Sahidic 1 Samuel – A Daughter Version of the Septuagint 1 Reigns [1 ed.] 9783666540578, 9783525540572

143 57 2MB

English Pages [257] Year 2017

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Sahidic 1 Samuel – A Daughter Version of the Septuagint 1 Reigns [1 ed.]
 9783666540578, 9783525540572

Citation preview

De Septuaginta Investigationes (DSI) Edited by Anneli Aejmelaeus, Kristin De Troyer, Wolfgang Kraus, Emanuel Tov In Co-operation with Kai Brodersen (Erfurt, Germany), Cécile Dogniez (Paris, France), Peter Gentry (Louisville, USA), Anna Kharanauli (Tbilisi, Georgia), Armin Lange (Wien, Austria), Alison Salvesen (Oxford, GB), David Andrew Teeter (Cambridge, USA), Julio Trebolle (Madrid, Spain), Florian Wilk (Göttingen, Germany)

Volume 8

Vandenhoeck & Ruprecht

Elina Perttilä

Sahidic 1 Samuel – A Daughter Version of the Septuagint 1 Reigns

Vandenhoeck & Ruprecht

With 84 Tables

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data available online: http://dnb.d-nb.de. ISSN 2198-1140 ISBN 978-3-666-54057-8 You can find alternative editions of this book and additional material on our Website: www.v-r.de © 2017, Vandenhoeck & Ruprecht GmbH & Co. KG, Theaterstraße 13, D-37073 Göttingen/ Vandenhoeck & Ruprecht LLC, Bristol, CT, U. S. A. www.v-r.de All rights reserved. No part of this work may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without prior written permission from the publisher.

Acknowledgements This publication is based on my 2013 University of Helsinki dissertation. I would like to thank my supervisors, Prof. Anneli Aejmelaeus and Prof. Raija Sollamo. My cordial thanks are due to the pre-examiners, Prof. Anne Boud’hors and Prof. Melvin Peters for their comments. Prof. Boud’hors also served as opponent in the public defence of my thesis and offered many suggestions for improvement. I also wish to express my gratitude to Dr. Christian Askeland, Dr. Frank Feder, Dr. Tuukka Kauhanen, Marketta Liljeström, Prof. Antti Marjanen, Ivan Miroshnikov, Jessi Orpana, Christian Seppänen, Dr. Alin Suciu, Dr. Ulla Tervahauta-Helin, Timo Tekoniemi, Miika Tucker, Dr. Kirsi Valkama, Hanna Vanonen and Dr. Raimund Wirth.

Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.1. The Septuagint and its daughter versions . . . . . . . . . . . . . . . . . . . . . . 11 1.2. Septuagint of 1 Samuel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1.3. Manuscripts of Sahidic 1 Samuel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.4. Sahidic 1 Samuel and its research history . . . . . . . . . . . . . . . . . . . . . . 25 1.5. Aims and methods of this study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2. Description of the translation technique . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Clause connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Questions and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2. Occurrences of clauses in Greek . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3. Coordinated clauses and their renderings . . . . . . . . . . . . . . . . 2.1.3.1. καί-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3.1.a) Renderings of καί-clauses . . . . . . . . . . . . . . . . . . . . 2.1.3.1.b) καί-clause rendered asyndetically . . . . . . . . . . . . . 2.1.3.1.c) καί-clause rendered with ⲁⲩⲱ . . . . . . . . . . . . . . . . . 2.1.3.1.d) καί-clause rendered with ⲇⲉ . . . . . . . . . . . . . . . . . . 2.1.3.1.e) καί-clause rendered with other connectives . . . . . 2.1.3.1.f) καί-clause without equivalent . . . . . . . . . . . . . . . . . 2.1.3.2. Asyndetic clauses and their renderings . . . . . . . . . 2.1.3.2.a) Asyndeton rendered with asyndeton . . . . . . . . . . . 2.1.3.2.b) Asyndeton rendered with a conjunction . . . . . . . 2.1.3.3. Interrogative clauses and their renderings . . . . . . 2.1.3.4. Other coordinated clauses . . . . . . . . . . . . . . . . . . . . 2.1.3.4.a) ὰλλά-clauses and their renderings . . . . . . . . . . . . . 2.1.3.4.b) γάρ-clauses and their renderings . . . . . . . . . . . . . . 2.1.3.4.c) δέ- and οὐδέ-clauses and their renderings . . . . . . 2.1.3.4.d) ἤ-clauses and their renderings . . . . . . . . . . . . . . . . . 2.1.3.4.e) οὖν-clauses and their renderings . . . . . . . . . . . . . . . 2.1.3.4.f) οὕτως-clauses and their renderings . . . . . . . . . . . . 2.1.3.4.g) πλήν-clauses and their renderings . . . . . . . . . . . . . 2.1.4. Subordinate clauses and their renderings . . . . . . . . . . . . . . . . 2.1.4.1. Relative clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4.2. ἐάν -clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4.3. εἰ-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32 32 32 34 34 34 35 36 40 45 53 58 64 64 68 70 75 75 79 80 82 83 85 85 86 86 90 93

8

Contents

2.1.4.4. ἕως-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 2.1.4.5. καθώς- and καθά-clauses . . . . . . . . . . . . . . . . . . . . 98 2.1.4.6. μήποτε-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 2.1.4.7. ὅπως-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 2.1.4.8. ὅτε- and ὅταν-clauses . . . . . . . . . . . . . . . . . . . . . . . . 101 2.1.4.9. ὅτι-, διότι- and καθότι-clauses . . . . . . . . . . . . . . . . 101 2.1.4.10. ὡς-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 2.1.4.11. ὥστε-clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 2.1.5. Non-finite clauses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 2.1.5.1. Participium coniunctum and its renderings . . . . . 112 2.1.5.2. Genetivus absolutus and its rendering . . . . . . . . . . 116 2.1.5.3. Infinitive constructions and their renderings . . . . 117 2.1.6. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 2.2. Additions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 2.2.1. Questions and method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 2.2.2. Additions based on a grammatical difference . . . . . . . . . . . . . 124 2.2.2.1. Completion of an incomplete clause . . . . . . . . . . . 124 2.2.2.2. Relative clause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 2.2.2.3. An auxiliary word in translation . . . . . . . . . . . . . . . 126 2.2.2.4. Possessive article . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 2.2.3. Logical additions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 2.2.3.1. Addition of direct or indirect object . . . . . . . . . . . 127 2.2.3.2. Addition of a second verb . . . . . . . . . . . . . . . . . . . . 129 2.2.3.3. Addition of the subject . . . . . . . . . . . . . . . . . . . . . . . 130 2.2.3.4. Introduction of a speech . . . . . . . . . . . . . . . . . . . . . . 130 2.2.4. Stylistic additions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 2.2.4.1. An adverb, a prepositional phrase or an adjective 132 2.2.4.2. A complement to the subject or the object . . . . . . 133 2.2.4.3. Explanation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 2.2.4.4. Interjection or a short comment . . . . . . . . . . . . . . . 135 2.2.4.5. Other additions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 2.2.5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 3. Affiliations of the Sahidic manuscripts and the Greek Vorlage of the translator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 3.1. Questions and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 3.2. Textual analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 3.2.1. SaMAV 7: 8–8: 1, 9: 21–10: 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 3.2.2. SaMI 12: 4–5,10–11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 3.2.3. SaMAU 14: 24–32 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 3.2.4. SaMAB 17: 33– 44 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 3.2.5. SaMAFJ 29: 5–9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 3.2.6. SaMAJ 30: 21–24 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 3.2.7. SaMAB 31: 1–13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202

Contents

9

3.3. Special cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 3.3.1. Text-critical notes in the clause connections chapter . . . . . . 218 3.3.2. Corrections towards the MT . . . . . . . . . . . . . . . . . . . . . . . . . . . 220 3.4. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 5. Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 6. Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 6.1. Collation of SaI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 6.2. Collation of SaJ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 6.3. Collation of SaU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240 6.4. Collation of SaY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 6.5. Greek manuscripts and their groupings . . . . . . . . . . . . . . . . . . . . . . . 244 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 Index of Biblical References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

1. Introduction This study considers the Sahidic Coptic version of 1 Samuel (1 Sam) and how it may best be used in Greek textual criticism.1 The Sahidic translation is a daughter version of the Septuagint. The Septuagint itself is the Greek translation of the Hebrew Bible, rendered beginning in the 3rd century BCE, probably in Alexandria. These Jewish translators rendered the Pentateuch into Greek, and, later, additional texts from the Hebrew corpus were translated and incorporated into this corpus. The translation of 1 Sam probably happened sometime in the 2nd century BCE.2 The text of this original translation is referred to as the Old Greek. Since no manuscript preserves this original text, the OG must be recovered through textual criticism. Further studies on the Septuagint, including retroversions to Hebrew and comparison with other textual witnesses, should be based on the OG text. The Septuagint was later adopted by early Christians as their Scripture, whereas Jews eventually replaced it.3

1.1. The Septuagint and its daughter versions The Septuagint has ancient daughter versions in Latin, Coptic4, Ethiopic, Armenian and Georgian.5 Additionally, there was the Syro-Hexapla (a Syriac version of the Hexapla), but only fragments of 1 Sam survive from it.6 1 2

3

4

5 6

1 Samuel is also referred to as 1 Kingdoms, like 1 Reigns (1 Regnorum), in the Septuagint. I use the name Samuel in order to avoid confusion with 1 Kings. For more information about the origin of the Septuagint, see Jan Joosten, “Reflections”. Concerning the dating of the 1 Sam translation, Anneli Aejmelaeus (“Corruption or Correction”, 16) dates some revisions to the MT to the 1st century BCE with no parallel in the Septuagint of 1 Sam. Thus, the LXX of 1 Sam was rendered before that time. Since the Pentateuch was rendered first, the rendering of 1 Sam happened perhaps not before the second century BCE. Emanuel Tov, Textual Criticism, 143. The modern situation concerning the OT canon is complicated. The protestant reformers reverted to translations based upon the Masoretic Hebrew Bible and its canon, while the Roman and Eastern Churches continued to prefer the Septuagint and related translations. See also John J. Collins, Introduction to the Hebrew Bible, 2–7. For a more thorough discussion, see Lee Martin McDonald, The Biblical Canon, and his useful table on pp. 443–4. Coptic is the latest phase of Egyptian, whose written history started already in the late 4th millennium BCE. For a concise description of the setting and history of the EgyptianCoptic language, see Grossman/Richter, The Egyptian-Coptic language. Basically, daughter versions are translations that have the Septuagint as their source text. Marketta Liljeström, “Looking for Fragments of the Syrohexapla”.

12

Introduction

The two first mentioned, Latin and Coptic, are important for the textual history of the Greek text since they were translated before most extant Greek manuscripts were copied.7 Therefore, through these versions, we might have access to older text traditions than the text in our preserved Greek manuscripts. The Old Latin version is preserved only fragmentarily, and therefore the Coptic becomes even more important. The translation of 1 Sam into Sahidic Coptic is usually dated to the 3rd century CE.8 The Sahidic translation of the Septuagint of 1 Sam was made by Christians. Obviously, Christianity came to Egypt early, but our knowledge of the early stages begins only with Demetrius’ rise as bishop of Alexandria in 189 CE.9 There are Greek fragments of the New Testament coming from Egypt, dated to the 2nd – 3rd century that indicate the presence of Christianity at that time. However, the dating of these fragments is not absolute and it is mostly impossible to get a grasp of 2nd- century Christianity with the help of the papyri, be they literary or documentary.10 If the origin of Christianity in Egypt remains obscure, so also do the first Coptic Bible translations. We do not know exactly when and where the first translations arose, nor is it clear in which order the books of the Septuagint were translated into Coptic.11 At the moment, the research situation and research questions

7

For more information on the Latin version, see Eugene Ulrich, The Dead Sea Scrolls and the Origins of the Bible, 275–89. 8 Paul E. Kahle (Bala’izah, 11) contends that the translation of the Sahidic OT happened “not later than the late 3rd century”. He bases this dating on manuscript evidence. Gregor Emmenegger (Der Text des Koptischen Psalters, 12–13) dates the Sahidic versions of the Psalter to the 3rd century, and the Bohairic ones in the 4th – 5th century. Frank Feder (Biblia Sahidica, 3) dates at least the beginning of the Sahidic OT translation in the 3rd century, based on the earliest fragments from the 4th (or even 3rd) century and the assumed time span between the original translation and these fragments. John Barton Payne (Critical and comparative, 17) has an early dating: he dates some parts of the Sahidic Bible before the end of the 2nd century, and the rest of it before Origen’s Hexapla c. 245. However, the arguments of Payne concerning the lack of recensional readings in Sahidic 1 Sam and the needs of missionaries are not as convincing and accurate as preserved manuscripts when it comes to the dating of the translation. 9 Roger S. Bagnall, Early Christian Books, 4–5. The main source of information is Eusebius’s Church History, and according to Bagnall, “it does not give one the impression that Eusebius knew a lot”[about the time before Constantine, 313 CE]. 10 See Bagnall, Early Christian Books, 7–11. 11 The Coptic language had many dialects. The literary dialects are Sahidic, Bohairic, Achmimic, Lycopolitan, Mesokemic and Fayyumic. Besides these dialects, there are some minor dialects. For more information, see Wolf-Peter Funk, “Dialects Wanting Homes: A Numerical Approach to the Early Varieties of Coptic”. For a short introduction to the New Testament manuscripts in different Coptic dialects, see Askeland, “The Coptic versions of the New Testament”. 1 Sam was translated into the Sahidic dialect of Coptic, which was the primary literary dialect beginning from the 3rd century until the rise of Bohairic in the 8th-9th centuries. From the 12th century onwards, the Sahidic version was replaced by Bohairic. For the Fayyumic version, from which one tiny fragment has survived from 1 Sam 25, it is difficult to know whether the entire 1 Sam ever existed. For more information about preserved Fayyumic literature, see Anne Boud´hors, “Manuscripts and

The Septuagint and its daughter versions

13

are taking shape, and new initiatives and questions arise. There is a consensus among scholars that mostly the dating of Coptic manuscripts has been haphazard, and datings based on paleographical arguments need reassessment.12 The earliest witness of Sahidic 1 Samuel has been dated to the 4th century.13 The use of the daughter versions in the textual criticism of the Greek text has many problems that need attention.14 One must consider that each translation also has its own textual history.15 Variant readings stem from both the copying processes in Greek and in the daughter versions. Intentional changes also occur in the textual traditions, either as corrections according to some other text form or because of linguistic issues in the relevant languages. The textual history of the Greek text needs to be evaluated separately from the textual history of the Coptic text.16 At the same time, however, one has to consider the overall picture to avoid assumptions that do not fit into the framework. The framework, as long as Coptic is used in textual criticism of the LXX, is the Greek text and its textual history. With regard to the Greek text and its history, it is presumed in this study that our extant Greek witnesses all descend from one common hypothetical ancestor —the Old Greek, that is, the original Greek translation. Thus, the aim of studying the Greek text and its history is to trace back the history of the text and to reconstruct the OG on the basis of the extant manuscripts as well as the daughter versions.

12

13

14

15 16

Literature in Fayoumic Coptic”. For some reason, she does not mention the tiny fragment of 1 Sam but states that there are no preserved fragments from historical books. For an overview of the situation in different dialects, see Feder, “Coptic Translations”. Concerning the questions urgent in dating and paleography, see Feder, “Die koptische Übersetzung des Alten und Neuen Testamentes im 4. Jahrhundert”. Aaron Michael Butts, “P.Duk.inv. 797”, 10. This fragment has the siglum sa 77 in Karlheinz Schüssler’s Die koptischen Bibeltexte 1: 1, 75. Schüssler references a possible 3rd-century date. Butts himself proposes a dating to the 4th century. J.W. Wevers (“The Use of Versions for Text Criticism: The Septuagint”, 15–20) provides three propositions concerning versions and their use in textual criticism of the source language: 1) Before a version can be used text critically the nature and limitations of the target language in contrast to those of the source language must be thoroughly understood. 2) By version one can only mean the actual translation itself, and not later corruption or revisions of it. 3) One must fully understand just how and from what points of view this translation was done by a particular translator. For an example from the Ethiopic version, see Michael A. Knibb, “The Greek Vorlage of the Ethiopic text of Ezekiel”, 416–17. The same principle is found in Ulrich, The Dead Sea Scrolls and the Origins of the Bible, 276–8.

14

Introduction

1.2. Septuagint of 1 Samuel The Greek translation of 1 Sam has been described as one of the most literal in the Septuagint.17 The literalness of this translator is most apparent in some particular areas, as pointed out by Aejmelaeus. First, one notes the use of the conjunction καί as the rendering of the Hebrew conjunction ‫ו‬ together with the rare use of the connective δέ. Second, the translator has used apodotic conjunctions abundantly, against normal Greek idiom.18 Third, the use of participium coniunctum is very limited in 1 Sam, suggesting that the translator did not render larger units at a time. The other participial construction, genetivus absolutus, occurs fairly frequently in Greek 1 Sam as compared to its frequency in the Pentateuch.19 Besides gen.abs., the translator’s use of various verbal forms further reveals his tendencies. His use of the Greek historical present and the imperfect as well as contextual renderings show that he is not exhaustively literal. As Aejmelaeus concludes, this Greek translator proceeded by rendering one small unit at a time, but within such a unit, he could be quite free and use idiomatic Greek expressions.20 The Greek translator mostly proceeded on a word-for-word basis, and this enables a retroversion into Hebrew more easily than in those books where the translator rendered larger units with freedom. In those passages in 1 Sam where variation exists between the LXX and the MT, the differences probably do not stem from the translator who usually follows his Vorlage literally. The Septuagint of 1 Sam was translated from a Hebrew text at times different from the MT. This view has two bases: 1) the translation technique behind the Septuagint and 2) Hebrew Samuel manuscripts found at Qumran. The fragments of three Samuel scrolls found at Qumran confirm that there are readings where the LXX and a Qumran reading agree against the reading of the MT.21 According to Aejmelaeus, a consensus exists among scholars that the MT of 1 Sam “contains numerous grave 17 Ilmari Soisalon-Soininen (Infinitive, 177–8, 186) puts 1 Sam in the group that holds an intermediate position between the most slavish and most free translations, but within this group, 1 Sam is nearer to the literal than the free extreme. Raija Sollamo (Renderings of Hebrew Semiprepositions, 286–7) has a grouping based on the renderings of the Hebrew semiprepositions, and she places 1 Sam in the third out of four groups, the fourth group consisting of the most literal and slavish translations. For more literature, see A. Aejmelaeus, On the Trail, 128, note 23. 18 Aejmelaeus, On the Trail, 128–9. Typically, Greek καί does not occur at the start of the apodosis, when the apodosis follows the protasis. When this does occur against the general tendency of the language, this is referred to as an apodotic καί. 19 Aejmelaeus, On the Trail, 134–5. 20 Aejmelaeus, On the Trail, 136–41. 21 Samuel fragments from Qumran are published in Cross, F.M./Parry, D.W./Saley, R.J./ Ulrich, E.C., Qumran Cave 4. Before using fragmentary evidence from Qumran in comparison with the MT and LXX, see the critique of Tuukka Kauhanen (The Proto-Lucianic Problem, 165–6) concerning the reconstructions of the fragments.

Septuagint of 1 Samuel

15

errors and defects”. Additionally, the MT provides examples of deliberate editing.22 This makes the LXX a particularly helpful witness to the Hebrew text of 1 Sam. The textual history of the Greek text informs the textual history of the Sahidic. One should avoid proposing theories that do not cohere with the known history of the LXX. The Septuagint has been reworked, slightly or more thoroughly, several times during its history, and this complicates its study. In addition to the LXX, further Greek OT traditions like the καιγε, Theodotion, and Symmachus translations also play a role in the textual history and need to be taken into account in order to most fully comprehend the history of the LXX.23 Concerning textual history, the two terms recension and revision require clarification and careful use. These terms should always reference patterns of variants and variation which were consciously made and which can be ascribed to a common transmissional event.24 Aejmelaeus speaks of motives in this connection. One has to find the direction of the changes that were made during the recensional process.25 If one cannot find such a direction or common denominator, but only a few random changes, then the terms approximation and reformulation are appropriate. Approximation refers to a change towards some source language text, whereas a reformulation has no connection to a variant source text. Indisputably, Hexaplaric and Lucianic recensions are essential when the Septuagint of 1 Sam is concerned. In their very nature, these two recensions are different.26 The Hexaplaric recension was concerned with quantitative equivalence. Those Greek passages missing from the MT are marked with an obelos. Likewise, those passages that are present in the MT but missing from the LXX are supplied and also indicated by an asterisk. This recension arose as Origen’s ‘fifth column’ which itself was based on a comparison of the Greek and Hebrew witnesses available at the time.27 Additionally, this redaction produced doublets. If no Greek equivalent identifiable to the reviser existed for something in the Hebrew source, he inserted one into the

22 Aejmelaeus, “Corruption or Correction”, 6. 23 Aejmelaeus, “Corruption or Correction”, 3. 24 See also Ulrich, The Dead Sea Scrolls and the Origins of the Bible, 89–94. Tov, Textual Criticism, 160–1: “A witness reflecting a text-type or recension by definition should show a conscious effort to change an earlier text systematically in a certain direction.” 25 Aejmelaeus, “Corruption or Correction”, 4: “Such intentional features, wherever encountered, are particularly valuable for textual criticism, as they give us a clue of what can be expected of this textual witness and which explanations are available in individual cases. In case of repeated intentional changes toward a certain direction it is justified to speak of conscious editing of the text. In such cases, it may be even possible to discern the motive behind the changes.” 26 Aejmelaeus, “Kingdom at Stake”, 358 n. 16. See also her On the Trail, 245–6. 27 Sebastian P. Brock, Recensions, 170.

16

Introduction

Greek text. Some of the doublets result from Origen’s work, while others have some (probably) earlier origin.28 The Lucianic text contains several layers. First, there is an old base text that diverged from the rest of the Greek tradition early.29 Second, it preserves a significant number of corrections according to Hebrew.30 Third, some stylistic improvements concern grammar as well as lexis. Fourth, some changes aim at a text designed for public reading.31 The third and fourth features are those belonging to the recension proper. The Hesychian recension has sometimes been connected especially with the Sahidic text. Jerome has described the recension as follows: Alexandria and Egypt attribute their recension to Hesychius; from Constantinople to Syrian Antioch, the approved text is that of Lucian the Martyr, while in Palestine that of Origen as propagated by Eusebius and Pamphilus holds the field.32

The man Hesychius remains shrouded in mystery. Sidney Jellicoe advocated identification with an Egyptian martyr-bishop who died during the Diocletian persecutions.33 Léon Dieu, in his early study, suggests that the Sahidic text was translated from the Hesychian recension.34 Dieu himself does not state which manuscripts he includes within this group, but J. Barton Payne has, after a “process of elimination”, concluded that the group consists of the following manuscripts: M N(=V) 107 55 56 243 119 245 29 46 246.35 Partly in accordance with Dieu, Alberto Vaccari maintains that, in the historical books, the Hesychian recension is preserved in the following mss: M V 55 56 119 158. Vaccari substantiates this argument upon the dependencies between the Hesychian recension, Coptic versions and the Egyptian fathers.36 The theory of a Hesychian recension has encountered critiques from several scholars. Payne made a justified reservation by asking whether the manuscripts listed above constitute a distinct group.37 Sebastian Brock forcefully rejected the idea of a recension based on the arguments provided by Vaccari. First, he questions the argumentation for a recension in historical books, as Vaccari had studied prophetic books. Second, Brock did not find 28 Brock, Recensions, discusses the latter doublets on pp. 158–66. 29 Ibid., 306. 30 Ibid., 170–1. A. Aejmelaeus (“What Rahlfs Could not Know”, 89) mentions that the Lucianic recension acquired its corrections in Greek, mostly from various columns of the Hexapla. 31 Brock, Recensions, 298. 32 Cited here following Sidney Jellicoe, “Hesychian”, 409. 33 Jellicoe, “Hesychian”, 414. 34 L. Dieu, “Manuscrits Grecs”, 17–60. 35 Payne (Critical and comparative, 349) lists M, N, dhijnvb2, 46, 246. Payne also supposes a (707) and g (158) as belonging to this same group. Thus, it resembles Brock’s MN+ with some additional mss (Brock, Recensions, 19). 36 Vaccari, “Hesychian Recension”, 61. 37 Payne, Critical and comparative, 349.

Manuscripts of Sahidic 1 Samuel

17

“any close link between MN+ and Co[ptic] or Bo[hairic] on the one hand, or the scanty quotations of the Alexandrine fathers on the other”. If there is a need to find an affiliation for Coptic, Brock proposes Bb Eth for that purpose, but adds that in these manuscripts no recensional activity is found. He concludes: “In the textual tradition, as it comes down to us, of I Kms [= 1 Sam] there is no evidence for definite recensional activity outside the work of Origen and ‘Lucian’.”38 Actually, the whole discussion concerning the Hesychian recension should be inverted; one should begin with actual variants and manuscripts, and only if there is evidence of a recensional work, then proceed to naming. After Brock’s study, Aejmelaeus discerned additional recensional activity in Greek 1 Sam not connected to the name Hesychius, but rather to καιγεTheodotion. Previously, 1 Sam has been considered as unrelated to this recension, whose influence is usually thought to begin in 2 Sam.39 In 1 Sam, these approximations towards the MT appear sporadically, and typically in manuscript B. They include changes in vocabulary and omissions of plusses in the Greek text.40 In the passages discussed by Aejmelaeus so far, the Sahidic text does not usually follow the secondary readings. However, there are some passages in which the Sahidic text agrees with early corrections according to Hebrew, and these cases are of particular interest when tracing out the textual history of the Sahidic version.41 No direct influence of a Hebrew text is perceivable since the Sahidic does not correct the mistranslations of the Greek translator. Accordingly, Sahidic acquired its readings from Greek sources that contained these corrections towards the MT.

1.3. Manuscripts of Sahidic 1 Samuel Smaller or larger pieces of 24 manuscripts of Sahidic 1 Sam are extant, and only one among them is complete.42 This manuscript, for which I use the 38 Brock, Recensions, 33–4. 39 In Samuel-Kings, a recension called καιγε is widely recognized thanks to the discovery of the Naḥ al Ḥ ever Minor Prophets scroll and the work of Dominique Barthelémy (Les Devanciers d’Aquila). The books of Kingdoms are divided as follows (originally Thackeray’s division): α 1 Sam, ββ 2 Sam 1: 1–11: 1, βγ 2 Sam 11: 2- 1 Kgs 2: 11, γγ 1 Kgs 2: 12– 21: 43, γδ 1 Kgs 22 – 2 Kgs. Two of these sections, βγ (2Sam 11: 2 – 1Kgs 2: 11) and γδ (1Kgs 22 – 2Kgs), are representatives par excellence of the καιγε recension. This recension consists of early corrections towards the proto-Masoretic Hebrew text. These corrections are found in Vaticanus (B) and the majority of mss. 40 For examples, see Aejmelaeus, “Kingdom at Stake”, 354–9, 366. 41 These are analyzed in section 3.3.2. 42 I was informed in the IACS congress in Rome (September 2012) by Joost Hagen, that there are Sahidic fragments among the fragments that have been found in the excavations in Qasr Ibrim. He is currently working on these fragments. Unfortunately, at the moment I only know that these fragments exist, and what is even more important, some of them contain text from chapter 1 Sam 17 that is of special interest (see my section 3.2.4). One additional

18

Introduction

siglum SaM,43 is one of the Hamuli-manuscripts and belongs to the collections of The Morgan Library & Museum.44 SaM contains a colophon with the date 892/3.45 Although this manuscript preserves almost all of its original folios,46 the manuscript’s text often leaves much to be desired.47 James Drescher has edited the manuscript and also created an apparatus containing the different readings found in the fragments of other Sahidic manuscripts of 1 Sam that were known to him (SaA through SaT, in alphabetical order). In addition to the one complete manuscript, the following fragments of the Sahidic 1 Sam are preserved:48 SaA contains more than one third of the text: 6: 2–10; 6: 11–10: 3; 14: 17–32; 17: 31–44; 18: 28–22: 7; 22: 21–23: 14; 24: 21–25: 28; 28: 16–30: 1, 3–5; 30: 5–24; 30: 24–31: 13. The ten known fragments of this manuscript are now scattered across Europe.49 I have collated myself the fragments held in Vatican (6: 11–

43 44

45 46 47

48

49

manuscript not listed and used in my study is sa 158 according to Schüssler’s listing. This is an unedited papyrus manuscript in a poor state of preservation, and has preserved text from the following verses: 1 Sam 14: 50–52; 15: 1–3; 17(?); 28: 19–24; 29: 8–11; 30: 1. The fragments are kept in the Cambridge University Library, and K. Schüssler (Die koptischen Bibeltexte 2: 1, 95) has a detailed list of the references. He also mentions that Sarah J. Clackson has dated this manuscript to the 5th century. Schüssler has given the siglum sa 177lit to a bilingual Odes manuscript that contains the first verse from Song of Hannah (1 Sam 2: 1) in Sahidic and in Greek (Schüssler, Die koptischen Bibeltexte 2: 1, 129–30). In his edition, James Drescher marks this manuscript with M, while in Brooke/McLean this manuscript is marked Cw. I use Drescher’s sigla as index letters, e. g. SaM. The reference number is M567. See also Leo Depuydt, Catalogue of Coptic Manuscripts in the Pierpont Morgan Library. In Schüssler’s listing it has the siglum sa 25 (K. Schüssler, Die koptischen Bibeltexte 1: 2, 29–30). Schüssler states on p. 30, “Es handelt sich um eine unabhängige Übersetzung aus dem Griechischen”, without giving any further arguments. This appears very strange to me since such a conclusion needs solid argumentation. There are earlier fragments of Sahidic 1 Sam that clearly have the same Sahidic base text as this complete manuscript. For more information about the Hamuli-manuscripts and their significance, see Stephen Emmel, “The Library of the Monastery of the Archangel Michael at Phantoou (al-Hamuli)”. ⲁⲡⲟⲩ ⲇⲓⲱⲕⲗ [c. 7] ⲭ̅ⲑ̅ ‘year of Diocletianos 609’. This manuscript contains both 1 Samuel and 2 Samuel. The missing verses are 2 Sam 15: 20–30 (J. Drescher, Kingdoms, xii). I have read the actual manuscript, where the writing is badly damaged. The writing comes through the page to the other side, and occasionally the text is illegible, at least without ultraviolet light. I refer to the list of Drescher in his edition (Drescher, Kingdoms, ix-xiii) and provide here only the very basic information on each manuscript. My cordial thanks to prof. Schüssler who provided me with a list of 1 Sam manuscripts with their reference numbers in different collections. In Schüssler’s list, this manuscript has the siglum sa 24. K. Schüssler (Die koptischen Bibeltexte 1: 2, 20–6) gives the exact details with reference numbers in each library. The same statement that was given with the previous manuscript (SaM) is iterated here: “Es handelt sich um eine unabhängige Übersetzung aus dem Griechischen; sa 24 und sa 25 scheinen auf dieselbe Vorlage zurückzugehen.” What is the independence meant here? Obviously, these two manuscripts have a common source text according to Schüssler.

Manuscripts of Sahidic 1 Samuel

19

10: 3) and in Vienna (17: 31–44).50 This manuscript is dated to the 10th–11th century and once resided in the White Monastery. SaB is dated to the 11th–12th century, and contains such remarkable omissions that Gaston Maspero characterized it as a chronicle.51 Drescher transcribed this manuscript separately in its entirety, and the present study uses his edition.52 Additionally, I have collated the fragments held in the Vatican library (17: 33–19: 5) and Vienna collection (16: 2–8, 11–18). This manuscript comes from the White Monastery, and incorporates some Fayyumic features. In addition to 1 Sam (16: 2–8, 11–18; 17: 33–19: 5; 26: 7–25; 28: 3–25; 31: 1–13), the manuscript contains text from Judges and 2 Samuel. SaC, a parchment leaf, preserves the Song of Hannah, verses 2: 1–10. The fragment, edited by J. Schleifer, resides in the British Library.53 SaD is a bilingual Odes manuscript with a Greek and a Sahidic version of the Song of Hannah (2: 1–10). This manuscript is fragmentary, and only the beginnings of the lines are preserved. Walter Till and Peter Sanz edited this manuscript which belongs to the Vienna collection.54 I have also collated the manuscript myself. SaES,55 a papyrus manuscript, is preserved in several small fragments. It is dated to the 7th century based on paleography. Paul Kahle has assembled and edited the following verses: 2: 24–30; 3: 6–9; 6: 14–21; 14: 3, 5, 7, 10, 11; 15: 13, 15, 17, 19, 20; 21: 13–14; 22: 1–6; 24: 12, 15, 17–20. These fragments

50 The fragment in Vienna collection has been edited by Carl Wessely, Griechische und koptische Texte theologischen Inhalts, 71–2. After Wessely’s edition the fragment has been cleaned, and in its present state of preservation there is legible text in several points that Wessely had to reconstruct. The actual readings of the fragment are presented in my section 3.2.4. The Vatican fragment is part of Borgia’s collection. For more information about the collection, see Paola Buzi, “Stefano Borgia’s Coptic Manuscripts Collection”. 51 G. Maspero, Fragments de la version thébaine, 159. 52 Drescher, Kingdoms, 183–90. Schüssler (Die koptischen Bibeltexte 1: 2, 17–19) has assigned the siglum sa 23, and offers the reference numbers in three collections. 53 The reference number is BL Or. 4717(1). This manuscript does not yet have a number in Schüssler’s listing. The edition is found in J. Schleifer, Sahidische Bibel-Fragmente, 2–5. According to Walter Ewing Crum (Catalogue of the Coptic manuscripts in the British Museum, 4) and Schleifer, this leaf contains also the Odes of Moses (Deut 32: 30–43) and Habakkuk (Hab 3: 1–7). A detailed analysis of Hannah’s Song in Sahidic, an article of mine will soon be published, with the title “Sahidic Song of Hannah”. 54 Till, W./Sanz, P., Eine griechisch-koptische Odenhandschrift, 67–71. In the collections of ÖNB, the reference number is K 8706. The manuscript has the siglum sa 16lit according to Schüssler’s list (Schüssler, Die koptischen Bibeltexte 1: 2, 89). The edition of Till/Sanz reconstructs the missing text based on the manuscript known to them, SaC since SaM was not yet edited. 55 Schüssler (Die koptischen Bibeltexte 2: 1, 37–8) has been able to verify the possibility only mentioned by Drescher (Kingdoms, xi) “These fragments are very like those of S and it is not excluded that they all come from one and the same MS.” In Schüssler’s listing this manuscript has the siglum sa 134.

20

Introduction

belong to the collections of the Bodleian Library, Oxford.56 I have used Kahle’s edition and collated the fragments myself. Former SaS, now part of SaE adds the following verses to the text of the manuscript: 1: 18–2: 9; 5: 6– 11; 7: 15–8: 3; 8: 6, 8, 9; 9: 12–19; 12: 19–24; 13: 2–5, 6,8–9; 22: 6–13; 30: 3–11. The fragments belong to the collections of Corpus Christi College, Cambridge.57 These fragments have not been edited, and I have used Drescher’s apparatus where he cites all the variants deviating from SaM. SaF is a parchment leaf, dated to the 5th– 6th centuries. In addition to Kahle’s edition,58 I have collated this fragment myself from the manuscript. It preserves verses 29: 3–9. This fragment is kept in the Bodleian Library.59 SaG is a White Monastery parchment leaf, edited by Maspero.60 This fragment preserves verses 3: 8–20 and belongs to the collection of the Bibliothèque nationale, Paris.61 SaH is a parchment leaf of unknown origin and holding institution. Its editor, Oscar von Lemm, informs us that the leaf was offered to the Berlin Museum for purchase and that he used Prof. Schäfer’s copy for the edition. This leaf contains verses 8: 17–22.62 An early dating is proposed, based on the small page size (leaf 10 cm x8,5 cm, one column, 20 lines, 7 cm x 5,5 cm writing area).63 SaI is a fragment of a parchment leaf that belongs to the collections of the British Library and was edited by Schleifer.64 A small amount of text from four verses (12: 4–5, 10–11) is preserved. A dating to the 4th–5th century has been proposed.65 I have used a photo for the collation, and present my collation as an appendix since it differs slightly from Schleifer’s edition. SaJ is preserved in two parchment leaves, which are kept in the British Library.66 I have collated these fragments from a photo. The leaves differ 56 Kahle, Bala’izah, 301–11. 57 The reference numbers are listed in Schüssler, Die koptischen Bibeltexte 2: 1, 37–8. 58 Kahle, Bala’izah, 312–14. Schüssler (Die koptischen Bibeltexte 2: 1, 94) has given the siglum sa 157 to this manuscript, of which only one leave is known. 59 Its reference number is MS.Copt. e14, d188a, b (P). 60 Maspero, Fragments de la version thébaine. In Schüssler’s listing (Die koptischen Bibeltexte 2: 2, 94) this ms has the siglum sa 217L. 61 The reference number is Copte 1291 fol. 116. 62 Drescher (Kingdoms, xi) seems to have a typo since he has used Lemm’s edition but gives the beginning of the fragment as 8: 16. Schüssler has not yet published his siglum for this leaf. 63 Lemm, Sahidische Bibelfragmente, 2. 64 J. Schleifer, Bruchstücke, 14–15. In the British Library, the reference number is BL Or. 4916(3). Schüssler (Die koptischen Bibeltexte 2: 2, 57) has given the siglum sa 208 to this fragment. 65 Schüssler, Die koptischen Bibeltexte 2: 2, 57. 66 The reference number is BL Or. 4916(4). Schüssler has not yet published this item in his listing.

Manuscripts of Sahidic 1 Samuel

21

notably with regard to their preservation. The one (containing verses 30: 21– 24) is in a good condition and easily legible, whereas the other (verses 29: 5– 9) is wrinkled and difficult to read. Schleifer has published the first-mentioned leaf but not the latter. Therefore, I present my own collation as an appendix. There are also some differences between my collation and Schleifer’s in verses 30: 21–24. SaK has preserved two mutilated parchment leaves with only parts of some lines from verses 14: 49–52; 15: 1, 2, 21–26, 29, 30. These leaves are kept in Louvain and were edited by L.T. Lefort.67 SaN is a large parchment fragment of a katameros (Lectionary), edited by Maspero.68 This fragment contains text from verses 12: 5–8. It is kept in Paris.69 SaQ is a scala,70 dated by colophon to 1296 and 1310. The text contains phrases from 1–2 Sam. This manuscript belongs to the collections of the Bibliothèque nationale de France.71 SaR is also a scala, dated 1389. This manuscript contains the same phrases as SaQ, plus one extra from 2 Sam. This manuscript is also in Paris.72 SaS see SaE above. SaT is an ostracon that resides in the collection of the British Museum. It contains text from verses 16: 4–5.73 SaU 74 is probably the oldest fragment of Sahidic 1 Sam, dated to the 4th century.75 The leaf belongs to the collections of the Duke University Library with the reference number P.Duk.inv. 797.76 In Schüssler’s listing the manuscript has the siglum sa 77. The amount of text is remarkable, verses 14: 24– 67 L.T. Lefort, Les manuscrits coptes de l’université de Louvain, 49–52. Schüssler has not yet published a siglum for this manuscript. 68 Maspero, Fragments de la version thébaine, 155–6. 69 The reference number is B.N. Copte 12919. Drescher (Kingdoms, xii) mentions that he was unable to find the fragment in Paris. Schüssler (Die koptischen Bibeltexte 2: 2, 69–75) has given this ms siglum sa 212L. 70 A scala is a list of Coptic words and their Arabic counterparts. For more information, see the article of Werner Vycichl, “Sullam”. 71 The reference number is B.N. Copte 43. Schüssler has not yet published this scala in his lists, but he mentions it as a citation in his preliminary listing. The same also applies to the following manuscript, SaR. 72 The reference number is B.N. Copte 44. 73 Egypt Exploration Fund Ostraca, No 196. This is number 1 in W.E. Crum, Coptic Ostraca. In Schüssler’s listing this item has the siglum sa O 053. 74 This is the first fragment in my list that was not known to Drescher and, therefore, was not used in his edition. The same is true for the SaU–Z. The sigla continue from Drescher’s last witness, T, in alphabetical order. 75 F. Feder (“Koptische Übersetzung”, 69) lists the manuscripts that he dates to the 4th century, and this is one of them. 76 Formerly, P.Rob.inv. C. 1.

22

Introduction

46. I have collated this papyrus manuscript from a photo and include my own collation as an appendix. In a few passages, my deciphering differs from that of Aaron Michael Butts’s edition.77 SaV was published by Gerald Browne and Lucia Papini and dated to the 6th century by its editors.78 Two parchment leaves preserve verses 7: 8–8: 1 and 9: 21–10: 2. The manuscript belongs to the collection of the Istituto papirologico G. Vitelli, Florence.79 I have collated the fragment in a photo, and the edition is accurate. SaW refers to the parchment fragment P.Strasbourg copte 546 which contains verses 2: 35–36, 3: 6(?), 4: 3 and 4: 9. It belongs to the collections of BNU (Bibliothèque nationale et universitaire de Strasbourg) and was edited by Anne Boud’hors. The editor proposes a dating to the 5th–6th- centuries.80 SaX is the siglum for the fragment P. Strasbourg copte 319, containing verses 27: 8–9 and 28: 3b–5. The fragment belongs to the collections of BNU and was edited by Boud’hors.81 SaY belongs to the collections of Cambridge’s Corpus Christi College and it is not yet published.82 It contains text from 1 Sam 15: 12 and I have collated it from a photograph. My collation is found as an appendix. SaZ is kept in Manchester within the John Rylands Library. It offers text from verse 1 Sam 3: 6 and I have used a photograph since these fragments are not published. These tiny fragments have not featured in the current study because of their small size and broken nature. With some reconstructive work, perhaps they might yield a small amount of text. Bohairic fragments of some parts of 1 Sam are also known: 2: 1–10; 16: 1–13; 17: 17–54; 18: 6–9; 23: 26–28; 24: 1–23. I use the siglum BoA for these fragments published by Paul de Lagarde.83 An unpublished Psalms and Odes manuscript Barberiniani Orientali 2 contains four versions of the Song of Hannah (verses 2: 1–10) in four parallel columns, the Bohairic, Arabic,

77 Butts, “P.Duk.inv. 797”. 78 Browne, G.M./Papini, L., “Frammenti in copto dei Libri dei Re”. Drescher did not know about this fragment. 79 The reference number is PSI Inv.16 C. Schüssler (Die koptischen Bibeltexte 2: 2, 57) has given the siglum sa 208 to this manusccript. 80 A. Boud’hors, Catalogue des fragments coptes, 21–3. Schüssler has given the siglum sa 206 to this ms (Die koptischen Bibeltexte 2: 2, 54–5). This was not known to Drescher. 81 Boud’hors, Catalogue des fragments coptes, 24–5. In Schüssler’s listing, this ms hast he siglum sa 210 (Die koptischen Bibeltexte 2: 2, 59). Drescher did not know about this fragment. Boud’hors has not given a dating to this manuscript in her edition, but in personal correspondence she proposed a dating to the 6th century. 82 The reference number is CCC Ms 541 no. 38. 83 P. de Lagarde, “Bruchstücke der koptischen Übersetzung”, 63–72.

23

Manuscripts of Sahidic 1 Samuel

Ethiopic and Syriac.84 The siglum BoB refers to this fragment. One tiny fragment in Fayyumic survives, Oxford, Ms.Copt.e.162(P), containing text from 25: 31–34. I have taken into account those manuscripts that are direct witnesses to the Coptic biblical text. Therefore, citations in the Coptic literature are excluded. This is because of the complications that come with these indirect witnesses: before using a citation as a witness for the biblical text, one has to know the textual character of the citation. It can be a word-for-word citation from some biblical manuscript or quoted from memory, a reformulated citation, an allusion or anything between these ends.85 Below, the list outlines the preservation of 1 Sam in the Sahidic manuscripts discussed here. The complete manuscript SaM is present in every verse, and the list thus presents the additional manuscripts in those verses where SaM is not the sole witness. Ch.

Verses

Mss

Ch.

Verses

Mss

1 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 4 4 4 5 5

18–28 1 2 3–4 5–7 8 9 10 13 18–19 24–30 35–36 6 7 8–9 10–20 3 9 13 4 6–11

S CDS CS CDS CS CDS CDS CD QR QR E W EWZ E EG G W W QR QR S

6 6 7 7 7 8 8 8 8 8 8 8 8 9 9 9 9 10 10 12 12

2–13 14–21 1–7 8–14 15–17 1 2–3 4–5 6 7 8–9 10–16 17–22 1–11 12–19 20 21–27 1–2 3 4 5

A AE A AV ASV ASV AS A AS A AS A AH A AS A AV AV A I IN

84 In other parts of this manuscripts, there is also an Armenian column. For an introduction to this manuscript, see D.V. Proverbio, “BARB.OR. 2 (Psalterium Pentaglottum)”. A detailed analysis of mine of the Bohairic Hannah’s Song in this and other manuscripts will be published soon with the title “Bohairic Column in Barberiniani Orientali and other Bohairic manuscripts”. 85 The problems with Coptic citations are basically the same as in the case of patristic evidence in other languages, see Kauhanen, The Proto-Lucianic Problem in 1 Samuel, 26–8.

24

Introduction

Ch.

Verses

Mss

Ch.

Verses

Mss

12 12 12 13 13 13 14 14 14 14 14 14 14 14 14 14 15 15 15 15 15 15 15 15 16 16 16 16 16 16 17 17 17 17 17 17 17 17 17 18 18 19

6–8 10–11 19–24 2–4 6 8 3 5 7 10–11 14 17–23 24–32 33–46 49–50 51–52 1–2 12 13 15 17 19–20 21–26 29–30 2–3 4–5 6–8 11 12 13–18 5–7 18 31–32 33–44 45–48 49 50 51 52–54 6–27 28–29 1–5

N I S S S S E E E E QR A AU U K K K Y E E E E K K B BT B B BQR B QR QR A AB B BQR B BQR B B AB AB

19 20 20 20 21 21 21 22 22 22 22 22 23 24 25 25 25 26 26 26 26 26 27 27 28 28 28 28 28 29 29 29 29 30 30 30 30 30 30 31 31 31

6–24 1–18 19 20–43 1–12 13 14–15 (16) 1–4 5 6–7 8–13 21–23 1–14 21–23 1–16 17 18–28 4 7 8–19 20 21–25 8 9 2 3 4–5 6–15 16–25 1–2 3–4 5–9 10–11 1–2 2–11 12 13–20 21 25–31 1–2 3 4–13

A A AQR A A AQR A A AQR AS S A A A A AQR A QR BQR B BQR B QRX X QR BQRX BX B AB A AF AFJ A A AS AQR A AJ A AB ABQR AB

Sahidic 1 Samuel and its research history

25

1.4. Sahidic 1 Samuel and its research history Only a few studies have examined the nature of Sahidic 1 Sam and its Vorlage. In these studies, the questions have primarily addressed the Greek source text behind the Sahidic text. Likewise, these studies have considered the affiliation of the Vorlage of Sahidic 1 Sam to the extant Greek manuscripts. Dieu published the article “Le texte copte sahidique des livres de Samuel” in 1946.86 For this article, he had compared Rahlfs’ Greek text and the text of the Sahidic manuscript SaM in verses 17: 11–19: 1, a passage in which the Septuagint is considerably shorter than the MT. He concluded that the Vorlage of the Sahidic translator lacked Hexaplaric and Lucianic additions. This is, according to Dieu, the best way to explain the omission of verses 17: 55–18: 6 and the absence of many Lucianic variants. Later on, however, a scribe probably compared the Coptic text with the Greek text of his time and added some details here and there.87 Payne wrote his dissertation Critical and comparative study of the Sahidic Coptic texts of the First book of Samuel at Princeton in 1949. His study remains unpublished aside from a short derivative article from 1953.88 In his work, Payne begins with manuscripts SaM and SaA, and finds almost 1500 variations between these manuscripts. After excluding the variations that he considered unreliable for textual criticism, 670 variations, in his opinion, could be attributed to the Greek texts. These variations he classified as follows: A) 383 instances where SaM is closer to the unanimous testimony of the Greek evidence,89 B) 150 instances where SaA is closer to this Greek evidence,90 C) 137 instances where one is closer to one group within the Greek and the other is closer to another Greek group.91 As objective as the numbers might appear, Payne’s conclusions are not convincing because his argumentation lacks a sound methodology. Mostly, his work consists of lists in which he provides different kinds of examples without evaluating their importance.92 The most problematic issue in Payne’s work is the simplistic way in which he uses the text-critical principle lectio difficilior.93 This principle

86 Previously, Dieu had already written a more general article “Les Manuscrits Grecs des Livres de Samuel” on the Sahidic 1 Sam. 87 L. Dieu, “Le texte copte sahidique”, 452. 88 J.B Payne, “The Sahidic Coptic Text of I Samuel”. Thanks to Prof. Sebastian Brock’s personal archive, I have been able to use a microfilm copy of Payne’s dissertation. 89 Payne, Critical and comparative, 82–90. 90 Payne, Critical and comparative, 90–6. 91 Payne, Critical and comparative, 265–6. 92 The assessment of Payne above is not unique, as the same problem also occurs elsewhere. See the critical comment of Kauhanen concerning a comparable Princeton thesis on the Latin version of 1 Sam, Kauhanen, The Proto-Lucianic Problem, 131–2. 93 According to this rule, in the case of variants the priority should be given to the more difficult reading since texts are prone to simplification in the process of transmission.

26

Introduction

constructively explains typical errors that happen within textual transmission. Problematically, this kind of rule does not allow for other explanations, such as errors changing the text in ways other than what the rule assumes.94 Payne also, in a simplistic way, applies de Lagarde’s principle, designed for the textual criticism of the Septuagint in its relation to the Hebrew text, to the Coptic text. According to this principle, in the LXX the more original reading is the one that is further removed from Hebrew.95 It is not reasonable to use these rules every time one encounters a difference between the Sahidic manuscripts. The result is at best haphazard since not even clear mistakes are spared from having the lectio difficilior principle applied in Payne’s work. After comparing the variants, Payne draws the conclusion that SaM was corrected and amended to more strictly conform to a later Greek source, while the readings of SaA dated to an earlier time when texts were translated more freely.96 Payne mentions Hexaplaric and Lucianic as the most influential and thorough-going recensions with respect to Greek 1 Sam.97 In addition to these, he supposes pre-recensional revisions and associates these revisions especially with the “longer asterisked passages represented in Sahidic I Samuel, points of deficiency in the OG which must early have been restored in certain texts”.98 The approximations resulting from these pre-recensional activities were all incorporated into later Hexaplaric and Lucianic recensions. Thus, he considers the Sahidic version as representing an intermediate Greek source text, now lost, where some approximations already existed. Basically, Frank Feder expresses the same view concerning Sahidic Jeremiah. According to him, the Greek Vorlage of Sahidic Jeremiah was a preHexaplaric revision of the LXX, which itself was a precursor of Origen’s and Lucian’s recensions.99 It is problematic to suppose that Sahidic texts represent an intermediate source text, a “missing link” from the Old Greek via the Sahidic translator’s Vorlage to Hexaplaric and Lucianic recensions.100 More probably, the text resulted from a complex transmissional 94 See Aejmelaeus, “Corruption or Correction”, 3: “There, however, are so many exceptions to these rules—errors producing difficult wordings and parablepsis shortening the text— that they really are not of much use”. 95 Before using this rule on the Coptic text, one has to be sure that there were attempts to change an earlier translation to more accurately agree with the Greek text of that time. 96 Payne, Critical and comparative, 98–9. 97 There is a brief mention of a Hesychian recension but it is not taken up in his chapter “Sahidic Textual History”. 98 Payne, Critical and comparative, 392. 99 Feder, Biblia Sahidica, 77. And more: “So erscheinen O und L als Rezipienten der früheren Rezensionen bei Änderungen nach dem MT neben Sa und den Minuskeln”, Feder, Biblia Sahidica, 75. 100 One should avoid unnecessary theories whenever possible. In the present discussion, it seems probable that the Sahidic text has been sporadically supplemented with Hebrewbased corrections. However, there is neither evidence nor need to assume an intermediate text.

Sahidic 1 Samuel and its research history

27

history with limited, occasional changes after the initial translation. The view of Payne is also inaccurate on the basis of what we know about the history of the Greek text and the recensions in 1 Sam. According to Payne, the Vorlage of the earliest Sahidic was basically of the Old Greek type, containing some Lucianic and Hexaplaric readings, though neither Lucian nor Origen had yet produced their texts. In the definition of the OG, Payne follows John W. Wevers: B (Vaticanus), 509, 121 and the Ethiopic version.101 Payne states that the revision of the Old Coptic was derived from a mixed Greek text.102 Thus, in his opinion, SaM is a revision of the Old Coptic text behind SaA. Payne supposes “a constant process of pre-recensional revision”. Similarly, Feder (on Jeremiah) supposes some early recension(s) before the famous ones, and thinks that these were adopted and further revised according to the MT into the Hexaplaric and Lucianic recensions.103 Perhaps, these two scholars could have more constructively chosen their terminology. Obviously, both use the term recension when referring to changes that more precisely are labelled approximations or sporadic corrections towards the MT.104 In these cases, there are sporadic changes according to a variant Greek text closely conformed to the MT. After Payne’s study, Drescher, in his preface to the edition of the manuscript SaM, suggests that SaM and SaA represent basically one and the same translation from Greek. He considers the free renderings of SaA to be those of a revised version that suffered under poor copyists.105 With regard to the question of the Old Coptic and its revision, Drescher’s position contradicts that of Payne. Whereas Payne argues that SaA more faithfully represents the Old Coptic, Drescher considers SaM as a more accurate representative of the Coptic translation. According to Drescher, the priority of SaM is evident in the conformity of the earliest fragments to SaM. Drescher states that SaA and SaM have both undergone some independent revision, but he does not estimate the extent of these revisions.106 Clearly, the complexities of Sahidic 1 Sam have not yet been fully resolved and deserve further attention. A detailed and systematic analysis of the translation technique is necessary for further inquiry into the Vorlage of the Coptic text. It is not enough to compare different Coptic texts with each other and to count their differences. The description and investigation of a translation should begin with translation technique. This is the only way to distinguish the translator’s idiosyncracies from the features deriving from the Greek Vorlage. This methodical principle was developed by Ilmari Soi101 Payne, Critical and comparative, 139. 102 Payne, Critical and comparative, 390–4. 103 Feder, Biblia Sahidica, 74–5. 104 See, for example, Aejmelaeus, “Corruption or Correction”, 13; and Brock, Recensions, 21*, 303–4. 105 Drescher, Kingdoms, xviii. 106 Drescher, Kingdoms, xxv.

28

Introduction

salon-Soininen for the study of Septuagintal syntax, and several studies have used the principle.107 Moreover, since Payne and Drescher wrote their studies on Sahidic 1 Sam, scholars have identified several early Sahidic fragments, and such discoveries allow for further progress.

1.5. Aims and methods of this study The aim of this study is twofold. First, a chapter will examine the translation technique of the Sahidic translator. Second, a further chapter will apply this basic knowledge of the translation technique in the area of textual criticism in order to discover the affiliations of the Sahidic version (or its Greek Vorlage) with Greek textual traditions as well as the affiliations among the Sahidic manuscripts. This study will support the creation of a new critical edition of the Septuagint of 1 Sam for the Göttingen series,108 undertaken by Aejmelaeus. With respect to this edition, the primary goal is to identify the affiliations of the Sahidic version. This translation-technical study, however, will additionally allow for a more careful and accurate citation of the Sahidic within the critical apparatus. Naturally, secondary features of the Coptic should not be cited in support of Greek readings. Secondary features include those corruptions that emerged during transmission as well as variations related to the translator’s style and language (like rendering Greek καί into Coptic with ⲇⲉ). For the purposes of the forthcoming Göttingen edition, collations have been produced of all extant Greek manuscripts and daughter versions of 1 Sam.109 These collations facilitated comparison with all the existing Greek and versional variants. In the translation-technical section, clause connections and translator’s additions feature prominently. These foci were motivated by the tendency of the citations within the Brooke/McLean edition. If the Sahidic appears in their apparatus, its citation mostly concerns conjunctions or additions found in Sahidic.110 107 For more on this principle in Hebrew-Greek translations, see Soisalon-Soininen, Infinitive, 12–16. The studies that follow, more or less, this method are (in chronological order) Sollamo, Renderings of Hebrew Semiprepositions; A. Aejmelaeus, Parataxis; Seppo Sipilä, Between Literalness and Freedom; and Anssi Voitila, Présent et imparfait de l´indicatif dans le Pentateuque grec. 108 Septuaginta. Vetus Testamentum Graecum Auctoritate Academiae Scientiarum Gottingensis editum. 109 These collations were prepared by the Göttingen Septuaginta-Unternehmen. The Ethiopic daughter version, Old Latin manuscripts and marginal readings were collated by Aejmelaeus, the Syriac by Marketta Liljeström and the Georgian version by Prof. Anna Kharanauli. 110 Brooke/McLean, The Old Testament in Greek.

Aims and methods of this study

29

The first part of the translation-technical analysis (chapter 2.1) examines clause connections. The clause connections appear throughout the text and offer, therefore, a fruitful starting point for a translation-technical study. When working with this material, comparative analyses rely on clausal sectioning as a basis for comparable semantic units. Conjunctions, particles and asyndeton (lack of a connective) that have been employed to connect clauses, have an impact on how the text is understood as a whole, how the story line proceeds and how successive clauses relate to each other.111 Aejmelaeus and Seppo Sipilä studied this feature (clause connections) using the method of Soisalon-Soininen in their dissertations on the translation technique of the Septuagint.112 The comparative method used here resembles that used by Aejmelaeus and Sipilä, despite the fact that the languages differ. Their works evaluate Hebrew-Greek translations, whereas the present study has a Greek-Coptic translation as its object. One essential difference between a Hebrew-Greek and a Greek-Coptic translation-technical study (in the area of the LXX) is the fact that Greek has a long history as a written language before the translation of the LXX, whereas Coptic as a written language emerged more or less simultaneously with the translation of the Scriptures. Thus, there are no contemporary texts originally composed in Coptic to which the translations could be compared.113 The only texts available for comparison are either translations from Greek or texts composed significantly later than Sahidic 1 Sam. The second topic in the translation-technical part (chapter 2.2) is translator’s additions. Kirsten Malmkjær describes the use of corpora in translation studies, stating that “before a translation corpus can be searched, it has to be matched up clause for clause or sentence for sentence, insofar as this is possible”.114 Precisely in this manner, the present study of 1 Sam compares the Greek text and the Sahidic translation clause for clause. As a result, there are pairs comprised of source text and its rendering in the target text as well as some surplus in both texts. The surplus in the source text is labeled omission, whereas the surplus in the target text is seen as addition. Cases of the first-mentioned surplus (omissions) are analyzed in the chapter on clause connections, and cases of the last-mentioned (additions) constitute the corpus in the chapter on translator’s additions.

111 Aejmelaeus (On the Trail, 45–7) discusses these questions concerning a Hebrew-Greek translation-technical study. 112 Initially, the principles are described in Soisalon-Soininen, Infinitive, 12–16. These were expanded by Aejmelaeus, Parataxis, and Sipilä, Between Literalness and Freedom. 113 For an interesting comparison on the level of syntactic patterns, see Andrea Hasznos, Graeco-Coptica. She has studied final clauses, consecutive clauses, object clauses/infinitive constructions after verbs of exhorting, and subject clauses/infinitive constructions, and their structures in translated versus original Coptic literature in a restricted corpus (Matthew, John, Romans representing translations). 114 K. Malmkjær, Linguistics and the Language of Translation, 59.

30

Introduction

In chapter three, where affiliations are concerned, the following considerations have influenced the choice of analyzed passages. The passages in the first section (7: 8–8: 1; 9: 21–10: 2) are preserved by SaV, a sizable fragment with an early dating. Likewise, one of the earliest fragments, SaI, preserves the passages (12: 4–5, 10–11) discussed in the second section. The third section contains the passage (14: 24–32) from SaU which might be the earliest fragment of Sahidic 1 Sam. The fourth section discusses a passage (17: 33–44) that contains many important secondary readings. In chapter 17, the Greek and Hebrew texts repeatedly differ from one another, and the LXX tradition is likewise unstable. The fifth and sixth sections concern passages 29: 5–9 and 30: 21–24, found in SaMA as well as in a third (and sometimes fourth) manuscript. These passages were chosen because the relationship between SaM and SaA is so important. The passage discussed in the last section (1 Sam 31) was chosen because the entire chapter is preserved in the three most extensive manuscripts of Sahidic 1 Sam: SaA, SaB and SaM. One further criteria was to include passages from different parts of the book. Thus, even though the entire book is not covered, the various passages enable conclusions that are trustworthy in the whole book.115 I have analyzed the Sahidic and Greek texts in these seven sections in a detailed manner. These analyses describe the textual character of each Sahidic manuscript, and search for the existence of secondary readings and/or corruptions. In the cases where variants occur within the Sahidic tradition, the present study seeks to establish whether these readings relate to Greek variants or solely derive from Sahidic transmission. During the centuries that followed the original translation event, Greek readings theoretically further influenced the Sahidic text. Greek and Coptic texts might have been compared not only at the moment of the translation but also afterwards.116 Probably, scribes knew that the Coptic had been translated from a Greek source, and, therefore, one could revise the Coptic text according to the Greek. Since only one complete manuscript of the Sahidic 1 Sam has survived, the nature of later Greek influence is extremely difficult to ascertain. However, where fate has preserved parallel Sahidic witnesses, it is all the more important to make a careful analysis of these passages where we do have more than one manuscript, and the earliest fragments, of course, play a crucial role in this analysis. In order to facilitate discussion and analysis, a few terms require definition. Gideon Toury introduced the method known as Descriptive Translation Studies (DTS) in his monograph Descriptive Translation Studies – and beyond.117 Toury discusses several important elements that need attention when describing a translation. The terms ‘equivalence’ and ‘textual segment’ 115 Naturally, single small fragments not analysed here need to be studied separately before supposing anything about their textual character or using them in the textual criticism. 116 See also Brock, Recensions, 23. 117 G. Toury, Descriptive Translation Studies.

Aims and methods of this study

31

are both significant. Toury defines equivalence as a “functional-relational concept”, and continues: A descriptive study would always proceed from the assumption that equivalence does exist between an assumed translation and its assumed source. What remains to be uncovered is only the way this postulate was actually realized, e. g., in terms of the balance between what was kept invariant and what was transformed.118

Toury uses the terms ‘replaced segment’ and ‘replacing segment’ when discussing the units of comparative analysis.119 In my work, I use ‘equivalent’ as a synonym for ‘rendering’, assuming that the equivalent in the target text (the ‘replacing segment’) parallels the equivalent in the source text (‘replaced segment’). Additionally, Toury uses the term ‘norm’. A norm is a translational possibility that occupies a place on the spectrum between the extremes of ‘rule’ and ‘idiosyncrasy’. Norms can be placed on a continuum whose ends are strong and weak norms respectively. An initial norm, in Toury’s system, describes the translator’s initial decision in prioritizing source and target language norms. He explains: “Thus, whereas adherence to source norms determines a translation’s adequacy as compared to the source text, subscription to norms originating in the target culture determines its acceptability.”120 In the case of the Sahidic Bible translation, there are no extra-textual witnesses where the norms could be found. The norms must be evaluated within the translation.121

118 Toury, Descriptive Translation Studies, 86. 119 Ibid., 87–9. 120 Ibid., 54–7. The same theme is discussed by Soisalon-Soininen (Infinitive, 13–14), although with a different terminology. 121 According to Toury (Descriptive Translation Studies, 65–6), texts are primary witnesses, whereas any formulation of a norm is subject to different interests and modifications.

2. Description of the translation technique This chapter aims at a description of the translation technique of the Sahidic translator. Two issues serve as its material basis: clause connections and translational additions. Translation technique as a term is used to refer to a description of the translator’s activity, his way of rendering certain Greek expressions in Sahidic. Naturally, no consciously formulated and detailed “technique” in its present-day meaning is supposed to result from this study.

2.1. Clause connections In this chapter, I will present my analyses of clause connections in 1 Sam. This analysis will consider a variety of distinct semantic constructions including main clauses, subordinate clauses, participial constructions and infinitival constructions.

2.1.1. Questions and methods The corpus was collected on the basis of the hypothetical Greek Vorlage, the source language.1 Each clause was identified and written on a card, together with its Sahidic rendering. There are two subgroups in this chapter; the first comprised finite clauses and the second comprised non-finite clauses: participial and infinitival constructions. These two latter groups are included because they have been used to render Hebrew coordinated clauses in the LXX.2 The primary group of clauses reflects the Greek connectives used.3 Inside 1 2

3

The edition of Rahlfs was used since it is found in electronic form. See Aejmelaeus, Parataxis, 116. Her corpus contains those cases where a coordinated Hebrew clause is rendered with these constructions. I do not have Hebrew as my starting point, but the Greek text of 1 Sam. Thus, there are perhaps some more cases than there would have been had I collected only those cases where Hebrew shows coordination. I use the term connective as a hypernym for the terms conjunction and particle, referring to the word that is used to connect the clauses or somehow gives information concerning the relation between the clauses. Basically, I have included those Greek particles that have been used in a clause connective function in the Septuagint. Particles ἄρα, ἀτάρ, αὐτάρ, γε, δῆθεν, δήπου, δῆτα, μήν, μέντοι, οὔκουν, γοῦν, τε, τοι, τοιγαροῦν, τοίνυν do not occur in the Greek text of 1 Sam. Two particles that occur in 1 Sam, δή and μέν, are not included even though they sometimes have a connective function (Especially δή, see John

Clause connections

33

these groups, the material was further subgrouped according to Sahidic renderings. Participial constructions are discussed as their own subgroups, participium coniunctum and genetivus absolutus, regardless of the Greek connectives. In the following pages, I will describe how the translator of 1 Sam has rendered clauses beginning with various Greek connectives. Numerous examples illustrate my analyses and conclusions. When dealing with the examples, one should consider Malmkjær’s warnings concerning the limitations of corpus-based studies: “Corpora provide information about frequency, not about what is possible in language.”4 Accordingly, my aim is not to argue that certain Greek expressions have certain possibilities with which they can be rendered in Sahidic Coptic, but to describe how this particular Sahidic translator has rendered certain expressions. This approach is also needed according to Matthias Müller, whose forthcoming article systematically studies the different connectives used in Coptic: “Further research is needed specifically in regard of a quantitative approach to the various patterns as well as preference in different textual sorts or of the specific style of certain authors.”5 However, the present study has a different perspective on the issue because it is primarily concerned with Sahidic 1 Sam as a translation and as a witness to its Greek Vorlage. Text-critical notes follow numerous examples in this section. The most important of these appear together in chapter three (3.3.1), where the Vorlage of the Sahidic translation receives more attention. I refer to the Greek manuscripts and their groups according to the grouping of Aejmelaeus.6

4 5 6

Dewar Denniston, The Greek Particles, 238–40). In the Greek 1 Sam, the typical usage of δή is to render Hebrew emphatic particle ‫‘ ־נא‬surely’. In these instances, it does not function as a connective. In 1 Sam, there are 36 occurrences of δή. In 33 instances the Sahidic translator has not translated it separately, but it seems to be included in the verbal form (imperative or jussive). In two passages (16: 17, 23: 22) Sahidic renders it with ϭⲉ. In verse 25: 8 SaA renders it with ⲇⲉ, perhaps following Greek mss 379–530 that read δέ instead of δή. Malmkjær, Linguistics and the Language of Translation, 131. M. Müller, “Greek Connectors”, 1. I am grateful to Dr. Müller for letting me use his article before its publication. See the appendix 6.5.

34

Description of the translation technique

2.1.2. Occurrences of clauses in Greek καί-clauses Asyndetic clauses Interrogative clauses Other coordinate clauses Relative clauses Conditional clauses ὅτι, διότι, καθότι ὡς

Other subordinate clauses part.coni. gen.abs. inf. constructions

2144 467 115 55 125 72 224 33 37 146 15 30

The chart above summarizes the groups of clause connections present in 1 Sam. Not every connective appears in its own box but only those that occur more than 20 times. Those occurring less than 20 times have been calculated together and given in the numbers “Other coordinate clauses” and “Other subordinate clauses” respectively.

2.1.3. Coordinated clauses and their renderings The traditional way to present conjunctions and their use is to make a distinction between coordinating and subordinating conjunctions. Thus, I present my analyses of different connectives and their renderings according to this division. Besides this basic structure, I separate non-finite clauses into their own group. 2.1.3.1.καί-clauses Most of the clauses in 1 Sam (2144 instances) begin with the conjunction καί. This is a Greek coordinating conjunction. It is common in Greek texts,

connecting words, phrases, clauses and sentences. In the Septuagint, this conjunction is a standard equivalent for ‫ ו‬in Hebrew, and is, therefore, very common. According to Aejmelaeus, καί is used in ordinary coordination, and also in cases where the relation between clauses is adversative, consecutive, conditional or temporal.7 In Sahidic 1 Sam, this conjunction is mostly rendered with ⲁⲩⲱ, ⲇⲉ or asyndeton, only 70 out of 2144 passages are ren7

Aejmelaeus, Parataxis, 16–17. Blass/Debrunner/Rehkopf also list different nuances of this conjunction (Grammatik, §442).

Clause connections

35

dered with a different equivalent. In 19 passages, there is no equivalent for the καί-clause. Payne provided in his list “the most common translations” of καί, but his presentation is problematic since he does not give weight to the frequency of each translation. He lists ⲁⲩⲱ, ⲇⲉ and ⲙⲛ̄ as the most common, but without mentioning asyndeton in this connection, only later recognizing asyndeton as a special case that requires explanation.8 Likewise, Feder describes the renderings of this Greek conjunction in Corpus Ieremiae.9 According to him, this is one of the most productive sources when it comes to ‘false’ variants—i. e., those that emerged during the translation and not from a Greek Vorlage.10 Therefore, in order to accurately employ Coptic in Greek textual criticism, one requires a thorough analysis of the renderings of the clauses that begin with the conjunction καί.11 2.1.3.1.a) Renderings of καί-clauses asyndeton ⲁⲩⲱ ⲇⲉ ⲧⲉⲛⲟⲩ ϭⲉ ϭⲉ ⲟⲛ ⲁⲗⲗⲁ ϫⲉ ⲙⲛ ⲟⲩⲇⲉ ⲏ ⲙⲛⲛⲥⲱⲥ ϫⲉⲕⲁⲥ

8

920 times 707 times 401 times 36 times, of which 35 cases read καὶ νῦν in Greek 11 times 7 times 6 times 4 times 2 times 2 times once once once

Payne, Critical and comparative, 48–52. One must bear in mind that Payne lists all the conjunctions, whereas I have listed only those used to connect clauses. 9 Feder (Biblia Sahidica, 86–7) lists ⲁⲩⲱ, ⲙⲛ̄, ϩⲓ, ⲁⲗⲗⲁ, ⲇⲉ, ⲏ, ⲟⲩⲇⲉ and no element as the equivalents used for καί in Jeremiah, Lamentations, Epistle of Jeremiah and Baruch, depending on the context of each occurrence. 10 “… eine der ergiebigsten Quellen für übersetzungsbedingte Lesarten” (Feder, Biblia Sahidica, 86). 11 Christian Askeland has studied John’s gospel and its Coptic translations. He has written a subsection “The Sahidic and Classical Bohairic compared with the NA27” in C. Askeland, John’s Gospel, 12–22. In this section, he provides tables that illustrate equivalences of καί and δέ with their typical renderings. His tables convincingly show that the numbers do not match. Thus, one can conclude that Greek καί and δέ had several equivalents in the Sahidic and Classical Bohairic translations.

36

Description of the translation technique

2.1.3.1.b) καί-clause rendered asyndetically Asyndeton, lack of connective, is in Coptic a natural way to form sentences. According to Bentley Layton, linkage expressed by asyndeton is closer than that expressed by the conjunction ⲁⲩⲱ.12 The most prevalent rendering for καί-clauses is the group of asyndetic clauses: 920 cases. In some of these almost one thousand cases, the Greek Vorlage used by the translator possibly also read without the conjunction. In any case, the number of asyndetic readings is significant enough to argue compellingly that the translator often rendered καί-clauses with asyndetic clauses. The fact that καί-clauses have been rendered asyndetically has bearing on their citation in textual criticism; Coptic should not be used as an argument supporting the omission in the cases where the textual critic has to make a choice between καί and omission of it in the Greek text. However, this is typically done in Brooke-McLean,13 but it is misleading and in any case it does not prove anything about the affiliations of the Sahidic text. καί-clauses rendered asyndetically

Apodotic usage in Greek

the past tense ⲁⲡⲉϫⲉ /ⲡⲉϫⲁ⸗ conjunctive circumstantial future optative (future III) preterit nominal clause aorist ϣⲁⲣⲉinfinitive, complementary group others

14 cases of apodotic καί

385 190 180 104 17 13 7 514 4 4 1115

one apodotic καί (21: 5) 13 cases of apodotic καί 5 cases of apodotic καί 4 cases of apodotic καί one apodotic καί

This table lists the verbal forms that are used with asyndeton to render a καί-clause in Greek. In most cases, the simple past tense is employed, which is natural in a narrative context. In the following instances, I provide examples of the usual cases as well as of the cases where the translator has used equivalents differing from his own norm. The examples have been ordered according to the table above.

12 For instance, asyndetically conjoined verbs typically share the same subject. B. Layton, Grammar, §237. 13 E.g., Brooke/McLean in verses 3: 9, 6: 9, 10: 2, 10: 11, 12: 15, 14: 19, 14: 23, 20: 19. 14 25: 25, 17: 13, 14: 49, 4: 16, 28: 12 15 1: 15, 6: 8; 9: 10, 9: 12, 14: 28, 20: 11, 20: 19, 22: 8, 23: 2, 23: 27, 30: 1(apodotic).

Clause connections

37

13: 12 Νῦν καταβήσονται οἱ ἀλλόφυλοι πρός με εἰς Γάλγαλα, καὶ τοῦ προσώπου τοῦ κυρίου οὐκ ἐδεήθην· καὶ ἐνεκρατευσάμην καὶ ἀνήνεγκα τὴν ὁλοκαύτωσιν.

ⲧⲉⲛⲟⲩ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲁⲉⲓ ⲛⲁⲓ ⲉϩⲣⲁⲓ ⲉⲅⲁⲗⲅⲁⲗⲁ· ⲉⲙⲡⲓⲧⲉⲃϩ ̄ ⲡϩⲟ ⲙⲡϫⲟⲉⲓⲥ· ⲁⲓⲉⲅⲕⲣⲁⲧⲉⲩⲉ ⲙⲙⲟⲓ ⲁⲓⲧⲁⲗⲟ ⲛⲛⲉϭⲗⲓⲗ ⲉϩⲣⲁⲓ This is a typical example of asyndeton in a Sahidic text: no conjunctions, only asyndetic clauses with past tense verbal forms to render καί-clauses.16 14: 42–43

καὶ κατακληροῦται Ιωναθαν. καὶ εἶπεν Σαοὺλ πρὸς Ἰωναθάν Ἀπάγγειλόν μοι τί πεποίηκας·

ⲁ ⲡⲉⲕⲗⲏⲣⲟⲥ ⲉⲓ ⲉϫ ̅ ⲛ̄ ⲓⲱⲛⲁⲑⲁⲛ· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲓⲱⲛⲁⲑⲁⲛ·̄ ϫⲉ ⲙⲁⲧⲁⲙⲟⲓ· ϫⲉ ⲟⲩ ⲡⲉ ⲛⲧⲁⲕⲁⲁϥ There are 190 cases like this where asyndeton with an appropriate form of the verb ⲡⲉϫⲉ- is used to render καί and a verb of speaking. Asyndeton used with a conjunctive in Coptic is a natural equivalent for coordinating καί, and it is often used in 1 Sam (180 times).17 Feder maintains that, in the cases where a future or an imperative verbal form is followed by a conjunctive, classical Sahidic sources regularly eschew the conjunction ⲁⲩⲱ.18 However, Layton seems to support this only partially: “Extending a verb, ⲛⲧ ̄ ⲉ is optionally preceded by a conjunction such as ⲁⲗⲗⲁ (but), ⲁⲩⲱ (and) …, but mostly occurs without such a linking term.”19 29: 7

καὶ νῦν ἀνάστρεφε καὶ πορεύου εἰς εἰρήνην, καὶ οὐ μὴ ποιήσεις κακίαν ἐν ὀφθαλμοῖς τῶν σατραπῶν τῶν ἀλλοφύλων.

ⲧⲉⲛⲟⲩ ϭⲉ ⲕⲧⲟⲕ ⲛⲅⲃⲱⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣⲏⲛⲏ· ⲛⲅⲧⲙ̄ⲉⲓⲣⲉ ⲛⲟⲩⲕⲁⲕⲓⲁ ⲛⲁϩⲣⲁⲩ· ⲛⲛⲉⲥⲁⲧⲣⲁⲡⲏⲥ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ In this example, the translator has used conjunctive forms twice. The first imperative ἀνάστρεφε is rendered with the imperative, and the next two verbs, imperative and future in Greek, read with the Sahidic conjunctive. Conjunctive chains occur often; in 54 cases there is also a conjunctive in the preceding clause.20 28: 16 καὶ εἶπεν Σαμουήλ Ἵνα τί ἐπερωτᾷς με, καὶ Κύριος ἀφέστηκεν ἀπὸ σοῦ καὶ γέγονεν μετὰ τοῦ πλησίον σου; SaM ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲁϥ ϫⲉ ⲁϩⲣⲟⲕ ⲕϫ ̄ ⲛⲟⲩ ⲙⲙⲟⲓ· ⲉⲁ ⲡϫⲟⲉⲓⲥ ⲥⲁϩⲱⲱϥ ⲉⲃⲟⲗ ⲙⲙⲟⲕ ⲁϥϣⲱⲡⲉ ⲙⲛ̄ ⲡⲉⲑⲓⲧⲟⲩⲱⲕ

16 For examples from Coptic John, see Askeland, John’s Gospel, 25. 17 Examples from Coptic John are listed in Askeland, John’s Gospel, 25. 18 Feder, Biblia Sahidica, 87. According to Feder, the use of ⲁⲩⲱ in these cases is to be seen as a signal of a striving to imitate the Greek source text. 19 Layton, Grammar, §352b. 20 For examples from Coptic John, see Askeland, John’s Gospel, 24–5.

38

Description of the translation technique

SaA

ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲁϥ ϫⲉ ⲁϩⲣⲟⲕ ⲉⲕϫⲛⲟⲩ ⲙ̄ⲙⲟⲓ. ⲁⲡϫⲟⲉⲓⲥ ⲥⲁϩⲱⲱϥ ⲉⲃⲟⲗ ⲙ̄ⲙⲟⲕ ⲁϥϣⲱⲡⲉ ⲙⲛ̄ ⲡⲉⲧϩⲓⲧⲟⲩⲱⲕ. SaB ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲥⲁⲟⲩⲗ· ϫⲉ ⲁϩⲉ ⲁ ⲡϫⲟⲉⲓⲥ ⲥⲁϩⲱⲱϥ ⲉⲃⲟⲗ ⲙⲙⲟⲕ· This example shows circumstantial and past tense as renderings of καίclauses. The second καί-clause is rendered beginning with the prefix ⲉⲁ-, the circumstantial of the past tense. This clause continues the question ‘Why do you ask me, with the Lord having withdrawn himself from you?’ SaA reads without the converter ⲉ-, resulting in two past tenses one after another. This fits the overall tendency in SaA to shorten and simplify the story line as well as attesting to a significant number of scribal errors.21 SaB is totally liberated here. It abbreviates and retells the story with freedom so that one hardly recognizes the passage: ‘Samuel said to Saul “Yes, the Lord has withdrawn himself from you”.’ 14: 13 καὶ ἀνέβη Ἰωναθὰν ἐπὶ τὰς χεῖρας αὐτοῦ καὶ ἐπὶ τοὺς πόδας αὐτοῦ, καὶ ὁ αἴρων τὰ σκεύη αὐτοῦ μετ᾽αὐτοῦ· ⲁⲩⲱ ⲓⲱⲛⲁⲑⲁⲛ ⲁϥⲃⲱⲕ ⲛⲁⲩ ⲉϩⲣⲁⲓ ⲉⲣⲉ ⲧⲟⲟⲧϥ̄· ⲙⲛ ⲣⲁⲧϥ̄ ⲧⲁϫⲣⲏⲩ· ⲉⲣⲉ ⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲏⲩⲉ ̄ ⲟⲩⲏϩ ⲛⲥⲱϥ Here a nominal clause beginning with καί is translated with a circumstantial into Sahidic. In this context the circumstantial is best understood as an attributive circumstantial. Text-critical note: With respect to the Greek Vorlage, Sahidic probably supports the reading ὀπίσω αὐτοῦ (L CII s-489 =MT) instead of μετ᾽αὐτοῦ (all other mss, omitted in M). 10: 3–4

…καὶ ἕνα αἴροντα ἀσκὸν οἴνου· καὶ ἐρωτήσουσίν σε τὰ εἰς εἰρήνην,…

ⲁⲩⲱ ⲡⲕⲉⲟⲩⲁ ⲉⲣⲉ ⲟⲩϩⲱⲧ ⲛⲏⲣⲡ̄ ϩⲓϫⲱϥ ⲥⲉⲛⲁϣⲁϫⲉ ⲛⲙ ̄ ⲙⲁⲕ ⲉⲧⲃⲉ ϯⲣⲏⲛⲏ· In this case the καί-clause has asyndeton and a future verbal form as its rendering. In most of the clauses where a Greek καί-clause is rendered with the future verbal form in Sahidic, the Greek text reads an apodotic καί. There are altogether 17 clauses, of which 13 contain an apodotic καί. 20: 5

καὶ ἐγὼ καθίσας οὐ καθήσομαι μετὰ τοῦ βασιλέως φαγεῖν, καὶ ἐξαποστελεῖς με…

ⲁⲩⲱ ⲁⲛⲟⲕ ⲛϯⲛⲁⲧⲱϣ ⲁⲛ ⲉϩⲙⲟⲟⲥ ⲉϩⲣⲁⲓ ⲉⲟⲩⲱⲙ ⲙⲛ ⲡⲣⲣ̄ ⲟ̄ · ⲉⲕⲉϫⲟⲟⲩⲧ ⲉⲃⲟⲗ In this example, the rendering of the καί-clause is asyndeton with an optative verbal form in Sahidic. 2: 11

καὶ ἀπῆλθον εἰς Ἁρμαθάιμ· καὶ τὸ παιδάριον ἦν λειτουργῶν τῷ προσώπῳ Κυρίου…

ⲁⲩⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲁⲣⲙⲁⲑⲁⲉⲓⲙ· ⲡϣⲏⲣⲉ ϣⲏⲙ ⲛⲉϥϣⲙ̄ϣⲉ ϩⲓⲑⲏ ⲙⲡϫⲟⲉⲓⲥ 21 See my section 3.2.7.

Clause connections

39

Above, the rendering of the καί-clause consists of asyndeton with the preterit verbal form. 28: 12 Ἵνα τί παρελογίσω με; καὶ σὺ εἶ Σαούλ. ⲉⲧⲃⲉ ⲟⲩ ⲁⲕϫⲓ ⲕⲣⲟϥ ⲙⲙⲟⲓ̈ ⲛⲧⲟⲕ ⲡⲉ ⲥⲁⲟⲩⲗ· Here, the rendering of the καί-clause results in an asyndetic nominal clause in Sahidic. 3: 6

καὶ προσέθετο (+ἔτι M V CI 328 29 55 71 158 707) Κύριος (+ἔτι 121 f 68–122) καὶ ἐκάλεσεν (+ἔτι A O d-68 122 554) Σαμουὴλ Σαμουήλ· καί 2°] pr. καλέσαι τὸν σαμουήλ M V L C’ a s 29 55 71 158 244 245

318 707 ⲡϫⲟⲉⲓⲥ ⲇⲉ ⲁϥⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ̄ ⲛⲕⲉⲥⲟⲡ ⲉⲙⲟⲩⲧⲉ· ϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲥⲁⲙⲟⲩⲏⲗ In this example, the καί-clause is rendered with an infinitive and with the preposition ⲉ. Text-critical note: The Greek witnesses are particularly divergent in this verse. Lucianic manuscripts read a Hexaplaric approximation καλέσαι τὸν σαμουήλ after Κύριος as a doublet, but this is not found in Sahidic. The place of ἔτι varies in the manuscripts, as can be seen in the text above. Sahidic seems to render ἔτι with ⲛⲕⲉⲥⲟⲡ, but it is not possible to argue on the basis of the Sahidic for a certain placement of ἔτι in the Vorlage. After this sentence there is a Hexaplaric approximation καὶ ἀνέστη Σαμουήλ in A O 127 d-68 122 554 (without Σαμουήλ in 44–125). This approximation is also not attested in the Sahidic. 25: 34 …εἰ μὴ ἔσπευσας καὶ παρεγένου εἰς ἀπάντησίν μοι,… Sa ⲉⲛⲉ ⲙⲡⲉⲧⲱⲟⲩⲛ̄ ⲧⲁⲭⲏ ⲉⲉⲓ ⲉⲣⲁⲧ Fa ]ⲙⲡⲉⲕⲱⲗⲉⲙ ⲛⲧⲉⲓ̈ ⲉⲃⲁⲗ ⲉϩⲗⲉⲓ ̅ [ Here, the translator has formulated a natural Sahidic expression by rendering the aorist παρεγένου with the preposition ⲉ and an infinitive. In this example, the καί-clause coordinates with the εἰ-clause ‘if you had not hurried and come to meet me’. Sahidic renders the verb σπεύδω with ⲧⲱⲟⲩⲛ ⲧⲁⲭⲏ, with a common verb ‘to rise’ with a Greco-Coptic word ⲧⲁⲭⲏ ‘quickly’. Text-critical note: There is one small Fayyumic fragment for 1 Sam, and it preserves this verse partially; one can read this sentence, and the construction clearly differs from that found in the Sahidic version.22 Fayyumic offers a past tense 2nd person singular fem.neg. of the first verb like Sahidic, but with a different verb: ⲕⲱⲗⲉⲙ ‘to hasten, hurry’, which is an accurate render-

22 My cordial thanks to Anne Boud’hors who helped me in deciphering the letters on this wrinkled fragment.

40

Description of the translation technique

ing of σπεύδω. The second verb is a conjunctive 2nd person singular fem. in Fayyumic. It fits the context well, but differs from the infinitive used in Sahidic. The verb ⲉⲓ ‘to come’ is used in Sahidic with the preposition ⲉ ‘to, towards’, whereas in Fayyumic ⲉⲓ occurs with the adverb ⲉⲃⲁⲗ ⲉϩⲗⲉⲓ ̅ ‘came out to’. The meaning is approximately the same, but the construction differs. Given the small data sample, the Sahidic and Fayyumic texts appear to have arisen from independent translation events. 1: 15

καὶ οἶνον καὶ μέθυσμα οὐ πέπωκα, καὶ ἐκχέω τὴν ψυχήν μου ἐνώπιον Κυρίου·

ⲁⲩⲱ ⲙⲡⲓⲥⲉ ⲏⲣⲡ·̄ ⲟⲩⲇⲉ ⲥⲓⲕⲉⲣⲁ· ⲉⲓⲡⲱϩⲧ ⲛⲧⲁⲯⲩⲭⲏ ⲙ̄ⲡⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲙⲡϫⲟⲉⲓⲥ· In this example, a Sahidic focalizing present translates the Greek καί-clause. Here, the focalization probably emphasizes that it was the Lord to whom Hannah poured out her soul. 19: 22 καὶ ἐπορεύθη καὶ αὐτὸς εἰς Αρμαθαιμ ⲁϥⲃⲱⲕ ϩⲱⲱϥ ⲉϩⲟⲩⲛ ⲉⲁⲣⲙⲁⲑⲁⲉⲓⲙ There are six cases like the one above:23 the intensive pronoun ϩⲱⲱϥ renders καὶ αὐτός, the emphasizing pronoun.24 2.1.3.1.c) καί-clause rendered with ⲁⲩⲱ According to my calculations, there are 707 cases where a καί-clause is rendered with ⲁⲩⲱ as the connective. While many of the conjunctions used in Sahidic 1 Sam are of Greek origin, ⲁⲩⲱ is an indigenous Egyptian word. This word is originally an imperative of the verb ‘to put, add, place’.25 According to Layton, it is used to connect all kinds of entity terms,26 and Ariel Shisha-Halevy describes it as “a focusing additive conjunct”.27 In Walter Ewing Crum’s dictionary, one finds different usages of this conjunction such as: joining phrases, introducing narrative, introducing apodosis, joining nouns or adjectives.28 Feder describes the use of this conjunction as follows: ⲁⲩⲱ dient im klassischen Koptischen – in der Regel – zur Einführung einer neuen Sinneinheit, zur Anzeige eines neuen Faktums oder einer neuen, darauffolgenden Aktion. Sobald nur derselbe Gedanke (Erzählung) fortgeführt wird oder eine wie

23 The other similar cases are found in 8: 20, 14: 22, 19: 20, 19: 21 and 31: 5. 24 This same word is also used as an adversative connector; see section 2.1.3.1.e) for examples. 25 W. Vycichl, Dictionnaire Étymologique, 18. Wolfhart Westendorf (Handwörterbuch, 14) renders the imperative “füge hinzu!” For a more thorough analysis of this imperative, see William F. Edgerton, “Obsolescence of the Imperative Mood”, 66. 26 Layton, Grammar, §145, 231. 27 Ariel Shisha-Halevy, Coptic Grammatical Categories, 51. 28 W.E. Crum, Dictionary, 19–20.

41

Clause connections

auch immer geartete Anreihung von Fakten und Aktionen vorliegt, wird ⲁⲩⲱ vermieden (wie in der Anreihung von Nomina ⲙⲛ̄ erscheinen muss).29

In the Sahidic New Testament, ⲁⲩⲱ often occurs as a rendering of καί.30 The fact that ⲁⲩⲱ frequently occurs where the source text reads καί and only sporadically to render another Greek connective (δέ) might create a misguided inclination to see ⲁⲩⲱ as a standard equivalent of καί. If, however, one considers the statistics, it is evident that asyndeton occurs, at least in 1 Sam, more often than ⲁⲩⲱ to render clause connective καί. This can be partly explained by the different structure of Greek and Coptic as languages; Coptic can express with different verbal forms some relations that in Greek are expressed with conjunctions. καί-clauses rendered with ⲁⲩⲱ-clauses

Consecutive Coordination New turn in the story Different usages

320 216 146 26

Overlaps with ⲇⲉ

In the following instances, I provide examples of these subgroups. The examples proceed in the same order as they are listed above. 3: 8

καὶ εἶπεν Ἰδοὺ ἐγώ, ὅτι κέκληκάς με. καὶ ἐσοφίσατο Ἠλεὶ ὅτι Κύριος κέκληκεν τὸ παιδάριον.

ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ· ϫⲉ ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲓⲉⲓ ϫⲉ ⲕⲙⲟⲩⲧⲉ ⲉⲣⲟⲓ· ⲁⲩⲱ ϩⲏⲗⲉ̄ⲓ·̅ ⲁϥⲉⲓⲙⲉ ϫⲉ ⲡϫⲟⲉⲓⲥ ⲡⲉ ⲛⲧⲁϥⲙⲟⲩⲧⲉ ⲉⲡϣⲏⲣⲉ ϣⲏⲙ The largest subgroup, 320 instances, of all the clauses connected with ⲁⲩⲱ contains clauses with consecutive meaning (i. e., the conjunction relates the clause to a prior clause). These cases support Feder’s statement cited above. In verse 3: 8, the text relates how the Lord called young Samuel and how Eli after three times understood that it was the Lord who had called the boy. In most cases, the consecutive meaning is not as obvious as in this passage. 15: 19 …ἀλλ᾽ὥρμησας τοῦ θέσθαι ἐπὶ τὰ σκῦλα καὶ ἐποίησας τὸ πονηρὸν ἐνώπιον Κυρίου;

29 Feder, Biblia Sahidica, 89. Ludwig Stern (Koptische Grammatik, §592) states that with two verbs in the same temporal form this conjunction expresses “ein übergehen zu etwas neuem”. 30 ⲁⲩⲱ to render καί “passim” with no other Greek conjunctions mentioned: Michel Wilmet, Concordance II: 1, 28. I wonder if it really is the case that no other conjunctions in the NT are rendered with ⲁⲩⲱ. In 1 Sam there are some cases where δέ is rendered with ⲁⲩⲱ and in 2 Samuel there is one such case.

42

Description of the translation technique

ⲁⲗⲗⲁ ⲁⲕϯ ⲙⲡⲉⲕⲟⲩⲟⲓ ⲉϫⲛ ⲛⲉϣⲱⲗ ⲉϥⲓⲧⲟⲩ· ⲁⲩⲱ ⲁⲕⲉⲓⲣⲉ ⲙⲡⲟⲛⲏⲣⲟⲛ ⲙⲡⲉⲙⲧⲟ ⲉⲃⲟⲗ ⲙ̄ⲡϫⲟⲉⲓⲥ In this case, the consecutive meaning is secondary, and one can also see an epexegetical shade in this case;31 Saul has acted against the command of the Lord by retaining the spoils of victory, and, in this way, he has done what is evil before the Lord. There are 216 clauses where the same subject continues to act or the content of the clause is merely a list of actions, enumerated one after another. In these cases ⲁⲩⲱ seems to be a generic connective without any specialized meaning. These cases do not support what Feder describes as the typical and classic use of ⲁⲩⲱ.32 4: 19

καὶ ἤκουσεν τὴν ἀγγελίαν ὅτι ἐλήμφθη ἡ κιβωτὸς τοῦ θεοῦ καὶ ὅτι τέθνηκεν ὁ πενθερὸς αὐτῆς καὶ ὁ ἀνὴρ αὐτῆς,…

ⲁⲩⲱ ⲛⲧⲉⲣⲉⲥⲥⲱⲧⲙ ⲉⲡⲟⲩⲱ· ϫⲉ ⲁⲩϥⲓ ⲛⲧⲕⲟ̅ⲓⲃⲱⲧⲟⲥ ⲙⲡⲛⲟⲩⲧⲉ ⲁⲩⲱ ϫⲉ ⲁ ̄ ⲡⲉⲥϣⲟⲙ ⲙⲟⲩ· ⲙⲛ ⲡⲉⲥⲕⲉϩⲁⲓ· In this instance, ⲁⲩⲱ functions without any special meaning, only as a coordinator that punctuates two ϫⲉ-clauses, just as καί functions in Greek between two ὅτι-clauses. καὶ θέσθαι αὐτοὺς ἑαυτῷ χιλιάρχους … καὶ ποιεῖν σκεύη πολεμικὰ αὐτοῦ καὶ σκεύη ἁρμάτων αὐτοῦ· καὶ τὰς θυγατέρας ὑμῶν λήμψεται εἰς μυρεψοὺς καὶ εἰς μαγειρίσσας καὶ εἰς πεσσούσας·

8: 12–13

ⲉⲧⲣⲉϥϩⲟⲕⲟⲩ ⲟⲛ· ⲛϥⲁⲩ ⲛⲁϥ ⲛⲭⲓⲗⲓⲁⲣⲭⲟⲥ· … ⲁⲩⲱ ⲉⲧⲣⲉⲩⲙⲟⲩⲛⲅ ̄ ⲛⲛⲉϥⲥⲟⲧⲃⲉϥ ⲙⲙⲓϣⲉ· ⲁⲩⲱ ⲛⲥⲁⲧⲃⲉϥ ⲛⲛⲉϥϩⲁⲣⲙⲁ·ⲁⲩⲱ ⲛⲉⲧⲉⲛⲕⲉϣⲉⲉⲣⲉ ϥⲛⲁϫⲓⲧⲟⲩ ⲉⲧ̄ ⲣⲉⲩⲡⲉⲥⲧⲥⲟϭⲉⲛ· ⲁⲩⲱ ⲉⲧⲣⲉⲩⲥⲁϩⲧⲉ· ⲁⲩⲱ ⲉⲧⲣⲉⲩⲧⲱϭ In this verse, the conjunction ⲁⲩⲱ occurs in a list to enumerate the professions in which the future king will employ the daughters of Israel. This list was preceded by a similar one in the prior verse. 9: 13

…ὅτι οὗτος εὐλογεῖ τὴν θυσίαν, καὶ μετὰ ταῦτα ἐσθίουσιν οἱ ξένοι·

ϫⲉ ⲡⲁⲓ ⲡⲉⲧⲛⲁⲥⲙⲟⲩ ⲉⲧⲉⲑⲩⲥⲓⲁ· ⲁⲩⲱ ⲙⲛⲛⲥ̄ ⲁ ⲛⲁⲓ ⲛϣⲙ̄ⲙⲟ·̅ ⲥⲉⲛⲁⲟⲩⲱⲙ Here, the meaning of the latter clause is temporal, relating the actions to the prior narrative. The Greek expression μετὰ ταῦτα (Sa ⲙⲛⲛⲥ̄ ⲁ ⲛⲁⲓ) is the basis for the temporal reference. In this case, ⲁⲩⲱ only punctuates between clauses.

31 Actually many of the cases can be understood in different ways, and especially consecutive and epexegetical meanings often overlap. 32 Feder, Biblia Sahidica, 89.

Clause connections

43

23: 28 καὶ ἀνέστρεψεν Σαοὺλ μὴ καταδιώκειν ὀπίσω Δαυείδ, καὶ ἐπορεύθη εἰς συνάντησιν τῶν ἀλλοφύλων·

ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲕⲟⲧϥ ̄ ⲉⲡⲁϩⲟⲩ ⲉⲧⲙ̄ⲡⲱⲧ ⲛⲥⲁ ⲇⲁⲩⲉⲓⲇ ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲃⲱⲕ ⲉⲣⲁⲧⲟⲩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ In this verse, the translator has logically avoided a potential misunderstanding. Had he simply used asyndeton, David could be understood as the subject of the clause. Therefore, the proper name Saul is needed to make it clear that the subject is not David and the conjunction is required to make a small stop between these two clauses. This activity of the translator corresponds to Theo A. van der Louw’s statement that transformation has its origin in a literal rendering considered improper by the translator.33 31: 4

…μὴ ἔλθωσιν οἱ ἀπερίτμητοι οὗτοι καὶ ἀποκεντήσωσίν με καὶ ἐμπαίξωσίν μοι.

ⲙⲏⲡⲱⲥ ⲛⲥⲉⲉⲓ ⲛϭⲓ ⲛⲓⲁⲧⲥⲃⲃ̄ ⲉ· ⲛⲥⲉⲙⲟⲟⲩⲧ ⲁⲩⲱ ⲛⲥⲉⲥⲱⲃⲉ ⲛⲥⲱⲓ Here, the Sahidic reads three conjunctives in a chain. The first clause begins with the conjunction ⲙⲏⲡⲱⲥ, the second follows with asyndeton and the third uses the conjunction ⲁⲩⲱ. Actually, the third conjunction is redundant here, and it, for some unknown reason, disturbs the connection between the clauses, unnecessarily punctuating the last clause in the chain. Here, the meaning of ⲁⲩⲱ also has an epexegetical nuance. In 146 cases, ⲁⲩⲱ-clauses express a new turn in the story with the change of the subject. The change of the subject probably was one signal for the translator that often encouraged the use of ⲇⲉ. Thus, the use and meaning of these conjunctions seem to overlap in clauses like the following. 13: 16 …καὶ ὁ λαὸς οἱ εὑρεθέντες μετ᾽αὐτῶν ἐκάθισαν ἐν Γαβεὲ Βενιαμεὶν καὶ ἔκλαιον, καὶ οἱ ἀλλόφυλοι παρεμβεβλήκεισαν εἰς Μαχεμάς. ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲛⲧⲁⲩϩⲉ ⲉⲣⲟϥ ⲛⲙ̄ⲙⲁⲩ· ⲁⲩϩⲙⲟⲟⲥ ϩⲛ̄ ⲅⲁⲃⲁⲁ· ⲡⲕⲁϩ ⲛⲃⲉⲛⲓⲁⲙⲓⲛ ⲁⲩⲣⲓⲙⲉ· ⲁⲩⲱ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩⲧⲁⲁⲩ ⲉϩⲣⲁⲓ ⲉⲙⲁⲭⲙⲁⲥ̅ In prior narration, Saul, Jonathan and the people stayed in Geba and after that, the actions of the Philistines were related. One might expect ⲇⲉ here, since this connective is often used in coordination when the subject changes. 19: 20 καὶ εἶδαν τὴν ἐκκλησίαν τῶν προφητῶν, καὶ Σαμουὴλ εἱστήκει καθεστηκὼς ἐπ᾽ αὐτῶν. ⲁⲩⲱ ⲁⲩⲛⲁⲩ ⲉⲧⲉⲕⲕⲗⲏⲥⲓⲁ ⲛⲛⲉⲡⲣⲟⲫⲏⲧⲏⲥ· ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲉϥⲁϩⲉⲣⲁⲧϥ ϩⲓϫⲱⲟⲩ·

33 T.A. van der Louw, Transformations, 57.

44

Description of the translation technique

This example might alternatively have been translated with the particle ⲇⲉ. Saul had sent messengers to take David, and while they were going, they saw Samuel prophesying. This clause starts a new phase in the narrative. The remaining 26 cases are those where ⲁⲩⲱ is used to coordinate its clause to clauses with various other meanings: adversative, causal, concessive, epexegetical, interrogative, or temporal. Some examples serve to illustrate these different contexts where ⲁⲩⲱ appears. 24: 12 καὶ ἰδοὺ τὸ πτερύγιον τῆς διπλοΐδος σου ἐν τῇ χειρί μου· ἐγὼ ἀφῄρηκα τὸ πτερύγιον καὶ οὐκ ἀπέκταγκά σε·

ⲁⲩⲱ ⲉⲓⲥ ⲡⲧⲁⲡ ⲛⲧⲉⲕⲭⲗⲁⲙⲩⲥ ϩⲛ̄ ⲛⲁϭⲓϫ· ⲁⲛⲟⲕ ⲁⲓϥⲓⲧϥ·̄ ⲁⲩⲱ ⲙⲡⲓⲙⲟⲟⲩⲧⲕ In this case, the meaning is clearly adversative. 15: 6

μὴ προσθῶ σε μετ᾽ αὐτοῦ καὶ σὺ ἐποίησας ἔλεος μετὰ τῶν υἱῶν Ισραηλ

ϫⲉ ⲛⲛⲁⲧⲁⲕⲟⲕ ⲛⲧⲉϥϩⲉ· ⲁⲩⲱ ⲛⲧⲟⲕ ⲁⲕⲉⲓⲣⲉ ⲛⲟⲩⲛⲁ ⲙⲛ̄ ⲛⲉ̄ ϣⲏⲣⲉ ⲙⲡⲓⲏⲗ̄ ̄ This is an example of causal meaning. Saul warns the Kenites to leave before he destroys the place, and gives the reason why he does not want to kill the Kenites. 22: 17 …ὅτι ἡ χεὶρ αὐτῶν μετὰ Δαυείδ, καὶ ὅτι ἔγνωσαν ὅτι φεύγει αὐτός καὶ οὐκ ἀπεκάλυψαν τὸ ὠτίον μου·

ϫⲉ ⲧⲟⲟⲧⲟⲩ ⲥⲙⲟⲛⲧ ⲙⲛ̄ ⲇⲁⲩⲉⲓⲇ· ⲁⲩⲱ ⲛⲧⲟⲟⲩ ⲁⲩⲉⲓⲙⲉ ϫⲉ ϥⲡⲏⲧ ⲛⲧ ̄ ⲟⲟⲧ· ⲁⲩⲱ ⲙⲡⲟⲩⲧⲁⲙⲟⲓ In this passage, the reader encounters the reason for Saul’s execution of the priests. In the Greek, the καί coordinates two subordinate clauses, but the Sahidic translator has transformed the second subordinate clause into a main clause. The meaning of the ⲁⲩⲱ-clause is best understood as concessive; even though the priests knew what was going on, they did not inform Saul. Πῶς πορευθῶ; καὶ ἀκούσεται Σαοὺλ καὶ ἀποκτενεῖ με. ⲉⲓⲛⲁⲃⲱⲕ ⲛⲁϣ ⲛϩ̄ ⲉ· ⲁⲩⲱ ⲛⲧⲉ ⲥⲁⲟⲩⲗ ⲥⲱⲧⲙ̄ ⲛϥⲙ ̄ ⲟⲟⲩⲧ In this case, Samuel asks the Lord for more instructions. The construction in Coptic is highly interesting. The question “How shall I go?” is followed by an apodotic conjunctive.34 .

16: 2

34 Layton, Grammar, §356. See also Shisha-Halevy, Coptic Grammatical Categories, §7.2.6. Prof. Boud’hors drew my attention to this feature in private correspondence. My own question is what this example tells us about the Coptic translator. According to ShishaHalevy, this construction is often found in a rhetorical function, but is poorly documented. Therefore this instance is important It seems that we have here an example of the Coptic translator using a fitting expression in direct speech. However, one should try to find more such examples from the Coptic Bible translations in order to avoid building too much on a single case.

Clause connections

2: 29

45

καὶ ἵνα τί ἐπέβλεψας ἐπὶ τὸ θυμίαμά μου καὶ εἰς τὴν θυσίαν μου ἀναιδεῖ ὀφθαλμῷ; καὶ ἐδόξασας τοὺς υἱούς σου ὑπὲρ ἐμὲ…

ⲁⲩⲱ ⲉⲧⲃⲉ ⲟⲩ ⲁⲕϭⲱϣⲧ ⲉϫⲛ ⲧⲁⲑⲩⲥⲓⲁ ⲁⲩⲱ ⲉϫⲙ ⲡⲁϣⲟⲩϩⲏⲛⲉ ϩⲛ ⲟⲩⲃⲁⲗ ⲛⲁⲧϣⲓⲡⲉ· ⲁⲩⲱ ⲁⲕϯⲉⲟⲟⲩ ⲛⲛⲉⲕϣⲏⲣⲉ ⲉϩⲟⲩⲉ ⲉⲣⲟⲓ Here, ⲁⲩⲱ renders a καί-clause that continues the question of the previous clause (ἵνα τί) without repeating the interrogative. 14: 21–22

ἐπεστράφησαν καὶ αὐτοὶ εἶναι μετὰ Ἰσραὴλ τῶν μετὰ Σαοὺλ καὶ Ἰωναθάν. καὶ πᾶς Ἰσραὴλ οἱ κρυπτόμενοι ἐν τῷ ὄρει Ἐφράιμ, καὶ ἤκουσαν ὅτι πεφεύγασιν οἱ ἀλλόφυλοι,…

ⲁⲩⲡⲱⲱⲛⲉ ⲉⲧⲣⲉⲩϣⲱⲡⲉ ϩⲱⲟⲩ ⲙⲛ̄ ⲡⲓⲏⲗ̄ ·̅ ⲛⲉⲧⲏⲡ ⲉⲥⲁⲟⲩⲗ· ⲙⲛ ⲓⲱⲛⲁⲑⲁⲛ· ⲁⲩⲱ ⲡⲓⲏⲗ̄ ̅ ⲧⲏⲣϥ̄ ⲛⲉⲩⲗⲏϫ ϩⲙ̄ ⲡⲧⲟⲟⲩ ⲛⲉⲫⲣⲁⲓ̄ ⲙ̄· ⲁⲩⲥⲱⲧⲙ̄ ϫⲉ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩϭⲱⲧⲡ̄ This ⲁⲩⲱ-clause describes the circumstances, giving the background information for the appearance of the people who had been hiding. The Sahidic translator has understood καὶ πᾶς Ἰσραὴλ οἱ κρυπτόμενοι ἐν τῷ ὄρει Ἐφράιμ as a nominal clause35 and formulated two independent clauses: ‘All Israel was hiding in the mountains’ and ‘They heard that the Philistines escaped’. 10: 14 καὶ εἶπαν Ζητεῖν τὰς ὄνους· καὶ εἴδαμεν ὅτι οὐκ εἰσίν, καὶ εἰσήλθομεν πρὸς Σαμουήλ. ⲡⲉϫⲁϥ ⲛⲁϥ ϫⲉ ⲛⲧ ̄ ⲁⲛⲃⲱⲕ ⲉⲕⲱⲧⲉ̅ ⲛⲥⲁ ⲛⲉⲟⲟⲩ· ⲁⲩⲱ ⲛⲧ ̄ ⲉⲣⲉⲛⲛⲁⲩ ϫⲉ ⲙⲡⲉⲛϩⲉ ⲉⲣⲟⲟⲩ· ⲁⲛⲃⲱⲕ ⲉⲣⲁⲧϥ̅ ⲛⲥⲁⲙⲟⲩⲏⲗ Here, the ⲁⲩⲱ-clause is temporal as is indicated with the temporal verbal form used in Sahidic. The Greek text could as well be interpreted as causal ‘since we saw that they were not’ but the Sahidic does not open the way to this interpretation. 2.1.3.1.d) καί-clause rendered with ⲇⲉ The Greek connective δέ coordinates clauses. According to John Dewar Denniston, this conjunctive particle can express “pure connection ‘and’, or contrast ‘but’, with all that lies between”.36 Typically it is used in the second or third place in a sentence. The Greek translator of 1 Sam has used δέ rarely compared to the frequency in the Pentateuch.37 Probably one reason for this is the word order: this connective, unlike καί, cannot occupy the initial position in a clause. Here, one also notes a difference between the Greek translator and the later Sahidic translator; the latter had no difficulty using ⲇⲉ, regardless of the restrictions concerning the word order.

35 Like 376 in Greek, which leaves out the article οἱ and thus reads a periphrastic clause. 36 Denniston, Greek Particles, 162. 37 Aejmelaeus, On the Trail, 128–9.

46

Description of the translation technique

When used as an adversative connective, the meaning of δέ can be separated from ἀλλά, which is eliminating, whereas δέ more often is balancing between two ideas.38 In Coptic, this originally Greek connective is widely in use. The use of ⲇⲉ as a rendering of καί actually proves that this connective had wide range of use in Coptic texts. Within 1 Sam, the ⲇⲉ-particle regularly begins new sentences. Especially when something new appears, typically a new subject, the translator uses this connective. In my corpus, there are 401 cases where a καί-clause is translated with a ⲇⲉ-clause. Although in a few of those cases the translator may have theoretically had δέ in the Vorlage, the overall picture still remains clear: the ⲇⲉ -connective was for the translator a natural equivalent for καί. Concerning textual criticism, it is important to note that ⲇⲉ was often used as an equivalent for καί and it cannot be seen as an indicator of δέ in the Vorlage.39 In the following I present examples from 1 Sam. The typical and clear cases appear first, and then follow cases where variation occurs in Sahidic manuscripts. All in all, variation is present in 50 out of 401 cases. 13: 12–13…καὶ ἀνήνεγκα τὴν ὁλοκαύτωσιν. καὶ εἶπεν Σαμουὴλ πρὸς Σαουλ… ⲁⲓⲧⲁⲗⲟ ⲛⲛⲉϭⲗⲓⲗ ⲉϩⲣⲁⲓ· ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲉϩⲟⲩⲛ ϩⲛ ⲥⲁⲟⲩⲗ This is a typical example. First Saul speaks, and then Samuel. The change of the speaker is emphasized by ⲇⲉ. 28: 23 καὶ οὐκ ἐβουλήθη φαγεῖν· καὶ παρεβιάζοντο αὐτὸν οἱ παῖδες αὐτοῦ καὶ ἡ γυνή,… ⲁⲩⲱ ⲙⲡⲉϥⲟⲩⲉϣ ⲟⲩⲱⲙ· ⲛⲉϥϩⲙ̄ϩⲁⲗ ⲇⲉ ⲙⲛ̄ ⲧⲉⲥϩⲓⲙⲉ ⲁⲩⲁⲛⲁⲅⲕⲁⲍⲉ ⲙⲙⲟϥ In another typical example without Sahidic textual variants, the narration continues and a change of subject precedes the particle ⲇⲉ. In the passages where the same subject continues to act, the ⲇⲉ-connective is usually not used as an equivalent for καί. . This example also shows how a rare Greek word is translated with a more common one: ⲁⲩⲁⲛⲁⲅⲕⲁⲍⲉ for παρεβιάζοντο.40 3: 9

καὶ ἐρεῖς Λάλει κύριε, ὅτι ἀκούει ὁ δοῦλός σου. καὶ ἐπορεύθη Σαμουὴλ καὶ ἐκοιμήθη ἐν τῷ τόπῳ αὐτοῦ.

38 Denniston, Greek Particles, 165. 39 For some examples from Jeremiah, see Feder, Biblia Sahidica, 92. In Brooke/McLean an example is found in verse 20: 22 where one cannot know whether there was καί or δέ or asyndeton in the source text of the Sahidic translator. For examples from Coptic John, see Askeland, John’s Gospel, 30. 40 To be sure, there are no variants with ἀναγκάζω in Greek. Nagel mentions in his article that the use of a Greek word in Sahidic to render another Greek word tells about the distribution of that word (P. Nagel, “Einwirkung des Griechischen”, 336–7). Hasznos (GraecoCoptica, 11) gives some examples and sees that “clearly the translator employs the loanword known and used already in the Egyptian idiom for the given meaning”.

47

Clause connections

ⲉⲕⲉϫⲟⲟⲥ ϫⲉ ⲡϫⲟⲉⲓⲥ· ϣⲁϫⲉ ⲡⲉⲕϩⲙ̄ϩⲁⲗ ⲥⲱⲧⲙ̄· ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲁϥⲃⲱⲕ ⲁϥⲉⲛⲕⲟⲧⲕ ̄ ϩⲙ̄ ⲡⲉϥⲙⲁ In this verse there appears yet another example in which καί is rendered with ⲇⲉ. Because the underlined καί-clause begins a new scene and the subject changes from Eli to Samuel, the Sahidic translator has chosen the ⲇⲉconnective. The next καί-clause, on the other hand, continues the same scene with the same subject, and thus the translator opts for asyndeton. 17: 14 καὶ Δαυὶδ αὐτὸς ὁ νεώτατος καὶ οἱ τρεῖς οἱ μείζονες ἐπορεύθησαν ὀπίσω Σαούλ (L-19 509 f 55 158 318 =ΜΤ) αὐτὸς] om 19 488 | +ἐστιν A CI d 554 καί 1° – νεώτατος] δαδ δε ὁ νεώτερος CII s-488 M Sa ⲇ̄ⲁⲇ ̄ ̄ ⲇⲉ ⲡⲕⲟⲩⲓ ⲉⲣⲟⲟⲩ· ⲡϣⲟⲙⲛⲧ ̄ ⲇⲉ ⲛⲛ ̄ ⲟϭ ⲛⲥⲟⲛ· ⲁⲩⲃⲱⲕ ⲇⲉ ⲁⲩⲟⲩⲁϩⲟⲩ ⲛⲥⲁ ⲥⲁⲟⲩⲗ This entire verse did not belong to the original Septuagint, but was added later to conform the Greek text with the MT. Sahidic seems to translate a Greek reading without αὐτός and ἐστιν here since there is no verb and even no copula in the first clause in Sahidic, nor any equivalent for αὐτός. With regard to translation technique, one cannot argue for either of these two possible Greek source texts; the Sahidic could render a text like that found in 19 488 or a text similar to CII s-488. Both καί and δέ are frequently rendered with ⲇⲉ, and the difference between νεώτατος and νεώτερος is indistinguishable in Sahidic Coptic. The second καί-clause has been translated in a more formally literal fashion—καί is rendered with ⲇⲉ. However, the Sahidic text offers a translational abnormality insofar as it reads two connectives, the latter of which has no equivalent in Greek and is to be seen as redundant here. Having weighed these comparatively simple verses, this study will now consider more complicated cases where variants in Sahidic are at hand. Examples are presented in the same order they are listed in the table below. The variation may be classified under the following categories:41 SaM ⲇⲉ SaM ⲇⲉ SaM asyndeton SaA ⲇⲉ SaM ⲇⲉ SaM ⲇⲉ

SaA asyndeton SaA omits the whole clause ⲇⲉ in 1(+) Sahidic ms SaM some other connective SaA ⲁⲩⲱ SaS asyndeton

25 cases 7 cases 9 cases 4 cases 4 cases once

41 In the table, the main mss are included, i. e., SaM and SaA. SaS appears on the last line since only in this instance does variation occur between the main manuscript and a single fragment. Additionally, there are cases where either SaM or SaA is not alone in its reading, but a third manuscript is extant with a text supporting either SaM or SaA. Where relevant, this additional information appears in the examples below.

48

Description of the translation technique

Many of the readings with variants in Sahidic cannot be evaluated in such a way that a reading with or without a certain conjunction is to be preferred as a more original Sahidic translation and, accordingly, the other could be seen as a reformulation. These cases show that Sahidic manuscripts have undergone some changes independently during the transmission processes. SaA does not offer a translation for καί in 25 cases, and in an additional seven cases the whole clause is omitted.42 A later chapter in this study will offer a more detailed description of the extant Sahidic manuscripts of 1 Sam.43 It suffices here to note that SaA is prone to omissions and abbreviations. 20: 3 οὐκ ἔστιν τοῦτο. καὶ ἀπεκρίθη Δαυεὶδ τῷ Ἰωναθὰν… SaM ⲉⲛϥϣⲟⲟⲡ ⲁⲛ ⲣⲱ̅· ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁϥⲟⲩⲱϣⲃ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲓⲱⲛⲁⲑⲁⲛ· ⲉⲛϥϣⲟⲟⲡ ⲣⲱ ⲁⲛ· ⲇⲁⲩⲉⲓⲇ ⲡⲉϫⲁϥ ⲛⲓⲱⲛⲁⲑⲁⲛ SaA In this example, SaM reads a longer clause, whereas SaA is shorter. Both renderings of the καί-clause (asyndeton, ⲇⲉ) are often found in Sahidic 1 Sam, but, generally, the reading of SaA appears to be a simplification of the reading found in SaM.44 In nine cases45 where the Greek tradition unanimously agrees on the presence of the conjunction, SaM does not offer an equivalent for καί. In seven of these passages, SaA offers the connective ⲇⲉ. In one case, SaS has the particle ⲇⲉ, and, once, SaT offers the particle ⲇⲉ. 6: 20–21 καὶ πρὸς τίνα ἀναβήσεται κιβωτὸς Κυρίου ἀφ᾽ἡμῶν; καὶ ἀποστέλλουσιν ἀγγέλους πρὸς τοὺς … λέγοντες SaM ⲁⲩⲱ ⲉⲣⲉ ⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡϫⲟⲉⲓⲥ ⲛⲁⲃⲱⲕ ⲉϩⲣⲁⲓ̈ ⲛⲛⲓⲙ ⲛⲥⲗⲟ ϩⲁⲧⲏⲛ· ⲁⲩϫⲟⲟⲩ ̅ ⲛϩⲉⲛⲃⲁⲓϣⲓⲛⲉ ⲉⲣⲁⲧⲟⲩ … ⲉⲩϫⲱ ⲙⲙⲟⲥ ⲛⲁⲩ ϫⲉ ⲁⲩⲱ ⲉⲣⲉ ⲧϭⲓⲃⲟⲩⲧⲟⲥ ⲙ̄ⲡϫⲟⲉⲓⲥ ⲛⲁⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲛⲓⲙ ⲛⲥ̄ ⲗⲟ ϩⲁϩⲧⲏⲛ· ⲁⲩϫⲟⲟⲩ SaA ⲇⲉ ⲛϩ̄ ⲉⲛⲃⲁⲓϣⲓⲛⲉ ⲉⲣⲁⲧⲟⲩ … ⲉⲩϫⲱ ⲙ̄ⲙⲟⲥ· ϫⲉ ⲁⲩⲱ ⲉⲣⲉ ⲧⲕⲓⲃⲱⲧⲟⲥ ⲙ̄ⲡϫⲟⲉⲓⲥ ⲛⲁⲃⲱⲕ ⲉϩⲣⲁⲓ̈ ⲛⲛ SaE ̄ ⲓⲙ ⲛⲥⲗⲟ ϩⲁϩⲧⲏⲛ· ⲁⲩϫⲟⲟⲩ ⲇⲉ ⲛϩ̄ ⲉⲛϥⲁⲓ̈ϣⲓⲛⲉ ⲉⲣⲁⲧⲟⲩ … ⲉⲩϫⲱ ⲙ̄ⲙⲟⲥ ⲛⲁⲩ ϫⲉ In this example, there are no Greek variants concerning the conjunction. The fragment SaE also offers the connective in agreement with SaA. Text-critical note: In 6: 14–21, SaAE read occasionally against SaM, and, likewise, SaME read sometimes against SaA, but, on the whole, these are only small-scale discrepancies. Therefore one has to conclude that the texts of both SaA and SaM have undergone some changes during the course of their transmission. SaE possesses such a similar wording that it obviously belongs

42 43 44 45

6: 13, 8: 1, 19: 23, 20: 7, 25: 3, 30: 9, 30: 21. Section 3.2. See section 3.2. 6: 21, 9: 12, 16: 4, 19: 10, 20: 24, 20: 41, 21: 13, 22: 7, 23: 9.

