How do we know that a particular writing served as a rule of faith and practice in the life of Jews or Christians in antiquity?
Lines of Evidence
In the assessment of the canonical status of a particular writing, five lines of evidence deserve consideration.
(1) The existence of pesharim, midrashim, or other forms of commentary on a particular text is strong evidence for said text ruling faith and life in a given context. Thematic pesharim among the Dead Sea Scrolls comment on passages from the following writings; each, in other words, was understood as canonical in the functional sense: Deut, 2 Sam, Exod, Amos, Pss, Ezek, Dan, and Isa (4Q174); Deut, Num, and Josh (4Q175); Isa and Zech (4Q176); and Pss, Isa, Mic, Zech, Ezek, and Hos (4Q177). Continuous pesharim are attested for the following writings: Isa, Hos, Mic, Nah, Hab, Zeph, and Pss. The focus of these texts on actualization in contemporary figures and events, and their appeal to the above writings and no others, is comparable to the contents of a remark attributed to the risen Christ in Luke: “This is what I meant by saying while I was still with you, that everything written about me in the law of Moses and in the prophets and psalms was bound to be fulfilled” (24:44). Cf. Luke 3:3-6; 4:16-21; and Matt 11:2-6, 10. The evidence cited points to the conclusion that these writings and not others were thought to hold within them a blueprint for faith and practice.
In rabbinic Judaism, exegetical and homiletical energies focused on the books of Moses, with actualization on both the halakhic and aggadic levels. Midrashim with a strong halakhic emphasis include Mekhilta de_Rabbi_Ishmael (on Exod 12:1-23:19; 31:12-17; 35:1-3); the Mekhilta de Rabbi Shimon ben Yochai (on Exod 3:2, 7-8: 6:2; 12:1-24:10: 30:20-31:15; 34:12, 14, 18-26; 35:2); Sifra (on all of Lev); Sifre (on the legal portions of Num and Deut, and a portion of the narrative of the latter); and Sifre Zutta (on Numbers). The hallmark of strongly aggadic midrashim is actualization in terms of a metanarrative current at the time of composition. Examples, in chronological order of composition: Bereshit Rabbah; Eichah Rabbah; Wayiqra Rabbah; Pesiqta deRav Kahana and Pesiqta Rabbati (on the readings of the feasts and the special Sabbaths); Tanchuma (on the whole Pentateuch); Devarim, Bamidbar, and Shemot Rabbah; Midrash Yonah , and Midrash Tehillim. Traditional commentary on the Torah is collected in Midrash Hagadol; on the Pentateuch and the Five Scrolls, in Midrash Rabbah ; on the prophetic writings and Psalms, Proverbs, and Job, in Yalkut haMakhiri; on the entire Tanakh, in Yalkut Shimoni .
The exegetical and homiletical ouput of early Christianity was also prodigious. A few examples may illustrate. Hippolytus (c. 170-c. 236) is credited with commentaries On Genesis, On the Blessings of Isaac, Jacob, and Moses, On the Blessings of Balaam, On Judges, On Ruth, On the Books of Samuel; a homily on David and Goliath and a treatise on David and the Ark are also known; On the Psalms, On Proverbs, On Ecclesiastes, On the Song of Songs, On Parts of Ezekiel, On Daniel, On Matthew, and On Revelation. He (or another whose name is now lost) resignifies the Old Testament in terms of Christ and his Church.
Origen (c.184-c.254) sought to interpret the text along three different lines (On First Principles 4,2,4): literal, moral, and allegorical. He wrote commentaries on Song of Songs, Matthew, John, and Romans; scholia on the Octateuch (Gen - Ruth), Lamentations, and the Psalter; and homilies on the Octateuch (extant remains cover Genesis, Exodus, Leviticus, Numbers, and Joshua); Jeremiah, and the Song of Songs. In the same vein as Origen, Didymus the Blind (c. 319-98) commented on Genesis, the Octateuch and Kings (Gen – Ruth + 1 Sam – 2 Kgs), Zechariah, Psalms, Proverbs, Job, and Ecclesiastes; John, Acts, Romans, Corinthians, and the Catholic Epistles. He is also known to have commented on Isaiah, Matthew, and the Apocalypse, but nothing thereof has survived.
Jerome (c. 342-420) stands outs for his programmatic attention to the Hebrew, Aramaic, and Greek originals he translated and commented on. His first exegetical work, Treatises on the Psalms, is an adaptation of an earlier work by Origen; his commentaries on Philemon, Galatians, Ephesians, Titus, and Ecclesiastes contain more that is new. Jerome the text critic and philologian is evident in his mature works: the Hebrew Questions on Genesis, and commentaries on the major and minor Prophets, including Daniel. His commentaries on Matthew, Mark, and the Apocalypse are less original.
