CHAPTER 1 INTRODUCTION
J. W. Wevers and the Göttingen Septuagint Septuagint
While Swete recognized the importance of G for the recovery of the
Hexaplaric text, subsequent scholarly activity has not been equal to the manuscript’s significance. G has been collated for the critical editions of the Göttingen Septuagint edited by J.W. Wevers. The signs were notated as necessary but have not been a
78 Swete, An Introduction to the Old Testament in Greek, 137.
79 Swete, An Introduction to the Old Testament in Greek, 138.
80 Swete, An Introduction to the Old Testament in Greek, 138.
focus of study. Additionally, the Göttingen editions seek to establish of the LXX Urtext, not to reconstruct the Hexapla, though these two goals do overlap to a degree.
Wevers’ Text History
J. W. Wevers completed a series on the text history of the Greek
Pentateuch in addition to the Göttingen editions of the Pentateuch.81 Here one may find some initial catalogs of the Aristarchian signs. These catalogs provide the most detailed information to date about the Aristarchian signs in G. However, a couple problems emerge when one takes a closer look at Wevers’ work. First, some inaccuracies exist in his work. These will be examined in the respective chapters and sections. These inaccuracies may be traced to Wevers’ reliance on the collation books rather than examination of the manuscript itself. Second, Wevers’ catalogs are incomplete. The obelus section does not claim exhaustiveness. Instead, it tracks the loss of obelized text throughout the textual tradition. Additionally, Wevers’
apparatus does not treat the instances of metobelus misuse. This oversight needs to be corrected since the data on the metobelus provides better information than what is currently available in the LXX introductions.
Conclusion
This overview demonstrates the value of G for the debates about Origen’s labors and intentions regarding the Hexapla. The overview also shows that G has not been given the attention needed in scholarly discourse. In the remainder of this
81 John William Wevers, Text History of the Greek Genesis (Göttingen: Vandenhoeck &
Ruprecht, 1974); Wevers, Text History of the Greek Exodus (Göttingen: Vandenhoeck & Ruprecht, 1992); Wevers, Text History of the Greek Leviticus (Göttingen: Vandenhoeck & Ruprecht, 1986);
Wevers, Text History of the Greek Numbers (Göttingen: Vandenhoeck & Ruprecht, 1982); Wevers, Text History of the Greek Deuteronomy (Göttingen: Vandenhoeck & Ruprecht, 1978).
dissertation, I seek to address this oversight.
CHAPTER 2
AN INTRODUCTION TO CODEX COLBERTINUS-SARRAVIANUS
Before discussing specific sign usage in G, it will be helpful to have a description of the manuscript on its own terms. This description will set the context for the subsequent discussion of the Aristarchian signs.
Provenance
Codex Colbertinus-Sarravianus (G) is an uncial Greek manuscript dated to around the fourth to fifth century AD.1 It originally contained the Greek translation of the Octateuch, although significant portions of the original codex are no longer extant. G preserves more of Origen’s Aristarchian signs than any other extant Greek manuscript. Although the date for the manuscript post-dates Origen (c. 185–c. 254) by about 50–150 years, this manuscript preserves his intended product from his Hexaplaric work.2
The era in which G was produced overlapped with Eusebius of Caesarea (c. 260–c. 340 AD) and could have derived from his work.3 Eusebius, and his mentor, Pamphilus, copied Origen’s work and disseminated it.4 At present, no other scribal
1 Natalio Fernandez Marcos, The Septuagint in Context: Introduction to the Greek Version of the Bible (Boston: Brill, 2000), 211
2 Francesca Schironi, “P.Grenf. 1.5, Origen, and the Scriptorium of Caesarea,” Bulletin of the American Society of Papyrologists 52 (2015): 213–14.
3 Henry Barclay Swete, An Introduction to the Old Testament in Greek, ed. Henry St. John Thackeray (Cambridge: Cambridge University Press, 2009), 76; Fernandez Marcos, The Septuagint in Context, 211.
4 Jerome, Praefatio in librum Paralipomenon, ed. J. P. Migne, Patrologia Latina 28 (Paris:
Apud Garnier Fratres, 1845)
school is known for reproducing and disseminating Origen’s work at so early a date.