Clause connections

49

to the same Sahidic tradition with SaMA. The editor of SaE , Kahle, mentions the close relationship of the Pierpont Morgan manuscripts (here SaM) and earlier fragments, whereas the manuscripts from the White Monastery (here SaA) are on their own.46 SaE has been dated to the 7th century based upon paleographic evidence.47 20: 24 καὶ παραγίνεται ὁ μήν καὶ ἔρχεται ὁ βασιλεὺς ἐπὶ τὴν τράπεζαν τοῦ φαγεῖν. SaM ϩⲏ ⲇⲉ ⲙ̄ⲡⲉⲃⲟⲧ ⲛⲧⲉⲣⲉϥϣⲱⲡⲉ· ⲡⲣⲣ̄ ⲟ ⲁϥⲃⲱⲕ ⲉϩⲟⲩⲛ ⲉϫⲛ ⲧⲉⲧⲣⲁⲡⲉⲍⲁ ⲉⲟⲩⲱⲙ· SaA ⲁⲩⲱ ⲥⲟⲩⲁ̄ ⲙⲡⲉⲃⲟⲧ ⲁϥϣⲱⲡⲉ· ⲡⲣⲣ̄ ⲟ ⲇⲉ ⲁϥⲃⲱⲕ ⲉϫⲛ ⲧⲉⲧⲣⲁⲡⲉⲍⲁ ⲉⲟⲩⲱⲙ In this example, the Sahidic tradition preserves two distinct readings that are both related to the two occurrences of καί above. The initial Greek clause is rendered by both ⲇⲉ (SaM) and ⲁⲩⲱ (SaA). In the second clause, SaM reads asyndeton against SaA that offers the connective ⲇⲉ. The Sahidic witnesses also contain distinct constructions and vocabulary. SaM preserves a temporal clause followed by a main clause without any connective (‘And when the new moon came, the king came to the table to eat’), translating ὁ μήν as ϩⲏ ⲙ̄ⲡⲉⲃⲟⲧ ‘new moon’.48 SaA, however, offers two coordinated clauses (‘And the new moon came and the king came to the table to eat’), the first one with ⲁⲩⲱ and the second one with ⲇⲉ, translating ὁ μήν as ⲥⲟⲩⲁ ⲙⲡⲉⲃⲟⲧ ‘new moon’. In this verse, the temporal construction of SaM is more nuanced in terms of Coptic idiom, whereas SaA has coordinated these two clauses with the particle ⲇⲉ. In four passages SaA reads ⲇⲉ while SaM offers another connective.49 7: 10 SaM

καὶ ἦν Σαμουὴλ ἀναφέρων τὴν ὁλοκαύτωσιν, καὶ ἀλλόφυλοι προσῆγον εἰς πόλεμον ἐπὶ Ἰσραήλ·

ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲛⲉϥⲧⲁⲗⲟ ⲉϩⲣⲁⲓ ⲙⲡⲉϭⲗⲓⲗ· ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϩⲱⲟⲩ ⲁⲩⲉⲓ ⲉϩⲣⲁⲓ ⲉⲣⲱϥ ⲙⲡⲓⲏ̅ ⲗ̅ ̅ ⲉⲡⲟⲗⲩⲙⲉⲓ ⲛⲙ̄ⲙⲁⲩ ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲉⲛϥ̄ⲧⲁⲗⲟ ⲉϩⲣⲁⲓ ⲙ̄ⲡϥ̄ϭⲗⲓⲗ· ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲉⲓ ⲉⲣⲱϥ ⲙⲡⲓⲏ̅ ⲗ̅ ̅ SaA ⲉⲡⲟⲗⲩⲙⲉⲓ ⲛⲙ̄ⲙⲁⲩ Here, SaM uses the word ϩⲱⲟⲩ, while SaA translates ⲇⲉ. The word ϩⲱⲟⲩ is, according to Müller, an adversative connector of Egyptian origin.50 In this 46 Kahle, Bala’izah, 10–11. 47 Drescher, Kingdoms, xi. Kahle, Bala’izah, 301: “Written in rounded uncials of the seventh century”. 48 Crum, Dictionary, 640b. 49 In verse 23: 4 SaM reads ⲟⲛ, ϩⲱⲱϥ in 17: 41, 21: 1, and ϩⲱⲟⲩ in 7: 10. 50 See Müller, “Greek Connectors”, 9. According to him, the word was “originally agreement-sensitive but later grammaticalized as a 3rd person singular form”. Layton (Grammar, §235b) mentions only the 3rd person singular form in his listing of enclitic conjunctions. Shisha-Halevy (Coptic Grammatical Categories, 173–5) differentiates between two functions of this word: inclusive confrontation ‘he too, he…in his turn’ and particle ‘on the contrary’.

50

Description of the translation technique

case, both Sahidic manuscripts render the source text accurately. In the Sahidic text, the previous clause as well as the next clause read with the connective ⲇⲉ. 23: 3–4 SaM

καὶ πῶς ἔσται ἐὰν πορευθῶμεν εἰς Κεειλά; εἰς τὰ σκῦλα τῶν ἀλλοφύλων εἰσπορευσόμεθα; καὶ προσέθετο Δαυεὶδ ἐρωτῆσαι ἔτι διὰ τοῦ κυρίου·

ⲁⲩⲱ ⲁϣ ⲧⲉ ⲑⲉ ⲉⲧⲛⲁϣⲱⲡⲉ ⲙⲙⲟⲛ· ⲉⲛϣⲁⲛ ⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲧⲕⲉⲉⲓⲗⲁ· ⲛⲧ ̄ ⲉⲛⲧⲁⲁⲛ ⲉϩⲟⲩⲛ ⲉⲛϣⲱⲗ ⲛⲧ ̄ ⲉ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲇⲁⲩⲉⲓⲇ ⲟⲛ ⲁϥⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ ⲉϫⲛⲉ ⲡϫⲟⲉⲓⲥ SaA ⲁⲩⲱ ⲁϣ ⲧⲉ ⲑⲉ ⲉⲧⲛⲁϣⲱⲡⲉ ⲙ̄ⲙⲟⲛ ⲛϣ ̄ ⲁⲛⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲧⲕⲉⲉⲓⲗⲁ· ⲛⲧ ̄ ⲛⲧ ̄ ⲁⲁⲛ ⲉϩⲟⲩⲛ ⲛϣ ̄ ⲱⲗ ⲛⲛ ̄ ⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲇ̅ⲁⲇ ̅ ̅ ⲇⲉ ⲁϥⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ ⲉϫⲛⲉ ⲡϫⲟⲉⲓⲥ Here the Sahidic manuscripts differ when it comes to the connective. SaA has ⲇⲉ, and SaM offers ⲟⲛ ‘again, also, still, further’. Crum references ⲟⲛ as an equivalent for πάλιν, ἔτι, καί, οὐδέ and οὖν.51 In the Sahidic New Testament, ⲟⲛ is used to render καί (206 cases), πάλιν (108 cases), ἔτι (11 cases), οὐδέ (6 cases), οὖν (6 cases), δέ (5 cases) and occasionally some other words. In 63 passages, no corresponding Greek term for ⲟⲛ is recognizable. In many of Michel Wilmet’s examples from the New Testament, ⲟⲛ is not used alone, but together with ⲡⲁⲗⲓⲛ, ⲁⲩⲱ or ⲇⲉ.52 The Coptic word ⲟⲛ as an equivalent for ἔτι occurs only three times in 1 Sam,53 the more common equivalent is the Greek-Coptic loanword ⲉⲧⲓ. In conclusion, there are no translation technical arguments for the priority of either Sahidic manuscript in this passage. 21: 1

καὶ ἀνέστη Δαυεὶδ καὶ ἀπῆλθεν, καὶ Ἰωναθὰν εἰσῆλθεν εἰς τὴν πόλιν.

SaM ⲁⲩⲱ ⲇ̅ⲁⲇ ̅ ̅ ⲁϥⲧⲱⲟⲩⲛ ⲁϥⲙⲟⲟϣⲉ· ⲓⲱⲛⲁⲑⲁⲛ ϩⲱⲱϥ ⲁϥⲃⲱⲕ ⲉϩⲟⲩⲛ ⲉⲧⲡⲟⲗⲓⲥ SaA ⲁⲩⲱ ⲇ̅ⲁⲇ ̅ ̅ ⲁϥⲧⲱⲟⲩⲛ ⲁϥⲙⲟⲟϣⲉ· ⲓⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲁϥⲃⲱⲕ ⲉϩⲟⲩⲛ ⲉⲧⲡⲟⲗⲓⲥ In this case, SaM reads the enclitic conjunction ϩⲱⲱϥ which is semantically similar to ⲇⲉ,54 as can also be seen in this example since SaA reads the particle ⲇⲉ here. As a variant for ⲇⲉ, ϩⲱⲱϥ occurs three times in my material.55

51 Crum, Dictionary, 255b. 52 M. Wilmet, Concordance II: 2, 608–10. 53 In addition to this verse, consider also verses 18: 29 and 20: 17. In verse 18: 29, ⲟⲛ is found in SaA, while SaM offers ⲇⲉ. In verse 20: 17, SaM reads ⲓⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲟⲛ ⲁϥⲟⲩⲱϩ, and SaA omits ⲟⲛ, perhaps intentionally shortening the text. 54 Layton, Grammar, §235b. See also Müller, “Greek Connectors”, 8–9. Müller presents both of these as adversative connectors, ⲇⲉ of Greek origin and ϩⲱⲱϥ of Egyptian origin. 55 In addition to this verse, in verses 7: 10 and 17: 41. More cases with ϩⲱⲱ⸗ are found in section 2.1.3.2.b) dealing with Greek asyndeton and its renderings as well as in subsection 2.1.3.1.e) where καί rendered with other connectives is the topic.

Clause connections

51

In four cases56, the manuscript SaA has the conjunction ⲁⲩⲱ, whereas SaM reads ⲇⲉ (as below). No Greek variants are attested for the καί-conjunctions in the verses below. 23: 5

καὶ ἔφυγον ἐκ προσώπου αὐτοῦ, καὶ ἀπήγαγεν τὰ κτήνη αὐτῶν,…

SaM ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲡⲱⲧ ϩⲁ ⲧⲉⲩϩⲏ· ⲇ̅ⲁⲇ ̅ ̅ ⲇⲉ ⲁϥϥⲓ ⲛⲛⲉⲩⲧⲃⲛ̄ ⲟⲟⲩⲉ· SaA ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲡⲱⲧ ϩⲁ ⲧϥ̄ϩⲏ· ⲁⲩⲱ ⲇ̅ⲁⲇ ̅ ̅ ⲁϥϥⲓ ⲛⲛⲉⲩⲧⲃⲛ̄ ⲟⲟⲩⲉ· Here is an example where both Sahidic manuscripts render the source text accurately but with minor discrepancies. 5: 9

καὶ ἐγενήθη μετὰ τὸ μετελθεῖν αὐτὴν καὶ γίνεται χεὶρ Κυρίου…

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲙⲛⲛ ̄ ⲥⲁ ⲉⲧⲣⲉⲥⲡⲱⲱⲛⲉ· ⲧϭⲓϫ ⲙⲡϫⲟⲉⲓⲥ ⲁⲥϣⲱⲡⲉ… Two Greek stock phrases frequently appear throughout 1 Sam, and this verse exemplifies their use: the first stock phrase is καὶ ἐγενήθη, and the second is καὶ ἐγένετο. These phrases are commonly identified as Hebraistic features of the Septuagint,57 and the Sahidic translation of these phrases is noteworthy.58 The phrase καὶ ἐγενήθη occurs 36 times in 1 Sam. In 27 occurrences, the phrase is used impersonally with an expression of time. In 20 of these 27 occurrences, the equivalent in Sahidic is ⲁⲥϣⲱⲡⲉ ⲇⲉ ‘and it happened’ as in this example. This construction is understandable in Coptic, but a more sophisticated construction appears in those cases where the verb ⲁⲥϣⲱⲡⲉ is omitted as redundant and only a connective and a suitable form of the next verb occur in the Sahidic reading.59 However, the formally literal equivalent for this phrase is ⲁⲥϣⲱⲡⲉ ⲇⲉ, and only occasionally has the translator omitted the redundant verb. Another interesting feature concerning this expression is the use of the apodotic conjunction. Among the mentioned 20 cases, there is only one passage in Greek where an apodotic καί-conjunction is not used.60 In the Sahidic translation, one finds the opposite pattern—the apodotic conjunction is not present. 14: 18–19

ὅτι αὐτὸς ἦρεν τὸ εφοὺδ ἐν τῇ ἡμέρᾳ ἐκείνῃ ἐνώπιον Ἰσραήλ.

καὶ

56 20: 24, 23: 5, 30: 16 and 30: 18. 57 See Aejmelaeus, Parataxis, 24–6. 58 Despite the examples given below, these clauses are grouped and counted like the other clauses as well, according to the connectives used in Sahidic. 59 These cases are: 7: 2, 14: 19, 16: 23, 25: 20. In verse 30: 25, there seems to be a homoiarchon mistake, and, thus, this verse might not belong to this group. However, this passage does not have the redundant verb. In 25: 2, the rendering is an ⲁⲩⲱ-clause. The remaining case (1: 4) appears as an example below. 60 In verse 1: 12.

52

Description of the translation technique

ἐγενήθη ὡς ἐλάλει Σαοὺλ πρὸς τὸν ἱερέα, καὶ ὁ ἦχος ἐν τῇ παρεμβολῇ τῶν ἀλλοφύλων ἐπορεύετο…

ϫⲉ ⲛⲧⲟϥ ⲛⲉϥϥⲓ̈ ⲙⲡⲉⲫⲟⲩⲇ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ ϩⲓⲑⲏ ⲙⲡⲓⲏⲗ̄ ·̄ ⲉⲣⲉ ⲥⲁⲟⲩⲗ ⲇⲉ ϣⲁϫⲉ ⲙⲛ̄ ⲡⲟⲩⲏⲏⲃ· ⲟⲩⲛⲟϭ ⲛϩⲣⲟⲟⲩ̄ ⲉⲙⲁⲧⲉ ̄ ⲁϥϣⲱⲡⲉ ϩⲛ ⲧⲡⲁⲣⲉⲙⲃⲟⲗⲏ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· In this case, the equivalent for the redundant verb ἐγενήθη has been omitted from the translation.61 The Sahidic translator formulated the clause with the connective ⲇⲉ and circumstantial form of the next verb ‘to speak’. 5: 10 SaM

καὶ ἐξαποστέλλουσιν τὴν κιβωτὸν τοῦ θεοῦ εἰς Ἀσκάλωνα. καὶ ἐγενήθη ὡς εἰσῆλθεν κιβωτὸς θεοῦ εἰς Ἀσκάλωνα, καὶ ἐβόησαν οἱ Ἀσκαλωνεῖται…

ⲁⲩⲱ ⲁⲩϫⲟⲟⲩ ⲛⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡⲛⲟⲩⲧⲉ ⲉϩⲣⲁⲓ ⲉⲁⲥⲕⲁⲗⲱⲛ. ⲛⲣⲱⲙⲉ ⲇⲉ ⲛⲁⲥⲕⲁⲗⲱⲛ ⲁⲩⲱϣ ⲉⲃⲟⲗ … ⲁⲩⲱ ⲁⲩϫⲟⲟⲩ ⲛⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡⲛⲟⲩⲧⲉ ⲙⲡⲓⲏⲗ̄ ̄ ⲉϩⲣⲁⲓ ⲉⲁⲥⲕⲁⲗⲱⲛ· ⲁⲥϣⲱⲡⲉ SaS ⲇⲉ ⲛⲧ ̄ ⲉⲣⲉ ⲧⲕⲓⲃⲱⲧⲟⲥ ⲙⲡⲛⲟⲩⲧⲉ ⲃⲱⲕ ⲉϩⲟⲩⲛ ⲉⲁⲥⲕⲁⲗⲱⲛ ⲛⲣⲱⲙⲉ ⲛⲁⲥⲕⲁⲗⲱⲛ ⲁⲩⲱϣ ⲉⲃⲟⲗ … Above, the clause occurs only in SaS, whereas SaM has excluded it. In this rare instance, SaM is apparently corrupt,62 whereas another Sahidic manuscript preserves this passage. Here, SaM reads the connective ⲇⲉ at the beginning of the apodotic clause in Greek. However, in SaM the conjunction is not apodotic since the clause construction is different as a result of the omission (‘And they sent the ark of God to Askalon. And the people of Askalon cried out…’); there are two main clauses in SaM, the first connected with ⲁⲩⲱ and the second with ⲇⲉ. 1: 4

καὶ ἐγενήθη ἡμέρᾳ καὶ ἔθυσεν Ἐλκανά,…

ⲡⲧⲱϣ ⲇⲉ ⲙⲡⲉϩⲟⲟⲩ ⲛⲧⲉⲣⲉϥϣⲱⲡⲉ· ⲉⲗⲅⲁⲛⲁ ⲁϥϣⲱⲱⲧ ⲛⲧⲉϥⲑⲩⲥⲓⲁ In this case, the Sahidic translator constructed the first clause with an extra word (ⲡⲧⲱϣ ‘manner, fashion, affair, ordinance, condition’). Crum cites ⲡⲧⲱϣ as an explanatory or supplementary word in his dictionary, offering this case and others as examples.63 The phrase καὶ ἐγένετο occurs 15 times in my material. In 10 of those cases, the clause contains an element of time definition and another verb, and, thus, the verb ἐγένετο is actually redundant. In nine cases, the Greek tradition offers an apodotic καί-conjunction, but only once does the Sahidic translation preserve a formally literal parallel. Seven times,64 the Sahidic 61 In my grouping, this case belongs to subgroup 2.1.3.1.f: “καί-clause without equivalent”. 62 A homoioteleuton mistake, from ⲉⲁⲥⲕⲁⲗⲱⲛ to ⲉⲁⲥⲕⲁⲗⲱⲛ, has obviously happened in the copying. 63 Crum, Dictionary, 451. 64 5: 4 (the conjunction ⲁⲩⲱ), 9: 26, 13: 10, 23: 6, 24: 17, 25: 38. In verse 25: 37, the construction is also different; see the example above. The remaining three cases with the redundant verb are in verses 3: 2, 4: 18 and 8: 1.

Clause connections

53

translator has omitted the redundant verb and thus rendered the clause idiomatically. 9: 26

καὶ ἐκοιμήθη. καὶ ἐγένετο65 ὡς ἀνέβαινεν ὁ ὄρθρος, καὶ ἐκάλεσεν Σαμουὴλ τὸν Σαοὺλ…

ⲁϥⲉⲛⲕⲟⲧⲕ·̄ ⲡⲟⲩⲟⲉⲓⲛ ⲇⲉ ⲛⲧⲉⲣⲉϥⲥⲱⲣ· ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲁϥⲙⲟⲩⲧⲉ ⲉ ⲥⲁⲟⲩⲗ In this example, the translator has formulated a natural Sahidic clause. As the equivalent for καὶ ἐγένετο, only ⲇⲉ is used here. After this idiomatic rendering, an equivalent for apodotic καί surprisingly appears in the Sahidic translation—namely ⲇⲉ. Usually, these apodotic conjunctions do not occur in the Sahidic text of 1 Sam. 4: 17–18 …καὶ ἡ κιβωτὸς τοῦ θεοῦ ἐλήμφθη. καὶ ἐγένετο ὡς ἐμνήσθη τῆς κιβωτοῦ τοῦ θεοῦ, καὶ ἔπεσεν… ⲧⲕⲟ̅ⲓⲃⲱⲧⲟⲥ ⲙⲡϫⲟⲉⲓⲥ ⲁⲩϥⲓⲧⲥ̄· ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛⲧ ̅ ̄ ⲉⲣⲉϥⲣ ̄ ⲡⲙⲉⲉⲩⲉ ⲛⲧ ̄ ⲕⲟ̅ⲓⲃⲱⲧⲟⲥ ̅ ⲙⲡϫⲟⲉⲓⲥ· ⲁϥϩⲉ… Above, the translator has rendered the redundant verb. The apodotic καί has no equivalent in the Sahidic translation. Text-critical note: There are no Greek variants reading κύριος instead of θεοῦ even though Sahidic reads ⲡϫⲟⲉⲓⲥ ‘Lord’, the normal equivalent for κύριος. 25: 37 καὶ ἐγένετο πρωί ὡς ἐξένηψεν ἀπὸ τοῦ οἴνου Ναβάλ, ἀπήγγειλεν αὐτῷ ἡ γυνὴ αὐτοῦ τὰ ῥήματα ταῦτα,… ϩⲧⲟⲟⲩⲉ ⲇⲉ ⲛⲧ ̄ ⲉⲣⲉϥϣⲱⲡⲉ ⲁϥⲛⲏⲫⲉ ̄ ϩⲁ ⲡⲏⲣⲡ·̄ ⲁⲩⲱ ⲧⲉϥⲥϩ̄ ⲓⲙⲉ ⲁⲥⲧⲁⲩⲟ ⲉⲣⲟϥ ⲛⲛⲉϣⲁϫⲉ· In this example, the Sahidic construction differs slightly from the text of the Greek Vorlage. The Sahidic reads, ‘When the morning came, he sobered up from the wine, and his wife told him the words.’ A dependent clause with a temporal verbal form precedes an asyndetic main clause. The third clause begins a new main clause in Sahidic with the connective ⲁⲩⲱ. In Greek, there are two dependent clauses with a main clause after them. There is no apodotic conjunction in Greek, against the usual habit in 1 Sam, but only an asyndetic connection. 2.1.3.1.e) καί-clause rendered with other connectives Besides asyndeton, ⲁⲩⲱ and ⲇⲉ, other conjunctions are used to render καί. When compared with the three most common renderings, the numbers of different connectives are small—altogether there are 90 cases. Inside this 65 Greek manuscripts 44 and 125 omit ἐγένετο here, but there are no reasons to suppose that Sahidic is here dependent on the textual tradition of these manuscripts which often omit some words.

54

Description of the translation technique

group, a subgroup exists of 35 cases where the word pair καὶ νῦν is rendered with ⲧⲉⲛⲟⲩ ϭⲉ. In 55 cases, there are other connectives rendering a καίclause. These cases often demonstrate how the translator understood the Greek text, since, in most cases, the connective bears a more specific meaning than any of the three most common renderings of καί. I present the examples in order of frequency in the chart below. καὶ νῦν καί καί καί καί καί καί καί καί καί καί

ⲧⲉⲛⲟⲩ ϭⲉ ϩⲱ⸗, ϩⲱⲱϥ ϭⲉ ⲟⲛ ⲁⲗⲗⲁ ϫⲉ ⲙⲛ ⲟⲩⲇⲉ ⲏ ⲙⲛⲛⲥⲱⲥ ϫⲉⲕⲁⲥ

35 cases 20 cases 11 cases 7 cases 6 cases 4 cases twice twice once once once

The καὶ νῦν word pair is usually rendered with ⲧⲉⲛⲟⲩ ϭⲉ.66 There are 35 such cases and one case where ⲧⲉⲛⲟⲩ ϭⲉ renders a plain καί in Greek.67 In Greek 1 Sam, there are altogether 37 καὶ νῦν cases, and, thus, one can conclude that this phrase was rendered concordantly. The two cases where ⲧⲉⲛⲟⲩ ϭⲉ is not used to render καὶ νῦν appear in verses 23: 11 and 28: 15. In verse 23: 11, Sahidic reads ⲁⲩⲱ ⲧⲉⲛⲟⲩ, in 28: 15 only ⲧⲉⲛⲟⲩ.68 8: 8–9 …οὕτως αὐτοὶ ποιοῦσιν καὶ σοί. καὶ νῦν ἄκουε τῆς φωνῆς αὐτῶν·

ⲧⲁⲓ ⲧⲉ ⲑⲉ ⲉⲧⲟⲩⲉⲓⲣⲉ ⲙⲙⲟⲥ ⲛⲁⲕ ϩⲱⲱⲕ· ⲧⲉⲛⲟⲩ ϭⲉ ⲥⲱⲧⲙ̄ ⲛⲥⲁ ⲡⲉⲩϩⲣⲟⲟⲩ This is a typical example; all the cases with καὶ νῦν occur within direct discourse. One possible reason for the popularity of ⲧⲉⲛⲟⲩ ϭⲉ as the rendering of καὶ νῦν might be the formal similarity of the structure; these two Coptic words are able to mirror their Greek counterparts both in number and in their initial clausal position. 28: 15 καὶ οὐκ ἐπακήκοέν μοι ἔτι καὶ ἐν χειρὶ τῶν προφητῶν καὶ ἐν τοῖς ἐνυπνίοις· καὶ νῦν κέκληκά σε γνωρίσαι μοι τί ποιήσω. ⲙⲡⲉϥⲕⲟⲧϥ ϭⲉ ⲉⲥⲱⲧⲙ̄ ⲉⲣⲟⲓ· ⲟⲩⲇⲉ ϩⲛ̄ ⲧⲟⲟⲧⲟⲩ ⲛⲛⲉⲡⲣⲟⲫⲏⲧⲏⲥ ⲟⲩⲇⲉ ϩⲛ̄ ⲣⲁⲥⲟⲩ· ⲉⲧⲃⲉ ⲡⲁⲓ ⲧⲉⲛⲟⲩ ⲁⲓⲙⲟⲩⲧⲉ ⲉⲣⲟⲕ ⲉⲧⲣⲉⲕⲧⲟⲩⲛⲓⲁⲧ ⲉⲃⲟⲗ ⲙⲡⲉϯⲛⲁⲁⲁϥ

66 For a description of ϭⲉ, see example 27: 4 below. 67 In verse 25: 8. 68 Then there are two οὖν-conjunctions, in verses 19: 2 and 20: 31, that are rendered with ⲧⲉⲛⲟⲩ ϭⲉ. See the examples under οὖν.

Clause connections

55

Here, the Sahidic translation has a few words more than the Greek,69 and thus, in Sahidic, Saul says explicitly ‘because of this’ at the beginning of this sentence. This passage is direct speech, which often contains free renderings and minor additions, probably in order to make the narrative sound natural in the ears of the translator and his audience. There are 20 clauses where ϩⲱⲱ⸗ is used to render a καί-clause, and in 13 of these cases the word has a suffix other than 3rd person masc. singular.70 According to Crum, this word can be used in two ways: either with a suffix according to the subject of the clause or in the form ϩⲱⲱϥ regardless of the subject. This pronoun can be understood either as emphatic ‘self, also, for his part’ or as expressing contrast ‘but, on the other hand’.71 Layton lists only the indeclinable form ϩⲱⲱϥ among the conjunctions.72 Müller writes from the perspective of diachronic grammar, and according to him this word was first agreement-sensitive but later grammaticalized as a 3rd person singular form.73 28: 4

καὶ συναθροίζει Σαοὺλ πάντα ἄνδρα Ἰσραήλ,… ⲥⲁⲟⲩⲗ ϩⲱⲱϥ ⲁϥⲥⲱⲟⲩϩ ⲉϩⲟⲩⲛ ⲙⲛ ⲡⲓⲏⲗ̄ ̄ ⲧⲏⲣϥ ̄ This is a typical example: ϩⲱⲱϥ is used in a context where one could as well find the particle ⲇⲉ.

23: 17 καὶ σὺ βασιλεύσεις ἐπὶ Ἰσραὴλ καὶ ἐγὼ ἔσομαί σοι εἰς δεύτερον· ⲁⲗⲗⲁ ⲛⲧⲟⲕ ⲡⲉⲧⲛⲁⲣⲣ̄ ⲣⲟ̄ ⲉϫⲙ̄ ⲡⲓⲏⲗ̄ ·̄ ⲁⲛⲟⲕ ϩⲱ ⲛⲧⲁϣⲱⲡⲉ ⲛⲁⲕ ⲙⲙⲉϩⲥ̄ⲛⲁⲩ ϩⲓⲧⲟⲩⲱⲕ· In this example, the conjunction uses the 1st person singular form. Again, one could also find the particle ⲇⲉ here, since it often appears in cases where the subject changes. 27: 4

καὶ ἀνηγγέλη τῷ Σαούλ ὅτι πέφευγεν Δαυεὶδ εἰς Γέθ· καὶ οὐ προσέθετο ἔτι ζητεῖν αὐτόν.

ⲁⲩϫⲓ ⲡⲟⲩⲱ̄ ⲇⲉ ⲛⲥⲁⲟⲩⲗ ϫⲉ ⲁ ⲇⲁⲩⲉⲓⲇ ⲡⲱⲧ ⲉϩⲣⲁⲓ ⲉⲅⲉⲑ ⲙⲡⲉϥⲟⲩⲱϩ ϭⲉ ⲉⲧⲟⲟⲧϥ ⲉⲃⲱⲕ ⲉϯⲟⲩⲟⲓ ⲛⲥⲱϥ There are 11 cases where the translator has rendered a καί-conjunction with ϭⲉ. One essential difference between these conjunctions is the position; ϭⲉ is

69 There are no Greek variants. 70 There are six cases (8: 20, 14: 22, 19: 20, 19: 21, 19: 22, 31: 5) where ϩⲱⲱ⸗ is used as an emphasizing pronoun to render καὶ αὐτός/αὐτοί/ἡμεῖς. In these cases, the conjunction καί appears first at the beginning of the clause, followed by a verb and, finally, the adverbial expression. These cases belong to my subgroup 2.1.3.1.b: ‘καί-clause rendered asyndetically’ since there is no rendering for the first, connecting καί. 71 Crum, Dictionary, 651–2. 72 Layton, Grammar, §235b. 73 Müller, “Greek Connectors”, 9.

56

Description of the translation technique

enclitic, whereas καί is not. The indigenous Coptic word ϭⲉ is described by Wolfhart Westendorf as an enclitic particle. The Demotic equivalent has the semantic range ‘also, more, still, further’.74 Crum gives two meanings to the word ϭⲉ: ‘then, therefore, but’ and ‘again, once more’. The latter one is typically found in a negative phrase. Thus, it is a perfect match to Greek καί οὐκ ἔτι. In this example, ϭⲉ is an accurate rendering and suits the context well.75 3: 6

καὶ προσέθετο Κύριος καὶ ἐκάλεσεν Σαμουὴλ Σαμουήλ· καὶ ἐπορεύθη πρὸς Ηλεὶ τὸ δεύτερον…

ⲡϫⲟⲉⲓⲥ ⲇⲉ ⲁϥⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ ̄ ⲛⲕⲉⲥⲟⲡ ⲉⲙⲟⲩⲧⲉ· ϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲥⲁⲙⲟⲩⲏⲗ· ⲁϥⲡⲱⲧ ⲟⲛ ⲉⲣⲁⲧϥ̄ ⲛϩⲏⲗⲉⲓ̅ ·̄ ⲙⲡⲙⲉϩⲥⲉⲡ ⲥⲛⲁⲩ Here, the equivalent for καί is ⲟⲛ ‘again, also, still, further’, which is used seven times to render καί. In this and related instances, ⲟⲛ functions well as both a formal and functional parallel to Greek οὖν. Six times, καί is rendered with ⲁⲗⲗⲁ in the Sahidic.76 Four of the cases appear in direct speech. In five cases, ⲁⲗⲗⲁ begins an affirmative clause after an emphatic negation. 8: 3

καὶ οὐκ ἐπορεύθησαν οἱ υἱοὶ αὐτοῦ ἐν ὁδῷ αὐτοῦ, καὶ ἐξέκλιναν ὀπίσω τῆς συντελείας·

ⲛⲉϥϣⲏⲣⲉ ⲇⲉ ⲙⲡⲟⲩⲙⲟⲟϣⲉ ⲕⲁⲧⲁ ⲧⲉϥⲟⲑⲉ· ⲁⲗⲗⲁ ⲁⲩⲣⲟⲕⲧⲟⲩ ⲉⲃⲟⲗ ϩⲛ̄ ⲧⲙⲉ The Sahidic translator has understood that the boys turn away from the completion. The Vorlage of the translator probably had the mistaken Greek rendering συντελείας ‘completion’ and not the approximation πλεονεξίας ‘greediness’ since the object from which the sons recoil is unambiguously positive ‘truth’ in the Sahidic translation.77 30: 21 καὶ παραγίνεται Δαυεὶδ πρὸς τοὺς διακοσίους ἄνδρας τοὺς ἐκλυθέντας τοῦ πορεύεσθαι ὀπίσω Δαυεὶδ, καὶ ἐκάθισεν αὐτοὺς ἐν τῷ χειμάρρῳ τῷ Βοσορ,… Βοσορ] βεανα B, βαιανα 121, βεενα βοσορ 509) SaM ⲁⲩⲱ ⲇⲁⲩⲉⲓⲇ ⲁϥⲉⲓ ⲉⲣⲁⲧⲟⲩ ⲙⲡϣⲏⲧ ⲛⲣⲱⲙⲉ· ⲉⲧⲙ̄ⲡⲟⲩϭⲙϭⲟⲙ ⲉⲟⲩⲁϩⲟⲩ ⲛⲥⲁ ⲇ̅ⲁⲇ ̅ ̅ · ⲁⲗⲗⲁ ⲁϥⲑⲙ̄ⲥⲟⲟⲩ· ϩⲙ̄ ⲡⲉⲭⲓⲙⲁⲣⲟⲥ ⲛⲃⲉⲁⲛⲁ ⲁⲩⲱ ⲁ ⲇ̅ⲁⲇ SaA ̅ ̅ ⲉⲓ ⲉⲣⲁⲧⲟⲩ ⲙⲡϣⲏⲧ ⲛⲣⲱⲙⲉ· ⲉⲧⲉ ⲙⲡⲟⲩ ϣ̄ ϭⲙϭⲟⲙ ⲉⲟⲩⲁϩⲟⲩ ⲛⲥⲱϥ· ⲁⲩⲱ ⲁϥⲑⲙⲥⲟⲟⲩ ϩⲙ ⲡⲉⲭⲓⲙⲁⲣⲣⲟⲥ ⲛⲃⲉⲛⲉⲁ

74 Westendorf, Handwörterbuch, 480. 75 A similar rendering is found in 28: 15, see the example above. 76 8: 3; 12: 20; 12: 23; 23: 17; 24: 11 and 30: 21. For an example from Jeremiah see Feder, Biblia Sahidica, 93. 77 According to Brock (Recensions, 151) it is probably Hexaplaric. The word πλεονεξίας is found in 82–93–108mg-127 55 318 554txt. Both words are found in 107´ 106 158.

Clause connections

57

There are no Greek variant readings for ἀλλά. If there is revision, then it seems to be the case here that SaA is revised according to the Greek, and that, conversely, SaM offers a free rendering. However, SaA and SaM may differ because of some small-scale changes that happened during copying. The place name is ⲃⲉⲁⲛⲁ in SaM and ⲃⲉⲛⲉⲁ in SaA, and this is one small argument that connects the Sahidic translation with the Greek B-group in a secondary reading. Payne, in his dissertation, lists dozens of such passages that, according to him, show this dependence.78 The problem in his lists is that he counts all the passages where Sahidic and B-group read the same text, not taking into account whether the reading is original or secondary in Greek. I presume that passages in which the Sahidic translation and the Greek B-text both read without a Hexaplaric approximation are not necessarily related, but both texts have preserved the OG independently of each other. Only secondary readings in Greek unanimously show dependence of the Sahidic version. 1: 22

Ἕως τοῦ ἀναβῆναι τὸ παιδάριον, ἐὰν ἀπογαλακτίσω αὐτό, καὶ ὀφθήσεται τῷ προσώπῳ Κυρίου,…

ϯⲛⲁϭⲱ ϣⲁⲛⲧⲉ ⲡϣⲏⲣⲉ ⲕⲟⲩⲓ̈ ϫⲁⲧⲉ· ⲛⲧⲁⲟⲙϫϥ̄ ϫⲉ ⲉϥⲉⲟⲩⲱⲛϩ̄ ⲉⲃⲟⲗ ϩⲓⲑⲏ ⲙⲡϫⲟⲉⲓⲥ There are four cases where a ϫⲉ-clause is used to render a καί-clause. In this example, καί with a future verb is rendered by ϫⲉ and an optative in Sahidic. This is an idiomatic rendering in this context; ϫⲉ begins a clause expressing purpose here.79 A similar case is found in verse 11: 2 where the conjunction in Sahidic is ϫⲉⲕⲁⲥ. 2: 7

Κύριος πτωχίζει καὶ πλουτίζει, ταπεινοῖ καὶ ἀνυψοῖ. ⲡϫⲟⲉⲓⲥ ⲡⲉⲧϯ ⲛⲧ ̄ ⲙⲛⲧ ̄ ϩⲏⲕⲉ· ⲙⲛ ⲧⲙⲉⲧⲣⲙⲙⲁⲟ· ⲉϥϯ ⲙⲡⲉⲑⲃⲃⲓⲟ·̅ ⲙⲛ ⲡϫⲓⲥⲉ Here, the construction is different: in Greek the coordinations are between verbs, whereas, in Sahidic, the translator has coordinated the nouns. This explains why the translator has used ⲙⲛ ‘and, with’, which usually conjoins words, and not clauses or sentences.

30: 12 ὅτι οὐ βεβρώκει ἄρτον καὶ οὐ πεπώκει ὕδωρ τρεῖς ἡμέρας καὶ τρεῖς νύκτας

ⲉⲃⲟⲗ ϫⲉ ⲁϥⲣ ̄ ϣⲟⲙⲧ ⲛϩⲟⲟⲩ· ⲙⲛ̄ ϣⲟⲙⲧⲉ ⲛⲟⲩϣⲏ· ⲙⲡⲉϥⲟⲩⲉⲙ ⲟⲉⲓⲕ ⲟⲩⲇⲉ ⲙⲡⲉϥⲥⲉ ⲙⲟⲟⲩ· Twice the rendering of a καί-clause begins with ⲟⲩⲇⲉ in Sahidic.80 However, in both cases, there is a variant reading with οὐδέ in the Greek.81 As a result, 78 Payne, Critical and comparative, 141–53 “strong cases of textual relationship” and 153–7 “relatively weak cases of textual relationship”. 79 See also Müller, “Greek Connectors”, paragraph 4.2.6.2. 80 The other case is in verse 25: 7. 81 In verse 25: 7 οὐδέ is found in L, and in 30: 12 it is found in 707.

58

Description of the translation technique

a mild reservation remains as to whether the Sahidic translator actually rendered two καί-clauses with ⲟⲩⲇⲉ-clauses. In this example, there is also a discrepancy in the time definition, but the Greek tradition preserves no variants corresponding to these Sahidic readings. 26: 12 καὶ οὐκ ἦν ὁ βλέπων καὶ οὐκ ἦν ὁ γινώσκων καὶ οὐκ ἦν ὁ ἐξεγειρόμενος·

ⲛⲉⲙⲛ̄ ⲗⲁⲁⲩ ⲇⲉ ⲓⲱⲣϩ̄ ̄ ⲙⲙⲟⲟⲩ· ⲡⲉ· ⲁⲩⲱ ⲛⲉⲙⲡⲉ ⲗⲁⲁⲩ ⲉⲓⲙⲉ ⲏ ⲁϥⲛⲉϩⲥⲉ· This is the only passage where a καί-clause is rendered with the GreekCoptic loanword ⲏ. This translation fits the context in which one encounters a list of occurrences that did not come to pass. 15: 33–4

καὶ ἔσφαξεν Σαμουὴλ τὸν Ἀγὰγ ἐνώπιον Κυρίου ἐν Γαλγάλ. καὶ ἀπῆλθεν Σαμουὴλ εἰς Ἁρμαθάιμ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲕⲱⲛⲥ̄ ⲛⲁⲅⲁⲅ ⲙⲡⲉⲙⲧⲟ ⲉⲃⲟⲗ ⲙⲡϫⲟⲉⲓⲥ ϩⲛ̄ ⲅⲁⲗⲅⲁⲗⲁ· ⲙⲛⲛ ̄ ⲥ̄ ⲱⲥ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲃⲱⲕ ⲉⲁⲣⲙⲁⲑⲁⲉⲓⲙ· In this case there is a transition to a new phase in the story, and the connective ⲙⲛⲛ ̄ ⲥ̄ ⲱⲥ ‘after that’ is an appropriate interpretation of the function of καί here. 2.1.3.1.f) καί-clause without equivalent There are cases where a καί-clause has no equivalent in the Sahidic text.82 I have grouped the cases based on the reason for the absence of the clause. There are three subgroups: the first group consists of the passages where the translator has omitted a redundant verb, the second group contains those cases where the source text was different and the third group contains mistranslations as well as copying errors. In the first group, there are 11 cases where the translator has left a redundant verb without equivalent. In these passages, the Greek text has a verb such as ἔσται, ἐγενήθη, or ἐγένετο as well as a temporal expression. I present here one example of each verb. 10: 5

καὶ ἔσται ὡς ἂν εἰσέλθητε ἐκεῖ εἰς τὴν πόλιν, καὶ ἀπαντήσεις χορῷ προφητῶν…

ⲁⲩⲱ ⲉⲧⲉⲧⲛϣ ̄ ⲁⲛⲃⲱⲕ ⲉϩⲟⲩⲛ ⲉⲧⲡⲟⲗⲓⲥ· ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙ ̄ ̄ ̄ⲙⲁⲩ· ⲕⲛⲁⲧⲱⲙⲛⲧ ⲉⲩⲭⲟⲣⲟⲥ ⲙⲡⲣⲟⲫⲏⲧⲏⲥ· In this case the redundant verb ἔσται is left out in the translation.83 There is only the conjunction ⲁⲩⲱ and after that the 2nd person plural of a conditional verbal form. Sahidic does not render the apodotic conjunction after the conditional clause. 82 In addition, there are a few cases where it is difficult to say whether the omission results from the translator or from a difference in the Vorlage. 83 Other passages of this kind are in verses 3: 9, 17: 25, 23: 23 and 25: 30.

Clause connections

59

14: 19 καὶ ἐγενήθη ὡς ἐλάλει Σαοὺλ πρὸς τὸν ἱερέα, καὶ ὁ ἦχος ἐν τῇ παρεμβολῇ τῶν ἀλλοφύλων ἐπορεύετο πορευόμενος… ⲉⲣⲉ ⲥⲁⲟⲩⲗ ⲇⲉ ϣⲁϫⲉ ⲙⲛ̄ ⲡⲟⲩⲏⲏⲃ· ⲟⲩⲛⲟϭ ⲛϩⲣⲟⲟⲩ̄ ⲉⲙⲁⲧⲉ ̄ ⲁϥϣⲱⲡⲉ ϩⲛ̄ ⲧⲡⲁⲣⲉⲙⲃⲟⲗⲏ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· In this passage, the redundant verb ἐγενήθη has no Sahidic equivalent,84 although the Greek connective ⲇⲉ does appear. The Sahidic translator has used a circumstantial verbal form to render the Greek temporal ὡς-clause. Again, there is no equivalent for the apodotic conjunction in Sahidic. 13: 10 καὶ ἐγένετο ὡς συνετέλεσεν ἀναφέρων τὴν ὁλοκαύτωσιν, καὶ Σαμουὴλ παραγίνεται·

ⲛⲧⲉⲣⲉϥⲟⲩⲱ ⲇⲉ ⲉϥⲧⲁⲗⲟ ⲉϩⲣⲁⲓ ⲙⲡϩⲟⲗⲟⲕⲁⲩⲧⲱⲙⲁ· ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲓ̈· This case resembles the previous example; there is no equivalent for the redundant verb ἐγένετο in the beginning,85 only the connective ⲇⲉ. The verbal form in Sahidic is temporal, an idiomatic translation of a temporal ὡςclause. The second group contains those cases where the omission in Sahidic probably results from a different source text. Thus, in these passages the translator has not omitted anything, but has accurately rendered his Vorlage. There are six clear cases.86 3: 17

μὴ δὴ κρύψῃς ἀπ᾽ ἐμοῦ· τάδε ποιήσαι σοι ὁ θεὸς, καὶ τάδε προσθείη, ἐὰν κρύψῃς ἀπ᾽ ἐμοῦ ῥῆμα ἐκ πάντων τῶν λόγων…

ⲙⲡⲉⲣϩⲱⲡ ⲉⲣⲟⲓ· ⲉⲣⲉ ⲡⲛⲟⲩⲧⲉ ⲉⲓⲣⲉ ⲛⲁⲕ ⲉⲕϣⲁⲛϩⲉⲡ ⲟⲩϣⲁϫⲉ ⲉⲣⲟⲓ· ⲉⲃⲟⲗ ϩⲛ ⲛⲉϣⲁϫⲉ ⲧⲏⲣⲟⲩ The underlined words are missing in B b as well as in the Sahidic. However, SaM also preserves no rendering for τάδε. The Sahidic translator probably had a Vorlage similar to the B-group in this passage. Additionally, a skip of the eye has happened in the copying of SaM (‘Do not hide from me. May God do to you if you hide anything from me of all the words’) or one of its predecessors, since SaG has an equivalent for τάδε: ⲉⲣⲉ ⲡⲛⲟⲩⲧⲉ ⲣ ̄ ⲛⲁⲓ ⲛⲁⲕ ‘may God do these to you’. 16: 16 καὶ ἀγαθόν σοι ἔσται καὶ ἀναπαύσει σε. ⲛϥϯⲉⲙⲧⲟⲛ ⲛⲁⲕ In this case, a homoiarchon mistake from the first καὶ ἀ- to the second one may explain the omission. However, the Greek Vorlage of the Sahidic translator may not have contained the doublet attested in extant Greek manu84 Another similar passage lies in verse 25: 20. In verse 30: 25, one could also see this kind of reformulation, but there is more missing in this verse, and therefore I see homoioteleuton as a better explanation. See the example below. 85 Other cases occur in verses 5: 4, 24: 17 and 25: 38. 86 2: 26, 3: 17, 14: 1, 16: 16, 17: 9, 17: 17.

60

Description of the translation technique

scripts. Because the Hebrew text does not have a parallel for καὶ ἀναπαύσει σε, one could conjecture that the original Greek translation was καὶ ἀναπαύσει σε ‘it will give you respite’ preserved in Sahidic. As is usually the case with doublets of this kind, the clause καὶ ἀγαθόν σοι ἔσται more closely renders the MT and appears before the OG reading, exactly at the place where the copyist noticed that something is missing compared to the Hebrew source text.87 This latter explanation is problematic, because all Greek manuscripts preserve both clauses—only the Sahidic and Ethiopic traditions omit the first clause. The third group consists of those cases where the Sahidic text reads an error or a reformulation. In some cases, it is impossible to say whether the discrepancy results from intentional or unintentional activity. There are also cases where one can explain the discrepancy either as resulting from a different Vorlage or as a reformulation without connection to Greek variants. There are altogether 26 passages in the third group, and below I present examples of different cases in this group. In ten cases88, the omission is most convincingly explained with a homoioteleuton or homoiarchon mistake. 4: 13

καὶ ἦλθεν, καὶ ἰδοὺ Ἠλεὶ ἐκάθητο ἐπὶ τοῦ δίφρου…

ϩⲏⲗⲉ̅ⲓ ̅ ⲇⲉ ⲛⲉϥϩⲙⲟⲟⲥ ⲡⲉ ϩⲓϫⲛ ⲧⲉϥⲧⲁϭⲥ… In this case, it seems clear that a homoiarchon mistake was made: Sahidic offers as an equivalent for the first καί the particle ⲇⲉ, and, thus, the omitted words are exactly the ones that lie between the two ηλ- beginnings.89 This mistake occurred within the Greek tradition since the same logic does not apply to the Sahidic text. Whether this results from the Sahidic translator misreading his Vorlage or rendering an already erroneously copied Greek text is impossible to know. 9: 4 SaM

καὶ διῆλθον διὰ τῆς γῆς Σελχά, καὶ οὐχ εὗρον· καὶ διῆλθον διὰ τῆς γῆς Ἐασακέμ, καὶ οὐκ ἦν· καὶ διῆλθον διὰ τῆς γῆς Ἰακείμ, καὶ οὐχ εὗρον.

ⲁⲩⲱ ⲟⲛ ⲁⲩⲃⲱⲕ ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲁϩ ⲛⲥⲉⲗⲭⲁ· ⲙⲡⲟⲩϩⲉ ⲉⲣⲟⲟⲩ· ⲁⲩⲱ ⲟⲛ ⲁⲩⲃⲱⲕ ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲁϩ ⲛⲉⲥⲁⲕⲓⲙ· ⲙⲡⲟⲩϩⲉ ⲉⲣⲟⲟⲩ ⲁⲩⲃⲱⲕ ⲟⲛ ⲉⲃⲟⲗ ϩⲙ ⲡⲧⲟⲟⲩ ⲛⲥⲉⲗⲭⲁ ⲁⲩⲱ ⲉⲙⲡⲟⲩϩⲉ ⲉⲣⲟⲟⲩ· ⲁⲩⲃⲱⲕ ⲟⲛ ⲉⲃⲟⲗ SaA ϩⲙ ⲡⲧⲟⲟⲩ ⲛⲥⲁⲕⲉ ⲁⲩⲱ ⲙⲡⲟⲩϭⲛⲧⲟⲩ Here, the Vorlage of the Sahidic text probably contained three instances of the phrase καὶ οὐχ εὗρον90 and the omission results from a homoioteleuton 87 88 89 90

Aejmelaeus, “What Rahlfs Could not Know”, 87. 4: 13, 7: 14, 9: 4, 14: 19, 18: 10, 18: 17, 25: 12bis, 30: 25 and 31: 13. The earliest manuscripts did not differentiate between lowercase and capital letters. Greek manuscripts A B O b d-4464’ 460 554 read the second instance καὶ οὐκ ἦν.

Clause connections

61

mistake from the second to the third instance. I hold this as the most probable explanation, against Payne who sees this as an example of the translator’s desire for variety.91 Alternatively, one could assume a parablepsis from Εασακεμ to Ιακιμ on the basis of similar endings in these names.92 If one chooses this last alternative, then it is impossible to say whether the Vorlage read καὶ οὐχ εὗρον or καὶ οὐκ ἦν.93 18: 17 μὴ ἔστω χείρ μου ἐπ᾽ αὐτῷ καὶ ἐσται ἐπ᾽ αὐτόν/ ἐν αὐτῷ94 χεὶρ ἀλλοφύλων. ⲛⲛⲉⲧⲁϭⲓϫ ϣⲱⲡⲉ ϩⲓϫⲱϥ In this case, I assume a homoioteleuton mistake in the Sahidic text.95 This clause would have been translated something like ⲁⲩⲱ ⲧϭⲓϫ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϣⲱⲡⲉ ϩⲓϫⲱϥ ‘and the hand of the Philistines will be upon him’, and the previous clause reads ⲛⲛⲉ ⲧⲁϭⲓϫ ϣⲱⲡⲉ ϩⲓϫⲱϥ ‘let not my hand be upon him’. Thus a homoioteleuton-based periblepsis is a convincing explanation in this case. 25: 12 καὶ ἀπεστράφησαν τὰ παιδάρια Δαυεὶδ εἰς ὁδὸν αὐτῶν, καὶ ἀνέστρεψαν καὶ ἦλθον καὶ ἀνήγγειλαν τῷ Δαυεὶδ κατὰ τὰ ῥήματα ταῦτα.

ⲛϩⲙ̄ϩⲁⲗ ⲇⲉ ⲛⲇ ̄ ⲁⲩⲉⲓⲇ ⲁⲩⲕⲟⲧⲟⲩ ⲉⲣⲁⲧϥ ⲛⲇⲁⲩⲉⲓⲇ ⲛⲧ ̄ ⲉⲩϩⲓⲏ· ⲁⲩⲧⲁⲩⲟ ⲉⲣⲟϥ ⲛⲛ ̄ ⲉⲓϣⲁϫⲉ ⲧⲏⲣⲟⲩ· This example shows two missing καί-clauses. One can explain this omission with the help of a homoiarchon mistake within either the Greek or the Sahidic tradition. This same omission also appears in 509 44–107–610 in Greek, but this small omission probably does not demonstrate a real affiliation between Sahidic and the Greek variant readings. On the whole, this passage remains unclear. 30: 25 καὶ ἐγενήθη ἀπὸ τῆς ἡμέρας ἐκείνης καὶ ἐπάνω, καὶ ἐγένετο εἰς πρόσταγμα… ⲁⲩⲱ ⲁ ⲡⲉⲓⲧⲱϣ ϣⲱⲡⲉ ⲉⲩⲡⲣⲟⲥⲧⲁⲅⲙⲁ In this case, I consider a homoiarchon-based mistake as the best explanation for omission from the first to the second instance of καὶ ἐγέν-, bringing about the Sahidic translation ‘And this disposition became an ordinance’. The probability of a translation-technical omission is low since there are five instances in 1 Sam where the translator has rendered the phrase ἀπὸ τῆς ἡμέρας ἐκείνης.96

91 92 93 94 95 96

Payne, Critical and comparative, 38. For instance, ελσακειμ and ιακειμ in 509. Brock (Recensions, 218) assumes that Sahidic read the verb ‘to find’ in this case. Variation between ἐπ᾽ αὐτον/ ἐν αὐτω is no longer perceivable through Sahidic here. Thus also Drescher, Kingdoms, 43. 16: 13, 18: 9, 29: 3, 29: 6, 29: 8.

62 22: 7

Description of the translation technique

καὶ εἶπεν Σαοὺλ πρὸς τοὺς παῖδας αὐτοῦ τοὺς παρεστηκότας αὐτῷ. καὶ εἶπεν αὐτοῖς Ἀκούσατε δή,… καὶ εἶπεν αὐτοῖς (+ σαουλ L -93) A B L -93 b 68’] > rel = MT

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲛⲉϥϩⲙ̄ϩⲁⲗ ⲉⲧⲁϩⲉⲣⲁⲧⲟⲩ ⲙⲡⲉϥⲙ̄ⲧⲟ ⲉⲃⲟⲗ· ϫⲉ ⲥⲱⲧⲙ̄… This case, at first sight, seems clear: the Vorlage of Sahidic read like most of the Greek mss without the underlined clause. This clear-cut explanation is not that self-evident if one takes a closer look at this passage. The Sahidic translation often renders the verb ‘to say’ with only ϫⲉ, as in the next example. Thus, if the translator occasionally used such abbreviated renderings, one cannot exclude this explanation in this case. 10: 12 καὶ ἀπεκρίθη τις αὐτῶν καὶ εἶπεν… ⲁⲩⲱ ⲟⲩⲁ ⲛϩ̄ ⲏⲧⲟⲩ ⲁϥⲟⲩⲱϣⲃ ̄ ϫⲉ In this case, the translator has avoided unnecessary repetition; instead of ‘and said,’ the Sahidic translates only the conjunction ϫⲉ which introduces direct speech. 3: 10

καὶ ἦλθεν Κύριος καὶ κατέστη, καὶ ἐκάλεσεν αὐτὸν… ⲁⲩⲱ ⲡϫⲟⲉⲓⲥ ⲁϥⲉⲓ ϩⲁϫⲱϥ ⲁϥⲙⲟⲩⲧⲉ ⲉⲣⲟϥ This case can be understood either as a free rendering of Greek or as an omission of καὶ κατέστη and an addition of the word ϩⲁϫⲱϥ ‘by his head, by him’. I agree with Payne that, in this case, freedom in translation is the most feasible explanation.97 Perhaps, for theological reasons, the translator chose to avoid saying that God stood somewhere. The first verb, however, is ‘to come’ and this together with the preposition in Sahidic still gives the impression that God came to Samuel.

25: 23 καὶ εἶδεν Ἀβειγαία τὸν Δαυεὶδ καὶ ἔσπευσεν καὶ κατεπήδησεν ἀπὸ τῆς ὄνου,… SaM ⲁⲃⲓⲕⲓⲁ ⲇⲉ̄ ⲁⲥⲛⲁⲩ ⲉⲇⲁⲩⲉⲓⲇ· ⲁⲩⲱ ⲛⲧⲉⲩⲛⲟⲩ ⲁⲥⲃⲟϭⲥ ⲉⲃⲟⲗ ϩⲛ̄ ⲧⲉⲱ· SaA ⲁⲃⲓⲅⲉⲁ ⲇⲉ ⲁⲥⲛⲁⲩ ⲉⲇ̅ⲁⲇ ̅ ̅ · ⲁⲥϫⲓⲃⲟϭⲥ ⲉⲃⲟⲗ ϩⲓϫⲛ̄ ⲧⲉⲥⲉⲓⲱ· In this case the Sahidic rendering for καὶ ἔσπευσεν is ⲛⲧⲉⲩⲛⲟⲩ ‘forthwith’.98 The adverb is omitted in SaA following the overall tendency of this manuscript to preserve a more succinct translation.99 13: 8 SaS

καὶ οὐ παρεγένετο Σαμουὴλ εἰς Γάλγαλα, καὶ διεσπάρη ὁ λαὸς αὐτοῦ ἀπ᾽αὐτοῦ.

ⲙⲡⲉ ⲥⲁⲙⲟⲩⲏⲗ [

97 Payne, Critical and comparative, 40. 98 This same equivalent is used in verses 28: 24 and 28: 20. 99 See sections 3.2 and 3.4 for more detailed descriptions.

Clause connections

63

Although the text of SaM omits this verse, the fragmentary remains of SaS preserve the beginning of the passage: ‘Samuel did not…’. There are no Greek manuscripts with this omission, nor is there any reason for a Sahidic revisor to omit these words. Therefore, this passage probably once was in the Sahidic version, but some error was made in the copying of SaM or in its prehistory. 17: 1

καὶ συνάγονται εἰς Σοκχὼθ τῆς Ἰδουμαίας, καὶ παρεμβάλλουσιν ἀνὰ μέσον Σοκχὼθ…

ⲁⲩⲱ ⲁⲩⲉⲓ̈ ⲉϩⲣⲁⲓ̈ ⲉⲁ̅ ⲥⲥⲟⲭⲟⲑ ⲛⲧ ̄ ⲉ ϯⲇⲟⲩⲙⲁⲓ̈ⲁ· ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛⲁⲥⲥⲟⲭⲱⲑ In this verse, the equivalent for παρεμβάλλουσιν is missing. Its equivalent in verses 4: 1 and 13: 5 is ⲁⲩⲧⲁⲁⲩ ⲉϩⲣⲁⲓ ⲉ- ‘They betook themselves to’. Either the translator avoided repetition (ⲁⲩⲉⲓ̈ ⲉϩⲣⲁⲓ̈ ⲉ-̅ in the first and in the next clause ⲁⲩⲧⲁⲁⲩ ⲉϩⲣⲁⲓ ⲉ-) or a simple mistake occurred during copying. 28: 4

καὶ συναθροίζει Σαοὺλ πάντα ἄνδρα Ἰσραήλ, καὶ παρεμβάλλουσιν εἰς Γελβοῦε.

ⲥⲁⲟⲩⲗ ϩⲱⲱϥ ⲁϥⲥⲱⲟⲩϩ ⲉϩⲟⲩⲛ ⲙⲛ ⲡⲓⲏⲗ̄ ̄ ⲧⲏⲣϥ ̄ ⲉϩⲣⲁⲓ ⲉⲕⲉⲗⲃⲟⲩⲉ Here, the prior context probably facilitated the omission of this clause. The preceding passage reads ⲁⲩⲱ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲥⲱⲟⲩϩ ⲉϩⲟⲩⲛ ⲁⲩⲉⲓ ⲁⲩⲧⲁⲁⲩ ⲉϩⲣⲁⲓ ⲉⲥⲱⲙⲁⲛ ‘And the Philistines gathered together, they came, they betook themselves to Soman’, and the narrative is comprehensible even without an explicit mention of the encampment of the Israelites (‘Saul for his part gathered together with all Israel to Kelboue’). This example resembles verse 17: 1 where the same verb was omitted; see above.100 29: 10 καὶ ὀρθρίσατε ἐν τῇ ὁδῷ καὶ φωτισάτω ὑμῖν, καὶ πορεύθητε. SaM ϣ̄ⲣⲡ̄ ⲧ ̄ ⲉⲧⲛⲃ̄ ⲱⲕ ̄ ⲏⲩⲧⲛ̄ ϭⲉ ⲉⲧⲉϩⲓⲏ̄ ⲛⲥⲟⲟⲩⲧⲛ ⲛⲁϩⲣⲏⲧⲛ̄ ⲛⲧ SaA ϣⲉⲣⲉⲡ ⲧⲏⲩⲧⲉⲛ ⲛⲥ̄ ⲟⲟⲩⲧⲉⲛ ⲛⲧ ̄ ⲉⲧⲛⲃ̄ ⲱⲕ In this case, the translator understood φωτισάτω to be in an attributive relationship to ὁδῷ and provided as an equivalent ⲥⲟⲟⲩⲧⲛ ‘straight, upright’.101 Accordingly, either the Vorlage was without καί before the word, like L 554mg, or the translator elided the conjunction. Text-critical note: In this passage, the Sahidic text notably does not support the Lucianic text in its late form since just before these words there is the phrase ὡς ἄγγελος Θεοῦ ἐν ὁδῶ for which no parallel exists in the Sahidic. The Lucianic recension results from a longer process,102 and this explains why so few Lucianic readings exist in the Sahidic. These occasional common

100 One more passage with a similar omission is found in verse 17: 2. 101 Crum (Dictionary, 372b) gives this passage in his examples, and according to him ⲛⲥⲟⲟⲩⲧⲛ has the meaning ‘forthwith, just now’ but I do not see his interpretation as sufficient in this passage. 102 Kauhanen, Proto-Lucianic Problem, 14–15.

64

Description of the translation technique

readings stem both from the base text of the Lucianic recension and also from common secondary sources.103 2.1.3.2. Asyndetic clauses and their renderings Asyndeton means ‘lack of conjunction’. Greek texts are less likely to omit conjunctions, while in Coptic the tendency is well-established.104 Therefore, it is no surprise that the great majority of 467 asyndetic clauses in Greek 1 Sam have been translated into Sahidic by using asyndeton. Most of these clauses occur in direct discourse, asyndeton being quite rare in the narrative texts of the LXX.105 When comparing asyndetic clauses in Greek and in Coptic, it is important to note that Coptic expresses many functions with various verbal forms that in Greek are expressed with conjunctions. The Sahidic translator has used asyndeton not only where Greek has asyndeton, but also as the most common equivalent for καί-clauses.106 I have included in the number of asyndetic clauses all the cases where no connective is used to join a clause to the previous clause. Edwin Mayser makes the following exception: “Kein Asyndeton liegt vor, wenn ein Satz ohne Bindewort mit einem Demonstrativpronomen oder mit einem den Fortschritt des Gedankens bezeichnenden Verbum beginnt.”107 I have included these cases under asyndeton since I want to cover all the clauses, and this is the group to which they mainly belong.108 Within this large group, one must further discern how the translator has rendered his source text. 2.1.3.2.a) Asyndeton rendered with asyndeton There are 425 clauses where an asyndetic clause in Greek produces asyndeton in Sahidic. Almost all of these clauses occur in direct discourse, the majority of them beginning a speech; only 26 cases out of 425 occur in a narrative context. As a sub-group in direct discourse, there are 40 cases where 103 I see it as probable that there were Hexaplaric-type readings in the margins of some manuscripts, and Sahidic has picked up some of them. This is perceivable in a few cases where a correction according to Hebrew is present in Sahidic, but the same correction is not attested in any surviving Greek ms. 104 For asyndeton in Greek see Edwin Mayser, Grammatik II 3, §166; Denniston, Greek Particles, xliii-xlvii; and Blass/Debrunner/Rehkopf, Grammatik, §§458–463. For asyndeton in Coptic, see Layton, Grammar, §237. 105 Cf. Aejmelaeus, Parataxis, 83–4. She actually mentions legal texts, poetic passages and direct discourse as typical contexts of asyndeton. The two first mentioned are not present in 1 Sam, Hannah’s psalm, with its ten verses, being the only poetical passage. 106 See my analyses of the renderings of καί, subsection 2.1.3.1.b). 107 Mayser, Grammatik II/3, §166. Thus also R. Kühner/B. Gerth (Ausführliche Grammatik II/2, §546,2), who make a distinction between ostensible and actual asyndeton: “Es sind aber zweierlei Asyndeta wohl zu unterschieden: wirkliche und scheinbare”. 108 One could include interrogative and perhaps also relative clauses in this group, but they are treated as their own groups.

Clause connections

65

the context can be described as reported speech. In these passages, someone’s speech is cited or transmitted by someone else. The Sahidic translator is logical in his use of the word ϫⲉ: this word is used every time the speech begins, but only at the beginning. If there is a speech embedded within another speech, it is also introduced with this small word. In such a context, ϫⲉ is to be seen as a mere colon.109 Thus there is no connection to ὅτι recitativum when ϫⲉ is present, and Sahidic is not useful in textual criticism where this Greek feature is concerned. 15: 1

καὶ εἶπεν Σαμουὴλ πρὸς Σαούλ Ἐμὲ ἀπέστειλεν Kύριος χρῖσαί σε εἰς βασιλέα ἐπὶ Ἰσραήλ, …

ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲛⲥⲁⲟⲩⲗ ϫⲉ ⲡϫⲟⲉⲓⲥ ⲁϥⲧⲛⲛ ̄ ⲟⲟⲩⲧ ⲉⲧⲁϩⲥⲕ ̄ ⲛⲣⲣ̄ ⲟ ⲉϫⲙ̄ ⲡⲓⲏⲗ̄ ̄ This clause exemplifies asyndeton. A verb of saying precedes the clause, and the speech itself begins with asyndeton. There is also a small change in the word order in Sahidic, but the order used (‘the Lord sent me’) is the normal one in Sahidic and thus does not require any further attention. 11: 9

καὶ εἶπεν τοῖς ἀγγέλοις τοῖς ἐρχομένοις Τάδε ἐρεῖτε τοῖς ἀνδράσιν Ἰαβείς Αὔριον ὑμῖν ἡ σωτηρία διαθερμάναντος τοῦ ἡλίου.

ⲡⲉϫⲁϥ ⲛⲛⲉⲃⲁⲓϣⲓⲛⲉ ⲛⲧ ̄ ⲁⲩⲉⲓ· ϫⲉ ⲛⲁⲓ ⲛⲉⲧⲉⲧⲛⲛⲁϫⲟⲟⲩ ⲛⲛⲣⲱⲙⲉ ⲛⲓ̄ ⲁⲃⲓⲥ· ϫⲉ ⲣⲁⲥⲧⲉ ⲡⲟⲩϫⲁⲓ̈ ⲛⲁϣⲱⲡⲉ ⲛⲏⲧⲛ̄ ϩⲙ̄ ⲡⲧⲣⲉ ⲡⲣⲏ ϯ ⲛⲧⲉϥϩⲙ̄ⲙⲉ· In this example, the Sahidic translator has logically structured his text with ϫⲉ. It is used in both instances where a new speech act begins. The latter speech in this case, the message to be conveyed, is reported speech. 1: 2

καὶ τούτῳ δύο γυναῖκες· ὄνομα τῇ μιᾷ Αννα, …

ⲁⲩⲱ ⲡⲁⲓ ⲛⲉⲟⲩⲉⲛⲧⲁϥ ⲙⲙⲁⲩ ⲛⲥ̄ ϩⲓⲙⲉ ⲥⲛⲧ ̄ ⲉ· ⲡⲣⲁⲛ ⲛⲧⲟⲩⲉⲓ ⲙⲙⲟⲟⲩ ⲡⲉ ⲁⲛⲛⲁ This passage is one of eight clauses110 where a nominal clause in narration describing a person or a situation is rendered with asyndeton. In Sahidic, it is natural to use a nominal clause when describing a person. Text-critical note: Coptic nominal clauses always contain a copula. Therefore, Coptic offers no arguments when discussing whether the Greek is with or without the verb in these cases. 20: 2

οὐ μὴ ἀποθάνῃς· ἰδοὺ οὐ μὴ ποιήσῃ ὁ πατήρ μου ῥῆμα μέγα ἢ μικρὸν καὶ οὐκ ἀποκαλύψει τὸ ὠτίον μου·

ⲛⲅⲛⲁⲙⲟⲩ ⲁⲛ ⲉⲓⲥ ϩⲏⲏⲧⲉ (+ ⲅⲁⲣ SaA) ⲡⲁⲉⲓⲱⲧ ⲛϥⲛⲁⲉⲣ ⲗⲁⲁⲩ ⲁⲛ ⲛϣⲁϫⲉ· ⲟⲩⲇⲉ ⲟⲩⲛⲟϭ· ⲟⲩⲇⲉ ⲟⲩⲕⲟⲩⲓ· ⲉⲙⲡⲉϥϭⲟⲗⲡϥ·̄ ⲉⲛⲁⲙⲁⲁϫⲉ·

109 Vycichl, Dictionnaire Étymologique, 322. He also compares this word to ‫ לאמר‬in Hebrew. The translators of the LXX used ὅτι (recitativum) to render ‫ לאמר‬only four times (Aejmelaeus, On the Trail, 39). 110 1: 2, 8: 2, 9: 2, 14: 20, 14: 49, 17: 4 (twice), 17: 13.

66

Description of the translation technique

This is another example of asyndeton rendered with asyndeton. The translator renders the Greek expression ‘Look, behold’ with its usual equivalent, and afterwards the Sahidic word order differs from the Greek. The Sahidic reads ‘My father will not do anything’, in keeping with idiomatic Sahidic word order. Text-critical note: SaA has the conjunction ⲅⲁⲣ like L f 318 554 in Greek. This seems to be a stylistic or clarifying addition in the Lucianic manuscripts, not the OG. SaM has preserved the original Sahidic translation, whereas SaA probably acquired this addition from a Greek manuscript containing this reading. In several other instances, SaA also seems to exhibit Lucianic influence.111 When analyzing asyndetic clauses, one must consider whether there are Greek variants supporting the possibility of a conjunction. Naturally, the translator should not be held responsible for features coming from a different Vorlage. Usually, the Greek variants involve asyndeton and καί.112 I have not included all variants in the discussion since it is not necessary to mention all the cases where a single manuscript or some late manuscripts happen to have the conjunction. The number of both asyndetic clauses and καί-clauses is large enough that variation in a few cases does not change the overall picture. However, where a Greek variant involving καί and asyndeton is at hand, the Sahidic translator typically renders both asyndeton and καί with asyndeton. Therefore, the presence of asyndeton in Sahidic does not support asyndeton in its Vorlage. A typical example: 10: 9

καὶ ἐγενήθη ὥστε ἐπιστραφῆναι τῷ ὤμῳ αὐτοῦ ἀπελθεῖν ἀπὸ Σαμουηλ μετέστρεψεν αὐτῷ ὁ θεὸς καρδίαν ἄλλην· μετέστρεψεν] pr καί L 509 29 318)

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛⲧ ̄ ⲟⲟϣⲉ ⲉⲃⲟⲗ ϩⲓⲧⲟⲟⲧϥ· ⲥⲁⲟⲩⲗ ⲁ ̄ ⲉⲣⲉϥⲕⲧⲟ ⲛⲧⲉϥⲛⲁϩⲃⲉ ⲉⲙ ⲡⲛⲟⲩⲧⲉ ϯ ⲉⲣⲟϥ ⲛⲕⲉϩⲏⲧ· Here, Greek mss L 509 29 318 (=MT) have the apodotic conjunction καί. On the basis of translation technique, one cannot argue for either of the alternatives in the Vorlage of the Sahidic translator. Both asyndeton and καί would have been rendered with asyndeton.113 Occasionally, there are two or more variants in the Greek tradition for the Vorlage of the Sahidic translator. Considering Greek textual variants invol-

111 See section 3.4. 112 In my material, these cases appear under two headings: renderings of καί and renderings of asyndeton. 113 On the basis of these affiliations, one could argue that the Vorlage did not have the conjunction; in chapter 31, SaA includes readings also found in Lucianic mss whereas SaM has preserved the reading based on the OG.

Clause connections

67

ving asyndeton and conjunctions, Rahlfs’ edition offers 11 instances in which the determination of the Sahidic Vorlage is essentially impossible.114 I have included these cases in asyndeton even though they are not unambiguously such, but at least there is no obvious reason to place them elsewhere. Some of these cases show variation in Greek, and, in some cases, variation occurs in both the Greek and the Sahidic. The following examples stem from this group. 16: 1

Ἕως πότε σὺ πενθεῖς ἐπὶ Σαούλ, κἀγὼ ἐξουδένωκα αὐτὸν μὴ βασιλεύειν ἐπὶ Ἰσραηλ; πλῆσον (+ οὖν O) τὸ κέρας σου ἐλαίου, καὶ δεῦρο ἀποστείλω σε…

ϣⲁ ⲧⲛⲁⲩ ⲉⲕⲣϩ̄ ⲏⲃⲉ ⲛⲧ ̄ ⲟⲕ ⲉϫⲛ̄ ⲥⲁⲟⲩⲗ· ⲁⲩⲱ ⲁⲛⲟⲕ ⲁⲓⲥⲟϣⲃⲉϥ· ⲉⲧⲣⲉϥⲣⲣ̄ ⲣⲟ̄ · ⲧⲉⲛⲟⲩ ϭⲉ ⲙⲟⲩϩ ⲙ̄ⲡⲉⲕⲧⲁⲡ ⲛⲛⲉϩ· ⲛⲅ̄ ⲉⲓ ⲛⲧⲁϫⲟⲟⲩⲕ In Greek, only the O-group has the conjunction οὖν.115 In Sahidic, ϭⲉ is a standard equivalent for οὖν, supporting the presence of the conjunction οὖν in the Vorlage. However, the conjunction ϭⲉ has been added by the translator without any Greek parallel in eight instances. In the Sahidic New Testament, ϭⲉ occurs 401 times as an equivalent for οὖν, and 76 times without any Greek parallel.116 Translation technique offers no decisive argument in this passage. The translator has in any case added the word ⲧⲉⲛⲟⲩ ‘now’ without a Greek parallel. 30: 19 …καὶ ἕως πάντων ὧν ἔλαβον αὐτῶν, τὰ πάντα ἐπέστρεψεν Δαυείδ. τὰ πάντα] καὶ πάντα A B 509 120; πάντα O ⲁⲩⲱ ⲛⲉⲛⲧⲁⲩϥⲓⲧⲟⲩ ⲧⲏⲣⲟⲩ ⲉⲛⲕⲁ ⲛⲓⲙ ⲇⲁⲩⲉⲓⲇ ⲁϥⲕⲧⲟⲟⲩ̄ In this case, the Greek tradition preserves variant readings for τὰ πάντα. On the basis of translation technique, one cannot argue for either καί or asyndeton in the Vorlage. Brooke/McLean cite the Coptic for O as attesting the shortest variant, reading only πάντα.117 This decision is untenable since one cannot know whether the Sahidic read τὰ πάντα, καὶ πάντα or only πάντα in its Vorlage. 22: 23 κάθου μετ᾽ἐμοῦ, μὴ φοβοῦ· μή] pr καί L SaM ϩⲙⲟⲟⲥ ⲙⲡⲉⲓⲙⲁ ϩⲁϩⲧⲏⲓ̈· (+ ⲁⲩⲱ SaA) ⲛⲅ̄ ⲧⲉⲙⲣϩ̄ ⲟⲧⲉ In this passage, SaA offers the conjunction ⲁⲩⲱ, thus, possibly, sharing the variant in Lucianic manuscripts with καί. Notably and unusually, the Sahidic has used ⲁⲩⲱ before a conjunctive that is closely connected to a previous 114 3: 5(x2), 3: 6, 3: 9, 10: 9, 14: 6. 16: 1, 17: 23, 22: 23, 30: 19, 31: 4. 115 Brooke/McLean (Old Testament in Greek, 52) cite the Coptic in the apparatus for O with “uid”. 116 M. Wilmet, Concordance II: 3, 1681–3. 117 Brooke/McLean, Old Testament in Greek, 102.

68

Description of the translation technique

imperative. Because the Sahidic translation is defying its usual tendency, one could conclude that ⲁⲩⲱ was added later to bring Sahidic into line with a Greek manuscript. Feder has noted that later manuscripts of Jeremiah, especially Bohairic ones, are inclined to put ⲟⲩⲟϩ (= Sahidic ⲁⲩⲱ) in all the passages where Greek has καί.118 Since there are several passages where SaA apparently agrees with variants found in Lucianic manuscripts, it is probable that the copyist of SaA also added the conjunction here. SaM preserves the original Sahidic translation without conjunction. 2.1.3.2.b) Asyndeton rendered with a conjunction There are 31 cases where the Sahidic reads with a conjunction, and the Greek has asyndeton.119 This is a small number if one considers that there are altogether 467 asyndetic clauses in my material. There might be some more since in the cases where Greek shows variation between conjunction/ asyndeton,120 the Vorlage possibly had asyndeton. Still, the number of passages where the translator has added a connective is relatively small. Typically, this happens in direct speech. 20: 22 ἐὰν τάδε εἴπω τῷ νεανίσκῳ Ὧδε ἡ σχίζα ἀπὸ σοῦ καὶ ἐπέκεινα, πορεύου, ὅτι ἐξαπέσταλκέν σε Κύριος. ⲛⲉⲓϣⲁϫⲉ ⲇⲉ ⲉⲓϣⲁⲛϫⲟⲟⲩ ⲙⲡϩⲙ̄ϩⲁⲗ· ϫⲉ ⲉⲣⲉ ⲡⲥⲟⲧⲉ· ϩⲓ ⲡⲓⲥⲁ ⲙⲙⲟⲕ· ⲁⲩⲱ ⲥⲟⲩⲏⲩ ⲙⲙⲟⲕ· ⲉⲓⲉ ⲧⲱⲟⲩⲛ ⲛⲅⲃⲱⲕ ϫⲉ ⲁ ⲡϫⲟⲉⲓⲥ ϫⲟⲟⲩⲕ ⲉⲃⲱⲕ In Greek, as in Coptic, the apodosis of a conditional protasis typically begins without a conjunction. The other possibility in Sahidic is to use the particle ⲉⲓⲉ ‘then’. There are four passages where the translator has included ⲉⲓⲉ in the beginning of such an apodosis.121 This does not alter the meaning, but, rather, makes the text easier to follow. 17: 8

οὐκ ἐγώ εἰμι ἀλλόφυλος, καὶ ὑμεῖς Ἐβραῖοι τοῦ Σαούλ; ἐκλέξασθε ἑαυτοῖς ἄνδρα…

ⲙⲏ ⲁⲛⲟⲕ· ⲁⲛⲅ ⲟⲩⲫⲩⲗⲉⲥⲧⲓ̈ⲁⲓⲟⲥ ̅ ⲁⲛ· ⲛⲧⲱⲧⲛ̄ ⲇⲉ ⲛⲧⲉⲧⲛ̄ ⲛⲉϩⲉⲃⲣⲁⲓ̄ ⲟⲥ· ̅ ⲛϩⲙ̄ϩⲁⲗ ⲛⲥⲁⲟⲩⲗ· ⲥⲱⲧⲡ ϭⲉ ⲛⲟⲩⲣⲱⲙⲉ ⲉⲃⲟⲗ ⲛϩⲏⲧⲧⲏⲩⲧⲛ̄ Eight times122 asyndeton is rendered with the conjunction ϭⲉ. These cases occur in direct speech, as in this example, which is from Goliath’s speech to the Israelites. 118 Feder, Biblia Sahidica, 86–8. 119 Strictly speaking, 29 cases if one excludes two instances that have δή in the Greek text: 16: 17 and 23: 22. See below under the example of ϭⲉ. 120 I refer mainly to the 11 cases discussed previously. 121 20: 7 x2, 20: 21, 20: 22. 122 1: 27, 16: 16, 16: 17, 17: 8, 17: 36, 23: 22, 24: 11, 28: 18. Actually, in verses 16: 17 and 23: 22 the translator seems to have rendered the word δή with ϭⲉ unlike other 34 occurrences of δή, where the word has no separate rendering but its meaning seems to be included in imperative or jussive verbal form. In the Greek text, δή is mostly used to render Hebrew emphatic

Clause connections

69

Text-critical note: In Sahidic, the usual word for Philistines, ⲁⲗⲗⲟⲫⲩⲗⲟⲥ, is not used, but, instead, ⲫⲩⲗⲉⲥⲧⲓ̈ⲁⲓⲟⲥ. ̅ There are no similar variants in the Greek manuscripts. This word conforms to ‫תי‬ ִּ ְ ‫ש‬ ׁ ִ ‫פְל‬ ּ found in Hebrew, and is, therefore, an intriguing reading. It suggests that the Sahidic text has acquired this reading from a Greek source closer to the Hebrew than any of our preserved Greek manuscripts. 15: 23 ὅτι ἁμαρτία οἰώνισμά ἐστιν, ὀδύνην καὶ πόνους θεραφὶν ἐπάγουσιν·

ϫⲉ ⲡⲛⲟⲃⲉ· ⲟⲩϫⲓⲙⲁⲉⲓⲛ ⲡⲉ· ⲁⲩⲱ ⲟⲩⲉⲙⲕⲁϩⲛϩ̄ ⲏⲧ ⲙⲛ ⲟⲩϩⲓⲥⲉ· ⲉϣⲁⲩⲧⲣⲉ ⲟⲩⲙⲟⲕϩⲥ ̄ ϣⲱⲡⲉ. This is one of three passages in my corpus where it seems that the translator added the conjunction ⲁⲩⲱ in a case where Greek offers asyndeton.123 The Sahidic text reads ‘For sin is a divination. And grief and pain bring on grief’ has as the equivalent for the transliteration θεραφιν ⲟⲩⲙⲟⲕϩⲥ̄ ‘grief’, with the same etymology as the preceding ⲟⲩⲉⲙⲕⲁϩⲛϩ̄ ⲏⲧ. The Sahidic translator probably did not understand the Hebrew word for idols, and therefore supplied something that fit the context. The verbal form is present with focalizing conversion, focusing on ⲟⲩⲙⲟⲕϩⲥ ̄ here.124 Although the underlying Greek text remains obscure, no Greek manuscript supports the presence of a conjunction.125 In my corpus of clause connections, there are only three passages where the translator has added the conjunction ⲁⲩⲱ and the Greek text has asyndeton. These passages must have been somehow difficult for the translator. Accordingly, for the purposes of textual criticism, ⲁⲩⲱ can be recognized as a reliable indicator of a conjunction in the Greek text. 26: 11 μηδαμῶς μοι παρὰ Κυρίου ἐπενεγκεῖν χεῖρά μου ἐπὶ χριστὸν Κυρίου. ⲁⲛⲟⲕ ⲇⲉ ⲛⲛⲉ ⲡⲁⲓ ϣⲱⲡⲉ ⲙⲙⲟⲓ· ⲛⲧⲟⲟⲧϥ̄ ⲙⲡϫⲟⲉⲓⲥ· ⲉⲧⲣⲁⲉⲛ ⲧⲟⲟⲧ· ⲉϫⲙ̄ ⲡⲉⲭⲥ̅ ̅ ⲙⲡϫⲟⲉⲓⲥ

particle ‫‘ ־נא‬surely’. The Sahidic translator seems to interpret δή in verses 16: 17 and 23: 22 as a connective and in other verses as an emphatic particle. 123 The other two passages are in verses 22: 15 and 25: 22. In both of these cases, there is a question of clause division, and it seems that one reason for the addition is a different segmentation by the Sahidic translator. In verse 13: 13, the Sahidic reads ⲁⲩⲱ ⲧⲉⲛⲟⲩ ⲛⲑⲉ … Thus the translator seems to have added ⲁⲩⲱ in a clause where Greek had ὡς νῦν. 124 My interpretation differs from Drescher’s, who renders this “For sin is a divination and a pain and a suffering which are wont to cause a grief to happen.” 125 The following Greek mss read the nominative singulars ὀδύνη and πόνος: A B V O CI b d f 29 55* 71 244 460. I have the preliminary critical text of Aejmelaeus in my example, and the corresponding apparatus, without church fathers, looks like this: ὀδύνην – θεραφίν] αδικια και θεραφειν (-φιν 19’ 318 554*) οδυνην και πονους L 318 554mg | ὀδύνην CII a s 55c 158 245 707] pr αδικιας 55; οδυνη 55* rel Sa | πόνους CII a s 55c 158 707] πονον 245; πονος 55* rel Sa |.

70

Description of the translation technique

This is a free rendering in which the translator has formulated the first part as an independent clause in Sahidic: ‘And I, may it not happen to me’. In the cases where Greek reads negation expressed with only one word, the Sahidic translator is forced to use free formulations since there is no direct equivalent for such an expression in Sahidic.126 28: 15 καὶ εἶπεν Σαμουήλ Ἵνα τί παρηνώχλησάς μοι ἀναβῆναί με; καὶ εἶπεν Σαούλ Θλίβομαι σφόδρα,… ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲁϥ ϫⲉ ⲉⲧⲃⲉ ⲟⲩ ⲁⲕⲉⲛⲱⲭⲗⲉⲓ ⲛⲁⲓ̈ ⲉⲉ̅ⲓ ̅ ⲉϩⲣⲁⲓ· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲁϥ ϫⲉ ⲉⲃⲟⲗ ϫⲉ ϯϩⲏϣ ⲉⲙⲁⲧⲉ Here, the translator has rendered the text according to the context. The previous clause is a question ⲉⲧⲃⲉ ⲟⲩ ‘why’, and its answer accordingly begins in Sahidic with ‘because’. Significantly, in only four instances, the added connective is a Greek-Coptic loanword—ⲇⲉ thrice and ⲙⲏⲡⲱⲥ once.127 Thus the remaining 27 cases have indigenous Coptic connectives: ϭⲉ, ⲉⲓⲉ, ⲁⲩⲱ, ⲣⲱ, ⲉϣϫⲉ, ⲉⲃⲟⲗ ϫⲉ, ⲟⲛ, ⲉϣⲱⲡⲉ, ⲧⲁⲓ ⲧⲉ ⲑⲉ, ϫⲉⲕⲁⲥ, ϩⲱⲱϥ, and ϫⲉ.128 This might be related to the fact that most asyndeta occur in direct speech. Only two of all the passages where a conjunction is added occur within narration—ⲧⲁⲓ ⲧⲉ ⲑⲉ once and ⲇⲉ once. In conclusion, the translator probably used more familiar (or idiomatic) expressions in direct speech. This same phenomenon was observed by Aejmelaeus in relation to free translations and their occurrences in the LXX Pentateuch: free, idiomatic translations were often found in direct speech.129 2.1.3.3. Interrogative clauses and their renderings There are 115 interrogative clauses in my material. In 86 clauses, the interrogative pronoun τίς in correct form appears,130 and eight times some other interrogative word is used in Greek.131 In 11 passages, the question begins with a negation. In six clauses ἦ begins the question. Four times a direct

126 Another case with μηδαμῶς is found in verse 22: 15 where it is rendered ⲙⲡⲉⲣϭⲱⲛⲧ ⲡⲣⲣⲟ. These translations are analyzed in the section on additions, 2.2.2.1. Nathalie Bosson has collected different Coptic versions of the book of Jonah, and in her work there is a passage (Jonah 1: 14) where μηδαμῶς is rendered with ⲙ̂ⲡⲣϭ ̂ ⲱⲛⲧ̂ ⲉⲣⲟⲛ in three Sahidic witnesses, whereas in one Sahidic ms and two Bohairic mss the rendering is only ⲙ̂ⲡⲱⲣ/ⲙ̇ ⲫⲱⲣ. See Bosson, “Synopse des témoins coptes de Jonas”. 127 11: 8, 25: 8, 26: 11 and 9: 5. 128 ϭⲉ 1: 27, 16: 16, 16: 17, 17: 8, 17: 36, 23: 22, 24: 11, 28: 18; ⲁⲩⲱ 15: 23, 22: 15, 25: 22; ⲉⲓⲉ 20: 7bis, 20: 21, 20: 22; ϫⲉ 15: 6, 17: 10, 20: 3; ϩⲱⲱϥ 11: 7, 22: 23; ϫⲉⲕⲁⲥ 27: 11; ⲟⲛ 20: 13; ⲧⲁⲓ ⲧⲉ ⲑⲉ 2: 14; ⲉϣϫⲉ 21: 10; ⲉϣⲱⲡⲉ 25: 8; ⲣⲱ 20: 2; ⲉⲃⲟⲗ ϫⲉ 28: 15. 129 Aejmelaeus, Parataxis, 59. 130 Blass/Debrunner/Rehkopf, Grammatik, §298. 131 Twice ἕως πότε, twice ποῦ, twice πόθεν and once each πῶς and ποῖος.

71

Clause connections

question has no special marking as a question.132 The following table lists the renderings used in the Sahidic text. The last group of the list (‘varia’) contains those cases where varying renderings have been used. Greek

Coptic

Number

τίς τίς ἵνα τί ἵνα τί οὐκ/ οὐχί οὐκ/ οὐχί ἦ τί ὅτι

ⲟⲩ ⲛⲓⲙ ⲉⲧⲃⲉ ⲟⲩ ⲁϩⲣⲟ ⲙⲏ asyndeton with negation ⲉⲛⲉ ⲉⲧⲃⲉ ⲟⲩ interrogative interrogative varia

32 cases 23 cases 16 cases 2 cases 9 cases 2 cases 3 cases 4 cases 10 cases 3 cases 11 cases135

interrogative133 τίς + a preposition134 varia 29: 8

Τί πεποίηκά σοι καὶ τί εὗρες ἐν τῷ δούλῳ σου…

ⲟⲩ ⲡⲉ ⲡⲛⲟⲃⲉ ⲛⲧⲁⲓⲁⲁϥ ⲛⲁⲕ· ⲁⲩⲱ ⲟⲩ ⲡⲉ ⲡϩⲱⲃ ⲛⲧⲁⲕϩⲉ ⲉⲣⲟϥ ⲉϩⲟⲩⲛ ⲉⲡⲉⲕϩⲙ̄ϩⲁⲗ These two questions have ⲟⲩ as the equivalent for τί. This word is used in the meaning ‘what’ and ‘what kind of’, thus being an accurate rendering for τί.136 In addition to these two cases there are 30 similar clauses: ⲟⲩ as the rendering of τί or τίς. 25: 10 Τίς ὁ Δαυεὶδ καὶ τίς ὁ υἱὸς Ἰεσσαί; ⲛⲓⲙ ⲡⲉ ⲇⲁⲩⲉⲓⲇ· ⲁⲩⲱ ⲛⲓⲙ ⲡⲉ ⲡϣⲏⲣⲉ ⲛⲓⲉⲥⲥⲁⲓ̄ ̅ In these two clauses, the translator uses the interrogative pronoun ⲛⲓⲙ ‘Who?’, which is a natural equivalent for τίς. There are 23 cases where ⲛⲓⲙ is used as an equivalent for τίς. Sahidic ⲛⲓⲙ is indeclinable, but prepositions can be attached to it as in the following example: 6: 20

καὶ πρὸς τίνα ἀναβήσεται κιβωτὸς Κυρίου ἀφ᾽ἡμῶν;

ⲁⲩⲱ ⲉⲣⲉ ⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡϫⲟⲉⲓⲥ ⲛⲁⲃⲱⲕ ⲉϩⲣⲁⲓ̈ ⲛⲛⲓⲙ ̅

132 12: 3, 16: 4, 16: 11, 25: 11. According to Blass/Debrunner/Rehkopf (Grammatik, §440), these direct questions without special marking are such that one can answer with ‘yes’ or ‘no’. 133 Twice τίς (5: 10, 17: 8), twice ἕως πότε (1: 14, 16: 1), twice ποῦ (10: 14, 19: 22), twice πόθεν (25: 11, 30: 13) and once each πῶς and ποῖος (16: 2, 9: 18). 134 4: 3 κατὰ τί, 24: 15 ὀπίσω τίνος, 26: 15 διὰ τί. 135 1: 8, 12: 3, 16: 4, 16: 11, 20: 2, 21: 16, 24: 17, 25: 11, 26: 17, 26: 18, 28: 15. 136 In verse 21: 15, there must be an error in SaM since it reads causal ⲉⲧⲃⲉ ϫⲉ instead of ⲉⲧⲃⲉ ⲟⲩ which is, however, attested in SaA.

72

Description of the translation technique

In this example, the translator has rendered this question as one unit: the word order in Sahidic is the usual one. 2: 29

καὶ ἵνα τί ἐπέβλεψας ἐπὶ τὸ θυμίαμά μου καὶ εἰς τὴν θυσίαν μου ἀναιδεῖ ὀφθαλμῷ;

ⲁⲩⲱ ⲉⲧⲃⲉ ⲟⲩ ⲁⲕϭⲱϣⲧ ⲉϫⲛ ⲧⲁⲑⲩⲥⲓⲁ ⲁⲩⲱ ⲉϫⲙ ⲡⲁϣⲟⲩϩⲏⲛⲉ ϩⲛ ⲟⲩⲃⲁⲗ ⲛⲁⲧϣⲓⲡⲉ· Of the 18 passages reading ἵνα τί, the translator has used ⲉⲧⲃⲉ ⲟⲩ ‘Why?’ in 16 cases.137 This is a suitable equivalent—as is ⲁϩⲣⲟ⸗, which is used only twice138: καὶ ἵνα τί οὐκ ἐσθίεις; ⲁⲩⲱ ⲁϩⲣⲟ ⲛⲧⲉⲟⲩⲱⲙ ⲁⲛ· The meaning of the word ⲁϩⲣⲟ⸗ is ‘Why? What is the matter with …?’. Sahidic indicates the person when using the interrogative ⲁϩⲣⲟ⸗ through a pronominal suffix, and is, consequently, more specific than its Greek source text. Perhaps, this is one reason for the prevalence of ⲉⲧⲃⲉ ⲟⲩ as a rendering of ἵνα τί.

1: 8

17: 8

οὐκ ἐγώ εἰμι ἀλλόφυλος, καὶ ὑμεῖς Ἐβραῖοι τοῦ Σαουλ;

ⲙⲏ ⲁⲛⲟⲕ· ⲁⲛⲅ ⲟⲩⲫⲩⲗⲉⲥⲧⲓ̈ⲁⲓⲟⲥ ̅ ⲁⲛ· ⲛⲧⲱⲧⲛ̄ ⲇⲉ ⲛⲧⲉⲧⲛ̄ ⲛⲉϩⲉⲃⲣⲁⲓ̄ ⲟⲥ· ̅ ⲛϩⲙ̄ϩⲁⲗ ⲥⲁⲟⲩⲗ There are 11 passages where Greek has a question beginning with οὐχί or οὐκ. Blass/Debrunner/Rehkopf distinguish between the use of negation in questions; when the anticipated answer is positive, οὐχί or οὐ is used, and, when expecting a negative answer, μή is used.139 In Sahidic, ⲙⲏ is used as an initial attitude marker, to mark the clause as a question, but, unlike Greek, without presupposition of a negative answer. In Sahidic, the answer is anticipated by either a negative (supposing a positive answer) or an affirmative (supposing a negative answer) clause.140 The Sahidic translator uses ⲙⲏ nine times as in this example. In only two instances has the translator rendered an οὐχί/οὐκ-clause without ⲙⲏ, but just with a negative verbal form in Sahidic; one of the cases is given here:141 26: 14 καὶ τῷ Ἀβεννὴρ ἐλάλησε λέγων Οὐκ ἀποκριθήσῃ, Ἀβεννήρ; ⲁⲩⲱ ⲁϥϣⲁϫⲉ ⲙⲛ̄ ⲁⲃⲉⲛⲛⲏⲣ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ· ϫⲉ ⲛⲅⲛⲁⲉⲣⲟⲩⲱ ⲁⲛ ϩⲁⲣⲱⲓ ⲁⲃⲉⲛⲛⲏⲣ·

137 ⲉⲧⲃⲉ ⲟⲩ is also used to render διὰ τί in verse 26: 15 and κατὰ τί in verse 4: 3. 138 In verses 1: 8 and 28: 16. 139 Blass/Debrunner/Rehkopf, Grammatik, §427. 140 Layton, Grammar, §238. Thus also Anba Gregorius, “Greek Loan Words”, 84. 141 The other is in verse 21: 12.

Clause connections

73

Here, the rendering is accurate. The translator has added the word ϩⲁⲣⲱⲓ ‘to me’ as elsewhere in the Sahidic translation: the addressee is mentioned explicitly.142 10: 11 Τί τοῦτο τὸ γεγονὸς τῷ υἱῷ Κείς; ἦ καὶ Σαοὺλ ἐν προφήταις; ⲟⲩ ⲡⲉ ⲡⲉⲓϩⲱⲃ ⲛⲧⲁϥϣⲱⲡⲉ ⲙⲡϣⲏⲣⲉ ⲛϭⲓⲥ· ⲉⲛⲉ ⲥⲁⲟⲩⲗ ϩⲱⲱϥ ⲁϥⲱⲡ ϩⲛ ⲛⲉⲡⲣⲟⲫⲏⲧⲏⲥ In this verse, the Greek has either the disjunctive particle ἤ or the interrogative particle ἦ.143 In a text without accents, one cannot distinguish between the two possibilities.144 The Sahidic text has the interrogative particle ⲉⲛⲉ and seems to indicate ἦ in the Vorlage, or at least the translator understood it as such. There are six cases of ἦ/ἤ in 1 Sam: three are rendered as the conjunction ἦ,145 two read the clause with a different word order thus marking it as a question146 without any interrogative particle. The remaining passage is presented below: ἵνα τί εἰσηγάγετε αὐτὸν πρός με; ἦ ἐλαττοῦμαι ἐπιλήμπτων ἐγώ, …; ἦ] om. B O =MT; μη CII-242 s 370; ει 509 610;

21: 15–6

SaM ⲉⲧⲃⲉ ⲟⲩ ⲁⲧⲉⲧⲛⲛ ̄ ⲧϥ ̄ ⲛⲁⲓ ⲉϩⲟⲩⲛ· ⲏ ⲉⲓϣⲁⲁⲧ ⲁⲛⲟⲕ· ⲛⲣⲱⲙⲉ ⲉϥⲡⲟϣⲥ ̄ ⲛϩⲏⲧ· SaA ⲉⲧⲃⲉ ⲟⲩ ⲁⲧⲉⲧⲛⲏ ̄ ⲛⲧϥ̄ ⲛⲁⲓ ⲉϩⲟⲩⲛ. ⲙⲏ ⲉⲓϣⲁⲁⲧ ⲁⲛⲟⲕ ⲛⲣ̄ ⲱⲙⲉ ⲉϥⲗⲟⲃⲉ. This example147 contains two questions and is relevant because of the many related variants in the Greek tradition. In the Greek manuscripts, there is the same problem as in the previous example with ἦ/ἤ—in an unaccented text the difference is not perceivable.148 SaM understands the first word as a disjunctive ⲏ, like other passages where ἤ is used as the connective.149 The reading of SaA is ⲙⲏ. ⲙⲏ initiates, in this instance, a question, and an affirmative clause signals that a negative answer is expected. More often, ⲙⲏ is used in 1 Sam to render Greek questions that begin with negation.150 There is no firm conclusion concerning the Sahidic translation in this particular verse: it has ⲙⲏ or ⲏ, both of which render an interrogative clause. Text-critical note: Greek mss B and O omit the first word of the second question, like the Hebrew, but ms 376 has a longer homoioteleuton mistake, not just omission of this word. Greek mss 509 and 610 begin the question

142 See section 2.2.3.1. 143 J. Blomqvist/P.O. Jastrup (Grekisk grammatik, § 301) state that ἦ is used mainly in Classical Greek, and in Koine the corresponding orthography is εἰ. 144 Brooke/McLean read the word ἤ whereas Rahlfs has ἦ in this verse. 145 In addition to this verse, verses 10: 12 and 22: 15. 146 24: 17 and 26: 17, in the group ‘varia’. 147 Belongs to the group ‘varia’. 148 However, 376 has a homoioteleuton mistake, not only omission of this word. 149 See subsection 2.1.3.4.d). 150 See the example in verse 17: 8 below.

74

Description of the translation technique

with εἰ thus actually giving only another form of ἦ. The word μή is found in CII-242 s 370, but there are no other passages in 1 Sam where μή begins a question, and, therefore, this reading is suspicious when it comes to the Vorlage of the Sahidic. Τί ὅτι κλαίει ὁ λαός; ⲉⲧⲃⲉ ⲟⲩ ⲡⲗⲁⲟⲥ ⲣⲓⲙⲉ There are five clauses with τί ὅτι and two passages with τί.151 In these cases, the question is ‘Why?’. Sahidic renders these cases with ⲉⲧⲃⲉ ⲟⲩ. Six cases are like this, having ⲉⲧⲃⲉ ⲟⲩ as a rendering for τί ὅτι. In one passage, the translator has formulated the question with variation:

11: 5

20: 2

καὶ τί ὅτι κρύψει ὁ πατήρ μου τὸ ῥῆμα τοῦτο; οὐκ ἔστιν τοῦτο.

ⲁⲩⲱ ϭⲉ ⲉⲣⲉ ⲡⲁⲉⲓⲱⲧ ⲛⲁϩⲉⲡ ⲡⲉⲓϣⲁϫⲉ ⲉⲣⲟⲓ· ⲉⲛϥϣⲟⲟⲡ ⲁⲛ ⲣⲱ̄ Usually, τί ὅτι is translated ⲉⲧⲃⲉ ⲟⲩ, but not in this case.152 In this sentence, there is a semantic discrepancy between the Greek and Sahidic texts. In Greek, the question is ‘Why is it that my father should hide this matter? This is not possible.’153 The Sahidic translator has rendered this question ‘And will my father, then, hide this thing from me? It is not at all so.’ The gloss, ϭⲉ, in this context is fitting, since the question and its answer are understood as a conclusion of the preceding clauses. 19: 22 καὶ ἠρώτησεν καὶ εἶπεν Ποῦ Σαμουὴλ καὶ Δαυιδ; SaM ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥϣⲓⲛⲉ ̅ ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ ⲉⲣⲉ ⲥⲁⲙⲟⲩⲏⲗ· ⲛⲧⲱⲛ· SaA ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥϣⲓⲛⲉ ⲉϥϫⲱ ⲙ̄ⲙⲟⲥ ϫⲉ ⲉⲣⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲧⲱⲛ ⲙⲉⲛ ⲇ̅ⲁⲇ ̅ ̅. This is an example from the group “interrogative”; both Greek and Sahidic use interrogative words that mean ‘Where?’.154 There is a small error in SaM. It is the only witness that omits ‘David’ here. 28: 15 καὶ νῦν κέκληκά σε γνωρίσαι μοι τί ποιήσω. ⲉⲧⲃⲉ ⲡⲁⲓ ⲧⲉⲛⲟⲩ· ⲁⲓⲙⲟⲩⲧⲉ ⲉⲣⲟⲕ· ⲉⲧⲣⲉⲕⲧⲟⲩⲛⲓⲁⲧ ⲉ̄ⲃⲟⲗ ⲙⲡⲉϯⲛⲁⲁⲁϥ Here, the translator has rendered a Greek indirect question with an articulated relative construction.155 The Greek expression resembles an object clause here, and the rendering is accurate.

151 Additionally, in verse 26: 18, ὅτι τί seems to be rendered with ⲉⲧⲃⲉ ⲟⲩ, but there are textual variants that might affect this clause. 152 This belongs to the group ‘varia’. 153 Translation from NETS, Bernard A. Taylor, “Text of Reigns”, 262. 154 SaM reads ⲛⲧⲱⲛ and SaA ⲧⲱⲛ, but this does not affect the meaning (Crum, Dictionary, 417b). 155 Layton, Grammar, §411.

75

Clause connections

2.1.3.4. Other coordinated clauses 2.1.3.4.a) ὰλλά-clauses and their renderings The Greek connective ἀλλά is an adversative coordinating particle that expresses a difference in relation to the preceding clause either by negating or by restricting it.156 ἀλλά can be used both after a negative clause and after an affirmative clause. In Sahidic, the Greek-Coptic loanword ⲁⲗⲗⲁ functions basically in the same way as in Greek. The meaning of ⲁⲗⲗⲁ is ‘but rather, but, though’.157 My corpus contains 20 ἀλλά-clauses. Greek

Coptic

Occurrences

ἀλλά158 ἀλλʼ ἤ ἀλλά ἀλλά159 ἀλλά ἀλλʼ ἤ

ⲁⲗⲗⲁ ⲁⲗⲗⲁ ⲁϩⲉ ⲣⲱ ϭⲉ no equivalent

11 times 5 times once once once once

In Sahidic 1 Sam, ⲁⲗⲗⲁ serves as the equivalent for both ἀλλά and ἀλλʼ ἤ.160 Denniston says that this latter connective and its explanation “has been much discussed”.161 In the following, I will provide examples of different ἀλλά-clauses and their translations.

156 Denniston (Greek Particles, 1) describes its basic meanings with the terms “eliminative, objecting, and less frequently balancing.” 157 Layton, Grammar, §145, 493. Müller (“Greek Connectors”, 9): “It appears in clause-initial position with any sentence pattern and is used for all types of contrast and a discourse marker.” 158 In three cases, Greek variants involve ἀλλʼ ἤ. Since the same equivalent is used for both ἀλλά and ἀλλʼἤ, it is impossible to use translation technical arguments to say whether the Coptic translator had ἀλλά or ἀλλʼ ἤ in his Greek Vorlage. There are three cases where the majority of Greek mss read ὅτι, but ἀλλά or ἀλλʼ ἤ appears in other mss. I have included these cases within this subgroup since there are no cases where ὅτι is rendered with ⲁⲗⲗⲁ in Sahidic 1 Sam. All three of these passages are presented below. 159 There are Greek variants reading ἀλλʼἤ; see previous note. 160 This is observed also by Feder (Biblia Sahidica, 84) within the Sahidic Jeremiah. Denniston (Greek Particles, 3) gives examples of passages where ἀλλά is used in the same meaning (except) as ἀλλʼ ἤ. However, in the cases where ἀλλʼ ἤ is used in the meaning ‘except’ it usually has the preposition ⲛⲥⲁ as the equivalent in Coptic, but this usage does not occur in a clause connective function. 161 Denniston, Greek Particles, 24. According to Kühner/Gerth (Ausführliche Grammatik II/ 2 §534, 6), the latter is used where only ἀλλά or ἤ was to be expected, and this combination results from a parallel usage and confusion of οὐδὲν ἄλλο ἀλλά and οὐδὲν ἄλλο ἤ. Mayser (Grammatik II,3§164) mentions that after a negation ἀλλά functions with the same meaning as εἰ μή and πλήν and is usually accompanied by a pleonastic ἤ. Denniston (Greek Particles, 4) also gives examples where ἀλλά is used as a synonym for πλήν.

76

30: 2

Description of the translation technique

οὐκ ἐθανάτωσαν ἄνδρα καὶ γυναῖκα ἀλλ᾽ ᾐχμαλώτευσαν·

ⲙⲡⲟⲩⲙⲉⲟⲩⲧ ⲟⲩϩⲟⲟⲩⲧ ⲟⲩⲇⲉ ⲟⲩⲥϩⲓⲙⲉ ⲁⲗⲗⲁ ⲁⲩⲁⲓⲭⲙⲁⲗⲱⲧⲓⲍⲉ ⲙⲙⲟⲟⲩ This is a typical example: after a negative clause an affirmative clause introduced with ⲁⲗⲗⲁ. 1: 23

κάθου ἕως ἂν ἀπογαλακτίσῃς αὐτό· ἀλλὰ στήσαι Κύριος τὸ ἐξελθὸν ἐκ τοῦ στόματός σου.

ϩⲙⲟⲟⲥ ⲛⲉ ϣⲁⲛⲧⲉⲱⲙ̄ϫ ⲙ̄ⲡϣⲏⲣⲉ ⲕⲟⲩⲓ ⲁⲗⲗⲁ ⲡϫⲟⲉⲓⲥ ⲉϥⲉⲧⲁϩⲟ ⲉⲣⲁⲧϥ ⲙⲡⲉⲛⲧⲁϥⲉⲓ ⲉⲃⲟⲗ ϩⲛ̄ ⲧⲟⲩⲧⲁⲡⲣⲟ In this example, ἀλλά restricts the preceding clause. Elkana confirms the vow made by his wife, and leaves the matter between Hannah and God. Here, ⲁⲗⲗⲁ occurs after an affirmative imperative clause. In the clause headed by ⲁⲗⲗⲁ, the verbal form is optative both in Greek and in Sahidic. This clause shows that ⲁⲗⲗⲁ does not occur only after negation.162 16: 6

καὶ εἶπεν Ἀλλὰ καὶ163 ἐνώπιον Κυρίου χριστὸς αὐτοῦ.

ⲡⲉϫⲁϥ ⲉⲧⲃⲏⲏⲧϥ ϫⲉ ⲉⲓⲥ ⲡⲉⲭⲣⲥ ⲣⲱ ⲙⲡϫⲟⲉⲓⲥ ⲙⲡⲉϥⲉⲙⲧⲟ ⲉⲃⲟⲗ In this passage, the translator is quite free. ⲣⲱ is an enclitic particle with the meaning ‘but, on the contrary, not even, indeed’. According to Layton, ⲣⲱ marks, but does not link, clauses. He uses the term ‘initial attitude marker’.164 Usually, ⲉⲓⲥ acts as an equivalent for ἰδού in Greek, but in this case no such variant occurs in Greek manuscripts. The translator has formulated the sentence with a freedom typical of direct speech translations elsewhere in Sahidic 1 Sam since after ‘said’ Sahidic reads ‘concerning him: Look’ without a Greek source text.165 14: 29–30 …ὅτι ἐγευσάμην βραχὺ τοῦ μέλιτος τούτου. ἀλλ᾽ ὅτι εἰ ἔφαγεν ἔσθων ὁ λαὸς σήμερον τῶν σκύλων τῶν ἐχθρῶν αὐτῶν ὧν εὗρεν, ὅτι νῦν ἂν μείζων ἦν ἡ πληγή… om. εἰ A B O 82 CI 799 68 55 460 SaM ⲉⲃⲟⲗ ϫⲉ ⲁⲓⲧⲉⲡ ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲓⲉⲃⲓⲱ̅· ⲉⲓ̅ ⲉ ̅ ⲉⲛⲉ ⲛⲧⲁ ⲡⲗⲁⲟⲥ ϭⲉ ⲟⲩⲱⲙ· ϩⲛ ⲟⲩⲱⲙ· (sic) ⲙⲡⲟⲟⲩ· ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ ⲛⲛⲉⲩϫⲁϫⲉ…ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ̈ ⲡⲉ ⲉⲃⲟⲗ ϫⲉ ⲁⲓⲧⲉⲡ ⲡⲓϣⲏⲙ ⲛⲉⲃⲓⲱ· ⲉⲓⲉ ⲛⲧⲁ ⲡⲗⲁⲟⲥ ⲟⲩⲱⲙ ϩⲛ ⲟⲩⲱⲙ(sic) ⲉⲙⲡⲟⲟⲩ SaA ⲉⲃⲟⲗ ϩⲛ̄ ⲛϣ ̄ ⲟⲗⲉⲥ ⲛⲛⲉⲩϫⲁϫⲉ…ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ ⲡⲉ ⲉ]ⲃⲟⲗ ϫ̣ ⲉ ̣ [ⲁⲓⲧⲉⲡ ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ] ϩⲙ SaU ̣ ̄ ⲡⲉⲉⲓⲉⲃⲓⲱ· ⲉⲉⲓⲉ ⲉⲛ̣ⲉ ⲛ[ⲧⲁ ⲡⲗⲁⲟⲥ ϭⲉ ⲟⲩⲱⲙ ϩⲛ ⲟ]ⲩⲱⲙ ⲙ̄ⲡⲟⲟⲩ ⲉⲃ[ⲟⲗ] ϩⲛ ̣ … [ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ̈] ⲡⲉ 162 Stern (Koptische Grammatik, §597) mentions only the use after a negation. 163 Greek manuscripts L d 554 read instead of ἀλλὰ καί the words ἀλλʼ ἤ, but this does not make any difference on the Coptic side. 164 Layton, Grammar, §238. 165 For a detailed description of additions see section 2.2.

Clause connections

77

The Sahidic text reads a conditional clause,166 and after this, the apodosis of a contrafactual conditional clause: ⲛⲉⲣⲉ…ⲛⲁ-. SaM has as the equivalent for ἀλλά the conjunction ϭⲉ, which is rare, but not impossible as an equivalent.167 ϭⲉ is an enclitic conjunction with the meaning ‘then, therefore, again, yet’. In this case, ⲉⲓ̅ ⲉ ̅ is probably an interjection, strengthening the next word, since the conditional marker ⲉⲛⲉ is already present.168 The manuscript SaA lacks the conjunction ϭⲉ, retaining only the conditional clause and its apodosis. I see ⲉⲓⲉ in this manuscript as a variant form of the conditional ⲉⲛⲉ, otherwise there is no conditional protasis and no reason to have the apodosis either.169 Thus, SaA shortens the text a little by excluding some words.170 SaU is fragmentary, but preserves the beginning of the clause like SaM. In the lacuna, one could reconstruct either like Butts ⲛ[ⲧⲁ ⲡⲗⲁⲟⲥ ϭⲉ ⲟⲩⲱⲙ ϩⲛ ⲟ]ⲩⲱⲙ or ⲛ[ⲧⲁ ⲡⲗⲁⲟⲥ ⲟⲩⲱⲙ ϩⲛ ⲟⲩⲟ]ⲩⲱⲙ. In both alternatives, the number of letters would remain the same.171 SaM and SaA both attest to the erroneous ⲟⲩⲱⲙ ϩⲛ ⲟⲩⲱⲙ, but this does not necessitate a dependence; ⲟⲩⲟⲩⲱⲙ is an orthographic equivalent to ⲟⲩⲱⲙ. However, this shared reading, probably an early mistake in the Sahidic tradition, is the reason for the reconstruction here of ϭⲉ within the lacuna of SaU. 10: 19 καὶ εἴπατε Οὐχί, ἀλλ᾽ἢ ὅτι βασιλέα στήσεις ἐφ᾽ἡμῶν. om. ὅτι A O L 509 130–381 ⲁⲩⲱ ⲡⲉϫⲏⲧⲛ ϫⲉ ⲛⲧ ̄ ⲛⲛⲁϭⲱ ⲁⲛ ϩⲓⲛⲁⲓ ⲁⲗⲗⲁ ⲕⲱ ⲛⲟⲩⲣⲣ̄ ⲟ̅ ϩⲓϫⲱⲛ The translator does not distinguish between ἀλλά and ἀλλʼ ἤ when used as clause connectors in the meaning ‘but’. Here ⲁⲗⲗⲁ occurs after a negation to express the positive alternative. The one-word negation in Greek is paraphrased and expanded into a clause ‘we will not remain in this way’,172 and the beginning of the speech is introduced with ϫⲉ in its usual position. The translator has rendered the Greek source text accurately, and the result is a fluent Sahidic expression. Text-critical note: The question remains as to whether the conjunction ὅτι was in the Vorlage. Aejmelaeus is suspicious when it comes to the original LXX in cases like this, and considers it possible that ὅτι was added secondarily to make the text conform more closely to the Hebrew.173 Translation

166 Layton (Grammar, §498) calls this a “contrary-to-fact stipulation clause”. 167 It also occurs in the examples of Crum, Dictionary, 802. 168 Crum, Dictionary, 74b(c). 169 ⲉⲓⲉ, a variant form of ⲉⲛⲉ. Cf. Crum, Dictionary, 74b(d). 170 See section 3.3 for a description of the character of SaA. 171 According to Butts (“P.Duk.inv. 797”, 16), the only argument for the inclusion of ϭⲉ is space. See also my appendix 6.3 for a collation of this fragment. 172 See also subsection 2.2.2.1. 173 Aejmelaeus, On the Trail, 131–2. In these cases, the Hebrew texts reads either ‫ כּי‬or ‫כּי אּם‬ with an adversative meaning, offering a particular challenge to the Greek translator. The standard equivalent for ‫ כּי‬was ὅτι, but this Greek term is not appropriate for adversative

78

Description of the translation technique

technique offers no help when enquiring whether ὅτι was present in the Vorlage of the Sahidic translator. Had there been ὅτι in the Vorlage, the Sahidic text in its present form does not indicate it. 8: 7 SaM SaA

ὅτι οὐ σὲ ἐξουθενήκασιν, ἀλλ᾽ἢ ἐμὲ ἐξουδενώκασιν τοῦ μὴ βασιλεύειν ἐπ᾽αὐτῶν.

ⲛⲧⲟⲕ ⲁⲛ ⲡⲉ ⲛⲧⲁⲕⲥⲟϣϥⲕ ̄ ⲉⲧⲉⲙⲣⲣ̄ ⲣ̄ ⲟ ⲉϫⲱⲟⲩ ⲛⲧⲟⲕ ⲡⲉ ⲛⲧⲁⲩⲥⲟϣϥⲕ ̄ ⲉⲧⲉⲙⲣⲣ̄ ⲣ̄ ⲟ ⲉϩⲣⲁⲓ ⲉϫⲱⲟⲩ In this example, there is no equivalent for the ἀλλά-clause in Sahidic. The omission of the ἀλλά-clause probably relates to a homoioteleuton copying mistake: two times with a long verb in the 3rd person plural, and only a few words in between. A few Greek manuscripts contain the same omission, but the same error could happen within the Sahidic tradition. Thus, there is no need to identify the Sahidic translation as dependent upon these Greek manuscripts.174 The fact that both Sahidic manuscripts attest to this mistake reveals a common Sahidic ancestor and speaks against any thorough revision according to Greek after the mistake. Sahidic manuscripts also have further corruptions in this verse, concerning the person and the negation; the earliest form should be ⲛⲧⲟⲕ ⲁⲛ ⲡⲉ ⲛⲧⲁⲩⲥⲟϣϥⲕ ̄ ‘It is not you that they have despised’, SaM erroneously translates the verbal person ⲛⲧⲟⲕ ⲁⲛ ⲡⲉ ⲛⲧⲁⲕⲥⲟϣϥⲕ ̄ ‘It is not you that you have despised’, and SaA omits the negation ⲛⲧⲟⲕ ⲡⲉ ⲛⲧⲁⲩⲥⲟϣϥⲕ ̄ ‘It is you that they have despised’.

2: 16

καὶ εἶπεν Οὐχί, ὅτι (om d-106) νῦν δώσεις· ὅτι] om d-106; ἀλλ᾽ἤ L 318 554

ⲡⲉϫⲁϥ ⲛⲁϥ ϫⲉ ⲛϯ ̄ ⲓⲛⲁϭⲱ ⲁⲛ· ⲁⲗⲗⲁ ⲉⲕⲛⲁϯ ⲛⲁⲓ ⲧⲉⲛⲟⲩ In three passages (2: 16, 5: 5, 30: 22), Sahidic reads ⲁⲗⲗⲁ, but the reading of the Vorlage remains unclear.175 In all these cases there are Greek manuscripts attesting to ἀλλά. Here, the translator has paraphrased οὐχί as a clause.176 In Greek, some manuscripts offer ἀλλ᾽ἤ, but the suspicions referred to earlier concerning the translation of ὅτι in these kinds of clauses undermine the argument of a dependence.177 Possibly, the Sahidic translator had ὅτι in the Vorlage, and he rendered freely according to the content of the clause.

clauses. In this verse, the Hebrew orthography is especially problematic. The LXX supposes ‫‘ לא‬not’, whereas the MT reads ‫‘ לו‬him’. 174 Greek manuscripts 731 106 318. 175 In Rahlfs’ edition, these passages contain ὅτι. 176 See subsection 2.2.2.1. 177 Again, the rendering in Greek depends on how one interprets the Hebrew: with ‘him’ or ‘not’, and the continuation with either adversative or recitative conjunction, accordingly. Either Hebrew read ‘said to him (that) give’ or ‘said no but give’. Cf. the reference to verse 10: 19 above.

Clause connections

5: 5

79

διὰ τοῦτο οὐκ ἐπιβαίνουσιν οἱ ἱερεῖς Δαγὼν… ἕως τῆς ἡμέρας ταύτης, ὅτι ὑπερβαίνοντες ὑπερβαίνουσιν. ὅτι] ἀλλά L; om. the whole clause 44–107–125–610 (=MT)

ⲉⲧⲃⲉ ⲡⲁⲓ ⲛⲉⲟⲩⲏⲏⲃ ⲛⲇⲁⲅⲱⲛ ⲙⲉϩⲟⲩⲉϩ ⲣⲁⲧⲟⲩ ⲉϫⲛ …ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛϩⲟⲟⲩ· ⲁⲗⲗⲁ ϩⲛ ⲟⲩⲡⲱⲣϫ̄ · ⲉϣⲁⲩⲡⲉⲣⲉϣ ⲣⲁⲧⲟⲩ ⲉⲃⲟⲗ ϩⲛ̄ ⲧⲡⲉⲛⲛⲏ Here, some Greek mss offer the conjunction ἀλλά like the Sahidic. This fits the context, and the change could happen independently in Sahidic and Greek manuscripts. 30: 22 οὐ δώσομεν αὐτοῖς ἐκ τῶν σκύλων ὧν ἐξειλάμεθα, ὅτι ἀλλ᾽ἢ ἕκαστος τὴν γυναῖκα αὐτοῦ καὶ τὰ τέκνα αὐτοῦ ἀπαγέσθωσαν… om. ὅτι L 509 44–74–125 245 460 ⲛⲧⲛⲛⲁϯ ⲛⲁⲩ ⲁⲛ ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ· ⲛⲁⲓ ⲛⲧⲁⲛϫⲓⲧⲟⲩ· ⲁⲗⲗⲁ ⲙⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ ϫⲓ ⲛⲧ ̄ ⲉϥⲥϩⲓⲙⲉ ⲙⲛ̄ ⲛⲉϥϣⲏⲣⲉ ⲛⲥⲉⲧⲱⲟⲩⲛ In this passage, the Sahidic ostensibly follows the Greek manuscripts that read without ὅτι, but again, the context of direct speech and the question concerning the OG undermine this argument for dependence.178 The other difference, variation between ἀλλʼ ἤ (rel) and ἀλλά (L), is not distinguishable in Sahidic since it renders both expressions with ⲁⲗⲗⲁ in this context. 2.1.3.4.b) γάρ-clauses and their renderings γάρ is a causal conjunction that is used postpositively. According to Denniston, this conjunction is (1) confirmatory and causal and (2) explanatory.179 In 1 Sam, the Greek-Coptic loanword ⲅⲁⲣ functions similarly as in Greek. There are only two occurrences in Greek 1 Sam. For some reason, in both cases the Sahidic manuscript SaA omits the conjunction ⲅⲁⲣ.

