) associated in various ways with one or more textual clusters. 50 These texts are not marked by direct disputes, as one finds in Rabbinic literature. Rather, the difference is in emphasis, interpretation and formulation.
The heterogeneity of the collection raises questions about the meaning of “authoritative scriptures” and how we ascertain that they were so regarded. The concept is theoretically possible as an assertion of authority, but a text’s claim to authority, if it is to be meaningful, needs to be affirmed by one or more communities that recognize and accept that declaration. Otherwise, it is an empty claim with little significance. For instance, the “Reworked Pentateuch” presents itself as a rival version of the Torah. 51 This is a self-claim to authority by the implied author or scribe(s) who compiled it, but for it to be “authoritative” in a meaningful sense, it requires that the communities reflected in the sectarian scrolls or another community to accept that it is a or the Torah for them.
How does one ascertain the concept of “authority” in the scrolls? Several ways have been suggested: references to books (e.g., “book of the torah”); the use of the content of a writing (e.g., sabbath rules); and allusions to traditional symbols of Israel (e.g., tribal confederacy; Israel as stubborn heifer). Perhaps, the feature most revealing of scriptural authority is the technique of citation of source-texts in the sectarian scrolls. 52 Only texts that are eventually included in the canon are cited with an introductory formula. 53 The books quoted with an introductory formula include Genesis to Deuteronomy, several of the prophets, and some of the writings. There is some evidence that the Pentateuch was considered a closed category and the books of the prophets an open one. There is no evidence of a third division of the Writings, quite apart from collections of psalms. 54
Textual authority, however, was not restricted to the books that were eventually included in the Rabbinic canon. The Damascus Document regarded the book of Jubilees as an authoritative perush or explanation of the times found in the law of Moses (CD 16:1-3). 55 This is a dual authority of the Pentateuch and its explication of the time, seasons and feasts by the book of Jubilees. The traditional biblical texts serve as the authoritative source of the sectarian interpretation but they are also in turn defined by it. The biblical texts have a formal authority and the sectarian interpretation has contemporary authority, and the latter often governs the meaning of the former. Frequently, it is not what the biblical texts say that is ultimately authoritative but what the sectarian scrolls understood them to have meant.
For the sectarian communities reflected in the scrolls, textual authority is also to be found in their own sectarian writings, most notable of which are the different genres of the Pesher. 56 The pesher, like midrash, is both a genre and a method of exegesis. 57 1QpHab, for instance, quotes sequentially passages from the first two chapters of the prophecy of Habakkuk. With variations, it follows a structural pattern of biblical quotation or lemma + space + interpretative formula (often including the technical term pesher [‘interpretation’]) + sectarian comment.
On the one hand, this structure distinguishes between the source-text and the sectarian interpretation. Implied is that the source-text is authoritative, and the sectarian secondary. The reconstituted biblical text is comprehensible in a way that the sectarian interpretation without the biblical lemmata is not. Nevertheless, the sectarian comment is also authoritative, since the method and content of pesherite interpretation are divinely revealed. God enlightened the Teacher of Righteousness by giving him understanding and a method to interpret the prophetic oracles (1QpHab 2:8-10). He also revealed the content of the mysteries of the prophetic oracles to the sectarian leader (1QpHab 7:4-5). The method was replicated by the Teacher’s followers, and his interpretation was recorded in the scroll that we call Pesher Habakkuk. That the pesher, biblical lemma and sectarian comment together, was considered authoritative is evidenced by its quotation and use in the Damascus Document (Pesher on Isa 24:17 in CD 4:12-19//4Q266 frag. 3). Textual authority evident in the sectarian scrolls also seems to be graded, on a sliding scale that differentiates the prerogative of different kinds of writing.