Eusebius knew Origen’s work well.5 A manuscript like G corresponds to testimonies about the work of Eusebius and Pamphilus.6 The dates, the Origenic provenance, and the resulting text point to a Eusebian and Caesarean origin for G. However, the possibility of a Eusebian provenance for the manuscript must remain only a theory.
Unlike the Syro-Hexapla, G does not preserve any colophons providing the text history.7
G is the work of two original scribes, designated by Omont as librarius primus and librarius secundus.8 The first scribe produced Genesis—Leviticus and Deuteronomy. His work is characterized by closer attention to detail and less post- transcription editorial work. He also tended to include 27 lines per column of text.
The second scribe produced Numbers, Joshua—Judges. His work is characterized by less attention to detail and more post-transcription editorial work. His columns tended to include 28 lines. The characteristics of these scribes must be kept in mind when discussing the errors in the transmission of the Aristarchian signs.
The text of G was subject to editorial work. One is able to discern three strata of editorial work.9 The first stratum consists of corrections performed prima
5 Eusebius of Caesarea. Die Kirchengeschichte, Band 6 of Eusebius Werke, ed. Eduard Schwartz, Theodor Mommsen, and Friedhelm Winkelmann, Die griechischen christlichen Schriftsteller der ersten Jahrhunderte (Berlin: De Gruyter, 2010), VI.16.
6 Jerome, Praefatio in librum Paralipomenon.
7 For the importance of the colophons in understanding the history and intention of the text, see Peter J. Gentry, "The Aristarchian Signs in the Textual Tradition of LXX Ecclesiastes," in In the Footsteps of Sherlock Holmes: Studies in the Biblical Text in Honour of Anneli Aejmelaeus, ed. Kristin De Troyer, T. Michael Law, and Karketta Liljeström, Contributions to Exegesis and Theology 72 (Walpole, MA: Peeters, 2014).
8 Henry Omont, ed., Vetus Testamentus Graece Codicis Sarraviani-Colbertini quae supersunt in Bibliothecis Leidensi Pariesiensi Petropolitana phototypice edita (Leiden: A. W. Sijthoff, 1897), vi.
9 Swete notes three important editorial hands, summarizing the work of Tischendorf.
Swete, An Introduction to the Old Testament in Greek, 137. It is difficult to determine distinct editors in the manuscript due to space constraints in the text.
manus. Among these corrections are deletion dots placed above errant letters. These dots signify that the letters under them are erroneous and should not be read (e.g., Gen 34:30). The second stratum consists of corrections performed secunda manus.
These edits have a similar script to the original scribal hand. While editing the text is within the purview of this editorial strata, this layer specialized in correcting the errant Aristarchian signs. A third editorial stratum consists of a later Latin hand.
This editor used a miniscule script and notated the text in Latin (e.g., Gen 36:18).
This editor also notated the content of certain portions of the text, especially the different groupings of laws in Leviticus. The presence of editorial work, especially by a contemporary hand, comports with the testimony from the Syh colophons. Syh was translated from Greek manuscripts that were copied under the supervision of Pamphilus and Eusebius.10 These manuscripts were copied and then subsequently edited. G witnesses to this kind of editorial work. The books that demonstrate significant editorial work also have mistakes that the editors did not correct. It is possible that the incomplete nature of the editorial work provides a terminus ad quem for dating the manuscript. If Eusebius was the editor, his death could have meant the end of editorial activity and the publication of the manuscript.
G currently resides in three different libraries: Leiden, Paris, and St.
Petersburg. The library at Paris preserves the extant portions of Exodus, all of the extant portions of Leviticus except 24:9–27:16, Numbers 25:2–26:3 and 29:12–33 σύγκρισιν.11 The St. Petersburg library preserves one leaf which contains Judg 9:48–
10:6.12 The remaining leaves reside at the Leiden library.13 All of these leaves were