20: 30 Υἱὲ κορασίων αὐτομολούντων, οὐ γὰρ οἶδα ὅτι μέτοχος εἶ σὺ τῷ υἱῷ Ιεσσαι… SaM ⲡϣⲏⲣⲉ ⲛⲛⲉϩⲓⲟⲙⲉ ⲉⲧⲛⲏϫ ⲉⲃⲟⲗ ⲙⲏ ⲛϯⲥⲟⲟⲩⲛ ⲅⲁⲣ ⲁⲛ ϫⲉ ⲛⲧⲟⲕ ⲕⲟ ⲛϣⲃⲏⲣ ⲉⲡϣⲏⲣⲉ ⲛⲓⲉⲥⲥⲁⲓ̄ ̅ SaA ⲱ ⲡϣⲏⲣⲉ ⲛⲛⲉϩⲓⲟⲙⲉ ⲉⲧⲛⲏϫ ⲉⲃⲟⲗ ⲙⲏ ⲛϯⲥⲟⲟⲩⲛ ⲁⲛ ϫⲉ ⲕⲟ ⲛϣⲃⲏⲣ ⲉⲡϣⲏⲣⲉ ⲛⲓⲉⲥⲥⲁⲓ̄ ̅ Manuscript SaA omits the conjunction ⲅⲁⲣ without any support from the Greek tradition. The same manuscript also reads an interjection ⲱ ‘O!’ at the beginning, thus denoting a Greek vocative which has no equivalent in the Coptic language. SaA omits redundant ⲛⲧⲟⲕ ‘you’ later in the sentence.

178 See the reference to verse 10: 19 above. 179 Denniston, Greek Particles, 58. Thus also Mayser (Grammatik II, 3§164: 4), according to whom this conjunction is in most cases “begründend”, but also, in many cases, explanatory “nämlich”.

80

Description of the translation technique

28: 20 οὐ γὰρ ἔφαγεν ἄρτον ὅλην τὴν ἡμέραν καὶ ὅλην τὴν νύκτα ἐκείνην. SaM ⲙⲡⲉϥⲟⲩⲉⲙ ⲟⲉⲓⲕ ⲅⲁⲣ ⲙⲡⲉϩⲟⲟⲩ ⲙⲛ̄ ⲧⲉⲩϣⲏ ⲉⲧⲙⲙⲁⲩ SaA ⲙⲡⲉϥⲟⲩⲉⲙ ⲟⲉⲓⲕ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙⲙⲁⲩ Greek manuscripts A B O CII-242 b s 244 460 have γάρ like SaM. The manuscript SaA has a shorter text, omitting ⲅⲁⲣ but also some other words. Because this manuscript habitually omits and abbreviates without any support from the Greek tradition, SaA probably preserves a secondary reading ‘He did not eat bread on that day’ in this instance.180 2.1.3.4.c) δέ- and οὐδέ-clauses and their renderings In the Greek text of 1 Sam, the connective δέ or its negative form οὐδέ introduces ten clauses.181 This particle has the meaning ‘and’ as well as ‘but’, and Denniston maintains that it can denote all that lies between these two ends.182 In 1 Sam the Greek translator has used δέ only in passages where the meaning is clearly adversative.183 Greek

Coptic

Occurrences

δέ δέ οὐδέ οὐδέ οὐδέ

ⲇⲉ ⲁⲩⲱ ⲟⲩⲇⲉ ⲁⲩⲱ asyndeton

six times once once once once

In Sahidic this originally Greek connective is ubiquitous. Beside the occurrences of ⲇⲉ as an equivalent for δέ, it appears as an equivalent for καί. In Sahidic, the connective was used in normal coordination, without any clear adversative meaning according to Müller.184 However, in most cases something new, typically a new subject, appears in the clauses where ⲇⲉ is used.

180 See section 3.3 for a description of the characteristics of the Sahidic mss. 181 Seven times δέ and three times οὐδέ. In the Greek text of 2Samuel, δέ occurs 19 times. In the Sahidic text, 15 times the equivalent is ⲇⲉ, twice the equivalent is asyndeton, once ⲁⲩⲱ, and once the Sahidic mss are divided between asyndeton (SaM) and ⲇⲉ (SaA). In 2 Samuel, οὐδέ is used only thrice to connect clauses, and in all these passages ⲟⲩⲇⲉ is the Sahidic equivalent. 182 Denniston, Greek Particles, 162. Thus also Blass/Debrunner/Rehkopf, Grammatik, §447. 183 Aejmelaeus, On the Trail, 128–9. According to her, the translator of 1 Sam translates conjunctions in a formally literal manner compared to the translators of the Pentateuch. 184 Müller (“Greek Connectors”, 10) discusses Coptic ⲇⲉ from the perspective that it represents two Greek conjunctions taken into Coptic: δέ and τε. This is a situation where translated texts add to the information of how a word was understood and used in Coptic since the translator of 1 Sam used ⲇⲉ so often as the equivalent for καί.

Clause connections

81

In two clauses,185 the Sahidic equivalent is ⲁⲩⲱ. In both of these cases, Greek variants attest the presence of καί. Since the conjunction ⲁⲩⲱ is often an equivalent for καί, the possibility of a different Vorlage cannot be excluded. 13: 21 καὶ ἦν ὁ τρυγητὸς ἕτοιμος τοῦ θερίζειν· τὰ δὲ σκεύη ἦν τρεῖς σίκλοι εἰς τὸν ὀδόντα, … ⲡϫⲱⲱⲗⲉ ⲇⲉ ⲏ ⲡⲱϩⲥ̄ ⲛⲉⲁϥⲉⲓ ⲉⲃⲟⲗ ⲉⲟϩⲥ̄ϥ̄ ⲛⲉⲩⲥⲕⲉⲩⲏ ⲇⲉ ⲛⲉⲣⲉ ⲧⲟⲃϩⲉ ⲧⲟⲃϩⲉ ⲛⲁⲉⲣϣⲟⲙⲧⲉ ⲛⲥⲁⲧⲉⲉⲣⲉ ⲛϩⲣⲏϣⲉ This example illustrates the complexities of the δέ-clause and its rendering. In Greek, the conjunction lies between the article and the noun, whereas in Coptic this position is not possible. It can also be observed how the same particle ⲇⲉ is an equivalent for καί in the preceding clause. 24: 18 …ὅτι σὺ ἀνταπέδωκάς μοι ἀγαθά, ἐγὼ δὲ ἀνταπέδωκά σοι κακά. σὺ]+ μεν L 554 ϫⲉ ⲛⲧⲟⲕ ⲙⲉⲛ ⲁⲕⲧⲱⲱⲃⲉ ⲛⲁⲓ ⲛϩⲉⲛⲡⲉⲧⲛⲁⲛⲟⲩⲟⲩ ⲁⲛⲟⲕ ⲇⲉ ⲁⲓⲧⲱⲱⲃⲉ ⲛⲁⲕ ⲛϩⲉⲛⲡⲉⲑⲟⲟⲩ In this passage, the Sahidic text probably had a Vorlage similar to the text of L 554 which attests the Greek μεν – δέ construction. There are no further passages of this kind in 1 Sam. 16: 8

καὶ εἶπεν Οὐδὲ τοῦτον ἐξελέξατο κύριος.

ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ϫⲉ ⲙⲡⲉ ⲡϫⲟⲉⲓⲥ ⲥⲉⲧⲡ̄ ⲡⲉⲓⲕⲉⲧ In this passage, the story is about the sons of Jesse, how they came to Samuel when he was looking for the one that God had chosen to become king of Israel. The translator uses asyndeton with a negative perfect to render οὐδέ.186 It fits well with direct speech, and the result is an idiomatic rendering ‘Samuel said: “The Lord has not chosen this one”’. Note the situation in Bohairic: “οὐδέ is always transliterated, but μηδέ is usually rendered by οὐδέ followed by the negative Bo[hairic] prefix ⲙⲡⲉ.”187 15: 29 καὶ οὐκ ἀποστρέψει οὐδὲ μετανοήσει,… ⲁⲩⲱ ⲡϫⲟⲉⲓⲥ ⲛⲁⲕⲧⲟϥ ⲁⲛ ⲟⲩⲇⲉ ⲛϥⲛⲁϣⲛϩ̄ ⲧ̄ ⲏϥ ⲁⲛ In this example, two negative statements are conjoined. The first clause is rendered with ⲁⲩⲱ + negative future. According to Feder, ⲟⲩⲇⲉ is an equivalent for καὶ οὐ(κ) in cases where several negative clauses exist in a sequence.188 Actually, Feder’s statement does not seem to apply to 1 Sam:

185 30: 3 δέ, 25: 15 οὐδέ. 186 In Greek, ms 44 has a variant reading with only οὐκ. 187 Melvin Peters, Textual Character, 130. 188 Feder, Biblia Sahidica, 84. Waheeb Atella Girgis (Greek Words, 162–3) gives examples of ⲟⲩⲇⲉ as a rendering of καί, μή, οὐ μή, μηδέ and μήτε.

82

Description of the translation technique

only two instances, 25: 7 and 30: 12, read ⲟⲩⲇⲉ in such a context.189 In this example, the Sahidic translator has followed his source text strictly, using ⲟⲩⲇⲉ to render οὐδέ. 2.1.3.4.d) ἤ-clauses and their renderings In my corpus, there are six clauses introduced with the disjunctive conjunction ἤ. Five of these are rendered with ⲏ, and once there is only asyndeton in Sahidic. According to R. Kühner and B. Gerth, this Greek conjunction expresses difference, and it can be used where the meaning is ‘or’ as well as in comparative meaning ‘than, as’.190 In Sahidic, the conjunction ⲏ is used with the meaning ‘and, or’, and does not appear in comparative constructions.191 According to Feder, in Jeremiah ⲏ often renders καὶ μή and καὶ ἐάν when several questions or conditionals come in a sequence.192 12: 3

μόσχον τίνος193 εἴληφα ἢ ὄνον τίνος εἴληφα, ἢ τίνα κατεδυνάστευσα ὑμῶν…

ⲁⲓϫⲓ ⲙⲁⲥⲉ ⲛⲧⲉ ⲗⲁⲁⲩ ⲙⲙⲱⲧⲛ̄ ⲏ ⲁⲓϫⲓ ⲟⲩⲉⲓⲱ ⲛⲧⲉⲛ ⲗⲁⲁⲩ ⲙⲙⲱⲧⲛ̄ ⲏ ⲁⲓϫⲓ ⲗⲁⲁⲩ ⲙⲙⲱⲧⲛ̄ ⲛϫⲛⲁϩ This is part of a farewell speech in which Samuel asks the people to say if he has done something wrong to them. In this case, the context is that of a question, and ⲏ is used to introduce the next question in a succession of many. Four of six total occurrences of ἤ are found in this verse. 26: 10 Ζῇ Κύριος, ἐὰν μὴ Κύριος παίσῃ αὐτόν, ἢ ἡ ἡμέρα αὐτοῦ ἔλθῃ καὶ ἀποθάνῃ, ἢ εἰς πόλεμον καταβῇ… παίσῃ] παιδεύσῃ B Mmg V 46 a-799 b 55 244 342; παιδεύσει O 799 460 ⲡϫⲟⲉⲓⲥ ⲟⲛϩ·̄ ϫⲉ ⲉⲣⲉⲧⲙ̄ ⲡϫⲟⲉⲓⲥ ϯⲥⲃⲱ ⲛⲁϥ· ⲡⲉϥϩⲟⲟⲩ ⲛⲁⲉⲓ ⲛϥ̄ ⲙⲟⲩ ⲏ ⲛϥ̄ ⲃⲱⲕ ⲉⲃⲟⲗ· ⲉⲡⲡⲟⲗⲩⲙⲟⲥ… Two of six cases of ⲏ occur in this verse, and the first ἤ may not have existed in the translator’s Vorlage. In Sahidic, the construction of this sentence is clear. With an oath formula, David asserts that, if the Lord will not teach Saul, Saul’s day will come, and he will die, or he will go to war and be destroyed. Since there are no Greek mss that omit the first ἤ, its omission in 189 For example, asyndeton with a negative verbal form is used six times: 4: 20, 9: 4bis, 15: 3, 22: 8, 24: 12. 190 Kühner/Gerth, Ausführliche Grammatik II/2, §537. 191 Layton (Grammar, §145) does not mention a comparative function, nor is there evidence for such in 1 Sam. 192 Feder, Biblia Sahidica, 84–5. 193 Greek manuscripts L 55 add υμων here, and one could argue that Sahidic shares this reading. This addition, however, is typical of this translator, occurring within direct speech and explicitations. Therefore, the reading has essentially no text-critical value. See also section 2.2.4.2.

Clause connections

83

Sahidic probably results from the translator clarifying the sentence structure. Text-critical note: Sahidic renders the verb ‘teach, discipline’ in the stipulation clause, and thus shows that the source text contained this verb. It is, however, impossible to say which of the two alternative Greek forms, future or aorist subjunctive, was in the Sahidic translator’s source text. 2.1.3.4.e) οὖν-clauses and their renderings There are two clauses connected with the conjunction οὖν in the Greek text of 1 Sam. This connective is used as an inferential conjunction.194 Additionally, οὖν often functions with a temporal meaning to introduce a new unit in the story, ‘next, then’. If the difference between these usages is that inferential closes a unit and temporal begins a new one, then both passages in 1 Sam are inferential. However, the difference is not mentioned by J. Blomqvist and P.O. Jastrup, but only the usage as a conclusive particle.195 Both occurrences in 1 Sam have ϭⲉ as the equivalent. Most typically, the source text of ϭⲉ reads οὖν in the Sahidic New Testament.196 19: 2 SaM SaA

Σαοὺλ ζητεῖ θανατῶσαί σε· φύλαξαι οὖν αὔριον πρωί, … Σαούλ] + ὁ πατήρ μου A O L d f 318 554 φύλαξαι οὖν] καὶ νῦν φύλαξαι L

ⲥⲁⲟⲩⲗ ⲕⲱⲧⲉ ⲛⲥⲁ ⲙⲟⲟⲩⲧⲕ·̄ ⲧⲉⲛⲟⲩ ϭⲉ ϩⲁⲣⲕ ̄ ⲙ̄ⲡⲛⲁⲩ ⲛϩ̄ ⲧⲟⲟⲩⲉ ⲛⲣⲁⲥⲧⲉ ⲥⲁⲟⲩⲗ ⲡⲁⲓⲱⲧ ϣⲓⲛⲉ ⲛⲥ̄ ⲁ ⲑⲉ ⲙ̄ⲙⲟⲟⲩⲧⲕ·̄ ⲧⲉⲛⲟⲩ ϭⲉ ϩⲟⲣⲕ ⲉⲣⲟⲕ ⲙ̄ⲡⲛⲁⲩ ⲛϩ̄ ⲧⲟⲟⲩⲉ ⲛⲣ̄ ⲁⲥⲧⲉ ⲥⲁⲟⲩⲗ ⲡⲁⲓⲱⲧ ⲕⲱⲧⲉ ⲛⲥⲁ ⲑⲉ ⲙⲙⲟⲟⲩⲧⲕ· ⲧⲉⲛⲟⲩ ϭⲉ ϩⲟⲣⲕ ⲉⲣⲟⲕ ⲙⲡⲛⲁⲩ SaB ⲛϩⲧⲟⲟⲩⲉ ⲛⲣⲁⲥⲧⲉ· In this case, οὖν is used in inferential meaning. Lucianic manuscripts read καὶ νῦν, whereas other Greek manuscripts offer οὖν.197 In Sahidic, the equivalent is ⲧⲉⲛⲟⲩ ϭⲉ which can render either of these two expressions individually. Lucianic manuscripts contain Hexaplaric approximations in this verse. Both καὶ νῦν and ὁ πατήρ μου conform the text more closely to the MT. SaAB have added the approximation ‘my father’. There is also another peculiar feature in SaA, namely ϣⲓⲛⲉ ⲛⲥ̄ ⲁ,198whereas SaM and SaB

194 Denniston, Greek Particles, 425–6. 195 Blomqvist/Jastrup, Grekisk grammatik, §300, 5. According to Blass/Debrunner/Rehkopf (Grammatik, §451), this develops the story further, especially leading back to the main topic. 196 There are 401 cases in the New Testament. Next in the list come usages without a corresponding Greek term (76 cases), rendering ἐτί (36 cases) and rendering δέ (30 cases). Wilmet, Concordance II: 3, 1681–4. 197 A 799 omit the word. 198 In Drescher’s apparatus, he supplies ⲡⲁⲓⲱⲧ ϣⲓⲛⲉ ⲛ̄ⲥⲁ as the reading of SaA, having stated in the introduction that his apparatus was created with the help of the manuscript itself as well as with photos. Thus, Drescher’s edition is more reliable than P. Ciasca’s, where SaA reads ⲡⲁⲓⲱⲧ ⲕⲱⲧⲉ ⲛ̄ⲥⲁ. Drescher, Kingdoms, 55; P. Ciasca, Sacrorum Bibliorum, 176.

84

Description of the translation technique

read the verb ⲕⲱⲧⲉ ⲛⲥ̄ ⲁ.199 This detail is interesting since, according to Boud’hors, two textual lines of the Gospel of Mark differ concerning exactly these words ϣⲓⲛⲉ ⲛⲥ̄ ⲁ/ⲕⲱⲧⲉ ⲛⲥ̄ ⲁ as they render ζητέω; the first appears in a newer, revised textual line, and the second features in the older, northern textual line.200 Taking into consideration the overall nature of SaA,201 it is logical that it has an approximation according to some Greek variant and uses a verb that, at least in the Gospel of Mark, is used in the newer (revised) textual line. 20: 31 …οὐχ ἑτοιμασθήσεται ἡ βασιλεία σου· νῦν οὖν ἀποστείλας λαβὲ τὸν νεανίαν,… οὐχ – σου] οὐχ ἑτοιμασθήσῃ σὺ οὐδὲ ἡ βασιλεία σου L SaM ⲛⲅ̄ ⲛⲁⲥⲁⲁϩⲉ ⲁⲛ ⲛⲧⲟⲕ ⲙⲛ̄ ⲧⲉⲕⲙⲛⲧ ̄ ⲣⲣ̄ ⲟ·̄ ⲧⲉⲛⲟⲩ ϭⲉ ϫⲟⲟⲩ ⲉⲧⲣⲉⲩⲉⲛ ⲡϣⲏⲣⲉ ϣⲏⲙ ⲛⲅⲛ̄ ⲁⲥⲟⲟⲩⲧⲛ ⲁⲛ ϩⲛ̄ ⲧⲕⲙ SaA ̄ ⲛⲧ ̄ ⲣⲣⲟ· ⲧⲉⲛⲟⲩ ϭⲉ ϫⲟⲟⲩ ⲛⲅⲙⲟⲩⲧⲉ ⲉⲣⲟϥ In this example, both Sahidic texts read the same word pair ⲧⲉⲛⲟⲩ ϭⲉ as connective, an accurate rendering of νῦν οὖν. Text-critical note: In the first clause, the verb choice in these Sahidic manuscripts is different, but their conjugation is the same. Both use 2nd person singular forms. These same variants (ⲥⲁⲁϩⲉ/ⲥⲟⲟⲩⲧⲛ) occur in Luke 13: 13.202 SaM, ‘you will not be established, you and your kingdom’ agrees with ‘you’ as found also in Lucianic mss and the MT. SaA shows more freedom in its wording, ‘you will not be established in your kingdom’. The object of the next clause (young man) is expressed with a suffixed preposition in SaA, whereas SaM uses a formal equivalent for τὸν νεανίαν, ⲡϣⲏⲣⲉ ϣⲏⲙ. SaM also conforms more accurately to Greek with the verbs: ‘send’ and ‘bring’ against ‘send’ and ‘call’ in SaA. There are no variants concerning these verbs in Greek mss.

199 Other occurrences of this verb in 1 Sam are as follows: 9: 3 SaM ϣⲓⲛⲉ, SaA ⲕⲱⲧⲉ; 10: 2 SaMA ϣⲓⲛⲉ; 10: 14 SaM ⲕⲱⲧⲉ; 10: 21 SaM ⲕⲱⲧⲉ; 13: 14 SaM ⲕⲱⲧⲉ; 16: 16 SaMBL ⲕⲱⲧⲉ; 22: 23 SaMA twice ⲕⲱⲧⲉ; 23: 10 SaMA ⲕⲱⲧⲉ; 23: 14 SaM ⲕⲱⲧⲉ; 23: 15 SaM ⲕⲱⲧⲉ; 23: 25 SaM ϯⲟⲩⲟⲓ ⲛⲥⲁ; 24: 3 SaM ⲕⲱⲧⲉ; 24: 10 SaM ⲕⲱⲧⲉ; 25: 26 SaMA ⲕⲱⲧⲉ; 25: 29 SaM ⲕⲱⲧⲉ; 26: 2 SaM ⲕⲱⲧⲉ; 26: 20 SaM ⲕⲱⲧⲉ; 27: 1 SaM ⲕⲱⲧⲉ; 27: 4 SaM ϯⲟⲩⲟⲓ ⲛⲥⲁ; 28: 7 SaMB ⲕⲱⲧⲉ. Thus one can see a clear dominion of ⲕⲱⲧⲉ in SaM, but still with some exceptions. SaB uses only the verb ⲕⲱⲧⲉ in those three passages where it is extant. 200 A. Boud’hors, “L’evangile de Marc”, 56. In a more recent article (A. Boud’hors, “‘Chercher’ dans les textes coptes”, 8–10) she provides examples from the Old Testament, and this verse (1 Sam 19: 2) is one among them. Interestingly, the dividing line is not only between revised and non-revised text but mainly a local issue. The verb typically used in northern dialects is ⲕⲱⲧⲉ whereas southern dialects prefer ϣⲓⲛⲉ. The manuscripts from the White Monastery are divided concerning this feature, as is also the case in the Gospel of Mark (Boud’hors, “‘Chercher’ dans les textes coptes”, 5). 201 See section 3.4. 202 Crum, Dictionary, 380b.

Clause connections

85

2.1.3.4.f) οὕτως-clauses and their renderings In my material, eight clauses begin with οὕτως. Technically, οὕτως is an adverb, and thus these clauses could be placed under the heading asyndeton. I have made a subgroup of these cases. Seven times, the equivalent used for οὕτως is ⲧⲁⲓ ⲧⲉ ⲑⲉ. The expression ⲧⲁⲓ ⲧⲉ ⲑⲉ is used in two ways: either as a conjunction ‘likewise, so’ with a main clause, relative or circumstantial verbal form, or as the beginning of a main clause ‘so too’ in correlated comparison.203 In cases of correlated comparison, an adverbial clause precedes or follows the main clause. Twice in 1 Sam, the phrase appears in a correlated comparison.204 In five cases, the expression ⲧⲁⲓ ⲧⲉ ⲑⲉ functions as a conjunction. In one remaining case, the construction facilitates comparison, and the rendering of οὕτως is ϩⲓⲛⲁⲓ ϩⲱ ‘so also’. 15: 33 Καθότι ἠτέκνωσεν γυναῖκας ἡ ῥομφαία σου, οὕτως ἀτεκνωθήσεται ἐκ γυναικῶν ἡ μήτηρ σου·

ⲉⲃⲟⲗ ϫⲉ ⲁ ⲧⲉⲕⲥⲏϥⲉ ⲉⲣ ϩⲁϩ ⲛⲥ̄ ϩ̄ ⲓⲙⲉ ⲛⲁⲧϣⲏⲣⲉ· ⲧⲁⲓ ⲧⲉ ⲑⲉ ⲉⲧⲉⲣⲉ ⲧⲉⲕⲙⲁⲁⲩ ⲛⲁⲣ ̄ ⲁⲧϣⲏⲣⲉ ⲉⲃⲟⲗ ϩⲛ ⲛⲉϩⲓⲟⲙⲉ In this example of correlated comparison, the protasis is causal in Sahidic and thus indicates that the Vorlage probably contained καθότι—the other possibility being καθώς which is attested in L 554. In the apodosis, οὕτως is rendered with ⲧⲁⲓ ⲧⲉ ⲑⲉ ‘so too’. 9: 13

ὡς ἂν εἰσέλθητε τὴν πόλιν, οὕτως εὑρήσετε αὐτὸν ἐν τῇ πόλει…

ⲉⲧⲉⲧⲛϣ ̅ ⲉⲧⲛⲛ ̄ ⲁⲛⲉⲣ ⲡϩ̄ ⲟⲩⲛ ⲛⲧⲡⲟⲗⲓⲥ· ⲧⲁⲓ ⲧⲉ ⲑⲉ ⲉⲧ ̄ ⲁϩⲉ ⲉⲣⲟϥ ϩⲛ ⲧⲡⲟⲗⲓⲥ This example is in direct speech, and ⲧⲁⲓ ⲧⲉ ⲑⲉ begins the main clause after a subordinate clause. 26: 24 καὶ ἰδοὺ καθὼς ἐμεγαλύνθη ἡ ψυχή σου σήμερον ἐν ταύτῃ ἐν ὀφθαλμοῖς μου, οὕτως μεγαλυνθείη ἡ ψυχή μου ἐνώπιον κυρίου,… ⲁⲩⲱ ⲕⲁⲧⲁ ⲑⲉ ⲛⲧⲁ ⲧⲉⲕⲯⲩⲭⲏ ⲁⲓⲁⲓ ⲙⲡⲟⲟⲩ ⲛⲁϩⲣⲁⲓ· ϩⲓⲛⲁⲓ ϩⲱ ⲉⲣⲉ ⲧⲁⲯⲩⲭⲏ ⲁⲓⲁⲓ ⲙⲡⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲙⲡϫⲟⲉⲓⲥ Here the Greek text has a καθώς – οὕτως construction, which is rendered with ⲕⲁⲧⲁ ⲑⲉ – ϩⲓⲛⲁⲓ ϩⲱ in Sahidic, resulting in an accurate rendering. 2.1.3.4.g) πλήν-clauses and their renderings In seven cases, πλήν occurs in Greek 1 Sam (three of those with ὅτι). According to Blass/Debrunner/Rehkopf, this is an adversative conjunction. H.G. Liddell and R.S. Scott give various meanings for this word when used

203 Layton, Grammar, §505–7. 204 8: 8 and 15: 33.

86

Description of the translation technique

as a conjunction: ‘save that, only, however, but’.205 In Sahidic, this borrowed connective is, according to Müller, primarily adversative.206 ShishaHalevy presents this word in Coptic as a “paragraph opener, subtextual initial-boundary marker”, with the meaning ‘yet, however’.207 12: 23–24

καὶ δείξω ὑμῖν τὴν ὁδὸν τὴν ἀγαθὴν καὶ τὴν εὐθεῖαν. πλὴν φοβεῖσθε τὸν κύριον καὶ δουλεύσατε αὐτῷ ἐν ἀληθείᾳ…

ⲛⲧⲁⲧⲥⲁⲃⲱⲧⲛ̄ ⲉⲧⲉϩⲓⲏ ⲉⲧⲛⲁⲛⲟⲩⲥ ⲁⲩⲱ ⲉⲧⲥⲟⲩⲧⲱⲛ·̄ ⲡⲗⲏⲛ ⲁⲣⲓ̈ϩⲟⲧⲉ ⲛϩⲏⲧϥ ⲙⲡϫⲟⲉⲓⲥ· ⲛⲧ ̄ ⲉⲧⲛⲣ̄ ϩ̄ ⲙ̄ϩⲁⲗ ⲛⲁϥ ϩⲛ̄ ⲟⲩⲙⲉ· There are three other cases like this; the Greek conjunction πλήν occurs as such in Sahidic with the meaning ‘only’. In this case, it is difficult to identify any adversative meaning. …καὶ σῶσαι χεῖρά μου ἐμοί. πλὴν ὅτι ζῇ Κύριος ὁ θεὸς Ισραηλ ὃς ἀπεκώλυσέν με σήμερον…

25: 33–34

ⲁⲩⲱ ⲉⲧⲉⲙϫⲱϩⲙ̄ ⲛⲧⲁϭⲓϫ ⲛⲁⲓ· ⲁⲩⲱ ϥⲟⲛϩ ⲛϭⲓ ⲡϫⲟⲉⲓⲥ ⲡⲛⲟⲩⲧⲉ ⲙⲡⲓⲏⲗ· ⲡⲁⲓ ⲛⲧⲁϥⲕⲱⲗⲩ ⲙⲙⲟⲓ ⲙⲡⲟⲟⲩ̄ In this verse, the equivalent for πλὴν ὅτι is ⲁⲩⲱ. In verse 1: 5 ⲇⲉ is used, and in 8: 9 SaA reads ⲁⲩⲱ, whereas SaM reads ⲇⲉ. One can conclude that πλὴν ὅτι was rendered as a coordinating conjunction in these cases. Text-critical note: Theoretically, the Sahidic translator could have rendered a now-lost Greek variant in these cases. In this passage, Aquila and Theodotion as well as O preserve καί (added before πλὴν ὅτι). In verse 1: 5 Theodotion and Symmachus have ἀλλά. Verse 8: 9, however, has no noteworthy variants, except for the omission of ὅτι by V 44.

2.1.4. Subordinate clauses and their renderings This group consists of clauses that cannot stand alone. Thus, they need a main clause to which they may be subordinated. Such a subordinate clause can precede or follow its main clause, or it can be embedded within it. 2.1.4.1. Relative clauses In Greek, relative pronouns are declined with respect to gender, number, and case.208 Relative clauses can express final, qualitative, and consecutive as well as conditional meanings.209 In 16 cases the word ἐἄν/ἄν occurs in 205 Blass/Debrunner/Rehkopf, Grammatik, §449. Liddell, H.G./Scott, R.S., Lexicon, 1419. 206 Müller, “Greek Connectors”, 10–11. 207 Shisha-Halevy, Grammatical Categories, 60. 208 For a more detailed description with certain exceptions, see Blass/Debrunner/Rehkopf, Grammatik, §294. 209 Blass/Debrunner/Rehkopf, Grammatik, §377–80.

Clause connections

87

Greek relative clauses in 1 Sam.210 Coptic relative clauses are constructed in a fundamentally different manner than in Greek. Whereas Greek clauses are subordinated by a relative pronoun, Coptic relatives are subordinated by verbal conversion, the converter being placed just before the converted clause. In Coptic, a resumptive pronoun is often added at the end of the converted expression in order to assure the right interpretation of the clause. Layton lists four functions of relative conversion: the simple attributive construction (the star that they saw), the appositive attributive clause construction (Mary, who has worked among you, …), the explanatory relative clause (Your servant, that is, me) and the articulated attributive construction (the person who…).211 In my corpus, there are 125 relative clauses and three out of the four types mentioned by Layton are used; only explanatory relative clauses do not occur as renderings of Greek relative clauses.212 In ten passages the Sahidic translator has used a circumstantial to render a Greek relative clause,213 and in five cases there is a more complicated issue concerning the rendering of the relative clause.214 25: 35 καὶ ἔλαβεν Δαυεὶδ ἐκ χειρὸς αὐτῆς πάντα ἃ ἔφερεν αὐτῷ,… ⲁⲩⲱ ⲇⲁⲇ ⲁϥϫⲓ̈ ⲛⲧⲟⲟⲧⲥ ̄ ⲛⲉⲛⲕⲁ ⲛⲓⲙ ⲛⲧⲁⲥⲉⲛⲧⲟⲩ ⲛⲁϥ The sentence above offers a typical example of a simple attributive construction. There are 67 cases like this in my material. 20: 37 καὶ ἦλθεν τὸ παιδάριον ἕως τοῦ τόπου τῆς σχίζης οὗ ἠκόντιζεν Ιωναθαν·

ⲁⲩⲱ ⲡϩⲙ̄ϩⲁⲗ ⲁϥⲓ ̅ ⲉϩⲣⲁⲓ· ⲉϫⲙ̄ ⲡⲙⲁ ⲉⲧⲉⲣⲉ ⲡⲥⲟⲧⲉ ⲛϩⲏⲧϥ·̄ ⲡⲙⲁ ⲛⲧⲁ ⲓⲱⲛⲁⲑⲁⲛ ⲛⲟϫⲥ ̄ ⲉⲣⲟϥ In this clause, which also belongs to the subgroup ‘simple attributive construction’, one can see how the translator inserts ⲡⲙⲁ into clauses where οὗ is used to denote a certain place.215 12: 16 …καὶ ἴδετε τὸ ῥῆμα τὸ μέγα τοῦτο ὃ ὁ κύριος ποιήσει… SaM ⲛⲧⲉⲧⲛⲛ ̄ ⲁⲩ ⲉⲡⲉⲓⲛⲟϭ ⲛϣⲁϫⲉ ⲡⲁⲓ ⲉⲧⲉⲣⲉ ⲡϫⲟⲉⲓⲥ ⲛⲁⲁⲁϥ̄ This is one of 16 clauses where the Sahidic translator has used an appositive attributive construction to render a relative clause. Surprisingly, in 6 cases, there is a difference between Sahidic manuscripts: manuscript SaM reads this

210 This marks the clause as a conditional relative clause, cf. Blomqvist/Jastrup, Grekisk grammatik, § 297, 13. One example is given below, verse 14: 42. 211 Layton, Grammar, §400. 212 Explanatory relative clauses do occur in some additions, see, for example, 9: 24, 17: 3, 17: 20, 17: 34, 24: 3. These are analyzed in section 2.2.4.3. 213 2: 34, 2: 35, 12: 21, 17: 36, 20: 31, 22: 23, 23: 22, 25: 7, 25: 11, 25: 16. 214 1: 28, 9: 6, 10: 19, 17: 25, 26: 21. 215 There are such cases in verses 3: 3, 6: 18, 9: 10 (only in SaA), 10: 5, 14: 4, 14: 47, 19: 3, 20: 37, 23: 13, 23: 22, 26: 5, 29: 4 and 29: 10.

88

Description of the translation technique

appositive attributive construction whereas manuscript SaA reads a simple attributive construction.216 The appositive attributive construction occurs only three times in SaA217 since, in addition to the six variant readings mentioned above, there are seven passages where SaA is lacunose. There is no obvious reason in the content of the clauses to explain why SaA sometimes uses the appositive construction but more frequently uses the simple attributive construction. This inconsistency can be most convincingly explained through variations arising in the transmission of the Sahidic translation; minor variants accrued over the course of centuries of scribal activity. This feature helps us in drawing a more detailed picture of the Sahidic manuscripts; also, some other patterns show that SaA favors simplified expressions.218 Perhaps this change resulted from the evolution of the Coptic language, but for that conclusion one needs clearer examples where the change and its direction are indisputable. 8: 9

…καὶ ἀπαγγελεῖς αὐτοῖς τὸ δικαίωμα τοῦ βασιλέως ὃς βασιλεύσει ἐπ᾽αὐτούς.

SaMS ⲁⲩⲱ ⲛⲅⲧ ⲙ̄ⲡⲣⲣ̄ ⲟ· ⲡⲁⲓ ⲉⲧⲛⲁⲣⲣ̄ ⲣ̄ ⲟ̄ ⲉϩⲣⲁⲓ ⲉϫⲱⲟⲩ· ̅ ̄ ⲁⲩⲟ ⲉⲣⲟⲟⲩ ⲙⲡⲇⲓⲕⲁⲓ̄ ⲱⲙⲁ SaA ⲁⲩⲱ ⲛⲅ̄ ⲧⲁⲩⲟ ⲉⲣⲟⲟⲩ ⲙ̄ⲡⲇⲓ̈ⲕⲁⲓⲱⲙⲁ ⲙ̄ⲡⲣⲣ̄ ⲟ· ⲉⲧⲛⲁⲣⲣ̅ ⲣ̅ ⲟ ⲉϩⲣⲁⲓ ⲉϫⲱⲟⲩ. In this example, SaA has the simple attributive construction, and SaMS attests the appositive attributive clause construction. 14: 42 Βάλετε ἀνὰ μέσον ἐμοῦ καὶ ἀνὰ μέσον Ἰωναθὰν τοῦ υἱοῦ μου· ὃν ἂν κατακληρώσηται Κύριος, ἀποθανέτω. ⲛⲉϫ ⲕⲗⲏⲣⲟⲥ ϩⲛ̄ ⲧⲁⲙⲏⲧⲉ· ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛⲓⲱⲛⲁⲑⲁⲛ ⲡⲁϣⲏⲣⲉ· ⲁⲩⲱ ⲡⲉⲧⲉⲣⲉ ⲡϫⲟⲉⲓⲥ ⲛⲁⲉⲓⲛⲉ ⲙ̄ⲡⲉⲕⲗⲏⲣⲟⲥ ⲉϫⲱϥ· ⲉϥⲉⲙⲟⲩ This is one example of 27 articulated attributive constructions that are found in my material. In 17 cases the fragmentary manuscript SaA does not preserve the text. In the remaining 10 clauses, variation occurs four times: three times SaM has the articulated attributive construction where SaA has a simple attributive and once the variation is vice versa.219 21: 3

Μηδεὶς γνώτω τὸ ῥῆμα περὶ οὗ ἐγὼ ἀποστέλλω σε…

ⲙⲡⲉⲣⲧⲣⲉ ⲗⲁⲁⲩ ⲉⲓⲙⲉ ⲉⲡϣⲁϫⲉ· ⲡⲉϯⲛⲁϫⲟⲟⲩⲕ ⲁⲛⲟⲕ ⲉⲧⲃⲏⲏⲧϥ Here, SaA reads the underlined word ⲉϯⲛⲁϫⲟⲟⲩⲕ, thus attesting the simple attributive construction. SaM has an articulated attributive construction. 29: 5

οὐχ οὗτος Δαυεὶδ ᾧ ἐξῆρχον ἐν χοροῖς λέγοντες Ἐπάταξεν Σαοὺλ…

216 6: 17, 8: 9, 8: 11, 14: 30, 21: 9, 30: 22. 217 In verses 8: 18, 10: 2 and 25: 25. 218 See sections 3.2 and 3.4. 219 14: 24, 20: 19, 21: 3, and vice versa in 29: 5. In the following passages, both Sahidic manuscripts read the articulated attributive: 14: 42, 19: 3, 25: 8, 30: 18, 30: 19, 31: 11.

Clause connections

89

SaM

ⲙⲏ ⲙⲡⲁⲓ ⲁⲛ ⲡⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲧⲁⲩⲙⲟⲟϣⲉ ϩⲓϩⲏ ⲙⲙⲟϥ ⲛϭⲓ ⲛⲉⲧⲭⲟⲣⲉⲩⲉ̅ ⲉⲩϫⲱ ⲙⲙⲟⲥ ϫⲉ ⲥⲁⲟⲩⲗ ⲁϥⲡⲁⲧⲁⲥⲥⲉ SaA ⲙⲏ ⲉⲙⲡⲁⲓ ⲁⲛ ⲡⲉ ⲇ̅ⲁⲇ ̅ ̅ ⲡⲉⲛⲧⲁⲩϩⲱⲥ ϩⲓⲑⲏ ⲙ̄ⲙⲟϥ ⲛϭⲓ ⲛⲉⲧⲭⲟⲣⲉⲩⲉ ⲉⲩϫⲱ ⲙ̄ⲙⲟⲥ ϫⲉ ⲥⲁⲟⲩⲗ ⲁϥⲡⲁⲧⲁⲥⲥⲉ In this passage, SaA has the articulated attributive construction and SaM attests the simple attributive. There is also a difference in vocabulary: the verb in SaA is ϩⲱⲥ ‘to sing, make music’ and in SaM ⲙⲟⲟϣⲉ ‘to walk, go’. Text-critical note: The Vorlage behind the verbs is clear: SaM follows the unanimous Greek testimony. SaA, however, seems to render a Greek source text nearer to Hebrew with the verb ‘to sing’.220 25: 7

…κείρουσίν σοι νῦν οἱ ποιμένες σου οἳ ἦσαν μεθ᾽ἡμῶν ἐν τῇ ἐρήμῳ, …

SaM ⲛⲉⲕϣⲟⲟⲥ ̄ ⲥⲉϩⲱⲱⲕⲉ ⲛⲁⲕ· ⲛⲁⲓ ⲉⲩⲛⲙ̄ⲙⲁⲛ ϩⲛ̄ ⲧⲉⲣⲏⲙⲟⲥ ⲛⲉⲕϣⲟⲟⲥ ̄ ϩⲱⲱⲕⲉ ⲛⲛⲉⲕⲉⲥⲟⲟⲩ· ⲛⲁⲓ ⲉⲛⲉⲩⲛⲙⲙⲁⲛ ϩⲛ̄ ⲧⲉⲣⲏⲙⲟⲥ SaA This is one of nine passages where a relative clause is rendered with a circumstantial.

Text-critical note: In this example, the manuscripts employ different tenses within the circumstantial clause: SaM has present, whereas SaA has preterit. The Greek has an imperfect. Both the subject and the object of the verb ϩⲱⲱⲕⲉ ‘to shear’ are different in the Sahidic manuscripts; SaM reads ‘your shepherds, they shear for you’, whereas SaA reads ‘your shepherds shear your sheep’. The Sahidic has no equivalent for ‘now’ as found in some Greek mss.221 I do not posit a dependence here, however, since the word ‘now’ already occurs as the first word in this verse, and thus it might have been omitted as redundant here. 20: 31 ὅτι πάσας τὰς ἡμέρας ἃς ὁ υἱὸς Ιεσσαι ζῇ ἐπὶ τῆς γῆς οὐχ ἑτοιSaM

μασθήσεται ἡ βασιλεία σου

ϫⲉ ⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲉⲣⲉ ⲡϣⲏⲣⲉ ⲛⲓ̈ⲉⲥⲥⲁⲓ̄ ·̅ ⲛⲁⲁⲁⲩ ⲉϥⲟⲛϩ̄ ⲛϩ̄ ⲏⲧⲟⲩ ϩⲓϫⲙ̄ ⲡⲕⲁϩ· ⲛⲅ̄ ⲛⲁⲥⲁⲁϩⲉ ⲁⲛ ⲛⲧⲟⲕ ⲙⲛ̄ ⲧⲉⲕⲙⲛⲧ ̄ ⲣⲣ̄ ⲟ ̄ SaA ϫⲉ ⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲉⲧϥⲛ ̄ ⲁⲁⲩ (sic) ⲉϥⲟⲛϩ ̄ ⲛϭⲓ ⲡϣⲏⲣⲉ ⲛⲓ̄ ⲉⲥⲥⲁⲓ ⲛⲅⲛ̄ ⲁⲥⲟⲟⲩⲧⲛ ⲁⲛ ϩⲛ̄ ⲧⲕⲙ ̄ ⲛⲧ ̄ ⲣⲣⲟ SaM renders the Greek relative with a circumstantial and SaA uses a simple attributive relative construction.222 This instance demonstrates an additional tendency in the translation of SaA; this text favors simple attributive constructions as renderings of Greek relative clauses.

220 This passage is analyzed in detail in section 3.2.5. 221 127* d 554 707. 222 Both forms are possible after a definite antecedent of time (Layton, Grammar, §407).

90

Description of the translation technique

Text-critical note: Here the word order also differs between the Sahidic manuscripts. The word ἅς is omitted in some Greek manuscripts223 but this does not affect the Sahidic readings. On the whole, SaA is shorter here, a typical feature in this manuscript.224 1: 28

κἀγὼ κιχρῶ αὐτὸν τῷ κυρίῳ πάσας τὰς ἡμέρας ἃς ζῇ αὐτός, χρῆσιν τῷ κυρίῳ.

ⲁⲛⲟⲕ ϭⲉ ϩⲱ ϯⲛⲁⲧⲁⲁϥ ⲛⲧ ̄ ⲁⲓⲟ̄ ⲙⲡϫⲟⲉⲓⲥ ⲛⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲙⲡⲉϥⲱⲛϩ̄ ⲉⲧⲣⲉϥϣⲱⲡⲉ ϩⲁⲣⲁⲧϥ ⲙⲡϫⲟⲉⲓⲥ. This is one example of five cases where the rendering of a relative clause is not straightforward in the Sahidic. In this passage the rendering is accurate but simplifies the relative clause ‘all the days that he will live’ into a prepositional phrase ‘all the days of his life’. 2.1.4.2. ἐάν-clauses Thirty clauses are introduced with the conjunction ἐάν. This conjunction, actually a compound of εί+ἄν, begins a conditional clause in Greek.225 Blass/Debrunner/Rehkopf mention two usages in Classical Greek: an eventual event (Eventualis) and an iterative event (Iterativus). It also states that different usages occur mixed in later phases of the Greek language.226 Blomqvist/Jastrup list future event and iterative event (in present and future) as being introduced by this word, but also mentions mixed cases.227 The clause headed by this conjunction is called the protasis, and is subordinate in relation to its main clause, called the apodosis which can precede or follow it. In ten clauses, the connective καί is also present, and, in two cases, the particle δέ is found.228 Typically, Greek καί does not occur at the start of the apodosis, when the apodosis follows the protasis. When this does occur against the general tendency of the language, this is referred to as an apodotic καί. If the conjunction is in the conditional clause and the apodosis comes later, the conjunction can be understood as linking either two subordinate clauses or the whole construction to the context. According to Blass/Debrunner/Rehkopf, concessive meaning can be expressed with κἄν, a compound of καὶ ἐάν.229 Müller describes different 223 509 245 707 (and 318, which also omits the following article). 224 See section 3.4. 225 Kühner/Gerth, Ausführliche Grammatik II/2, §575. Relative clauses reading with ἐάν/ ἄν are grouped among relative clauses, section 2.1.4.1. 226 Blass/Debrunner/Rehkopf, Grammatik, §371. 227 Blomqvist/Jastrup, Grekisk grammatik, §290. 228 In 1 Sam καί ἐάν is found in 2: 16, 2: 25, 6: 9, 14: 10, 14: 41, 17: 9bis, 20: 7, 20: 9 and 20: 14. ἐὰν δέ is found in 12: 15 and 17: 9. 229 Blass/Debrunner/Rehkopf, Grammatik, §374. This paragraph describes concessive conditionals. Müller addresses two categories used for concessive clauses: those belonging to adverbial clauses or to conditional clauses. M. Müller, “Contrast in Coptic”, 139–40.

Clause connections

91

concessives in his article and lists concessive constructions found in Sahidic: ⲕⲁⲓⲡⲉⲣ, ⲕⲁⲓⲧⲟⲓ, ⲕⲁⲛ, ⲉϣϫⲉ and circumstantial verbal form.230 There are a few passages that have καί ἐάν, but there are no concessive conditionals expressed with καί ἐάν or κἄν in my material.231 The Sahidic translator has rendered most of the conditional clauses with conditional verbal form: 25 out of 30 cases. The remaining renderings are twice circumstantial, twice conjunctive and once imperative. 2: 25

ἐὰν ἁμαρτάνων ἁμάρτῃ ἀνὴρ εἰς ἄνδρα, καὶ προσεύξονται ὑπὲρ αὐτοῦ πρὸς κύριον·

ⲉⲣϣⲁⲛ ⲟⲩⲣⲱⲙⲉ ⲉⲣⲛⲟⲃⲉ ⲉⲩⲣⲱⲙⲉ ϣⲁⲩⲧⲉⲃⲉϩ ⲡϫⲟⲉⲓⲥ ⲉϫⲱϥ In this example, the Sahidic translator has used a conditional to render the Greek conditional expression. Since Coptic has a special verbal form with a conditional meaning, no conjunction is needed to render the Greek conditional clause. The apodosis begins with καί in Greek but the Sahidic text lacks an equivalent. 3: 9

καὶ ἔσται ἐὰν καλέσῃ σε, καὶ ἐρεῖς Λάλει, κύριε…

ⲁⲩⲱ ⲉϥϣⲁⲛⲙⲟⲩⲧⲉ ⲉⲣⲟⲕ· ⲉⲕⲉϫⲟⲟⲥ ϫⲉ ⲡϫⲟⲉⲓⲥ ϣⲁϫⲉ This is an idiomatic rendering for the Greek expression where the first verb, a Hebraistic vestige of the Greek translation, is unnecessary in Greek as well as in Sahidic. As usual, Sahidic does not render the apodotic conjunction. …καὶ οὐκ ἔστιν μοι ἀγαθόν ἐὰν μὴ σωθῶ εἰς γῆν ἀλλοφύλων… ⲛⲧⲉⲧⲙ̄ ⲡⲉⲧⲛⲁⲛⲟⲩϥ ϣⲱⲡⲉ ⲙⲙⲟⲓ ⲉⲓϣⲁⲛⲧⲙⲧⲱⲟⲩⲛ ⲛⲧⲁⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ In this instance, the translator has used the negative form of the Sahidic conditional. The translator rendered two verbs for the one in the Vorlage. This is especially typical with the verb ⲧⲱⲟⲩⲛ ‘to arise, raise, carry’, as above.232

27: 1

17: 9

καὶ ἐὰν δυνηθῇ πρὸς ἐμὲ πολεμῆσαι καὶ ἐὰν πατάξῃ με, καὶ ἐσόμεθα ὑμῖν εἰς δούλους·

ⲁⲩⲱ ⲉϥϣⲁⲛϭⲙϭⲟⲙ ⲉⲙⲓϣⲉ ⲛⲙ̄ⲙⲁⲓ ⲛϥ̄ ⲡ ̄ ⲁⲧⲁⲥⲥⲉ ⲙ̄ⲙⲁⲓ ⲧⲛⲛ ̄ ⲁϣⲱⲡⲉ ⲛⲏⲧⲛ̄ ⲛϩⲙ̄ϩⲁⲗ In this example, the first clause is rendered in the same way as in the previous examples, with a conditional verbal form, whereas the second conditional clause is rendered with a conjunctive in Sahidic. The conjunctive continues the line of thought formulated in the previous clause.233 Thus, the 230 Müller, “Contrast in Coptic”, 141. 231 In 2 Samuel 18: 12, ⲕⲁⲛ occurs where the source text reads either καί or καί ἐάν. Since the reading with καί is found in the καίγε-text, the reading καί ἐάν now found in the Lucianic text is probably closer to the Vorlage of the Sahidic translator. 232 Cf. Drescher, Kingdoms, xv. I deal with all these cases in section 2.2.3.2. 233 According to Layton (Grammar, §351), “it signals only nexus and sequel after what came before”.

92

Description of the translation technique

translator has used a natural Sahidic formulation in this example. Here, the first word ⲁⲩⲱ can be understood as a coordinator that joins all three of these clauses to the previous text. The Greek text has an apodotic καί at the beginning of the apodosis, but the translator has formulated an expression without any conjunction, only a future verbal form with the apodosis. 1: 22

Ἕως τοῦ ἀναβῆναι τὸ παιδάριον, ἐὰν ἀπογαλακτίσω αὐτό, καὶ ὀφθήσεται τῷ προσώπῳ Κυρίου, … ἕως] pr οὐκ ἀναβήσομαι L

ϯⲛⲁϭⲱ ϣⲁⲛⲧⲉ ⲡϣⲏⲣⲉ ⲕⲟⲩⲓ ϫⲁⲧⲉ ⲛⲧⲁⲟⲙϫϥ ̄ ϫⲉ ⲉϥⲉⲟⲩⲱⲛϩ̄ ⲉⲃⲟⲗ ϩⲓⲑⲏ ⲙⲡϫⲟⲉⲓⲥ In this case, the Coptic conjunctive renders a Greek conditional clause. The Coptic conjunctive form above does not derive from the Greek conditional form, but, instead, serves to reproduce the Coptic ‘until’ conjugation (ϣⲁⲛⲧⲉ-) in the first clause. The translator has reformulated this expression by writing ϯⲛⲁϭⲱ ‘I will stay’ at the beginning of Hannah’s speech, thus making the line of thought clear. In the Lucianic text, οὐκ ἀναβήσομαι appears at the beginning of this verse. One can interpret ϯⲛⲁϭⲱ in van der Louw’s terms as an antonymic translation of this Lucianic reading.234 Alternatively, this could be an addition without any connection to the Lucianic text, inspired by the direct speech and smoothing the flow of the narrative. Obviously, the Greek source text was confusing in this verse, and the Lucianic recension and the Sahidic translation clarified the grammar so that it would be more comprehensible. 12: 14 ἐὰν φοβηθῆτε τὸν κύριον καὶ δουλεύσητε αὐτῷ …, καὶ ἦτε καὶ ὑμεῖς καὶ ὁ βασιλεὺς ὁ βασιλεύων ἐφ᾽ὑμῶν ὀπίσω Κυρίου πορευόμενοι·

ⲁⲣⲓϩⲟⲧⲉ ⲛϩⲏⲧϥ̄ ⲙⲡϫⲟⲉⲓⲥ ⲛⲧⲉⲧⲛⲣ̄ ϩ̄ ⲙ̄ϩⲁⲗ ⲛⲁϥ … ⲁⲩⲱ ⲛⲧ ̄ ⲱⲧⲛ̄ ⲙⲛ̄ ⲡⲉⲧⲛⲣ̄ ⲣ̄ ⲟ·̅ ⲉⲧⲛⲁⲣⲣ̄ ⲣ̅ ⲟ ⲉ̄ϫⲛ̄ ⲧⲏⲩⲧⲛ·̄ ⲛⲧⲉⲧⲛϣ ̄ ⲱⲡⲉ ⲉⲧⲉⲧⲛⲟ̄ ⲩⲏϩ ⲛⲥ̄ ⲁ ⲡϫⲟⲉⲓⲥ In this example, the Sahidic text clearly renders a Greek conditional clause with an imperative. In the Greek of verses 12: 14–15, there are two chains of conditionals; in verse 12: 14 a conditional sequence functions with a positive tone while in verse 12: 15 a second chain operates with a negative tone. The Sahidic translator has formulated this and the following clauses so that they have imperatives that suit the context.235 The use of verbs in this passage also shows that the Sahidic translator dealt with καὶ ἦτε καὶ ὑμεῖς καὶ ὁ βασιλεὺς ὁ βασιλεύων ἐφ᾽ὑμῶν ὀπίσω κυρίου πορευόμενοι as one segment; he rendered periphrastic ἦτε … πορευόμενοι with a Sahidic peri234 Van der Louw, Transformations, 65: “An antonymic translation is a transformation in which an SL [= source language] element is rendered by its TL [= target language] antonym plus a negation”. 235 Van der Louw (Transformations, 72) simply calls this kind of transformation “change of syntactic function”.

Clause connections

93

phrastic construction, the conjunctive of the verb ϣⲱⲡⲉ ‘to be’ and a circumstantial.236 20: 9

…ὅτι συντετέλεσται ἡ κακία παρὰ τοῦ πατρός μου τοῦ ἐλθεῖν ἐπὶ σέ, καὶ ἐὰν μή εἰς τὰς πόλεις σου, ἐγὼ ἀπαγγελῶ σοι. μή] + η A B

ϫⲉ ⲁ ⲡϫⲓⲛϭⲟⲛⲥ̄ ϫⲱⲕ ⲉⲃⲟⲗ ⲛⲧⲟⲟⲧϥ ⲙⲡⲁⲉⲓⲱⲧ· ⲉⲧⲣⲉϥⲉⲓ ⲉϩⲣⲁⲓ ⲉϫⲱⲕ· ⲉⲛϥϣⲟⲟⲡ ⲁⲛ ϩⲛ̄ ⲛⲉⲕⲡⲟⲗⲓⲥ̄· ⲁⲛⲟⲕ ϯⲛⲁⲧⲁⲙⲟⲕ ⲉⲡⲁⲓ Here, Sahidic apparently reads like A B in Greek, having the verb after μή. In any case, Sahidic grammar requires a verb, and thus the Sahidic verb cannot be understood as representative of a Greek parallel. The translator used a negative circumstantial to render the Greek negative conditional.237 2.1.4.3. εἰ-clauses In the present corpus, 42 clauses begin with the conjunction εἰ. This is a conditional conjunction meaning ‘if’, although its meaning varies with different verbal forms. When used with an indicative verb, εἰ expresses a ‘real’ condition ‘if’. In this case, the speaker does not indicate the certainty of the condition’s fulfillment. Blomqvist/Jastrup call this an “objective event.”238 When used with an indicative of a historical tempus, εἰ has the meaning of irrealis—a condition that is contrary to fact or an unfulfillable condition. When used with an optative verb, εἰ conveys potentiality.239 The condition is not fulfilled, but its fulfilment is possible. In these potential cases, the main clause has the modal particle ἄν.240 Another usage with the optative covers iterative past events.241 When used in indirect questions, this word has the meaning ‘whether, if’.242 The chart below lists different Sahidic renderings for εἰ-clauses. In the following section, I will present different renderings used by the Sahidic translator.

236 Layton, Grammar, §427. 237 Layton (Grammar, §497) calls this ‘undifferentiated cause’ (factual presupposition or general stipulation). 238 The term used is objektiva fallet (objective case), Blomqvist/Jastrup, Grekisk grammatik, §290, 4. 239 Kühner/Gerth, Ausführliche Grammatik II/2, §573, 574, 576. 240 Blomqvist/Jastrup, Grekisk grammatik, §290, 7. Their term is potentiala fallet. 241 Blomqvist/Jastrup, Grekisk grammatik, §290, 6. The term is iterative fallet. If an iterative event happens in the present or future, the conditional clause is introduced by ἐάν and a conjunctive verbal form. These cases belong to the previous subgroup according to my grouping. 242 Blass/Debrunner/Rehkopf, Grammatik, §§371, 368. Blomqvist/Jastrup, Grekisk grammatik, §288.

94

Description of the translation technique

Sahidic renderings for εἰ-clauses ⲉⲛⲉ ⲉϣⲱⲡⲉ question, without special marking ⲉϣϫⲉ conditional circumstantial preterit (apodosis) causative infinitive 9: 11

17 cases243 13 cases 5 cases 2 cases 2 cases once once once

καὶ λέγουσιν αὐταῖς Εἰ ἔστιν ἐνταῦθα ὁ βλέπων;

ⲡⲉϫⲁⲩ ⲛⲁⲩ ϫⲉ ⲉⲛⲉ ⲡⲉⲧⲛⲁⲩ ⲉⲃⲟⲗ ⲙⲡⲉⲓⲙⲁ In 16 cases, a question begins with εἰ, and the Sahidic translation has ⲉⲛⲉ as in this example.244 ⲉⲛⲉ marks the clause as a question.245 Most of these εἰclauses begin the question in direct speech. In some cases, however, the question began already in the previous clause, and an εἰ-clause continues the question. 14: 41 τί ὅτι οὐκ ἀπεκρίθης τῷ δούλῳ σου σήμερον; εἰ ἐν ἐμοὶ ἢ ἐν Ἰωναθὰν τῷ υἱῷ μου ἡ ἀδικία; εἰ] ἤ A B V O C’-52,313 527 509 d-44, 68, 120, 122 f s-130, 314 29 55 71 245 460 707 ⲉⲧⲃⲉ ⲟⲩ ⲙ̄ⲡⲉⲕϣⲁϫⲉ ⲙⲡⲟⲟⲩ ⲙⲛ̄ ⲡⲉⲕϩⲙ̄ϩⲁⲗ· ⲉⲛⲉⲣⲉ ⲡⲛⲟⲃⲉ ⲉⲣⲟⲓ· ⲏ ⲓⲱⲛⲱⲑⲁⲛ ⲡⲁϣⲏⲣⲉ· Here is another example of εἰ introducing a question, rendered with ⲉⲛⲉ in Sahidic. The Sahidic word ⲉⲛⲉⲣⲉ is a compound of ⲉⲛⲉ+ⲉⲣⲉ. Text-critical note: Some Greek readings offer ἤ instead of εἰ: A B V O C’-52,313 527 509 d-44, 68, 120, 122 f s-130, 314 29 55 71 245 460 707. The Sahidic translator is obviously rendering εἰ. Possibly, however, the Sahidic translator understood the η in his unaccented text as ἦ, not ἤ. 25: 34 …εἰ μὴ ἔσπευσας καὶ παρεγένου εἰς ἀπάντησίν μοι, τότε εἶπα Εἰ ὑπολειφθήσεται τῷ Ναβὰλ ἕως φωτὸς τοῦ πρωὶ οὐρῶν πρὸς τοῖχον. τότε εἶπα Εἰ] οὐκ ἄν L 245 554mg

ⲉⲛⲉ ⲙⲡⲉⲧⲱⲟⲩⲛ̄ ⲧⲁⲭⲏ ⲉⲉⲓ ⲉⲣⲁⲧ ⲉⲛⲉⲙⲛ̄ ⲗⲁⲁⲩ ⲉⲛⲉⲧⲏⲡ ⲧⲏⲣⲟⲩ ⲉⲛⲁⲃⲁⲗ ⲛⲁⲥⲉⲉⲡⲉ ϣⲁⲡⲛⲁⲩ ⲛϩ̄ ⲧⲟⲟⲩⲉ ̄ ⲉϥⲣⲙ ̄ ⲏ ⲉϩ̄ ⲟⲩⲛ̄ ⲉⲧϫⲟ

243 Sixteen times with a question and once in an irreal clause. 244 In verse 15: 32 there is ⲉⲓⲉ, an alternate form of ⲉⲛⲉ. Cf. Crum, Dictionary, 56b, 74b. 245 Layton, Grammar, §511.

Clause connections

95

In this example, two εἰ-clauses occur within the Greek text. The first clause has a negation whereas the second does not. In the Sahidic, the first clause is clear; it expresses the conditional clause with a negative verbal form ‘if you had not come quickly’. The second clause contains a Hebraism relating to the oath context, and, thus, the εἰ-clause should be interpreted as ‘that not’.246 Sahidic reads ‘there would be none left from those belonging to Nabal’. It apparently is dependent on Greek variants: L 245 554mg read οὐκ ἄν instead of τότε εἶπα εἰ.247 In this instance, either the Sahidic translator was dependent on the οὐκ ἄν reading, or he adapted his version according to the context—after the protasis, he added a contextually appropriate apodosis. 20: 29 καὶ νῦν εἰ εὕρηκα χάριν ἐν ὀφθαλμοῖς σου, διασωθήσομαι δὴ καὶ ὄψομαι τοὺς ἀδελφούς μου·

ⲧⲉⲛⲟⲩ ϭⲉ ⲉϣⲱⲡⲉ ⲁⲓϭⲉⲛ ⲭⲁⲣⲓⲥ ⲛⲁϩⲣⲁⲕ ⲉⲓⲉ ϯⲛⲁⲧⲱϣ ⲉⲃⲱⲕ ⲛⲧⲁⲛⲁⲩ ⲉⲛⲁⲥⲛⲏⲩ This corpus contains 13 clauses where εἰ is rendered with ⲉϣⲱⲡⲉ ‘if’.248 In five cases, the context is that of an oath.249 According to Blass/Debrunner/ Rehkopf, a clear Hebraismus is at hand when εἰ is used in oaths and declarations with the meaning ‘that not’.250 Naturally, these clauses would have been problematic for the Sahidic translator, and frequent strange renderings are to be expected. According to Coptic grammar, the conjunction ⲉϣⲱⲡⲉ occurs in the protasis of circumstantial and conditional statements.251 In my corpus, however, this conjunction only occurs in main clauses. Wolf-Peter Funk distinguishes between ⲉϣⲱⲡⲉ that, according to him, expresses an open condition ‘if it happens that’ and ⲉϣϫⲉ that expresses a closed condition ‘if it is true that’.252 3: 14

ὤμοσα τῷ οἴκῳ Ἠλεί Εἰ ἐξιλασθήσεται ἀδικία οἴκου Ἠλεὶ ἐν θυμιάματι καὶ ἐν θυσίαις ἕως αἰῶνος.

ⲁⲓⲱⲣⲕ ̄ ⲉϫⲉⲙ ⲡⲏⲓ ⲛϩ̄ ⲏⲗⲉ̅ⲓ ̅ ϫⲉ ⲉϣⲱⲡⲉ ⲥⲉⲛⲁⲕⲱ ⲉⲃⲟⲗ ⲙⲡϫⲓⲛϭⲟⲛⲥ ⲙⲡⲏⲓ ⲛϩⲏⲗⲓ ϩⲛ̄ ⲟⲩϣⲟⲩϩⲏⲛⲉ ⲁⲩⲱ ϩⲛ ϩⲉⲛⲑⲩⲥⲓⲁ ⲛϣⲁ ⲉⲛⲉϩ

246 Blass/Debrunner/Rehkopf, Grammatik, §454. Brock (Recensions, 251) cites the passages where the Lucianic text has changed this Hebraism, another instance: 1 Sam 17: 55. There are still many uncorrected passages. 247 There are other variants in Greek; for example, 509 reads οὐχ instead of the second εἰ. 248 Etymologically, this word comes from the circumstantial form of the verb ϣⲱⲡⲉ (ⲉⲥϣⲱⲡⲉ); W.-P. Funk, “On a semantic typology”, 412. 249 3: 14, 14: 45, 19: 6, 25: 22, 28: 10. In verse 26: 10, there is an oath, but the Greek conjunction used is ἐάν. 250 Blass/Debrunner/Rehkopf, Grammatik, §454: “Eindeutiger Hebraismus ist εἰ in Schwüren und Beteuerungen ‘dass nicht’.” 251 Layton, Grammar, §493. 252 Funk, “On a semantic typology”, 397–9. See also Müller, “Contrast in Coptic”, 146.

96

Description of the translation technique

This clause is part of an oath. Coptic has no true passive form, although the 3rd person plural frequently functions for this purpose, as in the instance above. 30: 15 καὶ εἶπεν πρὸς αὐτὸν Δαυείδ Εἰ κατάξεις με ἐπὶ τὸ γεδδοὺρ τοῦτο; ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲁϥ ϫⲉ ⲕⲛⲁⲉϣϫⲓⲧ ⲉϫⲉⲙ ⲡⲓⲕⲉⲇⲇⲟⲩⲣ· In this example, the translator has rendered the question using a 2nd person singular future and the auxiliary verb ⲉϣ ‘to be able to’. Unlike the Greek source, the Sahidic translation has no explicit grammatical marking to denote a question. Therefore, the Sahidic can be understood not only as a question but also as a command. There are five passages where a question beginning with εἰ is rendered thus into Sahidic, without any special interrogative signal.253 6: 3 SaM

καὶ εἶπαν Εἰ ἐξαπεστέλλετε ὑμεῖς τὴν κιβωτὸν διαθήκης Κυρίου θεοῦ Ἰσραήλ, μὴ δὴ ἐξαποστείλητε αὐτὴν κενήν,…

ⲡⲉϫⲁⲩ ⲛⲁⲩ ϫⲉ ⲉϣϫⲉ ⲧⲉⲧⲛⲛ ̄ ⲁϫⲟⲟⲩ ⲛⲧ ̄ ⲱⲧⲛ̄ ⲛⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡϫⲟⲉⲓⲥ ⲡⲛⲟⲩⲧⲉ ⲙⲡⲓⲏⲗ̄ ̄ ⲙⲡⲉⲣϫⲟⲟⲩⲥ ⲉⲃⲟⲗ ⲉⲥϣⲟⲩⲉⲓⲧ ⲡⲉϫⲁⲩ ϫⲉ ⲉϣϫⲉ ⲧⲉⲧⲛⲁϫⲟⲟⲩ ⲛⲧϭⲓⲃⲟⲩⲇⲟⲥ ⲛⲧⲇⲓⲁⲑⲏⲕⲏ ⲙ̄ⲡϫⲟⲉⲓⲥ ⲡⲛⲟⲩⲧⲉ SaA ⲙ̄ⲡⲓⲏ̅ ⲗ̄ ̄ ⲙⲡⲣϫⲟⲟⲩⲥ ⲉⲃⲟⲗ ⲉⲥϣⲟⲩⲉⲓⲧ In two cases, the translator has rendered an εἰ-clause with an ⲉϣϫⲉclause.254 This connective begins a ‘factual presupposition’ clause, meaning ‘since, if (as seems to be true)’.255 Text-critical note: In this case, SaA has an equivalent for διαθήκη like all Greek manuscripts except 29. The word ⲛⲁⲩ ‘to them’ without Greek equivalent is omitted. SaA also omits the rendering of ὑμεῖς, ⲛⲧ ̄ ⲱⲧⲛ.̄ SaA seems to omit some words here and there to shorten the text, and this happens regardless of the Greek text in many passages.256 The omission of διαθήκη in SaM is not typical of this manuscript, and I regard it as an error; in a chain of many words beginning with ⲛⲧ-, one was overlooked or ignored. 24: 20 εἰ εὕροιτό τις τὸν ἐχθρὸν αὐτοῦ ἐν θλίψει καὶ ἐκπέμψαι αὐτὸν ἐν ὁδῷ ἀγαθῇ, καὶ Κύριος ἀνταποτείσει αὐτῷ ἀγαθά,… ⲉⲣϣⲁⲛ ⲟⲩⲁ ϩⲉ ⲉⲡⲉϥϫⲁϫⲉ ϩⲛ̄ ⲟⲩⲑⲗⲓⲯⲓⲥ̅· ⲛϥⲟⲩⲟⲟⲧϥ̄ ⲉⲃⲟⲗ ϩⲛ̄ ⲟⲩϩⲓⲏ ⲉⲛⲁⲛⲟⲩⲥ· ⲡϫⲟⲉⲓⲥ ⲛⲁⲧⲱⲱⲃⲉ ⲛⲁϥ ⲛϩⲉⲛⲡⲉⲧⲛⲁⲛⲟⲩⲟⲩ·

253 The other four are in verses 14: 6, 14: 37, 18: 23, and 21: 9. 254 Another instance with this conjunction is 20: 8. 255 Layton, Grammar, §495. It can also begin a concessive clause; see Müller, “Contrast in Coptic”, 148–50. 256 See my chapter 3.4.

Clause connections

97

In this example, an optative Greek clause with εἰ is rendered by a conditional verbal form in Sahidic. In verse 17: 35, the translator has rendered εἰ and historical tempus with a conditional verb.257 6: 9

καὶ ὄψεσθε, εἰ εἰς ὁδὸν ὁρίων αὐτῆς πορεύσεται κατὰ Βαιθσάμυς, αὐτὸς πεποίηκεν ἡμῖν τὴν κακίαν ταύτην…

ⲛⲧⲉⲧⲛϯ ̄ ϩⲧⲏⲧⲛ̄ ⲉⲡⲉⲥⲙⲟⲉⲓⲧ· ⲉⲥⲛⲁⲧⲁⲁⲥ ⲉϩⲣⲁⲓ ⲉⲧⲉⲥϩⲓⲏ· ⲕⲁⲧⲁ ⲧⲉϩⲓⲏ· ⲛⲃⲏⲑⲥⲁⲙⲏⲥ· ⲉϣⲱⲡⲉ ⲥⲉ· ⲛⲧⲟϥ ⲡⲉ ⲛⲧⲁϥⲉⲓⲛⲉ ⲛⲁⲛ ⲛⲧⲉⲓⲛⲟϭ ⲛⲕⲁⲕⲓⲁ· In this case, there are more words in the Sahidic than in the Greek. The εἰclause is rendered with a circumstantial, without an equivalent for ὁρίων in the Sahidic.258 24: 7

Μηδαμῶς μοι παρὰ Κυρίου, εἰ ποιήσω τὸ ῥῆμα τοῦτο τῷ κυρίῳ μου, …

ⲛⲛⲉ ⲡⲉⲓϩⲱⲃ ϣⲱⲡⲉ ⲙⲙⲟⲓ ⲛⲧⲟⲟⲧϥ̄ ⲙⲡϫⲟⲉⲓⲥ· ⲉⲧⲣⲁⲉⲓⲣⲉ ⲙⲡⲉⲓϣⲁϫⲉ ⲙⲡϫⲟⲉⲓⲥ· This is the only case where an εἰ-clause is translated with a causative infinitive ‘to make me to do this thing to my lord’. 2.1.4.4. ἕως-clauses In my corpus, 11 clauses are introduced by the conjunction ἕως ‘until, during’. According to Mayser, ἕως was first used as a conjunction, and the use as a preposition developed later.259 As a preposition, ἕως can function in local and temporal senses. As a conjunction, the word possesses only a temporal meaning. It can express finality as well as simultaneity.260 In three of the clauses from my material, ἕως accompanies the modal particle ἄν.261 Sahidic has a formal equivalent for temporal ἕως: the limitative verbal form ϣⲁⲛⲧⲉ- ‘until’. According to Layton, “ϣⲁⲛⲧⲉ- expresses the limiting event beyond which the main event no longer continued, continues, or will continue”.262 When used as a preposition, ἕως is rendered with ϣⲁ ‘till, at, by, to’. 11: 11 καὶ ἔτυπτον τοὺς υἱοὺς Ἀμμὼν ἕως διεθερμάνθη ἡ ἡμέρα· ⲁⲩϩⲓⲟⲩⲉ ⲛⲥⲁ ⲛϣⲏⲣⲉ ⲛⲁⲙⲙⲁⲛ ϣⲁⲛⲧⲉ ⲡⲣⲏϯ ⲛⲧ ̄ ⲉϥϩⲙⲙⲉ

257 Payne (Critical and comparative, 48) does not mention this way of translating εἰ in his list of typical translations. 258 Brooke/McLean do not cite Sahidic in their apparatus with the Greek ms 707 omitting this word. 259 Mayser, Grammatik II,2, §133. 260 Blass/Debrunner/Rehkopf, Grammatik, §383, 455. 261 Blomqvist/Jastrup (Grekisk grammatik, §293, 2) mention that when referring to a single event in the future the clause has ἄν and subjunctive. 262 Layton, Grammar, §349.

98

Description of the translation technique

In this case, the Greek construction is simple: a conjunction at the head of the clause and a finite verb.263 The Sahidic translator uses the verbal form ϣⲁⲛⲧⲉ- which is an idiomatic rendering. 1: 23

…κάθου ἕως ἂν ἀπογαλακτίσῃς αὐτό· ϩⲙⲟⲟⲥ ⲛⲉ ϣⲁⲛⲧⲉⲟⲙ̄ϫ ⲙ̄ⲡϣⲏⲣⲉ ⲕⲟⲩⲓ Three cases in my material are like this one: ἕως ἄν with a Greek subjunctive. In Sahidic, an idiomatic solution is to use the ϣⲁⲛⲧⲉ- conjugation base. It is used in the meaning ‘until such time as’. 30: 4

…καὶ ἔκλαυσαν, ἕως ὅτου οὐκ ἦν ἐν αὐτοῖς ἰσχὺς ἔτι κλαίειν. ⲁⲩⲣⲓⲙⲉ· ϣⲁⲛⲧⲟⲩⲗⲟ ⲉⲩⲉϣϭⲙ̄ϭⲟⲙ ⲉⲣⲓⲙⲉ· There are two cases of ἕως ὅτου.264 In this passage the verbal form in Greek is indicative, and in the other passage a conjunctive is used. The Sahidic translator uses the ϣⲁⲛⲧⲉ- conjugation base in both verses. 14: 36 …καὶ διαρπάσωμεν ἐν αὐτοῖς ἕως διαφαύσῃ ἡ ἡμέρα,… ⲛⲧ ̄ ⲉⲛϩⲁⲣⲡⲁⲍⲉ ⲉⲃⲟⲗ ⲛϩⲏⲧⲟⲩ· ϣⲁ ⲡⲛⲁⲩ ⲛϩⲧⲟⲟⲩⲉ·̄ In this case, the translator has rendered the ἕως-clause with the prepositional expression ‘until the hour of dawn’. 2.1.4.5. καθώς- and καθά-clauses Eleven clauses are introduced with καθώς, and two clauses with καθά.265 These connectives mean ‘according as, just as’. Primarily, these terms occur in comparative cases, although Blass/Debrunner/Rehkopf also list καθώς together with ὡς as initiating a clause “ebenso wie, da”, with a weak causal meaning.266 In Sahidic, the usual rendering for καθώς and καθά is ⲕⲁⲧⲁ ⲑⲉ (ⲕⲁⲧⲁ ⲧϩⲉ) ‘according to the way’. It is used with a circumstantial or a relative clause.267 26: 20 ὅτι ἐξελήλυθεν ὁ βασιλεὺς Ἰσραὴλ ζητεῖν τὴν ψυχήν μου, καθὼς καταδιώκει ὁ νυκτικόραξ ἐν τοῖς ὄρεσιν. τὴν ψυχήν μου] ψύλλον ἕνα L 554mg SaM ϫⲉ ⲁϥⲕⲁ ⲣⲁⲧϥ̄ ⲉⲃⲟⲗ ⲛϭⲓ ⲡⲣⲣ̄ ⲟ̅ ⲙⲡⲓⲏⲗ̄ ·̄ ⲉⲕⲱⲧⲉ ⲛⲥⲁ ⲧⲁⲯⲩⲭⲏ· ⲉⲓⲟ ⲛⲑⲉ ⲛⲟⲩⲡⲏⲓ ⲛⲟ̄ ⲩⲱⲧ· ⲕⲁⲧⲁ ⲑⲉ· ⲉⲧⲉⲣⲉ ⲡⲃⲁⲓ ϯⲟⲩⲟⲓ ϩⲛ̄ ⲛⲉⲛⲧⲟⲩⲉⲓⲏ̄ SaB ⲡⲣⲣⲟ ⲙⲡⲓⲏⲗ̄ ̄ ⲁⲕⲕⲁ ⲣⲁⲧⲕ ⲉⲃⲟⲗ· ⲁⲕⲉⲓ ⲉⲕⲱⲧⲉ ⲛⲥⲁ ⲧⲁⲯⲩⲭⲏ· ⲁⲛⲟⲕ ⲟⲩⲟⲩϩⲟⲟⲣ ⲉϥⲙⲟⲟⲩⲧ· ⲁⲛⲟⲕ ⲟⲩⲡⲏⲓ ⲛⲟⲩⲱⲧ· ⲉϥϯⲟⲩⲟⲓ ϩⲉⲛ ⲉⲛⲧⲟⲩⲉⲓⲏ 263 The same construction is also found in 14: 36 and 15: 18. 264 22: 3 and 30: 4. 265 καθώς in 4: 9, 20: 3, 20: 13, 20: 35, 23: 11, 24: 14, 24: 20, 25: 26, 26: 20, 26: 24, and 28: 17 and καθά in 8: 7 and 17: 20. 266 “Als satzeinleitende Konjunktion kann ὡς und namentlich καθώς begründenden Sinn haben ‘ebenso wie = da’.” Blass/Debrunner/Rehkopf, Grammatik, §453. 267 Layton, Grammar, §493, 505.

Clause connections

99

This example is a simile using descriptive language. The rendering of the καθώς-clause is introduced by ⲕⲁⲧⲁ ⲑⲉ, and the animal mentioned in Greek as well as in SaM is ‘night raven’. Text-critical note: Notably, the Sahidic text contains as doublet the reading

ψύλλον ἕνα. The phrase occurs before καθώς in SaM, and elsewhere in SaB.

This doublet begins in the text of SaM with the circumstantial construction ⲉⲓⲟ ⲛⲑⲉ ‘me being like’. This is an approximation towards the Hebrew and probably of Hexaplaric origin. SaB is independent, reading ‘The king of Israel, you have come forth. You have come to seek after my life. I am a dead dog. I am a single flea stooping in the mountains.’ The phrase ‘I am a dead dog’ is placed before the mention of the flea. The comparison with a night raven is not found in SaB. Thus, both Sahidic witnesses attest to the correction according to the Hebrew here. 17: 20 καὶ ἔλαβε καὶ ἀπῆλθε καθὰ ἐνετείλατο αὐτῷ Ἰεσσαι… ⲁϥϫⲓ ⲛⲛⲉⲛⲕⲁ ⲁϥⲃⲱⲕ ⲕⲁⲧⲁ ⲑⲉ ⲛⲧⲁ ⲉⲓⲥⲥⲁⲓ̄ ̅ ϩⲱⲛ ⲉⲧⲟⲟⲧϥ ̄ This example was not part of the original Septuagint, but was added later to conform to the MT. The Sahidic translator adds the direct object after the first verb without any Greek variant support.268 2.1.4.6. μήποτε-clauses μήποτε appears twice269 as a conjunction in my material. This Greek term

means ‘lest … ever’. Blass/Debrunner/Rehkopf describe this connective as typical when expressing concern. 4: 9

κραταιοῦσθε καὶ γίνεσθε εἰς ἄνδρας ἀλλόφυλοι μήποτε δουλεύσητε τοῖς Εβραίοις μήποτε A O d-68, 122 554*] οπως μη rel270

271 ⲧⲱⲕ ⲙⲙⲱⲧⲛ̄ ⲛⲧⲉⲧⲛⲉ̄ ⲣϫⲱⲱⲣⲉ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ϫⲉ ⲛⲛ ̄ ⲉⲧⲛⲉ̄ ⲣϩⲙ̄ϩⲁⲗ ⲛⲛⲉϩⲉⲃⲣⲁⲓ̄ ⲟⲥ· In this case, the Sahidic translator uses ϫⲉ with a negative optative, which results in an exact rendering. Müller mentions avertive (negative purpose) clauses as introduced with ⲙⲏⲡⲟⲧⲉ or ⲙⲏⲡⲱⲥ and only additionally to these,

268 A typical feature in Sahidic, see section 2.2.3.1 on such additions. 269 4: 9 and 23: 22. 270 In this case it is possible that the critical text of Aejmelaeus will read ὅπως μή, not μήποτε. If this is the case, then the Sahidic translator has rendered the OG as in many passages in the NT where ϫⲉ/ϫⲉⲕⲁⲥ with conjunctive is used to render ὅπως μή (Matth 6: 18, Acts 20: 16, 1 Cor 1: 29). 271 Drescher has a typographical error here: ⲛ̄ⲛⲉⲧⲛ̄ⲉⲣⲙ̄ϩⲁⲗ. The manuscript clearly reads ⲛ̄ⲛⲉⲧⲛ̄ⲉⲣϩⲙ̄ϩⲁⲗ.

100

Description of the translation technique

is an optional ϫⲉ used.272 In 1 Sam, however, one clause is introduced by ⲙⲏⲡⲟⲧⲉ and the other with a plain ϫⲉ. 2.1.4.7. ὅπως-clauses In 1 Sam, four clauses are introduced with the conjunction ὅπως. This conjunction begins a subordinated final clause.273 The conjunction expresses purpose ‘for, (in order) that’. According to Liddell/Scott, ὅπως functions as a conjunction of manner as well as a final conjunction.274 The cases in my corpus are typical examples of the usage in final meaning. It is used with subjunctive in Greek. The Sahidic translator has twice used the final conjunction ϫⲉⲕⲁⲥ ‘that, in order that’, twice asyndeton. 9: 6

καὶ νῦν πορευθῶμεν, ὅπως ἀπαγγείλῃ ἡμῖν τὴν ὁδὸν ἡμῶν… ⲧⲉⲛⲟⲩ ϭⲉ ⲙⲁⲣⲟⲛ ⲛⲁϥ ϫⲉⲕⲁⲥ ⲉϥⲉⲧⲁⲙⲟⲛ ⲉⲡⲧⲱϣ ⲛⲧⲛϩⲓⲏ Here, the meaning is clear in both languages: a Greek final clause is rendered with a final clause in Sahidic, introduced by ϫⲉⲕⲁⲥ.275 Interestingly, Andrea Hasznos has concluded that this clause pattern in Sahidic, the conjunction ϫⲉⲕⲁⲥ with an optative (future 3 in Hasznos’ terms), is only used in Coptic translational literature. Accordingly, she sees it as deriving from the syntactic pattern of the Greek original.276

13: 9

Προσαγάγετε ὅπως ποιήσω ὁλοκαύτωσιν καὶ εἰρηνικάς·

ⲁⲩⲉⲓⲥ ⲛⲧⲁⲧⲁⲗⲟ ⲉϩⲣⲁⲓ ⲛⲛⲉϭⲗⲓⲗ ⲙⲛ ⲛⲉⲑⲩⲥⲓⲁ ⲛⲓ̈ⲣⲏⲛⲓⲕⲟⲛ In this case the translator has rendered the first verb as an imperative and the second as a conjunctive. A conjunctive form can be used to express final meaning, and thus, the result is an expression without double-marking.277

272 Müller, “Greek Connectors”, 51. In some passages, however, the construction used is ϫⲉⲕⲁⲥ with the negative conjunctive (2 Sam 17: 13 and several passages in the NT). 273 Blass/Debrunner/Rehkopf, Grammatik, §369. Blomqvist/Jastrup mention that ὅπως can also be used at the beginning of an independent, i. e., main, clause. In these cases, its context is that of an exhortation (Blomqvist/Jastrup, Grekisk grammatik, §285: 3). In my material, there are no such cases. 274 Liddell/Scott, Lexicon, 565. 275 Another such case is found in verse 6: 5. 276 Hasznos, Graeco-Coptica, 70, 74. 277 Another asyndeton occurs in verse 15: 15 where the translator has used ⲉ- and the infinitive in final meaning. The term double-marking is used by Müller, “Contrast in Coptic”, 45. It refers to a practice that Sahidic expresses some interclausal relations through verbal inflections plus a connective (of Greek origin), even though the relation would be clear without any connective.

Clause connections

101

2.1.4.8. ὅτε- and ὅταν-clauses There are three cases with the conjunction ὅτε and two with the conjunction ὅταν.278 The temporal conjunction ὅτε is rendered asyndetically with a temporal verbal form279 in Sahidic. Thus there is no double-marking in these passages. The conjunction ὅταν is also temporal, and it is rendered with conditional verbal forms in both passages. 1: 12

καὶ ἐγενήθη ὅτε ἐπλήθυνεν προσευχομένη ἐνώπιον Κυρίου, καὶ Ἠλεὶ ὁ ἱερεὺς ἐφύλαξεν τὸ στόμα αὐτῆς.

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛⲧ ̄ ⲉⲣⲉⲥⲱⲥⲕ ̄ ⲉⲥϣⲗⲏⲗ ⲙⲡⲉⲙⲧⲟ ⲉⲃⲟⲗ ⲙⲡϫⲟⲉⲓⲥ· ϩⲏⲗⲓ ⲡⲟⲩⲏⲏⲃ· ⲁϥϯⲧⲟⲟⲧϥ· ⲉⲣⲉⲛⲣⲱⲥ Here, the translator has rendered accurately. There is an equivalent for the redundant first phrase καὶ ἐγενήθη. In some passages, this phrase is not translated but only the conjunction and the next verb.280 2.1.4.9. ὅτι-, διότι- and καθότι-clauses My corpus contains 219 clauses that have ὅτι as connective, four clauses with the conjunction διότι and once καθότι.281 According to the grammars, the Greek conjunction ὅτι has various meanings: ‘that’ beginning an object clause, and a causal meaning ‘because’. One special usage, derived from the first-mentioned usage, introduces direct speech. In such passages, ὅτι equals our colon.282 In the Ptolemaic papyri studied by Mayser, the usage beginning an object clause is more common than the usage in causal meaning.283 Aejmelaeus has described the use of this conjunction in the LXX under the headings ὅτι recitativum and ὅτι causale.284 She describes and discusses ὅτι causale in the Pentateuch and other parts of the Septuagint, and concludes that, in the LXX, this conjunction also covers the function usually ascribed to the conjunction γάρ. Thus, ὅτι may introduce a clause with only a nuance of causality, more like an explanation or motivation. This usage of ὅτι results from the fact that Hebrew ‫ כי‬was rendered with ὅτι even in those passages where its meaning is indirectly or weakly causal.285 Interestingly, the Sahidic 1 Sam translator sometimes uses ⲅⲁⲣ as equivalent for ὅτι. 278 1: 12, 5: 4 and 6: 6. In verse 5: 4, variants in Greek read ὡς instead of ὅτε: L O. This does not make any difference on the Coptic side. The rendering can still be ⲛ̄ⲧⲉⲣⲉ-. ὅταν is found in 10: 7 and 17: 34. 279 Layton (Grammar, §344) calls this the precursive ⲛ̄ⲧⲉⲣⲉ-. 280 See subsection 2.1.3.1.f). 281 διότι in 17: 36, 21: 5, 28: 18, 30: 24. καθότι in 15: 33. 282 Mayser, Grammatik II.3, §155a; Kühner/Gerth, Ausführliche Grammatik II/2, §551, 4. 283 “Weit seltener als die … Substantivsätze nach den Verb. dicendi, sentiendi, declarandi” (Mayser, Grammatik II.3, §159,II). 284 Aejmelaeus, On the Trail, 11–41. 285 Aejmelaeus, On the Trail, 14–16. Aejmelaeus uses the term ‘indirect causality’ while Sipilä (Between Literalness and Freedom, 143) has the term ‘weak causality.’

102

Description of the translation technique

The conjunction ϫⲉ is an accurate equivalent for ὅτι. This indigenous Egyptian word stems from a verb meaning ‘to say’. According to Werner Vycichl, this word introduces names and designations, phrases and direct and indirect discourse.286 Westendorf adds to Vycichl’s list final, causal and consecutive meanings when used as a conjunction.287 In the Sahidic New Testament, ϫⲉ renders ὅτι in 1149 passages, and 20 times it occurs as the equivalent for διότι.288 The conjunction ϫⲉ functions as a causal conjunction ‘because’ to begin an object clause and to introduce direct speech. This function of introducing direct speech commonly occurs in Sahidic texts. In the Sahidic New Testament, ϫⲉ appears 2165 times without any equivalent (e. g. ὅτι) in the Greek text.289 Crum also mentions an explicative use of ϫⲉ, and in the Sahidic New Testament ϫⲉ is used thrice to render γάρ.290 Thus, the different usages of Greek ὅτι are all covered by ϫⲉ. In addition to these usages, ϫⲉ can be used with a final meaning to render ἵνα, and with a conditional meaning to render εἰ.291 However, the translator of 1 Sam has used not only ϫⲉ to render ὅτι, but, often, an additional word to further specify the relationship between clauses. In the following, I present some examples of different renderings of ὅτι-clauses in 1 Sam. Different renderings of ὅτι-clauses in 1 Sam causal and explanative ὅτι object clause ‘that’ ὅτι, varia διότι καθότι

107 times 21 times 5 times 70 16 times292 4 times once

ϫⲉ ⲉⲃⲟⲗ ϫⲉ ⲅⲁⲣ ϫⲉ ⲉⲃⲟⲗ ϫⲉ/ ϫⲉ/ ⲇⲉ/ ⲉⲃⲟⲗ ϫⲉ

286 Vycichl, Dictionnaire Étymologique, 322. He also compares this word to ‫ לאמר‬in Hebrew. The translators of the LXX used ὅτι (recitativum) to render ‫ לאמר‬only a few times (Aejmelaeus, On the Trail, 39). 287 Westendorf, Handwörterbuch, 410. 288 Wilmet, Concordance II: 3, 1544–5. 289 Wilmet, Concordance II: 3, 1545. This number includes not only those passages where ϫⲉ introduces direct speech, but also the passages where it introduces a biblical citation. 290 Crum, Dictionary, 746–7. For the number concerning the NT, see Wilmet, Concordance II: 3, 1543. 291 In the Sahidic New Testament, ϫⲉ occurs as the equivalent for ἵνα 203 times and as the equivalent for εἰ 34 times (Wilmet, Concordance II: 3, 1543–4). 292 In one case (24: 7), there is a homoioteleuton mistake in Sahidic, resulting in the omission of the rendering of a ὅτι-clause. In three cases (14: 30, 21: 4 and 25: 30), there was probably a different Vorlage, without ὅτι. The remaining cases in this group are found in 1: 8, 3: 9, 3: 10, 13: 19, 14: 29, 15: 35, 20: 17, 20: 26, 21: 6, 23: 17, 24: 6, and 28: 22.

Clause connections

6: 19

103

καὶ ἐπένθησεν ὁ λαός, ὅτι ἐπάταξεν Κύριος ἐν τῷ λαῷ πληγὴν μεγάλην σφόδρα.

ⲡⲗⲁⲟⲥ ⲇⲉ ⲁϥⲉⲣϩⲏⲃⲉ· ϫⲉ ⲁ ⲡϫⲟⲉⲓⲥ ⲉⲓⲛⲉ ⲛⲟⲩⲛⲟϭ ⲙⲡⲗⲩⲅⲏ ⲉⲙⲁⲧⲉ ⲉϫⲙ ⲡⲗⲁⲟⲥ In 107 cases, the meaning of ὅτι is either clearly causal as here, or more like an explanation. In both cases, the equivalent is ϫⲉ. Notably, in the following example, ⲅⲁⲣ appears in the Sahidic without a Greek parallel. I cannot see any difference in meaning between ϫⲉ and ⲅⲁⲣ in the clauses where Greek reads ὅτι to introduce an explanation. Obviously, the Sahidic translator assumed that ὅτι expresses a weakly causal relation, not just strictly causal subordination. 15: 26 Οὐκ ἀναστρέφω μετὰ σοῦ, ὅτι ἐξουδένωσας τὸ ῥῆμα Κυρίου,… ⲛϯⲛⲁⲕⲧⲟⲓ ⲁⲛ ⲛⲙ̄ⲙⲁⲕ· ⲉⲃⲟⲗ ϫⲉ ⲁⲕⲥⲉϣϥ̄ ⲡϣⲁϫⲉ ⲙⲡϫⲟⲉⲓⲥ In 21 passages, the translator has chosen the phrase ⲉⲃⲟⲗ ϫⲉ ‘because, for’, and thus restricted the number of possible interpretations when compared to simple ϫⲉ. ⲉⲃⲟⲗ ϫⲉ is a causal conjunction ‘because’ that begins a subordinate clause in Sahidic.293 In the Sahidic NT, ⲉⲃⲟⲗ ϫⲉ renders ὅτι, διὰ τό, ἐπεί, ἀνθ᾽ὧν, διότι, ἐπειδή and καθότι.294 In this verse, the meaning is obviously causal. The only καθότι in 1 Sam is also translated with ⲉⲃⲟⲗ ϫⲉ. 9: 20

μὴ θῇς τὴν καρδίαν σου αὐταῖς, ὅτι εὕρηνται·

ⲙⲡⲉⲣⲕⲁ ⲡⲉⲕϩⲏⲧ ϩⲓⲱⲟⲩ· ⲁⲩϩⲉ ⲅⲁⲣ ⲉⲣⲟⲟⲩ In five cases,295 the translator used the borrowed conjunction ⲅⲁⲣ as the equivalent for ὅτι. No Greek variants involve γάρ. 27: 4

καὶ ἀνηγγέλη τῷ Σαούλ ὅτι πέφευγεν Δαυεὶδ εἰς Γέθ·

ⲁⲩϫⲓ ⲙⲡⲟⲩⲱ̅ ⲇⲉ ⲛⲥⲁⲟⲩⲗ ϫⲉ ⲁ ⲇⲁⲩⲉⲓⲇ ⲡⲱⲧ ⲉϩⲣⲁⲓ ⲉⲅⲉⲑ There are 70 cases like this: ὅτι introducing a ‘that’-discourse clause, rendered with ϫⲉ. 24: 6

καὶ ἐπάταξεν καρδία Δαυεὶδ αὐτόν, ὅτι ἀφεῖλεν τὸ πτερύγιον τῆς διπλοΐδος αὐτοῦ.

ⲡϩⲏⲧ ⲛⲇⲁⲩⲉⲓⲇ ⲁϥⲙ̄ⲕⲁϩ ⲉⲣⲟϥ· ⲉⲧⲃⲉ ϫⲉ ⲁϥϣⲱϥ ⲙⲡⲧⲁⲡ ⲛⲧ ̄ ⲉⲭⲗⲁⲙⲩⲥ ⲛⲥⲁⲟⲩⲗ

293 Layton, Grammar, §495. 294 Wilmet, Concordance II: 1, 68. 295 9: 16, 9: 20, 12: 21, 16: 12 and 25: 35. In one of these five cases, verse 25: 35, the Vorlage is not self-evident, but Sahidic seems to suppose a Vorlage similar to L 245 554mg which attest ὅτι. Alternatively, the conjunction was added in Sahidic without any Greek source. This is not impossible, since this passage is direct speech, and most additions occur exactly in this kind of context. Usually, the added conjunctions are of Egyptian origin, but ⲅⲁⲣ is quite firmly established in Coptic vocabulary since it also renders other Greek conjunctions, such as ὅτι.

104

Description of the translation technique

This clause is causal. Sahidic reads ⲉⲧⲃⲉ ϫⲉ ‘because’, which, in the Sahidic NT, renders ὅτι, διὰ τό, ἐπεί, ἀνθ᾽ὧν, πρὸς τό and εἴπερ.296 In 1 Sam, it renders a ὅτι-clause only once. 1: 8

καὶ εἶπεν αὐτῇ Τί ἐστίν σοι ὅτι κλαίεις;

ⲡⲉϫⲁϥ ⲛⲁⲥ ϫⲉ ⲁϩⲣⲟ ϩⲓⲛⲁⲓ ⲧⲉⲣⲓⲙⲉ This is an example of a free translation of a ὅτι-clause. The context here is that of a conversation and this seems to have encouraged the translator to use an idiomatic expression ‘Why do you weep thus?’ 14: 29 ἰδὲ δὴ ὅτι εἶδον οἱ ὀφθαλμοί μου, ὅτι ἐγευσάμην βραχὺ τοῦ μέλιτος τούτου. δὴ ὅτι CI f 381] ὅτι 125; διοτι rel ⲁⲛⲁⲩ ⲉⲑⲉ ⲛⲧⲁ ⲛⲁⲃⲁⲗ ϫⲱⲧⲉ ⲉⲃⲟⲗ ϫⲉ ⲁⲓⲧⲉⲡ ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲓⲉⲃⲓⲱ̄ In this example, there are two ὅτι-clauses. The first one is rendered with ⲉⲑⲉ + relative clause ‘See how my eyes penetrated’.297 The Sahidic translation reads idiomatically, again in direct speech. The conjunction used in the second clause in Sahidic is causal ‘because’. Text-critical note: Concerning the first clause, it is not possible to argue for διότι or ὅτι or an omission in the Vorlage of the Sahidic translator. In this

case, the rendering is free, and one cannot retrovert the source text. The translator has favored functionality over formality in the rendering of this passage. 15: 35 καὶ οὐ προσέθετο Σαμουὴλ ἔτι ἰδεῖν τὸν Σαοὺλ ἕως ἡμέρας θανάτου αὐτοῦ, ὅτι ἐπένθει Σαμουὴλ ἐπὶ Σαούλ·

ⲁⲩⲱ ⲙⲡⲉ ⲥⲁⲙⲟⲩⲏⲗ ϭⲉ ⲕⲟⲧϥ̄ ⲉⲛⲁⲩ ⲉⲥⲁⲟⲩⲗ ϣⲁ ⲡⲉϩⲟⲟⲩ ⲙⲡⲉϥⲙⲟⲩ· ⲛⲉⲣⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲉⲣϩⲏⲃⲉ ⲉϫⲛ̄ ⲥⲁⲟⲩⲗ This is one of eight passages where the Sahidic renders ὅτι asyndetically. The Sahidic text reads without conjunction, beginning with a preterit of the verb ‘grieve’. No Greek variants omit ὅτι. If one reads the text so that Samuel was grieving the death of Saul, then, logically, no equivalent for ὅτι is required. Such a translation would not fit the context, and this seems to be the case with the Sahidic translator. The other possibility is to connect the grieving with the loss of kingship, i. e., Samuel was grieving the fate of Saul, not yet his death. 20: 17 καὶ προσέθετο ἔτι Ἰωναθὰν ὀμόσαι τῷ Δαυείδ, ὅτι ἠγάπησεν ψυχὴν ἀγαπῶντος αὐτόν.

296 Wilmet, Concordance II: 1, 125. 297 Literally, ‘See the manner in which my eyes penetrated’.

Clause connections

105

Δαυείδ] +διὰ τὸ ἀγαπᾶν αὐτόν L; + εν τω ηγαπηκεναι αυτον A O d-68,122,125 554298 SaM ⲓⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲟⲛ ⲁϥⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ ⲉⲱⲣⲕ ̄ ⲛⲇ ̄ ⲉ ⲙⲙⲟϥ ⲁⲩⲱ ̄ ⲁⲩⲉⲓⲇ· ⲉⲃⲟⲗ ϫⲉ ϥⲙ ⲁϥⲙⲉⲣⲉ ̄ ⲧⲉⲯⲩⲭⲏ ⲙⲡⲉⲧⲙⲉ ⲙⲙⲟϥ ⲓⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲁϥⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ ⲉⲱⲣⲕ ̄ ⲛⲇ SaA ̄ ⲁⲩⲉⲓⲇ· ϫⲉ ϥ̄ⲙⲉ ⲙⲙⲟϥ ϫⲉ ⲁϥⲙⲉⲣⲉ ̄ ⲧⲉⲯⲩⲭⲏ ⲙⲡⲉⲧⲙⲉ ⲙⲙⲟϥ In this passage, the Sahidic manuscripts differ in their readings. Both attest the approximation towards the MT also found in several Greek mss, and both Sahidic mss have formulated the sentences in their own way. In SaM, the text reads ‘Jonathan swore again to David, because he loved him, and he loved the soul of the one who loves him.’ The equivalent used for ὅτι is ⲁⲩⲱ, which is neither causal nor introduces a ‘that’-clause. In SaA, the conjunction ϫⲉ is used twice, and thus it is not evident which meaning one should choose for it; ‘Jonathan swore again to David that he loves him because he loved the soul of the one who loves him’ or ‘Jonathan swore again to David because he loved him, for he loved the soul of the one who loves him’.

Text-critical note: It is impossible to say whether one (or both) of the Sahidic manuscripts is secondary. In any case, the Hebrew-like reading is probably secondary here, and it might be the explanation for the different constructions: the Sahidic manuscripts acquired the reading from a marginal note and redacted the text independently of each other. Perhaps, the approximation lay in the source text of the Sahidic translator, and both Sahidic mss (or their predecessors) have reformulated it independently. Occasionally, in 1 Sam the Sahidic text offers readings that agree with the Hebrew text.299 20: 26 καὶ οὐκ ἐλάλησεν Σαοὺλ οὐδὲν ἐν τῇ ἡμέρᾳ ἐκείνῃ, ὅτι εἶπεν Σύμπτωμα φαίνεται μὴ καθαρὸς εἶναι, ὅτι οὐ κεκαθάρισται. ⲁⲩⲱ ⲙⲡⲉ ⲥⲁⲟⲩⲗ ϫⲉ ⲗⲁⲁⲩ ⲛϣⲁϫⲉ· ϩⲙ̄ ⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ· ϫⲉ ⲛⲧⲁϥϫⲟⲟⲥ· ϫⲉ ⲙ̄ⲉϣⲁⲕ· ⲟⲩⲧⲱϣ ⲛϫ ̄ ⲱϩⲙ̄· ⲡⲉ ⲛⲧⲁϥϣⲱⲡⲉ ⲙⲙⲟϥ· ⲉⲧⲓ̈ ⲉϥϣⲟⲟⲡ ϩⲙ̄ ⲡϫⲱϩⲙ̄ ⲙ̄ⲡⲁⲧⲉϥⲉⲓⲁⲁϥ ⲉⲃⲟⲗ Here, the translation is longer than the source text. The first ὅτι is rendered unremarkably with ϫⲉ. The second ὅτι has no direct equivalent, although the translator has used a ‘not yet’- verbal conjugation to render the clause. σύμπτωμα φαίνεται μὴ καθαρὸς εἶναι reads ⲙ̄ⲉϣⲁⲕ· ⲟⲩⲧⲱϣ ⲛϫ ̄ ⲱϩⲙ̄· ⲡⲉ ⲛⲧⲁϥϣⲱⲡⲉ ⲙⲙⲟϥ· ⲉⲧⲓ̈ ⲉϥϣⲟⲟⲡ ϩⲙ̄ ⲡϫⲱϩⲙ̄ ‘Perhaps it is a matter of uncleanness that has befallen him, he being still in the uncleanness’. In this clause, the rendering is an antonymic translation when compared to the source text: to render μὴ καθαρός ‘not clean’ the translator has chosen ‘unclean’.

298 Ms 44 (in d-group) leaves out the following ὅτι-clause and thus does not contain a doublet. 299 See section 3.3.

106 21: 6

SaM

Description of the translation technique

καὶ ἔδωκεν αὐτῷ Ἀβιμέλεχ ὁ ἱερεὺς τοὺς ἄρτους τῆς προθέσεως, ὅτι οὐκ ἦν ἐκεῖ ἄρτος, ὅτι ἀλλ᾽ἢ ἄρτοι τοῦ προσώπου,… ἄρτος] + ἕτερος L 554 ὅτι 2º] om L b 328 44–107–610 56 381 71 554300

ⲁⲩⲱ ⲁⲭⲓⲙⲉⲗⲉⲭ ⲡⲁⲣⲭⲓⲉⲣⲉⲩⲥ·̄ ⲁϥϯ ⲛⲁϥ ⲛⲛⲉⲟⲉⲓⲕ ⲛⲧ ̄ ⲉⲡⲣⲟⲑⲉⲥⲓⲥ· ⲉⲙⲛ̄ ϭⲉⲟⲉⲓⲕ ⲙⲙⲁⲩ̄ ⲛⲥⲁ ⲛⲉⲧⲕⲏ ⲉϩⲣⲁⲓ ϩⲓⲑⲏ ⲙⲡϫⲟⲉⲓⲥ· ⲁⲩⲱ ⲁⲭⲓⲙⲉⲗⲉⲭ ⲡⲟⲩⲏⲏⲃ· ⲁϥϯ ⲛⲁϥ ⲛⲛⲉⲟⲉⲓⲕ ⲛⲧ SaA ̄ ⲉⲡⲣⲟⲑⲉⲥⲓⲥ· ⲙⲛ̄ ϭⲉⲟⲉⲓⲕ ⲙⲙⲁⲩ̄ ⲛⲥⲁ ⲡⲟⲉⲓⲕ ⲉⲧⲕⲏ ⲉϩⲣⲁⲓ ϩⲓⲑⲏ ⲙⲡϫⲟⲉⲓⲥ· Although two ὅτι-conjunctions appear in the Greek, the Sahidic has an equivalent for only the first occurrence. It is not possible to know whether the other was omitted by the translator or was lacking in the Vorlage. The first ὅτι is rendered asyndetically, with a circumstantial verbal form that often has a causal meaning. Thus, there is no double-marking in this case. The circumstantial clause with causal meaning is common in Coptic, but, for some reason, it occurs in 1 Sam only seldom as a rendering for a ὅτιclause. SaA reads no circumstantial here, but omits ⲉ-. Text-critical note: The particle ϭⲉ ‘other’ seems to presuppose the reading with ἕτερος.301 Possibly this word was added because of the context: direct speech and immediate content of the clause. Interestingly, SaM reads ⲡⲁⲣⲭⲓⲉⲣⲉⲩⲥ̄ without a Greek variant.302 28: 22 …καὶ φάγε· καὶ ἔσται ἐν σοὶ ἰσχύς, ὅτι πορεύσῃ ἐν ὁδῷ. SaM ⲛⲅ̄ ⲟⲩⲱⲙ· ϫⲉ ⲉⲣⲉ ⲟⲩϭⲟⲙ ϣⲱⲡⲉ ⲙⲙⲟⲕ ϫⲉ ⲉⲕⲛⲁⲧⲁⲁⲕ ⲉⲧⲉϩⲓⲏ ⲙⲙⲟⲟϣⲉ ⲛⲅⲟ̄ ⲩⲱⲙ ϫⲉ ⲉⲣⲉ ⲟⲩϭⲟⲙ ⲛⲁϣⲱⲡⲉ ⲛⲁⲕ. ϫⲉ ⲕⲛⲁⲙⲟⲟϣⲉ ϩⲛ̄ ⲧⲉⲕϩⲓⲏ SaA Here, the meaning of ὅτι is probably explanative, ‘for you will go on the way’.303 Sahidic has ϫⲉ, which fits the context well. When reading only the Sahidic text one probably understands two successive ϫⲉ-clauses with a final meaning ‘so that strength comes to you to make yourself on the road’.304 28: 17–18

καὶ δώσει αὐτὴν τῷ πλησίον σου τῷ Δαυείδ. διότι οὐκ ἤκουσας φωνῆς Κυρίου…

ⲛϥⲧⲁⲁⲥ ⲙ̄ⲡⲉⲑⲓⲧⲟⲩⲱⲕ· ⲡⲉⲧⲛⲁⲁⲁϥ ⲉⲣⲟⲕ ⲇ̄ⲁⲇ ̄ ̄ · ⲉⲃⲟⲗ ϫⲉ ⲙⲡⲉⲕⲥⲱⲧⲙ̄ ⲛⲥⲁ ⲡⲉϩⲣⲟⲟⲩ ⲙ̄ⲡϫⲟⲉⲓⲥ

300 In O, there is a longer omission including this word. 301 Wilmet, Concordance II: 3, 1681–4. 302 Is there some connection with Mark 2: 26, where this story is referred to with the reading ‘high priest’ but with the proper name Abiathar? The New Testament was probably translated before 1 Sam, and thus the translator may have been familiar with the story. Only Mark and Sahidic 1 Sam speak of the ‘high priest’ in this connection. Other witnesses read ‘priest’. 303 Rendered thus in NETS by Taylor, “Text of Reigns”, 269. 304 Brooke/McLean (Old Testament in Greek, 98) have an erroneous marking, SaA does not omit what they maintain, the phrase καὶ ἔσται ἐν σοὶ ἰσχύς.

Clause connections

107

In four passages διότι appears as the connective.305 The Sahidic text offers the causal ⲉⲃⲟⲗ ϫⲉ ‘because’. 15: 33 Καθότι ἠτέκνωσεν γυναῖκας ἡ ῥομφαία σου, οὕτως ἀτεκνωθήσεται ἐκ γυναικῶν ἡ μήτηρ σου·

ⲉⲃⲟⲗ ϫⲉ ⲁ ⲧⲉⲕⲥⲏϥⲉ ⲉⲣ ϩⲁϩ ⲛⲥ̄ ϩ̄ ⲓⲙⲉ ⲛⲁⲧϣⲏⲣⲉ· ⲧⲁⲓ ⲧⲉ ⲑⲉ ⲉⲧⲉⲣⲉ ⲧⲉⲕⲙⲁⲁⲩ ⲛⲁⲣ ̄ ⲁⲧϣⲏⲣⲉ ⲉⲃⲟⲗ ϩⲛ ⲛⲉϩⲓⲟⲙⲉ In this example, the translator uses ⲉⲃⲟⲗ ϫⲉ ‘because’ for καθότι. The subordinate clause, a protasis, precedes its apodosis. Because ⲉⲃⲟⲗ ϫⲉ is a typical equivalent for causal ὅτι, the use of a causal conjunction here is not extraordinary. Usually, an apodosis beginning with ⲧⲁⲓ ⲧⲉ ⲑⲉ ‘thus, so too’ coordinates a protasis with the corresponding meaning ‘just as’.306 2.1.4.10. ὡς-clauses In Greek, the multipurpose conjunction ὡς can be used in many different contexts: it can begin a causal, temporal, final, consecutive, conditional, substantive or comparative clause.307 Takamitsu Muraoka considers the comparative-modal usage as historically original.308 My corpus contains 33 clauses headed by this conjunction. The following chart presents the different usages and their renderings. Greek ὡς ὡς ὡς ὡς ἄν ὡς ὡς

Sahidic temporal temporal comparative ‘that’ varia310

temporal verbal form free rendering309 ⲛⲑⲉ, ⲉⲑⲉ cond./circumstantial ϫⲉ

11 times once 11 times 5 times 2 times 3 times

24: 17 καὶ ἐγένετο ὡς συνετέλεσεν Δαυεὶδ τὰ ῥήματα ταῦτα λαλῶν πρὸς Σαούλ, καὶ εἶπεν Σαούλ… ⲛⲧⲉⲣⲉ ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲟⲩⲱ̅ ⲉϥⲧⲁⲩⲟ ⲛⲛⲉⲓϣⲁϫⲉ ⲉϩⲟⲩⲛ ϩⲛ̄ ⲥⲁⲟⲩⲗ· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲁϥ There are 13 clauses where the translator has understood ὡς as a temporal conjunction. In 11 of these 13 cases, the verbal form used in Sahidic is temporal without any conjunction. Because the Sahidic temporal form expli305 The other three are in 17: 36, 21: 5 and 30: 24. 306 Layton, Grammar, §506. 307 Blomqvist/Jastrup, Grekisk grammatik, §280. 308 T. Muraoka, “Use of ΩΣ”, 53–4. 309 25: 37. 310 2: 16, 13: 13, 14: 7. All three are presented below.

108

Description of the translation technique

citly subordinates a temporal clause, any additional equivalent for ὡς would be redundant. In this case, the translator has not rendered the superfluous verb ἐγένετο. The result is an idiomatic expression with temporal verbal form and ⲇⲉ as an equivalent for καὶ ἐγένετο ὡς.311 25: 37 καὶ ἐγένετο πρωί ὡς ἐξένηψεν ἀπὸ τοῦ οἴνου Ναβάλ, ἀπήγγειλεν αὐτῷ ἡ γυνὴ αὐτοῦ τὰ ῥήματα ταῦτα, … ἐγένετο] pr ὡς V ὡς] pr και L ϩⲧⲟⲟⲩⲉ ⲇⲉ ⲛⲧ ̄ ⲉⲣⲉϥϣⲱⲡⲉ ⲁϥⲛⲏⲫⲉ ̄ ϩⲁ ⲡⲏⲣⲡ·̄ ⲁⲩⲱ ⲧⲉϥⲥϩ̄ ⲓⲙⲉ ⲁⲥⲧⲁⲩⲟ ⲉⲣⲟϥ ⲛⲛⲉϣⲁϫⲉ In this example the clause has a different nuance in the Sahidic. The Greek says ‘It happened early when Nabal sobered up from the wine…’. The Sahidic text reads ‘When dawn came about, he sobered from the wine’. The Sahidic translator restructured this series of clauses to improve the flow. Although the second clause has ὡς, the translator uses the temporal for the first clause. The second clause has a Coptic perfect, even though ὡς subordinates it in the Vorlage. Thus, the third clause, ἀπήγγειλεν αὐτῷ in Greek is the apodosis for the subordinate clause and in Sahidic (ⲁⲩⲱ ⲧⲉϥⲥϩ̄ ⲓⲙⲉ ⲁⲥⲧⲁⲩⲟ ⲉⲣⲟϥ ‘and his wife told him’) an independent main clause coordinated with ⲁⲩⲱ with the previous main clause. Text-critical note: The Vorlage of the Sahidic translator possibly read ὡς before the first verb, like V in Greek. Brooke/McLean cite Sahidic in this connection. They also cite Sahidic as omitting the conjunction καί before ὡς. This latter citation does not indicate anything about the Vorlage of the Sahidic translator since he frequently used asyndeton as the rendering of καί. 6: 6

καὶ ἵνα τί βαρύνετε τὰς καρδίας ὑμῶν, ὡς ἐβάρυνεν Αἴγυπτος καὶ Φαραὼ τὴν καρδίαν αὐτῶν;

ⲁⲩⲱ ⲉⲧⲃⲉ ⲟⲩ ⲁⲧⲉⲧⲛϯ ̄ ⲉⲛϣⲟⲧ ⲛⲛ ̄ ⲉⲧⲛϩⲏⲧ· ⲛⲑⲉ ⲛⲧⲁ ⲫⲁⲣⲁⲱ ⲙⲛ ⲛⲉⲣⲙ̄ⲛⲕ̄ ⲏⲙⲉ ϯⲉⲛϣⲟⲧ ⲛⲛⲉⲩϩⲏⲧ In this example, ὡς is comparative-modal and represents, according to Muraoka, the most typical use of ὡς in the Greek Bible.312 In 1 Sam, this usage occurs 11 times.313 This verse preserves the conjunction ⲛⲑⲉ and a relative verbal form. There are ten cases where the equivalent in Sahidic is ⲛⲑⲉ ‘just as, in the manner of’, which can be used with a relative or circumstantial verbal form.314 Additionally, in verse 24: 11 there is ⲉⲑⲉ. 311 More such cases are in subsection 2.1.3.1.f). 312 Muraoka, “The Use of ΩΣ”, 54–5. Muraoka’s generalization may be valid in a wider context, but, in 1 Sam, the comparative-modal function is not the most frequently occurring usage of this conjunction. 313 In one of these cases (17: 43) the conjunction in Greek is ὡσει. 314 Layton, Grammar, §505.

Clause connections

15: 2

109

Νῦν ἐκδικήσω ἃ ἐποίησεν Ἀμαλὴκ τῷ Ἰσραήλ, ὡς ἀπήντησεν αὐτῷ ἐν τῇ ὁδῷ ἀναβαίνοντος αὐτοῦ ἐξ Αἰγύπτου.

ⲧⲉⲛⲟⲩ ϯⲛⲁϫⲓ ⲕⲃⲁ ⲙⲡⲁⲙⲁⲗⲏⲕ· ⲛⲛⲉⲛⲧⲁϥⲁⲁⲩ ⲙⲡⲓⲏⲗ̄ ·̄ ⲛⲑⲉ ⲛⲧⲁϥⲉⲓ ⲉⲃⲟⲗ ϩⲏⲧϥ̄ ϩⲛ̄ ⲧⲉϩⲓⲏ· ⲉϥⲛⲏⲩ ⲉⲃⲟⲗ ϩⲛ̄ ⲕⲏⲙⲉ· In this example, the translator has chosen ⲛⲑⲉ. Here, as well as in verses 24: 19 and 26: 23, Muraoka describes ὡς as epexegetical,315 and NETS renders this clause ‘how he encountered him’.316 The Sahidic translator has used ⲛⲑⲉ ‘just as, in the manner of’, introducing a comparison of some kind. This is one passage where one can understand the conjunction ὡς in different ways, and even a temporal meaning seems possible.317 9: 13

ὡς ἂν εἰσέλθητε τὴν πόλιν, οὕτως εὑρήσετε αὐτὸν ἐν τῇ πόλει πρὶν ἀναβῆναι αὐτὸν εἰς Βαμά…

ⲉⲧⲉⲧⲛϣ ̄ ⲁⲛⲉⲣ ⲡϩ̄ ⲟⲩⲛ ⲛⲧⲡⲟⲗⲓⲥ· ⲧⲁⲓ ⲧⲉ ⲑⲉ ⲉ̄ⲧⲉⲧⲛⲛ ̄ ⲁϩⲉ ⲉⲣⲟϥ ϩⲛ ⲧⲡⲟⲗⲓⲥ· ⲉⲙⲡⲁⲧⲉϥⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲃⲁⲙⲁ ὡς ἄν occurs five times in 1 Sam.318 The translator has rendered four of these with a conditional verbal form, and the fifth is presented below. 2: 13

καὶ ἤρχετο τὸ παιδάριον τοῦ ἱερέως ὡς ἂν ἡψήθη τὸ κρέας,…

ⲁⲩⲱ ϣⲁϥⲓ̈ ⲛϭⲓ ⲡϣⲏⲣⲉ ⲙⲡⲟⲩⲏⲏⲃ· ⲉⲣⲉ ⲡⲁϥ ⲛⲁⲡⲓⲥⲉ In this case, the translator has used a circumstantial future to render a ὡς ἄν-clause. This is an example of an adverbial circumstantial.319 18: 15 καὶ εἶδεν Σαοὺλ ὡς αὐτὸς συνίει σφόδρα,… ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲛⲁⲩ ϫⲉ ⲛⲧⲟϥ ⲇⲁⲩⲉⲓⲇ ϥⲟ ⲛⲣⲉⲙⲛϩⲏⲧ ⲉⲙⲁⲧⲉ In this example, as well as in verse 13: 11, ὡς begins a substantive clause ‘that’. In both cases the preceding verb is ‘to see’. 2: 16

Θυμιαθήτω πρῶτον ὡς καθήκει τὸ στέαρ, καὶ λαβὲ σεαυτῷ…

ⲙⲁⲣⲉ ⲧⲉⲑⲩⲥⲓⲁ ϣⲱⲡⲉ ⲕⲁⲗⲱⲥ· ⲁⲩⲱ ⲛⲅϫⲓ ⲛⲁⲕ In this instance, the ὡς-clause has only one word as its rendering in Sahidic, namely the Greco-Coptic word ⲕⲁⲗⲱⲥ. The Sahidic text is less specific in its description of the offering. Unlike the Greek, the Sahidic does not mention which part should be offered first. 315 Muraoka, “The Use of ΩΣ”, 61–2. Muraoka also comments on verse 28: 19, but, unfortunately, I am not able to discern whether he understands ὡς as epexegetical or as equaling a relative pronoun in this verse. 316 Taylor, “ Text of Reigns”, 259. 317 In Sahidic, ⲛⲑⲉ can also introduce an adverbial clause that expresses time. See Elanskaya’s article on the aspects of meaning of ⲛⲑⲉ in Coptic. Unfortunately, this article is in Russian. My thanks to Ivan Miroshnikov, who gave me the right page with English summary. Elanskaya mentions (p.33–4) as examples of temporal use Luke 12: 58, 20: 37 and 24: 32. In my view, at least the first and thid of her examples show ⲛⲑⲉ with temporal meaning. 318 2: 13, 9: 13, 10: 2, 10: 5 and 20: 12. 319 Layton, Grammar, §421.

110

Description of the translation technique

13: 13 …ὅτι οὐκ ἐφύλαξας τὴν ἐντολήν μου ἣν ἐνετείλατό σοι Κύριος· ὡς νῦν ἡτοίμασεν Κύριος τὴν βασιλείαν σου ἕως αἰῶνος ἐπὶ Ἰσραήλ.

