The Scribes and Correctors of Codex Vaticanus: A Study on the Codicology, Paleography, and Text of B(03) Jesse R. Grenz St Edmund’s College, University of Cambridge October 2021 This thesis is submitted to the Faculty of Divinity of the University of Cambridge for the degree of Doctor of Philosophy ii Declaration This thesis is the result of my own work and includes nothing which is the outcome of work done in collaboration except as declared in the preface and specified in the text. It is not substantially the same as any work that has already been submitted before for any degree or other qualification except as declared in the preface and specified in the text. The Faculty of Divinity Degree Committee have extended my word-limit to 100,000 words. This thesis does not exceed the limit. iii For Iane Marcela, meu amor and my parents, Jonathan and Susan iv v Abstract In this thesis, I investigate the production of Codex Vaticanus (B[03]) through the lens of its scribes and earliest correctors. While this manuscript, which contains the Greek Old and New Testaments, is recognized as one of the most important witnesses for Septuagint scholars and New Testament textual critics, there has yet to be any thorough examination of the scribes and their copying patterns. In other manuscripts, such as Codex Sinaiticus, it has been shown that knowledge of the scribal habits is necessary for understanding the overall textual value of the witness. This thesis, therefore, assesses the material, paratextual, and paleographic evidence concerning the number of scribes responsible for copying B(03), followed by an evaluation and comparison of their work. Part I of the study examines the physical structure of B(03) and then its formatting and paratexts. When these features are combined, the identification of overlapping irregularities can be used to flag potential loci of scribal transition. Part II follows a new division of scribes in B(03) and examines the types and methods of correction throughout the codex. While there are several ways to compare the quality of copying in a manuscript, this project concentrates on the types of errors that provoked corrections by either the scribes themselves or their colleagues. In doing so, the investigator learns more about the individual scribes beyond the first layer of writing, as the work of a scribe often involved correcting one’s own copying and, occasionally, that of another. By comparing the corrections in the work of each scribe, this project ends with an evaluation of the one scribe responsible for copying the New Testament. The conclusions emphasize the importance of understanding this manuscript as a whole pandect with both surprising consistency and telling irregularities. vi vii Table of Contents Acknowledgments Abbreviations ix xi 1. Introduction Date Provenance Contents Vaticanus and Sinaiticus Dictation and the Exemplar(s) The Scribes of B(03) Plan of this Study 1 2 6 10 11 14 17 19 PART I: CODICOLOGY, PARATEXTS, AND PALEOGRAPHY 2. Codicology: The Material Make-Up of B(03) The Study of a Multiple-Text Codex Material Blueprint of the Codex Summary 25 26 29 34 48 3. Paratextual Features and the Problem of Paleography Paratextual Features The Problem of Paleography Conclusion 51 51 82 106 PART II: THE EARLIEST CORRECTIONS 4. The Scribes and Early Correctors of the Codex Previous Research on the Correctors A Survey and Typology of the Earliest Corrections Summary 111 111 128 143 5. Early Corrections in 1 Kingdoms and Psalms 1 Kingdoms Psalms Conclusion 145 145 173 188 viii 6. Early Corrections in the New Testament The Gospels The Acts of the Apostles The Catholic Epistles The Corpus Paulinum Conclusion 191 192 221 233 241 256 7. Conclusions The Codex Paleography and Paratexts The Scribes and Early Correctors Limitations and Future Avenues of Research 259 259 260 261 263 Appendices 265 Bibliography 287 ix Acknowledgments Space does not permit the mention of all who have supported me in researching and writing this thesis. First, I want to express my sincerest gratitude to Dr. Dirk Jongkind, who acted as both my masters and doctoral supervisor. Through the years, he has offered his patience, careful attention, and kindness to me and my project. Likewise, I owe thanks to E. Randolph Richards, who inspired me as an undergraduate student to pursue the study of textual criticism. I am indebted to numerous conversations with Patrick Andrist and his expertise on the codicology and paleography of Codex Vaticanus. Furthermore, I thank Timothy Sailors for inviting me to present at the special session on Vaticanus, during the Annual Meeting of the International Society of Biblical Literature, in Rome (4 July 2019). Here, I received valuable feedback from Fr. Stephen Pisano, shortly before his passing, and Claudia Montuschi, the director of the Manuscript Department of the Biblioteca Apostolica Vaticana. Without friendship, this research may have remained incomplete, for “whatever is enjoyable is turned into weariness when friends are absent” (Aquinas, De regno 1.10.77). In particular, I want to thank Austin Stevenson, alongside Katherine Stevenson and my goddaughter, Sophia. Their kindness, joy, and encouragement are boundless. I have also learned much from conversations with Daniel Stevens, Elijah Hixson, Zachary Cole, and Jonathan Platter. The Tyndale House community is an irreplaceable source of dialogue and friendship, including, but not limited to, Andrew Keenen, Caleb Howard, Diego Dy Carlos, Pete Myers, Michael Dormandy, and Kaspars Ozoliņš. Most importantly, I thank my family, who are constant in their love, encouragement, and generosity. My sister, Phoebe, inspires me with thoughtfulness and compassion. I am grateful to meu pai e minha mãe, Junior and Suerda, for loving me as their own. Finally, this thesis is dedicated to my wife, Iane, and parents, Jonathan and Susan. My father is an example of humility and dedication unlike any other, and my mother instilled in me the determination to ask hard questions. Amor, this project is the fruit of your selflessness and care for me through the years. May it be considered a success if I have shown you even a fraction of the generosity in return. θϲ̅ ̅βοηθοϲ του εργου τουτου ☧ x xi Abbreviations B–M The Old Testament in Greek: According to the Text of Codex Vaticanus, Supplemented from Other Uncial Manuscripts, with a Critical Apparatus Containing the Variants of the Chief Ancient Authorities for the Text of the Septuagint, eds. Alan England Brooke and Norman McLean ECM Editio Critica Maior EDG Etymological Dictionary of Greek LBG Lexikon zur byzantinischen Gräzität LDAB Leuven Database of Ancient Books, http://www.trismegistos.org/ldab. NA28 Novum Testamentum Graece, 28th rev. ed., eds. Barbara Aland, Kurt Aland, Eberhard Nestle, Erwin Nestle, Johannes Karavidopoulos, Carlo M. Martini, and Bruce M. Metzger. NTVMR New Testament Virtual Manuscript Room, http://ntvmr.uni-muenster.de. SBLGNT The Greek New Testament SBL Edition, ed. Michael W. Holmes THGNT The Greek New Testament, Produced at Tyndale House, Cambridge, ed. Dirk Jongkind. TuT Text und Textwert der griechischen Handschriften des Neuen Testasments All other abbreviations of primary sources and secondary literature, follow Billie Jean Collins et al., eds., The SBL Handbook of Style, 2nd ed. (Atlanta: SBL, 2014). Unless otherwise stated, all translations of ancient and modern languages are my own. In this thesis, images of Codex Vaticanus are taken from the DigiVatLib (DVL), which are freely accessible and downloadable: https://digi.vatlib.it/view/MSS_Vat.gr.1209. xii CHAPTER 1 Introduction “Once again, it is my strongest wish that a young researcher would produce, on Codex B, a monograph comparable to that of Milne and Skeat on Sinaiticus.” – Paul Canart1 “Here again that blessed manuscript from the Papal library is brought forward against us.” – Erasmus of Rotterdam2 Whether or not one agrees with the animosity of Erasmus’s statement, it is certainly true that readers of the Greek Old and New Testaments are regularly confronted by the great Codex Vaticanus. Indeed, contemporary readers, who are sympathetic to Erasmus’ plight, might now prefer the designation, “blessed manuscript of Westcott and Hort, the Nestle-Aland, or Rahlfs’ Septuaginta.”3 For, on most pages of these critical editions, the siglum “B” stands in defense of the printed Greek text. And yet, while there is a general recognition of the codex, even among non- specialists, it has an enigmatic past filled with many thwarted attempts by scholars to study it. Codex Vaticanus graecus 1209, 4 currently residing in the Biblioteca Apostolica Vaticana (BAV), is one of the two earliest Greek pandects of the Christian Bible. It was Wettstein, who first used the letter “B” as a designation for the codex in his 1751–1752 edition of the Greek NT.5 Until recently, this was the preferred siglum for editions of the Septuagint and the Greek NT. A notable exception to this was von Soden, who preferred the designation “δ1.”6 However, it has now become standard to refer to the NT portion of Vaticanus by its Gregory-Aland number, 03. 7 To these designations, we can add the Trismegistos number 62316, LDAB number 3479, and the Pinakes 1 Canart, “Notice,” 43 n. 67. 2 Erasmus, Annotations on Luke 23:46; Translation from Krans, “Erasmus,” 462. 3 Hoskier preferred the charge “Hortian heresy” for those who relied on “the crooked path pursued by the MS B.” Hoskier, Codex B, 1:i. 4 This inventory number had already been associated with the codex since at least the library index of 1612. Carlo Martini and Paul Canart claim the classification originated around 1600, but do not cite any of the inventories. See the published notes of Giovanni Mercati in Giacomo Cardinali, “Vicende Vaticane,” 401–424 (414); Canart and Martini, Introduction, 7. 5 Wettstein, Ἡ Καινὴ Διαθήκη, 1:23; cf. Metzger and Ehrman, The Text, 161; Castelli, Wettstein, 7. 6 von Soden, Die Schriften, 1:102. 7 See, for example, the published volumes of the Editio Critica Maior. CHAPTER 1 2 diktyon number 67840.8 As the scope of this project deals with both the Greek OT and NT in the codex, we adopt the combined siglum “B(03).”9 Although B(03) generally retains its prestigious status in Septuagint and NT scholarship, there is a surprising dearth of studies solely devoted to the codex, especially as it relates to its material and scribal context. In particular, the brief appendix on B(03) in Milne and Skeat’s Scribes and Correctors of the Codex Sinaiticus (1938) has generated a scholarly consensus on the scribes of B(03) that has yet to be examined in any detail.10 Nonetheless, this chapter’s opening quotation from Paul Canart, former vice prefect of the BAV, illuminates the need for a detailed study concerning the physical structure of the codex, along with its paratextual and paleographic features. Drawing on these observations not only clarifies questions around the production of B(03), but also allows us to better understand the different scribes in the codex as well as the quality of their copying and subsequent editing throughout both the Greek OT and NT. Before undergoing such an investigation, however, it is pertinent to summarize several historical and introductory features of B(03). Inquiry into the date, provenance, contents, dictation, and exemplars of the codex is integral to understanding the context and scope of its production. We cannot hope to understand the scribes and correctors of B(03) without first approaching these matters. As the entire study will bring further clarity and, in some cases, further caution to such issues, much of the following discussion will be left open and revisited in subsequent chapters. 1.1 Date The antiquity of B(03) is certainly its most recognized quality throughout the history of research. Already on 18 June, 1521, Paolo Bombace wrote to Erasmus that he found the text of 1 John in the Vatican library, “written in very ancient characters.”11 Taking Bombace at his word, Erasmus too cited B(03), favoring his rejection of the Comma Johanneum (1 John 5:7), as “a very ancient manuscript.”12 Yet, an approximate date of the codex was not published until the Sixtine edition of 8 https://www.trismegistos.org/text/62316; https://pinakes.irht.cnrs.fr/notices/cote/67840/. 9 This combined siglum with GA number in round brackets was already employed by Léon Vaganay in 1933. Christian-Bernard Amphoux made this more explicit in his updating of Vaganay’s work, but rendered the new siglum with a full-stop rather than brackets. Vaganay, Initiation, 23–24, 106; Vaganay and Amphoux, An Introduction, 14. 10 Milne and Skeat, Scribes, 87–90. 11 Epistle 1213, translated in Erasmus, Correspondence, 248 ll. 74–75; Krans, “Erasmus,” 451, cites this as “the very moment in history that Codex Vaticanus is first brought up in New Testament text-critical matters.” 12 Apologia resp. Iac. Lop. Stunica, translated in Krans, “Erasmus,” 452; cf. Annotations on 1 John 5. INTRODUCTION 3 the Septuagint (1587). In the Praefatio ad lectorem, B(03) is described as having “large letters” (maioribus litteris) and specifically dated 1200 years before the edition’s publication (ante millesimum ducentesimum annum) and before the time of Jerome (ante tempora B. Hieronymi). Patricia Easterling has highlighted the surprising precision with which the editors assigned the fourth-century date to the codex, since the modern science of paleography (associated with Montfaucon) had not yet been developed.13 While this approximate date would become the consensus in current scholarship, it was far from stable in the opinions of early critics. Even before the publication of the Sixtine edition, Cardinal Sirleto claimed in his notes that B(03) originated in the ninth century.14 Sirleto played an important role in examining the Greek manuscripts for the edition, but his opinion on the date was not followed by the editors. 15 Likewise, in the seventeenth century, Denis Amelote (1687–1688) criticized Erasmus’ claim that the Comma was absent in the oldest manuscript in the Vatican (i.e., B[03]), since he had personally seen it in the oldest Greek manuscript.16 Richard Simon (1689) was happy to accept the fourth-century date of the codex, but Bartolocci (2 November, 1669) allowed for some uncertainty, claiming instead that the codex was written more than a millennium before (piu di 1000 anni che e scritto).17 To be sure, there were many around this time that believed the codex originated in the third or even second century.18 However, this optimism shifted in later years as critics in the eighteenth century regularly suggested a date between the fifth and seventh centuries. Most notably, Bernard de Montfaucon (1739) proposed a fifth or sixth-century date, based partly on the absence of original accents.19 In one of the earliest sample pseudo-facsimiles of B(03), Giuseppe Bianchini (1749) includes the description, “scriptus videtur ineunte Saeculo V. Iesu Christi.”20 Writing in 1699, Louis Ellies du Pin 13 Easterling, “Before Palaeography,” 182. 14 Sirleto’s Annotations are present in Vat. lat. 6134, which has not yet been digitized. However, see the examination in Höpfl, Sirlets, 39 n. 2; cf. Pisano, “L’histoire,” 111. 15 Mandelbrote, “Manuscripts Meet,” 259. 16 Assuming he did not misread B(03), he is clearly referring to a different manuscript he believed was older. Amelote, Nouveau Testament, 2:104; cited in McDonald, Biblical Criticism, 149. 17 Bartolocci, “Notes.” 18 See Cardinali, “Vicende Vaticane,” 390 nn. 252–254. 19 Montfaucon, Bibliotheca, 3. 20 Bianchini, Evangeliarium, cdxciii (Tabula I); cf. Michaelis, Introduction, 2:345. CHAPTER 1 4 claimed that B(03) was older than a thousand years, apparently since it is missing section numbers or titles that conform to the Eusebian apparatus (see Chapter 2).21 Nonetheless, there remained some who, like Jacques Le Long (1709), regarded the codex as “not truly ancient, nor of good esteem.”22 What was missing from these attempts to date B(03) was any clear criteria. Instead, critics made general remarks concerning the large majuscule letters, followed by the absence of early accents and Eusebian section numbers. From the beginning, comparison with the Vulgate and patristic citations provided a significant anchor for situating B(03) in its historical context. Christoph Matthaeus Pfaff (1709) suggested one of the first explicit paleographic comparisons with the third-century inscription on a statue of Hippolytus.23 A century later, J. Leonhard Hug (1810) provided the most significant early treatment of the codex’s antiquity.24 Among other evidence, Hug compared the hand of B(03) with a newly unrolled Herculaneum papyrus of Philodemus’ De musica from the first century BC (LDAB 3653). He adds to this, the absence of ornamentation, original accents, Eusebian or Euthalian divisions, the phrase εν εφεϲω (Eph 1:1), and the presence of irregular section numbers in the Pauline corpus (see Chapter 2).25 From this, Hug concludes that B(03) belongs “to the earliest period of the fourth century.”26 After Tischendorf’s discovery of Codex Sinaiticus (א[01]) in the 1840s, attention shifted to the comparative dating of the two codices (see §1.4). Since he had also assigned א(01) to the fourth century, the question shifted to which came first.27 However, in 1967 Guglielmo Cavallo provided the most recent and sustained argument for a date range of 328–373, with a preference for a date 21 du Pin, Dissertation, 1:258–259. 22 Le Long, Bibliotheca sacra, 339; cited in Bianchini, Evangeliarium, cdxcii. 23 The reference is almost certainly to the seated statue of a figure associated with St. Hippolytus, held in the Vatican Library (see E05385 in the Cult of Saints in Late Antiquity database). Pfaff, Dissertatio, 55–57; Cf. Hichtel, Exercitatio, 8–9; translated in Michaelis, Introduction, 2:344; http://csla.history.ox.ac.uk/record.php?recid=E05385. 24 Hug, De antiquitate. 25 Hug, De antiquitate; cf. Hug, Introduction, 1:262–267. 26 Additional arguments for the early date of B(03) include the brevity of titles (e.g., κατα µαθθαιον) and the thinness of the parchment. Granville Penn also argued for an early date based on the proper placement of the Altar of Incense in Hebrews 9:1–5 (cf. Exod 30:1–10). Hug, Introduction, 1:266; Penn, Annotations, 32; Taylor, Emphatic New Testament, 50–51; MacMillan, Roman Mosaics, 366. 27 Tischendorf, Sinaiticum, xxix–xxxiii; Tischendorf, Vaticanum, xxviiii–xxxi; Tischendorf, Appendix codicum, xi–xii; cf. Abbot, “Antiquity,” 189–200. INTRODUCTION 5 around 350.28 His contribution was the establishment of an evolutionary model of the Greek Biblical Majuscule, the bookhand of B(03). In this model, our codex represents the pinnacle of the canonical bookhand, which coincides with the fourth century.29 Cavallo also presents the early fourth-century papyri, P. Lond. Lit. 33 and P. Beatty IV (LDAB 1259 and 3160), as slightly earlier comparisons to the hand of B(03). Interestingly, the date range of 328–373 is based on the episcopacy of Athanasius and the apparent dependence of the order of books compared to those listed in his Festal Letter of 367 (see §1.3).30 Cavallo’s methodology has not gone uncriticized, but a fourth-century date of B(03) remains the consensus.31 In a forthcoming article, Brent Nongbri has criticized the earliest dating of א(01) based on the cursive ανω and κατω notes in some of the corrections. While Milne and Skeat confidently dated these to the fourth century—probably the first half—Nongbri has also found parallels in the early fifth century.32 Since similar notes can be found in B(03) (see §1.4 and Chapter 4), this may call for further caution against restricting the date of the codex to less than a century. Cavallo’s preference for a date circa 350 is likely too precise, and the use of Athanasius’ episcopacy as a date range is unconvincing. In Chapters 5 and 6, we will see that the high proportion of corrections of the orthographic interchange ει-ι, fits well with the fourth-century documentary papyri. 33 In summation, B(03) is unlikely to predate Emperor Constantine, though many have argued it was ordered by him (see §1.2); nor is it likely that a codex as significant as B(03) could evade the addition of Eusebian section numbers if it had been produced far into the fifth century. Nevertheless, I will continue to refer to the fourth century throughout this study as a shorthand for the age of production. 28 Cavallo, Ricerche, 52–56. 29 Cavallo and Maehler, Greek Bookhands, 34. 30 Cavallo, Ricerche, 55. 31 See the summary of early criticisms by José O’Callaghan, Peter J. Parsons, Jean Irigoin, and Nigel G. Wilson in Orsini, Studies, 57–59; More recent criticism has come from Askeland, “Dating,” 457–489; Nongbri, “Palaeographic Analysis,” 84–97. 32 Milne and Skeat, Scribes, 62; Nongbri, “The Date,” (forthcoming); cf. Cole, “The Date,” (forthcoming). 33 See Stolk, “Itacism,” 690–697. CHAPTER 1 6 1.2 Provenance As with the date of B(03), numerous places of origin have been proposed for the codex (Table 1). The earliest attempts to locate the production of B(03) centered on southern Italy, Rome, or “the west.” This, however, was largely dependent on the theory of “Latinization”—that the text of B(03) was brought into conformity with the Vulgate (see Chapter 4).34 Already, as readings from the codex were brought against the edition of Erasmus, he criticized the manuscript of being corrected toward the Latin manuscripts.35 Critics like John Mill followed this reasoning to argue that a Latin scribe (a Latino scriba) was responsible for copying B(03).36 The theory of “Latinization” developed from the fragmentary nature of the readings known to Erasmus, Mill, and Wettstein, but Andreas Birch had rejected it based on extensive collations.37 Likewise, Hug’s study on the antiquity of B(03) pushed the date of the codex to a time before Jerome’s Vulgate.38 It is noteworthy, therefore, that Hort, having demonstrated the superiority of B(03), also believed in the Roman provenance. The arguments for such a conclusion include the apparently Western orthography in words like ιϲακ or ιϲτραηλ(ειτηϲ), the word-order χριϲτοϲ ιηϲουϲ in Paul, and the shared numerical divisions in B(03) with Codex Amiatinus and other Vulgate manuscripts.39 Amphoux attempts to account for the clear similarities between B(03) and the Vulgate, while also acknowledging a connection to Athanasius of Alexandria, by placing the production around 340, shortly after Athanasius fled to Rome.40 Finally, although Caspar René Gregory seems to prefer Caesarea as the place of origin, he mentions in passing that the parchment appears to be western, but leaves this unsubstantiated (see Chapter 2).41 34 See also Amphoux, “Les circonstances,” 162–164. 35 For example, Annotations on Luke 10:1; cf. Krans, “Erasmus,” 463–469. 36 Mill, Novum Testamentum, 163. 37 The problem was exaggerated since the readings sent to Erasmus were often selected to show B(03)’s agreement with the Vulgate against his edition. Likewise, Wettstein had apparently been refused access to readings from Richard Bentley, which he had hoped would invalidate the codex altogether. Wettstein, Novum Testamentum, 1:24; Birch, Quatuor Evangelia, xxiii; Michaelis, Introduction, 346–348; Pisano, “L’histoire,” 109. 38 See also Amphoux, “Les circonstances,” 163–164. 39 See, however, Giurisato’s more recent comparison of both early and late numeration in B(03) with that of Amiatinus. Westcott and Hort, Introduction, 264–267; Giurisato, “Atti degli Apostoli,” 211–227. 40 Amphoux, “Les circonstances,” 157–176. 41 Gregory, Canon and Text, 345. INTRODUCTION 7 TABLE 1: THE PROVENANCE OF B(03)42 Italy or “the West” Egypt (Alexandria) Caesarea Mill (1707), Simon (1708), Wettstein? (1751), Ceriani (1864), Hort (1881), Gregory (1907)?, Amphoux (2009) Hug (1810), Tregelles (1856), Vercellone (1860), Rahlfs (1899), Traube (1907), Lake (1918), Martini (1966), Cavallo (1967), Birdsall (1970), Metzger (1991), Bogaert (1999, 2009), Andrist (2009) Aejmelaeus (2020) Harris (1893), Robinson (1895), Milne and Skeat (1938), Lake (1908), Gregory (1907)?, Zuntz (1995), Elliott (2004), Grafton and Williams (2006), Dormandy (2020)? The youngest theory concerning the provenance of B(03) is that of Caesarean origin. It seems that the first association of B(03) with Caesarea was made by J. Rendel Harris in a paper from 1884 and substantiated in the appendix of his Stichometry (1893).43 Kirsopp Lake defended this position in 1908, but shifted his opinion with the publication of his facsimile of א(01) in 1911, and subsequently claimed, “the case for their [א(01) and B(03)] origin in Egypt rather than Caesarea is too strong to be put to one side.” 44 T. C. Skeat has provided the most thorough argument for Caesarean provenance of א(01) and B(03).45 It is essential to the argument that the two codices were copied in the same scriptorium and likely share the same scribe for portions of the text (see §1.4). 46 Consequently, although the arguments for a Caesarean origin are almost exclusively in reference to א(01), they are assumed to apply equally to B(03).47 Michael Dormandy has argued strongly that א(01) was one of the fifty Bibles ordered by Constantine (Vit. Const. 4.36), and is inclined to think that the similarities with B(03) outweigh their differences.48 However, as we will see below and throughout this study, the differences between the two codices have not been fully appreciated.49 42 Question marks indicate those who are undecided or have proposed more than one possible location. 43 Harris cites the earlier claim by Ceriani that א(01) was copied in Caesarea, but B(03) originated in Magna Graecia (southern Italy). Since Harris believes the two were copied in the same location, he advances the claim that B(03) was also from Caesarea. Harris, Stichometry, 71–89. 44 Lake, The Text, 14–15; Lake and Lake, Codex Sinaiticus, x–xv; Lake, “Manuscripts,” 34. 45 Milne and Skeat, Scribes, 66–69; Skeat, “Sinaiticus,” 193–237. 46 As Elliott puts it plainly, “the important point of all this is that whatever we say about the provenance of Codex Sinaiticus must also apply to Codex Vaticanus and vice versa.” Elliott, “T.C. Skeat,” 72; cf. 70. 47 Elliott helpfully summarizes the arguments in favor of Caesarea, of which only one applies directly to B(03)—the section numbers in Acts share some similarities with the Euthalian sections, linked to Caesarea. Elliott, “T.C. Skeat,” 74. 48 On whether the passage is referring to complete pandects, see Dormandy, “Pandects,” 21–36. 49 It should also be noted that Dirk Jongkind has leveled a number of criticisms against Milne and Skeat’s arguments for the Caesarean origin of א(01). Jongkind, Scribal Habits, 253–254. CHAPTER 1 8 Furthermore, the absence of Eusebian sections and the difference in contents and order of books with Eusebius are strong arguments against Caesarean origin of B(03).50 Skeat has suggested their absence may have been influenced by the pressure to produce quickly for Constantine.51 In the following study, it will become clear that the complex layers of production in B(03) do not justify the notion of a rushed procedure. That there was time to add numerous types of early marginalia in the codex (Chapter 3), and at least two early layers of correction ensure that production was not too rushed for Eusebius’ apparatus to be added. This leaves the final and most persistent provenance theory—Egypt (Alexandria). 52 Certainly, there have been many unconvincing arguments leveraged in favor of Egyptian provenance. The presence of the so-called “Coptic mu” ( ) and the omega, shaped like an anchor, were often cited as evidence, but are now said to be common forms throughout the Greco-Roman world.53 Hug confidently associated the origin of B(03) with “an Egyptian Calligraphist” on account of the orthography, which had similarities with Greco-Coptic texts.54 This orthographic argument is inconclusive as it is based on limited papyrological evidence, which has only survived in Egypt.55 Dormandy is correct to note the fading significance of geography in scholarly opinions concerning manuscript relations and versional evidence.56 However, if there is any historical value in these relationships, we could point to the strong connection between B(03) and P75, the latter being almost certainly from Egypt.57 Anneli Aejmelaeus has recently argued that Origen must have used a text like B(03), which was the traditional text of Egypt from the beginning of the third century. 50 Knust and Wasserman highlight two geographical variants from Eusebius’ text of John, namely βηθανια for βηθαβαρα in Eusebius (John 1:28) and βηθϲαιδα for βηθζαθα (John 5:2). Whether or not these two variants alone are enough to say Eusebius was not behind the production B(03) is not clear to me. Yet, this fact is not completely irrelevant to the absence of Eusebian sections. Andrist, “Le milieu,” 229, 236; Knust and Wasserman, To Cast, 187. 51 Eusebius was apparently concerned about the charge of vanity, if he included his own work at the expense of timeliness. Skeat, “Sinaiticus,” 226; For a similar argument, see Dormandy, “Pandects,” 31. 52 See already Hug, De antiquitate, 12–14; Hug, Introduction, 1:266–267. 53 Gardthausen, Griechische Palaeographie 2:249; Milne and Skeat, Scribes, 66; Cavallo, Ricerche, 56; Jongkind, Scribal Habits, 87; cf. Elliott, “T.C. Skeat,” 76. 54 Hug, Introduction, 1:266; 55 Dormandy, “Pandects,” 29; cf. Tregelles, Introduction, 164; Ropes, Beginnings, xxxc; Birdsall, “Vaticanus,” 33. 56 Dormandy, “Pandects,” 31. 57 On their relationship, see Chapter 6. Problema; Porter, “Papyrus Bodmer XV (P75),” 363–376; On the provenance of P75, see the recent discussion in Nongbri, God’s Library, 157–168. INTRODUCTION 9 Moreover, the “B-text” of 1 Kingdoms is represented in the region by the Ethiopic and has some affinity to the Coptic version (see Chapter 5).58 The most persuasive argument for Egyptian provenance is found in the similarity of the contents and order of books in Athanasius’s 39th Festal Letter (367).59 It is well known that the contents are almost identical, apart from the inclusion of Wisdom, Sirach, Esther, Judith, and Tobit, which Athanasius distinguished from the κανονιζοµενα (see §1.3). 60 However, even these are included together in B(03) and in the same order as Athanasius. Additional evidence for the distinctiveness of the αναγινωϲκοµενα in B(03) may find partial support in the absence of early section numbers for these five books (cf. §3.1.1).61 Henry Swete is right to highlight the similarity of B(03) with other canon lists in the east and west, but there remains a unique connection to Athanasius. According to Patrick Andrist, B(03) and the Festal Letter are the only texts before the fifth century, which include the five OT αναγινωϲκοµενα, but omit the Maccabees.62 Proponents of the Caesarean theory have yet to provide a satisfactory explanation for this unprecedented agreement.63 Andrist suggests the theory that Eusebius ordered the production of B(03) to have a copy of the opponent’s Bible, but he ultimately prefers an Alexandrian provenance with Caesarean influence in the late fourth century. 64 Amphoux, on the other hand, has accounted for the relationship by placing the production of B(03) in Rome, while Athanasius was present (c. 340). Most importantly, there are numerous compelling reasons to doubt that א(01) and B(03) were 58 Aejmelaeus is in the company of Rahlfs and Swete, but advances the notion that this text represents the first Christian recension. Zuntz argued that a variety of exemplars with different text forms would have been needed in Caesarea to fulfil Constantine’s order. Elliott believes this could, therefore, explain B(03)’s affinity with Egyptian manuscripts. Aejmelaeus, “New Perspectives”; Zuntz, Lukian, 44; Elliott, “T.C. Skeat,” 73. 59 On the letter, see Gallagher and Meade, The Biblical Canon, 118–129; Andrist describes the ordering of B(03) as the traditional Egyptian canon. Athanasius’s innovation in the contents was the distinction between κανονιζοµενα and the αναγινωϲκοµενα—Wisdom, Sirach, Esther, Judith, Tobit, Didache, and Shepherd. Andrist, “Le milieu,” 239–240. 60 Alfred Rahlfs appears to be the first to have made this argument. Rahlfs, “Alter und Heimat”; See the table of contents in Skeat, “Sinaiticus,” 213. 61 This is not entirely obvious, since early section numbers are also absent in Genesis–Numbers and Job. Grenz, “Textual Divisions,” 16. 62 Thus, it seems unlikely that we can attribute the omission of Maccabees to mere oversight, as David Parker suggests. Swete, An Introduction, 219; Andrist, “Le milieu,” 239–240; Parker, An Introduction, 72. 63 I am, however, in agreement with Skeat that the comparison of the alternative Pauline letter order, found in the section numbers of B(03) (see Chapter 3), with the Coptic order of the 39th Festal Letter is inconclusive. Skeat, “Sinaiticus,” 212–214; pace Lake, “Manuscripts,” 35. 64 He highlights the exile of bishop Gelasius from Caesarea to Egypt, and his possible accompaniment by two Caesarean scribes. Andrist, “Le milieu,” 246–247. CHAPTER 1 10 copied in the same scriptorium. There is, therefore, little necessity to assume that the production of B(03) occurred in Caesarea. 1.3 Contents We have already described B(03) as a pandect, including the Greek OT and NT. However, the entire codex is no longer preserved, as it is missing Genesis 1–46:28a, 1 Kingdoms 2:5–7, 10–13, Psalms 105:27–137:6b, and everything that followed Hebrews 9:14a (Table 2). There is no codicological evidence to suggest that the manuscript originally contained any of the Maccabees (see n. 61) or the Prayer of Manasseh. TABLE 2: CONTENTS OF B(03) Old Testament65 New Testament [Gen 1–46:28a] Gen 46:28b–50 Ex Lev Num Deut Josh Judg Ruth 1–4 Kgdms [1 Kgdms 2:5–7, 10–13] 1–2 Chr 1–2 Esd Psalms [Pss. 105:27–137:6b] Ps. 151 Prov Eccl Song Job Wisd. Sirach Esther Judith Tobit Hosea Amos Micah Joel Obadiah Jonah Nahum Habakkuk Zeph Haggai Zechariah Malachi Isaiah Jeremiah Baruch Lamentations Epistle of Jer Ezekiel Daniel Matt Mark Luke John Acts James 1–2 Peter 1–3 John Jude Romans 1–2 Cor Galatians Ephesians Philippians Colossians 1–2 Thess Heb 1–9:14a [Heb 9:14b–13] [Pastoral Epistles] [Philemon] [Revelation] Proposed: [Apostolic Fathers] [Didache & Shepherd] [1 Clement] There have been a variety of opinions about what followed the letter to the Hebrews. Since the Pauline corpus, including Hebrews, follows the Catholic Epistles, we are left to wonder if the Pastoral Epistles and Philemon appeared before Revelation. However, even the original presence of Revelation has been questioned. David Parker prefers not to include B(03) in his list of seven complete Greek Bibles, because of the uncertainty around the ending of the codex.66 In part, this 65 Content in brackets indicate lacunae or missing books, whether known or hypothesized. 66 Parker, An Introduction, 72. INTRODUCTION 11 conclusion is based on Keith Elliott’s omission of B(03) in his list of complete NTs.67 However, Elliott is elsewhere clear that B(03) did originally contain Revelation.68 Furthermore, Skeat criticized the notion that “a manuscript of the entire Greek Bible written in the middle of the fourth century would have omitted the Pastorals.”69 The strong connection to Athanasius’ Festal Letter provides further evidence that the Pastoral Epistles and Philemon would have been copied between Hebrews and Revelation. When the missing pages of Hebrews and Revelation were recopied in the fifteenth century (GA 1957), the Pastorals and Philemon were not restored. While Skeat blamed this on a hasty restoration effort, it is more likely that the epistles were omitted because the majority of manuscripts in the fifteenth century contain them before Hebrews.70 If the restorers were only aware of this order, they might have assumed that B(03) was lacking the epistles from the beginning. Based on the list of αναγινωϲκοµενα in Athanasius, Rahlfs thought that the Didache and Shepherd originally followed Revelation.71 Hugh Macmillan, on the other hand, prefers the text of 1 Clement, probably by analogy to Codex Alexandrinus (A[02]).72 Kurt and Barbara Aland remain agnostic about the additional contents, but suggest that some texts from the Apostolic Fathers were present.73 This question must remain open, but if additional texts were present after Revelation, the close affinity with Athanasius would raise the prospect of the Didache and Shepherd. 1.4 Vaticanus and Sinaiticus It is clear from the previous sections that discussion concerning B(03)’s relationship to א(01) is unavoidable. The two manuscripts contain a striking number of resemblances, many of which will be highlighted throughout this study. Both are large pandects copied in the bookhand known as Biblical Majuscule (see Chapter 3), and although B(03) is written in three columns per page and 67 Elliott, “The Distinctiveness,” 153; Parker, An Introduction, 72. 68 For example, Elliott considers B(03) along with א(01) and Alexandrinus (A[02]) as the earliest complete manuscripts, which “agree in including the same 27 books.” Likewise, when he calls B(03) “virtually complete,” he is referring to the absence of 1–4 Maccabees and Manasseh, not Revelation. Elliott, “Manuscripts,” 618, 627–628. 69 Skeat appears to include Philemon in the Pastorals, since he does not mention its omission earlier. Skeat, “Vaticanus,” 133; cf. Nongbri, “Pauline Letter,” 90. 70 The exceptions to this ordering include GA 1729, 1947, 1978, 1996, and 2201. Skeat admits this was also a likely factor. Skeat, “Vaticanus,” 133. 71 Rahlfs, “Alter und Heimat,” 72–79; cf. Nestle, Introduction, 60. 72 Macmillan, Roman Mosaics, 369. 73 Aland and Aland, The Text, 109. CHAPTER 1 12 א(01) four columns, they share a two-column format in the Poetic books of the Greek OT.74 However, the question remains whether these similarities should be accounted for by direct connection, or whether the similarities have been amplified by the lack of contemporary parallels. I have already noted my hesitation to assume the shared origin of the two codices, as scholars have often depended too heavily upon א(01) for evidence concerning the provenance of B(03). In his Appendix codicum celeberrimorum, Tischendorf first argued that א(01) and B(03) not only originated together, but even shared one of their scribes.75 His comparison of the two codices led him to argue that the NT of B(03) was copied by scribe D of א(01).76 According to Tischendorf, this identification can be made on account of the similar use of xi with a curved tail ( ), line-fillers, marginal sigla (↓↑), nomina sacra, colons at the end of each book, and similar orthography. On the other hand, Tischendorf does note that א(01) is copied in noticeably larger letters than B(03), and in forty-eight lines-per-column, rather than forty-two. Milne and Skeat agreed on the close relationship between the two codices but criticized Tischendorf’s identification of scribe D with that of the NT (their scribe B) in B(03).77 Rather, they associate scribe D with their scribe A in B(03), who copied portions of the Greek OT (see §1.6). In support of this claim, they present the following parallels: similar (1) colophons and coronides (especially Mark in א[01] and Deuteronomy in B[03]), (2) use of line-fillers, (3) paragraphing, (4) and spelling of ιϲχυει; the appearance of (5) cursive ανω and κατω indicators in certain marginal corrections, and (6) the inverted pyramid-shaped corrections (see Chapter 4).78 Following the conclusion of Milne and Skeat, Versace adds to these similarities the bent oblique stroke of the και-compendium ( ) in scribe D and the marginal corrections of B(03).79 In Chapter 3, we will find that scribe A of B(03) often uses an ornamental mu 74 Canart claims that the resemblance in writing is so strong that one has to appeal to paratextual or extra- paleographic evidence to distinguish them. Grafton and Williams suggest that the use of two, three, and four columns reflects the context of Caesarea with Eusebius, where many innovations involving the column originated. Canart, “Notice,” 39; Grafton and Williams, Christianity, 220–221; cf. Dormandy, “Pandects,” 28–29. 75 Tischendorf, Appendix codicum, ix–xi; cf. Tischendorf, Vaticanum, xxi–xxiii; Milne and Skeat, Scribes, 89. 76 Tischendorf believed scribe D copied Tobit, Judith, the first part of 4 Maccabees, and six folios of the NT. Tischendorf, Sinaiticum, xxi; Tischendorf, Appendix codicum, x; cf. Jongkind, Scribal Habits, 9. 77 cf. Head, “Scribe D,” 134–135. 78 Milne and Skeat, Scribes, 89–90. 79 Versace, Marginalia, 17–18 n. 24. INTRODUCTION 13 with a long curved tail. This too can be found especially in the work of scribe D in א(01).80 However, Milne and Skeat are cautious to assume that scribes D and A are the same, since there remains some difference. Rather, they conclude, “the identity of the scribal tradition stands beyond dispute.”81 Paul Canart considered this conclusion Skeat’s strongest contribution to the study of B(03). In support of a shared copying location, Canart adds the unusual ruling patterns found in both codices.82 Unfortunately, Canart does not clarify this point with any examples. By comparing his schematic chart of the ruling in B(03) with that of Milne and Skeat in א(01), I have found three patterns of agreement.83 The first can hardly be called unusual as it simply assigns one line of text to a single ruled-line. The remaining two patterns are significant as they provide five or three single- text lines, followed by a number of double-text lines and a single-text line at the end of the column.84 While this agreement may be surprising, there are more ruling variations than similarities in the two codices. Milne and Skeat outline numerous irregularities in א(01), which do not appear in Canart’s schematics of B(03), including the ruling of three and four-text lines (e.g., Quire 79 f. 1v; f. 240r). They also demonstrated that א(01) was lined horizontally across complete bifolios. This does not appear to be the case in B(03), where the horizontal lines do not transgress the far left and right bounding lines (see Chapter 2). What Canart highlights as unusual ruling may in fact have been more widespread in large codices than the extant evidence allows us to conclude. There remain, however, other reasons for the rejection of the apparent connection between א(01) and B(03). As Knust and Wasserman put it, “There are simply too many differences between these two manuscripts to make them products of the same editorial initiative.”85 We have already 80 Head, “Scribe D,” 130. 81 Additionally, Elliott rightly notes that the two codices are virtually alone in this omitting Mark 16:9–20, even if the scribe of B(03) hesitated in doing so. We will see in Chapter 3 that the blank column on p. 1303B was likely unintended by those who planned the production of B(03), and rather indicates, as Elliott suggests, some hesitation on the part of the scribe. Milne and Skeat, Scribes, 90; Elliott, “T.C. Skeat,” 71; cf. Harris, Stichometry, 73 82 Canart, “Notice,” 39. 83 Canart, “Notice,” 31; Milne and Skeat, Scribes, 76–78. 84 Milne and Skeat give the codes “.1.1.1.1.221.1.” and “.1.1.222.1.” to represent these patterns in א(01). Canart presents the standard Leroy formula for these patterns as “Xa5b2” and “Xa3b2.” Leroy, Réglure, xxiv. 85 They do, however, leave open the possibility of a shared provenance. Knust and Wasserman, To Cast, 186, 189. CHAPTER 1 14 mentioned the significant differences in their contents and order of books.86 Although there is some agreement in ruling, B(03) does not appear to have been ruled across bifolios as in א(01). Likewise, B(03) was bound in quinions (quires of 5 sheets) rather than quaternions like א(01) and A(02) (see Chapter 2). In his codicological comparison of seven pandects, Patrick Andrist has shown the “great architectural diversity” between the codices, with very few constants.87 It is unlikely, therefore, that Skeat’s economic proposal for the codicological differences between א(01) and B(03) can sufficiently explain such diversity.88 Similarly, the differences in section numbering and texts are often cited as reasons to believe the two codices represent different production settings.89 The exception to this is the numbering in Acts, which only differs in five locations between the two codices and is largely overlapping with the Euthalian sections.90 The case is far from closed, but we must heed Parker’s warning against exaggerating connections in a context with fragmentary remains.91 As nearly all of the compelling arguments for the provenance of B(03) suggest, the great deal of movement between locales in the fourth century could easily allow for a broad “scribal tradition” (in Skeat’s understanding) without necessitating shared scribes or even provenance.92 1.5 Dictation and the Exemplar(s) In their study of א(01), Milne and Skeat present the argument that the codex was copied through dictation.93 Nearly two decades later, Skeat gave two lectures extending their argument in favor of dictation theory.94 Space does not permit a full summary of his arguments, but it is important to note that the strongest evidence has been criticized from a variety of perspectives. For example, 86 Westcott and Hort, Introduction, 214; Knust and Wasserman, To Cast, 186; Aejmelaeus, “New Perspectives,” 6. 87 Andrist, “Au croisement,” 71–72; cf. Andrist, “Le structure,” 11–37. 88 Certainly, the smaller sized parchment sheets can be explained as a more economical alternative to those of א(01). Skeat, “Sinaiticus,” 220–228. 89 Westcott and Hort, Introduction, 214; MacMillan, Roman Mosaics, 370; Bogaert, “Le «Vaticanus»,” 136. 90 Robinson, who suggests a Caesarean origin of B(03), claims א(01) and B(03) received these numbers “quite independently of one another.” Robinson, Euthaliana, 37; Elliott, “T.C. Skeat,” 74; Jongkind, Scribal Habits, 121. 91 Parker, Textual Scholarship, 73–74. 92 This is especially the case in Andrist’s reconstruction, where Caesarean scribes could have accompanied bishop Gelasius to Alexandria. Milne and Skeat are careful to admit that the lack of contemporary evidence means that “no amount of similarity can be used to decide origin.” Andrist, “Le milieu,” 246–247; Milne and Skeat, Scribes, 66. 93 Milne and Skeat, Scribes, 55–59. 94 Published in Skeat, “Dictation,” 3–32. INTRODUCTION 15 James Royse and Dirk Jongkind have argued that many of the orthographic errors leveraged by Skeat could also be explained through visual copying, including internal dictation and obscurities in the exemplar. 95 A recent study on duplicate papyri by the same scribes has also confirmed that orthographic variation occurred through visual copying.96 Moreover, one of the most repeated arguments of Skeat is the apparently nonsensical combination of numerals in place of the correct reading οκτακιϲχιλιοι in 1 Maccabees 5:20.97 In response, Zachary Cole has shown that this example is based on a misrepresentation of the evidence. He shows that the fourth, irregular character in the numeral is the cursive form of the standard /γ̅ (3000); the two other examples of this character were corrected to this standard form (1 Macc 9:5; 11:44).98 Cole agrees with Jongkind’s suggestion that the cursive form, which is isolated in 1 Maccabees, was inherited from the exemplar through visual copying.99 I have not found any explicit reference to the dictation of B(03) in Skeat’s work, but his claims concerning the shared scribe and scriptorium would seem to imply such a conclusion. Pierre- Maurice Bogaert, on the other hand, has concluded from his study, “l’orthographe de B est satisfaisante. B a été copié, non dicté.”100 We will examine the orthography in later chapters of this study, but recent work has shown far more sophistication in the orthography of B(03), which is not easily accounted for by dictation.101 For further clarity, it will be necessary to study the mise en page of the codex and how the mechanics of copying were affected by the physical parameters of B(03). M. A. Dain made the argument against dictation based on the complexity of copying the text according to a pre-arranged mise en page.102 Interestingly, Skeat acknowledges Dain’s argument for 95 Royse, Scribal Habits, 83–90; Jongkind, Scribal Habits, 251. 96 It should be added that Yuen-Collingridge and Choat helpfully remind the reader that the majority of literary references to dictation refer to composition rather than reproduction. They also critique the notion that desks for holding opened manuscripts did not exist in the Roman world. Yuen-Collingridge and Choat, “Copyist,” 828–829. 97 See Milne and Skeat, Scribes, 57; Skeat, “Four Years’ Work,” 114; Skeat, “Dictation,” 17. 98 Cole, “Paleographical Problem,” 103–107. 99 Jongkind, Scribal Habits, 252; Cole, “Paleographical Problem,” 106. 100 Bogaert, “Le «Vaticanus»,” 137. 101 Already in 1966, Carlo Martini exposed important orthographic agreements between P75 and B(03), suggesting some relationship in the ancestor. Martini, Problema, 86–122; cf. Canart and Martini, Introduction, 11; See recently Williams, “Semitic Long /i/,” 15–26; Williams, “When Does ϲυν- Assimilate?,” 429–438; Jongkind, “Redactional Elements,” 231–245. 102 Dain, Les manuscrits, 22. CHAPTER 1 16 “medieval manuscripts with their carefully drawn margins and lines.”103 B(03) may not be medieval, but it certainly contains carefully pricked and lined pages (Chapter 2). So too, we will find a number of significant features like lists and genealogies which have been copied with a modified structure, best accounted for by visual copying. To this we can add the consistent wordbreaks, unique abbreviations, and errors, which were often caused by the line or column-breaks in B(03). In part, our evaluation of the relevant data for visual copying or dictation theory is dependent on the exemplar(s) of the codex. However, without access to the same manuscripts used by the scribes, it is rare that a precise description of their format can be produced.104 There is general agreement that the production of B(03) was something pioneering or even experimental and, consequently, we cannot be certain that a single exemplar was used for the whole codex. 105 Michaelis, on the other hand, believed that the exemplar was obviously not an autograph precisely because it did not contain “single gospels, or epistles, but the whole canon divided into sections.”106 Furthermore, Jongkind has recently argued that the ancestor of B(03) was a carefully edited master copy.107 Since his evidence is confined to the NT, this may be the extent of the master copy. Yet, it is also possible that the scribes carefully collected and edited several exemplars to generate uniformity throughout while preserving the unique marginalia and divisions in particular sections of the codex. This study will, therefore, highlight consistencies and also inconsistencies in the codicology, paleography, and corrections in the codex, as they may indicate features inherited from the exemplars. Those who have attempted to answer the question of layout in the exemplars concentrate on the line lengths. Hort proposed line-lengths of twelve to fourteen letters, since longer omissions usually consist of this number of characters or multiples of it.108 Based on other omissions, Albert Clark has suggested shorter lengths of ten to twelve letters. 109 Metzger, on the other hand, 103 Skeat, “Dictation,” 14. 104 See, however, the description and reconstruction of the exemplar of N(022), O(023), and Σ(042) in Hixson, Scribal Habits, 254–255, 271–307. 105 Westcott and Hort, Introduction, 267–269; Kreuzer, “B or not B,” 272; cf. Elliott, “T.C. Skeat,” 67; Andrist, “Au croisement,” 76. 106 Michaelis, Introduction, 2:345, n. s. 107 Jongkind, “Redactional Elements”. 108 Westcott and Hort, Introduction, 233–234. 109 Clark, Primitive Text, 33. INTRODUCTION 17 reproduced the lines of B(03)’s exemplar from the omission of κοϲµου αλλ ινα τηρηϲηϲ αυτουϲ εκ του (thirty letters) in John 17:15, which was caused by homoeoteleuton with lines of fifteen letters.110 Gregory Paulson’s recent study of singular readings in Matthew has argued for line-lengths of ten letters, since two of the singular omissions consist of ten letters, one singular addition involves twenty letters, and another addition repeats a word from ten letters before.111 The proposed line- lengths, therefore, vary from ten to fifteen letters, based on the types of omissions that occur in the codex. Still, it is not clear that the lines of the exemplars were always the cause of omission, since line or column-breaks in B(03) could also have triggered scribal error. We will return to this question after examining the corrections of omission in Chapters 5 and 6. 1.6 The Scribes of B(03) Since Milne and Skeat published their brief appendix on the scribes of B(03), it has become customary to refer to two scribes as copyists of the entire manuscript.112 However, before their research on the codex, opinions about the number of scribes varied from one to four (Table 3). Both Frederic Kenyon (1898) and Edward Thompson (1912) only noted one scribe in the whole manuscript.113 Eberhard Nestle (1901), following the earlier work of Ezra Abbot, thought there were at least two scribes in the Greek OT portion of B(03).114 Tischendorf (1867) argued for three scribes, with one responsible for the whole NT.115 Ludwig Traube (1907) and James Ropes (1926) believed that there were three or four scribes responsible for B(03).116 This spectrum of results, from one to four scribes, forces us to ask whether we should take Milne and Skeat’s conclusion for granted. Indeed, they themselves claim that “were it not for the absolute evidence of the colophons one might be tempted to suspect a third hand.”117 In his recent study of the marginalia in B(03), Pietro Versace dissents from the current consensus, indicating that his experience with the codex leads 110 Metzger and Ehrman, The Text, 253. 111 Paulson, Scribal Habits, 56 n. 66. 112 Milne and Skeat, Scribes, 87–90; cf. Cavallo, Ricerche, 53; Canart, “Notice,” 25; Andrist, “Au croisement,” 17. 113 Kenyon, Our Bible, 135; Thompson, Introduction, 200. Andreas Birch also suggests that the Greek OT and NT were copied by the same hand. Birch, Kritisk Beskrivelse,51–52. 114 Nestle, Introduction, 61. 115 Tischendorf, Vaticanum, xxi n. 2; Tischendorf, Appendix codicum, ix; Lake, Text, 14; Metzger, Manuscripts, 74. 116 Traube, Nomina sacra, 66f.; Ropes, Beginnings, xxxviii. 117 Milne and Skeat, Scribes, 89. CHAPTER 1 18 him to believe there may have been more than two scribes, though the conclusion rests outside the purview of his project.118 TABLE 3: THE NUMBER OF SCRIBES IN B(03) 1 Scribe 2 Scribes 3 Scribes 3 or 4 Scribes Birch (1785) Kenyon (1898) Thompson (1912) Milne and Skeat (1938) Tischendorf (1867) Lake (1908, 4th ed.) Versace (2018)? Traube (1907) Ropes (1926) The discrepancy between these opinions is rooted in the difficulty of paleographic analysis in B(03). Milne and Skeat admit: In the Sinaiticus the shapes of individual letters have… proved to be the least satisfactory criterion between hands… It is doubly so of the Vaticanus, where the original script has been almost entirely overwritten by a later hand… The investigator is therefore more than ever thrown back upon the other and less subjective tests which have proved so effective in discriminating the scribes of the Sinaiticus.119 Indeed, the reinking of the entire manuscript in the tenth or eleventh century and again in the sixteenth century renders the traditional method of identifying scribal hands inadequate. 120 Because of this, we cannot rely solely on the handwriting of the scribes and must look to other possible evidence, such as codicology, paratexts, and spelling. However, it is important to note that the “script has been almost entirely overwritten” and, therefore, we do find unreinforced text, which reveals the original hand of the scribes. This will be crucial to the latter part of Chapter 3. In 1872, Ezra Abbot was the first to explicitly identify a change of scribes in 1 Kingdoms 19:11, based on a discrepancy in paragraphing, line–fillers, and nomina sacra. 121 Abbot counts 1441 occurrences of ekthesis—the projection of the first letter of a new section into the margin—in the first 294 pages of the codex (Gen 46:28–1 Kgdms 19:11a), with only two examples in the following 290 pages (1 Kgdms 19:11b–2 Esdras). Likewise, he noted a large difference in the use of line-fillers (>) and the abbreviation of πνευµα and ιϲραηλ between the two halves of 1 Kingdoms. Ludwig Traube and Milne and Skeat agreed with Abbot, but differed in what follows. Traube based his analysis on the use of nomina sacra throughout the codex and found four scribes, or possibly three (Table 4).122 118 Bogaert is also open to there being more than two scribes. Versace, Marginalia, 10 n. 8; Bogaert, “Le «Vaticanus»,” 137–138. 119 Milne and Skeat, Scribes, 87. [Italics added] 120 On the date of the reinforcement, see Versace, Marginalia, 7, 67; cf. Chapter 2. 121 Abbot, “Antiquity," 189–200. 122 Traube, Nomina Sacra, 66–67. INTRODUCTION 19 TABLE 4: THE SCRIBES OF MILNE AND SKEAT AND OF TRAUBE Ludwig Traube Milne and Skeat B1: pp. 1–334 (Gen 46:28–1 Kgdms 19:11) B2: pp. 335–674 (1 Kgdms 19:11–Ps 77:71) B3: pp. 675–1244 (Ps 77:72–Matt 9:5) B4/or B2: pp. 1245–1518 (Matt 9:5–Heb 9:14)123 Scribe A: pp. 41–334 (Gen 46:28–1 Kgdms 19:11) Scribe B: pp. 335–624 (1 Kgdms 19:11–2 Esdras) Scribe A: pp. 625 – 944 (Psalms–Tobit) Scribe B: pp. 945 – 1518 (Hosea–Heb 9:14) Milne and Skeat, on the other hand, identified only two scribes in the whole codex.124 As we saw above, they believed that scribe A, rather than Tischendorf’s suggestion of scribe B, was possibly the same scribe as scribe D in א(01). They too examined nomina sacra, but also utilized colophons, titles, line-fillers, paragraphing, and orthography as tools for identifying the scribes. However, as we will see in Chapter 3, the colophons were the decisive evidence for two scribes in the opinion of Milne and Skeat. It is the discrepancy between Traube and Milne and Skeat on the division of scribes in Psalms that will occupy a significant portion of this study. 1.7 Plan of this Study The cautious conclusions in the preceding discussion set the course for this thesis. That B(03) was a complete pandect from the fourth—or even early-fifth—century leads us to approach the codex as an early attempt to materialize the canonical consciousness of the early church.125 Therefore, the many observable discontinuities in the manuscript are often witness to the growing pains of such a procedure. Furthermore, the visual nature of the copying process implies that the scribes of B(03) were occasionally influenced by the particularities of the exemplars, and yet free to incorporate their distinctive habits according to the constraints of the layout. Finally, while B(03) and א(01) can no longer be assumed to share either scribes or provenance, their similarities remain important for mutually understanding the context of each codex. Although we prefer the Egyptian or Italian provenance of B(03), any theory of origin must accept that external influence from other regions possibly played a role in production, whether from Egypt to Rome or Caesarea to Alexandria. As is clear from our title and the preceding pages, this project aims to complete and modify the initial analysis of B(03) set forth in the appendix of Milne and Skeat’s Scribes and Correctors of the Codex Sinaiticus. Likewise, many of their methods for studying the scribes of א(01) will be imitated throughout, though often in more detail. While Milne and Skeat cover an impressive 123 On Traube’s break at Matthew 9:5, see §3.2.5. 124 Milne and Skeat, Scribes, 87–89. 125 What Wallraff has called the “Kanonbegriff” (Canon concept). Wallraff, Kodex und Kanon, 48. CHAPTER 1 20 amount of material, their study of א(01) is confined to a mere eighty-six pages. This has led to the charge that the two offer more of an impression concerning the scribes and correctors, leaving the reader to either trust their claims or collect the data themselves.126 The following pages provide numerous images, tables, charts, and transcriptions to assist readers in evaluating the various arguments. Unlike Milne and Skeat with א(01), I have not had the opportunity to perform an autopsy of the physical codex.127 I am, therefore, dependent on the 1999 and, to a lesser extent, the 1904–1907 photo-facsimile editions, including the online images provided by the BAV.128 Part I of this study concentrates on codicological, paleographic, and paratextual questions as they relate to the scribes of B(03). In Chapter 2, we present the material nature of the codex through the lens of structural codicology. Recent developments in the field suggest that aligning structural irregularities help us understand the layers of production in the manuscript, and possibly assist the identification of scribe changes. This is certainly the case in B(03), where a number of observable discontinuities occur at 1 Kingdoms 19:11, the verse in which scholars have agreed a new scribe assumed responsibility. It is likely, therefore, that codicological features can assist in identifying other scribe changes in B(03). Following the insights from structural codicology, Chapter 3 analyzes the paratexts and paleography of the codex as they relate to the scribes and the earliest layers of the production phase. We saw already that textual divisions, paratexts, and nomina sacra have been the key for scholarly divisions of the scribes. However, numerous other features remain unexamined and a comprehensive comparison is still necessary. It will be shown that utilizing paratexts can assist paleographic analysis of the hands, where the original handwriting is still visible. A summary of codicological, paratextual, and paleographic features will allow us to set forth a modified division and number of scribes, bringing together conclusions from both Traube and Milne and Skeat. Part II follows the proposed new division of scribes by examining early corrections throughout B(03). While there is no claim to the comprehensiveness of emendations in the codex, it is argued that an examination of the types of early corrections can provide some insight into the 126 Malik, “Corrections,” 212. 127 I am grateful to have been invited to view a single sheet of B(03) (pp. 1490, 1499; Gal 3:9–4:6, Phil 1:1–28) in the Sistine Hall, BAV (2 July 2019). 128 Bibliorum SS. graecorum Codex Vaticanus 1209 (cod. B), 4 vols (1904–1907); Codex Vaticanus B. Bibliothecae Apostolicae Vaticanae Codex Vaticanus Graecus 1209 (1999); Online images: https://digi.vatlib.it/view/MSS_Vat.gr.1209. INTRODUCTION 21 quality of original copying, as well as the overall editing behavior underlying the production of B(03). Chapter 4 introduces the various methods of correction through the story of scholarly interaction with the codex. While there was far less initial discussion concerning the scribes of B(03), early scholars already understood that the manuscript was corrected in various stages. After examining the methods of correction, namely the way information was added to the codex to indicate the intended addition, removal, or alteration of text, Chapters 5 and 6 survey the early corrections in 1 Kingdoms, Psalms, and the NT. Drawing on the conclusions of Part I, the aim of Chapter 5 is to compare the types and frequency of corrections between two scribes within the same book. Since the scribe of the NT is also responsible for copying roughly half of 1 Kingdoms and Psalms, the comparison with another scribe or scribe(s) in these books provides the foundation for a review of NT corrections in Chapter 6. The conclusions from Part II support the often-repeated description of the NT scribe as generally careful and mechanical, while having a propensity to omit text. However, the earliest corrections, including those executed by the scribe, provide further information surrounding the production and editorial context of B(03). For clarification, the primary means of referencing the codex is through the modern pagination, hand-copied in the upper corners of each page.129 Since the pagination was added after portions of the manuscript were lost, this does not reflect the original structure of the codex (see Chapter 2). When referencing a specific column or marginal notation connected to a column, the page number is directly followed by a column-letter, “A,” “B,” or “C” (e.g., p. 1512B). In some instances, when a specific line is required for ease of reference the citation will be followed by “l.” or “ll.” and the line number(s) (e.g., p. 1512B l. 14). Since Part I is largely concerned with material observations, this reference will be enough. However, when discussing the text of B(03), the biblical reference will follow the page number and column letter (e.g., p. 1512B, Heb 1:3). Readers can consult all images through the BAV digital library or the NT images with transcriptions on the NTVMR.130 129 Versace dates the addition of the pagination to the sixteenth century. Versace, Marginalia, 66. 130 https://digi.vatlib.it/view/MSS_Vat.gr.1209; https://ntvmr.uni-muenster.de/manuscript-workspace. Part I: Codicology, Paratexts, and Paleography CHAPTER 2 Codicology: The Material and Make-Up of B(03) Before we begin discussing the text of B(03), we will spend the following two chapters discussing the non-textual features of the manuscript. This procedure has become common practice in recent manuscript studies, with particular interest in the material codex and its relationship to the scribes.1 Even more important for our study is the work of Paul Canart (1927–2017) on the codicology of B(03), originally published in the introduction to the 1999 facsimile.2 As the former vice-prefect of the Vatican Library, Canart’s study is the fruit of in-depth study with the physical artifact, as well as years of experience in the field of codicology and paleography. However, the brevity of his chapter leaves our study with many unanswered questions. For example, while Canart helpfully outlines some of the quire irregularities in B(03), he does not explore their implications on the production of the codex, nor does he explain the differences between his reconstruction and the manuscript’s current structure (see below, §2.3.1). The second limitation, which is particularly important for this study, is his lack of attention to the scribes of B(03). While he points to the probability that Milne and Skeat were right about the two scribes of B(03),3 he does not present the implications that his codicological study might have on the work of the copyists. Why, however, is a codicological study important to understanding the scribes of a manuscript? One’s answer to this question will be largely influenced by how one views the manuscript as an entity. J. P. Gumbert helpfully notes the twofold nature of the codex: (1) the “material face” and (2) the text, an “immaterial object,” which “cannot reach us except clad in a body.”4 While recent methods such as the Coherence-Based Genealogical Method extract the text from its embodied state,5 this study is concerned with understanding “the ‘body language’ of the 1 Parker, Codex Bezae; Jongkind, Scribal Habits; Smith, Alexandrinus; Malik, P.Beatty III; Ebojo, “A Scribe”; Hixson, Scribal Habits. 2 Canart, “Notice,” 19–45. 3 Canart, “Notice,” 25. 4 Gumbert, “Fifty Years,” 506. 5 This is seen primarily in the distinction between the terms “witness” and “manuscript.” While it is important to remember that young manuscripts can contain old readings, this does not mean the text can be so easily liberated from its material form. For these terms, see Mink, “Contamination,” 143; Cf. Gurry, Critical Examination, 4–5. CHAPTER 2 26 book” before understanding the text it contains.6 In what follows, I present a description of the codicological method I am adopting and a survey of the material make-up of B(03). 2.1 The Study of a Multiple-Text Codex As codicology has developed into its own discipline (distinct from paleography and philology),7 there have been increasing attempts to produce a unified terminology of the book and its study. This effort is complicated by the variety of book cultures (e.g., Arabic, Armenian, Hebrew, Greek) and the scholarship surrounding each of them.8 This is not the place to survey the various attempts at a universal terminology, but only to highlight the points of contact with our own study. 2.1.1 What is Codicology? Until now, I have used the term “codicology” in its plain sense, but it should be asked what we mean by the term and, in practice, what is its goal? Specialists in the field often note that “codicology” was first printed in M. A. Dain’s Les manuscrits (1949),9 while his contemporary, Charles Samaran, had used the term “codicography,” clearly as an adaptation from “paleography.” 10 Neither of these French scholars, however, went as far as François Masai, who argued that codicology ought to be considered an archeological discipline, distinct from paleography and art history. 11 Almost any definition of codicology given today maintains this archeological aspect, i.e. codicology as “the archeology of the book.”12 However, the study of handwritten books does not end with production, but continues to analyze the after-life and conservation of the manuscripts.13 The present study might as well follow Gumbert’s more specific term “material codicology,” which is primarily “concerned with the manuscript book as a material object and a craftsman’s 6 Gumbert inherits this body language from Traube, who writes that the two “things belong to each other like body and soul.” Traube, “Zur Paläographie,” 8; Cited in Gumbert, “Fifty Years,” 508. 7 Only within the past century has codicology moved from an “auxiliary science” of history to an independent field of research. Cf. García, Introducción, 17; Touwaide, “Codicology and Paleography,” 266. 8 For this problem see Friedrich and Schwarke, “Introduction,” 25ff.; cf. Maniaci et. al., “Codicology,” 71. 9 Dain, Les manuscrits, 71; Gumbert, “Fifty Years,” 506. 10 German scholars already had the term, Handschriftenkunde (e.g. Traube’s Zur Paläographie und Handschriftenkunde), but this “was not seen as a separate discipline.” Gumbert, “Fifty Years of Codicology,” 505; Cf. García, Introducción, 20; Touwaide, “Codicology and Paleography,” 300; Agati, Manuscript Book, 21. 11 Masai, “Paléographie et codicologie,” 292. Cf. García, Introducción, 21; Agati, Manuscript Book, 26. 12 Agati, Manuscript Book, 22; Masai, “Paléographie et codicologie”; See also Canart, Lezioni; Gullick, “Codicology.” 13 Johnston and Van Dussen eds., Medieval Manuscript, 4. CODICOLOGY 27 product.”14 In studying the scribes of B(03), we are likewise interested in the production of a material object. Yet, this chapter is equally, if not more, concerned with viewing and understanding the scribes through the physical constraints of the manuscript. 2.1.2 Structural Codicology It should be highlighted that current scholarship has produced multiple types of codicology. A search through the handbooks on codicology will demonstrate the popularity of Quantitative and Comparative approaches to the discipline.15 However, with the publication of La syntaxe du codex in 2013, a full-scale Structural or Stratigraphic approach has now been proposed. In this study, Patrick Andrist, Paul Canart. and Marilena Maniaci emphasize the subtle complexities of the handwritten book16 and then propose a way forward in reconstructing the history from observable discontinuities in the codex.17 While their bold approach is focused on later medieval manuscripts, this layered understanding of manuscript production and history is significant for a study of the scribes of B(03), as it acknowledges that the codex is not simply a product, but also a process.18 Indeed, Andrist has already applied this methodology to B(03), א(01), A(02), and other biblical manuscripts.19 Although our research on the codicology of B(03) has occurred independently, I am indebted to the methodology set out in La syntaxe du codex,20 and our conclusions are largely complementary. While the complex terminology of La syntaxe du codex will often be impractical for our study, the method of looking to observable discontinuities will be pertinent in studying the 14 Gumbert, “Fifty Years,” 507. 15 Quantitative Codicology is concerned primarily with the economic context of manuscript production, unlike Andrew Smith’s quantitative analysis of Codex Alexandrinus (A(02)). Comparative Codicology is best exemplified by the recent volume, Comparative Oriental Manuscript Studies (COMSt), which devotes nearly 200 pages to codicological comparisons across ten ancient book cultures. This study reveals a significant amount of shared book practices, transcending language and culture. García, Introducción, 28; Bausi and Sokolinski, eds., Comparative; Agati, Manuscript Book, 28–40. 16 Gumbert had recognized this complexity much earlier and has also attempted to produce a coherent terminology for describing the codex as made up of “Codicological Units.” Andrist, et al., are concerned that Gumbert’s language is too ambiguous. Gumbert, “L’unité codicologique,” 4–8; Gumbert, “Codicological Units,” 17–42. 17 Andrist et al., La syntaxe, 8. 18 Friedrich and Schwarke, “Introduction,” 6. 19 Andrist, “Le structure,” 11–37. Since writing this chapter, Andrist has published an additional article with further implications for our study: Andrist, “Au croisement,” 3–106. 20 Andrist has since provided many fruitful insights into the codicology of B(03) through personal contact. CHAPTER 2 28 scribes of B(03).21 This method is especially relevant for manuscripts that contain more than one text and were copied by more than one scribe. There are many terms used to describe manuscripts with an accumulation of texts (“complex,” “miscellaneous,” or “composite”), but the term “Multiple-Text Manuscript” (MTM) has recently been proposed in an attempt to avoid ambiguity.22 B(03) certainly fits within this category, although there is some question as to the relevance of the term MTM for groups of works that were designated as canonical or “closed.”23 Yet, with the inclusion of both Greek OT and NT books, B(03) is best described as an MTM that originally involved a single Production Unit (UniProd)—“parts of a codex which are the result of the same act of production”—but now is composed of two.24 Our primary interest is in “UniProd a” since “UniProd b” was not added until the fifteenth century. This UniProd is composed of multiple phases of production, in which various discontinues arise and can be useful for identifying the work of the scribes, and scribal transitions. These discontinuities are primarily identified by various “units”—Support-Material Units (UniMat), Modular Units (UniMod), Mise en page Units (UniMep), and Hand Units (UniMain)25 . Elsewhere Maniaci has commented on the “modular structure” of the codex, in contrast to the scroll, “which made it possible to modify the original configuration by adding or subtracting leaves or quires or changing their order.”26 The following discussion will approach the modular structure of B(03) through the aforementioned units as a tool for describing the make-up of the manuscript and the physical parameters set for the scribes. 2.1.3 Practice and Limitations 21 Fundamentally, we are searching for different results in our respective studies. The authors use the analogy of “syntax” in contrast to “morphology,” because they are ultimately seeking a broader historical timeline of individual manuscripts. I am more concerned with the morphology of the codex and how it relates to the work of the scribes. Andrist et al., La syntaxe, 9. 22 Friedrich and Schwarke, “Introduction,” 1–26. 23 For example, Maniaci has argued that a codex with the four gospels and the Eusebian canons should not be considered an MTM. Sokolinski, “Conference Reports,” 88. 24 Andrist et al., La syntaxe, 59. 25 Andrist, et al. also include Ruling Units (UniRégl), Marks-of-Succession Units (UniMarq), Content Units (UniCont), and Writing Units (UniÉcri), which are distinct from Hand Units. These categories are less helpful in B(03), and do not play as significant a role in understanding the scribes. Andrist et al., La syntaxe, 83–108; For a list of related criteria, see Nyström, Containing Multitudes, 60–61. 26 Maniaci, “Medieval Codex,” 28. CODICOLOGY 29 The greatest hindrance in this examination is the lack of access to the physical manuscript. It is not uncommon to hear a codicologist claim that such a study is not possible without touching the parchment itself. However, in the case of B(03) this objection does not appear entirely valid. Indeed, there are many codicological observations that can be made through various other avenues. But in acknowledging my limitations, I note the following: (1) The measurements I give are not from the manuscript itself, but from the 1999 Vatican facsimile, which reproduces the original structure as closely as possible; (2) I will not make final judgements concerning the species of the parchment, and (3) apart from general observations, I will not make arguments based purely on the color of the parchment or inks. On the other hand, consistent use of the 1999 facsimile will allow for an internal consistency of measurements that can also be more readily checked by others. A second advantage of using the facsimile is its codex form. Unfortunately, since the preparation of two photo-typical facsimiles in 1889–1890 and 1904–1907 the codex has been unbound for the sake of preservation, with each sheet currently held in separate folders.27 The facsimile, however, is structured in the form as it was before the unbinding. In addition to the facsimile, the use of high-resolution images from the Vatican Library allows us to identify “bridge marks” across the fold of the bifolios. Gumbert classifies “bridge marks” as “spots or tears, which continued from one leaf to another,”28 but in this context I also include marginal notes and symbols that can be seen on the connected folio. Through these images I am able to check and confirm the quire arrangements found in the facsimile. 2.2 Material Any codicological analysis must begin with an examination of the materials from a variety of perspectives. Among these approaches, Canart includes the thickness of the medium used, follicle patterns, color, and material defects (ancient and modern).29 While a few of these can only be answered through an autopsy of the manuscript, we may still offer the following material description. 27 Cardinali, “Vicende Vaticane,” 399–400; cf. Andrist et al., La syntaxe, 79; Canart and Martini, Introduction, 7. 28 Gumbert, “Fifty Years,” 513–14. 29 Canart, Lezioni, 51–52; translated in Agati, Manuscript Book, 63. CHAPTER 2 30 2.2.1 Parchment B(03) was produced using parchment sheets, which are ubiquitously described as “very fine” and “delicate.” 30 Indeed, the parchment is so fine that Tischendorf describes it as transparent (perlucida),31 an indication that it was well prepared.32 As far as I am aware, there have not been any recent attempts to identify the species of animal(s) used to produce the codex. Since Tischendorf, it has often been said, with hesitation, that the animal used for both א(01) and B(03) was the antelope.33 This claim has since received criticism by Milne and Skeat, who claimed that it was likely composed of sheepskin and goatskin.”34 More recently, however, Gavin Moorhead, through follicle analysis, has concluded that א(01) came from calf and sheep skins. 35 This, of course, does not invalidate the claim that B(03) was written on antelope parchment, but the lack of DNA and microscopic testing across various regions makes it difficult to prove.36 The University of York has recently initiated the “Codex” project, in which they use non-invasive methods to test collagen molecules, through “matrix-assisted laser desorption/ionization time-of-flight mass spectrometry” (MALDI-TOF MS).37 With the aim of building a large “DNA data matrix,” they will be able to extract loose collagens from the surface of any manuscript and identify the animal species by its unique peptide mass.38 It is hopeful that an examination of this nature and a microscopic analysis of the follicle patterns would answer our questions about the parchment species of B(03). In the manuscript’s current state, B(03) has experienced discoloration and contains many stains, possibly from wax droppings and exposure to water. In relation to its original state, Canart 30 Mercati attributes some of this thinning to the history of rebinding, after its arrival in the Vatican. On Mercati’s comments, see Cardinali, “Vicende Vaticane,” 400; Birch, “Nachricht,” 140; Hug, Introduction, 1:262; Scrivener, Introduction, 106; Kenyon, Our Bible, 134; Gregory, Textkritik, 1:32–33; Gregory, Canon and Text, 343; Canart, “Notice,” 21. 31 Tischendorf, Vaticanum, xvii. 32 Canart and Martini, Introduction, 7. 33 Scrivener, Introduction, 1:106; Kenyon, Books and Readers, 86–87. 34 Milne and Skeat, Scribes, 70. 35 Moorhead, “Parchment Assessment.” 36 Recent DNA tests on the Dead Sea Scrolls have suggested the use of gazelles as a source of parchment. Gazelle parchment or leather was the preferred material, specified in most recipes, for Jewish love magic. D. W. Parry et al., “New Technological Advances,” 506; cf. Jongkind, Scribal Habits, 33 n. 17; Saar, Jewish Love Magic, 96–100. 37 For an introduction, see Liyanage and Lay, “An Introduction to MALDI‐TOF MS,” 39–60. 38 Doorn, “Zooarchaeology by Mass Spectrometry (ZooMS),” 7998–8000. For a basic description of the project, see https://www.york.ac.uk/archaeology/research/current-projects/codex/ CODICOLOGY 31 notes that “in rare cases, a skin defect, or an accident of preparation, gave rise to a small circular hole.”39 These holes are often called “maker’s holes,” but are not as rare as Canart suggests. There are at least seventy-three folios with either punctures or blemishes that directly affected the copying of the scribes.40 This does not include the numerous marginal holes, which could have developed in the parchment making process (e.g., pp. 427C/428A). The great leanness of the parchment made B(03) particularly susceptible to lacerations, some of which may have been caused by raised scar tissue that had been scraped off during preparation. The imperfections then became exaggerated as the parchment was tensioned. While the maker’s holes in א(01) are mostly located in the margins and appear less frequently,41 B(03) can contain up to four in a single column (p. 667B, p. 669A). On pp. 287C/288A there is one hole which has affected the copying of six lines (measuring 27mm at its largest angle). In order to handle this imperfection, the scribe terminated the lines early and added 19 line-fillers to cover the remaining space (Figure 1). FIGURE 1: PARCHMENT IMPERFECTION (P. 287C) The most significant imperfection for our study occurs on pp. 69A/70C. This maker’s hole only affects two lines of copying, in which the scribe has chosen to terminate the lines early. On p. 69A l. 31, the scribe abbreviates the word πατροϲ as a nomen sacrum (π̅ρϲ̅;̅ Exod 18:4). Of the twenty-three occurrences of the noun πατηρ in Exodus (all with human referents), this is the only incident of abbreviation. Instead of splitting the word on both sides of the hole, as is the common practice of the scribes, this is a prime example of how the physical imperfections of B(03) influenced the copying of the scribe. We will discuss the significance of these abbreviations in the following 39 Canart, “Notice,” 21. 40 On the various ways scribes handled pre-existing imperfections on papyrus and parchment, see Jones, “Avoiding Imperfections,” 371–83. 41 Moorhead, “Parchment Assessment.” On the parchment quality of א(01), see also Pattie, “The Creation,” 64; Jongkind, Scribal Habits, 34. CHAPTER 2 32 chapter. Finally, it should be noted that there is no obvious attempt to camouflage these imperfections by placing blemished leaves closer to the center of the quires.42 2.2.2 Inks It is difficult to discuss the types of ink used in copying B(03), since much of the manuscript was re- inked at a later date. There have been a variety of dates proposed for this reinking, but it is commonly placed in the tenth or eleventh century.43 The light ochre ink of the reinforcement is slightly darker than that of the original writing, though it can often be difficult to distinguish the hands based on color alone (Figure 2, p. 1492B).44 This challenge will feature heavily in our final two chapters on the corrections in B(03). On top of this reinforced text lies a third layer of ink added in the sixteenth century.45 This fresh reinking is best recognized by its dark-black ink and minuscule writing (e.g., p. 1257A). To these two layers we might add a third hand who occasionally reinked portions of text in an upright Ogival Majuscule hand (pp. 1090B, 1208A, 1289B). Versace dates this hand to the ninth century.46 The reinkings in B(03) cause problems for anyone who wishes to examine the handwriting and ink of the original scribes. However, in Chapter 3, we will find that detailed analysis of the codex presents numerous instances of unreinforced writing, which can range from individual letters (most often the final nu) to multiple lines (often because of dittography). FIGURE 2: TWO REINKERS (PP. 221B, 1492B) When spotted, the original ink color can be described as light brown or apricot.47 It should be noted that any attempt at visually comparing the ink colors is highly speculative, since, as Mark Clarke 42 cf. Agati, Manuscript Book, 68. 43 For this date, see Versace, Marginalia, 43–51, 78, 258–268; cf. Westcott and Hort, Introduction, 270; Canart and Martini, Introduction, 8. 44 Versace, who has personally examined the manuscript, describes the ink as “ocra chiaro.” Versace, Marginalia, 43. 45 This third hand was already recognized by Hug, Introduction, 1:263. On the date, see Versace, Marginalia, 67–68. 46 Versace, Marginalia, 28–31 (especially 31). 47 Payne and Canart, “The Originality,” 107. CODICOLOGY 33 explains, “there are many examples of pigments that share a color and appearance, but which are in fact chemically different.”48 While most appearances of the original ink are easily seen, there are many places where only faint traces are left. There are cases of early and late erasures in B(03), but the fading of ink is the primary motivation for re-inking.49 Along with the apricot ink, there is also the appearance of red ink for Greek section numerals (only in Isaiah through the Gospels) and occasionally for paragraphoi. Whether these were original to the scribes of the main text will have to be explored later. Finally, after the twelfth century, red, blue, and green inks were added as banners to ornament the beginning of each book in the codex.50 Versace has dated the banners to the sixteenth century, as the same hand appears to have reinked some text over the sixteenth- century reinforcement (p. 138A l.1).51 As for the type of inks used, it is not possible to give a definite answer without testing the chemical compounds. Nevertheless, Canart has noted that “the overly acidic ink has more than once gnawed, even punctured the parchment.”52 It is not entirely clear whether this corrosive effect was caused primarily by the original ink or the later reinkings. Still, the original ink was most likely a metallic ink. The presence of metals in the ink (usually iron(II) sulfate) often catalyzes lipid peroxidation in parchment, leading to deterioration in the area where ink has been applied.53 This conclusion finds support in the historical transition from using carbon inks on papyri to metallic inks (e.g. iron-gall) on parchment, like in א (01) and A(02).54 As for the additions in red, the ink was possibly made from lead oxide (Pb3O4) known as minium or red lead.55 This is recognized best at 48 Clarke, “Manuscript Pigments,” 39. 49 Metzger, Manuscripts, 74. 50 Canart provides a terminus post quem of the twelfth century since the opening banner of Daniel is displaced on account of the late title added in a twelfth-century hand. Canart, “Notice,” 33. 