The Pharisaic Canon and the Letters of Paul
In the middle of the first century CE, the Pharisee turned apostle to the gentiles wrote several letters preserved in the New Testament. Saul, later known as Paul, evidently became a convert on the Damascus Road and he devoted the rest of his life to planting churches in various places around the Mediterranean. Paul was not interested in defining his canon, but in his letters we find evidence that he was following a canon that was consistent with the Pharisaic canon. 58 In the so-called Hauptbrief or capital letters, he cited from all the books of the Pharisaic canon except for the Song of Songs, Ruth, Esther, and Ezra-Nehemiah. When he did so, he used formulas to introduce these quotations (e.g., “as it is written”). Paul used introductory formulas when he was quoting a text that was eventually included in the canon. He does not always use formulas to introduce his quotations, but when he does use them, he is almost invariably citing texts that were included in the Pharisaic/Rabbinic canon. He describes these biblical texts as “holy scriptures” (graphai hagiai [γραφαί ἁγίαι] Rom 1:2), a description unique in the New Testament, and an exact translation of the Rabbinic expression kitvey ha-qodesh (כתבי הקודש).
The Pharisaic Canon at the end of the First Century
By the end of the first century CE, there was a canon in all but the strictest sense. The combined attestation of Josephus, 4 Ezra, Mishnah Yadayim, and Bryennios list is strong evidence for the emergence of the canon. There continued to be disputes about a few books, notably Qohelet, Song of Songs, Ruth, and Esther (cf. m.Yadayim 3.5; bMeg. 7a), but the Pharisaic canon was held up by Josephus as the canon of all Jews, while not entirely accurate was a credible generalization. 59
This canon was the canon of the Pharisees that was adopted by the Rabbis in their re-founding of Judaism after 70. The Pharisees became the dominant force in Jewish society during the revolt against Rome, and they were the majority group at Javneh and it was their canon that became the canon of Rabbinic Judaism. 60 This canon was not the canon of other Jews. Presumably, most other Jews accepted it, since the books included on this list were also considered authoritative by other non-Pharisaic communities, like the Essenes. 61
Indicative Logic and the Formation of the Canon
The books included in the canon were not selected by the Pharisees using criteria that externally measured the authority and suitability of each composition. Rabbinic literature associates the inspiration of the holy spirit with the books that were included in the canon. The claim is that prophecy ceased with the death of the last prophets, Haggai, Zechariah, and Malachi (Tosefta Sotah 13:2; Tosefta Yadayim 2:14; b.Meg. 7a. Cf. 1 Macc 9:27, 4:26, 14:41; Joseph, Ag. Ap. 1.38-41; 2 Esdras 14). Other communities in the late Second Temple period did not believe that prophecy had ceased and contended that other non-biblical books (e.g., Jubilees, 1 Enoch, 1QpHab, Gal. 3:1-4:13) were also divinely inspired. The nub of the problem is that belief in the divine inspiration of a work is not a criterion at all, but the construct of a community. 62
The concept of “selection” is problematic in another sense. How are canonical books different from those compositions left out of the canon? For instance, how is the selected book of Exodus different from the book of Jubilees? Both recount the giving of the law on Mount Sinai and include regulations concerning the observance of rituals, feasts and festivals. How is the canonical Deuteronomy different from the non-canonical Temple Scroll? A logic based on identifying criteria leaves this fundamental question unanswered. One needs to turn to another way of thinking, an indicative logic.
Based on the analytical philosophy of Ludwig Wittgenstein, I have suggested that a better way of thinking about the definition of a “biblical book” is to use the concept of family resemblances. This approach does not require that a biblical book has an essence different from non-biblical books, implying that the boundaries between them are blurred. Thus, for instance, the book of Genesis does not have an essence different from the Genesis Apocryphon or the book of Enoch. They are not mutually exclusive, and they differ in degree and not kind.