10 Gentry, "Did Origen Use the Aristarchian Signs in the Hexapla?," 137.
11 Swete, An Introduction to the Old Testament in Greek, 137.
12 Swete, An Introduction to the Old Testament in Greek, 137.
13 Swete, An Introduction to the Old Testament in Greek, 137.
photographed by Omont and are presented in one volume, which is the source of the present project.14 Prior to their residence in these libraries, the manuscript pages passed through the hands of several collectors. Those at Leiden are known to have been owned by Claude Sarràve (d. 1651), Jaques Mentel, and Isaac Voss (d. 1681), before being purchased from Voss’s heirs by the Leiden library.15 The Paris leaves once comprised a portion of Henri Memme’s (d. 1596) library.16 They found their way into the hands of J. B. Colbert in 1732, from whom they arrived at the Paris library.17 The St. Petersburg leaf passed through the hands of F. Pithaeus, Desmarez, Montfaucon, and Dubrowsky, before coming to rest at its present location.18 This leaf is in a poor state of preservation.
Greek Script of G
The Greek of G is written in an uncial script. The lunate sigma (Ϲ) is used throughout the manuscript rather than the majuscule Σ. The manuscript is written scripta continua. The Aristarchian signs, except when added or altered by
correctors, are incorporated within the text. The scribes show sensitivity to the literary units within the text since the script occasionally terminates before the normal end of a line. The manuscript has examples of indenting and hanging lines to signal the beginning of a new portion of text. The manuscript has special characters as well. Numbers are not spelled out. They are written using the Greek alphabet with an overline. The same phenomenon obtains for the nomina sacra, which are spelled out using enough letters to specify the case and are written with an overline.
14 Omont, Vetus Testamentus Graece Codicis Sarraviani-Colbertini.
15 Swete, An Introduction to the Old Testament in Greek, 137.
16 Swete, An Introduction to the Old Testament in Greek, 137.
17 Swete, An Introduction to the Old Testament in Greek, 137.
18 Swete, An Introduction to the Old Testament in Greek, 137.
The nomina sacra demonstrate Christian influence (e.g., ις for Ἰησοῦς, the Greek spelling of both Joshua and Jesus). They are used somewhat indiscriminately: they do not always specify a divine person, nor is their use necessarily confined to important figures in sacred history (e.g., the fathers Abraham, Isaac, and Jacob). In this the scribes exercised caution lest they inadvertently utilized the full spelling for a sacred name. In spite of this caution, they do miss some of the sacred names. The obelus tradition occasionally demonstrates the antiquity of the nomina sacra in the text as well as potential confusion resulting from their use. If a nomen sacrum was used for someone other than a divine being, one may find a clarifying statement (e.g., Gen 32:5 [6]).19 The nomina sacra and the alphabetic letters are noted in the present work with the overline.
The Greek text of G has examples of itacism and otacism throughout.
These phenomena are more prominent in Numbers, Joshua, and Judges. Itacism and otacism occur in the text proper and in the Aristarchian readings. Therefore, they were probably introduced by the scribes. If they were native either to the text or the Aristarchian readings, one would expect to find them in only one of those places.
Itacism is more common than otacism by far. Generally, itacism occurs when a diphthong is reduced to one vowel. The vowel may be one of the members of the diphthong. In addition to itacism and otacism, one may also find etacism (αι»ε).
Otacism in G involves confusion of omicron and omega.
Layout of G
G is preserved in a codex format. The left margin on the recto page and the right margin on the verso page are narrower than the facing margin due to the binding of the codex. The text is written on the recto and verso sides of each page.
19 In the sentence και απεστειλα αναγγειλαι τω κω μου ÷ ησαυ ↙, ησαυ clarifies κω.
The pages are approximately 10”x9”.20
The text is written in two columns on each side of each leaf. Each column contains 27 lines (scribe one) or 28 lines (scribe two).21 Each line has between 13–
18 letters. The range is adjusted depending on the scribe, the number and frequency of Aristarchian signs, and any indenting/hanging lines and early line termination. It is common for words to break over lines. This fact introduces some complications to the use and accuracy of the Aristarchian signs. These complications will be
addressed in due course.
The columns do not fill the entire page. The top margin leaves room for approximately five additional lines of text. The bottom margin leaves room for approximately eight lines of text. The area between the columns is approximately five letters wide. The wide margin (i.e., the non-binding margin) is approximately two-thirds the width of a column. The narrow margin is approximately three letters wide. These extra spaces do not contain prima manus commentary, although enough space exists to insert additional information.