51 Skeat, however, dates the banners to the fifteenth century as a part of the restoration of the codex, including the addition of the supplementary leaves in Genesis, Psalms, Hebrews, and Revelation. Versace, Marginalia, 71–73; Skeat, “Vaticanus,” 125–126. 52 Canart, “Notice,” 21. 53 Skeat claims that it is the production of sulfuric acid which causes this deterioration. However, Gerhard Banik notes that this deterioration is far more likely caused by oxidation than acids. In fact, Christopher Woods claims that “parchment appears to be more resilient to this acidity than paper.” This does not disqualify the notion that multiple factors are at play in this kind of erosion. Skeat, “Book-Production,” 40; Banik, “Ink Corrosion,” https://irongallink.org/igi_index22a4.html; Woods, “Conservation,” 206; cf. Florian, “T Deterioration,” 40. 54 Woods, “Conservation,” 206; Milne and Skeat, Scribes, 79; Smith, Alexandrinus, 40. 55 Gumbert, Words for Codices, 21. CHAPTER 2 34 places where the orange-red has darkened almost to black (see να̅̅; p. 1244a).56 Without further testing, these observations can only provide suggestions as to the likely material used in producing B(03). However, it is important to appreciate the influence that these materials, particularly the parchment, had on the scribes. FIGURE 3: ERASURE (P. 626B) AND POSSIBLE INK DAMAGE (P. 1381B) 2.3 Blueprint of the Codex This parchment was used to create a multi-quire codex, constructed by stacking five large-sheets and folding them down the center to create a natural quire of ten folios (twenty pages). This five- sheet format (quinions) is the governing configuration, apart from a few instances that reflect either discontinuity in the original production or inconsistency on the part of later rebinders (see Quire Irregularities). The use of quinions was somewhat rare, especially as the quaternion became the standard format in both the Greek and Latin traditions.57 In B(03), as in most codices, the sheets in the quires are arranged according to “Gregory’s Law,” that is, beginning on the flesh side and alternating so that like surface faces like (i.e., flesh/flesh and hair/hair).58 The following discussion will examine how these quires were constructed and prepared for writing. At each point, we will discuss the observable discontinuities, which may or may not help us identify patterns of the individual scribes. 2.3.1 Quires and Folios We begin by discussing the quires of B(03), since they are rightly described as “the basic constitutive unit of the codex.”59 As noted earlier, the modular nature of the codex allows for any quire or group of quires to exist independently of the others. Conversely, groupings that were originally independent can later be joined to others and form a multiple-text volume. This inevitably means that any attempt to describe the quire structures of B(03) must not assume that the condition in 56 Clemens and Graham, Introduction, 25; cf. Smith, Alexandrinus, 39. 57 Turner gives examples of eight quinion papyrus codices and three parchment codices (including B[03]). Other important biblical manuscripts made of mostly quinions include the purple codices Rossanensis Σ(042) and Petropolitanus N(022), as well as Codex Marchalianus (LXX Prophets). Turner, The Typology, 64; Parker, Codex Bezae, 8; Maniaci et al., “Codicology,” 196; Scrivener, Introduction, 1:105; On the codicology of the purple codices, see Hixson, Scribal Habits, 526–530. On whether or not P47 was composed of quinions, see Malik, P.Beatty III, 28–30. 58 Canart, “Notice,” 21; Gregory, “The Quires,” 27–32. 59 Maniaci et al., “Codicology,” 78; cf. Gumbert, “C Catalogue,” 61; Andrist et al., La syntaxe, 50. CODICOLOGY 35 which the codex appeared in the Vatican Library was the same as the original production. 60 However, the more significant observation is that by using the multi-quire format, several scribes were able to copy a codex simultaneously.61 Therefore, one area of interest will be quire boundaries as a potential signal of scribal transition, especially when these boundaries align with other observable discontinuities in mise en page.62 Before B(03) was last unbound, the Vatican Library possessed 733 folios (pp. 41–694; 707– 1518) of the original manuscript contained in 76 quires. However, following the medieval quire numbers, written in the bottom right-hand corner of the page, Patrick Andrist believes that these folios originally involved 74 quires. 63 In his opinion, this discrepancy was likely caused by the rebinding process, in which some loose folios were brought together to create artificial bifolios.64 We will discuss shortly whether or not there is sufficient evidence of this sort of operation. Today, there are an additional 35 folios of supplementary writing that accompany the original 733, making a current total of 768 folios or 1536 pages.65 Most handbooks erroneously give a folio count of 759, because they do not include the 9 supplementary folios that follow Hebrews 9:14.66 This number is problematic because it excludes the last supplementary folios, yet includes the first 20 supplementary folios (Gen 1–46:28) and the 6 supplementary folios covering Psalms 105:27–137:6. The proper number of extant original folios is 591 (617 with supplements) in the Greek OT and 142 (151 with supplements) in the NT. In addition, Canart mentions that there are two extra folios numbered I–IV at the beginning of the codex; one unnumbered folio that completes the final quire, and six unnumbered guard folios that were likely at the beginning of B(03), when it was bound.67 60 Andrist, “Physiognomy,” 551. 61 Andrist et al., La syntaxe, 50. 62 This does not necessitate, however, that scribes only copy separate quires. See Jongkind, “One Codex,” 121– 36. 63 Andrist, “La structure,” 12. 64 Patrick has shared this opinion with me through personal contact. For the distinction between natural and artificial bifolios (with two types), see Andrist et al., La syntaxe, 49. 65 Canart and Martini, Introduction, 7. 66 Scrivener, Introduction, 1:106; Gregory, Textkritik, 1:32; Swete, An Introduction, 127; Hatch, Manuscripts, plate XIV; Finegan, Manuscripts, 127; Metzger, Manuscripts, 74. 67 The original location of these guard folios is uncertain, as they were not numbered when the manuscript was dismembered. For this reason, they were randomly placed in the facsimile. The folios can be seen in images 1543– 1554 of the online edition. Canart, “Notice,” 19; Canart and Martini, Introduction, 7. Andrist et al., La syntaxe, 79. CHAPTER 2 36 Canart correctly notes that “despite the indications of the facsimile of 1904–07, we do not find trace of ancient numeration of the quires, conforming to the ancient composition.”68 There are, however, two later additions of quire numbers to the codex. The most commonly spotted quire numbers are those found in the bottom right-hand corner of the quire’s first page. These are written as both Greek and Roman/Arabic hybrid numerals,69 and the first visible occurrence, marked as quire four [δ̅], is found on p. 53. As we will determine later, this numeration is behind by one when compared to the original structure of B(03) (i.e., quire [5]).70 Turner, on the other hand, mentions a different set of quire signatures that occasionally appear at the “center of lower margin.”71 These numbers are clearly not original as they are written in Arabic numerals (e.g., p. 1505; [οζ̅]̅; [78]). It is not evident, whether these numbers were present throughout the entire codex, since most of them were cropped when the manuscript was trimmed for rebinding. This must mean, however, that they were added at some point before the binding process.72 Regardless, neither set of quire signatures has any claim to originality, nor do they provide anything more than secondary evidence for the earliest structure of B(03). Foliation There is, however, one piece of evidence that helps provide insight into the ancient structure of the codex: the earlier foliation. This set of Greek numerals is located on the upper left-hand corner of the verso, though most of them have been cut by bookbinders and many others have been damaged by deterioration.73 The location of these numbers on the verso is quite rare in Greek manuscripts, leading Gregory to declare that they were “not from a Greek but a Semite,” because this is “where 68 In the upper margins of the facsimile, you will see ancient pagination written when visible, or supplied in brackets when reconstructed. The beginning of each quire is also marked with Greek numerals, but only in brackets. Canart, “Notice,” 20. 69 This kind of hybridization happened frequently during the transition from using Roman numerals to Arabic numerals, in the West. Crosby, The Measure of Reality, 115. 70 Here I use brackets with the Greek numeral to note the number written in the bottom corner of the page, while brackets with Arabic numerals are used for the reconstructed quire number. The number given also follows the reconstruction of Canart and Andrist. Canart, “Notice,” 21; Andrist, “La structure,” 20; Andrist, “Au croisement,” 22. 71 Turner, The Typology, 78. 72 It might be possible that these numbers were added to assist the binders and were intended to be cut out, after the quires were placed in their proper order. 73 Canart, “Notice,” 19; Canart and Martini, Introduction, 7. CODICOLOGY 37 the Semites put their numbers.”74 This explanation seems unlikely, since it is in reference to codices that were read from right to left. However, Andrist and Maniaci have both suggested that this position is evidence that the numbers are not for foliation, but for numbering the openings.75 For Andrist, this means that folio one verso (f.1v) might have been marked “2” instead of “1,” and any reconstruction of the quantity of missing folios from the surviving numerals would have to be one less. While this may be the case, it is not always true that manuscripts with numbered openings mark f.1v with “2.” One late example is the Vernon manuscript (Oxford, Bodleian Library, MS Eng. poet. a. 1), which, based on its table of contents, has been shown to indicate openings on the verso with red Roman numerals.76 However, rather than being one numeral ahead in sequence of the folio, it is actually matching (e.g., f.104v is marked CIIII). Wendy Scase, who led the Vernon Manuscript Project, notes that we have likely overlooked manuscripts using opening numbers, because we do not have the table of contents to aid in the distinction from folio numbers.77 It remains uncertain, therefore, if B(03) contains opening numbers, which are either matching the folio number or ahead by one. For the sake of calculation, we will follow the opening numbers as equivalent to the folios, though the evaluations can be modified to fit the alternative numeration (- 1 folio). It is unlikely that these numerals have any claim to originality in the codex.78 Although Versace does not appear to mention the opening numbers, the hand more consistently betrays the distinctive features of the tenth/eleventh-century reinker.79 Therefore, the opening numbers only provide evidence of the structure of B(03), as it was around six centuries later. According to the marginal notations of the 1904–1907 facsimile, the first fully visible number is ΛΗ on p. 52. However, it was not visible to me in either this facsimile or the 1999 facsimile (including online images). While there are many remnants of the Greek numerals, the first complete example I have found is on p. 74 Gregory, Canon and Text, 344. 75 Andrist notes that this “is found in certain western medieval manuscripts,” but does not give any examples. Andrist, “La structure,” 16; Maniaci et al., “Codicology,” 199. 76 Scase, “‘Looke This Calender’,” 294. 77 Scase, “‘Looke This Calender’,” 295. 78 Maniaci et al., “Codicology,” 199. Canart and Martini, Introduction, 7. 79 Versace terms this hand B18. See more on Versace’s classification in Chapter 4; cf. Versace, Marginalia, 43– 50. CHAPTER 2 38 460, reading CΜΑ (241). By dividing the page number in half (230) we can identify an eleven-folio disparity—or ten folios, if we accept Andrist’s conclusion—between the original folio quantity and the supplementary material. Therefore, in returning to the beginning of the codex, we can calculate the number of folios missing before p. 41, by the following formula: 𝑥 = !" + 11a x= ancient opening number/number of folios y= every even page number a= -1 folio (Andrist) #$" + 11 = 31a FIGURE 4: CALCULATING MISSING FOLIOS This same equation can be used to reconstruct missing opening numbers, up to the supplementary material in the Psalter (pp.695–706). After these additional folios, the first visible number is Τ̣ϤΕ (395) on p. 760. By dividing 760 in half (380) we now notice a fifteen-folio difference between the current pagination and the original. Since the missing portion of the Psalms was appended with a ternion (three sheets/six folios) the earlier disparity grew by four folios (two sheets). The same content was originally contained in one quinion and was probably lost at a point when the manuscript was unbound. Therefore, the new formula for reconstructing the original opening numbers (after p. 706) is as follows: 𝑥 = !" + 15a x= ancient opening number/number of folios y= every even page number a= -1 folio (Andrist) FIGURE 5: CALCULATING FOLIOS AFTER PSALMS (P. 706) The last original page of B(03) is p. 1518, where the text ends at Hebrews 9:14. Following the above equation, this would make it originally f. 774v or p. 1548. This conclusion can be confirmed by counting folios from the last visible opening number ΨΝΘ (759), on p. 1488. How many folios did B(03) originally contain, then? The answer is altogether dependent upon which books were contained in the last folios of the codex. Elliott claims that B(03) originally had a minimum of 1600 pages (80 quinions).80 If the current form of B(03) contains 774 folios (equal to 77.4 quinions), then Elliott estimates the rest of the codex comprised around two and a half quires. This number is given from the belief that the manuscript is “virtually complete,” and 80 In comparison, he estimates that א(01) had 1460 pages and A(02) had 1640. The estimation of 80 quinions was already made in 1860 by Carlo Vercellone. However, he incorrectly claims that there are 73 original quires, rather than the 74 that Andrist calculates. Vercellone, Dell’ antichissimo, 12 n. 1; Elliott, “Manuscripts,” 618. CODICOLOGY 39 originally contained the Pastorals Epistles, Philemon, and Revelation. 81 Since we possess the supplementary leaves of Hebrews (after 9:14) and Revelation, we have two ways to test this estimation. First, we can use the ratio of original folios to supplementary folios in Genesis (i.e., 31/20 or 1.55) as a way of estimating how many pages were used for Hebrews and Revelation.82 Second, we can test this outcome by using the average number of characters-per-page (2134 characters)83 and comparing this to the number of characters in Hebrews 9:14b–13:25 and Revelation.84 This final step will also be used to estimate how many pages were needed to contain the Pastorals and Philemon. TABLE 5: CALCULATING THE MISSING PAGES AFTER HEBREWS First Estimation Second Estimation Hebrews 9:14bff. 1.55 × 2 = 3.1 folios 6.2 pages 11698 ÷ 2134 = 𝟓. 𝟒𝟖 pages Revelation 1.55 × 7 = 10.85 folios 21.7 pages 47388 ÷ 2134 = 𝟐𝟐. 𝟐 pages Pastorals and Philemon X 21409 ÷ 2134 = 𝟏𝟎. 𝟎𝟑 pages Totals 27.9 pages 1575.9 pages (without Pastorals and Philemon) 27.68 pages 1575.68 pages (without Pastorals and Philemon) 37.71 pages 1585.71 pages (with Pastorals and Philemon) While these numbers are only estimations, it seems unlikely that B(03) would have needed as many as 1600 pages. As we will discover in our discussion of quire discontinuities, there are at least three folios that interfere with the pattern of ten-folio quires. This means that even if B(03) was originally composed of 80 quires, its total number of pages would be, at most, 1594. One may conclude that more folios were required if, like א(01) and A(02), a certain number of the Apostolic Fathers’ texts were also included in the codex (see Chapter 1). 81 Elliott, “Manuscripts,” 628. 82 The supplementary leaves in Psalms are copied in a three-column format rather than two. This factor significantly affects the ratio of supplementary folios to original folios (10/5 or 2). 83 This number is averaged from the NTVMR transcriptions of pp. 1514–1518 (character count: 2162; 2205; 2150; 2254; 2229). While the actual average of these numbers is 2200, I noticed that the SBL Greek text gave an average of 3% less characters-per-page of B(03) than the transcriptions. For the purpose of comparison with the SBLGNT, the average character count is 2134 (2200 × .97). 84 These numbers are taken from the SBLGNT. The figures include punctuation; much of which are accounted for by the middle dots in the B(03) transcription. CHAPTER 2 40 Quire Irregularities As I have mentioned already, the modular nature of B(03) and its history of unbinding and rebinding make it particularly difficult to identify the manuscript’s original structure from the work of later rebinders.85 That the codex had its own history in an unbound state, is seen most clearly in pp. 1133–1136, the center sheet of quire [59]. This sheet immediately stands out for its small dimensions and the considerable amount of faded text occurring on its outside pages (pp. 1133, 1136). This, of course, can happen to any folio, yet it is surprising to find in the center of a quire (the most protected of the sheets), while the remaining folios are well preserved. In the inner pages of the sheet (pp. 1134–1135), we find Latin text in the lower and upper margins: quaere supra folio praecedens (p. 1134) Reponendum post sequens folium (p. 1135) It is clear, then, that this sheet was once separated from its quire, slightly damaged, and later returned to quire [59]. Cardinali believes Pedro Chacón was responsible for relocating this sheet and copying the marginal notes in the sixteenth century.86 While fortunately this sheet was never lost, and can easily be returned to its original positioning, there are three other irregularities deserving our attention. Here, we are most concerned with how these discontinuities might give insight into the work of the scribes. The first two examples are identical in nature and happen in the first and last extant quires of B(03) (pp. 41–54; [4]; pp. 1505–1518; [78]). As Canart notes, these two quires only contain seven folios each. According to his reconstruction, quire [4] would have been a quinion, containing all of folios 29–38. Thus, there would have likely existed three quires before it, two quinions and one quaternion (28 folios).87 How Canart is able to know that quire [4] was once a quinion, and not the quaternion in his reconstruction, is unclear. What can be suggested, based on the number of original folios (38), is that quires [1–4] possibly contained three quinions and one quaternion. Thus, it seems equally probable that the odd number of folios in quire [4] is the result of one folio being cut from a quaternion. In the end, however, neither reconstruction is able to exclude the possibility that there 85 For the various bindings of B(03), see Cardinali, “Vicende Vaticane,” 394–400. 86 Similar notes are present in Deuteronomy, in the lower margins of pp. 227–228. Cardinali, Pedro Chacón, 63. 87 Canart, “Notice,” 20. CODICOLOGY 41 also existed two unnumbered folios at the beginning of B(03), resulting in four complete quinions up to folio [38].88 Upon further investigation into the first and last quires (as it was before unbinding), I noticed that these were no longer one quire each, but two. In both of these instances, the cut folio was not found adjacent to the lacuna, but in between the two quires of two sheets (binions) (Figure 6). Therefore, what is necessary for both my reconstruction and Canart’s is that all folios from pp. 41–54 must have been separated from their bifolio, and artificially bound to a different folio, at a later time. The likelihood of this is difficult to gauge without examining all of the folds in person since the images do not provide indications of such procedure on every bifolio. However, one is able to see evidence on pp. 49/50–51/52, 1507/8–1509/10, 1511/12, and 1517/1518 that loose folios have been artificially attached. Nonetheless, while quires [4] and [78] do represent irregularity in the common structure of B(03), this is more easily attributed to the damage occurring at the beginning and end of the codex, than to the intentional work of the scribes.89 Current Structure (with Artificial Attachments) FIGURE 6: THE FIRST AND LAST QUIRES The third irregularity is significant, because it occurs at the same location that Milne and Skeat detect a change in hands, from scribe A to B (at p. 945, Hos 1:1).90 According to Canart, quire [49] (pp. 927–944) was a quinion whose last folio was left blank and cut out.91 Thus, after the nine-folio quire, Hosea begins the whole of the Prophets on a new quire (i.e. [50]). As Andrist points out, this 88 Andrist, “La structure,” 16 n. 13. 89 It seems highly unlikely that quires [4] and [78] were originally two binions each and that the only two examples of this structure occur at the place where the codex has lost adjacent quires and contains several artificially combined folios. 90 Milne and Skeat, Scribes, 88. 91 Canart, “Notice,” 20. CHAPTER 2 42 marks a break between the two “modular units” in B(03) — where a complete text and quire end at the same location.92 The agreement between the observations of Milne and Skeat and those of Andrist, indicates a plausible place of scribal transition at Hosea 1:1. There is, however, one problem that is not acknowledged: Canart’s description does not match the manuscript’s current arrangement. Instead, quire [49] appears as a standard quinion, with the first folio of Hosea (pp. 945–46) acting as the closing folio of the quire. Following this arrangement, we find four irregular quires [50–53] that alternate between quaternion and senions (six sheets). The last of these quires, while made of six sheets, only contains eleven folios (see Figure 7 and Appendix). FIGURE 7: QUIRES [49–53]—CURRENT STRUCTURE AND PROBABLE ORIGINAL STRUCTURE An immediate clue to the original arrangement of these gatherings might be the quire numbers that were mentioned earlier. According to the current structure, each quire number is off, and they only become accurate again after this erratic set of quires. However, if the quires followed Canart’s reconstruction, then all the numberings would be properly placed. While this appears to be an easy solution, these numberings should not be perceived as without error. On both p. 53 and p. 1383 the quire numbers appear one folio early. As we discussed earlier, the numbers are also lagging behind by one (e.g. [δ]=[5]), better reflecting the structure after it was rebound than before it.93 The clearest solution to this problem must be Gregory’s Law. In many cases, the parchment has been scraped and pumiced so fine that it is difficult to distinguish the flesh from the hair side. However, there is enough from the images to label each side and to deduce the original arrangement (see Figure 7; f= Flesh and h= Hair). According to the current organization, only one of the five 92 The other break in modular units is found between Jude and Romans (p. 1445). There is also a break between 2 and 3 Kingdoms, but this may be more coincidence than intentional, since the books are combined as 1–4 Kingdoms. Andrist, “La structure,” 16–17. 93 It is likely that this irregularity was created after the quire signatures were added. CODICOLOGY 43 quires follows Gregory’s Law, by starting with the flesh side. While it is not impossible for a quire to begin on the hair side, it seems unlikely that it would have originally happened here in such a high concentration. In addition, we may also identify which folio was originally cut by finding the location in which the hair and flesh side meet. We find this abnormality at pp. 944–945, where Hosea begins. Thus, we may agree with Canart that the cut folio was originally attached to pp. 927/928. 94 Consequently, on codicological grounds, Milne and Skeat may also be correct in identifying the work of two scribes at this location. Andrist has advanced the argument for a potential two-volume structure of B(03), with the second half beginning at Hosea, the opening of the Prophets.95 The apparently intentional trimming of the last folio in the previous quire certainly seems to indicate a possible ending to a first volume, and the result of such procedure would leave two codices of a similar thickness. There is, however, one clue that has been overlooked. Unlike the rest of the books found within B(03), the first page of Hosea is missing an original title. Instead, the later reinker added his or her own title above the first column (see Figure 8). Instead, what is present on the first page of Hosea, unlike the other books, is a running title. This is significant, because the running titles in B(03) are reserved for the second recto page of each book and are never found on the title page (see Chapter 3).96 What might this indicate? Canart describes the folio before Hosea as left blank and subsequently cut from its quire.97 Is it possible, however, that the folio was originally intended to begin the book of Hosea? Dirk Jongkind has argued that in א(01) the title to 1 Maccabees may have been added before the main text was copied, rather than later. 98 Likewise, it seems probable that quire [50] was prepared beforehand to receive the second, and not the first, folio of writing for the book of Hosea. This would best explain the presence of a running title in the place of an opening title. 94 Andrist has recently confirmed this conclusion through examination of the folios in the BAV. He presented this conclusion in a paper given at iSBL in Rome, 2019. 95 Andrist, on the other, does not assume B(03) ever ended up as two volumes; only that the structure would allow for this possibility, even after production. Andrist, “Au croisement,” 17–22. 96 Exceptions to this are found when the running titles are split between the verso and recto of an opening (e.g., κατα] [µαθθαιον on pp. 1236–37). We will discuss the various patterns of running titles and their relationship to production in the following chapter. 97 Canart, “Notice,” 20. 98 Jongkind, Scribal Habits, 42–43. CHAPTER 2 44 FIGURE 8: THE TITLES OF HOSEA AND AMOS (PP. 945, 954) The evidence of this running title does not necessarily disprove Andrist’s theory, but rather pushes the editorial decision for a two-volume structure after the quire had been prepared for copying. Still, the likely change of scribes at Hosea and their simultaneous copying may well have been enough reason to begin a new quire, at the cost of one spare folio. In Milne and Skeat’s paradigm, only the change of scribes at Psalm 1 (from B to A) does not coincide with a quire break. If they are correct, this is the only example of two scribes working successively on the same quire, with the transition taking place at the half-quire mark.99 While we have spent most of our energy looking at quire irregularities in the codex, the overall picture that one gets from B(03) is one of great care and consistency. Most of these inconsistencies cannot be blamed on the scribes themselves, but on those who were forced to put back the pieces, when it was later rebound. Nevertheless, there are additional ways that we may examine the structure of the codex. 2.3.2 Pages, Columns, and Lines Canart and Martini describe B(03) as “practically square” in its current state, with the dimensions of 270 x 270 mm..100 This basic description is also given by Turner, who classifies the codex in his Group III, “Large, ‘Square’” parchment codices.101 For this reason, most measurements are given as identical in both breadth and height (see Table 6). We have already mentioned that B(03) was trimmed for rebinding, making it difficult to know much about the original dimension of the codex. As it remains today, B(03) is smaller in breadth and height than א(01), A(02), and Codex Ephraemi Rescriptus (C[04]), but larger than Codex Bezae (D[05] ).102 99 Bogaert shows that the Poetic and Wisdom books, Prophets, and NT all begin on either a new quire or folio six of a quire. Therefore, it is possible that the half-quire was also an important marker in the production of B(03). Bogaert, “Le «Vaticanus»,” 145. 100 Canart and Martini, Introduction, 7. 101 Turner, The Typology, 27. 102 א(01) is 38.1 X 34.5 cm.; A(02) is 32 X 26.3 cm.; C(04) is 31.4–32.5 x 25.6–26.6 cm.; D(05) is 26 X 21 cm.. Metzger, Manuscripts, 76; 86; 88; Hatch, Manuscripts, plate XX. CODICOLOGY 45 TABLE 6: PROPOSED DIMENSIONS OF B(03) Canart & Martini Average 270 x 270 mm. (27 x 27 cm.) Scrivener 10 x 10.5 inches (25.4 x 26.7 cm.) Hatch 27–28 cm. x 27–28 cm. Kenyon 10 x 10.5 inches (25.4 x 26.7 cm.) Aland & Aland 27 x 27 cm Metzger 27.5 x 27.5 cm. Finegan 10.8 x 10.8 inches (27.4 x 27.4 cm.) Turner 25.4 x 26.7 cm.; 27 x 27 cm. Each page of the codex is copied out in three columns,103 with the exception of the two-column format that is found in the Poetic books. On p. 893 we find a single example of the hybrid two and three-column format. Along with this example, Andrist identifies the common practice of starting a new book on the next available column, rather than on a fresh page. However, he locates four discontinuities, where one or more columns are left blank in order to start a new book on the following page: between (1) 2 Esdras and Psalms (p. 624), (2) Tobit and the Book of Twelve (p. 944), (3) Daniel and Matthew (p. 1234), and between (4) Mark and Luke (p. 1303).104 Knowing this, it is important to note that two of the three changes in scribes, according to Milne and Skeat, align with these discontinuities (i.e. (1) and (2)). Most handbooks give the number of lines-per-column as forty to forty-four, with around sixteen to eighteen letters-per-line,105 and some clarify that the NT consists of forty-two lines-per- column.106 This range, however, gives the impression that the lining is more haphazard than it really is. Scrivener gives the most precise description by noting that the standard number of lines is forty- two, with two exceptions: (1) Genesis–1 Kingdoms 19:11 is copied on forty-four lines-per-column; (2) pp. 535–554 (quire [29]) are copied with forty lines-per-column.107 The Poetic books are also copied in columns of forty-two lines, where the scribe(s) clearly avoided transgressing the designated 103 Hug, followed by Tregelles, describes the opening of B(03), with six narrow columns on the two pages, as mimicking the format of a book-roll. Hug, Introduction, 1:263; Tregelles, An Introduction, 160. 104 Andrist, “La structure,” 18. 105 Exceptions can be found where genealogies or lists are given. See also the compressed lines at the end of Philippians, which contain up to twenty letters-per-line (p. 1502B). 106 Gregory, Canon and Text, 343; Hatch, Manuscripts, plate XIV; Canart and Martini, Introduction, 8; Metzger, Manuscripts, 74. See also, Harris’ discussion of line-counts in Autographs, 1–52. 107 Scrivener, Introduction, 1: 107; Andrist, “Au croisement,” 22. CHAPTER 2 46 rulings. This is clear from the placement of διαψαλµα in the Psalms. Throughout the Psalter, the word διαψαλµα is always given its own line. The exception to this rule is found in instances where the scribe is forced to decide whether to add the word to a new column, or to improvise a new line. Instead, the scribe(s) chose to copy the word on the previous text-line, while leaving a space of about three letters in-between (e.g., pp. 627A, 664A). The overall consistency of the line count in B(03), stands in contrast to the Gospels in A(02).108 However, we should also acknowledge the inconsistencies mentioned by Scrivener. The first of these is found between pp. 334–335, where the line count goes from forty-four to forty-two. This shift also occurs at a break in quires [21] and [22], suggesting that the gatherings were lined separately before copying. What is most striking, however, is that this is also the location where Abbot, followed by Milne and Skeat, recognizes the first change in scribes (scribe A to B); this identification is accomplished without reference to the quire-break or line-count. 109 These two factors help sustain their conclusion. As we will find in the following chapter, a scribal transition at 1 Kingdoms 11:19 has the most support from both codicological and paratextual evidence. Nonetheless, the appeal to line-count cannot be the sole factor in identifying the work of scribes since most of B(03) was copied in forty-two lines. Likewise, a second inconsistency found in quire [29] (40 lines-per-column), is best explained as an anomaly in preparation, as the quires were lined beforehand and not necessarily by the scribes of the text.110 Since the scribes of B(03) were careful to follow the format of the prepared quire, it is the pricking and lining that played an important role in the copying of the codex. According to Maniaci, quires were usually prepared by pricking and then, following these marks, the folios were ruled. While the pricking in B(03) is found inside the outer columns of the folio, later manuscripts placed these marks near the edge of the page.111 Both types of pricking can be seen in B(03), since the original pages are pricked in the outer columns (see p. 624A),112 but the supplementary quires show 108 Smith, Alexandrinus, 50. 109 Milne and Skeat, Scribes, 87; Abbot, “Antiquity,” 189–200. 110 Gregory notes the “lines drawn in the parchment…probably betray the hands of different workmen.” Gregory, Canon and Text, 344. 111 Some portions of א(01) also contain this style of pricking. Maniaci et al., “Codicology,” 197. 112 On p. 626A l. 42, the scribe left an abnormal space in the middle of the word κοι_λιαϲ to avoid copying over the pricking. CODICOLOGY 47 holes in the outer margins (see p. 2). Not only do we find pricking as guides for the horizontal lines, but Canart identifies the presence of pricking to guide the vertical lines as well (see the top and bottom margins of p. 625).113 There is a third set of prickings, found only in the Pauline letters, but they do not function as “ruling pricking.” These holes are found in the outer margin, and can be seen clearest, working backwards from p. 1518, until they dissipate in Romans. It is unclear what these holes are for, since the usual prickings are still found in these pages (e.g., p. 1510), but their presence in only the Pauline letters may suggest an intention to group the quires together. After the folios were pricked, vertical and horizontal lines were drawn, likely on the flesh side with a dry-point pen. 114 The standard ruling in B(03) is with six vertical “bounding lines,” demarcating three columns, and eight “bounding lines” in the Poetic books. In the case of the eight vertical lines, four are used to outline each of the two columns, providing a wide boundary and a narrow boundary for poetic structuring of the text (Figure 9). Prose Poetry Hybrid (p. 893) FIGURE 9: COLUMN BOUNDING-LINES While the bounding lines are both easier to see and more consistently drawn, there is much more variation in the patterns of horizontal text-lines. Within the discipline of codicology and manuscript cataloguing, there have been multiple attempts to provide a thorough coding that can describe all the physical features of the manuscript in “as few words as possible.”115 This coding usually consists of a mixture of letters and numbers to describe the type and amount of lining in a codex. Canart, however, has shown the difficulty in describing B(03) with the existing systems of coding.116 The 113 Canart, “Notice,” 21. 114 Canart and Martini, Introduction, 8. 115 Agati, Manuscript Book, 200. 116 Canart, “Un X,” 53–59. A B C A B A C CHAPTER 2 48 trouble arises from the alternating of two written lines-per-text-line (this configuration is designated with an “X” in Leroy’s coding) to one written line-per-text-line. In his chapter on B(03), Canart summarizes his findings with a helpful table of the variety of ruling schemes.117 While the configuration of the rulings might vary frequently, there is surprising consistency in the overall dimensions of the lining. Below, I provide a table of the standard measurements for folios in both column formats:118 FIGURE 10: RULING DIMENSIONS While we have already noted that the scribes were not likely responsible for these rulings, it is important to recognize the quality of this production, as well as the influence that the prepared quire had on the scribes of B(03). Commencing above the first line,119 the scribes were consistent in only copying the number of lines that were traced for them. In the following chapter, we will examine further how the scribes handled word divisions and the columns’ bounding lines. 2.4 Summary In many ways, this chapter has set out a particular scope for considering B(03). That is, by understanding the codex through structural analysis, we are able to look at each individual feature, whether physical or paratextual, and appreciate them in their respective layers. While this can at times complicate matters further, it also allows us to avoid conflating qualities that developed in 117 Canart, “Notice,” 31. 118 The outer margins were affected by trimming. 119 Kerr places the eventual transition from writing above the line to below the line in the thirteenth century in England. Kerr, “‘From ‘above top line’,” 13–16. A B C? ? ? ? 4mm 175-179mm 53-54mm 70-71mm 17-18mm 193-198mm A B? ? ? ? 4mm 175-179mm 57mm 88mm 17-18mm 194-195mm 15-16mm CODICOLOGY 49 the codex through a variety of production processes. Indeed, even portions of B(03) that emerged before the manuscript was finished exhibit diverse threads of manufacturing and the multifaceted nature of manuscript production. For this reason, we began our study of the scribes of B(03) with a structural archeology of the book. This is done primarily by looking at the observable discontinuities in the codex. However, as we discovered, there are multiple factors which complicate this process. One of these factors is the missing folios at the beginning, middle, and end of B(03). While we may be able to reconstruct how many folios are missing from the beginning of the codex, it is less easy to estimate the material and contents after Hebrews 9:14. The second challenge to this project is the adjustments made to B(03) in the rebinding process of the fifteenth century and subsequent rebindings. In the end, Canart’s reconstruction seems to make the most sense of the data but is complicated by the fact that he does not acknowledge the variations in the post-restoration structure. Of particular interest to our study is the presence of two quire breaks, where Milne and Skeat have postulated a change of scribes; both of these are accompanied by additional discontinuities. The first is found between pp. 334–335, where Milne and Skeat identify the change from scribe A to scribe B. Not only does p. 335 begin on a new quire, a factor which would potentially allow both scribes to copy simultaneously, but we also found a change in ruling, from forty-four lines-per-column to forty-two. While this is not enough evidence in itself, it is potentially a codicological indication of transition. We will look to the textual and paratextual evidence for confirmation in the following chapters. However, the transition from quires of forty-four to forty- two lines-per-column, to a single quire of forty lines, and back to quires of forty-two lines suggests that the scribes were not in charge of lining the parchment. Instead, the scribes were constrained to the physical parameters set before them. This material limitation can also be seen when observing how the scribes handled imperfections in the parchment. In most cases, the scribes chose to copy around the holes, but we also saw one instance of the scribe copying a nomen sacrum (π̅̅ρϲ̅)̅ for a human referent, in order to avoid splitting the word. In the following chapter, we will discuss further how the scribes chose to apply and divide the text within the boundaries of the page. The second important quire break is found before Hosea (between pp. 944–945). Even though, in its current form, p. 945 is the last folio of quire [49], Gregory’s Rule reveals that it was originally the first folio of quire [50]. This is one of three occurrences where a new quire and a new CHAPTER 2 50 book begin simultaneously. This is not likely a coincidence, since the third column of p. 944 is left blank and, as a rule, the scribes began new texts on the very next column. The transition from Tobit to Hosea is where Milne and Skeat identify the last change of scribes (A to B), and while we are not yet able to tell if the scribes were working simultaneously, the break in quires suggests that this was a possibility. On the other hand, Milne and Skeat identify the second change of scribes (B to A) between 2 Esdras and Psalms (pp. 624–625), where no quire break is present. Their reasoning for this is not codicological, even though other observable discontinuities can be seen as supporting this division. Like the previous example, there are blank columns before the start of the Poetic books. This irregularity interrupts the continuous column pattern seen throughout the codex and marks the transition from a three to a two-column format. It seems likely that this decision was made to avoid mixing formats, as in א(01), even though we also saw that this type of hybridization did occur at the end of the Poetic books. If Milne and Skeat are correct, this would be the one example where two scribes worked together on a single quire. Therefore, both the method of copying simultaneously and successively could be seen in B(03). In this chapter, we have realized that not only is the codex a product of the scribe, but, in many ways, the work of the scribe is a product of the physical boundaries of the codex. We will return to these features throughout our study, as we move toward the paratextual and textual features attributed to the scribes of B(03). CHAPTER 3 Paratextual Features and the Problem of Paleography Now that we have examined the physical structure of B(03), our study may proceed to the paratextual and paleographic features of the codex. Along the way, this chapter will confront the major arguments for and against the various scribal divisions proposed by scholars (see Chapter 1). Since paleographic analysis of B(03) is problematic (see §3.2), paratextual examination is often seen as the key to understanding which scribe was responsible for copying various portions of the manuscript. While the majority of contemporary scholars accept the divisions of Milne and Skeat, I contend that their brief, yet important, discussion of paratextual features has been hastily adopted. Even if their conclusions are correct, they ought to be confirmed by further examination and a critical eye to their particular method. As seen in the previous chapter, Milne and Skeat did not take advantage of the codicological data, even when there was compatibility with their divisions. This chapter will progress in a threefold manner. First, by surveying the paratextual features of B(03), in relation to the work of the scribes, I will argue that the clearest location of scribal transition is found at 1 Kingdoms 19:11 (p. 334–345). This point is hardly contentious, although there has been surprisingly little evidence presented in favor of such an opinion. Second, this clear example will be used as an anchor to set out a framework for paleographic analysis of the scribes. Finally, all the evidence will be weighed in favor of a new division of scribes in B(03), including an additional scribe. This modification initially arises independently from Traube’s study, but ultimately finds support in the nomina sacra: namely, that scribe B was relieved of copying in Psalm 77:71 (p. 675), rather than at the opening of the Psalter, and that a third scribe (scribe C) carried out the copying of Psalm 77:71b through Tobit (For a comparison, see Table 9). 3.1 Paratextual Features The paratextual features of B(03) are of prime importance in identifying the work of the scribes.1 While Milne and Skeat exposed the pitfalls of detecting the scribes of א(01) paleographically, it is even more problematic in B(03), since the original handwriting is rarely visible. Thus, the two 1 For Milne and Skeat, the coronis, in particular, was “an infallible criterion.” Milne and Skeat, Scribes, 28. CHAPTER 3 52 scholars set out in search for criteria that could by-pass the problem of paleography;2 their solution was to look at paratextual features. In their study, they give pride-of-place to the colophons, but also rely on titles, line-fillers, and paragraphing (including ekthesis and paragraphoi). The following sections will both summarize the arguments of Milne and Skeat and contribute more data and criteria, which are missing from their study. Before moving to the main section of this chapter, one recent study on the marginalia should be noted. Near the end of my research for the chapter, Pietro Versace’s study, I marginalia del codex Vaticanus, was released. It is encouraging, therefore, that my findings often lead to similar conclusions, as this immense resource regularly provides further clarity on the issues. While this project is restricted to the earliest period of B(03)’s existence, Versace provides a history of the codex through the various layers of marginalia. He does so by identifying thirty-eight hands (B1–38) spanning from the fourth to sixteenth century. Since his study is solely interested in the marginal annotations, he does not distinguish between original scribes. Versace’s B1 is, therefore, equal to “all the interventions that can be attributed to the hand of the copyist (B1), or that are in any case related to the time when the manuscript was produced.”3 Likewise, the additions made by Versace’s B2 can occasionally be attributed to the main scribes, though he prefers not to specify. The last early hand identified by Versace is B3, who did not copy any of the main text, but contributed to the corrections and other marginalia in the codex. The current study provides an analysis of Versace’s early hands, alongside other paratextual and paleographic features, in order to distinguish the various scribes at play.4 3.1.1 Textual Divisions We begin by describing the various methods of textual division in B(03). I have written at greater length concerning these divisions, and so, this section represents a summary of my findings. 5 Although Milne and Skeat were primarily concerned with the use of ekthesis and the paragraphos, one also finds intralinear spacing, two sets of Greek section numerals, as well as colons (·) and dicolons (:). Rather than assuming, as many do, that all of these devices were introduced by the 2 Milne and Skeat, Scribes, 87. 3 Versace, Marginalia, 10. 4 Versace himself suggests that the two-scribe hypothesis would likely be modified after such an investigation. Versace, Marginalia, 10 n. 8. 5 Grenz, “Textual Divisions,” 1–22. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 53 scribes themselves, we must first acknowledge that only two are necessarily associated with the copying process—the intralinear spaces and ektheses.6 Spacing and Ekthesis In their appendix, Milne and Skeat state that scribe “A in prose always begins a new paragraph with a new line.”7 The result of such a maneuver is often a blank space at the end of the previous line. The two scholars continue their description of the first scribe by noting that, in new paragraphs, the initial letter is projected into the margin (i.e., ekthesis).8 In contrast, scribe B is described as, “in the Historical books…, indifferent whether he begins his new paragraph within the line after a punctuation space, or with a new line.”9 This, in turn, means that the second scribe varies in the use of ekthesis. As noted in Chapter 1, ekthesis played a significant role in Ezra Abbot’s identification of the first scribal transition at 1 Kingdoms 19:11. In B(03), Abbot tallied 1441 ektheses in the first 294 pages (pp. 41–334; Gen 46:28–1 Kgdms 19:11). However, in the subsequent 290 pages (pp. 335–624; 1 Kgdms 19:12–2 Esdras), there are only two examples found. This discovery led to the conclusion that “the natural inference is, that we have in the part of the MS. beginning with page 335 the hand of a different scribe.”10 The consistency of this shift, both within 1 Kingdoms and across the books on both sides of the divide, makes a change of scribes more likely than a change of exemplar. Furthermore, the quire break [18/19] and the transition from forty-four to forty-two lines-per- column, supports the identification of this division.11 Unfortunately, the stark difference in the use of ekthesis does not appear elsewhere in the codex, and therefore cannot be used to identify other places of transition. While scribe B does not use protrusion as frequently as scribe A, ekthesis occurs 6 Pierri, “Accentazione,” 141. 7 Milne and Skeat, Scribes, 87. 8 Milne and Skeat, Scribes, 88. 9 Milne and Skeat, Scribes, 88. 10 Abbot, “Antiquity,” 194; cf. Milne and Skeat, Scribes, 87. 11 The material break is more significant than the line changes, as the scribe was not clearly responsible for ruling the quires (Chapter 2). CHAPTER 3 54 frequently throughout the prophetic books and the beginning of the NT—both thought to be the work of scribe B.12 In Matthew, and less so in Mark and Luke, Dirk Jongkind has suggested that the combination of ekthesis and intralinear spacing (including paragraphos) represent a hierarchy, with the former delimiting major sections and the latter minor sections. 13 This conclusion has been followed in Wim de Bruin’s study of Isaiah 1–12 and John Olley’s study of Ezekiel.14 Olley adds the helpful observation that thirty-two of the minor divisions do not include paragraphoi, which “suggests insertion by a later scribe.” 15 Emmanuel Tov identifies a similar system in the Judean Desert documents, where intralinear spacing marks closed Masoretic sections and a space followed by a line-break indicates open Masoretic sections.16 Nevertheless, this hierarchy often fades in and out throughout B(03), and it cannot be easily used to distinguish one scribe from the other. The regular formatting of the lines in B(03) is occasionally modified for lists and genealogies (Appendix B). Rather than indicating new sections, ektheses occasionally highlight lists or repeated phrases, such as the Decalogue and the Matthean beatitudes. In other examples, intralinear spacing or the premature start of a new line can be used to separate listed elements or generations. The most notable formation involves the division of each line into two parts: an initial word followed by a space (of various widths) and a name (p. 76A, Exod 23:23; p. 211C, Deut 14:12–18; p. 917C, Jdt 8:1; pp. 1309C–1310B, Luke 3:23–28). In most cases, the first element of the line is justified with the left bounding line, while the second has a tendency to shift slightly to the right in consecutive lines (opposite “Maas’ Law”). Paragraphoi While only ekthesis and spacing are necessarily the work of the copyist, many also believe that the paragraphos—“a marginal sign indicating change of speaker in drama, corresponding sections in a 12 For example, Edward Glenny identifies twenty-one paragraphs that have been marked by ekthesis in Hosea. In the NT, I have counted eighty-four ektheses, with only seven occurring after Luke. Glenny, Hosea, 25; Grenz, “Textual Divisions,” 9. 13 Jongkind, Scribal Habits, 96. 14 de Bruin, “Interpreting Delimiters, 75; Olley, Ezekiel, 41. 15 Olley, Ezekiel, 41. 16 Tov, Scribal Practices, 145. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 55 chorus, or a division for other reasons between sections of text”17—was regularly added to the text by the scribes.18 Elsewhere, I have argued against their origin in scribendo.19 I do not presume that they are inevitably late additions,20 rather that they represent a layer of activity distinct from the copying of the main text.21 First, I have outlined five different forms of paragraphoi in the codex. Frequently, their variety is best explained if they were added at various times by different scribes or readers.22 For example, there were several red paragraphoi added along with the red Greek section numerals, after the copying of the text (see Section Numbers). While this does not indicate the later addition of all paragraphoi, it does imply that some of them were. Second, there is a common occurrence of redundant paragraphoi situated over existing ektheses. The book of Ruth contains five paragraphoi, all of which coincide with ektheses. Consequently, there are no examples of the typical intralinear space and paragraphos combination. Why, then, would a scribe add a paragraphos to text that he or she has already highlighted through ekthesis? A third observation would be the virtual absence of paragraphoi in Judith and the Sirach Prologue. Out of these two texts, there are only two examples of paragraphoi (Figure 11). The first of these is almost certainly a later addition, and it can hardly function as a section divider, since it appears in the margin and not between two lines (p. 918C).23 The second paragraphos is forked with an elongated diagonal stroke. While this marker could be original, it does not appear with any clear, internal division. The virtual absence of paragraphoi in Judith can also be seen at the end of the work, where only the coronis is present, without a final paragraphos (p. 930B). 17 Dickey, Ancient Greek Scholarship, 250. 18 See B1, in Versace, Marginalia, 10. 19 Scrivener already doubted the originality of the paragraphoi. More recently, Jan Krans has claimed they are likely a later addition. Grenz, “Textual Divisions,” 9–13; Scrivener, Introduction, 1:108; Krans, “Paragraphos,” 254; cf. Macmillan, Roman Mosaics, 367. 20 However, see the tenth or eleventh-century marginal correction on p. 725B, which contains a late paragraphos. It is difficult to distinguish this paragraphos, based on form and color, from those in the main text. 21 Likewise, Milne and Skeat attribute the paragraphoi in א(01) to a revision of the codex by scribe A. Milne and Skeat, Scribes, 37. 22 However, in the Poetic books, two forms of paragraphoi are clearly used in combination with one another. Milne and Skeat mention eleven forms in א(01), all of which originated with the same hand. Milne and Skeat, Scribes, 37–38. 23 The position of this paragraphos imitates that of the obeloi found in Isaiah and elsewhere (see below). CHAPTER 3 56 FIGURE 11: PARAGRAPHOI IN JUDITH (PP. 918C, 929A) Finally, there are an overwhelming number of instances where paragraphoi are present without intralinear spacing, or, conversely, where spaces are lacking paragraphoi. In the first case, the addition of a paragraphos appears to be a correction to the internal divisions copied by the scribe. We may be able to see a similar, though slightly different, example of correction with paragraphoi in Numbers 33:6–48 (pp. 186–87). In these verses, the repeated phrase “και απηραν…” occurs forty- one times. All but five of these instances are marked by ektheses. It seems clear that the remaining five examples were intended to have ektheses as well, but scribe A simply left a space to mark the division. The accompanying paragraphoi were, therefore, most likely added to correct the divisions and imitate the ektheses. While these may well have been added in scribendo, following the pattern of intralinear space + paragraphos, the evidence provided above points to their addition as a later correction. Indeed, there is no way of knowing, with certainty, when the paragraphoi were added to B(03). The presence of paragraphoi in documentary papyri and ostraca shows that scribes could, and did, add them while copying any kind of text.24 However, this does not prove that the scribe who copied a manuscript was always responsible for the existing paragraphoi. My point is this: If, as I suspect, the paragraphoi in B(03) were secondary to the original copying phase, they may still have been added very early on (possibly before the codex left its place of production). Even then, their addition would nevertheless represent a distinct level of production from the original copying. 25 Therefore, any attempt to reconstruct an early system of divisions based on the paragraphoi cannot assume that they represent the same tradition as the internal divisions of B(03). 24 Mugridge, Early Christian Texts, 77–78. 25 See also Schmid’s approach to the diplai and paragraphoi in א(01). Although he attributes one set of diplai and paragraphoi to the copying process (Acts 26:23), the rest belong to separate operations in the original production unit. Schmid, “Quellenangaben,” 91. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 57 Section Numbers In addition to spacing, ekthesis, and paragraphoi, there appears in the manuscript a nearly unique system of numbered divisions (Versace: B1 and B3), described as “the oldest system which is known to us.”26 Our task here is only to provide a summary of the section numerals as they relate to the scribes of the codex. Therefore, comments on the nature of this system of division, as it relates to other manuscripts and traditions, will be left to a minimum.27 Although B(03)’s numerals do appear in some editions of the NT, such as Tregelles’ (1857) and the NA18, their uniqueness has often meant that the numbers are ignored in critical editions.28 These section numbers signify an important addition to B(03), as they create an opportunity for reference, which the previous set of divisions did not provide. In the ninth century an additional layer of numbered divisions was added (Versace: B7 following the guidance of B6), but this is not of importance for our study.29 Instead, we will focus on the early numerals (what I have elsewhere called Old-Numeral Hands 1 and 2 or Old-NumH) in relation to the scribes of B(03). While most scholars have assumed the prima manus addition of the paragraphoi, there has not been the same agreement over the Old-Numeral Hands (see Table 7). I will briefly summarize my arguments for why these numerals were added in a layer of production subsequent to the main text. However, after further analysis of the paleography of B(03) and comparison with the work of Versace, my conclusions will be more nuanced than previously stated.30 26 Metzger, Manuscripts, 40; cf. Scrivener, Introduction, 1:56. 27 On this aspect, we await Charles Hill’s forthcoming monograph. See, for example, Hill, “Rightly Dividing,” 217–238. 28 This sentiment is explicit in Edwards, “Hermeneutical Significance,” 415 n. 6, who denies the significance of B(03)’s divisions, in favor of those found in A(02) and the succeeding tradition. The only known manuscript to contain the same section numbering is Codex Zacynthius (Ξ[040]), in the text of Luke. See Houghton and Parker, “The Gospel of Luke in the Palimpsest,” 36–39. 29 Grenz, “Textual Divisions,” 18–20; Versace, Marginalia, 31–34. 30 Grenz, “Textual Divisions,” 14–18. CHAPTER 3 58 TABLE 7: THE ORIGINALITY OF EARLY NUMERALS IN-PRODUCTION31 POST-PRODUCTION32 Abbot (1872), Swete (1900), Ropes? (1926), Duplacy (1976), Amphoux (1997), Bogaert (1999), Olley (2009), Hill (2015), Versace (2018) Wagstaffe (1739), Westcott and Hort (1896), von Soden? (1911), Martini (1968), Skeat (1999), Pisano (1999), Auld (2005), Goswell (2011), Glenny (2013) First, as Abbot noted, the numerals were “not made by the original scribe, but by one who preferred in some places a different division into paragraphs.”33 While we may not want to assume with Abbot that they were not copied by the original scribe, it is clear that the numerals often mark different divisions than those present in the text itself (i.e., ekthesis and spacing). For example, we find two places in Matthew where paragraphoi and numerals in the same red ink were added to mark a clear division, where no internal delimitation is found (p. 1238C, Matt 4:24; p. 1243B, Matt 8:5). Furthermore, we find instances of intralinear spacing, which were subsequently modified with the addition of red paragraphoi and numerals (e.g., p. 1275B, Matt 27:38). These examples signal divisions where the apricot paragraphoi were absent and the numerator thought it necessary to distinguish the break from other instances of interlinear spacing. The second piece of evidence is the change in numeral hands. In his edition of B(03), Tischendorf distinguished two hands responsible for copying the early section numerals, with a clear switch at Luke 22 on p. 1345.34 Moreover, on the previous page, it appears that the other numerator missed section ρλ̅̅ε ̅(135) and the new hand amended the omission (p. 1344A). If we trace the previous hand back, we find an earlier change in Isaiah 45:1 (p. 1045A) and then again for the book of Daniel.35 Therefore, what I have called Old-NumH2 (Versace: B1) copied the section numbers from Isaiah 45–Ezekiel and Matthew–Luke 22, while Old-NumH1 (Versace: B3) copied 31 Ropes states that they could be either contemporary or just as likely later. Ropes, Beginnings, xli. 32 Already in 1739, Thomas Wagstaffe doubted the originality of these numerals. After examining the manuscript in Rome, he states in his notes that “I am of [the] opinion that both these divisions [the early and young numerals in Exodus] are of a much later date than the era of the original manuscript.” Wagstaffe, “Some account,” f102v. 33 Hill has suggested that Matthew contains two system of division, “one with chapter numbers and another with ekthesis.” Abbot, “Antiquity,” 190; Hill, “Siglum,” 19 n. 91. 34 For Tischendorf, this change was indicated by the transition to the so-called “Coptic” form of writing. Tischendorf, Vaticanum, xxviiii. 35 As far as I can tell, Versace is the first to note that the numerals in Greek Daniel were written by the same hand as Luke 22:24–Hebrews. This is difficult to tell because they have been overwritten by the ninth-century numeral- hand (Versace: B7). Versace, Marginalia, 176. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 59 Deuteronomy–Psalm 11, Proverbs–Isaiah 44:21, Daniel, and Luke 22:24–Hebrews (see Table 8). Skeat describes the numbers of Old-NumH1 as “semi-cursive” with “no attempt to reproduce uncial forms,” and, therefore, “it immediately becomes obvious that they are not the work of either of the two scribes of the manuscript.”36 Versace similarly describes B3 as not writing in “Biblical Majuscule” and whose traits are “often undulated and curved” (Figure 12).37 TABLE 8: SCRIBES AND NUMERAL HANDS (H = OLD-NUMH) 1. Milne and Skeat (1938) Gen–1 Kgs 19:11 1 Kgdms 19:11–2 Esd Ps–Tob Hos–New Testament Scribe A Scribe B Scribe A Scribe B 2. A New Division of Scribes Gen–1 Kgs 19:11 1 Kgdms 19:11–Ps 77:71 Ps 77:71–Tob Hos–New Testament Scribe A Scribe B Scribe C Scribe B 3. Division of Numeral Hands Gen–Num Deut–Ps 1138 Ps 11–151 Prov–Isa 44 Isa 45–Ezek Daniel Mt–Lk 22 Lk 22.24–Heb — H1/B3 — H1/B3 H2/B1 H1/B3 H2/B1 H1/B3 In contrast to Old-NumH1, Versace associates our Old-NumH2 with his B1, the copyists or a contemporary hand.39 These numerals are copied in the same Biblical Majuscule as the main text. While certainty is not possible, we find some similarities in the hand of scribe B and Old-NumH2 (see §3.2.4). Even if assigned to one of the original scribes, it must have happened at a time after the copying had finished. Old-NumH1 (Duet 8) (Isa 42) (Luke 22) (Old-NumH1= undertext) Old-NumH2 (Matt 7) (Isa 45) (Luke 19) FIGURE 12: CHARACTERISTIC ALPHA/MU/RHO IN OLD-NUMERAL HANDS 36 Skeat, “Sinaiticus,” 212. 37 Versace, Marginalia, 76. 38 The last numeral by Old-NumH1 is κ̅η̅ in Psalm 11. 39 Versace, Marginalia, 13–14, 75. CHAPTER 3 60 By consulting Table 8, our proposal concerning scribal transitions, and that of Milne and Skeat, can be compared with the changes in numeral hands. While scribe B alone is responsible for copying all of Isaiah and Luke, the numerals of both Old-NumH1 and 2 are present in each. If the scribes were copying these numerals in scribendo, we would expect to see coinciding changes in the hands. Instead, the presence of Old-NumH1 (Versace: B3), who was not responsible for the main text, indicates that the sections were added following the copying phase. The final piece of evidence against the notion of simultaneous copying with the text is their relationship to other marginalia, like the diplai and corrections (see §3.1.3 and Chapter 6). There are at least four places in Matthew where we find numerals written over or around pre-existing marginalia (Figure 13). FIGURE 13: NUMERALS AND DIPLAI IN MATTHEW OF B(03) (PP. 1236C, 1248C, 1249C, 1252C) It will be argued later that although there are places where it seems diplai were added after the numerals, this may have more to do with the way the scribes copied diplai in B(03) than with an attempt to avoid the section numbers. In addition to these examples, we also find places where the red ink has bled over to its facing page (Figure 14). This is more likely to happen when the scribe quickly applies numerals to copied leaves, turning the pages before the ink dries. If, on the other hand, the numerals were added in scribendo, there is considerably more time for the ink to dry and prevent bleeding. FIGURE 14: RED INK BLEEDING (P. 1356C/1357A; P. 1376C/1377A) PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 61 There is one possible exception to this argument. Versace claims that the 150 Psalm numbers were added by the scribes during copying.40 In my earlier article, I assumed these numbers were copied after the first attempt at dividing the book, which ceased on p. 630. However, this notion was based on the conflation of Psalm numeration and section numbering. As we will find later, there is a case to be made for the in scribendo copying of Psalm numbers. Although Old-NumH1 began copying section numbers to the book of Psalms, it is possible that the pre-existing system of Psalm numbers rendered these sections superfluous. Space does not permit us to discuss further the relationship between the textual divisions in B(03), but we have attempted to sift through the earliest strata of delimitation found in the codex. The only layer that can confidently be attributed to the scribes of B(03), in scribendo, is the spacing and ektheses, which are inherent in the copying process. Indeed, although the addition of most paragraphoi and section numerals are likely very early, they represent additional layers and attempts at dividing the text of B(03), which cannot be assumed to go back to the same tradition of delimitation. Changes in the frequency of ekthesis helped identify a scribe change at 1 Kingdoms 19:11, but this has been less helpful in finding transitions elsewhere. 3.1.2 Line-Fillers Another way in which Abbot was able to distinguish two scribes in 1 Kingdoms was by their use or avoidance of line-fillers. While not all scribes cared for the aesthetic of a justified right margin, those who did could apply a variety of techniques. For one, a scribe might increase the kerning between letters so that the last letter of each line is even (ex. P.Hercul. 1423, LDAB 3558; see col. 8).41 The other method used to justify the right margin was the filling sign. This symbol appears in a variety of forms, including a dash with raised dot ∸ (cf. PSI 1200, LDAB 3770),42 a simple line — (cf. P46; P.Oxy. 11.1377, LDAB 761; P.Oxy. 2.230, LDAB 647),43 an apostrophe ’ (P66), and, in our codex, a diple or wedge > (P75; P.Oxy. 52.3672, LDAB 3785).44 40 Versace, Marginalia, 13. 41 Turner and Parsons, Greek Manuscripts, 5 n. 12. 42 Turner and Parsons, Greek Manuscripts, 5 n. 12. 43 In the case of P46, this line tends to occur only at the last line of a book. There are at least two examples of a bar line-filer (—) in B(03) (p. 582A l. 35; p. 1229A l. 1). Ebojo, “A Scribe,” 178; Turner and Parsons, Greek Manuscripts, 5 n. 12. 44 See also the examples of left-pointing wedge fillers < (P.Vindob. G.26010 + G.29283 + G.29782; LDAB 4184). cf. Johnson, Bookrolls and Scribes, 332. CHAPTER 3 62 Milne and Skeat describe scribe A’s use of fillers as very frequent in prose sections.45 We saw this habit clearly in Judges (pp. 287C/288A), where scribe A copied nineteen fillers to compensate for the space around a hole in the parchment. There is also an anomalous instance of scribe A using the wedge to fill both the final space of a line and an initial space in the following line (p. 193A). In contrast, Milne and Skeat characterize the work of scribe B as avoiding fillers “in the Historical books and the New Testament,” while noting the sporadic use in the Prophets.46 In general, this is a fair assessment of their two scribes, and the stark contrast can even be seen in 1 Kingdoms alone. Over 365 fillers occur in the first twenty-four pages (pp. 309–334), but only two appear in the final nineteen pages (pp. 335–353)—both on p. 353.47 In sum, there are seven line-fillers in 1 Kingdoms 19:11–2 Esdras (290 pages) and twelve in the NT (284 pages). The dramatic change in 1 Kingdoms, along with other portions copied by Milne and Skeat’s scribe B, again affirms our confidence in a transition of scribes at 1 Kingdoms 19:11. Furthermore, as v. 11 is split between two pages and two scribes, it is noteworthy that scribe A leaves a small space with a line-filler at the end of the page, and scribe B finishes the verse on a new quire (i.e. [p. 334C] αγγελουϲ > | ειϲ οικον δαυειδ [p. 335A]). In contrast, Milne and Skeat’s description of scribe B as sporadically using line-fillers in the Prophets requires modification. While it is true that they do not appear with anywhere near the frequency of those in scribe A’s text, I have found seventy-six fillers in the Minor Prophets (58 pages). Following these books, the first nine pages of Isaiah (pp. 1002–1010) contain seven fillers, while the rest of Isaiah through Daniel (pp. 1011–1234) does not display any. As mentioned already, the NT only contains twelve line-fillers, but occupies more than four times the number of pages as the Minor Prophets. This anomaly is difficult to explain, but an average of 1.3 fillers per page is more consistent than sporadic.48 45 Milne and Skeat, Scribes, 87. 46 Milne and Skeat, Scribes, 88. 47 Here, I am not including the wedge forms (>>—) that appear in the final line of some books. While they look identical to the line-fillers, we will consider them as a part of the coronis design (see §3.1.4). One reason for not considering them primarily as fillers is that they often do not actually fill the remaining line. 48 A related feature that is influenced by the use of line-fillers is the scribes’ decision to end a line with a nu or a macron. While there is little consistency between the pages of 1 Kingdoms, scribe A shows a noticeably higher frequency of written nus at line ends than scribe B. In the first seven pages of 1 Kingdoms, scribe A copied seventy-four macrons and eighty-five nus at line ends (74/85), including those with fillers. Scribe B, on the other hand, prefers macrons at a rate of ninety-one over sixty-two nus at line ends (91/62). PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 63 FIGURE 15: NUMBER OF LINE-FILLERS PER PAGE IN B(03)49 Before moving on, it is important to discuss what these statistics can and cannot tell us. First, they may indicate a likely change of scribes at 1 Kingdoms 19:11 and again between Tobit and Hosea. The regular appearance of line-fillers in the Minor Prophets does not necessitate that a scribe other than B copied them, though it is interesting to note a comparable pattern with ektheses. From 1 Kingdoms 19:11 through 2 Esdras (290 pages) we find only two ektheses and eleven fillers, while, in the Minor Prophets, there are 114 ektheses and 76 fillers (58 pages). However, this correlation fades in Isaiah through Luke, where there are frequent examples of ekthesis and little to no fillers. The rest of the NT is comparable to 1 Kingdoms 19:11–2 Esdras, with only seven ektheses and nine fillers. If, as is likely, scribe B was responsible for copying Hosea through the NT, the best explanation for this rise and fall in line-fillers is probably scribal fatigue. Since the individual books of the Minor Prophets are shorter, we do not see a decline in frequency. However, scribal fatigue seems to have set in after the first ten pages of Isaiah. What we cannot know from these figures is whether scribe A was certainly responsible for copying Psalms–Tobit. Since the Poetic books do not contain line-fillers, on account of the stichometric formatting, we cannot point to the pattern of either scribe for help.50 In contrast, there is a high quantity of fillers in Esther–Tobit, but not to the same degree as those in Genesis–1 Kingdoms 19:11 (see Figure 15). Therefore, while the examination of these line-fillers does give credence to the suspicions of Milne and Skeat—that there is a scribe change at 1 Kingdoms 19:11 and at Hosea 1:1—we are left to wonder whether scribe A is certainly responsible for copying Psalms– Tobit, or if there is enough difference to warrant the argument for another scribe. 49 Numbers–Judges are consistent with the work of scribe A; each book contains just over 100 fillers in the first eight or nine pages (avg. 11–12 fillers/page). 50 There are line-fillers on p.792A l.20; p.797A l.12; p.821A l.24; p.828A l.24; p.854A l.1. 13.4 10 14 0.02 1.3 0.04 7.8 Exod–Lev Ruth 1 Kgdms 1- 19:11 1 Kgdms 19:11– 2 Esd Esth–Tob Minor Prophets Isaiah–NT Scribe A Scribe B Scribe C CHAPTER 3 64 3.1.3 Diplai In contrast to the line-fillers, one paratextual feature that does appear frequently in the NT of B(03) is the diple (>), or διπλῆ ἀπερίϲτικτοϲ (un-dotted diple).51 While the symbol held a variety of functions in Homeric scholarship, this wedge-shaped sign eventually became used by some scribes to mark citations.52 Patrick McGurk identified an anonymous commentary on Plato’s Theaetetus (P.Berol. 9782, LDAB 3764; II) as an early example of this method.53 In this papyrus roll, each line, containing a citation from Plato’s text, is marked with a diple, while the beginning and last lines are often bracketed by paragraphoi. — > xxxxxxxxxxx > xxxxxxxxxxx > xxxxxxxxxxx — Two early Christian examples of marking biblical quotations with diplai are found in P.Oxy. 3.405 (LDAB 2459; Irenaeus, adv. Haer. 3.9.3; II/III)54 and the Christian homily, P.Mich. 18.764 (LDAB 562; II/III).55 While it does not appear that any of our earliest NT papyri use diplai for OT quotations, we have considerable evidence of this practice in the majuscules, including B(03).56 Both Schmid and Alexander Stokowski have surveyed the use of diplai in the NT of B(03).57 Stokowski provides a list with seven additional diplai, which are not mentioned by Schmid. He also identifies the only two examples of diplai in the OT: the final three lines of 3 Kingdoms (p. 442A) and Tobit 2:6 (p. 932A).58 The markings at the end of 3 Kingdoms were clearly placed there to emphasize that the text actually belongs to 4 Kingdoms 1:1. While we have noted that the main 51 Osann, Anecdotum romanum, 3; cited in Gardthausen, Griechische Palaeographie, 2:411. 52 Turner, Greek Papyri, 117; Johnson, Bookrolls and Scribes, 341; Montana, “Hellenistic Scholarship,” 134. 53 McGurk, “Citation Marks,” 4; cf. Schmid, “Die Diplé,” 78. 54 McGurk notes that C. H. Roberts was the one who pointed this out to him. McGurk, “Citation Marks,” 4; Schmid, “Die Diplé,” 78; Blumell, Lettered Christians, 210; Andrist, "la citation," 95–98. 55 Blumell and Wayment, Christian Oxyrhynchus, 291 §81. 56 Schmid, “Die Diplé,” 79. On diplai as citation marks, see Isidore of Seville, Etymologiae 1.21,13; cf. Smith, Alexandrinus, 208. 57 Schmid, “Vaticanus,” 99–114; Stokowski, “Diplé-Auszeichnungen," 93–114. 58 Stokowski, “Diplé-Auszeichnungen,” 96; See also the list in Versace, Marginalia, 90–92. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 65 purpose of diplai in B(03) is to mark citations, this example is likely meant for deletion.59 The reduplication of the first verse of the following book also occurs at the end of 1 Kingdoms (p. 353B– C) and 1 Chronicles (p. 521C). However, neither of these two are accompanied by diplai. 2 Kingdoms 1:1 is copied as if it were the proper ending to 1 Kingdoms, with no hint of correction, but 2 Chronicles 1:1 was boxed and crossed out after the reinking (on these reduplications, see Chapter 5). The form of the diplai in 3 Kingdoms (Scribe B) do not resemble those of the NT (Scribe B) and are likely later additions. The diplai at Tobit 2:6 (p. 932A–B) direct attention to a quotation from Amos 8:10, following the introductory phrase εµνηϲθην τηϲ προφητειαϲ αµωϲ καθωϲ ειπεν. The form of these markers is also different from those in the NT of B(03), as they resemble the shape of a seven (7). Since Tobit was copied by our proposed scribe C or scribe A, this difference could be explained by the distinctive style of the scribes, or as a later addition altogether. However, since the shape of the diplai do not share any resemblance with the line-fillers on the page, it is likely that they were added by a subsequent hand and there is, therefore, little value in the diplai for comparing the scribes.60 The question remains whether the rest of the diplai in the NT are the work of scribe B or subsequent hands. Since most of the markers were not reinforced, Schmid, Payne and Canart all argue that the fading ink and color are sufficient evidence of their originality.61 I have already expressed my hesitation towards this reasoning alone, but I do believe that there is still sufficient evidence to suggest a similar conclusion for many of the diplai. Indeed, Schmid is careful to state that they belong to the “ersten Produktionsphase,” and therefore does not assume that the scribe copied them in scribendo.62 Moreover, Versace classifies the diplai under his B1 and B3 categories. Those from the latter hand are larger and copied with thinner, curved strokes.63 59 Hixson has noted this function by the corrector of N(022) at Matthew 16:4. In this case, there are deletion dots above the letters, which accompany the diplai. Hixson, "Sixth-Century Manuscript Family," 543. 60 Versace agrees with the secondary nature of the diplai in Tobit. It is worth noting that Smith, likewise, did not find the diplai helpful in distinguishing scribes in A(02). Versace, Marginalia, 90, cf. n. 69; Smith, Alexandrinus, 210. 61 Payne and Canart, “Distigmai," 213; Schmid, “Vaticanus,” 99. 62 Schmid, “Vaticanus,” 99. 63 Schmid, on the other hand, appeals to the uniformity of the diplai in B(03) as a caution against distinguishing the hands responsible. In many cases, Versace’s criteria are less noticeable. Versace, Marginalia, 12–13; Schmid, “Vaticanus,” 112. CHAPTER 3 66 Importantly, Schmid identified a set of “summary marks” (summarische Markierungen) in Hebrews and 2 Corinthians 6:16–18. These summary marks indicate citations, without providing diplai for each line of the quotation, contrary to the normal practice.64 For example, the catena citation in 2 Corinthians 6:16–18 (p. 1481C) opens and closes with a diple, leaving the intermediate lines void of markings. Schmid suggests, therefore, that a possible intention was for a later hand to fill in the remaining lines.65 Although Versace recognizes B3 as responsible for the two diplai on p. 1481C, there are other citations where both hands B1 and B3 apparently contributed marginal wedges (p. 1491B–C).66 Together, these examples indicate that there was not a single process of marking citations in B(03), and that the earliest diplai may have only been cursory additions for later refinement. We noted already that the Greek section numerals must have been added subsequently to the initial copying, partly due to their unusual placement over certain pre-existing diple. However, some might counter this argument by appealing to occasions where the position of diplai seems to have been adjusted in order to avoid obscuring a numeral (Figure 16). a. p. 1236A ll. 4–10 b. p. 1311A ll. 2–4 c. p. 1311A ll. 36–39 d. p. 1392A ll. 20–26 e. p. 1464A ll. 5–11 FIGURE 16: DIPLAI POSITIONS AND GREEK SECTION NUMERALS I include three examples of this feature, along with two occurrences of similar diple positions, but without a numeral. Examples a, d, and e all have the final diple in a position to the right of the previous diplai; a and e have numerals to the left of the marker. On the other hand, b and c provide cases on the same column of final diplai in a position left of the previous markers. In cases a, b, and e, it is certainly possible that the above citation marks betray the priority of the numerals. Based on 64 Schmid, “Vaticanus,” 109–110. 65 Schmid, “Vaticanus,” 110. 66 In this example, Versace suggest that B1 copied three diplai at the end of column B and one in column C, line four. After this, B3 supplied the missing three diplai in lines one through three of column C. Versace, Marginalia, 13. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 67 form and positioning, Versace associates the diplai of b with his B3 (same as our Old-NumH1), while the section number originated with B1 (Old-NumH2).67 In e, likewise, he suggests the diplai came after the numeral, but this time B3 added the numeral.68 However, as examples c and d show, scribe B (Versace: B1) occasionally adjusts the positioning of the final diple, even without a present numeral. Thus, the positioning of the diplai alone cannot be a reason for suggesting an origin other than with the scribe. While some diplai were likely added by Versace’s B3 or a subsequent hand, others, like their counterpart line-fillers, originated with scribe B, and may even have been present in the exemplars. 3.1.4 Stichometry It has rarely been noted that B(03) contains stichometric markings in the margins of 1–4 Kingdoms and Isaiah.69 Surprisingly, in Bogaert’s own calculations of the stichometry in B(03), he is unaware of these ancient numbers.70 It is unfortunate that these do not appear more consistently throughout the manuscript, but their presence in 1–2 Kingdoms provides us with another opportunity to compare the two scribes. It is not surprising that Swete would claim that these numbers were “written prima manu”71 since there appears to be little reason to add stichoi after the text is finished. However, Versace convincingly identifies the hand of the numerals as B3.72 They are, therefore, of little help in identifying changes in scribes throughout the codex. The system of line counting found in B(03) is known as “partial stichometry,”73 in which a scribe supplies marginal numerals at set intervals (often by stichoi of 100).74 When this system is present in B(03), the average number of lines per stichos is 215.61 (Figure 17). 67 Versace, Marginalia, 91. 68 Versace, Marginalia, 92 n. 81. 69 Harris, Stichometry, 59–63; Swete, An Introduction, 348. Harris claims that Nestle made the first reference to these stichoi in a fly-sheet entitled Separatabdruck a. d. Corresp.-Blatt für die Gelehrten und Realschulen from 1883, though I have not been able to find this. See, however, Nestle, Introduction, 48; Before Nestle, see mention of the numerals in Fabiani, Commentarius, 45ff. 70 “Il n'y a pas de stichométrie dans B.” Bogaert, “Le «Vaticanus»,” 140, cf. 148, 153. 71 Swete, An Introduction, 348. 72 Versace, Marginalia, 23. 73 Gardthausen, Griechische Palaeographie 2:81. 74 There is at least one instance of stichoi marked at increments of 20 lines (P.Mich. 6.390). Turner and Parsons, Greek Manuscripts. In other instances, every tenth or twentieth line is marked with marginal dots. Houston, Inside Roman Libraries, 10. CHAPTER 3 68 FIGURE 17: STICHOMETRY IN B(03)75 While we might prefer a better word than “partial” to describe this system of enumeration, it is particularly fitting for B(03) since the full number of stichoi is only found in 2 Kingdoms ( or 2,300 marks the final line).76 In 1 Kingdoms, 3–4 Kingdoms, and Isaiah there are one or more stichoi missing before the end of the work (remaining lines: 3 Kingdoms: 64; 4 Kingdoms: 698; Isaiah: 189). In 1 Kingdoms, the final number is ω̅ (800), found in 1 Kingdoms 11 (p. 322C), leaving the remaining 3,974 lines unmarked.77 There are a number of reasons for including stichometry in a manuscript. First, they regularly reflect the professional context of a manuscript’s production, since scribes were often paid by the number of stichoi they copied.78 This does not, however, imply a scriptorium context, because scribes could also be hired privately.79 Second, stichometry was used to ensure the completeness or quality of a copied work. 80 However, Turner and Parsons provide a third explanation for the appearance of stichometry in some manuscripts: “One cannot be quite sure that such a figure might not itself in time come to be regarded as part of the paradosis.”81 Although they proceed to reject 75 The points on the graph in the 400s mark locations with one missing stichos, and the one point at 666 marks a location missing two stichoi. 76 Turner and Parsons use “current stichometry” to describe this system. Turner and Parsons, Greek Manuscripts, 16. 77 This number accounts for the transition from forty-four to forty-two lines-per-column at p. 335. 78 Gamble, Books and Readers, 277 n. 129. Ohly, Stichometrische, 86–90. Turner and Parsons, Greek Manuscripts, 1. 79 Turner, Greek Papyri, 95; Haines-Eitzen, Guardians, 88. 80 See the reference to this practice in the Cheltenham canon (before AD 365). Gallagher and Meade, Canon Lists, 192. 81 Turner and Parsons, Greek Manuscripts, 16. 0 100 200 300 400 500 600 700 0 500 1000 1500 2000 2500 3000 3500 4000 # o f li ne s b et we en st ich oi Stichometric Number 1 Samuel 2 Samuel 1 Kings 2 Kings Isaiah PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 69 this option, it is likely the best explanation for the presence of stichometry in B(03).82 That these numbers originated with a hand other than the one in the main text suggests they were viewed as an important part of the transmission, rather than a reflection of the scribe’s labor.83 It is hard to tell if this was the intention behind copying the numbers into B(03), but it is noteworthy that many of them were later reinked by the re-enforcer, who occasionally left the surrounding section numbers untouched (p. 417). Therefore, it is possible that the reinker considered the stichoi as a part of the paradosis, which needed preservation. 3.1.6 Titles, Running Titles, Subscriptions, and Tailpieces The final paratextual category to be examined is the identification of books using titles. Like many printed books, B(03) labels the beginning of each book using a title, while also providing the title on every other page, so that no matter which page one opens to, the book is quickly identified. Similarly, although rarer in printed books, the title is again written as a subscription at the end of each book. These various forms of titles have a clear purpose, but it is their relationship to the work of the scribes that we are most interested in. Were the titles copied by the same scribe as the main text? Were they copied before or after? It is to these questions that we now turn. Titles and Running Titles Simon Gathercole has claimed that B(03) is “probably the most consistent of all the great uncials in its presentation of titles.”84 This is unsurprising, given the level of consistency we have already observed in the manuscript. Initial titles are written above the first column of each book. Since, as we have already noted, books often start on the next available column, titles may appear directly above any of the three columns and are not confined to a particular place on the page. The only book that is missing an original title is Hosea. Instead, the tenth or eleventh-century reinforcer, who reinked the other titles, added one above the first column of the book (Figure 18). That this title is not original is clear, because there is no trace of underlying ink and a running title is already present on the page (see discussion below). 82 Thompson maintains that manuscripts with stichometry “more frequently transmit the measurements of the archetypes.” Thompson, An Introduction, 67. 