Family resemblances is a concept based on a biological analogy of the genetic makeup (DNA code) that manifests itself in the physical attributes of a family. Individual characteristics or traits are not all the same in a family, but the genetic information for eye, hair and skin colour; the shape of the nose, face and head; and height and body-type are passed on and contribute to family resemblances. 63 Each familial characteristic (e.g., blue eye colour) could also be found in others who are not biologically related to the family. Members of a family, therefore, are similar to each other without having individual traits that are unique to them. It is the combination of distinctive traits that contribute to the resemblance of family members. 64
The logic is not based on a set of criteria against which the biblical books are measured. It is non-essentialist and points to the combined traits that are shared between the biblical books that identify them as a group, without needing to show that individual features are unique to them. It does not mean that any book is similar to any other writing, any more than any family member is similar to any other person. 65 Each book or person is unique in its literary and genetic make-up respectively, but this uniqueness does not mean that individual features or traits are unique.
On this approach, one can identify the books of the canon to have the following traits: they recount the story of Israel as a nation from the beginning and through history; they attest to the content of beliefs and practices, including rites, customs, laws, and teachings of Israelite/Judaean religion; and they often, but not invariably, attribute to David or Solomon as author of the biblical books.
Footnotes
1
Timothy H. Lim, “An Indicative Definition of the Canon” in When Texts are Canonized ed. Timothy H. Lim with Kengo Akiyama (Atlanta: Brown Judaic Studies, 2017), pp. 3-12.
2
For a critical review of the literature, see my, “The Literature of Early Judaism” in Early Judaism and its Modern Interpreters: Second Edition ed. Matthias Henze and Rodney Werline (Atlanta: SBL, 2020), pp. 257-80.
3
In the prologue to the books of Samuel and Kings, Jerome noted three ways of counting the biblical books. There are 22 books of the Old Testament canon corresponding to the “elementary characters” of the Hebrew alphabet, 5 of Moses, 8 of the Prophets, and 9 of the Hagiographa. This is the count of the majority. He notes, however, that some people separate the book of Ruth from Judges, and Lamentations from Jeremiah, thus there are 24 books in the canon, the same number as the twenty-four elders who adored the lamb in the book of Revelation (Rev 4:4, 5, 10; 5:8; 11:16; and 19:4).
4
Epiphanius makes this explicit when he names the first five books according to their Greek and Hebrew names (Pan. 9.2.1). What we know about the formation of the Samaritan Pentateuch is discussed in my article, “The Emergence of the Samaritan Pentateuch” in Reading the Bible in Ancient Traditions and Modern Editions: Studies in Textual and Reception History in Memory of Peter W. Flint, eds. Andrew Perrin, Kyung Baek, and Daniel Falk (Atlanta: SBL Press, 2017), pp. 89-104.
5
Robert Kraft, “Para-mania: Beside, before and beyond Bible Studies” JBL 126:5 (2007): 10, who defines canon in Greek codicological terms, including the Old and New Testaments (see my, “Literature of Early Judaism”, pp. 260-1).
6
In the third century CE, Origen and Africanus discussed three criteria: whether the Old Testament should be restricted to books that were used in the synagogue; whether books used by the church, but not included in the Jewish canon, should be included; and whether only books written in Hebrew should be included. These were clearly the criteria as understood in the third century (see my, “Indicative Definition”, pp. 1-3).
7
So, Craig Evans, “Jesus and the Beginnings of the Christian Canon” in When Texts are Canonized ed. Timothy H. Lim with Kengo Akiyama (Atlanta: Brown Judaic Studies, 2017), pp. 95-107.
8
It was in the legend of the translation of Jewish law into Greek that the Rabbis found their requirement that holy scriptures had to be written on a scroll of leather, in square script (which they called ‘the Assyrian script’), and in ink. Not all Rabbis held this view. According to the stammaim, the halakha follows Rabban Gamaliel who permitted books of holy scriptures to be written in Greek. See my, “Rabbinic Concept of Holy Scriptures as Sacred Objects” in Scribal Practices and the Social Construction of Knowledge in Antiquity, Late Antiquity and Medieval Islam ed. Myriam Wissa (Leuven: Peeters, 2018), pp. 127-141.
9
The communities reflected in the sectarian Dead Sea Scrolls did not restrict their authoritative scriptures to the biblical books. See below.