Theodore of Mopsuestia (c. 350-429) attended to the historical sense and commented on Genesis, Exodus, and the other books of the Pentateuch; the Psalms, Job, and Ecclesiastes; the major and minor Prophets; Luke, John, Acts, and Paul’s major and minor epistles. Theodoret of Cyrrhus (c. 393-c. 466) wrote quaestiones on the Octateuch, Kingdoms, and Chronicles, and commentaries on the major and minor Prophets, the Psalms, Song of Songs, and Paul’s epistles.
It is noteworthy that among the extant examples of the chief genres of exegetical literature of the first six centuries - continuous commentaries, scholia, and quaestiones – Ben Sira, Wisdom of Solomon, Judith, Tobit, and 1-2 Maccabees are not treated, nor for that matter, are 1 Esdras, 2 Esdras (= Ezra-Nehemiah), or Esther. Commentaries on Chronicles and Baruch are extant from one author only (Theodoret). On the other hand, comment by three authors on the Song of the Three is extant (Hippolytus, Origen, and Theodoret); on Susanna, Bel, and the Dragon, by two (Hippolytus and Origen).
The fact that writings commented on more than once in extant exegetical literature cover the entire collection of writings now contained in the Tanakh with the exception of Esther, Chronicles, and Ezra-Nehemiah, plus the additions to Daniel, but, as far as inherited Jewish literature is concerned, nothing beyond, requires an explanation. In my view, the preponderance of attention paid to the Octateuch, Kingdoms, Prophets (including Daniel and Lamentations), Psalms, Job, Proverbs, Ecclesiastes, and Song of Songs reflects the fact that precisely said works, ab antico or more recently, had previously attained pride of place in Jewish tradition.
To be sure, canonicity and the fact of forming the base text for a commentary stand in partial correlation only. The book of Daniel may illustrate. In rabbinic Judaism, commentaries were not written on Daniel, nor was it read in the synagogue, but it would be wrong to conclude that it lacked authority or was not foundational to teaching. It was. Pre-rabbinic, extra-rabbinic, and rabbinic Jewish eschatological hopes depended heavily on the four kingdoms scheme of Daniel (the fourth kingdom now Rome). The book’s importance as a source of exempla of heroic faith is also well-attested. 
It is sometimes possible to be specific about the debt Christian exegesis owed to Jewish exegesis. Comparing the Pentateuch with the Prophets, it has been noted, “Clement of Alexandria’s quotations or references stand in the proportion of 5 to 3.” In the New Testament, the situation is reversed. An explanation lies near at hand. The focus of Clement (d. 215) on the Pentateuch, the treatment of the hexaemeron (six days of creation) by Hippolytus, Origen, and Basil of Caesarea (c. 329-379), and of the life of Moses by Clement and Gregory of Nyssa (d. 385), depends on Hellenistic Jewish exegetical tradition. The exegetical opus of Philo in particular was treasured in Christian antiquity. Philo’s writings, along with those of Josephus, exercised a profound influence on Christian tradition and became part and parcel of it.
Alongside writings enjoying pride of place among Jews and Christians alike, additional writings came to serve as quarries of prooftexts for the establishment of truth and as a goad to action. An early catechetical work, Testimonia ad Quirinum, may illustrate. Attributed to Cyprian of Carthage (d. 258) in the manuscript tradition, it provides scriptural proofs for a series of traditional teachings. Book One intends to demonstrate the nullity of the Jewish faith; Book Two treats christology; Book Three, the correct comportment of the believer. The pattern of citations is remarkable.
Book One cites the Old Testament profusely, 74x total: the Pentateuch, 13x; Josh-2 Kgs, 5x; Isa, 25x; Jer, 13x, Ezek, 1x; Dan, 1x; Pss, 9x; Prov, 1x; Ezra-Neh, 1x. New Testament confirmation is added on a few occasions from Matthew, Luke, John, Acts, and Paul’s Epistles. No other writings are cited.
Book Two cites both OT and NT profusely, and across a wider spectrum of texts. A christology is built up out of passages from the same books as before, except Ezra-Neh, with a preponderance of proofs from Isa and Pss. But, besides Prov 8 and 9, Ben Sira 24, Baruch 3 (under the title “Jeremiah”), and Wisdom 2 appear. The same NT books are cited as in Book One, with the addition of 1 Peter, 1 John, and a host of references to the Apocalypse of John.