83 Since section numbers were already provided for these books in B(03), they do not seem to function as referencing marks. cf. Damschen, “Stichometry;” Parsons, “Stichometry.” 84 Gathercole, “Titles,” 41. CHAPTER 3 70 FIGURE 18: TITLE OF HOSEA (10TH–11TH CENTURY) The initial titles of B(03) often take the shortened form, such as κατα µαθθαιον (B(03)*) rather than ευαγγελιον κατα µατθαιον (D(05)), or ιηϲουϲ (Β(03)) for ιηϲουϲ υιοϲ ναυη (A(02)) (see Appendix C). However, as we will see later, there are few instances where the initial title is short, while the end- title is expanded (e.g., ιηϲουϲ/ιηϲουϲ υιοϲ ναυη; pp. 238, 270). In the case of Lamentations, the initial title originally read θρηνοι, similar to the running titles of the book, but a later hand added ιερεµιου, in order to match the expanded end-title. All of the initial titles occupy a single line except for 1 Chronicles and 1 Thessalonians, where the numeral (α̅) is copied on a line of its own (common in the end-titles). The hand of these titles is in the same Biblical Majuscule script as the main text.85 Versace categorizes them under B1 and, therefore, most likely by the main scribes or a contemporary hand.86 The NA28 and the ECM identify the initial titles as B1, which is “roughly contemporaneous with B,”87 and therefore mark the titles as omissions in B* (the original hand). While it is unclear what evidence led the editors to change the identification of the initial titles from first-hand in NA27 to a correction in NA28/ECM, the following discussion reveals some arguments that may favor their decision. One of the difficulties with discerning the hands of the initial and final titles is that they have been re-inked; in the case of the initial titles, not always with precision (Figure 19). The running titles, however, have been mostly left untouched. Initial Title End-Title FIGURE 19: INITIAL AND END–TITLES IN NUMBERS (PP. 138, 191) 85 Canart, “Notice,” 24; Versace, Marginalia, 10; Versace, “Some marginalia,” 2. 86 Versace, Marginalia, 10–11; cf. Gathercole, “Titles,” 41. 87 NA28, 59*. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 71 David Parker helpfully notes that א(01), B(03), and D(05) are unique among the Greek manuscripts from before 500, because they contain running titles—a common feature in the Latin tradition.88 In B(03), the running titles do not appear until Deuteronomy 18 (p. 215). The remaining books contain running titles, except Psalms, the Prologue of Sirach, 2–3 John, and 2 Thessalonians.89 It is conceivable that the scribes did not find it necessary to copy these titles for the Psalter, since the word ψαλµοϲ appears in the incipit of most psalms. However, their absence in Genesis–Numbers complicates this explanation. The standard positioning of the running titles in B(03) is in the middle of the top margin, above the center column.90 Unless a book begins on a recto, the running titles appear on the first recto and continue on every other page. The exception to this pattern is found in certain books with two-part titles: 1 Kingdoms–2 Esdras, the Gospels, and the Pauline corpus (including Hebrews). In these books, both the verso and the recto are used (Figure 20; See also Appendix C). βαϲιλειων] [α̅ —] [αµωϲ β̅ ιουδα] [— κατα] [µαθθαιον προϲ] [γαλαταϲ προϲ] [θεϲϲαλονεικειϲ α̅ FIGURE 20: SAMPLE RUNNING TITLES IN B(03) ( VERSO] [RECTO ) For 1 Kingdoms–2 Esdras, the configuration follows name on the verso, and numeral on the recto.91 This pattern is not used for other numbered books in B(03). All of the Minor Prophets are numbered one through twelve, but the name and numeral are combined as a running title on the recto. This is also the case for 1 Peter–1 John. Another exception to the standard pattern is Jude, where the only running title is found on the verso (p. 1444).92 For the Gospels, the configuration is κατα on the verso, and name on the recto. Intriguingly, Luke has both an initial title and the κατα on the first page. While Matthew, Mark, and John all begin 88 Parker lists only these three manuscripts as having running titles, out of sixty-seven. I have also consulted the available images for the papyri published after 1992. These too, when the upper margins are extant, do not contain running titles. Gathercole lists five Syriac manuscripts and one Coptic manuscript that contain running titles in the Gospels. Parker, Codex Bezae, 16–19; Gathercole, "Titles," 75–76. 89 Of these five books, Psalms is the only one long enough for us to expect running titles. 90 Milne and Skeat note that the pages of א(01) are occasionally provide ruling for the running titles. I have not noticed a parallel example in B(03). Milne and Skeat, Scribes, 74, 76–78. 91 Peter Myers has helpfully noticed the discrepancy between the titles and main text in the spelling of Esdras. The titles consistently give the spelling εϲδραϲ, while εϲραϲ is found in the main text. Myers, “Transcriptions,” 195–96. 92 Since Jude occupies a single folio, this may be an example of compensation as there was no second recto for the addition of running titles. CHAPTER 3 72 on a recto, and thus have their first running title on the following opening, Luke begins on a verso (p. 1304). Although we would expect to see the first κατα of the Lukan running title two pages later (p. 1306), it is actually present on the first page of the gospel. It is tempting, therefore, to postulate that the beginning of Luke was meant to fill the famous empty column at Mark’s ending (p. 1303C), making p. 1304 the second page and first verso of Luke. In order to substantiate this claim, we must answer the question of when the running titles were added (see below). The Pauline epistles follow the same pattern as the Gospels—προϲ on the verso, and the recipients (including numerals) on the recto. Similar to Luke, Galatians is exceptional with the προϲ running title on the first page of the letter. There are few instances in B(03) where a running title is present on the final page of a book (a recto), which coincides with the initial title of the following book (Obadiah/Jonah, p. 973; Luke/John, p. 1349; Romans/1 Corinthians, p. 1461), but on the whole is avoided. Running titles were copied in a single line, with the exception of 1 Thessalonians, which copied the numeral (α̅) below the recipients (recto). So then, when were the running titles added to B(03)? Versace helpfully reports one instance of an irregular title, which may shed light on the question. At the bottom of p. 425, one is able to see a partially erased running title (βαϲιλειων), appearing upside down and reversed.93 This example suggests that the title had been added before the main text on the page.94 In light of the titular irregularity on p. 425, the following evidence can be understood as favoring the addition of running titles before the main text: FIGURE 21: UPSIDE DOWN TITLE IN LOWER MARGIN (P. 425) 1. Deuteronomy 18 (p. 215): We have already noted that the first running titles appear in Deuteronomy, but not at the beginning of the book. Instead, they appear on p.215, which happens to commence a new quire ([13]). The introduction of titles congruently with a quire 93 Versace, Marginalia, 11. 94 The title βαϲιλειων only occurs on the verso, but p. 425 is a recto in the quire. Therefore, the sheet had received the running title at point when p. 425 could have been rotated to become a verso. One could explain this as an erroneous addition by the scribe, who did not realize the text-filled page was upside-down. However, based on the evidence of Hosea’s running title, this argument seems less probable. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 73 break and in the middle of a book suggests that the titles were added before the manuscript was bound and possibly before the main text was added. If copied after the main text, one would expect the running titles to appear at the beginning of Deuteronomy. 2. 1 Kingdoms 19:11 (p. 335): Although following a consistent title pattern (βαϲιλειων] [α̅), 1 Kingdoms loses the running titles from p. 335 through the end of the book. This change, like the one in Deuteronomy, aligns with a quire break ([18–19]). 3. Hosea (p. 945): It has already been mentioned that the beginning of Hosea is lacking an original title, and that there is, instead, a running title. This irregularity coincides with a quire break and a cut folio from the quire [49], discussed in Chapter 2. Not only is there a missing folio, but the book of Tobit ends with an empty column (p. 944C).95 Therefore, it is tempting to claim that the running title of Hosea was added to the first page (a recto) of a new quire, because the text of Hosea was meant to begin either on p. 944C, after Tobit, or on the last folio of quire [49]. The scribe, however, began copying Hosea on a new quire [50], which already had the running title, ωϲηε, in the center of the upper margin.96 4. Isaiah (p. 1045): There appears to be a change of hand in the running titles of Isaiah on p. 1045, the beginning of quire [60]. However, it is likely that the new hand erased an existing running title on the page, as the previous hand copied the subsequent titles on pp. 1047, 1049. The new hand, therefore, does not consistently take over until p. 1051. 5. Luke (p. 1304): Similar to Hosea, the opening of Luke contains the κατα of the running title, this time accompanied by an initial title. While p. 1304 does not begin a new quire, it is the last page of quire [67]. Famously, the last column of p. 1303, where Mark’s gospel ends, is empty.97 If we have a circumstance similar to Hosea, it is possible that the running titles of Luke were added before the main text, with the beginning of the gospel planned to begin at p. 1303C and p. 1304 intended as the second page and first verso. 95 If the cut folio was left blank, this would have resulted in seven empty columns between the two books. 96 The evidence of the running title is not mentioned in Andrist’s theory concerning the two volumes of B(03). If my suggestion is correct, quire [50] was not prepared as the beginning of a new volume, but as a continuation of quire [49]. Nevertheless, the decision to begin Hosea on a new quire, as the opening of a second volume, could have occurred after the quire had been prepared with running titles. Andrist, “Au croisement,” 20. 97 Both א(01) and B(03) have irregularities at the end of Mark (ending at 16.8). Elliott has hypothesized that the NT scribe of B(03) left the blank column at p. 1303C, because of hesitation over the long ending. Elliott, “Twelve Verses," 256–257; Elliott, “T.C. Skeat,” 287–288; Jongkind, Scribal Habits, 45–46; Milne and Skeat, Scribes, 9–11. CHAPTER 3 74 When combined, this evidence points in favor of the priority of most, if not all, of the running titles to the main text. The reason for the absence of an initial title for Hosea, remains unclear. While the running titles were possibly added first, it is likely that the initial titles were added simultaneously or slightly after the main text. The irregularity of Hosea may support the latter option and the designation B1 rather than B* in the NA28/ECM might, therefore, be correct.98 If the running titles were added before the copying of the main text, can they help us in any way identify changes in scribes? In isolation, they have not been decisive. Yet, when we compare the running titles to the evidence already collected, there are two points of interest: 1 Kingdoms 19:11 and Hosea 1:1 (see examples 2 and 3 above). In both cases, irregularities in running titles align with changes in codicology and other paratextual features. Moreover, when analyzing the paleography (§3.2.5), there seem to be affinities between the hands of the main text and the titles. Even if the titles (all or a few) were added first, they continue to confirm the evidence for scribe changes at 1 Kingdoms 19:11 and Hosea 1. Whether or not the absence of running titles in the Psalter confirms a change at the beginning of Psalms or somewhere else in the book has yet to be decided. Subscriptions: End-Titles and Colophons Andrew Smith helpfully classifies the element at the end of each book as a “tailpiece,” including the end-titles, colophons, coronides, and artwork.99 However, I will first discuss the subscriptions in B(03), which contain the end-titles and, in Paul, the colophons. I do not intend this to mean that both end-titles and colophons were copied together, because, as we will see, this is not the case. Instead, the term is used to distinguish the textual portion from the design piece. Because the coronides were most important for Milne and Skeat’s argument, these will be discussed last. While there is occasion to doubt the relationship between the initial titles and the copying of the main text, there appears to be more certainty concerning the end titles.100 In contrast to the initial titles, the ECM attributes the subscriptions to the original scribes.101 Like the main text, the end titles are written in “Biblical Majuscule,” but with a preference for the reduced “rounded” 98 Hug had already noted the secondary nature of the titles, but included the running titles in this conclusion. Hug, Introduction, 264. 99 Smith, Alexandrinus, 128. 100 Gathercole, “Titles,” 41. 101 The NA28 does not mention subscriptions in the apparatus. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 75 letters—epsilon, theta, omicron, and sigma.102 This preference is also seen in the initial and running titles. Another feature that appears occasionally in the three titular forms is the omega ancora with a long central stoke (Figure 22).103 FIGURE 22: OMEGA ANCORAE (PP. 484, 713, 976) Generally, the end titles of B(03) match both the initial and running titles. However, there are a few instances where they have been expanded: Genesis, Joshua, Psalms, Sirach, and Lamentations (see Appendix C). End titles of the numbered books, 1 Kingdoms–2 Esdras, 1 Peter–3 John, follow a similar arrangement as those found in Homeric manuscripts, with the first line containing the name in the genitive, and the second line with a numeral.104 The most common configuration for the Gospels and Pauline epistles is κατα or προϲ with the name on a second line. However, since both Luke and Philippians end at the end of a column, their end-titles are written in the lower margin with the complete title written on a single line. While 1–2 Corinthians are written as προϲ κορινθιουϲ on the first line and the numeral on the second, the end-titles of 1–2 Thessalonians are copied in three lines. The separation between the final line of text and the end-title can vary from one to seven empty lines. The tailpiece of Psalms appears after Psalm 150, with the end-title βιβλοϲ ψαλµων ρν̅ ̅ (initial title: ψαλµοι).105 Psalm 151 appears on the next page without a section number or end-title. There is, however, a coronis after both Psalm 150 and 151. Following the end-titles of the Pauline epistles, one finds colophons describing the provenance of each letter (see Appendix C). These colophons were clearly added later as serifs are visible in the epsilons and gammas. The alphas as well as the upsilons are also noticeably different from the main text and titles (Figure 23). Versace classifies the colophons as B17 (9th CE), even though Tischendorf assigned them to the sixth century.106 Versace prefers the later dating, since the earlier 102 Versace, Marginalia, 10–11. 103 Milne and Skeat, Scribes, 25. 104 Schironi, To Mega Biblion, 20. 105 Compare with א(01), which has the end-title ψαλµοι δ̅α̅δ̅ ρ̅ν̅α̅ after Psalm 151. Also, see A(02) for the end-title ψαλµοι ρ̅ν̅ και ιδιογραφοϲ α̅ after Psalm 151. 106 Tischendorf, Vaticanum, 227 n.2. CHAPTER 3 76 date would result in an isolated example of additions to B(03) during the sixth century.107 Tregelles, and Hug before him, also noted the second-hand nature of the colophons, but claimed that “even these additions are so ancient that they differ from those introduced by Euthalius and adopted by the early copyists in general.”108 FIGURE 23: THE COLOPHON OF 1 CORINTHIANS (P.1477) The only differences between B(o3) and the Euthalian argumenta, however, seem to be the attribution of 2 Corinthians specifically to Philippi rather than generally to Macedonia, and 2 Thessalonians to Athens instead of Rome.109 Therefore, it is not clear that these colophons must predate the adoption of the Euthalian tradition. At the same time, we might find Versace’s leap from fourth-century to ninth-century marginalia puzzling. Still, an eighth or ninth-century date seems preferable.110 What, then, can the end-titles tell us about the scribes of B(03)? Milne and Skeat noted that “in the lines framing the subscriptions [scribe] A never uses the tailed bar ”, and that scribe “B frequently, if not invariably, uses the tailed bar.”111 More recently, Gathercole has contributed to their argument, noticing that the end-titles of scribe A are frequently framed with a “wavy line or tilde shape.” This wavy line does not appear in any end-titles of books attributed to scribe B.112 107 He does admit, however, that it is very difficult to date these colophons. Versace, Marginalia, 42. 108 Tregelles, An Introduction, 160; Hug, Introduction, 264. 109 See the text and translations in Blomkvist, Euthalian Traditions, 73ff. 110 It is likely, as Versace notes, that the later hand of the colophon was attempting to mimic the earlier hand of the end-title. This intentionality makes dating the hand difficult and could be why Tischendorf dated the colophons earlier. Versace, Marginalia, 42. 111 Milne and Skeat, Scribes, 87–88. 112 Gathercole, “Titles,” 42. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 77 FIGURE 24: FRAMED END-TITLES (PP. 304, 353) These patterns do seem to indicate some change in hands, but are not necessarily conclusive as to which scribes are responsible. If a “wavy line” is the defining feature, we may well have more than one responsible scribe; it is just not scribe B. Coronides, Final Paragraphoi, and Final Line-Fillers It is fitting to end our discussion of the paratextual features with an examination of the coronides and final paragraphoi in B(03). Following the suggestion of Schironi, we will distinguish between the two markers, “even though they might share a common origin.” 113 When present, the final paragraphoi can take the simple straight–form (—), but are most often forked ( or ).114 B(03) also uses a variety of patterned line-fillers. When these fillers are present, they most frequently take the form of one to three diple-shaped line-fillers and an obelos (e.g., >>–). Occasionally, the last diple and the obelos are combined and look very similar to the forked-paragraphos ( ). An additional variation of this final line-filler contains a dicolon between the diplai (p. 1202B; >:>:>–). This only happens in books where similar dots can be found in the coronis. The decision of how many diplai were used does not appear to be dependent on the remaining space in the line.115 The text and the line-fillers are usually separated by a dicolon, but, as we discussed earlier, it is unclear when these were added.116 The second form of final line-filler is what Milne and Skeat call the “running spiral” ( ).117 This running spiral can appear either at the end of the line or in the horizontal 113 Schironi, “Book-Ends,” 698; Schironi, To Mega Biblion, 10. 114 The end of Job is unique in having three final paragraphoi ( ). In rolls, the forked paragraphos became the standard for bookends by the first-century CE. Schironi, “Book-Ends,” 698. 115 Since the text of Jude ends with a complete line, the final line-filler is given a line of its own, under the final paragraphos (p. 1444C). The line-filler at the end of Tobit looks like a paragraphos that is forked on both ends ( ). 116 The final line of Judith contains two colons and hybrid line-filler/forked paragraphos (: : ). 117 Milne and Skeat, Scribes, 87. CHAPTER 3 78 cross-section of the coronis.118 Milne and Skeat also noticed that this form of line-filler only appears in the work of scribe A, except for a reversed variation in the coronis of Romans ( ).119 Johnson defines a coronis as “An elaborate marginal sign that marks a major point of division in the text, such as the end of a work.”120 The word κορωνιϲ is evidently derived from κορωνη (crow) because of its bent or hooked form. 121 We are even occasionally left with colophons or epigrams written on behalf of the coronis: • “I am the coronis, guardian of letters.” (P. Lond. Lit. 11: LDAB 1957)122 • “I am the coronis, teacher of the divine doctrine.” (H[015]: LDAB 7152)123 • “I, the coronis announcing the final lap, the most trustworthy guardian of the enclosure of written sheets.” (Anthologia Palatina 12.257)124 From these notes, it is clear that a main function of the coronides is to limit a text, so that no more is added or omitted.125 In B(03) there are coronides following Psalm 150 and Psalm 151 (pp. 713–714). This technique, along with the aforementioned end-title, “Book of 150 Psalms,” ensures that Psalm 151 remains distinct from the Psalter. Every book in the codex ends with a coronis, except for 3 Kingdoms, 2 Chronicles, and Ephesians. Both 3 Kingdoms and Ephesians have forked paragraphoi under the last line, and all three books contain final line-fillers (>>–). These instances of absent coronides are puzzling. None of the three are lacking in space for a coronis. And even if space were limited, the letter to the Philippians, which terminates on the last line of a column (p. 1502B), confirms that the scribes were comfortable adding a coronis in the lower margin. While the rare absence of coronides may point to 118 Only Genesis and Joshua have the running spiral in both the last line and the coronis. 119 Milne and Skeat, Scribes, 88. 120 Johnson, Bookrolls and Scribes, 341; This type of coronis is to be distinguished from the sign that marks crasis in Greek (❜). Dickey, Greek Scholarship, 244; see also Canart, Lezioni, 106; Cribiore, Writing, 83; Mugridge, Christian Texts, xix. 121 See the well-known example of a bird shaped coronis in P.Berol. inv. 9875. Schironi, To Mega Biblion, 16; Bobichon, Le lexicon, 51; cf. EDG, s.v. κορωνη. 122 Malik, P.Beatty III, 1 n. 1. 123 Blomkvist, Euthalian Traditions, 16. 124 Schironi, To Mega Biblion, 16 n.35. 125 Turner and Parsons, Greek Manuscripts, 13. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 79 their secondary nature, it is unclear in either case—whether added simultaneously with the main text or shortly thereafter—why a scribe would have left these bookends unornamented.126 The main shape of the coronides in B(03) constitutes a pair of two intersecting lines. To the left of the final column, the vertical line runs parallel to the bounding line, extending upward from one to five lines of main text and downward, at most, eleven empty-lines. The horizontal line intersects with the vertical, but only protrudes shortly to the left (at most the width of one letter). To the right of the vertical line, it extends no more than half-way across the final column and appears around one to four lines below the end of the main text. The form of the coronides can be broken down further into three categories with variations (Figure 25). The first (type-A), is the simplest and most common, and is present in the work of scribes A, B, and our proposed C. Type-A coronides use a variation of dots, diplai, and running spirals for both vertical and horizontal lines. The top and the bottom of the vertical lines are marked with a capital and a base, which are usually mirrored ( and ).127 Type-B coronides have diplai and the Ƨ- shape of the capitals in the horizontal line. The vertical lines are made from thin horizontal strokes that move from narrow at the top and bottom to wide in the center. Similar to type-A, capitals and bases are often present in this second category. Type-B is only found in the work of scribe B, and does not appear until Hosea. The final category, type-C, is the most elaborate and uncommon of the coronides (8x). The vertical lines of type-C coronides are drawn as narrow columns composed of entwined or braided lines. In the work of scribe B, the columns are found in the same place as the vertical lines of types A and B, but, in scribe A (2x), they are directly under the first letter of the line (pp. 137C, 191A). The horizontal line is also ornate in scribe B, but is composed of diplai or lacking altogether in scribe A. Type-C coronides are absent in the work of our proposed scribe C. 126 All of the extant bookends in א(01) contain coronides. Scrivener believed the end-pieces in B(03) were secondary additions. However, he did not provide much support for this claim. Scrivener, Introduction, 1:106. 127 It is from this Ƨ-shape in the capitals and bases that the early coronis developed. Stephen, “The Coronis,” 3– 4; cf. Smith, Alexandrinus, 128; See P.Sorbonne Inv. 2272b (later 3rd BCE) in Turner and Parsons, Greek Manuscripts, 74– 75 (plate 40). CHAPTER 3 80 Type-A Type-B Type-C FIGURE 25: CATEGORIES OF CORONIDES: TYPES A-C (PP. 304B, 1493A, 1382A) I have already noted that it is difficult to tell when, in the production, B(03) was furnished with coronides. Yet, it is very important for Milne and Skeat that they and the main text were copied by the same scribe. They state that the coronis “amounts to his [the scribe’s] signature,” and appeal to them as the “infallible criterion” for identifying the scribes of the main text. 128 It is even more significant for B(03), because they admit that “were it not for the absolute evidence of the colophons [coronides] one might be tempted to suspect a third hand.” 129 As discussed in Chapter 1, the similarity of coronis designs led them to suggest a shared scribe between Vaticanus and Sinaiticus (scribes A and D respectively).130 It is noteworthy that red ink is occasionally used in the coronides, but only in the books which have red section numerals (Isaiah–John). The exception to this is the book immediately before Isaiah (Malachi) and immediately after John (Acts). It is likely, therefore, that B(03) received the red ornamentation, if not the whole end-piece, after the main text was finished and during the addition of the red numerals.131 The fact that the type-A coronides in this section alternate between red and apricot dots suggests that they were copied simultaneously and not later embellished with red. If added slightly after the copying phase, it is not guaranteed that the same scribe copied the coronis as the main text. This may better explain the absence of coronides in 3 Kingdoms, 2 Chronicles, and Ephesians. 128 Milne and Skeat, Scribes, 27–28. 129 Milne and Skeat, Scribes, 89. 130 Milne and Skeat, Scribes, 89–90. 131 In the coronides of Lamentations, Ezekiel, John, and Acts, one finds staurograms (⳨) in red ink. Giambattista De Rossi examined these staurograms in search of the origin and date of the manuscript. Since then, this line of evidence has not been followed in any detail. de Rossi, “Monogrammatica,” 135–140. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 81 FIGURE 26: RED AND APRICOT CORONIS IN MATTHEW (P. 1277B) Nonetheless, Milne and Skeat identified some compelling patterns that do appear to align with changes we have already noted in the codicology and paratextual features. At the end of this chapter, we will have to ask whether or not “the absolute evidence” of the coronis was properly interpreted by Milne and Skeat. 3.1.7 Summary In the preceding discussion of paratextual features in B(03), we have attempted to wade through the various layers of the earliest marginalia.132 Unlike Versace’s larger project, we are most interested in which features can be attributed to the scribes of the main text, and whether or not there are patterns that assist the identification of scribe changes. When it comes to the textual divisions in the codex, it was argued that only the spacing and ektheses can be confidently attributed to the scribes. Indeed, it was in observing the ektheses that Ezra Abbot first noted a change of scribes at 1 Kingdoms 19:11, a position that has been corroborated by the codicological evidence and, now, in the running titles. However, this change in practice of ekthesis becomes less consistent in later parts of the manuscript and is, therefore, of less use. Line-fillers were also examined, as there is a large shift in frequency after 1 Kingdoms 19:11. Yet, when we turned to Esther–Tobit, we noticed that the frequency did not match either scribes A or B. If line-fillers are indicative of the scribes, then this could possibly point to a third hand (scribe C). Finally, we discussed the patterns and irregularities in the various titles and coronides. The discrepancy in running titles at 1 Kingdoms 19:11 and the beginning of Hosea confirmed Milne and Skeat’s suspicion of scribe transitions. However, since running titles are not present in the Psalter, they could not confirm a change at Psalm 1:1. The 132 We could, of course, discuss the so-called marginal distigmai (··) in B(03), but the recent analysis of Versace suggests they were the work of a sixteenth–century hand, his B33, which likely copied the Arabic, Vulgate chapter numbers. The distigmai were then reinked shortly after their addition, along with other scholia that Versace situates in the sixteenth century. Versace, Marginalia, 66, 68–70. See the previous debates in Payne and Canart, “Originality”; Payne and Canart, “Distigmai"; Niccum, “The Voice,” 245; Head, “Marginalia”; Gravely, "Sigla," 66–71; Fellows, “Distigme- Obelos," 246–51; Krans, “Paragraphos,” 252–57; Knust and Wasserman, To Cast, 127 n. 99. CHAPTER 3 82 coronides, the scribes’ signature, were also used to identify Milne and Skeat’s two scribes. But before we let this be the final word on the number of scribes, we must first turn to the paleography of B(03). 3.2 The Problem of Paleography In Chapter 1, we introduced Milne and Skeat’s diagnosis of the problem in B(03): In the Sinaiticus the shapes of individual letters have… proved to be the least satisfactory criterion between hands… It is doubly so of the Vaticanus, where the original script has been almost entirely overwritten by a later hand…133 Following this remark, they begin their analysis of the codex without a second glance at the paleography of the scribes. One should recognize how the results of their study of א(01) affect the decision not to investigate a particular feature in B(03). 134 Indeed, they are correct that the reinforced ink does obscure most of the original handwriting,135 but it does not follow that if the original ink were visible that it would still “be the least satisfactory criterion between hands,” as it was in א(01). We must also recognize that, while most of the text was reinked, there remains a significant number of examples where the reinforcer left the original ink untouched. This evidence has been left untapped and must be analyzed before we can decide whether or not it is useful for identifying the scribes. The following sections will discuss some of the paleographic features in B(03) and how they have been used to identify and depict the scribes. After this, the framework constructed from the codicology, paratextual and paleographic features will guide our analysis of the script in places with unreinforced text. The results of this study point to the likely presence of a third scribe, which, while noted by earlier scholars, Milne and Skeat were not able to identify on account of their limited method. 3.2.1 Nomina Sacra The origin and function of nomina sacra (sg. nomen sacrum; ns) have been a factor of considerable debate in recent decades. 136 These issues are of less importance for this study and need not be 133 Milne and Skeat, Scribes, 87. [Italics added] 134 This also explains the lack of discussion over the running titles in B(03), since they noticed some variation between the scribes of the main text and of the running titles in א(01). 135 Swete laments over the instaurator for “spoiling its [the text’s] original beauty.” Metzger also makes this complaint. In 1739, Wagstaffe describes the appearance of the over-writing as “in some places… wretched and very modern.” Swete, An Introduction, 128; Metzger, Manuscripts, 74. Wagstaffe, “Some account,” 102r. 136 Roberts, Manuscript, 26–48; Tuckett, “‘Nomina Sacra’,” 431–458; Hurtado, Artifacts, 95–134; Luijendijk, Greetings, 57–78; Nasselqvist, Public Reading, 33–34. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 83 detailed again.137 It is clear, however, from the rapidly growing collection of studies on individual manuscripts, or groups of manuscripts, that examination of nomina sacra is an important factor in understanding the copyists.138 We progress with caution in light of Parker’s warning that “a study which seeks too much information from the nomina sacra is of questionable value. One must remain more than doubtful as to whether they convey any theological significance. But they do betray something of a manuscript’s antecedents and of its scribe’s own habits.”139 Milne and Skeat based their brief discussion of B(03)’s nomina sacra on Ezra Abbot’s analysis of the contraction πνευµα (π̅να̅̅, π̅νϲ̅,̅ π̅νι̅)̅ in the early portion of the manuscript. Abbot noted that “In the first 294 pages [Gen 46:28–1 Kgdms 19:11] of the Vatican MS… πνευµα occurs forty-two times, in forty of which it is contracted; in the next 290 pages [1 Kgdms 19:11–2 Esd] it occurs forty- one times, in forty of which it is not contracted.”140 This, again, confirms Abbot’s suspicion that there is a change of scribes at 1 Kingdoms 19:11. Although he does not give the statistics, Abbot also notes that the same pattern can been seen with the word ιϲραηλ (only as ιϲ̅λ̅̅, except for ιη̅̅λ̅ at Ps 146:2).141 Specifically, of the fifty-three occurrences of ιϲραηλ in 1 Kgdms 1:1–19:11, only two are written plene (ns: 51/53), both of which appear in the same column (p. 323A). From the next page until the end of 1 Kingdoms there is not a single abbreviation in twenty-nine occurrences of ιϲραηλ (ns: 0/29). Similar frequencies of abbreviation are present in Exodus (ns: 152/173)142 and Leviticus (ns: 54/66). The exception to this pattern is the remaining leaves of Genesis, which do not contain any abbreviations of ιϲραηλ in its 20 occurrences (ns: 0/20). It seems likely that changes in pattern between books (like that from Genesis to Exodus) are caused mostly by different exemplars. We see similar deviations in Numbers–Joshua, where ιηϲουϲ (ιϲ̅,̅ ιυ̅̅, ιν̅)̅ is mostly abbreviated in Numbers (ns: 8/11), never abbreviated in Deuteronomy (ns: 137 For lists and variations of the nomina sacra, see Traube, Nomina Sacra, 88–121; Paap, Nomina Sacra; O’Callaghan, Nomina Sacra, 41–70; Aland, Repertorium, 1:420–428. 138 Parker, Codex Bezae, 97–106; Jongkind, Scribal Habits, 62–84; Smith, Alexandrinus, 219–225; Malik, P.Beatty III, 60–64; Ebojo, “A Scribe,” 323–366. 139 Parker, Codex Bezae, 106. [italics added] 140 Abbot, “Antiquity," 196 n.*. 141 Abbot, “Antiquity," 196 n. *. 142 Cf. Gurtner, Exodus, 8. CHAPTER 3 84 0/10),143 and mostly plene in Joshua (ns: 11/171). However, nine of the eleven nomina sacra in Joshua are copied at line endings, and eight appear in the conflated form: ιη̅̅ϲ ̅and ιη̅̅ν ̅(Table 9).144 Unless B(03) was copied from exemplars with similar line endings, these are likely to be the occasional decision of the scribe to avoid splitting the word. We do not find the conflated form of ιηϲουϲ anywhere else in the codex. This is particularly fascinating since the form appears in the book of ιηϲουϲ and the occurrences are in clear reference to Joshua the son of Nun (ναυη). While it is tempting to say that the scribe was intentionally using the suspended form because he was consciously distinguishing the sacredness of the name, this cannot be the case. For in Numbers, ten of the eleven references to ιηϲουϲ are qualified by his relationship to his father, Nun; seven of the ten are copied as nomina sacra in the regular contracted form.145 TABLE 9: NOMINA SACRA (ΙΗΣΟΥΣ) IN JOSHUA Line-Endings Total ιϲ̅ ̅/ ιυ̅̅ / ιν̅̅ 1 3 ιη̅̅ϲ ̅/ ιη̅̅ν̅ 8 8 plene 160 Abbot notes that we can also find abbreviations of δαυειδ (only as δ̅α̅δ̅) in B(03), but he does not mention how frequently or by which scribes.146 If he had, it would have provided additional support for his identification of a scribe change at 1 Kingdoms 19:11. Before this transition, we find δαυειδ abbreviated four times at line breaks, presumably to avoid separating the name; there are no examples of δαυειδ being separated across lines.147 In contrast, after the change of hands (p. 335), there are no occurrences of the nomen sacrum and thirty-four examples of δαυειδ split at line breaks (always as δαυ/ειδ).148 An attempt to avoid abbreviation is visible on p. 339B l. 6, which extends into the margin so that δαυειδ could be copied in full at the line-end. Again, the reserved use of abbreviations at line breaks can plausibly be attributed to the scribe, and, while four occurrences 143 One occurrence of plene ιηϲουϲ appears in a column-end correction at p. 233A. Versace attributes this to his hand B2, the diorthotes. Versace, Marginalia, 75, 123. See Chapter 4 for more discussion on the corrections. 144 Traube only counts seven occurrences of the conflated form. See Joshua 10:24 (2x), 42; 15:13; 22:1; 24:2, 21, 30. Traube, Nomina sacra, 66–67. 145 Contra Paap, who states that ιηϲουϲ is abbreviated “only if sacral” in B(03). Paap, Nomina Sacra, 119–120. 146 Abbot, “Antiquity," 196. 147 Statistically, it is unfortunate that David does not appear as a character until 1 Kingdoms 16. Therefore, δαυειδ appears only 57 times in scribe A (4x at line ending), but 199 times in Scribe B (34x at line ending) of 1 Kingdoms. 148 See p. 334C l.20 for a nomen sacrum and the next page, p.335C l.16, for a split δαυ/ειδ. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 85 may not be significant, it conforms to a pattern that has already been observed with πνευµα and ιϲραηλ– i.e., abbreviation before 1 Kingdoms 19:11, and plene afterwards. It should be noted that this distinction is not carried out through the rest of B(03), and abbreviations of πνευµα, ιϲραηλ, and δαυειδ can be found elsewhere in the work of scribe B.149 Other Nomina Sacra in B(03) Peter Malik has produced a helpful survey of the nomina sacra in the NT portion of B(03) (Table 10). Rather than repeat this information, it seems useful to summarize and compare Malik’s data with that of the Greek OT. TABLE 10: SUMMARY OF NOMINA SACRA IN THE NT OF B(03) 150 nomina sacra plene χριϲτοϲ 468 2 θεοϲ 1109 11 ιηϲουϲ 837 5 κυριοϲ 624 30 πνευµα 13151 332 ιϲραηλ 1 63 πατηρ 1 397 υιοϲ 0 359 ανθρωποϲ 0 500 µητηρ 0 80 δαυειδ 0 54 ουρανοϲ 0 214 ϲωτηρ 0 14 ιερουϲαληµ/ ιεροϲολυµα 0 0 72 62 ϲταυροϲ/ ϲταυροω 0 0 24 45 Malik summarizes his findings as follows:152 1. “Vaticanus exhibits an impressively consistent system of abbreviating the main four nomina sacra (θεόϲ, κύριοϲ, ἰηϲοῦϲ, and χριϲτόϲ)”; 149 For example, δαυειδ is abbreviated at line-ends twice on p. 414Β (3 Kgdms 18–19). 150 Adapted and corrected from Malik, “Nomina Sacra,” 97–101. Number of occurrences in the hundreds may vary slightly. 151 Malik only found nine occurrences of abbreviated πνευµα. It is likely that although he noted the three forms (π̅ν̅α̅, π̅ν̅ϲ,̅ π̅ν̅ι)̅, he only searched the transcription for π̅ν̅ι ̅(6x) and π̅ν̅α̅ (3x). The genitive, π̅ν̅ϲ,̅ appears four times, giving a total of 13 nomina sacra. One of the occurrences of the genitive abbreviation occurs in the last line of Philippians (p.1502B). Since πνευµα is rarely abbreviated in the NT, and there is no room to write out πνευµατοϲ without starting an entirely new column, this can likely be attributed to the scribe, rather than the exemplar. 152 Malik, “Nomina Sacra,” 101–102 CHAPTER 3 86 2. “[T]he overall impression is that of a scribe who copied his Vorlage with extreme care and consistency, while consciously following the meaning of the text being copied”; 3. "Thirdly, the use of nomina sacra in Vaticanus may also shed light on the manuscript’s (or its Vorlage’s) provenance. C.H. Roberts has observed that early Latin manuscripts exhibit consistent employment of only the four core nomina… Vaticanus reflects precisely this Tendenz, which could render additional support to the hypothesis of the Western provenance…” While the Greek OT portion of B(03) also maintains consistent abbreviation of θεοϲ and κυριοϲ, we cannot say as much for ιηϲουϲ and χριϲτοϲ. TABLE 11: ΙΗΣΟΥΣ AND ΧΡΙΣΤΟΣ IN THE OT OF B(03) nomina sacra plene ιηϲουϲ 26 238 χριϲτοϲ 27 11 The frequency of abbreviated ιηϲουϲ is much lower than in the NT. This may be expected since the referent is not Jesus but Joshua. However, as we have already noted, Joshua is often abbreviated even when he is specified as the “son of Nun”. Χριϲτοϲ, spelled χρειϲτοϲ in B(03), is regularly abbreviated in the Greek OT, but often written plene. Two of the eleven plene examples are in the genitive plural (1 Chr 16:22; Ps 104:15). Although Rahlfs’ edition prints the accusative plural χριϲτουϲ at Habakkuk 3:13, B(03) has an accusative singular written plene. None of the five occurrences of χριϲτοϲ in Leviticus are abbreviated. While Malik’s first conclusion is still warranted, the Greek OT of B(03) demonstrates greater irregularity. This too is expected, since we are observing the work of more than one scribe in the Greek OT. Malik’s second conclusion is difficult to verify from the nomina sacra in the Greek OT alone. The lack of abbreviated ιϲραηλ in the extant portion of Genesis, in contrast to the consistent contraction in Exodus–1 Kingdoms 19:11, may support the conclusion that the scribe carefully copied the exemplar—that is, if the exemplar was different from Genesis to Exodus. Yet, the numerous differences in abbreviation throughout 1 Kingdoms, which also align with changes in codicological and paratextual features, indicate the work of two separate scribes and their different approaches to nomina sacra, not their different exemplars. Nevertheless, whether or not the scribe abbreviated a word or expanded a nomen sacrum from the exemplar does not tell us much about the overall carefulness of the scribe. This question will be examined in detail in Part II of this project. The most alluring conclusion of Malik’s is the observation that B(03)’s almost exclusive abbreviation of the “four core nomina” might lend support to the hypothesis of Western or Roman provenance (see Chapter 1). While this evidence may still support the hypothesis for the exemplars PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 87 of B(03), this is clearly unhelpful in the Greek OT. As we have seen already, ιϲραηλ and πνευµα are regularly abbreviated in the work of scribe A, with δαυειδ also written as a nomen sacrum at line- ends. In fact, of all the words in Table 10 the only two that are never abbreviated in B(03) are ϲωτηρ and ϲταυροϲ/ϲταυροω.153 Ουρανοϲ is copied once as ου̅̅νο̅ν̅ ̅ at Ps 145:6 (p. 711B). I have found three occurrences of υιοϲ as a nomen sacrum: υ̅ιυ̅̅ at 1 Kingdoms 1:1 (p. 309B), υ̅ϲ ̅at Numbers 2:29 (p. 141A), and υ̅ϲ ̅at Ecclesiastes 10:17 (p. 761A). However, the first two examples are corrections (see Chapter 6). Μητηρ is abbreviated six times (as µ̅η̅ρ,̅ µ̅ρϲ̅,̅ µ̅ρα̅̅), although Traube only counts four;154 πατηρ is abbreviated eighteen times (as π̅η̅ρ,̅ π̅ρϲ̅,̅ π̅ρι̅,̅ π̅ρα̅̅). Traube gives a count of fourteen abbreviations of ιερουϲαληµ (as ιλ̅̅η̅µ̅ and ιλ̅̅µ̅) and thirty of ανθρωποϲ (as α̅νο̅ϲ̅,̅ α̅νω̅̅, α̅νο̅ι̅,̅ α̅νο̅υ̅̅ϲ)̅.155 One example of ιλ̅̅η̅µ̅ (p. 254A, Josh 12:10) occurs in a list format, where the abbreviation is clearly dependent on the scribe’s desire to avoid copying over a line break. The appearance of these nomina sacra in the Greek OT, and the change in frequency of abbreviations of ιηϲουϲ and χριϲτοϲ, suggests that the potential influence of the Latin tradition that Malik finds in the NT, does not hold true throughout the whole manuscript. However, we will return to these less common nomina sacra at the end of the chapter to see if they indicate any more about the scribes of B(03). 3.2.2 Numerical Abbreviations Nomina sacra are not the only form of abbreviation found in B(03). The recent work of Zachary Cole has brought to scholarly attention the importance of numbers and numerical abbreviations in NT Greek manuscripts. Cole summarizes one of the chief conclusions of his study as follows: Christian scribes deliberately adapted the alphabetic numeral system and—with surprising consistency— avoided the abbreviation of certain categories of numbers, all with the aim to produce codices that could be read aloud with ease and without ambiguity—what I term a distinctive “Christian number-writing technique.” 156 By categories of numbers, Cole specifies that scribes of NT Greek Manuscripts typically avoided the abbreviation of the number “one,” ordinal numbers, numbers with inflected forms, and those in the thousands.157 153 This is also noted in Traube, Nomina sacra, 66–67. 154 Psalms 138:13; Proverbs 1:8; 15:20; 20:9; Ecclesiastes 5:14; Sirach 4:10. 155 Traube, Nomina sacra, 66–67. 156 Cole, Numerals, 227. 157 Cole, Numerals, 221. CHAPTER 3 88 In his examination of the NT in B(03), Cole finds only one example of a numerical abbreviation (p. 1283C, Mark 5:13). This singular occurrence is intriguing because the scribe seems to have originally copied the number ϥ̅β̅ (92) or /ιβ̅̅ (10,002), but a later hand corrected it to /β̅ (2000; see Chapter 6).158 The numeral also appears near the end of a line, followed by a και-compendium and an epsilon that protrudes into the right margin. However, whether or not this had an effect on the scribe’s decision is difficult to tell. Furthermore, Cole identifies one example of a singular reading in B(03), which likely points to the presence of abbreviations in the exemplars, at least for the book of Acts.159 He ends by describing B(03) as “the earliest example (fourth century) of a NT majuscule manuscript that consistently avoids using numerical shorthand” and calling this habit “an intentional policy.”160 Does this description hold up for the manuscript as a whole? Cole is careful to note that the Greek OT portion “is not rigorous in using longhand number forms,” and gives eight examples of numerical abbreviation in Numbers alone.161 As we have already discussed, Numbers was copied by scribe A, so this could be attributed simply to differences between two scribes. In examining Genesis 46 through 2 Esdras, I have found 43 examples of numerical abbreviation: thirty-six in scribe A and six in scribe B.162 There are no abbreviations in the whole of the Psalter, with the next example appearing in Ecclesiastes 11:2 (p. 761B). 158 Cole suggests ιβ̅̅ as the first-hand reading, but he does not explain the irregular spacing in front of the numeral (see p. 208). Cole, Numerals, 99. 159 In Acts 27:37, B(03) reads πλοιω ωϲ εβδοµηκοντα εξ. Cole convincingly suggests that the exemplar read πλοιωϲο̅ϛ̅̅, giving the number διακοϲιαι εβδοµηκοντα εξ (with the initial text). Cole, Numerals, 99. 160 Tischendorf, however, identified this correction with a tenth or eleventh-century hand: B3. On Tischendorf’s correctors, see Chapter 4. Cole, Numerals, 100; Tischendorf, Vaticanum, 49. 161 Cole, Numerals, 98 n. 28. 162 A seventh abbreviation appears in Ezra 2:38 (p. 595B) as a correction, protruding left into the margin. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 89 TABLE 12: NUMERICAL ABBREVIATIONS IN GENESIS–2 ESDRAS SCRIBE A Exod 15:27 ιβ̅̅ Num 29:32 ιδ̅̅ Exod 26:18 κ̅ Num 31:5 ιβ̅̅ Lev 23:39 ζ ̅ Num 31:39 φ̅ Num 1:21 φ̅ Num 33:9 ιβ̅̅ Num 7:2 ιβ̅̅ Deut 1:23 ιβ̅̅ Num 7:3 ιβ̅̅ Josh 21:4 ιγ̅̅ Num 7:84 (2x) ιβ̅̅; ιβ̅̅ Judg 3:31 χ̅ Num 7:87 (2x) ιβ̅̅; ιβ̅̅ Judg 14:10 ζ ̅ Num 17:6 ιβ̅̅ Judg 20:24 β̅ Num 26:31 χ̅ Judg 20:25 (3x) β̅; η̅ και ι ̅ Num 26:47 χ̅ SCRIBE B Num 29:2 ζ ̅ 1 Esd 2:10 /β̅υ̅ Num 29:13 ιδ̅̅ Ezr 2:37 /α̅ πεντηκοντα δυο Num 29:15 ιδ̅̅ Ezr 2:38 (n. 180) /α̅ διακοϲιοι τεϲϲερακοντα επτα Num 29:17 (3x) ιβ̅̅; β̅; ιδ̅̅ Ezr 2:58 το̅β̅̅ Num 29:20 (2x) ια̅̅; ιδ̅̅ Neh 7:33 ρ̅ (correction?) Num 29:23 ιδ̅̅ Neh 7:41 /α̅ϲ ̅τεϲϲαρακοντα επτα Num 29:26 (2x) β̅; ιδ̅̅ Neh 7:70 /α̅ Neh 10:32 ζ ̅ While this shows a higher tendency by scribe A to abbreviate numbers, the seven examples in scribe B do not reveal the same “intentional policy” as in the NT. To be sure, seven abbreviations (all in 1– 2 Esdras) is hardly overwhelming, and still displays a preference for longhand numbers. But the examples above also break three of Cole’s four restrictions (there are no abbreviations of “one”). There are three ordinal abbreviations (Judg 20:24, 25; Neh 10:32); three numerals with inflected endings in the accusative (Judg 3:31; Neh 7:70, 10:32) and one feminine (1 Esd 2:10). Values in the thousands are abbreviated five times (1 Esd 2:10; Ezr 2:37, 38; Neh 7:41, 70). Although Brooke–McLean give the reading of B(03) in Nehemiah 7:33 as ναβιαα ρ ̅πεντηκοντα δυο (152), and the Gottingen edition prints ναβια α̅ ρ ̅πεντηκοντα δυο, it is likely that the first hand reading is actually ναβιααρ πεντηκοντα δυο (52)—the MT reads נבו אחר חמשים ושנים .163 However, when the breathings and accents were added to the codex, the reinker added a circumflex above 163 Hanhart recognized this original reading in B(03). Pete Myers gives a reconstructed text ם נבי אחר המשי ושנים , even though he copies the transcription of B(03) from the Göttingen edition. Hanhart, 2. Esrabuches, 348. Myers, “Transcriptions,” 481. CHAPTER 3 90 the final alpha and a macron above the rho, which was then separated by two middle dots (Figure 27). The resultant reading agrees with the number εκατον πεντηκοντα δυο in A(02), א(01), and 58.164 FIGURE 27: ΝΑΒΙΑΑ·Ρ ̅·ΠΕΝΤΗΚΟΝΤΑ ΔΥΟ (P.614C, NEH 7:33) That such a correction is probable is made clear in the same column (p.614C l. 22), where we find another numeral correction at Nehemiah 7:38. Here, ϲανανατ is corrected to ϲανανα /γ̅, with part of the final tau (T) erased to make a gamma (Γ). In verse 34, we also find a parallel to ναβι-ααρ ( נבו אחר ) with ηλαµ-ααρ ( עילמ אחר ).165 This makes it clear that the scribe of 2 Esdras did not originally intend the final rho to be a numeral, but it was later corrected to one. The only other witness to the number fifty-two is Rahlfs 55, a close relative of B(03). According to Hanhart, Rahlfs 122 (15th cent.) is an abschrift of B(03), but it was likely copied after the correction was made and so reads 152.166 Again, it is difficult to know whether the numerals that do appear in B(03) were copied from the exemplars or created by the scribe. The descendants of Pashur (1,247) are numbered three times in the Greek OT (1 Esd 5:25; Ezr 2:38; Neh 7:41). The first time it appears in B(03) it is written longhand as χειλιοι διακοϲιοι τεϲϲερακοντα επτα (p. 582B). The following occurrence at Ezr 2:38 is copied as /α̅ διακοϲιοι τεϲϲερακοντα επτα, though the numeral is likely an early correction, as it projects into the left margin (p. 595B). The final example is copied /α̅ϲ ̅τεϲϲερακοντα επτα, with the numeral appearing as the last two characters of the line (p. 614C). In the final two occurrences, therefore, the numerals appear presumably because of omission at a line-break or, as in Nehemiah 7:41, line restrictions, and were likely the creation of the scribe or an early corrector. As a whole, the scribes of B(03) use numeral abbreviations sparingly. From Genesis 46 through 2 Esdras, I found forty-three examples, twenty-five of which appear unsurprisingly in the book of Numbers—containing nearly seven-hundred cardinal numbers. The OT numerals do not entirely follow the restrictions set out by Cole, but it follows that scribe B remains less likely to abbreviate numbers than scribe A. This result largely confirms the evidence that Cole has found in 164 See Myers, “Transcriptions,” 481. 165 Hanhart, 2. Esrabuches, 249. 166 On the relationship between B(03), its abschrift (122), and 55, see Hanhart, 2. Esrabuches, 13, 249; Hanhart ed., Esdrae liber II, 30. Myers names the ancestor of this group B55. Myers, “Transcriptions,” 47. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 91 the NT of B(03). The question remains, whether the patterns of abbreviation change in the latter part of the OT, where I have suggested the presence of a third scribe. If Milne and Skeat were correct about the return of scribe A in Psalms–Tobit, we might expect to see a similar habit of numeral writing. However, this is not the case, as there only appear to be four numerals over the 320 pages (p. 761B, Eccl 11:2; p. 770B, Job 1:17; p. 890A, Sir 49:10; p. 896B, Esth 2:12). The first two numerals occur at line-ends, the latter of which is possibly a correction, on account of the irregular spacing before the numeral (p. 770B, Job 1:17). Since there are no numerals in the text of Psalms, they cannot serve as criterion for distinguishing a scribal division in the book. Moreover, the rare practice of abbreviating numbers in Psalms–Tobit corresponds to that of scribe B and complicates the comparison. If scribe A was responsible for these books, one would have to conclude that there was a change in habit or that the numerals originated in the exemplars. 3.2.3 Ligatures and Unique Letter-forms In addition to nomina sacra and Greek numerals, the scribes of B(03) regularly use other forms of abbreviation called ligatures. The two most frequent ligatures in B(03) are the και-compendium ( ) and the combining of letters through shared vertical strokes (e.g., ΝΗ → ), both of which are employed by all of the scribes. Although less frequent, the word µου is occasionally abbreviated at the end of lines ( ). The rarest abbreviation is found at line-ends in the first half of Psalms with words ending in -ται ( ).167 There is very little difference in the use of the και-compendium between scribe A and B. Although this abbreviation can occur anywhere in the line, it is most frequently found at line endings. It is typically used for the word και, but also appears in words such as δικαιοϲ (δι ον; -ω on pp. 645Β, 657Β). In the twenty-six pages from 1 Kingdoms 1:1 to 19:11, I have found fifty-six occurrences of the και-compendium at line-ends (2.2/pg.), and in the remaining nineteen pages of 1 Kingdoms, forty-nine occurrences (2.6/pg.). There is, however, a distinct form of the και- compendium in the latter half of Psalms through Tobit. In this unique form, the oblique stroke extending down from the kappa ends with either a curl or straight line to the right. 167 Similar to this abbreviation, there is one example of τηϲ copied as τ’ in a correction at Isaiah 35:10 (p. 1034B; Scribe B). CHAPTER 3 92 FIGURE 28: THE TAILED ΚΑΙ-COMPENDIUM IN PSALMS-TOBIT (PP. 940C, 921B) While this is not the only form of the abbreviation in Psalms–Tobit, it appears regularly and with a particular concentration in Esther–Tobit. Yet, after Tobit, the tailed form ceases to appear and the regular form of scribes A and B becomes the standard. There are twenty occurrences of the regular και-compendium at line-ends in the ten pages containing Hosea (2/pg.)—a similar frequency to that in 1 Kingdoms. This change from Tobit to Hosea, again, aligns with the scribe change at Hosea 1:1. However, the fact that the tailed και-compendium does not appear in the work of scribe A, from Genesis 46 to 1 Kingdoms 19:11,168 might be evidence of a different hand. When the scribes of B(03) reached the end of a line, they occasionally combined letters by their vertical strokes to save space.169 The most common letter that is built upon is nu, though eta and mu also appear as the base of letter combinations. From the base later, there are examples of added gammas, etas, kappas, mus, and nus. These combinations can also span multiple letters, the longest of which contains four letters and a µου-ligature (Figure 29). FIGURE 29: FOUR-LETTER COMBINATION AND ΜΟΥ-LIGATURE (P. 839A) From 1 Kingdoms 1:1–19:11, there are twelve two-letter combinations (11 νη; 1 µη), with only one example after the scribe change (1 νη). While scribe B can combine letters, there are far fewer examples in these sections of the codex. I did not find any instances in 2 Esdras, though three ligatures are present (3 νη) in Hosea, where scribe B started copying again. There are fourteen examples of letter combinations in Psalms (9 νη; 1 νµ; 1 νην; 1 ηγ; 2 ν ). 168 If anything, the και-compendiums in this portion have slight serifs to the left. 169 For the same feature in א(01), see Jongkind, Scribal Habits, 84–87; Likewise, I have found one example of a µου-ligature copied directly above the final letter of a line, in order to avoid further extension into the margin (p. 640A l. 15). PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 93 I briefly mention the abbreviation of words ending in -ται ( ),170 because I have only found ten examples, all of which appear in the first half of the Psalter.171 While it is possible that others have gone unnoticed, their concentration in pp. 638–662 and their absence in the rest of Psalms is further evidence in favor of a scribal division later in the book, rather than at the opening. Our final abbreviation is the µου-ligature,172 which takes two forms in B(03). The first, as shown above, uses the standard mu of the Biblical Majuscule bookhand ( ; Figure 29), and the second takes the shape of the so-called “Coptic mu” ( ).173 While scribes A and B use both styles of mu in their writing, they do show different preferences in how they copy the µου-ligatures. We are fortunate that these variations are not obscured by the reinforcement of text. In Genesis 46 through 1 Kingdoms 19:11 (294 pages), I have found only three examples of the µου-ligature, and an additional two in early corrections. Two of the ligatures have the standard mu ( ; pp. 239C, 273A) and one appears to be a conflation of both forms, with two vertical strokes, but a curved center (p. 90B). The two corrections on p. 89B and p. 121B appear in the Coptic style. The overall impression is that scribe A prefers not to use the µου-ligature. However, when it does appear, it is in the standard style or a mutation of it. We will return to whether or not this pattern returns for the later texts, where either scribe A or a third scribe was responsible. Scribe B, on the other hand, uses the µου-ligature ten times, from 1 Kingdoms 19:11 through 2 Esdras (290 pages). This time, every occurrence takes the form of the Coptic mu ( ).This consistent change in style and frequency provides additional support for a change of scribes at 1 Kingdoms 19:11. I have also found seventeen examples of the ligature, in the Coptic style, through the whole of the NT. Yet, when we look at the first half of Psalms, this pattern does not change, as if it were the work of scribe A or a third scribe. Instead, the pattern continues with the Coptic style ligature at an even higher frequency. From p. 625 to p. 674 (Psalms 1–77; 50 pages), there are at least sixty-three occurrences of the µου-ligature, and all except for two are in the Coptic style (pp. 650A, 170 For this abbreviation, see Thompson, An Introduction, 81; Kenyon, “Abbreviations,” 128; Gardthausen, Griechische Palaeographie 2:327; Canart, Lezioni, 88. 171 There is also one correction with the -ται abbreviation (p. 1246C; Matt 10:14). 172 This is regularly used for the word µου or εµου, but can also be used to complete words that end in µου. See examples of κοϲ on pp. 1373B, 1428C (2x), and 1467B; also, νο on p.1489B and οικτειρ on p.1505a. 173 Both forms are also present in א(01). Jongkind, Scribal Habits, 87. CHAPTER 3 94 656B).174 It is not until p. 678 (Psalm 83) that we find a switch to the standard mu form of the ligature. From Psalm 83 to 150 (27 pages)—Psalms 106–137 are not extant— there are fourteen examples of the standard form µου-ligature. The 180 pages from Proverbs through Sirach contain nineteen µου- ligatures in standard form, though only one appears in Esther–Tobit (p. 900A). This decline in frequency is comparable to that in Genesis 46–1 Kingdoms 19:11. After the ligature in Esther, the next occurrence is not until Isaiah 42 (p. 1041C), in the Coptic style. Four pages later there is another µου- ligature, but in the standard form (p. 1045A). Thus, from Hosea (the return of scribe B) to the end of Daniel (290 pages) there are thirteen examples of abbreviated µου: ten in the Coptic style and three with the standard mu. This is nearly the same frequency as that found in 1 Kingdoms 19:11–2 Esdras, but also demonstrates that scribe B could use the standard µου-ligature. If, indeed, the µου-ligature gives any indication of the individual scribes, then the lack of change in the first half of Psalms may be evidence that the scribes did not change until later in the book. Although the -ται ( ) ligature does not occur with much frequency, its appearance is restricted to the first half of Psalms and may therefore corroborate the evidence of the µου-ligature. We have already noted how Milne and Skeat’s dependence on the coronides limited their ability to be precise in locating scribe changes in B(03).175 They too, however, admit that scribe B may have continued copying into the Psalter.176 It is noteworthy that though Milne and Skeat were aware of Traube’s scribal divisions in B(03), they do not evaluate his argument for a transition of scribes in Psalm 77, rather than at the opening of the book.177 In the concluding sections of this chapter we will examine a final piece of evidence that has not yet been explored, in hopes of answering this question. It is not until we bring all the evidence back together that we will see a clearer picture of the scribes of B(03) and their division of labor. 3.2.4 Paleography of Un-reinforced Text What we have found repeatedly, in the last two chapters of this study, is that the evidence clearly points to the first change of scribes (from A to B) at 1 Kingdoms 19:11 (p. 335). While this is hardly 174 Both of these use the standard mu. They are, however, in the same column with two µου-ligatures in the Coptic style and so it is especially clear that these are anomalies. 175 This is why the only scribe changes they located independently were at the beginning of books (i.e., Psalm 1:1 and Hosea 1:1). 176 Milne and Skeat, Scribes, 89. 177 They cite Traube’s division of hands at Matthew 9:5. Milne and Skeat, Scribes, 89 n. 1. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 95 contested, the amount of data, whether codicological, paratextual, or paleographic, has not been comprehensively addressed until now. Indeed, it may not all be needed to identify two scribes in 1 Kingdoms, but it remains an important anchor for any identification of scribe changes throughout the rest of the codex. If this is so, we may at last be able to ask the paleographic question: Is there a noticeable difference in hands between these two sections? We have already noted how the manuscript’s reinking complicates the answer to this question. But is there any way around this? With the right technology, experts might be able to by- pass the newer ink to see what lies below, but for now we must rely on the many places throughout the codex where the original ink has been left untouched.178 I have collected over four-hundred screenshots of individual letters, words, or complete verses that were not re-inked by the reinforcers. This does not include all of the many examples of untouched epsilons or moveable nus. The longest sample of unreinforced text is from p. 199C (Deut 5:14) with seventy-three characters, and another sample on p. 1479B (2 Cor 3:15–16) contains sixty-eight untouched characters (Figure 30). FIGURE 30: SAMPLES OF UNREINFORCED TEXT (P. 199C, 1479B) Since the clearest evidence for a change of scribes is at 1 Kingdoms 19:11, my examination began by comparing unreinforced text on both sides of the divide. From this, I quickly noticed a striking, yet consistent, difference in the way the two scribes copied their alphas, lambdas, and sometimes deltas—there are far fewer unreinforced examples of the latter. 178 See the original hand collage created by the Vatican Library: https://spotlight.vatlib.it/greek- paleography/feature/biblical-majuscule. After completing this chapter, the Vatican website added a brief page examining the two-scribe hypothesis of Milne and Skeat from the perspective of the unreinforced text. They conclude that “the evidence from the script(s) of Vat. gr. 1209 is not incompatible with the two-scribe hypothesis, but that it does not contribute greatly to confirm it; certainly, it seems unlikely that the hypothesis would ever have been formulated on the basis of the script(s) alone.” https://spotlight.vatlib.it/greek-paleography/feature/observations-on-the-original- scribe-s-of-vat-gr-1209 CHAPTER 3 96 Alphas, Deltas, Lambdas Scribe Α ScribeB Scribe A Scribe B Scribe A Scribe B FIGURE 31: ALPHAS, DELTAS, AND LAMBDAS IN SCRIBES A AND B In Amy Myshrall’s contribution to the study of scribes in א(01), the same three letters were found to be indicative of different scribes in the manuscript.179 My identification of the three letters in B(03) was independent of her study, though it may be validated by her claims. However, more recently, Dan Batovici has criticized Myshrall’s theory of an additional scribe based on this paleographic analysis, as the distinguishing features can be found in both her proposed B1 and B2 scribes. The additional features, which she supplies in support of her theory, are “either irrelevant for distinguishing two scribes, or are peculiar enough to point to one scribe,” according to Batovici.180 We must, therefore, be careful with relying too heavily upon this paleographic evidence alone. Nevertheless, it is necessary to test this data against what we have previously gathered to see if it provides additional clarity to the scribal divisions in B(03). Starting with the general orientation of all three letters, it is clear that scribe A tends to slope further to the left than scribe B. In the alphas, deltas, and lambdas, this graphic pattern results in right obliques noticeably longer than the left. While it is not always the case, the left obliques of these letters in scribe A do not usually meet with the top of the right obliques. The two strokes virtually always meet in the work of scribe B. Fortunately, the orientation of these letters is often maintained by the reinforcer and is more perceptible when looking at the page as a whole (Figure 322). 179 Myshrall, “Fourth Scribe?,” 139–148. 180 Batovici, “Two B Scribes,” 197–206. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 97 FIGURE 32: GRAPHIC ORIENTATION OF ALPHAS, DELTAS, AND LAMBDAS: SCRIBES A AND B (PP. 140A, 389B) Out of the three letters, the alphas provide the most consistent evidence and contain the addition of a distinct crossbar in scribes A and B. The crossbars of scribe A are almost always straight, while those in scribe B are curved. We have noted already that the running titles of 1 Kingdoms stop at the same place as the scribe change (p. 335). But when the earlier titles of 1 Kingdoms are compared with those in 2 Kingdoms, the distinction between the alphas and lambdas is still recognizable (Figure 33). FIGURE 33: RUNNING TITLES IN 1–2 KINGDOMS (PP. 334, 362) After tracing these distinctive letters through the codex, from 1 Kingdoms 19:11 and on, I did not notice a change until p. 676A, in the unreinforced word διαψαλµα.181 While this change does take place in Psalms, it is not at the beginning, but in Psalm 79. However, p. 676 is the verso of the folio, and when we turn back to the recto we find that p. 675 starts with the last two verses of Psalm 77, the same location where Traube identified a change of scribes.182 Unlike the beginning of Psalms, this page happens to be the start of a new quire [36]. FIGURE 34: UNREINFORCED ΔΙΑΨΑΛΜΑ IN PSALMS (PP. 671B, 676A) 181 Unreinforced alphas, deltas, and lambdas appear frequently in Psalms because the word διαψαλµα is left untouched, most of the time. I have collected fifty-four examples of unreinforced διαψαλµα before p. 676, and nineteen up to the end of Psalms. 182 Traube, Nomina sacra, 66–67. CHAPTER 3 98 If paleographic analysis of the unreinforced text is valuable for distinguishing hands, then it appears that this is where the change of scribes occurs. This conclusion is not without supporting evidence, as hinted at in our discussion of the different µου and the -ται ( ) ligatures in Psalms. Yet, when comparing the alphas, deltas, and lambdas that appear after p. 675, they did not seem to match those of either scribe A or B (Figure 35). If this is true, then it is possible that the change is not back to scribe A, but to our proposed scribe C. Scribe A Scribe B Scribe C FIGURE 35: THE ALPHAS, DELTAS, LAMBDAS OF THREE SCRIBES? The main distinguishing feature of this hand is the upright orientation of letters and the crossover at the pinnacle of the three distinct letters. While the vertical orientation can be seen, in spite of the reinforcement, there are also places with complete lines of untouched ink, where this can be clearly observed (Figure 36). A similar crossover happens occasionally in the work of both scribes A and B, but it is particularly pronounced and ubiquitous in this section. FIGURE 36: UNREINFORCED TEXT IN SIRACH (P. 835A) After tracing this hand through the Poetic books and Esther-Judith-Tobit, a hand that looks like scribe B returns in Hosea. This confirms Milne and Skeat’s identification of scribe change at Hosea 1:1 (p. 945). Based on the hand alone, it appears that a third scribe (scribe C) was responsible for copying Psalm 77:71b–Tobit (270 pages), meaning scribe B was the copyist of 1 Kingdoms 19:11– PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 99 Psalm 77:71a (340 pages). In agreement with Milne and Skeat, scribe B returned to copying from Hosea through the NT. Whether the conclusion of a third scribe is warranted needs further examination, but the change of hand at Psalm 77:71 is not dependent on it. The Tailed-Mu In addition to the evidence supplied by the alphas, deltas, and lambdas, we have one more distinctive paleographic feature that can be of some help. The frequent occurrence of the letter mu at line-ends provided the scribes of B(03) an opportunity to extend the last stroke of the letter into the right margin. How a particular scribe did so appears to be based on personal preference, opening up the possibility for the examination of patterns between scribes. In most cases, the reinforcers faithfully retraced these tailed-mus. Nevertheless, they occasionally failed to follow the distinctive features of the specific scribe, complicating our analysis. Again, starting with the anchor at 1 Kingdoms 19:11, I compared the occurrences of the tailed- mu across the break. There are over two hundred examples from Genesis 46–1 Kingdoms 19:11, and just over one hundred from then until Psalm 77:71. The clear difference between scribes A and B (Figure 37) is the preference of A to fade or curl the tail downwards, and that of B to extend straight across or curl upwards. While there are some exceptions in scribe A, I have only found five in scribe B (pp. 380C, 407B, 474A, 495B, 594B). These, however, are not clearly the work of the scribe but may be the reinforcer’s addition to the original ink. We do find final mus with a straight tail in scribe B, where the reinker later curled the line downwards (pp. 366C, 376A, 388C, 422B). This makes it difficult to tell whether the five examples are really exceptions. From Hosea–Daniel, there are twelve occurrences of the tailed-mu, five of which have downward curls—p. 1233C clearly shows this in the undertext. This is a strange change in pattern for scribe B, but the majority of examples are unmistakably extending upwards. CHAPTER 3 100 Scribe A Scribe B Scribe C FIGURE 37: TAILED-MU IN SCRIBES A AND B (TOP: PP. 121B, 451B; BOTTOM: PP. 52C,183 482B); CF. SCRIBE C (P. 682B) We are left, then, with Psalm 77:71–Tobit. The tailed-mu appears far less frequently in this section than in the work of either scribe A or B—only twelve examples in 270 pages. All twelve of these occurrences are copied with straight or slightly ascending tails (Figure 37). The severe decline in frequency and the lack of scribe A’s distinctive feature in this portion of the manuscript gives additional support to the hypothesis of a third scribe. Combining the evidence of the unreinforced alphas, deltas, lambdas and the tailed-mu, with the ligatures—mainly the και-compendium and both the µου and -ται ( ) ligature—it is possible to identify more precisely the change of scribes at Psalm 77:71 (p. 675), instead of the beginning of the book. It also exposes the possibility that a third scribe was responsible for the following pages. The uniquely tailed και-compendiums (Figure 28) that were briefly mentioned do not appear until after this break (first on p. 682B), and cease to appear after Tobit (last on p. 943C). However, in order to test the theory of a third scribe, we must return to the evidence that first revealed the change between scribes A and B in 1 Kingdoms 19:11. 3.2.5 Codicology, Paratextual Features, and Nomina Sacra Revisited Codicology In the previous chapter, it was noted that both the proposed changes of scribes at 1 Kingdoms 19:11 (p. 335) and at Hosea 1:1 (p. 945) aligned with observable discontinuities in the production of the codex—namely, changes in line-count and irregular quire structures. These discontinuities also coincided with quire breaks, which could allow for simultaneous copying. Yet, when it came to Milne and Skeat’s proposed break at Psalm 1:1 (p. 625), there was no quire break. Although the presence of a nearly empty page before the start of Psalms could support a scribe change, this likely 183 Here, and in the example to the right, it is possible to see the faint extension of the undertext beyond that of the reinking. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 101 originated in an attempt to avoid mixing the three-column, prose format of 2 Esdras with the two columns of the Poetic books. Therefore, of the three suggested locations of scribe change, this was deemed the least convincing on codicological grounds. However, the return to Traube’s earlier proposal for Psalm 77:71 as the place of transition finds support in the quire break at p. 675 [36]. Although we cannot assume the scribes were responsible for ruling all of their own quires (see §2.3.2), the presence of a third scribe in Psalm 77:71–Tobit would imply that scribe A only copied text with forty-four lines-per-column. Paratextual Features Earlier in this chapter, the line-fillers in B(03) were said to reveal the presence of two scribes in 1 Kingdoms. Scribe A used line-fillers frequently, from Genesis 46–1 Kingdoms 19:11, averaging up to fourteen per page in 1 Kingdoms. Scribe B, on the other hand, only used them twice in 1 Kingdoms. Apart from rare examples, line-fillers do not appear in the Poetic books, so it is not possible to test their frequency until Esther. When examining Esther–Tobit, we noticed that they occur at a higher rate than in the work scribe B, but a lower rate than in that of scribe A. This deviation can be attributed to a change of habit for either scribe A or B, but it likely supports the other evidence in favor of a third scribe. It is unclear whether the scribe of the main text copied the running titles.184 In fact, the evidence presented above seems to indicate they appeared on the quires before the main text. Nevertheless, the running titles before 1 Kingdoms 19:11 appear to match the hand of scribe A, while those in 2 Kingdoms betray the hand of B. If this is any indication of the hand of the main text, we can compare the running titles of 2 Esdras with those of Proverbs (Psalms lacks running titles). FIGURE 38: RUNNING TITLES OF 2 ESDRAS AND PROVERBS (PP. 622, 715) In the titles of 2 Esdras, the distinctive curl of the crossbar in the alphas is clear. It is more difficult to tell with the first running title in Proverbs. While the first alpha is similar to those in Psalm 77:71– Tobit, the second one looks more like the alpha of scribe A. It possible, then, that scribe A is responsible for this portion of the manuscript, and happens to copy two different forms of alphas, 184 See §3.1.6, for comments on the change of hand in Isaiah. CHAPTER 3 102 deltas and lambdas.185 Turning to the main text of Proverbs, it is clear, in spite of the reinking, that the scribe prefers to copy the alphas we identified as coming from a third hand. However, what the examination of unreinforced text did not show, is that this scribe tends to slant the alphas, deltas, and lambdas to the left, when approaching the end of a line. The result of this slant is an alpha comparable to that of scribe A. Therefore, the vertical orientation of the text remains a distinctive of this portion of the text, but it is not strictly maintained along the right bounding lines of the columns. This may well explain the difference between the two alphas in first running title of Proverbs. Still, it remains difficult to identify the running titles of Proverbs with a particular scribe. A unique rounded alpha, which is not found in the unreinforced ink of the main text, can be found regularly in these running titles. FIGURE 39: ROUNDED ALPHA IN RUNNING TITLES (PP. 731, 1339A) These alphas look similar to those found in Old-NumH2 (Figure 39), but may simply be an additional form of alpha belonging to the main scribes. Nonetheless, the evidence of the running titles remains inconclusive. Unfortunately, the lack of these titles in the Psalter does not give us the opportunity to test the change at Psalm 77:71. We did, however, mention Versace’s conclusion that the Psalm numbers were copied in scribendo. While we have not yet provided any evidence in support of this, we may now be able to examine the paleography of the numerals across the proposed scribe change. At least three changes across this break can be identified. First, the numerals in Psalms 71–77 are copied with regular sized omicrons, while those in Psalm 78–79 appear with a compressed and raised omicron. Second, the stigmas have a distinct form in each half of the book (Figure 40). 185 This is the conclusion in Batovici’s critique of Myshrall’s take on א(01). PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 103 Psalm 1–77 Psalm 78–150 FIGURE 40: DISTINCT STIGMAS IN PSALMS (PP. 658B, 671A, 679B, 686B) Finally, the mu in Psalms 40–49 appears in the Biblical Majuscule hand, while those in Psalms 140– 149 are in the “Coptic” style. While these changes do not prove that the Psalm numbers were copied in scribendo, the difference in hand aligns with our proposed scribe change in Psalm 77. Although the transitions in section-numeral hands never align with a change of scribes in B(03), we have noticed coinciding changes in the hand of the main text and in Psalm numbers, which strongly indicates the responsibility of two scribes for copying the Psalter. Nomina Sacra We now return to our survey of the nomina sacra in B(03). It is important to remember that it was Traube’s analysis of the nomina sacra that led him to identify a change of scribes in Psalm 77, as well as the presence of a third scribe (possibly a fourth). We must, therefore, compare the nomina sacra before and after p. 675 to note any inconsistencies due to a transition in scribes. If the nomina sacra support a change in Psalm 77, then the next step is to compare the abbreviations in Genesis 46–1 Kingdoms 19:11 with those in Psalm 77:71–Tobit, to see if the results indicate the work of scribe A or another scribe altogether. In some cases, the frequency of abbreviated words is too high to be comprehensive without a full transcription of the Greek OT. In those circumstances, a sampling is taken from the whole, noting that the evidence is consistent with what is found in the rest of a section. The following chart combines the collected data of this study with that of Traube. CHAPTER 3 104 TABLE 13: NOMINA SACRA IN PSALMS (INCLUDING COMPARISON WITH SCRIBE A)186 NS Ps 1–Ps 77.71a Ps 77.71b–Ps 150 Gen 46–1 Kgdms 19:11 πνευµα o/15 7/12 42/44 ιϲραηλ 4/25 7/17 Exod 152/173; Lev 54/66 ανθρωποϲ 1/57 5/30 (Traube: 26x in Ps 77– Tob) Traube: 4/254 ιερουϲαληµ 0/3 3/5 Traube: 9/15 µητηρ 0/7 1/5 (6x in Ps 77– Tob) 0/68 ουρανοϲ 0/33 1/46 0/76 υιοϲ 0 0 (1x inEccl) 0 Starting with the break at p. 675, the difference between the first and second half of Psalms quickly becomes clear. In Psalms 1–77, the lack of abbreviated words in Table 13 is comparable to that in scribe B’s portion of 1 Kingdoms and the NT. Again, there is a strong contrast in the abbreviation of πνευµα, and a higher frequency of abbreviated ιϲραηλ in the latter half of Psalms. 187 The words ιερουϲαληµ, µητηρ, and ουρανοϲ are only abbreviated after p. 675. Although the sample of nomina sacra is admittedly limited, the consistent difference between the two halves of the Psalter supports the change in the middle of the book, rather than the beginning. Can the nomina sacra help identify the scribe of Psalm 77–Tobit as scribe A or a third hand? To begin, there are two nomina sacra that are unique to this portion of the manuscript: µητηρ and ουρανοϲ. The only time ουρανοϲ is ever abbreviated in B(03) is at Psalm 145:6 (p. 711B). This nomen sacrum likely originated with the scribe rather than the exemplar as it occurs at a line-end, terminating at the inside bounding line. The word µητηρ is only abbreviated once in Psalms, but can be found as a nomen sacrum another five times from Proverbs–Sirach. In addition to these two unique abbreviations, the word ανθρωποϲ is abbreviated significantly more often here than in Genesis 46–1 Kingdoms 19:11. According to Traube, scribe A only copies ανθρωποϲ as a nomen sacrum four times.188 In the thirty extant occurrences of the word in Psalms 77:71–150, I have found five abbreviations. Traube adds to this count twenty-one occurrences in Proverbs–Tobit. Thus, the two 186 This table does not include the main four nomina sacra, giving special attention to the distinguishing abbreviations in B(03). 187 This frequency is not consistently maintained after Psalms. Adams notes that ιϲραηλ appears only once as a nomen sacrum out of eighteen occurrences in Baruch. Adams, Baruch, 29. 188 This is in line with what I have found in Genesis–Numbers (NS 3x). Traube, Nomina sacra, 66–67. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 105 unique nomina sacra, ουρανοϲ and µητηρ, and the noticeably higher preference for abbreviated ανθρωποϲ may well point to the work of a third scribe. On the other hand, scribe A tends to abbreviate the words πνευµα and ιϲραηλ more frequently than the scribe of the latter half of Psalms. While the two words appear significantly more often in Genesis–1 Kingdoms, there is an impressive consistency in abbreviation that is hardly comparable to the end of Psalms. It should be noted that υιοϲ is only abbreviated three times in the whole manuscript, twice in scribe A and once in Psalm 77–Tobit. However, as mentioned above, the two examples in 1 Kingdoms 1:1 (p. 309B) and Numbers 2:29 (p. 141A) are corrections, making the abbreviation at Ecclesiastes 10:17 (p. 761A) the only original nomen sacrum of υιοϲ. A last word must be said about Traube’s proposed fourth scribe and the possible change at Matthew 9:5. Although he mentions a B4 (his siglum), he combines the data with his B2 (scribe B), ultimately making the distinction hard to gauge.189 Unfortunately, Traube’s reliance on the nomina sacra did not allow him to catch a change at Hosea 1:1. Instead, his scribe B3, our scribe C, continued copying until Matthew 9:5 (p. 1245). While this does happen to align with a quire break [65], I have not found any other variation to point to a scribe change. Indeed, the unreinforced text from both sides of the break are undistinguishable. Milne and Skeat note that Traube identifies this break based on the high frequency of abbreviated πνευµα in the first four chapters of Matthew—five nomina sacra, with only eight in the rest of the NT.190 Although these occur successively, we can also find the nomen sacrum later in Matthew (Matt 22:43, 27:50), and in Mark 1:8 (see also 12:36). Therefore, nine of the thirteen examples of abbreviated πνευµα in the NT are found between Matthew and Mark, rendering Traube’s evidence inconclusive. Coronides Again, we end this discussion with the coronides, “the absolute evidence” of the scribes, according to Milne and Skeat.191 They did, indeed, find compelling patterns in the work of scribes A and B, but we must now re-examine the coronides in Psalms–Tobit against those in Genesis–Ruth. Although subtle, there are consistent differences between the two sections, which make the presence of a third scribe possible, in spite of a common style. 189 Traube, Nomina sacra, 66–67; cf. Ropes, Beginnings, xxxviii. 190 Milne and Skeat, Scribes, 89. 191 Milne and Skeat, Scribes, 89. CHAPTER 3 106 First, all of scribe A’s final text-lines contain what Milne and Skeat call a “running spiral” ( ), which acts as an ornamental line-filler. While the running spiral is also present in the coronides of Psalms–Tobit, they never appear as line-fillers at the end of the text. In contrast, only Genesis and Joshua contain running spirals in both the final line and the cross-bar of the coronides. The second difference between the two sections is that scribe A occasionally uses an ornate column (type-C) for a coronis (2/8 coronides), but the scribe of Psalm 77:71–Tobit does not (0/12 coronides). Again, these minor differences may not be definitive on their own. But when combined with the codicological, paratextual, and paleographic evidence, the probability that a scribe C was responsible for copying part of the codex becomes clear. In Chapter 1, we made a passing reference to the similarities between the coronis of Deuteronomy in B(03) (scribe A) and of Mark’s gospel in א(01) (scribe D). Although Milne and Skeat used this similarity to suggest a shared scribe, Parker has rightly noted enough variation between the two end-pieces, including the titles, to conclude that they are not necessarily from the same scribe.192 Since the shared design of the coronides is relatively simple, it is certainly possible to imagine them appearing independently of one another.193 If this is the case, then it is even more plausible that two different scribes, one dependent on another, copied very similar coronides in the same manuscript. 3.3 Conclusion The past two chapters have explored a variety of material and visual features in B(03) with hopes of identifying unique patterns in each of the scribes. We have used codicology, paratextual features, and paleography to help identify how many scribes were responsible for copying the codex and where they alternated responsibility in production. This has occasionally led to inconclusive results, as the scribes did not always maintain consistency in their distinguishable copying habits, or because we were unable to attribute specific features to the scribes or their exemplars. Still, it was shown that the majority of the evidence, when combined, seems to support much of what Milne and Skeat concluded about the scribes. The frequency of ektheses and line-fillers certainly suggests 192 The two most significant differences are the distance between the horizontal ornamentation and the end- title in א(01), and the extension of this design across the whole width of the column. Parker, Textual Scholarship, 73. 