10
So, Philip Alexander who argues for the openness of the canon because the order of the books of the Prophets and Writings continued to fluctuate in manuscripts, and stabilization occurred only with the emergence of the great codices of St. Petersburg and Aleppo. But the fluctuation of the order is a secondary feature, see my Formation of the Jewish Canon, pp. 180-1. See now his “Textual Authority and the Problem of the Biblical Canon” in Is There a Text in this Cave? ed. A. Friedman, C. Hempel, and M. Ciota (Leiden: Brill, 2017), pp. 42-68.
11
Of Jewish origin and preserved in a Christian manuscript, see my Formation of the Jewish Canon, pp. 41-3, 193-94; and “Indicative Definition”, pp. 1-3.
12
For a discussion of the earliest canonical notices and lists, see my, Formation of the Jewish Canon, pp. 35-53; and “Literature of Early Judaism”.
13
Frants Buhl, Kanon und Text des Alten Testaments (Leipzig: Faber, 1891); G. Wildeboer, Die Entstehung des alttestamenlichen Kanons (Gotha, 1891); and especially H. E. Ryle, The Canon of the Old Testament (London: MacMillan and Co., 1892).
14
Heinrich Graetz (1871) was the first to label Javneh a ‘council’, and it is possible that he may have derived this idea from Baruch Spinoza. Graetz proposed that the final Writings section was assembled in two stages, first by the Pharisees and Sadducees in 65 CE, and then by the “synode” of Javneh. For him, the final closing of the canon took place only with the closing of the Mishnah in 189 CE. For a discussion of this topic, see my “Literature of Early Judaism”, pp. 262-3.
15
It was argued that the Samaritans, when they separated from the Jews, took with them the Pentateuch, which they considered canonical. Ryle dated this schism to 432 BCE, and inferred that the Pentateuch must have been closed sometime in the 5th century (Canon of the Old Testament, p. 93). The sources of 1 Kgs 17 and Josephus’ retelling are biased. Scholars, notably Magnar Kartveit and the late Gary Knoppers, consider the Samaritans as northern Israelites. For the implications for canon formation, see Lim, Formation of the Jewish Canon, pp. 18-20.
16.
The Old Testament Canon of the New Testament Church and its Background in Early Judaism (London: SPCK, 1985).
17.
The Oracles of God: the Old Testament Canon (Saint Louis: Concordia, 1999).
18.
Oracles of God: Perceptions of Ancient Prophecy in Israel after the Exile (Darton, Longman & Todd, 1986).
19.
“Canon and Content” in When Texts are Canonized ed. Timothy H. Lim with Kengo Akiyama (Providence, RI: Brown Judaic Studies, 2017), pp. 82-3.
20.
The best known of such disputes, of course, took place during the Reformation when Protestants rejected the Bible of the Catholics and returned to the Hebrew Bible. Martin Luther himself had an idiosyncratic and largely hostile view about the Old Testament. The re-opening of the canon continues to be raised from time to time.
21.
I am not counting comparative studies of non-biblical literature that may be earlier.
22.
See, for instance, Manfred Oeming, “The Way of God: Early Canonicity and the ‘Nondeviation Formula’” in When Texts are Canonized ed. Timothy H. Lim with Kengo Akiyama (Atlanta: Brown Judaic Studies, 2017), pp. 25-43.
23.
“The Formation of the Biblical Canon in Rabbinic Judaism” in The Canon of Scripture in Jewish and Christian Tradition ed. Philip S. Alexander and Daniel Kaestli (Lausanne: Editions du Zebre, 2007), p. 64.
24.
For an analysis of Melito’s letter, cited by Eusebius, Ecclesiastical History 4.26, and its relevance for dating the closing of the Rabbinic canon, see my Formation of the Jewish Canon, pp. 35-38
25.
Josephus corroborates this understanding, dividing early Jewish literature into just two categories of scripture and non-scripture (Ag. Ap. 1.38-41; see Lim, “Literature of Early Judaism”, pp. 266-69). The earliest canonical notices and lists are discussed in Formation of the Jewish Canon, pp. 35-53, 191-94.