Like Book Two, Book Three cites both OT and NT profusely, but across an even wider spectrum of texts. Job, Prov, and Eccl are cited often; Ben Sira, more often still. Wisdom, Tobit, 1-2 Maccabees, Susannah, and the Song of the Three are also cited; Wisdom and Tobit, several times. From the NT, Mark, 1-2 Timothy, Titus, and 2 Peter make their appearance. On the other hand, Esther, Judith, Hebrews, James, and Jude do not, in the Testimonia or elsewhere, in Cyprian’s writings. Evidently, none of these works, if they were known, were deemed suitable to appeal to in the African church tradition Testimonia reflects.
Output in the field of scriptural exegesis waned by the end of the patristic age. Anthologies of exegesis culled from earlier works were compiled from the 6th cent. forward. Called catenae, the form they commonly took in the Greek tradition was the following, as Carmelo Curti observes: “the biblical text, written in larger letters, is in the centre of the page, against the inside margin; round it on the three outside margins are arranged the exegetical extracts; more rarely, they occupy all four margins, in which case the scriptural text is in the exact centre of the page.” A core canon is once again in evidence. Procopius of Gaza (465-c. 530), the first attested catenist, collected comment on the Octateuch, Kingdoms, Ecclesiastes, and Song of Songs. His most illustrious successor, Nicetas of Heraclea (11th-12th cent.), composed catenae on Job, Psalms, the major Prophets, the Gospels, and the Pauline epistles.
Non-inclusion in a core set of authoritative writings notwithstanding, quotes from a text in teaching documents nonetheless imply that said text was deemed suitable for purposes of instruction in a given time and place. For example, attributed and unattributed quotes from Hebrew Ben Sira occur in the Talmuds and other rabbinic documents. The example is instructive. A book might be deemed worthy of study but not considered a basis for establishing a direction of thought or action, nor fit to read in assembly. In a nutshell, that is what discussions of Ben Sira in the Talmuds demonstrate.
On the other hand, attributed quotes of Ben Sira, sometimes introduced by a phrase like “scripture says,” are found in teaching documents of the early church. The example of Testimonia ad Quirinum was already given. To be sure, Athanasius (Festal Letter, 39), Rufinus (Comm. in Symb. Ap. 37-38), and Jerome (Prologue to the Books of Solomon [Vulg.]) move in a different direction. According to them, Ben Sira might be deployed in catechesis, and they acknowledge it was read in the churches, but they deny that it should be used to establish a point of dogma. Others, however, dissented, Augustine foremost among them (De Doct. Chr. 2, 8; De Civ. Dei, 18, 20, 1). For Augustine, and Cyprian before him, Ben Sira and Wisdom were just as useful as Proverbs for establishing a direction of thought or action . A canon agreed upon by all Christians, even by those of roughly the same tradition, did not exist in antiquity.
The Liturgical Use of Authoritative Literature
(2) Examples of the use of an excerpt from a book in the liturgy of one or more branches of Judaism or of one or more branches of Christianity. A distinction must be made between attributed and unattributed use. The unattributed use of parts of the Hebrew Ben Sira tradition in Jewish liturgy to the present is one thing. The attributed use in worship of a passage of 2 Baruch (78.1-86.1) to which lectionary manuscripts of the Syriac church tradition attest is another. Express quotation and public proclamation are strong indications of the sense that Syriac-speaking Christians had that God continued to speak to them through said 2 Baruch.
In Jewish practice, after a reading from the Torah and a reading from the Prophets (which might be skipped), the homily might begin with a petiḥta or proem in which a verse from the Psalter or Proverbs is interpreted first of all. Thus, when preaching on the text, “And Abraham was old” (Gen 24:1), the following verse might be quoted: “The hoary head is a crown of glory; it shall be found in the way of righteousness” (Prov. 16:31) (Genesis Rabbah ad loc). The body of the sermon consisted of halachic and aggadic exposition of details of the Torah passage. The Ḥatimah or conclusion might be eschatological in nature and depend on the Haftarah. Quoted biblical verses and rabbinic dicta form the warp and woof of early sermons; in later times, citations from Aristotle, other philosophical material, or the Kabbalah are attested.
In Christian practice, a Gospel passage was often but not necessarily the read out and preached text. The practice of reading through an entire Old or New Testament book on successive occasions is also attested. More often than not, preaching method followed patterns taken over from inherited Jewish tradition, transposed, of course, onto specifically Christian themes.
The texts of greatest authority in a given setting, it stands to reason, are to be identified with the texts most often preached. A study of extant sermons in the various traditions with a view to identifying the texts which were expounded most frequently, and in what context, remains to be done.