193 Though admittedly less ornate, the coronides of certain Homeric papyri (P.Cair. inv. 3675, P.Oxy. 4 771, P.Ryl. 1 53, P.Amh. inv. G 202) share a similar design to those in the work of scribe B (e.g., Micah, p. 968A). See the plates in Schironi, To Mega Biblion, 85–209. PARATEXTUAL FEATURES AND THE PROBLEM OF PALEOGRAPHY 107 a change of scribes at 1 Kingdoms 19:11 (p. 335) and Hosea 1:1 (p. 945), since scribe B utilizes them far less than scribe A. However, in Esther–Tobit, the frequency did not match either of the scribes outlined by Milne and Skeat. The discontinuity in running titles also corroborated a scribal transition in 1 Kingdoms 19, though it was argued that many, if not all, of the running titles were already present in the quires, before the main text. Since they do not appear in Psalms, they did not assist in identifying a scribe change. However, when the methodology of Milne and Skeat hindered them from being more precise about the division in Psalms, we found that additional evidence shifted the second change of scribes back to Psalm 77:71 (p. 675). On codicological grounds, this location is more fitting than Psalm 1:1 (p. 625), since p. 675 begins a new quire. Both 1 Kingdoms 19:11 and Hosea 1:1 begin a new quire, which suggests that this was most likely an editorial prerequisite for the transition of labor between scribes. Whether or not this also means that the scribes worked simultaneously is complicated by the fact that two of the divisions occur within a single verse (1 Kgdms 19:11a/b; Psalm 77:71a/b). TABLE 14: SUMMARY OF DISTINGUISHING FEATURES IN SCRIBE CHANGES 1 Kingdoms 19:11 (p. 335) Psalm 77:71 (p. 675) Hosea 1:1 (p. 945) - New Quire [19] - 44 to 42 lines-per-column - Loss of running titles - Loss of abbreviated πνευµα, ιϲραηλ, and δαυειδ - Loss of ekthesis - Loss of line-fillers - Loss of numerals - Change in tailed-mu - Change in alpha, delta, and lambda - Change in µου-ligature - Change in coronis - New Quire [36] - Unique nomina sacra: ουρανοϲ, µητηρ, and υιοϲ - Unique και-compendium - Loss of -ται-ligature - Change of hand in Psalm numbers - Change in alpha, delta, and lambda - Change in µου-ligature - Change in coronis - New Quire [50] - Missing folio in [49] - Irregular running title; lack of initial title - Loss of line-fillers - Loss of unique και- compendium - Change in alpha, delta, and lambda - Change in µου-ligature - Change in coronis Using all of the available evidence, this chapter challenged the notion that paleographical analysis of the hands in B(03) was hopeless or inadequate for distinguishing the scribes. After collecting images of all the unreinforced text, we were able to compare the hands in 1 Kingdoms, where we had the most reason to suppose a scribal transition. The clear difference in the alphas, deltas, and CHAPTER 3 108 lambdas allowed us to trace the hands into Psalter, where another change occurred in Psalm 77:71. However, this hand did not resemble that of either scribe, but of a third scribe. Thus, we reintroduced the possibility of a scribe C, a conclusion first proposed by Traube and one Milne and Skeat were tempted to accept—were it not for the ultimate proof of the coronides. However, even with the end-pieces of B(03), we found subtle differences in Psalms–Tobit, which did not appear in Genesis–Ruth. Fortunately, other paleographic evidence, like the nomina sacra, ligatures, and tailed-mu, provided further support for a hand other than scribe A. Therefore, Part II of this study will progress from the proposed, three-scribe outline by examining the earliest corrections in the codex and what they inform us about the work of each scribe. Part II: The Earliest Corrections CHAPTER 4 The Scribes and Early Correctors of the Codex In the first part of this project, we focused our attention on B(03) as a material artifact with a complicated history of paratextual features. The primary intention was to use this evidence for the identification of the number of scribes responsible for inscribing the text of the manuscript, as well as to set the limits for each scribe’s work. While not conclusive, the weight of the evidence revived the argument for a third scribe— something that has not been explored since before the work of Milne and Skeat. This chapter, along with Chapters 5 and 6, will explore the earliest corrections made to the text and their affiliation with our three scribes. Researchers have long acknowledged the need for this examination.1 Fortunately, Pietro Versace’s recent work on the marginalia has done much to provide data for the analysis of these corrections, but many questions remain. To accomplish this study, we begin by outlining the various attempts at identifying those responsible for the many corrections in B(03). In some cases, the same hands have also been attributed to other marginalia in the codex. Therefore, although we treat corrections separately as a textual phenomenon, they belong mechanically to the features outlined in Part I of this study. Taking the lead from previous examinations of the corrections, we will survey and provide a typology of the earliest emendations in B(03). The primary interest in early corrections does not nullify the importance of understanding the later correctors as both preserving and obscuring the earlier layers. The conclusions from this chapter and the following two will help solidify the new enumeration of scribes, identify particular copying habits of each scribe and their role in correcting the manuscript, and shed light on the scribal milieu in which B(03) was crafted. 4.1 Previous Research on the Correctors Already with Erasmus, B(03) was recognized (or accused) as a corrected manuscript. Erasmus, without having seen the manuscript for himself, lodged the criticism of Latinization against the codex, claiming that there was an agreement between the western and eastern churches to bring 1 Ropes, Beginnings, xl; Canart and Martini, Introduction, 8; Canart, “Notice,” 25. For an extensive bibliography of studies on corrections in manuscripts other than B(03), see Malik, P.Beatty III, 72–73 n. 5. CHAPTER 4 112 Greek manuscripts into conformity with the Vulgate.2 Although his theory—sometimes called the Foedus cum Graecis—was proven to be dubious, the accusation against B(03) of Latinization continued with scholars like Mill and Wettstein. 3 It is clear that these criticizers of B(03) are referring to a modern project of correction, but a layer of corrections nonetheless. In the Sixtine edition of the Septuagint (1587), we find references to marginal corrections, but few details about their origin.4 It is not until Richard Simon’s critique of this edition, where we find mention of the explicit distinction between the original text and the corrections. In a letter to John Ernest Grabe (1692), who was editing the Septuagint text from A(02), Simon warns not to make the same mistake that the Roman editors did, by confusing readings of B(03) in the original hand and those in a second hand.5 Simon does not, however, go into further detail about the date of this second hand. In Rulotta’s collation (1729) for Richard Bentley, we first hear of the antiquity of some corrections in B(03). At the end of the collation, Rulotta describes many of the deletions, and interlinear or marginal corrections as from a “vetustissimo Calligrapho.”6 He goes on to say that “almost all variant readings or interlinear emendations to the text are completely the same antiquity as the text itself.”7 Rulotta’s collation distinguished between original readings and corrections, by adding the corrected letters supralinearly or by striking out the original letters, imitating the format of B(03). He also indicated corrections using the phrase “ad marginem” (Trinity College, Cambridge Ms. B.17.20, f150r–f155r; f157r). In some instances, Rulotta even mimicked the marginal corrections by placing the lemniskos (˙/.) as a cross-referencing mark in the main text and in the margin. It appears that he distinguished between prima manus corrections and secunda manus corrections by choosing to underline those of the second hand.8 4.1.1 Thomas Wagstaffe 2 Annotations on Luke 10:1; Epistle 2905; Contra Morosos, 41; See translations in Krans, “Erasmus,” 461, 466–467, 469. 3 Sider, Erasmus, 345 n. 1486; Tregelles, An Introduction, 164; Pisano, “L’histoire,” 109. 4 See, Daniel 5:4 (p. 1219C): “quae addita sunt margini libri Vaticani”; cf. the reference to the marginal gloss in Sirach 28:14 (p. 864B). 5 The letter is printed in Simon, Bibliothèque critique, 1:280; Cf. Pisano, “L’histoire,”114. 6 Rulotta, “Collatio,” 16; cf. An-Ting Yi, et. al., “Prolegomena,” 322–338. 7 “[Item] fere omnes variantes lectiones sive potius emendationes Textus interlineares esse ejusdem penitus antiquitatis ac idem ipse Textus.” See Rulotta, “Collatio,” 16. 8 See Rulotta, “Collatio,” where the two hands are distinguished through a two-column format. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 113 In 1739, Thomas Wagstaffe produced detailed notes on B(03),9 while visiting the Vatican library.10 Starting on f.101v of his notes, Wagstaffe begins describing the reinking of the codex. After considering the various corrections made with the new ink, he then goes on to propose that “the curators of Sixtus V. Septuagint were the persons that caused this ms. to be renewed and corrected in the manner it now appears” (f. 102r). Wagstaffe comes to this conclusion on the basis of agreements with the Sixtine edition and the reinked text of B(03). He does, however, admit that there are places where the editors did not choose the reinked text. In addition, Wagstaffe does not explain why the editors of the Sixtine edition of the Septuagint would continue this project into the NT portion of the codex. Yet, after examining a few notes on marginal corrections in the Sixtine edition, Wagstaffe corrects his earlier suggestion: “it appears very evidently that the characters of the ms. were renewed before it came into the hands of the editors” (f. 103r). Again, the discussion up to this point seems to revolve around the clearly later corrections that were made by the reinforcer. However, Wagstaffe does not leave his description of the codex here. After discussing the reinforcer, he gives what appears to be the first detailed description of the types of corrections found in B(03) (on the following symbols, see §4.2). This includes “marginal additions” which “have the lemnisci and other ancient markers affixed to them” (f. 102v). In places where the original text is marked for deletion, Wagstaffe reproduces the text with round hooks to mark the boundaries of the correction. Like Rulotta’s collation, he also transcribes supralinear corrections, by adding the corrected letters above the original readings. At 4 Kingdoms 10:8, Wagstaffe copies two corresponding, S-shaped sigla above the words πυληϲ (in the text) and πολεωϲ (in the margin), also noting that this happens elsewhere. In Deuteronomy 2:5, he gives an extended discussion of the marginal correction εν κληρω δεδωκα τοιϲ. Here, he suggests that the first-hand reading had τοιϲ in place of ενκ, but was subsequently “rubbed out,” smudging the parchment. Wagstaffe reproduces another form of marginal correction, which highlights visual error on the part of the scribe at line breaks (parablepsis). In his transcription of 4 Kingdoms 2:13, we see how scribe B skipped from ε|πεϲτρεψεν to ε|ϲτη, on account of a line break (though Wagstaffe does not attribute the error to any reason; Figure 41). In the following two chapters, we will find numerous other examples of line breaks, which triggered omissions from same to same. 9 I am grateful to An-Ting Yi for providing access to his personal images of the manuscript. 10 Wagstaffe, “Some account”; Pisano, “L’histoire,” 115. CHAPTER 4 114 FIGURE 41: HAPLOGRAPHY IN 4 ΚΙΝGDOMS 2:13 (P. 444B) The only methods of correction that appear to be missing from Wagstaffe’s notes are the use of deletion dots and oblique strokes to cancel unwanted letters. On a few occasions, Wagstaffe comments on the hand and ink of the corrections. As we will see in the following section, it is often difficult to know whether the small letters at the end of lines were corrections or a regular feature of the scribe’s hand. Wagstaffe criticizes the editors of the Sixtine edition for its note on LXX Psalm 64:10 (p. 662B), since they seemed to mistake the word ϲου as a correction, with its compressed letters. But Wagstaffe is right to note that the scribes of B(03) are capable and often do write in both larger and smaller letters. On the other hand, he does suggest that the nonsense readings γρ|µατοειϲαγωγειϲ (p. 230B, Deut 29:10) and εχ| (p. 235C, Deut 32:41) resulted from the line-end omissions of αµ and θροιϲ, “and that they were supplied afterwards by another hand” (f. 106v). Wagstaffe remains hesitant, however, to attribute corrections to the scribes themselves, even when the ink and hand look identical to the unreinforced text. In Exodus 29:18 (p. 84A), the word θυµίαµα is corrected to θυϲίαϲµα, with both sigmas written supralinearly. Wagstaffe remarks that “the ink of the two ϲ… seems to be as much decayed as the letters of the word itself that are left unrestored” and that the letters are “in the ancient form.” Considering this similarity, he still concludes, “it is probable that these two ϲ were added by the person that retouched the text,” though “one might suspect the correction to be before the renewal” (f. 105r). In at least one location, he does attribute antiquity to a correction (p. 394B, 2 Kgdms 24:22). Here, the word ανενεγκετω is corrected to ανενεγκατω, with the alpha written above the word; both the original text and the correction are left untouched. He knows of its age, “because the α above the line is of the same colour with the other letters” (f. 107r). While I suspect he is correct about both corrections, the similarity of his comments do not make it clear how he distinguished early from late emendations. While Wagstaffe does not explicitly attribute any of the corrections to the first hand of B(03), his notes are important for their witness to early scholarly attention to the corrections in the codex. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 115 4.1.2 J. Leonhard Hug After Wagstaffe, the majority of comments about corrections in B(03) were bound to discussions of the reinking. For some, this was mentioned by way of criticism. 11 Others, like Andreas Birch, considered the corrections to be recent because the corrected readings agreed with the “younger codices.” 12 However, in the winter of 1809–1810, J. Leonhard Hug opportunely studied the manuscript in Paris, since Napoleon had filched it from Rome.13 His De antiquitate codicis Vaticani, published in April of 1810, defended the fourth-century date of B(03) and gave attention to the corrections in the codex. The clearest example of this interaction can be seen in his brief discussion of the marginal correction εν εφεϲω at Ephesians 1:1 (p. 1493B). Here, Hug explicitly states that this correction was made a prima manu.14 Although Hug was later criticized by Tischendorf and Gregory for classifying this particular correction as prima manus,15 his consideration of B(03)’s earliest corrections remains significant. Like Wagstaffe, Hug identified numerous markers of correction in the codex. The two sigla he classifies as a prima manu are the deletion dots (not recorded by Wagstaffe) and the inverted commas or round hooks.16 While he is not explicit as with Ephesians 1:1, Hug appears to suggest the marginal corrections at Matthew 13:52, 14:5, 16:4, and 22:10 are also from the first hand.17 This claim—that the scribes of B(03) corrected their own work—will be an important idea in the following discussion. 4.1.3 Tischendorf, Eduard von Muralt, and Cardinal Mai In the time between Hug’s study and Tischendorf’s visit to Rome in 1843, B(03) was returned to the Papal library. While he was there, Tischendorf was only allowed to study the manuscript for two 11 Le Long, Bibliotheca sacra, 160; cf. Bianchini, Evangeliarium quadruplex, cdxcii; Michaelis, Introduction, 2:348. 12 Birch, Variae lectiones, 18; cf. Birch, Kritisk Beskrivelse, 55–57. 13 Stunt, Tregelles, 66. 14 Hug, De antiquitate, 21; cf. Hug., Introduction, 1:266. 15 Tischendorf classified this correction as from a third hand, claiming that there was nothing of the form from the first hand to be found in it. Tregelles believed this correction to be either prima manus “or at least a very early hand.” Both Tischendorf and Tregelles made trips to see B(03), but were very limited in their investigation. Tischendorf, Vaticanum, 48; See his earlier comments in Tischendorf, “Biblisch kritisches”, coll. 962–63; Gregory, “Prolegomena,” 362; Tregelles, An Introduction, 161; Tregelles, An Account, 156. 16 Hug, De antiquitate, 9. 17 He does not mention the S-siglum, accompanying all four of these corrections. Hug, De antiquitate, 9. CHAPTER 4 116 days.18 Still, on July 25th of that year, he published a five-column entry in Allgemeine Kirchenzeitung, where he discussed various features of B(03) and even mentioned a forthcoming “facsimile” of the codex.19 Tischendorf, again, published on B(03) in 1847, where he compared his rapidly prepared collation with those of Bartolocci, Birch, Bentley, Cardinal Mai, and Lachmann’s 1842 edition of the NT. 20 The inconvenience which Tischendorf faced was primarily the result of Cardinal Mai’s coinciding attempt (1828–1838) at producing an edition of the NT of B(03).21 Mai’s edition was not released until 1857, but Tischendorf was given access to the printing by Mai himself.22 In both the 1843 and 1847 publications, Tischendorf’s main discussion of corrections in B(03) involves the marginal note at Ephesians 1:1 (εν εφεϲω). In response to Hug’s prima manus classification, Tischendorf claims that anyone who has “eyes that are familiar with paleography” will realize that Hug’s assertion was wrong.23 It would be another twenty-years before Tischendorf published his “facsimile,” Novum Testamentum Vaticanum (1867), wherein he identifies the hand of this marginal correction as B3. Apart from Ephesians 1:1, Tischendorf gives little attention to the earliest corrections in his initial publications on the codex.24 He does, however, make reference to the round “hooks” (classified as prima manus by Hug), in both the 1843 and 1847 publications. Yet, contra Hug, Tischendorf claims these “signs of non-validation” (Zeichen der Nichtgeltung) are from a second hand.25 Cardinal Mai’s infamous edition of Vaticanus was finally released in 1857, and the NT again in 1859.26 In these editions, Mai does not explicitly distinguish the layers of correction: there are only prima manus and secunda manus readings. Any correction to the manuscript is identified as “2. m.,” and only when Mai chose to print a correction (he often prefers the orthography of the reinker), 18 Tischendorf, Novum Testamentum Graece, 7th ed., cxliii; cf. Gregory, Canon and Text, 346. 19 Tischendorf, “Biblisch kritisches”, coll. 962–63. 20 He refers to Lachmann’s “grössern Ausgabe.” Tischendorf, “Nachricht,” 135. 21 Gregory, “Prolegomena,” 363. 22 Mai, Codex Vaticanus, iii. 23 Here again, he mentions his “facsimile”: “Augen, die der Paläographie kundig sind, werden von selber dieses Resultat aus meinem Facsimile gewinnen.” Tischendorf, “Nachricht,” 133. 24 This includes his Novum Testamentum Graece, 7th ed.. 25 Tischendorf, “Biblisch kritisches”, coll. 962; Tischendorf, “Nachricht,” 151. 26 Gregory states “it was the worst thing he ever did, and he knew it.” Gregory, Canon and Text, 346. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 117 does he give “1. m.” for the original reading.27 According to the publisher’s note at the beginning of the 1859 edition, Mai often ignored corrections in the first edition—something that Tischendorf did not hesitate to criticize. 28 We first find an explicit distinction between corrections in the comparison between the collations of Mai and Birch, given at the end of the 1859 printing. According to the printed table, Birch identified a different hand as responsible for the marginal correction at Matthew 10:37 (p. 1247C), while Mai thought this correction originated with the first hand.29 Turning to this passage in the edition, we find the marginal reading printed in the main text and lacking any indication of a correction. Such imprecision is what led to Tischendorf’s 1867 pseudo-facsimile, subtitled post Angeli Maii aliorumque imperfectos labores ex ipso codice. Before returning to Tischendorf, it is important to mention Eduard von Muralt, who, in 1844, was permitted to see B(03) for three days.30 In the preface of his Greek NT (1848), Muralt gives a list of twelve features in the codex. In the tenth section, he briefly mentions three types of corrections: 1) errores primae manus manifesti… a secunda manu correctis; 2) ubi ipsa prima manus se emendavit; 3) ubi secunda manus a communi differt.31 Muralt gives eight examples of prima manus corrections, but does not provide any reasons for their identification as such.32 The exceptions are the spelling corrections of ακουϲατε → ακουϲετε at Matthew 13:14 (p. 1271A) and ειδαµεν → ειδοµεν at 25:37 (p. 1349C). He supposes that these changes belong to the hand of the original scribe because ακουϲατε is followed by the future βλεψετε, while ειδαµεν is originally spelled ειδοµεν in the following two occurrences. Therefore, it is likely that the original scribe noticed this discrepancy while copying and made the interlinear correction. The problem with previous discussions of the correctors in B(03) is the restriction to two hands—one early hand (whether the scribe or a contemporary) and a late re-inker. Therefore, 27 At the end of his prolegomena to the The Greek Testament (1863), Henry Alford published a list of readings in B(03), which he procured from personal inspection of the codex two years prior. In this list, Alford utilizes Mai’s notations, “1. m.” and “2. m.,” but uses the sigla B1 and B2 in the apparatus of the main text. Many of his notes are corrections to Mai, with a preference for prima manus corrections. 28 In the 1859 edition, Carlo Vercellone appears to have significantly modified Mai’s earlier work. Mai, Codex Vaticanus, iii; cf. Epp, “Tischendorf,” 575 n. 54. 29 “sunt plane ab eadem manu in marg. cod. p.1247.” Mai, Codex Vaticanus, 199. 30 Muralt, Novum Testamentum, xxxv; cf. Gregory, “Prolegomena,” 363. 31 Muralt, Novum Testamentum, xxxii. 32 p. 1246C, Matt 10:14; p. 1247C, Matt 10:37; p. 1251C, Matt 13:14; p. 1271A, Matt 25:37; p. 1349C, John 1:4; p. 1454A, Rom 9:16; p. 1479C, 2 Cor 3:18; p. 1500C, Phil 2:23 CHAPTER 4 118 anything that could not be identified as prima manus, was relegated to secunda manus and was inevitably conflated with the later reinforcement.33 With the publication of Tischendorf’s pseudo- facsimile, we find the first classification of B(03)’s correctors into three hands (B1, B2, and B3). From February 28th to March 12th, 1866, Tischendorf was given forty-two hours with the manuscript in Rome.34 In the following year, Tischendorf published his long-awaited “facsimile”, and two-years later an appendix with emendations appeared.35 Ultimately, this printing was hardly a facsimile, but an edition of the NT of B(03) with nineteen pseudo-facsimile pages (see n. 34). Regardless, his Novum Testamentum Vaticanum remains a crucial study on the codex. In the prolegomena, he states that three correctors can be distinguished: B1) the original scribe,36 B2) a contemporary corrector writing in “what is the most ancient style,” and B3) who also added the breathings and accents and is dated to the tenth or eleventh century, based on the minuscule script.37 Tischendorf admits that it can be difficult to distinguish between B1 and B2, and so begins with the surest examples of prima manus corrections. For example, scribe B occasionally realizes a mistake in the middle of, or right after, copying a word.38 At Matthew 13:8 (p. 1251B), the scribe originally copied ειϲ at the end of a line (traces of the iota and sigma remain), but subsequently realized the word was meant to be επι. This must have happened immediately since the following 33 This was noted by Westcott and Hort: “confusion between the second and third hands of B has led to much error.” Westcott and Hort, Introduction, 270; Hurtado highlights the importance of distinguishing first-hand corrections from contemporary and later corrections, since all three types indicate attitudes to the text. The first-hand corrections exhibit the intention of the scribe to produce an accurate copy, while contemporary corrections reveal the editorial context of a manuscript, and later corrections indicates aspects of its reception. Hurtado, Artifacts, 186. 34 Tischendorf, Vaticanum, ix–x; Gregory, “Tischendorf,” 178; Gregory, Canon and Text, 347. 35 Tischendorf, Appendix Vaticani. 36 Ropes seems to misunderstand Tischendorf’s classification when he states, “[t]he earliest corrections (B1 and in part B2) are doubtless those of the diorthotes…” For Tischendorf, B1 is the original scribe, and B2 is the diorthotes. In Gregory’s prolegomena to Tischendorf’s eighth edition, the siglum B* is presented instead of B1. Ropes, Beginnings, xl; cf. Tischendorf, Vaticanum, xxiii–xxiv; Gregory, “Prolegomena,” 360. 37 There appears to be some confusion over the date of Tischendorf’s B3 corrector. In his edition of Vaticanus it is clear that he believes it to be a tenth or eleventh-century hand, while in volume one of his eighth edition he gives a sixth or seventh-century date. Gregory clarifies in the prolegomena that B3 is from the tenth or eleventh century. Tischendorf, Vaticanum, xv (for the date of B3, see xxvii); Tischendorf, Editio octava, x; Gregory, “Prolegomena,” 360. However, the Alands follow the sixth or seventh-century date, when they equate the NA26 corrector B2 (6th/7th) with Tischendorf’s B3. Aland and Aland, The Text, 108. Scrivener claims that this hand has been dated to the eighth century, but, in his fourth edition, notes Tischendorf’s later dating. Scrivener, Introduction, 1:107. 38 Tischendorf, Vaticanum, xxiii. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 119 line begins with πι (transcription: ειϲ̣|̣πι). 39 Tischendorf also identifies places of erasure and rewriting by B1 (e.g., p. 1252A, Matt 13:24; p. 1261A, Matt 19:20; p. 1265A, Matt 22:15), though many examples are difficult to confirm from the images. Yet, even when no trace of the scriptio inferior remains, it is often possible to see the effects of scraping or rubbing on the parchment. Finally, he describes seven marginal corrections of B1, which are marked with corresponding S-shaped sigla (noted by Wagstaffe). We will discuss this type of correction in more detail below and in the following chapters, but it is noteworthy that Tischendorf thought these readings were derived from another exemplar (ex alio codice), though still copied prima manus. Tischendorf goes on to discuss the corrections of the diorthotes (B2), who was contemporary (scriptori aequalis) with the original scribes.40 By copying compressed letters at the end of lines, B2 is responsible for correcting some places where the scribe omitted words at line-breaks (p. 1246C, Matt 10:14; p. 1292C, Mark 10:19). Furthermore, Tischendorf classifies marginal corrections with the lemniskos (˙/.) and ancorae (↑↓) as B2 (p. 1239B, Matt 5:16; p. 1247C, Matt 10:37; p. 1285B, Mark 6:17). This same corrector also added interlinear corrections (p. 1344C, Luke 22:39), omission dots (p. 1420A, Acts 25:15; p, 1433C, 1 Pet 4:18), oblique cancelation strokes (p. 1360C, John 7:28), and made erasures (p. 1240A, Matt 5:27; ερρεθη → ερρηθη). As mentioned above, Tischendorf did not believe the round hooks (circumdedit uncis), which enclose unwanted text, were from the first hand. Instead, they are attributed to B2 (p. 1370C, John 13:14; p. 1479B, 2 Cor 3:15). We briefly mention here the instaurator (Tischendorf’s B3), whose main project was the reinforcement of faded ink in the codex. However, in the process of reinking, B3 made corrections to the text. This is clearest in places where the reinker left the original text to fade, preferring to reinforce a marginal reading (p. 1248C, Matt 11:19). According to Tischendorf’s paradigm, the hand of B3 is unmistakable, when it reverts to correcting in a minuscule hand (p. 1257A, Matt 16:19, 20). However, in other cases, it is not always clear if a correction originated with the reinker or if it is a reinforced reading of B1 or B2. As mentioned above, Hug identified the reading εν εφεϲω (p. 1493B, Eph 1:1) as prima manus, while Tischendorf strongly claimed the correction originated with B3. Like 39 For further examples of in scribendo corrections, see Chapters 5 and 6. 40 Tischendorf, Vaticanum, xv. Elsewhere he claims, “B2 a B* aetate vix differt.” Tischendorf, Editio octava, x; cf. Gregory, “Prolegomena,” 360. CHAPTER 4 120 B2, the reinker added interlinear corrections (p. 1275C, Matt 27:45; p. 1391C, Acts 7:32), often indistinguishable from the earlier hands (p. 1408B, Acts 17:22).41 In addition to his three correctors, Tischendorf identified a sixth-century hand, responsible for the colophons in the Pauline epistles, as well as a thirteenth-century hand, which copied the correction and famous marginal note at Hebrews 1:3 (p. 1512B).42 While these additional marginalia do not disqualify Tischendorf’s paradigm of corrections, it begs the question whether there were really only three correctors. It should be noted that already in 1888, Ceriani allowed for the possibility of more correctors when he stated that corrections may be written by the “first hand, immediate diorthotes, or any subsequent hand,” but the re-inker hinders our ability to distinguish them.43 4.1.4 Vercellone, Cozza-Luzi, Fabiani, and Sergio Shortly after Tischendorf’s pseudo-facsimile was published, work began on a new Roman edition of B(03). It was widely acknowledged, even by Pope Pius IX, that Mai’s edition needed replacing.44 The decision for Cardinal Vercellone to lead this project, with the help of Friar Cozza-Luzi, was already made by April, 1866.45 The desire for Rome to have a new edition seems to have been heightened by Tischendorf’s plans to produce his own pseudo-facsimile. Vercellone was particularly concerned over whether Tischendorf would publish in Rome or Leipzig (Vat. Lat. 14042, f. 654). 46 Under Vercellone and Cozza-Luzi, the first volume to emerge was the typographical facsimile of the NT (vol. 5), published in 1868. After the death of Vercellone (d. 1869), volumes one through four were published in consecutive years by Cozza-Luzi, G. Sergio, and Canon E. Fabiani (1870–1872). However, the preface and commentary by Fabiani were not printed for another nine years after volume four (1881). 41 Tischendorf often applies the abbreviation ut vdtr to corrections that seem to have originated from the hand of B2, but are entirely covered by the reinking of B3 and may not be distinguishable paleographically. 42 αµαθεϲτατε και κακε αφεϲ τον παλαιον µη µεταποιει – “Most ignorant and wicked one, leave the old reading, do not change it.” Tischendorf, Vaticanum, xxvii n. 2. 43 Ceriani, “L’antico testament,” 545. 44 Tisserant, “Notes," 239. 45 For the various correspondences, which are preserved in Vat. Lat. 14042, see Tisserant, “Notes,” 237–248 (on 9 and 15 April, see 242). 46 Tisserant, “Notes,” 239 n. 2. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 121 In his prolegomena on B(03), Fabiani outlines the four correctors (B1–4) employed throughout the commentary. 47 In contrast to Tischendorf, Fabiani does not use a siglum to distinguish the original text (B*) from the prima manus corrections (B1). In the commentary, B1 identifies the original reading and corrections made by the scribe. Therefore, while B1 is frequently seen in the commentary, there are few corrections attributed to this hand. In one example, Fabiani attributes the correction of ανδεϲ to ανδρεϲ in 1 Kingdoms 29:2 (p. 350A) to B1. Unlike the more common interlinear corrections of B2 and B3, this correction is made by fitting the rho between the two letters. Fabiani goes on to define B2 as coming after the scribe, but before the diorthotes.48 Contrary to Tischendorf, who identifies B2 as the diorthotes, Fabiani reserves this title for B3. His description of B2 is vague, because the siglum encapsulates all corrections appearing between the fourth and fourteenth centuries. Fabiani describes the additions as “uncial” in character and in a yellow ink.49 In the commentary, we find that B2 made both interlinear (p. 322C, 1 Kgdms 11:11) and marginal corrections (p. 322C, 1 Kgdms 11:13). The marginal corrections appear on their own, or with corresponding sigla: the S-shaped sigla (signum variae lectionis), lemniskoi (˙/.), and ancorae (↑↓). B2 is also responsible for deletion dots (puncto) and oblique cancelation strokes (obelo). The work of B2 extends through the fourteenth century, because, unlike Tischendorf, Fabiani dates B3 to the fifteenth century. This is the “diorthotes” who reinked the manuscript.50 Not only is this hand dated later, but we also find a name attributed to the scribe: Clement the Monk. At the end of both Deuteronomy and 2 Esdras, we find the signature + κληµηϲ µοναχοϲ +, written in a hand Fabiani dates to the fifteenth century.51 However, this hasty association between the “black ink” of the signature and of the reinforcement has rightly received criticism,52 while Versace prefers a thirteenth-century date for the signatures (Versace: B29).53 Fabiani’s B3 is responsible for many 47 Fabiani, Prolegomena, xvii ff. 48 “eorum qui post illum et ante diorthotam generalem.” Fabiani, Prolegomena, xvii. 49 Fabiani, Prolegomena, xvii. 50 It is clear that Fabiani and Tischendorf have different understandings of the role of the diorthotes. 51 Fabiani, Prolegomena, xvii. 52 Scrivener, Introduction, 1:107 n. 136; Ropes, Beginnings, xli. 53 Versace, Marginalia, 60. CHAPTER 4 122 orthographic changes, either through lack of reinforcement (ει → ι) or interlinear corrections (τεϲϲερεϲ → τεϲϲαρεϲ); a number of scholia, and the supplementary leaves. Finally, Fabiani introduces a fourth group of correctors (B4). These hands “sprinkled corrections” (correctiones adspergerent), added other notes and scholia, and reinforced places where the ink of B3 had already faded.54 In the commentary, Fabiani often describes the hand of B4 as crude (rudi). On p. 402Α, it is noted that B4 used the lemniskos (˙/.), like our earlier hands. However, in this example there is no corresponding lemniskos in the margin. While B4 will not be significant for our examination of the earliest correctors, the identification of corrections post- reinking is important.55 4.1.5 Pietro Versace After Tischendorf and Fabiani, very little attention was given to the correctors in the codex. One finds summaries of Tischendorf’s conclusions in the work of Westcott and Hort, and that of Ropes, but there is little development beyond his outline. For instance, both Hort and Swete claim that B(03) does not seem to have been corrected in the time from the fourth century to the tenth or eleventh century—in other words, between Tischendorf’s B2 and B3.56 Although Ropes is clear that “[t]he designations [B1–3] are to be regarded as referring to groups of correctors, rather than to individuals,”57 B2 is still restricted to the fourth century rather than spanning centuries, as is the case of Fabiani’s B2. Ceriani, on the other hand, did show concern for the various corrections made before the reinforcement, and was unsatisfied with Tischendorf’s conclusions or Swete’s willingness to follow them. Instead, he agrees with Fabiani that it is better to be vague about B2, since it is unlikely that we can identify the various hands underneath the reinking.58 It is not until Pietro Versace’s recent work on the marginalia that we find a robust attempt at outlining the various hands and correctors in B(03). Even then, it is important to note that the corrections are not his only interest, nor does he include intracolumnar corrections in his 54 Fabiani, Prolegomena, xviii. 55 In Chapter 2, we identified two reinkings in the codex, following the original observation of Fabiani and expounded by Versace. 56 Westcott and Hort, Introduction, 270; cf. Swete, The Old Testament, 1:xix. 57 Ropes, Beginnings, xl. 58 Ceriani, “L’antico testament,” 545–46. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 123 examination. 59 It will be the task of the following chapters to test the coherence of Versace’s framework against the various corrections in the main text, and in relation to the original scribes. He is clear from the outset of the project that his concern is not with the scribes of the main text, and it is, therefore, important for us to compare Versace’s early hands with the three scribes outlined in Part I of this study. Although he has identified thirty-eight hands (B1–38), only a few were responsible for corrections; others copied additions like lectionary markings and scholia. It is important to note that while Versace categorizes the marginalia by “hands”, they often do not represent a single hand, but a collective group of additions, which could have originated from multiple scribes. The first three hands (B1–3) are the only fourth-century contributors to the marginalia in B(03). We have already encountered these early hands in our discussion of the paratextual features, including the section numerals, diplai, and stichometry. Following Tischendorf and Fabiani, Versace’s B1 is used to designate the additions of the scribes. This collective hand is responsible for marginal corrections marked with the S-shaped siglum and written in Biblical Majuscule. Although they are not corrections, B1 also copied the Hexaplaric asterisks and obeloi—critical signs that reflect a scholarly milieu—which are occasionally accompanied by notes clarifying their meaning.60 Versace makes a convincing case that these Hexaplaric sigla originated in the exemplar.61 Versace describes B2 as the “ancient diorthotai, who reread the manuscript immediately after it was written and produced numerous corrections.”62 Like those of B1, the B2 corrections are a collective group of hands, which always write in “Biblical Majuscule.”63 Because of the reinkings (by B18 and B37), it is not possible to clearly distinguish the number of B2 diorthotai in B(03).64 In this collective category, we find marginal corrections marked by lemniskoi (˙/.), or ancorae and lemniskoi (↑↓ and ˙/.).65 In addition, B2 makes marginal corrections that appear at line or column-breaks. In the 59 He does, however, mention some forms of erasure, which occur in the main text (see below). 60 Versace lists the examples of “ου κʹπʹεβρʹ (ου κειται παρ εβραιω) + obelo” and all occurrences of corrections with the S-shaped siglum by B1 in marginalia, 83–90. 61 The note on p. 1054C refers to a reading that is not found in B(03). Versace, Marginalia, 12. 62 Versace, Marginalia, 75, cf. 15. 63 Versace, Marginalia, 15 n. 16. 64 Versace, Marginalia, 15 n. 17. 65 Sometimes these ancorae are accompanied by the directional notes ανω and κατ(ω). However, these directional notes were not copied in the typical Biblical Majuscule of B2. Versace, Marginalia, 75. CHAPTER 4 124 former case, B2 copied omitted text, by extending the writing into the right margin and sometimes finishing it in the left margin, on the following line.66 According to Versace, the corrector followed the same practice at column breaks, by copying supplementary lines below the last line of the column. The largest correction involves over five lines of compressed writing (p. 233A), with nearly double the letters of a standard line. B2 is also responsible for expunging unwanted text with deletion dots, oblique strokes, or round hooks.67 In this way, Versace follows Tischendorf and Fabiani against earlier scholars who associated these correction methods with the first hand. Finally, B2 is responsible for corrections that appear over erasures (p. 233A l. 44), and possibly for transpositions, indicated by interlinear, oblique strokes (p.127B l. 37).68 Although these corrections are attributed to the diorthotai, Versace admits that it cannot be ruled out that the original scribes contributed some of these corrections.69 In the following chapters, we will find reason to believe that numerous B2 corrections originated with our three scribes. The final set of fourth-century corrections originated with B3, who did not write in the Biblical Majuscule hand, but with “undulated or curved” strokes. 70 This hand added marginal corrections, using the S-shaped siglum,71 lemniskoi (˙/.), and ancorae–lemniskoi (↑↓ and ˙/.); and, like B2, made corrections at line-ends. Furthermore, Versace informs us that B3 added supralinear corrections and made expunctions with oblique cancelation strokes and deletion dots. He gives two examples of expunction that are distinguishable from those of B2, because they are accompanied by marginal writing.72 On p. 1087B, µα|ναϲϲαιου is corrected to µναϲϲαιου, by adding deletion dots above mu and alpha and writing a mu to the left of the new line. The second example reveals a complex correction, combining four different methods (p. 1375C): line-end additions, deletion dots, oblique 66 Versace also associates the inverted “pyramidal” correction in Daniel 9:27 (p. 1227C)—a similar shape to those in א(01)—with B2. However, instead of continuing from the column, this correction is centered in the right margin and is connected to the column by an obelos—there are no corresponding sigla. 67 Versace, Marginalia, 15. 68 Versace, Marginalia, 75, cf. 16. 69 Versace, Marginalia, 15 n. 17. Elsewhere, he has noted that this category was made primarily to distinguish a production layer and not necessarily to classify hands. Versace, “Some Marginalia.” 70 Versace, Marginalia, 18. 71 Versace, Marginalia, 138 n. 327. 72 Versace, Marginalia, 19 n. 31, 76. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 125 cancelation strokes, and interlinear additions (Figure 42). Since B3 was responsible for copying part of the early section numbers (see Chapter 3), but not the text (see n. 103), Versace suggests that the corrections by this hand likely originated from the same exemplar as the delimitation markers.73 FIGURE 42: MULTIPLE METHODS OF CORRECTION BY B3 (P. 1375C) As we have already mentioned, Versace’s classification of hands does not include any additions between the fourth and ninth centuries. Therefore, his next corrector (B4) is dated to the ninth century. Although Versace is not certain in every case, B4 likely made three marginal corrections, which are marked by lemniskoi (˙/.): δαυ → δαυ(ειδ) (p. 347C; 1 Kgdms 27:1), νεαν → νεαν(ιαν) (p. 362A; 2 Kgdms 6:1), and καπινειν → κα(τα)πινειν (p. 979A; Hab 1:13). While not a correction, B4 also rewrote part of Amos 7:16–17 in the lower margin (p. 960B), because the parchment had become defective and the text obscured. This rewriting is linked by lemniskoi (˙/.).74 Versace proceeds to identify two marginal corrections in Isaiah that appear to be made by the hand B5, also from the ninth century. Both of these small corrections appear with the S-shaped siglum (pp. 1012A, 1013C), and only make- up one or two letters to correct the spelling of a word: απεϲτρα → απεϲτρα(φη), and περδαλιϲ → π(α)ρδαλιϲ. In the second example, it appears that B5 or a later hand attempted to erase the original epsilon.75 Still in the ninth century, Versace identifies the hand B9 who added one correction at the column end (similar to B2) on p. 460A. This correction begins with the copying of the article τον over the original και-compendium of the previous line, and then continues on a new line with the omitted text from 4 Kingdoms 11:2. 76 The next ninth-century corrector is B13, who made ten corrections in 3–4 Kingdoms, 2 Chronicles, and 1–2 Esdras. These marginal corrections are accompanied by a form of the S-shaped siglum and mostly mark spelling changes; only one of the changes comprises more than a single word (p. 478C, 4 Kgdms 22:19).77 73 Versace, Marginalia, 75–76. 74 Versace, Marginalia, 203. 75 Versace, Marginalia, 212, cf. n. 113. 76 Versace, Marginalia, 254. 77 Verscace distinguishes this hand from the fourth-century hands, because of the split kappa, the “apexes” on the horizontal bar of the tau and at the end of the zeta. However, he admits that the B13 correction at 2 Chronicles 28:10 may have been made first by B2. Versace, Marginalia, 40, 257. CHAPTER 4 126 After these four, ninth-century correctors,78 Versace introduces the first reinforcer (B18), who worked on the codex in the tenth or eleventh century. Since this reinforcer contributed far more to the codex in corrections and marginalia than the later reinker (B37), references to the “reinker” or “reinforcer” are to B18 unless stated otherwise. Nevertheless, it is important for readers to remember that at least two layers of reinking are present in the codex. Unlike B4, 5, 9, 13, B18 worked through the entire manuscript, and among many other additions, “inserted along the margins numerous supplements and corrections, drawn from the antigraph.”79 Versace admits that it cannot be known for certain, whether this was a single scribe or multiple, but claims that the consistent graphic characteristics in the notes suggests one hand.80 Many of the corrections in the main text are changes of spelling, which are intended to realign the codex with “classical orthography.”81 The reinforcer also edited word divisions at line-breaks by leaving the last letter of a line to fade and rewriting it on the next line—or vice versa. The methods of marginal correction used by B18 include corresponding lemniskoi (˙/.), ancorae–lemniskoi (↑↓ and ˙/.), or lemniskoi with horizontal bars (÷). However, some marginal corrections do not have any cross-reference marks. Versace gives one example of a word order correction in which B18 wrote the letters alpha through delta over the words to indicate their proper order (p. 813B).82 Like B2, this corrector also copied omitted text at line ends and column ends. One of these corrections is the disputed, marginal addition of the words εν εφεϲω (p. 1493B, Eph 1:1), which was first identified as prima manus by Hug and subsequently identified as Tischendorf’s B3.83 Finally, in order to expunge apparent errors in the text, B18 seems to have used oblique cancelation strokes (p. 188B l. 20) and deletion dots (p. 195C l. 33).84 78 Here, I use the term “corrector” in a loose sense—namely, anyone who attempted to correct a perceived error in the main text. None of the four correctors from the ninth century attempted anything like a full project of correction, a diorthosis, on B(03). 79 Versace, Marginalia, 78. 80 Versace, Marginalia, 45. 81 Versace, Marginalia, 78. 82 Versace, Marginalia, 50; cf. p. 725B, where B18 added an ancora-lemniskos correction in the upper-margin and clarified the place of omission by adding the letters alpha through delta next to each line in respective order. 83 Versace, Marginalia, 264. 84 The use of deletion dots by B18 is clearly visible when the corrector changed word divisions at line-breaks: e.g., κοϲ|̇ ϲµον (p. 1257B ll. 17–18) and εµπροϲ|̇ ϲθεν (p. 1257C ll. 2–3). However, this appears to be from a different hand than the one who corrected the word πρεϲ| ϲβυτερων on the same page (p. 1257A ll. 30–31). Versace, Marginalia, 46. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 127 However, B18 is not Versace’s final corrector. Like Fabiani, Versace also recognizes editorial activity after the codex was reinked. In the second half of the twelfth century, the hand B19 added numerous scholia, one correction in Genesis 50:11 (p. 46A; ταδ to αταδ), seven in Psalms, and nine in Matthew and Luke.85 This hand is associated with John Camatero, the Patriarch of Constantinople at the very end of the twelfth century.86 Notably, B19 is the hand responsible for the minuscule corrections that led Tischendorf to date his B3 to the tenth or eleventh century.87 These corrections mostly involve the addition of a pronoun or article, and only twice exceed two words. They are relatively insignificant, when compared to the rest of the scholia added by B19, but remain important as these too were reinked or corrected by an even later hand. In the thirteenth century, Versace identifies two more hands (B20 and B21), who made three additions in Exodus 8:14; 9:25; 11:2 (pp. 56A, 58A, 60B).88 The first two are marked with lemniskoi (˙/.), and the third with a small cross. Also among the thirteenth-century corrections is the famous note in Hebrews 1:3, written by Versace’s B27, who criticizes the reinforcer (B18) for changing the reading from φανερων to φερων (p. 1512B). This hand, then, rewrote the letters alpha and nu in the main text. Since many of the marginal additions after B18 were reinked, Versace proposes a sixteenth- century collective category, B37. This hand, however, did not correct the text, apart from a few examples. One marginal addition by B37 is the και-compendium at Matthew 16:19 (p. 1257A l. 13). Yet, the ink for this correction does not match the dark black of the second reinforcer, but a lighter ink, which Versace also attributes to the sixteenth century.89 While our main interest is in the earliest corrections of the fourth century, Versace’s thorough work reveals much about the history of the codex through its many layers of correction. One important pattern to notice is the use of similar cross-referencing sigla by the different hands: the lemniskos (˙/.) was adopted by B2, 3, 4, 18, 20, 21 and the S-shaped siglum by B 1, 3, 5, 13. This means that we cannot rely simply on the sigla for identifying the corrector, but must look for other distinctive 85 While Versace breaks them down into two categories—“corrections” and “integrations” to the text—they are all attempts to correct the text. Versace, Marginalia, 277, 287–88. 86 This observation is significant for the later history of B(03), before its arrival in Rome. Mazzucchi, Per la storia medieval, 135; cf. Versace, Marginalia, 56. 87 Tischendorf, Vaticanum, xxvii. 88 Versace, Marginalia, 289. 89 Versace, Marginalia, 63 n. 122. CHAPTER 4 128 features. As Versace makes clear, the process of isolating correctors is further frustrated by our reinkers, who often conceal the telling features of earlier hands. So far, then, we have summarized the various correctors identified in previous scholarship. However, there have been few attempts to clarify, in one place, all the methods and meanings of correction in B(03). TABLE 15: SUMMARY OF VIEWS ON CORRECTORS IN B(03) Tischendorf Fabiani Versace Swete LXX NA28/ECM THGNT90 Additional Sigla Production (4th CE) B* B1 B1 B* B* B* B1 B1 B1 B1 B1 Bs (C. E. Hill) B2 B2 B3 Ba B1/03C1 B1 Intermediary Period (4th–10th CE) B2 B4 B5 B9 B13 B2/03C2 (6th–7th CE)91 B2? Instaurator (10th–11th CE) B3 B2 B18 B2 or B2a/2b Scholia (12th CE) B2 B19 13th Century B2 B20 & B21 B27 Second Instaurator (15th–16th CE) B3 B37 Bb B4 Bc 4.2 A Survey and Typology of the Earliest Corrections One recent attempt to outline the corrections in B(03) was undertaken by Charles Hill.92 While his interest lies primarily with the NT corrections accompanied by the S-shaped siglum—what he calls the S-siglum—Hill briefly summarizes the other methods of early corrections in B(03). By early corrections, we mean those which appear in the fourth century: either Tischendorf’s B1–2, Fabiani’s B1 (sometimes B2), or Versace’s B1–3. In summarizing these corrections, Hill broadly categorizes the various methods as either adding or deleting text from the codex. Since most correction methods fit well within this framework, the ensuing discussion is likewise divided, saving the exceptions until the end. Furthermore, it is worth nothing that the following sections are primarily concerned with how information is added or removed from the manuscript through corrections. It will be the task 90 In the THGNT, B1 corrections originate from either the scribe or a colleague. There is some ambiguity over when the editors understand B2 corrections to have been added. It is possible that this siglum is meant to encompass all later corrections that cannot be attributed to the time and place of production. See Jongkind, An Introduction, 58. 91 See also, Schmid et. al., The New Testament, 8. 92 Hill, “Siglum,” 1–22. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 129 of Chapters 5 and 6 to provide a summary and evaluation of what the corrections actually indicate about the implied errors in the text and how they reflect the work of our three scribes. Although this survey cannot be comprehensive, some forms of correction lend themselves to more complete examination than others. It will become readily clear that there are challenges to detecting certain types of correction and to identifying the hand that created them. Nevertheless, this outline of correction methods in B(03) allows one to observe patterns of correction throughout the entire codex. Like the paratextual and paleographic features examined in Part I, correction methods appear with varying frequency as we progress through the manuscript. 4.2.1 Methods of Correction by Addition The first group of corrections involves additions to the text without corresponding sigla.93 Many of these additions are small corrections of a few characters or individual words, which appear above or between letters:94 corrections of this type will be indicated with the symbol “\/”, for supralinear additions, or “/\”, for compressed letters on the line. We find this method of correction already on the first extant page of B(03): π\λε/ιονι (p. 41A, Gen 46:29). 95 As seen in Versace’s work, other additions to the text appear at column-ends (“≞”)96 or line-ends (“]txt[”). TABLE 16: COLUMN-END AND LINE-END CORRECTIONS IN VERSACE B2 B3 Column-end (≞) 11 0 Line-end (]txt[) 241 14 Column-end corrections are rare, but easily spotted in the manuscript. As shown in Chapter 2, the line count is very consistent in B(03), drawing attention to any unevenness caused by these corrections. Ten of the eleven column-end corrections occur in the work of scribe A, and one in the work of scribe C. This nearly exclusive form of correction in scribe A’s text suggests that he or she was more likely than the others to omit text when transitioning columns or pages (p. 124C). Although Versace identifies column-end corrections with B2, our first example (p. 73B, Exod 21:31) 93 Hill, “Siglum,” 1. 94 In some places the replacement letters are stacked above the corrected letter. See p. 197C l. 18, for the correction ευκληρον → ευ̸\νγ/κληρον (γ above ν). 95 Swete identifies this correction as Bab. For Swete’s classification of correctors see Table 15. 96 The “m,” here, indicates the main text of the column. CHAPTER 4 130 clearly exhibits the dominant features of scribe A (Figure 43): the word εαν is copied with ekthesis, as it is elsewhere on the page; the alphas and deltas are similar in form to those of scribe A; and finally, the correction ends with a line-filler, the method used primarily by this scribe (see Chapter 3). If this is not a prima manus correction, then the corrector was intentionally mimicking the hand of scribe A. FIGURE 43: COLUMN-END CORRECTION (P.73B, EXOD 21:31) In contrast to the multiple examples in scribe A, the single column-end correction in the work of scribe C (p. 688A, Ps 100:3) consists of only two words and, thus, favors our judgement that there is a third scribe in Psalm 77–Tobit. We will return to the paleography of these corrections in the following chapter (see pp. 147–48). The first line-end correction appears on the second extant page of B(03): ]ηυξηθησαν | και[ (p. 42C, Gen 47:27).97 We will see in the following chapter that some line-end corrections, which Versace associates with his B2, are likely the work of the original scribes. Occasionally, such emendations do not actually begin from the line-end, but as supralinear corrections that continue into the margins (Figure 44). Rather than using a cross-reference siglum after the word ϲου in Deuteronomy 24:19, the corrector (Versace: B2) began copying directly above the line. This correction ends with the word ϲου in the margin, at which point the reader would return back to the ουκ in the main text. However, it is more common for the scribes to utilize cross-reference symbols and copy the full correction in the margin. FIGURE 44: MARGINAL CORRECTIONS IN DEUTERONOMY 24 (P. 223) Hill introduces the three forms of correction which are accompanied by “reciprocating sigla, or signes-de-renvoi.”98 The first siglum is the “dotted obelus” or lemniskos (˙/.), which is used “to mark 97 The vertical bar in the transcription indicates a line break, with the following word appearing to the left of the column and the new line. 98 Hill adopts this term from Neudorf and Liu at ArchBook (http://drc.usask.ca/projects/arch- book/signes_de_renvoi.php); cf. Muzerelle’s definition at Codicologia (http://codicologia.irht.cnrs.fr). Hill, “Siglum,” 2 n. 5. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 131 short omissions and their restorations.” 99 It should be noted, however, that not all of these corrections can be characterized by “short omission.” For example, the correction on p. 228A (Deut 28:31) constitutes fourteen words and a total of fifty-five characters—equal to three lines of a standard column—including abbreviations. 100 Many of these extended corrections follow the inverted “pyramidal schema” (Figure 45), like those in א(01).101 Although both the horizontal (÷) and oblique lemniskos (˙/.) are present in B(03), the earliest correctors prefer the latter. 102 In five instances, I have found an undulated form of the lemniskos, similar to those found in א(01) (Figure 45).103 FIGURE 45: UNDULATED LEMINSKOS IN B(03) (P. 213A) AND א(01) (F. 241V; © BRITISH LIBRARY) In every case, the reinker left the siglum to fade, and on p. 223A (Figure 48), an oblique lemniskos was added as a substitute. Only the marginal sigla appear to be undulated, while the corresponding lemniskoi are oblique. These lemniskoi are not to be confused with the S-shaped siglum, which sometimes occur horizontally, with dots (see below). As mentioned in the previous section, Versace identifies two early groups of lemniskoi corrections: those by B2 and B3. Unfortunately, it is not clear in every case how he distinguishes the two. Although we will deal with this problem more in the following chapters, it will be helpful to offer some description of the patterns and differences between Versace’s two groups. 99 In addition, this siglum is also called an obelos periestigmenos. Hill, “Siglum,” 1. For other uses of the lemniskos see Turner and Parsons, Greek Manuscripts, 14 n. 75; McNamee, Sigla, 18. 100 Other large corrections, accompanied by lemniskoi, can be found on pp. 213A, 223A, 230A, 943C. 101 Milne and Skeat, Scribes, 40. 102 Examples of the horizontal lemniskos by Versace’s B2 include pp. 144A (column only), 458C, 498C. 103 See pp. 213A, 223A, 224A, 228A, 230A. CHAPTER 4 132 TABLE 17: LEMNISKOS CORRECTIONS IN VERSACE104 B2 B3 Lemniskoi (˙/.) 122 24 FIGURE 46: CHART OF LEMNISKOS CORRECTIONS IN B(03) From the above table and figure, it is clear that Versace’s B2 is responsible for the majority of the lemniskos corrections. However, this disparity fades in the NT, with B2 using the siglum thirteen times, and B3 eleven times. Versace’s B2 preferred copying these corrections to the right of the column, only placing 28 of the 122 to the left. On the other hand, B3 does not show any clear preference (10x right/14x left). In most cases, the position is not dependent on the location of the omission in the line nor the space in the margin (the outside margins provide the most space). While it is likely that B3 represents another “scriptorium” hand, which did not copy any of the main text,105 Versace is not equally clear whether B2 goes back to the scribes or a different hand. Here, I will mention one example which seems to be attributable to one of our three scribes. On p. 581C (1 Esd 5:9), we find a lemniskos correction in the work of scribe B betraying the hand of scribe C (Figure 47). 104 Data for Table 17 and Figure 46 can be found in Versace, Marginalia, 114–18, 134–35. 105 As seen already, the hand of B3 does not copy in the Biblical Majuscule hand of the main text. This alone does not disqualify B3 from copying the any of the codex, as a scribe could adopt different hands for various purposes. However, it is clear from the corrections of Versace’s B1 (and some of B2) that the scribes of B(03) executed marginal corrections in “Biblical Majuscule.” Furthermore, there are no clear parameters in the work of B3, which could be used to align the corrections and marginalia (i.e., section numbers, diplai, stichometry) with the work of our three scribes. 0 5 10 15 20 25 30 Ge ne sis Ex od us Le vit icu s Nu m be rs De ut er on om y 1 K in gd om s 2 K in gd om s 3 K in gd om s 4 K in gd om s 1 C hr on icl es 2 C hr on icl es 1 E sd ra s 2 E sd ra s Ps alm s Pr ov er bs Sir ac h Ju di th To bi t Ho se a Am os Ha ba kk uk Isa iah Je re m iah Ba ru ch La m en ta tio ns Ez ek iel Da ni el M at th ew M ar k Lu ke Jo hn Ac ts 1 P et er 1 C or in th ian s He br ew s B2 B3 THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 133 FIGURE 47: SCRIBE C CORRECTION IN 1 ESDRAS 5:9 (P. 581C); COMPARE WITH P. 840B Because this correction is full-sized, rather than compressed, it is easier to compare the paleography with the main text. In the marginal correction, there are two features unique to the scribe of Psalm 77–Tobit: the και-compendium with a hooked oblique and the alpha with vertical orientation and looped apex (see Chapter 3). While this is only one example, it would not be surprising to find that scribes were occasionally responsible for corrections in the work of others.106 The next form of marginal correction involves the utilization of inverted arrows called ancorae (↑↓).107 While these may look like arrows pointing to the place of correction, the directional indicator is actually the shaft.108 Hill rightly describes the function of ancorae corrections as “mainly for longer restorations,”109 though one may spot short additions of only two or three words (pp. 94A, 176C). Like the lemniskos corrections, Versace attributes this method of correction to both B2 and B3 (Table 18). Furthermore, it is clarified that the ancorae are usually accompanied by lemniskoi in B(03), and as result we should refer to them as ancorae–lemniskoi (↑↓ and ˙/.) corrections.110 This combined siglum allows for greater precision in locating the omission. Since the ancorae only appear next to the line of the error, the lemniskos is copied supralinearly to the exact location of omission. In most cases, the corresponding ancora and lemniskos occur at the beginning and end of the marginal corrections. 106 This was shown to be the case in א(01) by Milne and Skeat. See especially the corrections of scribe D in Milne and Skeat, Scribes, 9–11, 41–42; cf. Jongkind, Scribal Habits, 46–48, 247; Malik, “Corrections,” 247–248, 250. 107 Hill, “Siglum,” 1. 108 The sigla in B(03) mostly have a pointed head. See an example of the curved ancora on p. 190C. McNamee, Sigla, 11–13; Hill, “Siglum,” 1 n. 3; For the use of the ancorae and lemniskos in literary papyri, see Daris, “PSI 977,” 7–22. 109 Hill, “Siglum,” 1. 110 Admittedly, the hybrid Latin and Greek designation is not entirely satisfactory. There is at least one example of an ancorae–lemniskoi correction, where ancorae are used on both sides of the column and the marginal correction (p. 200B). There are three ancorae corrections, which lack lemniskoi: p. 194B, Deut 2:22, p. 198B, Deut 4:34; p. 198C, Deut 4:42. CHAPTER 4 134 TABLE 18: ANCORAE–LEMNISKOS CORRECTIONS IN VERSACE111 B2 B3 Ancorae–Lemniskoi (↑↓ and ˙/.) 70 3 FIGURE 48: CHART OF ANCORAE-LEMNISKOI CORRECTIONS IN B(03) The majority of ancorae–lemniskoi corrections by B2, and all three of B3, are copied in the lower margin and only seven at the top of a page. In these seven emendations, the error occurs in the top- half of the column. 112 Although we will further discuss the attribution of ancorae–lemniskoi corrections to the three scribes in the following chapter, it is notable that they are virtually absent from the NT. In support of the directional ancorae, some corrections are accompanied by the words ανω and κατω to instruct the reader where to look (Figure 49). Versace classifies these marginalia with B2, but acknowledges the similarities of hand with B3. Because the writing of ανω and κατω is not in Biblical Majuscule it is difficult to know why B2 is a likely candidate, especially since every other addition is in this hand. The only clear reason for Versace’s conclusion is the assumed connection with the ancorae-lemniskoi of B2, and the presence of similar annotations in א(01). Milne and Skeat attributed the cursive ανω and κατω notes to the scribes of א(01) and utilized them as an anchor for dating the manuscript in the fourth century.113 As mentioned in Chapter 1, this commonality is one of Versace’s proposed links between the two codices.114 111 Data for Table 18 and Figure 48 can be found in Versace, Marginalia, 118–22, 135. 112 The ancora-lemniskos on p. 204A marks an omission at the twenty-second of forty-four lines. 113 Milne and Skeat, Scribes, 62; cf. Nongbri, “The Date.” 114 Versace, Marginalia, 17. 0 5 10 15 Exo du s Lev itic us Nu mb ers De ute ron om y Jos hu a Jud ges 4 K ing dom s 1 C hro nic les 2 C hro nic les 2 E sdr as Psa lm s Ecc les ias tes Job Sir ach Est her Isa iah Eze kie l Da nie l Ma tth ew B2 B3 THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 135 FIGURE 49: ΑΝΩ AND ΚΑΤΩ NOTES (P. 1219C, DAN 5:4) Hill’s primary interest is in the corrections marked by the S-siglum, the last of the three signes-de- renvoi. This method was “used by the scribes mainly, though not quite exclusively, for marking short, variant readings.”115 Hill does not, however, consider these to be corrections in the proper sense, but legitimate, alternative readings.116 Expanding the work of Tischendorf, who identified five prima manus S-siglum corrections in Matthew, Hill argues that the eight occurrences in the gospel could indicate the presence of a second exemplar with alternative readings (see §4.1.3).117 We will examine this argument further in Chapter 6, but will now discuss the forms of the S-siglum in B(03). Versace assigns this method of correction to both B1 and B3 (Table 19). In the first set of additions by B1, Versace suggests, contrary to Hill, that these were copied from the same exemplar as the main text. 118 Although the typical form of the S-siglum is vertical, it can also appear horizontally and with double dots.119 On p. 238B (Josh 1:6), this change of orientation appears within the same correction, where the supralinear siglum is vertical and the marginal one is horizontal. Versace also identifies thirty-two marginal sigla that are unaccompanied by text, thirteen of which 115 Hill defends Tischendorf and Versace’s conclusion that the S-siglum was utilized by the scribes of the main text, through the correction at Ezekiel 4:14 (p. 1147A). Here, he explains that the scribe wrote the main text with irregular spacing in anticipation of an alternative reading (_εω_λον). In the margin, Hill argues the same scribe copied the reading βεβελον, with the corresponding S-siglum. However, through this argument, he criticizes Versace’s identification of an erasure in verse 14, claiming that “there are no traces of any earlier letters in the spaces.” Upon further investigation, Versace’s claim seems to make the most sense of the correction. The image provided by Hill does appear to show traces of a beta in the gap. While the vertical stroke may also align with a stroke from the verso (p. 1148C), the two rounded portions of the beta are still visible. This would confirm Versace’s claim that βεβελον was erased and εωλον was copied using the shared letters from the original reading. Hill is correct that the typical erosion of the parchment from erasure is not visible here, and Versace agrees that the marginal addition is prima manus. This is a clear example where multispectral imagining (MSI) could provide more definitive evidence. Hill, “Siglum,” 2, 4; Versace, Marginalia, 88 n. 57. 116 Hill does distinguish some uses of the S-siglum from the “standard” practice of the scribes (e.g., p. 76A, Exod 23:23). Hill, “Siglum,” 3 n. 7, 20. 117 Pace Hill, Tischendorf does note the marginal reading at Matthew 11:19, but does not consider it to be prima manus. Tischendorf, Vaticanum, 14; Hill, “Siglum,” 10; cf. Tischendorf, Editio octava. 118 This was suggested to him by Carlo Maria Mazzuchi. Versace, Marginalia, 11 n. 10. 119 Versace, Marginalia, 11. CHAPTER 4 136 mark proper names in the main text. In addition, one finds isolated, supralinear S-sigla, where marginal siglum and text are either absent, erased, or covered by later marginalia. As seen in Figure 50, we do not find any B1, S-siglum corrections in the books where scribe changes occur,120 and only thirteen appear in the NT. TABLE 19: S-SIGLUM CORRECTIONS IN VERSACE121 B1 B3 S-siglum Corrections (S) 159 18 FIGURE 50: CHART OF S-SIGLUM CORRECTIONS IN B(03) While the S-sigla of B1 predominantly mark single word corrections, there are a few larger examples (p. 394A, 2 Kgdms 24:17).122 Like the previous methods of addition, the S-siglum corrections provide further information to the text. However, they are distinct in that they do not indicate missing readings, but alternative ones (see Chapter 6). The siglum is used once for correcting a transposition (p. 367C, 2 Kgdms 10:17),123 and, possibly, once to mark a place where the text is not found in the Hebrew (p. 1066C, Jer 3:8).124 For the majority of corrections, there does not seem to be a consistent method of copying. A different word form might be completely recopied in the margin (ευδοκηϲεν 120 There is only one example of an S-siglum correction by Versace’s B1 in the work of scribe C (p. 745A, Prov 26:24), and Versace admits that this one may even be attributed to B3 instead. Versace, Marginalia, 88 n. 48. 121 Data for Table 19 and Figure 50 can be found in Versace, Marginalia, 83–89, 137. 122 This correction involves two changes. The first change is the substitution of ηδικηϲα for ηµαρτηκα, and is likely the reason for the S-siglum. The second is a correction of omission that resulted from haplography after the repeated phrase εγω ειµι. Thus, this marginal correction amounts to six words. 123 δαυειδ απεναντι ϲυριαϲ → ϲυρια απεναντι δ̅α̅δ̅ 124 Versace claims that the position next to each other does not mean the marginalia are connected. Versace, Marginalia, 90. 0 10 20 30 Exo du s Lev itic us Nu mb ers Jos hu a Jud ges 1 K ing dom s 2 K ing dom s 3 K ing dom s 4 K ing dom s 2 C hro nic les 1 E sdr as 2 E sdr as Psa lm s Pro ver bs Wi sdo m Na hum Isa iah Jer em iah Eze kie l Da nie l Ma tth ew Luk e Joh n Ac ts B1 B3 THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 137 → ευδοκηϲει; p. 1059B, Isa 62:4), or corrected simply by adding the correct letters in the margin (αυτου → αυ s[των]; p. 474A, 4 Kgdms 19:19). In 4 Kingdoms 10:8 (p. 458B), the reading την θυραν τηϲ πυληϲ πολεωϲ was corrected by adding the S-siglum above πυληϲ and the marginal reading πολεωϲ (Figure 51). It is unlikely that the scribe or corrector intended the reading to become την θυραν τηϲ πολεωϲ πολεωϲ. Therefore, the S- siglum must indicate that the reading πολεωϲ is given as an alternative, not a correction, to the combined reading πυληϲ πολεωϲ. If this were a proper correction, we would expect to find a simple deletion of πυληϲ, using one of the methods discussed below. FIGURE 51: S-SIGLUM CORRECTION IN 4 KGDMS 10:8 (P. 458B) In total, the spelling of proper names constitutes twenty-four, or nearly one quarter, of the S-siglum corrections by B1.125 It is noteworthy that these additions often change their spelling, but not the names themselves.126 This feature provides some insight into the characteristic concern for spelling by the scribes of B(03) (see Chapters 5 and 6). Like the other signes-de-renvoi, the S-siglum is also found in א(01).127 Milne and Skeat give an example of a S-siglum correction by scribe D at Philippians 3:10 (f. 284r); another can be found at Mark 12:20 (f. 225r).128 However, Milne and Skeat do not consider this symbol S-shaped, but as a “caret identical in shape with D’s filling mark (>).”129 While most of the examples listed by Milne and Skeat do imitate line fillers, the correction at Mark 12:20 has more in common with the s-sigla of B(03). It may be that, at least in א(01), the S-shape developed from the earlier and more attested diple shape (Figure 52). There is at least one instance of a similar siglum in B(03), though this may be the result of fading ink (p. 474A, 4 Kgdms 19:19; Figure 53).130 125 This number does not include changes of case in a proper name, like λιβανω → του λιβανου (p. 253C, Josh 12:7). It does, however, include the thirteen isolated sigla mentioned above. 126 However, see εξ ιερειχω → εκ γαλγαλων (p. 443C, 4 Kgdms 2:1). 127 Hill, “Siglum,” 9. 128 Thanks to Dirk Jongkind for directing my attention to this correction. 129 Milne and Skeat, Scribes, 42–43. 130 In addition to א(01), Hill also identifies one S-siglum correction in P75 and one translational gloss in P72 using mirrored s-sigla (S Ƨ). Hill, “Siglum,” 7–8. CHAPTER 4 138 FIGURE 52: S-SIGLUM IN B(03) AND א(01): P. 1253B; FF. 284R, 284V (©BRITISH LIBRARY) FIGURE 53: DIPLE SIGLA IN B(03) (P. 474A, 4 KGDMS 19:19) 4.2.2 Methods of Deletion In contrast to the methods of addition, we find four methods of correction which are employed to remove unwanted or accidental information: (1) supralinear deletion dots (˙˙˙); (2) round hooks (‘ ’), which mark the beginning and end of a correction; (3) oblique cancelation strokes (///); and (4) erasures.131 Some corrections are marked using methods (1) and (2) (p. 467B, 4 Kgdms 15:35), and others using methods (1) and (3) (p. 138A, Num 1:5). We cannot, however, be certain that they always originate from the same hand.132 As mentioned before, Tischendorf and Fabiani attribute the round hooks to another fourth-century hand, and Versace notes that both B2 and B3 use methods (1) and (2). The reinker (Versace: B18) is also responsible for some of the deletion dots, oblique cancelation strokes, and erasures. This outline is further complicated by the fact that some deletions may simply be reinforcements of earlier corrections, while others originated with the reinker (p. 164A, Num 16:47). The most frequent method of deletion in the work of scribe A is the use of deletion dots. These can be utilized to delete a single letter or nearly complete lines of text (p. 55A, Exod 7:24). While the most common format is for each letter to receive one, supralinear dot, we also find examples of additional dots over the letters (p. 167B, Num 19:15). In the work of scribe A, round hooks 131 Certainly, the use of dots, hooks, and dashes constitutes some manner of added information to the codex. However, their express purpose is to delete text, even when the corrections are accompanied by other methods of addition (i.e., supralinear text). On deletions, see also Hill, “Siglum,” 2. 132 For a similar anthology of deletion methods in literary papyri (especially in P.Oxy. 24 2404 + P.Laur. inv. III/278; LDAB 69), see Montanari, "Correcting a Copy,” 1–16. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 139 are reserved for deletions of multiple lines and only appear twice, from my counting.133 One use of the round hooks is for the elimination of dittography (p. 209B, Deut 12:18), while the other is simply for deleting unwanted text (p. 199C, Deut 5:14). Many of the oblique cancelation strokes appear to be from the reinker, since they are often darker than the original ink, and they are regularly accompanied by supralinear corrections from the reinker. 134 One possible early deletion with oblique strokes can be seen on p. 263A (Josh 19:51). Similar to scribe A, the work of scribe B contains numerous corrections with deletion dots. However, in contrast to scribe A’s text, there are also many deletions using round hooks. I have found twenty-nine instances of this method in the work of scribe B, and except for three (p. 671A, Ps 75:4; p. 1129A, Bar 2:18; p. 1426A, Jas 1:3), all indicate cases of dittography.135 In one instance (p. 1488B, Gal 1:11), το ευαγγελιον is copied three times over, with the first two occurrences marked redundant by hooks. While the two sets of round hooks in scribe A’s work mark multiple lines, those in scribe B’s text can range from a single word (p. 1087A, Jer 20:10) to four lines (p. 1479B, 2 Cor 3:13). This rise in deletions suggests that scribe B was more susceptible to errors from parablepsis than scribe A. Deletions through oblique cancelation strokes are rare in the work of scribe B (p. 441B, 3 Kgdms 22:39; p. 1479B, 2 Cor 3:13), and the reinforcer appears responsible for the majority. We also find deletions by supralinear dots, round hooks, and oblique cancelation strokes in the work of scribe C.136 While the most frequent method is the use of deletion dots, there are at least nine corrections with round hooks. Three of these mark short deletions that were not copied from dittography (p. 857B, Sir 22:26; p. 868A, Sir 30:38; p. 926A, Jdt 13:13) and are accompanied by deletion dots. A fourth correction (p. 830A, Wis 18:14) may be the consequence of a scribal leap backwards to the words περιεχουϲηϲ τα παντα, since ‘τα̇̇ π̇α̇ντ̇α̇̇’ is recopied after the word µεϲαζουϲηϲ in the following line (Figure 54). 137 The five remaining corrections using round hooks mark larger 133 Turner and Parsons call these hooks περιγραφαι (not to be confused with paragraphoi). Turner and Parsons, Greek Manuscripts, 16. 134 This technique of deletion is called διαγραφειν. Turner and Parsons, Greek Manuscripts, 16. 135 There are two cases of dittography that are marked with round hooks below the base line (p. 394A, 2 Kgdms 24:17; p. 1058B, Isa 60:21), but they appear to be the work of the reinker. 136 While the convention is to assign one oblique stroke to each letter, I have found one example where a single cancelation stroke passes through two letters (p. 941A, Tob 11:16). 137 The repetition is omitted in both א(01) and A(02). CHAPTER 4 140 reduplications of text. 138 If Psalm 77–Tobit was copied by a third scribe, this may indicate a marginally higher tendency to recopy text than scribe A, but far less often than scribe B. We must keep in mind that scribe B copied the largest portion of the manuscript (915 pp.), while scribe C copied the least (277 pp.). If, as is likely, scribe A began copying from Genesis 1:1, then he or she copied around 356 pages. FIGURE 54: SCRIBAL LEAP BACKWARDS (P. 830A, WIS 18:14) Following the suggestion of Tischendorf and Fabiani, one might see the secondary nature of the round hooks in examples where earlier corrections are made to the repeated texts (p. 467B, 4 Kgdms 15:35; p. 1479B, 2 Cor 3:15–16).139 In 4 Kingdoms 15:35, the first writing is corrected εθυϲια\ζ/εν, only to be marked later with deletion dots and round hooks (Figure 55).140 The reinking was given to the second εθυϲιαζεν, which was copied correctly the first time. On p. 1479B (2 Cor 3:15–16), we find another example of supralinear correction in reduplicated text (περιε̸\αι/ρεται). This correction was likely made before the text was marked as redundant with round hooks. However, as we will see in Chapter 6, there are strong paleographic reasons to associate the supralinear correction with the reinker. The dark ink surrounding the dittography belongs to the second of Versace’s reinkers (Figure 55). FIGURE 55: SUPRALINEAR CORRECTIONS IN DITTOGRAPHY (P. 467B, 4 KGDMS 15:35 ; P. 1479B, 2 COR 3:15–16) This late corrector did not, however, spot dittography in every case (p. 1375C, John 17:18; p. 1394A, Acts 8:26; p. 1442C, 2 John 1:10; p. 1505C, Col 4:2). In these examples, round hooks are missing and the reinker covered the first occurrence of a word or phrase, while the second was untouched. This 138 See p. 835A, Sir 3:5; p. 836A, Sir 3:26; p. 867A, Sir 30:23; p. 914B, Jdt 6:7; p. 920A, Jdt 9:2. 139 See also p. 835A, Sir 3:4. 140 Although less likely, it is not out of the question that scribe B was capable of catching both the spelling mistake and the resulting dittography. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 141 deviates from the routine of reinforcing the second occurrence, when early or late round hooks are present. In addition to these three methods of deletion, we find many corrections by erasure in B(03). 141 Because of the additional layers of ink in the manuscript and the deterioration of the parchment over time, it can be difficult to identify these erasures. In papyrus manuscripts, erasures were often executed using a sponge.142 However, many of B(03)’s observable erasures appear to have been achieved through scraping with a penknife (Figure 56). As we have already noted, some of these erasures were important to Tischendorf, because they represented in scribendo corrections, where the original scribe erased and recopied text without disrupting the format of the line (see §4.1.3). Conversely, the other early correctors are likely to have been responsible for many of these erasures (Versace: B2). Eventually, the ink had sufficiently faded that the reinker could copy directly over the text without needing to erase (p. 1195A l.27, Ezek 41:15). However, late erasures from both reinkers are also visible throughout the codex (p. 627A, Ps 6:11). A complete examination of erasures in B(03) would require MSI and access to the manuscript in the BAV.143 Therefore, our study will deal cautiously with those erasures in books with scribe changes and in the NT. FIGURE 56: ERASURES WITH SCRAPED PARCHMENT (P. 202A); LARGE ERASURE AND COLUMN-END CORRECTION (P. 298C LL. 37–46) In general, what do these methods of deletion say about the types of error being corrected? The clearest case can be made for the round hooks, which usually correct instances of dittography. However, the evidence seems to favor the conclusions of Tischendorf, Fabiani, and Versace—that these hooks were added by correctors other than the scribes. The deletion dots, which are used 141 Hill briefly mentions this method in “Siglum,” 20 n. 95. 142 Turner and Parsons, Greek Manuscripts, 16; Canart, Lezioni, 79; Cribiore identifies one school exercise that may have been partially erased with a wet finger. Cribiore, Writing, 95 n. 194. 143 One can see the large number of erasures from Swete’s apparatus on the OT of B(03) (abbreviated as sup ras). However, it is not always clear how they were identified, as the recent images often show little trace of erasure. CHAPTER 4 142 frequently, appear to be the default method for deleting unwanted text, while oblique cancelation strokes are used to delete letters or words that need to be corrected with supralinear additions. As Tischendorf argued, erasures can be helpful for identifying in scribendo corrections, but these are often challenging to find. 4.2.3 Other Methods of Correction We must briefly mention two last methods of correction, which cannot be neatly categorized with the additions or deletions. The first involves the correction of letters to numerals. As discussed in Chapter 3, this occasionally happens in the Greek OT, and once in the NT with a correction of one numeral to another (/ιβ̅̅ (or ϥ̅β̅) → /β̅) (p. 1283C, Mark 5:13). In Nehemiah 7:33 (p. 614C), we argued that the numeral ρ ̅ was actually a correction from ναβιααρ to ναβιαα ρ,̅ which is followed by the correction ϲανανατ → ϲανανα /γ̅ (p. 614C, Neh 7:38), a few lines later. The latter was executed by erasing part of the tau’s cross-bar to form a gamma, then adding an oblique stroke and a macron. In this way, the correction towards a numeral can include both addition and deletion. Finally, corrections of word order occasionally occur through the use of supralinear, oblique strokes (e.g., p. 127B, Lev 21:18; p. 195A, Deut 2:37; p. 680B, Ps 87:10). Versace identifies this method with B2, but admits that at least one example (p. 1163B, Ezek 19:9) belongs to the tenth or eleventh century (B18). 144 These corrections work by placing incrementally more strokes above words, according to their intended order. Corrections of transpositions with oblique strokes can also be seen on ff. 36r and 273v of א(01), but are associated with a subsequent corrector, Ca.145 While the last two forms of correction, numerals and transpositions, are rare in comparison to the other methods, they represent another of the many options available to the scribes and earliest correctors of B(03). 144 Versace, Marginalia, 15–16 n. 20, 75 145 Milne and Skeat, Scribes, 46–47. THE SCRIBES AND EARLY CORRECTORS OF THE CODEX 143 TABLE 20: SUMMARY OF CORRECTION METHODS IN B(03) Corrections by Addition - supralinear additions: \/ - line-end or column-end additions: ]txt[ or ≞ - lemniskos corrections: ˙/. - ancorae–lemniskoi corrections: ↑↓ and ˙/. - S-siglum corrections: S Corrections by Deletion - deletion dots: ˙˙˙ - round hooks: ‘ ’ - oblique cancelation strokes: /// - erasure Other Methods of Correction - corrections to numerals - corrections of transposition: / + // + ///, etc. 4.3 Summary In this chapter, we have examined the modern reception of B(03) from the vantage point of its corrections. While the scribes of B(03) did not receive much attention until Tischendorf, it had long been recognized that the manuscript was corrected throughout history. At times this project was understood negatively, as an attempt to conform to the Vulgate (Erasmus, Mill, and Wettstein). Yet, others recognized the antiquity and, therefore, the value of some corrections early on (Rulotta, Wagstaffe, and Hug). One assumption that remained until Tischendorf was the bipartite division of correctors between the original scribe or diorthotes and the late reinker. Likewise, Tischendorf’s paradigm of three correction layers would eventually need to be expanded. The Roman editors who were contemporary to Tischendorf acknowledged the likelihood of various corrections made throughout the fourth to fifteenth centuries, but categorized them all under a single siglum (B2). It was Pietro Versace’s recent work that first attempted to give a comprehensive outline of the marginal hands in B(03), including the correctors. However, this survey left us with questions surround the relationship of the earliest correctors to the scribes outlined in Chapter 3. Can we identify prima manus corrections in the codex? What might contemporary corrections—whether from the copyists themselves or a diorthotes—tell us about the three scribes and their copying? Our typology of corrections in B(03) provides us with opportunities to observe broad correspondences with our division of scribes. The evidence provided in Versace’s work has shown that all examples of corrections at column-ends are found in the work of scribe A, apart from one in scribe C’s work. Although Versace classifies column-end corrections as B2, at least one of these has a clear affinity with the hand of scribe A (p. 73B, Exod 21:31). While scribes A and B have many S- siglum corrections in their text, we find only one in Scribe C. It is clear from the use of round hooks CHAPTER 4 144 that scribe B was more prone to rewrite text (dittography) than the other scribes, but we still find more examples from scribe C than A. In the following chapters, we will explore paleographic questions in more detail. However, the clear example of scribe C’s lemniskos correction in the work of scribe B (p. 581C, 1 Esd 5:9) opens possibilities for further investigation into the way the earliest scribes and correctors interacted with one another’s work. When compared to the Greek OT, Figures 46, 48, and 50 reveal the relatively few marginal corrections in the NT. This dwindling in the NT is reminiscent of other features we observed in Part I of our study: the constraining of nomina sacra to the core four; the absence of numeral abbreviations (except Mark 5:13); and the fading use of ekthesis after Luke. When taken together, these variations must reflect some of the complexities involved in producing a volume like B(03). The remaining two chapters will explore these complexities through a comparison of corrections between our three scribes in 1 Kingdoms, Psalms, and the NT. CHAPTER 5 Early Corrections in 1 Kingdoms and Psalms In the last chapter, we summarized previous attempts to outline the correctors of B(03) and produced a typology of early corrections in the manuscript. The recent work of Pietro Versace provides a comprehensive description of the marginalia in the codex, including corrections. However, his investigation avoids speculating about the relationship of the earliest corrections to the scribes of B(03). It is clear that Versace’s fourth-century hand B3, was not responsible for copying any of the main text, but is confined to corrections and section numerals. Yet, Versace’s B2 remains a vague category, which likely contains a mixture of corrections that were made by the scribes and by a diorthotes. As we saw in Chapter 4, we may also find corrections by one scribe in the work of another. All three of these editorial figures—scribes, diorthotai, and colleague scribes—can be seen as members of Versace’s category B2. Milne and Skeat’s study of the correctors in א(01) warns of the difficulty in assigning corrections to a particular hand.1 Since many of the corrections in B(03) are short additions of individual letters and words, or deletion, any identification of a specific scribe is tentative. Even when extensive writing is present, it is copied in a smaller hand and in condensed form. On account of this difficulty, we will resort to the methodology developed in Chapter 3, and compare corrections in books with scribe changes—1 Kingdoms and Psalms. If space permitted, this study might also benefit from a comparison across the break at Tobit–Hosea. Yet, the change of books and genre at Tobit–Hosea confronts us with additional uncertainties and likely reduces the cogency of our comparative method. Constraining ourselves to 1 Kingdoms and Psalms still gives us the opportunity to compare all three of our scribes. Chapter 6 will draw on the conclusions from this study, primarily those concerning scribe B, and develop them further in our analysis of the NT of B(03). 