26
Detailed analyses of each of these lists is found in my Formation of the Jewish Canon, ch. 3. See now, Juan Carlos Ossandon, The Origins of the Canon of the Hebrew Bible: An Analysis of Josephus and 4 Ezra (Leiden: Brill, 2019), who questions the Pharisaic nature of Rabbinic Judaism.
27
Authority, of course, can be vested in the person (Moses, Aaron, priests, etc), a ritual performed by a community or a story of a people repeated orally from one generation to the next (cf. Deut 26:4-9). Textual authority confers authority on the written word, a “scripturalization” that implies the rise of literacy and the recording of material that could be verified by others. In Ezra-Nehemiah, ככתוב (“as it is written”, Neh 8:15) or ככתוב בתרורה (“as it is written in the torah”, Ezra 3:2; Neh 10:35, 37) refers to sources external to Ezra-Nehemiah that the implied readers could verify. In Rabbinic Judaism, “Torah” likewise has a range of meanings, and it could refer to the books of Moses and the entire Hebrew Bible, including the oral tradition that accompanies and interprets it.
28
Genesis (Neh 9:6, 7-8); Exodus (Neh 9:9-11, 12-21); Leviticus (Ezra 3:4; 6:19-22; Neh 8:14-17; 10:32; 13:15-22), Numbers (Neh 9:12-22); Deuteronomy (Ezra 3:4; 6:19-22; Neh 10:32; 13:1-2; 13:25) and Joshua (Neh 9:23-25; cf. 9:26-37).
29
For a discussion of the canonical process in the Persian period, see my Formation of the Jewish Canon, pp. 54-73. I have suggested that the Samaritan Pentateuch was likewise formed in the fourth century BCE, in relation to the building of the cultic site on Mount Gerizim (“Emergence of the Samaritan Pentateuch”).
30
2 Macc 2:13-15, often thought to be important evidence, does not refer to the founding of a Maccabean library. The correct translation of v. 14 is that Judas collected all the books that “had fallen to pieces on account of the war” (see Formation of the Jewish Canon, pp. 114-17).
31
Rabbinic tradition traces the origins of the targum to an earlier period, to Ezra reading the Torah to the gathered Judeans by the Water Gate in Neh 8. See my Formation of the Jewish Canon, pp. 62-69 for the textual and historical complexities of this passage.
32
Often called “Letter of Aristeas”, this document is not a letter, but a charter myth. Scholars used to call this fictional author “Pseudo-Aristeas”, but it makes little sense to use the prefix, meaning “false”, when “Aristeas” himself is unknown. He presents himself as a Greek in Ptolemy’s court, but is likely to have been a Jew, judging by what he knows and says about Judaism. See now, Benjamin Wright, The Letter of Aristeas: Aristeas to Philocrates or On the translation of the Law of the Jews (Berlin: de Gruyter, 2015).
33
It was not written by a contemporary author, as evidenced by slips betraying a later composition. The clause “even now” (§182) suggests that the origins of arrangements regarding food and drink is to be found at an earlier time. The explanatory gloss of how kings conducted royal business uses the past tense “used to administer” (§28). Demetrius never served as librarian of Alexandria. And the denouement of the story did not mention the presentation of the translated rolls to the library (§308-311). For the positive evidence of authorship and dating of Aristeas, see my, Formation of the Jewish Canon, pp. 73-85.
34
Based on the historical information in Philo’s account. See Timothy H. Lim, “The Idealization of the Ptolemaic Kingship in the Legend of the Origins of the Septuagint” in Times of Transition. Judea in the Early Hellenistic Period ed. Sylvie Honigman, Christophe Nihan, and Oded Lipschits (University Park: Eisenbrauns, 2021), pp. 231-39.