ⲉⲃⲟⲗ ϫⲉ ⲙⲡⲉⲕϩⲁⲣⲉϩ ⲉⲧⲁⲉⲛⲧⲟⲗⲏ· ⲛⲧ ̄ ⲁ ⲡϫⲟⲉⲓⲥ ϩⲱⲛ ⲙⲙⲟⲥ ⲉⲧⲟⲟⲧⲕ·̄ ⲁⲩⲱ ⲧⲉⲛⲟⲩ· ⲛⲑⲉ ⲛⲧⲁ ⲡϫⲟⲉⲓⲥ ⲥⲃⲧ ̄ ⲉ ⲧⲉⲕⲙⲛⲧ ̄ ⲣⲣ̄ ⲟ· ⲉⲧⲣⲉⲥϣⲱⲡⲉ ⲛⲁⲕ ϣⲁⲃⲟⲗ Here, ⲛⲑⲉ is the equivalent for ὡς. Since there are no variant readings with καί, the translator probably added the conjunction ⲁⲩⲱ to transfer this expression into Sahidic. Usually, ⲧⲉⲛⲟⲩ ‘now’ is used with either ⲁⲩⲱ or ϭⲉ to begin a clause and this could explain the addition of ⲁⲩⲱ. Without the conjunction ⲁⲩⲱ to establish a new sentence, the word ‘now’ could easily be interpreted with the previous sentence. In his study of the transformations in the LXX, van der Louw states “Behind each transformation stands a literal rendering that has been rejected.”320 It is useful to consider what could be the literal rendering in those cases where the translator has reformulated the text. In many cases it soon becomes obvious why the translator has made the reformulation. Thus, in this verse ⲁⲩⲱ is added from the need to join ‘now’ in the Sahidic to the sentence it belongs to according to the Greek text. 14: 7

Ποίει πᾶν ὃ ἐὰν ἡ καρδία σου ἐκκλίνῃ· ἰδοὺ ἐγὼ μετὰ σοῦ, ὡς ἡ καρδία σοῦ καρδία μοῦ. μοῦ] + εκκλινον σεαυτον L

ⲁⲣⲓ ⲡⲉⲧⲉⲣⲉ ⲡⲉⲕϩⲏⲧ ⲟⲩⲁϣϥ̄ ⲧⲏⲣϥ̄ ⲉ̅ⲁⲁϥ· ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲛⲟⲕ ϯⲟⲩⲏϩ ⲛⲥⲱⲕ· ⲉⲣⲉ ⲡⲁϩⲏⲧ ⲧⲏϭ ⲉⲡⲉⲕϩⲏⲧ Here, the equivalent for the ὡς-clause is a circumstantial clause. Sahidic reads ‘Look, I will follow you, my heart being joined to your heart’. The translator has added a verb, not only in the rendering of the ὡς-clause but also in the preceding clause.321 Text-critical note: Variants in Greek add εκκλινον, assimilating this clause to the previous clause, which also has this verb. Sahidic also adds ⲧⲏϭ, but it is not an equivalent for εκκλινον. Since Sahidic needs a verb here, one should not connect the Sahidic to these Greek variants. 2.1.4.11. ὥστε-clauses There are only two clauses introduced with this conjunction (ὥστε ‘so that, in order to’). Both cases in 1 Sam are presented below.322

320 Van der Louw, Transformations, 57. 321 The Sahidic translator has to add a verb or at least a copula in incomplete clauses; see section 2.2.2.1. 322 Payne (Critical and comparative, 49) in his list gives ϫⲉⲕⲁⲥ as the most common translation of ὥστε without mentioning that this is only one of two occurrences. On this basis, one cannot maintain it being typical.

Clause connections

3: 11

111

Ἰδοὺ ἐγὼ ποιῶ τὰ ῥήματά μου ἐν Ἰσραήλ· ὥστε παντὸς ἀκούοντος αὐτὰ, ἠχήσει ἀμφότερα τὰ ὦτα αὐτοῦ. ὥστε] om B b

ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲛⲟⲕ· ϯⲛⲁⲉⲓⲣⲉ ⲛⲛⲁϣⲁϫⲉ ϩⲙ̄ ⲡⲓⲏⲗ̄ ̄ ϫⲉⲕⲁⲥ ⲟⲩⲟⲛ ⲛⲓⲙ ⲉⲧⲛⲁⲥⲱⲧⲙ̄ ⲉⲣⲟⲟⲩ ⲛⲧⲉ ⲛⲉϥⲙⲁⲁϫⲉ ⲙⲟⲩⲧⲉ ⲉⲃⲟⲗ ὥστε can express both purpose and result, and thus context determines its interpretation. In this example, the translator has rendered a final clause. This example also illustrates the statement of Blass/Debrunner/Rehkopf that the boundary between final and consecutive meaning is often no longer visible.323 The conjunction used in Sahidic is ϫⲉⲕⲁⲥ ‘that, in order to’, which expresses purpose. This is a typical case, also mentioned by Layton in his examples: a ϫⲉⲕⲁⲥ-clause of purpose has the conjunctive verbal form if there is an adverbial clause between the conjunction and the verb.324 10: 9

καὶ ἐγενήθη ὥστε ἐπιστραφῆναι τῷ ὤμῳ αὐτοῦ ἀπελθεῖν ἀπὸ Σαμουήλ, μετέστρεψεν αὐτῷ ὁ θεὸς καρδίαν ἄλλην·

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛⲧ ̄ ⲟⲟϣⲉ ⲉⲃⲟⲗ ϩⲓⲧⲟⲟⲧϥ· ⲥⲁⲟⲩⲗ ⲁ ̄ ⲉⲣⲉϥⲕⲧⲟ ⲛⲧⲉϥⲛⲁϩⲃⲉ ⲉⲙ ⲡⲛⲟⲩⲧⲉ ϯ ⲉⲣⲟϥ ⲛⲕⲉϩⲏⲧ Here, the Sahidic translator has chosen precursive ⲛⲧ ̄ ⲉⲣⲉ-, which expresses a temporal meaning ‘when, after’.325 A temporal translation fits here since καὶ ἐγενήθη has this nuance. The translator has substituted a pronoun for the name Samuel, and thereby explicitated the name ‘Saul’ in order to avoid saying that it was Samuel’s heart that was changed.

2.1.5. Non-finite clauses I have grouped the participial constructions, participium coniunctum and genetivus absolutus, under their own headings. I prefer this grouping because participial constructions as such often serve the same functions as subordinate clauses introduced by various conjunctions. In septuagintal Greek, participles occur less often and in more restricted functions than in texts originally composed in Greek.326 In my material, 146 passages with participium coniunctum and 15 passages with genetivus absolutus appear. Another small group consists of infinitive constructions. I have included those infinitives that are introduced with a preposition. There are altogether 30 cases in my material.

323 Blass/Debrunner/Rehkopf, Grammatik, §391. 324 Layton, Grammar, §355. 325 Layton, Grammar, §344. 326 Aejmelaeus, On the Trail, 8–9.

112

Description of the translation technique

part.coni. part.coni. part.coni. part.coni. part.coni. part.coni. part.coni. gen.abs. gen.abs. gen.abs. ἐν τῷ + inf. ἕως τοῦ + inf. πρίν + inf. μετά + inf. δία + inf. ὥστε + inf.

circumstantial ϩⲛ̄ ⲟⲩ-construction without equivalent past tense ϫⲉ temporal verbal form varia circumstantial temporal verbal form varia

See section 2.1.4.11 above

96 cases 20 cases 6 cases 5 times 4 times 3 times 12 times 8 times 4 times 3 times 17 cases 5 cases 4 cases thrice once once

2.1.5.1. Participium coniunctum and its renderings Participium coniunctum is a participle that belongs together with its main clause or to one noun of the main clause. A participle can be placed before or after the predicate.327 There is a wide range of meanings that can be expressed with part.coni. Aejmelaeus enumerates the functions of part.coni. in the LXX as follows: pleonastic, modal, temporal, conditional, concessive, final and causal. The pleonastic use is common in the LXX: two verbs in successive clauses describe one and the same activity, and are coordinated in the Hebrew source text.328 The Coptic language has a suitable equivalent for the part.coni., namely the circumstantial verbal form. This Coptic construction is used to form adverbial, completive, sequential and attributive clauses.329 Like part.coni., the circumstantial can precede or follow its main clause. It is no surprise that more than half of the part.coni. constructions are rendered with this circumstantial form: 96 passages out of 146. 23: 1

καὶ ἀπηγγέλη τῷ Δαυεὶδ λέγοντες…

ⲁⲩϫⲓ ⲡⲟⲩⲱ ⲇⲉ ⲛⲇⲁⲩⲉⲓⲇ ⲉⲩϫⲱ ⲙⲙⲟⲥ ϫⲉ This is one of 54 passages where the participle of λέγω is rendered with ⲉ[ⲥ/ ϥ/ⲩ]ϫⲱ ⲙⲙⲟⲥ ϫⲉ. This is almost a stereotyped rendering, since there are only 327 Blass/Debrunner/Rehkopf, Grammatik, §417. 328 Aejmelaeus, On the Trail, 6. See also Blass/Debrunner/Rehkopf (Grammatik, §418–419) for more detailed descriptions. 329 Layton, Grammar, §415, 421.

Clause connections

113

58 cases of part.coni. constructions with λέγω in 1 Sam.330 The converse, however, is not valid since the phrase ⲉ⸗ϫⲱ ⲙⲙⲟⲥ ϫⲉ is also used to render other expressions,331 and thus it does not necessarily indicate a participle of λέγω in the Vorlage. 20: 38 καὶ ἀνεβόησεν Ἰωναθὰν ὀπίσω τοῦ παιδαρίου αὐτοῦ λέγων… SaM ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲟⲛ ⲁϥⲱϣ ⲉⲃⲟⲗ· ⲟⲩⲃⲉ ⲡⲉϥϩⲙ̄ϩⲁⲗ ⲉϥⲱϣ ⲉⲃⲟⲗ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ ϫⲉ ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲁϥⲱϣ ⲉⲃⲟⲗ· ⲟⲩⲃⲉ ⲡⲉϥϩⲙ̄ϩⲁⲗ ϫⲉ SaA Here, SaA and SaM both possess secondary features. SaA omits ⲟⲛ ‘again’ as well as ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ ‘saying to him’. There are three passages where λέγων is rendered by ϫⲉ. However, SaA has probably omitted some words as it often does.332 SaM has a confusion resulting in the doubling ⲉϥⲱϣ ⲉⲃⲟⲗ ⲉϥϫⲱ ⲙⲙⲟⲥ ‘crying out, saying’. There are many variants on the Greek side, and SaM may have acquired the phrase ‘crying out’ from a marginal reading. 15: 32 καὶ προσῆλθεν πρὸς αὐτὸν Ἀγὰγ τρέμων. ⲁⲩⲱ ⲁⲅⲁⲅ ⲁϥϯ ⲙⲡⲉϥⲟⲩⲟⲓ ⲉϩⲟⲩⲛ ⲉⲣⲟϥ ⲉϥⲥⲧⲱⲧ This is a typical example of a modal part.coni. in Greek. The circumstantial here is adverbial, describing the action of the clause preceding it: ‘Agag came to him trembling’.333 25: 29 καὶ ἀναστήσεται ἄνθρωπος καταδιώκων σε καὶ ζητῶν τὴν ψυχήν σου,… ⲁⲩⲱ ⲉⲣϣⲁⲛ ⲣⲱⲙⲉ ⲧⲱⲟⲩⲛ ⲉϥⲡⲏⲧ ⲛⲥⲱⲕ· ⲁⲩⲱ ⲉϥⲕⲱⲧⲉ ⲛⲥⲁ ⲧⲉⲕⲯⲩⲭⲏ Here the function of the Greek participles is modal, additionally indicating finality. ⲁⲩⲱ as a coordinator between two circumstantials is quite strong; one would rather expect to find the second verb with the conjunctive verbal form, without any conjunction. 21: 7

καὶ ἐκεῖ ἦν ἓν τῶν παιδαρίων τοῦ Σαοὺλ ἐν τῇ ἡμέρᾳ ἐκείνῃ συνεχόμενος Νεεσσαράν ἐνώπιον Κυρίου,…

ⲛⲉⲣⲉ ⲟⲩⲁ ⲇⲉ ⲛⲛ ̄ ⲉϩⲙ̄ϩⲁⲗ ⲛⲥⲁⲟⲩⲗ ⲙⲙⲁⲩ· ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ ⲉϥϭⲏⲡ· ⲉⲙⲛⲧ ̄ ϥ̄ ⲡⲁⲣⲣⲏⲥⲓⲁ ⲙⲙⲁⲩ ⲛⲁϩⲣⲁϥ ⲙⲡⲉϥϫⲟⲉⲓⲥ Here, the transliteration of the Hebrew, Νεεσσαράν, used in Greek seems to have presented a problem for the Sahidic translator.334 The translator has rendered συνεχόμενος with ⲉϥϭⲏⲡ ‘seized’ and added a second circumstan330 In four passages (7: 3, 10: 18, 11: 14 and 20: 38 in SaA), the equivalent is the conjunction ϫⲉ. 331 One typical case is καὶ εἶπεν rendered with ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ. 332 See section 3.4. Brooke/McLean (Old Testament in Greek, 71–2) give a mistaken citation of Sahidic in this passage. The word αὐτοῦ in this example is not omitted in SaA, against Brooke/McLean. 333 Layton, Grammar, §421. 334 Drescher’s (Kingdoms, 52) translation of the Sahidic is ‘not having confidence before his lord’.

114

Description of the translation technique

tial with an object and adverb, thus making an understandable sentence ‘he had no confidence’ out of the cryptic word νεεσσαραν. There are no Greek variant readings with παρρησία and it is intriguing that the translator has chosen a Greek word to render a problematic passage. 17: 41 καὶ ἐπορεύθη ὁ ἀλλόφυλος πορευόμενος καὶ ἐγγίζων πρὸς Δαυείδ. ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϩⲱⲱϥ· ⲁϥⲙⲟⲟϣⲉ ⲉϩⲟⲩⲛ ⲉⲇⲁⲩⲉⲓⲇ· ⲉϥϣⲟⲩϣⲟⲩ ⲙⲙⲟϥ This verse is not attested in the original LXX. Thus Sahidic reads an approximation towards the MT in this case. The Sahidic translator has rendered ἐπορεύθη πορευόμενος with only one verb, ⲁϥⲙⲟⲟϣⲉ ‘he went’. For the second part.coni., the translator has used the rendering ϣⲟⲩϣⲟⲩ ‘to boast’. Apparently, the translator read ἐγγίζων erroneously as a participle of the verb ἐγκαυχάομαι which is rendered with ϣⲟⲩϣⲟⲩ in Ps 73: 4.335 In 1 Sam 2: 3, καυχάομαι is also rendered with ϣⲟⲩϣⲟⲩ ‘to boast’. 30: 8

Καταδίωκε, ὅτι καταλαμβάνων καταλήμψῃ καὶ ἐξαιρούμενος ἐξελῇ.

ⲡⲱⲧ ⲛⲥⲱϥ· ϫⲉ ϩⲛ̄ ⲟⲩⲧⲁϩⲟ· ⲕⲛⲁⲧⲁϩⲟϥ· ⲁⲩⲱ ϩⲛ̄ ⲟⲩⲛⲟⲩϩⲙ̄· ⲕⲛⲁⲛⲟⲩϩⲙ̄ In 20 cases, a part.coni. is rendered by a ϩⲛ̄ ⲟⲩ-construction, thus indicating manner.336 In all except five cases, the same verb is used twice. The Hebrew expression behind this phenomenon is called paronomasia.337 In this example, two participles occur, and both are rendered with ϩⲛ̄ ⲟⲩ-. In these cases, both Greek and Sahidic read the participle of the same verb as the main verb. 20: 6

καὶ ἐρεῖς Παραιτούμενος παρῃτήσατο ἀπ᾽ἐμοῦ Δαυείδ… ⲉⲕⲉϫⲟⲟⲥ ⲛⲁϥ· ϫⲉ ϩⲛ̄ ⲟⲩⲟⲩⲱϣ ⲁϥⲁⲝⲓⲟⲩ ⲙⲙⲟⲓ· Here, the Greek text contains two occurrences of the verb παραιτέομαι338 whereas the Sahidic reads two different verbs: the first one ⲟⲩⲱϣ ‘to desire, to wish, to want’ and the second a Greco-Coptic verb different from the parallel Greek verb.339 The verb παραιτέομαι appears in Sahidic as a loanword, but, for some reason, the translator has not used it here.340

335 Crum, Dictionary, 604. 336 Layton, Grammar, §221. Shisha-Halevy, “Tautological Infinitive”. 337 In Hebrew, one verb may occur twice in a sentence: as an infinitive absolute and a predicate. This paronomasia expresses emphasis or reiteration. According to Aejmelaeus (On the Trail, 2), the frequency of part.coni. as a rendering of these paronomastic verbs is Hebraistic. This is a typical feature especially in the Books of Kingdoms. For a thorough study of the infinitive absolute and its renderings in the Pentateuch, see R. Sollamo, “The LXX Renderings of the Infinitive Absolute”. 338 No Greek variants offer different verbs. 339 Shisha-Halevy (“Tautological Infinitive”, 107–8) has included this verse in his examples. He has made “a representative inventory and consequent typology of eleven …kinds of formal sequence and relationship in the TI construction”. In this typology, this verse is

Clause connections

20: 3 SaM

115

καὶ ἀπεκρίθη Δαυεὶδ τῷ Ἰωναθὰν καὶ εἶπεν Γινώσκων οἶδεν ὁ πατήρ σου ὅτι εὕρηκα χάριν ἐν ὀφθαλμοῖς σου,…

ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁϥⲟⲩⲱϣⲃ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲓⲱⲛⲁⲑⲁⲛ· ϫⲉ ⲡⲉⲕⲉⲓⲱⲧ ⲥⲟⲟⲩⲛ ϫⲉ ⲁⲓϭⲉⲛ ⲭⲁⲣⲓⲥ ⲛⲁϩⲣⲁⲕ· ⲇⲁⲩⲉⲓⲇ ⲡⲉϫⲁϥ ⲛⲓⲱⲛⲁⲑⲁⲛ· ϫⲉ ⲡⲉⲕⲉⲓⲱⲧ ⲥⲟⲟⲩⲛ ϫⲉ ⲁⲓϭⲉⲛ ⲭⲁⲣⲓⲥ ⲛⲁϩⲣⲁⲕ· SaA There are six passages341 where Sahidic has omitted the participle without a Greek parallel. In this verse, both Sahidic manuscripts render γινώσκων οἶδεν with only one word ⲥⲟⲟⲩⲛ ‘to know’. This verse again shows the shortening tendency of SaA: ‘he answered saying’ found in SaM has become ‘he said’.342 14: 33 Ἡμάρτηκεν ὁ λαὸς τῷ κυρίῳ, φαγὼν σὺν τῷ αἵματι· ⲁ ⲡⲗⲁⲟⲥ ⲉⲣⲛⲟⲃⲉ ⲉⲡϫⲟⲉⲓⲥ ⲁϥⲟⲩⲉⲙ ⲁϥ· ⲙⲛ ⲡⲉϥⲥⲛⲟϥ In this and four other passages, the equivalent for part.coni. is the past tense in Sahidic. Thus, the structure in Sahidic has two past tense verbal forms linked with asyndeton, an expression abundantly used in 1 Sam.343 17: 39 καὶ ἐκοπίασεν περιπατήσας ἅπαξ καὶ δίς. ⲁⲩⲱ ⲛⲧⲉⲣⲉϥⲙⲟⲟϣⲉ ⲛϩⲏⲧⲟⲩ ⲛⲟⲩⲥⲟⲡ ⲁⲩⲱ ⲥⲛⲁⲩ ⲁϥϩⲓⲥⲉ Here and in two other passages, Sahidic has a temporal verbal form as the equivalent for part.coni. This is an accurate rendering for a temporal part.coni. 19: 3 SaM

καὶ ἐγὼ ἐξελεύσομαι καὶ στήσομαι ἐχόμενος τοῦ πατρός μου ἐν ἀγρῷ οὗ ἐὰν ᾖς ἐκεῖ,…

ⲁⲩⲱ ⲁⲛⲟⲕ ϯⲛⲁⲧⲱⲟⲩⲛ ϩⲁⲧⲉⲙ ⲡⲁⲉⲓⲱⲧ· ⲛⲧⲁⲉⲓ ⲛⲁⲕ ⲉⲧⲥⲱϣⲉ· ⲉⲡⲙⲁ ⲉⲧⲉⲕⲛⲁϣⲱⲡⲉ ⲛϩⲏⲧϥ·̄ In 12 cases mentioned in the table as ‘varia’,344 the translator has used a conjunctive, relative construction, infinitive, adverb, future or some other rendering for part.coni. In this verse as well as in verse 4: 18, the participle ἐχόμενος with a genitive noun means ‘near’. The translator has used a preposition in this case and the verb ⲉϥϩⲓⲧⲟⲩⲉ ‘to be near’ in verse 4: 18.

one among several “miscellaneous cases of no lexical and merely approximative semantic congruence”. 340 Luke 14: 18 reads ⲁⲩⲁⲣⲭⲓ…ⲉⲡⲁⲣⲁⲓⲧⲓ. In this verse, the meaning is ‘to refuse’. Perhaps the Greek-Coptic loanword has a more restricted meaning than its Greek counterpart, and this is why the translator did not use it here. 341 2: 25, 17: 41, 17: 53, 19: 20, 20: 3, 22: 22. 342 See section 3.4. 343 Here, renderings out of one language and into another resemble a pendulum: Hebrew parataxis is rendered into Greek with the help of conjunctions and participial constructions to make it acceptable Greek. This Greek text is then rendered into Coptic, and, again, long series of paratactic clauses appear without special connectives. 344 2: 3, 3: 21, 9: 5, 13: 17, 14: 19, 14: 43, 17: 16, 19: 3, 20: 5, 20: 31, 24: 8, 26: 25.

116

Description of the translation technique

14: 19 καὶ ἐγενήθη ὡς ἐλάλει Σαοὺλ πρὸς τὸν ἱερέα, καὶ ὁ ἦχος ἐν τῇ παρεμβολῇ τῶν ἀλλοφύλων ἐπορεύετο πορευόμενος… ⲉⲣⲉ ⲥⲁⲟⲩⲗ ⲇⲉ ϣⲁϫⲉ ⲙⲛ̄ ⲡⲟⲩⲏⲏⲃ· ⲟⲩⲛⲟϭ ⲛϩⲣⲟⲟⲩ ⲉⲙⲁⲧⲉ ̄ ⲁϥϣⲱⲡⲉ ϩⲛ̄ ⲧⲡⲁⲣⲉⲙⲃⲟⲗⲏ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· Here, the translator has rendered a Greek paronomastic participle with ⲉⲙⲁⲧⲉ ̄ ‘greatly, very’. This is an idiomatic translation of this participle. 2.1.5.2. Genetivus absolutus and its renderings The other participial construction analyzed here is genetivus absolutus. Unlike part.coni., this participle is more independent of the clause it belongs to. Usually, the subject of a gen.abs. plays no grammatical role in the main clause. According to Greek grammar, gen.abs. can express the same kinds of connections as part.coni.345 However, in the LXX gen.abs. is used in only some of its typical functions.346 There are 15 passages with gen.abs. in my material.347 2: 27

…ἀπεκαλύφθην πρὸς οἶκον πατρός σου, ὄντων αὐτῶν ἐν γῇ Αἰγύπτῳ δούλων τῷ οἴκῳ Φαραώ·

ⲁⲓϭⲱⲗⲡ ⲉⲃⲟⲗ ⲙⲡⲏⲓ ⲙⲡⲉⲕⲉⲓⲱⲧ ⲉⲩϩⲙ ⲡⲕⲁϩ ⲛⲕⲏⲙⲉ ⲉⲩⲟ ⲛϩⲙϩⲁⲗ ⲙⲡⲏⲓ ⲙⲫⲁⲣⲁⲱ Here, the participial construction is rendered with a circumstantial verbal form into Sahidic. In seven more passages, the translator has used the circumstantial—the most often used rendering for gen.abs. in 1 Sam. In his introductory grammar, Layton juxtaposes gen.abs. and the Sahidic circumstantial.348 9: 5

αὐτῶν ἐλθόντων εἰς τὴν Σεὶφ καὶ Σαοὺλ εἶπεν τῷ παιδαρίῳ αὐτοῦ… τήν] τήν γην L; γην A CI f 29 244

ⲛⲧⲟⲟⲩ ⲇⲉ ⲛⲧⲉⲣⲟⲩⲉⲓ ⲉⲡⲕⲁϩ ⲛⲥⲓⲫ· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲙⲡⲉϥϩⲙϩⲁⲗ This is one of four cases where gen.abs. is rendered by a temporal verbal form. Gerd Mink mentions temporal as the most popular way to render gen.abs. and describes gen.abs. and its renderings in Sahidic briefly in his article.349 His assessment is problematic because he only provides examples where gen.abs. is rendered with a temporal verbal form. It gives the wrong —at least in the case of Sahidic 1 Sam—impression of this being the predo345 Blass/Debrunner/Rehkopf, Grammatik, §417–23. 346 I. Soisalon-Soininen (Studien, 177): “Beinahe alle gen.abs. in der Septuaginta haben temporale Bedeutung, nur einige Male kommt er in konditionaler, in konzessiver oder in kausaler Bedeutung vor.” 347 Aejmelaeus (On the Trail, 135) lists 14 passages for 1 Sam. The one extra passage in my material did not belong to the original LXX. 348 B.A. Layton, Coptic in 20 lessons, §121. 349 G. Mink, “Koptischen Versionen”, 245–6.

Clause connections

117

minant equivalent.350 However, in 1 Sam the most popular rendering is not temporal but circumstantial. Text-critical note: The word ‘land’ is attested in Sahidic, but found only in a few Greek mss. A homoioteleuton mistake has probably occurred in numerous Greek mss, την and γην having such a similar appearance. 11: 9

Αὔριον ὑμῖν ἡ σωτηρία διαθερμάναντος τοῦ ἡλίου.

ⲣⲁⲥⲧⲉ ⲡⲟⲩϫⲁⲓ̈ ⲛⲁϣⲱⲡⲉ ⲛⲏⲧⲛ̄ ϩⲙ̄ ⲡⲧⲣⲉ ⲡⲣⲏ ϯ ⲛⲧⲉϥϩⲙ̄ⲙⲉ In this case, the meaning of the Greek participle is temporal, and the translator has employed a causative infinitive for that purpose. The result is an accurate rendering since, when used in this way, the form does not have any causative meaning, regardless of its name, but is purely temporal.351 22: 4

καὶ κατῴκουν μετ᾽αὐτοῦ πάσας τὰς ἡμέρας ὄντος τοῦ Δαυεὶδ ἐν τῇ περιοχῇ.

ⲁϥϭⲱ ϩⲁⲧⲏϥ ⲛⲛ ̄ ϩ̄ ⲟⲟⲩ ⲧⲏⲣⲟⲩ· ⲛⲧ ̄ ⲁ ⲇⲁⲩⲉⲓⲇ ⲁⲁⲩ ⲉϥⲡⲟϣⲥ ̄ Here, the translator has used a relative construction, which is natural. In this passage, only the structure is fitting; the content of the text differs significantly. In the Sahidic the clause that reads in Greek ‘all the days that David was in the stronghold’ is rendered ‘all the days that David was beside himself’. This is due to a different understanding of the word περιοχή, which can mean ‘section, enclosing, extent, fortification, straitness’.352 There are no Greek variants, and thus it is a mistranslation or a copying error. I have not been able to find a plausible candidate for the word that was changed.353 2.1.5.3. Infinitive constructions and their renderings In this chapter, I analyze infinitive constructions used with a preposition. These are used like dependent clauses, and therefore, they are included in this chapter on non-finite clauses.

350 A. Hasznos (“Die Struktur der Konsekutivsätze”, 43) cites Mink in her article: “Beim genitivus absolutus zum Beispiel gibt es einen grammatischen Zwang für den Übersetzer, das logische Subjekt auszudrücken, weil der genitivus absolutus meistens mit dem Temporalis oder dem Perfekt ins Koptische übersetzt wird – zwei Formen, bei denen das Subjekt ausgedrückt werden muss.” I have marked with bold the problematic part of the citation. One should at least make a restriction if this is meant to refer to the New Testament only. 351 Layton, Grammar, §363: an adverbial infinitive phrase meaning ‘While…’. 352 In the next verse, the same word is found in Greek, and in Sahidic the loanword ⲡⲉⲣⲓⲟⲭⲏ is used. 353 If the mistake happened on the Greek side, then the Coptic would be based on something similar to περιοχή. If the mistake happened on the Sahidic side, then some word similar to ⲉϥⲡⲟϣⲥ̄ could explain what has happened.

118

Description of the translation technique

ἐν τῷ + infinitive ἐν τῷ + infinitive ἐν τῷ + infinitive ἐν τῷ + infinitive ἐν τῷ + infinitive ἕως τοῦ + infinitive πρίν + infinitive μετά τό + infinitive διά τό + infinitive

ϩⲙ̄ ⲡⲧⲣⲉ temporal verbal form circumstantial verbal form conditional past tense ϣⲁⲛⲧⲉ⸗ ‘not yet’- conjugation ⲙⲛⲛⲥⲁ ⲉⲃⲟⲗ ϫⲉ

5 times 5 times 4 times 3 times once 5 times 4 times 3 times once

There are 18 cases of ἐν τῷ + infinitive in 1 Sam. In Greek, this construction has temporal meaning, and it expresses simultaneous activity. 21: 6

ἐν τῷ ἐξελθεῖν με εἰς ὁδὸν γέγονε πάντα τὰ παιδάρια ἡγνισμένα.

ϩⲙ̄ ⲡⲧⲣⲉⲛⲧⲁⲁⲛ ⲉⲃⲟⲗ ⲉⲧⲉϩⲓⲏ· ⲁ ⲛϣⲏⲣⲉ ϣⲏⲙ ⲧⲏⲣⲟⲩ ⲧⲃⲃⲟⲟⲩ· A ϩⲙ̄ ⲡⲧⲣⲉ-construction is used five times to render ἐν τῷ + infinitive.354 This Sahidic construction means ‘while …-ing’,355 and the result is an idiomatic rendering here. 16: 6

καὶ ἐγενήθη ἐν τῷ αὐτοὺς εἰσιέναι καὶ εἶδεν τὸν Ἐλιάβ,… ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛⲧⲉⲣⲟⲩⲉⲓ ⲉϩⲟⲩⲛ· ⲁϥⲛⲁⲩ ⲉⲉⲗⲓⲁⲃ· This is an example of a temporal verbal form used to render ἐν τῷ + inf. This is done five times,356 and it is a suitable rendering.

2: 19

καὶ ἀνέφερεν αὐτῷ ἐξ ἡμερῶν εἰς ἡμέρας ἐν τῷ ἀναβαίνειν αὐτὴν μετὰ τοῦ ἀνδρὸς αὐτῆς…

ⲁⲥϫⲓⲧⲥ ⲛⲁϥ ⲉϩⲣⲁⲓ ⲕⲁⲧⲁ ⲡⲧⲱϣ ⲛⲛⲉϩⲟⲟⲩ ⲉⲥⲛⲁⲃⲱⲕ ⲉϩⲣⲁⲓ ⲙⲛ ⲡⲉⲥϩⲁⲓ In four cases, a circumstantial verbal form renders ἐν τῷ + inf.357 Although the circumstantial expresses various meanings, in this clause the temporal sequence remains implicit within the context. 20: 15 οὐκ ἐξαρεῖς ἔλεός σου ἀπὸ τοῦ οἴκου μου ἕως τοῦ αἰῶνος· καὶ εἰ μή, ἐν τῷ ἐξαίρειν Κύριον τοὺς ἐχθροὺς Δαυείδ… εἰ μή] om L 245 554mg ⲛⲛ ̄ ⲉⲕϥⲓ ⲙⲡⲉⲕⲛⲁ ⲉⲃ̄ ⲟⲗ ϩⲙ̄ ⲡⲁⲏⲓ ⲛϣ ̄ ⲁ ⲉⲛⲉϩ· ⲁⲩⲱ ⲉⲣϣⲁⲛ ⲡϫⲟⲉⲓⲥ ϥⲓ ⲙⲙⲁⲩ ⲛⲛⲉϫⲁϫⲉ ⲛⲇ ̄ ⲁⲩⲉⲓⲇ·

354 In addition to this verse, 1: 7, 1: 26 (ϩⲙ ⲡⲁⲓ ⲉⲧⲣⲁϣⲗⲏⲗ), 4: 22, and 11: 2. 355 Layton, Grammar, §493, 52. 356 The other four are in verses 15: 6, 17: 24, 22: 8, and 23: 6. 357 The others are in 9: 9, 25: 2, and 25: 15.

Clause connections

119

A conditional verbal form occurs three times in the rendering of ἐν τῷ + inf.358 Conditional ⲉⲣϣⲁⲛ⸗ appears in clauses meaning ‘if, when, since, whenever’.359 This example is not a plain infinitive construction. Instead, a Greek conditional conjunction precedes it. The Sahidic translator has adapted the text so that ⲁⲩⲱ begins a new sentence ‘And when the Lord removes David’s enemies … Ionathan’s name shall not be removed from David’s house’. Since this passage has been paraphrased, one cannot determine whether εἰ μή lay in the source text of the translator. 10: 8

ἑπτὰ ἡμέρας διαλείψεις ἕως τοῦ ἐλθεῖν με πρὸς σέ, …

ⲟⲩⲛ̄ ⲥⲁϣϥ ⲛϩⲟⲟⲩ ⲛⲁⲟⲩⲉⲓⲛⲉ ⲉⲣⲟⲕ· ϣⲁⲛϯⲉⲓ ⲛⲁⲕ· In five passages, ἕως τοῦ + infinitive occurs in 1 Sam. This construction has a specific meaning, and it is rendered with a specific construction into Sahidic: ϣⲁⲛⲧⲉ⸗ is a limitative expression ‘until such time as’.360 All five cases are rendered with the same construction, resulting in an accurate, albeit stereotyped, rendering.361 There are four cases that have a πρίν + infinitive construction in 1 Sam. This is a temporal expression meaning ‘before’ in Greek.362 In the Sahidic text of 1 Sam, this construction is rendered with the ‘not yet’ conjugation. According to Layton, this Sahidic construction describes the past, saying what has not yet happened, but will probably happen.363 3: 7

καὶ Σαμουηλ πρὶν ἢ γνῶναι θεὸν καὶ ἀποκαλυφθῆναι αὐτῷ ῥῆμα Κυρίου. ἤ] om B 509

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ· ⲙⲡⲁⲧⲉϥⲥⲟⲩⲉⲛ ⲡⲛⲟⲩⲧⲉ·̄ ⲁⲩⲱ ⲙⲡⲉ ⲡϣⲁϫⲉ ⲙⲡϫⲟⲉⲓⲥ ϭⲱⲗⲡ̄ ⲉⲣⲟϥ· The word πρίν is reiterated before the second infinitive ἀποκαλυφθῆναι in all Greek witnesses save A B O b d-125 554. The first infinitive is rendered with ‘not yet’ in the Sahidic, while the second has been translated with a negative past tense.364

358 The other two: 16: 16 and 16: 23. In these two cases, there is no conditional conjunction in Greek. The context has motivated the use of a conditional verbal form in Sahidic. 359 Layton, Grammar, §346. 360 Layton, Grammar, §349. 361 The other four are found in 1: 22, 9: 13, 16: 11, and 19: 23. 362 Blomqvist/Jastrup, Grekisk grammatik, §293, 4. 363 Layton, Grammar, §336. 364 There is a lacuna in SaM so that ⲁⲩ[ⲱ ⲙⲡⲉ] is only partly visible. In theory, ⲁⲩ[ⲱ ⲙⲡⲁⲧⲉ] is also possible. Because I have neither calculated the letters nor measured the line, I have to trust Drescher’s reconstruction.

120 5: 9

Description of the translation technique

καὶ ἐγενήθη μετὰ τὸ μετελθεῖν αὐτὴν καὶ γίνεται χεὶρ Κυρίου ἐν τῇ πόλει, τάραχος μέγας σφόδρα·

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲙⲛⲛ ̄ ⲥⲁ ⲉⲧⲣⲉⲥⲡⲱⲱⲛⲉ· ⲧϭⲓϫ ⲙⲡϫⲟⲉⲓⲥ ⲁⲥϣⲱⲡⲉ ϩⲓϫⲱⲟⲩ· ⲁⲩⲛⲟϭ ⲉⲙⲁⲧⲉ ⲛϣⲧⲟⲣⲧⲣ ̄ ϣⲱⲡⲉ ϩⲛ̄ ⲧⲡⲟⲗⲓⲥ· In three passages, the construction μετά τό + infinitive appears in 1 Sam. The Sahidic translator has rendered this construction with the preposition ⲙⲛⲛⲥⲁ ‘after’. In two cases, the Greek expression occurs with a Coptic infinitive construction ‘after …-ing’, and once with a circumstantial conversion with the same meaning.365 In this case, the translator has not omitted the redundant verb ⲁⲥϣⲱⲡⲉ ‘it happened’ at the beginning.366 The rendering is accurate, but would be more idiomatic without the first verb. 15: 20 καὶ εἶπεν Σαοὺλ πρὸς Σαμουήλ Διὰ τὸ ἀκοῦσαί με τῆς φωνῆς τοῦ λαοῦ· καὶ ἐπορεύθην ἐν τῇ ὁδῷ… ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲥⲁⲙⲟⲩⲏⲗ· ϫⲉ ⲉⲃⲟⲗ ϫⲉ ⲁⲓⲥⲱⲧⲙ ⲛⲥⲁ ⲡⲉϩⲣⲟⲟⲩ ⲙⲡⲗⲁⲟⲥ· ⲁⲩⲱ ⲁⲓⲙⲟⲟϣⲉ ϩⲛ̄ ⲧⲉϩⲓⲏ… ̄ This is the sole occurrence of this construction in 1 Sam. It is rendered with the causal connective ⲉⲃⲟⲗ ϫⲉ ‘because’ in Sahidic, and the rendering is accurate. Both texts read the underlined expression with causal meaning.

2.1.6. Summary This chapter has surveyed a spectrum of linguistic phenomena, and therefore a concise summary is not possible. Nevertheless, there are some topics that deserve special attention with respect to Greek textual criticism. The most important issues are discussed in the following paragraphs. First of all, a tremendous number of καί-clauses are rendered asyndetically. Therefore, in those passages where the choice is between καί and asyndeton, an asyndetic connection in Sahidic does not indicate asyndeton in the Greek text. Another tendency of καί-clauses relates to the frequency of ⲇⲉ in their rendering. Logically, the presence of a certain Greek-Coptic connector should demonstrate that the Greek equivalent was in the source text of the Sahidic translator. This presupposition, however, does not apply to ⲇⲉ or to ⲁⲗⲗⲁ, ⲅⲁⲣ, ⲙⲏⲡⲟⲧⲉ, ⲟⲩⲇⲉ, or ⲏ. Additionally, in some passages a Greek noun or verb is rendered with another Greek noun or verb into Sahidic. The Sahidic translator renders the great majority (94%) of Greek asyndetic clauses asyndetically. If there is a connective added to the rendering of such a clause, it is usually of Egyptian origin, and this happens mostly in 365 Layton, Grammar, §493; Crum, Dictionary, 315a. 366 There are also cases where the verb is omitted in this kind of expression; see subsection 2.1.3.1.f).

Additions

121

direct speech. Thus, a conjunction within a Sahidic narrative context convincingly indicates that a connective existed in the Greek source text. The use of ὅτι recitativum is not perceivable through Sahidic. The translator of 1 Sam has systematically employed ϫⲉ to mark all direct speech. Accordingly, Sahidic is not useful in textual criticism if there are Greek variants involving ὅτι. Another feature not traceable with the help of Sahidic is the presence/absence of the verb ‘to be’ in some Greek clauses. A Sahidic translator will always render a Greek nominal clause with a Coptic nominal clause with a copula—regardless of the Greek source text. When considering relative clauses, SaA often uses simple attributive constructions (e. g. ⲡⲣⲣ̄ ⲟ ⲉⲧⲛⲁⲣⲣ̄ ⲣ̄ ⲟ) against SaM, which reads appositive attributive constructions (e. g. ⲡⲣⲣ̄ ⲟ ⲡⲁⲓ ⲉⲧⲛⲁⲣⲣ̄ ⲣ̄ ⲟ). This tendency may relate to diachronic grammar, but also reflects the overarching tendency of SaA; one can conclude that these reformulations show SaA simplifying the constructions. The issue of double-marking repeatedly arises in the evaluation of subordinate clauses. Sahidic expresses some interclausal relations through verbal inflections, and the relation would be clear without any connective. For some reason, there are several passages where double-marking occurs. This theme may deserve a more detailed study to determine whether this phenomenon is related to the diachronic development of Sahidic and the extent to which it occurs in other Coptic literature. There are only a few passages where the Bohairic version of 1 Sam is extant, and one tiny fragment of Fayyumic survives. According to the limited testimony of these witnesses, the translations appear to be different with respect to both their Vorlage and the translation technique employed. Because most of the Sahidic and Bohairic Old Testament remains unpublished, a great deal of work must be done before the larger issues of textual history and translation technique can be approached. It is probable that the situation varies from book to book, as is the case concerning the LXX itself, but there are not yet enough detailed analyses for any firm conclusions.

2.2. Additions 2.2.1. Questions and method One feature in describing a translator is the quantitative equivalence between the source text, the Vorlage, and the resulting translation. When comparing two fundamentally different languages such as Sahidic and Greek,367 it is natural that the number of words in the translated text does 367 W.-P. Funk (“Bemerkungen”, 168–70) mentions some essential features, like the inclination towards inflection/agglutination: “Während das Griechische als Paradebeispiel einer flektierenden Sprache gelten kann, ist das Koptische in seinem Charakter schwer zu bestimmen, trägt aber sicher wesentliche Züge einer agglutinierenden Sprache.”

122

Description of the translation technique

not equal that in the original. In this context, however, addition is not defined just by different numbers of words. Instead, addition means that there is in the target text some semantic element that is not present in the source text.368 This quantitative discrepancy might result from the translator’s translation technique, but, also, from various other causes. The possibility of a different Vorlage, mistakes in the process of copying the manuscripts and intentional changes must also be considered as explanations for such additions.369 My aim in this section is to identify those additions that we can suppose were made by the translator. This means that I have determined all the Sahidic words that do not have an equivalent in any Greek manuscript, in Hebrew or in any daughter version.370 This limitation reflects the assumption that in the excluded cases, the addition may not derive from the translator, but may, in fact, result from a common Greek source. Concerning the remaining additions, I assume that if there had been some different Vorlage for them, it would be extant through other early witnesses besides Sahidic. Additions may stem not solely from the translator, but also from some later copyist or editor. It is not possible to determine exactly which additions descend from these last-mentioned sources. In those cases where an addition is found only in some Sahidic ms(s), one has particular reason to consider this explanation. Some manuscripts may preserve a stricter and purer text while others may be more prone to later readings and alterations. Variants found in the more careful manuscript may be more highly regarded by those interested in the Greek Vorlage because these variants are more likely representative of the first translation. Van der Louw divides additions into three groups: grammatical, logical and stylistic additions.371 I have also divided my cases into the same three groups following him. According to van der Louw, these additions occur in the cases where formally literal translation seemed ungrammatical or unnatural.372 I think it is necessary to add a small corrective: the additions in the first two groups probably have this kind of motivation (grammatical and logical additions), but for the third group (stylistic additions) I would not assume such an impulse.

368 Van der Louw, Transformations, 74. 369 Aejmelaeus (On the Trail, 79–85) provides examples of various types of divergences between the Septuagint and its Hebrew Vorlage. 370 Those additions that have an equivalent in Hebrew are of special interest when piecing together the history of the Sahidic text. I will present them in section 3.3. One more restriction is also made: the adding of ϫⲉ to introduce direct speech is excluded since it is mostly the equivalent of a colon, not any semantic element. 371 The term ‘stylistic’ is mine; van der Louw (Transformations, 75) speaks of “additions that are designed to improve the source text and to adapt it to the taste of the target language readership”. 372 Van der Louw, Transformations, 74–5.

Additions

123

One more essential term when analyzing additions is explicitation. This occurs when an implicit facet of the source text is made explicit in the translation.373 There is ongoing discussion in translation studies concerning the term explicitation, its definition and usage. Kinga Klaudy distinguishes four types of explicitations: obligatory, optional, pragmatic and translationinherent.374 The first three types more or less coincide with the three groups that van der Louw mentioned earlier; obligatory explicitation coincides with grammatical addition, optional explicitation with logical addition and pragmatic explicitation with stylistic addition, although the latter two pairs have some overlap in their coverage.375 In this chapter, the term addition has a twofold usage: first it refers to an addition according to the definition “a semantic element that is not found in the source text” and second to a secondary plus in the text-critical meaning. The first definition actually works as a hypernym for the latter. My basic aim here is to find those additions that answer not only the first definition but also the second. Scholarly reconstructions of a Greek Vorlage should not rely on variations that probably derive from free translations, grammatical additions or copyist’s errors. As already mentioned, I have divided the readings studied here into three groups. The first group376 contains the additions that result from a grammatical difference between Greek and Sahidic. These additions reflect the different structures of the languages concerned. In this first group, the total number of cases is 87 (+7). In parentheses, I note the additions that are not present in one or more Sahidic mss, and, naturally, it is difficult to know whether these additions were made by the translator or a later copyist. The second group377 consists of additions that result from a different logic of languages. These additions are not as obligatory as the ones in the first group; there is no grammatical necessity to add the word(s). However, without these additions the text would give the impression of a deficiency or clumsiness. In this second group, there are 241 (+29) additions. In the third group,378 stylistic and interpretive additions, the additions are the most free with regard to their relation to the Greek source text. In this group, the total number of additions is 134 (+31). Within the three main groups, there are subgroups according to the type of addition.

373 Van der Louw, Transformations, 81. 374 K. Klaudy, “Explicitation”, 82–3. 375 Klaudy’s last type, namely translation-inherent explicitation, is the most abstract and confusing. In her article, she does not provide examples of this group. Viktor Becher particularly criticizes the fourth group; see V. Becher, “Abandoning the notion of ‘translationinherent explicitation’”. There is also more literature on the topic in the same article. 376 Table 1 below. 377 Table 2 below. 378 Table 3 below.

124

Description of the translation technique

I have not divided the subgroups. For example, all the cases where a second verb appears are in the second group ‘logical additions’, even though there might be cases that could equally belong to the third group ‘stylistic additions’. When making the groupings, one cannot avoid a degree of subjectivity. Some cases are open to various interpretations. To make my decisions transparent, I have cited verse numbers of the cases in each subgroup within the footnotes.

2.2.2. Additions based on a grammatical difference Table 1. Grammatical additions Completion of an incomplete clause Addition of an antecedent in a relative clause Addition of an auxiliary word to the translation Possessive article total

55 (+2)379 13 (+5) 10 9 87 (+7)

2.2.2.1. Completion of an incomplete clause The Sahidic translator usually transforms Greek verbless clauses into verbal clauses.380 This means that he usually adds a verb. Occasionally, however, the translator has added a copula in the Sahidic—an addition necessary with an ordinary Sahidic nominal clause.381 2: 3

…καὶ θεὸς ἑτοιμάζων ἐπιτηδεύματα αὐτοῦ. ⲁⲩⲱ ⲛⲧⲟϥ ⲡⲉ ⲡⲛⲟⲩⲧⲉ ⲉⲧⲥⲟⲃⲧⲉ ⲛⲛⲉϥϩⲉⲃⲏⲩⲉ In this example, the translator has added a pronoun and a copula in Sahidic. The result is a nominal clause. The Greek participle ἑτοιμάζων is rendered as a relative clause in Sahidic. 9: 12

…ὅτι θυσία σήμερον τῷ λαῷ ἐν Βαμα. ϫⲉ ⲉⲣⲉ ⲟⲩⲑⲩⲥⲓⲁ ϣⲟⲟⲡ ⲙⲡⲟⲟⲩ ⲙⲡⲗⲁⲟⲥ ϩⲛ̄ ⲃⲁⲙⲁ This is one of 13 cases where the translator has used the verb ϣⲟⲟⲡ ‘to be’ without Greek variants containing εἰμί. Accordingly, in the cases where 379 In parenthesis are those cases where some Sahidic ms(s) do not have the addition. 380 2: 3, 2: 16, 2: 18, 3: 4, 3: 5, 3: 6, 3: 8, 3: 14, 3: 16, 4: 4, 5: 9, 6: 4, 6: 9, 6: 20, 7: 8, 8: 19, 9: 6, 9: 12, 9: 20, 10: 2, 10: 7, 10: 12, 10: 14, 10: 19, 11: 8bis, 12: 3bis, 12: 12, 14: 20, 15: 4, 15: 22, 15: 24, 17: 30, 18: 19, 18: 25, 18: 28, 19: 16, 20: 6, 20: 7, 20: 9, 20: 25, 21: 9, 22: 15, 22: 17, 23: 7, 24: 5, 24: 15bis, 25: 6, 25: 36, 26: 19, 27: 10, 28: 14, 30: 26, (20: 1, 22: 6, 25: 3). 381 Cf. Feder, Biblia Sahidica, 97: “Passagen, die im Griechischen ganz ohne Verbform auskommen, müssen im Koptischen Ergänzungen erfahren, um überhaupt verständlich zu sein.”

Additions

125

Greek has variants with and without εἰμί, a Sahidic reading with ϣⲟⲟⲡ does not necessitate a verb in its source text.382 In 8 cases, the translator has converted a single negation383 in Greek into an entire clause. 6: 9

…αὐτὸς πεποίηκεν ἡμῖν τὴν κακίαν ταύτην τὴν μεγάλην· καὶ ἐὰν μή, καὶ γνωσόμεθα ὅτι οὐ χεὶρ αὐτοῦ ἧπται ἡμῶν, …

…ⲛⲧⲟϥ ⲡⲉ ⲛⲧⲁϥⲉⲓⲛⲉ ⲛⲁⲛ ⲛⲧⲉⲓⲛⲟϭ ⲛⲕⲁⲕⲓⲁ· ⲉⲥⲧⲙⲧⲁⲁⲥ ⲇⲉ ⲉⲡⲥⲁ ⲉⲧⲙ̄ⲙⲁⲩ ⲧⲉⲛⲛⲁⲉⲓⲙⲉ ϫⲉ ⲛⲧⲉϥϭⲓϫ ⲧⲉ ⲁⲛ ⲛⲧⲁⲥⲉⲓ ⲉϫⲱⲛ… Here, two conditional constructions follow one another, ‘if it goes this way, then he has done it, but, if not, then it is not his hand’. The second conditional is expressed in Greek with only a short negation ‘and if not’. The Sahidic translator has rendered this phrase ‘but if it does not go to that side’, thus adding a finite verb. 2.2.2.2. Relative clause The translator frequently adds an antecedent before a relative clause.384 Often in these cases, an implicit Greek antecedent is made explicit in the Sahidic translation before the relative clause. In Greek, the relative pronoun expresses number, gender and case, whereas in Sahidic there is no relative pronoun. Sahidic distinguishes a relative clause by verbal conversion, placing the converter just before the converted clause. Additionally, a resumptive pronoun often appears at the end of the converted expression in order to assure the right interpretation of the clause. 3: 12

ἐν τῇ ἡμέρᾳ ἐκείνῃ ἐπεγερῶ ἐπὶ Ηλι πάντα ὅσα ἐλάλησα εἰς τὸν οἶκον αὐτου

ϩⲙ̄ ⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ ϯⲛⲁⲉⲓⲛⲉ ⲉϫⲛ̄ ϩⲏⲗⲉⲓ·̅ ⲛϣⲁϫⲉ ⲛⲓⲙ ⲛⲧⲁⲓⲧⲁⲩⲟⲟⲩ ⲉϫⲙ̄ ⲡⲉϥⲏⲓ· In this example, the translator has added the noun ‘words’ as the antecedent of the following relative clause. 12: 7

…καὶ ἀπαγγελῶ ὑμῖν τὴν πᾶσαν δικαιοσύνην Κυρίου, ἃ ἐποίησεν ἐν ὑμῖν καὶ ἐν τοῖς πατράσιν ὑμῶν·

…ⲛⲧⲁⲧⲁⲩⲟ ⲉⲣⲱⲧⲛ·̄ ⲛⲧ ⲧⲏⲣⲥ̄ ⲙⲡϫⲟⲉⲓⲥ· ⲛⲉϩⲃⲏⲩⲉ ⲛⲧ ̄ ⲇⲓⲕⲁⲓ̄ ⲟⲥⲩⲛⲏ ̅ ̄ ⲁϥⲁⲁⲩ ⲛⲏⲧⲛ·̄ ⲙⲛ̄ ⲛⲉⲧⲛⲉ̄ ⲓⲟⲧⲉ̄·

382 Feder (Biblia Sahidica, 96) gives an example from Lamentations, where Coptic is cited in the apparatus of the Greek text. 383 This could also be termed rhetorical negation. 384 3: 3, 3: 12, 10: 5, 12: 7, 12: 24, 13: 14, 14: 4, 14: 47, 20: 32, 23: 13, 26: 1, 29: 8bis, (6: 18, 8: 9, 8: 11, 9: 10, 20: 39).

126

Description of the translation technique

Here, the antecedent is implicit in Greek. However, the Sahidic translator has to add an antecedent before the relative clause. Otherwise, the relative clause would seem odd in Sahidic, having a feminine word ⲧⲇⲓⲕⲁⲓⲟⲥⲩⲛⲏ as the explicit antecedent and at the same time a resumptive suffix in the plural. 2.2.2.3. An auxiliary word in translation In this group, the number of cases is ten, including words like ⲙⲁ ‘a place’ or ϣⲏⲣⲉ ‘a son’ that are used as an aid in translation.385 In these clauses, the added words do not bring anything extra to the content.386 1: 6

ὅτι οὐκ ἔδωκεν αὐτῇ Κύριος παιδίον, κατὰ τὴν θλῖψιν αὐτῆς…

ⲉⲃⲟⲗ ϫⲉ ⲛⲧⲟϥ ⲡϫⲟⲉⲓⲥ ⲙⲡⲉϥϯ ϣⲏⲣⲉ ⲛⲁⲥ· ⲕⲁⲧⲁ ⲡⲧⲱϣ ⲛⲧⲉⲥⲑⲗⲓⲯⲓⲥ· Here, the word ⲧⲱϣ ‘ordinance, destiny, fashion, affair, matter, thing, condition’ appears in the Sahidic. 13: 16 καὶ Σαοὺλ καὶ Ἰωναθὰν υἱὸς αὐτοῦ καὶ ὁ λαὸς οἱ εὑρεθέντες μετ᾽αὐτῶν ἐκάθισαν ἐν Γαβεὲ Βενιαμεὶν καὶ ἔκλαιον… Βενιαμείν] pr τῆς 82–93–127 ⲥⲁⲟⲩⲗ ⲇⲉ ⲙⲛ̄ ⲓⲱⲛⲁⲑⲁⲛ ⲡⲉϥϣⲏⲣ̄ ⲉ· ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲛⲧⲁⲩϩⲉ ⲉⲣⲟϥ ⲛⲙ̄ⲙⲁⲩ· ⲁⲩϩⲙⲟⲟⲥ ϩⲛ̄ ⲅⲁⲃⲁⲁ· ⲡⲕⲁϩ ⲛⲃⲉⲛⲓⲁⲙⲓⲛ ⲁⲩⲣⲓⲙⲉ· Here the word ‘land’ is added. Possibly, the translator had a text like the one attested by the Lucianic mss, ἐν Γαβαὰ τῆς Βενιαμίν, and, theoretically, a skip between γῆς and τῆς happened. Thus, Sahidic could have had γῆς/τῆς γῆς in its source text. 2.2.2.4. Possessive article The Sahidic translator often uses a possessive article where Greek reads a definite article.387 Sometimes, Greek manuscripts offer the possessive pronoun, but usually the possessive element only occurs in the Sahidic text.388 In my material, there are nine cases where the possessive article is attested only in the Sahidic text. Accordingly, if there are Greek variants with the definite article and others with the possessive pronoun, Sahidic with the possessive article does not necessitate a Greek reading with the possessive pronoun in its source text.

385 1: 6, 2: 36, 6: 4, 13: 15, 13: 16, 14: 6, 14: 11, 18: 19, 19: 21, 23: 15. 386 This group is very similar to the previous group of antecedents of relative clauses. 387 This seems to be a typical Sahidic phenomenon; see Mink, “Koptischen Versionen”, 221–2. It also occurs in the Sahidic Jeremiah (Feder, Biblia Sahidica, 98). Alin Suciu (“The Sahidic Version of Jacob of Serugh’s Memrā on the Ascension of Christ”, 61) has used this feature as an argument for a native Sahidic translator in that patristic writing. 388 3: 9, 3: 16, 4: 6, 5: 6, 7: 3, 11: 9, 12: 2, 13: 5, 25: 3.

127

Additions

4: 6

καὶ ἤκουσαν οἱ ἀλλόφυλοι τῆς κραυγῆς, καὶ εἶπον…

ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲥⲱⲧⲙ̄ ⲉⲡⲉⲩϩⲣⲟⲟⲩ ⲁⲩⲱ ⲡⲉϫⲁⲩ ϫⲉ This example resembles the examples given by Mink in his chapter on possessive relations. According to him, Sahidic makes the possessive relation explicit also in the cases where it is implicitly signified by the Greek definite article. This happens especially with bodyparts, terms expressing relation (like father and mother) as well as with personal relations.389 Mink mentions a name as a close personal belonging, and similarly, voice, mentioned in the example above, is likewise part of the person(s).

2.2.3. Logical additions The following cases display additions that are due to logic and striving for clarity, not just structural differences between the two languages. The following clauses would have been intelligible in Sahidic without these additions. Klaudy has the following description for her group of optional explicitations: “They are optional in the sense that grammatically correct sentences can be constructed without their application in the target language, although the text as a whole will be clumsy and unnatural.”390 In these clauses, the translator’s ambition to clarify and improve the text becomes obvious. One can also observe how implicit information in the source language is often made explicit in the translation.391 Table 2. Logical additions Addition of direct or indirect object Addition of a second verb Addition of a subject Introduction of a speech total

126 (+9) 84 (+15) 25 (+5) 6 241 (+29)

2.2.3.1. Addition of direct or indirect object In Sahidic, the normal word order requires the direct object and the indirect object immediately after a transitive verb.392 A very common addition 389 Mink, “Koptischen Versionen”, 233–5. 390 Klaudy, “Explicitation”, 82. 391 This phenomenon is often mentioned in translation studies. It seems to be a universal feature that translations are longer than their source texts. For more discussion, see Maeve Olohan/Mona Baker, “Reporting that in translated English”, 142, and the literature mentioned there. 392 The object is marked by a preposition or suffixed directly to the verb. Layton, Grammar,

128

Description of the translation technique

occurs in the following example. After verbs of saying, the translator often adds the addressee of the speech. This feature has one significant, oftenoccurring consequence when discussing a Greek text. If there are readings with and without αὐτῷ after a saying- verb, Sahidic should not be used as a witness for the reading with αὐτῷ even though the equivalent is found in the Sahidic text. 26: 6

καὶ εἶπεν Ἀβεσσά Ἐγὼ εἰσελεύσομαι μετὰ σοῦ.

ⲡⲉϫⲉ ⲁⲃⲉⲥⲥⲁ ⲛⲁϥ ϫⲉ ⲁⲛⲟⲕ ϯⲛⲏⲩ ⲛⲙ̄ⲙⲁⲕ This addition of a 3rd person singular masc. indirect object occurs 43 times in 1 Sam, and is also common in other Sahidic texts.393 In order to avoid misinterpretations, one has to pay special attention to each case before supposing a connection to a Greek reading with αὐτῷ. 9: 9

καὶ ἔμπροσθεν ἐν Ἰσραὴλ τάδε ἔλεγεν ἕκαστος ἐν τῷ πορεύεσθαι ἐπερωτᾶν τὸν θεόν…

ϫⲓⲛ ⲉϣⲟⲣⲡ̄ ⲇⲉ ϩⲙ̄ ⲡⲓⲏⲗ̄ ·̄ ⲛⲁⲓ ⲛⲉ ⲛϣ ̄ ⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ ϫⲟⲟⲩ ⲙⲡⲉⲑⲓⲧⲟⲩⲱϥ· ⲉⲩⲛⲁⲃⲱⲕ ⲉϣⲓⲛⲉ ⲙⲡⲛⲟⲩⲧⲉ In this example, every Greek word has a Sahidic equivalent. Additionally, the Sahidic text reads the indirect object ‘to his neighbour’ underlined above. 9: 25 καὶ διέστρωσαν τῷ Σαοὺλ ἐπὶ τῷ δώματι, … SaM ⲁⲩⲡⲱⲣϣ̄ ϩⲁⲣⲟϥ ⲛⲥⲁⲟⲩⲗ· ϩⲓ̈ ϫⲉⲛⲉⲡⲱⲣ ⲛϩ̄ ⲛϩ̄ ⲟⲉⲓⲧⲉ· ⲉⲛⲁϣⲱⲟⲩ SaA ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲡⲱⲣϣ ⲛⲥⲁⲟⲩⲗ ϩⲓϫⲛ̄ ⲧϫⲉⲛⲉⲡⲱⲣ All these mss continue with the phrase ‘he laid down to sleep’. SaM has the addition ‘many clothes’. Verbal forms differ between the Sahidic texts. SaM reads a 3rd person plural, which often translates passive forms in Sahidic, whereas SaA reads a 3rd person singular and adds the proper name Samuel.394 The verb ⲡⲱⲣϣ̄ ‘to spread’ functions both transitively and intransitively.395 The addition results from the transitive meaning, which §166a. All the cases of this subgroup: 1: 18, 2: 5, 2: 16tris, 2: 20, 2: 28, 2: 30, 3: 5bis, 3: 6bis, 3: 8, 3: 10, 3: 17, 3: 18, 4: 3, 4: 20, 5: 7, 5: 8bis, 5: 11, 6: 4, 6: 9, 6: 20, 7: 5, 7: 8, 8: 11, 9: 4bis, 9: 9, 9: 15, 9: 21, 9: 26, 10: 11, 10: 14, 10: 21, 10: 22, 11: 9, 11: 12, 12: 13, 12: 21, 12: 22, 13: 8, 13: 11, 14: 8, 14: 10, 14: 11, 14: 12, 14: 14, 14: 15, 14: 28, 14: 33tris, 14: 34bis, 14: 36bis, 14: 38, 14: 41, 14: 45, 15: 10, 15: 14, 15: 15, 15: 16, 15: 21, 15: 30, 15: 32bis, 16: 2bis, 16: 4, 16: 5, 16: 11, 16: 18, 16: 22, 17: 10, 17: 20, 17: 23, 17: 25, 17: 28, 17: 29, 17: 35, 17: 37, 17: 43, 17: 49, 18: 20, 18: 22, 18: 23, 18: 25, 19: 5, 19: 14, 19: 21, 19: 22, 20: 6, 20: 10bis, 20: 13, 21: 9bis, 22: 12, 22: 22, 23: 1, 23: 19, 23: 28, 24: 10, 24: 12, 24: 20, 25: 8, 25: 12, 25: 19, 26: 1, 26: 6bis, 26: 12, 26: 14, 26: 19, 26: 21, 28: 2, 28: 11, 28: 15, 28: 16, 28: 21, 30: 20, 30: 23, (7: 8, 9: 15, 9: 19, 9: 25, 22: 12, 25: 13, 25: 14, 30: 8, 30: 15). 393 For Jeremiah see Feder, Biblia Sahidica, 97–8. 394 See also section 3.2.1 where this verse is analyzed in detail with regard to different Sahidic and Greek readings. 395 Crum, Dictionary, 269b–70. He cites an example where both transitive and intransitive forms are used within one text, the transitive having the object ⲛⲛⲉϥϩⲟⲓⲧⲉ.

Additions

129

makes the object necessary. SaA has understood the verb intransitively, and thus the object is not needed. 2.2.3.2. Addition of a second verb In a number of cases, the translator adds a second verb to a clause where the Greek text has only one verb.396 This happens especially when the beginning of an activity is described. In these cases, the added verb is ⲧⲱⲟⲩⲛ ‘to rise’ or ⲕⲱⲧⲉ ‘to turn’.397 Sometimes, the translator has rendered a composite verb with two separate verbs. In some cases, a subtle nuance in Greek is expressed with an additional verb in Sahidic. 5: 8

Μετελθέτω κιβωτὸς τοῦ θεοῦ πρὸς ἡμᾶς· καὶ μετῆλθεν κιβωτὸς τοῦ θεοῦ εἰς Γέθθα.

ⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡⲛⲟⲩⲧⲉ· ⲙⲁⲣⲉⲥⲡⲱⲱⲛⲉ ⲛⲥⲉⲓ ϩⲁⲑⲏⲛ· ⲁⲩⲱ ⲧⲕⲟⲓⲃⲱⲧⲟⲥ ⲙⲡⲛⲟⲩⲧⲉ· ⲁⲥⲡⲱⲱⲛⲉ ⲉϩⲣⲁⲓ ⲉⲅⲉⲑ· In this example, the verb μετέρχομαι occurs twice. The first instance is rendered with two Sahidic verbs ‘let it turn and come’, while the second has only one Sahidic verb ‘it crossed over to’. Thus, it is not necessary to render μετέρχομαι with two verbs. The translator appears to have arbitrarily rendered μετέρχομαι with two Sahidic verbs, perhaps attempting to dramatize or clarify the narrative. This double occurrence offers no basis for the retroversion of the Greek Vorlage, and suggests caution in the evaluation of similar instances of extra Sahidic verbs. 8: 5

καὶ νῦν κατάστησον ἐφ᾽ἡμᾶς βασιλέα δικάζειν ἡμᾶς… ⲧⲉⲛⲟⲩ ϭⲉ ⲧⲱⲟⲩⲛ ⲛⲅ̄ ⲕⲁⲑⲓⲥⲧⲁ ⲉϫⲱⲛ ⲛⲟⲩⲣⲣ̄ ⲟ̄ ⲉⲧⲣⲉϥϯϩⲁⲡ ⲉⲣⲟⲛ Here, people have first informed Samuel how his sons misbehave, and then they ask Samuel to appoint a king for them, ‘Now, rise and appoint a king.’

9: 8 SaM SaA

Ἰδοὺ εὕρηται ἐν τῇ χειρί μου τέταρτον σίκλου ἀργυρίου, καὶ δώσεις τῷ ἀνθρώπῳ τοῦ θεοῦ,…

ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲓϭⲓⲛⲉ ⲛⲟⲩϭⲓⲥⲕⲓ̈ⲧⲉ ⲛⲥⲁⲧⲉⲉⲣⲉ ⲛϩⲁⲧ ϩⲛ̄ ⲧⲁϭⲓϫ· ⲉⲕⲉϫⲓⲧⲥ·̄ ⲛⲅ̄ ⲧⲁⲁⲥ ⲙⲡⲣⲱⲙⲉ ⲙⲡⲛⲟⲩⲧⲉ· ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲓϭⲓⲛⲉ ⲛⲟⲩϭⲓⲥⲕⲓ̈ⲧⲉ ⲛⲥⲁⲧⲉⲉⲣⲉ ⲛϩⲁⲧ ϩⲛ̄ ⲧⲁϭⲓϫ· ⲉⲕⲉϫⲓⲧⲥ̄· ⲙⲡⲣⲱⲙⲉ ⲙⲡⲛⲟⲩⲧⲉ·

396 1: 6, 1: 7, 1: 18, 1: 19, 1: 22, 2: 30, 2: 34, 3: 5, 3: 12, 4: 2, 4: 3, 4: 5, 4: 10, 4: 19, 5: 8, 6: 12bis, 6: 13, 6: 14, 7: 2, 8: 5, 8: 19, 9: 16, 9: 26bis, 10: 6, 10: 26, 10: 27, 12: 2, 13: 3, 13: 4bis, 13: 5, 13: 7, 13: 11, 13: 13, 13: 15, 14: 7, 14: 36, 15: 6, 15: 11, 16: 8, 17: 13, 17: 14, 17: 17, 17: 29, 17: 51, 19: 5, 19: 15, 20: 1, 20: 22bis, 20: 27bis, 21: 1bis, 22: 5, 23: 20, 23: 25bis, 24: 10, 24: 19, 25: 13, 25: 29, 25: 30, 25: 37bis, 25: 41, 26: 8bis, 26: 16, 27: 1bis, 27: 4, 28: 11, 28: 22, 28: 25, 29: 3, 29: 4, 30: 8, 30: 9, 30: 22, 31: 4, 31: 6, (6: 14, 8: 8, 9: 8, 9: 26, 20: 6, 20: 29bis, 20: 34, 20: 38, 20: 41, 22: 3, 22: 23, 28: 3, 31: 4bis). 397 Drescher (Kingdoms, xv) also mentions the habit of adding the verb ⲧⲱⲟⲩⲛ and enumerates the passages where both SaM and SaA contain this addition.

130

Description of the translation technique

In this case, the Greek verb is rendered with two verbs in SaM but with only one verb in SaA. With respect to the originality of these Sahidic readings, SaM has apparently preserved the earlier reading ‘take it and give it’. Usually, the Sahidic verb ϯ ‘to give’ renders Greek δίδωμι. The variant of SaA ‘take it’398 is explainable as an omission of a redundant verb, a typical tendency in this manuscript. 2.2.3.3. Addition of the subject Typical additions include cases where the translator adds a subject to a clause.399 Such an addition can include a proper name, a noun or a pronoun. In the 3rd person singular, Sahidic verbs explicitly mark gender, and thus are more specific than their Greek counterparts in these cases. Still, in the cases where two males are acting or discussing, the translator often adds proper names in order to distinguish who is acting or speaking. 3: 6

Ἰδοὺ ἐγώ, ὅτι κέκληκάς με· καὶ εἶπεν Οὐ κέκληκά σε·

ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲛⲟⲕ ⲁⲓⲉⲓ· ϫⲉ ⲁⲕⲙⲟⲩⲧⲉ ⲉⲣⲟⲓ· ⲡⲉϫⲉ ϩⲏⲗⲓ ⲛⲁϥ ϫⲉ ⲙⲡⲓⲙⲟⲩⲧⲉ ⲉⲣⲟⲕ· In this example, God is calling young Samuel, who first thinks that Eli is calling him. The translator felt it necessary to add the proper name. Arguably, the story is easier to follow in Sahidic, where the proper name is expressed explicitly.400 25: 42 καὶ πέντε κοράσια ἠκολούθουν αὐτῇ, καὶ ἐπορεύθη ὀπίσω τῶν παίδων Δαείδ, καὶ γίνεται αὐτῷ εἰς γυναῖκα. ⲁⲩⲱ ϯ ⲛϩⲙ̄ϩⲁⲗ ⲛⲧⲁⲥ ⲁⲩⲟⲩⲁϩⲟⲩ ⲛⲥⲱⲥ· ⲁⲩⲱ ⲛⲧⲟⲥ ⲁⲥⲟⲩⲁϩⲥ̄ ⲛⲥⲁ ⲛⲉϩⲙ̄ϩⲁⲗ ⲛⲇⲁⲩⲉⲓⲇ· ⲁⲥϣⲱⲡⲉ ⲛⲁϥ ⲛⲥϩ̄ ⲓⲙⲉ· Here, the pronoun ⲛⲧⲟⲥ ‘she’ is added without similar variants in Greek or other versions. The person is already explicit in the verb, and this addition can be considered stylistic. 2.2.3.4. Introduction of a speech In this group, I gather those cases where the phrase ⲉϥϫⲱ (sg.) / ⲉⲩϫⲱ (pl.) ⲙⲙⲟⲥ ϫⲉ ‘saying that’ is added before a speech.401 I have not tallied those 398 The verb used in SaA also has the meaning ‘to bring’, which perhaps fits the context of SaA better. 399 1: 8, 2: 17, 2: 23, 3: 6, 3: 19, 6: 7, 6: 14, 10: 2, 12: 19, 13: 9, 15: 29, 16: 8, 16: 9, 17: 8, 20: 3, 20: 34, 22: 14, 23: 5bis, 23: 26, 23: 28, 25: 42, 28: 21, 29: 6, 30: 25, (6: 11, 6: 19, 19: 23, 29: 8, 30: 15). 400 The translator also adds the verb ‘I came’ to the first clause and an addressee ⲛⲁϥ of the speech, but these additions belong to other categories: the first one to the group ‘completion of an incomplete clause’ and the second to the group ‘addition of direct or indirect object’, both above. 401 3: 4, 3: 8, 9: 9, 10: 24, 25: 21, 27: 10.

Additions

131

cases where only ϫⲉ is added since there are myriads of them, and the function more resembles that of a colon. Actually, there are no cases in 1 Sam where this word is lacking before direct speech. 3: 8

καὶ προσέθετο Κύριος καλέσαι Σαμουὴλ ἐν τρίτῳ· καὶ ἀνέστη… Σαμουήλ] + Σαμουηλ V L 799 158 ἀνέστη] + Σαμουηλ L 318

ⲁ ⲡϫⲟⲉⲓⲥ ⲟⲩⲱϩ ⲉⲧⲟⲟⲧϥ ⲉⲙⲟⲩⲧⲉ ⲙⲡⲙⲉϩϣⲟⲙⲛⲧ ⲛⲥⲟⲡ ⲉⲥⲁⲙⲟⲩⲏⲗ· ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ ⲥⲁⲙⲟⲩⲏⲗ· ⲁϥⲧⲱⲟⲩⲛ ⲟⲛ This is a typical case exemplifying the addition of a speech introduction. Here, the Sahidic may have a connection to a source like the Lucianic text. In my opinion, however, there is no reason to suppose this dependence since the Lucianic text preserves a different word order and an addition of the name ‘Samuel’ in the next clause. Additionally, there is no Greek source for the speech introduction. The translator has apparently rephrased this passage in his own way. 9: 9

…τάδε ἔλεγεν ἕκαστος ἐν τῷ πορεύεσθαι ἐπερωτᾶν τὸν θεόν Δεῦρο…

ⲛⲁⲓ ⲛⲉ ⲛϣ ̄ ⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ ϫⲟⲟⲩ ⲙⲡⲉⲑⲓⲧⲟⲩⲱϥ· ⲉⲩⲛⲁⲃⲱⲕ ⲉϣⲓⲛⲉ ⲙⲡⲛⲟⲩⲧⲉ ⲉⲩϫⲱ ⲙⲙⲟⲥ · ϫⲉ ⲁⲙⲏⲉⲓⲧⲛ… ̄ In this example, the translator has added the typical speech introduction with a plural verbal form.

2.2.4. Stylistic additions The additions in this group are the most arbitrary. This group contains descriptive elements that correct and explain. Many of these are motivated by the difference between source and target language cultures and habits. Klaudy posits group pragmatic explicitation, which is described as follows: Members of the target language cultural community may not share aspects of what is considered general knowledge within the source language culture and, in such cases, translators often need to include explanations in translations. For example, names of villages and rivers, or of items of food and drink which are well known to the source language community may mean nothing to the target language audience.402

402 Klaudy, “Explicitation”, 83.

132

Description of the translation technique

Table 3. Stylistic additions An adverb, a prepositional phrase or an adjective A complement to the subject or object An explanation of the previous word An interjection or a short comment Other additions total

45 (+8) 34 (+11) 22 (+2) 9 24 (+10) 134 (+31)

2.2.4.1. An adverb, a prepositional phrase or an adjective 20: 35 Καὶ ἐγενήθη πρωὶ καὶ ἐξῆλθεν Ἰωναθαν εἰς ἀγρόν, … ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲉϩⲧⲟⲟⲩⲉ̄ ⲙⲡⲉϥⲣⲁⲥⲧⲉ· ⲓⲱⲛⲁⲑⲁⲛ ⲁϥⲃⲱⲕ ⲉⲃⲟⲗ ⲉⲧⲥⲱϣⲉ· Here, the translator adds an adverbial phrase ‘tomorrow’. However, the clause would be understandable even without the addition.403 8: 6 Δὸς ἡμῖν βασιλέα δικάζειν ἡμᾶς· SaM ⲙⲁ ⲛⲁⲛ ⲛⲟⲩⲣⲣ̄ ⲟ ̄ ⲉⲃⲟⲗ ϩⲓϫⲱⲛ ⲉⲧⲣⲉϥϯϩⲁⲡ ⲉⲣⲟⲛ·̄ For the underlined phrase ⲉⲃⲟⲗ ϩⲓϫⲱⲛ, SaA reads ⲉϩⲣⲁⲓ ⲉϫⲱⲛ, and SaS reads ⲉⲃⲟⲗ ⲉϫⲱⲛ. This same phrase occurs several times in chapter eight.404 SaA mostly uses ⲉϩⲣⲁⲓ ⲉϫⲱ⸗, once ⲉϫⲛ ⲧⲏⲩⲧⲛ.̄ SaM varies among ⲉⲃⲟⲗ ϩⲓϫⲱⲛ, ⲉϫⲛ ⲧⲏⲩⲧⲛ,̄ ⲉϫⲱ⸗ and ⲉϩⲣⲁⲓ ⲉϫⲱ⸗. This variation shows how several Sahidic adverbial expressions are interchangeable, the meaning being almost identical. 30: 19 καὶ οὐ διεφώνησεν αὐτοῖς ἀπὸ μικροῦ ἕως μεγάλου καὶ ἀπὸ τῶν σκύλων καὶ ἕως υἱῶν καὶ θυγατέρων καὶ ἕως πάντων ὧν ἔλαβον αὐτῶν, τὰ πάντα ἐπέστρεψεν Δαυείδ.

ⲉⲙⲡⲉϥⲕⲁ ⲗⲁⲁⲩ ⲉⲡⲁϩⲟⲩ ⲛϩⲏⲧⲟⲩ· ⲛϫ ̄ ⲓⲛ ⲡⲉⲩⲕⲟⲩⲓ· ϣⲁ ⲡⲉⲩⲛⲟϭ· ⲁⲩⲱ ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ ⲧⲏⲣⲟⲩ· ⲙⲛ̄ ⲛⲉⲩϣⲏⲣⲉ· ⲙⲛ̄ ⲛⲉⲩϣⲉⲉⲣⲉ· ⲁⲩⲱ ⲛⲉⲛⲧⲁⲩϥⲓⲧⲟⲩ ⲧⲏⲣⲟⲩ ⲉⲛⲕⲁ ⲛⲓⲙ ⲇⲁⲩⲉⲓⲇ ⲁϥⲕⲧⲟⲟⲩ̄ In the example above, the translator has varied his renderings for several ἀπό – ἕως constructions. At the same time, he has added the word ⲧⲏⲣⲟⲩ ‘all’ and left ἕως twice unrendered. In van der Louw’s terms, this sentence would be an example of a translator improving the source text according to the taste of his readership;405 the translator has varied the constructions unlike Greek, which uses only ἀπό – ἕως here. 403 All cases of this subgroup: 1: 10, 2: 1, 2: 23, 4: 5, 5: 5, 5: 9, 6: 18, 7: 7, 8: 6, 8: 10, 8: 14, 9: 2, 10: 2bis, 11: 7, 12: 22, 14: 1, 14: 18, 15: 3, 16: 1, 17: 16, 17: 25, 19: 9, 20: 21, 20: 26, 20: 29, 20: 35, 21: 4, 22: 23, 23: 5, 23: 9, 23: 11, 23: 16, 23: 17, 23: 18, 25: 21, 25: 25, 26: 11, 26: 19, 26: 22, 27: 1, 28: 19, 30: 7, 30: 9, 30: 19, (7: 9, 19: 5, 19: 10, 19: 4, 22: 2bis, 23: 3, 23: 13). 404 In verses 5, 6, 7, 9, 10, 11, 22. 405 Van der Louw, Transformations, 75.

Additions

133

2.2.4.2. A complement to the subject or the object 1: 23

…κάθου ἕως ἂν ἀπογαλακτίσῃς αὐτό· ϩⲙⲟⲟⲥ ⲛⲉ ϣⲁⲛⲧⲉⲱⲙϫ̄ ⲙ̄ⲡϣⲏⲣⲉ ⲕⲟⲩⲓ In this case, the fact that Sahidic has no neuter gender might have influenced the translator. In Greek, the antecedent for the neuter personal pronoun is clear, since the only neuter in the whole sentence is the child. Nevertheless, in Sahidic the near context contains three masculine actors: Elkanah, child and God. Therefore, the explicitation of the child prevents a potential misunderstanding. Without this addition, the text would still be comprehensible, since the verb means ‘to wean from breastfeeding’. 2: 20

καὶ ἀπῆλθεν ὁ ἄνθρωπος εἰς τὸν τόπον αὐτοῦ.

ⲁⲩⲱ ⲉⲗⲅⲁⲛⲁ ⲡⲣⲱⲙⲉ ⲁϥⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲡⲉϥⲙⲁ· In this clause, the addition of the name ⲉⲗⲅⲁⲛⲁ is understandable because of the previous context.406 By using the proper name, the translator avoids confusion between Elkanah and the priest. This addition is due to the translator’s interest in clarifying the story’s plot. 8: 20

καὶ ἐξελεύσεται ἔμπροσθεν ἡμῶν, καὶ πολεμήσει τὸν πόλεμον ἡμῶν.

ⲁⲩⲱ ⲛϥⲙⲟⲟϣⲉ ϩⲓϩⲏ ⲙⲙⲟⲛ· ⲛϥⲙⲓϣⲉ ⲙⲛ ⲡⲡⲟⲗⲩⲙⲟⲥ ⲉⲧⲛⲁⲧⲱⲟⲩⲛ ⲉϫⲱⲛ Here, the Sahidic translator expresses the object in a more detailed way. The Greek says succinctly ‘our war’, whereas Sahidic uses a relative construction with a future verbal form. Because the conjunctive forms used here do not distinguish time, perhaps the Sahidic translator added this final phrase to specify the future time that is evident in the Greek source. 20: 1

SaM

καὶ εἶπεν Τί πεποίηκα, καὶ τί τὸ ἀδίκημά μου, καὶ τί ἡμάρτηκα ἐνώπιον τοῦ πατρός σου… καί 2º] ἤ L καί 3º] ἤ L 318

ⲡⲉϫⲁϥ ⲛⲁϥ ϫⲉ ⲟⲩ ⲡⲉ ⲛⲧⲁⲓⲁⲁϥ· ⲁⲩⲱ ⲟⲩ ⲡⲉ ⲡⲁϫⲓⲛϭⲟⲛⲥ ̄ ⲛⲧⲁⲓⲁⲁϥ· ⲏ ⲡⲁⲛⲟⲃⲉ· ⲟⲩ ⲡⲉ ⲙⲡⲉⲙⲧⲟ ⲉⲃⲟⲗ ⲙⲡⲉⲕⲉⲓⲱⲧ· ⲡⲉϫⲁϥ ⲛⲁϥ ϫⲉ ⲟⲩ ⲡⲉ ⲛⲧⲁⲓⲁⲁϥ· ⲁⲩⲱ ⲟⲩ ⲡⲉ ⲡⲁϫⲓⲛϭⲟⲛⲥ·̄ ⲁⲩⲱ ⲡⲁⲛⲟⲃⲉ· ⲟⲩ SaA ⲡⲉ ⲙⲡⲉⲙⲧⲟ ⲉⲃⲟⲗ ⲙⲡⲉⲕⲉⲓⲱⲧ· This is an example of a case where the translator has added both a copula and a relative verbal form. In this verse, David asks a tripartite question. The second question reads in Greek without a verb τί τὸ ἀδίκημά μου. In SaM, the text reads ⲟⲩ ⲡⲉ ⲡⲁϫⲓⲛϭⲟⲛⲥ ̄ ⲛⲧⲁⲓⲁⲁϥ ‘what is my iniquity that I have

406 All the cases of this subgroup: 1: 3bis, 1: 6, 1: 23, 1: 27, 2: 5, 2: 11, 2: 20, 6: 3, 8: 20, 9: 3, 9: 6bis, 9: 8, 10: 7, 10: 14, 11: 13, 12: 3, 12: 4, 13: 9, 17: 14, 17: 25, 17: 29, 18: 9, 18: 15, 19: 3, 19: 17, 22: 18, 23: 21, 24: 3, 25: 2, 25: 3, 27: 5, 28: 21, (6: 17, 6: 21, 7: 10, 19: 4, 19: 6, 19: 7, 19: 10, 22: 12, 25: 5, 28: 17).

134

Description of the translation technique

done’, whereas in SaA the text lacks the verbal form ‘that I have done’. I regard the reading of SaM as the original Sahidic translation, since SaA frequently shortens the text against other Sahidic witnesses. In Greek, manuscripts L 318 have the conjunction ἤ between the questions, not καί. In SaM, the conjunction at the beginning of the third question is ⲏ, and some could see a connection here. However, the conjunction καί is also rendered with ⲏ in verse 26: 12 and in some passages in Jeremiah.407 Additionally, only the third question begins with ⲏ in SaM, whereas in the Greek manuscripts the second (in L) and the third questions (in L 318) are thus introduced. Therefore, I see the conjunction ⲏ in Sahidic resulting from a striving towards variation, like the different word order in the very same clause. 2.2.4.3. Explanation After a difficult word, especially after loanwords, the phrase ⲉⲧⲉ ⲡⲁⲓ ⲡⲉ ‘that is’ introduces an explanation of the previous word. The other word used for the same purpose is the Greek loanword ⲏ ‘or’.408 In these cases, it is especially difficult to determine whether the addition was made by the translator or by a later editor of the text. Possibly, the translator used words known to him, but, later, the Greek words were not commonly known, and the need for an explanation became obvious. Notably, chapter 17 contains significantly more explanations than other chapters: there are six instances in this chapter, whereas other chapters show zero to two occurrences.409 5: 6

καὶ ἐξέζεσεν αὐτοῖς εἰς τὰς ναῦς, καὶ μέσον τῆς χώρας αὐτῆς ἀνεφύησαν μύες·

ⲁϥⲧⲁⲕⲟ ⲛⲛⲉⲩⲉϫⲏⲩ ⲁⲩⲱ ⲛⲧⲉⲩⲭⲱⲣⲁ· ⲏ ⲧⲉⲩⲥⲱϣⲉ· ⲁⲥⲧⲁⲩⲟ ⲉϩⲣⲁⲓ ⲛϩⲉⲛⲡⲓⲛ Here, the Sahidic text contains an addition of an Egyptian word after its Greek counterpart. This addition was made either by the translator or some later copyist or editor. 17: 3

καὶ Ἰσραὴλ ἵσταται ἐπὶ τοῦ ὄρους ἐνταῦθα, καὶ ὁ αὐλὼν ἀνὰ μέσον αὐτῶν.

ⲡⲓⲥⲣⲁⲏⲗ ⲇⲉ ⲁϥⲁϩⲉⲣⲁⲧϥ ϩⲓ ⲡⲓⲥⲁ ⲙⲡⲧⲟⲟⲩ· ⲁⲩⲱ ⲡⲁⲩⲗⲱⲛ ⲉⲧⲉ ⲡⲓⲁ ⲡⲉ· ⲛⲉϥϩⲛ̄ ⲧⲉⲩⲙⲏⲧⲉ· Here, the explanation appears after a Greek word. Again, a Greek word precedes a Sahidic explanation.

407 Feder, Biblia Sahidica, 87, 92–3. 408 All the cases in this subgroup: 5: 6, 7: 12, 9: 24bis, 13: 21, 17: 1, 17: 3, 17: 7, 17: 20, 17: 32, 17: 34, 21: 6bis, 22: 15, 23: 19, 24: 3, 25: 8, 25: 32, 27: 8, 28: 3, 29: 4, 30: 23, (7: 12, 10: 1). 409 See also section 3.2.4.

Additions

135

2.2.4.4. Interjection or a short comment410 1: 8

καὶ εἶπεν αὐτῇ Ἐλκανα ὁ ἀνὴρ αὐτῆς Ἅννα· καὶ εἶπεν αὐτῷ Ἰδοὺ ἐγώ, κύριε· καὶ εἶπεν αὐτῇ…

ⲉⲗⲕⲁⲛⲁ ⲇⲉ ⲡⲉⲥϩⲁⲓ ⲡⲉϫⲁϥ ⲛⲁⲥ ϫⲉ ⲁⲛⲛⲁ· ⲛⲧⲟⲥ ⲇⲉ ⲡⲉϫⲁⲥ ⲛⲁϥ ϫⲉ· ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁⲛⲟⲕ ⲡϫⲟⲉⲓⲥ ϯⲥⲱⲧⲙ̄ ⲉⲣⲟⲕ· ⲡⲉϫⲁϥ ⲛⲁⲥ ϫⲉ… In this sentence, there are several minor additions.411 The underlined comment ϯⲥⲱⲧⲙ̄ ⲉⲣⲟⲕ might be motivated by the lack of a Sahidic vocative; the translator has added ‘I am listening to you’. At the same event, the translator has formulated a whole sentence by adding the verb.412 9: 5

καὶ Σαουλ εἶπεν τῷ παιδαρίῳ αὐτοῦ τῷ μετ᾽αὐτοῦ Δεῦρο καὶ ἀναστρέψωμεν, μὴ ἀνεὶς ὁ πατήρ μου τὰς ὄνους…

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲙⲡⲉϥϩⲙϩⲁⲗ ⲉⲧⲛⲙⲙⲁϥ· ϫⲉ ⲁⲙⲟⲩ ⲛⲧⲉⲛⲕⲧⲟⲛ· ⲛⲧⲉⲛⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲡⲉⲛⲙⲁ ⲙⲏⲡⲱⲥ ⲛⲧⲉ ⲡⲁⲉⲓⲱⲧ ⲉϥϥⲓ ⲣⲟⲟⲩϣ ϩⲁ ⲛⲉⲟⲟⲩ· Typically, ⲕⲱⲧⲉ functions as a reflexive with a suffix, meaning ‘turn one’s self, return’ as here.413 The added clause clarifies where they will go: ‘let us go to our place’. 2.2.4.5. Other additions In this group, I have collected all the additions that do not belong to any of the above-mentioned groups. Mostly, these additions involve single words, but some include clauses.414 A few of these additions can be explained as errors or misreadings. 2: 25

ἐὰν ἁμαρτάνων ἁμάρτῃ ἀνὴρ εἰς ἄνδρα, καὶ προσεύξονται ὑπὲρ αὐτοῦ πρὸς Κύριον·

ⲉⲣϣⲁⲛ ⲟⲩⲣⲱⲙⲉ ⲉⲣⲛⲟⲃⲉ ⲉⲩⲣⲱⲙⲉ ϣⲁⲩⲧⲉⲃⲉϩ ⲡϫⲟⲉⲓⲥ ⲉϫⲱϥ ⲉⲧⲣⲉϥⲕⲱ ⲛⲁϥ ⲉⲃⲟⲗ Here, the Sahidic translator adds the contents of the prayer. This is a free addition, meant to clarify the text. 12: 25 καὶ ἐὰν κακίᾳ κακοποιήσητε, καὶ ὑμεῖς καὶ ὁ βασιλεὺς ὑμῶν προστεθήσεσθε. προστεθήσεσθε] απολεισθε L 158 318 554mg 410 1: 8, 9: 5, 9: 12, 9: 20, 12: 17, 19: 5, 22: 12, 23: 20, 25: 11. 411 Addition of the subject ⲛⲧⲟⲥ, twice ϫⲉ to introduce speech, and the underlined comment. 412 Thus, this example also belongs to the subgroup completion of an incomplete clause 2.2.2.1. 413 Crum gives both ⲕⲱⲧⲉ and ⲃⲱⲕ as equivalents for ἀναστρέφω, but he provides for the latter only one example with a qualitative verbal form. Crum, Dictionary, 29a, 125a. 414 All cases of this group: 2: 25, 4: 19, 5: 2, 5: 9, 5: 12, 6: 9, 6: 17, 8: 12, 9: 12, 9: 24, 12: 25, 14: 41, 15: 30, 17: 30, 17: 49, 17: 52, 18: 11, 21: 8, 24: 18, 25: 26, 25: 44, 26: 10, 26: 20, 30: 6, (14: 21, 14: 23, 17: 36, 19: 4, 20: 5, 20: 16, 20: 21, 20: 42, 22: 7, 23: 13).

136

Description of the translation technique

ⲁⲩⲱ ϩⲛ̄ ⲟⲩⲕⲁⲕⲓⲁ· ⲉⲧⲉⲧⲛϣ ̄ ⲁⲛⲉⲣⲛⲟⲃⲉ ⲛⲧⲱⲧⲛ·̄ ⲙⲛ̄ ⲡⲉ̄ⲧⲉⲛⲣⲣ̄ ⲟ̄· ⲡϫⲟⲉⲓⲥ ⲛⲁⲉⲓⲛⲉ ⲉϫⲛ̄ ⲧⲏⲩⲧⲛ̄ ⲛⲛⲉⲧⲛϫⲁϫⲉ. In this example, the translator has added the subject and the object. He was obviously troubled with the mistake in his source text, and has created his own solution. The preceding clause is a conditional, and this addition makes up its main clause. The basis for this addition lies in a mistranslation of the Hebrew source by the Greek translator. Instead of ‫‘ ספה‬to be swept away’, he rendered the verb ‫‘ אסף‬to do again’. Clearly the Sahidic translator is dependent neither on the variant found in Lucianic mss here nor on exactly the same sources of Hebrew-based corrections since these do not parallel the approximation found in the Sahidic.415 Sometimes the translator betrays through his translation his own cultural milieu. This is a sort of cultural adaptation. The way people used to eat serves as an example. In the first chapter, the translator twice adds the verb ‘to sit’ before eating: ‘Hannah did not want to sit down to eat’ (1: 7) and ‘She sat and ate with her husband’ (1: 18). In verse 16: 11, the translator adds the verb ⲟⲩⲱⲙ ‘to eat’ after ‘we shall not recline’. Apparently, with only the verb ‘recline’, the audience would not have inferred that the narrative entailed eating, because they used to sit down to eat.416 16: 11 Ἀπόστειλον καὶ λαβὲ αὐτόν, ὅτι οὐ μὴ κατακλιθῶμεν ἕως τοῦ ἐλθεῖν αὐτόν. ϫⲟⲟⲩ ⲛⲅ̄ ⲙⲟⲩⲧⲉ ⲉⲣⲟϥ· ϫⲉ ⲛⲧⲛⲛ ̄ ⲁⲛⲟϫⲛ̄ ⲁⲛ ⲉⲟⲩⲱⲙ ϣⲁⲛⲧⲉϥⲉⲓ̈· 20: 27 Τί ὅτι οὐ παραγέγονεν ὁ υἱὸς Ἰεσσαὶ καὶ ἐχθὲς καὶ σήμερον ἐπὶ τὴν τράπεζαν; ⲉⲧⲃⲉ ⲟⲩ ⲣⲱ ⲙⲡⲉ ⲡϣⲏⲣⲉ ⲛⲓⲉⲥⲥⲁⲓ̄ ·̅ ⲉⲓ ⲛⲥⲁϥ ⲙⲛ̄ ⲡⲟⲟⲩ ⲉϩⲙⲟⲟⲥ· ⲉⲧⲉⲧⲣⲁⲡⲉⲍⲁ· ⲉⲟⲩⲱⲙ· In this example, the translator has added both the verb ‘to sit’ and the verb ‘to eat’. Without these verbs, the clause would be grammatically correct in Sahidic.

2.2.5. Summary Sahidic 1 Sam preserves various sorts of additions. In the examples provided above in the group “Grammatical additions”, the translator is forced to use additional words to render his Vorlage accurately. Thus, generally, one should include them under the heading “accurate rendering”. They belong

415 According to Brock (Recensions, 264), the approximation comes from Symmachus. 416 In the Greco-Roman milieu, people would recline and eat, whereas pictures from ancient Egypt show people sitting and eating.

Additions

137

to the group of additions according to the definition “semantic element not present in the source text”, but they do not necessarily belong to the group of additions relevant to the textual criticism of the underlying Greek text. The usage of these features in Greek textual criticism is especially problematic. In these cases, external criteria might help to decide whether Sahidic should be cited or not. However, before using external criteria one has to know the actual dependencies of the Sahidic text. The danger of circular reasoning is not only theoretical but a reality here. One should cite only those features whose relationship to the Greek Vorlage is not suspect and ignore translational variants, like the ones analyzed here. Feder gives examples of such cases in Jeremiah where Sahidic is cited in the critical apparatus of the Greek text although it is about a translational variant.417 The second group “Logical additions” holds an intermediate position between two extremes. These additions are not necessary to avoid grammatical translationese, but they are needed to make the text more natural. Thus, the translator has not changed the content in these cases, but he has adapted it to better meet the idiom of his target language. The translator and his own influence become most obvious in the readings of the third group “Stylistic additions”. These cases show that the text was not so unmalleable that adding to the content was impossible. In some passages, the Sahidic text deviates from its Vorlage in a certain direction, for example describing in a more detailed manner by adding an adjective or an adverb. This group shows the typical additions made by this translator according to his own tastes, and these should be borne in mind when attempting to retrovert the Sahidic translation. Where an addition of this kind is present, the reading is not relevant to the reconstruction of the Greek Vorlage, be there corresponding Greek variants or not.

417 Feder, Biblia Sahidica, 95–9.

3. Affiliations of the Sahidic manuscripts and the Greek Vorlage of the translator 3.1. Questions and methods In this chapter, my aim is to describe textual characteristics of different Sahidic manuscripts. This is done by comparing them with each other as well as with Greek readings. Drescher, in his introduction to the edition of SaM, writes, “For a precise estimate of the Coptic-Greek relation an exhaustive, detailed collation of manuscripts would be necessary.”1 He continues: “To assess the exact extent of the further revision which M [= SaM] or A [= SaA] or both had independently is a laborious task, involving detailed confrontation of the Coptic with the Greek apparatus of variant readings.” This detailed confrontation is exactly my purpose here. The first issue is to determine whether the Sahidic version, now preserved by several manuscripts, is basically one translation or several independent translations. In order to prove that the extant Sahidic witnesses all derive from one translation event, one would need to find arguments that connect all the Sahidic witnesses together; a special rendering is one such argument, as is also a secondary reading in Sahidic, against Greek and other versions. Another task in the analyses is to distinguish secondary features in the Sahidic manuscripts—i. e. those features that arose in translation or transmission. This is required both in reconstructing the textual history of the Sahidic tradition and in textual criticism when the Sahidic version is used for tracking its Greek Vorlage. To avoid misinterpretations, secondary readings should not be cited in the textual criticism of the Vorlage. However, these secondary readings should not be ignored, and are, in fact, crucial for the reconstruction of the transmission of the Sahidic text. Basically, there are two kinds of secondary readings: intra-Sahidic variation and approximation towards some Greek Vorlage. Intra-Sahidic variation should be recognized and excluded from citation in Greek textual criticism. Sometimes Sahidic texts seem to have been emended or corrected against later Greek manuscripts, and in these cases one must determine whether the changes can be associated with a particular Greek manuscript or Greek manuscript group.2

1 2

Drescher, Kingdoms, v. This study adopts the Greek manuscript groupings that have been developed for the forthcoming critical edition of the Greek text of 1 Sam by Aejmelaeus. These appear as an appendix at the end of this study (6.5).

Questions and methods

139

In this chapter, the analyses concern both textual criticism and textual history. Textual criticism features within this conversation in two ways— with respect to the Greek text and with respect to the Sahidic text. A critical text of Sahidic 1 Sam is not a practical goal, since only one complete manuscript has survived. However, the discussion will note and comment on obviously secondary features of the Sahidic text in the analyzed passages. This chapter will also consider how this information can and should influence Greek textual criticism, although such issues do not arise in every passage. The results discussed in this chapter have been systematically applied to the apparatus criticus of Greek I Sam. This survey repeatedly considers whether Sahidic readings merit citation. Textual history draws wider lines and creates the framework for the textual criticism.3 Unfortunately, scholarship concerning the textual history of Sahidic 1 Sam, not to mention the entire Sahidic Old Testament, is almost nonexistent.4 According to an important methodological principle, secondary readings evidence a shared history.5 Michael Segal formulates this principle in the following way: Some manuscripts … do share some unique textual characteristics, and therefore can be grouped together as a family of manuscripts. By unique, I refer to secondary readings (the only way to prove textual affiliation between manuscripts), with reference to differences in content. Agreement between original readings only demonstrates that two manuscripts can both be traced to a common original author or scribe, but cannot be used to demonstrate a unique, genetic relationship between two texts.6

According to Payne’s dissertation on 1 Sam, the Sahidic translation is a solid member of the B-group, comprised of B (Vaticanus), 509 and 121 together with the Ethiopic version.7 Payne lists the cases where a Hexaplaric approximation is present neither in the Greek B-group nor in the Sahidic tradition as evidence for Sahidic belonging to the B-group, without ascertaining

3 4

5 6 7

For a differentiation between textual criticism and textual history, see Aejmelaeus, “Corruption or Correction”, 1–2. Schüssler has published manuscript lists of Sahidic biblical manuscripts, but there is still a huge amount of work remaining before anything like a comprehensive textual history can be written, be it on one book or the entire Old Testament. To mention some examples, the dissertations of Frank Feder on the Corpus of Jeremiah (Feder, Biblia Sahidica) and that of Gregor Emmenegger on the Psalter represent progress. However, Emmenegger himself also comments on the urgent need for a critical text of the Sahidic Psalter as well as accurate editions of the existing manuscripts (Emmenegger, Der Text des koptischen Psalters, 14– 15). Hany N. Takla’s dissertation is titled “An Introduction to the Coptic Old Testament”, but he only offers information on the manuscripts and their locations, without discussing any questions concerning the textual history. Aejmelaeus, “Corruption or Correction”, 4. M. Segal, “Text of the Hebrew”, 7. Payne, Critical and comparative, 139.

140

Affiliations of the Sahidic manuscripts and the Greek Vorlage

whether the Greek reading is original or secondary.8 Thus, he is actually trying to show a dependence based on original readings, since a Hexaplaric reading certainly is secondary, and lack of it refers to the original Greek reading. In my opinion, when a minority of Greek witnesses and Sahidic preserve the Old Greek, then it might tell something of the Vorlage of the Sahidic translator. However, this is not the case if there are many witnesses besides Sahidic and the B-group that read without the emendation. Payne’s lists contain many proper names whose argumentative value is insignificant, since inner-Sahidic variation often makes their evaluation impossible.9 For Payne, the only way to explain the sporadic appearance of Lucianic and Hexaplaric readings in the Sahidic text is to postulate a Greek manuscript with exactly the same text.10 There are two problems with this idea that need further attention. First, Payne understands the Lucianic text as a monolithic entity. This is problematic, since there are several layers within the Lucianic text: the old base text, the subsequent recensional readings, and the still-later Hexaplaric readings. It makes a difference in the explanation whether the reading in question is in the early base text and perhaps even the Old Greek, or whether it is a later Hexaplaric emendation. The second problem is that, for some reason, Payne does not allow for sporadic corrections/additions in the Sahidic text when explaining the appearance of Hexaplaric and Lucianic readings. However, Dieu in his work had already explained the situation without assuming a Greek manuscript with exactly the text that results when a retroversion of Sahidic is made. According to Dieu, the Sahidic translator possessed as source text a pre-recensional Greek text. Afterwards, someone, sporadically, supplied the Sahidic translation with the Greek recensional readings.11 Payne himself does not comment on Dieu’s study even though it appears in his bibliography. Brock has shown that no one Greek manuscript preserves the Hexaplaric readings of 1 Sam, because no direct descendent of the Hexapla survives among the extant manuscripts. Instead, these readings surface sporadically in various witnesses, mainly in O L D E.12 Since this sporadic attestation is evident in the Greek tradition, I see it as probable, at least worth testing, also in the Sahidic tradition. Various types of deviations exist between manuscripts, and all discrepancies are not equal when it comes to their weight as evidence. Some of the 8

Payne, Critical and comparative, makes a distinction between “relatively strong cases of textual relationship”, on pp. 141–53, and “relatively weak cases of textual relationship”, on pp. 153–7. 9 Particular care should be taken when dealing with SaA because this witness so often confuses the letters. 10 Payne, Critical and comparative, 392. 11 Dieu, “Le texte copte sahidique”, 452. 12 Brock, Recensions, 170–1. Sigla are those of Brock, and the following manuscripts belong to these groups: O (A 376 247) L (19, 82, 93, 108, 127) D (107, 370, 106, 120, 134, 554) E (52, 489, 92, 130, 314).

141

Textual analyses

variants in Sahidic manuscripts, like obvious differences in vocabulary or longer additions, are clearly connected to a particular Greek reading. Sometimes two, or even three different readings appear in the Sahidic manuscripts, and each of them renders the Greek expression accurately. In these cases, no decision is possible concerning the priority of the Sahidic readings. Most cases do not belong to the extremes mentioned above, complicating the examination of these variants. Before deciding in individual cases, one needs a framework. The approach used here has been expounded already in the previous chapters. In analyzing the following passages, I will objectively test the results of the preceding analyses. What kind of evidence for the Greek Vorlage remains after excluding the translational variants?

3.2. Textual analyses I have listed the arguments for choosing each of the following passages in the Introduction (section 1.5). In the following analyses, the text is divided according to verses. The preliminary critical Greek text of Aejmelaeus stands in the first box. If significant variants exist in the Greek manuscripts with bearing on the discussion of the Sahidic, they appear below the Greek text. Afterwards, one finds the Sahidic texts available for that verse.13 3.2.1. SaMAV 7: 8–8: 1, 9: 21–10: 2 7: 8 καὶ εἶπαν οἱ υἱοὶ Ἰσραὴλ πρὸς Σαμουήλ Μὴ παρασιωπήσῃς ἀφ᾽ἡμῶν τοῦ μὴ βοᾶν πρὸς Κύριον θεόν σου, καὶ σώσει ἡμᾶς ἐκ χειρὸς ἀλλοφύλων. + καὶ εἶπεν Σαμουήλ Μὴ μοι γένοιτο ἀποστῆναι ἀπὸ Κῡ θῡ μου τοῦ μὴ βοᾶν περὶ ὑμῶν προσευχόμενον 108mg CI 119mg 121mg d f s 29 55 71 158 244mg 245 318 554 SaM

SaA

SaV

ⲡⲉϫⲉ ⲛⲉ[ϣⲏⲣⲉ] ⲙⲡⲓⲏ̄ⲗ ̄ ⲉϩⲟⲩⲛ [ϩⲛ] ⲥⲁⲙⲟⲩⲏⲗ· ϫⲉ ⲙⲡⲉⲣⲕⲁⲣⲱⲕ ⲉ[ϫⲱ]ⲛ ⲛⲅϭⲱ ⲛⲟⲩ[ⲉ]ϣ ⲛⲱϣ ⲉϩⲣⲁⲓ ⲉⲡϫⲟⲉⲓⲥ ⲡⲉⲕⲛⲟⲩⲧⲉ· ⲛ̄ⲅⲛ̄ ⲁϩⲙⲉⲛ ⲛⲧⲟⲟⲧⲟⲩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲁⲩ̄ ϫⲉ ⲛⲛⲉⲥϣⲱⲡⲉ ⲙⲙⲟⲓ· ⲉⲧⲣⲁⲥⲁϩⲱⲱⲧ ⲉⲃⲟⲗ ⲙⲡϫⲟⲉⲓⲥ ⲡⲁⲛⲟⲩⲧⲉ· ⲉⲧⲙ̄ⲧⲣⲁⲱϣ ⲉϩⲣⲁⲓ ⲉⲣⲟϥ ⲉϫⲱⲧⲛ·̄

ⲡⲉϫⲉ ⲛ̄ϣⲏⲣⲉ ⲙ̄ⲡⲓ ̅ⲏ̄ⲗ ⲉϩⲟⲩⲛ ϩⲛ̄ ⲥⲁⲙⲟⲩⲏⲗ ϫⲉ ⲙ̄ⲡⲣ̄ϭⲱ ⲛⲟⲩϣⲱϣ ⲉϩⲣⲁⲓ ⲉⲡϫⲟⲉⲓⲥ ⲡⲕ̄ⲛⲟⲩⲧⲉ· ⲛ̄ϥⲛⲁϩ̄ⲙ ⲉⲧⲟⲟⲧⲟⲩ ⲛ̄ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ϫⲉ ⲛ̄ⲛⲉ ⲡⲁⲓ ϣⲱⲡⲉ ⲉⲧⲣⲁⲥⲁϩⲱⲓ ⲉⲃⲟⲗ ⲙ̄ⲡϫⲟⲉⲓ̈ⲥ ⲡⲁⲛⲟⲩⲧⲉ· ⲉⲧⲙ̄ⲧⲣⲁⲱϣ ⲉϩⲣⲁⲓ ⲉⲣⲟϥ ⲉϫⲱⲧⲉⲛ·

]ⲛ̄ⲱϣ ⲉ̣ϩⲣ̣ ⲁ ̣ ̣ⲓ ̣ [ⲉⲡϫⲟⲉⲓⲥ ⲡⲉⲕ]ⲛⲟⲩⲧⲉ ⲛ̄ⲅⲛ̄ ⲁϩ[ⲙⲉⲛ ⲛⲧⲟⲟⲧⲟⲩ ⲛ]ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲡⲉϫⲉ [ⲥⲁⲙⲟⲩⲏⲗ] ⲛⲁⲩ ϫⲉ ⲛ̄ⲛⲉ ⲡⲁⲓ ϣⲱⲡ[ⲉ ⲙⲙⲟⲓ] ⲉⲧⲣⲁⲥⲁϩⲱ̄ ⲓ̈ ⲉⲃⲟⲗ ⲙ̄ⲡϫⲟ[ⲉⲓⲥ ⲡⲁⲛⲟⲩ]ⲧⲉ· ⲉⲧⲙ̄ⲧⲣⲁⲱϣ ⲉⲃⲟⲗ· ⲉϩ[ⲣⲁⲓ ⲉⲣⲟϥ] ⲉϫⲱⲧⲛ·̄

13 Sahidic texts are written in full, although it is redundant in some verses. I have chosen this practice to be able to show which words are reconstructed in lacunose manuscripts.

142

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In the second clause, SaM reads ⲙⲡⲉⲣⲕⲁⲣⲱⲕ ⲉϫⲱⲛ ⲛⲅϭⲱ ‘be not silent on us nor remain’ where SaA only has ⲙ̄ⲡⲣϭ ̄ ⲱ ‘do not remain’. This is explained by a shortening in SaA since there are no Greek variants and SaM accurately renders the Greek. SaA has not only omitted the first verb, but also reformulated the second verb so as to fit in the clause. SaM has a negative imperative of the verb ⲕⲁ ⲣⲱ⸗ ‘be silent’ and next a conjunctive from the verb ϭⲱ ‘to stop, desist, remain’, whereas SaA has only the second verb ‘to remain’ as a negative imperative.14 The verb ⲛⲟⲩϩⲙ ‘deliver’ is 2nd person singular masculine in SaMV but 3rd person singular masculine in SaA. SaMV continue the thought of the previous clause ‘do not remain without crying out to the Lord, your God, and deliver us from the hand of the Philistines’. Deliverance is connected to Samuel’s prayer, and not expressed as the result of it. SaA accords with the Greek, and its reading also fits the context better. Apparently, SaMV have preserved the original Sahidic translation, whereas SaA has reformulated either according to the Greek or simply according to the content of the passage. All three Sahidic manuscripts have the addition found in 108mg CI 119mg 121mg d f s 29 55 71 158 244mg 245 318 554 in Greek, with the asterisk sign (ø) in 56.15 According to Brock, this early addition probably renders a lost Hebrew variant. He conjectures that this reading appeared in the fifth column with an obelos since it is not present in the MT.16 The agreement between Sahidic manuscripts and the early dating of SaV are important here since they show that the Sahidic tradition contained this addition early, not just in some later manuscripts. Variation in this clause within SaV reads ⲛⲛ ̄ ⲉ ⲡⲁⲓ ϣⲱⲡ[ⲉ ⲙⲙⲟⲓ] ⲉⲧⲣⲁⲥⲁϩⲱ̄ⲓ̈ ‘may this not happen to me that I withdraw’, whereas SaA reads ⲛⲛ ̄ ⲉ ⲡⲁⲓ ϣⲱⲡⲉ ⲉⲧⲣⲁⲥⲁϩⲱⲓ ‘may this not happen that I withdraw’ and ⲛⲛⲉⲥϣⲱⲡⲉ ⲙⲙⲟⲓ ⲉⲧⲣⲁⲥⲁϩⲱⲱⲧ ‘may it not happen to me that I withdraw’ appears in SaM. The verb ⲥⲟⲟϩⲉ ‘remove’ is mostly used reflexively, and it also fits this clause. The forms differ in the manuscripts. SaA and SaV have similar wordings, with only a small difference, the lack of ⲙⲙⲟⲓ ‘to me’ in SaA. SaM has its own wording without ⲡⲁⲓ ‘this’ and with a personal infix as subject, but with ⲙⲙⲟⲓ like SaV. This minor variation does not alter the meaning, but implies that small changes occurred during the copying of the Sahidic manuscripts. Probably the word ⲙⲙⲟⲓ was in the original Sahidic translation since it appears in SaMV and SaA typically omits such small words. Whether the original read ⲛⲛ ̄ ⲉ ⲡⲁⲓ ϣⲱⲡⲉ or ⲛⲛⲉⲥϣⲱⲡⲉ is impossible to know.

14 Crum (Dictionary, 803b) gives the verb ϭⲱ as the equivalent used in Achmimic Hab 1: 13 to render παρασιωπάω where Sahidic and Bohairic use the verb ⲕⲁⲣⲟ. This does not change the fact that SaA has abbreviated the text, but it shows that there was a redundancy in the text, and it was eliminated in SaA. 15 The addition is not found in A B V O L CII a 121txt 509 244txt 460 707. Greek 121mg has the plus without the first phrase καὶ εἶπεν Σαμουήλ. 16 Brock, Recensions, 70.

143

Textual analyses

In the last clause, a slight difference occurs without changing the meaning: SaV reads ⲉⲧⲙ̄ⲧⲣⲁⲱϣ ⲉⲃⲟⲗ ‘that I do not cry’ against SaMA that read without ⲉⲃⲟⲗ. The last word in Greek, προσευχόμενον, has no equivalent in the Sahidic witnesses. 7: 9 καὶ ἔλαβεν Σαμουὴλ ἄρνα γαλαθηνὸν ἕνα καὶ ἀνήνεγκεν αὐτὸν ὁλοκαύτωσιν σὺν παντὶ τῷ λαῷ τῷ Κυρίῳ· καὶ ἐβόησεν Σαμουὴλ πρὸς Κύριον περὶ Ἰσραήλ, καὶ ἐπήκουσεν αὐτοῦ Κύριος. SaM

SaA

SaV

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲛⲟⲩϩⲓⲉⲓⲃ ⲉϥϩⲁ ⲧⲉϥⲉⲣⲱⲧⲉ· ⲁϥⲧⲁⲗⲟϥ ⲉϩⲣⲁⲓ ⲛⲟⲩϭⲗⲓⲗ· ⲙⲛ̄ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ̄ ⲙⲙⲁϩ ⲙⲡϫⲟⲉⲓⲥ· ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲱϣ ⲉϩⲣⲁⲓ ⲉⲡϫⲟⲉⲓⲥ ⲉϫⲙ̄ ⲡⲓⲏ̄ⲗ·̄ ⲁ ⲡϫⲟⲉⲓⲥ ⲥⲱⲧⲙ̄ ⲉⲣⲟϥ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲛⲟⲩⲉϩⲓⲉⲓⲃ ϩⲁ ⲡϥ̄ⲉⲣⲱⲧⲉ· ⲁϥⲧⲁⲗⲟϥ ⲉϩⲣⲁⲓ ⲛ̄ϭⲗⲓⲗ· ϩⲁ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ̄ ⲙ̄ⲙⲁϩ ⲡϫⲟⲉⲓⲥ· ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲱϣ ⲉϩⲣⲁⲓ ⲉⲡϫⲟⲉⲓⲥ ⲉϫⲉⲙ ⲡⲓ ̅ⲏ̄ⲗ ⲧⲏⲣϥ̄· ⲁ ⲡϫⲟⲉⲓⲥ ⲥⲱⲧⲉⲙ ⲉⲣⲟϥ·

ⲁⲩⲱ ⲁⲥⲁⲙⲟⲩ[ⲏⲗ ϫⲓ ⲛⲟⲩ]ϩⲓⲉⲓⲃ· ⲉϥϩⲁ ⲡⲉϥⲉⲣⲱⲧⲉ· ⲁ[ϥⲧⲁⲗⲟϥ] ⲉϩⲣⲁⲓ̈ ⲛ̄[ⲟⲩ]ϭⲗⲓⲗ, ϩⲁ ⲡⲗⲁⲟ[ⲥ ⲧⲏⲣϥ̄] ⲙ̄ⲙⲁϩ ⲡϫⲟⲉⲓⲥ· ⲁⲩⲱ ⲥⲁⲙⲟ̣[ⲩⲏⲗ ⲁϥ]ⲱϣ […]17 ⲉϩⲣⲁⲓ̈ ⲉⲡϫⲟⲉ[ⲓⲥ ⲉϫⲙ] ⲡⲓ ̅ⲏ̅ⲗ·̅ ⲁⲩⲱ ⲡϫⲟⲉⲓⲥ ⲁϥⲥ[ⲱⲧⲙ ⲉⲣⲟϥ]

In the first sentence, the construction in SaMA agrees against that in SaV: SaMA present Samuel in extraposition and an infix as the subject while SaV has Samuel as the subject. In the same sentence, the descriptions of the lamb differ in Sahidic manuscripts: SaMV read with a circumstantial ⲉϥϩⲁ against SaA that reads a plain ϩⲁ ‘in, at’, but this does not alter the meaning. In the next clause, SaMV read ⲛⲟⲩϭⲗⲓⲗ ‘as an offering’ against SaA that reads ⲛϭ ̄ ⲗⲓⲗ ‘as (an) offering’. In the same clause, the preposition used with ‘the people’ differs. SaM has ⲙⲛ̄ ‘with’, whereas SaAV have ϩⲁ which is used as a variant for ⲙⲛ.̄ 18 There is also a spelling difference towards the end of the clause; SaAV have ⲙ̄ⲙⲁϩ ⲡϫⲟⲉⲓⲥ against SaM that attests ⲙⲙⲁϩ ⲙⲡϫⲟⲉⲓⲥ. The last clause begins with the conjunction ⲁⲩⲱ in SaV but asyndetically in SaMA. The construction also differs since SaV has ‘Lord’ in extraposition and a personal infix as subject, but SaMA have ‘Lord’ as the subject. This verse has several such cases where one cannot discern the original Sahidic wording. Although SaA has perhaps shortened the words ⲉϥϩⲁ to ϩⲁ and ⲛⲟⲩϭⲗⲓⲗ to ⲛϭ ̄ ⲗⲓⲗ, the origins of the other discrepancies are more difficult to reconstruct.

17 Editors suppose that between the brackets there was perhaps ⲉⲡϫⲟ, but I would propose ⲉⲃⲟⲗ, which is used in a similar meaning in the previous verse. Additionally, there are some traces of ink visible, and the last character cannot be ⲟ. 18 Crum, Dictionary, 20a.

144

Affiliations of the Sahidic manuscripts and the Greek Vorlage

7: 10 καὶ ἦν Σαμουὴλ ἀναφέρων τὴν ὁλοκαύτωσιν, καὶ ἀλλόφυλοι προσῆγον εἰς πόλεμον ἐπὶ Ἰσραήλ· καὶ ἐβρόντησεν Κύριος ἐν φωνῇ μεγάλῃ ἐν τῇ ἡμέρᾳ ἐκείνῃ ἐπὶ τοὺς ἀλλοφύλους, καὶ συνεχύθησαν καὶ ἔπταισαν ἐνώπιον Ἰσραήλ. ἐν φωνῇ μεγάλῃ] > ἐν L 98´ 509 f 71; post ἐκείνῃ tr L(-82)

SaM

SaA

SaV

ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲛⲉϥⲧⲁⲗⲟ ⲉϩⲣⲁⲓ ⲙⲡⲉϭⲗⲓⲗ· ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϩⲱⲟⲩ· ⲁⲩⲉⲓ ⲉϩⲣⲁⲓ ⲉⲣⲱϥ ⲙⲡⲓⲏⲗ̄ ·̄ ⲉⲡⲟⲗⲩⲙⲉⲓ ⲛⲙ̄ⲙⲁⲩ· ⲡϫⲟⲉⲓⲥ ⲇⲉ ⲁϥϯ ⲛⲟⲩϩⲣⲟⲩⲃⲁⲓ̈ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ· ϩⲛ̄ ⲟⲩⲛⲟϭ ⲛϩⲣⲟⲟⲩ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩⲱ ⲁⲩϣⲧⲟⲣⲧⲣ̄· ⲁⲩϭⲱⲧⲡ ̄ ⲙ̄ⲡⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲙ̄ⲡⲓⲏ̄ⲗ·̄

ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲉⲛϥ̄ⲧⲁⲗⲟ ⲉϩⲣⲁⲓ ⲙ̄ⲡϥ̄ϭⲗⲓⲗ· ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲉⲓ ⲉⲣⲱϥ ⲙ̄ⲡⲓⲏ̅ ⲗ̄ ⲉⲡⲟⲗⲩⲙⲉⲓ ⲛⲙ̄ⲙⲁⲩ· ⲡϫⲟⲉⲓⲥ ⲇⲉ ⲁϥϯ ϩⲣⲟⲩⲃⲁⲓ ⲙ̄ⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϩⲛ̄ ⲟⲩⲛⲟϭ ⲛ̄ϩⲣⲟⲟⲩ· ⲁⲩϣⲧⲟⲣⲧⲣ̄ ⲁⲩⲱ ⲁⲩϭⲱⲧⲡ̄ ⲙ̄ⲡⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲛⲛ̄ϣⲏⲣⲉ ⲙ̄ⲡⲓ ̅ⲏ̄ⲗ·

ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲛⲉϥⲧⲁⲗⲟ [ⲉϩⲣⲁⲓ ⲙ]ⲡⲉϭⲗⲓⲗ· ⲛⲁⲗⲗⲟⲫⲩⲗⲟ[ⲥ ϩⲱⲟⲩ] ⲁⲩⲉⲓ ̅ ⲉⲣⲱϥ ⲙ̄ⲡⲓⲏ̅ ⲗ̅ ·̅ ⲉⲡⲟⲗⲉ̣ⲙ̣[ⲉⲓ ⲛⲙ]ⲙⲁⲩ· ⲡϫⲟⲉⲓⲥ ⲇⲉ ⲁϥϯ ϩⲣⲟⲩ[ⲃⲁⲓ] ⲙ̄ⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ· ⲛⲟⲩⲛⲟϭ ⲛϩⲣⲟⲟⲩ, ⲉϫⲛ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩ[ⲱ ⲁⲩ]ϣⲧⲟⲣⲧⲣ̄· ⲁⲩϭⲱⲧⲡ̄ ⲙ̄ⲡⲉⲙⲧ[ⲟ] ⲉⲃⲟⲗ ⲙ̄ⲡⲓ ̅ⲏ̅ⲗ·̅

The verbal form in the first clause is preterit in SaMV, but focalizing preterit in SaA. SaA also reads ‘his offering’ against SaMV ‘the offering’. However, this reading has no bearing on a theoretical Greek Vorlage, since Sahidic often uses possessive articles where one would expect definite articles.19 Without changing the meaning, SaM offers in the second clause the adverb ⲉϩⲣⲁⲓ which does not appear in SaAV. In SaA, ⲇⲉ connects this clause to the previous clause, while SaMV connect with ϩⲱⲟⲩ ‘for their part’. SaAV read ⲁϥϯϩⲣⲟⲩⲃⲁⲓ ‘he thundered’ but SaM has ⲁϥϯ ⲛⲟⲩϩⲣⲟⲩⲃⲁⲓ̈ with the indefinite article. Here, the Sahidic manuscripts preserve distinct word orders with respect to the prepositional phrases; SaMV offer first ‘on that day’, then ‘with mighty voice’ and finally ‘against Philistines’, the same order as in Lucianic manuscripts.20 In SaA, the order is ‘on that day’, ‘against Philistines’ and ‘with mighty voice’, without Greek support. The text of Rahlfs follows the order of the majority of manuscripts and the MT: voice, day, Philistines. One cannot determine which of the orders in Sahidic is original: either SaA has changed the order or SaMV have changed it according to a variant similar to that now found in Lucianic manuscripts. If the order of SaMV is original, one can assume a Greek source text like the one in Lucianic manuscripts or suppose that the translator, without an exact Greek Vorlage, reorganized the events of his source text according to his own style. In the last two clauses, SaMV agree with each other against SaA. The conjunction is placed before the first verb in SaMV but before the second in SaA.