5.1 1 Kingdoms In Part I, we developed the argument originally made by Ezra Abbot for a transition of scribes at 1 Kingdoms 19:11 (p. 335). While he recognized this division on account of the abrupt decline in the use of ekthesis, we found support in the codicological data and other paratextual features. This 1 Milne and Skeat, Scribes, 40; Malik states with more assurance that “the majority of corrections [in Mark] can be assigned with relative confidence.” Malik, “Corrections,” 212. CHAPTER 5 146 conclusion led to further observations about the differences in paleography, such as the shape of the letters alpha, delta, lambda, and even the tailed-mu. Therefore, the break at p. 335 (1 Kgdms 19:11) will continue to function as an anchor for the comparison of early corrections in the work of scribes A and B. 5.1.1 The Text of the Four Books of Kingdoms in B(03) The scholarship on the Old Greek (OG) text of 1–4 Kingdoms, including all of its recensions and daughter versions, is extensive.2 Thus, space only allows us to summarize the most crucial elements for this examination. First, it is important to recognize that 1 Kingdoms fits within the larger collection of 1–4 Kingdoms, which are equivalent to the two Hebrew books, Samuel and Kings.3 In Chapter 3, we saw briefly that 1–2 Kingdoms, 3–4 Kingdoms, and 1–2 Chronicles were linked together by the recopying of verses. Therefore, 2 Kingdoms 1:1, 4 Kingdoms 1:1, and 2 Chronicles 1:1 appear twice in B(03): once at the end of the previous book, and again in its proper location. Emanuel Tov has argued from the evidence at Qumran that “large books like Samuel–Kings were usually not written in a single scroll,” and that 2 Kingdoms may have been contained in two separate rolls. 4 If this was the case, then we might assume that the original intention of reduplicating the first verse of the following work was to aid the readers as they transitioned between scrolls. This may also explain why we sometimes find textual variation between the two occurrences of the same verse (Figure 57) κατενιϲχυϲεν ϲαλω µων υιοϲ δαυειδ επι τηϲ βαϲιλειαϲ αυτου· και κ̅ς ̅ο θς̅ ̅αυτου µετ αυ του· και ηυξηϲεν αυτο ̅ ειϲ υψοϲ:>>>- και ενιϲχυϲεν ϲαλω µων υιοϲ δαυειδ επι την βαϲιλειαν αυτου και κ̅ς ̅ο θς̅ ̅αυτου µε τ αυτου και εµεγαλυ νεν αυτον ειϲ υψοϲ· FIGURE 57: 2 CHRONICLES 1:1 ON P. 521 AND P. 522 IN B(03) The best explanation for the variation of 2 Chronicles 1:1 is that the exemplar had different readings in both places. At the end of Greek literary rolls, these repeated catchphrases are called reclamantes 2 See the relevant bibliography in Hugo, “Basileion I and II,” 173–179; On the recensions of 1 Kingdoms, see the important work of Brock, The Recensions; cf. Kauhanen, Proto-Lucianic. 3 See b. Baba Bathra 14b. Although four titles were given to Samuel–Kings, it was common for Christian canon lists to count them as only two. Gallagher and Meade, Canon Lists, 67–69, 82. 4 Tov comes to these conclusions in an attempt to explain the complex alternation of translations in Samuel– Kings (from kaige to non-kaige). This reminds us that the division into four separate books was probably not original to the work (cf. Eusebius, Hist. 6.25). Tov, “The Methodology," 489–499; Hugo, “1–2 Kingdoms,” 128. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 147 (sg. reclamans).5 Why these were preserved in codices is unclear. Indeed, B(03) is not the only LXX manuscript to contain reclamantes at the end of one or two of these books. Still, it is the only manuscript, that I am aware of, which has it in all three (Table 21). TABLE 21: LXX MSS WITH REDUPLICATION IN 1 SAMUEL, 1 KINGS, OR 1 CHRONICLES 2 Kgdms 1:1 in 1 Kgdms 4 Kgdms 1:1 in 3 Kgdms 2 Chr 1:1 in 1 Chr B(03)122, M, 119, 158, 509, Aeth B(03), 19+108, 82, 93, 127, Aeth B(03), 83, 127 While 1 Kingdoms needs to be understood alongside the other three books of Kingdoms, it is well known that the text of B(03) is far from uniform. The lasting contribution of Thackeray leads specialists to divide 1–4 Kingdoms into five textual groups:6 α= 1 Kingdoms ββ= 2 Kingdoms 1:1–11:1 (9:13 or 10:5) βγ= 2 Kingdoms 11:2 (10:1 or 10:6)7–3 Kingdoms 2:11 γγ= 3 Kingdoms 2:12–21:43 γδ= 3 Kingdoms 22:1–4 Kingdoms Sections βγ and γγ were later identified as the kaige text, which was an intentional attempt at bringing the Greek text into conformity with the Hebrew.8 This variation in textual forms has been attributed to shorter roll lengths in the Hellenistic period. Thus, when codices of 1–4 Kingdoms were copied, they were dependent on rolls with varying text quality.9 Fortunately, our study will focus primarily on 1 Kingdoms, which consists of only one text group (group α).10 This portion of B(03), along with its allies, has generally been accepted as our best witness to the OG.11 However, it should 5 Schironi notes that the use of reclamantes at the end of rolls is absent after the second century CE. However, West gives one irregular example of in P. Oxy. IV 698 (early-III), in which the reclamans, taken from Book 2 of Xenophon’s Cyropaedia, occurs before the coronis and end-title of Book 1. This irregularity fits the format found in B(03). West, “Reclamantes,” 314–315; Schironi, “Book-Ends,” 700; Johnson, Bookrolls and Scribes, 342. 6 Thackeray, “Four Books of Kings,” 262–66. 7 J. D. Shenkel argued that βγ begins at 2 Kingdoms 10:1. Shenkel, Chronology, 113–20. There has also been an attempt to mark the beginning of βγ at 2 Kingdoms 10:6. See Wirth, Die Septuaginta, 199–201; Wirth, “Dealing with Tenses,” 185–97. Shenkel’s position was recently defended by Pinto, “The Beginning of the ΚΑΙΓΕ,” 14–33. 8 This identification was made by Barthélemy, Les devanciers, see especially pp. 91–109. On the origin and purpose of the καί γε particle and the kaige revision, see Aitken, “The Origins,” 21–40. 9 Kim, “Kleinrollensystem,” 231–242; cf. Kreuzer, “5.4 Septuagint (Samuel).” 10 This does not mean B(03) is free of kaige readings in 1 Kingdoms. Aejmelaeus, “Kaige Readings,” 169–184. 11 The B-text is represented by B(03), Rahlfs 121–509 (given as ms group b), and the Ethiopic version (Aeth). This grouping was already given by Wevers as Bya2Eth (Brooke-Mclean) in 1948. Wevers, "Hebrew Variants,” 46; cf. Perttilä, Sahidic 1 Samuel, 27; Kauhanen, Proto-Lucianic, 13; Aejmelaeus, “Kaige Readings,” 171 n. 5. CHAPTER 5 148 not be said that the text of the codex simply is the OG of 1 Kingdoms.12 Anneli Aejmelaeus’ work on the Göttingen edition has led her to conclude that B(03) is representative of the “first Christian revision” of 1 Kingdoms, which existed at the beginning of the third century.13 Finally, we must note the absence of Göttingen editions for 1–4 Kingdoms. These editions are underway, but will not be released in time for our analysis.14 Instead, we will rely on Rahlfs- Hanhart’s edition, Swete’s smaller or manual edition, 15 and the Cambridge edition of Brooke, Mclean, and Thackeray (B–M). The last two print B(03) as the main text, while Rahlfs-Hanhart occasionally departs from B(03). Still, Rahlfs-Hanhart often follows B(03) even in the kaige sections. Besides these editions, the commentary produced by Fabiani (Vol. 6), which accompanied the Roman edition, can provide support for the identification of corrections. In fact, it appears that Fabiani’s commentary provided much of the data for Swete’s apparatus.16 5.1.2 The Earliest Corrections in 1 Kingdoms In the previous chapter, we provided a typology of the earliest corrections in B(03), which focuses on how information is either added or deleted from the manuscript. This typology is important from the perspective of text production, but does not provide meaning for the individual corrections. For example, we noted the S-siglum is used for corrections by addition, but that this does not usually indicate additions to the text. Instead, the S-siglum is often used for substitutions or spelling changes in proper names. However, from our outline of correction methods, we are able to examine the earliest corrections and the supposed errors they represent in the text of 1 Kingdoms. The following section has two aims. The first is to identify possible connections with the three scribes of B(03). This will be done primarily through paleographic analysis of the corrections. Since the hands of the corrections are far more varied than in the main text, we must begin with a brief description of the graphic similarities and differences, which will either assist or hinder our 12 Law and Kauhanen, “Methodological Remarks,” 78, 88–89; Hugo, “Text History,” 7. Kreuzer calls this “a somewhat milder Hebraizing revision,” in the non-kaige sections of B(03). Kreuzer, “B or not B?,” 291, 295. 13 Aejmelaeus, “New Perspectives,” 20–21. 14 Aejmelaeus, “Challenges,” (forthcoming). I am grateful for her willingness to share a prepublication version of this chapter. 15 Eberhard Nestle provided Swete with important corrections to the first edition based on new photographs of B(03). Swete, The Old Testament, 1: xviii. 16 The apparatus and the commentary rarely diverge from one another, and Swete occasionally includes Fabiani’s sigla B1 rather than the standard Ba (e.g., 1 Kgdms 8:21, 22). EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 149 identifications of the correctors. It will not be possible to classify a hand in every instance, but it is likely that numerous corrections identified as B2 by Versace can be attributed to an original scribe. In the process of identification, we may also distinguish corrections made in scribendo from those made after the text was finished.17 The second aim is to identify what these early corrections say about the initial copying of the text and how they compare between scribe A and B in 1 Kingdoms. To make this comparison, we will employ the following categories of correction: Additions, Omission, Substitution, Orthography, Nonsense, and Text Division.18 These categories will also be helpful in refining our understanding of the methods of correction, developed in Chapter 4. Because of the reinking, our analysis will begin with a larger set of corrections, as many cannot be confidently assigned to the earliest correctors. After examining the text of 1 Kingdoms and comparing the work of Fabiani, Swete, and Versace, I have initially identified 147 corrections that potentially go back to the earliest correctors: Fabiani’s B1 and B2, Swete’s Ba, and Versace’s B1, B2, and B3. This number is approximate since there is often disagreement over who was responsible for a correction; many corrections of the same type are even attributed to different correctors without clarification. However, we may be able to eliminate some later corrections with confidence. It is clear that orthographic corrections, which were made by leaving the ink unreinforced, were made by the later reinker. These include the deletion of the final-nu, the epsilon in ει (long-i), and the mu in the future form of λαµβανω (ληµψοµαι → ληψοµαι). Likewise, it will become clear through the paleography that the majority of orthographic corrections by supralinear addition were also executed by the reinker (B18). Since these constitute the majority of corrections, the number of potentially early corrections in 1 Kingdoms diminishes by nearly one hundred (see Orthography; Appendix D). Although the designation B18 does not belong to our early correctors, it remains crucial for us to identify the editorial habits of this late hand, primarily as a way of distinguishing earlier hands. Since marginal corrections give us the clearest evidence of scribal hands, we will begin by comparing them across the work of scribes A and B. This examination will often take us beyond the bounds of 1 Kingdoms, when there is little or no evidence for us to isolate. In the process, we will 17 On this, see Royse, Scribal Habits, 115–116 n. 65; 365–67.; Malik, P.Beatty III, 74. 18 These categories have been used for the examination of corrections and singular readings in manuscripts. Royse, Scribal Habits, 103–704; Jongkind, Scribal Habits, 147–246; Malik, P.Beatty III, 78–101, 117–139. CHAPTER 5 150 mention both the method and category of correction. After analyzing the marginal corrections, we will look for other distinctive features of the scribes from the remaining corrections. I will use the sigla BA, BB, and BC to denote corrections made by one of our three scribes.19 For those corrections that are indistinguishable, we may attribute them to one of the three scribes or B2, the diorthotes. Finally, with some corrections we follow Versace’s B3, whose fourth-century hand is distinguishable by its undulated letters (see Chapter 4). Because of the large sample size, we will look at them based on the category of correction and highlight those of importance. There are comparatively few marginal corrections in 1 Kingdoms, with only six by Versace’s B2 and four by B3 (Table 22). On the other hand, there are twenty-nine early marginal corrections in 2 Kingdoms, twenty-six of which are S-siglum corrections by Versace’s B1. Three S-siglum corrections appear in 1 Kingdoms, but Versace is right to deny their association with the original scribes by marking them B3. As seen in Table 22, there are no examples of column-end corrections, B1 S-siglum corrections, or ancorae-lemniskoi corrections in 1 Kingdoms. TABLE 22: EARLY MARGINAL CORRECTIONS IN 1 SAMUEL Corrections B1 B2 B3 Line-End — 5 — Lemniskos — 1 1 S-siglum — — 3 However, if we look to the rest of the work of scribes A and B, we find sufficient evidence to compare the paleography in larger marginal corrections like column-end corrections and ancorae-lemniskoi corrections. Paleography of the Early Marginal Corrections In the previous chapter, we noted that column-end corrections are only present in the work of scribe A, with one exception at Psalm 100:3 (p. 688A; Scribe C). Seven of the ten column-end corrections in the work of scribe A are copied in the same hand as the main text, albeit more compressed with ligatures and abbreviations. This allows us to confidently identify the majority of these corrections with scribe A. The first column-end correction (p. 73B, Exod 21:31) is a strong model of a prima manus correction (Figure 58). It is clear from the alphas and delta that the copyist was either scribe A or one trying to imitate the scribe’s hand. In addition to the script, the correction follows scribe A’s preference in using ekthesis as well as line-fillers (Chapter 3). 19 Capital letters are used to distinguish them from Swete’s sigla of correctors. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 151 FIGURE 58: COLUMN-END CORRECTION; EXODUS 21:31 (P. 73B) Three of the ten column-end corrections are copied in a smaller hand that is less consistent in form and uses the small, raised sigmas and omicrons (Figure 59). The first two (p 189C, Num 35:25; p. 216C, Deut 19:8) generally maintain the distinctive alphas and deltas of scribe A, but the final correction (p. 233A, Deut 31:14), in many cases, does not. For example, when we look at the alphas of µαρτυριου and παρα in the correction’s second line, it is clear that they have the vertical orientation of those found in scribe C. While some of the alphas slant to the left, like those in the first two corrections, they do not have the same sharp points. This is partly obscured by the reinking. p.189C, Num 35:25 p.216C, Deut 19:8 p.233A, Deut 31:14 FIGURE 59: COLUMN-END CORRECTIONS IN SMALL HAND However, there is further evidence to suggest a different hand for the third correction. First, there are two και-compendia with the obliques curved to the right (Figures 59–60). This is clearly seen beneath the reinforcement, which did not trace the curve. As in Chapter 3, the curved και- compendia continue to be important evidence of the work of scribe C. We have already witnessed a scribe C correction in the work of Scribe B (p. 581C, 1 Esd 5:9), so this possible identification should not be excluded. The second piece of evidence, which will become clearer in other corrections throughout B(03), is the change from the so-called “Coptic” mu to the Biblical Majuscule mu. The corrections that share the most traits with scribe A prefer the “Coptic” mu, while those of other correctors prefer the Biblical Majuscule mu. CHAPTER 5 152 FIGURE 60: CURVED ΚΑΙ-COMPENDIA (P. 233A); CF. FIGURE 59 Therefore, it appears that the majority of column-end corrections from Genesis to 1 Kingdoms 19:11 are prima manus, apart from the largest correction (p. 233A, Deut 31:14). In fact, this correction shares some, but not all, of the features we have attributed to scribe C. While Figure 60 exhibits the vertical alphas of scribe C and και-compendia with curved obliques, the correction as a whole (Figure 50) does not follow this consistently. Therefore, if this correction does not belong to scribe C, we should attribute it to a diorthotes. The ancorae-lemniskoi corrections also provide some recognizable features of the original scribes. For example, three consecutive ancorae-lemniskoi corrections in the work of scribe A contain the distinctive ornamental mu, discussed in Chapter 3 (p 111b, Lev 11:15; p. 124B, Lev 19:9; p. 134B, Lev 26:6; Figure 61). In the main text, scribe A uses this form pervasively, while it is very rare from scribe B and absent from scribe C. In addition, we see the slanted alphas, deltas, and lambdas that are clearest in the work of scribe A. Similarly, the third correction contains a και-compendium with an elongated, left-curved oblique, which extends below the vertical stroke of the kappa. This form of the και-compendium only appears in the corrections to scribe A’s text, which tends to exhibit the same distinctive alphas, deltas, and lambdas. Finally, it is only in corrections to scribe A’s text that we find ancorae on the right side of the correction, instead of the left (five times). There is one additional example with ancorae on both sides of the main column and the correction (p. 200Β). The use of two ancorae also happens once in the work of scribe B (p. 1156C), but we do not find a single ancora to the right of corrections in the work of scribes B or C. Therefore, it is probable that, like the column-end corrections, most of the ancorae-lemniskoi were made a prima manu. p.111Β, Lev 11:15 p.124Β, Lev 19:9 p.134B, Lev 26:6 FIGURE 61: ORNAMENTAL MU IN ANCORAE-LEMNISKOI CORRECTIONS EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 153 However, not all of these corrections can be easily attributed to scribe A. Versace notes the irregular και-compendium in the ancora-lemniskos correction on p. 215B (Deut 17:17) and suggests that this may be from a different hand (Figure 62).20 Again, we find the και-compendium with the curved oblique similar to scribe C, and a distinct change in the alphas, deltas, and lambdas. In addition, the form of the mu has changed from the “Coptic” style to the Biblical Majuscule. While Versace only identifies this correction as possibly originating from a different hand, the following four ancorae- lemniskoi corrections seem to exhibit similar features (p. 217B, Deut 19:21; p. 223A, Deut 24:13; p. 227A, Deut 28:12; p. 232A, Deut 30:13; Figure 63). In fact, the last two corrections may also contain curved και-compendia, which the reinker did not retrace. FIGURE 62: CURVED ΚΑΙ-COMPENDIUM IN ANCORA-LEMNISKOS CORRECTION (P. 215B; DEUT 17:17) p.217B; Deut 19:21 p.223A; Deut 24:13 p.227A; Deut 28:12 p.232A; Deut 30:13 FIGURE 63: SCRIBE C CORRECTIONS IN DEUTERONOMY Therefore, it is better to assign these corrections to either scribe C or a diorthotes, who copied similar και-compendia. When compared to the ancorae-lemniskoi corrections in the work of scribe C, we see some striking similarities (Figure 64). Most of the και-compendia have faint curves in their obliques; the alphas, deltas, and lambdas are copied with a more vertical orientation; and both forms of the mu are used, even in the same correction. It is also notable that the shape of the ancorae better fits those of scribe C, as the shaft is usually separate from the base. 20 Versace compares this correction to one on p. 202C. Versace, Marginalia, 120 n. 148. CHAPTER 5 154 FIGURE 64: ANCORAE-LEMNISKOI CORRECTIONS IN SCRIBE C21 While the differences between scribes A and C corrections are less striking than the differences in the main text, it is still possible to distinguish their hands. In general, we have seen enough paleographic evidence to suggest that the majority of column-end and ancorae-lemniskoi corrections were made a prima manu. While scribe B does not have any striking paleographic features like scribe A (the ornate mu) and scribe C (the curved και-compendium), it is likely that the majority, if not all, of the ancorae-lemniskoi corrections are prima manus. The alphas, deltas, and lambdas are generally consistent with those of scribe B (Figure 65), even when the alphas do not have the curved cross-bar. FIGURE 65: ANCORAE-LEMNISKOI CORRECTIONS IN SCRIBE B22 It is possible that the four ancorae-lemniskoi corrections in Daniel were copied by a different hand, but it is difficult to say. The correction on p. 1233A (Bel 24) contains a και-compendium that appears to have a slightly curved oblique. In addition, we find the only three occurrences of the 21 Right to Left: p.707A, Ps 138:12; p. 708A, Ps 139:12; p. 761B, Eccl 11:1; p. 780B, Job 12:23; p. 792B, Job 27:15; p. 795A, Job 30:4; p.879A, Sir 40:9; p. 906B, Esth 9:19; p. 907B, Esth 10:3. 22 Right to Left: p. 452B, 4 Kgdms 7:1; p. 608B, 2 Esd 12:10; p. 638A, Ps 24:9; p. 1028C, Isa 29:15: p. 1216C, Dan 3:95. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 155 corresponding ανω and κατ(ω), which accompany the corrections (Chapter 4).23 If these guiding words were copied by the same hand as the correction, their concentration here may suggest a different hand. A few of the line-end corrections preceding these examples also contain curved και- compendia (pp. 1216C, 1217A, 1220Β, 1228C, 1232A?). When we include the unique marginal correction at Daniel 9:27 (Figure 66),24 the evidence suggests a different hand that is responsible for the majority of corrections in Daniel. FIGURE 66: PYRAMID SCHEME CORRECTION IN DANIEL 9 (P. 1227C) The evidence we have collected from the column-end and ancorae-lemniskoi corrections generally favors the identification of prima manus. However, there are multiple instances where a hand resembling scribe C, or a different hand altogether, was responsible for corrections in the work of scribes A and B. When possible, our examination of 1 Kingdoms, Psalms, and the NT will use these examples as a guide for distinguishing prima manus corrections from the subsequent fourth- century corrections. Addition The most infrequent corrections are those of additions—that is, from the perspective of the corrector. Only four were executed in 1 Kingdoms and all appear in the work of scribe B. 23 There is one occurrence of the κατ(ω) abbreviation in the work of scribe C (p. 906B), but this is different in form and lacks the corresponding ανω. 24 Of the three scribes, this correction best fits the pattern of scribe A. Not only are the pyramid scheme corrections confined to the work of scribe A, but this correction also contains line-fillers and has the slanted alphas and deltas. It is still unique, however, since it is not accompanied by signes-de-renvoi, but attached to the column by an obelos. If not from the hand of scribe A, it likely originated with the diorthotes (B2). CHAPTER 5 156 TABLE 23: CORRECTIONS OF ADDITIONS IN 1 KINGDOMS25 SCRIBE B(03) CORRECTOR 1 Kgdms 22:3; p. 339A B µωαβ ‘κ̇α̇ι ̇ει̇π̇̇εν̇̇ π̇ρ̇οϲ̇’̇ ‘β̇α̇ϲι̇λ̇̇εα̇̇ µ̇ω̇α̇β̇’ και ει πεν προϲ βαϲιλεα µωαβ B2 and/or BB 1 Kgdms 23:15; p. 341B B εν τη καινη ζειφ ‘τη̇̇ κ̇α̇ιν̇̇η̇’ B2 and/or BB 1 Kgdms 23:26; p. 342A B εκ µερουϲ του ορ̇̇ου̇̇ϲ ̇τουτου (2) B2 or BB 1 Kgdms 26:21; p. 347B B εν οφθαλµοιϲ ϲου εν τη ϲηµερον… B2 or B18 The first two additions resulted in dittography. On p. 339A, the phrase και ειπεν προϲ βαϲιλεα µωαβ is copied twice and was corrected by using round hooks and deletion dots over the first occurrence. The error was likely caused by a scribal leap backward to the previous phrase, which ended with the word µωαβ. The second addition (p. 341B, 1 Kgdms 23:15) was also corrected with round hooks and deletion dots, but this time over the second occurrence, since only the first writing of τη καινη makes grammatical sense.26 As we saw in the previous chapter, it is unlikely that the round hooks were made by a diorthotai or the reinker, after the the text was finished. However, it remains unclear whether or not the deletion dots were added in scribendo. The third correction (p. 342A, 1 Kgdms 23:26) was made using deletion dots above the word ορουϲ. However, the consequent reading, εκ µερουϲ του τουτου, is nonsense and was eventually corrected by the reinker to εκ µερουϲ τουτου.27 Whatever the reason may be for this error, it is possibly linked to the presence of the same phrase three lines before. The alternative explanation is that τουτου was meant to be corrected to ετερου (Rahlfs 247, 376), but the change was never executed. The use of deletion dots, without the later round hooks, might indicate that scribe B was responsible for this faulty correction. 25 To avoid confusion, transcriptions in Chapter 5 and 6 only include the specific correction being discussed, unless an adjacent correction is significant for its understanding. Therefore, orthographic corrections will often be missing from transcriptions of the other types of correction. The following transcriptions are intended to provide a detailed visual representation of each correction and its location in-line or in the margins. Corrections are printed in a smaller font to imitate their appearance, unless they were copied in a hand equal to the main text. While erasures with unknown text are represented in double brackets “⟦⟧”, corrections that have visible undertext are printed in single brackets above the corrected text, often followed by “vid” (cf. p.671A, Ps 75:8). To avoid unnecessary transcriptions, the remaining letters of a line are often represented with an ellipsis. 26 We may also classify this error as a nonsense reading. 27 The omission of του ορουϲ is also found in Rahlfs 19, 247, and 376. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 157 Finally, a cancelation stroke was used to delete the και-compendium at 1 Kingdoms 26:21 (p.347B). The omission of the second και in verse 21 is found in A(02), Rahlfs M, and N (cf. SMR sa 2007/Biblia Coptica sa 25, Aeth, and L).28 Swete identified this correction as Ba, since it is clear that the reinker (Bb) chose to reinforce the και-compendium, rather than to leave it untouched. However, it appears that the cancelation stroke was also reinforced, or possibly originated with the reinker (Figure 67). FIGURE 67: CANCELATION STROKE IN 1 KGDMS 26:21 (P. 347B) Omission In contrast to the few corrections of additions, there are seventeen corrections of omissions in 1 Kingdoms. Thirteen of these corrections occur in scribe A’s half of the text, and the remaining four in scribe B’s. If the early corrections say anything about the scribes, then scribe B’s slight tendency to add text directly contrasts scribe A’s omissions. TABLE 24: CORRECTIONS OF OMISSIONS IN 1 KINGDOMS SCRIBE B(03) CORRECTOR 1 Kgdms 1:1; p. 309B A … ⟦ ⟧ υ̅ιυ̅̅ [θο vid κε … Β2 or BA 1 Kgdms 2:20; p. 311C A … εχρηϲαϲ [τω κ̅ω̅ … Β2 or BA 1 Kgdms 3:4; p. 312C A … εκαλεϲεν [κ̅ϲ ̅ ϲαµουηλ … Β2 or BA 1 Kgdms 3:6; p. 312C A … \ / καθευδε… ΒΑ 1 Kgdms 3:19; p. 313Β A … απο [παντω ̅ Β2 or BA 1 Kgdms 5:7; p. 315A A θ̅υ̅ \ισ̅̅λ̅/ µεθ ηµων… Β2 or BA 1 Kgdms 6:2; p. 315C A … ποιηϲωµεν \τη/ κιβω τω… Β2 or BA 1 Kgdms 6:15; p. 316B A … τα χρυϲα \α/ και … ΒΑ 1 Kgdms 12:12; p.323C A … βαϲιλευϲει εφ ηµω ̅ και νυν… [ κ̅ϲ ̅ο θϲ̅ ̅ηµων βαϲιλευϲ ηµων Β2 or BA 1 Kgdms 14:30; p. 327A A …µει ζων \ην/ η πληγη… Β2 or BA 28 For the Lucianic text, see Taylor, Lucianic Manuscripts. CHAPTER 5 158 1 Kgdms 14:39; p. 327C A … οτι ζη [κ̅ϲ ̅ ο ϲωϲαϲ … Β2 or BA 1 Kgdms 14:40; p. 327C A … πα ̅ τι \ανδρι/ ιϲ̅λ̅̅ … BA 1 Kgdms 18:23; p. 334A A και ειπεν δαυειδ \ει/ κου φον … B2 or B18 1 Kgdms 20:38; p. 337C B ταϲ ϲχιζαϲ ˙/. προϲ τον κυριον ˙/. ηνεγκεν ταϲ εχιζαϲ ΒΒ 1 Kgdms 25:6; p. 343C B … υγιαινων \ / ο οικοϲ Β2 or BB 1 Kgdms 25:36; p. 345C B … και ιδου αυτω πο τοϲ… [τοϲ εν οικω αυ του ωϲ πο Β2 or BB 1 Kgdms 30:30; p. 352C B … βηρϲαβεε[και Β2 or BB Fourteen of the seventeen corrections are of single word omissions, and the three longer omissions (1 Kgdms 12:12, 20:38, 25:36) are all clearly the result of homoeoteleuton. The first of the corrections involves the second of three abbreviations of υιοϲ in B(03) (Chapter 3).29 The previous υιοϲ nomen sacrum also occurs in a correction (p. 141A, Num 2:29) and is one of the column-end corrections that we attributed to scribe A. Because this abbreviation is rare in B(03) and we have found a previous example of scribe A abbreviating υιοϲ in a correction, we might identify this correction as prima manus. The undertext of the erasure is no longer visible, but the faint outline of a theta and omicron (from θοκε) might be present. After the letters were erased and the nomen sacrum was added, scribe A (or a later corrector) recopied the theta and omicron in a small hand. Two of the corrections of omission involve the addition of κ̅ϲ ̅at line-ends (p. 312C, 1 Kgdms 3:4; p. 327C, 1 Kgdms 14:39), both of which are in scribe A’s text. Neither Fabiani nor Swete (B-M following) detected these line-end corrections. Instead, they found evidence of erasure in the preceding word of the first correction, εκαλεϲεν. 30 There appears to be some erosion of the parchment beneath the epsilon and kappa, but this is far from clear in the images. It is Versace, who first identified the nomen sacrum, following εκαλεϲεν, as a correction and the analogous correction at 1 Kingdoms 14:39.31 In both, the first vertical stroke of κ̅ϲ ̅is copied along the right bounding line or 29 It is not entirely clear that this is a correction. Its identification goes back to Fabiani’s commentary, who states that the nomen sacrum was copied super rasura. Swete and Brooke-Mclean also note the omission of υιου and its subsequent correction (super rasura), but they do not identify this with a particular hand (B?). Fabiani, Prolegomena, 45. 30 εκαλεϲ est sup. ras. a B2. Fabiani, Prolegomena, 46. 31 While Versace regularly mentions erasures that may be connected to the marginal corrections, he makes no comment concerning εκαλεϲεν at 1 Kingdoms 3:4. Versace, Marginalia, 129. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 159 to the right of it. Notably, there was no attempt on the part of scribe A or the corrector to compress the two letters.32 FIGURE 68: POSSIBLE OMISSIONS OF κ̅ϲ ̅AT A LINE-END (PP. 312C; P. 327C) Two other corrections in 1 Kingdoms involve the omission of the article; both of which are in the work of scribe A (p. 311C, 1 Kgdms 2:20; p. 315C, 1 Kgdms 6:2). The first of these is a line-end correction that is likely from the hand of scribe A.33 The corrector extended the line into the margin by adding a regular tau and a raised, compressed omega. A similar line-end can be found at the bottom of the same column (p. 311C l. 41), where the same article was originally copied; again, with the omega raised and compressed. The second correction (p. 315C, 1 Kgdms 6:2) was made by the supralinear addition of the article, τη. It is difficult to know whether this was made a prima manu or by a diorthotes. In the first column of the same page, the nomen sacrum ιϲ̅λ̅̅ is added above line forty- three (p. 315A, 1 Kgdms 5:7). This correction is most likely from the hand of scribe A, because of the scribe’s preference for abbreviating ιϲραηλ and the shape of the lambda (cf. the nomina sacra on p. 315B ll. 32, 39).34 Therefore, the proximity and form of the two supralinear corrections may suggest a shared origin in the hand of scribe A.35 Likewise, there are six more supralinear corrections of omission in 1 Kingdoms. Only one of these is found in the work of scribe B (p. 343C, 1 Kgdms 25:6). On p. 343C, the omission of και is corrected by adding a raised και-compendium above the line. Fortunately, there is an analogous correction in the work of scribe A (p. 312C, 1 Kgdms 3:6) with which we can compare hands (Figure 69). The primary difference between the two και-compendia is the size. The correction in scribe B (left image) is the largest, with the vertical stroke extending down from the base of the previous line to the top of the nu. This is striking since it seems to obstruct the downward stroke of the rho in the 32 The compression of terminal letters at line-ends is common in the main copying of all three scribes. 33 Versace does not detect this line-end correction. Fabiani and Swete identify this correction with their B2 and Ba correctors, respectively. Fabiani, Prolegomena, 45; cf. Versace, Marginalia, 129. 34 The phrase [του] θ̅υ̅ ιϲ̅λ̅̅ occurs five times on p. 315, and once as θ̅υ̅ κ̅υ̅ ιϲ̅λ̅̅. This is the only time ιϲ̅λ̅̅ was omitted and corrected. 35 Compare the line-end correction on p. 163C l. 38. CHAPTER 5 160 previous line. In contrast, the correction in scribe A’s work (right image) is centered between the two lines, and only the downward oblique extends into the line below. There is nothing distinctive about the και-compendia that allows us to confidently assign them to one of the scribes, but the noticeable difference between the same correction in the work of both scribes might point to their origin with each scribe. FIGURE 69: SUPRALINEAR ΚΑΙ CORRECTIONS (PP. 343C, 312C) The omission on p. 316B (1 Kgdms 6:15) of the relative pronoun was corrected by adding an alpha above the line. The single-letter omission was likely caused by homoeoteleuton, because the previous word χρυϲα ends in the same letter.36 While there is no trace of the underlying ink and it is possible that this correction originated with the reinker, the shape of the alpha corresponds well with those produced by scribe A (Figure 70). FIGURE 70: SUPRALINEAR ADDITION OF ALPHA IN SCRIBE A (P. 316B);37 SUPRALINEAR ADDITION OF ΑΝΔΡΙ (P. 327C) The supralinear addition of ανδρι at 1 Kingdoms 14:40 (p. 327C) also appears to have been made by scribe A.38 It is possible that the original error was the result of homoeoteleuton. The correction was clearly made before the breathing marks were added because the smooth breathing of ιϲ̅λ̅̅ was forced to the left of the iota (Figure 70). However, the shape of the alpha and delta also betray the hand of scribe A. 36 This corrected reading is only found in B(03) and 121, a member of the B-text group. 37 The white-balance has been adjusted to help show the clear traces of original ink in the surrounding text. 38 The omission is only found in B(03). The phrase παντι ανδρι ιϲραηλ is present in A(02), Rahlfs 19–108, 107–82, 29, 120, 127, 130, 134, 314, 509, the Armenian, Coptic, and Old Latin Versions. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 161 The three multiword omissions are all explained by homoeoteleuton: one was the fault of scribe A (p. 323C, 1 Kgdms 12:12) and two were by scribe B (p. 337C, 1 Kgdms 20:38; p. 245C, 1 Kgdms 25:36). While the line-end correction at 1 Kingdoms 12:12 contains a lambda that might betray the hand of scribe A, it is likely that a different hand was responsible for this correction (B2). From the line-end corrections that were clearly the work of scribe A, there is a clear preference for the “Coptic” mu. 39 This tendency aligns with the paleographic evidence from Scribe A’s ancorae-lemniskoi corrections (mentioned above). In contrast, the use of the Biblical Majuscule mu fits better with those line-end corrections that might have been made by a diorthotes or scribe C (cf. p. 219C l. 3). FIGURE 71: CORRECTIONS OF MULTI-WORD OMISSIONS IN 1 KINGDOMS (PP. 323C, 337C, 345C) It is far more difficult to identify the hand for either of the corrections in the work of scribe B (p. 337C, 1 Kgdms 20:38; p, 245C, 1 Kgdms 25:36). The correction at 1 Kingdoms 25:36 looks like it may have been made a prima manu, but there are hardly any distinguishing features. Indeed, at least one line-end correction in the work of scribe B appears to have been made by scribe C (See 5.2.2). Nevertheless, both corrections are likely from scribe B or B2. Substitution There are eleven corrections of substitutions in 1 Kingdoms: five in the work of scribe A and six in scribe B. However, only six of these can confidently be placed in the fourth century. Three of the corrections involve proper names, one is a change of pronouns, and five are substitutions of verbs. 39 See the corrections on pp. 89C l. 34; 158A l. 35; 181A l. 17; 199A l.14; the exception is p. 270B l. 43. CHAPTER 5 162 TABLE 25: CORRECTIONS OF SUBSTITUTIONS IN 1 KINGDOMS SCRIBE B(03) CORRECTOR 1 Kgdms 3:2; p. 312C A … εκαθ\ευδεν/η̸τ̸ο̸… B18 1 Kgdms 6:21;40 p. 317Α A … απεϲτρ\α/οφ\η/α ϲιν… Β18 1 Kgdms 10:19; p. 321C A ˙/. Κατα … βαϲι λεα ˙/.ϲτηϲειϲ … Β3 1 Kgdms 11:13; p. 322C A s ϲαουλ … s ϲαµουηλ… B3 1 Kgdms 14:27; p. 327A A … ϲ̸κηπ̸̇⟦⟧ριον… BA 1 Kgdms 19:22; p. 335B B ειπ\α/ε̸ν… B18 1 Kgdms 20:14; p. 336B B … κα\ν/ι̸ µεν B2 or BB 1 Kgdms 22:15; p. 340A B s µου s αυτου… Β3 1 Kgdms 23:7; p. 340C B [ϲαουλ]vid … ο δαυειδ… B2 or B18 1 Kgdms 23:13; p. 341A Β διαϲεϲω\ϲ/ται… B18 1 Kgdms 24:3; p. 342B B … s εδδαιεµ… s τηϲ θη ραϲ των ελαφων B3 Versace is right to identify the three S-siglum corrections, with their undulated form, in 1 Kingdoms as coming from the hand B3 (Figure 17).41 In 1 Kingdoms 11:13, the original reading ϲαµουηλ is found in V, 56, 82, 509 and the Armenian, while the corrected reading ϲαουλ is found in A(02), M, 120a?, 121b (along with other minuscules), the Coptic, Ethiopic, Old Latin, and Lucianic. The variation is easily explained by the repetition of the two names in the passage, and the fact that ϲαµουηλ contains all five letters in the name ϲαουλ.42 FIGURE 72: B3 S-SIGLUM CORRECTIONS IN 1 KINGDOMS (PP. 322C, 340A, 342B) It is also likely that the substitution απεϲτροφαϲιν → απεϲτραφηϲαν (p. 317A, 1 Kgdms 6:21) originated with B3 or an even later corrector. The supralinear letters alpha and eta are noticeably smaller than 40 Compare the alpha and eta in the B18 correction αδελφη on p. 765Β l. 13. 41 Versace, Marginalia, 137. 42 Interestingly, the reinker chose to correct the reading in the column by skipping over the letters mu and eta. Versace, Marginalia, 137 n. 315. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 163 the usual corrections of the scribes or B2. While the alpha has been obscured by fading and the reinking, it is similar to the one found in the marginal correction ϲαουλ (Figure 72). The marginal addition of κατα to the verb ϲτηϲειϲ was also added by B3 (p. 321C, 1 Kgdms 10:19).43 There may be an erasure beneath the eta,44 but this appears to be an imperfection in the parchment, which affected lines 11–13 of column C. Here, the corrector used a lemniskos, rather than the S-siglum because the substitution is made by adding on to the original text, instead of replacing it. The original reading agrees with A(02), while the corrected reading follows that of M, V, and the Lucianic text. Β18 was most likely responsible for the correction ειπεν → ειπαν at 1 Kingdoms 19:22. The original epsilon of the verb ending received a cancelation stroke, and a rounded, supralinear alpha was added. This rounded alpha may be covering an earlier correction, but there is no clear evidence for this. The corrected reading ειπαν is also found A(02), 29, 55, 121, 509, and others. Two corrections of substitution occur on p. 304. The first is the correction µου → αυτου (p. 340A, 1 Kgdms 22:15). The latter reading is well attested,45 while only the Ethiopic, a member of the B-text group, agrees with the original text of B(03). The second substitution (p. 340C, 1 Kgdms 23:7) is made with a combination of correction methods (Figure 73). FIGURE 73: CORRECTION OF A SUBSTITUTION IN 1 KINGDOMS 23:7 (P. 340C) While the undertext of this correction is not entirely visible, it almost certainly reads the name ϲαουλ. Either the diorthotes (B2) or B18 corrected this erroneous reading by covering over the original letters, deleting the omicron with cancelation stroke, and adding the necessary letters above the line. It is unlikely that the article was in the corrector’s exemplar. Rather, instead of deleting the lunar sigma from the name ϲαουλ, the corrector chose to complete the circle and make an omicron. Similarly, the substitution of ϲκηπτρον → κηριον (p. 327A, 1 Kgdms 14:27) was made by copying over the original text. In this case, scribe A likely erased the letters tau and rho, and then 43 Versace, Marginalia, 135; cf. 117. 44 This is given by Swete and B–M, but not by Fabiani. 45 A(02), N, and (A, Coptic, Old Latin) give the reading µου, while 509 has the reading ϲου. CHAPTER 5 164 recopied a rho followed by an iota; this way, the final omicron and nu could be left alone. The result is a small space between the deleted pi and the re-written rho (Figure 74). While a deletion dot is visible above the pi, it is unclear whether one was present above the sigma. It is possible that this correction was made in scribendo, but the re-spacing of the word suggests it was not made until after the line was finished. FIGURE 74: CORRECTION OF A SUBSTITUTION IN 1 KINGDOMS 14:27 (P. 327A) Finally, the correction on p. 342B (1 Kgdms 24:3) involves the location of Saul’s men in pursuit of David. The original reading, εδδαιεµ, provides a proper name for the cliff (προϲωπον). The many variations on this name appear to be related to the Hebrew י עלימ (mountain goats or ibexes).46 The substitution, which was added by B3, agrees with 509 (B-text group) and the Lucianic text in the translation τηϲ θηραϲ των ελαφων (the trap of deer).47 Orthography By far, the largest number of corrections mark orthographic variation.48 Apart from clear instances of corrections by the reinkers—where ink is left untouched—there are eighty-eight orthographic corrections in 1 Kingdoms (see Appendix D). Attention to the spelling in B(03) did not end when the text was fully copied, since the reinker was involved in systematic and often pedantic revisions.49 Apart from corrections of ε → ο, ι → ει, κ → χ, and ν → µ, the work of scribe B contains more orthographic corrections than that of scribe A. 50 Because of the large number of orthographic corrections, we are unable to discuss each case and will focus on examples from each type of interchange. 46 εδδαιεµ] ϲαδδαιεµ 121; ϲαδεµ N 707 56 243 119 29; αειµειν A(02)*; αειαµειν A(02)c. 47 Eusebius’ Onomasticon gives the spelling αιαλιµ and cites Aquila, who reads των ελαφινων, and Theodotion, reading των πετρων των ελαφων. 48 On the relationship between non-standard orthography and linguistic register in the Greek papyri, see Stolk, “Orthographic variation,” 299–326. 49 For example, the root word λειτουργεω is routinely spelt -λιτ- by all three scribes. All 156 occurrences of the shorter spelling were systematically corrected with the supralinear epsilon by the reinker. 50 This is largely, though not entirely, dependent on the frequency of certain words in each half of 1 Kingdoms. The lemma εξολεθρευω is spelled -ολοθ- by both scribes, but occurs six times in 1 Kingdoms 1:1–19:11 and only once after. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 165 The correction α → αι only occurs once in 1 Kingdoms, in the work of scribe B (p. 344B, 1 Kgdms 25:16). The supralinear addition of the iota is clearly older than the sixteenth-century reinforcement, as the reinker copied an iota to the right of the earlier correction (Figure 75). Interestingly, the verb ποιµαίνω is regularly spelled ποιµεν- and then corrected to ποιµαιν-. This is the only example of the original -αι- spelling, where the scribe omitted the iota.51 It is possible that the scribe noticed this error, but it remains likely that B18 was responsible for the correction. FIGURE 75: Α-ΑΙ CORRECTION IN 1 KINGDOMS 25:16 (P. 344B) There are five corrections of the interchange between αι and ε.52 The four corrections of αι → ε (1 Kgdms 15:9; 19:17; 20:5; 23:23) likely originated with B18, although the correction γνωτ\ε/α̸ι̸ (p. 341C; 1 Kgdms 23:23) contains cancelation strokes, which could indicate an earlier correction (Figure 76). The only concurrence of πεδιω in 1 Kingdoms was originally copied as παιδιω, but later corrected with the interchange of αι → ε (1 Kgdms 20:5). Regarding the spelling of πεδιω, scribe A copied the epsilon in all eleven examples in Genesis 46:28–1 Kingdoms 19:11, while scribe B uses both spellings through Psalm 77:71 (7x each). Finally, the correction εντεταλµε → εντεταλµαι (p. 338A, 1 Kgdms 21:2) is mostly obscured by the sixteenth-century reinking, but may well go back to the scribe B or B2. 51 Cf. Psalm 48:15 (p. 654A), where the original reading ποιµαινει is corrected to ποιµανει by deletion dot. 52 According to Gignac, this interchange is the second most frequent, next to that of ει-ι. However, the Trismegistos Irregularities database indicates that the interchange ο-ω is the second most frequent, with 10888 examples, while αι-ε interchange occurs a total of 4945 times. Gignac, A Grammar, 1:192; https://www.trismegistos.org/textirregularities/texirr_type_list.php [consulted 18/2/2021]. CHAPTER 5 166 1 Kingdoms 15:9; 19:17; 20:5; 23:2353 B18 Marginal Corrections54 FIGURE 76: ΑΙ → Ε ORTHOGRAPHIC CORRECTIONS BY B18 Thackeray has already noted the regularity of the spelling (εξ)ολεθρευω, -ευµα, -ευϲιϲ, in B(03) and the regular correction with the interchange ε → ο.55 These corrections reflect a later Byzantine preference for the omicron spelling.56 The original preference for the spelling -ολεθρευω, and the nearly complete correcting of the form in the whole codex suggests that the corrections are the work of B18.57 Support for this can be found with the line-end correction at Jeremiah 2:30 (p. 1066A), where the letters ολε are added to the left of line 27 (ολε[θρευων); like the other occurrences of this word, this correction was later re-corrected to ολ\ο/εθρευων. Τhe interchange between ι and ει is the most frequent orthographic variation in B(03). Joanne Vera Stolk has helpfully shown that ι for ει happens most frequently in the fourth century, although the opposite interchange is less frequent.58 It is often noted that B(03) contains a peculiar preference for ει to indicate the long /i/.59 While there is a certain level of consistency in this spelling, it is possible to find variation between the scribes and even within the work of the same scribe. This diversity and commonality in spelling can be viewed through the corrections in 1 Kingdoms. 53 Left to right: pp. 329A, 335A, 336A, 341C. 54 Left to right: pp. 744B, 867Β, 426A. 55 Thackeray notes that there are only twenty-two instances of -ολοθρ– out of 250 examples in the OT of B(03). The one example of εξολεθρευω in the NT is likewise corrected (Acts 3:23; p. 1386B). The exchange of ε → ο is attributed to the “assimilation of the vowels flanking the liquid.” Thackeray, A Grammar, 87–88; cf. Blass, New Testament Greek, 21. 56 See LBG, s.v. ἐξολόθρευϲιϲ, and LSJ, s.v. ἐξολεθρ-εύω. 57 Versace, Marginalia, 46. 58 Nearly 30% of the texts from this century were corrected by modern editors ι → ει. Stolk, “Itacism,” 691–692. 59 Tischendorf, Vaticanum, xxii; von Soden, Die Schriften, 2:909; Thackeray, A Grammar, 86; Ropes, Beginnings, xxxviii–xxxix; Martini, Problema, 112; BDF §23; For recent discussions, see Williams, “Semitic Long /i/,” 15–26; Jongkind, “Redactional Elements,” 241–43. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 167 While all three scribes consistently spell the adjective µικροϲ with -ει-, 60 subsequent correctors deleted the epsilons with supralinear dots. These deletion dots are missing twice in 1 Kingdoms but were left untouched by the reinker (p. 311C, 1 Kgdms 2:19; p. 352A, 1 Kgdms 30:19). Likewise, the majority of ει → ι corrections were made by B18, who did not reinforce the epsilons, and are therefore not included in Appendix D. However, the correction ε̇ι̸δε (p. 338B, 1 Kgdms 21:8) does appear to be an early correction by the diorthotes. Here, a deletion dot and cancelation stroke are used in combination, sharing a comparably light and faded ink. There is a stark contrast between scribes A and B in the number of ι → ει corrections, since B tends to give the longer spelling.61 For example, scribe A consistently spells the dative δυναµει without the epsilon, which later appears as a supralinear correction (e.g. 1 Kgdms 2:10).62 Although δυναµει does not appear in the second half of 1 Kingdoms, the -ει- spelling is consistent throughout scribe B’s contribution to the codex (e.g., 25/25 times in the NT).63 The same pattern occurs with the datives φρονηϲει (1 Kgdms 2:10; cf. 3 Kgdms 10:26) and οραϲει (1 Kgdms 16:12; cf. Ezek 8:3). Subsequent corrections to the short spelling were made by the reinker (compare the epsilons in Figure 77). 1 Kingdoms 2:10 bis; 16:12 B18 Marginal Corrections FIGURE 77: ORTHOGRAPHIC CORRECTIONS OF ι → ει BY B18 (LEFT: PP. 311A BIS, 331A; RIGHT: P. 545Α) The correction of πν\ε/ιγει (p. 331A, 1 Kgdms 16:15), on the other hand, appears to be an early correction by either scribe A or the diorthotes. We can compare this correction to a later epsilon correction on the same page (Figure 78). Apart from the clear difference in ink color, the first supralinear epsilon maintains the unimodular form, while the second does not. It is also clear that 60 Exceptions can be found in scribe A (Josh 22:19), B (Ps 72:2), and C (Ps 103:25). 61 See, however, Milne and Skeat’s brief discussion of ιϲχυει and οριον in Scribes, 89. 62 Thackeray notes the spelling of δυναµι and ιϲχυει but not the consistent change after 1 Kingdoms 19:11. Thackeray, A Grammar, 86. 63 The exceptions to this are found in Pss 20:14; 53:8; 58:12; 67:29; 73:13; Zech 4:6; Ezek 28:5. CHAPTER 5 168 πν\ε/ιγει was an early correction, because scribe A gives the long spelling επνειγεν five lines above (1 Kgdms 16:14). FIGURE 78: BA OR B2 CORRECTION (P. 331A LL. 28–29); CF. B18 CORRECTION (P. 331B LL. 24–24) The reinker copied supralinear a chi to correct αβειµελεχ → αχειµελεχ (p. 346B, 1 Kgdms 26:6), µοκλων → µοχλων (p. 341Α, 1 Kgdms 23:7), and cases of unaspirated ουκ. From examining these late corrections (B18), we find that scribe A had a slightly higher tendency to leave the kappa unaspirated (4/11) than scribe B (1/9).64 On the other hand, there is one example of a correction in scribe B to the unaspirated ουκ (p. 341B, 1 Kingdoms 23:19).65 Although Fabiani associates this correction with B2, it likely originated with the reinker B18. This conclusion is complicated since both readings were later reinked (Figure 79). Regardless, it is clear that the reinker was responsible for correcting unaspirated kappas throughout the whole codex. FIGURE 79: CORRECTION OF ASPIRATED ΟΥΧ (P. 341B; 1 KGDMS 23:19) The corrections of the spelling ν → γ and ν → µ are all examples of the unassimilated nu with ϲυν- and εν-,66 except for the unassimilated nu in απεκτανκα → απεκταγκα (p. 342C, 1 Kgdms 24:12). Five of the eleven corrections occur in 1 Kingdoms 28 with the lemma ενγαϲτριµυθοϲ. Two examples illustrate that these corrections were made by the reinker. First, the assimilated εγγαϲτριµυθουϲ can be found once in Isaiah 19:3 (Figure 80). This reading was subsequently corrected by a supralinear nu and a cancelation stroke (probably by the diorthotes). However, B18, who normally adds a supralinear gamma, reinked the original gamma and left the supralinear nu untouched. That the early corrector 64 Aspirated ουχ occurs 20 times in 1 Kingdoms. The form ουκ occurs seventy-three times in 1 Kingdoms 1–19:11a and fifty-nine times in the rest of the book. 65 Versace, Marginalia, 46. 66 See Williams, “When Does ϲυν- Assimilate?,” 429–438. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 169 thought it was necessary to revert εγγαϲτριµυθουϲ back to the unassimilated nu, highlights the early preference of the scribes and diorthotes. The second example of a late correction towards the assimilated nu is on p. 132B ll. 28–29 (Lev 25:16). At the end of line 28, scribe A or the diorthotes made a line-end correction from την κτηϲιν → την ενκτηϲιν. Following this, B18 corrected the line-end correction with a supralinear gamma.67 These examples provide evidence that the majority, if not all, of the corrections to assimilated nu were made by the reinker, rather than the early correctors.68 FIGURE 80: ASSIMILATE NU IN ISA 19:3 (P. 1019C); NON-ASSIMILATED NU IN LINE-END CORRECTION (P. 132B; LEV 26:16) Corrections of the interchange ο → ω occur with the perfect tense of οραω (εορακ-; 7x) and once with αθοωθηϲεται (p. 346C, 1 Kgdms 26:9). It is frequently recognized that these supralinear corrections were made by the reinker or a later corrector.69 Although both omicron and omega forms were used by the scribes A and B,70 the consistent corrections of ο → ω throughout the codex fit the profile of the reinker, who also corrected the unassimilated nu, unaspirated kappa, and the unassimilated epsilon in -ολεθρευω.71 Nonsense There are possibly nine early corrections of nonsense readings in 1 Kingdoms and two which are most likely later additions. However, it will become clearer in Psalms and the NT that the majority of nonsense readings were corrected by the reinker. We should not be surprised, then, if those undecided corrections turn out to be late. Three of the corrections involve the pronoun αυτοϲ, three correct the article, and two correct a proper name. Six nonsense corrections occur in the work of scribe A, and the remaining five in that of scribe B. 67 Versace, Marginalia, 125 n. 191. 68 See also Fabiani, Prolegomena, xviii. 69 Fabiani alternates his identification between B2 and B3, while Swete more consistently attributes the corrections to Bb. In the NT, the same corrections are identified as B3 by Tischendorf (Luke 1:22) and C2 in the ECM (1 John 1:1). 70 For εορακ-, see 1 Kgdms 10:24; 22:9. For εωρακ-, see Exod 3:9; 2 Kgdms 13:34. 71 Versace, Marginalia, 46. CHAPTER 5 170 TABLE 26: NONSENSE CORRECTIONS IN 1 KINGDOMS SCRIBE B(03) CORRECTOR 1 Kgdms 9:25; p. 320B A … την̸̇ πολει… Β18 1 Kgdms 12:14; p. 323C A … πορευοµεν\οι/ω̸ν̸ Β2 or Β18 1 Kgdms 13:16; p. 325A A … ιω\ν/αθαν… Β2 or B18 1 Kgdms 14:37; p. 327C A αυτου\ϲ/… Β2 or Β18 1 Kgdms 15:22; p. 329C72 A … επα κρ\ο/αϲιϲ… B2 1 Kgdms 15:23; p. 329C A …τα̸̇ο ρηµα… Β2 or B18 1 Kgdms 18:20; p. 333C A τοι\ϲ/ οφθαλµοιϲ… Β2 or BA 1 Kgdms 23:14; p. 341B B αυχµωδε\ι/ϲ̸… B2 or BB 1 Kgdms 25:1; p. 343B B … αυτο\ν/υ B18 1 Kgdms 27:1; p. 347C B … δαυ\ειδ/˙/. … ˙/. ειδ B4 & B18 1 Kgdms 29:2; p. 350A B … ανδ/ρ\εϲ… BB 1 Kgdms 30:4; p. 351A B … αυτ\ου/ων… B2 or BB The three nonsense corrections of the article are the consequence of scribe A’s error (p. 320B, 1 Kgdms 9:25; p. 329C, 1 Kgdms 15:23; p. 330C, 1 Kgdms 18:20). Scribe B does not appear to make this mistake in 1 Kingdoms. Both the correction την → τη πολει and τα → το ρηµα were executed by using cancelation strokes and deletion dots (Figure 81). The second example is accompanied by the addition of an omicron. FIGURE 81: NONSENSE CORRECTIONS OF THE ARTICLE (PP. 320B, 329C) In the first correction, both the deletion dot and cancelation stroke appear to come from the reinker, and there is no remnant of the undertext. However, an earlier omicron is clearly visible below the sixteenth-century reinking in the second correction and yet the slightly irregular shape may still 72 See the omicron corrections by B18, nine and twelve lines above: εξωλ\ο/θρ– (1 Kgdms 15:20); εξολ\ο/θρ– (1 Kgdms 15:21). EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 171 betray the hand of B18. On the other hand, the article on p. 330C (1 Kgdms 18:20) was likely corrected from τοι → τοιϲ by an early corrector (Figure 82). FIGURE 82: NONSENSE CORRECTIONS (P. 330C; 1 KGDMS 18:20); NONSENSE CORRECTIONS (P. 327C; 1 KGDMS 14:37) One of the three pronoun corrections occurs in the work of scribe A (p. 327C, 1 Kgdms 14:37), and the remaining two in that of scribe B (p. 343B, 1 Kgdms 25:1; p. 351A, 1 Kgdms 30:4). As there is no trace of earlier ink, the supralinear sigma was likely added to αυτου\ϲ/ by one of the reinkers on p. 327C (Figure 82). The same can be said of the correction on p. 343B (Figure 83). Here, the supralinear nu resembles those of B18 with an undulated oblique stroke. In contrast, the third pronoun (p. 351A, 1 Kgdms 30:4) is corrected αυτων → αυτου with a supralinear omicron and upsilon, which clearly resemble the early hand of scribe B or the diorthotes (Figure 84).73 FIGURE 83: NONSENSE CORRECTION (P. 343B; 1 KGDMS 25:1); CF. B18 CORRECTION (P. 426Α; 3 ΚGDMS 14:26) FIGURE 84: NONSENSE CORRECTION (P. 351A; 1 KGDMS 30:4); CF. B18 CORRECTION (P. 765Β; CANT 4:10) Likewise, the correction of ανδεϲ → ανδρεϲ on p. 350A (1 Kgdms 29:2) appears to be from an early hand, likely from scribe B (Figure 85). Rather than adding a supralinear rho, scribe B fit the vertical stroke between the delta and epsilon. This correction may have been made in scribendo. FIGURE 85: NONSENSE CORRECTION (P. 350A; 1 KGDMS 29:2); CF. SCRIBE B UNREINFORCED RHO (P. 342; 1 KGDMS 23:26) 73 The reinker avoids the compressed omicron. CHAPTER 5 172 Finally, we find the correction of δαυ → δαυειδ (p. 347C, 1 Kgdms 27:1), which was corrected twice— first by Versace’s B4 (9th CE) in the margin, and then supralinearly by the reinker.74 These correctors not only attempted to emulate the majuscule hand of the original scribes, but also maintained the long -ει- spelling of the name (Figure 86). Although this is a late correction, it is interesting to note that this clear error was missed by both scribe B and the diorthotes. FIGURE 86: NONSENSE CORRECTION (P. 347C; 1 KGDMS 27:1) Text Division There is a single clear correction of text division, which was made in scribendo by scribe B (p. 352B, 1 Kgdms 30:25). In this correction, the scribe began copying the word και without leaving a space after the previous word. However, after copying the kappa scribe B erased the letter and rewrote και following a space of nearly two letters. This is a clear example of the intentionality with which the scribes of B(03) divided the text. FIGURE 87: CORRECTION OF TEXT DIVISION (P. 352B; 1 KGDMS 30:25) 5.1.3 Summary Through an examination of the early corrections in 1 Kingdoms, we have seen similarities and differences between the two scribes of the text. For instance, it is only in the work of scribe B that we find corrections of additions, while scribe A appears to be more likely to omit text. Only three corrections of omission exceed one word (cf. 1 Kgdms 12:12, 20:38, 25:36) and are clearly the result of homoeoteleuton. The corrector B3 appears to have corrected the text four times, all of which marked substitutions. The majority of corrections are the result of orthographic preferences. Nearly a quarter of the eighty-nine corrections involve the change ι → ει in the work of scribe A, while scribe B often gives the longer spelling.75 The reinker was clearly responsible for the most systematic 74 Versace, Marginalia, 203; cf. p. 362A, 2 Kgdms 6:1. 75 Many of the longer spellings in scribe B were eventually corrected to the shorter spelling by the reinker, but we did not examine these late corrections (unless there were deletion dots or cancelation strokes). EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 173 correction of orthography in the codex, with special attention given to the assimilation of nu or the aspirated kappa. While there were relatively few nonsense readings in 1 Kingdoms, a few surprising errors seemed to have evaded the attention of the scribes or diorthotes (δαυ →δαυειδ; 1 Kgdms 27:1). It is likely that the many apparent erasures in 1 Kingdoms were in scribendo corrections of nonsense readings by the scribes.76 5.2 Psalms We turn now to the second scribe change, which is found at Psalm 77:71 (pp. 674–675). Because there is limited space and much of the following discussion would be repetitive, we will only summarize the most important features of the corrections in Psalms, especially as it relates to our two scribes and the evident distinction between scribe C and what we have seen of scribe A. While it is important for the argument of a third scribe to distinguish scribe A and C through the early corrections, it is our aim to expand the profile of scribe B, before turning to the NT. 5.2.1 The Text of Psalms in B(03) Like 1 Kingdoms, we are still awaiting the publication of a new critical edition of the Greek Psalter in the Göttingen series.77 Therefore, Rahlfs’ critical edition, Psalmi cum Odis, remains our most important resource for studying the text of Psalms; Fabiani’s commentary and Swete’s edition still provide the most comprehensive data for studying the corrections in B(03). Even with the advancements made in Rahlfs’ study of the Greek Psalter,78 B(03) remains the foundational witness like earlier editions.79 Adapting Friedrich Baethgen’s two “recensions” of the Greek Psalter (O1 and O),80 Rahlfs compared manuscripts and daughter versions to B(03) (= O) on the one hand, and L, the Lucianic Recension (= O1), on the other.81 Based on this comparison, he further divided the 76 Possible early erasures in 1 Kingdoms include: p. 311A, 1 Kgdms 2:10; p. 314B, 1 Kgdms 4:18; p. 314C, 1 Kgdms 5:3; p. 317A, 1 Kgdms 7:5; p. 317B, 1 Kgdms 7:10 (Scribe A); p. 335A, 1 Kgdms 19:17; p. 338C, 1 Kgdms 21:9; p. 341B, 1 Kgdms 23:17 (2x); p. 347C, 1 Kgdms 27:1; p. 348A, 1 Kgdms 27:5 (Scribe B). 77 The newly established project to publish the Editio critica maior is not expected to finish until 2040. Albrecht, “Report,” 204. 78 Rahlfs, Der Text. 79 Bons and Brucker, “Psalmoi,” 305. 80 O1 being the majority text (based on the collations of Holmes-Parsons) and O being the reziperten Text, the Sixtine edition (1587). Baethgen preferred the O1 “recension” because O was closer to the MT and apparently corrupt. Baethgen, "Der textkritische Werth," 407–408. 81 On the two groups see Rahlfs, “Prolegomena,”, §1.3, 3, 7; Rahlfs, Der Text, 39–40. This two-part division has ancient precedence in Jerome’s letter to Sunnia and Fretela (Letter 106, 2.2). Jerome, here, outlines two editions CHAPTER 5 174 witnesses into six groups: three early text forms, two late, and one mixed.82 For our purposes, it is important to note that the original text of B(03) was grouped with א(01) and the Bohairic version, and so located in Lower Egypt. However, Rahlfs categorizes the corrector of B(03) (Bc in Rahlfs) in the Lucianic “recension.” This follows only for the corrections made by later hands and does not include corrections by the scribes or diorthotes.83 One last point of interest is made by Thackeray, concerning the orthography in the Greek Psalter. After examining the spelling in the Pentateuch of certain majuscules, including B(03), he concludes that there was a practice of dividing books “for clerical purposes, into two nearly equal portions.”84 Thackeray utilizes three pieces of evidence to suggest the two-part division of Psalms dates back centuries before B(03) and even to the time of translation (3rd/2nd BCE): (1) The change of spelling in nouns from -ια → -εια (e.g. δυναϲτεια); (2) the interchange of ε→αι in the first half (e.g. - ϲθε → -ϲθαι); and (3) the presence of the syllabic augment in ευφραινω (ηυφρ.) in the first half.85 Notably, Thackeray identifies the division of Greek Psalms at Psalm 77, the same location we have noticed the change from scribe B to C.86 The question remains whether Thackeray has identified an inherited division from the original translation or simply a change of scribes in B(03). In fact, he is forced to admit that the break in Psalms has been “somewhat obliterated” through transmission in א(01) and A(02), leaving B(03) as the only witness to the bi-section of Greek Psalms at Psalm 77. While it is likely that some changes in orthography were indeed influenced by the exemplars, we have already found evidence in 1 Kingdoms that variations in orthography align with paleographic and paratextual changes in scribes. Thus, Thackeray’s arguments provide additional support for our identification of a scribe (editionem), the koiné/Lucianic and that present in the Hexaplaric codices, which is closer to the Hebrew. Ceulemans, “Antiochene Text,” 149–50; Kreuzer, “Jerome,” 78–80; Pietersma, “The Present State,” 15. 82 For a fuller discussion of Rahlfs’ method of grouping, see Boyd-Taylor et. al., “Manuscript Affiliation,” 98– 124; cf. Bons and Brucker, “Psalmoi,” 305–306. 83 Rahlfs, Psalmi, 7. 84 Thackeray, A Grammar, 65; This argument was made originally in two articles—“The Greek Translators of Jeremiah” and “The Greek Translators of Ezekiel”— and subsequently applied to Exodus, Leviticus, and Psalms in “The Bisection,” 88–98. 85 Thackeray, “The Bisection,” 91–92. 86 According to Thackeray, this also aligns with the Masoretic division at Psalm 78 (MT). Thackeray, “The Bisection,” 93. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 175 change at Psalm 77:71 (pp. 674–675). We will, therefore, proceed in our examination of early corrections across this division of scribes. 5.2.2 The Earliest Corrections in Psalms Addition Like 1 Kingdoms, the early corrections of addition are only present in the work of scribe B.87 Apart from the correction on p. 671A, the additions are all of a single word—one article, three conjunctions, and one noun. Unlike the multiword additions in 1 Kingdoms, the addition at Psalm 75:4 does not result in dittography. According to Rahlfs, this reading is only found in B(03) (cf. Pss 74:11, 75:4a). TABLE 27: CORRECTIONS OF ADDITIONS IN PSALMS SCRIBE B(03) CORRECTOR Psalm 35:7; p. 645A B κ̇α̇ι ̇η δικαιοϲυνη… B2 or B18 Psalm 36:34; p. 646B B … τη̇̇ν̇ γην B18 Psalm 41:12; p. 650A B … η̇ ψυχη… B18 Psalm 61:3; p. 661A B … ο̸̇ θεοϲ… B2 or B18 Psalm 65:5; p. 663A B δευτε τε̇κ̇̇ν̇α̇ και ιδετε… B2 or B18 Psalm 73:8; p. 669Β B δευτε κ̇α̇ι ̇καταπαυϲωµεν… B2 or B18 Psalm 74:9; p. 670B B κ̇α̇ι ̇πιονται παντεϲ… B2 or B18 Psalm 75:4; p. 671Α B ‘εκει ϲυνκλαϲει τα κερατα’ B2 or BB Psalm 76:15; p. 671Β B ϲυ ει ο θ̅ϲ ̅η̇µ̇ω̇ν̇ ο ποιων… B18 It is unlikely that all of these corrections actually originated with Scribe B or the diorthotes. As noted already, the attribution of deletion dots to a single hand is difficult. There are at least three corrections which appear to come from the reinker (p. 646B, Ps 36:34; p. 650A, Ps 41:12; p. 671B, Ps 76:15). The deletion dots of both corrections appear darker and finer than the others, and there is no evidence of lighter ink beneath (Figure 88). 87 The following table does not include the late correction of the reinker on p. 710B (Ps 144:6; Scribe C). I have also omitted five erasures of additions, because they were clearly made after the text was reinked (p. 626B, Ps 5:6; p. 627A, Ps 6:11; p. 638A, Ps 24:7; p. 638A, Ps 25:12; p. 664A, Ps 67:12). CHAPTER 5 176 FIGURE 88: B18 DELETION DOTS IN PSALMS; BB OR B2 DELETION DOTS Finally, the correction at Psalm 61:3 (p. 661A) involves the use of the deletion dot and cancelation stroke. Since this is the only correction of addition with a combination of methods, it is possible that it originated from an early hand and was recorrected by a later hand. Omission Unlike 1 Kingdoms, we find significantly more corrections of omission in scribe B’s portion of the Psalter.88 While there are fewer in scribe C, this is largely due to the lacuna of Psalms 105:27–137:6b.89 The two column-end corrections in Psalms are unique to scribe C, while the third is likely a later addition, made after quire [37] (replaced by pp. 695–706) went missing.90 In 1 Kingdoms, there were only three multiword omissions. The Psalter, on the other hand, contains eighteen multiword omissions. Eight of the single word omissions involve pronouns, two omit θεοϲ (not including the omission of ο θεοϲ). A number of these omissions are clearly the result of homoeoteleuton (p. 626A, Ps 4:5; p. 635B, Ps 635B; p. 638A, Ps 24:3; p. 662A, Ps 63:3; p. 707A, Ps 138:12; p. 713A , Ps 148:14). TABLE 28: CORRECTIONS OF OMISSION IN PSALMS SCRIBE B(03) CORRECTOR Psalm 4:5; p. 626A B …εν \ταιϲ/ καρδια\αιϲ υµων/ επ ταιϲ κοιταιϲ υµων… B2 or BB Psalm 20:5; p. 635Β B … ειϲ αιωνα \ ειϲ αιωνα του/ αιωνοϲ B2 or BB Psalm 24:3; p. 638Α B αιϲχυνθητωϲαν \⟦παντεϲ⟧vid/ οι ανοµουντεϲ B2 or BB B18 or 37 erased Psalm 24:9; B ↑ … πραειϲ εν κριϲει ˙/. BB 88 The following table does not include the four corrections, which are clearly the work of B18 (Ps 43:8, p. 650B; Ps 54:6, p. 657A; Ps 55:14, p. 658A; Ps 69:2, p. 666B). Likewise, Swete identified two line-end corrections that are not clearly corrections (p. 649B, Ps 40:7; p. 688B, Ps 101:3). Versace does not include either in his list. Versace, Marginalia, 260–61. 89 Scribe B’s half of Psalms contains 0.52 corrections of omission per page (26/50 pp.), while scribe C’s has 0.48 corrections per page (13/27 pp.). 90 Versace only identifies the column-end correction on p. 688A. While we have noted that column-end corrections are mainly in scribe A, these examples differ from the more substantial omissions of scribe A. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 177 p. 638Α παϲαι αι οδοι… ↓ διδαξει πραειϲ οδουϲ αυτου ˙/. Psalm 60:2; p. 660Β B ειϲακουϲον \ο θϲ̅/̅ τηϲ δεηϲεωϲ µου B2 or BB Psalm 63:3; p. 662Α B … ψυχη µ̅ου ↑ οιτινεϲ ηκονηϲαν ωϲ ροµφαιαν ˙/. ↓ εϲκεπαϲαϲ µε απο ϲυϲτροφηϲ πονηρευοµενων απο πληθουϲ εργαζοµενων αδικιαν ˙/. ΒC Psalm 64:6; p. 662B B … τηϲ γηϲ ετοιµαζων… [ εν θαλαϲϲη µακραν BC Psalm 66:4; p. 663B B εξοµολογηϲαϲθωϲαν \ϲοι/ λαοι ο θ̅ϲ ̅ εξοµολογηϲαϲθωϲαν \ϲοι/ λαοι παντεϲ B2 or BB Psalm 67:4; p. 664A B … ευφρανθητωϲαν [διαψαλµα B2 or BB Psalm 67:29; p. 665A B εντειλαι \ο/ θ̅ϲ ̅… B2 or BB Psalm 68:14; p. 665B B επακουϲον \µου/ εν αληθεια … B2 or B3 Psalm 68:18; p. 666A B ↑ µη αποϲτρεψηϲ το προϲωπον ϲου απο του παιδοϲ ϲου ↓ ϲου επιβλεψον επ εµε ˙/. οτι θλειβοµαι.. BB Psalm 70:12; p. 667A B ο θ̅ϲ ̅\µου/ µη µακρυνηϲ απ εµου B18 Psalm 70:22; p. 667B B και γαρ \εγω/ εξοµολογηϲοµαι … B3 or Β18 Psalm 72:12; p. 669Α B ˙/. ουτοι ιδου ˙/. αµαρτωλοι… B3 Psalm 72:16; p. 669Α B τουτο κοποϲ ˙/. εν αντιον µου ˙/. εϲτιν B2 or BB Psalm 72:18; p. 669Α B ↑ πλην δια ταϲ δολιοτηταϲ εθου αυτοιϲ ˙/. πωϲ εγενοντο… ↓ κατεβαλεϲ αυτουϲ εν τω επαρθηναι ˙/. B3 Psalm 73:4; p. 669Β B ↑ εν µεϲω τηϲ εορτηϲ ϲου ˙/. ωϲ ειϲ την… ↓ εθεντο τα ϲηµεια αυτων ϲηµεια και ουκ εγνωϲαν ˙/. B3 Psalm 73:12; p. 670A B ο δε \θϲ̅/̅ βαϲιλευϲ ηµων… B2 or BB Psalm 73:14; p. 670A B ↑ δρακοντων επι του υδατοϲ ˙/. εδωκαϲ αυτον… ↓ ϲυ ϲυνετριψαϲ ταϲ κεφαλαϲ του δρακοντοϲ ˙/. BB Psalm 74:6; p. 670Β B … κατα του \θυ̅̅/ αδικιαν B3 or B18 Psalm 75:8; p. 671A B [απο τηϲ οργηϲ ϲου vid]* απο τη̸\οτ/ε \η/ οργη ϲου Β18 Psalm 76:5; p. 671B B … φυλακαϲ ˙/. οι ˙/. παντεϲ εχθροι µου B3 Psalm 76:19; p. 672A B ↑ φωνη τηϲ βροντηϲ ϲου εν τω τροχω ˙/. εϲαλευθη … B3 CHAPTER 5 178 ↓ εφαναν αι αϲτραπαι ϲου τη οικουµενη ˙/. Psalm 77:65; p. 674B B ωϲ δυνατοϲ \ / κεκραιπαληκωϲ Β18 Psalm 84:4; p. 678Α C ˙/. παϲαν κατεπαυϲαϲ ˙/. την οργην ϲου B3 Psalm 88:40; p. 682A C l. 42 εβεβηλωϲαϲ ειϲ την γην το αγιαϲµα l. 43 αυτου ΒC Psalm 98:8; p. 688A C ο θ̅ϲ ̅\ϲυ/ ευειλατοϲ εγεινου αυτοιϲ Β18 Psalm 100:3; p. 688Α C l. 42 ου προεθεµην προ οφωαλµων µου l. 43 πραγµα παρανοµον Β2 Psalm 103:26; p. 691Β C ζωα µικρα µετα µεγαλων δρακων ουτοϲ… [εκει πλοια διαπορευ ονται Β2 or BC Psalm 104:27; p. 693A C εθετο \εν/ αυτοιϲ τουϲ λογουϲ… B3 or B18 Psalm 105:26; p. 694B C l. 42 του καταβαλειν αυτουϲ εν τη ερη l. 43 µω Β18 Psalm 138:12; p. 707A C ↑ και νυξ φωτιϲµοϲ εν τη τρυφη ˙/. ωϲ το ϲκοτοϲ … ↓ οτι το ϲκοτοϲ ου ϲκοτιϲθηϲεται απο ϲου και νυξ ωϲ ηµερα φωτιϲθηϲεται ˙/. Β2 or BC Psalm 139:12; p. 708Α C ↑ εν ταλαιπωριαιϲ ου µη υποϲτωϲι ̅ ˙/. ανδρα αδικον … ↓ ανηρ γλωϲϲωδηϲ ου κατευθυνθηϲεται επι τηϲ γηϲ Β2 or BC Psalm 142:3; p. 709B C ϲου παϲ ζων εταπ\ε/ινωϲεν ειϲ την… [οτι κατεδιωξεν ο εχθροϲ την ψυχην µου Β2 or BC Psalm 148:14; p. 713A C και υψωϲει κεραϲ λαου αυτου τοιϲ υιοιϲ ιϲραηλ … [υµνοϲ παϲι τοιϲ οϲιοιϲ αυτου BC The most significant feature of the corrections of omission in the Psalter is the presence of scribe C’s hand in the work of scribe B. Like the lemniskos correction at 1 Esdras 5:9, which we assigned to scribe C, the line-end correction at Psalm 64:6 contains a curved και-compendium and similar alphas (Figure 89). On the same page, we find an ancora-lemniskos correction (p. 662A, Ps 63:3), which is likely from the same hand.91 While B3 was only responsible for correcting substitutions in 1 Kingdoms, the same scribe corrected a few omissions in the Psalter, using lemniskoi, ancorae- lemniskoi, and possibly supralinear additions. 91 The apex of the alphas have the exaggerated cross-section that is commonly seen in scribe C’s hand. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 179 FIGURE 89: SCRIBE C CORRECTION IN THE WORK OF SCRIBE B (P.662B L. 7); CF. P.581C A number of corrections are likely from the hand of B18, although it is difficult to assign them with confidence. The supralinear addition of µου in Psalm 70:12 (p. 667A) appears to be a late addition when compared to the parallel correction two pages before (Ps 68:14; Figure 90). Notable is the change in mou and the full-sized omicron, rather than the compressed form.92 FIGURE 90: EARLY ΜΟΥ CORRECTION (PS 68:14, P. 665B); LATE ΜΟΥ CORRECTION (PS 70:12; P. 667A) The supralinear correction at Psalm 75:8 (p. 671A) must be from B18, since the final sigma of the original reading is left untouched, without deletion dot or cancelation stroke (τηϲ οργηϲ → τοτε η οργη). A darker cancelation stroke is used to substitute the first eta for an omicron and tau (τη̸ϲ → τοτε) and is similar to the color of a late addition in the next line (Figure 91). FIGURE 91: TWO B18 CORRECTIONS (P. 671A, PS 75:8–9) The addition of a και-compendium in Psalm 77:65 (p. 674B) likely originated from B18. A parallel for this correction can be seen on the third line of the ancora-lemniskos correction on p. 622C (2 Esd 23:5). It seems likely that that the later orthographic correction (ε → αι) in the following word was made by the same hand. Finally, the supralinear correction ϲυ (p. 688A, Ps 98:8) appears to be from a later hand, based on the narrow sigma and slanted upsilon. 92 For a parallel B18 ancora-lemniskos correction, see the addition of τουϲ οφθαλµουϲ µου απο δακρυων in the lower margin of p. 658A (Ps 55:14). CHAPTER 5 180 FIGURE 92: B18 CORRECTIONS OF OMISSION (PP. 674B, 688A; PSS 77:65, 98:8) Substitution Substitutions constitute a large portion of the corrections in Psalms.93 Including the substitutions that may in fact be from B18, we find fifty-two corrections in the work of scribe B and twelve in that of scribe C. With the lacuna in mind, there remains a noticeable contrast between the two scribes. Additionally, all four of the S-siglum corrections (B3) are found in Psalms 1–77 and each corrects a substitution. Fifteen corrections indicate changes of pronouns—from article to pronoun and pronoun to different pronoun or different case. The majority of the remaining substitutions involve the change of case, tense, or preposition in compound verbs. A significant number of substitutions appear to have come from the reinker, though it is not always clear. TABLE 29: CORRECTIONS OF SUBSTITUTION IN PSALMS LXX Psalms SCRIBE B(03) CORRECTOR Psalm 7:5; p. 627A B [αποπεϲοινvid] αποπεϲοιµ/ι\ αρα απο των… B2 or B18 Psalm 16:1; p. 631B Β ενωτιϲαι τη\ν/ϲ προϲευχη\ν/ϲ µου B2 or B18 Psalm 16:14; p. 632A Β s απο ολιγων κ̅ε ̅s απολυων γηϲ Β3 Psalm 17:17; p. 633A B προϲελαβε\το µε/ εξ υδατων… B2 or BB Psalm 21:9; p. 636A B ηλπιϲ\ε/αν επι κ̅ν̅ … B2 or BB Psalm 30:7; p. 641A B εµι\ϲ/µηϲαϲ τουϲ φυλαϲϲονταϲ… B18 Psalm 32:15; p. 642Β B ϲυνι\ων/ειϲ παντα τα εργα… B2 or B18 Psalm 38:6; p. 648Α B ιδου παλαι\ /αϲ εθου… B18 Psalm 38:7; p. 648A B πλην µατην ταραϲϲ\ε/ονται B18 Psalm 38:8; p. 648A B … υποϲταϲιϲ µου παρα ϲο\ι/υ εϲτι ̅ B18 Psalm 39:5; p. 648B B και ουκ ε\π/νεβλεψεν ειϲ µαταιοτηταϲ B18 Psalm 39:11; p. 648B B … την δικαιοϲυνην \ϲ/µου B2 or BB Psalm 66:4; p. 649A B ουκ ηδυναϲθη\ν/ του βλεπειν B2 or B18 93 The chart below does not include a number of corrections, which are clearly the work of B18 (Ps 47:3, p. 653A; Ps 56:7, p. 658B; Ps 96:1, p. 686B; Ps 97:6, p. 687B; Ps 101:12, p. 689A; Ps 102:17, p. 690B; Ps 137:8, p. 707A; ). EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 181 Psalm 44·14; p. 652A B … εϲ\ωθε/εβων B18 Psalm 45:6; p. 652B B … ο θ̅ϲ ̅τω \το προϲ πρωι· πρωι·/ προϲωπω B18 Psalm 47:13; p. 653Β B … και περι\λ/βα\β/λετε αυτη ̅ B18 Psalm 48:4; p. 653Β B … τηϲ καρδιαϲ µου ϲυνε\ϲειϲ/ϲι ̅ B3 or B18 Psalm 51:9; p. 656Α B και \εν/εδυναµωθη επι τη… B18 Psalm 52:7; p. 656Β B εν τω \ε/απ\ι/οϲτρεψαι κ̅ν̅… B18 Psalm 53:2; p. 656B B τουϲ ζειφαιουϲ και ειπε\ι/ν τω ϲαουλ… B18 Psalm 54:11; p. 657Α B επι τα τειχη ϲ̸α̇υτηϲ B2 or BB and Β18 Psalm 57:2; p. 658Β B ευθεια\ϲ/ κρινεται οι υιοι… B2 or B18 Psalm 57:6; p. 659Α B [φαρµακου τε]* φαρµακουτ\αι/ε φαρµακευοµενου B18 Psalm 57:7; p. 659Α B ο θ̅ϲ ̅ϲυνε̸̇τριψ\ι/ε̸ν τουϲ οδονταϲ αυτω ̅ B2 or BB and Β18 Psalm 61:13; p. 661Β B … του θ̅υ̅ και ϲ/ο\υ κ̅ε ̅το ελεοϲ B2 or B18 Psalm 62:2; p. 661B B εδιψηϲεν ϲ\ε/οι η ψυχη µου B18 Psalm 62:7; p. 661B B … εν τ\οιϲ/ω̸ ορθρ\οιϲ/ω̸ B2 or BB Psalm 63:9; p. 662A B και εξ\ηϲ/ο̸υ̸θενηϲαν αυτον Β18 Psalm 63:9; p. 662A B και εξουθενηϲαν \επ/ αυτο\υϲ/ν̸ B18 Psalm 65:15; p. 663Α B ποιη\ανοι/ϲω ϲοι βοαϲ… Β18 Psalm 67:9; p. 664Α B … του θυ του\το/ ϲινα B2 or B18 Psalm 67:19; p. 664Β B αν\ε/αβ\η/αϲ ειϲ υψοϲ ηχµαλωτευϲαϲ B2 or B18 Psalm 67:19; p. 664B B … δοµατα εν ανθρωπ\οιϲ/ω Β18 Psalm 67:22; p. 664B B εν πληµµελιαιϲ αυτ\ων/ο̸υ̸ B2 or B18 Psalm 68:4; p. 665A B s ελπιζειν [εγγιζειν]vid s ελπιζειν επι τον θ̅ν̅ µου B3 Psalm 68:7; p. 665B B µη αιϲχυνθειηϲαν επ εµ\ε/οι οι υποµε νοντεϲ… Β18 Psalm 68:7; p. 665B B µη εντραπιηϲαν επ εµ\ε/οι οι ζητουν τεϲ… B18 Psalm 68:10; p. 665B B … του οικου ϲου κατ\ε/αφα γε[ται] µε B18 Psalm 68:16; p. 665B B [ϲυνεχετω]vid µη δε ϲυ⟦⟧ϲχετω επ εµε φρεαρ… B18 Psalm 70:15; p. 667B B οτι ουκ εγνων s πραγµατιαϲ s γραµµατιαϲ B3 CHAPTER 5 182 Psalm 71:4; p. 668A B και ϲωϲειϲ ̇τουϲ υιουϲ… B2 or Β18 Psalm 72:10; p. 669Α B δια του\το/ επιϲτρεψει ο λαοϲ… B2 or ΒB Psalm 73:3; p. 669Β B οϲα\ϲ/ επονηρευϲατο ο εχθροϲ… B2 or Β18 Psalm 73:14; p. 670A B οτι επληρωθηϲαν οι εϲκοτ\ιϲ/ωµε νοι τηϲ γηϲ… Β18 Psalm 74:11; p. 670Β B … ϲυν\θ/κλαϲω Β18 Psalm 75:8; p. 671Α B [απο τηϲ οργηϲ ϲου] vid απο τη̸\οτ/ε \η/ οργη ϲου Β18 Psalm 75:10; p. 671Α B s τηϲ γηϲ s τη καρδια δ̅ιαψαλµα̅ B3 Psalm 76:7; p. 671B B … και [εϲκαλεν]vid … και εϲκαλον … και εϲκαλ\α/ον … και εϲκαλ\ε/ον * B2 or BB B3 B18 Psalm 77:9; p. 672B B … τοξο\ιϲ/ν̸ B2 or Β18 Psalm 77:10; p. 672B B ουκ εφυλαξαντο̇ ̇την διαθηκη …̅ B2 or B18 Psalm 77:26; p. 673A B και \α/επηρεν νοτον… Β2 Psalm 77:57; p. 674A B και \α/ε̸πεϲτρεψαν και ηϲυνθετηϲαν B2 or BB Psalm 83:11; p. 678B C οτι κρειϲϲω\ν/ ηµερα µια… Β18 Psalm 85:17; p.679Β C … οι µειϲουντεϲ \ϲ/µε και Β2 or BC Psalm 88:21; p. 681B C εν ελ\αι/ε\ω/ε̸ι̸ αγιω εχρειϲα… Β2 or BC and B18 Psalm 90:2; p. 683B C ερει τω \κ/θ̸̅ω̅ αντιληµπτωρ… Β2 or BC Psalm 91:6; p. 684A C ϲφοδρα εβα\θ/ρυνθηϲαν οι… Β18 Psalm 93:19; p. 685Β C … ϲου \ευφραναν/ ηγαπηϲαν την ψυχην… Β18 Psalm 97:1; p. 687Α C εϲωϲεν αυτ\ον/ω η δεξια αυτου Β2 or Β18 Psalm 97:4; p. 687Β C αλαλαξατε τω \κ/θ̸̅ω̅ παϲα η γη Β2 or B18 Psalm 101:20; p. 689Α C οτι εξεκυψ\ε/⟦α⟧ν εξυψουϲ αγιου… Β2 or B18 Psalm 138:16; p. 707B C το ακατεργαϲτον µου ειδοϲα̇ν̇ οι Β2 or BC Psalm 140:7; p. 708B C … τα οϲτα \αυτ/ηµων πα ρα τον αδην Β2 or BC Psalm 140:8; p. 708B C επι ϲ\οι/ε ηλπιϲα… B18 The four S-siglum corrections in Psalms (p. 632A, Ps 16:14; p. 665A, Ps 68:4; p. 667B, Ps 70:15; p. 671A, Ps 75:10) mark changes towards the Lucianic text, three of which also agree with the first hand of EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 183 א(01). These are clearly the work of B3, and the reinker follows the correction in three of them (cf. p. 632A, Ps 16:14).94 The graphic similarities between the first three S-siglum substitutions are most obvious in Psalm 68:4 (p. 665A), where the reinker changed the letters ΓΓ to ΛΠ (εγγιζειν → ελπιζειν) instead of reinking the marginal correction (Figure 93).95 FIGURE 93: B3 CORRECTION OF SUBSTITUTION; ΕΓΓΙΖΕΙΝ → ΕΛΠΙΖΕΙΝ (P. 665A; PS 68:4) Two noteworthy, early corrections occur on p. 671B (Ps 76:7) and p. 681B (Ps 88:21). The first concerns the aorist form of ϲκελλω. It is striking that the word was apparently corrected three times: twice in the fourth century (εϲκαλεν vid → εϲκαλον → εϲκαλαν) and then again by the reinker (εϲκαλαν → εϲκαλεν; Figure 94). It is unclear who made the first correction, but the second reflects the hand of B3. Until this point, it has been difficult to attribute any supralinear corrections to B3 with confidence. This example provides support for the presence of other supralinear corrections by the same hand. FIGURE 94: CORRECTIONS OF SUBSTITUTION IN PSALMS (P. 671B, PS 76:7; P. 681B, PS 88:21) The second substitution (p. 681B; Ps 88:21) was corrected twice, ελεει → ελεω → ελαιω (Figure 94). The initial correction involves cancelation strokes over the final epsilon and iota, along with the supralinear addition of an omega. Likewise, the second correction is supralinear, but is lacking a cancelation stroke over the epsilon. This absence of the cancelation stroke, the orthographic nature of the correction (cf. p. 684B, Ps 91:11), and the irregular shape of the supralinear alpha support the secondary nature of this part of the correction, likely from the hand of B18. The reading of B* is also found in Rahlfs 1219 and may be a harmonization to Isaiah 54:8 (cf. 1 Clem 18:1). However, the visual similarity between ελεει and ελεω (with ε-αι interchange) is probably responsible for the substitution. 94 Versace, Marginalia, 137 n. 321. 95 See also the correction on p. 632A, Ps 16:14; p. 667B, Ps 70:15; p. 671A, Ps 75:10. CHAPTER 5 184 Finally, we mention two late corrections by B18 (p. 648A, Ps 38:6; p. 653B, Ps 47:13). The first involves a substitution of παλαιαϲ → παλαιϲταϲ through the use of a supralinear stigma ligature ( ; Figure 99). While this may have been added by scribe B, the diorthotes, or B3, we have not found any early examples of this ligature in the text or early, marginal corrections. Neither have I found any B18 corrections with the stigma, but there are a number of other ligatures that suggest a level of comfort with using them, even in the majuscule hand.96 The substitution of περιβαλετε → περιλαβετε (p. 653B, Ps 47:13) is clearly the result of visual confusion between the lambda and beta and the semantic overlap between the two verbs.97 This substitution could be classified as B3 from the shape of the lambda, but the size of the letters and the shape of the beta suggest the corrector was B18 (Figure 95). Even if scribe B misread the exemplar, the similarities of the substitution could easily have gone unnoticed by the early correctors. However, the reinker is much more likely to have spotted the difference while tracing each letter. FIGURE 95: CORRECTIONS OF SUBSTITUTION IN PSALMS (P. 648A, PS 38:6; P. 653B, PS 47:13) Orthography Again, it is no surprise that spelling variation is the cause for the majority of corrections in Psalms. Apart from those corrections, which are made by leaving letters untouched (B18), there are 331 changes of orthography in the work of scribes B and C (see Appendix D). The standard method of correction is the supralinear addition of letters and, occasionally, the use of deletion dots and cancelation strokes. However, unlike in 1 Kingdoms, B18 appears to have made numerous 96 See the two different abbreviations of γραπτεον in the lower margin of p. 821A (Wis 12:5) and the upper margin of p. 826A (Wis 15:12). For a later example (Versace’s B27), see the stigma in the famous marginal note in Hebrews 1:3 (p. 1512B)–αµαθεϲτατε. 97 Not only are they used in different manuscripts to translate נקף in the verse, they are also used to translate תבק —περιβαλλω in Job 24:8 and Lam 4:5; e.g., περιλαµβανω in Gen 48:10 and Eccl 4:5. EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 185 orthographic corrections by writing over the original text.98 Except for one, all of these overwritten corrections are of the interchange ε-ο, with the root εξολοθρ-/εξωλοθρ-.99 In contrast to those made by B18, there is one overwritten, orthographic correction that was made in scribendo by scribe B (p. 642B, Ps 32:1). The correction from πρεπι → πρεπει was clearly made as the scribe was copying, since the following letter is also an iota (Figure 96). If the correction was not made in scribendo, then the original reading would have to be a nonsense reading: πρεπιι αινεϲιϲ. Although possible, it is more likely that the scribe stopped after incorrectly writing an iota and, instead of erasing it, covered part of the vertical stroke with the curve of the epsilon; then scribe B proceeded to copy the correct iota. FIGURE 96: ORTHOGRAPHIC CORRECTION IN PSALMS (P. 642B; PS 32:1) Space does not permit a full examination of the orthographic corrections and, similar to 1 Kingdoms, it is often difficult to attribute them to an early or late hand. However, we have already made the case that the majority of corrections of the unassimilated nu, the unaspirated kappa, and the ε → ο interchange of (εξ)ολεθρευω were made by B18. Likewise, the corrections of τεϲϲερεϲ → τεϲϲαρεϲ are the work of the reinker. 100 This correction only appears once in Psalms (p. 686A, Ps 94:10: τεϲϲαρακοντα), but is found consistently throughout the codex. See, for example, the marginal addition of τεϲϲερακοντα by B3 on p. 1232A, which was subsequently corrected to τεϲϲ\α/ερακοντα.101 Most importantly, several interchanges confirm our conclusion that there are two scribes who copied the Psalter. The corrections αι → ε, ι → ει, ν → γ, and ρ → ρρ all reflect a difference in spelling preference between scribes B and C. While minor differences can be explained by changes in word occurrence or the lacuna in Psalms 105:27–137:6b, the difference between 147 and 26 corrections of ι → ει indicates scribe B’s preference for writing certain words with only the iota. 98 While the dark ink may indicate that this was the hand of Versace’s B37, the over-writing is not in the minuscule hand associated with the late reinker. 99 The exception can be seen in the overwritten correction εγλεκτουϲ → εκλεκτουϲ (p. 693B; Ps 104:43). Compare with the substitution εγγιζειν → ελπιζειν above (Figure 93). 100 Fabiani, Prolegomena, xix; cf. Pisano, “The Text,” 87; Versace, Marginalia, 46. 