35
Very little, if any, use is made of the book of Genesis with only one allusion to human propensity (§277) that could possibly refer to the concept of the good and evil inclination based on Gen 2:7. Alternatively, “the entire law” could refer to a “tetrateuch”, consisting of books from Exodus to Deuteronomy, since the book of Genesis is history and does not fit the description of Law, according to Philo’s understanding of the Pentateuch (Moses 2.45-65). Note: in scholarly designation, ‘tetrateuch’ usually refers to the first four books of the Pentateuch, from Genesis to Numbers.
36
Harry Orlinsky characterized this scene as the re-enactment of Sinai with three key elements: the community of Jews is used in an official sense symbolizing “the Jewish people”; the seventy-two translators are reminiscent of the elders of the twelve tribes and the priests who witnessed the event at Sinai; and the name of the high priest in Jerusalem, Eleazar, called to mind the third son of Aaron in the biblical account (“The Septuagint as Holy Writ and the Philosophy of the Translators” HUCA 46 [1975]:89-114).
37
The date of the grandfather’s Wisdom is inferred from the likely age of the adult grandson (between 25 and 40 years old), the number of years that separates one generation from the next (40-50 years), and the internal evidence of the praise of the high priest Simon (219-196 BCE; Sir 50:1-21), who was already considered a figure of the past (“in his days”, Sir 50:1, 3), and the absence of any reference to the hellenizing policies of Antiochus Epiphanes IV (175-164 BCE). See my, Formation of the Jewish Canon, pp. 95-97.
38
The Wisdom of Ben Sira is one of the most textually complex works of ancient Jewish literature. Originally written in Hebrew and then translated into Greek in the second century BCE, its textual tradition attests to two Hebrew recensions (HT1 and an expanded one, HT2) and two Greek recensions (GI and GII). See Benjamin Wright, No Small Difference: Sirach’s Relationship to its Hebrew Parent Text (Atlanta: SBL, 1989); and Jean Sébastian-Rey and Jan Joosten, The Texts and Versions of the Book of Ben Sira: Transmission and Interpretation (Leiden: Brill, 2011).
39
Primarily as mediated by the grandson in his translation. Ms B preserves only Sir 38:24-27.
40
The grandson refers to the law, the prophets, and the other books three times in the Prologue. In all three occasions, the expression of the law and the prophets/prophecies remain invariable. By stark contrast, the expression used for the implied third category varies in all three instances: “the others that followed them”; “the other books of our ancestors”; and “the rest of the books”. See my, Formation of the Jewish Canon, pp. 94-106.
41
The grandson makes this plain in the Prologue when he recommends to scribes (“lovers of learning”) to read his grandfather’s book of Wisdom with goodwill and attention, so that they may make progress in living according to the law.
42
Unlike Sir 39:1-3, the Prologue recommends only Israelite literature.
43
It is often thought that chs. 44-49 is separate from the appendix of ch. 50, but the praise of Simon is the culmination of the chapters, and not a secondary section of additional material, as he is lauded as “the great one of his brothers” and “the splendour of his people” (50:1; see my, Formation of the Jewish Canon, pp. 104-6).
44
For an itemization of the biblical passages in the laus patrum, see Appendix 5, Formation of the Jewish Canon. The discussion can be found on pages 102-106 of the same work.
45
Simon is also referenced in Josephus, Ant. 12.224-225 and 3 Macc 2.
46
This was the judgement of William Foxwell Albright in a letter of 15 March 1948 to the American Schools of Oriental Research, Jerusalem School (cited in John Trever, “The Discovery of the Scrolls”, The Biblical Archaeologist 11.3 [1948]: 55).
47
The identification of the sectarians of the scrolls with the Essenes remains, but there is an emerging consensus that there were several communities not only at Qumran but dispersed throughout Judaea (see John J. Collins, Beyond the Qumran Community: The Sectarian Movement of the Dead Sea Scrolls [Grand Rapids, MI: Eerdmans, 2010]). There is good evidence that date the origins of the sectarian community, once thought to belong to the Maccabean period, are to be found in the first century BCE, during the settlement of the east by Ptolemy to the Herodian period (see my identification of the “wicked priests” [sic] as the last three high priests of the Hasmonean dynasty in The Earliest Commentary on the Prophecy of Habakkuk [Oxford: OUP, 2020]), pp. 19-34).