Not all church oratory was scripture-based. The traditional oratory of the ancient church includes collections of homilies, sermons, and hagiographies. Generally speaking, one or the other of the three genres occupied the central place in liturgical celebration. A homily, by definition, took its cue from a scriptural text. Sermons are another matter. Their point of departure was often a specific topic or occasion. What texts topical and occasional sermons appeal to is a subject matter of interest in its own right. A hagiographical oration took its cue from traditional accounts of the acts, passion, and/or martyrdom of saints. By this and other means, early Christianity appropriated, reconfigured, and supplanted a prevailing feature of pagan antiquity, the commemoration and veneration of the dead.
The evidence provided by Jewish and Christian lectionary traditions is of great interest. For Christian tradition, Kevin Edgecomb provides a wealth of primary data. The evidence in hand for both Jewish and Christian tradition reflects realities of a relatively late period (4th cent. ce forward). Cases of direct borrowing from Jewish into Christian tradition are rare; I point out an example elsewhere. In Christian tradition, earlier evidence reflects a sparser and more circumscribed use of excerpts from writings which achieved deuterocanonical status than later evidence. Jewish tradition also read entire and excerpted a larger subset of the literature of the Tanakh in post-Talmudic than in earlier times.
Lectionary traditions demonstrate the need to distinguish between formal and actual canons. The old Slavonic tradition may illustrate. The canon in this tradition, insofar as it is reconstructible, was as inclusive as the Greek tradition to which it is indebted. Extant witnesses to the prophetologium, the Slavonic lectionary of Old Testament liturgical lections, attest however to a selective deployment of its contents. Lections drawn from the Octateuch and the Prophets are numerous, e.g., Genesis and Isaiah. Wisdom of Solomon, however, is equally or better represented (2:1, 10-22; 3:1-9; 4:1, 9-15; 5:15-6:3; 6:11, 17-18, 21-23; 7:15-16, 21-11, 26-27, 29-30; 8:2-4, 7-9, 17-18, 21-9:5; 9:10-11, 14; 10:9-10, 12; 15:1; 16:13). Baruch 3:36-4:4 appears as a lection. Both Psalms and Odes (an anthology of songs excerpted from a full range of canonical literature, the anthology is traditional and taken over from Greek tradition), the Prayer of Azariah and the Song of the Three Youth included, are prominently deployed. But lections from Ruth, 1-2 Kingdoms, 1-2 Chronicles, 1-2 Esdras, 1-4 Maccabees, Esther, Song of Solomon, Ben Sira, Qohelet, Lamentations, and Epistle of Jeremiah are unattested.
Patterns of Explicit Quotation
(3) Patterns of explicit quotation in Jewish and Christian literature bear on the question. Unfortunately, investigations have seldom been conducted with methodological rigor. Devorah Dimant lists and classifies explicit quotations of authoritative texts in late Second Temple period narrative literature. Across the corpus (which she fails, unfortunately, to delimit), quotations from the Pentateuch preponderate, for a total of 13 out of 19x. Amos, Isaiah, and Ezekiel are each quoted explicitly once; Psalms twice; and Proverbs once. The literature with explicit quotations she covers: Tobit, Judith, Susanna, Baruch, and 1-4 Maccabees. 4 Maccabees, however, as Dimant notes, is not a proper narrative, but a philosophical encomium. It is also the latest writing in the sequence, perhaps mid. 1st cent. ce. It quotes from authoritative literature more often than all the others (9x), and over a broader range of texts: the Pentateuch, the Prophets, Psalms, and Proverbs are all represented. The patterns Dimant explores are far from haphazard or coincidental.
The range of literature excerpted in late Second Temple period narrative is consonant with the range of literature excerpted for entirely different purposes in the Qumran pesharim. Patterns of scripture quotation in the New Testament point in the same direction. Aside from two or three outliers (4Q228 Frg. 1 i 9, which explicitly quotes Jubilees; Damascus Document [Cairo Geniza A] 16:1-3, which promotes the authority of Jubilees; Jude 14-15, which explicitly quotes a passage from Enoch), express quotation of authoritative literature in Jewish and Christian writings through the end of the first century of the current era is limited to the following corpus: the literature that later came to be contained in the first two divisions of the Tanakh (the Pentateuch and the Prophets), plus Psalms, Proverbs, and Daniel. Expressly quoted scripture through the end of the Second Temple period, so far as the evidence takes us, is limited to these texts.