G Within the Hexaplaric Tradition
G is an early witness to the Origenic tradition. However, it is not the earliest monolingual Hexaplaric manuscript scholars currently possess. Francesca Schironi has recently published findings from P.Grenf. 1.5.22 A comparison between these two manuscripts will prove salutary for later discussions of the Origenic tradition in G.
P.Grenf. is a late third–early fourth century AD papyrus manuscript.23 It
20 Swete, An Introduction to the Old Testament in Greek, 138.
21 Omont, Vetus Testamentus Graece Codicis Sarraviani-Colbertini, vi.
22 Schironi, “P.Grenf. 1.5.”
23 Schironi, “P.Grenf. 1.5,” 184.
contains a fragment of Ezek 5:12–6:3.24 The quality of the manuscript prevents a full accounting of the signs.25 However, two elements of this manuscript are worth noting. First is the layout. The margins are only wide enough for the signs. Since the margins are damaged, the signs are occasionally difficult to discern. The margins are not much wider than the space of a letter. This contrasts sharply with the wide margins of G. Even the binding margin of G is wider than the margin of P.Grenf.
Second, consider the signs. P.Grenf. does not have a sign for the metobelus. In the one place where the Aristarchian reading obviously ends, no sign exists to mark the termination of the reading.26 Codex Colbertinus-Sarrvianus, being produced approximately a century after P.Grenf., represents a development in the Origenic tradition by widening the margins and adding the metobelus to close readings.
The development of the metobelus will be treated in a later chapter. At present, it is necessary to discuss the development of the width of the margins between P.Grenf. and G. To do this, a comparison with the later Syro-Hexapla (Syh) manuscript of Arthur Vööbus is instructive. Vööbus published The Pentateuch in the Version of the Syro-Hexapla in 1976.27 This volume contains a facsimile of a Midyat manuscript, now known as Princeton University Scheide Library ms 150. Although not the only manuscript of Syh, it provides a helpful point of comparison with G. The folio leaves were 25.5cm x 16.5cm (approximately 10” x 6.5”).28 When the pages are open, the reader sees two columns. The non-binding margin is wide, as are the top
24 Schironi, “P.Grenf. 1.5,” 181.
25 Schironi, “P.Grenf. 1.5,” 208.
26 Schironi, “P.Grenf. 1.5,” 212.
27 Arthur Vööbus, The Pentateuch in the Version of the Syro-Hexapla: A Fac-simile Edition of a Midyat MS. Discovered in 1964, Corpus Scriptorum Christianorum Orientalum 45 (Louvain, Belgium: Imprimerie Orientaliste, 1975).
28 Vööbus, The Pentateuch in the Version of the Syro-Hexapla, 29.
and bottom margins on each page. The wide margin often contains notes, in some places resulting in overcrowding.29 The layout when the pages are open reminds one of the layout of G. G does not contain the marginal notes present in Syh. The edition of Vööbus is not a double-columned manuscript. However, it was copied onto pages that are narrower than G, so it is possible that the two columns from one page were written over two pages. The resulting layout is remarkably similar to G.
The Question of the Tetrapla
The foregoing discussion of the relationship between G and Syh raises the question of the Tetrapla. The nature and existence of the Tetrapla has generated significant debate among scholars.30 Some have come to doubt the existence of a separate Tetrapla.31 Nautin holds that the Tetrapla was the rough draft for the Hexapla.32 In this view, the Tetrapla was still a columned Bible. Fernandez Marcos states that the Tetrapla was the abbreviated Hexapla.33 Mercati argues that the Tetrapla consisted of the LXX and the Three and served to help Christians study the Old Testament.34
All of the aforementioned views have a columned Bible like the Hexapla in view. They all depend on one passage in Eusebius.35 Eusebius writes, “…in his own
29 Vööbus, The Pentateuch in the Version of the Syro-Hexapla, 37.
30 Fernandez Marcos, The Septuagint in Context, 207.
31 For instance, Dominique Barthelemy. Dominique Barthelemy, “Origène et le texte de l’ancient testament,” in Epektasis: Mélanges patristiques offerts au Cardinal Jean Daniélou (Paris:
Beauchesne, 1972), 247–261.