19 See Mink, “Koptischen Versionen”, 221–2. 20 L-82.

145

Textual analyses

Before Israel SaA adds ‘the sons (of)’, without similar Greek variant, and against other Sahidic manuscripts. This last feature might indicate some copying mistake also incorporating the beginning of the next verse. 7: 11 καὶ ἐξῆλθον ἄνδρες Ἰσραὴλ ἐκ Μασσηφὰθ καὶ κατεδίωξαν τοὺς ἀλλοφύλους, καὶ ἐπάταξαν αὐτοὺς ἕως ὑποκάτω τοῦ Βαιθχόρ. SaM

SaA

SaV

ⲛϣⲏⲣⲉ ⲇⲉ ⲙⲡⲓⲏ̄ⲗ ̄ ⲁⲩⲉⲓ ⲉⲃⲟⲗ ϩⲙ̄ ⲙⲁⲥⲏⲫⲁⲧ· ⲁⲩⲡⲱⲧ ⲛⲥⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩϩⲓⲟⲩⲉ ⲛ̄ⲥⲱⲟⲩ· ϣⲁⲛⲧⲟⲩⲉ̄ⲓ ̅ ⲉϩⲣⲁⲓ ⲉⲃⲁ̄ⲓ ̅ⲭⲣⲟⲑ

ⲁⲩⲉⲓ ⲉϩⲣⲁⲓ ⲉⲙⲁⲥⲥⲏⲫⲁⲧ ⲁⲩⲡⲱⲧ ⲛ̄ⲥⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩϩⲓⲟⲩⲉ ⲛ̄ⲥⲱⲟⲩ· ϣⲁⲛⲧⲟⲩⲉⲓ ⲉϩⲣⲁⲓ ⲉⲭⲉⲃⲣⲱⲛ.

ⲛ̄ϣⲏⲣⲉ ⲇⲉ ⲙ[ⲡⲓⲏⲗ] ⲁⲩⲉⲓ ̅ ⲉⲃⲟⲗ ϩⲛ̄ ⲙⲁⲥⲏⲫⲁⲑ· ⲁⲩⲡ[ⲱⲧ] ⲛ̄ⲥⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩϩⲓⲟⲩ[ⲉ ⲛⲥⲱ]ⲟⲩ· ϣⲁⲛⲧⲟⲩⲉⲓ ̅ ⲉϩⲣⲁⲓ̈ ⲉⲕⲉ[ⲃⲣⲱⲛ]

The ending of the previous verse and the beginning of this one show a mistake in SaA. Probably, a skip of eye occurred between the two ⲙⲡⲓⲏⲗ occurrences. Thus, there is no mention of ⲛⲛϣⲏⲣⲉ ⲙⲡⲓⲏⲗ ‘sons of Israel’ in this verse but only in the previous verse. In the first clause, SaA differs from other Sahidic witnesses and reads ‘they came in Massefat’ against other manuscripts that read ‘they came out from Massefat’. There are also Greek manuscripts that read like SaA ‘they entered’: 93 121 246 245. But 121 has omitted καὶ ἐξῆλθον ἄνδρες Ἰσραὴλ; probably a homoioteleuton-based periblepsis from Ἰσραήλ to Ἰσραήλ. In other Greek manuscripts, it is easy to assume an eye skip from ΕΚ to ΕΙⲤ as long as the script was uncial: one writes sigma as a curve, and kappa with one vertical and a curve. The variant of SaA may be explained by redaction against a Greek manuscript with EIS or by a copyist’s confusion regarding the implicit subject in this passage. In the last clause, SaAV seem to share a secondary reading ⲭⲉⲃⲣⲱⲛ. It is understandable as a copying error ⲃⲉⲭ -> ⲭⲉⲃ. However, since the word is partially a reconstruction in SaV this is not a sufficient argument without further arguments for a dependence. There are also many variants in the Greek tradition, but not in accordance with SaAV. 7: 12 καὶ ἔλαβεν Σαμουὴλ λίθον ἕνα καὶ ἔστησεν αὐτὸν ἀνὰ μέσον Μασσηφὰθ καὶ ἀνὰ μέσον τῆς παλαιᾶς, καὶ ἐκάλεσεν τὸ ὄνομα αὐτοῦ Ἀβενέζερ, Λίθος τοῦ βοηθοῦ, καὶ εἶπεν Ἕως ἐνταῦθα ἐβοήθησεν ἡμῖν Κύριος. SaM

SaA

SaV

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲛⲟⲩⲱⲛⲉ· ⲁϥⲧⲁϩⲟϥ ⲉⲣⲁⲧϥ ϩⲛ ⲧⲙⲏⲧⲉ ⲙⲙⲁⲥⲏⲫⲱⲛ· ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛ̄ⲧⲉⲧⲟⲩⲙⲟⲩⲧⲉ ⲉⲣⲟⲥ ϫⲉ ⲧⲡⲁⲗⲁ̄ⲓ ̅ⲁ̄·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲛⲟⲩⲱⲛⲉ ⲁϥⲧⲁϩⲟϥ ⲉⲣⲁⲧϥ̄ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲙ̄ⲙⲁⲥⲥⲉⲫⲁⲑ ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛ̄ⲧⲉⲧⲟⲩⲙⲟⲩⲧⲉ ⲉⲣⲟⲥ ϫⲉ ⲧⲡⲉⲗⲁⲓ̈ⲁ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ· ⲁϥϫⲓ ⲛ̄ⲟⲩ[ⲱⲛⲉ] ⲁϥⲧⲁϩⲟϥ ⲉⲣⲁⲧϥ̄ ϩⲛ̄ ⲧⲙⲏⲧ[ⲉ ⲙⲙⲁ]ⲥⲏⲫⲁⲛ· ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛⲧ[ⲉⲧⲟⲩ]ⲙⲟⲩⲧⲉ ⲉⲣⲟⲥ· ϫⲉ ⲧⲡⲁⲗⲁⲓⲁ·

146

Affiliations of the Sahidic manuscripts and the Greek Vorlage

ⲁϥⲙⲟⲩⲧⲉ ⲉⲡⲉϥⲣⲁⲛ· ϫⲉ ⲁⲃⲉⲛⲉⲍⲉⲣ ⲉⲧⲉ ⲡⲁⲓ ⲡⲉ ⲡⲱⲛⲉ ⲙⲡⲃⲟⲏⲑⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲉⲓⲙⲁ· ⲁ ⲡϫⲟⲉⲓⲥ ⲃⲟⲏⲑⲉⲓ ⲉⲣⲟⲛ·

ⲁϥⲙⲟⲩⲧⲉ ⲉⲡϥ̄ⲣⲁⲛ ϫⲉ ⲁⲃⲉⲛⲉⲍⲉⲣ ⲡⲱⲛⲉ ⲛⲃ̄ ⲟⲏⲧⲟⲥ ⲉϥϫⲱ ⲙ̄ⲙⲟⲥ· ϫⲉ ϣⲁϩⲣⲁⲓ ⲉⲡⲓ̈ⲙⲁ ⲁ ⲡϫⲟⲉⲓⲥ ⲃⲟⲏⲑⲉⲓⲁ ⲉⲣⲟⲛ.

[ⲁϥ]ⲙⲟ[ⲩ]ⲧ[ⲉ] ⲉⲡⲉϥⲣⲁⲛ ϫⲉ ⲁⲃⲉ[ⲛⲉⲍⲉⲣ] [ⲉⲧⲉ ⲡⲁⲓ ⲡⲉ ⲡⲱⲛⲉ] ⲙ̄ⲡⲃⲟⲏⲑⲟⲥ· [ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ] ϣⲁϩⲣⲁⲓ̈ ⲉⲡⲉⲓ̈ⲙⲁ [ⲁ ⲡϫⲟⲉⲓⲥ ⲃ]ⲟⲏⲑⲓ ⲉⲣⲟⲛ·

The place is called ⲙⲁⲥⲏⲫⲱⲛ in SaM, but ⲙⲁⲥⲥⲉⲫⲁⲑ in SaA. SaV is partly reconstructed, but the last letters more closely resemble the reading of SaM: ]ⲥⲏⲫⲁⲛ. All Sahidic witnesses read an explanation ⲛⲧ ̄ ⲉⲧⲟⲩⲙⲟⲩⲧⲉ ⲉⲣⲟⲥ ϫⲉ ‘which is called’ before ⲧⲡⲁⲗⲁⲓⲁ, against Greek and other versions. This connects Sahidic manuscripts together and implies a common ancestor. The place name ⲁⲃⲉⲛⲉⲍⲉⲣ, a transcription from Hebrew, has an explanation, which is introduced in SaMV in a typical way ⲉⲧⲉ ⲡⲁⲓ ⲡⲉ ‘that is’21 but without any connective in SaA. In Greek, 98–379–731mg(sub ø) 121mg 68’ 314– 488–489 read here ὁ σημαίνει, but there is no need to suppose a dependence since the Sahidic text often has these explanations without a Greek source.22 7: 13 καὶ ἐταπείνωσεν Κύριος τοὺς ἀλλοφύλους, καὶ οὐ προσέθεντο ἔτι ἐπελθεῖν εἰς ὅριον Ἰσραήλ· καὶ ἐγενήθη χεὶρ Κυρίου ἐπὶ τοὺς ἀλλοφύλους πάσας τὰς ἡμέρας τοῦ Σαμουήλ. SaM

SaA

SaV

ⲁⲩⲱ ⲡϫⲟⲉⲓⲥ̄· ⲁϥⲑⲃⲃⲓⲟ̅ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ̄· ⲙⲡⲟⲩⲱϩ ⲉⲧⲟⲟⲧⲟⲩ ⲉ̄ⲉⲓ̄ ̅ ⲉϫ ̄ ⲛ̄ ⲛⲉⲧⲟϣ ⲙⲡⲓⲏ̄ⲗ.̄ ⲁⲩⲱ ⲧϭⲓϫ ⲙⲡϫⲟⲉⲓⲥ ⲁⲥϣⲱⲡⲉ ⲉ̄ϫⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲛⲥⲁⲙⲟⲩⲏⲗ·

ⲁⲩⲱ ⲡϫⲟⲉⲓⲥ ⲁϥⲑⲃ̄ⲃⲓⲟ ⲛ̄ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ̄ ⲙⲡⲟⲩϣ ⲛ̄ⲧⲟⲟⲧⲟⲩ ⲉϫⲛ̄ ⲛⲁⲡⲏⲓ ⲙ̄ⲡⲓ ̅ⲏ̅ⲗ· ⲁⲩⲱ ⲧϭⲓϫ ⲉⲙⲡϫⲟⲉⲓⲥ ⲁⲥϣⲱⲡⲉ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛ̄ⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲛ̄ⲥⲁⲙⲟⲩⲏⲗ.

ⲁⲩⲱ ⲡϫⲟ[ⲉⲓⲥ ⲁϥⲑⲃ]ⲃⲓⲟ ⲛ̄ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲙⲡⲟⲩ]ⲱ̣ ϩ ϭⲉ ⲉⲧⲟⲟⲧⲟⲩ, ⲉⲉⲓ̈ ⲉϫⲛ̄ [ⲛⲉⲧⲟ]ϣ̣ ⲙ̄ⲡⲓ ̅ⲏ̅ⲗ·̅ ⲁⲩⲱ ⲧϭⲓϫ ⲙ̄ⲡϫⲟ[ⲉⲓⲥ ⲁⲥ]ϣⲱⲡⲉ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ [ⲛⲛⲉϩ]ⲟ̣ⲟⲩ ⲧⲏⲣⲟⲩ ⲛ̄ⲥⲁⲙⲟⲩⲏⲗ·

In this verse, the second clause reads in SaMV ‘they came no longer against Israelite territories’, SaV with the conjunction ϭⲉ, while SaM is asyndetic. There is a striking difference between the Sahidic witnesses: the reading of SaA, ‘they were not able to lay their hands on those of the house of Israel’ is unique, lacking Greek and versional support. It graphically resembles the reading of other Sahidic manuscripts, and therefore, I regard it as a corruption from a text like that in SaMV. There are also other passages where SaA has confused the letters.

21 The reconstruction in SaV fits the available space and is a typical way to introduce an explanation. 22 See section 2.2.4.3.

147

Textual analyses

7: 14 καὶ ἀπεδόθησαν αἱ πόλεις ἃς ἔλαβον οἱ ἀλλόφυλοι παρὰ τῶν υἱῶν Ἰσραήλ, καὶ ἀπέδωκαν αὐτὰς τῷ Ἰσραὴλ ἀπὸ Ἀσκάλωνος ἕως Ἀζόβ, καὶ τὸ ὅριον Ἰσραὴλ ἀφείλαντο ἐκ χειρὸς ἀλλοφύλων· καὶ ἦν εἰρήνη ἀνὰ μέσον Ἰσραὴλ καὶ ἀνὰ μέσον τοῦ Ἀμορραίου. Ἀσκάλωνος B b d-44, 106, 107, 125, 370, 610 554] ακαρων CII-242’; ακρων 707; rel Ἀκκαρών Ἀζόβ B CI 121 64–130 29 71 244*] Γέθ A O L d 314–488–489 318(γαιθ) 554; αζεβεγεθ 55

SaM

SaA

SaV

ⲁⲩⲱ ⲙⲡⲟⲗⲓⲥ ⲛⲧⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϥⲓⲧⲟⲩ ⲛⲧⲟⲟⲧⲟⲩ ⲛⲛ ̄ ϣ ̄ ⲏⲣⲉ ⲙⲡⲓⲏⲗ̄ ·̄ ϫⲓⲛ ⲁⲥⲕⲁⲗⲱⲛ ϣⲁ ⲉϩⲣⲁⲓ ⲉⲅⲉⲑ· ⲁⲩⲱ ⲡⲧⲟϣ ⲙ̄ⲡⲓⲏ̄ⲗ ̄ ⲁⲩϥⲓⲧϥ̄ ⲛⲧⲟⲟⲧⲟⲩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲛⲉⲣⲉ ⲟⲩⲉⲓⲣⲏⲛⲏ ⲇⲉ ϣⲟⲟⲡ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛⲛϣⲏⲣⲉ ⲙⲡⲓ ̅ⲏ̄ⲗ·̄ ⲁⲩⲱ ϩⲛ ⲧⲙⲏⲧⲉ ⲙⲡⲁⲙⲟⲣⲣⲁ̄ⲓ ̅ⲟⲥ·

ⲁⲩⲱ ⲙ̄ⲡⲟⲗⲓⲥ ⲛⲧ ̄ ⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϥⲓⲧⲟⲩ ⲛⲧ ̄ ⲟⲟⲧⲟⲩ ⲛⲛϣ ̄ ⲏⲣⲉ ⲙ̄ⲡⲓⲏ̅ ⲗ̄ · ϫⲓⲛ ⲉⲁⲥⲕⲁⲗⲱⲛ ϣⲁϩⲣⲁⲓ ⲉⲅⲉⲑ· ⲁⲩⲱ ⲛ̄ⲧⲟϣ ⲙ̄ⲡⲓ ̅ⲏ̄ⲗ ⲁⲩϥⲓⲧⲟⲩ ⲛ̄ⲧⲟⲟⲧⲟⲩ ⲛ̄ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲛⲉⲣⲉ ⲟⲩⲉⲓⲣⲩⲛⲏ ϣⲟⲟⲡ ϩⲛ̄ ⲧⲙⲏⲧⲉ

[ⲁⲩⲱ ⲙ]ⲡⲟⲗⲓⲥ, ⲉⲛⲧⲁ ⲛⲁ ̄ ⲗⲗⲟⲫⲩⲗⲟⲥ [ϥⲓⲧⲟⲩ] ⲛⲧ ̄ ⲟⲟⲧⲟⲩ ⲛⲛ ̄ ϣ ̄ ⲏⲣⲉ ⲙⲡⲓⲏ̅ ⲗ̅ ̅ [ϫⲓⲛ ⲁⲥⲕ]ⲁⲗⲱⲛ, ϣⲁϩⲣⲁⲓ̈ ⲉⲅⲁⲓⲑ· [ⲁⲩⲱ ⲡⲧ]ⲟϣ ⲙ̄ⲡⲓⲏ̄ⲗ· ⲁⲩϥⲓⲧϥ̄ ⲛ̄ⲧⲟ[ⲟⲧⲟⲩ ⲛⲛ]ⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲛⲉⲣⲉ [ⲟⲩⲉⲓⲣⲏⲛ]ⲏ ⲇⲉ ϣⲟⲟⲡ ϩⲛ̄ ⲧⲙⲏⲧⲉ [ⲙⲡⲓ]ⲏ̅ⲗ·̅ ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲙ̄ⲡⲁⲙⲟⲣⲣⲁ ̣ ̣ⲓ ̣ⲟ̣ⲥ·̣

ⲙ̄ⲡⲁⲙⲱⲣⲣⲁⲓⲟⲥ·

The place names in all Sahidic witnesses are the same: ‘from Askalon to Geth’. The first one reads without Hexaplaric approximation, but the second attests a Hexaplaric approximation. The same names are found in the dgroup and Greek 554. All three Sahidic witnesses preserve the same names, and thus these names were already in the Sahidic tradition before SaV was copied in the 6th century. The territory of Israel is singular in SaMV but plural in SaA. This difference is a small-scale variant with no relevance to the Greek Vorlage. It is concordant in the manuscripts since they also read the resumptive suffix accordingly in singular/plural. The last sentence is connected to the preceding text with ⲇⲉ in SaMV but asyndetically in SaA. A homoiarchon mistake appears in SaA from the first ϩⲛ̄ ⲧⲙⲏⲧⲉ ‘between, in the middle’ to the second. SaM reads ‘sons of Israel’ against ‘Israel’ in SaV and the Greek manuscripts.

148

Affiliations of the Sahidic manuscripts and the Greek Vorlage

7: 15 καὶ ἐδίκαζεν Σαμουὴλ τὸν Ἰσραὴλ πάσας τὰς ἡμέρας τῆς ζωῆς αὐτοῦ· ἐδίκαζεν] ἔκρινε L 318 554mg

SaM

SaA

SaV

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϯϩⲁⲡ ⲉⲡⲓⲏ̄ⲗ ̄ ⲛⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲙⲡⲉϥⲱⲛϩ̄· ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϯϩⲁⲡ ⲉⲡⲓⲏⲗ̄ ·̄

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϯϩⲁⲡ ⲉⲡⲓ ̅ⲏ̅ⲗ ⲛ̄ⲛⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲙ̄ⲡϥ̄ⲱⲛϩ̄·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϯϩⲁⲡ ⲉⲡⲓ ̅ⲏ̅ⲗ̅ ⲛ̣̄ⲛ̣ⲉϩⲟⲟⲩ ⲧⲏⲣⲟⲩ ⲙ̄ⲡⲉϥⲱ[ⲛ]ϩ̄·

This and the subsequent verse contain a copying mistake in SaM. Τhe scribe has written ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϯϩⲁⲡ ⲉⲡⲓⲏⲗ̄ ̄ ‘And Samuel was judging Israel’ twice, and not ⲁⲩⲱ ⲛⲉϥⲃⲏⲕ ⲧⲉⲣⲟⲙⲡⲉ ⲉϥ- ‘and he went yearly’ at the beginning of 7: 16 like other Sahidic manuscripts. The scribe’s eye probably skipped from the second ⲁⲩⲱ back to the first. The number of characters equals that of two lines in SaM, which might explain the confusion if the parent manuscript had the same line length. The verb used in this verse in all Sahidic witnesses is noteworthy. The translator has rendered the verb δικάζω with the verb ϯϩⲁⲡ ‘to judge’ and the corresponding nouns with ⲇⲓⲕⲁⲥⲧⲏⲥ and ⲇⲓⲕⲁⲓⲱⲙⲁ.23 The verb κρίνω used in Lucianic manuscripts24 coincides in 1 Sam with the Greco-Coptic verb ⲕⲣⲓⲛⲉ.25 Thus, one can safely reconstruct which one of these two verbs lay in the Vorlage of the Sahidic translator.26 7: 16 καὶ ἐπορεύετο κατ᾽ἐνιαυτὸν ἐνιαυτὸν καὶ ἐκύκλου Βαιθὴλ καὶ τὴν Γάλγαλα καὶ τὴν Μασσηφάθ, καὶ ἐδίκαζεν τὸν Ἰσραὴλ ἐν πᾶσι τοῖς ἡγιασμένοις τούτοις. SaM

SaA

SaV

ⲕⲱⲧⲉ ⲉⲃⲁ̄ⲓ ̅ⲑⲏⲗ· ⲙⲛ̄ ⲅⲁⲗⲅⲁⲗⲁ· ⲙⲛ̄ ⲙⲁⲥⲏⲫⲁⲧ·

ⲁⲩⲱ ⲛⲉϥⲃⲏⲕ ⲧⲉⲣⲟⲙⲡⲉ ⲉϥⲕⲱⲧⲉ ⲉⲃⲁⲓⲑⲏⲗ ⲙⲛ̄ ⲕⲁⲗⲁⲁⲧ.ⲙⲛ̄ ⲙⲁⲥⲥⲏⲫⲁⲧ·

ⲁⲩⲱ ⲛⲉϥⲃⲏⲕ ⲧⲣ̄ⲣⲟⲙⲡⲉ· [ⲉϥ]ⲕⲱⲧⲉ ⲉⲃⲁⲓⲑⲏⲗ· ⲙⲛ̄ ⲅⲁⲗⲅⲁⲗⲁ ⲙⲛ̄ ⲙⲁⲥⲏⲫⲁⲑ·

23 There is one exception with this verb. In verse 24: 16, the rendering of δικάσαι μοι ἐκ χειρός σου is ⲛϥϩⲙⲟⲟⲥ ⲉϩⲣⲁⲓ ⲛⲙⲙⲁⲕ. In verses 8: 2–3, δικαίωμα is rendered ϩⲁⲡ and δικαστής is rendered ⲣⲉϥϯϩⲁⲡ. 24 Here, manuscripts L 318 554mg. 25 The only exception is in the Song of Hannah, verse 2: 10, where κρίνω is rendered with ϯϩⲁⲡ. 26 This is important information when trying to find the OG in καιγε- and non-καιγε sections in 1–2 Samuel. Raimund Wirth is preparing his dissertation “Die Septuaginta der Samuelbücher. Eine Untersuchung der griechischen Samuelübersetzung in philologischer und theologischer Perspektive.” One topic in his research is the καιγε recension, its scope and description.

149

Textual analyses ⲉϥϯϩⲁⲡ ⲉⲡⲓⲏⲗ̄ ̄ ϩⲛ ⲛⲉⲙⲁ ⲧⲏⲣⲟⲩ ⲉⲧⲧ[ⲃ]ⲃ[ⲏ]ⲩ

ⲉϥϯϩⲁⲡ ⲉⲡⲓⲏ̅ ⲗ̅ ϩⲛ̄ ⲛⲙ ̄ ⲁ ⲧⲏⲣⲟⲩ ⲉⲧⲧⲃⲃ̄ ⲏⲩ.

ⲉϥϯϩⲁⲡ ⲉⲡⲓⲏ̅ ⲗ̅ ̅ [ϩⲛ] ⲛⲉⲙⲁ ⲧⲏⲣⲟⲩ ⲉⲧⲧⲃⲃ̄ ⲏⲩ.

The beginning of this verse is omitted in SaM, probably resulting from a copying mistake. SaAV render the Greek text accurately. There is a difference in place names: SaMV read ⲅⲁⲗⲅⲁⲗⲁ but SaA has ⲕⲁⲗⲁⲁⲧ (γαλααδ in O).27 SaM distinguishes between ⲅⲁⲗⲅⲁⲗⲁ and ⲅⲁⲗⲁⲁⲇ according to the Greek Vorlage.28 The reading of SaA is secondary here, resulting either from careless copying in this manuscript, or from the reading now attested by O in Greek. 7: 17 ἡ δὲ ἀποστροφὴ αὐτοῦ εἰς Ἁρμαθάιμ, ὅτι ἐκεῖ ἦν ὁ οἶκος αὐτοῦ, καὶ ἐδίκαζεν ἐκεῖ τὸν Ἰσραήλ· καὶ ῷκοδόμησεν ἐκεῖ θυσιαστήριον τῷ κυρίῳ. SaM

SaA

SaV

ⲡⲉϥⲙⲁ ⲛⲉⲥⲑⲁ ⲇⲉ̄ ⲡⲉ ⲁⲣⲙⲁⲑⲁⲉⲓⲙ· ⲉⲃⲟⲗ ϫⲉ ⲛⲉⲣⲉ ⲡⲉϥⲏⲓ ⲙ̄ⲙⲁⲩ̄· ⲁⲩⲱ ⲛⲉϥϯϩⲁⲡ ⲉⲡⲓⲏ̄ⲗ ̄ ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙ̄ⲙⲁⲩ· ⲁⲩⲱ ⲁϥⲕⲱⲧ ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙ̄ⲙⲁⲩ ⲛⲟⲩⲑⲩⲥⲓⲁⲥⲧⲏⲣⲓⲟⲛ ⲙⲡϫⲟⲉⲓⲥ·

ⲡⲉϥⲙⲁ ⲛ̄ⲥⲱⲟⲩϩ ⲇⲉ ⲡⲉ ⲉⲣⲙⲁⲑⲉⲙ ϫⲉ ⲛⲉⲣⲉ ⲡϥ̄ⲏⲓ ⲙ̄ⲙⲁⲩ· ⲁⲩⲱ ⲉⲛⲉϥϯϩⲁⲡ ⲉⲡⲓ̈ⲏⲗ̅ ̅ ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙ̄ⲙⲁⲩ ⲉⲁϥⲕⲱⲧ ⲛ̄ϩⲏⲧϥ̄ ⲛⲟⲩⲑⲩⲥⲓⲁⲥⲧⲏⲣⲓⲟⲛ ⲙ̄ⲡϫⲟⲉⲓⲥ.

[ⲡⲉϥ]ⲙⲁ ⲛⲥ̄ⲑⲟ ⲇⲉ ⲡⲉ ⲁⲣⲙⲁⲑⲁⲓⲙ· [ⲉⲃⲟⲗ] ϫⲉ ⲛⲉⲣⲉ ⲡⲉϥⲏⲓ̈ ⲙ̄ⲙⲁⲩ· [ⲁⲩⲱ ⲛ]ⲉϥϯϩⲁⲡ ⲉⲡⲓ ̅ⲏ̅ⲗ̅ ⲙ̄ⲡⲙⲁ ⲉⲧⲙ̄[ⲙⲁⲩ]· ⲁⲩⲱ ⲁϥⲕⲱⲧ ⲙ̄ⲡⲙⲁ ⲉⲧⲙ̄[ⲙⲁ]ⲩ ⲛⲟⲩⲑⲩⲥⲓⲁⲥⲧⲏⲣⲓⲟⲛ ⲙ̄ⲡϫⲟⲉⲓⲥ·

The first clause contains an intriguing variant in SaA in comparison with SaMV. SaA renders ἀποστροφὴ αὐτοῦ with ⲡⲉϥⲙⲁ ⲛⲥ̄ ⲱⲟⲩϩ ‘place of gathering’ against ⲡⲉϥⲙⲁ ⲛⲉⲥⲑⲁ/ⲛⲥ̄ⲑⲟ29 ‘retreat, lodging’ in SaMV. Here, the variation in SaA is apparently not necessarily a mistake, but rather a possible translation of the underlying Greek text. In this verse, the tendency to shorten is obvious in SaA: this manuscript omits the first word from ⲉⲃⲟⲗ ϫⲉ,30 the conjunction ⲁⲩⲱ is omitted before building, and ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙ̄ⲙⲁⲩ ‘in that place, there’ is expressed with ⲛϩ̄ ⲏⲧϥ̄ ‘there’. SaA has a tendency to add ⲉ- to verbal forms, as here with ⲉⲛⲉϥϯϩⲁⲡ and ⲉⲁϥⲕⲱⲧ. In this verse, these forms are best understood as focalizing forms that emphasize the place: ‘in exactly this place’ he judged and built.

27 In other occurrences of this proper name, Sahidic reads ⲅⲁⲗⲅⲁⲗⲁ in 10: 8, 11: 14, 11: 15, 13: 4, 13: 7, 13: 12, 13: 15bis, 15: 12, 15: 21and 15: 33. In these passages only SaM is available. In verse 13: 8 Sahidic omits the whole clause. 28 In verse 13: 7 there are ⲅⲁⲗⲅⲁⲗⲁ and ⲅⲁⲗⲁⲁⲇ in the same verse. In 2 Sam 17: 26, SaM reads ⲅⲁⲗⲁⲁⲇ whereas SaA has ⲕⲁⲗⲁⲁⲧ, the same word as here. 29 SaM shows Fayyumic influence in its spelling (Crum, Dictionary, 437a). 30 The editors read ⲇⲉ in SaV, but the text is not clearly visible, and as well one can read ϫⲉ. From the word pair ⲉⲃⲟⲗ ϫⲉ one can omit ⲉⲃⲟⲗ without changing the meaning.

150

Affiliations of the Sahidic manuscripts and the Greek Vorlage

8: 1 Καὶ ἐγένετο ὡς ἐγήρασεν Σαμουήλ, καὶ κατέστησεν τοὺς υἱοὺς αὐτοῦ δικαστὰς τῷ Ἰσραήλ. SaM

SaA

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛⲧⲉⲣⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲣ̄ϩⲗ̄ⲗⲟ· ⲁϥⲕⲁⲑⲓⲥⲧⲁ ⲛⲛⲉϥϣⲏⲣⲉ ⲛⲇⲓⲕⲁⲥⲧⲏⲥ ⲉϫⲙ̄ ⲡⲓ ̅ⲏ̄ⲗ·̄

SaV ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲛ̄ⲧⲉⲣⲉ ⲥⲁ[ⲙⲟⲩⲏⲗ] ⲣ̄ϩⲗ̄ⲗⲟ ⲁϥⲕⲁⲑⲓⲥⲧⲁ ⲛ̄ⲛⲉϥ[..

Here SaA omits the whole verse and continues in the next verse. There is no obvious explanation for this omission, whether a mistake or intentional activity. 9: 21 καὶ ἀπεκρίθη Σαοὺλ καὶ εἶπεν Οὐχὶ ἀνδρὸς υἱὸς Ἰεμειναίου ἐγώ εἰμι τοῦ μικροῦ σκήπτρου φυλῆς Ἰσραήλ; καὶ τῆς φυλῆς τῆς ἐλαχίστης ἐξ ὅλου σκήπτρου Βενιαμείν; καὶ ἵνα τί ἐλάλησας πρὸς ἐμὲ κατὰ τὸ ῥῆμα τοῦτο; SaM

SaA

SaV

ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥⲟⲩⲱϣⲃ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲥⲁⲙⲟⲩⲏⲗ· ϫⲉ ⲙⲏ ⲁⲛⲅ̄ ⲡϣⲏⲣⲉ ⲁⲛ ⲛⲟⲩⲣⲱⲙⲉ ⲛⲓⲉ̄ⲙⲓⲛⲁⲓⲟⲥ· ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲕⲟⲩⲓ ⲛⲕ̄ ⲗⲏⲣⲟⲥ ⲧⲉⲫⲩⲗⲏ ⲛⲃⲉⲛⲓⲁⲙⲓⲛ· ⲧⲉⲫⲩⲗⲏ ⲉⲧⲥⲏϣ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲉⲕⲗⲏⲣⲟⲥ ⲧⲏⲣϥ ⲙⲡⲓⲏ̄ⲗ·̄ ⲁⲩⲱ ⲉⲧⲃⲉ ⲟⲩ· ⲁⲕϣⲁϫⲉ ⲛⲙ̄ⲙⲁⲓ ⲕⲁⲧⲁ ⲡⲉⲓϣⲁϫⲉ·

ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥⲟⲩⲱϣϥ̄ ⲉϥϫⲱ ⲙ̄ⲙⲟⲥ ⲛⲥⲁⲙⲟⲩⲏⲗ· ⲁⲛⲟⲕ ⲡⲉ ⲡϣⲏⲣⲉ ⲛⲟⲩⲣⲱⲙⲉ ⲛⲉ̄ ⲙⲓⲛⲛⲁⲓⲟⲥ ⲡⲉ ⲉⲃⲟⲗ ϩⲉⲙ ⲡⲕⲟⲩⲓ ⲛⲕ̄ ⲗⲏⲣⲟⲥ ϩⲛ̄ ⲧⲉⲫⲩⲗⲏ ⲙ̄ⲡⲓⲏ̅ ⲗ̅ ·̅ ⲧⲉⲫⲩⲗⲏ ⲉⲧⲥⲏϣ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲉⲕⲗⲏⲣⲟⲥ ⲧⲏⲣϥ̄ ⲛ̄ⲃⲉⲛⲓⲁⲙⲓ̈ⲛ· ⲁϩⲣⲟⲕ ϭⲉ ⲉⲕϣⲁϫⲉ ⲛⲙ̄ⲙⲁⲓ ⲕⲁⲧⲁ ⲛⲓϣⲁϫⲉ·

ⲧⲏⲣϥ̄ ⲛ̄ⲃⲉⲛⲓⲁ[ⲙⲓⲛ· ⲁϩⲣⲟⲕ ϭⲉ] ⲁⲕϣⲁϫⲉ ⲛⲙ̄ⲙⲁ[ⲓ ⲕⲁⲧⲁ ⲛⲉⲓϣⲁϫⲉ]

In the first clause, both Sahidic witnesses add ‘to Samuel’ against Greek. In the next clause, SaM has a rhetorical question in accordance with the Greek manuscripts: ‘Am I not the son of a Ieminite man?’ SaA has formulated it as an affirmation ‘I am a son of a Ieminite man’. The next distinction between the Sahidic manuscripts organizes the situation the other way round: SaA reads in accordance with Greek first Israel and then Beniamin against SaM that positions Beniamin first and Israel afterwards.31 SaV has preserved only the latter but seems to be in accordance with SaA. The last question in this verse ‘Why did you speak?’ is formulated in different ways in SaA and SaM, both of which accurately render the Greek question. The editors of SaV have reconstructed the question following SaA, probably based on the accordance in the word Beniamin in the beginning of the lacuna. However, this reconstruction could as well follow SaM, which

31 There are no corresponding variants in Greek.

151

Textual analyses

agrees with SaV immediately after the lacuna, reading ⲁⲕϣⲁϫⲉ, not ⲉⲕϣⲁϫⲉ as SaA. There are 18–22 characters per line, and reconstruction following SaM makes 22 characters, and that following SaA 20 characters. Thus, both alternatives fit the avalaible space. The last line is reconstructed ⲛⲉⲓϣⲁϫⲉ ‘these words’ in SaV,32 following ⲛⲓϣⲁϫⲉ of SaA and not ⲡⲉⲓϣⲁϫⲉ ‘this word’ of SaM. Actually, one cannot prioritize either of these two when reconstructing SaV.33 9: 22 καὶ ἔλαβεν Σαμουὴλ τὸν Σαοὺλ καὶ τὸ παιδάριον αὐτοῦ καὶ εἰσήγαγεν αὐτοὺς εἰς τὸ κατάλυμα, καὶ ἔθετο αὐτοῖς τόπον ἐν πρώτοις τῶν κεκλημένων ὡσεὶ ἑβδομήκοντα ἀνδρῶν. SaM

SaA

SaV

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲛⲥⲁⲟⲩⲗ ⲙⲛ ⲡⲉϥϩⲙ̄ϩⲁⲗ· ⲁϥϫⲓⲧⲟⲩ ⲉϩⲟⲩⲛ ⲉⲡⲙⲁ ⲛϣⲱⲡⲉ· ⲁϥⲕⲱ ⲛⲁⲩ ⲉⲃⲟⲗ ⲛⲟⲩⲙⲁ ⲉϥϫⲟⲥⲉ· [ϩⲛ] ⲛⲛⲟϭ ⲛⲧⲁⲩⲧⲁϩⲙⲟⲩ ⲉⲩⲛⲁⲉⲣ ϣϥⲉ ⲛⲣⲱⲙⲉ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲛⲥ̄ ⲁⲟⲩⲗ ⲙⲛ̄ ⲡⲉϥⲙⲁ

ⲁⲩⲱ ⲥⲁⲙⲟ[ⲩⲏⲗ ⲁϥϫⲓ ⲛ]ⲥⲁⲟⲩⲗ ⲙⲛ̄ ⲡⲉϥϩⲙ̄ϩⲁⲗ [ⲁϥϫⲓⲧⲟⲩ]34 ⲉϩⲟⲩⲛ ⲉⲡⲙⲁ ⲛ̄ϣⲱⲡⲉ· ⲁ[ϥⲕⲱ ⲛⲁⲩ] ⲉⲃⲟⲗ ⲛ̄ⲟⲩⲙⲁ ⲉϥϫⲟⲥⲉ [ϩⲛ ⲛⲛⲟϭ] ⲉⲛⲧⲁⲩⲧⲁϩⲙⲟⲩ, ⲉⲩⲛⲁ[ⲉⲣ ϣϥⲉ ⲛ]ⲣⲱⲙⲉ·

ⲛϣⲱⲡⲉ· ⲁϥⲕⲱ ⲛⲁⲩ ⲛⲟⲩⲙⲁ ⲉϥϫⲟⲥⲉ ⲙⲛ̄ ⲛ̄ⲛⲟϭ ⲛ̄ⲧⲁϥⲧⲁϩⲙⲟⲩ ⲉⲩⲛⲁⲣ̄ ϣⲃⲉ ⲛ̄ⲣⲱⲙⲉ·

SaA has a mistake, probably resulting from homoiarchon. It reads ‘Samuel took Saul and his lodging’35 against other Sahidic and most Greek manuscripts that offer ‘Samuel took Saul and his servant and brought them into the lodging’. One wonders why an error this obvious was not corrected. SaMV have as equivalent for τίθημι36 ⲕⲱ ⲛⲁⲩ ⲉⲃⲟⲗ ‘set for them’, without difference in meaning when compared with ⲕⲱ ⲛⲁⲩ in SaA. ἐν πρώτοις has as its rendering [ϩⲛ] ⲛⲛⲟϭ ‘among the great’ in SaMV but ⲙⲛ̄ ⲛⲛ ̄ ⲟϭ ‘with the great’ in SaA.37 Both SaM and SaV are lacunose at the relevant point, and this makes the evaluation more problematic. However, usually the rendering for the preposition ἐν is ϩⲛ, and, therefore, the SaA variant reading is probably secondary, resulting from careless copying.

32 Browne/Papini, “Frammenti”, 188. 33 There are no variants in Greek. The word is singular. 34 The editors give the first character as visible but add a dot below. In the photo, nothing is visible and the fragment brokes before this word. 35 κατάλυμα αὐτοῦ in L 318 554. Possibly the possessive article existed in the Vorlage of SaA without any connection to the Lucianic reading, since a possessive article is so often used as a variant for the definite article in Sahidic. 36 In Greek, there are some variants. The majority of the manuscripts read ἔθετο; εταξεν is found in L 554mg. 37 No variants for the preposition occur in Greek manuscripts.

152

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In the last clause, τῶν κεκλημένων is rendered with the 3rd person plural ‘those who were invited’ in SaMV against SaA that reads a 3rd person singular ‘whom he had invited’. 9: 23 καὶ εἶπεν Σαμουὴλ τῷ μαγείρῳ Δός μοι τὴν μερίδα ἣν ἔδωκά σοι, ἣν εἶπά σοι Θεῖναι αὐτὴν παρὰ σοί. SaM

SaA

SaV

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲡⲉϫⲁϥ ⲙⲡⲙⲁⲕⲓⲣⲟⲥ· ϫⲉ ⲙⲁ ⲛⲁⲓ ⲛⲧⲙⲉⲣⲓⲥ ⲛⲧⲁⲓϫⲟⲟⲥ ⲛⲁⲕ ⲉⲧⲃⲏⲏⲧⲥ̄· ϫⲉ ⲕⲁⲁⲥ ϩⲁϩⲧⲏⲕ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲡⲉϫⲁϥ ⲙ̄ⲡⲙⲁⲅⲓⲣⲟⲥ· ϫⲉ ⲙⲁ ⲛⲁⲓ ⲛ̄ⲧⲙⲉⲣⲓⲥ ⲛ̄ⲧⲁⲓⲧⲁⲁⲥ ⲛⲁⲕ· ⲧⲉⲛⲧⲁⲓϫⲟⲟⲥ ⲛⲁⲕ ⲉⲧⲃⲏⲏⲧⲥ̄ ϫⲉ ⲕⲁⲁⲥ ϩⲁϩⲧⲏⲕ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩ[ⲏⲗ ⲡⲉϫⲁϥ] ⲙ̄ⲡⲙⲁⲅⲓⲣⲟⲥ ϫⲉ ⲙⲁ ⲛ[ⲁⲓ ⲛⲧⲙⲉ]ⲣⲓⲥ ⲉⲛⲧⲁⲓ̈ⲧⲁⲁⲥ ⲛⲁⲕ, ⲛ[ⲧⲁⲓϫⲟⲟⲥ] ⲛⲁⲕ ⲉⲧⲃⲏⲏⲧⲥ̄ ϫⲉ ⲕⲁ[ⲁⲥ ϩⲁϩⲧⲏⲕ]

In this verse, SaM evidences a homoiarchon mistake and omits the phrase ‘which I gave to you’. Greek manuscripts38 also preserve the same omission which can happen in Sahidic as well as in Greek. Since SaAV do not contain this mistake, the reading probably arose as a copying mistake in SaM, without connection to a Greek text. Later on, SaMV agree in their reading ⲛⲧⲁⲓϫⲟⲟⲥ against ⲧⲉⲛⲧⲁⲓϫⲟⲟⲥ in SaA, but this does not alter the meaning ‘which I said’. 9: 24 καὶ ἦρεν ὁ μάγειρος τὴν κωλέαν καὶ παρέθηκεν αὐτὴν ἐνώπιον Σαούλ· καὶ εἶπεν Σαμουὴλ τῷ Σαούλ Ἰδοὺ ὑπόλιμμα, παράθες αὐτὸ ἐνώπιόν σου καὶ φάγε, ὅτι εἰς μαρτύριον τέθειταί σοι παρὰ τοὺς ἄλλους· ἀπόκνιζε. καὶ ἔφαγεν Σαοὺλ μετὰ Σαμουὴλ ἐν τῇ ἡμέρᾳ ἐκείνῃ. ἦρεν] ηψησεν A B 247 b d-44 554* 707; εψησεν 376 44; ηγειρεν 246; ειρεν 158 313; ηψησεν καὶ ἦρεν V; ὕψωσεν Grabe’s conjecture.

SaM

SaA

SaV

ⲁⲩⲱ ⲡⲙⲁⲅⲓⲣⲟⲥ ⲁϥⲡⲓⲥⲉ ⲛⲧⲕⲟⲗⲉⲁ· ⲉⲧⲉ ⲡⲙⲉⲣⲟⲥ ⲡⲉ· ⲁϥⲕⲁⲁϥ ϩⲁⲣⲱϥ ⲛⲥⲁⲟⲩⲗ· ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲥⲁⲟⲩⲗ· ϫⲉ ⲉⲓⲥ ⲡⲉⲓⲥⲉⲉⲡⲉ· ⲏ̄ ⲡⲉⲓϣⲁⲩ ⲕⲁⲁϥ ϩⲁⲣⲱⲕ ⲛⲅⲟⲩⲱⲙ· ϫⲉ ⲛ̄ⲧⲁⲩⲕⲁⲁϥ ⲛⲁⲕ ⲉⲩⲙⲛ̄ⲧⲙⲉⲧⲣⲉ· ⲙⲡⲁⲣⲁ ⲛⲕⲟⲟⲩⲉ̄· ⲡⲱϣ ⲛⲁⲕ ⲛⲅⲟⲩⲱⲙ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ· ⲁϥⲟⲩⲱⲙ ⲙⲛ̄ ⲥⲁⲙⲟⲩⲏⲗ· ⲙ̄ⲡⲉϩⲟⲟⲩ ⲉⲧⲙⲙⲁⲩ·̄

ⲁⲩⲱ ⲡⲙⲁⲅⲓⲣⲟⲥ ⲁϥⲡⲓⲥⲉ ⲛ̄ⲧⲕⲉⲗⲓⲁ ⲉⲧⲉ ⲡⲙⲉⲣⲟⲥ ⲡⲉ· ⲁⲩⲱ ⲁϥⲕⲁⲁϥ ϩⲁⲣⲱϥ ⲛ̄ⲥⲁⲟⲩⲗ· ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛ̄ⲥⲁⲟⲩⲗ ϫⲉ ⲉⲓⲥ ⲡⲓⲥⲉⲉⲡⲉ ⲏ ⲡⲓϣⲁⲩ· ⲕⲁⲁϥ ϩⲁⲣⲱⲕ ⲛ̄ⲅⲟⲩⲱⲙ· ϫⲉ ⲛ̄ⲧⲁⲩⲕⲁⲁϥ ⲛⲁⲕ ⲉⲧⲙⲛ̄ⲧⲙⲛ̄ⲧⲣⲉ· ⲡⲁⲣⲁ ⲛⲕ̄ ⲟⲟⲩⲉ ⲡⲱϣ ⲛⲅ̄ ⲟⲩⲱⲙ. ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲟⲩⲱⲙ ⲙⲛ̄ ⲥⲁⲙⲟⲩⲏⲗ ⲙ̄ⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ.

ⲁⲩⲱ ⲡⲙⲁⲅⲓⲣⲟⲥ, ⲁϥⲡ[ⲓⲥⲉ ⲛⲧⲕⲟ]ⲗⲉⲁ· ⲉⲧⲉ ⲡⲙⲏⲣⲟⲥ ⲡⲉ [ⲁϥⲕⲁⲁϥ] ϩⲁⲣⲱϥ ⲛ̄ⲥⲁⲟⲩⲗ· ⲡⲉϫⲁϥ

38 Manuscripts f.

[ϫⲉ ⲉⲓⲥ] ⲡⲉⲓ̈ⲥⲉⲉⲡⲉ, ⲏ̅ ⲡⲉⲓ̈ϣⲁⲩ ⲕⲁ[ⲁϥ ϩⲁ]ⲣⲱⲕ ⲛⲅⲟ̄ ⲩⲱⲙ· ϫⲉ ⲛ̄ⲧ[ⲁⲩⲕⲁⲁϥ] ⲛⲁⲕ ⲉⲩⲙⲛ̄ⲧⲙⲛ̄ⲧⲣⲉ ⲡ[ⲁⲣⲁ ⲛⲕⲟ]ⲟⲩⲉ· ⲡⲱϣ ⲛⲁⲕ ⲛⲅⲟ̄ ⲩ[ⲱⲙ· ⲁⲩⲱ] ⲥⲁⲟⲩⲗ· ⲁϥⲟⲩⲱⲙ ⲙⲛ ⲥ[ⲁⲙⲟⲩⲏⲗ] ⲙ̄ⲡⲉϩⲟⲟⲩ ⲉⲧ[ⲙ]ⲙⲁⲩ·̄

153

Textual analyses

The first verb in all Sahidic manuscripts is ‘to cook’. This is not the Old Greek, but the original Greek rendering was ἦρεν. If this verb was ever changed to ὕψωσεν, that is not preserved in any Greek manuscript, although John Ernest Grabe has offered this as a conjectural reading.39 The source text of the Sahidic translator obviously read the verb ἕψω ‘to cook’.40 The reading with this verb is based on a mistake on the Greek side, from ὑψόω to ἕψω. Accordingly, the Vorlage of the Sahidic translator was similar to A B O b d 554* 707,41 attesting an early, καιγε-type correction according to the MT in Greek, but in an erroneous form. In this passage, it is logical that the verb ‘to cook’ is used when the context is that of meat and eating. All three Sahidic witnesses have a common ancestor. This is observable in the phrase ⲉⲧⲉ ⲡⲙⲉⲣⲟⲥ ⲡⲉ ‘that is part’, the double reading ⲡⲉⲓⲥⲉⲉⲡⲉ· ⲏ ⲡⲉⲓϣⲁⲩ ‘this remainder or this piece’ and finally in the verb ⲛⲅⲟⲩⲱⲙ ‘eat’ after ⲡⲱϣ ‘divide’ against unanimous Greek evidence.42 SaMA read ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲥⲁⲟⲩⲗ ‘Samuel said to Saul’ against SaV that reads shorter ⲡⲉϫⲁϥ ‘he said’. The reading of SaV is in accordance with 125 489 as well as with the MT. SaM reads ⲙⲡⲁⲣⲁ ⲛⲕⲟⲟⲩⲉ ̄ ‘in preference to the others’ where SaAV read ⲡⲁⲣⲁ ⲛⲕ̄ ⲟⲟⲩⲉ, but this is only a spelling difference.43 SaA leaves out ⲛⲁⲕ ‘to you’ after ⲡⲱϣ against SaMV. This word is not found in Greek, and this is either shortening in SaA or approximation towards Greek. 9: 25 καὶ κατέβη ἐκ τῆς Βαμὰ εἰς τὴν πόλιν· καὶ διέστρωσαν τῷ Σαοὺλ ἐπὶ τῷ δώματι, καὶ ἐκοιμήθη. SaM

SaA

SaV

ⲁϥⲓ ⲇⲉ ⲉⲃⲟⲗ ϩⲛ̄ ⲃⲁⲙⲁ· ⲉϩⲣⲁⲓ ⲉⲧⲡⲟⲗⲓⲥ· ⲁⲩⲡⲱⲣϣ̄ ϩⲁⲣⲟϥ ⲛⲥⲁⲟⲩⲗ· ϩⲓ̈ ϫⲉⲛⲉⲡⲱⲣ ⲛϩ̄ ⲛϩ̄ ⲟⲉⲓⲧⲉ· ⲉⲛⲁϣⲱⲟⲩ· ⲁϥⲉⲛⲕⲟⲧⲕ·̄

ⲁⲩⲉⲓ ⲇⲉ ⲉⲃⲟⲗ ϩⲛ̄ ⲃⲁⲙⲁ ⲉϩⲣⲁⲓ ⲉⲧⲡⲟⲗⲓⲥ· ⲥⲁⲙⲟⲩⲏⲗ ⲁϥⲡⲱⲣϣ ⲛⲥ̄ ⲁⲟⲩⲗ· ϩⲓϫⲛ̄ ⲧϫⲉⲛⲉⲡⲱⲣ ⲁϥⲛⲕ̄ ⲟⲧⲕ.̄

ⲁ[ϥⲓ ⲇⲉ ⲉⲃⲟⲗ] ϩⲛ̄ ⲃⲁⲙⲁ· ⲉϩⲣⲁⲓ̈ ⲉⲧⲡⲟⲗ̣[ⲓⲥ· ⲁⲩⲡⲱ]ⲣϣ̄ ϩⲁⲣⲟϥ ⲛⲥ̄ ⲁⲟⲩⲗ· ϩⲓ̈ϫⲛ̣ [ϫⲉⲛⲉ]ⲡⲱⲣ ⲁϥⲛⲕ̄ ⲟⲧⲕ ̄

39 See Rahlfs’ apparatus. 40 The rendering of Hebrew ‫ רום‬hiphil 3rd person singular derives from ὑψόω in 1 Sam 2: 10 and 2 Kings 2: 13, Ps 148: 14, Dan 12: 7; ἐπαίρω in Exod 7: 20, Num 20: 11; ἀναιρέω in Num 17: 2, ἀπάρχω in 2 Chron 35: 7. 41 Aejmelaeus (“Kingdom at Stake”, 357, 362) lists those Greek manuscripts that usually contain these early corrections: A B O b and some further minuscules, with some variation in individual cases. 42 LaM has manduca, but this is only a marginal reading, and without any Greek witnesses it is difficult to suppose a shared source for LaM and Sahidic manuscripts. This LaM is a marginal note, and the note ends here, so we cannot determine if there was another verb. Probably the Latin verb ‘eat’ is here as a result of a confusion with the previous sentence. 43 Cf. Layton (Grammar, §210), who gives ⲙ̄ⲡⲃⲟⲗ ⲛ̄- and ⲡⲃⲟⲗ ⲛ̄- as synonyms.

154

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In this verse, a difference occurs in the first word: SaA reads 3rd person plural ‘they came’ against SaM and Greek that read 3rd person singular ‘he came’. The editors of SaV have reconstructed a 3rd person singular, but this is only a guess.44 The next difference arises in the following clause. SaM reads 3rd person plural ‘they spread’ like the majority of Greek manuscripts. SaA has the 3rd person singular like Greek CII 55.45 Additionally, SaA places the proper name Samuel in extraposition ‘Samuel, he spread’ without similar Greek variants. The prefix indicating person is not preserved in SaV, but it does not mention Samuel. Sahidic manuscripts translate ‘on the roof’ in different ways in this and the following verse: SaM reads ϩⲓ (ⲧ)ϫⲉⲛⲉⲡⲱⲣ, SaA reads ϩⲓϫⲛ̄ ⲧϫⲉⲛⲉⲡⲱⲣ and SaV reads first ϩⲓ̈ϫⲛ̣ [ϫⲉⲛⲉ]ⲡⲱⲣ and then ϩⲓ ϫⲉⲛⲉⲡⲱⲣ. This variation has no connection to Greek but is inner-Sahidic and illustrates the small changes that have accumulated during the transmission. This expression is complicated, since the preposition in its longer form ends with the same characters (ϫⲉⲛ) with which the word ‘roof’ begins. SaM has an addition at the end, ‘many clothes’ without a similar addition in other witnesses. 9: 26 καὶ ἐγένετο ὡς ἀνέβαινεν ὁ ὄρθρος, καὶ ἐκάλεσεν Σαμουὴλ τὸν Σαοὺλ ἐπὶ τῷ δώματι λέγων Ἀνάστα, καὶ ἐξαποστελῶ σε· καὶ ἀνέστη Σαούλ, καὶ ἐξῆλθεν αὐτὸς καὶ Σαμουὴλ ἕως ἔξω. SaM

SaA

SaV

ⲡⲟⲩⲟⲉⲓⲛ ⲇⲉ ⲛⲧⲉⲣⲉϥⲥⲱⲣ· ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲁϥⲙⲟⲩⲧⲉ ⲉⲥⲁⲟⲩⲗ· ϩⲓ ⲧϫⲉⲛⲉⲡⲱⲣ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ· ϫⲉ ⲧⲱⲟⲩⲛ ⲛⲅⲉ̄ⲓ·̅ ⲛⲧⲁϫⲟⲟⲩⲕ ⲉⲃⲟⲗ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ· ⲁϥⲧⲱⲟⲩⲛ ⲁϥⲉⲓ ⲉⲡⲉⲥⲏⲧ ⲉϥⲙⲟⲟϣⲉ ⲉⲃⲟⲗ· ⲛⲧⲟϥ ⲙⲛ ⲥⲁⲙⲟⲩⲏⲗ·

ⲡⲟⲩⲟⲉⲓⲛ ⲇⲉ ⲛ̄ⲧⲉⲣϥ̄ⲥⲱⲣ· ⲥⲁⲙⲟⲩⲏⲗ ⲇⲉ ⲁϥⲙⲟⲩⲧⲉ ⲉⲥⲁⲟⲩⲗ ϩⲓϫⲛ̄ ⲧϫⲉⲛⲉⲡⲱⲣ ⲉϥⲛ̄ⲕⲟⲧⲕ̄· ⲉϥϫⲱ ⲙ̄ⲙⲟⲥ ⲛⲁϥ ϫⲉ ⲧⲱⲟⲩⲛ ⲁⲙⲟⲩ ⲧⲁϫⲟⲟⲩⲕ ⲉⲃⲟⲗ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲧⲱⲟⲩⲛ ⲁϥⲓ ⲉⲡⲉⲥⲏⲧ ⲁϥⲙⲟⲟϣⲉ ⲙⲛ̄ ⲥⲁⲙⲟⲩⲏⲗ.

ⲡⲟⲩⲟ[ⲉⲓⲛ ⲇⲉ ⲛ]ⲧⲉⲣⲉϥⲥⲱⲣ· ⲥⲁⲙⲟⲩⲏ[ⲗ ⲁϥⲙⲟⲩⲧⲉ] ⲉⲥⲁⲟⲩⲗ· ϩⲓ ϫⲉⲛⲉⲡⲱⲣ [ⲉϥϫⲱ ⲙ]46ⲙⲟⲥ ⲛⲁϥ· ϫⲉ ⲧⲱⲟⲩⲛ ⲛⲅ̣[ⲉ̄ⲓ·̅ ⲛⲧⲁ]ϫⲟⲟⲩⲕ ⲉⲃⲟⲗ· ⲁⲩⲱ ⲥⲁⲟⲩ[ⲗ· ⲁϥⲧⲱ]ⲟⲩⲛ· ⲁϥⲉⲓ ̅ ⲉⲡⲉⲥⲏⲧ· ⲁϥⲙ[ⲟⲟϣⲉ] [ⲉⲃⲟⲗ· ⲛⲧⲟϥ ⲙⲛ] ⲥⲁⲙⲟⲩⲏⲗ·

SaA adds one redundant ⲉϥⲛⲕ̄ ⲟⲧⲕ ̄ ‘as he lay’ after ‘on the roof’. This is probably because of the similar phrase in the previous verse. Samuel’s speech to Saul is expressed with two verbs in Sahidic manuscripts, SaMV read ⲧⲱⲟⲩⲛ ⲛⲅⲉ̄ⲓ ̅ ‘rise and come’ and SaA has ⲧⲱⲟⲩⲛ ⲁⲙⲟⲩ ‘rise and come’. This difference concerns the verbal form, not the meaning. Sahi44 Browne/Papini, “Frammenti”, 189. 45 In Greek: the majority read διέστρωσαν; διέστρωσε(ν) CII 55, έστρωσαν L f 318, δη έστρωσαν 244. 46 The editors have written one character in the middle of the word with a dot below, but in the photo one cannot see anything but the lacuna.

155

Textual analyses

dic has a specific imperative form for only some verbs, and the infinitival imperative has the same function.47 In this clause, SaMV read an infinitive and a conjunctive, whereas SaA has an infinitive and an imperative. Typically, in a chain of imperatives, the conjunctive verbal form occurs after the first infinitive, as in this case in SaMV. Both Sahidic variants are used constantly, and it is impossible to say which one is more original. Greek has only one verb ‘Get up’.48 All Sahidic witnesses read ⲁϥⲉⲓ ⲉⲡⲉⲥⲏⲧ ‘He came down’ before the last clause, without a Greek source text. In the last clause, SaAV render the verb with a past tense ⲁϥⲙⲟⲟϣⲉ ‘he went’ against SaM which reads ⲉϥⲙⲟⲟϣⲉ ‘going’. This is again inner-Sahidic variation. SaA has a shorter reading that lacks the adverb ⲉⲃⲟⲗ and ⲛⲧⲟϥ ‘you’, found in SaMV. 9: 27 αὐτῶν καταβαινόντων εἰς μέρος τῆς πόλεως, καὶ Σαμουὴλ εἶπεν τῷ Σαούλ Εἰπόν τῷ νεανίσκῳ καὶ διελθέτω ἔμπροσθεν ἡμῶν, καὶ σὺ στῆθι ὡς σήμερον καὶ ἄκουσον ῥῆμα θεοῦ. SaM

SaA

SaV

ⲛⲧⲉⲣⲟⲩⲉⲓ ⲇⲉ ⲉⲃⲟⲗ ⲉⲩⲥⲁ ⲛⲧⲡⲟⲗⲓⲥ· ⲥⲁⲙⲟⲩⲏⲗ ⲡⲉϫⲁϥ ⲛ̄ⲥⲁⲟⲩⲗ· ϫⲉ ⲁϫⲓⲥ ⲙⲡⲉⲕϩⲉⲣϣⲓⲣⲉ· ⲛϥⲙⲟⲟϣⲉ ϩⲓϩⲏ ⲙⲙⲟⲛ· ⲛⲧⲟⲕ ⲇⲉ ⲁϩⲉⲣⲁⲧⲕ ⲙⲡⲉⲓⲙⲁ ⲧⲉⲛⲟⲩ· ⲛⲅⲥⲱⲧⲙ̄ ⲉⲡϣⲁϫⲉ ⲙⲡⲛⲟⲩⲧⲉ.

ⲛ̄ⲧⲉⲣⲟⲩⲉⲓ ⲇⲉ ⲉⲃⲟⲗ ⲉⲩⲥⲁ ⲛ̄ⲧⲉ ⲧⲡⲟⲗⲓⲥ ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛ̄ⲥⲁⲟⲩⲗ ϫⲉ ⲁϫⲓⲥ ⲙ̄ⲡϩⲣ̄ϣⲓⲣⲉ ⲛ̄ϥⲙⲟⲟϣⲉ ϩⲓⲑⲏ ⲙ̄ⲙⲟⲛ· ⲛ̄ⲧⲟⲕ ⲇⲉ ⲁϩⲉⲣⲁⲧⲕ̄ ⲙ̄ⲡⲓⲙⲁ ⲛ̄ⲅⲥⲱⲧⲙ̄ ⲉⲡϣⲁϫⲉ ⲙ̄ⲡⲛⲟⲩⲧⲉ.

ⲛⲧⲉ[ⲣⲟⲩⲉⲓ ⲇⲉ ⲉⲃⲟ]ⲗ ⲉⲩⲥⲁ ⲛⲧⲡⲟⲗⲓⲥ· ⲥⲁ[ⲙⲟⲩⲏⲗ ⲡ]ⲉϫⲁϥ ⲛ̄ⲥⲁⲟⲩⲗ· ϫⲉ ⲁϫⲓⲥ [ⲙⲡϩⲉⲣ]ϣⲓⲣⲉ ⲛϥ̄ⲙⲟⲟϣⲉ ϩⲓϩⲏ ⲙ[ⲙⲟⲛ· ⲛ]ⲧⲟⲕ ⲇⲉ ⲁϩⲉⲣⲁⲧⲕ̄ ⲙⲡⲉⲓ̈[ⲙⲁ ⲧⲉⲛ]ⲟⲩ ⲛⲅⲥ̄ ⲱⲧⲙ ⲉⲡϣⲁϫⲉ [ⲙⲡⲛⲟⲩ]ⲧⲉ.

In this verse, a few syntactical elements in SaMV agree against SaA. The genitive is expressed with ⲛⲧⲡⲟⲗⲓⲥ ‘of the town’ in SaMV but ⲛⲧ ̄ ⲉ ⲧⲡⲟⲗⲓⲥ in SaA without affecting the meaning. In the second clause, the word order differs: SaMV read ‘Samuel, he said to Saul’ whereas SaA offers ‘Said Samuel to Saul’. Within the Greek tradition, Lucianic manuscripts49 read the same word order as SaA. The construction also differs in Sahidic witnesses. SaMV attest the proper name Samuel in extraposition and the verb with a personal infix as subject, whereas SaA offers the verb with Samuel as the subject. Both constructions fit this context, and arguing for the priority of one of them is impossible.

47 Layton, Grammar, §365–6. 48 However, the verb ‘come’ appears in the Armenian. Because this reading explicitates contextually implicit features, the variant does not necessary derive from a common Greek source behind the variant in Sahidic and Armenian. 49 L.

156

Affiliations of the Sahidic manuscripts and the Greek Vorlage

The word ‘young man’ has in SaM a possessive article, but in SaA a definite article. The editors have reconstructed SaV following SaA, but there is also space for a reconstruction following SaM; 20 characters following SaA and 22 characters following SaM. The discrepancy is insignificant since often the possessive article is only one form of definiteness in Sahidic.50 The phrase ὡς σήμερον is rendered ⲙⲡⲉⲓⲙⲁ ⲧⲉⲛⲟⲩ ‘in this place now’. Probably, Sahidic witnesses had in their Vorlage ὧδε σήμερον like 19–108 in Greek,51 since ὧδε is often rendered with ⲙⲡⲉⲓⲙⲁ/ⲉⲡⲉⲓⲙⲁ. The word ‘now’ is left out in SaA against SaMV. This does not affect the affiliation but reveals the shortening tendency of SaA. 10: 1 καὶ ἔλαβεν Σαμουὴλ τὸν φακὸν τοῦ ἐλαίου καὶ ἐπέχεεν ἐπὶ τὴν κεφαλὴν αὐτοῦ, καὶ ἐφίλησεν αὐτὸν καὶ εἶπεν αὐτῷ Οὐχὶ κέχρικέν σε Κύριος εἰς ἄρχοντα ἐπὶ τὸν λαὸν αὐτοῦ, ἐπὶ Ἰσραήλ; καὶ σὺ ἄρξεις ἐν λαῷ Κυρίου, καὶ σὺ σώσεις αὐτὸν ἐκ χειρὸς ἐχθρῶν αὐτοῦ κυκλόθεν. καὶ τοῦτό σοι τὸ σημεῖον ὅτι ἔχρισέν σε Κύριος ἐπὶ κληρονομίαν αὐτοῦ εἰς ἄρχοντα·

SaM

SaA

SaV

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲙⲡⲉⲫⲁⲅⲟⲥ· ⲏ̄ ⲡⲉϩⲛⲁⲁⲩ ⲛⲛ ̄ ⲉϩ· ⲁϥⲡⲁϩⲧϥ̄ ⲉϫⲛ̄ ⲧⲁⲡⲉ ⲛⲥⲁⲟⲩⲗ· ⲁⲩⲱ ⲁϥϯⲡⲓ ⲉⲣⲱϥ ⲡⲉϫⲁϥ ⲛⲁϥ· ϫⲉ ⲡϫⲟⲉⲓⲥ ⲁϥⲧⲁϩⲥⲕ·̄ ⲉⲧⲣⲉⲕⲉⲣⲁⲣⲭⲱⲛ ⲉϫⲙ̄ ⲡⲉϥⲗⲁⲟⲥ ⲡⲓⲏⲗ̄ ·̄ ⲁⲩⲱ ⲛⲧⲟⲕ ⲡⲉⲧⲛⲁⲣ̄ ⲛⲟϭ ⲉϫⲙ̄ ⲡⲗⲁⲟⲥ ⲙⲡⲓⲏ̄ⲗ·̄ ⲛ̄ⲅⲛⲁϩⲙⲉϥ ⲉⲧⲟⲟⲧⲟⲩ ⲛ̄ⲛⲉϥϫⲁϫⲉ ⲉⲧⲕⲱⲧⲉ ⲉⲣⲟϥ·52 ⲁⲩⲱ ⲡⲁⲓ ⲡⲉ ⲡⲙⲁⲉⲓⲛ ⲉⲧⲛⲁϣⲱⲡⲉ ⲛⲁⲕ· ϫⲉ ⲁ ⲡϫⲟⲉⲓⲥ ⲧⲁϩⲥⲕ ⲉⲧⲣⲉⲕⲣ̄ⲁⲣⲭⲱⲛ· ⲉϫⲛ ⲧⲉϥⲕⲗⲏⲣⲟⲛⲟⲙⲓⲁ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ ⲁϥϫⲓ ⲙ̄ⲡⲉⲫⲁⲅⲟⲥ ⲛⲛ ̄ ϩ̄ ⲁϥⲡⲁϩⲧϥ̄ ⲉϫⲛ̄ ⲧⲁⲡⲉ ⲛⲥ̄ ⲁⲟⲩⲗ· ⲁⲩⲱ ⲁϥϯⲡⲉⲓ ⲉⲣⲱϥ ⲡⲉϫⲁϥ ⲛⲁϥ· ϫⲉ ⲉⲓⲥ ⲡϫⲟⲉⲓⲥ ⲁϥⲧⲁϩⲥⲕ ̄ ⲉⲧⲣⲕⲣ̅ ̅ ⲁⲣⲭⲱⲛ ⲉϫⲛ̄ ⲡϥ̄ⲗⲁⲟⲥ ⲡⲓⲏ̅ ⲗ̅ ·̅ ⲁⲩⲱ ⲛ̄ⲧⲟⲕ ⲟⲛ ⲡⲉⲧⲛⲁⲣ̄ⲛⲟϭ ϩⲙ̄ ⲡⲗⲁⲟⲥ ⲙ̄ⲡⲓ ̅ⲏ̅ⲗ·̅ ⲛ̄ⲕⲛⲁϩⲙϥ̄ ⲉⲧⲟⲟⲧⲟⲩ ⲛ̄ⲛⲉⲩϫⲁϫⲉ ⲉⲧⲕⲱⲧⲉ ⲉⲣⲟⲟⲩ. ⲡⲁⲓ ⲡⲉ ⲡⲙⲁⲉⲓⲛ ⲉⲧⲛⲁϣⲱⲡⲉ ⲛⲁⲕ ϫⲉ ⲁ ⲡϫⲟⲉⲓⲥ ⲧⲁϩⲥⲕ̄ ⲉⲧⲣⲉⲕⲣ̅ⲣⲣ̅ ⲟ ⲉϫⲛ̄ ⲧϥ̄ⲕⲗⲏⲣⲟⲛⲟⲙⲉⲓⲁ·

ⲁⲩⲱ ⲥⲁⲙⲟⲩⲏⲗ· ⲁϥϫⲓ [ⲙⲡⲉⲫⲁⲅ]ⲟⲥ, ⲏ̄ ⲡⲉϩⲛⲁⲁⲩ ⲛⲛⲉϩ· [ⲁϥⲡⲁϩⲧϥ̄] ⲉϫⲛ̄ ⲧⲁⲡⲉ ⲛⲥ̄ ⲁⲟⲩⲗ· ⲁⲩ[ⲱ ⲁϥϯⲡⲓ ⲉ]ⲣⲱϥ· ⲡⲉϫⲁϥ ⲛⲁϥ· ϫⲉ [ⲉⲓⲥ ⲡϫⲟⲉⲓ]ⲥ ⲁϥⲧⲁϩⲥⲕ,̄ ⲉⲧⲣⲉⲕⲣⲁ̄ ⲣ[ⲭⲱⲛ ⲉϫ]ⲙ̣ ⲡⲉϥⲗⲁⲟⲥ ⲡⲓⲏ̅ ⲗ̅ ·̅ ⲁⲩⲱ [ⲛⲧⲟⲕ ⲡⲉ]ⲧⲛⲁⲣ̄ⲛⲟϭ ϩⲙ ⲡⲗⲁⲟⲥ [ⲙⲡⲓⲏ]ⲗ̣· ⲛⲅⲛ̄ ⲁϩⲙⲉϥ ⲉⲧⲟⲟⲧⲟⲩ [ⲛⲛⲉϥ]ϫⲁϫⲉ ⲉⲧⲕⲱⲧⲉ ⲉⲣⲟϥ· ⲁⲩⲱ [ⲡⲁⲓ ⲡⲉ ⲡ]ⲙ̣ ⲁ̣ⲉⲓⲛ ⲉⲧⲛⲁϣⲱⲡⲉ ⲛⲁⲕ [ϫⲉ ⲁ ⲡϫⲟⲉ]ⲓⲥ ⲧⲁϩⲥⲕ̄ ⲉⲧⲣⲉⲕⲣ̄ⲁⲣ[ⲭⲱⲛ ⲉϫⲛ] ⲧⲉ̣ϥⲕⲗⲏⲣⲟⲛⲟⲙⲓⲁ·

In the first clause, the Greek loanword ⲫⲁⲅⲟⲥ is used in Sahidic and supplied with an explanation ⲏ̄ ⲡⲉϩⲛⲁⲁⲩ ‘or the vessel’ in SaMV. Since there are many passages where SaA provides this kind of explanation, I see its omission as secondary here.53

50 51 52 53

See also Mink, “Koptischen Versionen”, 221–2. Greek variants are as follows: εὡς σήμερον b 68’, ωδε σημερον 19–108, ὡς > 244 460. Drescher begins verse 10: 2 here. All the passages are analyzed in section 2.2.4.3.

Textual analyses

157

SaA begins Samuel’s speech with ⲉⲓⲥ ‘Look’. In Greek, ἰδού is attested by 245 707.54 The editors of SaV have reconstructed ἰδού with ⲉⲓⲥ.55 I do not agree, since, with this word, there are 24 characters on that line and 20–22 characters on surrounding lines. Since SaM does not offer this word, it should not appear in the reconstruction. Brooke/McLean in their apparatus cite SaM as following those manuscripts that read ὅτι instead of οὐχί.56 Indeed, Sahidic does not contain a negation here, and the word ϫⲉ usually renders ὅτι. However, throughout the manuscript the translator uses ϫⲉ when introducing a speech. Therefore, the citation of Sahidic here gives the wrong impression of dependence, although SaM could as well be cited with Ethiopic and Latin as omitting the negation. The rendering of καὶ σὺ ἄρξεις ἐν λαῷ κυρίου has some interesting features in Sahidic. First, all Sahidic witnesses read against unanimous Greek evidence Israel, not Lord as the last word. This implies a common origin of Sahidic manuscripts. SaA reads ⲟⲛ ‘again, also, still’ against other Sahidic witnesses. The preposition used with the people is ϩⲙ̄ ‘in, at, on’ in SaAV but ⲉϫⲙ̄ ‘upon, over’ in SaM. The reading of SaM has its explanation in the surrounding phrases where the governance is expressed with the preposition ⲉϫⲛ. These variants result from inner-Sahidic variation, without connection to Greek. When SaMV refer to the people, they offer a different phrase than that found in SaA (ἐκ χειρὸς ἐχθρῶν αὐτοῦ): ⲛⲛ ̄ ⲉϥϫⲁϫⲉ ⲉⲧⲕⲱⲧⲉ ⲉⲣⲟϥ ‘his enemies who surround him’ in SaMV versus ⲛⲛ ̄ ⲉⲩϫⲁϫⲉ ⲉⲧⲕⲱⲧⲉ ⲉⲣⲟⲟⲩ ‘their enemies who surround them’ in SaA. This same variation, typical when the word ‘people’ is referred to, also occurs in the Greek, where 127 707 read αὐτῶν against αὐτοῦ of the majority.57 The following clause begins with the conjunction ⲁⲩⲱ in SaMV whereas SaA lacks it.58 In the rendering of the last words, εἰς ἄρχοντα, Sahidic manuscripts differ. SaMV read ⲉⲧⲣⲉⲕⲣⲁ̄ ⲣⲭⲱⲛ ‘that you rule’ against SaA which alone reads ⲉⲧⲣⲉⲕⲣⲣ̅ ⲣ̅ ⲟ ‘that you reign’. The reading of SaA reflects the context, since it is the usual way to speak of a king (Cf. 10: 16). In the Greek, no variants for ἄρχοντα exist. Probably, SaA is secondary in this case since it offers the more common expression. 10: 2 ὡς ἂν ἀπέλθῃς σήμερον ἀπ᾽ἐμοῦ, καὶ εὑρήσεις δύο ἄνδρας πρὸς τοῖς τάφοις Ῥαχὴλ ἐν τῷ ὁρίῳ Βενιαμεὶν ἁλλομένους μεγάλα, καὶ ἐροῦσίν σοι Εὕρηνται αἱ ὄνοι ἃς ἐπορεύθητε ζητεῖν· καὶ ἰδοὺ ὁ πατήρ σου ἀποτετίνακται τὸ ῥῆμα τῶν ὄνων, καὶ ἐδαψιλεύσατο δι᾽ὑμᾶς λέγων Τί ποιήσω ὑπὲρ τοῦ υἱοῦ μου;

54 There is ορα οτι in 98–379–731mg 554 mg. Α plain οτι as a variant for οὐχί is found in L 731txt f s-92 29 55 244 460, οτι οὐχί 799, νυν 376. 55 Browne/Papini, “Frammenti”, 190. 56 Brooke/McLean, Old Testament, 28. 57 Omission of αὐτοῦ d-68, 122 554. 58 καί unanimously in Greek.

158

Affiliations of the Sahidic manuscripts and the Greek Vorlage

SaM

SaA

SaV

ⲉⲕϣⲁⲛⲙⲟⲟϣⲉ ⲉ̄ⲃⲟⲗ ϩⲓⲧⲟⲟⲧ· ⲙ̄ⲡⲟⲟⲩ· ⲕⲛⲁϭⲓⲛⲉ ⲛⲣⲱⲙⲉ ⲥⲛⲁⲩ ϩⲁⲧⲉⲛ ⲡⲧⲁⲫⲟⲥ ⲛϩⲣⲁⲭⲏⲗ· ϩⲙ ⲡⲧⲟⲟⲩ ⲛⲃⲉⲛⲓⲁⲙⲓⲛ· ⲉⲩⲉⲓⲣⲉ ⲛϩⲉⲛⲛⲟϭ ⲛⲃⲟϭⲥ ϩⲁ ⲡⲣⲁϣⲉ· ⲁⲩⲱ ⲛⲧⲟⲟⲩ ⲥⲉⲛⲁϫⲟⲟⲥ ⲛⲁⲕ· ϫⲉ ⲁⲩϩⲉ· ⲉⲛⲉⲟⲟⲩ· ⲛⲁⲓ ⲛⲧⲁⲧⲉⲧⲛⲃ̄ ⲱⲕ ⲉϣⲓⲛⲉ ⲛⲥⲱⲟⲩ· ⲡⲉⲕⲉⲓⲱⲧ ⲁϥⲛⲉϩ ⲡⲣⲟⲟⲩϣ ⲉⲃⲟⲗ ⲛⲛⲉⲧⲃ̄ⲛⲟⲟⲩⲉ· ⲁⲩⲱ ϥϥⲓⲣⲟⲟⲩϣ ⲉⲙⲁⲧⲉ ⲉⲧⲃⲉ ⲧⲏⲩⲧⲛ· ⲉϥϫⲱ ⲙⲙⲟⲥ· ϫⲉ ⲟⲩ ⲡⲉϯⲛⲁⲁⲁϥ ⲉⲧⲃⲉ ⲡⲁϣⲏⲣⲉ.

ⲉⲕϣⲁⲛⲙⲟⲟϣⲉ ⲉⲃⲟⲗ ϩⲓⲧⲟⲟⲧ ⲙ̄ⲡⲟⲟⲩ· ⲉⲕⲛⲁϭⲛ̄ ⲣⲱⲙⲉ ⲥⲛⲁⲩ ϩⲁⲧⲙ̄ ⲡⲧⲁⲫⲟⲥ ⲛ̄ϩⲣⲁⲭⲏⲗ ϩⲙ̄ ⲡⲧⲟⲟⲩ ⲛ̄ⲃⲉⲛⲉⲓⲁⲙⲓⲛ· ⲉⲩⲉⲓⲣⲉ ⲛ̄ϩⲛ̄ⲛⲟϭ ⲛ̄ⲃⲟϭⲥ ϩⲁ ⲡⲣⲁϣⲉ· ⲁⲩⲱ ⲥⲉⲛⲁϫⲟⲟⲥ ⲛⲁⲕ ϫⲉ ⲁⲩϩⲉ ⲉⲛⲉⲟⲟⲩ ⲛⲁⲓ ⲛⲧ ̄ ⲁⲧⲉⲧⲛⲃ̄ ⲱⲕ ⲉϣⲓⲛⲉ ⲛⲥ̄ ⲱⲟⲩ· ⲡⲉⲕⲓⲱⲧ ⲁϥⲛϩ̄ ⲡⲣⲟⲟⲩϣ ⲉⲃⲟⲗ ⲉⲧⲃⲉ ⲛ̄ⲧⲃ̄ⲛⲟⲟⲩⲉ ⲁⲩⲱ ϥϥⲓⲣⲟⲟⲩϣ ⲙ̄ⲙⲁⲧⲉ ⲉⲧⲃⲉ ⲧⲏⲩⲧⲛ̄ ⲉϥϫⲱ ⲙ̄ⲙⲟⲥ· ϫⲉ ⲟⲩ ⲡⲉ ϯⲛⲁⲁⲁϥ ⲉⲧⲃⲉ ⲡⲁϣⲏⲣⲉ.

[ⲉⲕϣⲁⲛ]ⲙⲟⲟϣⲉ ⲉⲃⲟⲗ ϩⲓⲧⲟⲟⲧ [ⲙⲡⲟⲟⲩ] ⲕⲛⲁϭⲓⲛ̣[ⲉ] ⲛ̄ⲣⲱⲙⲉ ⲥⲛⲁⲩ· [ϩⲁⲧⲉⲙ ⲡ]ⲧⲁⲫⲟⲥ ⲛϩⲣⲁⲭⲏⲗ ϩⲙ ⲡⲧ ⲟ[ⲟⲩ ⲛⲃⲉⲛ]ⲓⲁⲙⲉⲓⲛ· ⲉⲩⲉⲓⲣⲉ ⲛ̄ϩⲉⲛ[ⲛⲟϭ ⲛⲃ]ⲟϭⲥ̄ ϩⲁ ⲡⲣⲁϣⲉ· ⲁⲩⲱ ⲛⲧ ̄ ⲟ[ⲟⲩ ⲥⲉⲛⲁ]ϫⲟⲟⲥ ⲛⲁⲕ· ϫⲉ ⲁⲩϩⲉ ⲉⲛⲉ[ⲟⲟⲩ· ⲛ]ⲁ̣[ⲓ] ⲛⲧⲁⲧⲉⲧⲛⲃ̄ ⲱⲕ· ⲉϣⲓ[ⲛⲉ ⲛⲥ]ⲱⲟⲩ· ⲡⲉⲕⲉⲓⲱⲧ ⲇⲉ ⲁϥⲛⲉϩ [ⲡⲣⲟⲟⲩ]ϣ ⲉⲃⲟⲗ ⲛ̅ⲛⲧ ̅ ⲃ̄ⲛⲟⲟⲩⲉ· ⲁⲩⲱ [ϥϥⲓⲣⲟ]ⲟⲩϣ ⲙ̄ⲙⲁⲧⲉ ⲉⲧⲃⲉ ⲧⲏⲩ[

In the second clause, SaMV render a future verbal form and its object marked with the preposition ⲛ. SaA reads the verbal form with an initial ⲉ-. This can be understood as a marker of circumstantial or focalizing conversion of a future verbal form. If this is circumstantial, then there is no main clause yet for the preceding clause but the main clause comes in the following ⲁⲩⲱclause ‘And they will say to you’. If this is a focalizing conversion, then the focalized point probably is the adverbial expression ϩⲁⲧⲙ̄ ⲡⲧⲁⲫⲟⲥ ⲛϩ̄ ⲣⲁⲭⲏⲗ ‘by the tomb of Rachel’. Further, SaA connects the object directly to a prenominal state of the infinitive, without using the preposition like SaMV. τάφοις is plural, but Sahidic reads singular. However, Liddell/Scott give this noun in plural and mentions that it can refer to a single grave.59 ἐν τῷ ὁρίῳ Βενιαμείν is rendered ϩⲙ ⲡⲧⲟⲟⲩ ⲛⲃⲉⲛⲓⲁⲙⲓⲛ ‘in the mountain of Benjamin’ in Sahidic manuscripts. The translator has misread the Greek word ὅριον ‘boundary, region’ and rendered the word ὄρος ‘mountain’. This mistake connects all the Sahidic manuscripts. SaMV have the personal pronoun ⲛⲧⲟⲟⲩ ‘they’ before the verb ‘to say’ against Greek and SaA. Probably, SaA has omitted this intentionally since it fits the tendency towards shortening, a typical feature of this manuscript. SaV is the only manuscript that has the particle ⲇⲉ before ‘father’, SaMA have no connective. The Greek manuscripts attest καὶ ἰδού. Usually, this word pair is rendered with ⲉⲓⲥ or ⲉⲓⲥ ϩⲏⲏⲧⲉ in 1 Sam. Only twice is it rendered with ⲇⲉ: 11: 5 and 14: 26.60 It is difficult to determine whether SaV has

59 Liddell/Scott, Intermediate Lexicon, 794. 60 In verse 4: 13, there is probably a homoiarchon mistake in Sahidic, and thus, this verse does not provide a third example of rendering καὶ ἰδού with ⲇⲉ. See my subsection 2.1.3.7. f.

Textual analyses

159

preserved an existing rendering with ⲇⲉ or added it out of need for a change of focus in the story since there are also a few passages where καὶ ἰδού has no rendering. In the rendering of ἀποτετίνακται τὸ ῥῆμα τῶν ὄνων, SaMV read ⲁϥⲛⲉϩ ⲡⲣⲟⲟⲩϣ ⲉⲃⲟⲗ ⲛⲛⲉⲧⲃⲛ̄ ⲟⲟⲩⲉ ‘He has cast the care for the cattle’, with the preposition ⲛ, whereas SaA reads with the preposition ⲉⲧⲃⲉ. This does not affect the meaning, but shows that SaA has a tendency to vary the wording. All three Sahidic manuscripts read ⲉⲙⲁⲧⲉ/ⲙ̄ⲙⲁⲧⲉ ‘greatly, very’ without Greek attestation of σφόδρα.61 This Sahidic word is spelled in two ways without any difference in the meaning, and its presence here probably has its explanation in the meaning of the verbs; the Greek verb ἐδαψιλεύσατο has an emphatic meaning ‘to be anxious’, and the meaning of the Sahidic verb ‘to take care, thought’ is strengthened with the adverb. Results The Sahidic manuscripts studied here cover a long period of time. The earliest comes from the 6th century and the latest from the 10–11th century. One translation stands behind the stream of tradition that has produced these manuscripts, and they, therefore, share a common background.62 This is visible in their shared additions against all the other witnesses: 7: 12 an explanation before ⲧⲡⲁⲗⲁⲓⲁ, 9: 24 ⲉⲧⲉ ⲡⲙⲉⲣⲟⲥ ⲡⲉ ‘that is part’, the double reading ⲡⲉⲓⲥⲉⲉⲡⲉ· ⲏ ⲡⲉⲓϣⲁⲩ and the verb ⲛⲅⲟⲩⲱⲙ ‘eat’ after ⲡⲱϣ ‘divide’. In 9: 26, these Sahidic manuscripts render ἀνάστα with two verbs ‘rise and come’ and in the same verse ‘He came down’ without any Greek source text. In 10: 1, the Sahidic manuscripts translate ‘the people of Israel’, not ‘the people of the Lord’. Finally, in 10: 2, the Sahidic manuscripts share a mistranslation ‘mountain’ instead of ‘region’, and read ⲉⲙⲁⲧⲉ/ⲙ̄ⲙⲁⲧⲉ ‘greatly, very’ without a Greek equivalent. The feature that next calls for attention is the tendency of SaA towards shortening and simplifying the text. All of verse 8: 1 is omitted in SaA. In 7: 8, SaM reads ⲙⲡⲉⲣⲕⲁⲣⲱⲕ ⲉϫⲱⲛ ⲛⲅϭⲱ, but SaA abbreviates to ⲙ̄ⲡⲣϭ ̄ ⲱ. Further into the same verse, SaA omits ⲙⲙⲟⲓ. In 9: 21, SaA has reformulated a rhetorical question into an affirmation. In 7: 17, instead of ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙ̄ⲙⲁⲩ, SaA reads ⲛϩ̄ ⲏⲧϥ.̄ Omission of the conjunction is met in 7: 14, 7: 17 and 10: 1, and the adverb ⲉⲃⲟⲗ is left out in 7: 17, 9: 22 and 9: 26. Other small words omitted

61 Peters (Textual Character, 128–9) records omission and addition of σφόδρα/ⲉⲙⲁϣⲱ in his analysis of Bohairic Deuteronomy and its Greek source. According to him, even the place of this adverb is accurately preserved in his text. 62 In Schüssler’s listing, this manuscript has the siglum sa 25 (Schüssler, Die koptischen Bibeltexte 1: 2, 29–30). Schüssler (ibid., 30) states, concerning the manuscript SaM, “Es handelt sich um eine unabhängige Übersetzung aus dem Griechischen.” There are earlier fragments of Sahidic 1 Sam that clearly have the same Sahidic base text as this complete manuscript, and one has to ask what is the independence meant here.

160

Affiliations of the Sahidic manuscripts and the Greek Vorlage

are ⲛⲁⲕ in 9: 24, ⲛⲧⲟϥ in 9: 26 and ⲧⲉⲛⲟⲩ in 9: 27. In verse 7: 9, instead of ⲉϥϩⲁ, SaA reads ϩⲁ, and instead of ⲛⲟⲩϭⲗⲓⲗ it reads ⲛϭ ̄ ⲗⲓⲗ. In the same manuscript, SaA, there are several copying mistakes (7: 8, 7: 10–11, 7: 14, 9: 22). A special characteristic of SaA is its inclination towards adding ⲉ- before verbal forms (7: 10 ⲉⲛϥ̄ⲧⲁⲗⲟ instead of ⲛⲉϥⲧⲁⲗⲟ, 7: 17 ⲉⲛⲉϥϯϩⲁⲡ and ⲉⲁϥⲕⲱⲧ, 9: 21 ⲉⲕϣⲁϫⲉ, and 10: 2 ⲉⲕⲛⲁϭⲛ̄ against ⲕⲛⲁϭⲓⲛⲉ in SaMV), thus changing the actual form either to a circumstantial or to a focalized verbal form. When analyzing the variants encountered in these verses in different Sahidic manuscripts, the largest group consists of those passages where SaMV agree against SaA. The differences vary significantly with respect to small-scale issues as well as more remarkable details. One finds variation in number (7: 8, 7: 14, 9: 22, 9: 25, 10: 1), connective (7: 10, 10: 1), spelling (7: 12, 9: 27), definite article (9: 23), preposition (10: 2), verbal form (9: 26, 10: 2), single words (7: 10, 9: 26, 10: 2), word order (7: 10, 9: 27), contents (7: 12, 7: 13, 7: 17, 10: 1), and construction (9: 27). Usually, both alternatives in Sahidic render the Greek text accurately, and neither reading presents itself as the obvious original. The situation changes as soon as one notices that SaA, often alone, differs from the rest of the tradition in several passages. Repeatedly, these readings are best explained as later emendations. Obviously, SaA more often has reformulated the text, and thus more probably provides a secondary reading. When analyzing the passages where SaAV read against SaM, the outcome is that these mostly comprise small-scale differences: spelling (7: 11), preposition/adverb (7: 9, 7: 10, 9: 24, 10: 1), article (7: 10), mistake (7: 15–16, 9: 23). There are a couple of passages where more remarkable differences appear: word-order in 9: 21, in 9: 25 addition of an object, and in 9: 26 SaM reads ⲉϥⲙⲟⲟϣⲉ against ⲁϥⲙⲟⲟϣⲉ in SaAV. These passages show independent reformulations typical of Sahidic witnesses. In these cases, the wording of SaM differs from the other two manuscripts, but most of these changes do not affect the meaning. Probably, SaM is secondary in these passages. Passages where SaMA read against SaV do not make up a long list; adverb (7: 8), conjunction (7: 9, 10: 2), and construction (7: 9 twice). Taking into account the centuries that separate SaV from the other Sahidic manuscripts, it is difficult to say whether SaV is secondary in these wordings. Possibly, the change has happened during the transmission, after SaV was copied but before the parent manuscript(s) of SaAM. Passages where all Sahidic witnesses differ are few in number compared to the other kinds of cases above. These examples show that some minor variations arose during the transmission of the Sahidic manuscripts. These small changes represent inner-Sahidic variation, without connection to Greek variants and without changing the meaning substantially. On the whole, proper names and place names often show differences in spelling, and these are not listed here. In 7: 8, SaAV have an almost identical wording ⲛⲛⲉ ⲡⲁⲓ ϣⲱⲡⲉ ⲙⲙⲟⲓ. The only difference is an omission of ⲙⲙⲟⲓ in SaA. SaM

Textual analyses

161

reads ⲛⲛⲉⲥϣⲱⲡⲉ ⲙⲙⲟⲓ. In verse 7: 13, SaA has its own wording, probably presenting a corrupt version of the text found in other Sahidic manuscripts; SaM reads asyndetically, whereas SaV has the conjunction ϭⲉ. In 7: 14, SaM reads ⲛⲛϣⲏⲣⲉ ⲙⲡⲓⲏⲗ and SaV, together with Greek manuscripts, has ⲙⲡⲓⲏⲗ. In SaA, this phrase is omitted as a result of a homoiarchon-based periblepsis. In 9: 21, the last question is formulated in different ways in SaA and SaM, both of which accurately render the Greek question. In verses 9: 25–26, there are different prepositions in the phrase ‘on the roof’. In evaluating these variants, one has to take into account the graphical similarity of the words. Affiliations of the Sahidic manuscripts are detectable in some verses. All three Sahidic manuscripts contain the addition ‘And Samuel said, May it not befall me that I withdraw from the Lord, my God, not to cry to him for you’ in 7: 8, attested by CI d f s 29 55 71 158 245 318 554, with the sign ø in 56. In 7: 11, SaA has ‘they came into Massefat’ against SaMV which read ‘they came out from Massefat’, the same variant as 93 121 246 245. In verse 7: 14, ‘from Askalon to Geth’, ασκαλωνος is found in B b d-106 44–107–125– 610 370 554 and Γέθ in A O L d 314–488–489 318 (γαιθ) 554. Thus, the first place name occurs in its OG form whereas the second appears with an adapted form in all Sahidic manuscripts. In verse 7: 16, SaMV with the majority of Greek manuscripts have ⲅⲁⲗⲅⲁⲗⲁ against SaA ⲕⲁⲗⲁⲁⲧ and O (γαλααδ). In 9: 24, SaMA offer ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲛⲥⲁⲟⲩⲗ against SaV that reads the shorter ⲡⲉϫⲁϥ. The reading of SaV is in accordance with 125 489 as well as with the MT. In 9: 27, the Sahidic witnesses probably had in their Vorlage ὧδε σήμερον like 19–108 in Greek. In verse 10: 1, SaA begins Samuel’s speech with ⲉⲓⲥ. In Greek, ἰδού is attested by 245 707. In 10: 1, the verb referring to the people is singular in SaMV against plural in SaA and 127 707 in Greek. This is a typical translational feature when referring to the people, and does not support any Greek reading. Moreover, these two manuscripts do not form a group, and therefore the variation probably occured independently in these three witnesses. SaA is the latest of these Sahidic manuscripts, and it thus contains more secondary readings than the other two manuscripts. Additionally, this manuscript shortens and simplifies its text, and this complicates its use in textual criticism; one must consider the variants in the light of SaA’s typical secondary features before citing it. SaM and SaV, on the other hand, are comparably more trustworthy witnesses of the Sahidic tradition. Some secondary features lie in SaM; it occasionally adds small words like prepositions, and contains a few scribal errors. If there is an addition of one or two words in the Greek variants, one should check whether this kind of addition occurs on a regular basis thoughout SaM before citing SaM in the apparatus.

162

Affiliations of the Sahidic manuscripts and the Greek Vorlage

3.2.2. SaMI 12: 4–5,10–11 12: 4 καὶ εἶπαν πρὸς Σαμουήλ Οὐκ ἠδίκησας ἡμᾶς καὶ οὐ κατεδυνάστευσας καὶ οὐκ ἔθλασας ἡμᾶς, καὶ οὐκ εἴληφας ἐκ χειρὸς οὐδενὸς οὐδέν. SaM

SaI

ⲡⲉϫⲁⲩ ⲧⲏⲣⲟⲩ ⲉϩⲟⲩⲛ ϩⲛ̄ ⲥⲁⲙⲟⲩⲏⲗ· ϫⲉ ⲙⲡⲉⲕϫⲓⲧⲛ̄ ⲛϭⲟⲛⲥ̄ ⲛⲗⲁⲁⲩ ⲁⲩⲱ ⲙⲡⲉⲕⲙⲟⲕϩⲉⲛ· ⲛⲗⲁⲁⲩ· ⲁⲩⲱ ⲙⲡⲉⲕⲑⲙ̄ⲕⲉ ⲗⲁⲁⲩ· ⲁⲩⲱ ⲙⲡⲉⲕϫⲓ ⲗⲁⲁⲩ· ⲛⲧⲉⲛ ⲗⲁⲁⲩ ⲛⲣⲱⲙⲉ·

ⲁⲩ]ⲱ [ ⲙⲡ]ⲕ[ⲙⲟ]ⲕϩⲛ̄ ϩⲛⲗ̄ ⲁⲁⲩ ⲁⲩⲱ ⲙⲡⲕⲑⲙ̄ⲕⲉ ⲗⲁⲁⲩ· ⲁⲩⲱ ⲙⲡⲕϫⲓ ⲗⲁⲁⲩ ⲛⲧ ̄ ⲛ̄ ⲗⲁⲁⲩ ⲛⲣⲱⲙⲉ·

SaMI read with four negations ‘you have not’, like most Greek manuscripts, against the MT and Lucianic text where only three negations are present. SaM adds ⲧⲏⲣⲟⲩ ‘all’ in the first clause, like 44–125 in Greek. A small difference arises in the third clause where SaM has ⲙⲡⲉⲕⲙⲟⲕϩⲉⲛ ⲛⲗⲁⲁⲩ ‘you have not afflicted us in anything’ and SaI reads ⲙⲡⲉⲕⲙⲟⲕϩⲛ̄ ϩⲛⲗ̄ ⲁⲁⲩ, without difference in the meaning. Both expressions are suitable renderings for οὐ κατεδυνάστευσας in Greek, but with some additions. Upon close examination, there is no Greek equivalent for the 1st person plural suffix or for ⲛⲗⲁⲁⲩ/ϩⲛⲗ̄ ⲁⲁⲩ ‘in anything’. This kind of variation (ⲛ/ϩⲛ)̄ readily happens in copying. Since the change could occur in either direction, it is impossible to give priority to one of the readings. Both Sahidic witnesses have a common ancestor, which is perceivable in their common attestation of the small details mentioned above against the Greek. 12: 5 καὶ εἶπεν Σαμουὴλ πρὸς τὸν λαόν Μάρτυς Κύριος ἐν ὑμῖν καὶ μάρτυς χριστὸς αὐτοῦ ἐν τῇ ἡμέρᾳ ταύτῃ, ὅτι οὐχ εὑρήκατε ἐν χειρί μου οὐθέν· καὶ εἶπαν Μάρτυς. ἐν τῇ ἡμέρᾳ ταύτῃ L 29 707] pr σημερον A B M V O C -379 a 121 d f s 55 71 158 244 245 318 460 554; > 379.63

SaM

SaI

ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲉϩⲟⲩⲛ ϩⲙ̄ ⲡⲗⲁⲟⲥ ϫⲉ ⲡϫⲟⲉⲓⲥ ϥⲟ ⲛⲙⲛⲧ ̄ ⲣⲉ ϩⲛ ⲧⲉⲧⲛⲙⲏⲧⲉ ⲙⲡⲟⲟⲩ· ⲁⲩⲱ ⲡⲉϥⲭⲣⲏⲥⲧⲟⲥ ⲟ ⲙⲙⲛⲧ ̄ ⲣⲉ· ϫⲉ ⲙⲡⲟⲟⲩ ⲛϩⲟⲟⲩ ⲙⲡⲉⲧⲛϩ̄ ⲉ ⲉϩⲱⲃ ⲛϫⲓⲛϭⲟⲛⲥ̄ ϩⲛ̄ ⲛⲁϭⲓϫ· ⲡⲉϫⲁⲩ ⲧⲏⲣⲟⲩ· ϫⲉ ϥⲟ ⲙⲙⲛⲧⲣⲉ̄·

ⲡⲉϫⲉ ⲥⲁⲙⲟⲩⲏⲗ ⲉϩⲟⲩⲛ ϩⲙ̄ ⲡⲗⲁⲟⲥ ϫⲉ ⲡϫⲟⲉⲓⲥ ϥⲟ ⲙⲙⲛⲧⲣⲉ ϩⲛ ⲧⲉ̣ⲧⲛⲙ ̄ ⲏⲧⲉ ⲙⲡ̣ⲟⲟⲩ ⲁⲩⲱ ⲡ[ϥ]ⲭⲣⲓⲥⲧⲟⲥ ⲟ ⲙ[ⲙⲛ]ⲧⲣⲉ ϫⲉ ⲙ̣ [ⲡⲟⲟⲩ] ⲛϩⲟ[ⲟⲩ]…

63 There are differences concerning the words after σημερον, but in all these manuscripts there is first σημερον.

Textual analyses

163

The Greek texts differ with respect to the expression(s) ‘today’ and its place. Both Sahidic witnesses have the same word order, which does not formally follow any Greek manuscript. This demonstrates that one common ancestor stood behind the Sahidic manuscripts. The Sahidic has ‘today’ twice, first in the shorter form ⲙⲡⲟⲟⲩ and second placed before the rendering of καὶ μάρτυς χριστὸς αὐτοῦ. The second time the longer form ⲙⲡⲟⲟⲩ ⲛϩⲟⲟⲩ is used, and it appears after the equivalent of ὅτι. In 1 Sam, the translator has translated these Greek expressions logically; σήμερον is rendered ⲙⲡⲟⲟⲩ and ἐν ταύτῃ τῇ ἡμέρᾳ/ ἐν τῇ ἡμέρᾳ ταύτῃ is rendered ⲙⲡⲟⲟⲩ ⲛϩⲟⲟⲩ. Thus, perhaps, the word σήμερον was added not only to many Greek manuscripts, but also to the Sahidic version, only placed differently. This addition must have happened early, since it is widespread in Greek and also found in Sahidic. 12: 10 καὶ ἐβόησαν πρὸς Κύριον καὶ ἔλεγον Ἡμάρτομεν, ὅτι ἐγκατελίπομεν τὸν κύριον καὶ ἐδουλεύσαμεν τοῖς Βααλεὶμ καὶ τοῖς ἄλσεσιν· καὶ νῦν ἐξελοῦ ἡμᾶς ἐκ χειρὸς ἐχθρῶν ἡμῶν, καὶ δουλεύσομέν σοι.

SaM

SaI

ⲁⲩⲱ ⲁⲩⲱ (sic) ⲁⲩⲱϣ ⲉⲃⲟⲗ ⲉϩⲣⲁⲓ ⲉⲡϫⲟⲉⲓⲥ· ⲉⲩϫⲱ ⲙⲙⲟⲥ· ϫⲉ ⲁⲛⲉⲣⲛⲟⲃⲉ· ϫⲉ ⲁⲛⲕⲱ ⲛ̄ⲥⲱⲛ ⲙⲡϫⲟⲉⲓⲥ· ⲁⲩⲱ ⲁⲛⲣ̄ϩⲙ̄ϩⲁⲗ ⲛⲛⲉⲃⲁϩⲁⲗⲓⲙ· ⲙⲛ̄ ⲛⲉⲙⲁ ⲛϣⲏⲛ· ⲧⲉⲛⲟⲩ ϭⲉ ⲛⲟⲩϩⲙ̄ ⲙⲙⲟⲛ ⲉⲧⲟⲟⲧⲟⲩ ⲛ̄ⲛⲉⲛϫⲁϫⲉ· ⲁⲩⲱ ⲧⲉⲛⲛⲁⲣ̄ϩⲙ̄ϩⲁⲗ ⲛⲁⲕ·

…]ⲁⲛ̣[ⲉⲣⲛⲟⲃⲉ] ϫⲉ ⲁⲛⲕⲱ ⲛⲥⲱⲛ ⲙⲡϫⲟⲉⲓⲥ ⲁⲩⲱ ⲁⲛⲣϩⲙ̄ϩⲁⲗ ⲛⲃ̄ⲃⲁϩⲁⲗⲓⲙ ⲙⲛ̄ ⲙ̄̄ⲙⲁ ⲛϣⲏⲛ· ⲧⲉⲛⲟⲩ ϭⲉ ⲛⲟⲩϩⲙ̄ ⲙ̄ⲙⲟⲛ ⲛ̄ⲧⲟⲟⲧⲟⲩ ⲛ̄ⲛⲉⲛϫⲁϫⲉ ⲁⲩⲱ ⲧⲛ̄ⲛ̣ⲁⲣ̄ϩⲙ̄ϩⲁⲗ ⲛⲁⲕ·

In this verse, the Sahidic witnesses share an essentially identical text. The word ⲛⲛⲉⲃⲁϩⲁⲗⲓⲙ ‘Baals’ (a transcription of the Hebrew word) is spelled ⲛⲃⲃ̄ ⲁϩⲁⲗⲓⲙ in SaI. Layton mentions this type of assimilation as a feature of some early manuscripts.64 On the whole, SaI differs from SaM when it comes to consonantal assimilation. SaI assimilates ⲛ to the first letter of the word more often than SaM. Examples of this are ⲛⲙⲛⲧ ̄ ⲣⲉ/ⲙⲙⲛⲧⲣⲉ, ⲛⲉⲙⲁ/ⲙ̄ⲙⲁ. In the second to last clause, SaM has ‘rescue us out of the hand of our enemies’, with ⲉⲧⲟⲟⲧⲟⲩ for ‘hand’, whereas SaI reads similarly with ⲛⲧ ̄ ⲟⲟⲧⲟⲩ for ‘hand’. According to Crum, the verb ⲛⲟⲩϩⲙ̄ ‘to rescue, save’ is used with different prepositions without difference in the meaning.65 SaM uses this same verb in the next verse (where SaI is not preserved) with the preposition I ⲛⲧ ̄ ⲟⲟⲧⲟⲩ like Sa in this verse 12: 10.

64 Layton, Grammar, §21b. 65 Crum, Dictionary, 244a.

164

Affiliations of the Sahidic manuscripts and the Greek Vorlage

12: 11 καὶ ἀπέστειλεν Κύριος τὸν Ἰεροβαὰλ καὶ τὸν Βαρὰκ καὶ τὸν Ἰεφθάε καὶ τὸν Σαμουὴλ καὶ ἐξείλατο ὑμᾶς ἐκ χειρὸς ἐχθρῶν ὑμῶν τῶν κυκλόθεν· καὶ κατῳκεῖτε πεποιθότες. SaM

SaI

ⲁ ⲡϫⲟⲉⲓⲥ ⲇⲉ ϫⲟⲟⲩ ⲛϩⲓⲉⲣⲟⲃⲁⲁⲗ· ⲙⲛ̄ ⲃⲁⲣⲁⲕ· ⲙⲛ ⲓⲉⲫⲑⲁⲉ· ⲙⲛ ⲥⲁⲙⲟⲩⲏⲗ ⲁⲩⲱ ⲁϥⲛⲉϩⲙ̄ⲧⲏⲩⲧⲛ̄ ⲛⲧⲟⲟ̄ⲧⲟⲩ ⲛ̄ⲛⲉⲧⲛ̄ϫⲁϫⲉ̄ ⲉⲧⲕⲱⲧⲉ ⲉⲣⲱⲧⲛ·̄ ⲁⲩⲱ ⲛⲉⲧⲉⲧⲛϣ ̄ ⲟⲟⲡ ⲡⲉ· ⲉⲧⲉⲧⲛⲥ̄ ⲙⲟⲛⲧ ̄ ·̄

[ⲁⲡ]ϫⲟⲉⲓⲥ ⲇⲉ ϫⲟ[ⲟⲩ] ⲛ̄ϩⲓⲉⲣⲟⲃⲟ[ⲟⲗ ⲙ]ⲛⲃⲁⲣⲁⲕ [ⲙⲛ ⲓⲉ]ⲫⲑⲁⲉ ⲙ̄ⲛ[…

In this verse, only the beginning of SaI is extant, and no differences are present. Results The Sahidic manuscripts possess one and the same translation. This is discernible in the lists of malpractices that Samuel has not done, verse 12: 4, as well as in the renderings of ‘today’ and its placement in verse 12: 5. There are no identical Greek readings, and Sahidic witnesses are completely independent. This feature, together with almost identical wordings in Sahidic witnesses in these verses compellingly suggest that these texts stem from a common ancestor. SaI is one of the earliest fragments of Sahidic 1 Sam, dated to the 4th–5th centuries. Thus it antedates SaM by some 400–500 years. This shows that there is a long manuscript tradition behind SaM and that this text has been copied with remarkable faithfulness. In 12: 4, SaMI preserve four negations ‘you have not’ like most Greek manuscripts, against the MT and Lucianic text where only three negations appear. Within the same verse, SaM adds ⲧⲏⲣⲟⲩ in the first clause like 44–125 in Greek. In verse 12: 10, SaI contains an assimilation peculiar to early Coptic manuscripts. This confirms the early dating of this fragment. 3.2.3. SaMAU 14: 24–32 SaU is probably the earliest fragment of Sahidic 1 Sam. Its editor, Butts, dates it to the 4th–5th century based upon its single-column format, Lycopolitan orthography, and its resemblance of early hands in Bodmer papyri.66 SaU is a papyrus leaf and contains verses 14: 24–50. On the verso side, approximately 13 lines are illegible with only some isolated letters visible. For my comments concerning the reconstructions, see my collation 66 Butts, “P.Duk.inv. 797”, 9–10.

165

Textual analyses

attached to this work.67 In verses 14: 24–32, three manuscripts are extant in Sahidic: SaMAU. 14: 24 καὶ Σαοὺλ ἠγνόησεν ἄγνοιαν μεγάλην ἐν τῇ ἡμέρᾳ ἐκείνῃ, καὶ ἀρᾶται τῷ λαῷ λέγων Ἐπικατάρατος ὁ ἄνθρωπος ὃς φάγεται ἄρτον ἕως ἑσπέρας καὶ ἐκδικήσω τὸν ἐχθρόν μου· καὶ οὐκ ἐγεύσατο πᾶς ὁ λαὸς ἄρτου, SaM

SaA

SaU

ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥϣⲱⲡⲉ ϩⲛ̄ ⲟⲩⲛⲟϭ ⲙⲙⲛ̄ⲧⲁⲧⲥⲟⲟⲩⲛ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲣⲁ̄ ⲛⲁϣ ⲙⲡⲗⲁⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ· ϫⲉ ϥⲥϩⲟⲩⲟⲣⲧ̄ ⲛϭ ̄ ⲓ ⲡⲉⲧⲛⲁⲟⲩⲉⲙ ⲟ̄ⲉⲓⲕ ϣⲁ ⲡⲛⲁⲩ ⲛⲣⲟⲩϩⲉ· ϫⲉ ⲉⲓⲉϫⲓⲕⲃⲁ ⲙ̄ⲡⲁϫⲁϫⲉ· ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲇⲉ ⲧⲏⲣϥ̄ ⲙⲡⲉϥⲧⲉⲡ ⲟⲉⲓⲕ·

ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥϣⲱⲡⲉ ϩⲛ ⲟⲩⲛⲟϭ ⲙⲙ̄ⲛⲧ ̄ ⲁ ̄ ⲧⲥⲟⲟⲩⲛ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙⲙⲁⲩ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲣⲁⲛⲁϣ ⲙⲡⲗⲁⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ· ϫⲉ ϥⲥϩⲟⲩⲟⲣⲉⲧ ⲛϭⲓ ⲡⲣⲱⲙⲉ ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟⲉⲓⲕ ϣⲁ ⲡⲛⲁⲩ ⲛⲣⲟⲩϩⲉ· ϫⲉ ⲉⲓϫⲓ ⲕⲃⲁ ⲛⲛⲁϫⲁϫⲉ· ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ ⲙⲡϥⲧⲉⲡ ⲟⲉⲓⲕ·

ϥ̣ϫ̣ⲱ[ⲙⲙⲟⲥ ϫⲉ ϥⲥϩⲟⲩⲟⲣⲧ̄ ⲛϭ ̄ ⲓ .. ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟ]ⲉⲓⲕ ϣⲁ ⲡ̣[ⲛⲁⲩ ⲛⲣⲟⲩϩⲉ ϫⲉ … ]ϫ̣̄ ⲁ̣ϫⲉ· ⲁⲩⲱ ⲡⲗⲁ̣ⲟ[̣ ⲥ…]68 [ⲧⲏⲣϥ̄ ⲙⲡⲉϥⲧⲉⲡ ⲟⲉⲓⲕ

Both Sahidic manuscripts read the second clause with Saul like CI in Greek and the MT. The Hexaplaric origin of this reading is established by the asterisk present in Greek 731. However, this kind of addition is typical in the Sahidic version; proper nouns are often added without any Greek source69, and therefore it is not necessary to suppose a connection to the Hexaplaric reading in the Sahidic version. The first difference between the Sahidic manuscripts is in the relative construction: SaM reads, against SaA, without ⲣⲱⲙⲉ ‘man’, and with the definite article. No Greek variants omit ἄνθρωπος. Either SaM preserves a less literal reading and SaA has been corrected according to the Greek, or two distinct Sahidic versions have independently formulated their wordings. Even without the word ἄνθρωπος, SaM could have rendered the Greek, and there is no need to suppose any corruption in SaM. Because SaU has not preserved the first part of this verse, it does not support either of the readings.70 The verbal forms are different in the next clause. SaM uses the optative ⲉⲓⲉϫⲓⲕⲃⲁ ‘I shall avenge’ after ϫⲉ while SaA has the circumstantial/focalizing form of the same verb, ⲉⲓϫⲓ ⲕⲃⲁ. The verb in SaM is an accurate rendering of

67 Appendix 6.3. 68 There could be ⲇⲉ but I have left it out to avoid seeming connections between manuscripts. 69 See section 2.2.3.3. 70 With the word ⲣⲱⲙⲉ, there would be 33–34 letters (depending whether there was ⲟⲩⲉⲙ or ⲟⲩⲙ̄) and without the word 29–30. According to Butts, the original width of the line was ca. 35–37 characters. However, his reconstruction varies between 27 and 36 characters per line.

166

Affiliations of the Sahidic manuscripts and the Greek Vorlage

the Greek future,71 whereas the verb in SaA is not very suitable, be it circumstantial or focalizing. The object of the verb is singular in SaM but plural in SaA. However, there is no need to suppose that SaA follows Greek 318 460c in reading the plural, since the difference between singular and plural is so slight in Coptic. There are also other passages where plural and singular vary without connection to Greek. Neither of these readings is visible in SaU. In the last sentence, SaM reads with a double connective, which seems to be an error. According to Butts, SaU here attests to the reading of SaM, but his judgement is apparently based upon an uncertain manuscript reading.72 14: 25 καὶ πᾶσα ἡ γῆ ἠρίστα. καὶ ἰδού δρυμὸς ἦν μελισσῶνος κατὰ πρόσωπον τοῦ ἀγροῦ. ἰδού Eth] ιααλ B b; ιαρ A; ιααρ Ra; > O L As

SaM

SaA

SaU

ⲡⲕⲁϩ ⲇⲉ ⲧⲏⲣϥ̄ ⲛⲉϥⲁⲣⲓⲥⲧⲁ· ⲛⲉⲣⲉ ⲟⲩⲙⲁ ⲛϣⲏⲛ ⲇⲉ ⲛϯϩⲙⲉϥ ⲛⲉⲃⲓⲱ ϣⲟⲟⲡ ⲙ̄ⲡⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲛⲧⲥⲱϣⲉ·

ⲡⲕⲁϩ ⲇⲉ ⲛⲉⲁϥⲁⲣⲓⲥⲧⲁ· ⲛⲉⲟⲩⲛ ⲟⲩⲙⲁⲛϣⲏⲛ ⲛϯϩⲙⲉϥ ⲛⲉⲃⲓⲱ ϣⲟⲟⲡ ⲙⲡⲙⲧⲟ ⲉⲃⲟⲗ ⲛⲧⲥⲱϣⲉ·

ⲡⲕ]ⲁϩ ⲇⲉ ⲧⲏⲣϥ̄ ⲛⲉϥⲁ̣[ⲣⲓⲥⲧⲁ ⲛⲉ… ⲟⲩⲙⲁ ⲛϣⲏ]ⲛ ⲇⲉ ⲛϯϩⲙⲉϥ ⲛⲉ̄ ⲃ̣ⲓⲱ ϣ̣ⲟ̣ⲟ̣ⲡ̣ [ⲙ̄ⲡⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲛⲧⲥ]ⲱϣⲉ·

In the first clause, SaA lacks ⲧⲏⲣϥ̄ ‘all’ against SaMU and the Greek witnesses. For the verbal form, SaA has a preterit of the past tense ⲛⲉⲁϥⲁⲣⲓⲥⲧⲁ ‘had dined’, against SaMU that have the preterit ⲛⲉϥⲁⲣⲓⲥⲧⲁ ‘was dining’. In the second clause, SaA offers as a verbal form ⲛⲉⲟⲩⲛ ⲟⲩⲙⲁ … ϣⲟⲟⲡ ‘there was’, while SaM reads ⲛⲉⲣⲉ ⲟⲩⲙⲁ … ϣⲟⲟⲡ which does not change the meaning.73 One cannot prioritize one over the other concerning the originality of the reading. With regard to the connective in the second clause, SaMU contain ⲇⲉ, but SaA has asyndeton. In the Sahidic witnesses, there is no equivalent for the word ἰδού. This same word is omitted in L O. However, Greek variants for this word exist, and a transcription of Hebrew was perhaps present in the Vorlage of the Sahidic translator. Probably, the Sahidic translator either omitted ἰδού, or his source text was similar to L O that lack this word.

71 There is one similar rendering for the future in verse 27: 5. The most common rendering for the Greek future is a Sahidic future or conjunctive, depending on the context. 72 Butts (“P.Duk.inv. 797”, 13, 15) gives the letters with dots, but adds in the notes that “only the faintest traces of ⲇⲉ remain”. I am not able to see any remnants of ⲇⲉ, and therefore, the word is in square brackets. 73 SaU has a lacuna, here, and it is not possible to reconstruct the lost text since both options need approximately the same space.

167

Textual analyses

14: 26 καὶ εἰσῆλθεν ὁ λαὸς εἰς τὸν μελισσῶνα, καὶ ἰδοὺ ἐπορεύετο λαλῶν· καὶ ἰδοὺ οὐκ ἦν ἐπιστρέφων τὴν χεῖρα αὐτοῦ εἰς τὸ στόμα αὐτοῦ, ὅτι ἐφοβήθη ὁ λαὸς τὸν ὅρκον Κυρίου. ἰδού 2º > A O L d-74 554 (=MT)

SaM

SaA

SaU

ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲁϥⲧⲁⲁϥ ⲉϩⲣⲁⲓ· ⲉⲡⲙⲁ ⲛϯ ̄ ϩⲙⲉϥ ⲛⲉⲃⲓⲱ· ⲛⲉⲣⲉ ⲡⲗⲁⲟⲥ ⲇⲉ ⲙⲟⲟϣⲉ ⲉϥϣⲁϫⲉ· ⲉⲙⲛ ⲗⲁⲁⲩ ⲕⲧⲟ ⲛⲧ ̄ ⲉϥϭⲓϫ ⲉ̄ϩⲟⲩⲛ ⲉⲧⲉϥⲧⲁⲡⲣⲟ ϫⲉ ⲡⲗⲁⲟⲥ ⲁϥⲣϩ̄ ⲟⲧⲉ ⲛϩⲏⲧϥ ⲙⲡⲁⲛⲁϣ· ⲙⲡϫⲟⲉⲓⲥ·

ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲁϥⲧⲁⲁϥ ⲉϩⲟⲩⲛ ⲉⲡⲙⲁ ⲛϯϩⲙⲉϥ ⲉⲛⲉⲃⲓⲱ· ⲛⲉⲣⲉ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ ⲙⲟⲟϣⲉ ⲉϥϣⲁϫⲉ ⲉⲙⲛ ⲗⲁⲁⲩ ⲕⲧⲟ ⲛⲧϥϭⲓϫ ⲉϩⲟⲩⲛ ⲉⲧϥⲧⲁⲡⲣⲟ· ϫⲉ ⲡⲗⲁⲟⲥ ⲁϥⲣϩⲟⲧⲉ ϩⲏⲧϥ ⲙⲡⲁⲛⲁϣ ⲙⲡϫⲟⲉⲓⲥ·

ⲁⲩⲱ [ⲡ]ⲗ̣ⲁⲟⲥ ⲁϥⲧ̣ [ⲁⲁϥ ⲉϩ…74 ⲉⲡⲙⲁ ⲛϯ ̄ ϩⲙⲉϥ ⲛ]ⲉⲃⲓⲱ· ⲛⲉⲣⲉ ⲡⲗⲁⲟⲥ ⲇⲉ [ⲙⲟⲟϣⲉ75 ⲉϥϣⲁϫⲉ ⲉⲙⲛ ⲗⲁⲁ]ⲩ ⲕⲧⲟ ⲛⲧ ̄ ⲉϥϭⲓϫ ⲉϩ[ⲟ]ⲩ̣ⲛ̣ [ⲉⲧ]ⲉ̣ϥ̣ⲧ̣[ⲁⲡⲣⲟ76 ϫⲉ ⲡⲗⲁⲟⲥ ⲁϥ]ⲣϩ̣ ̄ ⲟⲧⲉ ϩⲏⲧϥ̄ ⲙ̄ⲡⲁⲛⲁϣ ⲙ̣ ⲡϫⲟⲉⲓⲥ̣

In the first clause, SaM and SaA preserve distinct prepositional phrases, although both are fair translations of the Greek. SaM and SaA both have the word ⲡⲙⲁ ‘place’ which has no equivalent in Greek, but is needed to formulate the expression in Coptic. In SaU, the preposition has been reconstructed, and one cannot determine which one was there since both have five letters. In the second clause, SaMU offer the particle ⲇⲉ, while SaA reads ⲧⲏⲣϥ ‘all’. SaA is independent; there are no corresponding Greek variants to support its reading. All three Sahidic manuscripts formulate this second clause with the subject ⲡⲗⲁⲟⲥ against unanimous Greek evidence, and this implies that one common ancestor lies behind these Sahidic manuscripts. The Sahidic has not rendered ἰδού which is attested once or twice in Greek manuscripts. The first appearance has an equivalent in the MT, but the second does not. Accordingly, an omission of Hexaplaric origin of the second ἰδού is found in A O L d-74 554. In the last clause, SaAU read ϩⲏⲧϥ ‘[be afraid] of’ while SaM has ⲛϩⲏⲧϥ, but this does not alter the meaning.