101 An argument could be made that this correction was actually made by B3, since the color of the ink is similar to the original. However, it is unlikely that the scribe miscopied the spelling of this single word correction. CHAPTER 5 186 However, this conclusion is complicated by the fact that the work of scribe B only received three corrections of ι → ει in 1 Kingdoms, while scribe A had a higher preference for copying only the iota. This divergence in scribe B’s spelling is partly explained by the high repetition of words in Psalms 1–77:71, such as ταπ\ε/ινοω (corrected 15x), δυναϲτ\ε/ια (7x), δυναµ\ε/ι (5x).102 In contrast, the closeness of scribe A’s ι → ει corrections (22x) with scribe C’s (26x) does not confirm Milne and Skeat’s hypothesis that scribe A was responsible for copying Psalms. While the spelling of both scribes was likely influenced by the exemplars, the orthographic corrections corroborate the scribe change at Psalm 77:71 (pp. 674–675), but provide little help in identifying the scribes. Nonsense Very few of the nonsense corrections can be assigned to early hands with confidence.103 Therefore, we are dependent on the corrections of the reinker for our examination. From the table below, the contrast between scribe B’s twenty-four nonsense readings and scribe C’s three is clearly visible. Only six corrections involve more than a single letter, and the majority of errors can be clearly attributed to visual confusion. TABLE 30: NONSENSE CORRECTIONS IN PSALMS LXX Psalms SCRIBE B(03) CORRECTOR Psalm 7:3; p. 627A B µη \ο/ντοϲ λυτρουµενου… Β18 Psalm 13:2; p. 630B Β τουϲ υιου/ϲ\ των ανθρωπων B2 or Β18 Psalm 15:4; p. 631B Β … των ονοµα\των/ αυτω ̅ B2 or Β18 Psalm 16:15; p. 632Β B τω προϲωπ\ω/ου σου B18 Psalm 17:30; p. 633B B [ϲυ] απο εν σοι ρυϲθηϲοµαι… B18 Psalm 17:51; p. 634Β B εωϲ αι\ω/νοϲ B18 Psalm 21:10; p. 636A B η ελπιϲ µου απο µαϲτρ̸̇ων τηϲ Β18 Psalm 21:22; p. 636B B ϲωϲον µε \ε/κ ϲτοµατοϲ λεοντοϲ B18 Psalm 24:11; p. 638A B κα\ι/ ιλαϲη τη αµαρτια… B18 Psalm 26:14; p. 639Β B ανδριζου και κραταιου\ϲ/θω καρδι|α B2 or B18 Psalm 30:13; B επ\ε/ληϲθην ωϲει νεκροϲ… B2 or BB 102 Except for a single occurrence of ταπεινωϲηϲ (1 Kgdms 26:9) these words do not appear in 1 Kingdoms 19:11b– 31:13. 103 Possible early erasures in Psalms include: p. 627B, Ps 7:13; p. 627B, Ps 7:15; p. 634A, Ps 17:44; p. 650A, Ps 41:10; p. 655B, Ps 50:9; p. 667B, Ps 71:1 (Scribe B); p. 691A, Ps 103:11; p. 707B, Ps 138:13; p. 710B, Ps 144:4 (Scribe C). EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 187 p. 641A Psalm 30:15; p. 641B B εγω δε επι ϲ\οι/α ηλπιϲα κ̅ε ̅ B18 Psalm 32:12; p. 642Β B µακαριο\ν/ϲ̸ το εθνοϲ… B2 or B18 Psalm 37:16; p. 647Β B ϲ\υ/οι ειϲακουϲη κ̅ε ̅ο θ̅ϲ ̅µου Β18 or B37 Psalm 37:21; p. 647Β Β ενδιεβαλλον µε επ\ε/ι κατεδιωκον Β18 Psalm 49:9; p. 654Β Β … εκ των ποιµνιων ϲου χιµ\α/ρρουϲ B2 or BB Psalm 58:1; p. 659Α B τω δαυειδ ειϲ /ϲ\τηλογραφιαν Β18 Psalm 59:1; p. 660Α B … ετι ειϲ /ϲ\τηλογραφι|αν Β18 Psalm 61:9; p. 661Α B … ενωπιον αυ\του/ ταϲ καρδιαϲ υµω ̅ B18 Psalm 64:3; p. 662Α B … παϲα ϲαρ/ξ\ ηξει B2 or BB Psalm 74:2; p. 670Β B εξοµολογηϲοµεθα και επικα\λε/ϲοµε|θα B18 Psalm 75:9; p. 671Α B εκ του ουρανου ηκο\υ/ντιϲαϲ κριϲιν Β18 Psalm 77:5; p. 672Α B ον ενετειλα\το/ τοιϲ πατραϲιν ηµων B2 or B18 Psalm 79:12; p. 676A C εξετεινεν τα ϲ̸\̇κ/ληµατα αυτηϲ Β2 or B18 Psalm 96:3; p. 687Α C και φλογιει κυκλω του/ϲ\ εχθρουϲ αυτου Β18 Psalm 144:13; p. 711Α C τηϲ βαϲιλιαϲ ϲου η βαϲιλεια ϲου βα\ϲι/λεια παντων Β18 A number of nonsense readings arose from adjacent words ending and beginning with the same letter(s) (p. 636B, Ps 21:22; p. 638A, Ps 24:11; p. 659A, Ps 58:1; p. 660A; Ps 59:1; p. 672A, Ps 77:5; cf. p. 662A, Ps 64:3).104 Scribal leaps forward are the likely cause of five nonsense readings (p. 632Β, Ps 16:15; p. 641B, Ps 30:15; p. 642Β, Ps 32:12; p. 661Α, Ps 61:9; p. 687Α, Ps 96:3). While many of the corrections appear to have originated with B18, one of the nonsense corrections was possibly made in scribendo by scribe B (p. 654B, Ps 49:9). Rather than adding the missing alpha supralinearly, the corrector, whether scribe B or the diorthotes, compressed it between the mu and rho of χιµαρρουϲ (Figure 97). It is unlikely to have been added by the reinker since the placement of the alpha avoids obstructing the first rho, which B18 rejected. 104 We could also include those examples where the letters are graphically similar, like ε and ϲ (p. 669B, Ps 73:3; p. 687A, Ps 96:3). CHAPTER 5 188 FIGURE 97: CORRECTION OF NONSENSE READING (P. 654B; PS 49:9) The last of scribe C’s nonsense readings is surprising since the word βαϲιλεια was copied three times on two consecutive lines. One would expect that the repetition of the word would ensure its proper spelling. However, the opposite effect may have occurred as the scribe was less careful copying the same word for the third time in a row. This correction like the majority of nonsense corrections was likely made by B18. Not only is there a striking difference in occurrence between scribe B and C, but the late nature of these corrections suggests that the earliest correctors were not as attentive to nonsense readings. 5.2.3 Summary While we could certainly say more about the corrections in the Psalter of B(03), the preceding discussion has identified a variety of important patterns of the early scribes and correctors of the codex. The most significant observation is the striking difference between the two halves of the book. The identified variations confirm our earlier arguments for two scribes in the Psalter, who transitioned at Psalm 77:71 (pp. 674–675). While the orthographic variations also confirmed this change, the irregularity between scribe B’s work in 1 Kingdoms and Psalms problematizes our ability to rely on spelling alone to identify the scribes. All corrections of additions are found in the work of scribe B, along with the majority of substitutions and nonsense readings. However, if we account for the lacuna in Psalm 105:27–137:6b, the corrections of omission appear at roughly the same rate between the two scribes. 5.3 Conclusion In the previous chapter, we set out a typology for the corrections in B(03). This chapter examined the use of these correction methods and their implications for understanding our three identified scribes. While a comprehensive study of corrections would also include an investigation into the scribe change between Hosea–Tobit, this study allowed us to make observations about all three scribes and ensured greater consistency by looking at single works with two scribes. While using the correctors of B(03) may not be the most consistent way to examine copying habits–there are bound to be additional copying errors that have gone undetected–the study of the earliest corrections gives EARLY CORRECTIONS IN 1 KINGDOMS AND PSALMS 189 insight into the types of errors our three scribes were prone to make and the type of editing activity B(03) received during or shortly after production. FIGURE 98: SUMMARY OF CORRECTIONS AND CORRECTORS IN THE 1 KINGDOMS AND PSALMS Precision is difficult when trying to identify the majority of corrections, especially supralinear additions. However, we have found clear examples of corrections by our three scribes, some of which were made in scribendo. Those corrections that do not contain any distinctive features may also have been made by the scribes of the text or by a diorthotes (B2). More surprisingly, we have found a small number of corrections that scribe C appears to have made in the work of scribe B (p. 662B, Ps 64:6) and some ancorae-lemniskoi and column-end corrections in scribe A (e.g., p. 227, Deut 28:12; p. 232A, Ps 30:13; p. 233A, Deut 31:14). In 1 Kingdoms and Psalms, there were a total of seven S- siglum corrections. All of these were made by Versace’s B3 and mark substitutions towards the Lucianic text, suggesting a different exemplar. 105 It is especially difficult to find B3’s supralinear corrections, but the thrice corrected reading at Psalm 76:7 (p. 671B) supports the conclusion that B3 did make some corrections within the columns (cf. p. 693, Ps 104:27). Since our examination of the NT will be concerned with the copying of scribe B. The following observations on the scribe, in comparison to scribes A and C, will be a platform from which to begin the next chapter. We noted on multiple occasions scribe B’s tendency to add text, resulting most frequently in dittography. In one instance, the scribe or diorthotes corrected the dittography, but this resulted in the nonsense reading εκ µερουϲ του τουτου (p. 342A, 1 Kgdms 23:26). While there was a similar number of nonsense readings between scribes A and B in 1 Kingdoms (six and five respectively), the contrast was much higher between scribes B and C in the Psalter (twenty- 105 Contra Ropes, Beginnings, xcv, who states that there is no influence of the Lucianic text in Psalms of B(03). 3 5 33 9 3 3 9 2 1 5 2 3 2 4 1 1 1 22 2 2 9 1 5 13 24 5 3 44 2 1 3 5 15 1 4 10 1 2 5 3 3 9 2 3 22 10 23 3 45 9 4 8 BA B2 or BA B2 B3 B2 or B18 B18 BB B2 or BB B3 B2 or B18 B4 or B18 B18 BB B2 or BB BC B2 B2 or B3 B3 B2 or B18 B3 or B18 B18 BC B2 or BC B2 B3 B2 or B18 B3 or B18 B18 1 Kgdms 1:1–19:11a 1 Kgdms 19:11b–31:13 Psalms 1:1–77:71a Psalms 77:71b–150:6 Scribe A Scribe B Scribe B Scribe C Addition Omission Substitution Text Division Nonsense CHAPTER 5 190 four and three respectively). It is likely that additional nonsense readings were corrected in scribendo through erasures, but these are difficult to analyze without an autopsy of the codex or MSI. The orthographic corrections in 1 Kingdoms and Psalms have shown that there is likely some interference from the exemplars on the spelling of scribe B. However, the variation can also be attributed to the vocabulary and frequency of specific words in each book. Overall, scribe B has the most orthographic corrections, suggesting that the scribe diverted the most from later orthographic preferences. This chapter leaves us with a number of questions about scribal corrections in the OT of B(03). While the intention was to survey the editorial activity in 1 Kingdoms and Psalms, there is plenty of room for further refinement of the corrections and their textual traditions. However, this must wait until critical editions of the two books have been published. The number of corrections in each book (nearly 700 between the two) and the lack of critical editions ensure that we are only able to see the basic impressions left by the scribes and early correctors. Yet, for our purpose, the comparison of the three scribes of B(03) gives a more comprehensive and nuanced foundation for an examination of scribe B’s work in the NT. CHAPTER 6 Early Corrections in the New Testament While the books of 1 Kingdoms and Psalms were copied by two scribes in B(03), we have not found any evidence in support of multiple scribes in the NT. Therefore, our final chapter will analyze the early corrections in the text of the NT, which was copied entirely by scribe B. The conclusions drawn from Chapter 5 will allow for a more precise examination of the many well-known corrections in the NT text of B(03), such as Romans 5:1 and Ephesians 1:1. We will also be able to examine the early corrections in relation to work that has been done on harmonizations and even singular readings in the codex. This chapter will survey how critical editions of the NT have used and confused the early correctors of B(03). Finally, the examination of the entire NT will allow for some comparison of the corpora, such as the types and frequency of corrections in the Gospels as they compare to the Pauline corpus. The following analysis divides the NT of B(03) into four sections: The Gospels, Acts, the Catholic Epistles, and the Pauline corpus (including Hebrews).1 While this may seem natural to some and artificial to others, we have found codicological and paratextual evidence to support some of these sections. 2 The grouping of Acts and the Catholic Epistles (sometimes called the Praxapostolos) is not followed here on account of the distinct textual history of Acts. The presence of the Editio Critica Maior (ECM) for these books will allow us to compare corrections with additional data from more manuscripts. The value of the Coherence-Based Genealogical Method (CBGM) for identifying relatives of B(03) will also be explored in Acts and the Catholic Epistles. Before we look at each section in detail, it will be useful to make a few initial observations about corrections in the NT as a whole. In Chapter 4, we highlighted the relative paucity of marginal 1 Following Epp’s suggestion that “the classifications for all manuscripts should really be structured separately for various sections of the New Testament, particularly for the Gospels, for the Pauline letters, for Acts and the General Epistles, and for Revelation…” Epp, “Significance,” 371. 2 For example, the exclusive use in the NT of red section numbers for the Gospels and the continuous sections in the Pauline corpus. The opening of Romans aligns with a quire break, which was found to be a rare occurrence in B(03) (see Chapter 2). CHAPTER 6 192 corrections in the NT compared to the Greek OT.3 This observation can be caried through the NT, as the number of early corrections decreases after the Gospels. While our examination of 1 Kingdoms and Psalms was primarily in conversation with Fabiani’s commentary and Swete’s edition, there have been numerous attempts at dividing the corrections in the NT. Therefore, my own examination of corrections in this chapter will be compared against Tischendorf’s edition of B(03), the NTVMR transcription, the IGNTP transcription (John), and the ECM editions of Acts and the Catholic Epistles (cf. Table 15 in Chapter 4).4 6.1 The Gospels It should not be a surprise that we find over half of the NT corrections in the Gospels (roughly 600 of the 1100 corrections collected). When combined, the text of the four Gospels covers 148 of the 284 pages of the NT. However, we also found reason to expect that the frequency of corrections would fade after the Gospels. In Chapter 4, the dwindling use of marginal corrections, especially the S- siglum corrections, was cause for our hypothesis that the remaining forms of correction would also decrease through the NT. 6.1.1 The Text of the Gospels in B(03)5 Before turning to the earliest corrections in the Gospels, it would be useful to summarize the data provided by Text und Textwert (TuT) on the Gospel text of B(03), as it relates to that of Nestle-Aland and the Majority Text. From this broad overview, it is possible to investigate the possible contribution of the early corrections to the current developments in editing the NT text. Apart from the recent release of Mark (cf. n. 4), we do not yet have the ECM volumes for the four Gospels, leaving the TuT data as the primary contributor to our initial impression of the Gospel text.6 The following table presents the number of Teststellen where B(03) agrees with NA28 and the Majority 3 See also Stevens, History and Text, 69, who notes that there are “very few scribal corrections or marginal notes” in B(03), when compared to א(01); cf. Pisano’s observation that Matthew contains more marginal notations than the rest of the NT, in “The Text,” 89. 4 The ECM edition of Mark was not released in time for a full account. The IGNTP transcription of John and a full transcription of the NT is available on the NTVMR: . 5 For brief descriptions of the text of each NT book in B(03), see Pisano, “The Text,” 87–96. 6 In spite of the disparities in the number of Teststellen for each book, the following table consistently reveals the high agreement of B(03) with the Nestle-Aland text. We find even higher levels agreement between B(03) and the Ausgangstext for the CBGM data in Mark, Acts, and the Catholic Epistles. See also the impressive agreement between B(03) and the initial text of John 1:1–42; 17:1–26; 20:1–31 and Romans chapters 1, 8, and 15 (slightly less so), in Dormandy, “Pandects,” 353–354. EARLY CORRECTIONS IN THE NEW TESTAMENT 193 Text. The release of Mark Phase 3.5 in the online CBGM allows for further comparison between B(03) and the initial text (A) at 5407 passages. TABLE 31: TEXT UND TEXTWERT: B(03), NA28, AND THE MAJORITY TEXT OF THE GOSPELS;7 MARK IN THE CBGM B(03)–NA28 B(03)–Majority Text B(03)–A Matthew 47/64 (73.4%) 6/64 (9.4%) —— Mark 147/189 (77.8%) 44/189(23.3%) 5233/5407 (96.8%) Luke 46/54 (85.2%) 1/54 (1.9%) —— John 1–10 123/153 (80.4%) 53/153 (34.6%) —— When speaking about the text of the four Gospels in B(o3), it is inevitable that we discuss text- types/clusters, recensions, and editorial activity. Like the B-text in 1 Kingdoms, B(03) has often been described as the primary witness to a text, variously described as “neutral,” 8 “Alexandrian,” 9 “Hesychian/Egyptian,”10 “Eastern,”11 or the “B-text cluster”12 in the Gospels. Furthermore, we have already noted the critique of Latinization in the text by Erasmus, Mill, and Wettstein (see Chapter 4). To varying degrees, these descriptions imply some level of intentionality, whether one simply imagines a carefully copied manuscript, or a more complex editorial process. The idea that B(03) is the product of a fourth-century (or slightly earlier) Alexandrian recension has long permeated discussions of the codex. 13 While Westcott and Hort heavily 7 In the TuT volumes, these data are found in the columns “Lesarten 2 und 1/2” (left column above) and “Lesarten 1 und 1/2” (right column above). Aland, et al., eds., Die Synoptischen Evangelien; Aland, et al., eds, Das Johannesevangelium. 8 Westcott and Hort, Introduction, 126–30. 9 Griesbach classified the initial portion of Matthew as a Western text, while the rest of the Gospels were considered Alexandrian. Tregelles attributes this incorrect classification to the absence of a published transcription at the time of Griesbach’s work. However, B(03) was not present in Griesbach’s earlier editions, starting in 1771, since, as Martini rightly notes, the codex was “still under suspicion of having been heavily interpolated by the Latin tradition.” Griesbach, Novum Testamentum, lxxv–lxxvi; Tregelles, An Introduction, 164; Martini, “Alexandrian Text,” 287; cf. Epp, “Textual Clusters,” 315 n. 28. 10 Hug, Einleitung, 171–172; Bousset, “Hesychius,” 92; cf. Epp, “Textual Clusters,” 317–318; Metzger and Ehrman, The Text, 187. 11 Lachmann, “Rechenschaft,” 831; cf. Epp, “Textual Clusters,” 321; Epp., “Critical Editions, Part 2,” 519. 12 Earlier identified as the “β text” or “B” text group. Kenyon, The Text, 204; Epp, “Significance,” 362; Epp, “Textual Clusters,” 342. 13 Martini traces the tradition back to Griesbach and then Hug, who discusses a third-century “Egyptian recension… which had the authority of the Church in Alexandria and Egypt.” However, Hug’s description of the “Egyptian recension” ultimately becomes one of the Hesychian recension, inherited from Jerome’s discussion of the Septuagint (Preface to Chronicles, 54ff.). While Jerome also mentions Lucian and Hesychius in relation to the NT (Ep. ad Damasum), it is clear that Hug’s three NT text-types (Egypt, Syria, and Palestine) were appropriated from Jerome’s Septuagint recensions (Hesychius, Lucian, and Origen). Hug, Introduction, 1:190–198; Bousset, “Hesychius”; Martini, CHAPTER 6 194 influenced the perception of B(03) as a “neutral” text, untainted by editorial interference,14 Kenyon noted in 1940 that scholarly opinions had already shifting towards Bousset’s conclusions—B(03)’s text is revised and can be localized to Egypt.15 However, since the publication of Martini’s study on P75 and B(03) in Luke and Calvin L. Porter’s on John, it has generally been accepted that any recension, if there was one, must be dated to the end of the second century or start of the third.16 Yet again, a possible challenge to this consensus was made by Brent Nongbri, who suggests that P75 could indeed be dated to the same century as B(03). 17 While accepting Nongbri’s conclusions, Tommy Wasserman has attempted to supplement the P75-B(03) relationship with an examination of P4, which he believes to be confidently dated to the start of the third century.18 If the text of P4 is close enough to B(03), it can push the P75-B(03) text back again into the second century.19 However, Nongbri’s careful investigation into the discovery of P4 challenges even the early date given to this codex.20 Problema, 18–20; Martini, “Alexandrian Text,” 155; Metzger and Ehrman, The Text, 187; Kenyon, “Hesychius”; Metzger, “Lucian,” 191 n. 1. The attribution of B(03) to Hesychius’ recension can already be found in the writings of Ernest Grabe (1705). Like the critique of Latinization by Erasmus, Grabe’s classification of the OT in B(03) as Hesychian was pejorative. For Grabe, Rome relied too heavily upon B(03), and he instead highlighted the superiority of the OT text of A(02), which he was editing. Marcos, The Septuagint, 241; Jellicoe, The Septuagint, 177; Keene, “Grabe,” 663. 14 Hort does admit, however, “it must not of course be assumed to follow that B has remained unaffected by sporadic corruption.” Westcott and Hort, Introduction, 150. 15 This change of opinion was largely the result of recent papyrus discoveries like the Chester Beatty papyri in the 1930’s, which had revealed new connections between B(03) and Egypt, while also complicating the concept of localized texts. Kenyon highlights the Egyptian papyri, which have either “Western,” mixed, or even “Caesarean” characteristics. While far less influential, the criticisms of the “neutral” text by H. C. Hoskier appeared around this same time. Kenyon, “Hesychius,” 248; Hoskier, Codex B, 1:416; For a summary of the responses to Hort by Bousset, von Soden, Hoskier, and Lagrange, see Martini, Problema, 27–34; cf. Wasserman, “Alexandrian Recension?,” 5–6. 16 Martini, Problema, 149–152; Porter, “Papyrus Bodmer XV (P75),” 363–376; cf. Fee, “P75, P66, and Origen,” 248– 279; Fee, The Significance, 194–93. 17 Nongbri, “Papyrus Bodmer XIV–XV,” 405–437; Nongbri, God’s Library, 199–202. 18 Wasserman cites the dates given by NA28 (200–250 CE) and Orsini and Clarysse (175–200 CE). Wasserman, “Alexandrian Recension?,” 9; Orsini and Clarysse, “Manuscripts,” 470. 19 While the papyri are the best witness to an early B-text, they are not the only early comparative material. B(03) is also closely related to earlier patristic sources and versional evidence. Epp, “Textual Clusters,” 340; cf. Knust and Wasserman, To Cast, 187 n. 39. 20 Nongbri, God’s Library, 247–268. EARLY CORRECTIONS IN THE NEW TESTAMENT 195 Therefore, we are left with at least the possibility that some measure of third or fourth- century “recensional” activity took place before the production of B(03).21 This, indeed, is what Dirk Jongkind has recently suggested. 22 While the editorial background of the codex challenges the notion of a “neutral” text, it does not deal the blow intended by earlier critics of B(03). Rather, as Kenyon and Jongkind have argued, those who prepared the text copied by the scribes of B(03) did not create a new text type, “but produc[ed] a carefully prepared representation of a text already existing.”23 What we know of the text also involves an assessment of the scribes, and statements concerning the skill of scribe B abound.24 As these assessments relate to the text of the Gospels, the most significant conclusions relate to the scribe’s avoidance of harmonization. While Lagrange demonstrated the near absence of harmonization in B(03),25 Cambry Pardee has provided the most definitive discussion of the issue in the Synoptics. At the end of his chapter on B(03), Pardee concludes that only ninety variants in the Synoptics likely involved harmonized material, the majority of which are widespread in the manuscript tradition. However, he clarifies that “there are only seven singular and eight sub-singular readings attributable to harmonization” and therefore “it 21 I am hesitant to use the word “recensional” here, since its connotation has transformed from earlier usage. As noted already, the word had been used to describe an edition created by an identified individual, such as Hesychius, and even authorized by ecclesial authorities. It has been common in Septuagint scholarship to refer to recensions and revisions towards the Hebrew without clear distinctions (cf. kaige and Lucianic). Likewise, the language of text types and recensions has often been indistinguishable in NT textual scholarship. McLay, “Recension and Revision,” 293–303; Holmes, “Codex Bezae,” 123–160. 22 Jongkind highlights five redactional features in the NT text of B(03), which, when combined, suggest “we have in B03 a copy of a text that was carefully prepared and done so with linguistic interest and competence…” While we will examine some of these redactional elements in more detail, “the five features are (1) the change from καθωϲ to καθαπερ in the expression καθωϲ γεγραπται, (2) the spelling ιωανηϲ, (3) the order χριϲτοϲ Ιηϲουϲ, (4) the omission of the article in ο Ιηϲουϲ, and (5) the representation of long /i/ with -ει-.” Jongkind, “Redactional Elements” 231–245 (quotations from pp. 234, 244). 23 Kenyon, “Hesychian,” 249–250; cf. Jongkind, “Redactional Elements,” 244–245; cf. Porter, “Papyrus Bodmer XV (P75),” 376. 24 See the classic and often-repeated description of Hort: “The final impression produced by a review of all the trustworthy signs is of a patient and rather dull or mechanical type of transcription, subject now and then to the ordinary lapses which come from flagging watchfulness, but happily guiltless of ingenuity or other untimely activity of brain, and indeed unaffected by mental influences except of the most limited and unconscious kind.” Westcott and Hort, Introduction, 237; cf. von Soden, Die Schriften, 2:907. 25 Lagrange, La critique rationnelle, 86, 99; Wasserman, “Criteria,” 589; Pisano, “The Text,” 87; Head, “The Early Text,” 119. CHAPTER 6 196 may be said that the scribe was not at all prone to creating harmonizing variants.”26 Still, in those instances when scribe B does appear to have harmonized, it was usually to the Gospel of Matthew.27 This observation provides further support for our earlier observation that Matthew was prioritized both textually and paratextually. Yet, it is rarely clear how or if corrections play a role in these evaluations of the scribe.28 For this, we return to Hill’s article on the S-siglum corrections in the NT. As noted in Chapter 4, this method of correction is not present in the NT, except for the Gospels.29 Hill, following the argument of Tischendorf, has proposed that the early S-siglum corrections in Matthew mark variant readings from a second exemplar.30 Five of the eight corrections in Matthew agree with the majority of witnesses, while the remaining three present a minority reading. Versace, on the other hand, has suggested that the S-siglum corrections of his B1 were found in the margin of the main exemplars.31 Hill’s conclusions regarding the alternative text of the S-siglum corrections conform to what we have learned from the B3 S-siglum corrections and their relation to the Lucianic text in 1 Kingdoms and Psalms (§5.3). However, Versace’s claim makes the most sense of the production phases of B(03). If, as both Hill and Versace believe, the B1 S-siglum corrections were copied in scribendo, it is more likely that they were present in the main exemplars. A parallel to this can be seen in the Hexaplaric sigla and notations in the Greek OT or the diplai in the NT.32 The S-siglum corrections of B3, on the other hand, would likely have originated from the alternative exemplar. 6.1.2 The Earliest Corrections in the Gospels Addition In our comparison of the three scribes in 1 Kingdoms and Psalms, we only found early corrections of addition in the work of scribe B. It is no surprise, then, that we find examples of this in the NT of 26 In Greg Paulson’s study of singular readings in Matthew, he concludes that there does not appear to be any singulars in B(03)’s text of Matthew, which conflate with or harmonize to remote parallels. Pardee, Harmonization, 278; Paulson, Scribal Habits, 56. 27 Matthean material is harmonized within the gospel itself, but also in Mark and Luke. Therefore, Pardee claims that Matthew “served as the horizon of expectation” for scribe B. Pardee, Harmonization, 278–279. 28 The main exception is Pardee, who regularly mentions corrections. 29 See the S-siglum at p. 1397C, Acts 10:37, where there is no longer a marginal reading. Instead, a corrector erased the original reading and copied over it (discussed below). 30 Hill, “Siglum,” 15; cf. Tischendorf, Vaticanum, xxiv. 31 Versace, Marginalia, 11. 32 Versace, Marginalia, 89–92. EARLY CORRECTIONS IN THE NEW TESTAMENT 197 B(03). There are seven early corrections of additions in the Gospels. One further correction could have originated with an early hand, but appears to be late.33 Four of the eight corrections are of single-word additions, while the longest addition is comprised of forty-three characters (p. 1370C, John 13:14). TABLE 32: CORRECTIONS OF ADDITIONS IN THE GOSPELS SCRIBE B(03) CORRECTOR Matt 12:32; p. 1250B B ανθρωπου ου̇̇κ̇ αφε θηϲεται… B2 or BB Matt 21:4; p. 1262C B ινα πληρωθη το ρηθε ̅ ❛δια του πληρωθη το❜ ❛ρηθεν❜ δια του προφη του… B2 Matt 26:57; p. 1273B B εφυγον ❛οι δε κρατηϲα ❜̅ ❛τεϲ τον ιν̅̅ εφυγον❜ οι δε κρατηϲαντεϲ τον ιν̅̅ B2 Luke 1:37; p. 1305B B ϲτειρα ❛οτι ουκ αδυνα❜ ❛τηϲ ει❜ οτι ουκ αδυνα τηϲ ει παρα του θ̅υ̅… B2 Luke 13:22; p. 1331A B διδαϲκων και πορεια ̅ ❛π̇ορ̇̇ει̇α̇̇ν̇❜ ποιουµενοϲ B2 and/or BB John 7:28; p. 1360C B ο̸ ̇ισ̅̅ και λεγων καµε… B18 John 13:14; p. 1370C Β γαρ ❛ει ουν εγω ενιψα❜ ❛υµων τουϲ ποδαϲ ο κ̅ς ̅ και ο διδαϲκαλοϲ❜ ει ου ̅ εγω ενιψα υµων τουϲ ποδαϲ ο κ̅ς ̅και ο διδαϲκα λοϲ και υµειϲ… B2 John 17:11; p. 1375B Β εν καθωϲ κ̇α̇ι ̇ηµειϲ… B2 or BB The addition of ουκ at Matthew 12:32 (p. 1250B) is a singular reading and was clearly influenced by the appearance of ουκ αφεθηϲεται three lines above.34 While this unique reading would imply that “whoever speaks a word against the Son of Man will not be forgiven” (Matt 12:31; B(03)*), it is unlikely that there was any Christological motivation for adding the negative particle.35 Regardless, the faint presence of deletion dots above the word suggests that this reading was rejected from the beginning, 33 This does not include three corrections which are clearly from a later hand: p. 1303A, Mark 16:1; p. 1318A, Luke 7:39; p. 1375C, John 17:18. 34 Hoskier classifies this as a solecism. However, he does not note that this reading was corrected early on, potentially by the scribe. Hoskier, Codex B, 1:14; For a list of singular readings in Matthew, see Paulson, Scribal Habits, 155–162. 35 See also the two additions of µη that were corrected with deletion dots and cancelation strokes (p. 1358A, John 6:35; p. 1373A, John 15:7). Paulson, Scribal Habits, 51 n. 41; cf. Hagner, Matthew 1–13, 345. CHAPTER 6 198 either by scribe B or a diorthotes. Likewise, one of the early correctors was likely responsible for the deletion of και at John 17:11 (p. 1375B).36 Five of the ten corrections of addition remedy cases of dittography. The faint, round hooks above the single word addition of πορειαν (p. 1331A, Luke 13:22) were likely added by a diorthotes, though the scribe may have added the deletion dots. The dittography was perhaps created because of homoeoarchon with the following word ποιουµενοϲ. This factor may also have caused the replication of the phrase πληρωθη το ρηθεν at Matthew 21:4 (p. 1262C), which resulted in the nonsense reading δια του πληρωθη το ρηθεν, rather than δια του προφητου. On p. 1305B (Luke 1:37), the dittography could have originated with the visual confusion over the ending of ϲτειρα with παρα. Interestingly, the round hooks used to correct the addition at Matthew 26:57 (p. 1273B) do not enclose the first occurrence of εφυγον, but the second. As a consequence, the corrected reading separates the verb from the originally adjacent subject, οι κρατηϲαντεϲ, with the space of twenty-five characters.37 The largest addition is a dittography of forty-three characters (p. 1370C, John 13:14) and there is no clear reason that caused the scribe to err.38 Apart from those corrections which were made by leaving the original ink untouched (cf. n. 28), at least one correction of an addition appears to be from the reinker. While Tischendorf and the NTVMR transcription assign the correction at John 7:28 (p. 1360C) to their early hands (C1 and B2 respectively), the IGNTP John transcription assigns it to C2. We have already challenged the sixth- century dating of this hand as a misunderstanding of Tischendorf (Chapter 4), but the editors are probably right to understand this as a late correction (Figure 99). FIGURE 99: B18 CORRECTIONS OF ADDITION IN THE GOSPELS (PP. 1360C) 36 Tischendorf is unclear on this, as he identifies the correction as B3 in the Commentarius, but as B2 et B3 later in the edition. Tischendorf, Vaticanum, xxxxi, 141. 37 The contrast created by the later reinking certainly helps the reader span this gap. 38 The IGNTP transcription assigns this to a later hand (C2) than Tischendorf (B2) and the NTVMR transcription (C1). EARLY CORRECTIONS IN THE NEW TESTAMENT 199 This correction consists of a deletion of the article before the name ιηϲουϲ. Both the deletion dot and cancelation stroke appear to have been made in a darker ink than the original. This correction in John is striking, as it has been demonstrated that B(03) has a tendency to omit the article in front of ιηϲουϲ.39 The anarthrous nominative ιηϲουϲ is found in v. 28 and v. 33 in P75, while B(03) provides the article in both examples.40 It is unlikely that scribe B mistakenly added the article, since, as we will see in the following section, it is more common to find corrections of omitted articles. Omission In the Gospels there are approximately fifty-seven early corrections (BB, B2, B3) of omission. This number, in contrast to the minimal corrections of addition, fits Hort’s description of scribe B as being prone “to drop petty words not evidently required by the sense.”41 However, in the last chapter we also noted that scribe B’s work contained fewer early corrections of omission than that of scribe A (1 Kingdoms) and roughly the same average as scribe C’s (Psalms). The majority of these corrections remedy single-word omissions, while the largest omission consists of forty-two characters (p. 1247C, Matt 10:37). The most common omissions involve the apparent loss of an article, pronoun, or conjunction. Five of the single-word omissions were corrected by B3, who mostly used lemniskoi, except for one line-end correction (p. 1375C, John 17:15–16). TABLE 33: CORRECTIONS OF OMISSIONS IN THE GOSPELS 42 SCRIBE B(03) CORRECTOR Matt 5:16; p. 1239B B ˙/. εργα τα καλα ˙/. και δοξαϲωϲι ̅ B3 Matt 5:18; p. 1239B B εωϲ \αν/ παντα γενηται B2 or BB Matt 7:9; p. 1242B B η τιϲ ˙/. εξ υµων ανθρω ποϲ… ˙/. εϲτιν B2 or BB Matt 7:24; p. 1243A B τουϲ λογουϲ ˙/. και ποιει ˙/. του τουϲ B2 or BB Matt 10:37; p. 1247C B ↑ αξιοϲ ˙/. και οϲ ου λαµβα νει τον ϲταυρον… ↓ ο φιλων υιον η θυγατερα υπερ εµε B2 or BB 39 Jongkind, “Redactional Elements,” 238–241; see earlier, Nevius, “Definite Article,” 81–85; Fee, "Definite Article,” 168–183. 40 Porter, “Papyrus Bodmer XV (P75),” 368–374. On the absence of the article in Luke of P75 and B(03), see Martini, Problema, 80; Birdsall, “Rational Eclecticism,” 43. 41 Westcott and Hort, Introduction, 236; cf. Paulson, Scribal Habits, 48, who claims that “there are more instances of omission than addition” in B(03)’s singular readings in Matthew. [italics original] 42 This chart does not include corrections that clearly originated with B18 or B37: p. 1257B, Matt 16:24; p. 1305A, Luke 1:25; p. 1305A, Luke 1:34; p. 1325B, Luke 10:38. CHAPTER 6 200 ουκ εϲτιν µου αξιοϲ ˙/. Matt 12:48; p. 1251A B ειϲιν οι αδελφοι \µου/ και ε κτειναϲ… B2 or BB Matt 13:17; p. 1251C Β … οτι πολλοι προφητ [αι]vid οι επεθυµηϲαν… [αι δι BB Matt 14:2; p. 1253C Β νεκρων και ˙/. αι δυναµειϲ ˙/. δια τουτο BΒ Matt 14:3; p. 1253C Β … και εν \τη/ φυλακη B18 Matt 14:13; p. 1254A B [εκειθε ]̅ vid ιϲ̅ ̅ανεχωρηϲεν εκειθε [ν ε ̅ B2 or BB Matt 14:30; p. 1254C B ˙/. ιϲχυρο ̅ δε τον ανεµον ˙/. εφο βηθη… B2 or BB Matt 14:36; p. 1255A B ˙/. αυτον εχονταϲ και παρεκα λουν ˙/. ινα µονον… B2 or BB Matt 16:17; p. 1257A B ⟦ ⟧ καριοϲ ει ϲιµων [βαριω] [να οτι BB Matt 19:17; p. 1260C B αγαθου \ειϲ/ εϲτιν ο αγαθοϲ B2 or BB Matt 21:43; p. 1264B B δια τουτο λεγω υµιν [οτι B2 or BB Matt 23:37; p. 1267C B … τα νοϲ ϲια ˙/. υπο ταϲ πτερυγαϲ ˙/. αυτηϲ B2 or BB Matt 25:40; p. 1271A Β … ενι του των των ˙/. ελαχιϲτων ˙/. αδελ φων των ΒB Matt 25:42; p. 1271A Β γαρ και \ουκ/ εδωκατε µοι… B2 or BB Matt 26:3; p. 1271B Β … και οι πρεϲβυτεροι ειϲ την αυλην… [του λαου B2 or BB Matt 26:4; p. 1271B B … δολω κρατηϲωϲιν ελεγον δε µη… [ απο κτεινω ϲιν B2 or BB Matt 26:13; p. 1271C B … αµην \δε/ λεγω υµιν B2 or BB Matt 27:12; p. 1274B B … υπο των αρχιερεων και \των/ πρεϲβυ τερων… B2 or B18 Matt 28:15; p. 1277A B … οι δε λαβον τεϲ \τα/ αργυρια εποιηϲα ̅ Β18 Mark 2:16; p. 1279C B … των αµαρτω λων και \των/ τελωνων… B2 or BB Mark 4:1; p. 1281C B αυτον ειϲ \το/ πλοιον… B2 or B18 Mark 4:16; p. 1282B Β … ϲπειροµενοι οταν ακουϲωϲι… [οι B2 or B18 Mark 5:28; p. 1284A Β αψωµαι \καν/ των ιµατιω ̅ Β18 Mark 6:17; p. 1285B Β ˙/. την γυ ναικα … δια ηρωδι αδα ˙/. φιλιππου… B2 or BB Mark 6:54; p. 1287A B … και εξελ θοντων ˙/. εκ του… ˙/. αυτω ̅ B2 or BB EARLY CORRECTIONS IN THE NEW TESTAMENT 201 Mark 8:10; p. 1289A B απελυϲεν αυτουϲ ̅ευθυϲ εµβαϲ αυτοϲ [και ΒΒ Mark 10:19; p. 1292C B ψευδοµαρτυρηϲηϲ τειµα… [µη αποϲτε ρηϲηϲ B2 Mark 10:29; p. 1293A B … ενεκεν εµου του ευαγγελιου… [ενε κεν Β2 or BB Mark 10:45–46; p. 1293C B … αν[]τι πολλω ̅ και εκπ[]ορευοµενου [ ερχονται ειϲ ιερειχω B2 or BB Mark 14:32; p. 1300A B … καθιϲατε \ωδε/ εωϲ Β2 or B18 Mark 15:4; p. 1301C B … ουκ α ποκρεινη ˙/. ιδε ποϲα ϲου ˙/. ουδεν B2 or BB Luke 9:59; p. 1323B Β … ο δε ειπεν \κ̅ε/̅ επιτρε ψον µοι… Β18 Luke 10:15; p. 1324B Β … µη εωϲ \του/ ουρα νου… B2 or BB Luke 10:27; p. 1324C B κ̅ν̅ τον θ̅ν̅ \ϲου/ εξ οληϲ… B18 Luke 10:37; p. 1325B B ο] ιϲ̅ ̅πορευου και ϲυ… B2 or BB Luke 11:42; p. 1327B B … την αγα πην \του θυ̅̅/ ταυτα δε… B2 or BB Luke 13:11; p. 1330C B … εχουϲα α ϲθενειαϲ \ετη/ δεκαοκτω B2 or B18 Luke 13:14; p. 1330C B … οτι εξ ηµεραι ειϲιν \εν αιϲ/ δει εργαζεϲθαι B2 or BB Luke 14:27; p. 1332C B οϲτιϲ ουν \ου/ βαϲταζει Β18? Luke 18:15; p. 1337C Β … τα βρε φη ινα \αυτων/ απτηται ιδον B2 or B18 Luke 18:19;43 p. 1337C Β ουδειϲ αγαθοϲ ει µη ειϲ θ̅ϲ… [ο̣ ̣ B2 or BB? Luke 19:25; p. 1339B Β ειπαν αυτω \κ̅ε/̅ εχει δεκα B2 or BB? Luke 19:40; p. 1340A Β υµιν \οτι/ εαν ουτοι… Β18? Luke 20:9; p. 1340C Β απεδηµηϲεν χρονουϲ και καιρω… [ικα νουϲ B2 or BB? Luke 20:13; p. 1340C Β του αµπελωνοϲ \τι ποιηϲω/ πεµ ψω τον υιον µου… B2 or BB? Luke 22:19; p. 1344A B … ποιειτε \ειϲ/ την εµην Β18? Luke 22:39; p. 1344C B … ηκο λουθηϲαν δε αυτω \ / οι µαθηται… B2 or BB? Luke 22:40; p. 1344C B µη \ειϲελθειν/ ειϲ πειραϲµον… Β18? Luke 23:6; p. 1346A B … ει \ο/ ανθρω ποϲ γαλειλαιοϲ εϲτιν B2 or BB? Luke 24:15; p. 1348A B και ϲυνζητειν \ / αυτουϲ B2 or BB 43 The two dots that underlie the line-end correction could either be subscripted deletion dots or a later distigme. Versace, Marginalia, 133 n. 285. CHAPTER 6 202 Luke 24:52; p. 1349B B χαραϲ ˙/. και ηϲαν… ˙/. µεγαληϲ B3 John 1:4; p. 1349C B ζωη ην το φωϲ ˙/. και το [erasure] ˙/. των ανθρωπων B3 John 1:13; p. 1349C Β … ουδε εκ θεληµατοϲ ϲαρκοϲ [ουδε εκ θεληµα τοϲ ανδροϲ B2 or BC John 1:14;44 p. 1350A Β χαριτοϲ \ / αληθειαϲ Β18 John 3:34; p. 1353B Β ˙/. το π̅να̅̅ … ου γαρ εκ µετρου δι δωϲιν ˙/. ο πατηρ αγαπα Β3 John 4:1; p. 1353C B ποιει και βαπτιζει \η/ ιωα νηϲ… B2 or BB John 4:3; p. 1353C B … και απηλ θεν ˙ /. ειϲ την γαλειλαιαν ˙/. παλιν B2 or BB John 4:39–40; p. 1355A B εποιηϲα \ωϲ/ ουν \ϲυν/ηλθον ου ̅ B2 or BB John 10:25; p. 1365C B ο] ιϲ̅ ̅ειπον υµιν… B2 or BB John 12:15; p. 1369A B µη φοβου \η/ θυγατηρ… B2 or B18 John 12:18; p. 1369A B τουτο \ / υπηντηϲεν… B2 or BB John 14:10; p. 1372A B … α εγω \λεγω/ υµιν… Β18 John 17:15–16; p. 1375C B … ινα α ρηϲ αυτουϲ εκ το[]υ ⟦πο⟧ ν̇η̇ρ̇ου̇̇ εκ του \π/κ̸ο\νη/ϲ̸\ρ/µ̸ου ουκ ειϲιν… [κοϲµου αλλ ινα τηρη ϲηϲ αυτουϲ [εκ του κοϲµου Β3 John 18:36; p. 1377B Β οι εµοι ηγωνιζοντο \αν/ ι να µη... B2 or BB In the corrected Synoptic omissions, Pardee identifies four which may have occurred because of harmonization (p. 1243A, Matt 7:24, cf. Luke 6:47; p. 1271A, Matt 25:40, cf. Matt 25:45; p. 1292C, Mark 10:19, cf. Matt 19:18//Luke 18:20; p. 1293C, Mark 10:46//Matt 20:29) and one correction towards a harmonization (p. 1271C, Matt 26:13//Mark 14:9).45 However, of these five corrections, only one cannot also be explained by homoeoteleuton (p. 1271C, Matt 26:13). Since this appears to be the most common reason for omission in B(03) and harmonization exerted minimal force in the copying of the codex, we might take homoeoteleuton to be the better explanation for four of these five omissions. The corrections to the above omissions were all made by early correctors. We can further specify that scribe B likely added the marginal correction of αδελφων µου των (p. 1271A, Matt 25:40), 44 Both the NTVMR and the IGNTP transcriptions mark this correction as C1. However, Tischendorf was probably correct that the και-compendium originated with B18, his B3. 45 Pardee, Harmonization, 229, 245, 261–262. EARLY CORRECTIONS IN THE NEW TESTAMENT 203 since the “Coptic” style µου-ligature, preferred by this scribe, is used (Figure 100). We will see the same omission in Romans 9:3, which again indicates accidental omission rather than harmonization. FIGURE 100: SCRIBE B'S ΜΟΥ-LIGATURE AND MARGINAL CORRECTION (PP. 1271A, 665B) Additional corrections, likely made by scribe B, include those on p. 1251C, Matt 13:17; p. 1253C, Matt 14:2; p. 1257A, Matt 16:17; p. 1289A, Mark 8:10; p. 1293C, Mark 10:46. The alpha in the line-end addition of και δικαιοι (p. 1251C, Matt 13:17), betrays the hand of the scribe B (Figure 101). That there is little evidence of erasure beneath -οι- (originally -[α]ι-) in the following line, may indicate that the correction was made before the ink had dried on the parchment.46 By comparison with the prima manus, S-siglum corrections (B1), the lemniskos correction on p. 1253C (Matt 14:2) was also an intervention by scribe B (Figure 102). Two columns before this correction, scribe B added the words λεγει αυτοιϲ, accompanied by the S-siglum. Since the dative pronoun is already present in the main text, the reinker left it untouched, leaving the original hand visible. If we compare this correction with the lemniskos correction, δια τουτο, we find a very similar hand in the successive letters -υτο-, bearing in mind the obstruction caused by the reinking of the latter. Versace identifies the και on p. 1289A l. 5 (Mark 8:10) as a line-end correction from B2.47 Since there is a textual division, the και fits within the line and does not extend far into the margin. Based on the resemblance with the και directly above the correction, it seems likely that this was made a prima manu. Scribe B was undoubtedly responsible for several of the corrections marked “B2 or BB”, but there are few distinguishing features to assist their identification. FIGURE 101: LINE-END CORRECTION (P. 1251C, MATT 13:17) 46 The omission is a singular reading in B(03). It is possible that the omission was caused by the line-break and homoeoteleuton with the ending -ται and the conjunction και. 47 Versace, Marginalia, 133. CHAPTER 6 204 FIGURE 102: SCRIBE B CORRECTIONS (P. 1253A, MATT 13:52; P. 1253C, MATT 14:2) One of the corrections of omission may come from scribe C (p. 1349C, John 1:13). In this line-end correction, the vertical orientation of the alphas, deltas, and lambdas resembles that of our third scribe (Figure 103). Since this is a sub-singular reading, the omission is best explained as stemming from the repetition of the phrase ουδε εκ θεληµατοϲ.48 FIGURE 103: SCRIBE C CORRECTIONS (P.1349C, JOHN 1:13; P. 780B, JOB 12:23) In Chapter 4, we noted a complex correction of omission, which involved the use of numerous correction methods (p. 1375C, John 17:15–16). Here, scribe B omits twenty-three characters because of the repetition of a subjunctive ending in -ηϲ, followed by the words αυτουϲ εκ του (Figure 104). The confusion is amplified by the recurrence of the phrase εκ του κοϲµου five times within seven lines of B(03). While the deletion dots and cancelation strokes look as if they originated with B18, because of the dark ink, the supralinear and marginal additions reveal the hand of B3.49 FIGURE 104: B3 CORRECTION OF OMISSION (P. 1375C, JOHN 17:15–16) Some of these omissions, on the other hand, were corrected by the reinker (B18). Yet, it is striking that this late hand was relatively restrained in expanding the text of B(03) in the Gospels. While the medieval reinker was thoroughly involved in correcting nonsense readings and orthography (see below and Chapter 5), we might also expect extensive corrections of omission, since as Dean Burgon noted, B(03) “is found to omit at least 2877 words” in the Gospels, when compared to the 48 Tischendorf cites the first hand of GA 17, Eusebius, Athanasius, and Chrysostom. 49 Because of the dark ink and the irregular hand of B3, Tischendorf and the NTVMR transcription identify the corrector as the reinker. However, the IGNTP transcription classifies the correction as C1. cf. Versace, Marginalia, 136. EARLY CORRECTIONS IN THE NEW TESTAMENT 205 Received Text.50 That there are fewer corrections of omission might suggest a more conservative approach behind the reinker’s agenda—to preserve and standardize the text, but not to thoroughly alter it. Substitution Corrections of substitutions make up an even larger number in the Gospels.51 Forty-six of these corrections possibly originated from an early hand. However, there are also a significant number of substitutions that were corrected by B18. Thus, it is not always clear, especially with single letter corrections, whether an earlier correction underlies the reinked text. The following examples of substitution regularly involve the change of preposition, the case of nouns and adjectives, or the tense, voice, and mood of verbs. Twice we find a substitution of proper names in the Gospels (p. 1310A, Luke 3:24; p. 1351A, John 1:42). TABLE 34: CORRECTIONS OF SUBSTITUTIONS IN THE GOSPELS SCRIBE B(03) CORRECTOR Matt 6:34; p. 1242A B … γαρ αυριον µε ριµνηϲει \ε/αυτηϲ αρκε τον… B18? Matt 7:14; p. 1242C B … δι αυτηϲ ο̸τ̇ι δε… B18 Matt 8:24; p. 1244B B … \α/⟦υ̸⟧πο των κυµατων… B2 or BB Matt 9:30; p. 1246Α B … ενεβρειµη\ϲατο/θη B18? Matt 11:19; p. 1248C B S τεκνω ̅ S εργων αυτηϲ… BΒ Matt 13:8; p. 1251B Β … αλλα δε επεϲεν ε⟦ιϲ⟧ πι την γην… BΒ Matt 13:13; p. 1251C Β ϲυνι\ω/ουϲιν… B18 Matt 13:14; p. 1251C Β … ακουϲ\ε/ατε… B18 Matt 13:24; p. 1252A Β [ελα[?]εν] vid … παρεθηκεν αυτοιϲ Β2 or BΒ Matt 13:52; p. 1253Β Β S λεγει αυτοιϲ … ο δε S ειπεν αυ τοιϲ… BΒ Matt 14:5; p. 1253C Β S οτι S επει ωϲ προφητην … BΒ Matt 14:19; p. 1254Β Β ωδε αυτουϲ και κελευ [ε]vid ϲατ̸̇ϲ τουϲ οχλουϲ… Β2 or B18 50 Burgon, Revision Revised, 11. 51 The following table does not include corrections which clearly originated with B18 or another late hand: p. 1247B, Matt 10:25; p. 1247B, Matt 10:25; p. 1251B, Matt 13:4; p. 1257C, Matt 17:8; p. 1294Α, Mark 10:48; p. 1305Β, Luke 1:36; p. 1323C, Luke 10:6; p. 1324B, Luke 10:15; p. 1368C, John 12:13. CHAPTER 6 206 Matt 16:4; p. 1256Β Β S επιζη τει S αιτει και ϲηµειον… BΒ Matt 16:20; p. 1257Α Β … τοτε S επετει µηϲεν… S διεϲτει λατοvid BΒ Matt 21:18; p. 1263B B επαναγ̸̇α̸̇γων ειϲ την πολιν… Β2 or BΒ Matt 22:10; p. 1265A B ϲυνηγαγον πανταϲ [ο ϲ]ουϲ ευρον πονηρουϲ Β2 or BΒ Matt 22:10; p. 1265A B … και ε πληϲθη S ο νυµφων S ο γαµοϲ BΒ Matt 23:26; p. 1267A B … και το ε κτοϲ αυτ\ων/ο̸υ̸ καθαρο ̅ Β2 or BΒ Matt 25:41; p. 1271A B … εξ ευω νψµων S πορευεϲθε… ⟦S υπα γετε⟧ ΒB Matt 27:4; p. 1274A B … αιµα S αθωο ̅ S δι ο ̅ BΒ Matt 27:13; p. 1274Β B ειϲ \π/οϲα ϲου… Β2 or BΒ Matt 27:35; p. 1275Β B … δε αυτον διεµεριϲα ̅ [το Β2 or BΒ Mark 3:12; p. 1281Α B … αυτον φανερο ̅ ποιη̇ϲ ̇ω̸ϲιν… Β18 Mark 4:21; p. 1282B B … ουχ ινα \ε/υ π\ι/ο την λυχνιαν… Β2 or B18 Mark 4:28; p. 1282C B χορτον ειτ\α/εν ϲταχυ ̅ ειτ\α/εν πληρεϲ ϲειτοϲ Β18 Mark 4:38; p. 1283A B … και \δι/εγειρουϲι ̅ Β2 or BΒ Mark 5:13; p. 1283C B [⟦ ⟧β̅] την θαλαϲϲαν ωϲ /β̅ … B18 Mark 5:38; p. 1284C B … και αλαλαζονταϲ πολλαϲ̇ ̸και… Β2 or BΒ Mark 6:4; p. 1285Α B ϲυγγενευ̸̇ϲιν αυτου Β2 or BB Mark 6:33; p. 1286Α B και ε\πε/γνωϲαν πολλοι… B18 Mark 6:39; p. 1286Β B αυτοιϲ ανακλειθ̇η̇ναι Β2 or BB Mark 6:39; p. 1286Β B [εν]vid … επ\ι/ τω χλωρω χορ τω… Β2 or B18? Mark 9:30; p. 1291Β B … εξελθοντεϲ \παρ/επο ρευοντο δια τηϲ γαλει λαιαϲ… Β18 Mark 14:61; p. 1301Α B … ο\ϲ/ δε εϲιωπα Β2 or BB Mark 15:6; p. 1301C B … ενα δεϲµιον ον π\ε/αρητουντο… Β18 Mark 15:42; p. 1302C B … ο εϲτιν προ\ϲ/ ϲαββατο ̅ Β18 Luke 1:17; p. 1304Β B και αυτοϲ προϲ̸ε̇λευ ϲεται ενωπιον αυτου Β2 or BB Luke 1:17; p. 1304Β B εν πνευµατι και δυνα µει ηλει\ου/α επιϲτρεψαι Β18? EARLY CORRECTIONS IN THE NEW TESTAMENT 207 Luke 2:13; p. 1307A/Β B l. 42 … ουρα || l. 1 ν\ι/ου… Β2 or B18 Luke 2:44;52 p. 1308Β B εν τοιϲ ϲυγγενευ̸̇ϲιν B18 Luke 3:1; p. 1308C B S βαϲι λειαϲ … πεντεκαιδε κατω τηϲ S ηγεµονιαϲ ΒΒ Luke 3:1; p. 1308C B S ορει νηϲ S ιτουραιαϲ και τραχω νειτιδοϲ… ΒΒ Luke 3:24; p. 1310A B [ηλει] του η̸̇λευει Β2 or BB Luke 5:29; p. 1313C B … οι ηϲαν µετ αυ [του]vid των κατακειµενοι Β2 or BB Luke 6:38; p. 1315C B … γαρ µετρω µε τρειτε \αντι/µετρεηθηϲεται Β2 or BB Luke 8:13; p. 1319Α B \ο/α̸υτοι ριζαν ουκ… B18? Luke 8:54; p. 1320C B τηϲ χειροϲ αυτη\ϲ/ εφω νηϲεν λεγων… B18 Luke 9:18; p. 1321C B … κατα µοναϲ ϲυν η̸̇ν̸̇τ̸̇ηϲαν αυτω… [η B18 Luke 9:23; p. 1322A B … ερχεϲθαι ❛α̸̇π̇❜αρνη ϲαϲθω εαυτον… Β2 or B18 Luke 10:34; p. 1325A B … και \επ/εµε[µε]ληθη… B18 Luke 11:16; p. 1326A B … δε πει ραζοντεϲ ϲηµειου̸̇ [ ̅ B18 Luke 13:7; p. 1330B B … και τ\η/ο̸ν \γην/ τ̸ο̸π̸ο̸ν̸… Β2 or BB Luke 13:15; p. 1330C B … και απα\γα/γων… B18 Luke 13:28; p. 1331B B οταν οψ\η/εϲθη αβρααµ Β2 or B18 Luke 13:31; p. 1331B B … τη \ηµε/ω̸ρα προϲηλθαν Β2 or B18? Luke 14:16; p. 1332B B … εποιει δειπνο ̅ µεγα\ν/̈ και εκαλεϲεν… Β2 or B18? Luke 14:18; p. 1332B B και εχω αναγκη\ν/ εξελ θων… B18? Luke 15:4; p. 1333B B και απολεϲ\αϲ/η̸ εξ αυτω ̅ Β2 or BB Luke 16:1; p. 1334B B … οικονοµ\ν/ουϲ και… B18 Luke 16:15; p. 1335A B … οτι το εν ανθρω π\οιϲ/ω̸ υψηλον… B18? Luke 19:16; p. 1339A B … κ̅ε ̅η µνα\ϲ/ ϲου δεκα B18? Luke 21:14; p. 1342B B θε\ϲθ/τε ουν εν ταιϲ καρ διαιϲ υµων… B18? Luke 22:30; p. 1344B B και καθη\ϲε/ϲθ\αι/ε επι θρο νων… B18? 52 This correction may also be classified as orthographic. CHAPTER 6 208 Luke 24:15; p. 1348A B και ϲυνζητειν αυτου̇ϲ Β2 or BB John 1:13; p. 1349C B αλλ εκ θυ εγεν\ν/ηθηϲα ̅ B18? John 1:15; p. 1350A B λεγων ουτοϲ ην ο\ν/ ειπ\ο/ω ̅ B18 John 1:42; p. 1351A B ιω[α]ν\α/[ου] ϲυ κληθηϲη B18 John 4:40; p. 1355A B εποιηϲα \ωϲ/ ουν \ϲυν/ηλθον ου ̅ Β2 or BB John 4:42; p. 1355A B … δια την S λαλια ̅ S ΒΒ John 5:7; p. 1355C B εγω αλλοϲ προϲ̸ ̇εµου Β2 or BB John 7:3; p. 1360A B θεωρηϲ\ω/ουϲιν ϲου… B18 John 8:39; p. 1362C B … τα εργα του αβρααµ ε] ποιειτε νυν δε… B18 John 8:54; p. 1363B B λεγετε οτι θ̅ϲ ̅\η/υµων ε ϲτιν… B18 John 8:56; p. 1363B B … ινα ει̇δη την ηµεραν… Β2 or B18 John 10:29; p. 1366A B ο πατηρ µου ο\ϲ/ δεδωκε ̅ Β2 or B18? John 11:27; p. 1367A B [πιϲτευω] … εγω \πε/πιϲτευκα ο τι ϲυ… BB John 12:40; p. 1370A B οφθαλµουϲ και \π/επω ρω\κ/ϲεν αυτων… Β18 John 14:10; p. 1372A B ου πιϲτευϲ̸ε̇ιϲ οτι εγω Β2 or B18? John 19:31; p. 1379A B [εκεινη] … η ηµερα εκεινου Β18 John 20:22; p. 1380B B … ειπων εν S εφυϲηϲε S BB John 20:23; p. 1380B B αφει\ε/ονται αυτοιϲ… Β18 All ten of the NT’s S-siglum corrections appear as substitutions in the Gospels.53 Rather than the hand of B3, as in 1 Kingdoms and Psalms, these corrections often bear the resemblance of scribe B or a diorthotes. The substitutions ειπεν αυτοιϲ → λεγει αυτοιϲ (p. 1253B, Matt 13:52) and ηγεµονιαϲ → βαϲιλειαϲ (p. 1308C, Luke 3:1) both share similar alphas and lambdas with the scribe (Figure 105). The very first S-siglum correction can likely be attributed to scribe B, as its presence in the margin 53 Two further S-sigla appear in John, but they do not have an alternative reading (p. 1355A, John 4:42; p. 1380B, John 20:22). While an S-siglum appears in Acts 10:37 (p. 1397C), it is not accompanied by an alternative reading or a marginal S-siglum. EARLY CORRECTIONS IN THE NEW TESTAMENT 209 forced the slightly later addition of the numeral below the normal position (p. 1248C, Matt 11:19).54 Most agree that the corrected reading τεκνων arose from a harmonization with Luke 7:35.55 However, two other s-sigla appear to be correcting harmonizations in the main text. Pardee argues that the corrections επει → οτι (p. 1253C, Matt 14:5) and επετειµηϲεν → διεϲτειλατο (p. 1257A, Matt 16:20)56 were made against harmonization with Matthew 21:46, in the first instance, and Mark 8:30//Luke 9:21 in the latter.57 If, as we have suggested, both readings were present in the exemplar, then scribe B was not the creator of the harmonization. The correction ο νυµφων → ο γαµοϲ (p. 1265A, Matt 22:10) is particularly challenging as the main text appears to have the easier reading (“wedding hall”), while the marginal reading (“wedding”) is slightly less specific, but supported by the majority of manuscripts.58 It is noteworthy that the article, already present in the main text, is recopied in the margin. This feature supports the view that the S-siglum marginalia provide alternative readings rather than proper corrections.59 FIGURE 105: S-SIGLUM CORRECTIONS IN THE GOSPELS (P. 1248C, MATT 11:19; P. 1253B, MATT 13:52; P. 1308C, LUKE 3:1) Along with the prima manus corrections, which utilize the S-siglum, there are at least two in scribendo corrections of substitution (Figure 106). The first correction ειϲ → επι (p. 1251B, Matt 13:8) 54 There has been little agreement about the date of this correction. Versace and Hill agree that this correction likely came from the scribe, while Tischendorf originally identified the reading with B3, but changed it to B2 in his 8th edition. The NTVMR transcription associates the reading with C1, while the NA28 and the UBS committee identified it as B2 (6th/7th c.). The THGNT also places this marginal correction with their later B2. It is possible that the correction was added by a diorthotes, but it certainly was added before the red section numbers in the fourth-century. Tischendorf, Vaticanum, 14; Versace, Marginalia, 89; Hill, “Siglum,” 10 n. 52. 55 Pardee identifies the opposite feature in א(01), which has the reading εργων in both Matthew and Luke. Metzger, Textual Commentary, 24; Hill, “Siglum,” 10; Pardee, Harmonization, 232. 56 This correction was made twice. Apparently scribe B or a diorthotes added the correction in the margin, but it was later scratched out. Versace’s B37 later rewrote the correction over the original reading in column A. Versace does not provide a reading for the B1 correction, since it is hardly visible. Therefore, the marginal reading could be an unknown variant. Versace, Marginalia, 89; Hill, “Siglum,” 13. 57 However, Hill has argued against the harmonization of Matthew 16:20 to Mark 8:30//Luke 9:21, since Matthew was likely dependent on Mark’s text, and since this would be Matthew’s only use of the verb διεϲτελλοµαι. Unfortunately, this does not explain the origin of the marginal reading. If the original reading is not a harmonization to Mark, then it was likely a harmonization to Matthew’s other uses of the verb επιτιµαω. Pardee, Harmonization, 233, 235; Hill, “Siglum,” 13. 58 Metzger, Textual Commentary, 47; Hill, “Siglum,” 14. 59 Thus, it is appropriate that the THGNT prints γαµοϲ with νυµφων as a diamond reading. CHAPTER 6 210 was executed by partially erasing the iota and sigma, then starting the next line with pi and iota. It was clearly made in scribendo since the corrected reading fits appropriately within the line of the original text. Although Pardee does not mention this variant, the original reading is presumably a harmonization to Mark 4:8//Luke 8:8 (… επεϲεν ειϲ την γην). The correction πιϲτευω → πεπιϲτευκα (p. 1367A, John 11:27) was also made in scribendo. Immediately after copying the present tense form of the verb, scribe B added the supralinear reduplication of pi and epsilon, followed by the perfect ending. While the kappa was added over the original omega, there are no signs of erasure beneath the alpha. The regular spacing in the line suggests that the correction was made before the following word (οτι) was copied. Admittedly, the original form of the supralinear letters has been obscured by the reinker and cannot be assigned confidently to the hand of scribe B on paleographic grounds. FIGURE 106: IN SCRIBENDO CORRECTIONS OF SUBSTITUTION (P. 1251B, MATT 13:8; P. 1367A, JOHN 11:27) Another correction which may have originated with scribe B is found in the genealogy of Luke (p. 1310A, Luke 3:24). Here, the name ηλει is corrected to λευει, through the deletion of the eta and the addition of epsilon and iota (Figure 107).60 Rather than adding a supralinear upsilon, the corrector built the new letter off of the stem of the original iota. Since the name ηλει appears two lines before, on the previous page, the first reading was likely an error caused by the turning of the page and the visual similarity between the two names.61 If this error was not corrected by scribe B, then it was made by an early diorthotes. FIGURE 107: SUBSTITUTION OF PROPER NAME (P. 1310A; LUKE 3:24) 60 The NTVMR transcription follows Tischendorf in giving an original nonsense reading ηλειει. However, it is more likely that the second -ει- was a part of the correction, since the name ηλει was just copied by the scribe, two names earlier. The unique spacing of the genealogy ensures that there was plenty of space for the corrector to copy additional letters without needing to compress or raise them. 61 In this case, it should also be classified as a nonsense reading. EARLY CORRECTIONS IN THE NEW TESTAMENT 211 While numerous corrections of substitution were made by the B18, the correction of the numeral in Mark 5:13 (p. 1283C) deserves mentioning. In Chapter 3, we discussed the single occurrence of a numeral in the NT. While Zachary Cole is right to highlight the uniqueness of the numeral in the NT of B(03), it is actually the reinker who gives the reading /β̅ rather than the scribe (Figure 108).62 Cole’s suggestion that the original reading was ιβ̅̅ is plausible, but the remains of the erasure suggest that the numeral may have been ϥ̅β̅ or /ιβ̅̅. 63 While the importance of “twelve” in the NT is undeniable, the higher number of ninety-two or 10,002 makes more sense in reference to the swine- herd. All three possible readings appear to be unique, and the origin of the numeral likely lies in the exemplar. FIGURE 108: B18 NUMERAL CORRECTION (P. 1283C; MARK 5:13); CF. P. 1120C Orthography As in 1 Kingdoms and Psalms, the majority of corrections in the Gospels are orthographic. The following table only examines corrections which could have originated with the early correctors. However, our work in the previous chapter revealed that there are very few orthographic corrections which likely originated in the fourth century. Instead, even those corrections which use deletion dots or cancelation strokes often appear to have originated with the reinker (e.g., µει̇κροϲ). As a result, the following evaluation is in some ways incomplete, as it does not include other corrections by B18, when the reinker chose to leave letters untouched. This is mostly the case with corrections of ει → ι. However, since our focus is on the scribes and earliest correctors, we will use the present, albeit limited data to discuss spelling habits of scribe B.64 Collectively, the largest number of corrections come from the interchange ι-ει. The higher frequency of ει → ι corrections than in 1 Kingdoms and Psalms is due to the frequency of words like µεικροϲ, which are usually corrected with a deletion dot or cancelation stroke. It is not entirely clear 62 The online ECM incorrectly cites B(03) as reading διϲχιλιοι. Tischendorf classifies the correction as B3, but does not give the original reading. Cole, Numerals, 99. 63 Through personal correspondence, Cole has mentioned a number of papyri with horn-like strokes to mark numerals in the thousands. 64 Since we are concerned with the main text, the tables below do not include orthographic corrections of titles, which can be found in Matthew or Colossians, for example. CHAPTER 6 212 why this method was used by B18, since it was common practice to leave the unwanted letter to fade. Tischendorf and the NTVMR transcription present three ει → ι corrections which may have been corrected by the scribe or diorthotes (p. 1284B, Mark 5:29; p. 1344Β, Luke 22:27; p. 1363Β, John 8:56). Although the deletion dot over the epsilon of ειαται shares a similar color to the original ink, the irregularly high diaeresis over the iota may be a sign of a late addition (p. 1284B, Mark 5:29). In Luke 22:27, scribe B gives the only NT example of ουχι with the long /i/ spelled -ει-. Since it is a one-time occurrence, the correction could have originated with the scribe or diorthotes. However, if it was an early correction, the reinking of the deletion dot obscures it.65 Another group of orthographic corrections which may have originated with scribe B or a diorthotes are the corrections of ερρεθ̣η → ερρηθη in Matthew 5. Tischendorf documents the correction five times with the designation B2 (p. 1240, Matt 5:27, 31, 33, 38, 43), and the NTVMR increases this collection by one (p. 1239C, Matt 5:21), cited as C1. 66 However, there is no clear evidence that the extra correction in the NTVMR transcription originally read ερρεθη. 67 The remaining five corrections show signs of erasure, and the form of the etas appear to match those of scribe B. The same correction of ερρεθη → ερρηθη occurs in Jonah 3:7, Susannah 27, Romans 9:12, 26, and Galatians 3:16 (see below), but these were clearly made by B18, who copied supralinear etas, instead of overwriting the epsilon.68 If, therefore, scribe B or a diorthotes did make the corrections in Matthew, it is likely that the scribe preferred the epsilon spelling, but may have found the eta spelling in the exemplar. Unlike 1 Kingdoms and Psalms, the interchange αι-ε is very frequent in the Gospel corrections (41x combined). The interchange regularly affects the endings of second-person plural active and middle-passive verbs (-τε and -ϲθε), third-person singular and plural middle-passive verbs (-ται), and middle-passive infinitives (-ϲθαι). These verb forms occur over five-times more in the Gospels than in 1 Kingdoms and the extant portions of Psalms. Thus, the disparity in corrections cannot be easily associated with either the spelling of scribe B or the exemplars. A large number of the α → ο corrections highlight the frequency of the α-conjugation in thematic aorists. This spelling 65 For this reason, Tischendorf cites this correction as B2(vid) et B3 66 Hort does not seem to recognize these corrections when distinguishing the spelling in the Gospels from Paul. Westcott and Hort, Introduction, 166. 67 I thank An-Ting Yi for highlighting this discrepancy before I had started research for this chapter. 68 The epsilon spelling is not corrected in 2 Kingdoms 5:6 or Hosea 1:10. EARLY CORRECTIONS IN THE NEW TESTAMENT 213 is rare in both 1 Kingdoms or Psalms and is never corrected by B18.69 Therefore, the α-conjugation was likely in the exemplar(s) of the Gospels rather than a preference of scribe B. The corrections of εορακαϲ → εωρακαϲ are mostly made by the reinker, who transformed the original omicrons into narrow omegas (Figure 109). FIGURE 109: ORTHOGRAPHIC CORRECTION OF ΕΩΡΑΚΑC (P. 1363B, JOHN 8:57) As we saw in the previous chapter, scribe B shows a higher tendency to leave nus unassimilated, compared to the other two scribes. This pattern continues in the Gospels, especially with the corrections of ν → γ, which were made by B18. Similarly, the unaspirated ουκ is corrected six times in the Gospels.70 While the aspirated ουχ is never corrected to the unaspirated form, we do find a single correction of εξ → εκ in (p. 1236B, Matt 2:6). The most common correction in proper names concerns the interchange of τ-θ, especially with geminates. B18 regularly makes these corrections by supralinear addition, though the corrector will occasionally overwrite the old letter (θ → τ; p. 1308B, Luke 2:39). The preference of B(03) to present the name ιωαννηϲ with a single nu is well documented.71 Jongkind has recently suggested that the -ν- spelling was an editorial decision behind the production of B(03), and that the few cases of -νν- reflect scribe B’s personal spelling preference, which has occasionally resurfaced.72 Martini has demonstrated some similarities between P75 and B03 in their spelling of ιωανηϲ.73 Lastly, the two spellings of Nazareth were harmonized by B18 into the form ναζαρετ. Twice in the Gospels, the augmented form of pluperfect ιϲτηµι is corrected to the unaugmented form (p. 1346A, Luke 23:10; p. 1361A, John 7:37). Hort is certain that the original form is ιϲτηκει since even B(03), with its “habitual addiction” of ει for ι, has the unaugmented form five 69 Cf. 1 Kingdoms 6:19; 19:20 and Psalms 34:21. 70 The NA28 prints the aspirated ουχ forty-one times in the Gospels. 71 Westcott and Hort, Introduction, 159; cf. BDF §40; Blass, Philology of the Gospels, 75, cited in Parker, Codex Bezae, 109. 72 He also argues that the -ν- spelling may reflect knowledge of the underlying Hebrew, since it is “philologically more correct.” Jongkind, “Redactional Elements,” 236. Cf. Ropes’ statement “The great significance of B[03] lies in the general soberness of its text (except in the proper names) and its relative freedom from deliberate revision.” Ropes, Beginnings, xcii. 73 Martini, Problema, 98. CHAPTER 6 214 times in the NT.74 Three of the remaining seven occurrences in the NT are corrected from ει → ι with cancelation strokes; once accompanied by a deletion dot (p. 1361A, John 7:37). In Luke 23:10, the reading ειϲτηκειϲαν was corrected twice, to the unaugmented form and back to the original (Figure 110). However, this does not guarantee that the correction originated with the scribe or diorthotes, as Tischendorf, Hort, and the NTVMR transcription assume. The use of a cancelation stroke could have been added by the reinker and then covered over by B37. FIGURE 110: ORTHOGRAPHIC CORRECTION OF ΕΙϲΤΗΚΕΙϲΑΝ IN LUKE 23:10 (P. 1346A) Transposition While there are three corrections of transpositions in the Gospels, only one appears to have originated with scribe B (p. 1248B, Matt 11:9). This correction was clearly made in scribendo, since the line has been unaffected. After the verb ιδειν was initially copied, scribe B erased it and copied the word προφητην, before recopying ιδειν. The diaeresis and effects of the erasure are still visible (Figure 111). Since scribe B never finished copying the two words in the first attempt, this could be read as an initial omission of προφητην. However, the transposition has support from a number of witnesses.75 TABLE 35: CORRECTIONS OF TRANSPOSITIONS IN THE GOSPELS SCRIBE B(03) CORRECTOR Matt 7:17; p. 1242C B // / καρπουϲ ποιει καλουϲ Β 18 Matt 11:9; p. 1248B B ⟦ιδειν⟧vid … προφητην ιδειν… B B John 8:37; p. 1362B B // / … ζητειτε µε απο /// κτειναι οτι ο λογοϲ… B2 or Β18 74 The editors print the unaugmented form in all fourteen occurrences in the NT. Westcott and Hort, Introduction, 162. 75 NA28 lists the following as witnesses to the corrected reading of B(03): א(01)*, W(032), Z(035), etc. However, the majority of manuscripts give the original reading of B(03). There is an error in the apparatus, which cites B2 as agreeing with B*. Clearly the later corrector (B18) followed the corrected reading. EARLY CORRECTIONS IN THE NEW TESTAMENT 215 FIGURE 111: CORRECTION OF TRANSPOSITION IN THE GOSPELS (P. 1248B; MATT 11:9) The other two corrections of transposition use consecutive, supralinear strokes to reorder the text. Tischendorf did not identify the first correction, but associates the third with an early hand. It is likely, however, that the reinker executed both. Nonsense Although corrections of nonsense readings are a dominant factor in the Gospels, the majority appear to have been made by B18. The following table provides many apparent B18 corrections, except those in which the correction was made by leaving text untouched. 76 While some may have originated early, with little trace remaining, the table illustrates the high density of late corrections of nonsense readings in contrast to the few that were made by the scribes or a diorthotes. Of the seventy-one nonsense corrections listed below, thirty-nine involve the addition or omission of a single letter, and twenty-two involve two letters. As we saw in the previous chapter, these small errors went undetected by the early correctors. However, the nature of the reinking process meant that they were easily noticed by the reinforcer. TABLE 36: CORRECTIONS OF NONSENSE READINGS IN THE GOSPELS SCRIBE B(03) CORRECTOR Matt 3:12; p. 1237C B πυρι αϲβε\ϲ/τω τοτε... Β18? Matt 6:32; p. 1242A B ο ουρανιοϲ οτι χρη\ζε/τε B2 or Β18 Matt 10:14; p. 1246C B ϲτραφητω και οϲ αν υ]µαϲ µη δε ακουϲη… [µην δεξη BB Matt 10:19; p. 1247A B … δε παραδωϲιν υ µαϲ µη µεριµ\ν/ηϲητε Β18 Matt 10:22; p. 1247A B … οδε υ ποµε\ι/ναϲ ειϲ τελοϲ… Β18 Matt 12:1; p. 1249A B … των ϲπορι µων ο\ι/ δε µαθηται… Β18 Matt 12:33; p. 1250B B δενδρον \ϲ/απρον και το ̅ καρπον… B2 or Β18 Matt 13:15; p. 1251C B … του του και τοι\ϲ/ ωϲιν… B2 or Β18 Matt 13:30; p. 1252B B το κατακαυϲαι αυτα⟦ϲ⟧ Β18 or Β37? Matt 13:32; p. 1252B B και καταϲκηνο\υ/ιν εν Β18 76 Likewise, it excludes corrections which have no possibility of being early: p. 1280Β, Mark 2:26; p. 1283C, Mark 5:13; p. 1317C, Luke 7:33; p. 1325B, Luke 10:41. CHAPTER 6 216 Matt 15:32; p. 1256A B … προϲκα λεϲαµενοϲ του\ϲ/ µαθη ταϲ αυτου… B2 or Β18 Matt 17:10; p. 1258A B ηλειαν δ\ε/ι ελθειν πρω τον… B18 Matt 17:23; p. 1258B B και τη τρι\τη/ ηµερα… B18 Matt 18:32; p. 1260A B … αφηκα ϲοι ε π\ε/ι παρεκαλεϲαϲ… B18? Matt 20:15; p. 1261C B [εξι] ϲοι ουκ εξεϲτιν µοι BB Matt 21:33; p. 1264A B και εξεδ\ο/ετ\ο/ε αυτον… B18 Matt 21:38; p. 1264B B εν εαυτο\ι/ϲ ουτοϲ… B18 Matt 21:38; p. 1264B B [ν] ο κληρο\νο/µοϲ δευτε… B18 Matt 21:46; p. 1264C B … ζητου ̅ τεϲ αυτον ⟦ε⟧κρατηϲα[ι vid B18? Matt 26:59; p. 1273B B … ολον εζη τουν ψευδοµαρτυρ\ι/αν B18 Matt 26:63; p. 1273C B … του θ̅υ̅ του ζω/ν\τοϲ B18 Matt 27:1; p. 1274A B … πρωιαϲ δε γ/εν\οµε νηϲ ϲυµβουλιον… B18 Matt 27:6; p. 1274B B … ειϲ τον κορ βα\να/ν επει τιµη… B18 Matt 27:45; p. 1275C B γην ε\ωϲ/ ωραϲ ενατηϲ… B18? Matt 28:3; p. 1276C B ην δε η ειδε\α/ αυτου… B18 Mark 1:14; p. 1278A B µε\τα/ το παραδοθην… B18 Mark 1:36; p. 1278C B και \οι/ µετ αυτου και… Β18 Mark 1:38; p. 1279A B … και λεγει αυτοι\ϲ/ αγωµεν… B18? Mark 2:3; p. 1279B B … παραλυ τικον αιροµεν\ο/ω̸ν υπο τεϲϲαρων… B18 Mark 2:12; p. 1279C B … και ευθυϲ αρα\ϲ/το ̅ B2 or ΒΒ Mark 3:3; p. 1280B B τω ανθρωπω τω τη ̅ χειραν̇ εχοντι ξηραν B18? Mark 3:4;77 p. 1280B B … και \γ/λ̸ει αυτοιϲ… [λε B2 or ΒΒ Mark 6:1; p. 1284C B … και εξη\λ/θεν εκειθε ̅ B2 or ΒΒ Mark 6:22; p. 1285C B ει\ϲ/ελθουϲηϲ τηϲ θυγα τροϲ αυτου… B18? Mark 7:5; B … αλλα κο\ι/ναιϲ χερ B18 77 Alternatively, the original reading was κα| λει αυτοιϲ, since the iota in και is smaller than usual. In this case, the correction would be καλει → και λεγει with the line-end addition of [ι λε. EARLY CORRECTIONS IN THE NEW TESTAMENT 217 p. 1287B ϲιν εϲθιουϲιν… Mark 7:18; p. 1287C B και υµειϲ αϲυν\ε/τοι εϲτε B2 or Β18? Mark 7:21; p. 1288A B … οι δι\α/λογιϲµοι οι κα κοι εκπορευονται… B2 or ΒΒ Mark 7:23; p. 1288A B ⟦ν⟧vid κοινοι τον ανθρωπο ̅ B2 or Β18 Mark 8:18; p. 1289B B … και ω τα εχοντε\ϲ/ ουκ ακουε τε… B2 or Β4 Luke 2:37; p. 1308A B ουκ αφ⟦ε⟧ιϲτα\το/ του ιερου B2 or ΒΒ Luke 3:17; p. 1309C B … πυρι αϲβε\ϲ/τω… B2 or B18 Luke 5:18; p. 1313A B και ιδου ανδρεϲ ε̸φε ροντεϲ επι… B2 or ΒΒ Luke 5:27; p. 1313B B και µε\τα/ταυτα εξηλθε ̅ B18 Luke 7:24; p. 1317B B καλαµον υπο ανεµου ϲαλευοµεν\ον/ αλλα… B18 Luke 7:34; p. 1317C B ανθρωπου ε\ϲ/θιων και πεινων… B18? Luke 7:35; p. 1317C B … και \ε/δικαιωθη η ϲο φια απο παντων… B18 Luke 8:28; p. 1319C B µη µε βα\ϲα/νιϲηϲ παρηγ γειλεν γαρ… B2 or B18 Luke 9:26; p. 1322A B οϲ \γ/αρ αν επαιϲχυνθη B2 or B18 Luke 11:46; p. 1327B B … οτι φορτιζε\τε/ τουϲ ανθρωπουϲ φορτια B18 Luke 14:3; p. 1331C B λεγων εξεϲτ\ιν//ι\ τω ϲαββα τω θεραπευϲαι… B2 or ΒΒ; B18 Luke 14:10; p. 1332A B ˙/. νωτερο ̅ … προϲανα βηθι α ˙/. τοτε εϲται… B2 or ΒΒ Luke 14:31; p. 1333A B … ει δυνατοϲ ε ϲ\τ/ιν εν δεκα χειλιαϲιν B18 Luke 16:1; p. 1334B B διαϲκορπιζων τα υπ και φωνηϲαϲ αυτον [αρχο ̅ τα αυτου BB or B2 Luke 17:23; p. 1336C B [ωδε]vid υµιν ιδου εκει η ιδου ωδε… B2 or B18 Luke 19:29; p. 1339C B ειϲ ιεροϲολψµα και ε γε\νε/το ωϲ ηγγιϲεν… B18 Luke 19:29; p. 1339C B [καλουµενο ]̅vid το οροϲ το καλουµενο [\ν/ ελαιω ̅ B2 or BC Luke 19:35; p. 1339C B … επι τον πωλον επεβι\βα/ϲαν τον ιν̅̅… B18 Luke 21:12; p. 1342B B επι βα\ϲι/λειϲ και ηγεµο ναϲ… B2 or B18 Luke 22:65; p. 1345B B [ϲ] …πολλα βλα ϲφηµουντεϲ ελεγον B2 or B18 John 1:49; B … ϲυ βαϲιλευ\ϲ/ ει του ιϲ B18 CHAPTER 6 218 p. 1351B ραηλ… John 6:19; p. 1357C B … ουν ω\ϲ/ ϲταδι ουϲ εικοϲι πεντε… B18 John 6:25; p. 1358A B … και µ̸̇η̸̇ ευροντεϲ αυ τον περαν τηϲ θαλαϲ ϲηϲ… B2 and/or ΒΒ John 7:38; p. 1361A B ο πιϲτευων ει\ϲ/ εµε… B18 John 7:43; p. 1361B B … ϲχι\ϲ/µα ουν εγενε το εν τω οχλω… B18 John 9:29; p. 1364B B … οτι µωυϲει λελα [λη κεν ο θ̅ϲ…̅ B18 John 10:10; p. 1365B B … ουκ ερ [ν]vid χεται \ε/ι µη ινα κλεψη BB John 11:11; p. 1366C B … εϲτιν εν αυτω τ̸α̸υ̸ ταυτα ειπεν… B2 or B18 John 13:7; p. 1370C B µε\τα/ ταυτα λεγει αυτω B18 John 15:7; p. 1373A B εαν µ̸̇η̸̇ µεινητε εν ε µοι και τα ρηµατα µου B2 or B18 John 15:9; p. 1373A B µεινατε ε\ν/ τη αγαπη τη εµη… B2 or B18 John 18:17; p. 1376C B … λεγει ου\ν/ τω πετρω… B18 Still, there are some nonsense readings that were likely corrected by scribe B. The line-end addition of µην δεξηται | υ contains the abbreviated -ται, which we have only found in scribe B’s work in Psalms (Chapter 3).78 Albert Clark seems to suggest that this error was caused by homoeoarchon, between µη and [υ]µαϲ, since the exemplar apparently had lines of ten to twelve letters.79 At least two nonsense readings were corrected in scribendo.80 In Matthew 20:15 (p. 1261C), scribe B copied εξι before overwriting the iota with an epsilon to complete the word εξεϲτιν (Figure 112).81 The sigma could not have been added before the correction of ι → ε was made. As we saw in Chapter 5, scribe B made a similar correction in Psalm 32:1 (p. 642B). Likewise, the correction of ιν → ει in John 10:10 (p. 1365B) appears to have been made in scribendo, as the spacing of the subsequent 78 The correction is certainly early, since the final nu of µην was left untouched by the reinker. 79 Clark, The Primitive Text, 33. 80 It likely that early erasures were corrections of nonsense readings, though it is usually unclear what was originally copied. These erasures include: p. 1261Α, Matt 19:20–22; p. 1312B, Luke 5:2; p. 1345B, Luke 22:57; p. 1373C, John 15:26. 81 Tischendorf does not recognize a correction here and therefore omits it from his list of in scribendo corrections. EARLY CORRECTIONS IN THE NEW TESTAMENT 219 text is written as if the original reading was already corrected (Figure 112).82 Presumably, scribe B skipped over the words ει µη and began copying ινα. However, after completing the nu the scribe noticed the error and partially erased it. The addition of a supralinear epsilon before the iota completed the word ει and a mu was copied sharing the final stroke of the original nu. FIGURE 112: IN SCRIBENDO CORRECTION OF NONSENSE READINGS (P. 1261C, MATT 20:15; P. 1365B, JOHN 10:10) Other nonsense corrections may have originated with scribe B, but they lack any telling features. The line-end correction in Luke 16:1 (p. 1334B), τα υπ → τα υπαρχοντα αυτου, looks similar to the line- end addition of διαψαλµα in Psalm 67:4 (p. 664A; Scribe B; Figure 113). This could mean that they were both came from the hand of scribe B or the same diorthotes. Like the correction above, Clark cites this example as evidence for ten to twelve letters-per-line in the exemplar. 83 Unlike the previous correction, the reason for this omission cannot be clearly explained. FIGURE 113: LINE-END CORRECTIONS IN SCRIBE B (P. 664A, PSALM 67:4; P. 1334B, LUKE 16:1) There is, however, one correction that shares characteristics with the hand of scribe C (p. 1339C, Luke 19:29). In the line-end addition of ελαιων, the alpha and lambda exhibit a similar vertical orientation and a hint of the looped apex can be seen behind the reinking of the alpha (Figure 114). It is also distinguished from other line-end corrections, since the original macron has been covered with a supralinear nu. FIGURE 114: LINE-END CORRECTION OF NONSENSE READING (P. 1339C, LUKE 19:29) Two further corrections deserve mention. The first is the early correction of εξηθεν → εξηλθεν (p. 1284C, Mark 6:1), which was accomplished by the supralinear addition of a lambda. The error was 82 Tischendorf, Vaticanum, xxiii. 83 Clark, The Primitive Text, 33. CHAPTER 6 220 possibly caused by the similarity with the following word εκειθεν. On p. 1289B (Mark 8:18) the correction εχοντε → εχοντεϲ was either corrected by an early corrector or Versace’s B4, a ninth- century hand, which copied numerous marginal glosses.84 Although he does not recognize B4 in the main text it appears that this is the same hand that has reinked part of the text (see Chapter 2). In the same column as the correction, this hand retraced portions of sixteen lines in an Upright Ogival Majuscule Style (Figure 115). However, it is more likely that the correction originated with scribe B or a diorthotes and was later retraced by B4, since the sigma is unimodular rather than oval-shaped. FIGURE 115: COMPARISON OF B4 HAND WITH NONSENSE CORRECTION (P. 1289B, MARK 8:18; CF. P. 751B, ECCL 2:11) Like we observed in Psalms, the word βαϲιλευϲ is misspelled twice by scribe B (p. 1342B, Luke 21:12; p. 1351B, John 1:49). In both instances, B18 corrected the error with supralinear additions. The reinker also made the correction κορβαν → κορβαναν (p. 1274B, Matt 27:6), which Pardee mentions as a possible harmonization to Mark 7:11, but rightly concludes is a simple error.85 6.1.3 Summary How do the corrections in the Gospels compare to what we know of scribe B in 1 Kingdoms and Psalms? In all three datasets, corrections of addition are rare. They mark significant cases of dittography in a few places, but minor additions include unwanted articles, conjunctions, and particles. Corrections of omission are more frequent in the Gospels than in 1 Kingdoms, but are similarly common in Psalms. However, only once is the omission large enough to warrant the use of ancorae-lemniskoi (p. 1247C, Matt 10:37). The preceding discussion has also shown that we cannot follow Hill’s suggestion for a new siglum, styled Bmg, both because of the multitude of marginal correction methods and hands, but also the close relationship between marginal and intracolumnar 84 Versace, Marginalia, 23–28, 76, 189–203. 85 Pardee is not sure whether the scribe or a later corrector made the change. Pardee, Harmonization, 246. EARLY CORRECTIONS IN THE NEW TESTAMENT 221 corrections.86 Substitutions constitute one of the largest categories of correction in the Gospels. Scribe B is likely responsible for the S-siglum corrections, but also for a number of substitutions in the main text (eg. p. 1367A, John 11:27). More significant is the suggestion that the S-siglum corrections originated from the margins of the exemplars.87 If this was so, those who prepared the manuscripts from which B(03) was copied considered these readings as genuine alternatives to be included with the text. Finally, we noted that B18 was more likely to identify nonsense readings than the early correctors. The number of these corrections corresponds with the high frequency in Psalms (especially in comparison to scribe C). While this is not a study of singular readings in B(03), we can compare our results with Paulson’s list of singulars in Matthew.88 Among the ninety-five readings he identifies as singular in B(03), there are only eleven corrections, which can confidently be attributed to an early hand.89 This is not surprising, given the majority of identified singulars are nonsense readings. As in Chapter 5, the observation that the earlier correctors rarely amended nonsense readings presents a portrait of the early hands as more concerned with detecting additions, omissions, and, to a lesser extent, substitutions. Furthermore, those singulars which were left uncorrected by early or late hands confirm the conclusions made from the corrections themselves: scribe B often copied the α- conjugation in thematic aorists and the single nu spelling of ιωαννηϲ, while having a tendency to omit more frequently than add.90 6.2 The Acts of the Apostles 6.2.1 The Text of Acts in B(03) While there has been much discussion on B(03)’s text of Acts, it has centered on comparisons with the “Western” or “D-text” of Acts. 91 Since B(03) is seen as the primary representative of the 86 Hill, “Siglum,” 19. 87 The absence of alternative readings with the s-sigla in John 4:42 and 20:22 (pp. 1355A, 1380B) could mean that they originated with the main exemplars of B(03), but the marginal reading was no longer visible or desired. 88 Paulson, Scribal Habits, 155–162. 89 Matthew 5:16; 10:14; 12:32; 12:48; 13:17; 14:13; 16:17; 19:17; 21:4; 26:3; 27:13. An additional nine are identified as early corrections by Paulson, but these are less certain: Matthew 6:32; 12:1; 12:33; 13:15; 13:30; 13:48; 15:32; 21:46; 27:45. 90 Though Paulson notes that the additions constitute more words than the omission in Matthew. Paulson, Scribal Habits, 45. 91 Pisano, “The Text,” 91; There are many criticisms of the D-text concept. With the arrival of the ECM Acts volumes came further challenges to the text-type theory altogether. Georg Gäbel prefers a trajectory model and uses the term “Bezan Trajectory” instead. Gäbel, “‘Western Texts,’” 83–136; cf. Wachtel, “On the Relationship,” 137–148. CHAPTER 6 222 Alexandrian text, and D(05) represents the “Western” text, numerous studies have progressed from a basic comparison of the two codices.92 The superiority of B(03) in Acts was argued for by Ropes on the basis of its relationship to the papyri, the evidence of careful copying, and the absence of Hexaplaric influence in most of the OT.93 Ropes also compared the number of singular readings in B(03) with those found in א(01), A(02), C(04), and 81, through which he found that B(o3) had the least.94 Our earliest reference to the text of Acts in B(03) is found in Erasmus’ Annotations, where he notes a unique reading in B(03)—at the time it was the only witness to the reading καυδα instead of κλαυδα (Acts 27:16).95 At one point, it had been suggested that there was a unique connection between the Ethiopic text and B(03), though James Royse has shown that the agreements are far from unique.96 The publication of ECM Acts provides additional opportunities to discuss the text of Acts in B(03). The greatest benefit for our study is the use of the ECM apparatus for analyzing corrections in the codex.97 Moreover, the CBGM data offers ample evidence for understanding the text of B(03) in relation to the edited text of Acts. In phase 1 of the CBGM, B(03) is shown to agree with the NA28/UBS5 text in 7525/7796 variation units (96.5%), making it the closest witness to the 92 Ropes provided an early edition with the text of both manuscripts on facing pages. Boismard and Lamouille also use B(03) as the primary text to compare against their reconstructed “Western” text, “except when it offers obvious errors or when there is a very high probability that it has suffered the influence of the TO [Occidental Text].” Likewise, Jenny Read-Heimerdinger compares the text-length of both manuscripts as representing the difference between Alexandrian and Western traditions. However, Eldon Epp rightly warns against the assumption that D(05) is “in any fashion coincident with the early text of the D-Cluster [“Western”/D-text] or as a closely approximate representative of the D-Text.” Ropes, Beginnings; Boismard and Lamouille, Text occidental, 1:122; Read-Heimerdinger, “Texts of Acts,” 245– 261; cf. Rius-Camps and Read-Heimerdinger, Message of Acts; Epp, “Text-Critical Witnesses,” 629. 93 He also claimed that there were less idiosyncrasies in B(03) than the rest of the Old Uncial group—his term for the Alexandrian text group. Ropes, Beginnings, cclvi. 94 It should be noted that these “singular” readings are understood differently from current scholarship, which follows and adapts E. C. Colwell’s definition—“readings without other manuscript support.” Ropes counts a total of 223 singulars in B(03), although 133 of these readings find some support from Greek manuscripts outside of the “Old Uncial” group. Ropes, Beginnings, cclvi–cclx; Colwell, “Scribal Habits,” 108; cf. Royse, Scribal Habits, 65–73; Jongkind, Scribal Habits, 140–143; Malik, P. Beatty III (P47), 115–117. 95 This reading is now supported by P74, א(01)C2, and 1175. Jan Krans notes that Erasmus likely knew the text from Sepúlveda’s list of readings. Erasmus, Annotations on Acts 27:16; Krans, “Erasmus,” 462; cf. Elliott, “Eclectic Textual Study,” 12. 96 Royse targets the argument of Montgomery, “Ethiopic Text,” 169–205; Royse, “Ethiopic Support,” 258–262. 97 With some exceptions, the majority of B(03)’s corrections are accounted for in the 7,446 variation units. EARLY CORRECTIONS IN THE NEW TESTAMENT 223 Ausgangstext (A).98 While the ECM volumes reflect a new appreciation for the Byzantine text, the results of phase 3 in the CBGM reveal a slightly higher percentage of agreement between B(03) and the new A (96.7% or 6976/7212 variation units) (Figure 116).99 FIGURE 116: PERCENTAGE OF AGREEMENT BETWEEN B(03) AND A IN ACTS Of the fifty-two departures from the NA28/UBS5 text, ECM Acts prints sixteen new readings in agreement with B(03); thirteen of which also agree with Byz.100 Only once does a change of readings agree with the corrected text of B(03) (Acts 8:31; B18), and in Acts 20:5 the editors of the ECM follow the first-hand reading προϲελθοντεϲ instead of the B18 correction προελθοντεϲ, which is found in NA28/UBS5. Of the 155 split readings found in ECM Acts, three are supported by both the original reading and the corrected reading (Acts 2:7; 9:34; 20:13), while a fourth is split between the corrected reading and alternative reading not found in B(03) (Acts 4:21). 6.2.2 The Earliest Corrections in Acts Addition TABLE 37: CORRECTIONS OF ADDITIONS IN ACTS SCRIBE B(03) CORRECTOR Acts 28:8; p. 1424B B … και προϲ ευξαµενοϲ ❛ευ̇̇ξα̇̇µ̇εν̇̇οϲ̇❜̇ B18 Only one correction marks an occurrence of addition, which resulted in dittography. This copying error was perhaps caused by a homoeoteleuton with the previous word προϲ. Although the correction 98 B(03) is followed by P74 (95.1%) and א(01) (94.7%). In the initial phase, the hypothetical A shares the same text as NA28/UBS5. Wachtel, “Notes,” 28–29; For the number of agreements between A and its closest relatives in phase 1, see http://intf.uni-muenster.de/PreCo/PreCoActs/Acts_Clusters.html 99 Wachtel, “Notes,” 30; For the number of agreements between A and B(03) in phase 3, see http://intf.uni- muenster.de/cbgm/actsPh3/Comp4.html; On the renewed appreciation for the Byzantine text, see Wachtel, “Notes,” 30–31; Wasserman and Gurry, A New Approach, 10–11. 100 Acts 1:15; 2:5, 20; 3:13 (2x); 7:7, 22; 9:21; 10:40; 11:22; 14:3; 16:11; 20:5, 6; 27:8, 23. The siglum Byz represents “all manuscripts agreeing with the majority text at least at 90% of all variant passages in Acts...” Therefore, Byz has become a substitute for the “Majority Text” (𝔐) in the ECM. Wachtel, “Introduction,” 19. 