48
Josephus’ description of a broadly bipartite collection of authoritative scriptures converges well with what we know from the sectarian scrolls (see my Formation of the Jewish Canon, pp. 148-55).
49
See, e.g., the library of Alexandria and its inclusion of the translated laws of the Jews. Disagreement among the corpus manifests itself in different conceptions (e.g., epistemology) rather than direct disputes.
50
See Timothy H. Lim and John J. Collins, “Introduction” in The Oxford Handbook of the Dead Sea Scrolls (Oxford: OUP, 2010). That the Dead Sea Scrolls is a scribal collection (so Sidnie Crawford, Scribes and Scrolls at Qumran [Grand Rapids: Eerdmans, 2019]) is not incompatible with the view of a heterogenous collection, since they explain the content and scribal practice of the scrolls respectively.
51
It is now regarded as a pentateuchal text (see Michael Segal, “4QReworked Pentateuch or 4QPentateuch?” in The Dead Sea Scrolls: Fifty Years after Their Discovery, 1947-1997 ed. Lawrence H. Schiffman, Emanuel Tov and James VanderKam [Jeruslaem: Israel Exploration Society, 2000], pp. 391-99).
52
Scholars argue about the different meanings of ‘sectarian’, but there is wide agreement that the scrolls of the Serekh, the Damascus Document, Pesharim, Hodayot, 4QInstructions, and Enoch are the writings of communities that distinguished themselves from common Judaism by their distinctive teachings.
53
So Johann Lust, “Quotation Formulae and Canon in Qumran” in Canonization and Decanonization ed. Arie van der Kooij and Karel van der Toorn (Leiden: Brill, 1998), p. 68.
54
Detailed discussions are found in the following: “Authoritative Scripture and the Dead Sea Scrolls” in The Oxford Handbook of the Dead Sea Scrolls ed. Timothy H. Lim and John J. Collins (Oxford: OUP, 2010), pp. 303-22; “The Alleged Reference to the Tripartite Division of the Hebrew Bible” RevQ 20.1 (2001): 23-37; and “The Writings in the Hellenistic and Roman Period” in The Oxford Handbook of the Writings of the Hebrew Bible ed. Donn F. Morgan (Oxford: OUP, 2018), pp. 33-48.
55
This is not the same as Jubilees’ own understanding of its own status as coeval and complementary to the first law, the Torah of Moses. See Lim, Formation of the Jewish Canon, pp. 131-35.
56
The continuous and thematic pesher (see my Pesharim [London: Bloombury, 2000]). Some also identify the “isolated pesher” as a genre; although, this could be alternatively understood as the quotation of a pesher.
57
For midrash see my, “The Origins and Emergence of Midrash in Relation to the Hebrew Scriptures” in The Midrash. An Encyclopedia of Biblical Interpretation in Formative Judaism eds. Jacob Neusner and Alan J. Avery-Peck (Leiden: Brill Academic Publishers, 2004), Vol. 2: 595-612.
58
It is often said that Paul’s “Bible” was the Septuagint, but this characterization misunderstands textual classification. He most commonly quotes from the uniform text of the MT and LXX (see my, Formation of the Jewish Canon, pp. 165-68).
59
Formation of the Jewish Canon, pp. 48-49.
60
Josephus states that so great was the Pharisees’ power over the multitudes that when they spoke against the king or high priest, they were immediately believed (Ant. 13.288).
61
Exceptional are the Samaritans who have fewer books.
62
“Indicative Definition”, pp. 3-12. Other suggested criteria (like the original language of Hebrew and use of a book in the synagogue or church) are ex post facto rationalizations.
63
Similar combinations of physical traits, usually among the same ethnic group, could result in coincidental family resemblance.
64
It goes without saying that not all family members have all the traits!
65
Indicative logic can also show that some non-biblical writings (e.g., documents used in transacting daily life) are not similar to the biblical books.