74 This is in the lacuna, but the spacing suggests that there was a five-letter word, as is the case in both Sahidic manuscripts. 75 Butts marks this word with dots, but I am not able to decipher any letters here. 76 Butts (“P.Duk.inv. 797”, 13, 16) makes a notion of uncertainty, but transcribes these words as ⲉ[ⲧⲉ]ϥⲧⲁⲡⲣⲟ, obviously overlooking the ϩ that is clearly visible before the lacuna.

168

Affiliations of the Sahidic manuscripts and the Greek Vorlage

14: 27 καὶ Ἰωναθὰν οὐκ ἀκηκόει ἐν τῷ ὁρκίζειν τὸν πατέρα αὐτοῦ τὸν λαόν· καὶ ἐξέτεινεν τὸ ἄκρον τοῦ σκήπτρου αὐτοῦ τοῦ ἐν τῇ χειρὶ αὐτοῦ καὶ ἔβαψεν αὐτὸ εἰς τὸ κηρίον τοῦ μέλιτος, καὶ ἐπέστρεψεν τὴν χεῖρα αὐτοῦ εἰς τὸ στόμα αὐτοῦ, καὶ ἀνέβλεψαν οἱ ὀφθαλμοὶ αὐτοῦ. SaM

SaA

SaU

ⲓⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲙⲡⲉϥⲥⲱⲧⲙ̄ ϫⲉ ⲁ ⲡⲉϥⲉⲓⲱⲧ ⲧⲁⲣⲕⲉ ⲡⲗⲁⲟⲥ· ⲁⲩⲱ ⲓⲱⲛⲁⲑⲁⲛ ⲁϥⲥⲟⲩⲧⲛ̄ ϩⲧⲏϥ ⲉⲃⲟⲗ ⲙⲡⲉϥϭⲉⲣⲱⲃ ⲉⲧϩⲛ̄ ⲧⲉϥϭⲓϫ· ⲁϥⲥⲟⲡϥ̄ ϩⲙ ⲡⲙⲟⲩⲗϩ ⲛⲉⲃⲓⲱ̄ · ⲁϥⲕⲧⲟ ⲛⲧ ̄ ⲉϥϭⲓϫ ⲉϩⲟⲩⲛ ⲉⲣⲱϥ· ⲁ ⲛⲉϥⲃⲁⲗ ⲛⲁⲩ ⲁⲩϫⲱⲧⲉ·

ⲓⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲙⲡϥⲉⲓⲙⲉ ϫⲉ ⲁ ⲡϥⲓⲱⲧ ⲧⲁⲣⲕⲉ ⲡⲗⲁⲟⲥ· ⲁⲩⲱ ⲓⲱⲛⲁⲑⲁⲛ ⲁϥⲥⲟⲟⲩⲧⲛ ⲛϩⲧⲏϥ ⲙⲡϥϭⲉⲣⲱϥ ⲉⲧ ϩⲛ ⲧϥϭⲓϫ· ⲁϥⲥⲟⲡϥ ϩⲙ ⲡⲙⲟⲩⲗϩ ϩⲙ ⲡⲉⲃⲓⲱ· ⲁϥⲕⲧⲉ ⲧϥϭⲓϫ ⲉϩⲟⲩⲛ ⲉⲧϥⲧⲁⲡⲣⲟ ⲁ ⲛϥⲃⲁⲗ ⲛⲁⲩ ⲁⲩϫⲱⲧⲉ·

[ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲙⲡⲉ]ϥⲥⲱⲧ̅ⲙ̅ ϫⲉ ⲁⲡⲉϥⲉⲓⲱⲧ ⲧ̣ ⲁⲣⲕⲉ [ⲡⲗⲁⲟⲥ ⲁⲩⲱ ⲓ̈ⲱⲛ]ⲁ̣ⲑⲁⲛ ⲁϥⲥⲟⲩⲧⲛ̄ ϩⲧⲏϥ ⲉⲃⲟⲗ ⲙ̄[ⲡⲉϥϭⲉⲣⲱⲃ ⲉⲧϩⲛ]̄ ⲧ̣ ⲉϥϭⲓϫ ⲁϥⲥⲟⲡϥ̄ ϩⲙ̄ ⲡⲙⲟⲩⲗ̣ϩ ̄ [ⲛⲉⲃⲓⲱ ⲁϥⲕⲧⲟ ⲛⲧ ̄ ⲉ]ϥ̣ϭⲓϫ ⲉϩⲟⲩⲛ ⲉⲣⲱϥ ⲁⲛⲉϥⲃⲁ̣ⲗ ⲛ̣ [ⲁⲩ ⲁⲩϫⲱⲧⲉ]77

In the first clause, SaMU have the verb ⲥⲱⲧⲙ̄ ‘to hear’ whereas SaA offers ⲉⲓⲙⲉ ‘to understand, to know’ without corresponding Greek variants. SaMU offer ⲁϥⲥⲟⲩⲧⲛ̄ ϩⲧⲏϥ ⲉⲃⲟⲗ ⲙⲡⲉϥϭⲉⲣⲱⲃ ‘He stretched forth the tip of his staff’ with a prenominal verb and an object, against SaA that has ⲁϥⲥⲟⲟⲩⲧⲛ ⲛϩⲧⲏϥ ⲙⲡϥϭⲉⲣⲱϥ, an absolute verb with the preposition ⲛ̄ before the object. This does not affect the meaning, but shows that SaA has its own wording against SaMU. The same is true of the adverb ⲉⲃⲟⲗ in the same expression, present in SaMU but not in SaA. In the same clause, all Sahidic witnesses read ‘Jonathan’ against all the Greek witnesses. SaM formulates ⲁϥⲥⲟⲡϥ̄ ϩⲙ ⲡⲙⲟⲩⲗϩ ⲛⲉⲃⲓⲱ̄ ‘He dipped it into the honeycomb’ against ⲁϥⲥⲟⲡϥ ϩⲙ ⲡⲙⲟⲩⲗϩ ϩⲙ ⲡⲉⲃⲓⲱ ‘He dipped it into the honeycomb in the honey’ in SaA. Butts argues that the spacing of SaU suggests the reading of SaM, and it seems probable.78 In the penultimate clause, SaM and SaA again have different constructions marking the object; this time SaA offers the prenominal state of the verb, and SaM uses the preposition ⲛ̄ to mark the object. SaU has a lacuna. In the same clause, the word for ‘mouth’ is ⲣⲟ in SaMU, but ⲧⲁⲡⲣⲟ in SaA. According to Feder, in the manuscripts of Jeremiah the word ⲣⲟ was systematically redacted to ⲧⲁⲡⲣⲟ. Feder theorizes that an old-fashioned word had been replaced by a more contemporary word.79 In 1 Sam however, this change is not systematic.80 77 I agree with Butts (“P.Duk.inv. 797”, 16) that the reconstruction given here makes the line considerably too long to be correct. Without any Coptic variants, however, it would be pure guesswork to propose another reconstruction. 78 Butts, “P.Duk.inv. 797”, 16. In the line above, the reconstruction seems secure; the characters match so that there are ca. 12 characters before ϭⲓϫ on this line. With the reconstruction according to SaM, there are 14 characters, but according to SaA there would be 16 characters. 79 Feder, Biblia Sahidica, 59. 80 See verses 17: 34 and 17: 35 below.

169

Textual analyses

14: 28 καὶ ἀπεκρίθη εἷς ἐκ τοῦ λαοῦ καὶ εἶπεν Ὁρκίσας ὥρκισεν ὁ πατήρ σου τὸν λαὸν λέγων Ἐπικατάρατος ὁ ἄνθρωπος ὃς φάγεται ἄρτον σήμερον· καὶ ἐξελύθη ὁ λαός. SaM

SaA

SaU

ⲟⲩⲁ ⲇⲉ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲗⲁⲟⲥ ⲛⲉϥϫⲱ ⲙⲙⲟⲥ̄ ⲛⲁϥ· ϫⲉ ϩⲛ̄ ⲟⲩⲁⲛⲁϣ· ⲡⲉⲕⲉⲓⲱⲧ· ⲁϥⲧⲁⲣⲕⲉ ⲡⲗⲁⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ· ϫⲉ ϥⲥ̄ϩⲟⲩⲟⲣⲧ ⲛϭⲓ ⲟⲩⲣⲱⲙⲉ ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟⲉⲓⲕ ⲙⲡⲟⲟⲩ· ⲡⲗⲁⲟⲥ ⲇⲉ ⲁϥⲥⲱϣⲙ̄

ⲟⲩⲁ ⲇⲉ ⲉⲃⲟⲗ ϩⲙ ⲡⲗⲁⲟⲥ ⲁϥⲟⲩⲱϣϥ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ ϫⲉ ϩⲛ ⲟⲩⲁⲛⲁϣ ⲁ ⲡⲕⲓⲱⲧ ⲧⲁⲣⲕⲉ ⲡⲗⲁⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ ϥⲥϩⲟⲩⲟⲣⲧ ⲛϭⲓ ⲡⲣⲱⲙⲉ ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟⲉⲓⲕ ⲙⲡⲟⲟⲩ· ⲡⲗⲁⲟⲥ ⲇⲉ ⲁϥⲥⲱϣⲙ̄·

[ⲟⲩⲁ ⲇⲉ ⲉⲃⲟ]ⲗ ϩⲙ̄ ⲡⲗⲁⲟⲥ ⲁϥⲟⲩⲱϣ[ϥ ⲉϥ]ϫⲱ [ⲙⲙⲟⲥ ⲛⲁϥ ϫⲉ ϩⲛ̄ ⲟⲩ]ⲁ̣ⲛⲁϣ ⲡⲉⲕⲉⲓⲱⲧ [ⲁϥⲧⲁⲣⲕⲉ ⲡⲗⲁⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ] ϫⲉ ϥⲥϩⲟⲩⲟⲣⲉⲧ ⲛϭ ̄ ⲓ ⲡ̣ⲣⲱ ̣ ̣ ⲙ̣ ⲉ ̣ [ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟⲉⲓⲕ] ⲙⲡⲟⲟⲩ ⲡⲗⲁⲟⲥ ⲇⲉ ⲁ̣ϥ̣ⲥⲱ[ϣⲙ̄

In the first clause, there is one important difference: SaAU use the structure ⲁϥⲟⲩⲱϣϥ ⲉϥϫⲱ ‘He answered, saying’ against SaM that offers ⲛⲉϥϫⲱ ‘He said’. In Greek, the verb is ἀπεκρίθη ‘to answer’. Usually, SaA shortens in the same way as SaM here. However, if this longer reading was the original Coptic, as is the case in other similar passages,81 then SaAU do not share a secondary reading, but rather SaM contains a later alteration.82 Additionally, Greek manuscript 44 reads the beginning with the verb ‘to say’. The omission of the verb ‘answer’ is logical, since no question anticipated an answer. Instead, the verb ‘answer’ only corresponds to the preceding action. The verbal forms in the second clause differ slightly; SaMU construct the clause with an extraposited term and a personal infix ⲡⲉⲕⲉⲓⲱⲧ· ⲁϥⲧⲁⲣⲕⲉ ‘your father, he charged’ against SaA which has a past tense with a nominal subject ⲁ ⲡⲕⲓⲱⲧ ⲧⲁⲣⲕⲉ ‘your father charged’.83 This does not alter the meaning, but offers an alternative formulation.84 Additionally, SaM defies Coptic grammar by using a relative converter with the indefinite article. Usually, a relative as a simple attributive is expressed with the circumstantial converter whenever the antecedent has an indefinite article.85 Here, SaA uses the definite article and the same seems to be the case with SaU. Butts believes that SaU could have contained either of the readings, although traces of the definite article ⲡ are visible. The papyrus has a lacuna between ⲛϭⲓ and ⲡⲣⲱⲙⲉ, and the separate papyrus parts are currently situated in a slightly misleading manner. One should place the parts closer to each other, so that ⲱ, ϥ and ⲙ on the lines above are complete.

81 See, for example, 23: 4, 25: 10 and 29: 9. 82 The same wording as in SaM is attested by the Ethiopic version. 83 The reading of SaU is partly a reconstruction, but there is enough text visible: there is no ⲁ before ⲡⲉⲕⲉⲓⲱⲧ. 84 All Sahidic witnesses read tautological infinitive in this verse (ϩⲛ︥ ⲟⲩⲁⲛⲁϣ…ⲧⲁⲣⲕⲉ, which is mentioned in Shisha-Halevy’s examples as belonging to the group “lexical grammaticalsemantic divergence”, Shisha-Halevy “Tautological Infinitive”, 108–10. 85 Layton, Grammar, §404.

170

Affiliations of the Sahidic manuscripts and the Greek Vorlage

After this operation, there would actually no longer be space for the indefinite article ⲟⲩ. 14: 29 καὶ ἔγνω Ἰωναθὰν καὶ εἶπεν Ἀπήλλαχεν ὁ πατήρ μου τὴν γῆν· ἰδὲ δὴ ὅτι εἶδον οἱ ὀφθαλμοί μου, ὅτι ἐγευσάμην βραχὺ τοῦ μέλιτος τούτου. SaM

SaA

SaU

ⲁⲩⲱ ⲓⲱⲛⲁⲑⲁⲛ ⲁϥⲉⲓⲙⲉ ⲡⲉϫⲁϥ ϫⲉ ⲡⲁⲉⲓⲱⲧ ⲁϥⲑⲉⲙⲕⲉ ⲡⲕⲁϩ· ⲁⲛⲁⲩ ⲉⲑⲉ ⲛⲧⲁ ⲛⲁⲃⲁⲗ ϫⲱⲧⲉ ⲉⲃⲟⲗ ϫⲉ ⲁⲓⲧⲉⲡ ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲓⲉⲃⲓⲱ̄ ·

ⲁⲩⲱ ⲓⲱⲛⲁⲑⲁⲛ ⲁϥⲓⲙⲉ· ⲡⲉϫⲁϥ ϫⲉ ⲡⲁⲓⲱⲧ ⲁϥⲑⲙⲕⲉ ⲡⲗⲁⲟⲥ· ⲁⲛⲁⲩ ⲉⲑⲉ ⲛⲧⲁ ⲛⲁⲃⲁⲗ ϫⲱⲧⲉ· ⲉⲃⲟⲗ ϫⲉ ⲁⲓⲧⲉⲡ ⲡⲓϣⲏⲙ ⲛⲉⲃⲓⲱ·

[ⲁ]ⲩ̣[ⲱ ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲁϥⲉⲓⲙ]ⲉ ⲡⲉϫⲁϥ ϫⲉ ⲡⲁ[ⲉⲓⲱⲧ ⲁϥ]ⲑ̣ⲙ̣ ̄[ⲕⲉ ⲡ… ⲁⲛⲁⲩ ⲉ]ⲑⲉ ⲛⲧⲁ ⲛⲁⲃⲁⲗ ϫⲱ̣ [ⲧⲉ ⲉ]ⲃⲟⲗ ϫ̣ ⲉ ̣ [ⲁⲓⲧⲉⲡ ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ] ϩⲙ ̣ ̄ ⲡⲉⲉⲓⲉⲃⲓⲱ·

In the third clause, SaM and Greek manuscripts read ‘land’, but SaA alone offers ‘people’. The last phrase is ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲓⲉⲃⲓⲱ̄ ‘a little of this honey’ in SaMU, but ⲡⲓϣⲏⲙ ⲛⲉⲃⲓⲱ ‘this little honey’ in SaA. The reading of SaU is largely reconstructed, but in addition to ϩⲙ̄ just after the lacuna, the space is a strong indicator that the reading of SaU conforms to that of SaM.86 The reading of SaA is best explained by the inclination towards shortening in this manuscript. The reading of SaMU is nearer to the Greek text. 14: 30 ἀλλ᾽ὅτι εἰ ἔφαγεν ἔσθων ὁ λαὸς σήμερον τῶν σκύλων τῶν ἐχθρῶν αὐτῶν ὧν εὗρεν, ὅτι νῦν ἂν μείζων ἦν ἡ πληγὴ ἐν τοῖς ἀλλοφύλοις. εἰ] η 707; > A B O 82* CI 799 68 55 460 ἔφαγεν] pr καί L 554

SaM

SaA

SaU

ⲉ̄ⲓⲉ̄ ⲉⲛⲉ ⲛⲧⲁ ⲡⲗⲁⲟⲥ ϭⲉ ⲟⲩⲱⲙ· ϩⲛ ⲟⲩⲱⲙ ⲙⲡⲟⲟⲩ· ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ ⲛⲛⲉⲩϫⲁϫⲉ· ⲛⲁⲓ ⲛⲧⲁϥϩⲉ ⲉ̄ⲣⲟⲟⲩ ⲧⲉⲛⲟⲩ· ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ̈ ⲡⲉ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ·

ⲉⲓⲉ ⲛⲧⲁ ⲡⲗⲁⲟⲥ ⲟⲩⲱⲙ ϩⲛ ⲟⲩⲱⲙ ⲉⲙⲡⲟⲟⲩ ⲉⲃⲟⲗ ϩⲛ ⲛϣⲟⲗⲥ ⲛⲛⲉⲩϫⲁϫⲉ· ⲛⲧⲁϥϩⲉ ⲉⲣⲟⲟⲩ ⲧⲉⲛⲟⲩ· ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ ⲡⲉ ϩⲓϫⲛ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ·

ⲉⲉⲓⲉ ⲉⲛⲉ ⲛ[ⲧⲁ ⲡⲗⲁⲟⲥ ϭⲉ ⲟⲩⲱⲙ ϩⲛ ⲟ]ⲩⲱⲙ ⲙ̄ⲡⲟⲟⲩ ⲉⲃ[ⲟⲗ] ϩⲛ̄[… ⲛⲛⲉⲩϫⲁϫⲉ ⲛⲁⲓ87] ⲛⲧⲁϥϩⲉ ⲉⲣⲟⲟⲩ ⲧⲉⲛⲟⲩ̣ [ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ̈] ⲡⲉ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫ[ⲩ]ⲗⲟⲥ

86 Thus, also, Butts, “P.Duk.inv. 797”, 16. He does not indicate that the upper curve of ϩ in ϩⲙ̄ is also visible, but instead marks ϩ as reconstructed. 87 Butts (“P.Duk.inv. 797”, 16) marks this word with dots, admitting that it is “very uncertain”. I place the word in square brackets since I am not able to decipher any of the letters. There seems to be enough space for this word.

171

Textual analyses

In the first clause, SaMU have ⲉ̄ⲓⲉ̄ ⲉⲛⲉ ⲛⲧⲁ ‘Surely, if…were/had’ against SaA that reads ⲉⲓⲉ ⲛⲧⲁ ‘If…were/had’. The clause construction is a contrafactual conditional clause with ⲉⲛⲉ ⲛⲧⲁ in the stipulation clause and ⲛⲉⲣⲉ-… ⲛⲁ-… ⲡⲉ ‘…would…’ in the main clause. ⲉⲓⲉ can be understood as strengthening the next word, like in SaMU or as meaning the same as ⲉⲛⲉ, like in SaA here.88 This explains the fact that SaA omits ⲉⲛⲉ; if ⲉⲓⲉ is understood as a conditional conjunction, ⲉⲛⲉ is redundant. In the Greek tradition, there is variation in this sentence that is not relevant to these Sahidic variants. Butts reconstructs ϭⲉ ‘then’ within the first clause in SaU on the basis of space, and following SaM.89 Perhaps, however, SaU did not have the simplified reading of SaMA ϭⲉ ⲟⲩⲱⲙ ϩⲛ ⲟⲩⲱⲙ ‘then had eaten in eating’ but the longer version ⲟⲩⲱⲙ ϩⲛ ⲟ ⲩ ⲟⲩⲱⲙ, which would require the same space. An adverb of manner is expressed with a ϩⲛ ⲟⲩ-construction in Sahidic.90 In this case, SaMA lack the first two characters, probably as a result of a homoiarchon mistake (ϩⲛ ⲟⲩⲟⲩ becomes ϩⲛ ⲟⲩ). However, the shared reading of SaMA is the reason that I have followed Butts’ reconstruction with ϭⲉ. SaM reads ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ ‘from the plunder’, where SaA has ⲉⲃⲟⲗ ϩⲛ ⲛϣⲟⲗⲥ. These nouns are synonyms, and both of them are used in SaM and SaA within the same verse, 30: 20, to render Greek σκῦλον. In the last clause, SaMU have the preposition ⲉϫⲛ,̄ while SaA offers ϩⲓϫⲛ. The basic meaning of these two is similar ‘on, upon’, and one cannot determine which was the original Sahidic reading. 14: 31 καὶ ἐπάταξεν ἐν τῇ ἡμέρᾳ ἐκείνῃ ἐκ τῶν ἀλλοφύλων ἐν Μαχεμάς, καὶ ἐκοπίασεν ὁ λαὸς σφόδρα. ἐκ] > L 527 318

SaM

SaA

SaU

ⲁⲩⲱ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ· ϩⲙ ⲙⲁⲭⲙⲁⲥ· ⲡⲗⲁⲟⲥ ⲇⲉ ⲁϥϩⲓⲥⲉ ⲉⲙⲁⲧⲉ·

ⲁⲩⲱ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϩⲙ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ ϩⲙ ⲙⲁⲭⲙⲁⲥ· ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ ⲁϥϩⲓⲥⲉ ⲉⲙⲁⲧⲉ

[ⲁⲩⲱ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲉⲃⲟⲗ ϩⲛ91 ̄ ⲛⲁⲗⲗⲟ]ⲫⲩⲗⲟⲥ ⲙ̄[ⲡⲉ]ϩ̣ [ⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ ϩⲙ ⲙⲁⲭⲙⲁⲥ ⲡⲗⲁⲟⲥ ⲇ]ⲉ ⲁϥϩⲓⲥ[ⲉ ⲉⲙⲁ] ⲧ̣ ⲉ ̣

According to SaA, all the Philistines were smitten ‘in that place’. SaMU more accurately follow the Greek by saying that some of the Philistines were smitten on that day. The preposition ἐκ does not appear in Greek manuscripts L 527 318, but, still, the meaning does not accord with that of SaA. SaA offers ϩⲙ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ ‘in that place’ in the same clause against ⲙⲡⲉϩⲟⲟⲩ

88 89 90 91

Crum, Dictionary, 74–5. Butts, “P.Duk.inv. 797”, 16. Layton, Grammar, §221. In this line, the spacing suggests the reconstruction with ⲉⲃⲟⲗ ϩⲛ̄.

172

Affiliations of the Sahidic manuscripts and the Greek Vorlage

ⲉⲧⲙ̄ⲙⲁⲩ ‘on that day’ in SaMU and the Greek. There are no similar Greek variants, and this probably constitutes an error in SaA, the characters are almost the same, and both expressions fit into this context. In the last clause, SaMU have the particle ⲇⲉ where SaA has ⲧⲏⲣϥ ‘all’. There is a similar case in verse 14: 26 above. In Greek, there is no reading corresponding to the variant of SaA. 14: 32 καὶ ἐκλίθη ὁ λαὸς εἰς τὰ σκῦλα· καὶ ἔλαβεν ὁ λαὸς ποίμνια καὶ βουκόλια καὶ τέκνα βοῶν, καὶ ἔσφαξεν ἐπὶ τὴν γῆν· καὶ ἤσθιεν ὁ λαὸς σὺν τῷ αἵματι. SaM

SaA

SaU

ⲁⲩⲱ ⲡⲗⲁⲟⲥ· ⲁϥⲧⲁⲁϥ ⲉⲛϣⲱⲗ· ⲁ ⲡⲗⲁⲟⲥ ϫⲓ ⲛϩⲛ̄ⲟϩⲉ ⲛⲉⲥⲟⲟⲩ· ⲙⲛ̄ ϩⲉⲛⲁϩⲟⲟⲩ· ⲁⲩⲱ ϩⲛ̄ⲕⲧⲏⲣ ⲛⲉϩⲉ· ⲁⲩϣⲱⲱⲧ ⲙⲙⲟⲟⲩ ϩⲓϫⲙ̄ ⲡⲕⲁϩ· ⲁⲩⲱ ⲡⲗⲁⲟⲥ· ⲛⲉϥⲟⲩⲱⲙ ⲙⲛ̄ ⲡⲉϥⲥⲛⲟϥ·

ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ ⲁϥⲧⲁⲁϥ ⲛϣⲱⲗ· ⲁ ⲡⲗⲁⲟⲥ ϫⲓ ⲛ ϩⲛⲉⲥⲟⲟⲩ ⲙⲛ ϩⲛⲉϩⲟⲟⲩ ⲙⲛ ϩⲛⲕⲧⲏⲣ· ⲁⲩϣⲱⲱⲧ […

[ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲁϥⲧⲁⲁϥ ⲉⲛ]ϣⲱⲗ· ⲁⲡⲗⲁⲟⲥ ϫⲓ ⲛϩⲛ̣ⲁ̣ϩⲉ̣ ̣ [ⲛⲉⲥⲟⲟⲩ ⲙⲛ̄ ϩⲉⲛⲁ]ϩⲟⲟⲩ ⲁⲩⲱ ϩⲉⲛⲕⲧ̣ ⲏ̣ⲣ̣ ⲛ[ⲉϩⲉ ⲁⲩϣⲱⲱⲧ ⲙⲙⲟⲟⲩ ϩⲓϫⲙ̄] ⲡⲕⲁϩ· ⲁⲩⲱ [ⲡ]ⲗ̣ⲁⲟ[ⲥ ⲛⲉϥⲟⲩⲱⲙ ⲙⲛ̄ ⲛⲉϥⲥⲛⲟϥ]

In the first clause, SaA complements ‘the people’ with the attribute ⲧⲏⲣϥ ‘all, entire’, against SaMU which lack the complement.92 SaA often adds ⲧⲏⲣϥ after the people without a Greek source. In this chapter, this addition also occurs in verses 14: 26 and 14: 31. Probably, this is due to the frequency of the phrase πᾶς ὁ λαός, which is rendered ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ. In the second sentence, SaA clearly abbreviates, omitting words unnecessary for understanding the narrative; ⲛϩⲛⲟ̄ ϩⲉ ⲛⲉⲥⲟⲟⲩ ‘flocks of sheep’ thus becomes ⲛ ϩⲛⲉⲥⲟⲟⲩ ‘sheep’ and ϩⲛⲕ̄ ⲧⲏⲣ ⲛⲉϩⲉ ‘calves of ox’ becomes ϩⲛⲕⲧⲏⲣ ‘calves’. The text of SaU is difficult to reconstruct here since it seems to read ⲛϩⲛⲁϩⲉ instead of ⲛϩⲛⲟ̄ ϩⲉ of SaM.93 Additionally, the line reconstructed [ⲛⲉⲥⲟⲟⲩ ⲙⲛ̄ ϩⲉⲛⲁ]ϩⲟⲟⲩ ⲁⲩⲱ ϩⲉⲛⲕⲧⲏⲣ̣ ⲛ is a bit short, lacking 3–6 characters when compared with other lines of the papyrus.94 Results The data in the survey above cogently supports the premise that these three Sahidic Samuel witnesses share a common Sahidic predecessor. Typically, the wordings are practically identical, and additional examples illustrate particularly peculiar agreements. In 14: 26, all three Sahidic manuscripts men92 The reading of SaU is based on the requirement of space. Nothing is legible there. 93 Crum (Dictionary, 64) gives ⲁϩⲏ as a Fayyumic form of ⲉϩⲉ ‘ox, cow’. ⲁϩⲉ as such is one form of the particle that is spelled ⲉϩⲉ and ⲁϩⲏ (Fayyumic). The word ⲟϩⲉ ‘flock, fold’ is also spelled ⲁϩⲓ in Fayyumic. 94 According to Butts (“P.Duk.inv. 797”, 16), the line lacks ca. 6 characters, but taking into consideration the varying script, it could be said that 6 characters is a maximum.

Textual analyses

173

tion the subject ⲡⲗⲁⲟⲥ explicitly against unanimous Greek evidence with only a 3rd person singular verbal form. Later, in the same verse, Sahidic lacks any rendering for the word ἰδού, attested once or twice in Greek, depending on the manuscript. In 14: 27, all Sahidic witnesses again mention the subject explicitly, against all the Greek witnesses, which only allude to the subject through a 3rd person singular verb. Some relationships among the Sahidic Samuel manuscripts seem clear. There are numerous passages where SaMU agree against SaA.95 These pertain to conjunction (14: 25, 14: 26, 14: 31), preposition/adverb (14: 27, 14: 30 ), abbreviated phrase (14: 29, 14: 30, 14: 32), attribute (14: 25, 14: 32), verbal form (14: 25, 14: 27, ), construction (14: 27, 14: 28, 14: 31), and vocabulary (14: 27). In 14: 32, the text of SaU is difficult to reconstruct, but at least it is not shortened like SaA. In three passages, SaAU agree against SaM.96 In verse 14: 26, SaAU have ϩⲏⲧϥ where SaM reads ⲛϩⲏⲧϥ without altering the meaning. In 14: 28, SaAU have ⲁϥⲟⲩⲱϣϥ ⲉϥϫⲱ against SaM which reads ⲛⲉϥϫⲱ. Probably, SaAU preserve the original Sahidic reading, and SaM has been redacted. Otherwise, it would be hard to explain a connection between SaAU in only one detail. In 14: 28, SaM uses a relative converter with the indefinite article. SaA attests the definite article, and the same seems to be the case with SaU. This last case is probably a mistake in SaM, since it defies the usual practice in Sahidic by using a relative converter with the indefinite article. There are no passages where SaAM agree against SaU. Butts mentions one uncertain reading in verse 14: 26 as belonging under this heading.97 However, there is a minor mistake in his edition, and thus his conclusion does not hold here. Affiliations of Sahidic manuscripts to the Greek tradition are of special interest, but there are not many cases available in these verses. In verse 14: 24, SaMA offer the proper name Saul like CI in Greek and the MT. The Hexaplaric origin of this reading in Greek manuscripts is established by the asterisk found in 731 in Greek. However, in many cases, Sahidic manuscripts add the subject explicitly without any Greek source, and, therefore, I see this connection as a coincidence.98 In 14: 24, the object of the verb is singular in SaM, but plural in SaA. However, there is no need to suppose that SaA follows Greek 318 460c in reading the plural since the interchange of singular and plural can happen easily in Coptic. In other passages, plural and singular vary without connection to Greek. Neither of these readings in

95 Butts (“P.Duk.inv. 797”, 11, n. 38) lists 17 passages where SaU disagrees with SaA. I have compared SaMU and SaA, and have 18 fairly certain cases. I mention here only those cases where no connection to a Greek manuscript is evident in SaA. 96 Butts (“P.Duk.inv. 797”, 11, n. 36) mentions only the first two cases. The third case cited above does not appear in his list because he has deciphered the text of SaU differently. 97 Butts, “P.Duk.inv. 797”, 11. 98 See section 2.2.3.3 for a comprehensive list and analysis of those cases.

174

Affiliations of the Sahidic manuscripts and the Greek Vorlage

14: 24 is visible in SaU. In verse 14: 25, the Sahidic manuscripts offer no equivalent for ἰδού. This same word is omitted in L O. This omission in Sahidic might be a redaction against some Greek manuscript with the same omission. However, the Sahidic text is more prone to addition, and usually does not omit anything. In 14: 28, the Ethiopic version attests the same wording as SaM. Additionally, Greek manuscript 44 reads the beginning with the verb ‘to say’, omitting the verb ‘to answer’. This agreement is almost certainly best explained by the context of the narrative and the translator’s desire to create a coherent discourse; the dialogue incorporates no question in need of an answer, but, instead, one person has commented on an action. 3.2.4. SaMAB 17: 33– 44 17: 33 καὶ εἶπεν Σαοὺλ πρὸς Δαυίδ Οὐ μὴ δύνῃ πορευθῆναι πρὸς τὸν ἀλλόφυλον τοῦ πολεμῆσαι μετ᾽αὐτοῦ, ὅτι παιδάριον εἶ σύ, καὶ αὐτὸς ἀνὴρ πολεμιστὴς ἐκ νεότητος αὐτοῦ. SaM

SaA99

SaB

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲉϩⲟⲩⲛ̄ ϩⲛ ⲇⲁⲩⲉⲓⲇ· ϫⲉ ⲛⲅⲛⲁⲉϣϭⲉⲙϭⲟⲙ ⲁⲛ ⲉⲃⲱⲕ· ⲉⲙⲓϣⲉ ⲙⲛ̄ ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ϫⲉ ⲛ̄ⲧⲕ ⲟⲩϣⲏⲣⲉ ϣⲏⲙ ⲛ̄ⲧⲟⲕ ⲡⲏ ⲇⲉ ⲟⲩⲣⲱⲙⲉ ⲡⲉ ⲛⲣⲉϥⲙⲓ̈ϣⲉ ⲛ̄ϫⲓⲛ̄ ⲧⲉϥⲙⲛ̄ⲧⲕⲟⲩⲓ·

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲉϩⲟⲩⲛ ϩⲛ ⲇ̅ⲁⲇ ̅ ̅ ϫⲉ ⲛ̇ⲕⲛⲁϣ̇ ϭⲙϭⲟⲙ ⲁⲛ ⲉⲃⲱⲕ ⲉⲃⲟⲗ ⲉⲙⲓ̈ϣⲉ ⲙⲛ̇ ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ϫⲉ ⲛⲧⲕ ⲟⲩϣⲏⲣⲉ ϣⲏⲙ ⲛ̇ⲧⲟⲕ ⲡⲁⲓ ⲇⲉ ⲟⲩⲣⲱⲙⲉ ⲛ̇ⲣϥ̇ⲙⲓϣⲉ ⲡⲉ ϫⲓⲛ ⲉⲧϥ̇[ⲙ]ⲛ̅ⲧ̅ⲕⲟⲩⲓ̈·

…ⲧⲉϥⲙⲛ̄ⲧⲕⲟⲩⲓ·

πορευθῆναι πρὸς τὸν ἀλλόφυλον τοῦ πολεμεῖν μετ᾽αὐτοῦ ‘go to the Philistine to fight with him’ is rendered ⲉⲃⲱⲕ· ⲉⲙⲓϣⲉ ⲙⲛ̄ ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ‘go to fight with this Philistine’. SaM and SaA share the same wording, differing in unison slightly from the Greek. Since the Sahidic reading did not arise from a reading in any extant Greek manuscript, the Sahidic readings are therefore best explained as having arisen from the same translation event. This rendering shows that the Sahidic translator read and rendered οὐ μὴ δυνήσῃ πορευθῆναι πρὸς τὸν ἀλλόφυλον τοῦ πολεμεῖν μετ᾽αὐτοῦ as one unit. This kind of information is significant when we describe the translator and try to determine the length of the passages that he translated as a unit. In the last sentence, some small discrepancies between SaM and SaA arise. In these cases, one cannot prioritize one Sahidic reading over the other so far as their originality is concerned. There is variation in the demonstrative 99 In this section (3.2.4), I have consulted the actual fragment of SaA, together with Wessely’s edition (Wessely, Griechische und koptische Texte theologischen Inhalts). Obviously, the fragment has been cleaned after Wessely’s time, and there is more legible text than he had quite exactly one hundred years ago.

175

Textual analyses

pronouns: ⲡⲏ in SaM and ⲡⲁⲓ in SaA. Crum gives the English equivalents ‘this’ for ⲡⲁⲓ and ‘that’ for ⲡⲏ. In the same sentence, SaM and SaA order the particle ⲡⲉ differently. SaM places this copula between a noun and its attribute ⲟⲩⲣⲱⲙⲉ ⲡⲉ ⲛⲣⲉϥⲙⲓ̈ϣⲉ, whereas SaA sets the copula after the noun and its attribute ⲟⲩⲣⲱⲙⲉ ⲛⲣⲉϥⲙⲓ̈ϣⲉ ⲡⲉ. In the last phrase, SaA and SaM preserve different prepositions. SaA reads ϫⲓⲛ ⲉ-, whereas SaM reads ⲛϫ ̄ ⲓⲛ.̄ This variation does not affect the meaning, only the wording of the expression ‘from his youth’. 17: 34 καὶ εἶπεν Δαυὶδ πρὸς Σαούλ Ποιμαίνων ἦν ὁ δοῦλός σου τῷ πατρὶ αὐτοῦ ἐν τῷ ποιμνίῳ· καὶ ὅταν ἤρχετο ὁ λέων καὶ ἡ ἄρκος καὶ ἐλάμβανεν πρόβατον ἐκ τῆς ἀγέλης, SaM

SaA

SaB

ⲇ̄ⲁⲇ ̄ ̄ ⲇⲉ ⲡⲉϫⲁϥ ⲛⲥⲁⲟⲩⲗ· ϫⲉ ⲁⲛⲟⲕ ⲡⲉⲕϩⲙ̄ϩⲁⲗ· ⲛⲉⲓⲙⲟⲟⲛⲉ ⲛⲛⲉⲥⲟⲟⲩ ⲙⲡⲁⲉ̅ⲓⲱⲧ· ⲁⲩⲱ ̅ ⲉⲣϣⲁⲛ ⲡⲙⲟⲩⲓ ⲙⲛ ⲧⲁⲣⲝ ⲉ̅ⲓ.̅ ⲛⲧⲉⲣⲱϥ ϫⲓ ⲟⲩⲉⲥⲟⲟⲩ ⲛϩⲏⲧⲟⲩ·

[ⲇⲁ]ⲩⲉⲓⲇ ⲇⲉ ⲡⲉϫⲁϥ ⲛ̇[ⲥ]ⲁⲟⲩⲗ· ϫⲉ ⲡⲕϩⲉⲙ[ϩ]ⲁⲗ ⲉⲧⲉ ⲁⲛⲟⲕ ⲡⲉ [ⲛ]ⲉⲓⲙⲟⲟⲛⲉ ⲛ̇ⲛⲉⲥⲟⲟⲩ [ⲙ]ⲡⲁⲓⲱⲧ· ⲁⲩⲱ ⲣ̇[ϣ]ⲁⲛ ⲡⲙⲟⲩⲓ ⲙⲛ̇ ⲧⲁ[ⲣ]ⲝ ⲉⲓ· ⲛⲧⲉⲣⲱϥ ϫⲓ ⲉⲛⲟⲩⲉⲥⲟⲟⲩ ⲉⲃⲟⲗ ⲛϩⲏⲧⲟⲩ·

ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲥⲁⲟⲩⲗ· ϫⲉ ⲁⲛⲟⲕ ⲡⲉⲕϩⲙϩⲁⲗ ⲛⲉⲓⲙⲟⲟⲛⲉ ⲛⲛⲉⲥⲟⲟⲩ ⲙⲡⲁⲉⲓⲱⲧ· ⲉⲣϣⲁⲛ ⲡⲙⲟⲩⲓ· ⲙⲛ ⲧⲁⲣⲝ ⲉⲓ ⲛ̄ⲧⲉ ⲣⲱϥ ϫⲓ ⲛⲟⲩⲉⲥⲟⲟⲩ ⲉⲃⲟⲗ ⲛ̄ϩⲏⲧⲟⲩ·

In the first clause, SaMA construct their sentence similarly with the connective ⲇⲉ, a verb with a personal infix as subject and the name David in extraposition. SaB omits the conjunction and reads the verb with a nominal subject. All three Sahidic witnesses use the word ⲁⲛⲟⲕ ‘I’ where the Greek tradition has ὁ δοῦλός σου. Naturally, the conjugation of the Greek and Sahidic verbs depend on the nature of this subject. Sahidic witnesses have a 1st person singular, whereas Greek has a 3rd person singular. Accordingly, in the same clause, Sahidic witnesses offer ‘my father’, where Greek reads ‘his father’. These features show that the translator did not try to imitate the source text word for word, but rendered ad sensum. The last sentence begins with the conjunction ⲁⲩⲱ ‘and’ in SaMA but without it in SaB, as is often the case in this manuscript. πρόβατον ἐκ τῆς ἀγέλης is rendered ⲟⲩⲉⲥⲟⲟⲩ (ⲉⲃⲟⲗ) ⲛϩⲏⲧⲟⲩ· ‘sheep from them’ in Sahidic against ‘sheep from the herd’ in Greek.

176

Affiliations of the Sahidic manuscripts and the Greek Vorlage

17: 35 καὶ ἐξεπορευόμην ὀπίσω αὐτοῦ καὶ ἐπάταξα αὐτὸν καὶ ἐξέσπασα ἐκ τοῦ στόματος αὐτοῦ· καὶ εἰ ἐπανίστατο ἐπ᾽ἐμέ, καὶ ἐκράτησα τοῦ φάρυγγος αὐτοῦ καὶ ἐπάταξα καὶ ἐθανάτωσα αὐτόν. SaM

SaA

SaB

ϣⲁⲓⲁⲛⲅⲧ ̄ ·̄ ⲉⲃⲟⲗ ⲉⲡⲁϩⲟⲩ ⲙⲙⲟϥ· ⲛⲧⲁⲣⲁϩⲧ ⲡⲙⲟⲩⲓ· ⲛⲧⲁϥⲓ ⲉⲃⲟⲗ ϩⲛ̄ ⲧⲉϥⲧⲁⲡⲣⲟ· ⲙⲡⲉⲛⲧⲁ ⲡⲙⲟⲩⲓ ⲧⲟⲣⲡϥ̄· ⲉϥϣⲁⲛⲕⲧⲟϥ ⲇⲉ ⲉⲣⲟⲓ· ϣⲁⲓⲙⲉϩⲧⲟⲟⲧ ϩⲛ ⲧⲉϥϣⲟⲩⲱⲃⲉ̄· ⲛⲧⲁⲁⲁϥ ⲛϭ ̄ ⲱⲃ ⲛⲧⲁⲙⲟⲟⲩⲧϥ̄·

ϣⲁⲓⲡⲱⲧ ϩⲓⲡⲁϩⲟⲩ ⲙ̇ ⲙⲟϥ ⲧⲁⲣⲱϩⲧ ⲙ̄ⲡ̣[ⲙⲟⲩⲓ] ⲛⲧⲁϥⲓ ⲉ[ⲃⲟⲗ ϩ]ⲛ ⲧϥⲧⲁⲡⲣⲟ ⲙⲡⲛ̇ⲧⲁϥⲧⲟⲣⲉⲡϥ· ⲉϥϣⲁⲛⲕⲧⲟϥ ⲇⲉ ⲉⲣⲟⲓ ϣⲁⲓⲙⲉϩⲧⲟⲟⲧ ϩⲛ̇ ⲧϥ̇ϣⲟⲩⲱⲃⲉ ⲧⲁⲁϥ ⲛ̇ϭⲱϥ [ⲧ]ⲁⲙⲟⲟⲩⲧϥ̇·

ϣⲁⲓⲟⲛⲅⲧ ϩⲓⲡⲁϩⲟⲩ ⲙⲙⲟϥ ⲧⲁⲣⲱϩⲧ ⲙ̄ⲡⲙⲟⲩⲓ ⲧⲁϥⲓ ⲉⲃⲟⲗ ϩⲛ ⲣⲱϥ ⲙ̄ⲡⲉⲛⲧⲁϥⲧⲟⲣⲡⲉϥ· ⲉϣϥⲁⲛⲕⲧⲟϥ ⲉⲣⲟⲓ̈ ϣⲁⲓⲟⲛⲅⲧ ϩⲓⲡⲁϩⲟⲩ ⲙⲙⲟϥ ⲧⲁⲣⲱϩⲧ ⲙⲡⲙⲟⲩⲓ· ⲧⲁϥⲓ ⲉⲃⲟⲗ ϩⲛ ⲣⲱϥ ⲙⲡⲉⲛⲧⲁϥⲧⲟⲣⲡⲉϥ·

The first verb in this verse is ⲱⲛⲕ ‘to leap’ in SaMB. SaM adds the adverb ⲉⲃⲟⲗ.100 SaA uses a more common verb ⲡⲱⲧ ‘run, go’, being obviously secondary here, since the usual direction of changes is from rare words to more common ones. The second clause offers the verbal form ⲛⲧⲁⲣⲁϩⲧ ‘to strike’ in SaM, but ⲧⲁⲣⲱϩⲧ in SaAB. The same feature recurs with the next verbs ⲛⲧⲁϥⲓ/ⲧⲁϥⲓ ‘to take’, ⲛⲧⲁⲁⲁϥ/ⲧⲁⲁϥ ‘to do’ and ⲛⲧⲁⲙⲟⲟⲩⲧϥ/̄ ⲧⲁⲙⲟⲟⲩⲧϥ ̄ ‘to kill’. The readings parse the same, since the conjunctive of the 1st person singular has the variant orthographies, ⲛⲧⲁ- and ⲧⲁ-.101 The Sahidic witnesses complement the verb ⲣⲱϩⲧ ‘to strike’ with the object ‘bear’, while Greek has only αὐτόν. SaMA render Greek στόμα with the word ⲧⲁⲡⲣⲟ, while SaB uses ⲣⲟ. In verse 14: 27 above, the word used for ‘mouth’ is ⲣⲟ in SaMU but ⲧⲁⲡⲣⲟ in SaA. According to Vycichl, ⲧⲁⲡⲣⲟ means ‘mouth, entry, bit, tip, point, peak’.102 ⲣⲟ has the meanings ‘mouth, door, gate, cutting edge of a weapon’.103 In Sahidic Jeremiah, a later manuscript has systematically changed ⲣⲟ to ⲧⲁⲡⲣⲟ.104 Minimally, one can conclude that SaA has the word ⲧⲁⲡⲣⲟ in both passages. The Sahidic translator typically complements his sentences with the direct object.105 In this verse, Sahidic manuscripts add the object after the phrase ‘out of its mouth’, without any Greek parallel. SaM has ⲙⲡⲉⲛⲧⲁ ⲡⲙⲟⲩⲓ 100 Crum (Dictionary, 524b) gives the form of the verb ⲁⲛⲅ ̄ found in SaM with the siglum Saf and thus indicates that it belongs to Fayyumic features in Sahidic. 101 Layton, Grammar, §351. 102 Vycichl, Dictionnaire Étymologique, 219. 103 Vycichl (Dictionnaire Étymologique, 171) makes a comparison with the Hebrew expression ‫פי־החרב‬. In the Sahidic New Testament, two common usages are found for these two words: as equivalent for θύρα and as equivalent for στόμα. ⲣⲟ is used 35 times for θύρα and 33 times for στόμα, (Wilmet, Concordance II: 2, 679). ⲧⲁⲡⲣⲟ is used only once for θύρα and 45 times for στόμα, (Wilmet, Concordance II: 2, 936). 104 Feder, Biblia Sahidica, 59. 105 See my section on additions, 2.2.3.1.

177

Textual analyses

ⲧⲟⲣⲡϥ̄ ‘what the lion had seized’ while SaAB attest ⲙ̄ⲡⲉⲛⲧⲁϥⲧⲟⲣⲡⲉϥ ‘what it had seized’. 17: 36 καὶ τὸν λέοντα καὶ τὴν ἄρκον ἔτυπτεν ὁ δοῦλός σου, καὶ ἔσται ὁ ἀλλόφυλος ὁ ἀπερίτμητος ὡς ἓν τούτων· οὐχὶ πορεύσομαι καὶ πατάξω αὐτὸν, καὶ ἀφελῶ σήμερον ὄνειδος ἐξ Ἰσραήλ; διότι τίς ὁ ἀπερίτμητος οὗτος ὃς ὠνείδισεν παράταξιν θεοῦ ζῶντος; ἀλλόφυλος] + ουτος L 530 135 158 554

SaM

SaA

SaB

ⲡⲙⲟⲩⲓ ϭⲉ ⲙⲛ̄ ⲧⲁⲣⲝ· ⲡⲉⲕϩⲙ̄ϩⲁⲗ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟⲟⲩ· ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ϩⲱⲱϥ ⲛⲁⲧⲥ̄ⲃⲃⲉ· ⲛⲁⲉⲣ ⲑⲉ ⲛⲟⲩⲁ ⲛⲛⲁⲓ· ϯⲛⲁⲃⲱⲕ ϭⲉ ⲟⲛ· ⲉⲃⲟⲗ ⲉⲣⲟϥ· ⲛⲧⲁⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟϥ· ⲛⲧⲁϥⲓ ⲛ̄ⲧⲉϥⲁⲡⲉ· ⲛⲧⲁϥⲓ ⲙⲡⲟⲟⲩ ⲛⲟⲩⲛⲟϭⲛⲉϭ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲓⲏ̄ⲗ·̄ ϫⲉ ⲟⲩ ϩⲱⲱϥ ⲡⲉ ⲡⲓⲁⲧⲥ̄ⲃⲃⲉ̄· ⲉⲁϥⲛⲟϭⲛⲉϭ ⲙⲡⲙⲏⲏϣⲉ ⲙⲡⲛⲟⲩⲧⲉ ⲉⲧⲟⲛϩ̄·

ⲡⲙⲟⲩⲓ̈ ⲙⲛ ⲧⲁⲣⲝ ⲡⲉⲕϩⲙ̇ ϩⲁⲗ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟⲟⲩ ⲡⲁⲓ[ⲁ]ⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲁⲧⲥⲃ[ⲃ]ⲉ ⲛⲁⲣ ⲑⲉ ⲛⲟⲩⲁ ⲙⲙⲟⲟⲩ ϯⲛⲁⲃⲱⲕ ϭⲉ ⲉⲃⲟⲗ ⲉⲣⲟϥ ⲧⲁⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟϥ ⲧⲁ[ϥⲓ ⲙ̄]ⲡⲟⲟⲩ ⲛⲟⲩⲛⲟϭⲛ[ⲉϭ ⲉⲃⲟ]ⲗ ϩⲙ ⲡⲓ ̅ⲏ̅ⲗ·̅ ϫ̣ [ⲉ ⲟⲩ]106 ϩⲱⲱϥ ⲡⲉ ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϫⲉ ⲉϥⲉⲛⲟϭⲛϭ̇ ⲙⲡⲗⲁⲟⲥ ⲙ̇ ⲡⲛⲟⲩⲧⲉ ⲉⲧⲟⲛϩ̇

ⲡⲙⲟⲩⲓ ⲙⲛ̄ ⲧⲁⲣⲝ ⲡⲉⲕϩⲙ̄ϩⲁⲗ ⲡⲉⲛⲧⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟⲟⲩ· ⲡⲉⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ϩⲱⲱϥ ⲛⲁⲧⲥⲃⲃⲉ ϥⲛⲁⲉⲣ ⲑⲉ ⲛⲟⲩⲁ ⲛⲛⲁⲓ· ϯⲛⲁⲃⲱⲕ ⲉⲃⲟⲗ ⲉⲣⲟϥ ⲛⲧⲁⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟϥ· ⲧⲁϥⲓ ⲙⲡⲛⲟϭⲛⲉϭ ⲙ̄ⲡⲓⲏ̅ⲗ·̅ ⲙⲡⲟⲟⲩ ⲛϩⲟⲟⲩ·

In the second clause, some Greek manuscripts have ὁ ἀλλόφυλος and some ὁ ἀλλόφυλος οὗτος. SaM reads ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ where ⲡⲓ- can be understood as a variant form of the definite article or as an affective demonstrative.107 SaA has ⲡⲁⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ and SaB reads ⲡⲉⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ, a variant spelling of the form used in SaM. I would hesitate to cite Sahidic in the apparatus as belonging together with the variant Greek reading with demonstrative. Mink has argued that Sahidic demonstratives often are used in renderings of Greek definite articles.108 Sahidic witnesses render οὐχὶ πορεύσομαι without negation, and SaMA add the conjunction ϭⲉ ‘then’. The fact that all three Sahidic witnesses read without negation connects these manuscripts together and makes it probable that the connective at some stage belonged to the basic text of SaB. SaM is the only manuscript that reads ⲛⲧⲁϥⲓ ⲛⲧ ̄ ⲉϥⲁⲡⲉ ‘I will take his head’ after ⲛⲧⲁⲡⲁⲧⲁⲥⲥⲉ ⲙⲙⲟϥ ‘I will strike him’. There is no explanation for this addition from the Greek tradition. Perhaps this phrase was added under the

106 Wessely in his edition (Griechische und koptische Texte theologischen Inhalts, 71) has reconstructed ϫ[ⲉ ⲛⲓⲙ] ϩⲱⲱϥ but it does not fit into the space. Therefore, I have reconstructed following SaM, whose reading fits in. 107 Layton, Grammar, §58. 108 Mink, “Koptischen Versionen”, 221–2.

178

Affiliations of the Sahidic manuscripts and the Greek Vorlage

influence of verse 17: 46 (καὶ ἀποκτενῶ σε καὶ ἀφελῶ τὴν κεφαλήν σου ἀπὸ σου), explicitating the fate of the Philistine. SaM renders τίς ὁ ἀπερίτμητος οὗτος strictly (ⲟⲩ ϩⲱⲱϥ ⲡⲉ ⲡⲓⲁⲧⲥⲃ̄ ⲃⲉ ̄

‘Who is this uncircumcised?’), while SaB omits this question altogether, and SaA reads ⲟⲩ ϩⲱⲱϥ ⲡⲉ ⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ‘Who is this Philistine?’ without similar Greek variants. ὃς ὠνείδισεν παράταξιν θεοῦ ζῶντος is either rendered freely or following the Lucianic reading ὅτι ὠνείδισε παράταξιν Θεοῦ ζῶντος, with the conjunction ὅτι and not a relative clause. SaM (ⲉⲁϥⲛⲟϭⲛⲉϭ ⲙⲡⲙⲏⲏϣⲉ ‘since he has reviled the crowd’) uses a circumstantial of the simple past, which is a typical rendering for a clause headed by ὅτι. SaA (ϫⲉ ⲉϥⲉⲛⲟϭⲛϭ̄ ⲙⲡⲗⲁⲟⲥ ‘since he has reviled the people’) attests ϫⲉ and an optative. Only, SaA offers ⲗⲁⲟⲥ instead of ⲙⲏⲏϣⲉ ‘crowd’. 17: 37 Κύριος ὃς ἐξείλατό με ἐκ χειρὸς τοῦ λέοντος καὶ ἐκ χειρὸς τῆς ἄρκου, αὐτὸς ἐξελεῖταί με ἐκ χειρὸς τοῦ ἀλλοφύλου τοῦ ἀπεριτμήτου τούτου. καὶ εἶπεν Σαοὺλ πρὸς Δαυείδ Πορεύου, καὶ ἔσται Κύριος μετὰ σοῦ. init] pr και ειπεν δαυιδ (sub ( 243mg) A O L C´-243txt d s-64´ 55 71 245s 158 554; και ειπεν αυτω παλιν 509 χειρός] στοματος L 55 158; του στοματος 554 SaM

SaA

SaB

ⲁⲩⲱ ⲇⲁⲩⲉⲓⲇ ⲡⲉϫⲁϥ ϩⲓⲛⲁⲓ· ϫⲉ ⲡϫⲟⲉⲓⲥ ⲡⲉ ⲛⲧⲁϥⲛⲁϩⲙⲉⲧ ⲉⲧⲧⲁⲡⲣⲟ̄ ⲙⲡⲙⲟⲩⲓ· ⲁⲩⲱ ⲛⲧⲟⲟⲧϥ̄ ⲛⲧⲁⲣⲝ· ⲛⲧⲟϥ ⲟⲛ ⲡⲉⲧⲛⲁⲛⲁϩⲙⲉⲧ ⲛ̄ⲧⲟⲟⲧϥ̄ ⲙⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲁⲧⲥ̄ⲃⲃⲉ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲡⲉϫⲁϥ ⲛⲇⲁⲩⲉⲓⲇ· ϫⲉ ⲃⲱⲕ ⲉⲃⲟⲗ ⲉⲣⲟϥ· ⲁⲩⲱ ⲉⲣⲉ ⲡϫⲟⲉⲓⲥ ϣⲱⲡⲉ ⲛⲙ̄ⲙⲁⲕ·

ⲁⲩⲱ ⲁϥⲕⲧⲟϥ ⲟⲛ ⲛϭⲓ ⲇ̅ⲁⲇ ̅ ̅ ⲡⲉϫⲁϥ ϫⲉ ⲡϫⲟⲉⲓⲥ ⲡⲉ ⲛⲧⲁϥⲛⲁϩⲙⲉⲧ ⲉⲃⲟⲗ ϩⲛ̇ ⲧⲧⲁⲡⲣⲟ ⲙⲡⲙⲟⲩⲓ̈ ⲁⲩⲱ ⲉⲧⲟⲟⲧϥ̇ ⲛ̇ⲧⲁⲣⲝ· ⲛⲧⲟϥ ⲟⲛ ⲡⲉⲧⲛⲁⲛⲁϩⲙⲉⲧ ⲉⲧⲟⲟⲧϥ̇ ⲙⲡⲉⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲁⲧⲥⲃⲃⲉ ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲡⲉϫⲁϥ ⲛⲇ̅ⲁⲇ ̅ ̅ ϫⲉ [ⲃⲱ]ⲕ ⲉⲃⲟⲗ ⲉⲣⲟϥ ⲉⲣⲉ ⲡϫⲟⲉⲓⲥ ⲛⲁϣⲱⲡⲉ ⲛⲉⲙⲙⲁ[ⲕ]

ⲡⲛⲟⲩⲧⲉ ⲡⲉⲛⲧⲁϥⲛⲁϩⲙⲉⲧ ⲉⲡⲙⲟⲩⲓ ⲙⲛ ⲧⲁⲣⲝ ⲛⲧⲟϥ ⲉⲧⲛⲁⲛⲁϩⲙⲉⲧ ⲉⲧϭⲓϫ ⲙⲡⲓⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲇⲁⲩⲉⲓⲇ· ϫⲉ ⲃⲱⲕ ⲉⲃⲟⲗ ⲉⲣⲟϥ ⲉⲣⲉ ⲡϫⲟⲉⲓⲥ ϣⲱⲡⲉ ⲛⲙ̄ⲙⲁⲕ.

In SaMA, the first words parallel the Hexaplaric approximation καὶ εἶπε Δαυίδ attested by A O L CI-243txt CII 509 d s-64 381 55 71 245s 158 554, whereas SaB does not attest to this approximation. The omission of SaB began already in the previous verse, and therefore, it does not display a more original Sahidic text. The constructions are significantly different in SaAM. SaA uses an ⲛϭⲓ-construction and seems to have an equivalent for παλιν (ⲁϥⲕⲧⲟϥ ⲟⲛ ‘he turned again’) like Greek manuscript 509. Both SaM and SaA resemble L 55 158 554 ‘from the mouth of a lion’ but with different Sahidic prepositions. This same variant is also present in Theodotion, demonstrating that Sahidic and Lucianic texts had common

179

Textual analyses

sources from which they have acquired some variants. The expression ‘from the hand of’ in the next clause is ⲛⲧⲟⲟⲧϥ ̄ in SaM but ⲉⲧⲟⲟⲧϥ ̄ in SaA. These differences show that SaAM have independently formulated the details of their texts. SaB has its own abbreviated wording, ‘saved from lion and bear’. It uses another word for ‘hand’, ϭⲓϫ, and employs a relative construction without a definite article against SaAM. Some Greek manuscripts (19 f) read like SaB without ‘uncircumcised’. However, SaB is considerably free in this verse, omitting so many words without any Greek support that one could hardly posit a connection based on the omission of this one word. 17: 38 καὶ ἐνέδυσεν Σαοὺλ τὸν Δαυεὶδ μανδύαν καὶ περικεφαλαίαν χαλκῆν περὶ τὴν κεφαλὴν αὐτοῦ, περί] pr εθηκεν 799; επι A 799 d; επεθηκεν επι L 158 554; εθηκεν επι V αὐτοῦ] + και ενεδυσεν αυτον θωρακα L 158 318 554mg

SaM

SaA

SaB

ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥϯ ϩⲓⲱⲱϥ ⲛ̄ⲇ̄ⲁⲇ ̄ ̄ · ⲛⲟⲩϣⲧⲏⲛ ⲛ̄ⲣⲃ̄ ⲧ ̄ ·̄ ⲁⲩⲱ ⲟⲩⲡⲉⲣⲓⲕⲉⲫⲁⲗⲁ̄ⲓ ̅ⲁ ⲛϩⲟⲙⲛⲧ· ⲁϥⲧⲁⲁⲥ ⲉϫⲛ̄ ⲧⲉϥⲁⲡⲉ· ⲁⲩⲱ ⲁϥϯ ϩⲓⲱⲱϥ ⲛⲟⲩϩⲱⲕ·

ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥϯ ϩⲓⲱ[ⲱϥ] ⲛⲇ̅ⲁⲇ ̅ ̅ ⲛⲟⲩϣⲧⲏⲛ̣ ⲉⲛⲉⲣⲃ̣ⲏⲧ· ⲁⲩⲱ [ⲟⲩ]ⲡⲉⲣⲓⲕⲉⲫⲁⲗⲁⲓⲁ ⲛ̇ϩⲟⲙⲛ̅ⲧ ⲁϥⲧⲁⲁⲥ ϩⲓ ̣ϫⲛ̇ ⲧϥ̇ⲁⲡⲉ· ⲁⲩⲱ ⲁ[ϥ]ϯϩⲓⲱⲱϥ ⲛⲟⲩϩⲱ[ⲕ]

ⲁ ⲥⲁⲟⲩⲗ ϯ ϩⲓⲱⲱϥ ⲛⲇⲁⲩⲉⲓⲇ ⲛⲟⲩϣⲧⲏⲛ ⲛⲉⲣⲃⲧ· ⲁϥⲙⲟⲣϥ ⲛⲟⲩϩⲱⲕ ⲟⲩⲡⲉⲣⲓⲕⲉⲫⲁⲗⲉⲁ ⲉϫⲉⲛ ⲧⲉϥⲁⲡⲉ·

In this verse, SaMA follow a Greek text that contains corrections towards the Hebrew. ⲁϥⲧⲁⲁⲥ ⲉϫⲛ̄ ⲧⲉϥⲁⲡⲉ ‘he put it on his head’ is a formally literal rendering of (ἐπ)έθηκεν ἐπὶ τὴν κεφαλὴν αὐτοῦ. The last clause καὶ ἐνέδυσεν αὐτὸν θώρακα (ⲁⲩⲱ ⲁϥϯ ϩⲓⲱⲱϥ ⲛⲟⲩϩⲱⲕ ‘And he put on him a breastplate’) is only present in L 158 318 554mg in Greek. SaB preserves the same reading as the other Sahidic witnesses, but afterwards follows some textual variation: The clause begins ⲁϥⲙⲟⲣϥ ⲛⲟⲩϩⲱⲕ ‘He girded a breastplate’ but the continuation ⲟⲩⲡⲉⲣⲓⲕⲉⲫⲁⲗⲉⲁ ⲉϫⲉⲛ ⲧⲉϥⲁⲡⲉ ‘a helmet on his head’ does not fit grammatically. Probably, this results from an attempt to formulate the text logically, first putting other equipment on David and the helmet last. The scribe changed the verb to ⲙⲟⲩⲣ ‘to gird’ but did not notice the need for at least a connective before ⲟⲩⲡⲉⲣⲓⲕⲉⲫⲁⲗⲉⲁ.

180

Affiliations of the Sahidic manuscripts and the Greek Vorlage

17: 39 καὶ ἔζωσεν τὸν Δαυεὶδ τὴν ῥομφαίαν αὐτοῦ ἐπάνω τοῦ μανδύου αὐτοῦ· καὶ ἐκοπίασεν περιπατήσας ἅπαξ καὶ δίς, ὅτι ἄπειρος ἦν. καὶ εἶπεν Δαυεὶδ πρὸς Σαούλ Οὐ μὴ δύνωμαι πορευθῆναι ἐν τούτοις, ὅτι οὐ πεπείραμαι· καὶ ἀφαιροῦσιν αὐτὰ ἀπ᾽αὐτοῦ. μανδύου] θώρακος L 158 318 554mg καὶ ἐκοπίασεν περιπατήσας] εχωλαινε δαδ εν τω βαδιζειν εν αυτοις L 158 554mg ἀφαιροῦσιν] περιείλατο L 158 554c

SaM

SaA

SaB

ⲁⲩⲱ ⲧⲉϥⲥⲏϥⲉ· ⲁϥⲙⲟ̄ⲣⲥ̄ ⲉ̄ϫⲛ̄ ⲡϩⲱⲕ ⲛⲇⲁⲩⲉⲓⲇ· ⲁϥϫⲱⲣⲁϩ ⲇⲉ ⲛϭ ̄ ⲓ ⲇⲁⲩⲉⲓⲇ· ⲛⲧ ̄ ⲉⲣⲉϥϫⲱⲛⲧ ⲉⲙⲟⲟϣⲉ ⲛ̄ϩⲏⲧⲟⲩ· ϫⲉ ⲛⲉϥϫⲟⲛⲧ ⲁⲛ ⲙⲙⲟⲟⲩ· ⲁⲩⲱ ⲛⲧⲉⲣⲉϥⲙⲟⲟϣⲉ ⲛϩⲏⲧⲟⲩ ⲛⲟⲩⲥⲟⲡ ⲁⲩⲱ ⲥⲛⲁⲩ ⲁϥϩⲓⲥⲉ· ⲡⲉϫⲉ ⲗⲁⲩⲉⲓⲇ (sic) ⲛ̄ⲥⲁⲟⲩⲗ· ϫⲉ ⲛϯⲛⲁⲉϣⲙⲟⲟϣⲉ ⲁⲛ· ⲉⲣⲉ ⲛⲁⲓ ϩⲓ̈ⲱⲱⲧ· ϫⲉ ⲛϯⲧⲏⲏⲡ ⲁⲛ ⲉⲣⲟⲟⲩ̄· ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁϥⲕⲁⲁϥ ⲕⲁϩⲏⲩ ⲙⲙⲟⲟⲩ·

ⲁⲩⲱ ⲧϥ̇ⲥⲏⲃⲉ ⲁϥⲙ̣ [ⲟ]ⲣⲉⲥ ⲉϫⲉⲙ ⲡϩⲱⲕ· ⲛⲇ̅ⲁⲇ ̅ ̅· ⲁϥϫⲟⲩϩⲉ ⲇⲉ̣ [ⲛϭ ̄ ]ⲓ ⲇ̅ⲁⲇ ̅ ̅ ⲛⲧⲉⲣϥϫⲱ[ⲛⲧ109 ⲉ]ⲙⲟⲟϣⲉ ⲛ̇ϩⲏⲧⲟⲩ̣ [ⲛⲉ]110ϥϫⲟⲛⲧ ⲁⲛ ⲙⲙⲟⲟⲩ ⲁⲩⲱ ⲛ̇ⲧⲉⲣϥ̇ⲙⲟⲟϣⲉ ⲛϩⲏⲧⲟⲩ ⲛⲟⲩⲥⲟⲡ ⲙⲛ ⲥⲛⲁⲩ ⲁϥϩⲓⲥ̣ⲉ ⲡⲉϫⲉ ⲇ̅ⲁⲇ ̅ ̅ ⲛ̇ⲥⲁⲟⲩⲗ ϫⲉ ⲛϯⲛⲁϣⲙⲟⲟϣⲉ ⲁⲛ ⲉⲣⲉ ⲛⲁⲓ ϩⲓⲱⲱⲧ111 ϫⲉ ⲛϯⲧⲏⲡ ⲉⲣⲟⲟⲩ ⲁⲛ· ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁϥⲕⲁⲁϥ ⲕⲁϩⲏⲩ ⲙ̇ ⲙⲟⲟⲩ·

ⲁⲩⲱ ⲧⲥⲏⲃⲉ ⲁϥⲙⲟⲣⲥ ⲉϫⲙ ⲡⲉϥϩⲱⲕ· ⲁϥϫⲟⲩϩⲉ ⲛϭⲓ ⲇⲁⲩⲉⲓⲇ ⲛⲧⲉⲣⲉϥⲙⲟⲟϣⲉ ⲛϩⲏⲧⲟⲩ ⲛⲟⲩⲥⲟⲡ ⲁⲩⲱ ⲥⲛⲁⲩ ⲁϥϩⲓⲥⲉ· ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ϫⲉ ⲛ̄ϯⲛⲁⲉϣⲙⲟⲟϣⲉ ⲁⲛ ⲉⲣⲉ ⲛⲁⲓ ϩⲓⲱⲱⲧ· ϥⲓⲧⲟⲩ ⲙ̄ⲙⲁⲩ· ⲁⲩⲱ ⲁ ⲇⲁⲩⲉⲓⲇ112 ⲕⲁⲁϥ ⲕⲁϩⲏⲩ ⲙⲙⲟⲟⲩ.

SaMA contain similar additions in this verse. They agree with L 158 318 554mg that the sword was girded on the breastplate.113 This is implicit in the narrative, since the previous verse ended with the breastplate in Sahidic (and L 158 318 554mg in Greek). These manuscripts attest the doublet ⲁϥϫⲱⲣⲁϩ ⲇⲉ ⲛϭ ̄ ⲓ ⲇⲁⲩⲉⲓⲇ· ⲛⲧ ̄ ⲉⲣⲉϥϫⲱⲛⲧ ⲉⲙⲟⲟϣⲉ ⲛϩ̄ ⲏⲧⲟⲩ ‘He limped, namely David, when he tried to move with those’ (καὶ ἐχώλαινε Δαυὶδ ἐν τῷ βαδίζειν ἐν αὐτοῖς), situated before the rendering of ὅτι ἄπειρος ἦν (ϫⲉ ⲛⲉϥϫⲟⲛⲧ ⲁⲛ ⲙⲙⲟⲟⲩ ‘for he was not used to those’). Afterwards, the rendering of καὶ ἐκοπίασεν περιπατήσας ἅπαξ καὶ δίς (ⲁⲩⲱ ⲛⲧⲉⲣⲉϥⲙⲟⲟϣⲉ ⲛϩⲏⲧⲟⲩ ⲛⲟⲩⲥⲟⲡ ⲁⲩⲱ ⲥⲛⲁⲩ ⲁϥϩⲓⲥⲉ ‘And when he moved with those once and 109 Wessely has a typographical error, the ms reads ⲛⲧⲉⲣϥ-, not ⲛⲧⲉⲣⲉϥ-. When it comes to the next word, Wessely has reconstructed [ⲙ]ⲙⲟⲟϣⲉ but I prefer [ⲉ]ⲙⲟⲟϣⲉ, following SaM. 110 Wessely has reconstructed ⲛ̄ϩⲏⲧⲟ[ⲩ ⲁ]ϥϫⲟⲛⲧ, but this reading appears too short. However, the reading of SaM has too many characters anyway to fit in the fragment. In the fragments present state, the ypsilon of ⲛ̄ϩⲏⲧⲟⲩ is partly visible, and thus the space for reconstruction is quite clearly defined. The space allows reconstructing for example ⲛ̄ϩⲏⲧⲟⲩ̣ [ⲉⲁ]ϥϫⲟⲛⲧ /ⲛ̄ϩⲏⲧⲟⲩ̣ [ⲛⲉ]ϥϫⲟⲛⲧ or ⲛ̄ϩⲏⲧⲟⲩ̣ [ϫⲉ]ϥϫⲟⲛⲧ. 111 This clause is a parade example of the cleaning that the fragment has undergone: Wessely had the following reconstructions ⲛϯⲛⲁ[ϣⲙ]ⲟⲟϣⲉ ⲁⲛ ⲉⲣⲉ ⲛⲁⲓ [ϩⲓⲱ]ⲱⲧ. In its present state, the fragment is easily legible in this point. 112 In the actual manuscript, it looks odd that ⲉ in this word is written with a capital letter, partly in the margin, as if starting a new word or paragraph. 113 θώρακος in L 158 318 554mg.

Textual analyses

181

twice, he fell’) follows.114 The phrase ὅτι ἄπειρος ἦν (ϫⲉ ⲛⲉϥϫⲟⲛⲧ ⲁⲛ ⲙⲙⲟⲟⲩ ‘for he was not used to those’) in SaMA lies before ἅπαξ καὶ δίς,115 probably because the addition of the doublet necessitated a reformulation of the sentence. This shows that the Sahidic text does not render just one Greek text but combines material from at least two sources. During the transmission of the Sahidic text, secondary Greek readings influenced the translation as it was partly redacted against later Greek manuscripts. The last clause in Sahidic contains the proper name David against all the other witnesses. The use of a 3rd person singular could indicate a connection to a text similar to that in L 158 554c. Because the addition of David shows a certain freedom in translation in this clause, the Sahidic is not reliable for such minutia.116 Both Greek verbs, ἀφαιρέω and περιαιρέω are rendered with the same verb in Sahidic.117 SaB is shorter than SaMA. In the first clause, it offers ⲡⲉϥϩⲱⲕ ‘his breastplate’ without mentioning David. SaMA read in the first clause ⲡϩⲱⲕ ⲛⲇⲁⲩⲉⲓⲇ and, accordingly, the subject cannot be David but must be Saul, as is the case in the OG also. In SaB, there is no equivalent for ἐν τῷ βαδίζειν ἐν αὐτοῖς … ὅτι ἄπειρος ἦν (ⲛⲧ ̄ ⲉⲣⲉϥϫⲱⲛⲧ ⲉⲙⲟⲟϣⲉ ⲛϩ̄ ⲏⲧⲟⲩ· ϫⲉ ⲛⲉϥϫⲟⲛⲧ ⲁⲛ ⲙⲙⲟⲟⲩ). It could have been left out on purpose but a homoiarchon mistake from ⲛⲧ ̄ ⲉⲣⲉϥϫⲱⲛⲧ ‘when he tried’ to ⲛⲧⲉⲣⲉϥⲙⲟⲟϣⲉ ‘when he moved’ is more likely. This could explain why SaB has awkwardly omitted a part of the text. In other passages, SaB abbreviates and freely narrates, but not as artlessly as here; after ⲁϥϫⲟⲩϩⲉ ⲛϭ ̄ ⲓ ⲇⲁⲩⲉⲓⲇ ‘He limped, namely David’, one would expect continuation. The clause as it stands now is awkward, a torso. SaB reads ϥⲓⲧⲟⲩ ⲙ̄ⲙⲁⲩ ‘Take these off’ where other manuscripts read ϫⲉ ⲛϯⲧⲏⲏⲡ ⲉⲣⲟⲟⲩ̄ ⲁⲛ ‘since I am not used to those’. Possibly ϥⲓⲧⲟⲩ ⲙ̄ⲙⲁⲩ results from a doublet or a marginal note concerning two verbs (both meaning ‘to take away, remove’) found in Greek manuscript tradition: ἀφαιρέω and περιαιρέω.

114 Bohairic reads ⲟⲩⲟϩ ⲁϥϧⲟⲥⲓ ⲉϥⲙⲟϣⲓ ⲛⲟⲩⲥⲟⲡ ⲛⲉⲙ ⲥⲛⲁⲩ ⲉϥⲕⲉϩ ⲉⲣⲱⲟⲩ ⲁⲛ thus rendering the OG accurately without the additions found in the Sahidic manuscripts. 115 V L C’ a 121mg f s 29 55 71 158 244 245 318 460 554mg 707 put it after ἅπαξ καὶ δίς. 116 ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁϥⲕⲁⲁϥ ⲕⲁϩⲏⲩ ⲙⲙⲟⲟⲩ would be in the 3rd person plural ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁⲩⲕⲁⲁϥ ⲕⲁϩⲏⲩ ⲙⲙⲟⲟⲩ. The latter clause reads the expression as passive ‘David, he was stripped off from those’. 117 Crum, Dictionary, 101a.

182

Affiliations of the Sahidic manuscripts and the Greek Vorlage

17: 40 καὶ ἔλαβεν τὴν βακτηρίαν αὐτοῦ ἐν τῇ χειρὶ αὐτοῦ, καὶ ἐξελέξατο ἑαυτῷ πέντε λίθους λείους ἐκ τοῦ χειμάρρου καὶ ἔθετο αὐτοὺς ἐν τῷ καδίῳ τῷ ποιμενικῷ τῷ ὄντι αὐτῷ εἰς συλλογὴν, καὶ σφενδόνην αὐτοῦ ἐν τῇ χειρὶ αὐτοῦ· καὶ προσῆλθεν πρὸς τὸν ἄνδρα τὸν ἀλλόφυλον. SaM

SaA

SaB

ⲁϥϫⲓ ⲙⲡⲉϥϭⲉⲣⲱϥ ⲉϩⲣⲁⲓ ⲉⲛⲉϥϭⲓϫ· ⲁⲩⲱ ⲁϥⲥⲱⲧⲡ ⲛⲁϥ ⲛϯⲟⲩ ⲛⲱⲛⲉ ϩⲙ̄ ⲡⲓ̈ⲁ· ⲉⲩⲟⲩⲟϫ̄ · ⲁϥⲕⲁⲁⲩ ⲉϩⲣⲁⲓ ϩⲙ̄ ⲡⲉϥϫⲓⲗⲗⲏⲥ̄· ⲉϣⲁϥⲙⲟⲟⲛⲉ ⲛϩⲏⲧϥ̄ ⲉϥϣⲟⲟⲡ ⲛⲁϥ· ⲉⲣⲉ ⲧⲉϥⲥ̄ⲫⲉⲛⲧⲱⲛⲉ ϩⲛ̄ ⲧⲉϥϭⲓϫ· ⲁϥⲙⲟⲟϣⲉ ⲉⲃⲟⲗ ⲉ̅ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ̄·

ⲁϥϫⲓ ⲙⲡⲉϥϭⲉⲣⲱϥ ⲉⲧϥϭⲓϫ· ⲁⲩⲱ ⲁϥⲥⲱⲧⲡ ⲛⲁϥ ⲛϯⲟⲩ ⲛⲱⲛⲉ ϩⲙ ⲡⲉⲓⲁ ⲉⲩⲟⲩⲟϫ· ⲁϥⲕⲁⲁⲩ ⲉϩⲣⲁⲓ [ⲉⲧ]118ϥϫⲉⲗⲗⲏⲥ· ⲉϣⲁϥ[ⲙⲟⲟ]119ⲛⲉ ⲛ̇ϩⲏⲧϥ̇· ⲉⲣⲉ [ⲧ]ϥⲥⲫⲉⲛⲧⲟⲛⲏ ϩⲛ̇ [ⲧ]ϥϭⲓϫ120· ⲁϥⲙⲟⲟ[ϣ]ⲉ ⲉⲃⲟⲗ ⲉⲡⲁⲗⲗⲟⲫⲩ[ⲗⲟ]ⲥ

ⲁϥϫⲓ ⲙⲡⲉϥϭⲉⲣⲱϥ ⲉϩⲣⲁⲓ· ⲉⲛⲉϥϭⲓϫ· ⲁϥⲥⲱⲧⲡ ⲛⲁⲩ ⲛϯⲟⲩ ⲛⲱⲛⲉ ⲉⲩⲟⲩⲱϭ ⲉⲩⲥⲉϩⲥⲱϩ· ⲁϥⲧⲁⲁⲩ ⲉϩⲣⲁⲓ ⲉⲡⲉϥϫⲓⲗⲗⲉⲥ· ⲁⲩⲱ ⲧⲉϥⲥⲫⲉⲛⲇⲟⲛⲏ· ⲁϥⲙⲟⲟϣⲉ ⲉϩⲟⲩⲛ ⲉⲡⲁⲗⲗⲟⲫⲓⲗⲟⲥ.

In the first clause, SaMB have ‘the hands’ in the plural against the unanimous testimony of other witnesses, both Coptic and Greek. The Sahidic manuscripts often disagree on the number of personal nouns such as ϭⲓϫ ‘hand’ without parallel variation in the Greek tradition. The place where the stones were gathered is rendered with ⲡⲓ̈ⲁ̄ ‘valley’ in SaMA but omitted from SaB. SaMA also share a similar word order with the location before the adjective. In the second clause, the adjective λείους ‘smooth’ has a variant τελείους ‘perfect’ in Greek B. SaMA have ⲉⲩⲟⲩⲟϫ̄ ‘be whole, safe, sound’ as the equivalent for τελείους.121 Thus, SaMA seem to presuppose the reading of B. Interestingly, SaB has a doublet, first ⲟⲩⲱϭ, an alternate form of ⲟⲩⲟϫ̄, found in SaMA and after that ⲥⲉϩⲥⲱϩ ‘smooth’ that is used to render λεῖος.122 Sahidic witnesses differ in their renderings of καὶ ἔθετο αὐτοὺς ἐν τῷ καδίῳ τῷ ποιμενικῷ τῷ ὄντι αὐτῷ εἰς συλλογὴν. SaB has the shortest form ‘he put them in his pouch’, SaA reads ‘he put them in his pouch with which he used to shepherd’, and SaM reads ‘he put them in his pouch with which he used to shepherd, which he had with him’. There are no Greek variants corresponding to these different readings and, therefore, I regard this as 118 This is Wessely’s reconstruction, and it seems fitting. 119 Here, I follow SaM against Wessely, who reads ⲉϣⲁϥ[ϯⲱ]ⲛⲉ with half of the omega inside square brackets. Both readings fit the available space exactly, but I prefer an existing reading to a mere conjecture. 120 Wessely has reconstructed [ⲧⲉ]ϥϭⲓϫ but in the beginning of this verse the same word is spelled ⲧϥϭⲓϫ. Therefore, I have left the ⲉ out from the reconstruction. 121 Crum, Dictionary, 334a, 511b. In the Sahidic NT, this word is used to render ἵλεως, ὁλόκληρος, ὑγιαίνειν and ὑγιής (Wilmet, Concordance II: 2, 1128). 122 Crum, Dictionary, 386b.

183

Textual analyses

inner-Sahidic variation, SaA and SaB abbreviate the text as in several other passages. SaB is corrupt in the next sentence. ⲁⲩⲱ ⲧⲉϥⲥⲫⲉⲛⲇⲟⲛⲏ ‘and his sling’ does not have a verbal counterpart—something has been lost. All Sahidic witnesses render τὸν ἄνδρα τὸν ἀλλόφυλον with ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ‘the Philistine’. 17: 41123 καὶ ἐπορεύθη ὁ ἀλλόφυλος πορευόμενος καὶ ἐγγίζων πρὸς Δαυίδ καὶ ὁ ἀνὴρ ὁ αἴρων τὸν θυρεὸν αὐτοῦ ἔμπροσθεν αὐτοῦ

omit B V a b f 64–381 29 55 71 244 245 460 707 SaM

SaA

SaB

ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϩⲱⲱϥ· ⲁϥⲙⲟⲟϣⲉ ⲉϩⲟⲩⲛ ⲉⲇⲁⲩⲉⲓⲇ· ⲉϥϣⲟⲩϣⲟⲩ ⲙⲙⲟϥ· ⲉⲣⲉ ⲡⲉⲧϥⲓ ⲙⲡⲉϥⲑⲩⲣⲟⲛ ϩⲓϩⲏ ⲙⲙⲟϥ·

ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ [ⲇⲉ]·124 ⲁϥⲙⲟⲟϣⲉ ⲉϩⲟⲩⲛ [ⲉⲣⲟ]ϥ· ⲉϥϣⲟⲩϣⲟⲩ ⲙⲙⲟϥ·

The whole verse is missing.

This verse is secondary, it did not belong to the original LXX. Its Hexaplaric origin in Greek manuscripts is clear, since manuscript 127 has preserved the asterisk sign ※. SaB does not contain this verse. SaM agrees with the text found in A O L C’ d s-64 381 158 318 554, and SaA reads the first part like SaM, but does not have the second part ‘his shield-bearer before him’. SaMA descend from a common ancestor, since they share a mistranslation; ἐγγίζων is rendered ⲉϥϣⲟⲩϣⲟⲩ ⲙⲙⲟϥ ‘vaunting himself’. This reading might stem from a misread participle of ἐγκαυχάομαι ‘to boast, be proud’ instead of ἐγγίζομαι ‘come near, approach’. The former is often rendered with ϣⲟⲩϣⲟⲩ in Sahidic.125

123 This verse, secondary in the LXX, is not present in B V a b f 64’ 29 55 71 244 245 460 707. 124 The reading with ⲇⲉ is based on a reconstruction, but it seems reasonable. In the line following the next line, the same construction ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ is extant. 125 Crum, Dictionary, 604a. In 1 Sam 2: 10, there are several occurrences of καυχάομαι rendered with the verb ϣⲟⲩϣⲟⲩ.

184

Affiliations of the Sahidic manuscripts and the Greek Vorlage

17: 42 καὶ εἶδεν Γολιὰδ τὸν Δαυεὶδ καὶ ἠτίμασεν αὐτόν, ὅτι αὐτὸς ἦν παιδάριον καὶ αὐτὸς πυρράκης μετὰ κάλλους ὀφθαλμῶν. init] pr καὶ ἐπέβλεψεν ὁ ἀλλόφυλος A O L C’ s-64 381 158 318 (ὁ ἀλλόφυλος > CII); καὶ ἐπέβλεψεν Γολιὰθ ὁ ἀλλόφυλος καὶ εἶδεν τὸν… d 554 SaM

SaA

SaB

ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁϥϭⲱϣⲧ·̄ ⲁϥⲛⲁⲩ ⲉⲇⲁⲩⲉⲓⲇ· ⲁϥⲕⲱⲙϣ̄ ⲛⲥⲱϥ̄ ϫⲉ ⲛⲉⲩϣⲏⲣⲉ ϣⲏⲙ ⲡⲉ· ⲉϥⲟ ⲙⲙⲣϣ ⲉⲣⲟϥ· ̄ ̄ · ⲛⲥⲁⲉ̄ⲓⲉⲃⲁⲗ ̅

ⲡ̣ⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁϥϭⲱϣ̇ ⲧ ⲁϥⲛⲁⲩ ⲉⲇ̅ⲁⲇ ̅ ̅ ⲁϥⲕⲱⲙ̇ ϣ ⲛ̣̇ⲥⲱϥ ϫⲉ ⲛⲉⲟⲩϣⲏⲣⲉ ϣⲏⲙ ⲡⲉ ⲙⲙⲉⲣⲓϣ ⲉϥⲟ ⲛⲥⲁⲓⲉⲃⲁⲗ

The whole verse is missing.

SaB does not have this verse. SaMA attest to the approximation according to Hebrew καὶ ἐπέβλεψεν ὁ ἀλλόφυλος since they have two verbs that both mean ‘to see’. However, SaMA have formulated the sentence so that the name Goliath is not mentioned, but only ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ at the beginning. Thus, SaMA are not directly dependent on the variant found in any Greek manuscript, but instead agree with the MT more closely. Either this is a reformulation in Greek or Sahidic manuscripts, or the source behind these approximations was different. 17: 43 καὶ εἶπεν ὁ ἀλλόφυλος πρὸς Δαυείδ Ὡσεὶ κύων ἐγώ εἰμι, ὅτι σὺ ἔρχῃ ἐπ᾽ἐμὲ ἐν ῥάβδῳ καὶ λίθοις; καὶ εἶπεν Δαυείδ Οὐχί, ἀλλ᾽ἢ χείρω κυνός. καὶ κατηράσατο ὁ ἀλλόφυλος τὸν Δαυεὶδ ἐν τοῖς θεοῖς αὐτοῦ. Ὡσεί] μή L 318 554c; ω συ O καὶ εἶπεν Δαυείδ Οὐχί, ἀλλ᾽ἢ χείρω κυνός] om A O L

SaM

SaA

SaB

ⲁⲩⲱ ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲡⲉϫⲁϥ ⲉϩⲟⲩⲛ ϩⲛ̄ ⲇⲁⲩⲉⲓⲇ· ϫⲉ ⲉⲓⲟ̅ ⲁⲛⲟⲕ ⲛ̄ⲑⲉ ⲛⲟⲩⲟⲩϩⲟⲣ· ϫⲉ ⲛⲧⲟⲕ ⲉⲕⲛⲏⲩ ⲉⲃⲟⲗ ⲉⲣⲟϥ· ϩⲛ̄ ⲟⲩϭⲉⲣⲱⲃ· ⲙⲛ̄ ϩⲉⲛⲱⲛⲉ· ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲁϥ̄ ϫⲉ ⲁⲣⲏⲩ ̄ ⲕϩⲟⲟⲩ ⲉ̄ⲡⲉⲩϩⲟⲣ· ⲁⲩⲱ ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁϥⲥⲁϩⲟⲩ ⲛ̄ⲇⲁⲩⲉⲓⲇ ϩⲛ̄ ⲛⲉϥⲛⲟⲩⲧⲉ̄·

ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲡⲉϫⲁϥ ⲛ̇ⲇ̅ⲁⲇ ̅ ̅ ϫⲉ ⲉⲓ[ⲟ]126 ⲁⲛⲟⲕ ⲛ̇ⲑⲉ [ⲛⲟ]ⲩⲟⲩ̇[ϩⲟ]ⲟⲣ· ϫⲉ ⲁⲕⲉⲓ ⲛ̇ⲧ[ⲟⲕ] ⲉⲃⲟⲗ ⲉⲣⲟⲓ ⲙⲛ̇ ⲟⲩϭ̣ [ⲉ]ⲣⲱⲃ ⲙⲛ̇ ϩⲛ̇ ⲱⲛ[ⲉ] ⲡⲉϫⲉ ⲇ̅ⲁⲇ ̅ ̅ ⲛⲁϥ ϫ[ⲉ] ⲁⲣⲏⲩ ⲛ̇ⲧⲟⲕ ⲕϩ[ⲟⲩ] ⲉⲛ̇ϩⲟ̇ ⲩⲟ ⲉⲡⲉⲩϩⲟ[ⲟⲣ] ⲁⲩⲱ ⲡⲁⲗⲗⲟⲫⲩⲗ[ⲟⲥ] ⲁϥⲥⲁϩⲟⲩ ⲛ̄ⲇ̅ⲁⲇ ̅ ̅ ϩⲛ̣ ⲛⲉϥⲛⲟⲩⲧⲉ·

ⲡⲉϫⲉ ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲇⲁⲩⲉⲓⲇ ϫⲉ ⲙⲏ ⲁⲛⲟⲕ ⲟⲩⲟⲩϩⲟⲟⲣ ⲉⲕⲛⲏⲩ ⲉⲃⲟⲗ ⲉⲣⲟⲓ ⲙⲛ ⲡⲉⲓ̈ϭⲉⲣⲱⲃ ⲙⲛ ⲛⲉⲓⲱⲛⲉ· ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲁϥ̄ ϫⲉ ⲕϩⲟⲟⲩ ⲙⲡⲁⲣⲁ ⲡⲉⲩϩⲟⲟⲣ· ⲁⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲥⲁϩⲟⲩ ⲛⲇⲁⲩⲉⲓⲇ

126 Wessely has reconstructed the word ⲉⲓ[ⲥ] but as well it can be ⲉⲓ[ⲟ] since the last character is broken exactly at the point where the difference between ⲉ and ⲥ lies. I prefer the existing reading of SaM since it also fits the context better.

Textual analyses

185

SaM renders formally the above Greek text. The only free formulation is ⲁⲣⲏⲩ̄ ‘perhaps, if’ for οὐχί ἀλλ᾽ἤ. Additionally, one finds ⲉⲣⲟϥ ‘to it’ instead of ⲉⲣⲟⲓ ‘to me’ in the third clause. SaA shares basically the same text with SaM but with some different minor elements. SaA has the particle ⲇⲉ, while SaM has ⲁⲩⲱ. SaA reads ⲡⲉϫⲁϥ ⲛⲇ ̄ ̅ ⲁⲇ ̅ ̅ ‘He said to David’ where SaM has ⲡⲉϫⲁϥ ⲉϩⲟⲩⲛ ϩⲛ̄ ⲇⲁⲩⲉⲓⲇ. Both of these expressions are used to render εἶπεν+πρός. Drescher mentions that Coptic versions of the New Testament and Pentateuch do not render εἶπεν+πρός with ⲉϩⲟⲩⲛ ϩⲛ̄ in a context like this.127 Therefore, if one of the Sahidic witnesses is secondary in this case, it is SaA, since this text uses the more common expression. Two other features demonstrate the secondary nature of SaA. First, ϩⲛ̄ ⲟⲩϭⲉⲣⲱⲃ ‘with a rod’ in SaM strictly renders ἐν ῥάβδῳ, whereas ⲙⲛ̄ ⲟⲩϭⲉⲣⲱⲃ ‘with a rod’ of SaA is understandable as a variant of ϩⲛ̄ ⲟⲩϭⲉⲣⲱⲃ. Second, ⲕϩⲟⲟⲩ ⲉⲛϩⲟⲩⲟ ‘you are much worse’ does not render any Greek variant, but is probably a corruption, or perhaps exaggeration, resulting from ⲕϩⲟⲟⲩ ⲉ̄ⲡⲉⲩϩⲟⲣ ‘you are worse than a dog’ of SaM. SaB reformulates the story by omitting conjunctions and other small words wherever possible (ⲁⲩⲱ/ⲇⲉ, ⲛⲑⲉ, ϫⲉ, ⲁⲣⲏⲩ)̄ . The resultant clauses are more straightforward in this manuscript: ⲡⲉϫⲉ ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲛⲇⲁⲩⲉⲓⲇ ‘said the Philistine to David’, ⲙⲏ ⲁⲛⲟⲕ ⲟⲩⲟⲩϩⲟⲟⲣ ‘Am I a dog?’, ⲕϩⲟⲟⲩ ⲙⲡⲁⲣⲁ ⲡⲉⲩϩⲟⲟⲣ ‘You are worse than a dog’. ⲙⲏ ⲁⲛⲟⲕ ⲟⲩⲟⲩϩⲟⲟⲣ at first seems to connect SaB to the reading of L 318 554c μὴ κύων εἰμὶ ἐγώ. However, it seems more plausible that ⲙⲏ is used in SaB because of its suitability in the context. This term arises frequently in direct discourse to mark a clause as a question,128 and thus makes the narrative easier to follow. In SaB, there is either a homoioteleuton mistake from David to the next David (in verse 17: 44) or a plain abbreviation. If one assumes a homoioteleuton-based parablepsis, then it is necessary to read the first clause in verse 17: 44 with David as seems to have been the case in SaA.129 All Coptic witnesses have an equivalent for καὶ εἶπεν Δαυιδ Οὐχί ἀλλ᾽ἢ χείρω κυνός and thus do not attest the Hexaplaric omission.

127 Drescher, Kingdoms, xvi. 128 Layton, Grammar, §511. 129 This fragment of SaA ends with the words ⲉϩⲟⲩⲛ ϩⲛ, but the next word probably would have been David.

186

Affiliations of the Sahidic manuscripts and the Greek Vorlage

17: 44 καὶ εἶπεν ὁ ἀλλόφυλος πρὸς Δαυείδ Δεῦρο πρός με καὶ δώσω τὰς σάρκας σου τοῖς πετεινοῖς τοῦ οὐρανοῦ καὶ τοῖς κτήνεσιν τῆς γῆς. κτήνεσιν] θηρίοις A V 376 L C’ a d s 29 55 158 244 460 554

SaM

SaA

ⲉⲣⲉ ⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ϫⲱ ⲙⲙⲟⲥ ⲉⲣⲟϥ· ϫⲉ ⲁⲙⲟⲩ ⲛⲁⲓ· ⲛⲧⲁϯ ⲛⲛⲉⲕⲁⲃⲟⲩⲓ̈ ⲛⲛⲉϩⲁⲗⲁⲧⲉ ⲛⲧⲡⲉ· ⲁⲩⲱ ⲛⲛⲉⲑⲩⲣⲓⲟⲛ ⲙ̄ⲡⲕⲁϩ

ⲉⲣⲉ ⲡⲁⲗⲗⲟⲫⲩⲗⲟ̣ⲥ ̣ ϫⲱ ⲙⲙⲟⲥ ⲉϩⲟⲩⲛ ϩⲛ […

SaB ϫⲉ ⲁⲙⲟⲩ ⲛⲁⲓ ⲛⲧⲁϯ ⲛⲛⲉⲕⲁϥ ⲛⲛⲉⲛϩⲁⲗⲁⲁⲧⲉ ⲛⲧⲡⲉ· ⲁⲩⲱ ⲛⲉⲕⲥⲁⲣⲝ ⲛⲛⲉⲑⲩⲣⲓⲟⲛ ⲙⲡⲕⲁϩ.

The fragment of SaA breaks in the middle of the first clause. SaB reads in the previous verse ⲁⲡⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲥⲁϩⲟⲩ ⲛⲇⲁⲩⲉⲓⲇ ‘The Philistine cursed David’ and this verse 17: 44 continues that sentence. Possibly, a homoioteleuton mistake from David in the previous verse to David here occurred in this verse. In the third clause, SaM reads the plural form of ‘flesh’ whereas SaB reads the article in the plural but the noun itself in the singular. Only a few Coptic nouns possess special plural forms, and this might explain errors relating to these terms. The next noun ‘birds’ is spelled correctly, but the article in SaB is ⲛⲉⲛ-, perhaps due to this manuscript’s Fayyumic birthplace. SaB adds ⲛⲉⲕⲥⲁⲣⲝ ‘your flesh’ before ⲛⲛⲉⲑⲩⲣⲓⲟⲛ ‘to the beasts’ without similar Greek variants. In the Greek tradition, the expression τοῖς κτήνεσιν ‘to the animals (domesticated animals)’ has the variant reading τοῖς θηρίοις ‘to the animals (beasts)’. SaMB have ⲑⲩⲣⲓⲟⲛ, probably rendering θηρίοις, since there is a tendency to render κτῆνος with ⲧⲃⲛⲏ ‘cattle, beast’ in the Sahidic New Testament as well as in Sahidic Jeremiah.130 Therefore, SaMB can be cited with those Greek manuscripts that read τοῖς θηρίοις. Further, Bohairic has ⲛⲓⲑⲏⲣⲓⲟⲛ ⲛⲧⲉ ⲡⲕⲁϩⲓ, exactly like SaMB. Usually, Bohairic follows Greek manuscript B in this chapter, but not in this case.131 Is the reading of B with κτήνεσιν secondary? This could be a correction towards Hebrew, attested by B 247 b f 71 245 318 707. This combination of manuscripts (B and its companions b, 247 from O-group, with some additional minuscules) is typical in those cases where early approximations are found.132 In this verse, the 130 Luke 10: 34, Acts 23: 24, 1 Cor 15: 39, Jer 28: 62, Jer 9: 9. 131 During the study, the Bohairic version was a side track, even though it did not end up in the actual study. Based on the preliminary remarks in 1 Sam 17, the Bohairic version seems to be a solid member of the b-group, together with the Ethiopic version. The Song of Hannah (1 Sam 2: 1–10) is preserved in two Bohairic manuscripts. These two texts and their relation to the Sahidic version of 1 Sam 2: 1–10 will be analyzed in my forthcoming article on the Bohairic Song of Hannah. 132 See Aejmelaeus, “Corruption or Correction”, 13; and eadem, “Kingdom at Stake”, 362.

Textual analyses

187

Hebrew tradition has ‫בהמה‬, which is usually rendered by κτῆνος, and this parallelism further supports the possibility of a Hebrew-based correction in the manuscripts that have κτήνεσιν. Results All three Sahidic witnesses share a common Sahidic ancestor. This is the simplest and most compelling way to explain the uniform wordings (e. g., 17: 33, 17: 34, 17: 36, 17: 40) with some variation in comparison with Greek. The translator has not rendered word-for-word, but has considered larger units, as οὐ μὴ δυνήσῃ πορευθῆναι πρὸς τὸν ἀλλόφυλον τοῦ πολεμεῖν μετ᾽αὐτοῦ which was translated as one unit in verse 14: 28, shows. The Hexaplaric approximations were already present in the Sahidic tradition before the (grand)parent133 manuscript of SaMA since in verse 17: 41 SaMA share an erroneous translation. There are several passages where SaMA agree against SaB. These agreements pertain to construction (17: 34, 17: 37, 17: 39), conjunction (17: 34, 17: 36), vocabulary (17: 35, 17: 37), and Hexaplaric approximation (17: 37, 17: 39). These agreements, together with SaB’s peculiar readings demonstrate that SaMA are nearer to the original Sahidic tradition, whereas SaB contains a large number of corruptions and emendations. In some passages, SaAB agree against SaM. These concern construction (17: 35), addition (17: 36), and conjunction (17: 37). One of these appears in an addition that is only present in Sahidic manuscripts, and another is an addition extant only in SaM. The third follows the shortening tendency of both SaA and SaB. In two passages, SaMB agree against SaA. These agreements concern verb (17: 35) and number (17: 40). The verb is secondary in SaA, having been emended to a more common one. The agreement in number concerns a case where SaMB read singular ‘hand’ against the plural ‘hands’ in SaA. Variation of number in hand/hands is a typical inner-Sahidic phenomenon, frequently witnessed in the manuscripts. In the discussed verses, SaA closely resembles the text of SaM; typically they use the same formulations. A typical feature of SaA is abbreviation (17: 40), and another is erroneous copying, as in verse 17: 43. However, when compared to the variants in chapter 31, one observes that SaA in chapter 17 more closely resembles the text of SaM than in chapter 31, i. e., there are fewer mistakes and abbreviations in chapter 17. Does this perhaps indicate that the scribe was more familiar with chapter 17? Or that this chapter was revised and complemented with some Hebrew-based corrections? Among the manuscripts studied here, SaB is peculiar because it is based on the same text but transmits it in a shorter (17: 37, 17: 39, 17: 40, 17: 43) and 133 I.e. the Sahidic manuscript from which SaMA descended.

188

Affiliations of the Sahidic manuscripts and the Greek Vorlage

often corrupted (17: 38, 17: 39, 17: 40) form. Typical features include omissions of conjunctions and grammatically simpler clause constructions (17: 39, 17: 43). In verse 17: 43, an example demonstrates how Sahidic manuscripts contain variant readings that have arisen during the transmission history (innerSahidic variation). Small changes involving conjunction/asyndeton occur occasionally in the manuscripts. The problem is when to postulate an affiliation between manuscripts based on these shared details (conjunction, construction). To what extent can one regard them as mere coincidences? In several passages, one cannot say which one of the Sahidic variants is more original, and this hinders decisions concerning the affiliations, since only shared secondary readings unambiguously show dependences. Chapter 17 contains dozens of secondary readings in Greek,134 and is a treasure trove for evaluating secondary features in the Sahidic text. These readings reveal the influence of the Greek tradition on the later transmission of the Sahidic text, as these later readings entered the Sahidic tradition. Therefore, the affiliations in these verses are of special significance. In 17: 37, there is a Hexaplaric approximation in SaMA with variation. Combined with the approximation, SaA agrees with 509 in one detail. If this detail is a later emendation, then it explains why SaA has also emended the Hexaplaric reading. If the Hexaplaric reading was inserted later, the scribe of SaM probably omitted the detail at the same time. Theoretically, SaM and SaA have acquired these variants independently, and this explains their differences. However, this assumption does not hold in other cases examined here. In 17: 38–39, SaMA preserve additions present only in L 158 318 554mg (=MT), and therefore some common source must lie behind these variants. In 17: 40, SaMA render the adjective τελείους like B in Greek against λείους in other Greek manuscripts.135 SaB reads a doublet. In verse 17: 41, which is on the whole secondary, SaAM share a mistranslation: ἐγγίζων is rendered ⲉϥϣⲟⲩϣⲟⲩ ⲙⲙⲟϥ ‘vaunting himself’. This shows that a Sahidic manuscript existed in the transmission history that had this alternate reading since this kind of error would probably not occur independently in these Sahidic manuscripts. SaB does not have this verse, but it also lacks the next one. Therefore, one cannot conclude that its Vorlage was free from this Hexaplaric variant and hence was more original. In 17: 42, SaMA contain an emended approximation. As such, the Sahidic text agrees more closely with the MT than the Greek text. Either the Sahidic scribe emended the text so that the name Goliath is not there (thus avoiding

134 The principal source for these secondary readings is the existence of different Hebrew texts of this passage (David and Goliath) that were read and copied (Ulrich, The Dead Sea Scrolls and the Origins of the Bible, 36–7). Accordingly, there came to be Greek texts rendered from one Vorlage and complemented according to another Hebrew source. 135 Could this be Old Greek or only a mistake in B?

189

Textual analyses

redundancy), or the Greek source text lacked the proper name.136 In 17: 43, SaB perhaps follows the reading now found in L 318 554c μὴ κύων εἰμὶ ἐγώ. The text of SaB, however, is not formally literal in this passage, and this variant could also stem from the copyist without any Greek source. In the same verse, all Sahidic witnesses have preserved a passage that Greek A O L have omitted based on Hexaplaric influence.137 In 17: 44, SaMB apparently render the noun θηρίοις like A V 376 L C’ a d s 29 55 158 244 460 554, thus preserving the OG against the B-group and a few minuscules that contain an approximation.138 In conclusion, the Sahidic manuscripts preserve agreements with the Btext as well as with Greek manuscripts corrected towards the MT. Perhaps, the source text of the Sahidic translator was similar to B in its early form, without those καιγε-type corrections now attested by B. After the initial translation, some additions were assimilated into the Sahidic tradition. This happened before the textual lines of SaM and SaA diverged. The Greek affiliations of SaB are difficult to discern because it has such a special character: abbreviations, its own wordings, and omissions without any source text. At least some of the emendations, however, are attested also by SaA. Several of the Hebrew-based corrections also appear in the Lucianic manuscripts. Other secondary features of the Lucianic text are not present in Sahidic, and therefore a direct connection is an invalid explanation. Probably, the Sahidic and Lucianic traditions acquired their Hexaplaric-type corrections from common sources. 3.2.5. SaMAFJ 29: 5–9 29: 5 οὐχ οὗτος Δαυεὶδ ᾧ ἐξῆρχον ἐν χοροῖς λέγοντες Ἐπάταξεν Σαοὺλ ἐν χιλιάσιν αὐτοῦ, καὶ Δαυεὶδ ἐν μυριάσιν αὐτοῦ; ἐξῆρχον] εξηρχοντο 121 d 244 245 554txt 707; εξηλθον V ἐν χοροῖς λέγοντες] αἱ χορεύουσαι λέγουσαι L 121 318 554mg

SaM

SaA

SaF 139

ⲙⲏ ⲙⲡⲁⲓ ⲁⲛ ⲡⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲧⲁⲩⲙⲟⲟϣⲉ ϩⲓϩⲏ ⲙⲙⲟϥ ⲛϭⲓ ⲛⲉⲧⲭⲟⲣⲉⲩⲉ̄ ⲉⲩϫⲱ

ⲙⲏ ⲉⲙⲡⲁⲓ ⲁⲛ ⲡⲉ ⲇ̅ⲁⲇ ̅ ̅ ⲡⲉⲛⲧⲁⲩϩⲱⲥ ϩⲓⲑⲏ ⲙ̄ⲙⲟϥ ⲛϭ ̄ ⲓ ⲛⲉⲧⲭⲟⲣⲉⲩⲉ ⲉⲩϫⲱ

…[ⲉⲩ]ϫⲱ

SaJ

136 Perhaps it is too far-fetched to suppose a Greek text that was copied according to the Hexaplaric signs, and thus did not contain the proper name found in Lucianic manuscripts. 137 Perhaps, the Hexaplaric signs were not understood or were ignored by the Sahidic copyists. 138 The fragment of SaA breaks before this reading. 139 I have used the edition of Kahle (Bala’izah) and also checked the manuscript myself.

190

Affiliations of the Sahidic manuscripts and the Greek Vorlage

ⲙⲙⲟⲥ· ϫⲉ ⲥⲁⲟⲩⲗ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ϩⲛ̄ ⲛⲉϥϣⲟ· ⲇ̅ⲁⲇ ̅ ̅ ϩⲱⲱϥ ϩⲛ ⲛⲉϥⲧⲃⲁ·

ⲙ̄ⲙⲟⲥ ϫⲉ ⲥⲁⲟⲩⲗ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ϩⲛ̄ ⲛϥ̄ϣⲟ ⲇ̅ⲁⲇ ̅ ̅ ϩⲱⲱϥ ϩⲛ̄ ⲛϥ̄ⲧⲃⲁ.

[ⲙⲙⲟⲥ· ϫⲉ ⲥ]ⲁ[ⲟⲩⲗ] ⲁϥⲡⲁⲧⲁⲥ[ⲥⲉ ϩ]ⲛ̄ ⲛⲉϥϣ[ⲟ] [ⲇⲁⲩⲉⲓⲇ140 ϩⲱ]ⲱ[ϥ ϩⲛ ⲛⲉϥⲧⲃⲁ]

[..]ⲛ[..] [..ϩ]ⲱⲱϥ ϩⲛ ̣ ̣̄ ⲛⲉ[ϥⲧⲃⲁ]

In this verse, one remarkable difference distinguishes the Sahidic witnesses. The verb used in SaM in the second clause is ⲙⲟⲟϣⲉ ‘to walk, go’, an erroneous rendering of the underlying Greek text. The translator has rendered as if the verb were ἐξέρχομαι ‘come out’. Actually, the verb is ἐξάρχω ‘to begin (a hymn)’ and it is corrected in SaA, where the verb is ϩⲱⲥ ‘to sing’. Notably, the SaA variant corresponds to the Hebrew more closely, since the verb in Hebrew is ‘to sing’. Perhaps, this is only a copying mistake in SaM since the verb ‘to come out’ fits the context very well and is a common verb. Both SaM and SaA render the next phrase so that there is a formally literal rendering of αἱ χορεύουσαι as found in L 121 318 554mg.141 29: 6 καὶ ἐκάλεσεν Ἀγχοὺς τὸν Δαυεὶδ καὶ εἶπεν αὐτῷ Ζῇ Κύριος, ὅτι εὐθὴς σὺ καὶ ἀγαθὸς ἐν ὀφθαλμοῖς μου, καὶ ἡ ἔξοδός σου καὶ ἡ εἴσοδός σου μετ᾽ἐμοῦ ἐν τῇ παρεμβολῇ· καὶ ὅτι οὐχ εὕρηκα κατὰ σοῦ κακίαν ἀφ᾽ἧς ἡμέρας ἥκεις πρός με ἕως τῆς σήμερον ἡμέρας, καὶ ἐν ὀφθαλμοῖς τῶν σατραπῶν οὐκ ἀγαθὸς σύ. ὅτι 2º] om. L

SaM

SaA

SaF

SaJ

ⲁⲭⲟⲩⲥ142 ⲇⲉ ⲁϥⲙⲟⲩⲧⲉ ⲉⲇⲁⲩⲉⲓⲇ143 ⲡⲉϫⲁϥ ⲛⲁϥ· ϫⲉ ⲡϫⲟⲉⲓⲥ ⲟⲛϩ̄· ϫⲉ

ⲁⲭⲟⲩⲥ ⲇⲉ ⲁϥⲙⲟⲩⲧⲉ ⲉⲇ̄ⲁⲇ ̄ ̄ ⲡⲉϫⲁϥ ⲛⲁϥ ϫⲉ ⲡϫⲟⲉⲓⲥ ⲟⲛϩ̄ ⲁⲩⲱ

ⲁ[ⲅ]ⲭⲟ[ⲩⲥ ⲇⲉ] ⲁϥⲙⲟ[ⲩⲧⲉ ⲉ]ⲇⲁⲩⲉⲓⲇ [ⲡ]ⲉϫⲁϥ ⲛⲁϥ· ϫⲉ ⲡϫⲟⲉⲓⲥ ⲟⲛ[ϩ̄] ϫⲉ

ⲁⲅⲭⲟⲩⲥ ⲇⲉ ⲁ146ϥⲙ̣ [ⲟⲩⲧⲉ ⲉ]ⲇⲁⲩⲉⲓ ̣ⲇ̣ ⲡⲉϫ̣ ⲁ̣ϥ [ⲛ]ⲁ̣ϥ̣ ϫⲉ [ⲡϫⲟ]ⲉⲓⲥ ⲟⲛϩ ϫⲉ

140 Kahle (Bala’izah, 313) reconstructs here ⲇⲁⲩⲉⲓⲇ ⲇⲉ. I have omitted ⲇⲉ because it is not found in other Sahidic witnesses. 141 Perhaps Vaticanus attests a Hebraizing correction, ἐν as a rendering of ‫ב‬. 142 The spelling of the king’s name is ⲁⲭⲟⲩⲥ in SaAM but ⲁⲅⲭⲟⲩⲥ in SaFJ. Concerning the proper names, Sahidic manuscripts frequently differ, especially between dentals ⲧ ⲇ ⲑ and gutturals ⲕ ⲅ ⲭ ϭ. Thus, there is no reason to suppose any connection to Greek behind this difference. Brock (Recensions, 15) mentions that Coptic follows the secondary readings of the B-group especially in proper names. The fact that proper names so often have suffered from inner-Sahidic corruptions makes the argument less compelling. Among the lists of Payne, I have found one significant reading in 30: 21 where B b Sa Eth attest a secondary reading Βεανα. (Payne, Critical and comparative, 151). 143 The name David is in SaA in the abbreviated form ⲇ̅ⲁⲇ ̅ ̅ while SaMF attest the longer form ⲇⲁⲩⲉⲓⲇ. SaF has only long forms of David while SaM is inconsistent in the spelling of this name: it also has short forms, already in verses 9 and 11 in this very same chapter. SaA has only short forms in this fragment, but in the previous chapters long forms occur, for example, in verses 17: 45 and 18: 12. Short and long forms occur in the same verse (23: 9).

Textual analyses ⲕⲥⲟⲩⲧⲱⲛ ⲛⲧⲟⲕ ⲁⲩⲱ ⲛⲧ ̄ ⲕ ⲟⲩⲁⲅⲁⲑⲟⲥ ⲛⲁϩⲣⲁⲓ144· ⲁⲩⲱ ⲧⲉⲕϩⲓⲏ ⲛⲃⲱⲕ ⲉⲃⲟⲗ· ⲁⲩⲱ ⲧⲉⲕϩⲓⲏ̄ ⲛⲓ ⲉϩⲟⲩⲛ ⲛⲉⲥⲛⲙ̄ⲙⲁⲓ ϩⲛ̄ ⲧⲡⲁⲣⲉⲙⲃⲟⲗⲏ· ⲁⲩⲱ ϫⲉ ⲙⲡⲓϩⲉ ⲉϩⲱⲃ ⲉϥϩⲟⲟⲩ ⲉϩⲟⲩⲛ ⲉⲣⲟⲕ ·ⲛϫⲓⲛ ⲡⲉϩⲟⲟⲩ̣ ⲛⲧⲁⲕⲉⲓ ⲉϩⲟⲩⲛ ⲛⲁⲓ ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛϩⲟⲟⲩ· ⲛⲥⲁⲧⲣⲁⲡⲏⲥ ⲇⲉ ⲛⲥⲉⲧⲏⲧ ⲛϩ̄ ⲏⲧ ⲛⲙ̄ⲙⲁⲕ ⲁⲛ

ⲕⲥⲟⲩⲧⲱⲛ ⲛⲧⲟⲕ. ⲁⲩⲱ ⲛⲧ ̄ ⲕ ⲟⲩⲁⲅⲁⲑⲟⲥ ⲛⲛ ̄ ⲁϩⲣⲁⲓ. ⲁⲩⲱ ⲧⲕ̄ϩⲓⲏ ⲛⲉⲓ ⲉϩⲟⲩⲛ ⲙⲛ̄ ⲧⲉⲕϭⲓⲛⲓ ⲉⲃⲟⲗ ⲥⲥⲙⲁⲙⲁⲁⲧ. ⲁⲩⲱ ⲙ̄ⲡⲓϩⲉ ⲉⲩϩⲱϥ ⲉϩⲟⲩⲛ ⲉⲣⲟⲕ ϫⲓⲛ ⲉⲡⲉϩⲟⲟⲩ ⲛ̄ⲧⲁⲕⲓ ⲛⲁⲓ ⲉϩⲟⲩⲛ ϣⲁ ⲡⲟⲟⲩ ⲛ̄ϩⲟⲟⲩ. ⲛⲥⲁⲛⲇⲣⲁⲡⲏⲥ ⲛ̄ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲧⲏⲧ ⲛϩ̄ ⲏⲧ ⲛⲉⲙⲙⲁⲕ ⲁⲛ.

ⲕⲥⲟⲩ[ⲧ]ⲱⲛ ⲛⲧ ̄ ⲟⲕ· ⲁ[ⲩ]ⲱ ⲛⲧ ̄ ⲕ ⲟⲩⲁⲅⲁⲑⲟⲥ ⲛⲛ ̄ ⲁϩⲣⲁⲓ̈ ⲁⲩⲱ ⲧⲉⲕϩⲓⲏ ⲛ̄ⲃⲱⲕ ⲉⲃⲟⲗ· ⲁ[ⲩ]ⲱ ⲧⲉⲕϩⲓⲏ ⲛ̄ⲉ[ⲓ ⲉ]ϩⲟⲩⲛ ⲛⲉⲥ[ⲛⲙ̄]ⲙⲁⲓ̈ ϩⲛ̄ ⲧⲡ[ⲁⲣ]ⲉⲙⲃⲟⲗⲏ [ⲁ]ⲩⲱ ϫⲉ ⲙ̄ⲡⲓϩ[ⲉ ⲉϩⲱⲃ] ⲉϥϩⲟ[ⲟⲩ ⲉϩⲟⲩⲛ] ⲉⲣⲟ[ⲕ145 ⲛ̄ⲥⲁ]ⲧⲣ[ⲁⲡⲏⲥ ⲇⲉ ⲛ]ⲥ̣ⲉ[̣ ⲧⲏⲧ ⲛϩ̄ ⲏⲧ] ⲛⲙ̄ⲙⲁⲕ [ⲁⲛ]146

191 ⲕⲥⲟⲩⲧⲱⲛ ⲛⲧ ̄ ⲟⲕ ⲁⲩⲱ ⲛⲧ ̄ ⲕ̄ ⲟⲩⲁⲅⲁ[ⲑⲟ]ⲥ ⲛⲛ ̄ ⲁϩⲣⲁⲓ̈· ⲁⲩⲱ [ⲧⲉ]ⲕϩⲓ ̣ⲏ̣ ⲛ̄ⲃⲱⲕ ⲉ̣ⲃⲟⲗ [ⲁⲩⲱ ⲧⲉ]ⲕϩⲓ ̣ⲏ̣ [ⲛ]ⲉⲓ [ⲉϩⲟⲩⲛ ⲛⲉ]ⲥⲛ̄ [ⲙ]ⲙⲁ̣[ⲓ ϩⲛ̄ ⲧⲡⲁ]ⲣⲉ̣ ⲙⲃⲟ[ⲗⲏ ⲁⲩⲱ ϫⲉ ⲙ]ⲡⲓϩⲉ ⲉϩⲱⲃ ⲉϥϩⲟⲟ]ⲩ ⲉϩⲟⲩⲛ [ⲉⲣⲟⲕ ⲛϫⲓⲛ ⲡ]ⲉ̣ϩⲟⲟⲩ [ⲛⲧⲁⲕⲉⲓ ⲉϩ]ⲟⲩⲛ ⲛ̣[ⲁⲓ] […]ⲩⲟ[.]ⲛ̄[ϩ]ⲟ147 [ⲟⲩⲛⲥⲁⲧⲣⲁ]ⲡⲏⲥ ⲇⲉ [ⲛⲥⲉⲧⲏⲧ ⲛϩ̄ ⲏ]ⲧ ⲛⲙ ̄ ̄[ⲙⲁⲕ ⲁⲛ]

One significant difference arises with two expressions, namely, (1) ⲧⲉⲕϩⲓⲏ ⲛⲃⲱⲕ ⲉⲃⲟⲗ· ⲁⲩⲱ ⲧⲉⲕϩⲓⲏ̄ ⲛⲓ ⲉϩⲟⲩⲛ ‘your way of going out and your way of coming in’ in SaMFJ and the majority of Greek manuscripts, and (2) ⲧⲕϩ̄ ⲓⲏ ⲛⲉⲓ ⲉϩⲟⲩⲛ ⲙⲛ̄ ⲧⲉⲕϭⲓⲛⲓ ⲉⲃⲟⲗ ‘your way of coming in and your going out’ in SaA. In the latter expression, ⲧⲉⲕϭⲓⲛⲓ ⲉⲃⲟⲗ is formulated so that the word ‘way’ is not repeated as in other Sahidic witnesses. The same order as in SaA, first coming in and then going out, appears only in Greek 107 and the Ethiopic and Armenian versions. SaA also preserves the next word uniquely; SaMFJ render μετ᾽ἐμοῦ accurately by ⲛⲉⲥⲛⲙ̄ⲙⲁⲓ ‘with me’ but SaA makes it ⲥⲥⲙⲁⲙⲁⲁⲧ ‘blessed’ obviously confusing the letters. The reading of SaA is understandable only via the reading of SaMFJ, and, thus, it would be impossible to argue for the priority of

144 In the fourth line, SaM reads ⲛⲁϩⲣⲁⲓ where SaAFJ have ⲛ̄ⲛⲁϩⲣⲁⲓ. In this fragment, which contains verses 28: 16–30: 5, SaA always spells this word with double ⲛ. In the same verses, SaM also consistently reads with only one ⲛ. In SaF, the spelling is with double ⲛ. However, this is only a spelling difference and does not affect the meaning. According to Crum (Dictionary, 649b), the double ⲛ is found in older manuscripts. 145 Kahle (Balai´zah, 314) has three empty lines after this word. One could quite nicely reconstruct those lines according to the text of SaM. The reason this is not done here is the variant reading of SaA. The text of SaA is a bit too short to fit the space in SaF, but it would give the wrong impression of an agreement between SaMF if one reconstructs it. 146 The first ⲁ is much bigger than the other letters and written partly in the margin. 147 This line is impossible to reconstruct since the first visible letter cannot be ⲟ as one would expect but rather resembles ⲩ. To assume a mistake in the reconstruction is too risky, even though the most reasonable reading is ϣⲁⲉϩⲣⲁⲓⲉⲡⲟⲟⲩⲛ̄ϩⲟⲟⲩ (ⲟⲩ on the following line) found in SaM. Evidently, the reading of SaA, ϣⲁⲡⲟⲟⲩⲛ̄ϩⲟⲟⲩ, is too short, there being only 9 letters on the line.