Other texts, of course, were also treasured. Allusions to figures and events written up in other works, both within and beyond the limits of canons attested in later times, are scattered throughout late Second Temple Period literature. Sir 49:13 evokes the figure of Nehemiah in terms which prove familiarity with the relevant material in Ezra-Nehemiah; 49:14-16 idealizes the figures of Enoch, Joseph, Shem, Seth, Enosh, and Adam in accordance with extra-Pentateuchal tradition; Hebrews 11:35-38 evokes the Maccabean martyrs and the fate of Isaiah as recounted in popular legend; Jude 9 summarizes an event in the life of Moses recounted in the Testament of Moses. My point in drawing up this list is not to suggest that writings as diverse as Enoch, Life of Adam and Eve , Ezra-Nehemiah, 1-2 Maccabees, and the Testament of Moses belong on the same plane. I would side with rabbinic judgment on these books in each and every case. The point is another. It is clear that Ben Sira, the author of Hebrews, and Jude did not feel bound to stick to information contained in a collection of writings they otherwise recognized as authoritative when it came to describing events and figures of the past. That only makes them like, not different from, Josephus, the selfsame rabbis, and the ancients in general. It follows that Ben Sira’s laud of Nehemiah is a slender basis on which to conclude that Ezra-Nehemiah functioned canonically in the 2nd cent bce on a par with texts expressly quoted in that and the following two centuries.
Express quotations that involve an appeal to authority are an incontestable index of canonicity; allusions, in and of themselves, not so much. Popularity and canonicity do not correlate either. The case of Tobit may illustrate. Its attestation in multiple copies at Qumran (Aramaic and Hebrew) and across the entire sweep of ancient Christendom (Greek, Latin, Syriac, Coptic, Ethiopic, and Armenian) is an index of popularity. Nevertheless, few authors in Christian antiquity cite it to establish a point of doctrine, and when so used, often if not always in conjunction with other witnesses. Perhaps because it contradicted halacha on marriage, rabbinic Judaism chose not to transmit it. To be sure, Tobit attained canonical status in the historic Christian churches. Its popularity contributed to the achievement. But it is false to assume that since it was popular in a given period or place, it must have been canonical. Popularity and canonicity are two different things.
(4) Canon lists in Judaism and Christianity reflect the culmination of lengthy historical processes. They are attested relatively late in the game. They are important witnesses to the concept of canon, but are not as revealing as patterns of actual use of their contents. The oldest surviving Jewish list is the one found in the Bavli, Bava_Batra 14b-15a, a baraita attributed to the 2nd, but in its final form not necessarily earlier than the 4th cent. ce. Its components coincide exactly with those listed by the 4th cent Jerome. B. B. lists the Prophets and the Writings in an order rarely attested in later masoretic codices; the latter attest to a variety of orders. The order of the books in Christian traditions is different again, and subject to variation within the same tradition.
Handy overviews of historic Christian lists are found here and here. Note that the Ethiopic Orthodox canon, the Coptic Orthodox canon, and the Armenian Orthodox canon include books that, based on their attestation among the Dead Sea Scrolls, were understood to constitute authoritative revelation by one or more strands of Judaism before the Christian movement existed. The following books incontestably fall into this category: Enoch, Jubilees, and the Testaments of the Twelve Patriarchs.
It is probable that proponents of a variety of Jewish formations became followers of the Way and brought their diverse understandings of what constituted authoritative revelation with them. This is also the best explanation for the express quote of the book of Enoch in Jude 14-15. The author of Jude regarded the book of Enoch as authoritative revelation, as is the case to this day in the Ethiopian Orthodox Church. He explicitly quotes from it for that reason.
The Witness of the Pandects
(5) The great Bible manuscripts of antiquity also bear on the question. The Hebrew and Aramaic codices Aleppo (10th cent.) and Leningradensis (1009), the Greek Alexandrinus (5th cent.), the Syriac Ambrosianus (6th-7th cent.), and the Latin Amiatinus (8th cent.) come to mind. Except for the order in which the components of the second and third divisions of the Tanakh are presented, ancient Jewish Bibles are remarkable for their virtual textual identity. Ancient Christian Bibles, on the other hand, are remarkable for their inclusion of broad ranges of texts. None, furthermore, matches the other.
Alexandrinus contains the protocanonical writings, along with the usual additions to Jeremiah (Baruch and Epistle of Jeremiah, with Lamentation placed between them), Daniel (Song of the Three, Susanna, Bel and the Dragon), and Esther. Tobit and Judith are placed after Esther; then comes 1 Esdras (an edition of Ezra materials at variance with 2 Esdras [=MT Ezra-Nehemiah]) beside 2 Esdras, followed by 1-4 Maccabees. The Epistle to Marcellinus attributed to Athanasius, the Hypothesis of Eusebius (a table of contents of the Psalms), the Periochae of the Psalms (synopses of each Psalm), and the Canons of the Psalms (assignation of the Psalms to daily services) are prefaced to the Psalms. Psalm 151 also appears. The Odes, excerpted from surrounding books plus the Prayer of Manasseh, follow; then Wisdom of Solomon and Ben Sira. The New Testament concludes with 1 Clement (lacking 57:7-63); 2 Clement (up to 12:5a), Philemon, and Revelation. James, 1-2 Peter, 1-3 John, and Jude appear earlier, placed between Acts and Romans. Psalms of Solomon, no longer extant, comprised an appendix.