94 .55 95 .18 97 .48 94 .71 97 .5 9 9.0 5 95 .82 98 .04 97 .57 96 .95 96 .97 96 .22 96 .45 97 .69 98 .05 95 .93 96 .6 97 .1 97 .79 97 .56 97 .52 97 .71 93 .18 97 .67 97 .38 94 .14 96 .53 97 .62 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8CHAPTERS B(03) AND A IN A CTS CHAPTER 6 224 could have originated with an early hand, the deletion dots and round hooks appear in a slightly darker ink like to that of B18. Their presence over the second occurrence of ευξαµενοϲ, rather than the first, suggests that they were added after the reinker had already passed over the first occurrence. Omission As expected, corrections of omission appear more frequently than those of addition in Acts. While a number of omissions were corrected by the reinker, at least twelve were likely made by an early corrector. From the seventeen corrections listed below, only four remedy omissions of more than one word. Articles, conjunctions, and pronouns constitute ten of the single-word omissions in Acts; six of which may have been corrected by an early hand. The largest omission is thirty-three characters long and was one of four omissions corrected by B3. TABLE 38: CORRECTIONS OF OMISSIONS IN ACTS SCRIBE B(03) CORRECTOR Acts 1:7; p. 1382B B … ει πεν \ουν/ προϲ αυτου ϲου B18 Acts 1:19; p. 1383A B εκεινο τη \ιδια/ διαλεκτω B18 Acts 2:34; p. 1384C B δε αυτοϲ ειπεν \ο/ κ̅ϲ ̅τω κ̅ω̅ B18 Acts 3:21; p. 1386A B των] ϲτοµατοϲ των αγιων απ αιωνοϲ αυτου… B2 or B18 Acts 4:18; p. 1387A B επι τω ονοµατι \του/ ιυ̅̅… B18 Acts 5:38; p. 1389B B … και \τα/ νυν λεγω υµιν… B2 or BB Acts 7:42; p. 1392A B ˙/. εν τη ερηµω τεϲϲερακοντα ˙/. οικοϲ ιϲραηλ… B3 Acts 8:34; p. 1394B B … ο προφητηϲ λεγει [τουτο περι εαυτου… B2 or BB Acts 9:34; p. 1395C B … αινεα ιαται ϲε ιϲ \ο/ χϲ αναϲτηθι και ϲτρω ϲον ϲεαυτω… B18 Acts 11:24; p. 1399A B … και προϲετεθη οχλοϲ ικανοϲ ˙/. εξηλθεν δε ειϲ ταρϲον… ˙/. τω κ̅ω̅ B2 or BB Acts 14:10; p. 1403A B … και η λατο \ / περιεπατει… B2 or BB Acts 14:26; p. 1403C B … κακειθεν ˙/. ειϲ αντιοχειαν… ˙/. απεπλευϲα ̅ B3 Acts 21:12; p. 1414A B … τοτε απεκρι θη \ο/ παυλοϲ… B2 or B18 EARLY CORRECTIONS IN THE NEW TESTAMENT 225 Acts 23:18; p. 1417C B εχοντα τι λαληϲαι \ϲοι/ επι λαβοµενοϲ… B2 or B18 Acts 23:28; p. 1418A B ˙/. κατηγαγον αυ τον ειϲ το ϲυνε δριον αυτων … δι ην ενεκαλου ̅ αυτω ˙/. ον ευρον … B3 Acts 25:24; p. 1420C B [αυτο ̅] βοωντεϲ µη δειν αυτο [ν ζην B2 or BB Acts 26:16; p. 1421B B ˙/. ϲτηθι αλλα αναϲτηθι ˙ /. επι τουϲ ποδαϲ ϲου…. B3 The omission of των in Acts 3:21 (p. 1386A) was likely amended by an early corrector. Although Tischendorf associated it with the reinker, Ropes and Versace identified the hand as B2.101 In this example, the raised omega seems to reflect an early hand (cf. p. 1399A, Acts 11:24). While it is possible that the article was omitted because of homoeoteleuton, it is also omitted in the first-hand of B(03)’s closest relatives, P74vid, א(01)*, 81, A(02), C(04), 1175, and 1739.102 The line-end addition of τουτο may have been copied by scribe B, but it is not entirely clear (p. 1394B, Acts 8:34). In Acts 11:24 (p. 1399A), scribe B omitted τω κ̅ω̅, which was subsequently corrected with lemniskoi. Versace identifies the corrector as B2, while the rest of the lemniskoi in Acts were copied by B3. The first two omissions of the article ο were clearly made by the reinker (p. 1384C, Acts 2:34; p. 1395C, Acts 9:34), but the third (p. 1414A, Acts 21:12) bears some resemblance with an early hand (Figure 117). FIGURE 117: CORRECTIONS OF THE OMITTED ARTICLE IN ACTS (P. 1384C, ACTS 2:34; P. 1395C, ACTS 9:34; P. 1414A, ACTS 21:12) In Acts 23:18 (p. 1417C), the singular omission of ϲοι may have been influenced by the ending of the preceding word, λαληϲαι. Although the ECM marks this correction as C1, Tischendorf is more hesitant, writing “B3 et iam2 ut vdtr add [ϲ]οι post λαλη[ϲ]αι”.103 The early identification may be correct, but the slightly elongated omicron does not imitate the standard compressed omicron of scribe B or the early diorthotes (Figure 118, cf. 117). 101 B2 was the diorthotes for Ropes. Ropes, Beginnings, 32; Versace, Marginalia, 134. 102 Listed in descending order of agreement. However, א(01) was later corrected to include the article. 103 It is not clear if Tischendorf’s ordering of B3 before B2 indicates more hesitation, since the normal ordering is B2 followed by B3. Tischendorf, Vaticanum, 183. CHAPTER 6 226 FIGURE 118: CORRECTION OF OMISSION (P. 1417C, ACTS 23:18) Similar to the Gospels, B3 was responsible for making four lemniskoi corrections of omission. It is unclear what caused scribe B to omit the phase εν τη ερηµω in Acts 7:42 (p. 1392A). While B3 potentially copied the correction from a second exemplar, the phrase is also present in B(03)’s closest relatives and, therefore, likely in the primary exemplar. The largest omission was apparently caused by a visual slip from αυτω to αυτων (p. 1418A, Acts 23:28),104 while the last omission occurred because of the repetition of ϲτηθι (p. 1421B, Acts 26:16). Substitution At first glance, substitutions constitute an even larger category of early corrections. However, out of the twenty-four corrections listed below, only thirteen can possibly date to the fourth century.105 The majority of substitutions are changes of case, number, tense, or voice. However, unlike in the Gospels, we do not find any completed S-siglum corrections in Acts (cf. p. 1397C, Acts 10:37). Already, we have noticed that two substitutions coincide with split readings in ECM Acts (p. 1383C, Acts 2:7; p. 1412B, Acts 20:13). TABLE 39: CORRECTIONS OF SUBSTITUTIONS IN ACTS SCRIBE B(03) CORRECTOR Acts 2:7; p. 1383C B … ουχι ιδου \α/παν τεϲ ουτοι ειϲιν… BB Acts 3:2; p. 1385B B … υπαρ χων εβαϲταζε\το/ ον ετι θουν καθ ηµεραν… B18? Acts 4:20; p. 1387A B … γαρ ηµειϲ α \ο/ειδ\ο/α µεν και ηκουϲαµεν… B18/B37 Acts 4:21; p. 1387B B … το πωϲ κολαϲω\νται/ϲι̇ν̇̇ αυτουϲ δια τον λαον… B2 or B18 Acts 5:21; p. 1388C B … παραγενο µενο\ϲ/⟦ι⟧ δε ο αρχιερευϲ οι ϲυν αυτω… B2 or BB Acts 7:10; B τον οικον \αυ/τουτ̸̇ο̸ν̸̇̇… BB 104 Ropes, Beginnings, 221. 105 Tischendorf suggests an additional prima manus substitution of απο → ειϲ (p. 1400B, Acts 12:25), but I have not found any remnants of a correction. EARLY CORRECTIONS IN THE NEW TESTAMENT 227 p. 1390C Acts 8:9; p. 1393B B το εθν̇οϲ τηϲ ϲαµαρει αϲ… B2 or BB Acts 10:37; p. 1397C B [κηρυ]vid … µετα το βαπτι s[γµα] ϲµα ο εκηρυξεν… B18 or earlier Acts 12:8; p. 1399C B και υποδ\η/υ̸ϲαι τα ϲανδα λια ϲου… B18 Acts 12:11; p. 1399C B … και ο πετροϲ εν \ε/αυτω γενοµενοϲ B18 Acts 12:13; p. 1400A B προϲ̸η̇λθε παιδιϲκη υ πακουϲαι… B2 or BB Acts 13:38; p. 1402A B οτι δια τουτο\υ/ υµιν… B2 or B18 Acts 13:44; p. 1402B B … τον λογον του \κ/θ̅υ̅ B18 Acts 20:5; p. 1412A B ουτοι δε προϲε̇λθον τεϲ εµενον… B2 or B18 Acts 20:13; p. 1412B B … ηµειϲ δε προϲε̇λθοντεϲ επι το πλοιον… B2 or B18 Acts 20:24; p. 1412C B … εµαυτω \ε/ωϲ τελει ωϲω τον δροµον µου B18 Acts 22:10; p. 1415C B … περι παντ ων εντετα\λ/κται ϲοι B18 Acts 23:7; p. 1417A B … τουτο δε αυτου λα λουντοϲ επ̇επ̇εϲε ϲτα ϲιϲ των φαρειϲαιων… B2 or B18 Acts 25:15; p. 1420A B … ειϲ ι εροϲολυµα ενεφανι ϲθ̇̇η̇ϲαν οι αρχιερειϲ… B2 or B18 Acts 25:25; p. 1420C B … αυτου δε του π̇α̇υ̇ \τ/λ̸ου επικαλεϲαµενου B18 Acts 26:13; p. 1421B B βαϲιλευϲ̸ ̇ουρανοθεν υ περ την λαµπροτητα… B18 Acts 27:14; p. 1422C B … ο καλουµενοϲ ευ [λ] ρ\υ/ακ\λ/υδων… B18 Acts 27:39; p. 1423C B δυναιντο ε\ξ/κϲ̸ωϲαι το πλοιον… B2 or B18 Acts 27:41; p. 1424A B … επ\ω/εκειλαν τη ̅ ναυν… B18 The first substitution was likely made by scribe B, since the alpha shares the same curved cross-bar (Figure 119). The ECM prints a split reading with the corrected reading above and the original CHAPTER 6 228 reading below (Acts 2:7). Since the closest relatives of B(03) read απαντεϲ,106 and the correction appears to be from the hand of scribe B, this was likely the reading in the exemplar. FIGURE 119: CORRECTION OF SUBSTITUTION BY SCRIBE B (P. 1383C, ACTS 2:7; P. 1390C, ACTS 7:10) Likewise, the substitution of τουτον → αυτου (p. 1390C, Acts 7:10) contains an alpha with a similar appearance. This singular reading may have arisen from the repetition of accusatives—ολον τον οικον—combined with the ending of αυτου. Deletion dots alone correct five substitutions and are therefore difficult to attribute to an early hand.107 It is likely that the two corrections of προϲελθοντεϲ → προελθοντεϲ were made by the reinker (p. 1412A, Acts 20:5; p. 1412B, Acts 20:13), since it is more common for B18 to make corrections of the preposition in compound verbs (though see p. 1400A, Acts 12:13). The latter correction is one of the split readings in ECM Acts, with the corrected reading on top and the first-hand below. The deletion dots in Acts 4:21 (p.1387A) were certainly added after the supralinear correction (Figure 120). However, it is unclear whether an early hand made the substitution κολαϲωϲιν → κολαϲωνται. Tischendorf identified the corrector as B3, but the alpha is clearly different from the B18 correction four lines below it. Therefore, it remains a possibility that an early hand added the supralinear letters, while the reinker added the deletion dots (cf. p. 1475B, 1 Cor 15:35) FIGURE 120: SUBSTITUTIONS OF QUESTIONABLE DATE (P. 1387A, ACTS 4:21; P. 1397C, ACTS 10:37) Finally, the substitution κηρυγµα → βαπτιϲµα (p. 1397C, Acts 10:37) was likely added some time after the production of the codex. Versace originally identified a trace of an S-siglum without a 106 P74vid, א(01), A(02), C(04), and 1739. However, 88 and 1175 do read παντεϲ. 107 p. 1393B, Acts 8:9; p. 1412A, Acts 20:5; p. 1412B, Acts 20:13; p. 1417A, Acts 23:7; p. 1420A, Acts 25:15. Although he does not mention the deletion dots in Acts 8:9, Tischendorf identifies two corrections from B2 and two from B3. EARLY CORRECTIONS IN THE NEW TESTAMENT 229 corresponding marginal siglum or variant reading, 108 and believes instead that scribe B or a diorthotes copied the new reading over the original.109 While this is indeed possible, the hand is noticeably different from the earlier scribes (Figure 120). First, the beta, alpha, iota, and sigma transgress the base-line. Even more striking are the serifs on the beta and the extended cross-bar on the pi. These features likely reflect a hand after the production of B(03), but maybe earlier than the first reinking. Orthography The orthographic corrections reflect the normal spelling patterns we have identified with scribe B. The highest frequency of corrections is found with the interchange of ι-ει, followed by the corrections of unassimilated nu. It is possible that the correction δειϲιδαµονεϲτερουϲ → δειϲιδαιµονεϲτερουϲ (p. 1408B, Acts 17:23) was made by an early hand, as the stroke of the under text shares a similar angle to the main text. Proper names are also regularly corrected by the reinker. As in the Gospels, ιωαννηϲ is spelled with a single nu and corrected once, while the geminate -θθ- is corrected to -τθ- in µαθθαιοϲ and µαθθιαϲ. The name ιϲαακ is copied incorrectly with a single alpha once in Acts 7:8 (p. 1390C) and likely corrected by an early hand to the usual spelling.110 Tischendorf also cautiously identifies B2 (ut vdtr) as responsible for the deletion dot and cancelation stroke in the correction µελιτηνη → µελιτη (p. 1424A, Acts 28:1). Transposition There is only one correction of transposition in Acts. This paucity of corrections corresponds to what we have found in the 1 Kingdoms, Psalms, and the Gospels. TABLE 40: CORRECTIONS OF TRANSPOSITIONS IN ACTS SCRIBE B(03) CORRECTOR Acts 21:5; p. 1413C B // / εγενετο εξαρτιϲαι ηµαϲ ταϲ ηµεραϲ B2 or Β18 108 Hill is less certain of the symbol. It is certainly irregular, since the S-siglum is not found above the initial portion of the word, but above the gamma and after a line-break. Versace, Marginalia, 89; Hill, “Siglum,” 20 n. 95. 109 Versace, Marginalia, 89 n. 62. 110 The supralinear alpha occurs in the same column as the correction τουτον → αυτου (Acts 7:10), which we have already associated with an early corrector. CHAPTER 6 230 The sequential, diagonal bars used to transpose εξαρτιϲαι and ηµαϲ may have been added early, as Tischendorf suggests (B2 nisi forte est B*). However, they are not copied like those bars which Versace claims are early (e.g., p. 127B). 111 The other examples of transposition bars were copied diagonally from the right, descending to the left. In this correction, the strokes appear to be copied in the other direction.112 It is more likely that the contrast in ink with the late reinforcement (B37) influenced Tischendorf’s opinion that the correction was executed in the fourth century. Therefore, I am more cautious than the editors of the ECM, who cite 03C1 as supporting the initial text. FIGURE 121: CORRECTIONS OF TRANSPOSITION (P. 127B, LEV 21:18; P. 1413C, ACTS 21:5) Nonsense As expected, the largest set of corrections, apart from orthographic changes, rectify nonsense readings.113 However, we have already observed in 1 Kingdoms, Psalms, and the Gospels, that the majority of nonsense readings are corrected by the reinker. Of these twenty-three corrections, eleven involve only a single letter, and nine amend errors of two letters. The omission of five letters created the largest nonsense reading, which was subsequently corrected by an early hand (p. 1403B, Acts 14:21). TABLE 41: CORRECTIONS OF NONSENSE READINGS IN ACTS SCRIBE B(03) CORRECTOR Acts 1:11; p. 1382C B ειϲ τ\ον/ ουρανον… Β18 Acts 2:36; p. 1384C B … αϲφα λωϲ ου\ν/ γεινωϲκετω παϲ Β18 Acts 4:32; p. 1387C B … και ουδε ειϲ τι των υπαρχοντω ̅ αυτω ελεγ\ε/ον ιδιον ειναι Β18 Acts 6:14; p. 1390B B … και αλλα ξει τα εθν̸̇η α παρεδω κεν ηµιν… Β18 Acts 7:32; p. 1391C B … εν τροµοϲ δε γε\νοµε/νοϲ µωυϲηϲ… Β18 111 Versace, Marginalia, 16. 112 This is similar to another example, which Versace suggests originated with the reinker (p. 1163B, Ezek 19·9). 113 The table below does not include the correction on p. 1386B (Acts 3:24), which was clearly made by the reinker or an even later hand. This correction is interesting, however, as the ending of προφηται seems to have been conformed to the plural article οι, resulting in προφητοι. EARLY CORRECTIONS IN THE NEW TESTAMENT 231 Acts 9:1; p. 1394B B … ο δε ϲαυλοϲ \ε/οτι εµπνεων απειληϲ Β2 or Β18 Acts 9:2; p. 1394C B l. 42 … προϲ ταϲ || l. 1 ⟦υ⟧ϲυναγωγαϲ… ΒB or B18 Acts 10:36; p. 1397C B … ευαγγελιζοµε νοϲ ειρη\νη/ν δια ιυ̅̅ χ̅υ̅… Β18 Acts 11:25; p. 1399A B … ειϲ ταρϲον ανα\ζη/ϲτηϲαι ϲαυ λον… Β18 Acts 13:1; p. 1400B B ηρωδου του \τε/τραρχου Β18 or 9th ce? Acts 13:13; p. 1401Α B … αν\α/ε̸ χθεντεϲ δε απο τηϲ πα φου… Β2 or ΒB Acts 13:46; p. 1402Β B … κρεινετε εαυτο\υ/ϲ Β2 or Β18 Acts 14:21; p. 1403B B εκεινην και µαθητευϲ ικανουϲ υπεϲτρεψαν [αντεϲ Β2 or ΒΒ Acts 15:1; p. 1403C B οτι εαν µη περι\τ/θ̸µη\θη/τε Β18 Acts 15:38; p. 1405Β B … και µη ϲυνελθο\ν/τα αυτοιϲ ειϲ το εργον… Β2 or ΒΒ Acts 17:23; p. 1408Β B … εγω κα ταγγελλω υµ\ι/ω̸ν… Β18 Acts 18:1; p. 1408C B µε\τα/ ταυτα χωριϲθειϲ… Β18 Acts 20:13; p. 1412B B … τον παυλον ουτωϲ γαρ διατεταγµενο\ϲ/ν η ̅ Β18 Acts 20:16; p. 1412B B … κεκρ\ικ/ει γαρ ο παυ λοϲ… Β2 or Β18 Acts 21:28; p. 1414C B κεκοινω\νη/κεν τον αγι ον τοπον τουτον… Β18 Acts 23:21; p. 1417C B … και νυν ε\ι/ϲιν ετοιµοι… Β2 or Β18 Acts 23:24; p. 1418A B … τον παυλον δι αϲω\ϲω/ϲι προϲ φηλικα… Β18 Acts 27:27; p. 1423Β B οι ναυται προϲα\νε/χειν Β18 Only eight of the nonsense readings reflect any of the distinguishable features of the early hands. The correction οτι → ετι (p. 1394B, Acts 9:1), through the supralinear addition of an epsilon, contains traces of the undertext. Whether the lighter ink is from B18 or an early hand is not clear, but the shape of the epsilon appears to follow the unimodular form of the fourth century (Figure 122).114 114 Tischendorf identifies the hand as B3, while Rope associates it with the scriptorium hand B2. The corrections on p. 1412B (Acts 20:16) and 1417C (Acts 23:21) also bear some resemblance to the early hands, but there is no remnant of undertext and the letters are too simple to distinguish. CHAPTER 6 232 FIGURE 122: CORRECTIONS OF NONSENSE READINGS IN ACTS 9 (P. 1394B/C) On the following column, we find another correction of a nonsense reading (p. 1394C, Acts 9:2). This time, the error seems to have been caused by the repetition of sigma in ταϲ ϲυναγωγαϲ and the transition between two columns (Figure 122). Since the upsilon at the beginning of column C was erased, Tischendorf associates the correction with the reinker. This would result in the first-hand reading ταϲ υϲυναγωγαϲ. However, it could be argued that the correction was made in scribendo. If so, scribe B noticed the error after copying the upsilon and began rewriting ϲυναγωγαϲ immediately after. On his account, Tischendorf does not explain what would have led the scribe to rewrite the word, if the error was not recognized from the beginning. Still, one would have expected scribe B to overwrite the first upsilon as we have seen in other in scribendo corrections. The supralinear addition of an alpha in the correction ανεχθεντεϲ → αναχθεντεϲ (p. 1401A, Acts 13:13) was likely supplied by an early corrector.115 If the curved cross-bar in the reinking is faithful to the original shape, it may even reflect the hand of scribe B. Tischendorf classifies the correction of εαυτοϲ → εαυτουϲ as B2 (p. 1402B, Acts 13:46). Here, the missing upsilon is added above the letters with the stem extending between the appropriate letters (Figure 123). Surprisingly, a similar correction is made on the previous column, which Tischendorf identifies as B3 (cf. above, p. 1402A, Acts 13:38). It is unclear how Tischendorf differentiated the two, apart from the difference in the late reinker’s handling of the correction. Therefore, it is possible that both are early corrections, though the narrow angle of the cone in both upsilons fits better with B18. FIGURE 123: CORRECTIONS WITH COMPRESSED UPSILONS (P. 1402A/B) 6.2.3 Summary There is little in the corrections of Acts that differs from what we have found in the Greek OT and the Gospels. A number of early corrections are particularly important for understanding the text of B(03). For example, it is likely that scribe B corrected the reading from παντεϲ to απαντεϲ (p. 1383C, Acts 2:7), because that was the original reading in the exemplar. This is the reading found in nearly 115 contra Tischendorf (B3). EARLY CORRECTIONS IN THE NEW TESTAMENT 233 all of B(03)’s closest relatives, and we can see a particular habit of omitting single letters in the corrections of nonsense readings. Likewise, the correction of the singular reading τουτον to αυτου (p. 1390C, Acts 7:10) was potentially executed by the scribe. The surround text of the first-hand reading provides an explanation for its origin with scribe B, rather than in the exemplar. While we are used to seeing evidence of B3 corrections of substitutions, likely from a second exemplar, Acts only provides clear evidence of marginal corrections of omission by B3. These corrections, often of singular omissions, do not provide any substantial indication of the nature of this second exemplar. 6.3 The Catholic Epistles 6.3.1 The Text of the Catholic Epistles in B(03) As with our study of Acts, we are fortunate to have the ECM2 edition of the Catholic Epistles at our disposal. 116 Since the changes to the text already occurred in the NA28, we will compare the ECM2/NA28 with the NA27 and how the new decisions reflect or oppose the text of B(03) in the Catholic Epistles. Since the NA28 does not print split readings—diamond readings are used instead—the agreement with the B(03) differs slightly from what is printed in the ECM2. The NA28 differs from the NA27 in twenty-four variant units, five of which are split readings in ECM2.117 FIGURE 124: COMPARISON OF CHANGES IN NA28 WITH B(03) There are a total of eleven changes towards the text of B(03), and one towards a correction by B18 (2 Pt 2:15), while twenty-one changes diverge from the codex. From these new agreements, B(03) and the Byzantine text only share readings twice (2 Pt 2:20; 2 John 5). It should also be noted that seven of the new agreements were already bracketed readings in NA27 and were subsequently omitted in NA28, following B(03). 116 The first installment of the ECM for the Catholic Epistles was re-done after significant developments in the CBGM. Our analysis comes from the second edition (ECM2), which was published in 2013 and corrected in 2014. On these developments, see ECM2 IV/1, 31–32*; Gurry, A Critical Examination, 17–21. 117 The ECM2 does not list the change from αλλα → αλλ in 1 Peter 2:25. ECM2 also includes the split reading in James 1:22, but the top reading, the reading in NA28, is the same as NA27. 5 9 10 4 2 1 31 3 2 1 1 0 3 J A S 1 P E T 2 P E T 1 J O H N 2 J O H N 3 J O H N J U D E NA28 ≠ NA27 B(03) + NA28 ≠ NA27 CHAPTER 6 234 Having taken these changes into account, we still find a very high percentage of agreement between B(03) and A (Figure 125).118 The alterations made between ECM1 and ECM2 resulted in a slight drop from 96.86% (2803/2894) to 96.00% (2852/2971) agreement.119 Yet, this total percentage still lands B(03) in the position of closest descendant to A in the Catholic Epistles. It is only in 2 Peter and Jude, where we find another witness that is closer to A than B(03). In 2 Peter, GA 665 is one one-thousandth of a percent closer to A (95.814%; B(03) = 95.813%).120 However, 665 contains just over half of the comparable variant passages found in B(03), rendering this slight advantage inconclusive. In Jude, GA 81 is closer to A by 1.005%. Although the ECM1 understands this as an indication that B(03) “loses its exceptional position in Jd,” Tommy Wasserman has rightly criticized the overconfidence of this statement.121 The relatively few shared variation units (only 199) in Jude results in a larger divide between B(03) and 81. However, the latter witness only agrees with A at two more units of variation than B(03).122 FIGURE 125: PERCENTAGE OF AGREEMENT BETWEEN B(03) AND A IN THE CATHOLIC EPISTLES123 Sakae Kubo produced a similar study to Martini’s, having investigated the relationship between B(03) and P72 in 1–2 Peter and Jude.124 Out of the 337 compared readings (1 Pt–174, 2 Pt–111, and Jude–52), Kubo identified 95 variants: 60 prior readings in P72, 34 in B(03), and one “neither."125 He 118 See also Terrance Callan’s statement, “There are so few distinctive features of 2 Peter in Vaticanus that reading it differs very little from reading the probable original text.” Callan, “Earliest Copies,” 47. 119 For a comparison of the data, see Genealogical Queries Versions 1.0 and 2.0: http://intf.uni- muenster.de/cbgm/index_en.html. 120 Consequently, if we exclude A, this makes 665 the closest relative to B(03) in 222 units of variation (91.89%). The CBGM considers 665 the only other potential ancestor of B(03) on account of a single extra prior variant in 665. 121 “Notes on the Reconstruction,” in ECM1 IV/1.4, 36; Wasserman, Jude, 25. 122 81 agrees with A in 191/199 passages (95.980%), while B(03) agrees in 189/199 passages (94.975%). 123 Agreement in variant units for each book: James–715/740; 1 Peter–648/685; 2 Peter–389/406; 1 John–719/744; 2 John–100/102; 3 John–92/95; Jude 189/199; Total–2852/2971. 124 Kubo, P72 and Vaticanus. 125 Kubo, P72 and Vaticanus, 24, 150. 96.62 94.60 95.81 96.64 98.04 96.84 94.98 96.00 J A S 1 P T 2 P T 1 J N 2 J N 3 J N J D T O T A L B(03) AND A IN THE C ATHOLIC E PISTLES EARLY CORRECTIONS IN THE NEW TESTAMENT 235 concludes that the common text of P72 and B(03) “is almost always superior to any other opposing combinations,” but also that B(03) is “not so free of ‘improvements’ of the primitive text.”126 As a result of this conclusion, Kubo generally prefers the text of P72 in places where the two manuscripts disagree. The CBGM data behind the ECM2 certainly confirms that, apart from A, B(03) is the closest relative of P72.127 However, the opposite cannot be said, as P72 is the 87th relative of B(03) in 1 Peter, 2nd in 2 Peter, and 116th in Jude. 128 Contrary to Kubo’s argument, the text-critical decisions undergirding the CBGM reveal that B(03) generally contains the prior readings in places of variation (Table 42). TABLE 42: PRIOR AND POSTERIOR READINGS IN B(03) AND P72 (CBGM)129 ≠ B(03) > P72 B(03) < P72 UNCLEAR NO RELATION 1 Peter 121 80 22 13 6 2 Peter 38? 25 7 7 1 Jude 36 24 5 5 2 Kubo concedes this possibility when he admits, “one’s methodology and criterion in selecting a genuine reading will determine the final results.” 130 It appears that the much higher rate of comparison between the two witnesses in the CBGM allows for such modification. 6.3.2 The Earliest Corrections in the Catholic Epistles Addition There are only two corrections of addition, which may have originated with an early hand.131 Unlike the additions in other books, neither of these two resulted in dittography. TABLE 43: CORRECTIONS OF ADDITIONS IN THE CATHOLIC EPISTLES SCRIBE B(03) CORRECTOR James 1:3; p. 1426A B δοκιµιον υµων ❛τηϲ❜ ❛πιϲτεωϲ❜ κατεργαζε ται… Β2 or Β18 1 Peter 4:18; p. 1433C B και ει ο δικαιοϲ µολιϲ ϲωζεται ο δ̸̇ε̸̇ αϲεβηϲ Β2 or ΒB 126 Kubo, P72 and Vaticanus, 152, 154. 127 1 Pt–564/685 (82.34%); 2 Pt–378/416 (90.87%); Jd–153/189 (80.95%). However, it should be noted that 81 shares the same number of agreements with P72 in Jude, and A(02) is equally close in 1 Peter (568/690; 82.32%). 128 665 agrees with B(03) 91.89% (204/222), while P72 agrees 90.87% (378/416). The significant difference in units of variation weakens the comparison. 129 Results taken from http://intf.uni-muenster.de/cbgm2/PotAnc5.html. 130 Kubo, P72 and Vaticanus, 152. 131 This does not include the erasure on p. 1444A (Jude 9), which was executed after the reinking. CHAPTER 6 236 The deletion of τηϲ πιϲτεωϲ in James 1:3 (p. 1426A) may have originated with the reinker (Tisch: B3; ECM2: C2), but it is also possible that the round hooks were reinforced after an earlier corrector added them (NTVMR: C1). B(03)’s ten closest relatives all share the first-hand reading, while the ECM cites only Didymus and the Latin (F text-type) as supporting the corrected reading. The addition of δε in 1 Peter 4:18 (p. 1433C) was likely corrected by an early hand, using deletion dots and cancelation strokes. The closest relative of B(03), which gives the first hand-reading, is 0142 (70th in number of agreements). Omission Compared to the Gospels and Acts, there are very few corrections of omission. However, the numbers are more analogous to the corrections in 1 Kingdoms (only four). Here, all three omissions can be explained by homoeoteleuton, and only one of the corrections could conceivably have been made by scribe B (p. 1441B, 1 John 4:21). TABLE 44: CORRECTIONS OF OMISSIONS IN THE CATHOLIC EPISTLES SCRIBE B(03) CORRECTOR James 2:4; p. 1427A B µου \ου/ διεκριθητε εν ε αυτοιϲ… Β18 1 Peter 1:1; p. 1430B B βιθυ νιαϲ … αϲιαϲ ˙/. κατα προγνωϲιν θ̅υ̅ πατροϲ Β3 1 John 4:21; p. 1441B B απ αυτου ινα ο αγαπω ̅ τον αδελφον αυτου [το ̅θν̅ ̅ αγαπα Β2 or ΒB The first omission was almost certainly amended by the reinker, since the omicron is no smaller than the upsilon, and the stem of the latter does not extend far from the base of the writing (Figure 126). Since the ECM2 marks the corrected reading with a “Z,” for addition, they believe it is not clearly a correction or an alternative reading.132 Klaus Wachtel describes the original reading as still standing in conformity with the context.133 However, given scribe B’s tendency to omit small words because of homoeoteleuton, the previous word µου may be the more likely cause of the omission.134 132 “Introduction,” in ECM2 IV/1, 29. 133 Wachtel, Der byzantinische Text, 223. 134 See corrections of omitted ου in Matthew 1:25 (p. 1236A) and Luke 14:27 (p. 1332C). The only close relative, which omits the negative particle, is 1852 (6th). The remaining ten closest relatives agree with the corrected reading of B(03). EARLY CORRECTIONS IN THE NEW TESTAMENT 237 FIGURE 126: CORRECTIONS OF OMISSIONS IN THE CATHOLIC EPISTLES (PP. 1427A, 1430B, 1441B) The omission of και βιθυνιαϲ (p. 1430B, 1 Pet 1:1) appears to be a singular reading with partial support from 1175, which omits both αϲιαϲ and βιθυνιαϲ. Versace has correctly associated the marginal correction with B3.135 Although there is a supralinear lemniskos in the column, the corresponding marginal siglum is either missing or covered by the late ornamentation (Figure 126). Again, the best explanation for this omission is homoeoteleuton with the repeated ending -ιαϲ: γαλατιαϲ, καπαδοκιαϲ, and αϲιαϲ. It is noteworthy that scribe B listed the three locations with small spaces between each placename. The fact that the omission occurred despite this unique spacing may suggest that it was not so in the exemplar, but a novelty of scribe B. The final omission in 1 John 4:21 (p. 1441B) was likely corrected by scribe B or a diorthotes (Figure 126). The last two alphas in this line-end correction appear to have curved cross-bars, similar to those of scribe B. However, this feature could derive more from the reinker than the original hand. The ECM2 cites this as a lacuna, but with a first-hand vid reading: 03C1(*V). Indeed, it is likely that the error arose from homoeoteleuton with τον αδελφον.136 Substitution There are five corrections of substitution in the Catholic Epistles; only one of which could have originated with an early hand (p. 1435C, 2 Pet 2:1). All of these corrections involve additions, omissions, or substitutions of individual letters, rather than complete words. TABLE 45: CORRECTIONS OF SUBSTITUTIONS IN THE CATHOLIC EPISTLES SCRIBE B(03) CORRECTOR James 2:3; p. 1427A B … ϲυ ϲτηθι η καθου ε κει \ε/υ̇π\ι/ο ̇το υποποδιο ̅ Β18 James 5:4; p. 1429B B … υµων ο α\π/φ\ε/υϲτε ρηµενοϲ αφ υµων Β 18 2 Peter 2:1; p. 1435C B … επαγον τεϲ \ε/αυτοιϲ ταχινην Β2 or Β18 1 John 5:20; p. 1442A B … ι να γεινωϲκ\ω/οµεν το ̅ αληθινον… Β18 3 John 4; B … µειζοτεραν του Β18 135 Versace, Marginalia, 135. 136 Without the final nu the omission fits the 10–12-character line-lengths, which Clark proposed in the exemplar of B(03). Clark, The Primitive Text, 33. CHAPTER 6 238 p. 1443A των ουκ εχων̸ χαριν ινα ακουω… The first correction covers the substitution υπο → επι by adding a supralinear epsilon and iota (p. 1427A, Jas 2:3). While deletion dots are present above the original upsilon and omicron, the latter dot is off-centered, indicating that the supralinear iota was added first. This could mean that an early hand made the correction, but the supralinear epsilon betrays the hand of B18. The correction αφυϲτερηµενοϲ → απεϲτερηµενοϲ was also made by B18, using supralinear letters (p. 1429B, Jas 5:4). Both readings are printed on split lines in ECM2, with the corrected reading on top. The last two corrections of substitution also appear to come from B18. The change from the indicative to subjunctive, γεινωϲκωµεν (1442A, 1 John 5:20), could also be understood as an orthographic correction of ο → ω (see below) or a nonsense reading.137 The singular reading εχων was likely corrected to εχω by the reinker, by using a cancelation stroke over the nu. It is possible that scribe B copied the nu because of the previous word τουτων. Finally, the correction of αυτοιϲ → εαυτοιϲ (p. 1435C, 2 Pet 2:1) may have been made by an early hand. At first glance, the epsilon follows the shape of those made by B18. However, the remains of the undertext bear some resemblance with the early hands (Figure 127). In support of an early corrector, we find that the majority of B(03)’s closest relatives read εαυτοιϲ, while the nearest descendant with the first-hand reading is 1243 (28th in agreements). The accidental omission of single letters has already proven to be the most common error in the nonsense readings of scribe B. FIGURE 127: B2 OR B18 CORRECTION OF SUBSTITUTION (P. 1435C, 2 PET 2:1) Orthography The orthographic corrections reveal similar spelling patterns as in the Gospels and Acts. The two most common interchanges are of ε-αι and ι-ει, while the correction ο → ω is relatively more frequent than in the previous books.138 Interestingly, the opposite correction ω → ο has only occurred once so 137 On the present indicative following ινα, see BDF §369; cf. Man, “Textual Significance,” 90. 138 See also the correction γεινωϲκοµεν → γεινωϲκωµεν in the previous section (1442A, 1 John 5:20). EARLY CORRECTIONS IN THE NEW TESTAMENT 239 far in the NT. In contrast to the previous books, there are far fewer corrections of the unassimilated nu. Tischendorf and the NTVMR identify one of the two corrections of aspirated ουχ as coming from an early hand, possibly even scribe B (p. 1428C, Jas 4:2). Here, what was apparently a chi has been erased with a kappa reinscribed. While it is certainly possible that the kappa came from an early hand, it is noticeably narrower than the unimodular kappa of the Biblical Majuscule. If an early corrector had added the letter, we might expect the scribe to have adequately filled the available space with a standard kappa. The one orthographic correction of a proper name attests to the normal spelling practice of using a theta in place of the standard tau in Semitic names (cf. γεννηϲαρεθ → γεννηϲαρετ; §6.1.2). In Jude 14 (p. 1444B), the verb επροφητευϲεν is corrected by B18 to the double augmented form επροεφητευϲεν, through the use of a supralinear epsilon. This correction occurs five times in the Gospels, except with the prefixed augment left untouched. In Jude 14, the initial augment was reinked, but subsequently deleted with double deletion dots (possibly from B37). Nonsense Much like the previous sections, corrections of nonsense readings comprise the majority of corrections, apart from orthography. Of the nine corrections, only two could have been made by an early hand. All nine corrections involve no more than two letters. This conforms to the pattern of nonsense corrections in the Gospels and Acts. TABLE 46: CORRECTIONS OF NONSENSE READINGS IN THE CATHOLIC EPISTLES SCRIBE B(03) CORRECTOR James 3:6; p. 1428A B … η γλωϲ ϲα καθιϲ\τα/ται εν τοιϲ µελεϲιν ηµων… Β18 James 5:3; p. 1429B B … ωϲ πυρ εθ\ηϲ/αυ ριϲατε εν εϲχαταιϲ… Β18 James 5:17; p. 1430A B … επι τηϲ γηϲ ενιαυ του/ϲ\ τρειϲ… Β18 1 Peter 4:3; p. 1433B B αϲελγειαιϲ επιθυµι αιϲ οινοφλυγι\α/ο̸ιϲ κω µοιϲ… Β2 or Β18 2 Peter 2:18; p. 1436B B υπερογκα γαρ µαταιο [τη τηϲ φθεγγοµενοι… Β18 1 John 5:6; p. 1441B B το πνευµα \εϲ/τιν̣ το µαρ τυρουν οτι το πνευµ̣α̣ εϲτιν η αληθεια… Β2 or Β18 2 John 12; p. 1442C B και ϲτοµα προ\ϲ/ ϲτοµα Β18 3 John 6; B … ου\ϲ/ καλωϲ ποι Β18 CHAPTER 6 240 p. 1443A ηϲειϲ… 3 John 14; p. 1443B B … και ϲτο µα προ\ϲ/ ϲτοµα… Β18 The first correction, which may have originated with an early corrector, involves the supralinear addition of an alpha and a cancelation stroke to correct οινοφλυγιοιϲ → οινοφλυγιαιϲ (p. 1433B, 1 Pet 4:3). The error was likely caused by assimilation to the ending of the following word κωµοιϲ, even though the previous word ends in -αιϲ. While there is little remaining from the undertext, the overall shape of the alpha betrays an early hand (Figure 128). Likewise, the cancelation stroke is noticeably thicker than those which appear to have come from B18. FIGURE 128: CORRECTION OF NONSENSE READING IN THE CATHOLIC EPISTLES (P. 1433B, 1 PET 4:3; P. 1441B, 1 JOHN 5:6) In 1 John 5:6 (p. 1441B), the nonsense reading τιν is corrected to εϲτιν by what appears to be an early hand. Unfortunately, damage to the parchment has further complicated identification of the corrector. However, both the epsilon and sigma bear resemblance with the early corrections, including the fact that the latter is more raised and compressed than the former. Still, we cannot be fully confident of this identification. Two similar errors occur in James 3:6 (p. 1428A) and 2 Peter 2:18 (p. 1436B). The two instances involve the omission of the penultimate syllable with the same letters as the final syllable: καθιϲταται and µαταιοτητηϲ. This repetition of letters was likely the cause of scribe B’s error. The reinker was responsible for copying the omitted letters, first supralinearly and then at a line-end.139 The latter omission may have also been influenced by the line break. Finally, the idiom ϲτοµα προϲ ϲτοµα is miscopied as ϲτοµα προ ϲτοµα in both occurrences (2 John 12; 3 John 14). In the previous sections we have already noted that many of the nonsense readings are the result of single-letter omissions, especially when followed by the same letter. As with the majority of nonsense corrections, the sigmas were later added supralinearly by B18. 6.3.3 Summary 139 Versace identifies the line-end correction (p. 1436B, 1 Pet 2:18) as coming from B18. Versace, Marginalia, 264. EARLY CORRECTIONS IN THE NEW TESTAMENT 241 Although there are far fewer corrections in the Catholic Epistles than in the previous sections, this is due to the brevity of the collection.140 Nonetheless, these few corrections confirm what we have learned about corrections in the other work of scribe B. That there is only one more correction of omission than addition is slightly irregular. However, we also noted that the two corrections of addition were not a response to the common occurrence of dittography. In this way, we see the preservation of normal patterns of correction in the Catholic Epistles. Apart from orthography, nonsense readings, followed by substitutions, prompt the majority of corrections. However, B18 is responsible for nearly all of the corrections of nonsense readings and substitutions. Lastly, the paucity of transpositions in the Gospels and Acts is supported by the total absence of such corrections in the Catholic Epistles. 6.4 The Corpus Paulinum What remains of the NT in B(03) ends with the Pauline corpus. As discussed in the Introduction, the text no longer contains Philemon or the Pastoral epistles, breaking off after Hebrews 9:14. That there is an early, consecutive numbering system throughout the letters and that they begin on a new quire (see Chapters 2 and 3), confirms the letters, including Hebrews, should be treated together on internal grounds.141 6.4.1 The Text of the Corpus Paulinum in B(03) It is generally accepted that B(03) is a main representative of the “Alexandrian” text in the Pauline corpus (see 6.1.1).142 In Gunther Zuntz’s programmatic study of the epistles, he further categorizes B(03) as a member of the “proto-Alexandrian” group, along with P46, 1739, the Sahidic, and the Bohairic, supported by Clement and Origen.143 However, it is also well known that the codex has a 140 The Catholic Epistles comprise nineteen pages in B(03), while the Gospels cover 148 pages. 141 For the argument that B(03) was produced from a copy of a two-volume edition, see Stevens, “The Two- Volume Archetype,” 102–126; cf. Chapter 3. I am not interested here in B(03)’s place in the discussions concerning the development of the Pauline collection. See, for example, Trobisch, Letter Collection; Gamble, “Pauline Corpus.” 142 Royse, “Text of Paul,” 178; Pisano, “The Text,” 94; Jongkind, “Pauline Corpus,” 221; Even those who reject text- type theories agree that B(03) “is one of the most valuable witnesses to the text where it is present.” Parker, An Introduction, 257; cf. Henry Sanders’ warning against dependence on B(03) “as a sufficient authority for the Alexandrian text, when all other members of the group are opposed.” Sanders, Τhird-century papyrus, 26. 143 Metzger also included the distinction between proto and later Alexandrian groups in his introduction to the NT text. However, unlike Zuntz, he included א(01) in the proto-Alexandrian group. This division of the Alexandrian witnesses was removed in the fourth edition by Ehrman. Zuntz, Epistles, 156, 241; Metzger, The Text, 216; cf. Metzger and Ehrman, The Text, 312–313. CHAPTER 6 242 “distinctly Western element” in the Pauline corpus.144 Hort regularly mentions this feature in B(03), calling it an “exceptional intrusion of an alien element into the Pauline text of B.”145 This negative assessment of certain readings can also be seen in Metzger’s Textual Commentary, where it is said that B(03) “not infrequently displays a strand of Western contamination,” and in those places “should not be overevaluated.”146 On the other hand, Zuntz highlights the “Western” readings in B(03) and P46 as ancient and pre-Alexandrian, rather than a corruption from “secondary witnesses.” 147 These readings disappeared from the “later Alexandrian” tradition (א(01), A(02), C(04), 33) and reappeared in both Eastern and Western witnesses. In his study on the text of Galatians, Stephen Carlson is “even more optimistic about the value of “Western” readings when they are supported by both P46 and B.”148 For our purpose, it is important to note that very few corrections occur in the readings identified as “Western,” with support from B(03) and P46. Between 1 Corinthians and Hebrews, Zuntz lists twenty-six readings in which B(03) agrees with Western readings, whether pure “Western”, “Western-plus,” or “Western-plus-Byzantine.”149 In only three instances do we find a correction. The first two corrections move away from the “Western” reading (1 Cor 1:6; 9:9), while the last one brings B(03) into conformity with “Western” witnesses and P46 (1 Cor 14:39). Likewise, Carlson lists seven important “Western” readings with support from P46 and B(03), yet none of these align with corrections in the codex (see n. 143). The “Western” readings in B(03) will continue to be evaluated as the ECM editions for the Pauline corpus are produced. However, Grant Edwards’ recent PhD thesis on the text of 2 144 Metzger, Manuscripts, 74; cf. Sanday and Headlam, Romans, lxxii; Nongbri, “Pauline Letter Manuscripts,” 90. 145 However, Hort uses this feature in Paul to highlight the relative absence of the “Western” influence in the rest of B(03). Westcott and Hort, Introduction, 228–229, cf. 105, 165–166, 210, 240. 146 In other units of variation, B(03) is cited as containing “a Western expansion” (Rom 14:21) and a “predominantly Western reading” (2 Thess 3:6), as well as accompanying the “chiefly Western” witnesses (Rom 15:31). Metzger, Textual Commentary, 542. 147 Zuntz, Epistles, 96, 142, 156–57; cf. Holmes, “Sixty Years After,” 110. 148 As examples, he includes the readings in Galatians 1:3, 1:17, 1:18, 2:20, 4:3, 4:19, and 6:2. Carlson, Galatians, 248. 149 Western-plus readings are those which have “non-Western ‘side-attestation’,” while Western-plus- Byzantine readings are the Majority Text readings, when opposed by the “non-Western Old Uncials.” Zuntz, Epistles, 85, 96, 124. EARLY CORRECTIONS IN THE NEW TESTAMENT 243 Thessalonians affirms the primary importance of B(03).150 This does not mean the codex will be consistent throughout the corpus, as Michael Dormandy has shown the quality of its text in Romans is slightly inferior to א(01), C(04), and even A(02). 151 It is inevitable that the CBGM, with its reinvigorated appreciation for the Byzantine text, will come to slightly different conclusions than Edwards and Dormandy. However, we have already seen that in both Acts and the Catholic Epistles, this approach has affirmed many of the earlier conclusions concerning the text of B(03). Since the ECM volumes are not yet available, the TuT volumes remain the most important tool for understanding the relationship between B(03), the NA28 text, and other witnesses to the Pauline corpus (Table 47).152 While the percentage of agreement with the NA28 text is noticeably smaller than that of ECM Acts and Catholic Epistles, it does not vary significantly from the TuT results in the Gospels. One will also notice that the number of Teststellen decreases considerably after 1 Corinthians, affecting the rate of agreement in books like 2 Thessalonians.153 TABLE 47: TEXT UND TEXTWERT: B(03), NA28, AND THE MAJORITY TEXT OF THE CORPUS PAULINUM154 B(03)–NA28 B(03)–Majority Text Romans 36/47 (76.5%) 12/47 (25.5%) 1 Corinthians 49/59 (83%) 13/59 (22%) 2 Corinthians 23/26 (88.4%) 4/26 (15.4%) Galatians 15/17 (88.2%) 3/17 (17.6%) Ephesians 13/18 (72.1%) 2/18 (11.1%) Philippians 9/11 (81.8%) 3/11 (27.3%) Colossians 9/10 (90%) 1/10 (10%) 1 Thessalonians 4/5 (80%) 0/9 (0%) 2 Thessalonians 2/4 (50%) 150 B(03) has a pregenealogical coherence of 97.22% with Edwards’ reconstructed text, even though he changed the NA28 readings twice, against the reading in the codex. Edwards, “2 Thessalonians,” 252. 151 Dormandy notes that B(03) “departs from the initial text almost twice as often as [C]04” and that harmonizations and linguistic improvements constitute the majority of the variations. We may add to this that two of Jongkind’s five editorial features in B(03) are especially apparent in Romans—the change of καθωϲ γεγραπται to καθαπερ γεγραπται and the order χριϲτοϲ ιηϲουϲ. Dormandy, “Pandects,” 358; Jongkind, “Redactional Elements,”234–38. 152 Certainly, studies like that of Carlson on Galatians provide more detailed information on individual manuscripts and their relatives in particular books. However, they are less practical for this sort of comparison. On the cautious use of TuT data in the Pauline corpus, see Schmid and Morrill, “Editorial Activity,” 361–383; cf. Carlson, Galatians. 153 There is also variation in the location of Teststellen in each book. For example, 26 of the 47 Teststellen in Romans are found in the last three chapters, while none are found in Romans 1–4. Schmid and Morrill, “Editorial Activity,” 376. 154 In the TuT volumes, these data are found in the Vorbemerkung and the Ergänzungsliste of Aland ed., II. Die Paulinischen Briefe. CHAPTER 6 244 Hebrews 1–9:14 13/15 (86.6%) 1/15 (6.7%) Nevertheless, in the TuT volumes of the Pauline corpus, B(03) is occasionally surpassed in agreements with the NA28 by א(01) and C(04). For Romans in particular, this result independently corroborates Dormandy’s conclusion about the value of א(01) and C(04) over B(03). 155 Still, it remains to be seen whether the early corrections in the Pauline corpus have any effect on the questions surrounding the text of the epistles. 6.4.2 The Earliest Corrections in the Corpus Paulinum Addition All four corrections of addition occur in the first three letters of the Pauline corpus. The first and the last constitute the largest examples of corrected dittography in the NT. In one instance, a single- word addition did not result in dittography (p. 1474A, 1 Cor 14:39), and it finds support from a few important witnesses (א[01], A[02], 1739). None of the four corrections certainly originated in the fourth century. TABLE 48: CORRECTIONS OF ADDITIONS IN THE CORPUS PAULINUM SCRIBE B(03) CORRECTOR Romans 4:4; p. 1448B B τω δε εργαζοµενω ❛ο❜̇ ❛µιϲθοϲ ου λογιζεται❜ ❛κατα χαριν αλλα κατα❜ ❛οφειληµα· τω δε µη❜ ❛εργαζοµενω❜ ο µιϲθοϲ Β2 or Β18 1 Cor 13:7; p. 1472C B τη αληθεια ❛παντα ϲτε❜ ❛γει❜ παντα ϲτεγει παν τα πιϲτευει παντα ελ πιζει παντα υποµενει Β2 or Β18 1 Cor 14:39; p. 1474A B … ωϲτε αδελ φοι µ̇ου̇̇ ζηλουτετο Β2 or Β18 2 Cor 3:15; p. 1479B B … ηνικα αν ανα γεινωϲκηται µωυϲηϲ ❛καλυµµα επι την καρ❜ ❛διαν αυτων κειται· ηνι❜ ❛κα δ αν επιϲτρεψη προϲ❜ ❛κ̅ν̅ περι\αι/ε̸ρειται το❜ κα λυµµα επι την καρδι αν… B18 The first correction of addition involves a single deletion dot and rounded hooks on five lines of repeated text (p. 1448B, Rom 4:4). The corrector may have originally intended to use deletion dots 155 Dormandy, “Pandects,” 360. EARLY CORRECTIONS IN THE NEW TESTAMENT 245 before noticing the extent of the error. It is unlikely that the dot came from a different hand, since the rest of the error would have hardly gone unnoticed to a corrector. While the NTVMR transcription associates the correction with C1, Tischendorf believes that the reinker was responsible for the deletion. The latter seems to be the better suggestion, since we find a late medial dot in the deleted text. We would not expect to find this if the lines had already been clearly corrected with the round hooks. A parallel correction appears in 2 Corinthians 3:15 (p. 1479B), where the orthographic correction περιερειται → περιαιρειται occurs twice, on account of the dittography. A late medial dot is also included in the deleted portion of text. Tischendorf identifies his B3 as responsible for the round hooks and orthographic correction, while the NTVMR distinguishes two hands (C2a/C2b). It is possible that the same hand made the orthographic correction before noticing the dittography. At first glance the supralinear alpha looks closer to an early hand than the standard alpha of B18. However, a comparable example can be found in the B18 correction on p. 765B (Song 4:40; Figure 129).156 In this correction, the more common rounded alpha precedes an alpha that is close in form to the correction in 2 Corinthians 3:15. Likewise, Tischendorf may be right that the round hooks in 1 Corinthians 13:7 (p. 1472C) were also the work of B18. FIGURE 129: THE LETTER ALPHA IN B18 CORRECTIONS (P. 765B, SONG 4:40; P. 1479B, 2 COR 3:15) One of the early hands could be responsible for the single word deletion of µου in 1 Corinthians 14:39 (p. 1474A). Tischendorf, on the other hand, is likely correct in associating the dark deletion dots with the reinker. We have already seen that the corrected reading of this text is one of the few examples in Zuntz’s study, where B(03) is brought into conformity through correction with P46 and the “Western” witnesses, D(06)*, F(010), and G(012). Zuntz, however, was not convinced by this reading.157 Omission Fourteen corrections in the Pauline corpus amend omissions. Ten of these corrections could have originated with an early hand, three of which exhibit the features of B3. The largest omission consists 156 Both corrections seem to be examples of the reinker’s attempt to imitate an earlier hand. The first, however, shows significant variation in copying the other alphas (see p. 765B, Song 4:40). 157 Zuntz, Epistles, 179. CHAPTER 6 246 of three words or thirteen letters (p. 1453C, Rom 9:3). Ten of the fourteen corrections remedy single word omissions. TABLE 49: CORRECTIONS OF OMISSIONS IN THE CORPUS PAULINUM SCRIBE B(03) CORRECTOR Romans 8:24; p. 1453A B βλεπει τιϲ \τι/ ελπιζει ει δε ο ου βλεποµεν… Β2 or Β18 Romans 9:3; p. 1453C B … ειναι αυτοϲ εγω α πο του χ̅υ̅ υπερ των [αδελφων µου τω ̅ ϲυγγενων µου… Β3 Romans 9:8; p. 1453C B τουτ εϲτιν \οτι/ ου τα τε κνα… Β2 or ΒB Romans 10:17; p. 1455B B [ρ] … η δε ακοη δια ρη µατοϲ χ̅υ̅… ΒB 1 Cor 6:9; p. 1465C B … οτι αδικοι θ̅υ̅ βαϲι λειαν ˙/. κληρονοµηϲου ϲιν… ˙/. ου Β3? 1 Cor 11:3; p. 1470B B … η κε φαλη \ο/ χ̅ϲ ̅εϲτιν… Β18 2 Cor 12:11; p. 1486C B γαρ ωφειλον υ\φυ/µων ϲυνιϲταϲθαι… Β18 Gal 6:3; p. 1492C B … ει γαρ δοκει τιϲ ειναι \τι/ µηδεν ων φρε ναπατα εαυτον… B2 or Β18 Eph 1:1; p. 1493B B τοιϲ αγιοιϲ τοιϲ ουϲιν και πιϲτοιϲ… [εν ε φεϲω Β18 1 Thess 4:1; p. 1508C B … των αγιων αυτου το] λοιπον αδελφοι ερω τωµεν… B2 or Β18 1 Thess 4:1; p. 1508C B … των αγιων αυτου λοιπον \ουν/ αδελφοι ερω τωµεν… Β18 1 Thess 4:4; p. 1508C B … ειδεναι [ενα εκαϲτον υµων… B2 or Β18 Heb 5:4; p. 1515Α B ουκ εαυτω \τιϲ/ λαµβανει B2 or Β18 Heb 7:5; p. 1516C B ˙/. τουτ ε ϲτιν κατα τον νοµον ˙/. τουϲ αδελφουϲ αυτων… Β3 Scribe B certainly corrected one of the omissions in scribendo (p. 1455B, Rom 10:17). After initially skipping the preposition δια, the scribe began copying the word ρηµατοϲ. Scribe B caught the error immediately after copying the rho and remedied it by overwriting the rho with a delta. Apart from this overwriting, the regular length of the line is unchanged, signaling the correction’s origin in scribendo. Versace has also identified two marginal corrections as coming from his B2, the scribe or EARLY CORRECTIONS IN THE NEW TESTAMENT 247 diorthotes (p. 1508C, 1 Thess 4:1, 4).158 The latter reading was likely omitted because of the shared letters in ειδεναι and ενα. Tischendorf, however, believed the two corrections came from the reinker. Although the form of these marginal corrections is obscured by the reinking, it seems probable that Versace is correct. Four supralinear corrections of omission include the letters tau and iota and may have originated with an early hand (Figure 130). The addition of οτι (p. 1453C, Rom 9:8) and τιϲ (p. 1515A, Heb 5:4) have the raised omicron and sigma, which is characteristic of the early hands. The two additions of τι are less clear (p. 1453A, Rom 8:24; p. 1492C, Gal 6:3). Tischendorf classifies the former as a B2 correction, while the latter is said to have originated with his B3. However, both corrections are probably from the same hand as the crossbar of each tau meets the middle height of the iota. The slight tilt of the second crossbar could be the result of the reinking, though it is not absent in other early corrections. FIGURE 130: CORRECTIONS OF OMISSION CONTAINING TAU AND IOTA (P. ) The corrector B3 was likely responsible for three corrections of omission. The first amends a scribal error that was clearly caused by homoeoteleuton at the line-break (p. 1453C, Rom 9:3). This omission parallels that in Matthew 25:40 (p. 1271A), which Clark cites as evidence for an ancestor containing lines of ten to twelve letters (not including the macron of the final nu). However, the earlier error was possibly corrected by scribe B, while this one clearly came from B3. Furthermore, the omission of τουτ εϲτιν in Hebrews 7:5 (p. 1516C), was likely triggered by homoeoarchon with the following word τουϲ. On p. 1465C (1 Cor 6:9), it is not entirely certain that the lemniskos correction ου was added by B3, because it is obscured by the reinking.159 As discussed in Chapter 4, the omission and subsequent correction of εν εφεϲω (p. 1493B, Eph 1:1) was one of the earliest features debated in B(03). Leonhard Hug first identified this line-end 158 Versace, Marginalia, 134. 159 Versace, Marginalia, 135 n. 303. CHAPTER 6 248 correction as prima manus, but Tischendorf significantly opposed the early dating.160 More recently, Lynn Cohick has cited B(03) as containing the place-name in its “earliest editorial changes.” 161 However, she corrects this assessment in her newest commentary, following the tenth or eleventh- century dating.162 There can be little doubt that Tischendorf and, subsequently, Versace are correct in attributing this correction to the reinker.163 Substitution Substitutions are the reason for eighteen corrections in the Pauline corpus.164 However, only six of the eighteen may have come from early hands. Fifteen of these corrections involve a single letter, while the largest constitutes three. The interchange of ηµειϲ-υµειϲ accounts for six corrections of substitution, a feature that has played a minor role in the previous sections (cf. p. 1363B, John 8:54; p. 1408B, Acts 17:23). Likewise, the substitution θεοϲ → χριϲτοϲ occurs three times and only in Paul (cf. θυ̅̅ → κ̅υ̅; p. 1402B, Acts 13:44). Another significant substitution involves three corrections from subjunctive to indicative in the third-person plural; twice with the verb εχω (p. 1449B, Rom 5:1; p. 1492C, Gal 6:10 [2x]). TABLE 50: CORRECTIONS OF SUBSTITUTIONS IN THE CORPUS PAULINUM SCRIBE B(03) CORRECTOR Romans 5:1; p. 1449Β B … ει ρηνην εχ\ο/ωµεν προϲ τον θ̅ν̅… Β18 Romans 14:18; p. 1458C B … δουλω ̅ [θ] τω χ̅ω̅ ευαρεϲτοϲ τω θ̅ω̅… Β18 1 Cor 1:6; p. 1461C B ⟦θ̣⟧ το µαρτυριαν του χ̅υ̅ B2 or Β18 1 Cor 9:9; p. 1468B B γεγραπται ου \φ/κ\ι/ηµω ϲειϲ βουν αλοωντα Β18 2 Cor 1:21; p. 1478Α B … ο δε βε βαιων υµαϲ ϲυν υµι ̅ ειϲ χ̅ν̅ και χρειϲαϲ \η/υ̸µαϲ B2 or ΒB 2 Cor 5:12; p. 1480C B … διδον τεϲ \υ/ηµιν καυχηµατοϲ Β18 2 Cor 9:3; p. 1483C B καυχηµα \η/υ̸µων το υ περ υµων κενωθη… Β18 160 Interestingly, there is no mention of the correction in Westcott and Hort, Introduction, 123–124, or Hort, Prolegomena to St Paul’s Epistles, 75–81. 161 Cohick, Ephesians, 9. 162 Cohick, Letter to the Ephesians, 26–27 n. 98. 163 Versace, Marginalia, 264. 164 The following table does not include the clearly late correction on p. 1493C (Gal 6:11). EARLY CORRECTIONS IN THE NEW TESTAMENT 249 2 Cor 9:4; p. 1483C B … ηµειϲ ινα µη λεγωµεν \υ/η̸µειϲ εν τη υποϲταϲει ταυτη Β18 Gal 3:28; p. 1490C B … ουκ ενι αρ ϲεν και θηλυ \α/παντεϲ γαρ υµειϲ ειϲ εϲτε εν χ̅ω̅ ιυ̅̅… B2 or Β18 Gal 6:10; p. 1492C B … αρα ου ̅ ωϲ καιρον εχ\ο/ωµεν ερ γαζ\ο/ωµεθα το αγαθον Β18 Col 4:3; p. 1505C B … λαληϲαι το µυ τηριον του \χ/θ̸̅υ̅ δι ον και δεδεµαι… B2 or Β18 1 Thess 1:8; p. 1507A B …η πι ϲτιϲ υµων η προϲ το ̅ θ̅ν̅ εξεληλυθεν ωϲτε µη χρειαν εχειν \η/υ̸̇µαϲ λαλειν… B2 or Β18 1 Thess 3:9; p. 1508Β B … δυναµεθα τω θ̅ω̅ ανταποδουναι περι \υ/η̸µων… Β18 Heb 1:3; p. 1512B B φ⟦αν⟧ερων τε τα παντα Β18 Heb 6:7; p. 1515C B … γη γαρ ηπι ουϲα τον επ αυτη\ν/ϲ̸ ερ χοµενον… B2 or Β18? Heb 6:10; p. 1516A B … και τηϲ αγαπηϲ η\ν/ϲ ενεδειξα ϲθε… Β18 Heb 8:7; p. 1518A B … ουκ αν \δευ/ετε ραϲ εζητειτο τοποϲ Β18 It is difficult to identify the hand of the correction θ̣υ̅̅ → χ̅υ̅ in 1 Corinthians 1:6 (p. 1461C). The parallel correction in Romans 13:18 was likely made by B18, who overwrote the original theta without clearly erasing it. However, the former correction decisively erased the theta leaving no trace of the original letter (Figure 131). The chi that appears over the erasure is also more comparable to those of scribe B. Therefore, Tischendorf identifies both corrections as B3, but allows for the possibility of an earlier hand in 1 Corinthians 1:6. The third substitution of θυ̅̅ → χ̅υ̅ on p. 1505C (Col 4:3) differs from the previous two, since it was corrected through supralinear addition. There are clearly two layers to the supralinear chi, but the form of the cancelation stroke suggests it is more likely to have originated with B18. FIGURE 131: Θ ̅Υ ̅ → Χ ̅Υ ̅ CORRECTIONS IN THE CORPUS PAULINUM (PP. 1458C, 1461C, 1505C) CHAPTER 6 250 Likewise, the correction παντεϲ → απαντεϲ (p. 1490C, Gal 3:28) could have originated with an early or late hand. In this example, the alpha appears to follow the form of the scribes and early correctors (Figure 132). However, we can also find some parallel with the B18 alpha, mentioned above (Figure 129). Only א(01) and A(02) share the corrected reading of B(03), and this may well favor the earlier dating of the correction.165 FIGURE 132: ΠΑΝΤΕC → ΑΠΑΝΤΕC CORRECTION IN GALATIANS 3:28 (P. 1490C) In the Pauline Corpus, we find six corrections of the interchange ηµειϲ-υµειϲ, with three examples of each substitution. The majority of these substitutions occur in contexts where the same pronoun is repeated, suggesting that context influenced the scribe. From the six corrections, Tischendorf identifies two as coming from an early corrector (p. 1478A, 2 Cor 1:21; p. 1508B, 1 Thess 3:9). However, the latter correction of ηµων → υµων with its slanted upsilon is more likely to have originated with the reinker (Figure 133). We might add the correction υµαϲ → ηµαϲ (p. 1507A, 1 Thess 1:8) as a potentially early correction, since the two deletion dots seem to have been added around a pre- existing eta. If this is an early correction, it is unlikely to have come from the same hand as the early correction in 2 Corinthians 1:21 (p. 1478A), since the cancelation stroke crosses the left oblique rather than the stem of the upsilon, and the eta is noticeable wider and higher. FIGURE 133: ΗΜΕΙC-ΥΜΕΙC CORRECTIONS IN THE CORPUS PAULINUM (PP. 1478A, 1507A, 1508B) One of the most discussed substitutions is found in Romans 5:1 (p. 1449B). 166 In B(03), the subjunctive εχωµεν is corrected to the indicative with a supralinear omicron. There is hardly any doubt that this correction and the two parallel corrections in Galatians 6:10 (p. 1492C) originated with the reinker. However, Loretta Man has recently questioned the value of the original and corrected readings in B(03) by examining the other ο-ω corrections in the NT.167 She concludes her 165 The NTVMR transcription gives C1, while Tischendorf has B3. 166 For example, see Tregelles, An Account, 156; Weiss, Paulinischen Briefe, 44–45; Scrivener, Introduction (1961 ed.), 447; Sanday and Headlam, Romans, 120; Fitzmyer, Romans, 395. 167 Man, “Textual Significance,” 70–93. EARLY CORRECTIONS IN THE NEW TESTAMENT 251 study by stating, “the prima manus tends to have an incorrect reading. Therefore, the B* reading in Rom 5,1 is questionable.”168 However, Man’s suggestion that B(03) misheard εχοµεν is unconvincing, as we have already challenged the notion of dictation for the codex (see Chapter 1). Nevertheless, the correction of ω → ο in Romans 5:1 stands out, since the majority of corrections involving ο-ω interchanges are in the opposite direction. Orthography In the Pauline corpus, there is little variation from the standard patterns of orthographic correction in the work of scribe B. Corrections of ι → ει constitute the single largest category of correction, with an additional eight corrections of ει → ι, using either cancelation strokes or deletion dots.169 Unique to the Pauline corpus is the unusually high number of corrections of the unassimilated nu. While this has always been a significant category in B(03), it is exaggerated by Paul’s preference for vocabulary with the ϲυν- prefix.170 The orthographic corrections α → αυ, γ → ν, ζ → ϲ, ι → η, ν → ϲ, ου → ω, ω → ου, π → φ, and φ → π are all unique to the Pauline corpus in the NT and scribe B’s portions of 1 Kingdoms and Psalms.171 It is unclear whether any of these corrections were made by the scribe or early correctors. It is certainly possible that some of the ει → ι corrections originated with an early hand, which used deletion dots and cancelation strokes. However, we have already seen a consistent pattern of B18 corrections with deletion dots over the word µεικροϲ. An early hand may have added the ε → αι corrections in 1 Corinthians 7:15 (p. 1466C) and Philippians 1:18 (p. 1499B), based on the shape of the alphas (Figure 134). However, the cancelation stroke in the first correction appears to be late and the shape of the second alpha is completely covered by the reinking, making certainty difficult. FIGURE 134: Ε → ΑΙ CORRECTIONS IN THE CORPUS PAULINUM (PP. 1466C, 1499B) 168 Man, “Textual Significance,” 91. 169 Again, this number does not include corrections that necessarily originated from the reinker, by leaving the text untouched. 170 Dunn, Theology of Paul, 401–404; cf. Campbell, Union with Christ, 228–236. 171 TM Text Irregularities gives 331 examples of the interchange α for αυ. Alternatively, the correction αναϲταρουνταϲ → αναϲταυρουνταϲ (p. 1515C, Heb 6:6) could be understood as a correction of a nonsense reading. CHAPTER 6 252 All of the corrections of proper names occur in Romans and Colossians.172 The neighboring city of λαοδικεια is lacking the epsilon all four times in Colossians. The reinker added the missing letter by compressing the epsilon between the kappa and iota in all four cases. Following the other orthographic corrections, the spelling of κεγχρεαι is present without the assimilated nu. Twice the aorist of ευδοκεω is found with the augment, and was subsequently corrected by B18. Nonsense Apart from orthography, nonsense readings comprise the majority of corrections in the Pauline corpus. Of the thirty corrections of nonsense readings listed below, only twelve could possibly have originated with the fourth-century hands.173 Twelve of the thirty involve errors of a single letter, and the remaining eighteen comprise two or three letter errors. For example, the word διακονοϲ is misspelled twice, both times with a missing alpha (p. 1457C, Rom 13:4; p. 1485C, 2 Cor 11:15). As we have already seen, parablepsis is likely the main cause of the nonsense readings.174 TABLE 51: CORRECTIONS OF NONSENSE READINGS IN THE CORPUS PAULINUM SCRIBE B(03) CORRECTOR Romans 2:7; p. 1446B B … τοιϲ µεν κα θ υποµ\ον/ην εργου αγα θου… B2 or Β18 Romans 3:28; p. 1448A B ουν δικαιουϲθαι πι ϲτει α\ν/ρθρωπον χωριϲ εργων νοµου… B2 or Β18 Romans 8:30; p. 1453A B … ου\ϲ/ δε προωριϲεν τουτουϲ και εκαλεϲεν… Β18 Romans 9:29; p. 1454B B ωϲ ϲοδοµα αν εγενη [θη \µ/θεν και ωϲ γοµορρα… Β18 Romans 11:1; p. 1455B B … φυληϲ βενιαµ\ειν/ ουκ απωϲατο B2 or ΒB Romans 11:6; p. 1455C B … χαριϲ επ\ε/ι το εργον ουκετι εϲτιν χαριϲ… Β18 Romans 11:24; p. 1456B B … οι κατα φυϲιν εν κεντριϲθηϲον\ται/ τη ι δια ελαια… Β18 Romans 13:4; p. 1457C B γαρ δι\α/κονοϲ εϲτιν ϲοι B2 or ΒΒ 172 See also the nonsense reading βενιαµ for βενιαµιν in Romans 11:1 (p. 1455B). 173 The following table does not include the early erasure in 1 Corinthians 14:4 (p. 1473A). The NTVMR transcribes the original nonsense reading as εαυτον οικ̣ο̣κοδοµει. Although this would make sense of the two-letter empty space between οι and κοδοµει, the erasure goes back to the alpha of εαυτον. Therefore, the space was caused by the overwriting of a shorter text, while the letters of κοδοµει had already been copied on the previous line. 174 See especially, p. 1456B, Rom 11:24; p. 1460B, Rom 16:7; p. 1462A, 1 Cor 1:11; p. 1464B, 1 Cor 4:6; p. 1475B, 1 Cor 15:39; p. 1480A, 2 Cor 4:15; p. 1485C, 2 Cor 11:15; p. 1494A, Eph 1:19; p. 1504C, Col 3:4; p. 1517A, Heb 7:16. EARLY CORRECTIONS IN THE NEW TESTAMENT 253 Romans 16:7; p. 1460B B αϲπα\ϲα/ϲθε ανδρονεικο ̅ Β18 1 Cor 1:11; p. 1462A B γαρ µοι περι υµων α δελφοι µο\υ/ι υπο των χλοηϲ… Β18 1 Cor 4:6; p. 1464B B … ειϲ εµαυ τον και α\πο/ πολλων… Β18 1 Cor 4:15; p. 1464C B … δια του \ευ/αγγελι ου εγω υµαϲ εγεννη ϲα… Β18 1 Cor 5:11; p. 1465B B … η ειδωλο\λα/τρηϲ η λουδοροϲ η µεθυϲοϲ B2 or ΒΒ 1 Cor 9:11; p. 1468B B [ε] υµιν τα πνευµατικ [α ε ϲπειραµεν… Β18 1 Cor 15:35; p. 1475B B … δε ϲωµατι ερχον\ται/ α φρων ϲυ ο ϲπειρειϲ B2 or ΒΒ 1 Cor 15:39; p. 1475B B ου πα\ϲα/ ϲαρξ η αυτηϲ ϲαρξ Β18 2 Cor 1:5; p. 1477B B … τα παθηµατα του χ̅υ̅ ει\ϲ/ ηµαϲ… Β18 2 Cor 1:16; p. 1478A B και παλιν απο \µα/κεδονιαϲ Β18 2 Cor 4:15; p. 1480A B ⟦τ⟧ ϲυν υµιν \τα/ γαρ παντα B2 or ΒΒ 2 Cor 6:3; p. 1481B B ινα µη µω\µη/θη η διακονι α… Β18 2 Cor 11:15; p. 1485C B ωϲ δι\α/κονοι δικαιοϲυ νηϲ… B2 or Β18 Gal 2:16; p. 1489B B εαν µη δια πιϲτε\ω/ϲ χ̅υ̅ ιυ̅̅ Β18 Gal 5:17; p. 1492B B … µη α \εα/ν θελητε ταυτα B2 or ΒB Gal 5:21; p. 1492B B φθονοι·ι ̇µεθαι κωµοι Β18 Eph 1:19; p. 1494A B … τουϲ πιϲτευον [η] ταϲ κατα την ενεργει αν… Β18 Phil 1:30; p. 1500A B τον αυτον αγωνα εχο ̅ [τεϲ οιον ειδετε… Β18 Col 3:4; p. 1504C B η ζωη \η/µων… Β18 Heb 7:16; p. 1517A B … κατα νοµον εντο ληϲ \ϲ/αρκινηϲ… Β18 Heb 8:11; p. 1518A B … οτι παντεϲ ειδ\ηϲ/ουϲιν µε απο µει κρου… B2 or Β18 Heb 9:4; p. 1518B B αγιων \ε/χουϲα την κιβω τον τηϲ διαθηκηϲ… B2 or Β18 Tischendorf identifies the first correction (p. 1446B, Rom 2:7) as coming from an early hand. However, the reinking has obscured the form of the nu. It is also striking that the omicron is equal CHAPTER 6 254 in size with the nu, unlike the more common raised and compressed nu of the early hands. Whether or not this is an early correction, it is probable that the following correction αρθρωπον → ανθρωπον (p. 1448A, Rom 3:28) came from the same hand.175 In Romans 11:1 (p. 1455B), the name βενιαµιν is misspelled as βενιαµ. This error does not happen anywhere else in B(03), and was possibly influenced by the name αβρααµ in the line before. Like the previous corrections, the supralinear addition of ειν is obscured by the reinking. One might expect that the reinker would have preferred the -ι- spelling of βενιαµιν, but the -ει- spelling is reinforced in every occurrence.176 However, an early hand is preferred based on the remnants of the undertext. The same could be said of the correction χουϲα → εχουϲα in Hebrews 9:4 (p. 1518B). Versace has included the correction ερχον → ερχονται (p. 1475B, 1 Cor 15:35) in his B2 category.177 We can compare this to a parallel correction in Romans 11:24 (p. 1456B), for which B18 is surely responsible (Figure 135). The article τη likely influenced the omission of the verb ending in Romans, while it is less clear what caused the error in 1 Corinthians. FIGURE 135: CORRECTIONS OF NONSENSE READING ENDING IN -ΤΑΙ (P. 1456B, ROM 11:24; P. 1475B, 1 COR 15:35) Furthermore, this correction can be compared to five others which involve the addition of a supralinear alpha. An early hand likely corrected the nonsense reading δικονοϲ (or δικονοι) to διακονοϲ twice (p. 1457C, Rom 13:4; p. 1485C, 2 Cor 11:15). The error in 2 Corinthians was probably influenced by the following word δικαιοϲυνηϲ. Tischendorf and the NTVMR transcription claim the first nonsense reading was corrected early, while the second originated with the reinker (Figure 136). However, the only clear difference between these two is that the latter was completely overwritten, while traces of undertext in the former remain. If we accept Versace’s early date for the 175 Tischendorf classifies this correction as B3, but this is most likely because there is no visible undertext. The form of the nu is even closer to the early lemniskoi corrections than the previous correction. 176 Swete identifies a correction of ι → ει in 2 Esdras 21:4 (p. 597B), but this does not appear to be a genuine correction. Rather, the epsilon is compressed because of the line-end. 177 Versace, Marginalia, 134. EARLY CORRECTIONS IN THE NEW TESTAMENT 255 correction in 1 Corinthians 15:35, these two are also likely to have come from the hand of scribe B or a diorthotes. FIGURE 136: CORRECTIONS OF ΔΙΑΚΟΝΟC IN THE CORPUS PAULINUM (PP. 1457C, 1485C) In 1 Corinthians 5:11 (p. 1465B), an early corrector amended the nonsense reading ειδωλοτρηϲ by adding a supralinear lambda and alpha before the tau (Figure 137). The error was likely prompted by the previous syllable -λο- (cf. µωθη → µωµηθη; p. 1481B, 2 Cor 6:3). If this were a late correction, we would expect to see a curve at the top of the lambda. The nonsense reading ταρ was corrected to τα γαρ by erasing the left half of the tau’s crossbar and adding the letters tau and alpha above the line (p. 1480B, 2 Cor 4:15). The simplicity of the correction tau to gamma highlights the ease in which the two characters can be confused, leading to this nonsense reading. The supralinear alpha is far more elongated than usual, but Tischendorf is likely correct to identify this with an early hand. Likewise, the last correction of αν → α εαν (p. 1492B, Gal 5:17) probably originated with an early corrector (Figure 137). FIGURE 137: SUPRALINEAR CORRECTIONS OF NONSENSE READINGS IN THE CORPUS PAULINUM (PP. 1465B, 1480B, 1492B, 1518B) Finally, the correction of ειδουϲιν → ειδηϲουϲιν in Hebrews 9:4 (p. 1518B) exhibits the characteristics of an early hand. The reinking and likely some fading have obscured the original form of the correction. However, one only needs to look at the final letters of ειδωλολατρηϲ (p. 1465B, 1 Cor 5:11) to see the similarities with the original, compressed eta and sigma (Figure 137). 6.4.3 Summary The earliest corrections in the Pauline corpus conform to the broader patterns of correction in the work of scribe B. Corrections of omission occur with more frequency than those of addition. Yet, we have also seen that the largest examples of dittography in the NT occur in the first three epistles. In general, substitutions prompt even more corrections than omissions. However, only a few of these CHAPTER 6 256 can be confidently assigned to the fourth century. This results in a higher frequency of early corrected omissions than of substitutions. Corrections of orthography are mostly attributed to B18, but still highlight the spelling habits of scribe B. In particular, we saw the high frequency of ϲυν- prefixed words in the Pauline corpus with the unassimilated nu, which is common to the scribe. Likewise, nonsense readings comprise a majority of corrections in Paul, as in the rest of the NT. These minor errors are normally caused by parablepsis and were mostly corrected by the reinker. In contrast to the rest of the NT, there is an abnormally high number of nonsense corrections, which may have been executed by scribe B or a diorthotes. 6.5 Conclusion This broad study of the early corrections in the NT is far from exhaustive. Erring on the side of caution has led to a number of unresolved corrections (listed as “B2 or B18”). This category often marks unclear supralinear corrections or deletions using the various methods outlined in Chapter 4. While the previous chapter compared the corrections in the work of our three scribes, this chapter has extensively examined corrections to the work of scribe B. In each of the four sections of the NT, we have found strong similarities in the types and frequency of corrections. These commonalities also have parallels in 1 Kingdoms 19:11b–31:13 and Psalm 1:1–77:71a. As in the case of Psalm 32:1 (p. 642B) and 1 Kingdoms 30:25 (p. 352B), we find clear examples of in scribendo corrections by scribe B.178 The most common form of in scribendo correction occurs after the scribe begins copying a word and notices that the previous word had been skipped. Scribe B amends this error by overwriting the proper word, usually before the following word had been fully copied. Apart from orthography, nonsense readings and substitutions prompt the majority of corrections in the NT. On a few occasions, the nonsense readings were corrected by an early hand. However, B18 seems to have amended the errors at a higher rate because of the nature of the reinker’s task. Since most of these readings were influenced by parablepsis and involve the omission of a single letter, scribe B and the diorthotes were more likely to overlook them. The reinker, on the other hand, was bound to notice such errors after reinforcing each letter of the text. Furthermore, corrections of substitution evidently reveal some decline in early corrections after the Gospels 178 See p. 1248B, Matt 11:9; p. 1251B, Matt 13:8; p. 1261C, Matt 20:15; p. 1365B, John 10:10; p. 1367A, John 11:27; p. 1455B, Rom 10:17. EARLY CORRECTIONS IN THE NEW TESTAMENT 257 (Figure 138). Also missing from this category are the S-siglum corrections by B3, which we found in 1 Kingdoms and Psalms. These marginal corrections had a tendency towards Lucianic readings, leading to the suggestion that they originated from a second exemplar. However, those S-siglum corrections by scribe B were likely present in the main exemplars of B(03). This conclusion seems to make most sense of their nearly exclusive appearance in Matthew and high frequency in Joshua, 2–4 Kingdoms, Jeremiah, and Ezekiel. In Chapter 3, we saw that a number of Hexaplaric sigla were also inherited from the exemplars, including the abbreviation ου κ’ π’ εβρ’ (§4.1.5). Therefore, we should not be surprised that the main exemplars of B(03) contained marginal readings, especially substitutions. Whether we prefer the readings εργων or τεκνων (p. 1248C, Matt 11:19), νυµφων or γαµοϲ (p. 1265A, Matt 22:10), ηγεµονιαϲ or βασιλειαϲ (p. 1308C, Luke 3:1), it appears that the those involved in preparing B(03) viewed these variations as genuine alternatives worth transmitting. FIGURE 138: SUMMARY OF CORRECTIONS AND CORRECTORS IN THE NT Frequent reference to the reinker (B18) has been crucial for distinguishing early from late hands. It also provides a base of comparison for the early correctors throughout the NT (Figure 138). In the Gospels, scribe B and the diorthotes (B2) are responsible for correcting the majority of omissions. One could argue that this evidence exhibits a heightened awareness by the early correctors that scribe B was prone to short omissions. While they no doubt overlooked errors throughout, especially those involving a single letter (cf. nonsense corrections), the attention to omission may give us reason to question the oversimplified view of the scribe as carelessly overlooking text. 3 35 39 5 7 11 5 4 3 5 1 1 3 5 4 14 21 11 35 2 3 7 12 4 5 11 3 10 1 15 41 3 4 16 2 7 6 5 1923 72 2 33 89 11 15 34 2 4 12 8 18 35 BB B2 or BB B2 or BC B3 B2 or B4 B2 or B18 B18 BB B2 or BB B3 B2 or B18 B18 B2 or BB B3 B2 or B18 B18 BB B2 or BB B3 B2 or B18 B18 Gospels Acts Catholic Epistles Corpus Paulinum Addition Omission Substitution Transposition Nonsense CHAPTER 6 258 In the NT, B3 only made corrections of omission, all of which likely originated from another exemplar. However, this does not imply that all of the omissions were inherited from the primary exemplars of scribe B. The clearest example of this feature is the omission of αδελφων µου των in Romans 9:3 (p. 1453C), with the parallel omission in Matthew 25:40 (p. 1271A). Both are clearly the result of homoeoteleuton, but only the first was caught by scribe B. The latter omission was corrected by B3, not necessarily because there was a different reading in the exemplar, but because the corrector caught the omission by comparison with the other exemplar. On two occasions, we have found reason to believe that our scribe C, or a corrector with a similar hand, was responsible for a correction in the NT (p. 1339C, Luke 19:29; p. 1349C, John 1:13). Both examples occur at line-ends, analogous to what we observed in Psalm 64:6 (p. 662B). This does not guarantee that scribe C was responsible for all three of these corrections, nor is it certain that scribe C copied only these three. However, the evidence suggests that there was some editorial cooperation between the scribes. Finally, in our discussion of the so-called “Western element” in Paul, we noted that there was only one clear example from Zuntz’s study, where B(03) was corrected to a “Western” reading, agreeing with P46 (p. 1474A, 1 Cor 14:39). However, even this correction likely originated with the reinker and has little impact on our understanding of the early text of B(03). Only after the remaining editions of the ECM are published will we be able to understand the relationship between the early corrections and B(03)’s closest relatives. In the Catholic Epistles, we found a number of corrections that brought the text of B(03) into conformity with its closest relatives (e.g., p. 1430B, 1 Pet 1:1; p. 1435C, 2 Pet 2:1). In Acts, however, we also saw early corrections away from B(03)’s relatives (e.g., p. 1386A, Acts 3:21). Along with the lack of ECM editions for the Gospels and Pauline corpus, our study was again confronted with the need for autopsy of the codex and the potential gains of new multi-spectral images of B(03). CHAPTER 7 Conclusion It is an unavoidable truth that any examination of a codex with both the scope and importance of B(03) will face limitations on various fronts. Nevertheless, the previous chapters have set out a framework for identifying and comparing the scribes and early correctors who produced B(03). Not only do the layers of correction inform us concerning the reception of the codex, but they also provide a way for comparing the work of each scribe and understanding the editorial context of B(03). Along the way, we uncovered numerous codicological, paratextual, and paleographic features which help us better appreciate the skill with which the entire codex was fashioned, as well as the important relationship that existed between the physical parameters of the manuscript and the copy-work of the scribes. 7.1 The Codex The adoption of structural codicology has resulted in numerous important observations in B(03). Indeed, while the temptation is often to conflate the layers of production when studying a manuscript, this stratified view of the codex cautions against such maneuvers. In the case of B(03), we find a codex produced with great care and consistency. However, the attempt to create the finest parchment possible unintentionally led to a high frequency of “maker’s holes,” which the scribes regularly chose to copy around. In one instance, scribe A even copied a rare nomen sacrum to avoid splitting the word πατροϲ across an imperfection (p. 69A, Exod 18:4). Nevertheless, the carefully stacked quires of five parchment sheets, ruled with forty-two lines-per-column, usually presented the scribes with a consistent format for copying. Structural codicology provides a framework for recognizing observable discontinuities in a manuscript, whether in material or mise en page. Such discontinuities are particularly noticeable when the remainder of the manuscript exhibits the consistency of B(03). Therefore, when quires change to forty-four or even forty lines-per-column, there may be some significance attributed to these irregularities. More importantly, when various observable discontinuities align in a single location, we found useful evidence to suggest a change in scribes. This was the case in 1 Kingdoms 19:11 (pp. 334– 335) and Hosea 1:1 (pp. 944–945). While both scribe changes were already identified by Milne and Skeat, they had not sufficiently appreciated the irregularities in quire structure, line-counts, and CHAPTER 7 260 running titles, which align with breaks in the quires. Although we argued that many, if not all, of the running titles were added to the quires before the main text, the irregularities still signaled a transition in production that was more likely to align with a change of scribes. The importance of quire breaks appears to have been lost on Milne and Skeat when studying B(03). For this reason, among others, they misidentified the second of three scribe changes at Psalm 1:1. Unlike the other two, this supposed transition occurs in the middle of a quire. Rather, through paleographic and paratextual analysis, we found agreement with Ludwig Traube that the change occurred in Psalm 77:71, where there is indeed a quire break. While the clean break at Hosea 1:1, aligning with an irregular quire break, would suggest that the scribes could have copied simultaneously, the first two divisions (1 Kgdms 19:11; Ps 77:71) occur in the middle of a verse. Therefore, simultaneous copying would have only been possible if the scribes shared the exemplars, or if the quires of the exemplars were also divided at these verses. 7.2 Paleography and Paratexts One of the greatest challenges to the study of B(03) is the endeavor to identify the scribes by their hands. For one thing, the overall consistency in the production of the codex is similarly reflected in the shared graphic features of the writing. Likewise, in their study of א(01), Milne and Skeat judged this form of paleographic analysis to be inconclusive for identifying the copyists. To make matters worse, a tenth or eleventh-century scribe decided to reinforce the fading ink of B(03), eclipsing many of the distinctive features of each hand. However, this study offered a solution to the dilemma by collecting and studying all the examples of unreinforced text in the codex. When utilized alongside other methods of comparison (i.e., codicological and paratextual), we found that each of the scribes had a distinct way of writing their alphas, deltas, and lambdas. Furthermore, the hand after Psalm 77:71, attributed to scribe A by Milne and Skeat, actually reveals unique features, which are better explained by the existence of a third scribe. While this conclusion cannot be fully sustained from the unreinforced text alone (though see §7.4), numerous other differences in paratexts, paleography, and corrections support a third hand. Notable among the other paleographic features are the distinct και-compendia of scribe C and the tailed-mus of scribe A. As Part I of this study revealed, the paratextual features of B(03) have played a significant role in understanding the earliest layers of production. They have a complex relationship to each other and the main text: some, like the line-fillers, Hexaplaric sigla, and diplai (with exceptions), 261 were likely copied simultaneously with the text, either as scribal creations or as received from the exemplars; others, like the early section numbers or paragraphoi, were added in subsequent phases of production. While many of the additions from the latter group were meant to supplement the earliest layer, they could also be used to correct and clarify. Likewise, the paratexts are spatially related as later additions were either placed around or over earlier ones. Therefore, upon returning to the copied pages of B(03), the scribes themselves or their colleagues acted as interpreters of what was already on the page. The complexity of the paratexts hinders one’s attempts to distill a single source or tradition from them, and yet, their shared presence on the folios exhibits the ambition of the scribes and editors to produce a manuscript that is easy to read, reference, and maybe even recopy. 7.3 The Scribes and Early Correctors What then can we say about scribes A, B, and C, along with their colleagues or contemporaries, B2 and B3? It is noteworthy that the scribe who copied the entire NT is responsible for copying the largest portion of the codex (Table 52). Since scribe B produced nearly triple the number of pages as the other two scribes, it is possible that he or she was considered the most efficient or best-trained copyist. TABLE 52: THE DIVISION OF SCRIBES IN B(03) Scribe A Scribe B Scribe C Scribe B Genesis–1 Kgdms 19:11a 1 Kgdms 19:11b–Ps 77:71a Ps 77:71b–Tobit Hosea–Heb 9:14a 354–356 pages 341 pages 277 pages 574 pages For example, from the corrections, we noted that scribe A is the only scribe who tended to omit, and then correct, text at column-ends. Moreover, scribe B shows a higher reservation over the use of nomina sacra and numeral abbreviations than scribe A. While this may not reflect better training, it does conform to what Zachary Cole has termed the “Christian number-writing technique.” 1 Nevertheless, our investigation did not lead to a clear hierarchy of the scribes. Some features in B(03) occur with varying degrees of frequency throughout the codex, even by the same scribe. As a result, we are occasionally left with the impression that scribal fatigue occurred in the latter portions of the scribes’ work. In the NT, and parts of the Greek OT, this is seen in the use of ekthesis and intralinear spacing. Scribe B utilized ekthesis in Matthew with some 1 Cole, Numerals, 33. CHAPTER 7 262 frequency to produce a hierarchy of divisions between major and minor sections. However, this fades significantly in Mark and the rest of the NT (cf. Isaiah 1–12). Likewise, scribe B avoids line- fillers until Hosea, after which they appear frequently through the Minor Prophets, only to dwindle out at the beginning of Isaiah. This feature may also be reflected in the corrections as the overall presence of marginal emendations fades into the NT. This study has encountered two notable characteristics in B(03), which may reveal something of the editorial context of the codex. First, we have noticed a concentration of features in the Gospel of Matthew that are absent or diminish in the rest of the NT. As mentioned above, this is the only NT book to have received a hierarchical set of divisions through ekthesis and spacing. Moreover, Matthew received the majority of marginal corrections, including the only ancora- lemniskos correction in the NT. Pardee also demonstrated that the few harmonizations in B(03) move towards the Matthean text. This special attention was likely prompted by the popularity of the Gospel in the first centuries of Christianity.2 However, the diminishing of such features in later books may be indicative of a time before they had received similar editorial attention. Scribal fatigue is also likely to have played a role in this tendency, though it may not explain all of the variations we have observed. The prima manus, S-siglum corrections (B1) in B(03) are the second notable addition to the codex. In the NT, these too are largely confined to Matthew and consistently represent alternative readings, analogous to the ECM’s split readings or diamond readings in NA28 and the THGNT. While many of the other marginal corrections could have originated in the comparison with different manuscripts, we judged these S-siglum corrections to have come from the main exemplars. Since nearly one-quarter of these B1 corrections mark spelling variations of proper names in the whole codex, they better reflect the concerns of a scholarly edition, rather than liturgical use. Although we were able to cover well-known corrections such as those in Romans 5:1 and Ephesians 1:1, their late hands did not provide much clarity on the readings of the exemplar. Regarding the absence of εν εφεϲω, the early corrections of omission suggest that scribe B, the diorthotes, or B3 often caught such omissions. A parallel can be found in the marginal addition of βιθυνιαϲ by B3 in 1 Peter 1:1. 2 Massaux, Influence, vols. 1–3; cf. Bird, The Gospel, 303. 263 Lastly, we should mention the way editions handle the correctors of B(03). It is significant that the current editions of the Nestle-Aland and ECM rarely depart from the earlier work of Tischendorf. In general, therefore, there is a lack of clarity over how the editors decide on the corrections when Tischendorf remained undecided (cf. his classification “B2 vel B3”). Moreover, there are inconsistencies in the classification of B1 S-siglum corrections in NA28: εργων → τεκνων (p. 1248C, Matt 11:19) is marked B2, while ειπεν αυτοιϲ → λεγει αυτοιϲ (p. 1253B, Matt 13:52) is designated B1. This study has confirmed Versace’s recent contribution, by distinguishing another early corrector (B3). In the B3, S-siglum corrections of 1 Kingdoms and Psalms, we found a clear affinity with the Lucianic text. This signals an alternative tradition to the corrections of the scribes and our B2, the diorthotai. Therefore, critical editions would benefit from distinguishing two early correctors, the scribe or diorthotai and B3. 