192

Affiliations of the Sahidic manuscripts and the Greek Vorlage

SaA in this case.148 SaA omits the words ϩⲛ̄ ⲧⲡⲁⲣⲉⲙⲃⲟⲗⲏ against the unanimous evidence of other Sahidic witnesses and Greek. The next sentence, ⲁⲩⲱ ϫⲉ ⲙⲡⲓϩⲉ ⲉϩⲱⲃ ⲉϥϩⲟⲟⲩ ⲉϩⲟⲩⲛ ⲉⲣⲟⲕ ‘And I have not found any bad thing in you’ in SaMFJ, is ⲁⲩⲱ ⲙ̄ⲡⲓϩⲉ ⲉⲩϩⲱϥ ⲉϩⲟⲩⲛ ⲉⲣⲟⲕ ‘And I have not found anything in you’ in SaA. Probably, in a chain of four similar words beginning with ⲉ, the first two have intermingled in SaA thus producing a unique reading without mentioning ‘bad’. SaA omits ϫⲉ, the usual equivalent of ὅτι, like L. SaA is independent of this variant reading in this case since the tendency of SaA to omit offers the simplest explanation. Further, SaA reads the first part of the oath ⲁⲩⲱ ⲕⲥⲟⲩⲧⲱⲛ ⲛⲧⲟⲕ ‘and you are straight, you’ and not ϫⲉ ⲕⲥⲟⲩⲧⲱⲛ ⲛⲧⲟⲕ ‘that you are straight, you’ like SaMFJ. To omit the second ϫⲉ is consistent in SaA. ⲛϫⲓⲛ ⲡⲉϩⲟⲟⲩ̣ ⲛⲧⲁⲕⲉⲓ ⲉϩⲟⲩⲛ ⲛⲁⲓ ‘from the day you came in to me’ in SaM parallels ϫⲓⲛ ⲉⲡⲉϩⲟⲟⲩ ⲛⲧ ̄ ⲁⲕⲓ ⲛⲁⲓ ⲉϩⲟⲩⲛ in SaA. The meaning does not change; only the construction and word order differ slightly. This phrase lies in a lacuna in SaF and SaJ.149 The expression ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛϩⲟⲟⲩ ‘until this day’ in SaMFJ appears as ϣⲁ ⲡⲟⲟⲩ ⲛϩ̄ ⲟⲟⲩ ‘until this day’ in SaA. SaJ contains a reconstruction here, but the reading of SaA is too short for the lacuna in SaJ. The change in SaA does not change the meaning, but shows again the tendency to shorten in SaA. The last sentence is ‘The satraps of the Philistines are not satisfied with you’ in SaA and ‘But the satraps, they are not satisfied with you’ in SaMFJ. When compared with other Sahidic manuscripts, SaA omits the conjunction and adds the word ⲛⲛ ̄ ⲁⲗⲗⲟⲫⲩⲗⲟⲥ as an attribute of the satraps. While the Satraps are the subject in SaA, SaMFJ have formulated the clause with ‘satraps’ in extraposition and a 3rd person plural infix as the formal subject.150 This sentence is rendered especially freely in the Sahidic text: καὶ ἐν ὀφθαλμοῖς τῶν σατραπῶν οὐκ ἀγαθὸς σύ ‘and in the eyes of the satraps you are not good’ is rendered ⲛⲥⲁⲧⲣⲁⲡⲏⲥ ⲇⲉ ⲛⲥⲉⲧⲏⲧ ⲛϩ̄ ⲏⲧ ⲛⲙ̄ⲙⲁⲕ ⲁⲛ ‘but the satraps are not satisfied with you’.151 The B-group in Greek has no negation in this sentence but Coptic presupposes the negation in its Vorlage.

148 However, Payne (Critical and comparative, 85) has done this in his dissertation. He follows the rule ‘lectio difficilior’, and here it is true that the erroneous reading is also the more difficult one. 149 There are a couple of letters visible in SaJ but they do not perfectly match with the readings attested by other witnesses. 150 The spelling of Greek words is not consistent in SaA where satraps is spelled ⲥⲁⲛⲇⲣⲁⲡⲏⲥ (ⲥⲁⲛⲧⲣⲁⲡⲏⲥ 6: 12). 151 In verse 29: 9 the same Greek expression, for some reason, is rendered formally literally.

193

Textual analyses

29: 7 καὶ νῦν ἀνάστρεφε καὶ πορεύου εἰς εἰρήνην, καὶ οὐ μὴ ποιήσεις κακίαν ἐν ὀφθαλμοῖς τῶν σατραπῶν τῶν ἀλλοφύλων. SaM

SaA

SaF

SaJ

ⲧⲉⲛⲟⲩ ϭⲉ ⲕⲧⲟⲕ ⲛⲅⲃⲱⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣⲏⲛⲏ· ⲛⲅⲧⲙ̄ⲉⲓⲣⲉ ⲛⲟⲩⲕⲁⲕⲓⲁ ⲛⲁϩⲣⲁⲩ152· ⲛⲛⲉⲥⲁⲧⲣⲁⲡⲏⲥ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ

ⲧⲉⲛⲟⲩ ϭⲉ ⲕⲧⲟⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣⲩⲛⲏ ⲛⲅⲃ̄ ⲱⲕ ⲛⲅⲧ ̄ ⲙ ⲉⲓⲣⲉ ⲛⲟⲩⲕⲁⲕⲓⲁ ⲛⲛ ̄ ⲁϩⲣⲁⲩ ⲛⲛ ̄ ⲥⲁⲛⲇⲣⲁⲡⲏⲥ ⲛⲛ ̄ ⲁⲗⲗⲟⲫⲩⲗⲟⲥ.

ⲧⲉⲛⲟⲩ ϭⲉ ⲕⲧⲟⲕ ⲛⲅⲃ̄ ⲱⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣⲏⲛⲏ· ⲛⲅⲧⲙ̄ⲉⲓⲣⲉ ⲛⲟ̄ ⲩⲕⲁⲕⲓⲁ ⲛⲛ ̄ ⲁϩⲣⲁⲩ· ⲛⲛ ̄ ⲥ̄ ⲁⲧⲣⲁⲡⲏⲥ ⲛⲛ ̄ ⲁⲗⲗⲟⲫⲩⲗⲟⲥ

[ⲧⲉⲛ]ⲟⲩ ϭ̣ ⲉ [ⲕⲧⲟⲕ ⲛⲅⲃⲱⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣ]ⲏⲛⲏ· [ⲛⲅⲧⲙⲉⲓⲣⲉ] ⲛ[̄ ⲟ]ⲩⲕⲁ[ⲕⲓⲁ ⲛⲛⲁϩ]ⲣⲁ ̣ ̣[ⲩ] ⲛ̣ⲛ ̄ ̣ⲥ̄ ⲁⲧ[ⲣⲁⲡⲏ]ⲥ̣ ⲛ̣ⲛ̣[ⲁⲗⲗⲟⲫⲩⲗⲟⲥ]

In this verse, one significant difference exists among the Sahidic manuscripts, the word order. SaMFJ offer ⲕⲧⲟⲕ ⲛⲅⲃⲱⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣⲏⲛⲏ ‘Turn and go peaceably’, while SaA has ⲕⲧⲟⲕ ϩⲛ̄ ⲟⲩⲉⲓⲣⲩⲛⲏ ⲛⲅⲃ̄ ⲱⲕ ‘Turn peaceably, go’. For some reason, SaA has its own word order.153 29: 8 καὶ εἶπεν Δαυεὶδ πρὸς Ἀγχούς Τί πεποίηκά σοι καὶ τί εὗρες ἐν τῷ δούλῳ σου ἀφ᾽ἧς ἡμέρας ἤμην154 ἐνώπιόν σου καὶ ἕως τῆς ἡμέρας ταύτης, ὅτι οὐ μὴ ἔλθω πολεμῆσαι τοὺς ἐχθροὺς τοῦ κυρίου μου τοῦ βασιλέως; SaM

SaA

SaF

SaJ

ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲁⲭⲟⲩⲥ ϫⲉ ⲟⲩ ⲡⲉ ⲡⲛⲟⲃⲉ ⲛⲧⲁⲓⲁⲁϥ ⲛⲁⲕ· ⲁⲩⲱ ⲟⲩ ⲡⲉ ⲡϩⲱⲃ ⲛⲧⲁⲕϩⲉ ⲉⲣⲟϥ ⲉϩⲟⲩⲛ ⲉⲡⲉⲕϩⲙ̄ϩⲁⲗ· ⲛϫⲓⲛ ⲡⲉϩⲟⲟⲩ ⲛⲧⲁⲓⲉⲓ ⲛⲁϩⲣⲁⲕ ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛ̄ϩⲟⲟⲩ· ϫⲉ ⲛⲛⲁⲉⲓ ⲉⲃⲟⲗ ⲁⲛⲟⲕ ⲉⲙⲓϣⲉ· ⲙⲛ ⲛⲉϫⲁϫⲉ ⲙⲡⲁϫⲟⲉⲓⲥ ⲡⲣⲣⲟ·

ⲡⲉϫⲉ ⲇ̄ⲁⲇ ̄ ̄ ⲛⲁⲭⲟⲩⲥ ϫⲉ ⲟⲩ ⲡⲉ ⲡⲁⲛⲟⲃⲉ ⲛⲧ ̄ ⲁⲓⲁⲁϥ ⲉⲣⲟⲕ·

ⲡⲉϫⲉ ⲇⲁⲩⲉⲓⲇ ⲛⲁⲅⲭⲟⲩⲥ ϫⲉ ⲟⲩ ⲡⲉ ⲡⲛⲟⲃⲉ ⲛⲧⲁⲓ̈ⲁⲁϥ ⲛⲁⲕ· ⲁⲩⲱ ⲟⲩ ⲡⲉ ⲡϩⲱⲃ ⲛⲧⲁⲕϩⲉ ⲉⲣⲟϥ ⲉϩⲟⲩⲛ ⲉⲡⲉⲕϩⲙ̄ϩⲁⲗ· ⲛϫⲓⲛ ⲡⲉϩⲟⲟⲩ ⲛⲧⲁⲓⲉⲓ ⲛ̄ⲛⲁϩⲣⲁⲕ ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛ̄ϩⲟⲟⲩ· ϫⲉ ⲛⲛⲁⲉⲓ ̅ ⲉⲃⲟⲗ ⲁⲛⲟⲕ ⲉⲙⲓϣⲉ· ⲙⲛ ⲛϫⲁϫⲉ ⲙⲡⲁϫⲟⲉⲓⲥ ⲡⲣⲣⲟ

[ⲡⲉ]ϫⲉ ⲇⲁⲩⲉ[ⲓⲇ] ⲛⲁ[ⲅⲭⲟⲩⲥ ϫ]ⲉ ⲟⲩ ⲡⲉ ⲡⲛⲟⲃⲉ̣ ⲛ̣ⲧ ̄ ⲁⲓ̈ⲁⲁϥ ⲛⲁⲕ· ⲁⲩⲱ ⲟⲩ ⲡ̣ⲉ ⲡϩⲱⲃ ⲛⲧⲁⲕϩⲉ[ⲉⲣⲟϥ] ⲉϩⲟⲩⲛ[ⲉⲡ]ⲉ[ⲕϩⲙϩⲁⲗ ⲛ]ϫ[ⲓⲛ ⲡⲉ]ϩⲟⲟ[ⲩ ⲛⲧⲁⲓⲉ]ⲓ ⲛ̄[ⲛⲁ]ϩⲣⲁ[ⲕ ϣⲁ ⲉϩⲣⲁ]ⲓ̈ ⲉ[ⲡⲟ]ⲟⲩ ⲛ̄[ϩⲟⲟⲩ ϫⲉ]ⲛ[ⲛⲁⲉⲓ] ⲉⲃⲟⲗ [ⲁⲛⲟⲕ ⲉⲙⲓϣⲉ] ⲙⲛ̄ ⲛ̄ϫ̣[ⲁϫⲉ ⲙⲡⲁϫⲟⲉⲓⲥ] [ⲡ]ⲣ̄ⲣⲟ

ϫⲓⲛ ⲉⲡⲉϩⲟⲟⲩ ⲛ̄ⲧⲁⲓⲉⲓ ⲉϩⲟⲩⲛ ⲉϩⲣⲁⲕ ϣⲁϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛ̄ϩⲟⲟⲩ. ⲉⲓⲛⲏⲩ ⲛⲉⲙⲙⲁⲕ ⲉⲃⲟⲗ ⲉⲡⲡⲟⲗⲉⲙⲟⲥ ⲉⲓⲙⲓϣⲉ ⲙⲛ̄ ⲛ̄ϫⲁϫⲉ ⲙ̄ⲡⲁϫⲟⲉⲓⲥ ⲙⲣ̄ⲣⲟ·

152 SaM reads ⲛⲁϩⲣⲁⲩ, where SaAF read ⲛ̄ⲛⲁϩⲣⲁⲩ. This is like ⲛⲁϩⲣⲁⲓ versus ⲛ̄ⲛⲁϩⲣⲁⲓ in the previous verse–only a spelling difference. SaJ is reconstructed here, and, therefore, I ignore it, although the reconstruction is consistent with the double ⲛ as above. 153 As in the previous verse ⲉϩⲟⲩⲛ ⲛⲁⲓ versus ⲛⲁⲓ ⲉϩⲟⲩⲛ. 154 With respect to the verb, Greek 44 has the reading ηλθον. The same meaning ‘come’ appears in the Ethiopic version. There is no reason to suppose Sahidic dependence on those witnesses.

194

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In verse eight, there are significant variants between Sahidic manuscripts. SaA shortens David’s question by excluding ⲁⲩⲱ ⲟⲩ ⲡⲉ ⲡϩⲱⲃ ⲛⲧⲁⲕϩⲉ ⲉⲣⲟϥ ⲉϩⲟⲩⲛ ⲉⲡⲉⲕϩⲙ̄ϩⲁⲗ ‘And what is the issue that you have found in your servant’ that appears in other Sahidic manuscripts. Here, the change consists of not only a simple omission but also an explicitation of ⲡⲛⲟⲃⲉ ‘the sin’ into ⲡⲁⲛⲟⲃⲉ ‘my sin’ and ⲛⲁⲕ ‘to you’ into ⲉⲣⲟⲕ ‘to you’. Apparently, SaA has intentionally emended and shortened the text. SaMFJ preserve (ⲛ)̄ ⲛⲁϩⲣⲁⲕ ‘before you’, while SaA reads ⲉϩⲟⲩⲛ ⲉϩⲣⲁⲕ as the equivalent for ἐνώπιόν σου. This does not alter the meaning, but shows how SaA again differs from all the other Sahidic witnesses. The rendering of ὅτι οὐ μὴ ἔλθω πολεμῆσαι is ϫⲉ ⲛⲛⲁⲉⲓ ⲉⲃⲟⲗ ⲁⲛⲟⲕ ⲉⲙⲓϣⲉ ‘that I will not come, I, to fight’ in SaMFJ. SaA has ⲉⲓⲛⲏⲩ ⲛⲉⲙⲙⲁⲕ ⲉⲃⲟⲗ ⲉⲡⲡⲟⲗⲉⲙⲟⲥ ⲉⲓⲙⲓϣⲉ ‘I come with you to the war, fighting’. Somehow, the negation and ϫⲉ have disappeared from SaA and ⲡⲟⲗⲉⲙⲟⲥ has appeared, all without parallel Greek variants. 29: 9 καὶ ἀπεκρίθη Ἀγχοὺς πρὸς Δαυείδ Οἶδα ὅτι ἀγαθὸς σὺ ἐν ὀφθαλμοῖς μου, ἀλλ᾽οἱ σατράπαι τῶν ἀλλοφύλων λέγουσιν Οὐχ ἥξει μεθ᾽ἡμῶν εἰς πόλεμον. Ἀγχούς] + καὶ εἶπε A V O L CI 342 ὀφθαλμοῖς μου] + καθως αγγελος Θεου A O L-93 82 242 488 σατράπαι] στρατηγοί Vc L-127 d 55 71 245 318 554 707 Eth

SaM

SaA

SaF

SaJ

ⲁⲩⲱ ⲁⲭⲟⲩⲥ ⲁϥⲟⲩⲱϣⲃ ⲉϥϫⲱ ⲙⲙⲟⲥ ⲛⲇ̄ⲁⲇ ̄ ̄· ϫⲉ ϯⲥⲟⲟⲩⲛ̄ ϫⲉ ⲛ̄ⲧⲕ ⲟⲩⲁⲅⲁⲑⲟⲥ ⲛⲧⲟⲕ ⲛⲁϩⲣⲁⲓ· ⲁⲗⲗⲁ ⲛⲛⲟϭ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲥⲉϫⲱ ⲙⲙⲟⲥ· ϫⲉ ⲛⲛⲉϥⲉϣⲉⲓ ⲉⲃⲟⲗ ⲛⲙ̄ⲙⲁⲛ̄ ⲉⲡⲡⲟⲗⲩⲙⲟⲥ

ⲁⲩⲱ ⲡⲉϫⲉ ⲁⲭⲟⲩⲥ ⲛⲇ̄ⲁⲇ ̄ ̄ ϫⲉ ϯⲥⲟⲟⲩⲛ ϫⲉ ⲛⲧⲕ ⲟⲩⲁⲅⲁⲑⲟⲥ ⲛⲧⲟⲕ ⲛ̄ⲛⲁϩⲣⲁⲓ. ⲁⲗⲗⲁ ⲛⲛⲟϭ ⲛ̄ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲥⲉϫⲱ ⲙⲙⲟⲥ ϫⲉ ⲛ̄ⲛϥϣ̄ ⲉⲓ ⲛⲉⲙⲙⲁⲛ ⲉⲡⲡⲟⲗⲉⲙⲟⲥ

ⲁⲩⲱ ⲁⲅⲭⲟⲩⲥ ⲁϥⲟⲩⲱϣⲃ ⲉϥϫⲱ ⲙ̄ⲙⲟⲥ ⲛⲇⲁⲩⲉⲓⲇ […155

[ⲁⲩⲱ ⲁⲅⲭⲟⲩⲥ] ⲁϥⲟⲩⲱ[ϣⲃ̄ ⲉϥϫⲱ ⲙ]ⲙⲟⲥ[ⲛⲇⲁⲩⲉⲓⲇ ϫⲉ ϯ]ⲥⲟⲟ[ⲩⲛ̄ ϫⲉ ⲛ̄ⲧⲕ ⲟⲩ]ⲁⲅⲁⲑ[ⲟⲥ ⲛⲧⲟⲕ]156

In this verse, SaMFJ presuppose the Hexaplaric reading with καὶ εἶπε before πρὸς Δαυείδ like A V O L CI 342.157 SaA has only the verb ‘to say’, and thus SaA is understandable as an abbreviation of the text similar to that in other Sahidic witnesses. This is more probable than supposing a connection

155 End of the fragment. 156 The fragment ends here. 157 An asterisk in the margin of 243.

Textual analyses

195

to Greek 44–125 that read like SaA ‘Achous said to David’158 since there are other similar passages where SaA makes this same omission without any Greek variants.159 The Sahidic version does not attest to the other Hexaplaric approximation in this verse, καθὼς ἄγγελος Θεου, present in A O L-82,93 242 488 after ἐν ὀφθαλμοῖς μου. This is a typical feature in Sahidic 1 Sam, which occasionally attests Hexaplaric readings. The next phrase ⲁⲗⲗⲁ ⲛⲛⲟϭ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲥⲉϫⲱ ⲙⲙⲟⲥ ‘But the commanders of the Philistines, they say’ presupposes στρατηγοί in the Vorlage of the Sahidic translator,160 not σατράπαι. According to Brock, the reading with στρατηγοί lay in the original LXX.161 In 1 Sam, the Sahidic translation is consistent. στρατηγοί is rendered with ⲛⲛⲟϭ ‘the great ones’ and σατράπαι with ⲥⲁⲧⲣⲁⲡⲏⲥ. Results One common translation lies behind the extant Sahidic witnesses. This is perceptible in their similar wordings, and through the free rendering in verse 29: 6: καὶ ἐν ὀφθαλμοῖς τῶν σατραπῶν οὐκ ἀγαθὸς σύ ‘and in the eyes of the satraps you are not good’ is rendered ⲛⲥⲁⲧⲣⲁⲡⲏⲥ ⲇⲉ ⲛⲥⲉⲧⲏⲧ ⲛϩ̄ ⲏⲧ ⲛⲙ̄ⲙⲁⲕ ⲁⲛ ‘but the satraps are not satisfied with you’. In several passages, SaMFJ agree against SaA. These variant readings differ in their length, from single words to whole clauses. The discrepancies concern the order of events (29: 6) and word order (29: 7). There are divergences in construction (29: 6, 29: 8, 29: 9), conjunction (29: 6), and contents (29: 6, 29: 8). Since there are no textually significant agreements where some other combination of Sahidic manuscripts occurs, SaA clearly deviates from the original Sahidic tradition more than the other three manuscripts studied here. SaA has shortened the text in many passages. There are both omissions (29: 6) and abbreviating reformulations (29: 6, 29: 8). SaA contains obvious mistakes. In verse 29: 6, ⲛⲉⲥⲛⲙ̄ⲙⲁⲓ of SaM has become ⲥⲥⲙⲁⲙⲁⲁⲧ ‘blessed’ in SaA. Later in the same verse, within a chain of four similar words beginning with ⲉ, the first two have been mixed up in SaA, thus producing a unique, erroneous reading. Of special interest are those passages where SaA shares a reading with the Greek d-group. Drescher in his introduction has mentioned that there are

158 Additionally, manuscript 71 reads καὶ εἶπεν αυτω αγχους. 159 For example, 25: 10. In verse 23: 4, the same omission appears in SaA and Greek 107–125– 610. Typically, the older manuscripts of the d-group (44 125 610) abbreviate. The purpose of these shortened manuscripts is difficult to discern. Perhaps, they were intended only for personal use? 160 στρατηγοί is found in Vc L-127 d 55 71 245 318 554 707 Aeth. 161 Brock, Recensions, 290.

196

Affiliations of the Sahidic manuscripts and the Greek Vorlage

some agreements between Greek 107 and SaA.162 The nature of these agreements remains unclear. In verse 29: 6, the order of events is at stake. Manuscript 107 in Greek attests the same order as SaA, first coming in and then going out.163 In verse 29: 9, SaMFJ presuppose the Hexaplaric reading with καὶ εἶπε before πρὸς Δαυεὶδ in the Vorlage like A V O L CI 342. SaA has only one verb in this phrase, the verb ‘to say’. Greek 44–125 read like SaA ‘Achous said to David’. In some passages, further affiliations of Sahidic manuscripts are discernible. Both SaM and SaA offer in verse 29: 5 a strict translation of αἱ χορεύουσαι as found in L 121 318 554mg. In the same verse, SaA might be dependent on a Greek source nearer to the MT in reading with the verb ‘to sing’. Another possible explanation is the influence of the parallel passage 18: 6–7, where the same verb ‘to sing’ occurs in Sahidic. In verse 29: 9, the Hexaplaric reading καὶ εἶπε precedes πρὸς Δαυεὶδ, attested in Sahidic. However, the Hexaplaric addition καθὼς ἄγγελος Θεου, present in A O L82,93 242 488 after ἐν ὀφθαλμοῖς μου is not present in Sahidic. In verse 29: 9, the phrase ⲁⲗⲗⲁ ⲛⲛⲟϭ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲥⲉϫⲱ ⲙⲙⲟⲥ ‘but the commanders of the Philistines, they say’ presupposes στρατηγοί in the Vorlage, like Vc L-127 d 55 71 245 318 554 707, and not σατράπαι. 3.2.6. SaMAJ 30: 21–24 30: 21 Καὶ παραγίνεται Δαυεὶδ πρὸς τοὺς διακοσίους ἄνδρας τοὺς ἐκλυθέντας τοῦ πορεύεσθαι ὀπίσω Δαυείδ, καὶ ἐκάθισεν αὐτοὺς ἐν τῷ χειμάρρῳ τῷ Βοσορ, καὶ ἐξῆλθον εἰς ἀπάντησιν Δαυεὶδ καὶ εἰς ἀπάντησιν τοῦ λαοῦ τοῦ μετ᾽αὐτοῦ, καὶ προσήγαγεν Δαυεὶδ ἕως τοῦ λαοῦ, καὶ ἠρώτησαν αὐτὸν τὰ εἰς εἰρήνην. SaM

SaA

ⲁⲩⲱ ⲇⲁⲩⲉⲓⲇ ⲁϥⲉⲓ ⲉⲣⲁⲧⲟⲩ ⲙⲡϣⲏⲧ ⲛⲣⲱⲙⲉ· ⲉⲧⲙ̄ⲡⲟⲩϭⲙϭⲟⲙ ⲉⲟⲩⲁϩⲟⲩ ⲛⲥⲁ ⲇ̄ⲁⲇ ̄ ̄ · ⲁⲗⲗⲁ ⲁϥⲑⲙ̄ⲥⲟⲟⲩ· ϩⲙ̄ ⲡⲉⲭⲓⲙⲁⲣⲟⲥ ⲛⲃⲉⲁⲛⲁ· ⲁⲩⲉⲓ ⲇⲉ ⲉⲃⲟⲗ ϩⲏⲧϥ̄ ⲛⲇ̄ⲁⲇ ̄ ̄ ⲁⲩⲱ ⲉⲃⲟⲗ ϩⲏⲧϥ̄ ⲙⲡⲗⲁⲟⲥ ⲉⲧⲛⲙ̄ⲙⲁϥ· ⲇⲁⲩⲉⲓⲇ ⲇⲉ ⲁϥⲡⲱϩ ⲉⲡⲗⲁⲟⲥ· ⲁⲩⲱ ⲁϥϣⲁϫⲉ ⲛⲙ̄ⲙⲁⲩ ϩⲛ ⲟⲩ̄ⲉⲓⲣⲏⲛⲏ·

ⲁⲩⲱ ⲁ ⲇ̅ⲁⲇ ̅ ̅ ⲉⲓ ⲉⲣⲁⲧⲟⲩ ⲙ̄ⲡϣⲏⲧ ⲛⲣⲱⲙⲉ· ⲉⲧⲉ ⲙⲡⲟⲩ ϣ̄ ϭⲙϭⲟⲙ ⲉⲟⲩⲁϩⲟⲩ ⲛⲥⲱϥ· ⲁⲩⲱ ⲁϥⲑⲙⲥⲟⲟⲩ ϩⲙ ⲡⲉⲭⲓⲙⲁⲣⲣⲟⲥ ⲛ̄ⲃⲉⲛⲉⲁ· ⲇ̅ⲁⲇ ̅ ̅ ⲁϥⲓ ϣⲁ ⲡⲗⲁⲟⲥ ⲁϥϣⲁϫⲉ ⲛⲙ̄ⲙⲁⲩ ϩⲛ̄ ⲟⲩⲉⲓⲣⲩⲛⲏ·

SaJ

]ⲉⲡⲗⲁⲟⲥ· ⲁⲩⲱ ⲁϥϣⲁϫⲉ ⲛ̄ⲙ̄ⲙⲁⲩ ϩⲛ̄ ⲟⲩⲉⲓⲣⲏⲛⲏ·

In the first clause, SaM reads with the proper name in extraposition and a 3rd person singular infix as subject, whereas SaA has only the proper name as 162 Drescher, Kingdoms, v. He also actually mentions manuscript 707 in this connection. 163 This order is found also in the Ethiopic and Armenian versions.

Textual analyses

197

subject. This does not affect the meaning, only makes SaA’s clause more straightforward. In the second clause, SaA offers ⲉⲧⲉ ⲙⲡⲟⲩ ϣ̄ϭⲙϭⲟⲙ ‘who were not able’, a relative past tense with negation, and additionally the redundant verbal auxiliary ϣ̄ - ‘to be able’. The construction used in SaM is the same except for ϣ̄ .164 The reading with ϭⲙϭⲟⲙ ‘to be able’ with a negation, used in both Sahidic witnesses, seems to presuppose the reading of A B O b with the verb ἐκλύω ‘become weary, lose courage, give up’. This reading seems to be a correction toward the MT (‘be too feeble, tired to …’). If this is an early correction, then Sahidic attests to this later Greek reading. Other Greek manuscripts read λείπω ‘to leave, lack’ with different prefixes. The end of the second clause also shows a difference. SaA uses the pronominal suffix ⲛⲥⲱϥ ‘behind, after him’ where SaM has the proper noun ⲛⲥⲁ ⲇ̄ⲁⲇ ̄ ̄ ‘behind, after David’. In the Greek tradition, 71 381 read ὀπίσω αὐτοῦ while the majority have the proper name David. In the third clause, SaM offers the conjunction ⲁⲗⲗⲁ where SaA reads ⲁⲩⲱ. Most of the Greek manuscripts have καί.165 In six cases, the Sahidic text of 1 Sam uses ⲁⲗⲗⲁ to render καί.166 Further in the same clause, SaM reads ϩⲙ̄ ⲡⲉⲭⲓⲙⲁⲣⲟⲥ ⲛⲃⲉⲁⲛⲁ and SaA ϩⲙ ⲡⲉⲭⲓⲙⲁⲣⲣⲟⲥ ⲛⲃⲉⲛⲉⲁ. Here, both Sahidic witnesses depend on the reading offered by the B-group in Greek: Vaticanus reads Βεανα, Greek 121 reads Βαιανα and Greek 509 a doublet Βεενα βοσωρ.167 Of all the readings in Payne’s lists, only this example demonstrates Sahidic dependence on the B-group.168 SaA has confused the letters slightly, as is frequently the case with the proper names in this manuscript. A homoiarchon mistake has occurred in SaA from ⲁⲩⲉⲓ ⲇⲉ ‘and they came’ to ⲇⲁⲩⲉⲓⲇ ⲇⲉ ‘and David’. Except for the first letter, these phrases are identical. However, this does not explain the reading of SaA as it stands now. In addition to this mistake, there is a variant without the particle ⲇⲉ and using the verb ⲉⲓ ϣⲁ- ‘come to’ instead of ⲡⲱϩ ⲉ- ‘attain to’. Perhaps the reading of SaA results from two successive stages, the first being clumsy copying and the second a tendency to shorten. SaJ has a lacuna for the verb, but it preserves ⲉ- as the last letter before ⲡⲗⲁⲟⲥ and thus it seems to agree with the reading ⲡⲱϩ ⲉ- of SaM.

164 Layton, Grammar, §184c. 165 L f 55 158 offer the relative pronoun οὕς. 166 See my subsection on clause connections, 2.1.3.1.e). 167 Two other occurrences of the name in verses 30: 9 and 30: 10 read ⲃⲟⲥⲟⲣ in Sahidic. 168 Payne offers extensive lists (Critical and comparative, 141–53 “relatively strong cases of textual relationship”, and 153–7 “relatively weak cases of textual relationship”) but unfortunately, they mostly contain readings where no secondary feature is shared by the Bgroup and the Sahidic text. Cases where Sahidic, together with the B-group, does not attest to a Hexaplaric reading are not valid arguments, since one should rely on the readings where secondary features are shared by the Sahidic witnesses and the B-group.

198

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In the last clause, Sahidic witnesses rely on a Hexaplaric approximation169 with the singular verbal form and a plural pronoun, attested by L 554 in Greek.170 Still, in the last clause SaA omits the particle ⲇⲉ against SaMJ and certain Greek manuscripts. 30: 22 καὶ ἀπεκρίθη πᾶς ἀνὴρ λοιμὸς καὶ πονηρὸς τῶν ἀνδρῶν τῶν πολεμιστῶν τῶν πορευθέντων μετὰ Δαυεὶδ καὶ εἶπαν Ὅτι οὐ κατεδίωξαν μεθ᾽ἡμῶν, οὐ δώσομεν αὐτοῖς ἐκ τῶν σκύλων ὧν ἐξειλάμεθα, ὅτι ἀλλ᾽ἢ ἕκαστος τὴν γυναῖκα αὐτοῦ καὶ τὰ τέκνα αὐτοῦ ἀπαγέσθωσαν καὶ ἀποστρεφέτωσαν. ὅτι 2º] om L 509 44–74–125 245 460

SaM

SaA

SaJ

ⲁⲩⲟⲩⲱϣⲃ̄ ⲇⲉ ⲛϭⲓ ⲣⲱⲙⲉ ⲛⲓⲙ ⲙⲡⲟⲛⲏ̄ⲣⲟⲥ· ⲁⲩⲱ ⲗⲟ̄ⲓⲙⲟⲥ̄ ⲛⲓⲙ ⲛⲣⲱⲙⲉ· ⲉⲃⲟⲗ ϩⲛ̄ ⲛ̄ⲣⲱⲙⲉ ⲛⲣⲉϥⲙⲓϣⲉ· ⲛⲧ ̄ ⲁⲩⲃⲱⲕ ⲉⲃⲟⲗ ⲙⲛ̄ ⲇⲁⲩⲉⲓⲇ ⲉⲩϫⲱ ⲙⲙⲟⲥ· ϫⲉ ⲉⲃⲟⲗ ϫⲉ ⲙⲡⲟⲩⲉⲓ ⲉⲃⲟⲗ ⲛⲙ̄ⲙⲁⲛ· ⲛⲧⲛⲛⲁϯ ⲛⲁⲩ ⲁⲛ ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ· ⲛⲁⲓ ⲛⲧⲁⲛϫⲓⲧⲟⲩ· ⲁⲗⲗⲁ ⲙⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ ϫⲓ ⲛ̄ⲧⲉϥⲥϩⲓⲙⲉ ⲙⲛ̄ ⲛⲉϥϣⲏⲣⲉ· ⲛⲥⲉⲧⲱⲟⲩⲛ ⲛⲥⲉⲃⲱⲕ·

ⲁⲩⲟⲩⲱϣϥ ⲛϭⲓ ⲛ̄ⲣⲱⲙⲉ ⲙⲡⲟⲛⲏⲣⲟⲥ ⲁⲩⲱ ⲛ̄ⲗⲟⲓⲙⲟⲥ ⲛⲓⲙ ⲛ̄ⲣⲱⲙⲉ ⲛⲣϥⲙⲓ̈ϣⲉ ⲛⲧ ̄ ⲁⲩⲃⲱⲕ ⲙⲛ̄ ⲇ̅ⲁⲇ ̅ ̅ · ⲉⲩϫⲱ ⲙ̄ⲙⲟⲥ ϫⲉ ⲉⲃⲟⲗ ϫⲉ ⲙ̄ⲡⲟⲩⲉⲓ ⲛⲙ̄ⲙⲁⲛ ⲉⲡⲡⲟⲗⲁⲓ̅ ⲙⲟⲥ· ⲛⲧ ̄ ⲛⲛ ̄ ⲁϯ ⲛⲁⲩ ⲁⲛ ϩⲛ̄ ⲛϣⲱⲗ ⲛⲧ ̄ ⲁⲛϣⲟⲗⲟⲩ· ⲁ̅ⲗⲗⲁ̅ ⲙⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ ϫⲓ ⲛ̄ⲧϥ̄ⲥϩⲓⲙⲉ ⲙⲛ̄ ⲛϥ̄ϣⲏⲣⲉ ⲛ̄ⲥⲉⲧⲱⲟⲩⲛ ⲛ̄ⲥ[ⲉ]ⲃⲱⲕ

ⲁⲩⲟⲩⲱϣⲃ̄ ⲇⲉ ⲛ̄ϭⲓ ⲣⲱⲙⲉ ⲛⲓⲙ ⲙ̄ⲡⲟⲛⲏⲣⲟⲥ· ⲁⲩⲱ ⲗⲟⲓⲙⲟⲥ ⲛⲓⲙ ⲛ̄ⲣⲱⲙⲉ ⲛ̄ⲣⲉϥⲙⲓϣⲉ ⲉⲛⲧⲁⲩⲃⲱ̣ ⲕ ⲉⲃ[ⲟⲗ ⲙⲛ]̄ ⲇⲁⲩⲉⲓⲇ ⲉⲩ[ϫⲱⲙⲙⲟ]ⲥ· ϫⲉ ⲉⲃⲟ[ⲗ] [ϫⲉ ⲙⲡⲟⲩⲉ]ⲓ ⲉⲃⲟⲗ[…]

The order of the adjectives in Sahidic in the first clause is the same as in Greek L (= MT)—first ⲡⲟⲛⲏⲣ̄ ⲟⲥ and then ⲗⲟⲓ̄ ⲙⲟⲥ.̄ Brock states that it is possible to find the Old Greek in Lucianic manuscripts in this case.171 If his assumption is correct, then this is only one passage where SaMAJ have preserved the OG together with L. In the first clause, SaMJ and all the Greek manuscripts preserve the connective, and only SaA omits it. In the same clause, SaMJ share the construction ⲣⲱⲙⲉ ⲛⲓⲙ ⲙⲡⲟⲛⲏⲣ̄ ⲟⲥ ‘every evil man’ against SaA that reads ⲛⲣ̄ ⲱⲙⲉ ⲙⲡⲟⲛⲏⲣⲟⲥ ‘evil men’. Further in the same clause, ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲣ̄ ⲱⲙⲉ ‘among the men’ is found in SaM against SaAJ. In the Greek tradition, manuscripts 530 44 omit τῶν ἀνδρῶν from the phrase τῶν ἀνδρῶν τῶν πολεμιστῶν τῶν πορευθέντων. However, I do not suggest a connection between SaAJ and those Greek

169 ἠρώτησεν αὐτούς = MT. According to Brock (Recensions, 157), this is a Hexaplaric approximation. 170 In L, both the verb and the pronoun are in the plural in 554. 171 Brock, Recensions, 125.

Textual analyses

199

manuscripts since SaJ is very early and there are no other hints of a dependence between SaA and SaJ. Perhaps someone added these words into SaM, since without these words the text actually says that all the men who were wandering with David were worthless. Thus, I see SaAJ as preserving the earliest Coptic here. τῶν ἀνδρῶν was rendered originally with the prefix ⲣⲉϥ- and the reading with ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲣ̄ ⲱⲙⲉ in SaM resulted from explicitation. In the following relative clause, SaMJ read the verb with ⲉⲃⲟⲗ against SaA, which omits it. In the middle of the verse, SaA contains the verb ⲙⲡⲟⲩⲉⲓ ‘they did not come’ without ⲉⲃⲟⲗ and with ⲉⲡⲡⲟⲗⲁⲓⲙⲟⲥ ‘to the war’ against SaMJ. Because no Greek equivalent having πόλεμος exists, this reading is probably no more than yet another peculiarity of SaA. In the following clause, SaM reads ⲛⲧⲛⲛⲁϯ ⲛⲁⲩ ⲁⲛ ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲉϣⲱⲗ ‘We will not give to them from the spoil’ and SaA without ⲉⲃⲟⲗ: ⲛⲧ ̄ ⲛⲛ ̄ ⲁϯ ⲛⲁⲩ ⲁⲛ ϩⲛ ⲛϣⲱⲗ. In keeping with its shortening tendency, SaA omits ⲉⲃⲟⲗ without altering the meaning, since ϩⲛ is used synonymously with ⲉⲃⲟⲗ ϩⲛ.̄ 172 The spoil is described in SaM with ⲛⲁⲓ ⲛⲧⲁⲛϫⲓⲧⲟⲩ ‘these that we took’ but in SaA with ⲛⲧ ̄ ⲁⲛϣⲟⲗⲟⲩ ‘what we despoiled’. This same phenomenon appears several times in my material. SaM uses the appositive attributive construction for relative clauses, while SaA uses simple attributive construction.173 This fits the tendency of SaA towards shortening and simplifying the text. The verb is different, but it is impossible to say whether ϫⲓ ‘to take’ in SaM is more original than ϣⲱⲗ ‘to despoil’ in SaA. The reading of SaA fits together with the preceding noun ϣⲱⲗ. The Sahidic text renders the last sentence ⲁⲗⲗⲁ ⲙⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ ϫⲓ ⲛⲧ ⲉϥⲥϩⲓⲙⲉ ⲙⲛ̄ ⲛⲉϥϣⲏⲣⲉ· ⲛⲥⲉⲧⲱⲟⲩⲛ ⲛⲥⲉⲃⲱⲕ ‘But that each one takes his wife ̄ and sons and they rise and go’ with three verbs instead of the two in the Greek text. The translator has added the jussive verb ⲙⲁⲣⲉ … ϫⲓ ‘to take’, and thus made the text easier to read. This peculiar agreement demonstrates that these Sahidic manuscripts had a common ancestor, and are not independent translations. The beginning of the sentence as it now stands seems to attest the reading without ὅτι like L 509 44–74–125 245 460. However, it is impossible to determine whether the Sahidic translator has omitted ὅτι as unreasonable or his Greek Vorlage did not have that word.174

172 Crum, Dictionary, 684a. 173 For more about these constructions, cf. Layton, Grammar, § 403–8. For an analysis of the cases present in my material, see section 2.1.4.1. 174 The Sahidic does not differentiate between ἀλλά and ἀλλ᾽ἤ when used as clause connectors. See also my analyses of these passages in 1 Sam, section 2.1.3.4.

200

Affiliations of the Sahidic manuscripts and the Greek Vorlage

30: 23 καὶ εἶπεν Δαυείδ Οὐ ποιήσετε οὕτως μετὰ τὸ παραδοῦναι τὸν Κύριον ἡμῖν καὶ φυλάξαι ἡμᾶς, καὶ παρέδωκεν Κύριος τὸν γεδδοὺρ τὸν ἐπερχόμενον ἐφ᾽ἡμᾶς εἰς χεῖρας ἡμῶν. SaM

SaA

SaJ

ⲡⲉϫⲉ ⲇ̄ⲁⲇ ̄ ̄ ⲛⲁⲩ ϫⲉ ⲛⲉⲧⲉⲧⲛ̄ⲛⲁⲉⲓⲣⲉ ⲁⲛ ϩⲓⲛⲁⲓ ⲙⲡⲉⲓⲧⲱϣ· ⲙⲛⲛⲥⲁ ⲉⲧⲣⲉ ⲡϫⲟⲉⲓⲥ· ⲧⲁⲁⲩ ⲉϩⲣⲁⲓ ⲉⲧⲟⲟⲧⲛ·̄ ⲛϥ̄ϩⲁⲣⲉϩ ⲉⲣⲟⲛ· ⲁⲩⲱ ⲡⲉⲓⲕⲉⲧⲇⲟⲩⲣ· ⲏ ⲡⲉⲓⲙⲏⲏϣⲉ ⲛⲧⲁϥⲉⲓ ⲉϩⲣⲁⲓ ⲉϫⲱⲛ· ⲉⲁϥⲧⲁⲁⲩ ⲉϩⲣⲁⲓ̈ ⲉⲧⲟⲟⲧⲛ·̄

ⲡⲉϫⲉ ⲇ̅ⲁⲇ ̅ ̅ ⲛⲁⲩ ϫⲉ ⲛ̄ⲧ̅ⲛⲛ ̅ ⲁⲉⲓⲣⲉ ⲁⲛ ⲙⲡⲉⲓⲧⲱϣ ϩⲓ ⲛⲁⲓ· ⲙⲛ̄ⲛⲥⲁ̄ ⲧⲣ̄ ⲉ ⲡϫⲟⲉⲓⲥ ⲧⲁⲁⲩ ⲉⲧⲟⲟⲧⲛ̄ ⲁϥ̄ ϩⲁⲣϩ ⲉⲣⲟⲛ· ⲁⲩⲱ ⲡⲉⲭⲉⲇⲇⲟⲩⲣ ⲙⲛ ⲡⲓⲙⲏⲏϣⲉ ⲛⲧⲁⲩⲉⲓ ⲉϫⲱⲛ ⲁϥⲧⲁⲁⲩ ⲉⲧⲟⲟⲧⲉⲛ

ⲙⲛ̄ⲛⲥ̄ ⲁ ⲧⲣⲉ ⲡϫⲟⲉⲓⲥ ⲧⲁⲁⲩ ⲉϩⲣⲁⲓ̈ ⲉⲧⲟⲟⲧⲛ̄ ⲛϥ̄ϩⲁⲣⲉϩ ⲉⲣⲟⲛ· ⲁⲩⲱ ⲡⲉⲓ̈ⲅⲉⲇ·ⲇⲟⲩⲣ· ⲏ ⲡⲉⲓ̈ⲙⲏⲏϣⲉ ⲉⲛⲧⲁϥⲉⲓ̈ ⲉϩⲣⲁⲓ̈ ⲉϫⲱⲛ· ⲉⲁϥⲧⲁⲁⲩ ⲉϩⲣⲁⲓ̈ ⲉⲧⲟⲟⲧⲛ·̄

In the second clause, SaM uses the 2nd person plural of the verb, and SaA uses the 1st person plural. In the Greek, the persons are 2nd person plural in the majority, and the 3rd person singular in other manuscripts.175 The variant of SaA improves the narrative, and results either from scribal alteration or a misspelling. In the next clause, SaA offers ⲙⲛⲛ ̄ ⲥⲁ̄ ⲧⲣ̄ ⲉ ⲡϫⲟⲉⲓⲥ ⲧⲁⲁⲩ ⲉⲧⲟⲟⲧⲛ̄ ⲁϥ̄ ϩⲁⲣⲉϩ ⲉⲣⲟⲛ ‘After the Lord gave them into your hands, he guarded us’ as an independent sentence, not subordinated as in SaMJ.176 The same phenomenon occurs later when SaA offers an independent sentence (as main clause ⲁϥⲧⲁⲁⲩ ⲉⲧⲟⲟⲧⲉⲛ ‘he gave them to our hands’) while SaMJ continue the previous sentence. SaA consistently omits ⲉϩⲣⲁⲓ against SaMJ—three times in this verse alone. This inner-Sahidic variant has no Greek parallel, and again demonstrates SaA’s tendency to abbreviate. In the second clause, the equivalent of οὕτως is ϩⲓⲛⲁⲓ ⲙⲡⲉⲓⲧⲱϣ ‘thus in this manner’ in SaM, but ⲙⲡⲉⲓⲧⲱϣ ϩⲓ ⲛⲁⲓ ‘in this manner, thus’ in SaA without difference in meaning. SaJ has a lacuna here. ⲡⲉⲓⲕⲉⲧⲇⲟⲩⲣ ⲏ ⲡⲉⲓⲙⲏⲏϣⲉ is an intriguing loanword that appears in all Sahidic witnesses with only minor spelling variation. The word is complemented with an explanation, introduced by ⲏ as is typical of explanations.177 SaA preserves the secondary variation ⲡⲉⲭⲉⲇⲇⲟⲩⲣ ⲙⲛ ⲡⲓⲙⲏⲏϣⲉ thus grammatically distinguishing two enemies and resulting in the need to change the following verb into the plural, against the singular in SaMJ and in all Greek manuscripts.178 175 376 19–82 509 610 488–762* 245 460 707*. 176 ⲙⲛⲛⲥⲁ ⲉⲧⲣⲉ- is equivalent to ⲙⲛⲛⲥⲁ ⲧⲣⲉ- (Layton, Grammar, §363). 177 These cases are analyzed in section 2.2.4.3 dealing with translator’s additions. 178 Payne (Critical and comparative, 88) cites this as an example where the person of the verb in SaM is corrected according to Greek and SaA preserves the earliest Coptic. Against Payne, Dresher (Kingdoms, xvii–xviii) has noted that SaA is corrupt in this verse with respect to the ⲏ-/ ⲙⲛ- variants.

201

Textual analyses

30: 24 καὶ τίς ὑπακούσεται ὑμῶν τῶν λόγων τούτων; ὅτι οὐχ ἧττον ὑμῶν εἰσιν· διότι κατὰ τὴν μερίδα τοῦ καταβαίνοντος εἰς πόλεμον οὕτως ἔσται ἡ μερὶς τοῦ καθημένου ἐπὶ τὰ σκεύη· κατὰ τὸ αὐτὸ μεριοῦνται.

SaM

SaA

SaJ

ⲁⲩⲱ ⲛⲓⲙ ϭⲉ ⲡⲉⲧⲛⲁⲥⲱⲧⲙ̄ ⲛⲥⲁ ⲛⲉⲧⲛϣⲁϫⲉ· ⲉⲧⲉ ⲛⲁⲓ ⲛⲉ· ϫⲉ ⲛⲉⲓⲣⲱⲙⲉ ⲛⲥⲉⲑⲃⲃ̄ ⲓⲏⲩ ⲁⲛ ⲉⲣⲱⲧⲛ·̄ ϫⲉ ⲕⲁⲧⲁ ⲡϣⲓ ⲛⲧ ̄ ⲧⲟ̄· ⲙⲡⲉⲛⲧⲁϥⲉⲓ ⲉⲃⲟⲗ ⲉⲡⲡⲟⲗⲩⲙⲟⲥ· ⲉⲥⲛⲁϣⲱⲡⲉ ϩⲱⲱϥ ⲛⲛⲉⲧϩⲙⲟⲟⲥ ϩⲁⲧⲛ̄ ⲛⲉⲥⲕⲏⲩⲉ· ⲉⲛⲛⲁⲡⲱϣ ⲉ̄ϫⲱⲟⲩ ⲕⲁⲧⲁ ⲡⲉⲓⲧⲱϣ·

ⲁⲩⲱ ⲛⲓⲙ ϭⲉ ⲡⲉ[ⲧⲛⲁⲥⲱ]ⲧⲙ ⲁⲩⲱ ⲛⲓⲙ ϭⲉ ⲡⲉⲧⲛⲁⲥⲱⲧⲙ [ⲛ]ⲥⲁⲛⲉⲧⲛ̄[ϣⲁϫⲉ ⲉⲧⲉ] ⲛ̄ⲥⲁ ⲛⲉⲧⲛ̄ϣⲁϫⲉ· ⲉⲧⲉ [ⲛ]ⲁⲓ̈ⲛⲉ[…] ⲡⲁⲓ ⲡⲉ ϫⲉ ⲛⲓⲣⲱⲙⲉ ⲛⲥⲉⲡⲟⲣϫ ⲉⲣⲱⲧⲛ ⲁⲛ ̄ [ⲕⲁ]ⲧⲁ ⲡϣⲓ ⲙⲡⲉⲛⲧⲁ[ϥⲉ]ⲓ ⲉⲃⲟⲗ ⲉⲡⲡⲟⲗⲉⲙⲟⲥ ⲧⲁⲓ ⲧⲉ ⲑⲉ ⲉⲧⲛⲁϣⲱⲡⲉ ϩⲱⲱϥ ⲛϭ ̄ ⲓ ⲧⲧⲟ ⲙ̄ⲡⲉⲧϩⲙⲟⲟⲥ ϩⲁⲧⲉⲛ ⲛⲉⲥⲕⲉⲩⲏ· ⲉⲩⲛⲁⲡⲱϣ ϩⲓϫⲱⲟⲩ ⲕⲁⲧⲁ ⲡⲉⲓⲧⲱϣ179

In the second clause, SaMJ offer plural ⲛⲁⲓ ⲛⲉ ‘these’, and SaA has singular ⲡⲁⲓ ⲡⲉ ‘this’. In Greek, some manuscripts180 read with sg. λόγον. SaA, however, is unrelated to these Greek witnesses, since it has the first expression ⲛⲉⲧⲛϣ ̄ ⲁϫⲉ ‘your words’ in the plural and only what comes as equivalent for τούτων is singular. In the next clause, SaM and SaA preserve different verbs: SaM like Greek ‘they are not lesser than you’ and SaA with the verb ⲡⲱⲣϫ ‘they will not share to you’. No variant in the Greek tradition can explain the reading of SaA, and, apparently, the scribe was not formally literal but has written something that fits the context by using the verb ‘to share, divide’. SaA contains a longer version of the last lines than SaM. Notably, SaA tends to shorten readings, but not here. In this case, SaA makes the text easier to read by adding the subject and a ⲧⲁⲓ ⲧⲉ ⲑⲉ ‘so too’-construction. Despite the formal differences, the meaning is about the same in both Sahidic manuscripts. The last verb ‘to share’ is a 1st person plural in SaM against SaA and unanimous Greek evidence of a 3rd person plural. Results One ancestor stands behind the Sahidic witnesses also in this passage. This is perceivable in uniform wordings, and especially in a shared reading against the Greek tradition (30: 22).

179 This last clause is missing in Maspero’s edition (Maspero, Fragments de la version thébaine, 157) but cited in Drescher’s apparatus (Kingdoms, 95). 180 τὸν λόγον τοῦτον 509 d-125 527 245 460; τον λόγον τούτων 488 158; τοὺς λόγους τούτους f; > O.

202

Affiliations of the Sahidic manuscripts and the Greek Vorlage

When evaluating the agreements, the largest group consists of cases where SaMJ agree against SaA. These concern vocabulary (30: 21, 30: 23), conjunction (30: 21, 30: 22), construction (30: 22, 30: 23), adverb (30: 22, 30: 23), and number (30: 24). The similarity of SaM and SaJ seems to belong to the original Sahidic text, or at least to the oldest currently attainable form. These two manuscripts witness a common text, but they are not dependent on each other in secondary features. SaAJ agree against SaM in only one prepositional phrase (30: 22). SaM should be recognized as secondary in this reading, and, thus, there is no dependence between SaAJ in a secondary reading. Because SaM’s reading is an emendation, the agreement of SaAJ is of little value in reconstructing the Sahidic textual tradition, except to say that the original translation probably contained their reading. In no passages do SaAM agree against SaJ. In these verses, SaA is clearly independent of SaMJ. Its tendency to shorten and simplify the text is particularly remarkable in this brief passage. In verse 30: 21, the actual state of SaA is best explained by two separate stages in its history. A careless copyist produced an error-laden text and a second scribe abbreviated the text. Perhaps, this later person tried to ameliorate the corrupted readings. The results—simplifying reformulations and omissions, as well as mistakes—are encountered repeatedly in this passage. Some affiliations are evident in the Sahidic manuscripts. In 30: 21, SaMA presuppose the reading now attested by Greek A B O b (=MT), and later in the same verse SaMA share the proper name with the Greek B-group, SaA with a spelling mistake. In 30: 21, SaA shares the same reading with Greek 71 381, but this same variant (proper name/pronominal suffix) also occurs elsewhere in Sahidic 1 Sam with no Greek manuscript support. Furthermore, in 30: 21, SaMAJ read the persons like L (=MT) ‘he spoke with them’. In verse 30: 22, there might be a connection to the reading now found in L 509 44–74–125 245 460, but only a weak one. The situation of the Sahidic in 30: 22 is problematic (although possible) because it is impossible to determine whether the Sahidic translator has omitted ὅτι as unreasonable or his Greek Vorlage did not have that word. 3.2.7. SaMAB 31: 1–13 31: 1 καὶ οἱ ἀλλόφυλοι ἐπολέμουν ἐπὶ Ἰσραήλ, καὶ ἔφυγον οἱ ἄνδρες Ἰσραὴλ ἐκ προσώπου τῶν ἀλλοφύλων, καὶ πίπτουσιν τραυματίαι ἐν τῷ ὄρει τῷ Γελβοῦε. SaM

SaA

SaB

ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ· ⲁⲩⲡⲟⲗⲩⲙⲉⲓ ⲙⲛ̄ ⲡⲓⲏⲗ̄ ·̄ ⲁⲩⲱ ⲛⲣⲱⲙⲉ ⲙⲡⲓⲏⲗ̄ ̄ ⲁⲩϭⲱⲧⲡ ϩⲁ ⲡϩⲟ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩϩⲉ ⲉⲩϣⲱⲱϭⲉ ϩⲙ̄ ⲡⲧⲟⲟⲩ ⲛⲕⲉⲗⲃⲟⲩⲉ·

ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲡⲟⲗⲉⲙⲉⲓ ⲙⲛ̄ ⲡⲓⲏⲗ̄ ̄ ⲁⲩϭⲱⲧⲡ ⲙⲡⲉⲙⲧⲟ ⲉⲃⲟⲗ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩϣⲱⲱϭⲉ ⲁⲩϩⲉ ϩⲙ̄ ⲡⲧⲟⲟⲩ ⲛⲕⲉⲗⲃⲟⲩⲉ·

ⲙⲛⲛⲥⲁ ⲛⲁⲓ ⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲡⲟⲗⲩⲙⲉⲓ ⲙⲛ ⲡⲓⲏ̅ ⲗ̄ ·̄ ⲁⲡⲗⲁⲟⲥ ϭⲱⲧⲡ ϩⲓ ⲑⲏ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ

203

Textual analyses

In the first clause, SaA agrees with SaM in preserving the connective ⲇⲉ. Among the Greek manuscripts, one introduces the sentence with οἱ ἀλλόφυλοι δέ.181 There is no reason to maintain that SaMA are dependent on this Greek reading, since the translator has used ⲇⲉ 407 times in a clause that reads with καί in Greek. καί-clauses were clearly often rendered with a ⲇⲉ-clause.182 SaB offers the independent reading ⲙⲛⲛⲥⲁ ⲛⲁⲓ ‘after these’. This same expression is used by SaB to begin a new pericope in 16: 14, 18: 6 and 28: 4. In these passages, other Sahidic witnesses do not attest to this expression, nor is there an equivalent Greek variant. With respect to these added discourse markers, SaB resembles a chronicle. Specifically, the narrative regularly contains pericopes that are laxly connected with an undefined temporal framework. In the following clauses, SaM explicitly says that ‘the men of Israel fled before the Philistines.’ Approximately the same content is found in SaA where it is said that ‘they fled in front of the Philistines.’ Either the words ⲁⲩⲱ ⲛⲣⲱⲙⲉ ⲙⲡⲓⲏⲗ̄ ̄ ‘and the men of Israel’ were omitted intentionally to abbreviate the text, or there was a skip of eye from the first to the second ⲡⲓⲏⲗ̄ .̄ A homoioteleuton mistake could only occur within the Coptic tradition, because of word order. SaB has ⲡⲗⲁⲟⲥ as the subject without corresponding Greek variants. In what follows, SaB leaves out the rest of this verse and the beginning of the next. Both SaM and SaA render πίπτουσιν τραυματίαι with two verbs, ϩⲉ ‘to fall’ and ϣⲱⲱϭⲉ ‘to wound, to smite’. SaM first reads a past tense of ϩⲉ and then a circumstantial of ϣⲱⲱϭⲉ. SaA has a past tense of ϣⲱⲱϭⲉ and then a past tense of ϩⲉ. The content in both Sahidic witnesses differs from that in the Greek, since Sahidic explicitly states that all the men of Israel were wounded and fell. The Greek describes only that ‘the wounded fell’. 31: 2 καὶ συνάπτουσιν ἀλλόφυλοι τῷ Σαοὺλ καὶ τοῖς υἱοῖς αὐτοῦ, καὶ τύπτουσιν ἀλλόφυλοι τὸν Ἰωναθὰν καὶ τὸν Ἀμιναδὰβ καὶ τὸν Μελχισά υἱοὺς Σαούλ. SaM

SaA

SaB

ⲁⲩⲱ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲕⲱⲧⲉ ⲉⲥⲁⲟⲩⲗ· ⲁⲩⲱ ⲉⲛⲉϥϣⲏⲣⲉ· ⲁⲩⲱ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲣⲱϩ̄ⲧ̄ ⲛⲓⲱⲛⲁⲑⲁⲛ· ⲙⲛ̄ ⲁⲙⲓⲛⲁⲇⲁⲃ· ⲙⲛ ⲙⲉⲗⲭⲓⲥⲁ· ⲛϣⲏⲣⲉ ⲛⲥⲁⲟⲩⲗ·

ⲁⲩⲱ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲕⲱⲧⲉ ⲉⲥⲁⲟⲩⲗ ⲙⲛ̄ ⲛϥ̄ϣⲏⲣⲉ· ⲁⲩⲣⲱϩⲧ ⲛⲓ̈ⲱⲛⲁⲑⲁⲙ· ⲙⲛ̄ ⲁⲙⲓⲛⲁⲇⲁⲃ· ⲙⲛ ⲙⲉⲗⲭⲉⲓⲥⲁ· ⲛϣⲏⲣⲉ ⲥⲁⲟⲩⲗ·

ⲁⲩⲙⲟⲩⲟⲩⲧ ⲛⲓⲱⲛⲁⲑⲁⲛ· ⲡϣⲏⲣⲉ ⲛⲥⲁⲟⲩⲗ

In the first clause, SaMA differ in the connective and the marking of the object. SaM reads the preposition ⲉ- before ‘his sons’ whereas SaA does not repeat the preposition. 181 Manuscript 44. 182 See my subsection 2.1.3.1.d) on this issue.

204

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In the second clause, SaM contains the conjunction and ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ in extraposition. SaA omits these two words with only a pronominal infix for the subject. In the Greek, 44 381 omit ‘the Philistines’ like SaA. These intraversional Sahidic variations, however, do not demonstrate anything concerning the history of the Sahidic textual tradition, since SaA also omits the conjunction unlike these two Greek manuscripts. In addition, a typical feature in SaA is shortening, and omissions are met constantly without any Greek source text. SaB is independent in this verse, merely describing Jonathan’s death with a more common verb than SaMA. This same phenomenon occurs in verses 31: 6 and 31: 8 where Greek and other Sahidic witnesses have ‘Saul and his three sons’ but SaB reads ‘Saul and his son Jonathan’. One can conclude that SaB is interested in Saul and Jonathan but not in Saul’s other sons and their destiny. It is fitting in a chronicle-like story to concentrate on the main characters and diminish the role of the others. SaB reads ‘son of Saul’ in singular like Greek B 509, but the entire sentence before these words differs considerably. Furthermore, at least in this verse, SaB preserves such a free formulation that this apparent agreement does not support a dependence between SaB and this Greek reading. 31: 3 καὶ βαρύνεται ὁ πόλεμος ἐπὶ Σαούλ, καὶ εὑρίσκουσιν αὐτὸν οἱ ἀκοντισταί, ἄνδρες τοξόται, καὶ ἐτραυματίσθη εἰς τὰ ὑποχόνδρια. SaM

SaA

SaB

ⲁⲩⲱ ⲡⲡⲟⲗⲩⲙⲟⲥ ⲁϥⲉⲛϣⲟⲧ ⲉϫⲛ̄ ⲥⲁⲟⲩⲗ· ⲁⲩϩⲉ ⲉⲣⲟϥ ⲛϭⲓ ⲛⲉⲧⲛⲉϫⲥⲟⲧⲉ· ⲁⲩⲛⲉϫⲥⲟⲧⲉ ⲉⲣⲟϥ· ⲁⲩⲣⲁϩⲧϥ̄ ⲉⲡⲥⲉⲗⲉⲡⲓⲛ ⲛϩⲏⲧϥ·

ⲁⲩⲱ ⲡⲡⲟⲗⲉⲙⲟⲥ ⲁϥⲛ̄ϣⲟⲧ ⲉϫⲛ̄ ⲥⲁⲟⲩⲗ ⲁⲩϩⲉ ⲉⲣⲟϥ ⲛ̄ϭⲓ ⲛⲉⲧⲛϫⲥⲟⲧⲉ·

ⲁ ⲡⲡⲟⲗⲉⲙⲟⲥ ⲛϣⲟⲧ ⲉϫⲉⲛ ⲥⲁⲟⲩⲗ ⲁϥⲉⲓ ⲉϩⲟⲩⲛ ⲉϩⲣⲁⲩ ⲛⲛⲉⲧⲛⲉϫⲥⲟⲧⲉ·

ⲁⲩⲣⲁϩⲧϥ̄ ⲉⲡⲥⲉⲗⲉⲡⲓⲛ ⲛϩⲏⲧϥ·

ⲁⲩⲣⲁϩⲧϥ ⲉⲡⲥⲉⲗⲉⲡⲓⲛ ⲛϩⲏⲧϥ

In the first clause, SaMA share the conjunction ⲁⲩⲱ while SaB omits it. The constructions in SaMA agree (past tense with an extraposited term and 3rd person singular infix) against SaB (past tense with the noun as subject). According to SaMA, the archers find Saul, but, according to SaB, Saul himself meets the arrow shooters. However, there are no Greek variants, and, thus, SaB is independent here. In SaM, ἄνδρες τοξόται is rendered with ⲁⲩⲛⲉϫⲥⲟⲧⲉ ⲉⲣⲟϥ ‘they shot arrows at him’. Perhaps, τοξόται was understood by the Sahidic translator as a participle of τοξεύω ‘to shoot arrows’.183 In SaAB, these words have no equivalent. In Greek, there are no omissions like SaAB, and thus the abbre183 In any case, this word is redundant in the Greek, and the Greek translator possibly did not understand the Hebrew expression as describing the weapon with which the men were shooting.

205

Textual analyses

viating tendency of these Sahidic manuscripts most cogently explains these Sahidic omissions. 31: 4 καὶ εἶπεν Σαοὺλ πρὸς τὸν αἴροντα τὰ σκεύη αὐτοῦ Σπάσαι τὴν ῥομφαίαν σου καὶ ἀποκέντησόν με ἐν αὐτῇ, μὴ ἔλθωσιν οἱ ἀπερίτμητοι οὗτοι καὶ ἀποκεντήσωσίν με καὶ ἐμπαίξωσίν μοι. καὶ οὐκ ἐβούλετο ὁ αἴρων τὰ σκεύη αὐτοῦ, ὅτι ἐφοβήθη σφόδρα· καὶ ἔλαβεν Σαοὺλ τὴν ῥομφαίαν αὐτοῦ καὶ ἐπέπεσεν ἐπ᾽αὐτήν. μή] μηποτε L-82 CII s-64 381 ἐπέπεσεν A B M V 127 CI 242 527 121 134 56 64’ 29 55 71 158 245 318 554mg 707vid] ἐπαισεν 509* 107’; ἐπεσεν rel.

SaM

SaA

SaB

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲙⲡⲉⲧϥⲓ̈ ⲛⲛⲉϥⲥⲕⲏⲩⲉ· ϫⲉ ⲧⲱⲕⲙ̄ ⲛⲧ ̄ ⲉⲕⲥⲏϥⲉ ⲛⲅ̄ ⲕⲟⲟⲛⲥ̅ⲧ̅ ⲛϩⲏⲧⲥ· ⲙⲏⲡⲱⲥ ⲛⲥⲉⲉⲓ ⲛϭⲓ ⲛⲓⲁⲧⲥⲃ̄ⲃⲉ· ⲛⲥⲉⲙⲟⲟⲩⲧ ⲁⲩⲱ ⲛⲥⲉⲥⲱⲃⲉ ⲛⲥⲱⲓ ⲡⲉⲧϥⲓ̈ ⲇⲉ ⲛⲛⲉϥⲥⲕⲏⲩⲉ ⲙⲡⲉϥⲟⲩⲱϣ· ⲉⲙⲟⲟⲩⲧϥ̄ ϫⲉ ⲁϥⲣ̄ϩⲟⲧⲉ ⲉⲙⲁⲧⲉ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥϫⲓ ⲛⲧⲉϥⲥⲏϥⲉ· ⲁϥⲃⲟϭϥ ⲉϫⲱⲥ·

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲙⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲉⲩⲏ· ϫⲉ ⲧⲱⲕⲙ̄ ⲛⲧ ̄ ⲉⲕⲥⲏⲃⲉ ⲛⲅ̄ ⲉⲓ ⲛⲅⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ ⲛϩⲏⲧⲥ· ⲙⲏⲡⲟⲧⲉ ⲛⲥⲉⲉⲓ ⲛ̄ϭⲓ ⲛⲓⲁⲧⲥⲃⲃⲉ ⲛ̄ⲥⲉⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ ⲁⲩⲱ ⲛⲥⲉⲥⲱⲃⲉ ⲛ̄ⲥⲱⲓ̈· ⲡⲉⲧϥⲓ ⲇⲉ ⲛ̄ⲛϥ̄ⲥⲕⲉⲩⲏ ⲙ̄ⲡϥⲟⲩⲱϣ ⲉⲙⲟⲩⲟⲩⲧ· ⲉⲙⲙⲟϥ ϫⲉ ⲁϥⲣ̄ϩⲟⲧⲉ ⲉⲙⲁⲧⲉ· ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥϥⲓ ⲛⲧⲉϥⲥⲏⲃⲉ ⲙⲁⲩⲁⲁϥ ⲁϥⲛⲟϫϥ̄ ⲉϩⲣⲁⲓ ⲉϫⲱⲥ· ⲁϥⲙⲟⲩ

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲙⲡⲉⲧϥⲉⲓ ϩⲁ ⲛⲉϥⲥⲕⲉⲩⲏ ϫⲉ ⲧⲱⲕⲉⲙ ⲛⲧⲉⲕⲥⲏⲃⲉ ⲛⲅⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ·

ⲡⲉⲧϥⲓ ⲇⲉ ϩⲁ ⲛⲉϥⲥⲕⲉⲩⲏ ⲙⲡⲉϥⲟⲩⲱϣ ⲉⲉⲛ ⲧⲟⲟⲧϥ· ⲉϫⲱϥ ⲉⲙⲟⲟⲩⲧϥ ⲁ ⲥⲁⲟⲩⲗ ϫⲓ ⲛⲧⲉϥⲥⲏⲃⲉ· ⲁϥⲛⲟϫϥ ⲉϫⲱⲥ ⲁϥⲙⲟⲩ

In the third clause, SaM reads ⲕⲟⲟⲛⲥ̅ⲧ̅ ‘pierce, slay’ which is used to render both the majority reading ἀποκεντέω and ἐκκεντέω present in Lucianic manuscripts.184 SaA offers two common verbs ⲛⲅ̄ ⲉⲓ ⲛⲅⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ ‘come and kill me’. SaB has ⲛⲅⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ ‘kill me’ like SaA, omitting ‘come’. No Greek manuscripts attest two verbs, and thus, SaA apparently has a free formulation here. The dependence of SaB on SaA remains uncertain since the verb used would be expected as the gloss for the more specific ⲕⲱⲛⲥ without necessitating reliance on a text like in SaA. SaB twice offers ⲡⲉⲧϥⲉⲓ ϩⲁ ⲛⲉϥⲥⲕⲉⲩⲏ ‘he who takes care of his equipment (= his armour-bearer)’ with the preposition ϩⲁ against SaMA which read ⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲏⲩⲉ ‘he who takes his equipment (= his armour-bearer)’. The different preposition does not alter the meaning but shows that the text of SaB has minor alterations pertaining to the style or idiolect of its copyist. SaM uses the conjunction ⲙⲏⲡⲱⲥ in this verse where SaA has ⲙⲏⲡⲟⲧⲉ. In the Greek tradition, the readings μή (majority) and μήποτε exist. Either the

184 ἐκκεντέω L-82. Crum, Dictionary, 112a.

206

Affiliations of the Sahidic manuscripts and the Greek Vorlage

Sahidic translator used ⲙⲏⲡⲱⲥ/ⲙⲏⲡⲟⲧⲉ to render μή185 or Sahidic manuscripts have picked up the conjunction from a variant Greek text.186 In the next clause, SaMA both contain the verb ⲙⲟⲩⲟⲩⲧ ‘to kill’, SaA with the preposition ⲙⲙⲟⲓ and SaM without the preposition, using the status pronominalis of the verb.187 Possibly the Sahidic manuscripts agree with L-82 which read ἀποκτείνωσιν. Perhaps the translator misread the source text since ἀποκτείνωσιν and ἀποκεντήσωσιν are graphically similar and contextually appropriate. In the next sentence, SaM adds that the armour-bearer did not want ‘to kill him’ ⲉⲙⲟⲟⲩⲧϥ.̄ SaA says the same, again with the preposition, ⲉⲙⲟⲩⲟⲩⲧ ⲉⲙⲙⲟϥ, while SaB has independently ⲉⲉⲛ ⲧⲟⲟⲧϥ ⲉϫⲱϥ ⲉⲙⲟⲟⲩⲧϥ ‘to bring his hand upon him to kill him’. Αfter ὁ αἴρων τὰ σκεύη αὐτοῦ, Greek manuscript 44 reads σπάσαι τὴν ρομφαίαν αὐτοῦ καὶ ἐκκέντησαι αὐτόν ‘to draw his sword and pierce him’. The Greek and Sahidic additions, however, are fundamentally different enough to eliminate any possible relationship between one another. The last sentence contains remarkable differences between the Sahidic witnesses. The beginning is rendered in SaMA with the conjunction ⲁⲩⲱ and Saul’s name in extraposition with the past tense and a 3rd person infix, against SaB that has the past tense without conjunction and the proper name as the subject. The verb in SaMB is ϫⲓ ‘to receive, take’ while SaA has ϥⲓ ‘to bear, carry, take’. However, this last-mentioned feature is not significant since the meaning and orthography are so similar that both could have resulted from inner-Sahidic variation. The Greek verb λαμβάνω is regularly rendered by ϫⲓ and ϥⲓ throughout 1 Sam and other Sahidic translations.188 The penultimate clause is ⲁϥⲃⲟϭϥ ⲉϫⲱⲥ ‘he leapt on it’ in SaM, whereas AB Sa preserve a different verb, ⲁϥⲛⲟϫϥ̄ ⲉϩⲣⲁⲓ ⲉϫⲱⲥ ‘he threw himself on it’ in SaA and similarly ⲁϥⲛⲟϫϥ ̄ ⲉϫⲱⲥ in SaB. Notably, SaA attests a longer reading than SaM, as was the case already twice within this same verse. Thus, the tendency of SaA to abbreviate the text, although obvious in some verses, is not entirely consistent. In these cases, one must remember that the readings of SaA are here compared with SaM, and this manuscript (SaM) also has some secondary readings. The last verb in this verse attests a significant similarity between SaA and B Sa . Both have ⲁϥⲙⲟⲩ ‘he died’ against SaM and the unanimous Greek tradition. If the word ⲁϥⲙⲟⲩ was in the original Sahidic translation, then it does

185 In verse 9: 5, there are similar variants. Greek reads μή and Sahidic witnesses as above. SaM ⲙⲏⲡⲱⲥ and SaA ⲙⲏⲡⲟⲧⲉ. Greek variants do not affect verse 9: 5, since μήποτε is present in only V f. 186 Hasznos (Graeco-Coptica, 20) has not found any passage where ⲙⲏⲡⲱⲥ/ⲙⲏⲡⲟⲧⲉ occurs in Sahidic without μήποτε in the Greek source text. She has analyzed Coptic Matthew, John and Romans (ibid., X). 187 The suffix of the 1st person singular is ø here. 188 Crum, Dictionary, 620.

207

Textual analyses

not necessitate a dependence between SaA and SaB in a secondary reading. Therefore, it would be necessary to presuppose abbreviation according to the Greek in SaM or only a free wording. On the other hand, if the reading with ⲁϥⲙⲟⲩ is secondary, then this shared feature of SaAB might demonstrate a dependence between them. In the end of the next verse, the same words are used to describe the destiny of the armour-bearer, ‘He leapt on his sword and died’. Perhaps verse five motivated the addition in SaAB in this verse, and this addition could have happened independently in both Sahidic manuscripts. 31: 5 καὶ εἶδεν ὁ αἴρων τὰ σκεύη αὐτοῦ ὅτι τέθνηκεν Σαούλ, καὶ ἐπέπεσεν καὶ αὐτὸς ἐπὶ τὴν ῥομφαίαν αὐτοῦ καὶ ἀπέθανεν μετ᾽αὐτοῦ· SaM

SaA

SaB

ⲛⲧⲉⲣⲉϥⲛⲁⲩ ⲇⲉ ⲛϭⲓ ⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲏⲩⲉ̄· ϫⲉ ⲁ ⲥⲁⲟⲩⲗ ⲙⲟⲩ· ⲛⲧⲟϥ ϩⲱⲱϥ ⲁϥⲃⲟϫϥ̄ ⲉ̄ϫⲉⲛ ⲧⲉϥⲥⲏϥⲉ· ⲁϥⲙⲟⲩ ⲛⲙ̄ⲙⲁϥ·

ⲛⲧⲉⲣϥ̄ⲛⲁⲩ ⲇⲉ ⲛϭⲓ ⲡⲉⲧϥⲓ ⲛ̄ⲛϥ̄ⲥⲕⲉⲩⲏ ϫⲉ ⲁ ⲥⲁⲟⲩⲗ ⲙⲟⲩ ⲛⲧⲟϥ ϩⲱⲱϥ ⲟⲛ ⲁϥⲛⲟϫϥ ⲉϫⲛ ⲧⲉϥⲥⲏⲃⲉ ⲁϥⲙⲟⲩ ⲛⲙ̄ⲙⲁϥ·

ⲛⲧⲉⲣⲉϥⲛⲁⲩ ⲛϭⲓ ⲛⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲉⲩⲏ ϫⲉ ⲁ ⲥⲁⲟⲩⲗ ⲙⲟⲩ· ⲁϥϥⲟϭϥ ⲉϫⲉⲛ ⲧⲉϥⲥⲏⲃⲉ ⲁϥⲙⲟⲩ ⲛⲙⲙⲁϥ·

In the first clause, SaMA agree, against SaB, in offering the particle ⲇⲉ. In the next sentence, SaMA read ⲛⲧⲟϥ ϩⲱⲱϥ ‘he himself’ and SaA continues with ⲟⲛ ‘still, further’. SaB lacks equivalent phrases. καὶ αὐτός is the usual equivalent for ⲛⲧⲟϥ ϩⲱⲱϥ, but no parallel for ⲟⲛ is found in the Greek tradition. Either this belonged to the original Coptic translation and SaM omitted it, or SaA added this word among many other free renderings. The same possibility applies in the following verse. For the penultimate clause, three different renderings appear in Sahidic: ⲁϥⲃⲟϫϥ̄ ⲉϫ ̄ ⲉⲛ ⲧⲉϥⲥⲏϥⲉ ‘he leapt on his sword’ in SaM, ⲁϥⲛⲟϫϥ ⲉϫⲛ ⲧⲉϥⲥⲏⲃⲉ ‘he threw himself on his sword’ in SaA, and ⲁϥϥⲟϭϥ ⲉϫⲉⲛ ⲧⲉϥⲥⲏⲃⲉ ‘he leapt on his sword’ in SaB.189 The reading of SaA could originate through erroneous copying, from ⲃⲟϫϥ to ⲛⲟϫϥ. Possibly SaA intentionally chose the same verb ⲛⲟϫϥ that was used in the previous verse. In the previous verse, SaAB had a similar wording, but this does not continue in this verse, SaB agreeing with SaM here. No connection to Greek variants is evident, since all the Sahidic witnesses seem to suppose (ἐπι)πίπτω.

189 There are two orthographical variants between SaM and SaB. Crum (Dictionary, 625b) offers examples of variation between ϥo- and ⲃⲟ- in Sahidic. The other variation, ϫ and ϭ seems to belong to Fayyumic features in SaM since ϫ is used only in Bohairic and Fayyumic forms. Notably, in the previous verse, SaM uses the Sahidic form ⲁϥⲃⲟϭϥ.

208

Affiliations of the Sahidic manuscripts and the Greek Vorlage

31: 6 καὶ ἀπέθανεν Σαοὺλ καὶ οἱ τρεῖς υἱοὶ αὐτοῦ καὶ ὁ αἴρων τὰ σκεύη αὐτου ἐν τῇ ἡμέρᾳ ἐκείνῃ κατὰ τὸ αὐτό. τὰ σκεύη αὐτου] + και παντες οι ανδρες αυτου A O L-82 CII-242 s-64´ 130 (=MT)

SaM

SaA

SaB

ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲙⲟⲩ ⲙⲛ̄ ⲡⲉϥϣⲟⲙⲧ̄ ⲛϣⲏⲣⲉ· ⲁⲩⲱ ⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲉⲩⲏ· ⲁⲩⲱ ⲛⲕⲉⲣⲱⲙⲉ ⲧⲏⲣⲟⲩ ⲉⲧⲛⲙ̄ⲙⲁϥ ⲁⲩⲙⲟⲩ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ·

ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲙⲟⲩ ⲙⲛ ⲡϥ̄ϣⲟⲙⲛ̄ⲧ̄ ⲛϣⲏⲣⲉ ⲁⲩⲱ ⲟⲛ ⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲉⲩⲏ ⲁϥⲙⲟⲩ· ⲁⲩⲱ ⲛⲕⲉⲣⲱⲙⲉ ⲧⲏⲣⲟⲩ ⲉⲧ ⲛⲙⲙⲁϥ ⲁⲩⲙⲟⲩ ⲙ̄ⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ·

ⲁ ⲥⲁⲟⲩⲗ ⲙⲟⲩ ⲙⲛ ⲓⲱⲛⲁⲑⲁⲙ ⲡⲉϥϣⲏⲣⲉ· ⲙⲛ ⲕⲉⲥⲛⲁⲩ ⲛⲧⲃⲁ ⲛⲣⲱⲙⲉ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙⲙⲁⲩ·

In the first clause, SaMA share the conjunction ⲁⲩⲱ, have ⲥⲁⲟⲩⲗ as an extraposited term and use a 3rd person singular infix as subject, against SaB which omits the conjunction and the proper name ⲥⲁⲟⲩⲗ as the subject. Further, SaMA have ⲁⲩⲱ ⲛⲕⲉⲣⲱⲙⲉ ⲧⲏⲣⲟⲩ ⲉⲧⲛⲙ̄ⲙⲁϥ ‘and all other men who were with him’ like A O L-82 CII-242 s-64,381,130 καὶ πάντες οἱ ἄνδρες αὐτοῦ (=MT). SaB is independent, stating that Saul died with his son Jonathan and 20,000 other men, but SaB also seems to depend on a text that tells about more men than only Saul, Jonathan and the armour-bearer. Against the Greek, but in accordance with each other, SaMA have the verb ⲁⲩⲙⲟⲩ ‘they died’ in connection with the approximation. This implies that this approximation entered the Sahidic at one time, not independently in SaA and SaM. SaA alone reads the verb ⲁϥⲙⲟⲩ ‘he died’ after ⲡⲉⲧϥⲓ ⲛⲛⲉϥⲥⲕⲉⲩⲏ ‘his armour-bearer’ and in the same clause, the particle ⲟⲛ after the conjunction ⲁⲩⲱ. Thus, in this verse, SaA does not abbreviate the text as is often the case, but produces a longer text than SaM. SaB has an independent wording that also differs consistently from other Sahidic witnesses as well as from the Greek witnesses. Sahidic witnesses omit the last phrase κατὰ τὸ αὐτό without similar Greek variants. 31: 7 καὶ εἶδον οἱ ἄνδρες Ἰσραὴλ οἱ ἐν τῷ πέραν τῆς κοιλάδος καὶ οἱ ἐν τῷ πέραν τοῦ Ἰορδάνου ὅτι ἔφυγον οἱ ἄνδρες Ἰσραήλ, καὶ ὅτι τέθνηκεν Σαοὺλ καὶ οἱ υἱοὶ αὐτοῦ, καὶ καταλείπουσιν τὰς πόλεις αὐτῶν καὶ φεύγουσιν· καὶ ἔρχονται οἱ ἀλλόφυλοι καὶ κατοικοῦσιν ἐν αὐταῖς. SaM

SaA

SaB

ⲁⲩⲛⲁⲩ ⲇⲉ ⲛϭⲓ ⲛⲉⲣⲱⲙⲉ ⲙⲡⲓⲏⲗ̄ ·̅ ⲛⲉⲧϩⲓ ⲡⲉⲕⲣⲟ ⲙⲡⲓⲁ. ⲁⲩⲱ ⲛⲉⲧϩⲓ ⲡⲉⲕⲣⲟ ⲙⲡⲓⲟⲣⲇⲁⲛⲏⲥ̄ ϫⲉ ⲛⲣⲱⲙⲉ ⲙⲡⲓⲏ̄ⲗ ̄ ⲁⲩϭⲱⲧⲡ̄ ⲁⲩⲱ ϫⲉ ⲁϥⲙⲟⲩ̄ ⲛϭⲓ ⲥⲁⲟⲩⲗ· ⲙⲛ̄

ⲁⲩⲛⲁⲩ ⲇⲉ ⲛϭ ̄ ⲓ ⲛⲣⲱⲙⲉ ⲙⲡⲓⲏⲗ̄ ̄ ⲛⲉⲧ ϩⲛ ⲡⲉⲕⲣⲟ ⲙⲡⲉⲓⲁ ⲁⲩⲱ ⲛⲉⲧ ϩⲛ ⲡⲉⲕⲣⲟ ⲙ̄ⲡⲓⲟⲣⲇⲁⲛⲏⲥ· ϫⲉ ⲛ̄ⲣⲱⲙⲉ ⲙⲡⲓⲏ̄ⲗ ̄ ⲁⲩⲡⲱⲧ ⲁⲩⲱ ϫⲉ ⲥⲁⲟⲩⲗ ⲁϥⲙⲟⲩ ⲙⲛ̄

The whole verse is missing.

209

Textual analyses ⲛⲉϥϣⲏⲣⲉ· ⲁⲩⲕⲁ ⲛⲉⲩⲡⲟⲗⲓⲥ ⲉⲃⲟⲗ· ⲁⲩⲡⲱⲧ ϩⲱⲟⲩ· ⲁⲩⲉⲓ ⲛϭⲓ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲟⲩⲱϩ ⲛϩ̄ ⲏⲧⲟⲩ.

ⲛϥϣⲏⲣⲉ· ⲁⲩⲕⲁ ⲛⲉⲡⲟⲗⲓⲥ ⲉⲃⲟⲗ· ⲁⲩⲡⲱⲧ ϩⲱⲟⲩ· ⲁⲩⲉⲓ ⲛϭⲓ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲟⲩⲱϩ ⲛϩ̄ ⲏⲧⲟⲩ·

In the first sentence, SaM uses the preposition ϩⲓ twice whereas SaA has the preposition ϩⲛ. There is no difference in the meaning in this case. Both prepositions render the Greek expression ἐν τῷ πέραν τῆς κοιλάδος. In the next clause, the verbs differ. SaM reads ⲁⲩϭⲱⲧⲡ̄ ‘to become defeated, overcome’ and SaA reads ⲁⲩⲡⲱⲧ ‘to flee’. Crum cites both verbs as equivalents for φεύγω.190 In the Sahidic New Testament, ϭⲱⲧⲡ̄ occurs only twice, rendering ἡττᾶσθαι and κλίνειν.191 The verb used in SaA, ⲡⲱⲧ, occurs in the NT 15 times rendering φεύγω and 11 times rendering τρέχω.192 Thus, clearly, the verb used in SaA is the most common equivalent in the Sahidic NT suggesting that SaA has changed a less common verb to a more common one.193 τέθνηκεν Σαούλ has different formulations in the Sahidic manuscripts. SaM uses a past tense with ⲛϭⲓ while SaA places the proper noun in extraposition before the verbal form. SaM parallels the Greek ‘their cities’, whereas SaA has ‘the cities’. This minor difference can easily emerge in either direction, from the possessive ⲛⲉⲩ- to definite ⲛⲉ- or the other way round.194 31: 8 καὶ ἐγενήθη τῇ ἐπαύριον, καὶ ἔρχονται οἱ ἀλλόφυλοι ἐκδιδύσκειν τοὺς νεκροὺς· καὶ εὑρίσκουσιν τὸν Σαοὺλ καὶ τοὺς τρεῖς υἱοὺς αὐτοῦ πεπτωκότας ἐπὶ τὰ ὄρη Γελβοῦε. νεκρούς] τραυματιας L-82 f 244 (=MT)

SaM

SaA

SaB

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲙⲡⲉϥⲣⲁⲥⲧⲉ ⲁⲩⲉⲓ ⲛ̄ϭⲓ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩⲃⲱϣ ⲛⲛⲉⲧⲙⲟⲟⲩⲧ· ⲁⲩϩⲉ ⲇⲉ ⲉⲥⲁⲟⲩⲗ ⲙⲛ̄ ⲡⲉϥϣⲟⲙⲧ ⲛϣⲏⲣⲉ̄· ⲉⲩⲣⲁϩⲧ·̄ ϩⲙ̄ ⲡⲧⲟⲟⲩ ⲛⲕⲉⲗⲃⲟⲩⲉ̄·

ⲁⲥϣⲱⲡⲉ ⲇⲉ ⲙⲡⲉϥⲣⲁⲥⲧⲉ ⲁⲩⲉⲓ ⲛϭⲓ ⲛⲁⲗⲗⲟⲫⲩⲗⲟ[ⲥ· ⲁⲩ]ⲃⲱϣ ⲛ̄ⲛⲉⲧⲙⲟⲟⲩⲧ [ⲁⲩϩⲉ] ⲉⲥⲁⲟⲩⲗ ⲙⲛ̄ ⲡϥ̄ϣⲟ[ⲙⲛ̄ⲧ]̄ ⲛϣⲏⲣⲉ ⲉⲩⲣⲁϩⲧ ϩⲙ ⲡ[ⲧⲟ]ⲟⲩ ⲛⲕⲉⲗⲃⲟⲩⲉ·

ⲙⲡⲉϥⲣⲁⲥⲧⲉ ⲇⲉ ⲁⲩⲉⲓ ⲛϭⲓ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩⲃⲉϣ ⲛⲉⲛⲧⲁⲩⲙⲟⲩ· ⲁⲩϩⲉ ⲉⲥⲁⲟⲩⲗ ⲙⲛ ⲓⲱⲛⲁⲑⲁⲙ ⲡⲉϥϣⲏⲣⲉ· ⲙⲛ ⲡⲗⲁⲟⲥ ⲉⲧ ⲛⲙⲙⲁϥ· ⲉⲩⲣⲁϩⲧ ϩⲁⲧⲙ ⲡⲧⲟⲟⲩ ⲛⲅⲉⲗⲃⲟⲩⲉ̄·195

190 Crum, Dictionary: ϭⲱⲧⲡ̄ 833b, ⲡⲱⲧ 274a. 191 Wilmet, Concordance II: 3, 1721. 192 Wilmet, Concordance II: 2, 638. 193 There is some graphical similarity in ⲁⲩϭⲱⲧⲡ̄ and ⲁⲩⲡⲱⲧ, and, theoretically, the change could happen unintentionally in SaA. 194 See also Mink, “Koptischen Versionen”, 221–2. 195 Concerning the last word, a marginal note lies at the bottom of the page in the manuscript: ⲁⲇⲁⲙ⸗ ⲉⲧⲁⲓ ⲉⲓⲙⲟϣⲉ ⲙⲙⲙⲉⲣⲉ ⲛⲥⲁϥ· ϩⲓϫⲛ ⲡⲧⲟⲟⲩ ⲉⲕⲉⲗϥⲟⲩⲉ.

210

Affiliations of the Sahidic manuscripts and the Greek Vorlage

In the first clause, SaB omits a redundant verb against SaMA that have it. This same phenomenon, omission of ⲁⲥϣⲱⲡⲉ ‘it became’ when rendering καὶ ἐγενήθη/καὶ ἐγένετο, occurs in other Sahidic witnesses within other passages.196 In the third clause, SaMA share the rendering ⲁⲩⲃⲱϣ ⲛⲛⲉⲧⲙⲟⲟⲩⲧ ‘they stripped the dead’ against ⲁⲩⲃⲉϣ ⲛⲉⲛⲧⲁⲩⲙⲟⲩ in SaB, but there is no difference in the meaning. All Sahidic witnesses attest a text independent of the variant Greek reading according to which ‘they came to strip the wounded’. In the same line, SaM has the connective ⲇⲉ, whereas SaAB omit it, as is often the case in these two manuscripts. The list of dead in SaMA contains Saul and his three sons while SaB lists Saul, Jonathan, his son and the people who were with him. No Greek variation corresponds to the reading of SaB. Already in verse 31: 6, SaB reads ‘Jonathan’ against ‘three sons’ in other Sahidic and Greek witnesses. In verse 31: 6, SaB had a free formulation of 20,000 other men. Here, the formulation is ‘with people who were with him’. Interestingly, SaAM do not mention the men that were with Saul (present in verse 31: 6), and this suggests that the Hexaplaric approximation in verse 31: 6 entered the Sahidic text after the original Sahidic translation. In the last sentence, the preposition is ϩⲙ̄ ‘in, at, on’ in SaMA but ϩⲁⲧⲙ ‘beside, with’ in SaB. In the Greek, variation exists between the prepositions ἐπί and ἐν as well as between singular and plural forms of ‘mountain’. Both Coptic prepositions can render both Greek prepositions.197 31: 9 καὶ ἀποκεφαλίζουσιν αὐτὸν καὶ ἐξέδυσαν τὰ σκεύη αὐτοῦ, καὶ ἀποστέλλουσιν αὐτὰ εἰς γῆν ἀλλοφύλων κύκλῳ, εὐαγγελιζόμενοι τοῖς εἰδώλοις αὐτῶν καὶ τῷ λαῷ αὐτῶν. ἀποκεφαλίζουσιν Mmg L-82 f 554mg] αποστρεφουσιν rel σκεύη αὐτοῦ]+ καὶ ἀποκόπτουσι τὴν κεφαλήν αὐτοῦ M V 82 CII a d s 71 158

244 342 460 707mg αὐτά A B O CI 121 29 71 318] > rel SaM

SaA

SaB

ⲁⲩⲡⲟⲟⲛⲉϥ ⲁⲩⲃⲟϣϥ̄ ⲛⲛⲉϥⲥⲕⲏⲩⲉ̄· ⲁⲩϫⲟⲟⲩ ⲇⲉ ⲉⲃⲟⲗ ⲙⲡⲉⲩⲕⲱⲧⲉ ⲧⲏⲣϥ̄· ⲙⲛ̄ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲉⲩⲧⲁϣⲉⲟⲉⲓϣ ⲛ̄ⲛⲉⲩⲉ̅ⲓ ̅ⲇⲱⲗⲟⲛ ⲙⲛ̄ ⲡⲉⲩⲙⲏⲏϣⲉ

ⲁⲩⲡⲟⲟⲛⲉϥ ⲁⲩⲃⲟϣϥ ⲛ̄ⲛϥ̄ⲥⲕⲉⲩⲏ· ⲁⲩϥⲓ ⲛⲧϥⲁⲡⲉ· ⲁⲩϫⲟⲟⲩⲥ ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲱⲧⲉ ⲧⲏⲣϥ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲉⲩⲧⲁϣⲉⲟⲉⲓϣ ⲛⲛⲉⲩⲉⲓⲇⲱⲗⲟⲛ ⲙⲛ ⲡⲉⲩⲙⲏⲏϣⲉ·

ⲁⲩⲃⲟϣⲟⲩ ⲛⲛⲉⲩϩⲟⲉⲓⲧⲉ ⲙⲛ ⲛⲉⲩⲥⲕⲉⲩⲏ ⲙⲙⲓϣⲉ· ⲁⲩϫⲟⲟⲥⲟⲩ ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ·

196 Thus 3: 9 and 25: 20, for example. For detailed analyses, see my subsection 2.1.3.1.d). 197 Crum, Dictionary, 428b and 683a. Brooke/McLean (Old Testament, 105) cite Sahidic after L, thus making a connection on the basis of singular against plural. I would not make this citation since it is a stable expression that is used in Sahidic.

Textual analyses

211

In the first clause, SaMA offer the verb ‘to turn’ whereas SaB is the only manuscript that omits it. SaB adds ‘their clothes and’ against other witnesses as well as an attribute ‘battle’ after ⲥⲕⲉⲩⲏ ‘equipment’. The number and person also differ in the beginning: SaMA say ‘they took his equipment’ but SaB offers the plural ‘they took their clothes and their battle equipment’. In the Greek, 44–107–610 read ἐξέδυσαν αὐτοὺς καὶ τὰ σκεύη αὐτοῦ, thus resembling the reading of SaB that has two conjoined objects for the verb. However, the reading of SaB is more specific in describing that they stripped off their clothes, and the last possessive pronoun is plural, not singular as in those Greek manuscripts. SaB does not describe the decapitation mentioned in Greek, but continues directly to the sending of the clothes and equipment. Probably, SaB has a peculiar wording that is independent of any Greek source text. Only this manuscript omits κύκλῳ εὐαγγελιζόμενοι τοῖς εἰδώλοις αὐτῶν καὶ τῷ λαῷ αὐτῶν. After the equivalent of τὰ σκεύη αὐτοῦ, SaA presupposes the reading καὶ ἀποκόπτουσι τὴν κεφαλήν αὐτοῦ attested by Greek M V 82 CII a d s 29 71 158 244 342 460 707mg. The secondary nature of this phrase in the LXX becomes obvious from the way its place varies within the manuscripts and their margins. This clause is a Hexaplaric approximation towards the ΜΤ198 and occurs either before καὶ ἀποστρέφουσιν199 or after τὰ σκεύη αὐτοῦ as in SaA and M V 82 CII a d s 71 158 244 342 460 707mg. My hypothesis is that the OG translation was ἀποκεφαλίζουσιν, now present in Mmg L-82 f 554mg; καὶ ἀποστρέφουσιν ‘they turned’ must be an early mistake since the reading has diffused so widely among the manuscripts. Only the base text of the Lucianic recension has escaped this error. An orthographic similarity lies between ἀποκεφαλίζουσιν and ἀποστρέφουσιν. In addition to this similarity, the use of the verb ἀποκεφαλίζουσιν meant that there was no need to mention the object, since it is part of the verb. The erroneous form ἀποστρέφουσιν was therefore in the source text of the Sahidic translator. Origen found no equivalent for the Hebrew reading ‘cut his head’ in his Greek text and added one. Thus, a doublet entered into Greek manuscripts that were corrected according to the Hexapla. SaM attests to the OG reading in its corrupt form whereas SaA attests a doublet with the Hexaplaric approximation and thus presupposes here some kind of revision with Greek after the original translation. Next, the text tells of sending, and the thing(s) sent depend on the previous sentence. According to SaA, the head was sent, but according to SaM, ‘they sent out … proclaiming’, and thus SaM seems to presuppose a Greek Vorlage without αὐτά. SaB reads ‘they sent them’. This refers to the equipment, since the bodies are brought to the city in the next verse. The question is whether SaB follows those Greek manuscripts that read with αὐτά. Prob-

198 ‫ויכרתו את־ראשו‬ 199 Thus Greek manuscript 29.

212

Affiliations of the Sahidic manuscripts and the Greek Vorlage

ably, however, SaB has the suffix because Coptic transitive verbs usually have explicit objects. This verb may also function intransitively, but according to Crum only occasionally.200 SaM renders this sending-sentence with the particle ⲇⲉ while SaAB lack ⲇⲉ. All Greek manuscripts attest καί. In 1 Sam, both ⲇⲉ and asyndeton often render καί. Therefore, translation technique offers no arguments in this case. However, considering the characteristics of these Sahidic manuscripts, SaAB have probably omitted the connective because of their tendencies to abbreviate. SaM reads ⲙⲡⲉⲩⲕⲱⲧⲉ ⲧⲏⲣϥ·̄ ⲙⲛ̄ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ‘all round them and the land of the Philistines’, thus preserving equivalents to both εἰς γῆν ἀλλοφύλων and κύκλῳ; SaA reads ϩⲙ ⲡⲕⲱⲧⲉ ⲧⲏⲣϥ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ‘all round the Philistines’, without ‘land’ and SaB ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ‘out of the land of the Philistines’, omitting any equivalent for κύκλῳ and the rest of the clause. Since no equivalent Greek variants exist, these omissions probably stem from inner-Sahidic variation. SaAB, which typically shorten the text, have both independently abbreviated in this case. 31: 10 καὶ ἀνέθηκαν τὰ σκεύη αὐτοῦ εἰς τὸ Ἀσταρτεῖον, καὶ τὸ σῶμα αὐτοῦ κατέπηξαν ἐν τῷ τείχει Βαιθσάμ. SaM

SaA

SaB

ⲛⲉϥⲥⲕⲏⲩⲉ̄ ⲇⲉ ⲁⲩⲕⲁⲁⲩ̄ ϩⲙ̄ ⲡⲉⲩⲙⲁ ⲛⲉⲓⲇⲱⲗⲟⲛ· ⲁⲩⲱ ⲡⲉϥⲥⲱⲙⲁ· ⲁⲩⲟⲃⲧϥ̄· ⲉⲡⲥⲟⲃⲧ·̄ ⲛⲃⲉⲧⲥⲁⲙ·

ⲡⲉϥⲥⲕⲉⲩⲏ ⲇⲉ ⲁⲩⲕⲁⲁⲩ ϩⲙ ⲡⲉⲩⲙⲁ ⲛⲉⲓⲇⲱⲗⲟⲛ· ⲁⲩⲱ ⲡϥⲥⲱⲙⲁ ⲁⲩⲟⲩⲟⲧⲃⲉⲃ ⲉⲃⲟⲗ ϩⲙ ⲡⲥⲟⲃⲧ ⲛⲃⲉⲑⲗⲉⲉⲙ·

ⲡⲥⲱⲙⲁ ⲇⲉ ⲛⲥⲁ̈ ⲟⲩⲗ ⲙⲛ ⲓⲱⲛⲁⲑⲁⲙ ⲡⲉϥϣⲏⲣⲉ· ⲙⲛ ⲡⲉϥϥⲓ ϩⲁ ⲛⲉⲩⲥⲕⲉⲩⲏ· ⲁⲩϥⲓⲧⲟⲩ ⲉϩⲣⲁⲓ ⲉⲧⲡⲟⲗⲓⲥ· ⲁⲩⲁϣⲧⲟⲩ ⲉⲡⲥⲟⲃⲧ ⲛⲃⲩⲑⲥⲁⲙⲏⲥ·

SaB differs from other witnesses by adding that the bodies were brought into the city. Probably, SaB has formulated this freely, improving upon the narrative discourse. In the previous verse, SaB had already noted that their equipment was sent to the land of the Philistines; thus, at the beginning of this verse, SaB does not mention the equipment again, against SaAM and most Greek witnesses. Only manuscript 246 omits the clause καὶ ἀνέθηκαν τὰ σκεύη αὐτοῦ εἰς τὸ Ἀσταρτεῖον, but this Greek omission is best explained as a homoiarchon-based skip of eye from the first to the second καί. In the last sentence, SaM preserves the verb ⲱϥⲧ/ⲱⲃⲧ ‘to fix, nail’, which corresponds to καταπήγνυμι. SaA has a unique wording with the verb ⲟⲩⲱⲧⲃ ‘to remove, transport’, thus stating ‘They removed his body from

200 Crum, Dictionary, 793a.

213

Textual analyses

the wall.’ SaA is obviously erroneous here since the next verse clearly states that the body was removed from the wall, and its mention here confuses the discourse. According to this erroneous reading, the body had never been placed on the wall, but had twice been removed from it. SaA has many copying mistakes in other passages, and this instance is best explained as an error from ⲟⲃⲧϥ̄ to ⲟⲩⲟⲧⲃⲉⲃ; confusion related to the clusters of ⲃ and ⲧ and an addition of ⲟⲩ. The last letter is ϥ in SaM but ⲃ in SaA. This change, reading ⲃ instead of ϥ, fits with the many other features from SaM that relate to its Fayyumic provenance. SaA comes from the White Monastery and we know that many codices that come from this monastery were produced in Fayyum.201 SaB has the verb ⲉⲓϣⲉ ‘to hang, suspend’, which differs from ⲱϥⲧ ‘to nail’ in SaM. The verb ⲱϥⲧ occurs only once in the Sahidic New Testament, to render προσηλόω ‘to nail’.202 The verb ⲉⲓϣⲉ occurs in the Sahidic NT five times: twice for κρεμάννυμι ‘to hang’, twice for σταυρόω ‘to crucify’ and once for προσπήγνυμι ‘to nail to’.203 Thus, the verb used in SaB is a more common one with respect to the pattern of the Sahidic New Testament. This fits well in the picture of SaB as a chronicle-like story. The suffix of the verb is singular in SaMA but plural in SaB. SaB independently describes several bodies without any Greek source. The place is ⲃⲉⲧⲥⲁⲙ in SaM as in the majority of Greek manuscripts, but ⲃⲩⲑⲥⲁⲙⲏⲥ in SaB.204 The reading ⲃⲉⲑⲗⲉⲉⲙ in SaA probably results from an error, either from ⲃⲉⲧⲥⲁⲙ or ⲃⲩⲑⲥⲁⲙⲏⲥ. In verse 31: 12, SaA attests to ⲃⲩⲑⲥⲁⲙⲏⲥ. Here, as elsewhere, this manuscript offers a problematic and confusing reading. Some obvious mistakes relate to parallel passages that have become corrupt and that remain uncorrected. 31: 11 καὶ ἀκούουσιν οἱ κατοικοῦντες Ἰαβις τῆς Γαλααδίτιδος ἃ ἐποίησαν οἱ ἀλλόφυλοι τῷ Σαούλ. SaM

SaA

SaB

ⲛⲉⲧⲟⲩⲏϩ ⲇⲉ ϩⲛ̄ ⲁⲃⲓⲥ ⲛⲧⲉ ⲅⲁⲗⲁⲁⲇ ⲁⲩⲥⲱⲧⲙ̄ ⲉⲛⲉⲛⲧⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲁⲩ ⲛⲥⲁⲟⲩⲗ

ⲛⲉⲧⲟⲩⲏϩ ⲇⲉ ϩⲛ ⲓⲁⲃⲓⲥ ⲛ̄ⲧⲉ ⲕⲁⲗⲁⲁⲧ ⲁⲩⲥⲱⲧⲙ ⲉⲛⲉⲛⲧⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲁⲩ ⲛⲥⲁⲟⲩⲗ

ⲛⲣⲱⲙⲉ ⲇⲉ ⲙⲡⲓⲏ̄ⲗ ̄ ⲉⲧⲟⲩⲏϩ ϩⲉⲛ ⲓⲁⲃⲓⲥ ⲛⲧⲉ ⲅⲁⲗⲁⲁⲇ ⲁⲩⲥⲱⲧⲙ ⲉⲛⲉⲛⲧⲁ ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲁⲩ ⲛⲥⲁⲟⲩⲗ·

In this verse, three Sahidic witnesses share essentially the same text. SaB adds an explanation that those living in Iabis are ‘men of Israel’. No similar variants exist in the Greek tradition.

201 H.N. Takla, “Biblical Manuscripts”, 160–2 202 Col 2: 14. 203 Wilmet, Concordance II: 1, 242. 204 βαιθσαμυς 92* a-527; βεθσαμυς 527; βαιθσαμαν 82; βαιθσαμη 342.

214

Affiliations of the Sahidic manuscripts and the Greek Vorlage

The place is ⲁⲃⲓⲥ in SaM but ⲓⲁⲃⲓⲥ in SaAB. Without any Greek manuscript omitting the first ⲓ,205 I see SaM as resulting from a minor error, either in hearing or writing. SaM has probably not experienced a systematic revision, but this is not completely certain. ⲅⲁⲗⲁⲁⲇ and ⲕⲁⲗⲁⲁⲧ are essentially the same since Sahidic witnesses of 1 Sam continuously alternate ⲅ/ⲕ and ⲧ/ⲇ.206 31: 12 καὶ ἀνέστησαν πᾶς ἀνὴρ δυνάμεως καὶ ἐπορεύθησαν ὅλην τὴν νύκτα, καὶ ἔλαβον τὸ σῶμα Σαοὺλ καὶ τὸ σῶμα Ἰωναθὰν τοῦ υἱοῦ αὐτοῦ ἀπὸ τοῦ τείχους Βαιθσάμ, καὶ φέρουσιν αὐτοὺς εἰς Ἰαβὶς καὶ κατακαίουσιν αὐτοὺς ἐκεῖ. SaM

SaA

SaB

ⲁⲩⲧⲱⲟⲩⲛ ⲛϭⲓ ⲛⲉⲣⲱⲙⲉ ⲧⲏⲣⲟⲩ ⲛϫⲱⲱⲣⲉ· ⲁⲩⲙⲟⲟϣⲉ ⲛⲧⲉⲩϣⲏ ⲧⲏⲣⲥ̄· ⲁⲩⲱ ⲁⲩϫⲓ ⲙⲡⲥⲱⲙⲁ ⲛⲥⲁⲟⲩⲗ· ⲁⲩⲱ ⲡⲥⲱⲙⲁ ⲛⲓⲱⲛⲁⲑⲁⲛ ⲡⲉϥϣⲏⲣⲉ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲥⲟⲃⲧ· ⲛⲃⲉⲑⲥⲁⲙ· ⲁⲩⲉⲛⲧⲟⲩ ⲉϩⲣⲁⲓ ⲉⲁⲃⲓⲥ· ⲁⲩⲱ ⲁⲩⲣⲟⲕϩⲟⲩ ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ·

ⲁⲩⲧⲱⲟⲩⲛ ⲛϭⲓ ⲛⲣⲱⲙⲉ ⲧⲏⲣⲟⲩ ⲛ̄ϫⲱⲱⲣⲉ ⲁⲩⲙⲟⲟϣⲉ ⲛⲧⲉⲩϣⲏ ⲧⲏⲣⲉⲥ ⲁⲩϫⲓ ⲙⲡⲥⲱⲙⲁ ⲛ̄ⲥⲁⲟⲩⲗ ⲁⲩⲱ ⲡⲥⲱⲙⲁ ⲛ̄ⲓⲱⲛⲁⲑⲁⲛ ⲡϥ̄ϣⲏⲣⲉ· ⲉⲃⲟⲗ ϩⲙ ⲡⲥⲟⲃⲧ ⲛ̄ⲃⲩⲑⲥⲁⲙⲏⲥ ⲁⲩⲏⲛⲧⲟⲩ (sic) ⲉϩⲣⲁⲓ ⲉⲓⲁⲃⲓⲥ ⲁⲩⲟⲣⲭⲉⲓ ⲉⲣⲟⲟⲩ ϩⲙ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ

ⲁⲩϫⲓ ⲛϩⲉⲛⲣⲱⲙⲉ ⲛϫⲱⲱⲣⲉ ϩⲉⲛ ⲧⲉⲩϭⲟⲙ ⲁⲩⲙⲟⲟϣⲉ ⲛⲧⲉⲩϣⲏ ⲧⲏⲣⲥ· ⲁⲩϥⲓ ⲙⲡⲥⲱⲙⲁ ⲛⲥⲁⲟⲩⲗ· ⲙⲛ ⲓⲱⲛⲁⲑⲁⲙ ⲡⲉϥϣⲏⲣⲉ: ⲁⲩⲛⲧⲟⲩ ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ· ⲁⲩⲉⲛⲧⲟⲩ ⲉϩⲣⲁⲓ ⲉⲡⲕⲁϩ· ⲙⲡⲓⲏ̄ⲗ ̄

In the first clause, Sahidic renders the men in plural like L 245s. However, this does not necessitate any connection, since the verbs are plural which easily explains the Sahidic plural subject.207 SaB renders the first sentence uniquely ‘They took men mighty in their power’ without any corresponding Greek variants. ⲁⲩⲱ ⲁⲩϫⲓ ⲙⲡⲥⲱⲙⲁ ⲛⲥⲁⲟⲩⲗ·ⲁⲩⲱ ⲡⲥⲱⲙⲁ ⲛⲓⲱⲛⲁⲑⲁⲛ ⲡⲉϥϣⲏⲣⲉ ‘And they took the body of Saul and the body of Jonathan, his son’ renders strictly καὶ ἔλαβον τὸ σῶμα Σαοὺλ καὶ τὸ σῶμα Ἰωναθὰν τοῦ υἱοῦ αὐτοῦ in SaM. SaA omits the first conjunction, but otherwise shares the same text as SaM. In SaB, the first verb is different, but is so similar in appearance and meaning as to be explained as a minor scribal error. More conscious activity is apparent in the emendation of ⲁⲩⲱ ⲡⲥⲱⲙⲁ ⲛⲓⲱⲛⲁⲑⲁⲛ ‘and the body of Jonathan’ into ⲙⲛ ⲓⲱⲛⲁⲑⲁⲙ ‘and Jonathan’, omitting the word ‘body’ before Jonathan. This minor change is an example of SaB abbreviating and freely altering its text. SaM preserves ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲥⲟⲃⲧ ⲛⲃⲉⲑⲥⲁⲙ ‘from the wall of Bethsam’. SaA shares the same reading, but with the place name ⲃⲩⲑⲥⲁⲙⲏⲥ. In the Greek tradition, the longer name appears as βαιθσαμυς a-527 121 and βεθσαμυς 527. SaB omits this phrase and renders the rest of this verse and the next verse independently without any corresponding Greek variant. 205 In verse 12, manuscript 246 omits the first ι. In this verse, however, as well as in verse 13, this same manuscript reads with ι. 206 Layton, Grammar, §39, describes these pairs as typically confused. 207 Greek A O 125 have the first verb in singular and the second in plural.

215

Textual analyses

The place is spelled ⲁⲃⲓⲥ in SaM but ⲓⲁⲃⲓⲥ in SaA and in the Greek as in the previous verse. The last sentence is ⲁⲩⲱ ⲁⲩⲣⲟⲕϩⲟⲩ ϩⲙ̄ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ ‘And they burned them in that place’ in SaM and ⲁⲩⲟⲣⲭⲉⲓ ⲉⲣⲟⲟⲩ ϩⲙ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ ‘they danced to them in that place’ in SaA. SaM renders Greek strictly, whereas SaA seems to result from a mistake in Sahidic transmission. The verb in SaA is ⲁⲩⲟⲣⲭⲉⲓ but this verb does not make sense, because the verb ὀρχέομαι means ‘to dance, leap, bound’. Possibly a change occurred from ⲕϩ to ⲭ, consonants that alternate in Sahidic. The change from ⲁⲩⲣⲟⲭⲟⲩ to ⲁⲩⲟⲣⲭⲉⲓ ⲉⲣⲟⲟⲩ requires further errors, but in chapter 29 such errors were present in SaA, and thus the explanation seems likely in this verse as well. 31: 13 καὶ λαμβάνουσιν τὰ ὀστᾶ αὐτῶν καὶ θάπτουσιν ὑπὸ τὴν ἄρουραν τὴν ἐν Ἰαβίς, καὶ νηστεύουσιν ἑπτὰ ἡμέρας. καὶ ἐγένετο μετὰ τὸ ἀποθανεῖν Σαοὺλ καὶ Δαυὶδ ἀνέστρεψεν τύπτων τὸν Ἀμαλήκ. SaM

SaA

SaB

ⲁⲩⲧⲟⲙⲥⲟⲩ ϩⲁ ⲧⲥⲓϯⲱϩⲉ ⲉⲧϩⲛ̄ ⲁⲃⲓⲥ· ⲁⲩⲱ ⲁⲩⲛⲏⲥⲧⲉⲩⲉ̄ ⲛⲥⲁϣϥ ⲛϩⲟⲟⲩ·

ⲁⲩⲧⲟⲙⲥⲟⲩ ϩⲁⲧⲛ ⲧⲥⲓ̈ϯⲱϩⲉ ⲛⲓⲁⲃⲓⲥ· ⲁⲩⲱ ⲁⲩⲛⲏⲥⲧⲉⲩⲉ ⲛ̄ⲥⲁϣϥ ⲛ̄ϩⲟⲟⲩ.

ⲁⲩⲧⲟⲙⲥⲟⲩ ϩⲙ ⲡⲙⲁ ⲉⲧⲙⲙⲁⲩ:

All three Sahidic witnesses attest to a homoiarchon mistake at the beginning of this verse. The first clause καὶ λαμβάνουσιν τὰ ὀστᾶ αὐτῶν has no equivalent in the Sahidic, probably caused by parablepsis from the first to the second καί. No similar Greek variants are extant, and therefore this similarity textually connects all three Sahidic witnesses, demonstrating that they have a common ancestor. This also shows that the Sahidic translations have not been systematically redacted against a Greek text. After the first verb, SaB renders the verse independently, omitting the place name and the mention of fasting. SaMA render the Greek strictly with only a little difference in the expression ‘under the ground’. Both manuscripts use the word ⲥⲓϯⲱϩⲉ ‘field’ which is a Fayyumic form of the word ⲥⲧⲉⲓⲱϩⲉ.208 SaMA do not render the last sentence of the Greek. In some manuscripts, this is the last verse of 1 Sam, while in others it is the first sentence in 2 Sam. SaMA arrange this verse as the first sentence in 2 Sam. Results This analysis has shown that one common ancestor stands behind all three Sahidic witnesses. This is the most compelling explanation for the similar 208 Crum, Dictionary, 89b. Actually Crum has the Fayyumic form written ⲥⲓⲧⲓⲱϩⲉ. The additional ⲓ between ⲥ and ⲧ distinguishes the Fayyumic and Sahidic forms.