Ambrosianus contains, beside the usual additions to Jeremiah and Daniel, Wisdom of Solomon, Ben Sira, and Judith; but not the additions to Esther known in Greek and Latin tradition, nor Tobit. It also contains Ps 151, 2 Baruch, 4 Ezra, 1-4 Maccabees, and Josephus, Jewish Wars, Book 6! The New Testament lacks 2-3 John, 2 Peter, Jude, and Revelation.
Amiatinus contains, beside the additions to Jeremiah, Daniel, and Esther known from Greek tradition; Tobit, Judith; Wisdom of Solomon, Ben Sira; and 1-2 Maccabees, but not 3-4 Maccabees known from Greek and Syriac tradition. 3 Esdras [=1 Esdras above], 4 Esdras (=5 Ezra, 4 Ezra, and 6 Ezra, in that order) and Psalm 151 are found in an appendix. It also contains prologues by Eusebius and Jerome to various books. It lacks the Epistle to the Laodiceans, a work widely attested in early Vulgate manuscripts. This last, once extant in Greek, was dismissed as inauthentic by Theodore of Mopsuestia (c. 350-429), but was often considered to be one of Paul’s letters in the West (e.g., by Gregory the Great [c. 540-604], Moralia in Job 35:20).
Each of these Bibles represents a different understanding of the limits of authoritative literature. The inclusion of books like Joseph and Aseneth and the Testament of the Twelve Patriarchs in Armenian manuscripts of the Bible attests no less to the fact that said writings were deemed worthy of study in the context of a larger authoritative corpus. Whether or not Joseph and Aseneth and the Testament of the Twelve Patriarchs were read in worship or taught to catechumens in the Armenian church are separate questions.
A canon is an expression of a community of faith's confidence that God vouchsafed the truth in a set of writings. God is thought of as the one who superintended not only the production of the texts, but their conservation, and this not apart from, but in conjunction with, the larger work of divine providence within tradition and the life of the bearers of tradition. At the same time, principles reemphasized by the Reformation are also valid, to wit, that scripture is meant to stand in judgment of those who read it, not the other way around; that the text itself is inspired, quite apart from the interpretation tradition provides it, and so on. It’s a both/and proposition.
A Three-Way Distinction
Distinctions made in antiquity by Athanasius and Augustine might be reappropriated today by Orthodox, Catholic, Anglican, Lutheran, Reformed, and evangelical Christians alike. With regard to Jewish literature inherited by the church, Athanasius advocated a three-way distinction: “canonical” books (those to which one may appeal to establish a point of doctrine); “read” books (those that are read in the churches, but which are not suitable to appeal to in establishing doctrine); and “apocryphal” books (literature unsuited to be read in the churches and unreliable on questions of doctrine, but still of interest for other purposes) (Festal Letter, 39). Augustine also advocated a three-way distinction. His category of “canonical scriptures” includes the “canonical” and the “read” scriptures of Athanasius, but he distinguishes between canonical scriptures received by some and those received by all: “among the canonical scriptures [the skillful interpreter of the sacred writings] will judge according to the following standard: to prefer those that are received by all the catholic churches to those which some do not receive” (De Doct. Chr. 2, 8, 12).
Applied to today’s context, Augustine’s counsel would result in a differentiation of use among the books in question at variance with the one he advocated. If the combined counsel of Athanasius and Augustine were heeded today, the books called deuterocanonical in Catholic tradition and their equivalents in other historic churches might continue be read in churches which have always done so, but said books would not be accorded the same authority as the others in teaching and preaching.
If the “twenty-two books” Jerome regarded as canonical and the text form (proto-MT) that served as Jerome’s primary but not exclusive point of departure for his translation of them came to be valued as “The Scriptures Held in Common” by Jews and Christians of all persuasions, the common good would be served. As I point out elsewhere, a straight-up translation of MT remains a desideratum; the Jewish Publication Society versions do not qualify. Notes in the margin of a translation of MT might indicate the more significant departures from it attested in Jewish tradition elsewhere (e.g., in ancient Aramaic and Greek translations).
At the same time, a study edition of the entire range of texts treated as authoritative scripture in Judaisms and Christianities past and present is a desideratum. A study edition of the statutory prayer tradition across confessional lines and the synagogue/church divide would also be welcome.
Public reading and preaching of excerpts from extra-canonical books among, Lutheran, Reformed, Methodist, Baptist, and Pentecostal Christians seems unlikely. In those contexts, it is nonetheless easy to imagine using non-canonical texts consonant with the witness of texts within the canon in the form of unison prayers, responsive readings, or illustration on appropriate occasions. Examples with precedent on their side include those cited in Testimonia ad Quirinum (see Appendix B).