7.4 Limitations and Future Avenues of Research While attempting to examine the entire codex leaves many avenues for detailed studies of individual books or sections of B(03), our primary limitation has been the presence of two reinkings to the main text, corrections, and marginalia. This element has often left us with unsatisfactory evidence for distinguishing early from late hands. Therefore, a common conclusion from each of the previous chapters is the need for multispectral images of the manuscript. Such a procedure is by no means a solution to all the problems caused by reinking. After all, B(03) is not a palimpsest, but a manuscript, nearly retraced in its entirety. However, the promise of MSI could still allow for easier detection of distinctive features underlying the reinkings. Another related factor, which has caused difficulty is the occasional similarity in color between the early hands and the tenth or eleventh-century reinking. The current images do not permit close distinctions between ink colors, though MSI may provide additional ways of measuring these differences. The alternative method for circumventing the problem of reinking would be the recent development of artificial-intelligence based writer identification. This operation has been successfully executed on the well-known Isaiah scroll (1QIsaa), resulting in the identification of two copyists.3 Again, the reinkings would limit this method’s precision when it comes to the minute differences in ink traces. However, advanced pattern recognition could provide comprehensive data on the general shape, angles, and orientation of the text and corrections in the work of each 3 Popović, et al., “Artificial intelligence”. CHAPTER 7 264 scribe. Our cursory comparison of the orientation of unreinforced text with reinked pages suggested that this was still maintained through the reinforcement. It is possible, then, that the AI could be calibrated with the unreinforced text to better understand the variations in the reinked pages of B(03). Therefore, the combination of MSI and AI based writer identification could provide more clarity to the three-scribe hypothesis presented above. While some questions have received answers throughout this study, many more must be asked of the great Vatican codex, as it continues to bear the load for current and future textual scholarship. APPENDIX A Codicological Outline and Observable Discontinuities Legend: Quire Numbers= Greek and Roman quire numerals • X= Missing or misplaced number (#) = Original number of folios in the quire Scribes= (M&S): Milne and Skeat (G): Grenz Modular Units= New text and new quire * = Possible UniMod change Quire Irregularities= Changes made in rebinding Content Pages Quires Original Quires Folios Scribe (M&S) Scribe (G) Modular Unit Quire Irreg. Material Unit UniMep 1–40 UniMat2 41–48 X 4 4 A A 49–54 4 3 A A 55–74 4 X 5 10 A A 75–94 5 6 10 A A 95–114 6 7 10 A A 115–134 7 8 10 A A 135–154 8 9 10 A A Mep1 155–174 9 10 10 A A 175–194 10 11 10 A A 195–214 11 12 10 A A 215–234 12 13 10 A A 235–254 13 14 10 A A 255–274 14 15 10 A A 275–294 15 16 10 A A 295–314 16 17 10 A A –1 Kgdms 19:11 315–334 17 18 10 A A 1 Kgdms 19:11– 335–354 18 19 10 B B UniMat1 355–374 19 20 10 B B –2Kgdms 375–394 20 21 10 B B * 3Kgdms– 395–414 21 22 10 B B * 415–434 22 23 10 B B Mep1 435–454 23 24 10 B B UniMod1 455–474 24 25 10 B B 475–494 25 26 10 B B Material Units • UniMat1= Original parchment • UniMat2= Replacement parchment Mise-en-page Units • Mep1 (Blue)= 3 Columns; 42 Lines • Mep1 (Orange)= 3 Columns; 44 Lines • Mep1 (Red)= 3 Columns; 40 Lines • Mep2= 2 Columns; 42 Lines • Mep1/2= Both 2 and 3 Columns; 42 Lines • Mep3= 1+ Empty Columns APPENDICES 266 495–514 26 27 10 B B 515–534 27 28 10 B B 535–554 28 29 10 B B Mep1 555–574 29 30 10 B B 575–594 30 31 10 B B Mep1 595–614 31 32 10 B B 2 Esd–Psalms 615–634 32 33 10 B/A B Mep1/2; 3 635–654 33 34 10 A B –Ps 77:71b 655–674 34 35 10 A B Mep2 Ps 77:71b– 675–694 35 36 10 A C 695–706 37 6 UniMat2 Mep1 707–726 37 38 10 A C 727–746 38 39 10 A C 747–766 39 40 10 A C Mep2 767–786 40 41 10 A C 787–806 41 42 10 A C 807–826 42 43 10 A C 827–846 43 44 10 A C 847–866 44 45 10 A C 867–886 45 46 10 A C Sir–Esth 887–906 46 47 10 A C Mep1/2 907–926 47 48 10 A C –Tobit 927–946 (927–944) 48 49 10 (8) A C Mep3 Hosea– 947–962 (945–964) 49 X 50 8 (10) B B 963–986 (965–984) 50 X 51 12 (10) B B 987–1002 (985–1004) 51 X 52 8 (10) B B 1003–1024 (1005–1024) 52 X 53 11 (10) B B 1025–1044 53 54 10 B B 1045–1064 54 55 10 B B 1065–1084 55 56 10 B B Mep1 1085–1104 56 57 10 B B 1105–1124 57 58 10 B B UniMat1 1125–1144 58 59 10 B B 1145–1164 59 60 10 B B 1165–1184 60 61 10 B B 1185–1204 61 62 10 B B UniMod2 1205–1224 62 63 10 B B Dan–Matt 1225–1244 63 64 10 B B Mep3 1245–1264 64 65 10 B B 1265–1284 65 66 10 B B 267 Mark–Luke 1285–1304 66 67 10 B B Mep3 1305–1324 67 68 10 B B 1325–1344 68 69 10 B B 1345–1364 69 70 10 B B Mep1 1365–1384 70 71 10 B B 1385–1404 71 X 72 10 B B 1405–1424 72 73 10 B B –Jude 1425–1444 73 74 10 B B Romans– 1445–1464 74 75 10 B B 1465–1484 75 76 10 B B 1485–1504 76 77 10 B B UniMod3 1505–1510 77 78 3 B B 1511–1518 78 X 78 4 B B 1519–1536 UniMat2 Mep1 APPENDIX B Lists and Genealogies in B(03)1 Lists - The Twelve Patriarchs (pp. 44A–45A, Gen 49:3–27) - The Decalogue (pp. 71C–72A, Exod 20:2–17) - The Nations that God Will Blot Out (p. 76A, Exod 23:23) - The Helpers from the Twelve Tribes (p. 138A, Num 1:5–15) - The Camps of Israel after Leaving Egypt (pp. 186B–187B, Num 33:5–48) - The Unclean Birds (p. 211C, Deut 14:12–18) - The Nations that Gathered against Joshua (p. 247C, Josh 9:1)2 - Five Kings of the Amorites (p. 250A, Josh 10:5) - Defeated Kings of Joshua (p. 254A, Josh 12:10b–22) - Five Rulers of the Philistines (p. 254B, Josh 13:3) - Pasturelands of Aaron’s Descendants (p. 264A, Josh 21:14–16) - The Golden Seat Offerings (p. 316C, 1 Kgdms 6:17) - The Chiefs of Edom (p. 485B, 1 Chr 1:51–54) - Blessings to the Lord (pp. 1215B–1216B, LXX Dan 3:52–90) - Beatitudes (p. 1239A, Matt 5:3–11) - Greetings to the Romans (p. 1460C, Rom 16:3–23) - Paul’s Dangers (p. 1486A, 2 Cor 11:26) - Diaspora Cities (p. 1430B, 1 Pt 1:1) Genealogies - The Genealogy of Moses and Aaron (p. 53B–C, Exod 6:14–26) - The Genealogy of David (p. 309C, Ruth 4:18–22) - The Patriarchal Genealogies (pp. 484A–495A, 1 Chr 1:1ff.) - The Genealogy of Judith (p. 917C, Jdt 8:1) - Matthew’s Genealogy of Jesus (p. 1235A–C, Matt 1:1–17) - Luke’s Genealogy of Jesus (pp. 1309C–1310B, Luke 3:23–28) 1 This chart does not include all examples of such formatting, as some lists are only partially distinguished by new formatting. Updated from Grenz, “Textual Divisions,” 6–7. 2 The lists in Joshua have been already identified in Auld, Joshua, x. APPENDIX C The Titles in B(03) Initial Title Running Title End-Title — γενεσιϲ κατα τουϲ εβδοµηκοντα εξοδοϲ εξοδοϲ λευειτικον λευειτικον αριθµοι αριθµοι δευτερονοµιον δευτερονοµιον (p.215) δευτερονοµιον ιησουϲ ιησουϲ ιησουϲ υιοϲ ναυη κριται κριται κριται ρουθ ρουθ ρουθ βασιλειων α̅ βασιλειων] [α̅ βασιλειων α̅ βασιλειων β̅ βασιλειων] [β̅ βασιλειων β̅ βασιλειων γ̅ βασιλειων] [γ̅ βασιλειων γ̅ βασιλειων δ̅ βασιλειων] [δ̅ βασιλειων δ̅ παραλειποµενων α̅ παραλειποµενων] [α̅ παραλειποµενων α̅ παραλειποµενων β̅ παραλειποµενων] [β̅ παραλειποµενων β̅ εσδραϲ α̅ εσδραϲ] [α̅ εσδραϲ α̅ εσδραϲ β̅ εσδραϲ] [β̅ εσδραϲ β̅ ψαλµοι βιβλοϲ ψαλµων ρν̅ ̅ παροιµιαι παροιµιαι παροιµιαι εκκλησιαστηϲ εκκλησιαστηϲ εκκλησιαστηϲ ασµα ασµα ασµα ιωβ ιωβ ιωβ σοφια σαλωµων σοφια σαλωµων σοφια σαλωµων προλογοϲ προλογοϲ σοφια σειραχ σοφια σειραχ σοφια ιησου υιου σειραχ εσθηρ εσθηρ εσθηρ ιουδειθ ιουδειθ ιουδειθ APPENDICES 272 τωβειτ τωβειτ τωβειτ ωσηε α̅ ωσηε α̅ αµωϲ β̅ αµωϲ β̅ αµωϲ β̅ µειχαιαϲ γ̅ µειχαιαϲ γ̅ µειχαιαϲ γ̅ ιωηλ δ̅ ιωηλ δ̅ ιωηλ δ̅ οβαειου ε ̅ οβαειου ε ̅ οβαειου ε ̅ ιωναϲ ς ̅ ιωναϲ ς ̅ ιωναϲ ς ̅ ναουµ ζ ̅ ναουµ ζ ̅ ναουµ ζ ̅ αµβακουµ η̅ αµβακουµ η̅ αµβακουµ η̅ σοφονιαϲ θ ̅ σοφονιαϲ θ ̅ σοφονιαϲ θ ̅ αγγαιοϲ ι ̅ αγγαιοϲ ι ̅ αγγαιοϲ ι ̅ ζαχαριαϲ ια̅̅ ζαχαριαϲ ια̅̅ ζαχαριαϲ ια̅̅ µαλακιαϲ ιβ̅̅ µαλακιαϲ ιβ̅̅ µαλακιαϲ ιβ̅̅ ησαιαϲ ησαιαϲ ησαιαϲ ιερεµιαϲ ιερεµιαϲ ιερεµιαϲ βαρουχ βαρουχ βαρουχ θρηνοι θρηνοι θρηνοι ιερεµιου επιστολη ιερεµιου επιστολη επιστολη ιερεµιου ιεζεκιηλ ιεζεκιηλ ιεζεκιηλ δανιηλ δανιηλ δανιηλ κατα µαθθαιον κατα] [µαθθαιον κατα µαθθαιον κατα µαρκον κατα] [µαρκον κατα µαρκον κατα λουκαν κατα] [λουκαν κατα λουκαν κατα ιωανην κατα] [ιωανην κατα ιωανην πραξειϲ αποστολων πραξειϲ πραξειϲ αποστολων ιακωβου επιστολη ιακωβου ιακωβου πετρου α̅ πετρου α̅ πετρου α̅ πετρου β̅ πετρου β̅ πετρου β̅ ιωανου α̅ ιωανου α̅ ιωανου α̅ ιωανου β̅ ιωανου β̅ ιωανου γ̅ ιωανου γ̅ ιουδα ιουδα (verso) ιουδα 273 προϲ ρωµαιουϲ προϲ] [ρωµαιουϲ (p.1461) προϲ ρωµαιουϲ εγραφη απο κορινθου προϲ κορινθιουϲ α̅ (p.1461) προϲ] [κορινθιουϲ α̅ προϲ κορινθιουϲ α̅ εγραφη απο εφεσου προϲ κορινθιουϲ β̅ προϲ] [κορινθιουϲ β̅ προϲ κορινθιουϲ β̅ εγραφη απο φιλιππων προϲ γαλαταϲ (p.1488) προϲ] (p.1488) [γαλαταϲ προϲ γαλαταϲ εγραφη απο ρωµηϲ προϲ εφεσιουϲ προϲ] [εφεσιουϲ προϲ εφεσιουϲ εγραφη απο ρωµηϲ προϲ φιλιππησιουϲ προϲ] [φιλιππησιουϲ προϲ φιλιππησιουϲ εγραφη απο ρωµηϲ προϲ κολασσαειϲ προϲ] [κολασσαειϲ προϲ κολασσαειϲ εγραφη απο ρωµηϲ προϲ θεσσαλονεικειϲ α̅ προϲ] [θεσσαλονεικειϲ α̅ προϲ θεσσαλονεικειϲ α̅ εγραφη απο αθηνων προϲ θεσσαλονεικειϲ β̅ προϲ θεσσαλονεικειϲ β̅ εγραφη απο αθηνων προϲ εβραιουϲ προϲ] [εβραιουϲ APPENDIX D Orthographic Corrections in B(03) 1.1 ORTHOGRAPHIC CORRECTIONS IN 1 KINGDOMS ORTH SCRIBE # EXAMPLES ø → ϲ Α 1 βερσεχθαν Β 0 — α → αι Α 0 — Β 1 ποιµαινοντεϲ αι → ε Α 1 εδεϲµατων Β 3 γνωτε; µε; πεδιω ε → αι Α 0 — Β 1 εντεταλµαι ε → ο Α 9 εξολοθρ– Β 3 εξολοθρ–; οµωµοκαµεν ει → η Α 1 απελευϲη Β 6 γνωϲη; ειϲαγηοχατε ει → ι Α 4 µικρ– Β 7 ιδε; µικρ– ι → ει Α 22 αγγειων; λειτουργων; οραϲει Β 3 πληµµελεια; υπολειψοµαι κ → γ Α 0 — Β 1 παγιδευειϲ κ → χ Α 4 ουχ Β 2 µοχλων; ουχ ν -γ Α 1 ϲυγχυϲιϲ Β 6 απεκταγκα; εγγαϲτριµυθ- ν → µ Α 2 ϲυµπτωµα Β 2 εµπεπηγοϲ ο → ω Α 3 εωρακα Β 5 αθωωθηϲεται; εωρακα χ → κ Α 0 — Β 1 ουκ 1.2 ORTHOGRAPHIC CORRECTIONS OF PROPER NAMES IN 1 KINGDOMS ΝΑΜΕ TYPE SCRIBE # EXAMPLES αχειµελεχ β → χ Β 1 αβειµελεχ → αχειµελεχ (1 Kgdms 26:6) βαιθωρων ø → ϲ Α 1 βαιθωρων → βαιθϲωρων (1 Kgdms 13:5) APPENDICES 276 2.1 ORTHOGRAPHIC CORRECTIONS IN PSALMS ORTH SCRIBE # EXAMPLES ø → ν B 2 δαµαλεϲιν; κατηϲχυνθηϲαν 1 C 0 — α → ε B 3 εξηρευν- C 1 ειϲελθετω α → ει B 1 ελεει C 0 — α → η B 1 κορην C 0 — αι → α B 1 ποιµανει C 0 — αι → ε B 27 κρινετε; -εϲθε; µε; πεδια C 0 — αι → ο B 0 — C 1 οικοδοµειτο γ → κ B 0 — C 2 εκλεκτ- ε → α B 0 — C 1 τεϲϲαρακοντα ε → αι B 13 επαινειται; ειπαιϲαν; ελαιον C 2 ελαιω ε → ει B 3 ηχρειωθηϲαν; καθιειται C 0 — ε → η B 3 ϲκοτοµηνη; ηβουληθην; επιλαθη C 0 — ε → ο B 9 εξωλοθρ-; εξολοθρ- C 7 εξωλοθρ-; εξολοθρ- ει → ε B 0 — C 1 ουκ ει → η B 1 πληρηϲ C 0 — ει → ι B 1 µικρου C 0 — η → α B 1 ανεωξεν C 0 — η → ε B 2 ευφραναϲ; υπερεδυναµωϲαν C 0 — η → ει B 2 ει; οψει C 1 οψει ι → ει B 147 (ε)ταπεινω-; γειτοϲιν; ειδον; δυναµει; ϲηµεια; θλιψειϲ 1 This is not properly a correction of a moveable-nu, since the nu is already present in the supralinear macron. 277 C 26 (ε)ταπεινω-; (ε)λειτουργ-; γειτοϲιν; ειδοϲαν; (προ)ειδεϲ κ → χ B 1 ουχ C 1 µοχλουϲ µ → µµ B 1 απερριµµαι C 0 — µ → ν B 0 — C 2 εν µεϲω [εµµεϲω]* ν → γ B 24 εγκαταλ(ε)ιπ-; εγκαυχα; ϲυγκλειϲον C 2 ϲυγκαθηϲθαι; εγγραπτον ν → λ B 3 ϲυλληµφθητωϲαν; ϲυλλυπουµενον; ϲυλλαµβανονται C 0 — ν → µ B 8 εµπαγ-; εµπαιγµων; ϲυµβιβω C 3 ϲυµπαρεγενετο; ϲυµπαραϲτηϲεται; ϲυµπροϲεϲται ν → ϲ B 2 ϲυϲϲειοντοϲ; ϲυϲϲειϲει C 0 — ο → ω B 3 αθωοιϲ; προωρωµην; ϲαλωµων C 0 — οι → ω B 0 — C 1 ωκτειρηϲεν ου → ω B 1 γηρωϲ 2 C 0 — ρ → ρρ B 12 ερρυϲ-; (επι/απο)ρριψ-; περρηϲιαϲοµαι C 0 — ϲ → ϲϲ B 0 — C 1 εννοϲϲευϲουϲιν υ → οι B 1 ηνοιξα C 0 — χ → κ B 3 ουκ C 0 — ω → ο B 3 αγαλλιαϲοµεθα; εξολοθρ- C 0 — 3.1 ORTHOGRAPHIC CORRECTIONS IN THE GOSPELS ORTH SCRIBE # EXAMPLES 2 This may also be considered a substitution from γεροϲ → γεραϲ in the genitive. APPENDICES 278 ø → α B 1 µετα αλληλων ø → ν B 11 ϲυλλεγουϲιν; εκειθεν; ειπεν; περαν α → αι B 1 αιγιαλον α → ε B 5 ανεπεϲαν; κατεβαινεν; ενι; ερευνατε α → η B 1 πληµµυραϲ α → ο B 21 προϲηλθον; ειδοµεν; ειδον; εγογγυζον; µεϲονυκτιον; ηλθον; ανευρον; ευρον; απεθανον; ευροµεν; ηλπιζοµεν; ελαβον; ειπον α → ω B 1 ανωγαιον αι → α B 1 καταβαινοντων αι → ε B 27 ηκουϲατε; δεξαϲθε; ευρηϲετε; αφετε; κληθητε; διαλογιζεϲθε; λεγετε; οϲφυεϲ; θεωρουντεϲ; κατιϲχυϲητε; εϲθητα; θυγατερεϲ; ϲε β → ββ; ττ → τ B 9 κραββατον γ → γγ B 1 µογγιλαλον ε → α B 12 τεϲϲαρακοντα; εκαθαριϲθη; αφηκατε; αγγαρευουϲιν; ειπαϲ; πληρωϲατε; εορακαϲ ε → αι B 14 κεραια; καθαριζονται; γινωϲκεται; προϲκαλειται; ενδυϲαϲθαι ; αναπεϲαι; κραιπαλη; εργαζεϲθαι; τεταρταιοϲ ε → ει B 3 ειπειν ε → η B 8 ερρηθη; πιϲτευητε; οψηϲθε ε → ι B 5 αλιειϲ; ευριϲκει ε → ο B 3 επελαθοντο; εξεδοτο ει → ε B 1 ϲυνετε ει → η B 4 οµολογηϲηvid; κληθηϲ; εφη; διψηϲη ει → ι B 30 µικροϲ; µοιχαλιϲ; ιαται; εξιϲταντο; ιϲχυϲαϲ; τριϲ; ουχι; ιδη ει → ιει B 4 πιειν; ϲυνιειναι η → α B 4 αναϲταϲει; µαχαιρα η → ε B 2 εκλειϲθη; επιθυµηϲεται η → ει B 3 ει; οικοδοµειϲθαι; υψωθηϲει θ → τ B 7 κατ ιδιαν ι → ε B 2 λεγεων ι → ει B 42 επει; οφειϲ; οικειακοι; ειϲτηκει; ορει; αγαπηϲειϲ; τετειµηµενου; ωφελει; καταλειψει; ιερατειαϲ; 279 τρειϲ; δανειζετε; πολει; χρεοφειλεται; ϲυγγενειϲ; εκλειπη; ηµιϲεια; ϲηµειον; επειϲελευϲεται; φαγειν; φιλονεικια; ειϲιν; µειϲειν; δειπνου; δειλιατω; αληθεια ι → η B 1 ληνον3 κ → χ B 7 ουχ; χιτωναϲ λ → λλ B 1 ανταλλαγµα ν → ø B 1 µακαριουϲιν ̇ ν → γ B 10 παλιγγενεϲια; ϲυγκαλουϲιν; εγγεγραπται; εγκακειν; ϲυγχρωνται ν → λ B 3 ϲυλλαλουντεϲ; ϲυλλυπουµενοϲ ν → µ B 8 ϲυµπορευονται; εµφοβων; ϲυµµαθηταιϲ; εµβρειµωµενοϲ; ϲυµπνιγει ξ → κ B 1 εκ ο → α B 4 δυναµαι; ειπον ο → ε B 1 αφειενται ο → ω B 9 µειζων; εϲτωϲ; εωρακεν οι → ω B 1 ωκοδοµηθη ου → ω B 1 θεωρηϲωϲιν ρ → ρρ B 13 ερριµµενοι; επιρραπτει; παρρηϲια; διαρρηξαϲ; προϲερρηξεν ϲ → ϲϲ B 1 περιϲϲον τ → θ B 1 αποκαθιϲτανει υ → οι B 1 διηνοιγεν ω → ο B 1 οφειλοµεν 3.2 ORTHOGRAPHIC CORRECTIONS OF PROPER NAMES IN THE GOSPELS ΝΑΜΕ TYPE SCRIBE # EXAMPLES βηθϲαιδα δ → θ B 1 βηδϲαιδα → βηθϲαιδα (Luke 10:13) βηθφαγη δ → θ; ø → ϲ B 1 2 βηδφαγη → βηθϲφαγη (Mark 11:1); βηθφαγη → βηθϲφαγη (Matt 21:2; Luke 19:29) βηθανια ø → ν B 2 βηθανια → βηθανιαν (Matt 21:17; Mark 11:1) γεννεϲαρετ θ → τ B 1 γεννηϲαρεθ → γεννηϲαρετ (Mark 6:53) γεθϲηµανι τ → θ B 1 γετϲηµανει → γεθϲηµανει (Mark 14:32) ιεροϲολυµα ø → ι B 1 εροϲολυµα → ιεροϲολυµα (Luke 13:22) ηλιαϲ α → ου B 1 ηλεια → ηλειου (Luke 1:17) ιαρετ τ → δ B 1 ιαρετ → ιαρεδ (Luke 3:37) ιωαννηϲ ει → η ν → νν B 3 1 ιωανει → ιωανη (Matt 11:4; Luke 7:18; Luke 7:22) 3 This may also be a nonsense reading in context. See LSJ, s.v. λίνον. APPENDICES 280 ιωανηϲ → ιωαννηϲ (Matt 3:1) µατταθιαϲ θ → τ B 1 µαθθαθιου → µατθαθιου (Luke 3:25) µαθθαιοϲ θ → τ B 4 µαθθαιοϲ → µατθαιοϲ (Matt 10:3) µαθθαιον → µατθαιον (Matt 9:9; Mark 3:18; Luke 6:15) µαθθατ θ → τ B 1 µαθθατ → µατθατ (Luke 3:29) µατθαν θ → τ B 2 µαθθαν → µατθαν (Matt 1:15, 2x) µωυϲηϲ η → ει B 1 µωυϲη → µωυϲει (Mark 9:4) ναζαρα α → ετ B 2 ναζαρα → ναζαρετ (Matt 4:13; Luke 4:16) ναζαρεθ θ → τ B 2 ναζαρεθ → ναζαρετ (Luke 2:39; Luke 2:51) ϲαµαρεια ι → ει B 1 ϲαµαριαϲ → ϲαµαρειαϲ (Luke 17:11) ϲαρεπτα π → φ; τ → θ B 1 ϲαρεπτα → ϲαρεφθα (Luke 4:26) 3.3 CORRECTIONS OF AUGMENT AND REDUPLICATION IN THE GOSPELS CHANGE SCRIBE # EXAMPLES Augment B 5 2 8 4 1 3 _προεφητευϲεν ε̸ιϲτηκειϲαν ηδυνατο ειργαϲατο ηµελλεν διηκονει Reduplication B 2 µεµνηϲτευµενη 4.1 ORTHOGRAPHIC CORRECTIONS IN ACTS ORTH SCRIBE # EXAMPLES ø → ν B 2 αρχουϲιν; π̣αϲιν ø → ϲ B 3 αχριϲ α → αι B 2 δειϲιδαιµονεϲτερουϲ; δοθηϲεται α → ε B 2 αναφανεντεϲ vid; τεθεραπευµενον α → η B 2 λυδδηϲ; οδηγηϲει α → ο B 1 παραγενοµενοϲ αι → ε B 4 βοηθειτε; εθεϲθε; εϲθητι; θορυβειϲθε β → ββ; ττ → τ B 2 κραββατον ε → α B 10 τεϲϲαρακοντα ε → αι B 4 αναιρειϲθαι; εξαιφνηϲ; παραιτουµαι; προϲερχεϲθαι ε → ι B 2 τεϲϲαριϲκαιδεκατη ε → ο B 2 διεδιδοτο; εξολοθρευθηϲεται ει → ι B 4 εϲθ_ι̇ειν; _̇ιαϲατο; _̸ιϲτηκειϲαν ; µ_̇ικρω ει → ιει B 3 πιειν; επιεικεια 281 η → α B 1 µαχαιρα ι → ει B 14 ατταλειαν; δειϲιδαιµονιαϲ; εξαλειφθηναι; επιδεικνυϲ; επικουρειων; καταλελειµµενοϲ; κολωνεια; παραχειµαϲιαν; πεπιϲτευκειϲαν; πολιτειαν; ϲυγγενειαϲ; ϲυνειδυιηϲ; ψηλαφηϲειαν ι → υ B 1 πρυµνα ν → ø B 1 ειπε_̇ ν → γ B 8 εγκαλουµενον; εγκατελειφθη; εγκοπτω; ϲυγκαθηµενοι; ϲυγκαταβαντεϲ ν → λ B 1 ϲυλλαληϲαϲ ν → µ B 7 εµµενειν; ϲυµπαραλαβειν; ϲυµπαροντεϲ; ϲυµπεριλαβων; ϲυµπληρουϲθαι ο → α B 1 δυναµενου ο → ε B 1 αρτεµωνα ο → ω B 3 ξυρηϲωνται; προωρωµην; ωκοδοµηϲεν ρ → ρρ B 4 αναντιρρητων; ερριψαν vid; περιρρηξαντεϲ ϲ → ϲϲ B 1 απαλλαϲϲεϲθαι υ → οι B 1 διηνοιγµενουϲ χ → κ B 1 ουκ 4.2 ORTHOGRAPHIC CORRECTIONS OF PROPER NAMES AND THE AUGMENT IN ACTS ΝΑΜΕ TYPE SCRIBE # EXAMPLES αδραµυττηνοϲ τ → ττ B 1 αδραµυντηνω → αδραµυττηνω (Acts 27:2) αλεξανδρινοϲ ει → η; ι → η B 1 1 αλεξανδρεινον → αλεξανδρηνον (Acts 27:6) αλεξανδρινω → αλεξανδρηνω (Acts 28:11) αρεοπαγιτηϲ ο → ω B 1 αρεοπαγειτηϲ → αρεωπαγιτηϲ (Acts 17:34) ιϲαακ α → αα B 1 ιϲακ → ιϲαακ (Acts 7:8) ιωαννηϲ ν → νν B 1 ιωανου → ιωαννου (Acts 1:22) µαθθαιοϲ θ → τ B 1 µαθθαιοϲ → µατθαιον (Acts 1:13) µαθθιαϲ θ → τ B 2 µαθθιαν → µατθιαν (Acts 1:23, 26) µελιτη νη → ø B 1 µελιτηνη → µελιτη (Acts 28:1) νικανωρ ο → ω B 1 νικανορα → νικανωρα (Acts 6:5) φηλιξ ι → η B 1 φιλιξ → φηλιξ (Acts 24:24) CHANGE SCRIBE # EXAMPLES Augment B 1 1 ειργαζοντο ηδυνατο APPENDICES 282 1 1 ηυχοντο προϲηυξατο 5.1 ORTHOGRAPHIC CORRECTIONS IN THE CATHOLIC EPISTLES ORTH SCRIBE # EXAMPLES ø → ν B 1 επιεικεϲιν α → ε B 3 ελαβετε; εξηρευνηϲαν; ερευνωντεϲ αι → ε B 6 βλαϲφηµουντεϲ; εξοµολογειϲθε; ορατε; ε → αι B 4 αντιταϲϲεται; εκδεχεται; επιτελειϲθαι; φευξεται ε → η B 1 νηφοντεϲ ει → ι B 4 κατ_ι̇ωται; µ_̇ικρον; ν_̸ικη; ν_̸ι̇κηϲαϲα ι → ει B 9 αλαζονεια; επιτηδεια; εριθειαν; κακοπαθειαϲ; µεγαλειοτητοϲ; ϲυνειδηϲιν; ταπεινωϲει; ωφελειαϲ ν → ø B 1 ειπε_̇ ν → γ B 3 εγκατοικων; ελεγξιν; ϲυγκληρονοµοιϲ ο → ω B 8 εωρακαµεν; πρωιµον ρ → ρρ B 2 επιρριψαντεϲ; ερρυϲατο χ → κ B 2 ουκ 5.2 ORTHOGRAPHIC CORRECTIONS OF PROPER NAMES AND THE AUGMENT IN THE CATHOLIC EPISTLES ΝΑΜΕ TYPE SCRIBE # EXAMPLES λωτ θ → τ B 1 λωθ → λωτ (2 Peter 2:7) CHANGE SCRIBE # EXAMPLES Augment B 1 1 επροεφητευϲεν ειργαϲαµεθα 6.1 ORTHOGRAPHIC CORRECTIONS IN THE CORPUS PAULINUM ORTH SCRIBE # EXAMPLES ø → ν B 1 ουϲιν ø → ϲ B 3 αχριϲ α → αι B 1 εκκληϲιαιϲ α → αυ B 1 αναϲταυρουνταϲ α → ε B 2 ανεξερευνητα; ερευνα α → ο B 1 ϲυναπαγοµενοι αι → ε B 4 δυνηϲεϲθε; εϲτε; χαριϲαϲθε γ → κ B 1 απεκδυϲει γ → ν B 1 ειλικρενειαϲ ε → α B 3 τεϲϲαρακοντα 283 ε → αι B 13 αιϲθητηρια; αϲπαζεται; γραφεϲθαι; ελαιον; περιαιρειται ε → η B 3 ερρηθη ε → ο B 1 παρεδιδοτο ει → ι B 8 εκρ_̸ινα; επιποθ_̸ιαν; µ_̇ικρα; ν_ι̇κω ει → ιει B 3 πιειν ζ → ϲ B 1 ϲβεννυτε η → ει B 1 διψει ι → ει B 52 αδιαλειπτωϲ; δεηϲειϲ; ερειθειαϲ; ερµηνειαν; ευχαριϲτει; ηχρειωθηϲαν; θηϲαυριζειν; κακοηθειαϲ; λειτουργια; λογειαι; µεθοδειαϲ; οφειλει; πεπειϲµαι; ϲπειρειϲ; υπολειµµα; φυϲει ι → η B 1 ληµµα λ → λλ B 1 παρηγγελλοµεν µ → µµ B 1 ϲυµµορφιζοµενοϲ ν → γ B 32 εγγεγραµµενη ; εγκακειν; εγκεντριϲαι; εγκοπην; εγκριναι; ϲυγγνωµην; ϲυγκαµψον; ϲυγκληρονοµοι; ϲυγκοινωνειτε; ϲυγκριναι ; ϲυγχαιρει ν → λ B 1 ϲυλλαµβανου ν → µ B 19 εµπεριπατηϲω; ϲυµβαϲιλευϲωµεν; ϲυµβιβαζοµενον; ϲυµµαρτυρει; ϲυµµετοχα; ϲυµµιµηται; ϲυµπαθηϲαι; ϲυµπαϲχει; ϲυµπολειται; ϲυµφηµι; ϲυµψυχοι ν → ϲ B 1 ϲυϲϲωµα ο → ω B 3 αγιωϲυνη; εωρακεν οι → ω B 1 επωκοδοµηϲεν ου → ω B 1 καταδουλωϲωϲιν4 π → φ B 1 αφ’ ων ρ → ρρ B 5 ερρυϲατο; παραρρυωµεν; παρρηϲιαν ϲ → ϲϲ B 1 γλωϲϲαιϲ τ → θ B 1 µεθ’ ορκωµοϲιαϲ φ → π B 2 επ’ ελπιδι; απιδω χ → κ B 1 ουκ ω → ο B 1 οφειλετε ω → ου B 1 ελεουντοϲ 4 This may be a substitution of the future active to an aorist subjunctive. APPENDICES 284 6.2 ORTHOGRAPHIC CORRECTIONS OF PROPER NAMES AND THE AUGMENT IN THE CORPUS PAULINUM ΝΑΜΕ TYPE SCRIBE # EXAMPLES αµπλιατοϲ ø → ν B 1 αµπλιατον → αµπλιαντον (Romans 16:8) αριϲτοβουλοϲ ο → ου B 1 αριϲτοβολου → αριϲτοβουλου (Romans 16:10) κεγχρεαι ν → γ B 1 κενχρεαιϲ → κεγχρεαιϲ (Romans 16:1) λαοδικεια ι → ει B 4 λαοδικια → λαοδικεια (Col 2:1; 4.13, 15, 16) µωυϲηϲ η → ει B 1 µωϲη → µωϲει (Romans 9:15) CHANGE SCRIBE # EXAMPLES Augment B 2 3 ευδοκηϲεν κατειργαϲατο (-ϲθη) APPENDIX E Quires [49–53]—Current Structure and Probable Original Structure (p.40) Bibliography Abbot, Ezra. “On the Comparative Antiquity of the Sinaitic and Vatican Manuscripts of the Greek Bible.” Journal of the American Oriental Society 10 (1872): 189–200. Adams, Sean A. Baruch and the Epistle of Jeremiah: A Commentary Based on the Texts in Codex Vaticanus. Septuagint Commentary Series. Leiden: Brill, 2014. Aejmelaeus, Anneli. “Challenges in Preparing the Critical Edition of 1 Samuel.” Septuagint and Coptic Studies. Edited by Felix Albrecht and Frank Feder. De Septuaginta Investigationes. Göttingen: Vandenhoeck & Ruprecht, forthcoming. ———. “Kaige Readings in a Non-Kaige Section in 1 Samuel.” Pages 169–84 in The Legacy of Barthélemy: 50 Years After Les Devanciers D’Aquila. Edited by Anneli Aejmelaeus and Tuukka Kauhanen. De Septuaginta Investigationes 9. Göttingen: Vandenhoeck & Ruprecht, 2017. ———. “New Perspectives on the Textual Character of Codex Vaticanus: A New Chapter in the Textual Criticism of the Septuagint, Especially the Books of Samuel.” Svensk Exegetisk Årsbok 85 (2020): 1–21. Agati, Maria Luisa. The Manuscript Book: A Compendium of Codicology. Translated by Colin W. Swift. Rev. and upd. Studia Archeologica 214. Roma: L’Erma di Bretschneider, 2017. Aitken, James K. “The Origins of ΚΑΙ ΓΕ.” Pages 21–40 in Biblical Greek in Context: Essays in Honour of John A. L. Lee. Edited by Trevor V. Evans and James K. Aitken. Biblical Tools and Studies 22. Leuven: Peeters, 2015. Aland, Barbara, Kurt Aland, Johannes Karavidopoulos, Carlo M. Martini, and Bruce M. Metzger, eds. Novum Testamentum Graece. 27th ed. Stuttgart: Deutsche Bibelgesellschaft, 1993. ———, eds. Novum Testamentum Graece. 28th ed. Stuttgart: Deutsche Bibelgesellschaft, 2012. Aland, Barbara, Kurt Aland, Gerd Mink, Holger Strutwolf, and Klaus Wachtel, eds. Novum Testamentum Graecum: Editio Critica Maior IV: Catholic Letters: Part 1: Text. 2nd ed. Stuttgart: Deutsche Bibelgesellschaft, 2013. ———, eds. Novum Testamentum Graecum: Editio Critica Maior IV: Catholic Letters: Part 2: Supplementary Material. 2nd ed. Stuttgart: Deutsche Bibelgesellschaft, 2013. Aland, Barbara, Kurt Aland, Gerd Mink, and Klaus Wachtel, eds. Novum Testamentum Graecum: Editio Critica Maior IV: Catholic Letters: Part 1: Text: Installment 4: The Second and Third Letter of John, the Letter of Jude. Stuttgart: Deutsche Bibelgesellschaft, 2005. Aland, Barbara, and Klaus Wachtel, eds. Text und Textwert der griechischen Handschriften des Neuen Testaments: V. Das Johannesevangelium: 1. Teststellenkollation der Kapitel 1-10. Arbeiten zur neutestamentlichen Textforschung 35/36. Berlin: De Gruyter, 2005. Aland, Kurt, ed. Repertorium der griechischen christlichen Papyri. I, Biblische Papyri, Altes Testament, Neues Testament, Varia, Apokryphen. Patristische Texte und Studien 18. Berlin: De Gruyter, 1976. BIBLIOGRAPHY 288 Aland, Kurt, and Barbara Aland. The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism. Translated by Erroll F. Rhodes. 2nd ed. Grand Rapids, MI: Eerdmans, 1995. Aland, Kurt, Barbara Aland, Klaus Wachtel, and Klaus Witte, eds. Text und Textwert der griechischen Handschriften des Neuen Testaments: IV. Die synoptischen Evangelien: 1. Das Markusevangelium. Arbeiten zur neutestamentlichen Textforschung 26/27. Berlin: De Gruyter, 1998. ———, eds. Text und Textwert der griechischen Handschriften des Neuen Testaments: IV. Die synoptischen Evangelien: 2. Das Matthäusevangelium. Arbeiten zur neutestamentlichen Textforschung 28/29. Berlin: De Gruyter, 2003. ———, eds. Text und Textwert der griechischen Handschriften des Neuen Testaments: IV. Die synoptischen Evangelien: 3. Das Lukasevangelium. Arbeiten zur neutestamentlichen Textforschung 30/31. Berlin: De Gruyter, 2003. Aland, Kurt, Annette Benduhn-Mertz, Gerd Mink, and Horst Bachmann, eds. Text und Textwert der griechischen Handschriften des Neuen Testaments: II. Die Paulinischen Briefe. 4 vols. Arbeiten zur neutestamentlichen Textforschung 16–19. Berlin: De Gruyter, 1991. Albrecht, Felix. “Report on the Göttingen Septuagint.” Textus 29 (2020): 201–20. Amelote, Denis. Le Nouveau Testament de Nostre Seigneur Jesus-Christ. 2 vols. Paris: Muguet, 1687. Amphoux, Christian-Bernard. “Les circonstances de la copie du Codex Vaticanus (Vat. Gr. 1209).” Pages 157–76 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. Andrist, Patrick. “Au croisement des contenus et de la matière : les structures des sept pandectes bibliques grecques du premier millénaire.” Scrineum Rivista 17 (2020): 3–106. ———. “La structure des Codex Vaticanus, Alexandrinus et Sinaiticus: questions ouvertes sur le canon, la fabrication et la circulation de ces bibles.” Bibliologia 40 (2015): 11–37. ———, ed. Le manuscrit B de la Bible (Vaticanus graecus 1209): introduction au fac-similé: actes du colloque du Genève (11 juin 2001): contributions supplementaires. Histoire du texte biblique = Studien zur Geschichte des biblischen textes 7. Lausanne: Éditions du Zèbre, 2009. ———. “Le milieu de production du Vaticanus graecus 1209.” Pages 135–55 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. ———. “The Physiognomy of Greek Contra Iudaeos Manuscript Books in the Byzantine Era: A Preliminary Survey.” Pages 549–86 in Jews in Byzantium: Dialectics of Minority and Majority Cultures. Edited by Robert Bonfil. Jerusalem Studies in Religion and Culture 14. Leiden: Brill, 2012. Andrist, Patrick, Paul Canart, and Marilena Maniaci. La syntaxe du codex: essai de codicologie structurale. Bibliologia 34. Turnhout: Brepols, 2013. Askeland, Christian. “Dating Early Greek and Coptic Literary Hands.” Pages 457–89 in The Nag Hammadi Codices and Late Antique Egypt. Edited by Hugo Lundhaug and Lance Jenott. Studien und Texte zu Antike und Christentum 110. Tübingen: Mohr Siebeck, 2018. 289 Baethgen, Friedrich. “Der textkritische Werth der alten Uebersetzungen zu den Psalmen.” Jahrbücher für protestantische Theologie 8 (1882): 407–8. Banik, Gerhard. “Ink Corrosion - Chemistry.” The Iron Gall Ink Website, 1998. https://irongallink.org/igi_index22a4.html. Barthélemy, Dominique. Devanciers d’Aquila: Première publication intégrale du texte des fragments du Dodécaprophéton. Vetus Testamentum, Supplements 10. Leiden: Brill, 1963. Bartolocci, Giulio. “Notes sur le manuscrit B de la Bible.” Supplément grec 53. Bibliothèque nationale de France, 1669. https://gallica.bnf.fr/view3if/ga/ark:/12148/btv1b525100460 Batovici, Dan. “Two B Scribes in Codex Sinaiticus?” The Bulletin of the American Society of Papyrologists 54 (2017): 197–206. Bausi, Alessandro, and Eugenia Sokolinski, eds. Comparative Oriental Manuscript Studies: An Introduction. Hamburg: Tredition, 2015. Bianchini, Giuseppe. Evangeliarium quadruplex Latinae versionis antiquae. Partis primae. Volumen secundum. Roma, 1749. Birch, Andreas. Kritisk Beskrivelse over graeske Haandskrifter af det Nye Testamente. Copenhagen: Schultz, 1785. ———. “Nachricht vom Vatikanischen Codex, nebst einigen kurzen Bemerkungen über andere handschriften in Rom.” Orientalische und exegetische Bibliothek 23 (1785): 138–63. ———. Quatuor Evangelia Græce, cum variantibus a textu lectionibus codd. MSS. Bibliothecæ Vaticanæ, Barberinæ, Laurentianæ, Vindobonensis, Escurialensis, Havniensis Regiæ, quibus accedunt lectiones versionum, Syrarum, Veteris, Philoxenianæ, et Hierosolymitanæ, jussu et sumtibus regiis. Copenhagen: J. F. Schultz, 1788. ———. Variae lectiones ad textum IV evangeliorum. Copenhagen: C. G. Proft, 1801. Bird, Michael F. Gospel of the Lord: How the Early Church Wrote the Story of Jesus. Grand Rapids, MI: Eerdmans, 2014. Birdsall, J. Neville. “The Codex Vaticanus: Its History and Significance.” Pages 33–40 in The Bible as Book: The Transmission of the Greek Text. Edited by Scot McKendrick and Orlaith O’Sullivan. London: British Library, 2003. Blass, Friedrich. Grammar of New Testament Greek. Translated by H. St J. Thackeray. London: Macmillan and Co., 1898. ———. Philology of the Gospels. London: Macmillan and Co., 1898. Blomkvist, Vemund. Euthalian Traditions: Text, Translation and Commentary. Texte und Untersuchungen 170. Berlin: De Gruyter, 2012. Blumell, Lincoln H. Lettered Christians: Christians, Letters, and Late Antique Oxyrhynchus. New Testament Tools, Studies, and Documents 39. Leiden: Brill, 2012. Blumell, Lincoln H., and Thomas A. Wayment, eds. Christian Oxyrhynchus: Texts, Documents, and Sources. Waco, TX: Baylor University Press, 2018. BIBLIOGRAPHY 290 Bobichon, Philippe. “Le lexicon: Mise en page et mise en texte des manuscrits hébreux, grecs, latins, romans et arabes.” École thématique (2009): 1–82. Bogaert, Pierre-Maurice. “Le «Vaticanus», Athanase et Alexandrie.” Pages 135–55 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. Boismard, M.-E., and A. Lamouille. Text occidental des Actes des apôtres: Reconstitution et rehabilitation. 2 vols. Paris: Editions Recherche sur les civilisations, 1984. Bons, Eberhard, and Ralph Brucker. “Psalmoi/The Book of Psalms.” Pages 297–316 in Introduction to the Septuagint. Edited by Siegfried Kreuzer, Translated by David A. Brenner and Peter Altman. Waco, TX: Baylor University Press, 2019. Bousset, Wilhelm. “Die Recension des Hesychius.” Pages 74–110 in Textkritische Studien zum Neuen Testament. Leipzig: J. C. Hinrichs, 1894. Boyd-Taylor, Cameron, Peter C. Austin, and Andrey Feuerverger. “The Assessment of Manuscript Affiliation within a Probabilistic Framework: A Study of Alfred Rahlfs’s Core Manuscript Groupings for the Greek Psalter.” Pages 98–124 in Old Greek Psalter: Studies in Honour of Albert Pietersma. Edited by Robert J. V. Hiebert, Peter J. Gentry, and Claude E. Cox. Journal for the Study of the Old Testament, Supplement Series 332. Sheffield: Sheffield Academic Press, 2009. Brock, Sebastian P. The Recensions of the Septuagint Version of 1 Samuel. Quaderni di Henoch 9. Torino: S. Zamorani, 1996. de Bruin, Wim. “Interpreting Delimiters: The Complexity of Text Delimitation in Four Major Septuagint Manuscripts.” Pages 66–89 in Studies in Scriptural Unit Division. Edited by Marjo C. A. Korpel and Josef M. Oesch. Pericope 3. Assen: Koninklijke Van Gorcum, 2002. Burgon, John William. The Revision Revised. London: William Clowes and Sons, 1883. Callan, Terrance. “Reading the Earliest Copies of 2 Peter.” Biblica 93 (2012): 427–50. Campbell, Constantine R. Paul and Union with Christ: An Exegetical and Theological Study. Grand Rapids, MI: Zondervan, 2012. Canart, Paul. “Le Vaticanus graecus 1209: notice paléographique et codicologique.” Pages 19–45 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. ———. Lezioni di paleografia e di codicologia greca. Città del Vaticano: Biblioteca Apostolica Vaticana, 1980. ———. “Un X qui risque d’être un x : réflexions sur la codification d’un type de réglure.” Gazette du livre médiéval 44 (2004): 53–59. Canart, Paul, and Carlo M. Martini. The Holy Bible: The Vatican Greek Codex 1209 (Codex B) Facsimile Reproduction by Order of His Holiness Paul VI. The New Testament. Introduction. Vatican City: Vatican, 1965. Cardinali, Giacomo. “Le vicende vaticane del codice B della Bibbia dalle carte di Giovanni Mercati : II : i prestiti e le cessioni esterne.” Pages 177–236 in Miscellanea Bibliothecae Apostolicae Vaticanae XXII. Studi e testi 501. Città del Vaticano: Biblioteca Apostolica Vaticana, 2016. 291 ———. «Qui havemo uno spagnolo dottissimo»: gli anni italiani di Pedro Chacón (1570ca.-1581): saggio di ricostruzione bio-bibliografica a partire da carteggi coevi. Studi e testi 513. Città del Vaticano: Biblioteca Apostolica Vaticana, 2017. Carlson, Stephen C. The Text of Galatians and Its History. Wissenschaftliche Untersuchungen zum Neuen Testament 2/385. Tübingen: Mohr Siebeck, 2015. Castelli, Silvia. Johann Jakob Wettstein’s Principles for New Testament Textual Criticism: A Fight for Scholarly Freedom. New Testament Tools, Studies, and Documents 62. Leiden: Brill, 2020. Cavallo, Guglielmo. Ricerche sulla maiuscola biblica. Studi e testi di papirologia 2. Florence: G. Vitelli, 1967. Cavallo, Guglielmo, and Herwig Maehler. Greek Bookhands of the Early Byzantine Period, A.D. 300- 800. Bulletin Supplement 47. London: University of London, Institute of Classical Studies, 1987. Ceriani, Antonio. “L’antico Testament in Greco Seconda i LXX.” Rendiconti. Reale Instituto Lombardo 21 Serie II. (1888): 540–49. Ceulemans, Reinhart. “Theodoret and the Antiochene Text of the Psalms.” Pages 149–64 in XV Congress of the International Organization for Septuagint and Cognate Studies: Munich, 2013. Edited by Wolfgang Kraus, Michaël N. van der Meer, and Martin Meiser. Septuagint and Cognate Studies 64. Atlanta, GA: Society of Biblical Literature, 2016. Clark, Albert C. The Primitive Text of the Gospels and Acts. Oxford: Clarendon Press, 1914. Clarke, Mark. “The Analysis of Manuscript Pigments. Why, What and How?” Gazette du livre médiéval 40 (2002): 36–44. Clemens, Raymond, and Timothy Graham. Introduction to Manuscript Studies. Ithaca, NY: Cornell University Press, 2007. Cohick, Lynn H. Ephesians: A New Covenant Commentary. New Covenant Commentary Series. Cambridge: Lutterworth Press, 2013. Cohick, Lynn H. The Letter to the Ephesians. The New International Commentary on the New Testament. Grand Rapids, MI: Eerdmans, 2020. Cole, Zachary J. “An Unseen Paleographical Problem with Milne and Skeat’s Dictation Theory of Codex Sinaiticus.” Journal of Biblical Literature 135 (2016): 103–7. ———. Numerals in Early Greek New Testament Manuscripts: Text-Critical, Scribal, and Theological Studies. New Testament Tools, Studies, and Documents 53. Leiden: Brill, 2017. ———. “The Date of Codex Sinaiticus; Revisiting Milne and Skeat’s Numerical Argument” (forthcoming). Colwell, Ernest C. “Method in Evaluating Scribal Habits: A Study of P45, P66, P75.” Pages 106–24 in Studies in Methodology in Textual Criticism of the New Testament. New Testament Tools, Studies, and Documents 9. Leiden: Brill, 1969. Cribiore, Raffaella. Writing, Teachers, and Students in Graeco-Roman Egypt. American Studies in Papyrology 36. Atlanta, GA: Scholars Press, 1996. Crosby, Alfred W. The Measure of Reality: Quantification and Western Society, 1250-1600. Cambridge: Cambridge University Press, 1997. BIBLIOGRAPHY 292 Dain, Marie-Alphonse. Les manuscrits. 3rd ed. Collection d’études anciennes. Paris: Les Belles Lettres, 2014. Damschen, Gregor. “Stichometry.” Brill’s New Pauly. http://referenceworks.brillonline.com/entries/brill-s-new-pauly/stichometry-e1123080 Daris, Sergio. “PSI 977: aggiunta minima.” Studia Papyrologica 7 (1968): 7–22. Dickey, Eleanor. Ancient Greek Scholarship: A Guide to Finding, Reading, and Understanding Scholia, Commentaries, Lexica, and Grammatical Treatises, from Their Beginnings to the Byzantine Period. American Philological Association Classical Resources Series 7. Oxford: Oxford University Press, 2006. van Doorn, Nienke L. “Zooarchaeology by Mass Spectrometry (ZooMS).” Pages 7998–8000 in Encyclopedia of Global Archaeology. Edited by Claire Smith. New York, NY: Springer, 2014. Dormandy, Michael Peter. “An Investigation into Textual Characteristics of the Early Greek Majuscule Pandects.” Ph.D. diss., University of Cambridge, 2020. Dunn, James D. G. The Theology of Paul the Apostle. Grand Rapids, MI: Eerdmans, 1998. Easterling, Patricia. “Before Palaeography: Notes on Early Descriptions and Datings of Greek Manuscripts.” Pages 179–88 in Studia Codicologica. Edited by Kurt Treu. Texte und Untersuchungen 124. Berlin: Akademie Verlag, 1977. Ebojo, Edgar Battad. “A Scribe and His Manuscript: An Investigation into the Scribal Habits of Papyrus 46 (P. Chester Beatty II - P. Mich. Inv. 6238).” Ph.D. diss., The University of Birmingham, 2014. Edwards, Grant G. “The Text and Transmission of 2 Thessalonians.” Ph.D. diss., University of Birmingham, 2019. Edwards, James R. “The Hermeneutical Significance of Chapter Divisions in Ancient Gospel Manuscripts.” New Testament Studies 56 (2010): 413–26. Elliott, J. Keith. “An Eclectic Textual Study of the Book of Acts.” The Book of Acts as Church History / Apostelgeschichte als Kirchengeschichte. Edited by Tobias Nicklas and Michael Tilly. Beihefte zur Zeitschrift für die neutestamentliche Wissenschaft 120. Berlin: De Gruyter, 2003. ———. “Manuscripts, the Codex and the Canon.” Pages 613–31 in New Testament Textual Criticism: The Application of Thoroughgoing Principles. Novum Testamentum, Supplements 137. Leiden: Brill, 2010. ———. “T.C. Skeat on the Dating and Origin of Codex Vaticanus.” Pages 65–78 in New Testament Textual Criticism: The Application of Thoroughgoing Principles. Novum Testamentum, Supplements 137. Leiden; Boston: Brill, 2010. ———. “The Distinctiveness of the Greek Manuscripts of the Book of Revelation.” Pages 145–55 in New Testament Textual Criticism: The Application of Thoroughgoing Principles. Novum Testamentum, Supplements 137. Leiden: Brill, 2010. ———. “The Last Twelve Verses of Mark: Original or Not?” Pages 253–74 in New Testament Textual Criticism: The Application of Thoroughgoing Principles. Novum Testamentum, Supplements 137. Leiden: Brill, 2010. 293 Epp, Eldon Jay. “Critical Editions and the Development of Text-Critical Methods, Part 2: From Lachmann (1831) to the Present.” Pages 517–54 in Perspectives on New Testament Textual Criticism, Volume 2: Collected Essays, 2006–2017. Novum Testamentum, Supplements 181. Leiden: Brill, 2021. ———. “Text-Critical Witnesses and Methodology for Isolating a Distinctive D-Text in Acts.” Pages 627–700 in Perspectives on New Testament Textual Criticism, Volume 2: Collected Essays, 2006– 2017. Novum Testamentum, Supplements 181. Leiden: Brill, 2021. ———. “Textual Clusters: Their Past and Future in New Testament Textual Criticism.” Pages 305– 68 in Perspectives on New Testament Textual Criticism, Volume 2: Collected Essays, 2006–2017. Novum Testamentum, Supplements 181. Leiden: Brill, 2021. ———. “The Significance of the Papyri for Determining the Nature of the New Testament Text in the Second Century: A Dynamic View of Textual Transmission.” Pages 345–81 in Perspectives on New Testament Textual Criticism: Collected Essays, 1962-2004. Novum Testamentum, Supplements 116. Leiden: Brill, 2005. Erasmus, Desiderius. The Correspondence of Erasmus: Letters 1122-1251 (1520-1521). Translated by R. A. B. Mynors. Annotated by Peter G. Bietenholz. Collected Works of Erasmus 8. Toronto: University of Toronto Press, 1988. Fabiani, Enrico, and Giuseppe Cozza-Luzi, eds. Bibliorum sacrorum Graecus Codex Vaticanus auspice Leone XIII. Pontifice Maximo cum prolegomena, commentarios et tabulas. Romae: Congregatio de Propaganda fide, 1881. ———, eds. “Commentarius.” Bibliorum sacrorum Graecus Codex Vaticanus auspice Leone XIII. Pontifice Maximo cum prolegomena, commentarios et tabulas. Romae: Congregatio de Propaganda fide, 1881. ———, eds. “Prolegomena.” Bibliorum sacrorum Graecus Codex Vaticanus auspice Leone XIII. Pontifice Maximo cum prolegomena, commentarios et tabulas. Romae: Congregatio de Propaganda fide, 1881. Fee, Gordon. “P75, P66, and Origen: The Myth of Early Textual Recension in Alexandria.” Pages 248– 79 in Bodmer Papyri, Scribal Culture, and Textual Transmission. Edited by Eldon Jay Epp. New Testament Tools, Studies, and Documents 63. Leiden: Brill, 2021. ———. “The Significance of Papyrus Bodmer II and Papyrus Bodmer XIV-XV for Methodology in New Testament Textual Criticism.” Ph.D. diss., University of California, 1966. ———. “The Use of the Definite Article with Personal Names in the Gospel of John.” New Testament Studies 17 (1971): 168–83. Fellows, Richard G. “Are There Distigme-Obelos Symbols in Vaticanus?” New Testament Studies 65 (2019): 246–51. Fernández Marcos, Natalio. The Septuagint in Context: Introduction to the Greek Version of the Bible. Translated by Wilfred G. E. Watson. Atlanta, GA: Society of Biblical Literature, 2009. Finegan, Jack. Encountering New Testament Manuscripts: A Working Introduction to Textual Criticism. London: SPCK, 1975. BIBLIOGRAPHY 294 Fitzmyer, Joseph A. Romans: A New Translation with Introduction and Commentary. The Anchor Bible 33. New York, NY: Doubleday, 1993. Florian, Mary-Lou E. “The Mechanisms of Deterioration in Leather.” Pages 36–57 in Conservation of Leather and Related Materials. Edited by Marion Kite and Roy Thomson. Butterworth- Heinemann Series in Conservation and Museology. Boston, MA: Elsevier Butterworth- Heinemann, 2006. Friedrich, Michael, and Cosima Schwarke. “Introduction — Manuscripts as Evolving Entities.” Pages 1–26 in One-Volume Libraries: Composite and Multiple-Text Manuscripts. Edited by Michael Friedrich and Cosima Schwarke. Studies in Manuscript Cultures 9. Berlin: De Gruyter, 2016. Gäbel, Georg. “‘Western Texts,’ ‘D-Text Cluster,’ ‘Bezan Trajectory,’ or What Else?” Pages 83–136 in Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 3: Studies. Edited by Holger Strutwolf, Georg Gäbel, Annette Hüffmeier, Gerd Mink and Klaus Wachtel. Stuttgart: Deutsche Bibelgesellschaft, 2017. Gallagher, Edmon L., and John D. Meade. The Biblical Canon Lists from Early Christianity: Texts and Analysis. Oxford: Oxford University Press, 2019. Gamble, Harry Y. Books and Readers in the Early Church: A History of Early Christian Texts. New Haven, CT: Yale University Press, 1995. ———. “The Formation of the Pauline Corpus.” The Oxford Handbook of Pauline Studies. Edited by Matthew V. Novenson and R. Barry Matlock. Oxford: Oxford University Press, 2020. García, Elisa Ruiz. Introducción a la codicología. 2nd ed. Madrid: Fundación Germán Sánchez Ruipérez, 2002. Gardthausen, Viktor Emil. Griechische palaeographie. 2 vols. Leipzig: Teubner, 1879. Gathercole, Simon J. “The Titles of the Gospels in the Earliest New Testament Manuscripts.” Zeitschrift für die neutestamentliche Wissenschaft und die Kunde der älteren Kirche 104 (2013): 33–76. Gignac, Francis Thomas. A Grammar of the Greek Papyri of the Roman and Byzantine Periods, Vol. 1 Phonology. Testi e documenti per lo studio dell’antichità 55. Milan: Istituto Editoriale Cisalpino–La Goliardica, 1976. Giurisato, Giorgio. “Atti degli Apostoli: le divisioni dei codici Vaticanus e Amiantinus.” Liber Annuus 61 (2011): 211–27. Glenny, W. Edward. Hosea: A Commentary Based on Hosea in Codex Vaticanus. Septuagint Commentary Series. Leiden: Brill, 2013. Grafton, Anthony, and Megan Williams. Christianity and the Transformation of the Book: Origen, Eusebius, and the Library of Caesarea. Cambridge, MA: Harvard University Press, 2008. Gravely, Edward D. “The Relationship of the Vaticanus Umlauts to Family 1.” Pages 54–72 in Digging for the Truth: Collected Essays Regarding the Byzantine Text of the Greek New Testament: A Festchrift in Honor of Maurice A. Robinson. Edited by Mark Billington and Peter Streitenberger. Norden: FocusYourMission, 2014. 295 Gravely, Edward D. “The Text Critical Sigla in Codex Vaticanus.” Ph.D. diss., Southeastern Baptist Theological Seminary, 2009. Gregory, Caspar René. Canon and Text of the New Testament. The International Theological Library. Edinburgh: T&T Clark, 1907. ———. Prolegomena. Vol. 3 of Tischendorf of Novum Testamentum Graece: Editio Octava Critica Maior. Leipzig: J. C. Hinrichs, 1894. ———. Textkritik des Neuen Testamentes. 3 vols. Leipzig: J. C. Hinrichs, 1900. ———. “The Quires in Greek Manuscripts.” The American Journal of Philology 7 (1886): 27–32. ———. “Tischendorf.” Bibliotheca Sacra 33 (1876): 153–93. Grenz, Jesse R. “Textual Divisions in Codex Vaticanus. A Layered Approach to the Delimiters in B(03).” TC: A Journal of Biblical Textual Criticism 23 (2018): 1–22. Griesbach, Johann Jakob. Novum Testamentum Graece. Textum Ad Fidem Codicum Versionum et Patrum Recensuit et Lectionis Varietatem. 2nd ed. London: Peter Elmsley, 1796. Gullick, Michael. “Codicology.” Grove Art Online, 2003. https://www-oxfordartonline- com.ezp.lib.cam.ac.uk/groveart/view/10.1093/gao/9781884446054.001.0001/oao- 9781884446054-e-7000018410 Gumbert, J. P. “C Catalogue and Codicology – Some Reader’s Notes.” Pages 57–70 in A Catalogue and Its Users. A Symposium on the Uppsala C Collection of Medieval Manuscripts. Edited by Monica Hedlund. Acta Bibliothecae R. Universitatis Upsaliensis 34. Uppsala: Acta Universitatis Upsaliensis, 1995. ———. “Codicological Units: Towards a Terminology for the Stratigraphy of the Non- Homogeneous Codex.” Segno e testo 2 (2004): 17–42. ———. “Fifty Years of Codicology.” Archiv für Diplomatik Schriftgeschichte, Siegel- und Wappenkunde 50 (2004): 505–26. ———. “L’unité codicologique, ou : à quoi bon les cahiers?” Gazette du livre médiéval 14 (1989): 4–8. ———. Words for Codices: A Codicological Terminology in English. An Attempt by J. P. Gumbert (State of May 2010), www.cei.lmu.de/extern/VocCod/WOR10-1.pdf; www.cei.lmu.de/extern/VocCod/WOR10-2.pdf; www.cei.lmu.de/ extern/VocCod/WOR10- 3.pdf. Gurry, Peter J. A Critical Examination of the Coherence-Based Genealogical Method in New Testament Textual Criticism. New Testament Tools, Studies, and Documents 55. Leiden: Brill, 2017. Gurtner, Daniel M. Exodus: A Commentary on the Greek Text of Codex Vaticanus. Septuagint Commentary Series. Leiden: Brill, 2013. Hagner, Donald A. Matthew 1-13. Word Biblical Commentary 33A. Dallas, TX: Word, Incorporated, 1993. Haines-Eitzen, Kim. Guardians of Letters: Literacy, Power, and the Transmitters of Early Christian Literature. Oxford: Oxford University Press, 2000. BIBLIOGRAPHY 296 Hanhart, Robert, ed. Esdrae liber II. Septuaginta: Vetus Testamentum Graecum 8,2. Göttingen: Vandenhoeck & Ruprecht, 1993. ———. Text und Textgeschichte des 2. Esrabuches. Mitteilungen des Septuaginta-Unternehmens 25. Göttingen: Vandenhoeck & Ruprecht, 2003. Harris, J. Rendel. New Testament Autographs. Supplement to the American Journal of Philology 12. Baltimore, MD: Isaac Friedenwald, 1882. ———. Stichometry. London: C. J. Clay and Sons, 1893. Hatch, William H. P. The Principal Uncial Manuscripts of the New Testament. Chicago: University of Chicago Press, 1939. Head, Peter M. “Some Observations on Various Features of Scribe D in the New Testament of Codex Sinaiticus.” Pages 127–37 in Codex Sinaiticus: New Perspectives on the Ancient Biblical Manuscript. Edited by Scot McKendrick, David C. Parker, Amy Myshrall, and Cillian O’Hogan. London: British Library, 2015. ———. “The Early Text of Mark.” Pages 108–20 in The Early Text of the New Testament. Edited by Charles E. Hill and Michael J. Kruger. Oxford: Oxford University Press, 2012. ———. “The Marginalia of Codex Vaticanus: Putting the Distigmai (Formerly Known as ‘Umlauts’) in Their Place.” Presented at the Society of Biblical Literature Annual Meeting. New Orleans, 2009. Hichtel, Johann Samuel. Exercitatio critica qua antiquitas et praestantia codicis Romani prae Alexandrino contra Millium adseritur. Jena: Litteris Hornianis, 1734. Hill, Charles E. “A Neglected Text-Critical Siglum in Codex Vaticanus and Its Import for the Matthean Text.” TC: A Journal of Biblical Textual Criticism 24 (2019): 1–22. ———. “Rightly Dividing the Word: Uncovering an Early Template for Textual Division in John’s Gospel.” Pages 217–38 in Studies on the Text of the New Testament and Early Christianity. Edited by Daniel M. Gurtner, Juan Hernández Jr., and Paul Foster. New Testament Tools, Studies, and Documents 50. Leiden: Brill, 2015. Hixson, Elijah. Scribal Habits in Sixth-Century Greek Purple Codices. New Testament Tools, Studies, and Documents 61. Leiden: Brill, 2019. Hixson, Elijah Michael. “The Gospel of Matthew in a Sixth-Century Manuscript Family: Scribal Habits in the Greek Purple Codices 022, 023 and 042.” Ph.D. diss., The University of Edinburgh, 2017. Holmes, Michael W. “Codex Bezae as a Recension of the Gospels.” Pages 123–60 in Codex Bezae: Studies from the Lunel Colloquium, June 1994. Edited by David C. Parker and Christian-Bernard Amphoux. New Testament Tools, Studies, and Documents 22. Leiden: Brill, 1996. ———. “The Text of the Epistles Sixty Years After: An Assessment of Günther Zuntz’s Contribution to Text-Critical Methodology and History.” Pages 89–113 in Transmission and Reception. Edited by Jeff W. Childers and David C. Parker. Texts and Studies (Third Series) 4. Piscataway, NJ: Gorgias Press, 2006. 297 Höpfl, Hildebrand. Kardinal Wilhelm Sirlets Annotationen zum Neuen Testament: eine Verteidigung der Vulgata gegen Valla und Erasmus. Biblische Studien 13. Freiburg: Herdersche Verlagshandlung, 1908. Hort, Fenton John Anthony. Prolegomena to St. Paul’s Epistles to the Romans and the Ephesians. London: Macmillan and Co., 1895. Hoskier, H. C. Codex B and Its Allies: A Study and an Indictment. 2 vols. London: Bernard Quaritch, 1914. Houghton, H. A. G., and David C. Parker. “The Gospel of Luke in the Palimpsest.” Pages 33–58 in Codex Zacynthius: Catena, Palimpsest, Lectionary. Edited by H. A. G. Houghton and David C. Parker. Texts and Studies (Third Series) 21. Piscataway, NJ: Gorgias Press, 2020. Houston, George W. Inside Roman Libraries: Book Collections and Their Management in Antiquity. Chapel Hill, NC: The University of North Carolina Press, 2017. Hug, Johann Leonhard. An Introduction to the Writings of the New Testament. Translated by Daniel Guildford Wait. 2 vols. London: Rivington, 1827. ———. De Antiquitate Codicis Vaticani Commentatio. Freiburg: Herder, 1810. ———. Einleitung in die Schriften des Neuen Testaments. 2 vols. Tübingen, 1808. Hugo, Philippe. “1–2 Kingdoms (1–2 Samuel).” Pages 127–46 in T&T Clark Companion to the Septuagint. Edited by J. K. Aitken. London: T&T Clark, 2015. ———. “Basileion I and II/1-2 Kingdoms/1-2 Samuel.” Pages 173–94 in Introduction to the Septuagint. Edited by Siegfried Kreuzer, Translated by David A. Brenner and Peter Altman. Waco, TX: Baylor University Press, 2019. ———. “Text History of the Books of Samuel: An Assessment of the Recent Research.” Pages 1–19 in Archaeology of the Books of Samuel: The Entangling of the Textual and Literary History. Edited by Philippe Hugo and Adrian Schenker. Vetus Testamentum, Supplements 132. Leiden: Brill, 2010. Hurtado, Larry W. The Earliest Christian Artifacts: Manuscripts and Christian Origins. Grand Rapids, MI: Eerdmans, 2006. Jellicoe, Sidney. The Septuagint and Modern Study. Oxford: Clarendon Press, 1968. Johnson, William A. Bookrolls and Scribes in Oxyrhynchus. Toronto: University of Toronto Press, 2004. Johnston, Michael, and Michael Van Dussen, eds. The Medieval Manuscript Book: Cultural Approaches. Cambridge Studies in Medieval Literature 94. Cambridge: Cambridge University Press, 2015. Jones, Brice C. “Scribes Avoiding Imperfections in Their Writing Materials.” Archiv für Papyrusforschung 61 (2015): 371–83. Jongkind, Dirk. An Introduction to the Greek New Testament, Produced at Tyndale House, Cambridge. Wheaton, IL: Crossway, 2019. BIBLIOGRAPHY 298 ———. “One Codex, Three Scribes, and Many Books: Struggles with Space in Codex Sinaiticus.” Pages 121–36 in New Testament Manuscripts: Their Texts and Their World. Edited by Thomas J. Kraus and Tobias Nicklas. Texts and Editions for New Testament Study 2. Leiden: Brill, 2006. ———. “Redactional Elements in the Text of Codex B.” Pages 231–45 in The Future of New Testament Textual Scholarship: From H. C. Hoskier to the Editio Critica Maior and Beyond. Edited by Garrick V. Allen. Wissenschaftliche Untersuchungen Zum Neuen Testament 417. Tübingen: Mohr Siebeck, 2019. ———. Scribal Habits of Codex Sinaiticus. Texts and Studies 5. Piscataway, NJ: Gorgias Press, 2007. ———. “The Text of the Pauline Corpus.” Pages 216–31 in The Blackwell Companion to Paul. Edited by Stephen Westerholm. Chichester: Wiley-Blackwell, 2011. Kauhanen, Tuukka. The Proto-Lucianic Problem in 1 Samuel. De Septuaginta Investigationes 3. Göttingen: Vandenhoeck & Ruprecht, 2012. Keene, Nicholas. “John Ernest Grabe, Biblical Learning and Religious Controversy in Early Eighteenth-Century England.” The Journal of Ecclesiastical History 58 (2007): 656–74. Kenyon, Frederic G. “Abbreviations and Symbols in Greek Papyri.” Pages 127–30 in Abbreviations in Greek Inscriptions: Papyri, Manuscripts and Early Printed Books. Edited by A. N. Oikonomides. Chicago, IL: Ares Publishers, 1974. ———. Books and Readers in Ancient Greece and Rome. Oxford: Clarendon Press, 1932. ———. “Hesychius and the Text of the New Testament.” Pages 245–50 in Mémorial Lagrange. Cinquantenaire de l’École biblique et archéologique française de Jérusalem (15 Novembre 1890-15 Novembre 1940). Paris: J. Gabalda, 1940. ———. Our Bible and the Ancient Manuscripts. 3rd ed. London: Eyre and Spottiswoode, 1898. ———. The Text of the Greek Bible: A Student’s Handbook. Repr. London: Duckworth, 1949. Ker, N. R. “‘From “above Top Line” to “below Top Line”: A Change in Scribal Practice.’” Celtica 5 (1960): 13–16. Kim, Jong-Hoon. “Vom hellenistischen Kleinrollensystem zum Kodex: Beobachtungen zur Textgestalt der griechischen Samuel- Und Königebücher.” Pages 231–42 in XIV Congress of the International Organization for Septuagint and Cognate Studies, Helsinki, 2010. Edited by Melvin K. H. Peters. Septuagint and Cognate Studies 59. Atlanta, GA: Society of Biblical Literature, 2013. Knust, Jennifer Wright, and Tommy Wasserman. To Cast the First Stone: The Transmission of a Gospel Story. Princeton: Princeton University Press, 2018. Krans, Jan. “Erasmus and Codex Vaticanus: An Overview and an Evaluation.” Annali di storia dell’esegesi 37 (2020): 447–70. ———. “Paragraphos, Not Obelos, in Codex Vaticanus.” New Testament Studies 65 (2019): 252–57. Kreuzer, Siegfried. “»… et a plerisque nunc loukianeios dicitur«: Jerome’s Statements on the Greek Biblical Texts and Modern Septuagint Scholarship.” Zeitschrift für die alttestamentliche Wissenschaft 130 (2018): 69–85. ———. “5.4 Septuagint (Samuel).” Textual History of the Bible, http://dx.doi.org/10.1163/2452- 4107_thb_COM_0005040000. 299 ———. “B or Not B? The Place of Codex Vaticanus in Textual History and in Septuagint Research.” Pages 277–98 in The Bible in Greek: Translation, Transmission, and Theology of the Septuagint. Septuagint and Cognate Studies 63. Atlanta, GA: Society of Biblical Literature, 2015. Kubo, Sakae. P72 and the Codex Vaticanus. Studies and Documents 27. Salt Lake City, UT: University of Utah Press, 1965. Lachmann, Karl. “Rechenschaft über seine Ausgabe des Neuen Testaments.” Theologische Studien und Kritiken (1830): 817—845. Lagrange, M.-J. La critique rationnelle. Introduction à l’étude du Nouveau Testament. 2nd ed. Vol. 2. Paris: Lecoffre, 1935. Lake, Kirsopp. “The Sinaitic and Vatican Manuscripts and the Copies Sent by Eusebius to Constantine.” The Harvard Theological Review 11 (1918): 32–35. ———. The Text of The New Testament. 4th, rev. ed. New York, NY: Edwin S. Gorham, 1908. Lake, Kirsopp, and Helen Lake. Codex Sinaiticus Petropolitanus: The New Testament, the Epistle of Barnabas and The Shepherd of Hermas. Oxford: Clarendon Press, 1911. Law, T. Michael, and Tuukka Kauhanen. “Methodological Remarks on the Textual History of Reigns: A Response to Siegfried Kreuzer.” Bulletin of the International Organization for Septuagint and Cognate Studies 43 (2010): 73–87. Le Long, Jacques. Bibliotheca sacra. 2 vols. Paris: Pralard, 1709. Leroy, Julien. Les types de réglure des manuscrits grecs. Paris: Éditions du centre national de la recherche scientifique, 1976. Liyanage, Rohana, and Jackson O. Lay, Jr. “An Introduction to MALDI‐TOF MS.” Pages 39–60 in Identification of Microorganisms by Mass Spectrometry. Edited by Charles L. Wilkins and Jackson O. Lay, Jr. Chemical Analysis 169. Hoboken, NJ: John Wiley & Sons, Inc., 2006. Luijendijk, AnneMarie. Greetings in the Lord: Early Christians and the Oxyrhynchus Papyri. Harvard Theological Studies 60. Cambridge, MA: Harvard University Press, 2008. Macmillan, Hugh. Roman Mosaics, or Studies in Rome and Its Neighborhood. Macmillan and Co.: London, 1888. Mai, Angelo. Codex Vaticanus. Η Καινη Διαθηκη. Novum Testamentum Graece ex antiquissimo codice vaticano. Repr. New York, NY: D. Appleton and Co., 1859. Malik, Peter. P.Beatty III (P47): The Codex, Its Scribe, and Its Text. New Testament Tools, Studies, and Documents 52. Leiden: Brill, 2017. ———. “The Earliest Corrections in Codex Sinaiticus: A Test Case from the Gospel of Mark.” The Bulletin of the American Society of Papyrologists 50 (2013): 207–54. ———. “The Nomina Sacra in the Marcan Portion of Codex Vaticanus: A Note on the Scribal Habits.” Biblische Notizen 175 (2017): 95–105. Man, Loretta H.Y. “The Textual Significance of Corrected Readings in the Evaluation of the External Evidence: Romans 5,1 as a Test Case.” Zeitschrift für die neutestamentliche Wissenschaft 107 (2016): 70–93. BIBLIOGRAPHY 300 Mandelbrote, Scott. “When Manuscripts Meet: Editing the Bible in Greek during and after the Council of Trent.” Pages 251–67 in For the Sake of Learning: Essays in Honor of Anthony Grafton. Edited by Ann Blair and Anja-Silvia Goeing. Scientific and Learned Cultures and Their Institutions 18. Leiden: Brill, 2016. Maniaci, Marilena. “The Medieval Codex as a Complex Container: The Greek and Latin Traditions.” Pages 27–46 in One-Volume Libraries: Composite and Multiple-Text Manuscripts. Edited by Michael Friedrich and Cosima Schwarke. Studies in Manuscript Cultures 9. Berlin: De Gruyter, 2016. Maniaci, Marilena, Alessandro Bausi, and Eugenia Sokolinski, eds. “Codicology.” Comparative Oriental Manuscript Studies: An Introduction. Hamburg: Tredition, 2015. Martini, Carlo M. Il problema della recensionalità del codice B alla luce del papiro Bodmer XIV. Roma: Pontificio Istituto Biblico, 1966. ———. “Is There a Late Alexandrian Text of the Gospels?” New Testament Studies 24 (1978): 285– 96. Masai, François. “Paléographie et codicologie.” Scriptorium 4 (1950): 279–93. Massaux, Édouard. Influence of the Gospel of Saint Matthew on Christian Literature Before Saint Irenaeus: Book 1: The First Ecclesiastical Writers. Edited by Arthur J. Bellinzoni. Translated by Norman J. Belval and Suzanne Hecht. New Gospel Studies 5/1. Macon, GA: Mercer University Press, 1990. ———. Influence of the Gospel of Saint Matthew on Christian Literature Before Saint Irenaeus: Book 2: The Later Christian Writings. Edited by Arthur J. Bellinzoni. Translated by Norman J. Belval and Suzanne Hecht. New Gospel Studies 5/2. Macon, GA: Mercer University Press, 1992. ———. Influence of the Gospel of Saint Matthew on Christian Literature Before Saint Irenaeus: Book 3: The Apologists and the Didache. Edited by Arthur J. Bellinzoni. Translated by Norman J. Belval and Suzanne Hecht. New Gospel Studies 5/3. Macon, GA: Mercer University Press, 1993. McDonald, Grantley. Biblical Criticism in Early Modern Europe Erasmus, the Johannine Comma, and Trinitarian Debate. Cambridge: Cambridge University Press, 2016. McGurk, Patrick. “Citation Marks in Early Latin Manuscripts. (With a List of Citation Marks in Manuscripts Earlier than A. D. 800 in English and Irish Libraries).” Scriptorium 15 (1961): 3–13. McLay, R. Timothy. “Recension and Revision: Speaking the Same Language with Special Attention to Lucian and Kaige.” Pages 293–303 in XIII Congress of the International Organization for Septuagint and Cognate Studies, Ljubljana, 2007. Edited by Melvin K. H. Peters. Septuagint and Cognate Studies 55. Atlanta, GA: Society of Biblical Literature, 2008. McNamee, Kathleen. Sigla and Select Marginalia in Greek Literary Papyri. Papyrologica Bruxellensia 26. Bruxelles: Fondation Égyptologique Reine Élisabeth, 1992. Metzger, Bruce M. A Textual Commentary on the Greek New Testament. 2nd ed. London: United Bible Societies, 1994. ———. “Lucian and the Lucianic Recension of the Greek Bible.” New Testament Studies 8 (1962): 189–203. 301 ———. Manuscripts of the Greek Bible: An Introduction to Greek Palaeography. Oxford: Oxford University Press, 1981. Metzger, Bruce M., and Bart D. Ehrman. The Text of the New Testament: Its Transmission, Corruption, and Restoration. 4th ed. Oxford: Oxford University Press, 2005. Michaelis, Johann David. Introduction to the New Testament. Translated by Herbert Marsh. 2nd ed. 2 vols. London: F. and C. Rivington, 1802. Mill, John. Hē Kainē Diathēkē : Novum Testamentum : cum lectionibus variantibus mss. exemplarium, versionum, editionum, SS. patrum & scriptorum ecclesiasticorum, & in easdem notis : accedunt loca scripturae parallela, aliaque exēgētika, & appendix ad variantes lectiones : praemittitur dissertatio, in qua de libris N.T. & canonis constitutione agitur, historia s. textus n. foederis ad nostra usque tempora deducitur, et quid in hac editione praestitum sit. Oxford: E Theatro Sheldoniano, 1707. Milne, H. J., and T. C. Skeat. Scribes and Correctors of the Codex Sinaiticus. London: British Museum for the Trustees, 1938. Mink, Gerd. “Contamination, Coherence, and Coincidence in Textual Transmission: The Coherence- Based Genealogical Method (CBGM) as a Complement and Corrective to Existing Approaches.” The Textual History of the Greek New Testament: Changing Views in Contemporary Research. Edited by Klaus Wachtel and Michael W. Holmes. Text-Critical Studies 8. Atlanta, GA: Society of Biblical Literature, 2011. Montana, Fausto. “Hellenistic Scholarship.” Pages 60–183 in Brill’s Companion to Ancient Greek Scholarship. Edited by Franco Montanari, Stefanos Matthaios, and Antonios Rengakos. Vol. 1 of Brill’s Companions to Classical Studies. Leiden: Brill, 2015. Montanari, Franco. “Correcting a Copy, Editing a Text. Alexandrian Ekdosis and Papyri.” Pages 1–16 in From Scholars to Scholia: Chapters in the History of Ancient Greek Scholarship. Edited by Franco Montanari and Lara Pagani. Trends in Classics - Supplementary Volumes 9. Berlin: De Gruyter, 2011. Montfaucon, Bernard de. Bibliotheca bibliothecarum manuscriptorum nova: Tomus primus. Paris: Briasson, 1739. Montgomery, James A. “The Ethiopic Text of Acts of the Apostles.” The Harvard Theological Review 27 (1934): 169–205. Moorhead, Gavin. “Parchment Assessment of the Codex Sinaiticus,” 2009. http://codexsinaiticus.org/en/project/conservation_parchment.aspx#note16. Mugridge, Alan. Copying Early Christian Texts: A Study of Scribal Practice. Wissenschaftliche Untersuchungen Zum Neuen Testament 362. Tübingen: Mohr Siebeck, 2016. Muralt, Eduard von. Novum Testamentum Graece ad fidem codicis principis Vaticani. Hamburg: J. A. Meissner, 1848. Myers, Peter Daniel. “A Grammar of Transcriptions in 2 Esdras: Text, Philology, Phonology.” Ph.D. diss., University of Cambridge, 2019. BIBLIOGRAPHY 302 Myshrall, Amy. “The Presence of a Fourth Scribe?” Pages 139–48 in Codex Sinaiticus: New Perspectives on the Ancient Biblical Manuscript. Edited by Scot McKendrick, David C. Parker, Amy Myshrall, and Cillian O’Hogan. London: British Library, 2015. Nässelqvist, Dan. Public Reading in Early Christianity: Lectors, Manuscripts, and Sound in the Oral Delivery of John 1-4. Novum Testamentum, Supplements 163. Leiden: Brill, 2016. Nestle, Eberhard. Introduction to the Textual Criticism of the Greek New Testament. Edited by Allan Menzies. Translated by William Edie. London: Williams and Norgate, 1901. Nevius, Richard C. “The Use of the Definite Article with ‘Jesus’ in the Fourth Gospel.” New Testament Studies 12 (1965): 81–85. Niccum, Curt. “The Voice of the Manuscripts on the Silence of Women: The External Evidence for 1 Cor 14:34-35.” New Testament Studies 43 (1997): 242–55. Nongbri, Brent. God’s Library: The Archaeology of the Earliest Christian Manuscripts. New Haven, CT: Yale University Press, 2018. ———. “Palaeographic Analysis of Codices from the Early Christian Period: A Point of Method.” Journal for the Study of the New Testament 42 (2019): 84–97. ———. “Pauline Letter Manuscripts.” Pages 84–102 in All Things to All Cultures: Paul Among Jews, Greeks, and Romans. Edited by Mark Harding and Alanna Nobbs. Grand Rapids, MI: Eerdmans, 2013. ———. “Reconsidering the Place of Papyrus Bodmer XIV–XV (𝔓75) in the Textual Criticism of the New Testament.” Journal of Biblical Literature 135 (2016): 405–37. ———. “The Date of Codex Sinaiticus.” The Journal of Theological Studies (forthcoming). Nyström, Eva. Containing Multitudes: Codex Upsaliensis Graecus 8 in Perspective. Studia Byzantina Upsaliensia 11. Uppsala: Acta Universitatis Upsaliensis, 2009. O’Callaghan, José. Nomina sacra in papyris graecis saeculi III neotestamentariis. Analecta Biblica 46. Rome: Biblical Institute Press, 1970. Ohly, Kurt. Stichometrische Untersuchungen. Leipzig: Harrassowitz, 1928. Olley, John W. Ezekiel: A Commentary Based on Iezekiēl in Codex Vaticanus. Septuagint Commentary Series. Leiden: Brill, 2009. Orsini, Pasquale. Studies on Greek and Coptic Majuscule Scripts and Books. Studies in Manuscript Cultures 15. Berlin: De Gruyter, 2019. Orsini, Pasquale, and Willy Clarysse. “Early New Testament Manuscripts and Their Dates: A Critique of Theological Palaeography.” Ephemerides Theologicae Lovanienses 88 (2012): 443–74. Osann, Friedrich, ed. Anecdotum Romanum de notis veterum criticis: Inprimis Aristarchi Homericis et Iliade Heliconia. Berlin: De Gruyter, 1851. Paap, A. H. R. E. Nomina Sacra in the Greek Papyri of the First Five Centuries A.D.: The Sources and Some Deductions. Leiden: Brill, 1959. Pardee, Cambry G. Scribal Harmonization in the Synoptic Gospels. New Testament Tools, Studies, and Documents 60. Leiden: Brill, 2019. 303 Parker, David C. “A Dictation Theory of Codex Bezae.” Pages 5–18 in Manuscripts, Texts, Theology: Collected Papers 1977-2007. Arbeiten zur neutestamentlichen Textforschung 40. Berlin: De Gruyter, 2009. ———. An Introduction to the New Testament Manuscripts and Their Texts. Cambridge: Cambridge University Press, 2010. ———. Codex Bezae: An Early Christian Manuscript and Its Text. Cambridge: Cambridge University Press, 1992. ———. “Scribal Tendencies and the Mechanics of Book Production.” Pages 173–84 in Textual Variation: Theological and Social Tendencies?: Papers from the Fifth Birmingham Colloquium on the Textual Criticism of the New Testament. Edited by David C. Parker and H. A. G. Houghton. Texts and Studies 6. Piscataway, NJ: Gorgias Press, 2008. ———. Textual Scholarship and the Making of the New Testament. Oxford: Oxford University Press, 2014. Parry, D. W., D. V. Arnold, D. G. Long, and Woodward. “New Technological Advances: DNA, Graphics and Imaging, Underground Radar.” Pages 496–515 in The Dead Sea Scrolls After Fifty Years: A Comprehensive Assessment. Edited by Peter W. Flint and James C. VanderKam. Leiden: Brill, 1998. Parsons, P. J. “Stichometry.” Oxford Classical Dictionary. http://classics.oxfordre.com/view/10.1093/acrefore/9780199381135.001.0001/acrefore- 9780199381135-e-6081 Pattie, T. S. “The Creation of the Great Codices.” Pages 61–72 in The Bible as Book: The Manuscript Tradition. Edited by John L. Sharpe and Kimberly Van Kampen. London: British Library, 1998. Paulson, Gregory S. Scribal Habits and Singular Readings in Codex Sinaiticus, Vaticanus, Ephraemi, Bezae, and Washingtonianus in the Gospel of Matthew. GlossaHouse Dissertation Series 5. Wilmore, KY: GlossaHouse, 2018. Payne, Philip B., and Paul Canart. “Distigmai Matching the Original Ink of Codex Vaticanus: Do They Mark the Location of Textual Variants?” Pages 199–226 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. ———. “The Originality of Text-Critical Symbols in Codex Vaticanus.” Novum Testamentum 42 (2000): 105–13. Penn, Granville. Annotations to the Book of the New Covenant: With an Expository Preface. London: J. Duncan, 1837. Perttilä, Elina. Sahidic 1 Samuel – A Daughter Version of the Septuagint 1 Reigns. De Septuaginta Investigationes 8. Göttingen: Vandenhoeck & Ruprecht, 2017. Pfaff, Christoph Matthaeus. dissertatio critica de genuinis librorum Novi Testamenti lectionibus. Amsterdam: Leeuwen, 1709. Pierri, Rosario. “Accentazione delle ossitone in rapporto ai segni di divisione nei vangeli in B (03).” Liber Annuus 64 (2014): 137–63. BIBLIOGRAPHY 304 Pietersma, Albert. “The Present State of the Critical Text of the Greek Psalter.” Pages 12–32 in Der Septuaginta-Psalter und seine Tochterübersetzungen: Symposium in Göttingen 1997. Edited by Anneli Aejmelaeus and Udo Quast. Abhandlungen der Akademie der Wissenschaften in Göttingen, Philosophisch-historische Klasse 230. Göttingen: Vandenhoeck & Ruprecht, 2000. du Pin, Louis Ellies. Dissertation préliminaire ou prolégomènes sur la Bible. 2 vols. Paris: Pralard, 1699. Pinto, Leonardo Pessoa Da Silva. “The Beginning of the ΚΑΙΓΕ Section of 2 Samuel.” Biblica 100 (2019): 14–33. Pisano, Stephen. “L’histoire du Codex Vaticanus B pendant quatre siècles: Les notes inédites du cardinal Mercati.” Pages 105–18 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. ———. “The Vaticanus Graecus 1209: A Witness to the Text of the New Testament.” Pages 77–97 in Le manuscrit B de la Bible (Vaticanus graecus 1209). Edited by Patrick Andrist. Histoire du texte biblique 7. Lausanne: Éditions du Zèbre, 2009. Popović, Mladen, Maruf A. Dhali, and Lambert Schomaker. “Artificial Intelligence Based Writer Identification Generates New Evidence for the Unknown Scribes of the Dead Sea Scrolls Exemplified by the Great Isaiah Scroll (1QIsaa).” PLOS ONE 16 (2021): 1–28. Porter, Calvin L. “Papyrus Bodmer XV (P75) and the Text of Codex Vaticanus.” Journal of Biblical Literature 81 (1962): 363. Rahlfs, Alfred. “Alter und Heimat der vatikanischen Bibelhandschrift.” Pages 72–79 in Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-Physikalische Klasse. Göttingen: Lüder Horstmann, 1899. ———. Der Text des Septuaginta-Psalters. Nebst einem Anhang, Griechische Psalterfragmente aus Oberägypten nach Abschriften von W.E. Crum. Repr. Septuaginta-Studien 2. Göttingen: Vandenhoeck & Ruprecht, 1979. ———. “Prolegomena.” Septuaginta: Societatis Scientiarum Gottingensis auctoritate. X. Psalmi Cum Odis. Edited by Alfred Rahlfs. Göttingen: Vandenhoeck & Ruprecht, 1931. Read-Heimerdinger, Jenny. “The «Long» and the «Short» Texts of Acts: A Closer Look at the Quantity and Types of Variation.” Revista Catalana de Teología 22 (1997): 245–61. Rius-Camps, Josep, and Jenny Read-Heimerdinger. The Message of Acts in Codex Bezae: A Comparison with the Alexandrian Tradition. 3 vols. Library of New Testament Studies 257, 302, 365. London: T&T Clark, 2004. Robinson, Joseph Armitage. Euthaliana: Studies of Euthalius, Codex H of the Pauline Epistles, and the Armenian Version. Texts and Studies 3. Cambridge: Cambridge University Press, 1895. Ropes, James Hardy. “The Beginnings of Christianity: Part I, The Acts of the Apostles: Vol. III, The Text of Acts.” Edited by F. J. Foakes-Jackson and Kirsopp Lake. 5 vols. London: Macmillan and Co., 1926. de Rossi, Giovanni Battista. “Della croce monogrammatica segnata nel codice greco Vaticano della Bibbia.” Pages 16–21 in Dell’antichissimo Codice Vaticano della Bibbia Greca. Roma: Nella Tipografia della Rev. Camera Apostolica, 1860. 305 Royse, James R. Scribal Habits in Early Greek New Testament Papyri. New Testament Tools, Studies, and Documents 36. Leiden: Brill, 2008. ———. “The Early Text of Paul (and Hebrews).” Pages 175–203 in The Early Text of the New Testament. Edited by Charles E. Hill and Michael J. Kruger. Oxford: Oxford University Press, 2012. ———. “The Ethiopic Support for Codex Vaticanus in Acts.” Zeitschrift für die neutestamentliche Wissenschaft 71 (1980): 258–62. Rulotta, Abbé. “Collatio Codicis Vaticani.” Pages 119–54 in Bentleii Critica Sacra. Edited by Arthur Ayres Ellis. Cambridge: Deighton, Bell, and Co., 1862. Saar, Ortal-Paz. Jewish Love Magic: From Late Antiquity to the Middle Ages. Magical and Religious Literature of Late Antiquity 6. Leiden: Brill, 2017. Sanday, William, and Arthur Cayley Headlam. A Critical and Exegetical Commentary on the Epistle to the Romans. 5th ed. The International Critical Commentary. Edinburgh: T&T Clark, 1908. Sanders, Henry A, ed. A Third-Century Papyrus Codex of the Epistles of Paul. Ann Arbor, MI: University of Michigan Press, 1935. Scase, Wendy. “‘Looke This Calender and Then Proced’: Tables of Contents in Medieval English Manuscripts.” Pages 287–306 in The Dynamics of the Medieval Manuscript: Text Collections from a European Perspective. Edited by Karen Pratt, Bart Besamusca, Matthias Meyer, and Ad Putter. Göttingen: V&R unipress GmbH, 2017. Schironi, Francesca. “Book-Ends and Book-Layout in Papyri with Hexametric Poetry.” Pages 695– 704 in Proceedings of the 25th International Congress of Papyrology: Ann Arbor, July 29-August 4, 2007. Edited by Traianos Gagos and Adam Hyatt. American Studies in Papyrology. Ann Arbor, MI: Scholarly Publishing Office, The University of Michigan Library, 2010. ———. To Mega Biblion: Book-Ends, End-Titles, and Coronides in Papyri with Hexametric Poetry. The American Studies in Papyrology 48. Durham, NC: American Society of Papyrologists, 2010. Schmid, U. B, W. J Elliott, and D. C Parker, eds. The New Testament in Greek IV, 2: The Gospel According to St. John - The Majuscules. New Testament Tools, Studies, and Documents 37. Leiden: Brill, 2007. Schmid, Ulrich. “Die Diplé: Einführung.” Pages 77–82 in Von der Septuaginta zum Neuen Testament: Textgeschichtliche Erörterungen. Edited by Martin Karrer, Siegfried Kreuzer, and Marcus Sigismund. Arbeiten zur neutestamentlichen Textforschung 43. Berlin: De Gruyter, 2010. ———. “Diplés Im Codex Vaticanus.” Pages 99–114 in Von der Septuaginta zum Neuen Testament: Textgeschichtliche Erörterungen. Edited by Martin Karrer, Siegfried Kreuzer, and Marcus Sigismund. Arbeiten zur neutestamentlichen Textforschung 43. Berlin: De Gruyter, 2010. ———. “Diplés Und Quellenangaben Im Codex Sinaiticus.” Pages 83–98 in Von der Septuaginta zum Neuen Testament: Textgeschichtliche Erörterungen. Edited by Martin Karrer, Siegfried Kreuzer, and Marcus Sigismund. Arbeiten zur neutestamentlichen Textforschung 43. Berlin: De Gruyter, 2010. Schmid, Ulrich B., and Bruce Morrill. “Editorial Activity and Textual Affiliation: The Case of the Corpus Paulinum.” Pages 361–84 in The New Testament in Antiquity and Byzantium. Edited by BIBLIOGRAPHY 306 H.A.G. Houghton, David C. Parker, and Holger Strutwolf. Arbeiten zur neutestamentlichen Textforschung 52. Berlin: De Gruyter, 2019. Scrivener, Frederick Henry Ambrose. A Plain Introduction to the Criticism of the New Testament. Edited by Edward Miller. 4th ed. 2 vols. London: George Bell & Sons, 1894. Shenkel, James Donald. Chronology and Recensional Development in the Greek Text of Kings. Harvard Semitic Monographs 1. Cambridge, MA: Harvard University Press, 1968. Sider, Robert D., ed. The New Testament Scholarship of Erasmus: An Introduction with Erasmus’ Prefaces and Ancillary Writings. Collected Works of Erasmus 41. Toronto: University of Toronto Press, 2019. Simon, Richard. Bibliothèque critique, ou recueil de diverses pièces critiques, dont la plûpart ne sont point imprimées, ou ne se trouvent que très-difficilement. 4 vols. Amsterdam: Jean Louis de Lormes, 1708. Sixtine edition. Ἡ παλαιὰ διαθήκη κατὰ τοὺς ἑβδοµήκοντα δι’ αὐθεντίας Ξύστου ε´, ἄκρου ἀρχιερέως ἐκδοθεῖσα. Vetus testamentum iuxta septuaginta ex auctoritate Sixti V. Pont. Max. editum. Romae: Ex typographia Francisci Zannetti, 1587. Skeat, T. C. “Early Christian Book-Production: Papyri and Manuscripts.” Pages 33–59 in The Collected Biblical Writings of T.C. Skeat. Edited by J. K. Elliott. Novum Testamentum, Supplements 113. Leiden: Brill, 2004. ———. “Four Years’ Work on the Codex Sinaiticus: Significant Discoveries in Reconditioned MS.” Pages 107–18 in The Collected Biblical Writings of T.C. Skeat. Edited by J. K. Elliott. Novum Testamentum, Supplements 113. Leiden: Brill, 2004. ———. “The Codex Sinaiticus, the Codex Vaticanus and Constantine.” Pages 193–237 in The Collected Biblical Writings of T.C. Skeat. Edited by J. K. Elliott. Novum Testamentum, Supplements 113. Leiden: Brill, 2004. ———. “The Codex Vaticanus in the Fifteenth Century.” Pages 122–34 in The Collected Biblical Writings of T.C. Skeat. Edited by J. K. Elliott. Novum Testamentum, Supplements 113. Leiden: Brill, 2004. ———. “The Use of Dictation in Ancient Book-Production.” Pages 3–32 in The Collected Biblical Writings of T.C. Skeat. Edited by J. K. Elliott. Novum Testamentum, Supplements 113. Leiden: Brill, 2004. Smith, W. Andrew. A Study of the Gospels in Codex Alexandrinus: Codicology, Palaeography, and Scribal Hands. New Testament Tools, Studies, and Documents 48. Leiden: Brill, 2014. Soden, Hermann von. Die Schriften des Neuen Testaments in ihrer ältesten erreichbaren Textgestalt hergestellt auf Grund ihrer Textgeschichte. 4 vols. Göttingen: Vandenhoeck & Ruprecht, 1902. Sokolinski, Eugenia. “Conference Reports: The Emergence of Multiple-Text Manuscripts Hamburg, 9 to 12 November 2016.” COMSt Bulletin 2 (2016): 87–88. Stephen, Gwendolen M. “The Coronis.” Scriptorium 13 (1959): 3–14. 307 Stevens, Chris S. History of the Pauline Corpus in Texts, Transmissions and Trajectories: A Textual Analysis of Manuscripts from the Second to the Fifth Century. Texts and Editions for New Testament Study 14. Leiden: Brill, 2020. Stevens, Luke J. “The Two-Volume Archetype of the Pauline Corpus.” Journal for the Study of Paul and His Letters 8 (2018): 102–26. Stokowski, Alexander. “Diplé-Auszeichnungen im Codex Vaticanus Graecus 1209 (B): Liste nebst einigen Beobachtungen.” Pages 93–114 in Textual History and the Reception of Scripture in Early Christianity: Textgeschichte und Schriftrezeption im frühen Christentum. Edited by Johannes de Vries and Martin Karrer. Septuagint and Cognate Studies 60. Atlanta, GA: Society of Biblical Literature, 2013. Stolk, Joanne Vera. “Itacism from Zenon to Dioscorus: scribal corrections of <ι> and <ει> in Greek documentary papyri.” Proceedings of the 28th Congress of Papyrology: Barcelona 1-6 August 2016. Edited by Alberto Nodar, Sofía Torallas Tovar, María Jesús Albarrán Martínez, Raquel Martin Hernández, and Irene Pajón Leyra. Barcelona: Universitat Pompeu Fabra, 2019. ———. “Orthographic Variation and Register in the Corpus of Greek Documentary Papyri (300 BCE–800 CE).” Pages 299–326 in Varieties of Post-Classical and Byzantine Greek. Edited by Klaas Bentein and Mark Janse. Trends in Linguistics. Studies and Monographs 331. Berlin: De Gruyter, 2021. Strutwolf, Holger, Georg Gäbel, Annette Hüffmeier, Gerd Mink, and Klaus Wachtel, eds. Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 1.1: Text. Stuttgart: Deutsche Bibelgesellschaft, 2017. ———, eds. Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 1.2: Text. Stuttgart: Deutsche Bibelgesellschaft, 2017. ———, eds. Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 2: Supplementary Material. Stuttgart: Deutsche Bibelgesellschaft, 2017. ———, eds. Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 3: Studies. Stuttgart: Deutsche Bibelgesellschaft, 2017. Stunt, Timothy C. F. The Life and Times of Samuel Prideaux Tregelles: A Forgotten Scholar. Christianities in the Trans-Atlantic World. Cham: Palgrave Macmillan, 2020. Swete, Henry Barclay. An Introduction to the Old Testament in Greek. Cambridge: Cambridge University Press, 1900. ———. The Old Testament in Greek According to the Septuagint. 4th ed. 3 vols. Cambridge: Cambridge University Press, 1909. Taylor, Bernard A. The Lucianic Manuscripts of 1 Reigns. Volume 1: Majority Text. Harvard Semitic Monographs 50. Atlanta, GA: Scholars Press, 1992. Taylor, John. The Emphatic New Testament, According to the Authorized Version. London: S. Bagster, 1854. Thackeray, H. St J. A Grammar of the Old Testament in Greek According to the Septuagint. Vol. I: Introduction, Orthography and Accidence. Cambridge: Cambridge University Press, 1909. BIBLIOGRAPHY 308 ———. “The Bisection of Books in Primitive Septuagint MSS.” The Journal of Theological Studies 9 (1907): 88–98. ———. “The Greek Translators of Ezekiel.” The Journal of Theological Studies 4 (1903): 398–411. ———. “The Greek Translators of Jeremiah.” The Journal of Theological Studies 4 (1903): 245–66. ———. “The Greek Translators of the Four Books of Kings.” The Journal of Theological Studies 8 (1907): 262–78. Thompson, Edward Maunde. An Introduction to Greek and Latin Palaeography. Oxford: Clarendon Press, 1912. Tischendorf, Constantin von. Appendix codicum celeberrimorum Sinaitici Vaticani Alexandrini: cum imitatione ipsorum antiqua manu scriptorum. Leipzig: Giesecke et Devrient, 1867. ———. Appendix Novi Testamenti Vaticani. Leipzig: Giesecke et Devrient, 1869. ———. “Biblisch Kritisches Sendschreiben an David Schulz.” Allgemeine Kirchenzeitung (1843): coll. 961-966. ———. “Nachricht vom vaticanischen Bibelcodex.” Pages 129–52 in Theologische Studien und Kritiken. Hamburg: Friedrich Perthes, 1847. ———. Novum Testamentum Graece. 7th ed. 2 vols. Leipzig: Sumptibus Adolphi Winter, 1859. ———. Novum Testamentum Graece: Editio Octava Critica Maior. 3 vols. Leipzig: J. C. Hinrichs, 1894. ———. Novum Testamentum Sinaiticum: sive Novum Testamentum cum epistula Barnabae et fragmentis Pastoris ex codice Sinaitico. Leipzig: Brockhaus, 1863. ———. Novum Testamentum Vaticanum: post Angeli Maii aliorumque imperfectos labores ex ipso codice. Leipzig: Giesecke et Devrient, 1867. Tisserant, S. E. Eugène. “Notes sur la préparation de l’édition en fac-simile typographique du ‘Codex Vaticanus’ (B).” Angelicum 20 (1943): 237–48. Touwaide, Alain. “Codicology and Paleography.” Pages 266–329 in Handbook of Medieval Studies: Terms - Methods - Trends. Edited by Albrecht Classen. Berlin: De Gruyter, 2010. Tov, Emanuel. Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert. Studies on the Texts of the Desert of Judah 54. Atlanta, GA: Society of Biblical Literature, 2009. ———. “The Methodology of Textual Criticism in Jewish Greek Scriptures, with Special Attention to the Problems in Samuel-Kings-The State of the Question.” Pages 489–99 in The Greek and Hebrew Bible: Collected Essays on the Septuagint. Vetus Testamentum, Supplements 72. Leiden: Brill, 1999. Traube, Ludwig. Nomina sacra: Versuch einer Geschichte der christlichen Kürzung. Munich: Beck, 1907. ———. Zur Paläographie und Handschriftenkunde. Munich: Beck, 1909. Tregelles, Samuel Prideaux. An Account of the Printed Text of the Greek New Testament: With Remarks on Its Revision Upon Critical Principles. London: Samuel Bagster and Sons, 1854. 309 ———. An Introduction to the Textual Criticism, Etc. of the New Testament. 10th ed. Vol. 4 of An Introduction to the Critical Study and Knowledge of the Holy Scriptures. Edited by Thomas Hartwell Horne and Samuel Davidson. London: Longman, Brown, Green, Longmans, & Roberts, 1856. Trobisch, David. Paul’s Letter Collection: Tracing the Origins. Minneapolis, MN: Fortress Press, 1994. Tuckett, Christopher M. “"Nomina Sacra”: Yes and No?” The Biblical Canons. Edited by J.-M. Auwers and H. J. de Jonge. Bibliotheca Ephemeridum Theologicarum Lovaniensium 163. Leuven: Leuven University Press/Peeters, 2003. Turner, E. G. Greek Papyri: An Introduction. Oxford: Oxford University Press, 1968. Turner, E. G., and P. J. Parsons. Greek Manuscripts of the Ancient World. 2nd ed., rev. and enl.. Bulletin Supplement 46. London: University of London, Institute of Classical Studies, 1987. Turner, Eric G. The Typology of the Early Codex. Philadelphia, PA: University of Pennsylvania Press, 1977. Vaganay, Léon. Initiation à la critique textuelle néotestamentaire. Paris: Librairie Bloud & Gay, 1934. Vaganay, Léon, and Christian-Bernard Amphoux. An Introduction to New Testament Textual Criticism. Translated by Jenny Read-Heimerdinger. 2nd ed., rev.. Cambridge: Cambridge University Press, 1991. Vercellone, Carlo. Dell’antichissimo Codice Vaticano della Bibbia Greca. Roma: Nella Tipografia della Rev. Camera Apostolica, 1860. Versace, Pietro. I marginalia del Codex Vaticanus. Studi e testi 528. Città del Vaticano: Biblioteca Apostolica Vaticana, 2018. ———. “Some Marginalia of the B Codex Vaticanus.” iSBL Annual Meeting. Rome, 2019. Wachtel, Klaus. Der byzantinische Text der Katholischen Briefe: Eine Untersuchung zur Entstehung der Koine des Neuen Testaments. Arbeiten zur neutestamentlichen Textforschung 24. Berlin: De Gruyter, 1995. ———. “Introduction.” Pages 18–27 in Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 1.1: Text. Edited by Holger Strutwolf, Georg Gäbel, Annette Hüffmeier, Gerd Mink, and Klaus Wachtel. Stuttgart: Deutsche Bibelgesellschaft, 2017. ———. “Notes on the Text of the Acts of the Apostles.” Pages 28–33 in Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 1.1: Text. Edited by Holger Strutwolf, Georg Gäbel, Annette Hüffmeier, Gerd Mink, and Klaus Wachtel. Stuttgart: Deutsche Bibelgesellschaft, 2017. ———. “On the Relationship of the ‘Western Text’ and the Byzantine Tradition of Acts – A Plea Against the Text-Type Concept.” Pages 137–48 in Novum Testamentum Graecum: Editio Critica Maior III: Acts of the Apostles: Part 3: Studies. Edited by Holger Strutwolf, Georg Gäbel, Annette Hüffmeier, Gerd Mink and Klaus Wachtel. Stuttgart: Deutsche Bibelgesellschaft, 2017. Wagstaffe, Thomas. “Some Account of the Famous Vatican MS, from Whence the Vatican Septuagint Was Printed.” 1739. MS Ges. 1162: Nazionale Centrale Vittorio Emanuele II, Rome. BIBLIOGRAPHY 310 Wallraff, Martin. Kodex und Kanon: Das Buch im frühen Christentum. Hans-Lietzmann-Vorlesungen 12. Berlin: De Gruyter, 2013. Wasserman, Tommy. “Criteria for Evaluating Readings in New Testament Textual Criticism.” Pages 579–612 in The Text of the New Testament in Contemporary Research: Essays on the Status Quaestionis. Edited by Bart D. Ehrman and Michael W. Holmes. 2nd ed. New Testament Tools, Studies, and Documents 42. Leiden: Brill, 2013. ———. The Epistle of Jude: Its Text and Transmission. Coniectanea Biblica New Testament Series 43. Stockholm: Almqvist & Wiksell Internat, 2006. ———. “Was There an Alexandrian Recension of the Living Text?” Pages 1–24 in Liturgy and the Living Text of the New Testament: Papers from the Tenth Birmingham Colloquium on the Textual Criticism of the New Testament. Edited by H. A. G. Houghton. Texts and Studies (Third Series) 16. Piscataway, NJ: Gorgias Press, 2018. Wasserman, Tommy, and Peter J. Gurry. A New Approach to Textual Criticism: An Introduction to the Coherence-Based Genealogical Method. Resources for Biblical Study 80. Atlanta, GA: Society of Biblical Literature, 2017. Weiss, D. Bernhard. Textkritik der paulinischen Briefe. Texte und Untersuchungen 14. Leipzig: J. C. Hinrichs, 1896. West, Stephanie. “Reclamantes in Greek Papyri.” Scriptorium 17 (1963): 314–15. Westcott, Brooke Foss, and Fenton John Anthony Hort. The New Testament in the Original Greek: Introduction. Appendix. London: Macmillan and Co., 1882. Wettstein, Johann Jakob. Ἡ Καινὴ Διαθήκη. Novum Testamentum Graecum editionis receptae cum lectionibus variantibus codicum mss., Editionum aliarum, Versionum et Patrum nec non commentario pleniore ex Scriptoribus veteribus Hebraeis, Graecis et Latinis historiam et vim verborum illustrante. tomus I. Continens quatuor Evangelia. Amsterdam: Dommer, 1751. Wevers, John Wm. “A Study in the Hebrew Variants in the Books of Kings.” Zeitschrift für die alttestamentliche Wissenschaft 61 (1948): 43–76. Williams, Peter J. “Semitic Long /i/ Vowels in the Greek of Codex Vaticanus of the New Testament.” Pages 15–26 in Semitic Linguistics and Manuscripts: A Liber Discipulorum in Honour of Professor Geoffrey Khan. Edited by Nadia Vidro, Ronny Vollandt, Esther-Miriam Wagner, and Judith Olszowy-Schlanger. Studia Semitica Upsaliensia 30. Uppsala: Acta Universitatis Upsaliensis, 2018. ———. “When Does Συν- Assimilate?” Pages 429–38 in The New Testament in Antiquity and Byzantium. Edited by H.A.G. Houghton, David C. Parker, and Holger Strutwolf. Arbeiten zur neutestamentlichen Textforschung 52. Berlin: De Gruyter, 2019. Wirth, Raimund. “Dealing with Tenses in the Kaige Section of Samuel.” Pages 185–97 in The Legacy of Barthélemy: 50 Years After Les Devanciers D’Aquila. Edited by Anneli Aejmelaeus and Tuukka Kauhanen. De Septuaginta Investigationes 9. Göttingen: Vandenhoeck & Ruprecht, 2017. ———. Die Septuaginta der Samuelbücher: Untersucht unter Einbeziehung ihrer Rezensionen. De Septuaginta Investigationes 7. Göttingen: Vandenhoeck & Ruprecht, 2016. 311 Woods, Christopher S. “The Conservation of Parchment.” Pages 200–224 in Conservation of Leather and Related Materials. Edited by Marion Kite and Roy Thomson. Butterworth-Heinemann Series in Conservation and Museology. Boston, MA: Elsevier Butterworth-Heinemann, 2006. Yi, An-Ting, Jan Krans, and Bert Jan Lietaert Peerbolte. “Prolegomena to Bentley’s Unfinished New Testament Project.” Novum Testamentum 62 (2020): 322–38. Yuen-Collingridge, Rachel, and Malcolm Choat. “The Copyist at Work: Scribal Practice in Duplicate Documents.” Pages 827–34 in Actes du 26e Congrès international de papyrologie. Genève 16-21 Août 2010. Edited by P. Schubert. Droz: Geneva, 2012. Zuntz, Günther. Lukian von Antiochien und der Text der Evangelien. Edited by Barbara Aland and Klaus Wachtel. Abhandlungen der Heidelberger Akademie der Wissenschaften, Philosophisch-Historische Klasse. Heidelberg: C. Winter, 1995. ———. The Text of the Epistles: A Disquisition Upon the Corpus Paulinum. Schweich Lectures. London: Oxford University Press, 1953. Bibliorum SS graecorum Codex Vaticanus 1209 (Cod. B): denvo phototypice expressus. 4 vols. Milan: U. Hoepli, 1904. Codex Vaticanus B. Bibliotheca Apostolicae Vaticanae Codex Vaticanus Graecus 1209. Roma: Istituto Poligrafico e Zecca dello Stato, 1999.