32 Pierre Nautin, Origène: sa vie et son œuvre, Christianisme antique 1 (Paris: Beauchesne, 1977), 342–343.
33 Fernandez Marcos, The Septuagint in Context, 207.
34 Giovanni Mercati, “Il Problema della Colonna II dell’Esaplo,” Biblica 28, no. 2 (1947):
212ff.
35 Eusebius, Die Kirchengeschichte, VI.6.
way having arranged the recension of Aquila, Symmachus, and Theodotion together with the recension of the Seventy in the Tetrapla (ἰδίως τὴν Ἁκύλου καὶ Συμμάχου καὶ Θεοδοτίωνος ἔκδοσιν ἅμα τῇ τῶν ἑβδομήκοντα ἐν τοῖς Τετρασσοῖς
ἐπισκευάσας).” These words, however, do not necessitate a columned Bible like that of the Hexapla. They merely require an “arranging together.” It is possible that Eusebius could be describing a version in which the LXX is the main text,
accompanied by the readings of Aquila, Symmachus, and Theodotion. If this is the case, one would expect to find a text with the LXX surrounded by sufficient space in order to include the readings from the three. In other words, if one was looking for the Tetrapla, one would expect to find a manuscript like G, but with the readings from the Three.
This returns the discussion to the comparison between G and Syh. Gentry has recently described the information the colophons of Syh provide.36 The
colophons tell the reader that Syh was translated from copies produced and corrected by Eusebius and Pamphilus.37 Syh was not translated directly from the fifth column of the Hexapla.38 The textual evidence of Syh confirms the data from the colophons: Syh is a distant member of the O´ group despite preserving an abundance of Aristarchian signs.39 Syh also preserves the readings from the Three.40 Since Syh preserves the Aristarchian signs, the readings from the Three, and a text that is only a distant witness to the fifth column, it is reasonable to conclude that the colophons
36 Peter J. Gentry, "Did Origen Use the Aristarchian Signs in the Hexapla?," in XV Congress of the International Organization of Septuagint and Cognate Studies(Atlanta: SBL, 2016), 133–148;
Gentry, "The Aristarchian Signs in the Textual Tradition of LXX Ecclesiastes.”
37 Gentry, "The Aristarchian Signs in the Textual Tradition of LXX Ecclesiastes," 466.
38 Gentry, "The Aristarchian Signs in the Textual Tradition of LXX Ecclesiastes," 466.
39 Gentry, "The Aristarchian Signs in the Textual Tradition of LXX Ecclesiastes," 466.
40 Gentry, "The Aristarchian Signs in the Textual Tradition of LXX Ecclesiastes," 466.
are correct: Syh was translated from a recension. If Syh was translated from a
recension, the recension would already have contained the readings from the Three.
In other words, this recension would have arranged together the LXX, Aquila, Symmachus, and Theodotion, agreeing with the words of Eusebius regarding the Tetrapla. Gentry is correct to conclude that, in some books, Syh was translated from the Tetrapla, not the Hexapla.41
Although G does not contain the readings of the Three in the margins, it is reasonable to conclude that G was a stage towards the Tetrapla. If Syh was
translated from the Tetrapla, the Tetrapla would have had to preserve both the Greek text and the readings from the Three. Such a version would have to have space to record all of the readings germane to the text. A manuscript such as P.Grenf.
would not suffice: it did not have the necessary marginal space to record anything more than an asterisk or obelus. However, a text like G has the marginal space to contain readings from the Three. The margins of G and the margins of the
manuscript from Vööbus bear strong resemblance to each other. G is a development within the Origenic tradition from earlier manuscripts and demonstrates that a version such as the Tetrapla was possible.
Conclusion
G is an uncial Greek codex from the fourth/fifth century AD It is the work of two scribes who showed varying degrees of care in their work. The manuscript shows evidence of editorial work, both near the time of transcription and by later hands. The text on each page is set in two columns with wide margins. It preserves a large quantity of Origen’s Aristarchian signs.
G represents a development in the Origenic tradition from P.Grenf. It has
41 Gentry, "Did Origen Use the Aristarchian Signs?," 146.