Within the bounds of the preaching canon as understood by rabbinic Judaism on the one hand and the churches of the Reformation on the other, it is possible to cite passages from non-canonical books as illustrations in a sermon based on a canonical text. It is also possible to include texts from outside the Bible in a prayerbook (a normal practice in traditions that make use of one). It is not possible to cite a non-canonical text in support of teaching unless it has independent support from a canonical text.
For my part, I will continue to teach and preach from a translation like ESV without the Apocrypha. Precisely those books Athanasius regarded as canonical, it might be remarked, are contained in such editions. (Proto-) MT, the form of the text that served as Jerome’s primary but not exclusive point of departure for his translation of the “twenty-two books,” furthermore, served the same purpose for the ESV translators. To be sure, I do not hesitate to depart from a standard translation in the text I presuppose and the nuances I highlight if my conscience so prods. My parishioners know I work from the Hebrew, Aramaic, and Greek. I fold in references to the wording of the original languages as the occasion permits.
I may never cite or expound on them for more than a sermon illustration, and I may never include more than a few lines from them as part of a unison prayer or responsive reading, but I still want both the Hebrew and Greek Ben Sira traditions in the next study Bible I purchase. On top of the other extra-canonical books printed in the superb NRSV study Bibles available today, add in Enoch, Jubilees, 2 Baruch, Psalms of Solomon, and Testament of the Twelve Patriarchs. The matrix from which rabbinic Judaism and early Christianity developed is documented by these writings in unique and illuminating ways. And if you, like me, hold a position close to Jerome’s, for whom “all the apocryphal books” contain “many faulty elements in them,” but still some “gold in the mud” (Epistle 107,12), I say to you as I do to myself: let the gold rush begin.
The pre-rabbinic and extra-rabbinic Jewish literature alluded to above is widely taught and studied today in rabbinical seminaries and in universities with a primarily Jewish student body in Israel, the United States, and beyond. Reappropriation proceeds apace. To re-accord it an authority on a level with what it once had in specific strands of pre-rabbinic or extra-rabbinic Judaism is not under discussion.
In a Christian context, among the writings inherited from Judaism, a strong case can be made for according the greatest authority to the books and text form Jerome referred to as the “Hebrew Truth.” These are the scriptures Jews and Christians of all persuasions hold in common. A case can also be made that neglect of the wider set of Jewish writings treasured in ancient Christian traditions impoverishes the intelligence of contemporary Christians.
The textual worlds of Judaism and Christianity are among the great cultural legacies of all time. Said worlds are shot through with a myriad tensions, but are united by a central conviction, namely, that God speaks, even now, through textually mediated past events. As we read, “It is not with our fathers that God sealed this covenant, but with us, the living, we who are here today” (Deut 5:3). In the repetition of these words, it is understood, God still speaks. The expected response is obedience to the one whose voice is heard.
Parts of the canon report God’s word. Other parts furnish models of obedience to that word. What incredible models they are: Job, Psalms, and Lamentations; the patriarchal narratives; the careers of Samuel, David, Elijah, and Elisha; Isaiah, Jeremiah, and Daniel; Ezra and Nehemiah; in the New Testament, the careers of Peter, Stephen, and Paul; the letters of Paul, James, and John. Even more incredible: examples of obedience furnished by those for whom the bat qol (the heavenly voice) was not constitutive to their reported experience: Qohelet, Esther, and Song of Songs.
The self-understanding of Judaism and Christianity stands or falls on belief in a God who speaks to “us, the living, we who are here today.” Apart from this belief, said faiths would never have fashioned canons in the first place.
Suggestions for Further Study
The notion of canon as a closed list of books, however defensible from a historical point of view, explains exactly nothing. What matters is the role texts play in life and discourse.
If a writing is inquired of – I use the mantic metaphor on purpose – in order to prove something, and more especially the truth of something, it is canonical. Interpretation becomes an act of divination.
The truth of something, furthermore, should not be defined narrowly. Truth has to do with what is true in all aspects of life. This is not the same thing as saying that scripture is a vademecum of history and astronomy, though of course it has been construed in that way, and some would so construe it today. It is saying that the contents of the canon describe the direction and inner dynamic of history, and the sun, moon, and stars as elements in God’s orderly creation.
But defining a text as canonical if it can be appealed to for the purpose of distinguishing truth from falsehood is also reductive. A collection of canonical writings does more than that. A canonical collection’s contents occupy the central place in the religious imagination of those who transmit it. It’s the place one goes to be still and know that the Lord is God. It’s the place one tastes the goodness of the Lord. The act of turning to a text and construing it as a lamp unto one’s feet is the central religious fact of both Judaism and Christianity. The more often a text is turned to in this sense, the more canonical it is.
The works listed below make points of their own.
Halbertal, Moshe. People of the Book: Canon, Meaning, and Authority. Cambridge: Harvard University Press, 1997.
Hughes, Aaron. “Presenting the Past: The Genre of Commentary in Theoretical Perspective.” Method & Theory in the Study of Religion 15 (2003) 148-168.
McEvenue, Sean. “The Old Testament, Scripture or Theology?” Interpretation 25 (1981) 229-242.
Neusner, Jacob. “Scripture and Tradition in Judaism. With Special Reference to the Mishnah.” Henoch 2 (1980) 285-305.
Neusner, Jacob. “Beyond Historicism, After Structuralism: Story as History in Ancient Judaism.” Henoch 3 (1981) 171-196.
Smith, Jonathan Z. “Sacred Persistence: Toward a Redescription of Canon,” in Approaches to Ancient Judaism: Theory and Practice 1. Ed. William Scott Green; Brown Judaic Studies 1; Atlanta: Scholars Press, 1978) 11-28. Reprinted in Imagining Religion: From Babylon to Jonestown (Chicago Studies in the History of Judaism; Chicago: University of Chicago Press, 1982) 36-52.
 E.g. the eschatological hopes of Josephus (Ant. 10.11.17), 4 Ezra (12:10-11), and 2 Baruch (39:1-8) depend on the four kingdom scheme. A satisfactory description of rabbinic eschatology has yet to be written, but it cannot be doubted that the same remark applies. See Robert P. Gordon, “The Targumists as Eschatologists,” Congress Volume, Göttingen (VTSup 29; Leiden: Brill, 1978) 131-48; and Uwe Glessmer, “Die ‘Vier-Reiche’ in der targumischen Literatur,” in The Book of Daniel: Composition and Reception (2 vols.; ed. John J. Collins and Peter W. Flint; VTSup 83; Leiden: Brill, 2001) 2:468-89. Glessmer (471, n. 5) cites M. Maher, Targum Pseudo-Jonathan: Genesis. Translated, with Introduction and Notes (The Aramaic Bible 1B; Edinburgh: T & T Clark, 1992) 60, n. 20: “The Midrashim frequently find a reference to the four kingdoms . . . in Gen 15:12; see Mekilta to Exod 20:18 (2, 268); Gen. R. 44:17; Exod. R. 51:7; Lev R. 13:5; PRE 28 (201).”
 William Horbury, “Old Testament Interpretation in the Writings of the Church Fathers,” in Mikra: Text, Translation, Reading and Interpretation of the Hebrew Bible in Ancient Judaism and Early Christianity (eds. Martin Jan Mulder and Harry Sysling; CRINT 2/1; Assen/Maastricht: Van Gorcum; Minneapolis: Fortress, 1990) 727-789; 763, citing William B. Tollinton, Clement of Alexandria: A Study in Christian Liberalism (2 vols.; London: Williams & Norgate, 1914) 2:198, n. 3.
 Carmelo Curti, “Catenae, Biblical,” in Encyclopedia of the Early Church (2 vols.; ed. Angelo di Berardino; tr. Adrian Walford; foreword and bibliographic amendments, W. H. C. Frend;New York: Oxford University Press, 1992 [1983-88]) 1:152-53; 152.
 Francis J. Thomson, “The Slavonic Translation of the Old Testament,” in The Interpretation of the Bible: The International Symposium in Slovenia  (ed. Jože Krašovec; JSOTSup 289; Sheffield: Sheffield Academic Press, 1998) 605-920.
 Devorah Dimant, “Use and Interpretation of Mikra in the Apocrypha and Pseudepigrapha,” in Mikra: Text, Translation, Reading and Interpretation of the Hebrew Bible in Ancient Judaism and Early Christianity (eds. Martin Jan Mulder and Harry Sysling; CRINT 2/1; Assen/Maastricht: Van Gorcum; Minneapolis: Fortress, 1990) 379-419; 385.
 Pandects or manuscripts containing all of the biblical books were rare. Portions of the Bible, e.g. the Pentateuch in Hebrew or the Gospels in Latin, were more common.
 A play on Ps 46:11. In its near context, Ps 46:11 is a call to warring nations to stop their strife. In its far context, all of scripture, it is a call to drop everything and listen.
 A play on Ps 34:9 (and 27:13). In its near context, Ps 34:9 is a joyful acknowledgment of Yhwh’s saving acts. In its far context, it describes what one experiences in reading scripture.