216

Affiliations of the Sahidic manuscripts and the Greek Vorlage

wordings and shared mistakes encountered in the Sahidic manuscripts. In verse 31: 1, SaMA say that all the men of Israel were wounded and fell whereas according to the Greek ‘the wounded fell’, in 31: 6 Sahidic manuscripts omit κατὰ τὸ αὐτό, and in 31: 13 καὶ λαμβάνουσιν τὰ ὀστᾶ αὐτῶν is without a Sahidic equivalent. These examples show, in addition to their common origin, that no systematic revision according to a Greek text has been conducted on the Sahidic text. When examining the discrepancies between Sahidic witnesses, the biggest group consists of the cases where SaMA agree against SaB. Some of the differences are only small, concerning connective (31: 3, 4, 5, 6), number (31: 9, 10) or preposition (31: 4, 8). With respect to these minor issues, the original Sahidic reading is difficult to determine. The more remarkable variants offer more insight. Obviously, SaB is secondary in the following, prominent variants: different constructions (31: 3, 4, 6, 8), omissions in SaB (31: 5, 8, 9, 10, 13), and additions (31: 9, 10, 11). There are additional passages where the content is no longer the same, as in verses 31: 6, 8, and 12. There are passages where SaAB agree against SaM, but in comparison to the previous paragraph, these are few. Omission (in 31: 3, 8) of a connective and a paraphrase in SaAB, but not necessitating a dependence between SaAB. A different verb appears in 31: 4. This difference, however, concerns only one character, and, in the next verse, SaMB attest to the same verb as SaM here, against SaA. In the same verse, SaAB add the verb ‘he died’. A spelling difference also occurs (31: 11, 12). SaM has ⲁⲃⲓⲥ where SaAB have ⲓⲁⲃⲓⲥ, obviously an error in SaM. SaAB derive from the White Monastery, and perhaps their textual line preserves some shared, secondary variants like the verb ‘he died’ in verse 31: 4. In only one passage, SaMB agree against SaA, and this one might result from a copying mistake in SaA. In verse 31: 5, SaMB read, with some orthographical variation, ⲁϥⲃⲟϫϥ̄ ⲉϫ ̄ ⲉⲛ ⲧⲉϥⲥⲏϥⲉ ‘he leapt on his sword’, and SaA has ⲁϥⲛⲟϫϥ ⲉϫⲛ ⲧⲉϥⲥⲏⲃⲉ ‘he threw himself on his sword’. The reading of SaA could originate through scribal error, confusing ⲃⲟϫϥ with ⲛⲟϫϥ. Possibly SaA intentionally reused the same verb ⲛⲟϫϥ that occurred in the previous verse. In any case, the text of SaMB preserves the original Sahidic reading, and, thus, this variant does not evidence dependence between SaMB. SaM is the most reliable witness for ascertaining the original Sahidic translation. In three instances, however, SaM preserves a minor mistake. One finds the name ⲁⲃⲓⲥ, not ⲓⲁⲃⲓⲥ as in other Sahidic and Greek witnesses. SaA usually shortens the Sahidic base text. In verse 31: 1, the words ⲁⲩⲱ ⲛⲣⲱⲙⲉ ⲙⲡⲓⲏⲗ̄ ̄ ‘and the men of Israel’ are missing, in verse 31: 2 SaM places the conjunction and ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ in extraposition, whereas SaA offers only a pronominal infix as subject and without connective. However, this tendency to shorten does not hold in all the details. In verses 31: 4, 5, 6 there are additions in SaA when compared with SaM. In verse 31: 9, SaA preserves a Hexaplaric addition against SaM. SaA also contains some obvious mistakes and misspellings that occurred during the Sahidic transmission (31: 7, 10,

Textual analyses

217

12). SaA repeatedly contains characteristically nonsensical readings. There are mistakes that affect each other, but the copyist has not corrected them. SaB demonstrates its individuality with several abbreviations, omissions and free formulations. A special feature in this chapter is the interest in the destiny of Saul and Jonathan as opposed to the other sons. In verse 31: 2, SaB tells only about Jonathan’s death and with a more common verb than SaMA, and, in verses 31: 6 and 31: 8, SaB reads ‘Saul and his son Jonathan’ against other witnesses. In verses 31: 4–5 two cases suggest the dependence of SaB on SaA. Especially, if more such cases arise in other chapters, then SaB is not only dependent on a Sahidic base text but on a base text that contains some secondary readings of SaA. In verse 31: 5, the reading of SaB (ⲁϥϥⲟϭϥ ⲉϫⲉⲛ ⲧⲉϥⲥⲏⲃⲉ) supposes the reading of SaM (ⲁϥⲃⲟϫϥ̄ ⲉϫ ̄ ⲉⲛ ⲧⲉϥⲥⲏϥⲉ) without the variant of SaA (ⲁϥⲛⲟϫϥ ⲉϫⲛ ⲧⲉϥⲥⲏⲃⲉ), and this shows that SaB is not directly dependent on SaA but on a former manuscript with a similar text. In verse 31: 9, SaB and SaA have different readings. SaM reads ⲙⲡⲉⲩⲕⲱⲧⲉ ⲧⲏⲣϥ̄·ⲙⲛ̄ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ, SaA reads ϩⲙ ⲡⲕⲱⲧⲉ ⲧⲏⲣϥ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ, and SaB ⲉⲃⲟⲗ ϩⲙ ⲡⲕⲁϩ ⲛⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ. Since there are no parallel Greek variants, these omissions probably arose through inner-Sahidic variation. SaAB, which typically shorten the text, have in this case both abbreviated. In 31: 9, only SaB omits κύκλῳ εὐαγγελίζοντες τοῖς εἰδώλοις αὐτῶν καὶ τῷ λαῷ αὐτῶν. In verse 31: 10, SaB uses a higher frequency verb as compared to SaM. This fits well into the picture of SaB as a chronicle-like story. In verses 31: 12–13, SaB renders the text independently without corresponding Sahidic or Greek variants. In a few passages, a connection to a certain Greek reading is discernible in Sahidic witnesses. These affiliations of the Sahidic manuscripts are of special interest. Understanding the textual affiliation of these manuscripts is fundamental to their use in textual criticism. In the first clause of 31: 1, SaMA have the particle ⲇⲉ, and Greek manuscript 44 offers οἱ ἀλλόφυλοι δέ. In the next verse, 44 381 omit ‘the Philistines’ like SaA. These are, however, only seeming connections since these features are also present in Sahidic without any Greek source. Without studying the translation technique, these connections seem obvious. Probably, this also explains why Drescher has supposed a special connection to the Greek d-group: he was familiar with the Sahidic text after editing SaM and checking other Sahidic manuscripts, but he did not study the translation technique.209 In verse 31: 4, SaMA seem to agree with readings present in Greek L-82. In the same verse, SaM adds ⲉⲙⲟⲟⲩⲧϥ̄ ‘to kill him’, SaA ⲉⲙⲟⲩⲟⲩⲧ ⲉⲙⲙⲟϥ and SaB ⲉⲉⲛ ⲧⲟⲟⲧϥ ⲉϫⲱϥ ⲉⲙⲟⲟⲩⲧϥ ‘to bring his hand upon him to kill him’. In the same passage, Greek manuscript 44 offers σπάσαι τὴν ρομφαίαν αὐτοῦ καὶ ἐκκέντησαι αὐτόν ‘to draw his sword and pierce him’. In verse 31: 6, SaMA preserve ⲁⲩⲱ ⲛⲕⲉⲣⲱⲙⲉ ⲧⲏⲣⲟⲩ ⲉⲧⲛⲙ̄ⲙⲁϥ like Greek A O L-82 CII-242 s-64,

209 Drecher, Kingdoms, v.

218

Affiliations of the Sahidic manuscripts and the Greek Vorlage

381,130

καὶ πάντες οἱ ἄνδρες αὐτοῦ, thus attesting this Hexaplaric approximation. SaB independently states that Saul died with his son Jonathan and an additional 20,000 men, but SaB also mentions more men than merely Saul, Jonathan and the armour-bearer. SaMA read the verb ⲁⲩⲙⲟⲩ ‘they died’ after this variant against the Greek witnesses. This shared addition shows that it arose within the Sahidic tradition before the textual lines behind SaMA split. In verse 31: 8, the list of the dead does not mention these men, and it implies that the addition entered after the original Sahidic translation. In verse 31: 9, SaA reads as a doublet the same Hexaplaric variant as Greek M V 82 CII a d s 29 71 158 244 342 460 707mg. In this passage, SaM lacks this approximation, and supposes the early mistake from ἀποκεφαλίζουσιν to ἀποστρέφουσιν present in most Greek manuscripts; only the base text of the Lucianic recension has avoided this mistake. In verse 31: 12, SaA attests the name ⲃⲩⲑⲥⲁⲙⲏⲥ. In the Greek tradition, the longer form of the name appears as βαιθσαμυς (a-527 121) and βεθσαμυς (527). In 31: 13, SaMA do not render the last sentence of the Greek. This verse constitutes the last sentence of 1 Sam in some manuscripts, but the first sentence in 2 Sam in others. SaMA arrange this verse as the first sentence in 2 Sam. Sometimes, all Sahidic witnesses differ. In several passages, SaB has its own wording, while SaMA show only a slight difference in their readings. In 31: 4, SaM reads ⲕⲟⲟⲛⲥⲧ ̅ ̅ ‘pierce, slay’ and SaA employs the high frequency verbs ⲛⲅ̄ ⲉⲓ ⲛⲅⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ ‘come and kill me’. No Greek manuscript parallels these two verbs of SaA. SaB uses ⲛⲅⲙⲟⲩⲟⲩⲧ ⲙⲙⲟⲓ ‘kill me’ like SaA but without ‘come’. Additionally, in 31: 4, SaM adds that the armour-bearer did not want ⲉⲙⲟⲟⲩⲧϥ ̄ ‘to kill him’, SaA has ⲉⲙⲟⲩⲟⲩⲧ ⲉⲙⲙⲟϥ ‘to kill him’ and SaB ⲉⲉⲛ ⲧⲟⲟⲧϥ ⲉϫⲱϥ ⲉⲙⲟⲟⲩⲧϥ ‘to bring his hand upon him to kill him’. In verse 31: 10, the place is called ⲃⲉⲧⲥⲁⲙ in SaM as in the majority of Greek witnesses but ⲃⲩⲑⲥⲁⲙⲏⲥ in SaB, more or less like Greek 82 a 92* 342. ⲃⲉⲑⲗⲉⲉⲙ in SaA probably results from a mistake, but it is hard to determine which one of the alternatives was there originally. In verse 31: 12, SaA attests to ⲃⲩⲑⲥⲁⲙⲏⲥ.

3.3. Special cases 3.3.1. Text-critical notes in the clause connections chapter Several text-critical considerations concerning possible affiliations of Sahidic manuscripts were presented in section 2.1. In the following paragraphs, these notes have been collected and organized. In a few passages, Sahidic manuscripts attest to a secondary Greek reading that closely resembles the MT. In the Greek tradition, these readings are often attested by L (with/without 318 554).210 In verse 14: 13, (only SaM pre210 In chapter 17, the manuscripts that most often contain the same reading as Sahidic manuscripts were L 158 318 554 when correcting according to the MT.

Special cases

219

sent in this verse) Sahidic renders ὀπίσω αὐτοῦ (L CII s-489 = MT) instead of μετ᾽αὐτοῦ (all other Greek mss, omitted in M). In verse 17: 4, SaM has as an attribute of Goliath ‘Amessaios’ which more closely conforms to the MT. In this verse, the same word is found only in one Greek manuscript (93). In verse 17: 23, which is secondary in the LXX, this attribute is present in several Greek manuscripts. In verse 20: 17, SaMA both have a Hebrewbased addition found in L (+διὰ τὸ ἀγαπᾶν αὐτόν) and in other words in A O d-68,122,125 554 (εν τω ηγαπηκεναι αυτον). The wordings in Sahidic more closely resemble the reading of L. At the same time, it is to be noted that SaMA do not agree in all details but show minor discrepancies in this verse.211 In 26: 20 (SaMB present) a doublet ‘a single flea’ is present in Sahidic. In Greek, ‘a single flea’ is found in L 554mg (= MT) but as a variant for τὴν ψυχήν μου, not as a doublet. In verse 21: 6, (SaMA present) an attribute ‘other’ is found in Sahidic as well as in L 554. However, this variant fits the context especially well, and it is possible that it is added in Sahidic without consulting any Greek source. Sometimes only SaA contains the variant reading. In verse 20: 2, this happens with the conjunction ⲅⲁⲣ which is present in Greek L f 318 554. In verse 29: 5, SaA has the verb ‘to sing’, closely resembling the MT. SaM contains an erroneous rendering; it renders the verb ἐξέρχομαι, not ἐξάρχω. Sahidic manuscripts do not contain all variants found in L 318 554, although these manuscripts frequently do share secondary Greek readings with the Sahidic tradition. In verse 3: 6, Greek manuscripts M V L C’ a s 29 55 71 158 244 245 318 707 attest a Hexaplaric correction καλεσαι τον σαμουηλ. In the same verse, another such correction και ανεστη σαμουηλ is present in A O 127 d-68,122 554. Sahidic lacks both of these corrections. In verse 29: 10, the phrase ως αγγελος θεου is present in L 55 554mg (L 554mg continue by adding εν οδω) but not found in Sahidic. This Hexaplaric variant differs from other cases since there is no source text in the MT. It seems that this variant has ended up in the wrong place since it actually belongs to the previous verse where it is found in Α L-82,93 242 488. Additionally, O has καθως αγγελος κ̅ῡ in verse 29: 9. In verse 17: 8, the Sahidic text has the attribute ⲫⲩⲗⲉⲥⲧⲓ̈ⲁⲓⲟⲥ instead of the usual ⲁⲗⲗⲟⲫⲩⲗⲟⲥ. In Hebrew, the word is ‫‘ פלשׁתי‬Philistine’ and Sahidic thus agrees with the MT more closely than any of the preserved Greek readings. 26: 10 Ζῇ Κύριος, ἐὰν μὴ Κύριος παίσῃ αὐτόν, ἢ ἡ ἡμέρα αὐτοῦ ἔλθῃ καὶ ἀποθάνῃ, ἢ εἰς πόλεμον καταβῇ… παίσῃ =MT] παιδεύσῃ B Mmg V 46 a-799 b 55 244 342; παιδεύσει O 799 460

211 The passage and the texts are provided among the examples in section 2.1.4.9. ὅτι-, διότι-, and καθότι-clauses.

220

Affiliations of the Sahidic manuscripts and the Greek Vorlage

ⲡϫⲟⲉⲓⲥ ⲟⲛϩ̄· ϫⲉ ⲉⲣⲉⲧⲙ̄ ⲡϫⲟⲉⲓⲥ ϯⲥⲃⲱ ⲛⲁϥ· ⲡⲉϥϩⲟⲟⲩ ⲛⲁⲉⲓ ⲛϥ̄ ⲙⲟⲩ ⲏ ⲛϥ̄ ⲃⲱⲕ ⲉⲃⲟⲗ· ⲉⲡⲡⲟⲗⲩⲙⲟⲥ… In this verse, the Sahidic renders the Greek text with the verb ‘to teach, discipline’. An equivalent for this verb appears in B Mmg V O 46 a b 55 244 342 460. In the Old Greek, the verb was ‘to strike, hit’. Sahidic renders a secondary Greek reading, probably an early error that happened before the base text of the Hexaplaric recension departed from the OG tradition since the mistake is found in O. 20: 31 …οὐχ ἑτοιμασθήσεται ἡ βασιλεία σου· νῦν οὖν ἀποστείλας λαβὲ τὸν νεανίαν, … οὐχ – σου] οὐχ ἑτοιμασθήσῃ σὺ οὐδὲ ἡ βασιλεία σου L = MT SaM ⲛⲅ̄ ⲛⲁⲥⲁⲁϩⲉ ⲁⲛ ⲛⲧⲟⲕ ⲙⲛ̄ ⲧⲉⲕⲙⲛⲧ ̄ ⲣⲣ̄ ⲟ·̄ ⲧⲉⲛⲟⲩ ϭⲉ ϫⲟⲟⲩ ⲉⲧⲣⲉⲩⲉⲛ ⲡϣⲏⲣⲉ ϣⲏⲙ ⲛⲅⲛ̄ ⲁⲥⲟⲟⲩⲧⲛ ⲁⲛ ϩⲛ̄ ⲧⲕⲙ SaA ̄ ⲛⲧ ̄ ⲣⲣⲟ· ⲧⲉⲛⲟⲩ ϭⲉ ϫⲟⲟⲩ ⲛⲅⲙⲟⲩⲧⲉ ⲉⲣⲟϥ Here, SaM reads ‘you will not be established, you and your kingdom’ and thus attests the Greek text found in L ‘neither you nor your kingdom shall be established’. SaA may be abbreviating freely or translating another Greek text in its wording ‘you will not be established in your kingdom’. Perhaps only SaM has been corrected according to a later Greek source to conform more closely to the MT. If this is the case, then SaA preserves the more original Sahidic reading in this passage. 22: 18 καὶ εἶπεν ὁ βασιλεὺς τῷ Δωηκ ἐπιστρέφου σὺ καὶ ἀπάντα εἰς τοὺς ἱερεῖς καὶ ἐπεστράφη Δωηκ ὁ Σύρος

ⲁⲩⲱ ⲡⲣⲣ̄ ⲟ ⲡⲉϫⲁϥ ⲛⲇⲱⲏⲕ ⲡⲓⲇⲟⲩⲙⲁⲓ̅ ⲟⲥ· ϫⲉ ⲃⲱⲕ ⲛⲧⲟⲕ· ⲛⲅⲙⲟⲩⲟⲩⲧ ̅ ⲛⲛ ̄ ⲉⲟⲩⲏⲏⲃ· ⲇⲱⲏⲕ ⲇⲉ ⲡⲥⲩⲣⲟⲥ· ⲁϥⲃⲱⲕ In this passage, the proper name Doeg is first complemented with the attribute ‘Idoumaios’ and later in the same verse with the attribute ‘Syros’. The first attribute (Idoumaios) does not appear in other witnesses within this same phrase, but within the next Doeg-phrase in the MT and in Greek L. Therefore, Sahidic seems to add this attribute in the wrong place.

3.3.2. Corrections towards the MT While collecting the material for section 2.2 on translator’s additions, I identified cases where an addition in Sahidic more closely conforms to the MT. In these cases, no Greek manuscript has preserved this variant, and the reading is only found in Sahidic. These cases are as follows: 13: 5

καὶ οἱ ἀλλόφυλοι συνάγονται εἰς πόλεμον ἐπὶ Ισραηλ εἰς πόλεμον] ‘to fight, wage war’ MT

ⲛⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲇⲉ ⲁⲩⲥⲱⲟⲩϩ ⲉϩⲟⲩⲛ ⲉⲩⲡⲟⲗⲩⲙⲟⲥ· ⲉⲙⲓϣⲉ ⲙⲛ̄ ⲡⲓⲏⲗ̅ ·̅

Special cases

221

Here, the Sahidic adds the verb ‘to fight’: ‘The Philistines gathered to the war, to fight with Israel.’ In the Greek, no equivalent exists for this verb,212 and Sahidic more closely conforms to the MT which reads ‘The Philistines assembled to fight with Israel.’ Thus, Sahidic has a doublet, first rendering the Greek text, and after that the reading of the MT. 14: 18 καὶ εἶπεν Σαουλ τῷ Αχια προσάγαγε τὸ εφουδ ὅτι αὐτὸς ἦρεν τὸ SaM

εφουδ ἐν τῇ ἡμέρᾳ ἐκείνῃ τὸ εφουδ] ‘Ark of God’ MT

ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲁⲭⲓⲃ· ϫⲉ ⲁⲩⲉ· ⲡⲉⲫⲟⲩⲇ ⲉⲡⲉⲓⲙⲁ· ϫⲉ ⲛⲧⲟϥ ⲛⲉϥϥⲓ̈ ⲙⲡⲉⲫⲟⲩⲇ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲁⲭⲓⲃ· ϫⲉ ⲁⲩⲉⲓⲥ ⲡⲉⲫⲟⲩⲇ ⲙⲡϫⲟⲉⲓⲥ ⲉⲡⲉⲓⲙⲁ ϫⲉ ⲛϥϥⲓ ⲛⲛ SaA ̄ ϥⲥ̄ ⲕⲉⲩⲏ ϩⲛ̄ ⲛⲉϩⲟⲟⲩ ⲉⲧⲙ̄ⲙⲁⲩ In this case, the possible correction according to the MT is only present in SaA. However, this verse should not be considered as a MT-based correction since this additional word ‘of God’ does not assimilate the entire sentence to the MT. In Hebrew, the object is not the ephod but the ark of God. Usually, this Hebrew word is rendered with ἡ κιβωτός, not with εφουδ. In this case, the addition in SaA is not a correction according to the MT but a free formulation in SaA. 18: 23 κἀγὼ ἀνὴρ ταπεινὸς καὶ οὐχὶ ἔνδοξος ⲁⲛⲟⲕ ⲇⲉ ⲁⲛⲅ ⲟⲩⲣⲱⲙⲉ ⲛϩⲏⲕⲉ· ⲉⲓⲥⲏϣ ⲉⲛϯⲧⲁⲓⲏⲩ ⲁⲛ In this case, Sahidic renders the second part twice, first with an affirmative form ⲉⲓⲥⲏϣ ‘I am despised’ and, second, with ⲉⲛϯⲧⲁⲓⲏⲩ ⲁⲛ ‘I am not honored’. In the MT, only an affirmative expression is present ‘I am despised’, whereas Greek renders it with οὐχὶ ἔνδοξος ‘not honored’. Symmachus has ατιμος for οὐχὶ ἔνδοξος, like Sahidic with the doublet ⲉⲓⲥⲏϣ. 20: 33 καὶ ἐπῆρεν Σαουλ τὸ δόρυ ἐπὶ Ιωναθαν τοῦ θανατῶσαι αὐτόν SaM ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥϥⲓ ⲉϩⲣⲁⲓ ⲛⲧⲉϥⲗⲟⲅⲭⲏ· ⲉⲛⲟϫⲥ· ⲛⲥ̄ⲁ ⲓⲱⲛⲁⲑⲁⲛ ⲉⲙⲟⲟⲩⲧϥ· SaA ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥϥⲓ ⲉϩⲣⲁⲓ ⲛⲧ ̄ ϥ̄ⲗⲟⲅⲭⲏ ⲁϥⲥⲁⲧϥ̄ ⲛⲥ̄ ⲁ ⲓⲱⲛⲁⲑⲁⲛ ⲉⲙⲟⲟⲩⲧϥ· Hebrew has the verb ‘to throw’, not ‘to lift’ like the Greek. Notably, SaMA have different readings here: SaM has ⲁϥϥⲓ ⲉϩⲣⲁⲓ ⲛⲧⲉϥⲗⲟⲅⲭⲏ· ⲉⲛⲟϫⲥ· ⲛⲥⲁ̄ ⲓⲱⲛⲁⲑⲁⲛ ⲉⲙⲟⲟⲩⲧϥ ‘He took up his spear to cast it at Ionathan to kill him’, and SaA offers ⲁϥϥⲓ ⲉϩⲣⲁⲓ ⲛⲧ ̄ ϥ̄ⲗⲟⲅⲭⲏ ⲁϥⲥⲁⲧϥ̄ ⲛⲥ̄ ⲁ ⲓⲱⲛⲁⲑⲁⲛ ⲉⲙⲟⲟⲩⲧϥ· ‘He took up his spear, threw it at Ionathan to kill him’. Possibly, the Sahidic manuscripts have independently expanded the text with this Hebrew-based correction. Likewise, SaA may have altered the text to a more straightforward construction.

212 There, however, are several passages in 1 Sam where the verb πολεμέω renders this Hebrew verb. See, for example, 4: 9, 12: 9 and 17: 9.

222 26: 7

Affiliations of the Sahidic manuscripts and the Greek Vorlage

καὶ ἰδοὺ Σαουλ καθεύδων ὕπνῳ ἐν λαμπήνῃ… καὶ Αβεννηρ καὶ ὁ λαὸς αὐτοῦ ἐκάθευδεν κύκλῳ αὐτοῦ

ⲁⲩⲱ ⲉⲓⲥ ⲥⲁⲟⲩⲗ ⲉϥⲛⲏϫ ⲉϥⲟⲃϣ̄ ϩⲛ̄ ⲗⲁⲙⲡⲏⲛⲏ·… ⲁⲩⲱ ⲁⲃⲉⲛⲛⲏⲣ ⲙⲛ ⲡⲉϥⲗⲁⲟⲥ· ⲉⲩⲛⲏϫ ⲉⲩⲟⲃϣ̄ ϩⲙ̄ ⲡⲉϥⲕⲱⲧⲉ The Greek has rendered the Hebrew verb ‘to lie’ with ‘to sleep’. Sahidic has both, first ‘to lie’ and then ‘to sleep’. In the first clause, the Sahidic translation is perhaps an idiomatic rendering of καθεύδων ὕπνῳ to say that Saul lay sleeping. The second occurrence of the same verbs, then, is either a free formulation based on the previous similar expression or a correction according to Hebrew ‘lie down’. Aejmelaeus has labeled one type of correction in the Greek text of 1 Sam as καιγε-type corrections. In these cases, B (Vaticanus) and the majority of Greek manuscripts contain a correction according to the MT. The articles of Aejmelaeus offer a detailed discussion on the Greek variants and their connection to Hebrew.213 The following paragraphs weigh those instances that concern the Sahidic version. 15: 11 μεταμεμέλημαι ὅτι ἔχρισα τὸν Σαοὺλ εἰς βασιλέα, ὅτι ἀπέστρεψεν ἀπὸ ὄπισθέν μου καὶ τοὺς λόγους μου οὐκ ἔστησεν. μεταμεμέλημαι] παρακέκλημαι in A B O 93mg-108mg 121*(vid) ἔχρισα] ἐβασίλευσα B O L b 244 460 Eth ἔστησεν A L] εφυλαξεν O CII 121 s Εth; ετηρησεν rel ⲁⲓⲉⲣϩⲧⲏⲓ ϫⲉ ⲁⲓⲧⲱϩⲥ̄ ⲛⲥⲁⲟⲩⲗ ⲉⲧⲣⲉϥϣⲱⲡⲉ ⲛⲣⲣ̄ ⲟ̄· ϫⲉ ⲁϥⲕⲧⲟϥ· ⲁϥⲥⲁϩⲱⲱϥ ⲉⲃⲟⲗ ⲙⲙⲟⲓ· ⲙⲡⲉϥϩⲁⲣⲉϩ ⲉⲛⲁϣⲁϫⲉ· In this verse, three καιγε-type corrections occur. The first is from the Old Greek μεταμεμέλημαι to παρακέκλημαι in A B O 93mg-108mg 121*(vid).214 Sahidic has rendered the OG: ⲁⲓⲉⲣϩⲧⲏⲓ ‘to repent, regret’ since the verb παρακαλέω is rendered with ⲥⲟⲡⲥⲡ ‘to comfort, entreat’ and ⲥⲟⲗⲥⲗ ‘to comfort’ in 1–2 Samuel. The second change in the same verse is from ἔχρισα to ἐβασίλευσα,215 and again Sahidic renders the OG: ⲁⲓⲧⲱϩⲥ̄ ‘I have anointed’. The usual rendering for βασιλεύω is, depending on the context, ⲣⲣ̄ ⲣ̄ ⲟ ‘to reign’ or ⲉⲓⲣⲉ ⲛⲣⲣ̄ ⲟ ‘to make king’ in 1–2 Sam. The third point of interest in this verse is ἔστησεν ‘to carry out’. In Sahidic, the verb used is ϩⲁⲣⲉϩ ‘to keep, guard, watch’, which renders both εφυλαξεν and ετηρησεν. According to Aejmelaeus, the variant ετηρησεν is based on an early mistake,216 and this variant was probably the reading in the Vorlage of the Sahidic translator. Sometimes the Sahidic text attests corrections towards the MT, but in this case, εφυλαξεν is not such a correction.

213 Aejmelaeus, “Kingdom at Stake”, 353–66; eadem, “Corruption or Correction?”, 1–17. 214 Aejmelaeus, “Kingdom at Stake”, 354–6. 215 Ibid., 357–8. 216 Ibid., 359.

Summary

223

15: 29 ὅτι οὐχ ὡς ἄνθρωπός ἐστιν τοῦ μετανοῆσαι· αὐτὸς ἀπειλήσει καὶ οὐκ ἐμμενεῖ; αὐτὸς – ἐμμένει] αὐτός A B O b -121mg d -44 (=MT); > L 44; ϫⲉ ⲛⲉϥⲟ ̅ ⲁⲛ ⲛⲑⲉ ⲛⲟⲩⲣⲱⲙⲉ ⲉⲧⲣⲉϥⲣϩ̅ ⲧⲏϥ ⲛⲧⲟϥ ⲉϥⲛⲁϭⲱⲛⲧ· In this verse, A B O b d-44 shorten the text according to the MT.217 Sahidic does not follow them, but instead renders the second verb. The Sahidic does not resemble any of the preserved Greek variants, but has its own version of the verse: ‘For he is not like a man, to repent, he, when angry.’ In this case, Sahidic does not parallel the Hebrew-based omission found in L. 1: 13

καὶ φωνὴ αὐτῆς οὐκ ἠκούετο καὶ εἰσήκουσεν αὐτῆς Κύριος. καὶ εἰσήκουσεν αὐτῆς Κύριος] om. A B O b f 55 245 707txt (=MT)

ⲉⲛⲥⲉⲥⲱⲧⲙ ⲁⲛ ⲉϩⲣⲟⲟⲩ ⲛⲧⲁⲥ· In this case, Sahidic renders a Greek text where the second part of the sentence is omitted according to the MT.218

3.4. Summary In the seven passages analyzed in section 3.2, one stream of tradition clearly stands behind these Sahidic manuscripts. This is perceivable in shared misrenderings, omissions, and additions, as well as reformulations against the Greek, the Hebrew and other daughter versions. The earliest Sahidic manuscripts, SaU and SaI, agree closely with later Sahidic manuscripts. Thus, the extant manuscript tradition goes back as far as the 4th–5th century. Upon close examination, there were clearly no thorough revisions according to Greek manuscripts. Otherwise, obvious mistakes in the Sahidic would have been corrected. Thus, I agree with Butts that these manuscripts (he refers to SaMAU) and their characteristics “challenge Drescher’s understanding of the textual transmission of Coptic biblical text, which he assumed had involved continual revision and correction”.219 The Sahidic textual tradition was translated from an Old Greek type of source text. During the transmission, different kinds of changes happened. First, there are inner-Sahidic variants that have their origin in copying mistakes. Such variation is nothing surprising, but rather a natural phenomenon as long as the texts were hand copied. In some cases, the mistake is of aural nature. After eliminating the unintentional variants, one can continue with the remaining variants. These constitute two distinct categories, free reformulations and Greek-based variant readings. Even though the difference 217 Ibid., 362. 218 Aejmelaeus, “Corruption or Correction”, 13. 219 Butts, “P.Duk.inv. 797”, 12.

224

Affiliations of the Sahidic manuscripts and the Greek Vorlage

between these categories is obvious, the designation of individual readings is not always clear. Here, Sahidic manuscripts diverge from each other. SaFJIMUV do not show nearly as many free reformulations as SaA, not to mention SaB. Greek-based variants (i. e., agreements with secondary Greek readings) are found sporadically in SaABFJMV. In most of the cases where such a Greek-based variant is found in Sahidic, the variant is present in all Sahidic witnesses for that passage. Occasionally, SaA alone attests to such a variant. In the passages studied here, there are no Greek-based variants in SaIU. However, these variants were not necessarily added later than these early manuscripts were copied since, in these passages, absolutely no Greek-based variants exist in any Sahidic manuscripts. SaM is a particularly pure witness to the original Sahidic translation, and the same can be said of SaFJIUV. SaM has small additions that explicitate some thoughts found implicit in the text. In a few cases, the information added to SaM comes from some other passage in nearby surroundings, not from the sentence in question (harmonization). A few copying mistakes appear throughout SaM, but, on the whole, its text is highly reliable, and has preserved the text already encountered in the oldest fragments some 500 years earlier well. The second-best preserved among Sahidic manuscripts is SaA, at least with regard to what remains of its text. This manuscript shares the same base text as SaM and its earlier companions. However, SaA has numerous corruptions and emendations. Most of the changes do not show a connection to a Greek manuscript/manuscript group, but occasionally such an affiliation is possible. These are as follows: in verse 7: 11 Greek manuscripts 93 121 246 245, in verse 7: 16 O, in verse 10: 1 245 707, in verse 10: 1 127 707, in verse 17: 37 509, in verse 29: 6 107 (Eth Arm), in verse 29: 9 44–125, in verse 29: 5 a Greek source closer to the MT than any preserved Greek manuscript, in verse 30: 21 71 381, in verse 31: 1 44 381, in verse 31: 9 M V 82 CII a d s 71 158 244 342 460 707mg, in verse 31: 12 a 121. Thus, no single Greek group may be identified with SaA. Instead, the variants are in agreement with various Greek manuscripts. Not even the nature of these Greek variants is coherent since some of them are closer to Hebrew while others show emendations of different kinds. Often, when there is a connection between SaA and member(s) of the d-group, it is an omission or abbreviation, and this poses a problem. Since SaA has numerous omissions without any Greek source, the connections probably do not show a real affiliation but only a coincidence. In his introduction, Drescher mentioned manuscript 707 as worth studying in comparison to SaA, and this connection is also evident here, as well as some connections to the variants of B-group and manuscript 245. Interestingly, discrepancies occur between different chapters of SaA. When compared with the text in chapters 7–9 and 29–31, chapters 14 and 17 are of significantly better quality, without the copying mistakes found abundantly in chapters 7–9 and 29–31. Perhaps, these chapters were revised,

Summary

225

at least chapter 17 where the story of David and Goliath is partly complemented with Hebrew-based additions. On the whole, SaA is already further removed from the initial Sahidic text with its many special readings and mistakes.220 When closely examining this manuscript’s text, its variations repeatedly produce a grammatically and narratively improved text with simpler constructions, more recent vocabulary and shortened wordings. SaB shares similar tendencies towards abbreviation, but SaB’s deviations are particularly extreme to the extent that it is justified to speak of a chronicle.221 The Sahidic tradition generally has been affected by some Hexaplarictype readings. Most of these are also encountered in L (with/without 318 554). However, only some Hexaplaric-type corrections present in these Greek manuscripts enter the Sahidic text. In a few cases (3.3.2 above), the reading encountered in Sahidic witnesses is nearer to the MT than any preserved Greek reading, and this shows that a now-lost Greek source(s) lie(s) behind some variants in the Sahidic manuscripts. One cannot know whether scribes working on the Sahidic tradition acquired these readings from one source, from several margins or from several manuscripts over the years. The Sahidic tradition contains some early corrections according to the MT, but it does not consistently reflect any particular Greek text type. This textual diversity is typical of early Greek manuscripts as well.222 Thus, Sahidic scribes have adapted their text to readings that were circulating in Egypt in the first few centuries CE. The same kinds of texts have been used by the Greek scribes who complemented the B-text with its καιγε-type corrections. The Sahidic tradition has one more peculiar feature. Usually, it has not omitted sections that are lacking in the MT, but instead collected missing parts. When Greek manuscripts that usually agree with Sahidic omit according to the MT, the Sahidic tradition will preserve the text against the MT.

220 Thus also Butts, “P.Duk.inv. 797”, 12. 221 Is this a typical phenomenon in the 11th century and onwards? In the Sahidic found in SaAB and also perceivable in the Greek d-group? But not in Bohairic 1 Sam or Deuteronomy. 222 See, for example, Aejmelaeus (“What Rahlfs Could Not Know”) for information on the 4th-century fragment 845 containing an early correction according to the MT as well as 707 and καιγε-type corrections in B.

4. Conclusions The aim of this study was twofold: first, to obtain a description of the translation technique of the Sahidic translator; second, to apply this basic knowledge of the translation technique in the area of textual criticism in order to discover the affiliations of the Sahidic version (or its Greek Vorlage) with Greek textual traditions and the affiliations among the Sahidic manuscripts. The translation technique was studied in chapter 2. In section 2.1, clause connections were analyzed. A clear majority of the clauses in Greek 1 Sam begin with the conjunction καί. Consequently, the dominant features in the clause connections corpus are asyndetic clauses and the abundant use of the connective ⲇⲉ that appear in renderings of the Greek καί-clauses. Thus, the translator did not strive towards formal equivalency but was in this respect more bound to the norms of his target language. For text-critical purposes, it must be emphasized that the Sahidic translator used Greek words as renderings of other Greek words. As tempting as it might seem to retrovert a Greek word in the Sahidic text with the corresponding Greek word in the supposed Vorlage, one should refrain from it unless a careful study of the translation technique justifies the retroversion. Interestingly, double-marking (interclausal relation expressed through a verbal inflection plus a connective) typically occurs in subordinate clauses. This theme may deserve a more detailed study to determine whether this phenomenon is related to the diachronic development of Sahidic and to what extent it occurs in other Coptic literature. In section 2.2, the purpose was to identify those additions that supposedly were made by the translator and thus to outline more clearly the translator’s attitude towards his source text and the translation’s formal equivalency with the source text. The term addition has a twofold usage in this chapter: first it refers to an addition according to the definition “a semantic element that is not found in the source text” and second to “a secondary plus in the text-critical meaning”. The aim was to find those additions that correspond not only to the former definition but also to the latter. To achieve this goal, the additions were identified and divided into three groups based on their obligatoriness; in the first group the Sahidic grammar requires the addition, in the second group the resulting expression would be illogical or clumsy without the addition, and in the third group the additions serve stylistic and epexegetical interests. As a result of the analyses in this section, one can conclude that the translator’s personal impetus is to be found among the most free additions, i. e., the stylistic additions. The stylistic additions are also the ones that most unambiguously deserve the charac-

Conclusions

227

terization as secondary plusses whereas the additions based on Sahidic grammar should be characterized as accurate translations. The first group consisted of 87 (+7) grammatical additions and the second group contained 241 (+21) logical additions. The third group of stylistic additions contained 134 (+31) additions. The numbers are difficult to compare with each other since it mostly depends on the Greek text which obligatory additions were needed in the translation. The numbers of the additions in the second and third groups show that the translator was more bound to the norms of the target language than to the source text. Formal equivalency was not his goal but he tried to produce a comprehensible text for his audience. It is of interest to compare the additions analyzed in section 2.2 to the typical features of the Lucianic recension. According to Brock, these include the insertion of proper names or pronouns to make the narrative explicit as well as the insertion of the verb εἰμί or γίνομαι.1 Since these additions occur in the Sahidic text without any Greek source, these features do not weigh as arguments when looking for potential dependences between Lucianic readings and the Sahidic version. Another issue is whether these features tell us something about the use of the Sahidic text. Concerning the Lucianic text, these features are seen as an indication of the public reading of the text.2 In the present study, the translation technique is described based on the material concerning clause connections and translator’s additions. However, these areas are only pieces of a larger puzzle and other topics would make the picture more reliable and detailed. For example, the renderings of different Greek tempora could be one area that deserves more attention. How did the Sahidic translator proceed with various Greek tempora? Was he consistently using certain renderings for certain tempora? Another area that is not touched in the present work but deserving attention is the lexicon. Is this translator striving towards concordant renderings or does he render according to the context in each passage? How does he use GrecoCoptic words, and how often do these appear in his text? In chapter 3, the topic was the Greek Vorlage and affiliations of the Sahidic manuscripts. First, the analyses have shown that one common translation event stands behind all Sahidic witnesses. This is the most compelling explanation for the similar wordings and shared mistakes encountered in the Sahidic manuscripts. The earliest Sahidic manuscripts agree closely with later ones. Thus, the conclusion follows that the extant manuscript tradition goes back as far as to the 4th – 5th century. Additionally, no thorough revision according to Greek manuscripts has been carried out. Otherwise, obvious mistakes in the Sahidic would have been corrected. One area of spe-

1 2

Brock, Recensions, 252. Ibid., 252. He also mentions use of the article and removal of Hebraisms as such indications.

228

Conclusions

cial interest is the relation of versions in different dialects of the Coptic language. This study touched briefly on the only extant fragment of the Fayyumic version of 1 Sam. Based on this small amount of material, it seems independent from the Sahidic version. Among the Sahidic manuscripts, the Hamuli manuscript SaM is the best when quality and reliability in Greek textual criticism are concerned. This complete manuscript of 1 Sam has preserved the Sahidic version with only some minor scribal mistakes and additions from the Greek tradition. The other two main witnesses of the Sahidic text, SaA and SaB, are of totally different character. Both of these contain numerous omissions and free rewordings, and SaA also contains dozens of copying mistakes. In a few cases, SaA alone provides a Hebrew-based correction in Sahidic. The Sahidic version has affiliations with varying combinations of Greek manuscripts, and no one group has predominance in this respect. In 1 Sam 17, the Sahidic version has mostly the same additions as L 158 318 554 but even in these passages the Sahidic text does not contain all of the additions found in these Greek manuscripts. Interestingly, in several cases where these Greek manuscripts show a Hexaplaric-based omission, Sahidic manuscripts do not omit but preserve the text. The Sahidic version seems to have been more prone to additions than omissions. Affiliations with the Greek d-group were supposed and offered for a detailed examination by Drescher in his introduction to the edition of SaM. In the present study, this possible affiliation was observed in several passages. However, the typical features of the d-group consist of omissions and abbreviations, and this makes the whole enterprise problematic. In numerous cases, it is especially SaA that seems to agree with the Greek d-group, and this Sahidic manuscript typically omits and abbreviates without any Greek source. Thus, the agreement in omissions does not show a real affiliation. In the introduction, the lack of studies on the textual history of the Sahidic version was touched upon. As a result of this study, the textual history of Sahidic 1 Sam has gradually taken shape. After the thorough text-critical analyses, there is no basis to suppose Sahidic to be one of the members of Payne’s b-group,3 as he thought some 60 years ago. More accurately, the Sahidic version has been rendered from an early Greek text with some details now found in the B-text. Later, the Sahidic tradition got additional corrections, partly the same as the Hexaplaric-type corrections in the Lucianic tradition. It is to be noted that the Sahidic also attests some Hebrewbased corrections that have not been preserved in any Greek manuscript. This fits the picture drawn by Aejmelaeus that no one Greek manuscript has preserved all the early Hebrew-based corrections but they are scattered

3

B b Eth.

Conclusions

229

in varying combinations in the Greek tradition. Accordingly, some of these readings have been preserved only in the Sahidic tradition. There is an urgent need for current, accurate editions of Sahidic biblical texts. This task suffers from the present situation: fragments are scattered in collections around the world, and even to reassemble a manuscript needs a huge effort, not to mention the study of its textual character. This is, however, a prerequisite for any further study of these texts and their relations to their Vorlagen and other Greek and Sahidic texts. Only after enough textual material has been recovered, can this promising area of comparative studies be extensively researched.4 In the present study, the text of the Sahidic Jeremiah-corpus, edited by Feder, was used for the comparisons. At the same time, however, it is self-evident that comparative material from other historical books as well as from the Sahidic Pentateuch would be most welcome. In the present study, the issue of versions in various Coptic dialects did not prevail because of the lack of attestation in other dialects apart from Sahidic. In other books of the Septuagint, however, this theme is among the most intriguing topics. How are the versions in various Coptic dialects related to each other and how do they affiliate with diverse Greek textual traditions? As an example of this kind of study is Christian Askeland’s monograph on the Gospel of John in Coptic.5 An intriguing area of further research is the plausible variation between different Sahidic translators. In the case of the Septuagint itself, one should not speak of one translation as if it were such, but diverse Septuagint translations in different books. In the same way, it is most probable that the Sahidic daughter version has several translators who have left their own “fingerprints” in the texts. With the help of these clues we might be able—sometime in the future—to describe the characteristic features of each translation in the Sahidic version of the Septuagint.

4

5

When writing this conclusion, I was happy to hear that there is a prospectual project in its initial stage by Prof. Heike Behlmer and Dr. Frank Feder. The project at Göttingen Academy of Sciences is entitled Digital Edition and Translation of the Coptic-Sahidic Old Testament. For its present state see http://coptot.manuscriptroom.com/home. Christian Askeland (“The Coptic versions of the New Testament”, 222), emphasizes the importance of codicological research on Coptic manuscripts but also calls for text-critical and translational analysis of published biblical texts. Askeland, John’s Gospel.

5. Bibliography Aejmelaeus, Anneli, “A Kingdom at Stake: Reconstructing the Old Greek – Deconstructing the Textus Receptus”, in Anssi Voitila/Jutta Jokiranta (ed.), Scripture in Transition: Essays on Septuagint, Hebrew Bible, and Dead Sea Scrolls in Honour of Raija Sollamo (Supplements to the Journal for the Study of Judaism 126; Leiden/Boston: Brill, 2008) 353–66. — “Corruption or Correction? Textual Development in the MT of 1 Samuel 1”, in P.A. Torijano Morales/A. Piquer Otero (ed.), Textual Criticism and Dead Sea Scrolls Studies in Honour of Julio Trebolle Barrera: Florilegium Complutense (Supplements to the Journal for the Study of Judaism 157; Leiden: Brill, 2012) 1– 17. — On the Trail of the Septuagint Translators: Collected Essays. Revised and Expanded Edition (Leuven/Paris/Dudley, Mass.: Peeters, 2007). — Parataxis in the Septuagint: A Study of the Renderings of the Hebrew Coordinate Clauses in the Greek Pentateuch (Annales Academiae Scientiarum Fennicae. Dissertationes Humanarum Litterarum 31; Helsinki: Suomalainen tiedeakatemia, 1982). — “What Rahlfs Could not Know: 1 Sam 14,4–5 in the Old Greek” in H. Ausloos/ B. Lemmelijn/J. Trebolle Barrera (ed.), After Qumran: Old and Modern Editions of the Biblical Texts – The Historical Books (Bibliotheca ephemeridum theologicarum lovaniensium 246; Leuven: Peeters, 2012) 81–93. Askeland, Christian, John’s Gospel: The Coptic Translations of its Greek Text (Arbeiten zur Neutestamentlichen Textforschung 44; Berlin/Boston: De Gruyter, 2012). — “The Coptic versions of the New Testament” in Bart D. Ehrman and Michael W. Holmes, eds. The text of the New Testament in contemporary research: essays on the status quaestionis, 2nd ed. Leiden: Brill, 2012) 201–230. Bagnall, Roger S., Early Christian Books in Egypt (Princeton/Oxford: Princeton University Press, 2009). Barthelémy, Dominique, Les Devanciers d’Aquila: première publication intégrale du texte des fragments du Dodécaprophéton trouvés dans le désert de Juda (Supplements to Vetus Testamentum 10; Leiden: Brill, 1963). Becher, Viktor, “Abandoning the notion of ‘translation-inherent’ explicitation: against a dogma of translation studies”, Across Languages and Cultures 11(1) (2010) 1–28. Blass, Friedrich/Debrunner, Albert, Grammatik des neutestamentlichen Griechisch: Bearbeitet von F. Rehkopf (Göttingen: Vandenhoeck & Ruprecht, 141976). Blomqvist, J./Jastrup, P.O., Grekisk grammatik (Århus: Akademisk Forlag, 1996). Bosson, Nathalie, “Synopse des témoins coptes de Jonas”, Journal of Coptic Studies 16 (2014) 1–46. Boud´hors, Anne, Catalogue des fragments coptes de la Bibliothèque Nationale et Universitaire de Strasbourg, I. Fragments bibliques. (Corpus Scriptorum Christianorum Orientalium 571; Louvain: Peeters, 1998).

Bibliography

231

— “‘Chercher’ dans les textes coptes: à propos d’une variation lexicale”, in J.P. Monferrer-Sala/Á. Urbán (ed.), Sacred Texts: Explorations in Lexicography (Studien zur romanischen Sprachwissenschaft und interkulturellen Kommunikation 57; Frankfurt: Peter Lang, 2009) 15–29. — “L’evangile de Marc en copte-sahidique: essai de clarification”, in David W. Johnson (ed.), Acts of the Fifth International Congress of Coptic Studies, Washington DC, 12–15 August 1992: Volume 2, Papers from the Sections, Part 1 (Rome: CIM, 1993) 53–65. — “Manuscripts and Literature in Fayoumic Coptic” in Gawdat Gabra (ed.), Christianity and Monasticism in the Fayoum Oasis: Essays from the 2004 International Symposium of the Saint Mark Foundation and the Saint Shenouda the Archimandrite Coptic Society in Honor of Martin Krause (Cairo/New York: The American University in Cairo Press, 2005). Brock, Sebastian P., The Recensions of the Septuagint Version of I Samuel. (Quaderni Di Henoch 9; Torino: Silvio Zamorani Editore, 1996). Brooke, A./McLean, N., The Old Testament in Greek. Vol. II. The Later Historical Books. Part I. I and II Samuel (Cambridge: Cambridge University Press, 1927). Browne, Gerald M./Papini, Lucia, “Frammenti in copto dei Libri dei Re”, Orientalia 51 (1982) 183–90. Butts, Aaron Michael, “P.Duk.inv. 797 (U) – I Kingdoms 14: 24–50 in Sahidic”, Le Muséon 118 (2005) 7–19. Buzi, Paola, “Stefano Borgia’s Coptic Manuscripts Collection and the ‘Strange Case’ of the Borgiano Copto Fund in the Vatican Library” in Paola Buzi/Delio Vania Proverbio (ed.), Coptic Treasures from the Vatican Library: A Selection of Coptic, Copto-Arabic and Ethiopic Manuscripts: Papers collected on the occasion of the Tenth International Congress of Coptic Studies (Studi e testi 472; Vatican: Biblioteca Apostolica Vaticana, 2012) 21–6. Ciasca, P. Augustini, Sacrorum Bibliorum fragmenta copto-sahidica Musei Borgiani, Vol. I (Romae: typis eiusdem S. Congregationis, 1885). Collins, John J., Introduction to the Hebrew Bible (Minneapolis: Fortress Press, 2004). Cross, F.M./Parry, D.W./Saley, R.J./Ulrich, E.C., Qumran Cave 4. XII: 1–2 Samuel (DJD XVII; Oxford: Clarendon Press, 2005). Crum, Walter Ewing, A Coptic Dictionary (Oxford: Clarendon Press, 1939, repr. Eugene, Oregon: Wipf and Stock Publishers, 2005). — Catalogue of the Coptic manuscripts in the British Museum (London: British Museum, 1905). — Coptic Ostraca from the collections of the Egypt Exploration Fund, the Cairo Museum and others (London: The Egypt Exploration Fund, 1902). Denniston, John Dewar, The Greek Particles (Oxford: Oxford University Press, 2nd edn 1950, repr. Bristol Classical Press, 1996). Depuydt, Leo, Catalogue of Coptic Manuscripts in the Pierpont Morgan Library (Leuven: Peeters, 1993). Dieu, Léon, “Les Manuscrits Grecs des Livres de Samuel”, Le Muséon 34 (1921) 17– 60. — “Le texte copte sahidique des livres de Samuel”, Le Muséon 59 (1946) 445–52. Drescher, James, The Coptic (Sahidic) Version of Kingdoms I, II (Samuel I, II). Ed. James Drescher (Corpus Scriptorum Christianorum Orientalium 313, Scriptores Coptici, tomus 35; Louvain: Secrétariat du Corpus SCO, 1970).

232

Bibliography

— The Coptic (Sahidic) Version of Kingdoms I, II (Samuel I, II). Translated by James Drescher (Corpus Scriptorum Christianorum Orientalium 314, Scriptores Coptici, tomus 36; Louvain: Secrétariat du Corpus SCO, 1970). Edgerton, William F., “Obsolescence of the Imperative Mood in Egyptian”, in Stephen R.K. Glanville (ed.), Studies Presented to F. LL. Griffith (London: Egypt Exploration Society, Oxford University Press, 1932) 61–8. Elanskaya, A. I., “Peculiarities of ⲛⲑⲉ-clauses in Coptic” Palestinskij Sbornik 15 (1966) 28–34. Emmel, Stephen, “The Library of the Monastery of the Archangel Michael at Phantoou (al-hamuli)”, in Gawdat Gabra (ed.), Christianity and Monasticism in the Fayoum Oasis (Cairo/New York: The American University in Cairo Press, 2005) 63–70. Emmenegger, Gregor, Der Text des koptischen Psalters aus al-Mudil: Ein Beitrag zur Textgeschichte der Septuaginta und zur Textkritik koptischer Bibelhandschriften, mit der kritischen Neuausgabe des Papyrus 37 der British Library London (U) und des Papyrus 39 der Leipziger Universitätsbibliothek (2013) (Texte und Untersuchungen zur Geschichte der altchristlichen Literatur 159; Berlin/New York: Walter de Gruyter, 2007). Feder, Frank, Biblia Sahidica: Ieremias, Lamentationes (Threni), epistula Ieremiae et Baruch (Texte und Untersuchungen zur Geschichte der altchristlichen Literatur 147; Berlin/New York: de Gruyter, 2002). — “1.4.2 Coptic Translations”, in Armin Lange (ed.), Textual History of the Bible (Brill Online, 2016). — “Die koptische Übersetzung des Alten und Neuen Testamentes im 4. Jahrhundert”, Hallesche Beiträge zur Orientwissenschaft 44 (2007) 65–93. Funk, Wolf-Peter, “Bemerkungen zum Sprachvergleich Griechisch-Koptisch”, in Peter Nagel (ed.), Graeco-Coptica: Griechen und Kopten im byzantinischen Ägypten (Halle (Saale): Martin-Luther-Universität, 1984) 147–80. — “Dialects Wanting Homes: A Numerical Approach to the Early Varieties of Coptic”, in Jacek Fisiak (ed.), Historical Dialectology, Regional and Social (Berlin: Walter de Gruyter, 1988). 149–92. — “On a semantic typology of conditional sentences”, Folia Linguistica 19 (1985) 365–413. Girgis, Waheeb Atella (later Gregorius Anba), “Greek Loan Words in Coptic (VII)”, Bulletin de la Société d’archéologie Copte 30 (1991) 77–92. — Greek Words in Coptic Usage (PhD thesis, University of Manchester, 1955; published by Society of Anba Gregorios, 2010). Grossman, Eitan/Richter Tonio Sebastian, “The Egyptian-Coptic language: its setting in space, time and culture”, Egyptian-Coptic Linguistics in Typological Perspective 55 (2015) 69–101. Hasznos, Andrea, “Die Struktur der Konsekutivsätze im koptischen Neuen Testament”, Enchoria 29 (2004/05) 32–43. — Graeco-Coptica: Greek and Coptic Clause Patterns (Göttinger Orientforschungen IV. Reihe Ägypten 52; Wiesbaden: Harrassowitz Verlag, 2012). Jellicoe, Sidney, “The Hesychian Recension Reconsidered”, Journal of Biblical Literature 82 (1963) 409–18. Joosten, Jan, “Reflections on the ‘Interlinear Paradigm’ in Septuagintal Studies”, in A. Voitila/J. Jokiranta (ed.), Scripture in Transition: Essays on Septuagint, Hebrew

Bibliography

233

Bible, and Dead Sea Scrolls in Honour of Raija Sollamo (Supplements to the Journal for the Study of Judaism 126, Leiden: Brill, 2008), 163–78. Kahle, Paul E., Bala’izah: Coptic texts from Deir el-Bala’izah in Upper Egypt, Volume 1. (London: Oxford University Press, 1954). Kauhanen, Tuukka, The Proto-Lucianic Problem in 1 Samuel (De Septuaginta Investigationes 3; Göttingen: Vandenhoeck & Ruprecht, 2012). Klaudy, Kinga, “Explicitation” in Mona Baker/Kirsten Malmkjær (ed.), Routledge encyclopedia of translation studies (London/New York: Routledge, 1998) 80–4. Knibb, Michael A., “The Greek Vorlage of the Ethiopic Text of Ezekiel”, in A. Voitila/J. Jokiranta (ed.), Scripture in Transition: Essays on Septuagint, Hebrew Bible, and Dead Sea Scrolls in Honour of Raija Sollamo (Supplements to the Journal for the Study of Judaism 126, Leiden: Brill, 2008), 413–21. Kühner, R./Gerth, B., Ausführliche Grammatik der griechischen Sprache. Zweiter Teil: Satzlehre. Zweiter Band (Hannover: Verlag Hahnsche Buchhandlung, 1904). Lagarde, Paul de, “Bruchstücke der koptischen Übersetzung des Alten Testaments” in Orientalia 1, (Abhandlungen der Gesellschaft der Wissenschaften zu Göttingen 24; Göttingen, 1879) 63–72. Layton, Bentley, A Coptic Grammar (2nd edn; Porta Linguarum Orientalium 20; Wiesbaden: Harrassowitz Verlag, 2004). — Coptic in 20 lessons: Introduction to Sahidic Coptic With Exercises & Vocabularies (Leuven/Paris/Dudley, Mass.: Peeters, 2007). Lefort, L.T., Les manuscrits coptes de l’Université de Louvain 1: Textes littéraires (Louvain: Bibliothèque de l’Université, 1940). Lemm, Oscar von, Sahidische Bibelfragmente, III (Bulletin de l’Académie Impériale des Sciences de Saint-Pétersbourg, 5e série, t. XXV, n. 4; St. Pétersbourg, 1906). Liddell, H.G./Scott, R.S., A Greek-English Lexicon: With a revised supplement (Oxford: Clarendon Press, 1996). — An intermediate Greek-English Lexicon founded upon the 7th edition of Liddell and Scott´s Greek-English Lexicon (Oxford: Clarendon Press, 71889). Liljeström, Marketta, “Looking for Fragments of the Syrohexapla: The Song of Hannah in Barberiniani Orientali 2 as a Test Case?”, Bulletin of the International Organization for Septuagint and Cognate Studies 40 (2007) 49–61. Louw, Theo A. van der, Transformations in the Septuagint: Towards an Interaction of Septuagint Studies and Translation Studies (Contributions to Biblical Exegesis and Theology 47; Leuven/Paris/Dudley, Mass.: Peeters, 2007). Malmkjær, Kirsten, Linguistics and the Language of Translation (Edinburgh Textbooks in Applied Linguistics; Edinburgh: Edinburgh University Press, 2005). Maspero, G., Fragments de la version thébaine de l’Ancien Testament (Mission archéologique française au Caire, Mémoires, vol. 6, fasc. 1; Paris, 1892). Mayser, Edwin, Grammatik der griechischen Papyri aus der Ptolemäerzeit: Band II 2: Satzlehre (Berlin/Leipzig: de Gruyter, 1934). — Grammatik der griechischen Papyri aus der Ptolemäerzeit. Band II 3. Satzlehre (Berlin/Leipzig: de Gruyter, 1934). McDonald, Lee Martin, The Biblical Canon: Its Origin, Transmission, and Authority (Peabody, Mass.: Hendrickson Publishers, 2007). Mink, Gerd, “Die koptischen Versionen des Neuen Testaments” in K.Aland (ed.), Die alten Übersetzungen des Neuen Testaments, die Kirchenväterzitate und Lektionare (Arbeiten zur neutestamentlichen Textforschung 5; Berlin/New York: Walter de Gruyter, 1972) 160–299.

234

Bibliography

Müller, Matthias, “Contrast in Coptic I. Concessive constructions in Sahidic”, LingAeg 17 (2009) 139–82. — “Greek Connectors in Coptic. A Contrastive Overview” (forthcoming). Muraoka, Takamitsu, “The Use of ΩΣ in the Greek Bible”, Novum Testamentum 7 (1964) 51–72. Nagel, Peter, “Die Einwirkung des Griechischen auf die Entstehung der koptischen Literatursprache”, in Franz Altheim/Ruth Stiel (ed.), Christentum am Roten Meer 1 (Berlin/New York: de Gruyter, 1971) 327–55. — “Old Testament”, in Aziz Suryal Atiya, (ed.), The Coptic encyclopedia 6 (Digital Publisher Claremont Graduate University, School of Religion, 1991). 1836a–40a. Olohan, Maeve/Baker, Mona, “Reporting that in translated English. Evidence for subconscious processes of explicitation?”, Across Languages and Cultures 1 (2000) 141–58. Payne, John Barton, Critical and comparative study of the Sahidic Coptic texts of the First book of Samuel (Princeton Theological Seminary: Unpublished dissertation, 1949). — “The Sahidic Coptic Text of I Samuel”, Journal of Biblical Literature 72 (1953) 51–62. Peters, Melvin, An Analysis of the Textual Character of the Bohairic of Deuteronomy (Society of Biblical Literature Septuagint and Cognate Studies 9; Missoula, Mont.: Scholars Press, 1979). Proverbio, D.V, “BARB.OR.2 (PSALTERIUM PENTAGLOTTUM)” in Paola Buzi/Delio Vania Proverbio (ed.), Coptic Treasures from the Vatican Library: A Selection of Coptic, Copto-Arabic and Ethiopic Manuscripts: Papers collected on the occasion of the Tenth International Congress of Coptic Studies (Studi e testi 472; Vatican: Biblioteca Apostolica Vaticana, 2012). 163–74. Schleifer, J., Bruchstücke der sahidischen Bibelübersetzung (Sitzungsberichte der Kais. Akademie der Wissenschaften in Wien, Philosophisch-Historische Klasse 170,1; Wien: Hölder, 1912). — Sahidische Bibel-Fragmente aus dem British Museum zu London (Sitzungsberichte der Kais. Akademie der Wissenschaften in Wien. Philosophisch-Historische Klasse. 162. Band, 6. Abhandlung; Wien, 1909). Schüssler, Karlheinz, Biblia Coptica: Die koptischen Bibeltexte: Das sahidische Alte und Neue Testament. Bd. 1, Lieferung 1: sa 1–20 (Wiesbaden: Harrassowitz, 1995). — Biblia Coptica: Die koptischen Bibeltexte: Das sahidische Alte und Neue Testament. Bd. 1, Lieferung 2: sa 21–48 (Wiesbaden: Harrassowitz, 1996). — Biblia Coptica: Die koptischen Bibeltexte: Das sahidische Alte und Neue Testament. Bd. 2, Lieferung 1: sa 121–184 (Wiesbaden: Harrassowitz, 2012). — Biblia Coptica: Die koptischen Bibeltexte: Das sahidische Alte und Neue Testament. Bd. 2, Lieferung 2: sa 185–260. Bearbeitet von Frank Feder/Hans Förster (Wiesbaden: Harrassowitz, 2015). Segal, Michael, “The text of the Hebrew Bible in light of the Dead Sea scrolls”, Materia Giudaica XII/1–2 (2007) 5–20. Shisha-Halevy, Ariel, Coptic Grammatical Categories: Structural Studies in the Syntax of Shenoutean Sahidic (Analecta Orientalia 53; Roma: Pontificium Institutum Biblicum, 1986). — ‘The “Tautological Infinitive” in Coptic: A Structural Examination’, Journal of Coptic Studies 1 (1990) 99–127.

Bibliography

235

Sipilä, Seppo, Between Literalness and Freedom: Translation technique in the Septuagint of Joshua and Judges regarding the clause connections introduced by ‫ ו‬and ‫( כי‬Publications of the Finnish Execetical Society 75; Helsinki: The Finnish Execetical Society/Göttingen: Vandenhoeck & Ruprecht, 1999). Soisalon-Soininen, Ilmari, Die Infinitive in der Septuaginta (Annales Academiae scientiarum Fennicae. Series B, tome 132,1; Helsinki: Suomalainen Tiedeakatemia, 1965). — Studien zur Septuaginta-Syntax. Anneli Aejmelaeus/Raija Sollamo (ed.) (Annales Academiae scientiarum Fennicae B 237; Helsinki: Suomalainen tiedeakatemia, 1987). Sollamo, Raija, Renderings of Hebrew Semiprepositions in the Septuagint (Annales Academiae scientiarum Fennicae B Diss 19; Helsinki: Suomalainen Tiedeakatemia, 1979). — “The LXX Renderings of the Infinitive Absolute Used with a Paronymous Finite Verb in the Pentateuch”, in Natalio Fernández Marcos (ed.), La Septuaginta en la investigación contemporánea (V Congreso de la IOSCS) (Textos y Estudios “Cardenal Cisneros” 34; Madrid: Instituto “Arias Montano”, 1985) 101–13. Stern, Ludwig, Koptische Grammatik (Leipzig: Weigel, 1880). Suciu, Alin, “The Sahidic Version of Jacob of Serugh’s Memrā on the Ascension of Christ”, Le Muséon 128 (2015) 49–83. Takla, H. N., “An Introduction to the Coptic Old Testament”, Coptica 6 (2007) 1– 115. — “Biblical Manuscripts of the Monastery of St. Shenoute the Archimandrite”, in G. Gabra/H.N. Takla (ed.), Christianity and Monasticism in Upper Egypt. Volume I. Akhmim and Sohag (Cairo, New York: The American University in Cairo Press, 2008) 155–67. Taylor, Bernard A., “The Old Greek Text of Reigns”, in A New English Translation of the Septuagint (Oxford: Oxford University Press, 2007) 244–70. Till, Walter/Sanz, Peter, Eine griechisch-koptische Odenhandschrift (Papyrus Copt. Vindob. K 8706) (Monumenta Biblica et Ecclesiastica 5; Rom: Päpstliches Bibelinstitut, 1939). Toury, Gideon, Descriptive Translation Studies – and beyond (The Benjamins Translation Library 4; Amsterdam/Philadelphia: John Benjamins Publishing Company, 1995). Tov, Emanuel, Textual Criticism of the Hebrew Bible (2nd edn; Minneapolis: Fortress Press, 2001). Ulrich, Eugene, The Dead Sea Scrolls and the Origins of the Bible (Studies in the Dead Sea Scrolls and Related Literature 1; Grand Rapids: Eerdmans, 1999). Vaccari, Alberto, “The Hesychian Recension of the Septuagint”, Biblica 46 (1965) 60–6.

Voitila, Anssi, Présent et imparfait de l´indicatif dans le Pentateuque grec: Une étude sur la syntaxe de traduction (Publications of the Finnish Exegetical Society 79; Helsinki: Finnish Exegetical Society, 2001). Vycichl, Werner, Dictionnaire Étymologique de la Langue Copte (Leuven: Peeters, 1983). — “Sullam”, in Aziz Suryal Atiya (ed.), The Coptic encyclopedia 8 (Digital Publisher Claremont Graduate University, School of Religion, 1991) A204a-A207a. Wessely, Carl, Griechische und koptische Texte theologischen Inhalts (Studien zur Palaeographie und Papyruskunde XV; Leipzig, 1914).

236

Bibliography

Westendorf, Wolfhart, Koptisches Handwörterbuch (Heidelberg: Carl Winter Universitätsverlag, 1965). Wevers, J. W., “The Use of Versions for Text Criticism: The Septuagint”, in Natalio Fernández Marcos (ed.), La Septuaginta en la investigación contemporánea (V Congreso de la IOSCS) (Textos y Estudios “Cardenal Cisneros” 34; Madrid: Instituto “Arias Montano”, 1985) 15–24. Wilmet, Michel, Concordance du Nouveau Testament sahidique. II. Les mots autochtones, 1 (Corpus Scriptorum Christianorum Orientalium 173; Louvain: Secrétariat du Corpus SCO, 1957). — Concordance du Nouveau Testament sahidique. II. Les mots autochtones, 2 (Corpus Scriptorum Christianorum Orientalium 183; Louvain: Secrétariat du Corpus SCO, 1958). — Concordance du Nouveau Testament sahidique. II. Les mots autochtones, 3 (Corpus Scriptorum Christianorum Orientalium 185; Louvain: Secrétariat du Corpus SCO, 1959).

6. Appendices 6.1. Collation of SaI 1 Sam 12: 4–5, 10–11 This fragment preserves four verses from Sahidic 1 Sam, verses 4–5 and 10– 11. The fragment is known as BL Or. 4916(3), and in Crum’s catalogue its reference number is 935.1 The fragment has been edited by Schleifer.2 This collation of mine is based on the photos acquired from the British Library. The fragment comes from Achmim.3 The material is parchment. The text on the recto is difficult to read; especially ⲡ and ⲛ are hard to differentiate. According to Crum, there were two columns.4 On the verso, there might be remnants of a letter on the left margin, line 8. The pages have 16 lines, each line containing 7–11 letters. The lines are ruled, which is especially clearly visible on the recto. If calculated with the help of SaM and an average of 9 letters per line, then there were 60 lines between these passages. With two columns this makes about 24–25 lines per page. This is reasonable since with only one column each page would have had 76 lines. recto, 1 Sam 12: 4–5 1 [ⲁⲩ]ⲱ [ⲙⲡ]ⲕ[ⲙⲟ] 2 ⲕϩⲛ̄ ϩⲛⲗ̄ ⲁⲁⲩ 3 ⲁⲩⲱ ⲙⲡⲕⲑⲙ5 4 ⲕⲉ ⲗⲁⲁⲩ· ⲁⲩ 5 ⲱ ⲙⲡⲕϫⲓ ⲗⲁⲁⲩ 6 ⲛⲧ ̄ ⲛ̄ ⲗⲁⲁⲩ ⲛⲣⲱ6 7 ⲙⲉ· ⲡⲉϫⲉ ⲥⲁ 8 ⲙⲟⲩⲏⲗ ⲉϩⲟⲩⲛ 9 ϩⲙ̄ ⲡⲗⲁⲟⲥ ϫⲉ 10 ⲡϫⲟⲉⲓⲥ ϥⲟ ⲙ 11 ⲙⲛⲧⲣⲉ ϩⲛ ⲧⲉ̣

1 2 3 4 5

6

12: 4

12: 5

Crum, Catalogue, 392. Schleifer, Bruchstücke. Crum, Catalogue, 392. Crum, Catalogue, 392. On this line, ⲡ looks like ⲛ and one could also read ⲛⲛⲕⲑⲙ-. However, it does not make any sense on this context and since ⲡ and ⲛ are so similar, I have deciphered the word as it is. The difference between ⲙ and ⲛ is impossible to see in this word because the letter is not well preserved. ⲱ is written with a small letter.

238 12 13 14 15 16

Appendices 7 ⲧⲛⲙ ̄ ⲏⲧⲉ ⲙⲡ̣[ⲟ] ⲟⲩ ⲁⲩⲱ ⲡ[ϥ] ⲭⲣⲓⲥⲧⲟⲥ ⲟ ⲙ[ⲙⲛ] ⲧⲣⲉ ϫⲉ ⲙ̣ [ⲡⲟⲟⲩ] ⲛϩⲟ[ⲟⲩ]…

verso, 1 Sam 12: 10–11 1 ⲁⲛ̣[ⲣⲛⲟⲃⲉ] 2 ϫⲉ ⲁⲛⲕⲱ ⲛⲥⲱⲛ 3 ⲙⲡϫⲟⲉⲓⲥ ⲁⲩ 4 ⲱ ⲁⲛⲣϩⲙ̄ϩⲁⲗ 5 ⲛⲃⲃ̄ ⲁϩⲁⲗⲓⲙ ⲛ8̄ 6 ⲙ̄ ⲙⲁ ⲛϣⲏⲛ· 7 ⲧⲉⲛⲟⲩ ϭⲉ ⲛⲟⲩ 8 ϩⲙ̄ ⲙ̄ⲙⲟⲛ ⲛⲧ ̄ ⲟ 9 ⲟⲧⲟⲩ ⲛⲛ ̄ ⲉⲛ 10 ϫⲁϫⲉ ⲁⲩⲱ ⲧⲛ̄ 11 ⲛ̣ⲁⲣϩ̄ ⲙ̄ϩⲁⲗ ⲛⲁⲕ· 12 [ⲁⲡ]ϫⲟⲉⲓⲥ ⲇⲉ ϫⲟ 13 [ⲟⲩ] ⲛϩ̄ ⲓⲉⲣⲟⲃⲟ 14 [ⲟⲗ ⲙ]ⲛ ⲃⲁⲣⲁⲕ 15 [ⲙⲛ ⲓⲉ]ⲫⲑⲁⲉ ⲙ̄ⲛ 16 […]

12: 10

12: 11

6.2. Collation of SaJ 1 Sam 29: 5–9, 30: 21–24 Two fragments, in different state of preservation, come from one manuscript with the reference number BL Or. 4916(4). In Crum’s catalogue their number is 936.9 The first fragment is in a poor condition, whereas the second is well preserved and easily legible. The fragments come from Achmim.10 Each line has 12–15 characters. Probably, the original page had 20 lines since the missing text between the recto’s last line and the verso’s first preserved line constitutes two lines.11 Schleifer (Bruchstücke, 15) reads ⲧⲛⲙⲏⲧⲉ and adds: “vor ⲛⲙⲏⲧⲉ fehlen vielleicht noch die Buchstaben ⲉⲧ, Spuren dafür sehe ich weder am Anfang dieser Zeile noch am Ende der vorangehenden”. I disagree here with Schleifer, and see clearly that after ⲧ on the previous line there comes a letter, and the beginning of this line reads ⲧⲛ-, not just ⲛ̄. 8 Layton (Grammar, §21b) mentions that the variant ⲃ̄- for ⲛ̄- is found in some early manuscripts. 9 Crum, Catalogue of the Coptic manuscripts, 392. 10 Ibid. 11 Calculated on the basis of SaM. 7

Appendices

recto, 1 Sam 29: 5–7 1 [..]ⲛ[..] 2 [..ϩ]ⲱⲱϥ ϩⲛ ̣ ̣̄ 3 ⲛⲉ[ϥⲧⲃⲁ] ⲁⲅⲭⲟⲩⲥ ⲇⲉ 4 ⲁ12ϥⲙ̣ [ⲟⲩⲧⲉ ⲉ]ⲇⲁⲩⲉⲓ ̣ⲇ̣ 5 ⲡⲉϫ̣ ⲁ̣ϥ [ⲛ]ⲁ̣ϥ̣ ϫⲉ [ⲡϫⲟ] 6 ⲉⲓⲥ ⲟⲛϩ ϫⲉ ⲕⲥⲟⲩⲧⲱⲛ 7 ⲛⲧ ̄ ⲟⲕ ⲁⲩⲱ ⲛⲧ ̄ ⲕ ̄ ⲟⲩ 8 ⲁⲅⲁ[ⲑⲟ]ⲥ ⲛⲛ ̄ ⲁϩⲣⲁⲓ̈· 9 ⲁⲩⲱ [ⲧⲉ]ⲕϩⲓ ̣ⲏ̣ ⲛⲃ̄ ⲱⲕ 10 ⲉ̣ⲃⲟⲗ [ⲁⲩⲱ ⲧⲉ]ⲕϩⲓ ̣ⲏ̣ 11 [ⲛ]ⲉⲓ [ⲉϩⲟⲩⲛ ⲛⲉ]ⲥⲛ̄ 12 [ⲙ]ⲙⲁ̣[ⲓ ϩⲛⲧ ̄ ⲡⲁ]ⲣⲉ̣ ⲙⲃⲟ 13 [ⲗⲏ ⲁⲩⲱ ϫⲉ ⲙ]ⲡⲓϩⲉ 14 [ⲉϩⲱⲃ ⲉϥϩⲟⲟ]ⲩ ⲉϩⲟⲩⲛ 15 [ⲉⲣⲟⲕ ⲛϫⲓⲛ ⲡ]ⲉ̣ϩⲟⲟⲩ 16 [ⲛⲧⲁⲕⲉⲓ ⲉϩ]ⲟⲩⲛ ⲛ̣[ⲁⲓ] 17 […]ⲩⲟ[.]ⲛ[̄ ϩ]ⲟ13 18 [ⲟⲩ ⲛⲥⲁⲧⲣⲁ]ⲡⲏⲥ ⲇⲉ 19 [ⲛⲥⲉⲧⲏⲧ ⲛϩ̄ ⲏ]ⲧ ⲛⲙ ̄ ̄14 20 [ⲙⲁⲕ ⲁⲛ ⲧⲉⲛ]ⲟⲩ ϭ̣ ⲉ *lower margin* verso, 1 Sam 29: 7–9 1 [ⲕⲧⲟⲕ ⲛⲅⲃⲱⲕ ϩⲛ̄ ⲟⲩ] 2 [ⲉⲓⲣ]ⲏⲛⲏ· [ⲛⲅⲧⲙⲉⲓⲣⲉ] 3 ⲛ[̄ ⲟ]ⲩⲕⲁ[ⲕⲓⲁ ⲛⲛⲁϩ]ⲣⲁ ̣ ̣[ⲩ] 4 ⲛ̣ⲛ ̄ ̣ⲥ̄ ⲁⲧ[ⲣⲁⲡⲏ]ⲥ̣ ⲛ̣ⲛ̣[ⲁⲗ] 5 [ⲗⲟⲫⲩⲗⲟⲥ ⲡⲉ]ϫⲉ ⲇⲁⲩ 6 ⲉ[ⲓⲇ] ⲛⲁ[ⲅⲭⲟⲩⲥ ϫ]ⲉ ⲟⲩ 7 ⲡⲉ ⲡⲛⲟⲃⲉ̣ ⲛ̣ⲧ ̄ ⲁⲓ̈ⲁⲁϥ 8 ⲛⲁⲕ· ⲁⲩⲱ ⲟⲩ ⲡ̣ⲉ ⲡϩⲱ 9 ⲃ ⲛⲧⲁⲕϩⲉ [ⲉⲣⲟϥ] ⲉϩⲟⲩⲛ 10 [ⲉⲡ]ⲉ[ⲕϩⲙϩⲁⲗ ⲛ]ϫ[ⲓⲛ ⲡⲉ] 11 ϩⲟⲟ[ⲩ ⲛⲧⲁⲓⲉ]ⲓ ⲛ[̄ ⲛⲁ] 12 ϩⲣⲁ[ⲕ ϣⲁ ⲉϩⲣⲁ]ⲓ̈ ⲉ[ⲡⲟ] 13 ⲟⲩ ⲛ[̄ ϩⲟⲟⲩ ϫⲉ] ⲛ[ⲛⲁⲉⲓ]

239

29: 5 29: 6

29: 7

29: 8

12 ⲁ is much larger than the other letters and written partly in the margin. 13 This line is impossible to reconstruct since the first visible letter cannot be ⲟ as one would expect but rather looks like ⲩ. To assume a mistake in the reconstruction seems too risky, even though the most reasonable reading is ϣⲁ ⲉϩⲣⲁⲓ ⲉⲡⲟⲟⲩ ⲛϩⲟⲟⲩ (ⲟⲩ on the following line) found in SaM. It is evident that the reading of SaA, ϣⲁ ⲡⲟⲟⲩ ⲛ̄ϩⲟⲟⲩ, is too short, giving only 9 letters on the line. 14 Little superlinear strokes as on the second line, recto, of the following fragment.

240

Appendices

14 ⲉⲃⲟⲗ [ⲁⲛⲟⲕ ⲉⲙⲓϣⲉ] 15 ⲙⲛ̄ ⲛϫ ̄ ̣ [ⲁϫⲉ ⲙⲡⲁϫⲟⲉⲓⲥ] 16 [ⲡ]ⲣⲣ̄ ⲟ [ⲁⲩⲱ ⲁⲅⲭⲟⲩⲥ] 17 15ⲁϥⲟⲩⲱ[ϣⲃ ̄ ⲉϥϫⲱ ⲙ] 18 ⲙⲟⲥ [ⲛⲇⲁⲩⲉⲓⲇ ϫⲉ ϯ] 19 ⲥⲟⲟ[ⲩⲛ̄ ϫⲉ ⲛⲧ ̄ ⲕ ⲟⲩ] 20 ⲁⲅⲁⲑ[ⲟⲥ ⲛⲧⲟⲕ] *lower margin*

29: 9

recto, 1Sam 30: 21–2216 1 ⲉⲡⲗⲁⲟⲥ· ⲁⲩⲱ ⲁϥϣⲁ 2 ϫⲉ ⲛⲙ ̄ ̄ⲙⲁⲩ ϩⲛ̄ ⲟⲩⲉⲓⲣⲏ 3 ⲛⲏ· ⲁⲩⲟⲩⲱϣⲃ ̄ ⲇⲉ 4 ⲛϭ ̄ ⲓ ⲣⲱⲙⲉ ⲛⲓⲙ ⲙ̄ⲡⲟ 5 ⲛⲏⲣⲟⲥ· ⲁⲩⲱ ⲗⲟⲓⲙⲟⲥ 6 ⲛⲓⲙ ⲛⲣ̄ ⲱⲙⲉ ⲛⲣ̄ ⲉϥ 7 ⲙⲓϣⲉ ⲉⲛⲧⲁⲩⲃⲱ̣ ⲕ 8 ⲉⲃ[ⲟⲗ ⲙⲛ]̄ ⲇⲁⲩⲉⲓⲇ ⲉⲩ 9 [ϫⲱ ⲙⲙⲟ]ⲥ· ϫⲉ ⲉⲃⲟ[ⲗ] 10 [ϫⲉ ⲙⲡⲟⲩⲉ]ⲓ ⲉⲃⲟⲗ[…]

30: 21 30: 22

verso, 1Sam 30: 23–2417 1 ⲙⲛⲛ ̄ ⲥ̄ ⲁ ⲧⲣⲉ ⲡϫⲟⲉⲓⲥ 2 ⲧⲁⲁⲩ ⲉϩⲣⲁⲓ̈ ⲉⲧⲟⲟⲧⲛ̄ 3 ⲛϥϩ̄ ⲁⲣⲉϩ ⲉⲣⲟⲛ· ⲁⲩⲱ 4 ⲡⲉⲓ̈ⲅⲉⲇ·ⲇⲟⲩⲣ· ⲏⲡⲉⲓ̈ 5 ⲙⲏⲏϣⲉ ⲉⲛⲧⲁϥⲉⲓ̈ ⲉ 6 ϩⲣⲁⲓ̈ ⲉϫⲱⲛ· ⲉⲁϥⲧⲁⲁⲩ 7 ⲉϩⲣⲁⲓ̈ ⲉⲧⲟⲟⲧⲛ·̄ ⲁⲩⲱ 8 ⲛⲓⲙ ϭⲉ ⲡⲉ[ⲧⲛⲁⲥⲱ]ⲧⲙ 9 [ⲛ]ⲥⲁ ⲛⲉⲧⲛ[̄ ϣⲁϫⲉ ⲉⲧⲉ] 10 [ⲛ]ⲁⲓ̈ ⲛⲉ[…]

30: 23

30: 24

6.3. Collation of SaU 1 Sam 14: 24–50 One large fragment, P.Duk.inv. 797, has preserved a significant amount of text. The fragment is described by Butts in his edition.18 15 A diple on the left margin. 16 A page number is preserved on the right, above the text: ⲧⲕⲋ. The last character of the page number is ⲉor ⲋ, but the fact that the next page is numbered ⲧⲕⲍ makes ⲋ more probable. 17 Page number ⲧⲕⲍ (327) preserved on the left, above the text. 18 Butts “P. Duk.inv. 797”, 7–9.

Appendices

recto, 1 Sam 14: 24–37 1 [ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲣⲁ̄ ⲛⲁϣ ⲙⲡⲗⲁⲟⲥ ⲉ]ϥ̣ϫ̣ⲱ[ⲙⲙⲟⲥ ϫⲉ 19 2 [ϥⲥϩⲟⲩⲟⲣⲧ̄ ⲛϭ ̄ ⲓ …. ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟ] ⲉⲓⲕ ϣⲁ ⲡ̣[ⲛⲁⲩ] 20 3 [ⲛⲣⲟⲩϩⲉ ϫⲉ ] ϫ̣ ̄ ⲁ̣ϫⲉ· ⲁⲩⲱ ⲡⲗⲁ̣ⲟ[̣ ⲥ ]21 4 [ⲧⲏⲣϥ̄ ⲙⲡⲉϥⲧⲉⲡ ⲟⲉⲓⲕ ⲡⲕ]ⲁϩ ⲇⲉ ⲧⲏⲣϥ̄ ⲛⲉϥⲁ̣[ⲣⲓⲥⲧⲁ] ̣ ̣[ⲙ̄23ⲡ] 5 [ⲛⲉ… ⲟⲩⲙⲁ ⲛϣⲏ]22ⲛ ⲇⲉ ⲛϯϩⲙⲉϥ ⲛⲉ̄ ⲃ̣ⲓⲱ ϣ̣ⲟ̣ⲟⲡ 6 [ⲙ̄ⲧⲟ ⲉⲃⲟⲗ ⲛⲧⲥ]ⲱϣⲉ· ⲁⲩⲱ [ⲡ]ⲗ̣ⲁⲟⲥ ⲁϥⲧ̣ [ⲁⲁϥ ⲉϩ…]24 7 [ⲉⲡⲙⲁ ⲛϯ ̄ ϩⲙⲉϥ ⲛ]ⲉⲃⲓⲱ· ⲛⲉⲣⲉ ⲡⲗⲁⲟⲥ ⲇⲉ [ⲙⲟⲟϣⲉ ⲉϥ] 25 8 [ϣⲁϫⲉ ⲉⲙⲛ ⲗⲁⲁ]ⲩ ⲕⲧⲟ ⲛⲧ ̄ ⲉϥϭⲓϫ ⲉϩ[ⲟ]ⲩ̣ⲛ̣ [ⲉⲧ]ⲉ̣ϥ̣ⲧ̣[ⲁⲡ] 9 [ⲣⲟ ϫⲉ ⲡⲗⲁⲟⲥ ⲁϥ]ⲣϩ̣ ̄ ⲟⲧⲉ ϩⲏⲧϥ̄ ⲙ̄ⲡⲁⲛⲁϣ ⲙ̣ ⲡϫⲟⲉⲓⲥ̣ 10 [ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲇⲉ ⲙⲡⲉ]ϥⲥⲱⲧ̅ⲙ̅ ϫⲉ ⲁⲡⲉϥⲉⲓⲱⲧ ⲧ̣ ⲁⲣⲕⲉ 11 [ⲡⲗⲁⲟⲥ ⲁⲩⲱ ⲓ̈ⲱⲛ]ⲁ̣ⲑⲁⲛ ⲁϥⲥⲟⲩⲧⲛ̄ ϩⲧⲏϥ ⲉⲃⲟⲗ ⲙ̄ 12 [ⲡⲉϥϭⲉⲣⲱⲃ ⲉⲧϩⲛ]̄ ⲧ̣ⲉϥϭⲓϫ ⲁϥⲥⲟⲡϥ̄ ϩⲙ̄ ⲡⲙⲟⲩⲗ̣ϩ ̄ 13 [ⲛⲉⲃⲓⲱ ⲁϥⲕⲧⲟ ⲛⲧ ̄ ⲉ]26ϥ̣ϭⲓϫ ⲉϩⲟⲩⲛ ⲉⲣⲱϥ ⲁⲛⲉϥⲃⲁ̣ⲗ ⲛ̣ 14 [ⲁⲩ ⲁⲩϫⲱⲧⲉ ⲟⲩⲁ ⲇⲉ ⲉⲃⲟ]ⲗ ϩⲙ̄ ⲡⲗⲁⲟⲥ ⲁϥⲟⲩⲱϣ[ⲃ ⲉϥ]ϫⲱ 15 [ⲙⲙⲟⲥ ⲛⲁϥ ϫⲉ ϩⲛ̄ ⲟⲩ]ⲁ̣ⲛⲁϣ ⲡⲉⲕⲉⲓⲱⲧ [ⲁϥⲧⲁⲣⲕⲉ ⲡⲗ] 16 [ⲁⲟⲥ ⲉϥϫⲱ ⲙⲙⲟⲥ] ϫⲉ ϥⲥϩⲟⲩⲟⲣⲉⲧ ⲛϭ ̄ ⲓ ⲡ̣ⲣⲱ ̣ ̣ ⲙ̣ ⲉ ̣ 17 [ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟⲉⲓⲕ]ⲙⲡⲟⲟⲩ ⲡⲗⲁⲟⲥ ⲇⲉ ⲁ̣ϥ̣ⲥⲱ[ϣⲙ̄ ⲁ]ⲩ̣[ⲱ] 18 [ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲁϥⲉⲓⲙ]ⲉ ⲡⲉϫⲁϥ ϫⲉ ⲡⲁ[ⲉⲓⲱⲧ ⲁϥ]ⲑ̣ⲙ̣ ̄[ⲕⲉ] 19 [ⲡ… ⲁⲛⲁⲩ ⲉ]ⲑⲉ ⲛⲧⲁ ⲛⲁⲃⲁⲗ ϫⲱ̣ [ⲧⲉ ⲉ]ⲃⲟⲗ ϫ̣ ⲉ ̣ 27 20 [ⲁⲓⲧⲉⲡ ⲟⲩϣⲏⲙ ⲉⲃⲟⲗ]ϩⲙ ̣ ̄ ⲡⲉⲉⲓⲉⲃⲓⲱ· ⲉⲉⲓⲉ ⲉⲛⲉ ⲛ 21 [ⲧⲁ ⲡⲗⲁⲟⲥ ϭⲉ ⲟⲩⲱⲙ ϩⲛ ⲟ]28ⲩⲱⲙ ⲙ̄ⲡⲟⲟⲩ ⲉⲃ[ⲟⲗ] ϩⲛ̄ 22 [ ⲛⲛⲉⲩϫⲁϫⲉ ⲛⲁⲓ] ⲛⲧⲁϥϩⲉ ⲉⲣⲟⲟⲩ ⲧⲉⲛⲟⲩ̣ 23 [ⲛⲉⲣⲉ ⲧⲉⲡⲗⲩⲅⲏ ⲛⲁⲁϣⲁⲓ̈] ⲡⲉ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫ[ⲩ]ⲗⲟⲥ 24 [ⲁⲩⲱ ⲁϥⲡⲁⲧⲁⲥⲥⲉ ⲉⲃⲟⲗ ϩⲛ̄ ⲛⲁⲗⲗⲟ]ⲫⲩⲗⲟⲥ ⲙ̄[ⲡⲉ]ϩ29 ̣ [ⲟⲟⲩ] 25 [ⲉⲧⲙ̄ⲙⲁⲩ ϩⲙ ⲙⲁⲭⲙⲁⲥ ⲡⲗⲁⲟⲥ ⲇ]ⲉ ⲁϥϩⲓⲥ[ⲉ ⲉⲙⲁ]30ⲧ̣ ⲉ ̣

241

14: 24

14: 25 14: 26

14: 27

14: 28

14: 29

14: 30

14: 31

19 Based on the space, there could be either ⲡⲉⲧⲛⲁⲟⲩⲉⲙ ⲟ̄ⲉⲓⲕ of SaM or ⲡⲣⲱⲙⲉ ⲉⲧⲛⲁⲟⲩⲉⲙ ⲟⲉⲓⲕ of SaA. 20 Both ⲉⲓⲉϫⲓⲕⲃⲁ ⲙ̄ⲡⲁϫⲁϫⲉ of SaM and ⲉⲓϫⲓ ⲕⲃⲁ ⲛⲛⲁϫⲁϫⲉ of SaA are possible reconstructions. 21 SaM reads here ⲇⲉ while SaA is without it. 22 There is space for either ⲛⲉⲣⲉ- of SaM or ⲛⲉⲟⲩⲛ- of SaA. 23 There is a supralinear stroke visible. 24 SaM has ⲉϩⲣⲁⲓ and SaA has ⲉϩⲟⲩⲛ. The spacing allows either one, and a five-letter word fits the spacing exactly. 25 In this line my reading differs from that of Butts. There is clearly a ϩ visible after ⲛ̄ⲧⲉϥϭⲓϫ ⲉ-. 26 In the line above, the reconstruction seems quite sure, and the characters match so that there are ca. 12 characters before ϭⲓϫ in this line. With the reconstruction according to SaM there are 14 characters, and according to SaA there would be 16 characters. Therefore I have reconstructed according to SaM. 27 Butts has marked ϩ as reconstructed, but its upper curve is visible on the fragment. 28 Butts reconstructs ϭⲉ in SaU in the first clause on the basis of the spacing, and following SaM. One should also take into account that perhaps SaU did not have the erroneous reading of SaMA ⲟⲩⲱⲙ ϩⲛ ⲟⲩⲱⲙ but the correct ⲟⲩⲱⲙ ϩⲛ ⲟ ⲩ ⲟⲩⲱⲙ which would require the same space. However, the shared reading of SaMA is the reason that I have followed Butts’ reconstruction with ϭⲉ. 29 The bottom curve of ϩ is visible. 30 The shape of the letter just before ⲧ does not seem to be ⲁ but it is nothing understandable in any case.

242 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 4937

Appendices

̣ [ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲁϥⲧⲁⲁϥ ⲉⲛ]ϣⲱⲗ· ⲁⲡⲗⲁⲟⲥ ϫⲓ ⲛϩⲛ̣ⲁ̣ϩⲉ̣ 31 [ⲛⲉⲥⲟⲟⲩ ⲙⲛ̄ ϩⲉⲛⲁ]ϩⲟⲟⲩ ⲁⲩⲱ ϩⲉⲛⲕⲧ̣ ⲏ̣ⲣ̣ ⲛ [ⲉϩⲉ ⲁⲩϣⲱⲱⲧ ⲙⲙⲟⲟⲩ ϩⲓϫⲙ̄] ⲡⲕⲁϩ· ⲁⲩⲱ32 [ⲡ]ⲗ̣ⲁⲟ[ⲥ] [ⲛⲉϥⲟⲩⲱⲙ ⲙⲛ̄ ⲛⲉϥⲥⲛⲟϥ ⲁ]ⲩⲧⲁⲙⲉ ⲥⲁⲟⲩⲗ [ⲇⲉ ⲉⲩϫⲱ ⲙⲙⲟⲥ ⲛⲁϥ ϫⲉ ⲁⲡⲗⲁ]ⲟⲥ ⲣⲛ̄ ⲟⲃⲉ ⲉⲡϫⲟ33 [ⲉⲓⲥ ⲁϥⲟⲩⲉⲙ ⲁϥ ⲙⲛ ⲡⲉϥⲥⲛⲟϥ] ⲁⲩⲱ ⲡⲉϫⲉ [ⲥ]ⲁ̣[ⲟⲩ]ⲗ̣ [ⲛⲁⲩ ϫⲉ ⲁⲛⲓ ⲟⲩⲛⲟϭ ⲛⲱⲛⲉ] ⲉⲡⲉⲉⲓⲙⲁ ⲉⲃⲟⲗ [ϩⲛ̄ ⲅⲑ̄ ⲁⲉⲓⲙ ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ] ⲇ̣ ⲉ ̣ ⲛ̣ⲁⲩ ϫⲉ ⲃⲱⲕ ⲉⲃⲟⲗ ϩⲙ̣ [ⲡⲗⲁⲟⲥ ⲛⲧⲉⲧⲛϫ ̄ ⲟⲟⲥ ⲛⲁ]ⲩ̣ ϫ̣ ⲉ ̣ ⲙⲁⲣⲉ ⲡⲟⲩⲁ ⲡⲟⲩⲁ [ⲙⲙⲱⲧⲛ̄ ⲉⲓⲛⲉ ⲁ̣ ⲙ]ⲡⲉϥⲙⲁⲥⲉ34 ⲁⲩⲱ ⲡⲉϥ [ⲉⲥⲟⲟⲩ ⲛϥ̄ ϣ ̄ ⲁⲁⲧϥ̄ ϩⲓϫⲙ̄ ⲡⲓ]ⲱⲛⲉ ϫⲉ ⲛⲛⲉⲧⲛ̄ [ⲉⲣⲛⲟⲃⲉ ⲉⲡϫⲟⲉⲓⲥ ⲉⲧⲣⲉⲧⲉⲧⲛ]ⲟⲩⲉⲙ ⲁϥ ⲙⲛ̄ ⲡⲉϥⲥⲛⲟϥ· [ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ̄ ⲡⲟⲩⲁ ⲡⲟⲩⲁ] ⲁϥⲉⲓⲛⲉ ⲙ̄ⲡⲉⲧⲛ.̄ . [ⲧⲟⲟⲧϥ ⲁⲩϣⲱⲱⲧ ⲙⲙⲟⲟⲩ ⲙ]ⲡⲙⲁ ⲉⲧ[ⲙⲙⲁⲩ] [ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥⲕⲱⲧ ⲙⲡⲙ]ⲁ ⲉⲧ̣ⲙ̣ⲙⲁⲩ ⲛⲟ[ⲩ]ⲑⲩⲥⲓ[ⲁ] [ⲥⲧⲏⲣⲓⲟⲛ ⲙⲡϫⲟⲉⲓⲥ ⲡⲁ]ⲓ̈ ⲡⲉ ⲡⲉⲑⲩⲥⲓⲁⲥⲧⲏⲣⲓⲟⲛ [ⲛⲧⲁ ⲥⲁⲟⲩⲗ ⲁⲣⲭⲓⲥⲑⲁⲓ ⲛ]ⲕⲟⲧϥ̄ ⲙⲡϫⲟⲉⲓⲥ ⲁⲩⲱ..35 [ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲙⲡⲗⲁⲟⲥ ϫⲉ ⲙⲁⲣ]ⲉ̣ⲛⲃⲱⲕ ⲉϫⲛ̄ ⲛⲁⲗⲗⲟⲫ̣ⲩ̣ [ⲗⲟⲥ ⲛⲧⲉⲩϣⲏ ⲛⲧ ̄ ⲉⲛϩⲁⲣ]ⲡⲁⲥ[ⲉ ⲉⲃⲟⲗ]ⲛϩ̄ ⲏⲧ̣ⲟⲩ̣ [ϣⲁ ⲡⲛⲁⲩ ⲛϩⲧⲟⲟⲩⲉ ⲛⲧⲉⲛⲧⲙ̄]ϣⲉ̣ϫ̣[ⲡ ⲟ]ⲩⲣⲱⲙⲉ [ⲉⲃⲟⲗ ⲛϩ̄ ⲏⲧⲟⲩ ⲡⲉϫⲁ]ⲩ ⲛⲁϥ ϫⲉ ⲡ̣ⲁ̣[ⲅⲁⲑⲟⲛ] [ⲧⲏⲣϥ ⲁⲣⲓϥ ⲛⲁϩⲣⲁⲕ ⲡⲉ]ϫⲉ ⲡⲟⲩ[ⲏⲏⲃ] ϫⲉ [ⲙⲁⲣⲉⲛϯ]36 [ⲙⲡⲉⲛⲟⲩⲟⲓ ⲉⲡⲉⲉⲓⲙⲁ]ⲉⲣⲁⲧϥ̄ ⲙ̄[ⲡⲛⲟⲩⲧⲉ ⲛⲧⲛϫⲛⲟⲩϥ] [ⲁⲩⲱ ⲥⲁⲟⲩⲗ ⲁϥϫ]ⲛⲉ [ⲡⲛⲟⲩⲧⲉ ϫⲉ ⲉⲛⲉ]

verso, 1 Sam 24: 37–45 138 [ⲧⲁⲃ]ⲱ̣ [ⲕ] ⲉⲣⲁⲧⲟⲩ ⲛⲛ[ⲁⲗⲗⲟⲫⲩⲗⲟⲥ ⲁⲩⲱ ⲛⲉⲕⲛⲁⲧⲁⲁⲩ ⲉϩⲣⲁⲓ] 2 [ ]ⲡ[ ]ⲁ[ ] 3 [ⲉⲧⲙⲙ]ⲁⲩ· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ [ⲛⲁⲩ ϫⲉ ⲁⲛⲓⲛⲉ ⲉⲡⲉⲓⲙⲁ ⲛⲛⲉⲧⲁⲅ] 4 [ⲙⲁ] ⲧⲏⲣⲟⲩ ⲙ̄ⲡⲓⲏ̅ ⲗ̄ ̄ ⲛⲧ ̄ ⲉⲧⲛⲉ̄ ⲓⲙⲉ [ⲁⲩⲱ ⲛⲧ ̄ ⲉⲧⲛⲛ ̄ ⲁⲩ ϫⲉ ⲛⲧⲁ] 5 ⲡⲉⲉⲓⲛⲟⲃⲉ ϣⲱⲡⲉ ϩⲛ̄ ⲛⲓⲙ ⲙ̄ⲡⲟ̣ⲟ[̣ ⲩ ϫⲉ ⲡϫⲟⲉⲓⲥ ⲟⲛϩ̄ ⲡⲉⲛ] 6 ⲧⲁϥⲧⲟⲩϫⲉ ⲡⲓⲏ̅ ⲗ̄ ̄ ϫⲉ ⲉϥϣⲁⲛⲧ[ⲁⲙⲟ]ⲛ̣ [ⲉⲧⲃⲉ ⲓⲱⲛⲁⲑⲁⲛ] 7 ⲡⲁϣⲏⲣⲉ ϩⲛ̄ ⲟⲩⲙⲟⲩ ϥⲛⲁⲙⲟⲩ ⲡ̣ⲉ ̣ [ⲁⲩⲱ ⲙⲡⲉ ⲗⲁⲁⲩ ϣⲁ]

14: 32

14: 33

14: 34

14: 35 14: 36

14: 37

14: 38 14: 39

31 This could be ⲟϩⲉ as Butts has it, but the shape of the letter more resembles an ⲁ. 32 There is something above the word but it is impossible to make out any letters. 33 There is something above the line, resembling ϫ and some other character but one cannot see it properly. 34 This word is corrupted in SaM to read ⲙⲡⲉϥⲣⲁⲥⲧⲉ. Drescher has proposed the correct form already in his edition, and now it is assured by this fragment. 35 ⲁ is clearly visible, after it the junction of the pieces seems to be a little inaccurate, thus hiding something under the join. 36 This line seems a bit long to be correct. 37 After this line there are at least two lines where nothing can be deciphered. However, there is no gap in the text. 38 Above this line there seems to be one written line, but it is not legible.

Appendices

8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

ϫ̣ ⲉ ⲉⲃⲟⲗ ϩⲙ̄ ⲡⲗⲁⲟⲥ ⲧⲏⲣϥ̅ ·̅ ⲡⲓⲏ[ⲗ ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ⲛⲛ ̄ ⲣⲱⲙⲉ ⲧⲏ] ⲣⲟ̣ ⲩ ⲙ̄ⲡⲓⲏ̅ ⲗ̄ ̄ ϫⲉ ⲛⲧⲱⲧⲛ̄ ⲧⲉⲧⲛⲁ̣[ϣⲱⲡⲉ ϩⲛ̄ ⲟⲩⲙⲛⲧϩⲙ̄ϩⲁⲗ] [ⲁ]ⲛⲟⲕ ϩⲱⲱⲧ ⲟⲛ ⲙⲛ̄ ⲓ̈ⲱⲛⲁⲑ[ⲁⲛ ⲡⲁϣⲏⲣⲉ ⲧⲉⲛⲛⲁϣⲱⲡⲉ ϩⲛ]̄ ̣ ̣ⲟ̣ⲥ ̣ ⲛ̣[ⲥⲁⲟⲩⲗ ϫⲉ ⲡⲁⲅⲁⲑⲟⲛ ⲉⲧ] ⲟⲩⲙⲛⲧ ̄ ϩⲙ̄ϩⲁⲗ· ⲡⲉϫⲉ ⲡ̣ⲗⲁ [ⲛⲁ]ϩⲣⲁⲕ ⲁⲣⲓϥ· ⲡⲉϫⲉ [ⲥⲁⲟⲩⲗ ϫⲉ ⲡϫⲟⲉⲓⲥ ⲡⲛⲟⲩⲧⲉ ⲙⲡⲓⲏⲗ] ⲉ̣ⲧ̣ⲃⲉ̣ ̣ [ⲟⲩ ⲙ̄]ⲡ39 ̄ ⲁ[ϫ]ⲉ ⲙ̣ ⲡⲟⲟⲩ [ⲙⲛ̄ ⲡⲉⲕϩⲙ̄ϩⲁⲗ ⲉⲛⲉⲣⲉ] ̄ ⲕϣ [ⲡⲛⲟⲃⲉ ⲉ]ⲣⲟⲓ ⲏ ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲡⲁϣⲏ[ⲣⲉ ⲡϫⲟⲉⲓⲥ ⲡⲛⲟⲩⲧⲉ ⲙ] ⲡⲓⲏⲗ̄ ̄ ϯⲙ̣ ⲁⲉⲓⲛ ⲛⲁⲛ ⲁ[ⲩⲱ] ⲉⲕϣ[ⲁⲛϫⲟⲟⲥ ϩⲓⲛⲁⲓ ⲧⲁⲓ ⲧⲉ ⲑⲉ] ⲉⲧⲛⲁϣⲱⲡⲉ ϯⲙⲁⲉⲓⲛ ⲡϫⲟⲉⲓⲥ [ⲉⲡⲉⲕⲗⲁⲟⲥ ⲡⲓⲏⲗ̄ ̄ ⲁⲩⲱ] ⲡ̣ⲉⲕ̣ ⲗⲏⲣⲟⲥ ⲁϥⲉⲓ ⲉϫⲛ̄ ⲓ̈ⲱⲛⲁⲑ̣ⲁ[ⲛ ⲙⲛ̄ ⲥⲁⲟⲩⲗ ⲡⲗⲁⲟⲥ ⲁϥ] ⲉⲓ ⲉ̣ⲃⲟ̣ ⲗ̣· ⲡⲉϫⲉ ⲥⲁⲟⲩⲗ ϫⲉ [ⲛⲉϫ ⲕⲗⲏⲣⲟⲥ ϩⲛ̄ ⲧⲁⲙⲏⲧⲉ] ⲁⲩⲱ [ϩⲛ̄ ⲧⲙⲏⲧ]ⲉ̣ ⲛ̣ⲓⲱ ̈ ⲛⲁⲑⲁⲛ ⲡ̣[ⲁϣⲏⲣⲉ ⲁⲩⲱ ⲡⲉⲧⲉⲣⲉ] ⲡϫⲟⲉⲓⲥ ⲛⲁⲉ̣ⲓ ̣ⲛ̣ⲉ ̣ ⲙ̄ⲡⲉ[ⲕⲗⲏⲣⲟⲥ ⲉϫⲱϥ ⲉϥⲉⲙⲟⲩ ⲡⲉ] ϫⲉ ⲡⲗⲁⲟⲥ [ⲉϩ]ⲟⲩⲛ [ϩⲛ̄ ⲥⲁⲟⲩⲗ ϫⲉ ⲛⲛⲉ ⲡⲉⲓϣⲁϫⲉ] [ϣⲱ]ⲡⲉ [ⲥⲁⲟⲩⲗ] ⲇⲉ ⲁϥ[ⲁⲛⲁⲅⲕⲁⲍⲉ ⲙⲡⲗⲁⲟⲥ ⲁⲩⲛⲉϫ] [ⲕⲗⲏ]ⲣⲟⲥ ϩⲛ̄ [ⲧⲉϥⲙⲏⲧⲉ ⲁⲩⲱ ϩⲛ̄ ⲧⲙⲏⲧⲉ ⲛⲓ̈ⲱⲛⲁⲑⲁⲛ ⲡⲉϥ] ϣ40[ⲏⲣ]ⲉ· ⲁ ⲡ[ⲕⲗ]ⲏⲣⲟ[ⲥ ⲉⲓ ⲉϫⲛ̄ ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲡⲉϫⲉ ⲥⲁ] ⲟⲩ̣[ⲗ] ⲛ̣ⲓ̄ ⲱ ̈ ⲛⲁ̣[ⲑⲁⲛ ϫⲉ ⲙⲁⲧⲁⲙⲟⲓ ϫⲉ ⲟⲩ ⲡⲉ ⲛⲧⲁⲕⲁⲁϥ] ⲁⲩⲱ ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲁϥⲧⲁ[ⲙⲟϥ ⲉϥϫⲱ ⲙⲙⲟⲥ ϫⲉ ⲛⲧⲁ] ⲓ ̣ⲥ̣ⲉⲡ̣ ⲛϩ̄ ⲧⲏϥ ⲙ̄ⲡ̣ⲁϭⲉⲣⲱⲃ̣ [ⲛⲟⲩⲕⲟⲩⲓ ⲛⲉⲃⲓⲟ ⲁⲓⲧⲟ] ⲡϥ̄ ⲁⲩⲱ ⲉⲓⲥ ϩⲏⲏⲧⲉ ⲁ̣ⲛ̣[ⲟⲕ ⲉⲓⲛⲁⲙⲟⲩ ⲉⲧⲃⲉ ⲡⲁⲓ ⲡⲉ] ϫⲉ ⲥⲁⲟⲩⲗ ⲛⲁϥ ϫⲉ ⲛⲁⲓ̈ [ⲛⲉⲧⲉⲣⲉ ⲡⲛⲟⲩⲧⲉ ⲛⲁⲁⲁⲩ ⲁⲩⲱ] ⲛⲁⲓ̈ ⲉϥⲉⲟⲩⲁϩⲟⲩ ⲉⲣⲟⲓ̈ [ϫⲉ ϩⲛ̄ ⲟⲩⲙⲟⲩ ⲕⲛⲁⲙⲟⲩ ⲙⲡⲟⲟⲩ] ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲡⲉϫⲁϥ ⲛ[̄ ⲥⲁⲟⲩⲗ ϫⲉ ⲉⲛⲉ ⲡⲟⲟⲩ ϩⲏ] ϥⲛⲁⲙⲟⲩ ⲛϭⲓ ⲡⲉⲛⲧⲁ[ϥⲉⲓⲣⲉ ⲙⲡⲉⲓⲛⲟϭ ⲛⲟⲩϫⲁⲓ ϩⲙ̄] ⲡⲓⲏ̅ ⲗ̄ ·̄ ⲡϫⲟⲉⲓⲥ ⲟⲛϩ̄ ̄ ϫ[ⲉ ⲉϣⲱⲡⲉ ⲟⲩⲛ̄ ⲟⲩⲃⲱ ⲛⲟⲩⲱⲧ ϩⲛ]̄ ⲧⲉϥⲁⲡⲉ ⲛⲁϩⲉ ⲉϩⲣⲁⲓ̈ [ⲉϫⲙ ⲡⲕⲁϩ ϫⲉ ⲡⲗⲁⲟⲥ ⲙⲡⲛ] ⲟⲩⲧⲉ ⲡⲉⲛⲧⲁϥⲉ̣ⲓ ̣ⲣⲉ ⲙ̣ [ⲡⲉⲓⲛⲟϭ ⲛⲛⲟⲃⲉ ⲁⲩⲱ ⲡⲗⲁⲟⲥ ⲁϥϣ] ⲗⲏⲗ ⲉϩⲣⲁⲓ̈ ⲉⲡⲛ̣[ⲟⲩⲧⲉ ⲉϩⲣⲁⲓ ⲉϫⲛ̄ ⲓ̈ⲱⲛⲁⲑⲁⲛ ⲙⲡⲉϩⲟⲟⲩ ⲉⲧⲙ] [ⲙⲁⲩ] ⲁⲩⲱ ⲙ̄ⲡ[ⲉϥⲙⲟⲩ ⲥⲁⲟⲩⲗ ⲇⲉ ⲁϥⲃⲱⲕ ⲉⲣⲁⲧⲟⲩ ⲛⲛⲁⲗⲗⲟ]

243 14: 40

14: 41

14: 42

14: 43

14: 44

14: 45

6.4. Collation of SaY 1 Sam 15: 1241 1 ⲁⲩⲱ ⲥⲁⲙ̣ [ⲟⲩⲏⲗ ⲁϥϣⲟⲣⲡ̄ ϥ]̄ 2 ⲁϥⲃⲱⲕ ⲉⲣⲁ̣[ⲧϥ ⲙⲡⲓⲏⲗ42 ⲁⲩⲧⲁ] 39 There is a mistake in Butts here; there is no ⲉ in ⲡⲕ̄-. 40 The bottom curve of ϣ is visible; it seems to be the first character on this line. 41 This manuscript has the reference number CCC Ms 541 no. 38. The fragment contains text only on one side, the other side has an ornament decoration and lines that seem to mark the margins of the page. 42 SaM has ⲙⲡⲉϥϩⲧⲟⲟⲩⲉ after Israel, but it does not fit in here. Therefore, I have not reconstructed it. La116 is the only other manuscript with this same omission.

244 3 4 5 6

Appendices

ⲙⲉ ⲥⲁⲟⲩⲗ43 ⲉⲩ̣[ϫⲱ ⲙⲙⲟⲥ ϫⲉ] ⲥⲁⲙⲟⲩⲏⲗ ⲁϥ[ⲃⲱⲕ ⲉϩⲣⲁⲓ ⲉⲡⲕⲁⲣⲙⲏⲗⲟⲥ]44 ⲁⲩⲱ [ⲥ]ⲁⲙⲟ[ⲩⲏⲗ45 ⲁϥⲃ̣[

6.5. Greek manuscripts and their groupings46 Codices : ABMV Fragments: 842 2328 – 242; 246–8.12–13.18–20 845 1316–18.20–21; 1323 – 141.3–4; 188–25 846 2411–17; 2420 – 2520; 3112 – fin libri 867 46.9.13.15–16 Manuscript groups : O = 247–376 L = 19–82–93–108–127 19’ = 19–108 CI = 98-(243)-379–731 98’ = 98–379 CII = 46–52–236–242–313–328–530 46’ = 46–52 242’ = 242–328 C’ = CI + CII a = 119–527–799 b = 121-509* d = 44–68–74–106–107–120–122–125–134-(370)68’= 68–122 120’= 74–106–120–134-(370) 134’= 120–134 107’= 44–107–125–610 f = 56–246 s = 64–92–130–314–381–488–489-(762) 64’= 64–381 488’= 488–489

610

43 SaM has the connective ⲇⲉ here, after Saul. SaY is the only manuscript that leaves out the connective, Greek unanimously attests καί. 44 Obviously, this line is too long if it is reconstructed according to the text of SaM. Perhaps, one can leave out ⲉϩⲣⲁⲓ, but this is pure quesswork without any manuscript support. 45 Here, SaM reads ⲁⲩⲱ ⲉⲓⲥ ⲥⲁⲟⲩⲗ but SaY does not agree with it. SaY agrees with those manuscripts (B 376 b h 130 244 460 La116) that do not contain the word ‘look’ but only the conjunction. There are no Greek manuscripts in agreement with SaY concerning the proper name. Interestingly, La116 and LaM read, like SaY, the proper name Samuel. 46 This is Aejmelaeus’ preliminary grouping of the manuscripts for the critical edition.

Appendices

245

Manuscripts without grouping: 29 71 244 245 460 707 55 158* 318* (342) 554* bold – those representing the shorter text in Chs. 17 – 18 * – mixti

Abstract The Sahidic version of 1 Sam is an ancient daughter version of the Septuagint. Because the Sahidic translation was made before most of the Greek manuscripts we know were copied, it potentially contains ancient readings no longer preserved or only faintly attested in the Greek tradition. This study considers the Sahidic version of 1 Sam as a translation and how it may best be used in Greek textual criticism. The aim of this study is twofold. First, one chapter examines the translation technique of the Sahidic translator. Second, based on this knowledge of the translation technique, a further chapter analyzes the affiliations between the Sahidic manuscripts as well as the affiliations between the Sahidic version and Greek traditions. In the translation-technical section, clause connections and translator’s additions feature prominently. These foci were motivated by the tendency of the citations given in the Brooke/McLean edition of the Greek text. If the Sahidic appears in their apparatus, the citation mostly concerns conjunctions or additions found in the Sahidic. The first part of the translation-technical analysis examines clause connections. Clause connections appear throughout the text and offer, therefore, a fruitful starting point for a translation-technical study. When working with this material, comparative analyses rely on clausal sectioning as a basis for comparable semantic units. The second topic in the translation-technical part is translator’s additions. In the chapter concerning the affiliations of the Sahidic text, detailed textual analyses prevail. Seven passages were selected in which several Sahidic manuscripts have preserved the text. Additional criteria in selecting the passages were the age of the fragments and the coverage of different parts of the book. These analyses describe the textual character of each Sahidic manuscript, and search for the existence of secondary readings and/or corruptions. In the cases where variants occur within the Sahidic tradition, this study seeks to establish whether these readings relate to Greek variants or solely derive from Sahidic transmission. This study supports the creation of a new critical edition of the Septuagint of 1 Sam for the Göttingen series. With respect to this edition, the primary goal is to identify the affiliations of the Sahidic version. This translation-technical study, however, will additionally allow for a more careful and accurate citation of the Sahidic version within the critical apparatus.

Index of biblical references Ex 7: 20

153n

Num 17: 2 20: 11

153n 153n

Deut 32: 30–43

19n

1 Samuel 1: 2 1: 3 1: 4 1: 5 1: 6 1: 7 1: 8 1: 10 1: 12 1: 13 1: 14 1: 15 1: 18 1: 18–2: 9 1: 19 1: 22 1: 23 1: 26 1: 27 1: 28 2: 1 2: 1–10 2: 3 2: 5 2: 7 2: 10 2: 11 2: 13 2: 14 2: 16 2: 17 2: 18

65, 65n 133n 52 86 126, 126n, 129n, 133n 118n, 129n, 136 71n, 72, 72n, 102n, 104, 130n, 135, 135n 132n 51n, 101, 101n 223 71n 36n, 40 128n, 129n, 136 20 129n 57, 92, 119n, 129n 76, 98, 133, 133n 118n 68n, 70n 87n, 90 18n, 132n 19, 22, 186n 114, 115n, 124, 124n 128n, 133n 57 148n, 153n, 183n 38, 133n 109, 109n 70n 78, 90n, 107n, 109, 124n, 128n 130n 124n

2: 19 2: 20 2: 23 2: 24–30 2: 25 2: 26 2: 27 2: 28 2: 29 2: 30 2: 34 2: 35 2: 35–36 2: 36 3: 2 3: 3 3: 4 3: 5 3: 6 3: 6–9 3: 7 3: 8 3: 8–20 3: 9 3: 10 3: 11 3: 12 3: 14 3: 16 3: 17 3: 18 3: 19 3: 21 4: 1 4: 2 4: 3 4: 4 4: 5 4: 6 4: 9 4: 10 4: 13 4: 16

118 128n, 133, 133n 130n, 132n 19 90n, 91, 115n, 135, 135n 59n 116 128n 45, 72 128n, 129n 87n, 129n 87n 22 126n 52n 87n, 125n 124n, 130n 67n, 124n, 128n, 129n 22, 39, 56, 67n, 124n, 128n, 130, 130n, 219 19 119 41, 124n, 128n, 130n, 131 20 36n, 46, 58n, 67n, 91, 102n, 126n, 210n 62, 102n, 128n 111 125, 125n, 129n 95, 95n, 124n 124n, 126n 59, 59n, 128n 128n 130n 115n 63 129n 22, 71n, 72n, 128n, 129n 124n 129n, 132n 126n, 127 22, 98n, 99, 221n 129n 60, 60n, 158n 36n

248 4: 17–18 4: 18 4: 19 4: 20 4: 22 5: 2 5: 4 5: 5 5: 6 5: 6–11 5: 7 5: 8 5: 9 5: 10 5: 11 5: 12 6: 2–20 6: 3 6: 4 6: 5 6: 6 6: 7 6: 8 6: 9 6: 11 6: 11–10: 3 6: 12 6: 13 6: 14 6: 14–21 6: 17 6: 18 6: 19 6: 20 6: 20–21 6: 21 7: 2 7: 3 7: 5 7: 7 7: 8 7: 8–8: 1 7: 9 7: 10 7: 11 7: 12 7: 13 7: 14 7: 15 7: 15–8: 3

Index of biblical references 53 52n, 115 42, 129n, 135n 82n, 128n 118n 135n 52, 59n, 101n 78, 79, 132n 126n, 134, 134n 20 128n 128n, 129, 129n 51, 120, 124n, 132n 52, 71n 128n 135n 18 96, 133n 124n, 126n, 128n 100n 101n, 108 130n 36n 36n, 90n, 97, 124n, 125, 128n, 135n 18, 130n 18 129n 48n, 129n 129n, 130n 19, 48 88n, 133n, 135n 87n, 125n, 132n 103, 130n 71 f, 124n, 128n 48 48n, 133n 51n, 129n 113n, 126n 128n 132n 124n, 128n, 141 f, 159, 160, 161 22, 30, 141 132n, 143, 160 49, 49n, 50n, 133n, 144 f, 160 145, 160, 161, 224 134n, 145 f, 159, 160 146, 160, 161 60n, 147, 159, 160, 161 148, 160 20

7: 16 7: 17 8: 1 8: 2 8: 2–3 8: 3 8: 5 8: 6 8: 7 8: 8 8: 8–9 8: 9 8: 10 8: 11 8: 12 8: 12–13 8: 14 8: 17–22 8: 18 8: 19 8: 20 9: 2 9: 3 9: 4 9: 5 9: 6 9: 8 9: 9 9: 10 9: 11 9: 12 9: 12–19 9: 13 9: 15 9: 16 9: 18 9: 19 9: 20 9: 21 9: 21–10: 2 9: 22 9: 23 9: 24 9: 25 9: 26 9: 27 10: 1 10: 2

148, 160, 161, 224 149, 159, 160 48n, 52n, 150, 159 65n 148n 56, 56n 129, 129n 20, 132, 132n 78, 98n 20, 85n, 129n 54 20, 86, 88, 88n, 125n 132n 88n, 125n, 128n 135n 42 132n 20 88n 124n, 129n 40n, 55n, 133, 133n 65n, 132n 84n, 133n 60, 60n, 82n, 128n 70n, 115n, 116, 135, 135n, 206n 87n, 100, 124n, 133n 129, 129n, 133n 118n, 128, 128n, 130n, 131 36n, 87n, 125n 94 36n, 48n, 124, 124n, 135n 20 42, 85, 109, 109n, 119n 128n 103n, 129n 71n 128n 103, 103n, 124n, 135n 128n, 150 f, 160, 161 21, 30, 141, 159, 160 151 f, 159, 160 152, 160 87n, 134n, 135n, 152 f, 159, 160, 161 128, 128n, 153 f, 160, 161 52n, 53, 128n, 129n, 154 f, 159, 160, 161 155 f, 160, 161 134n, 156 f, 159, 160, 161, 224 36n, 84n, 88n, 109n, 124n,

Index of biblical references

10: 3 10: 3–4 10: 5 10: 6 10: 7 10: 8 10: 9 10: 11 10: 12 10: 14 10: 16 10: 18 10: 19 10: 21 10: 22 10: 24 10: 26 10: 27 11: 2 11: 5 11: 7 11: 8 11: 9 11: 11 11: 12 11: 13 11: 14 11: 15 12: 2 12: 3 12: 4 12: 4–5 12: 5 12: 5–8 12: 7 12: 9 12: 10 12: 10–11 12: 11 12: 12 12: 13 12: 14 12: 14–15 12: 15 12: 16 12: 17 12: 19 12: 19–24 12: 20

130n, 132n, 156n, 157 f, 159, 160 19 38 58, 87n, 109n, 125n 129n 101n, 124n, 133n 119, 149n 66, 67n, 111 36n, 73, 128n 62, 73n, 124n 45, 71n, 84n, 124n, 128n, 133n 157 113n 77, 78n, 79n, 87n, 124n 84n, 128n 128n 130n 129n 129n 57, 118n 74, 158 70n, 132n 70n, 124n 65, 117, 126n, 128n 97 f 128n 133n 113n, 149n 149n 126n, 129n 71n, 82, 124n, 133n 133n, 162, 164 20, 30, 162 162 f, 164 21 125, 125n 221n 163, 164 20, 30, 162 164 124n 128n 92 92 36n, 90n, 92 87 135n 130n 20 56n

12: 21 12: 22 12: 23 12: 23–24 12: 24 12: 25 13: 2–5 13: 3 13: 4 13: 5 13: 6 13: 7 13: 8 13: 8–9 13: 9 13: 10 13: 11 13: 12 13: 12–13 13: 13 13: 14 13: 15 13: 16 13: 17 13: 19 13: 21 14: 1 14: 3 14: 4 14: 5 14: 6 14: 7 14: 8 14: 10 14: 11 14: 12 14: 13 14: 14 14: 15 14: 17–32 14: 18 14: 18–19 14: 19 14: 20 14: 21 14: 21–22 14: 22 14: 23 14: 24 14: 24–32 14: 24–46 14: 24–50

249 87n, 103n, 128n 128n, 132n 56n 86 125n 135, 135n 20 129n 129n, 149n 63, 126n, 129n, 220 f 20 129n, 149n 62, 128n, 149n 20 100, 130n, 133n 52n, 59 109, 128n, 129n 37, 149n 46 69n, 107n, 110, 129n 84n, 125n 126n, 129n, 149n 43, 126, 126n 115n 102n 81, 134n 59n, 132n 19 87n, 125n 19 67n, 96n, 126n 19, 107n, 110, 129n 128n 19, 90n, 128n 19, 126n, 128n 128n 38, 218 f 128n 128n 18 132n, 221 51 f 36n, 51n, 59, 60n, 115n, 116 65n, 124n 135n 45 40n, 55n 36n, 135n 88n, 165 f, 173, 174 30, 164, 165 21 f 164

250 14: 25 14: 26 14: 27 14: 28 14: 29 14: 29–30 14: 30 14: 31 14: 32 14: 33 14: 34 14: 36 14: 37 14: 38 14: 41 14: 42 14: 42–43 14: 43 14: 45 14: 47 14: 49 14: 49–52 14: 50–52 15: 1 15: 1–3 15: 2 15: 3 15: 4 15: 6 15: 10 15: 11 15: 12 15: 13 15: 14 15: 15 15: 16 15: 17 15: 18 15: 19 15: 20 15: 21 15: 21–26 15: 22 15: 23 15: 24 15: 26 15: 29 15: 30 15: 32 15: 33 15: 33–34 15: 35

Index of biblical references 166, 173, 174 158, 167, 172, 173 168, 173, 176 36n, 128n, 169, 173, 174, 187 102n, 104, 170, 173 76 f 88n, 102n, 170 f, 173 171 f, 173 172, 173 115, 128n 128n 98, 98n, 128n, 129n 96n 128n 90n, 94, 128n, 135n 87n, 88, 88n 37 115n 95n, 128n 87n, 125n 36n, 65n 21 18n 21, 65 18n 21, 109 82n, 132n 124n 44, 70n, 118n, 129n 128n 129n, 222 22, 149n 19 128n 19, 100n, 128n 128n 19 98n 19, 41 f 19, 120 128n, 149n 21 124n 69, 70n 124n 103 21, 81, 130n, 223 21, 128n, 135n 94n, 113, 128n 85, 85n, 101n, 107, 149n 58 102n, 104

16: 1 16: 1–13 16: 2 16: 2–8 16: 4 16: 4–5 16: 5 16: 6 16: 8 16: 9 16: 11 16: 11–18 16: 12 16: 13 16: 14 16: 16 16: 17 16: 18 16: 22 16: 23 17 17: 1 17: 2 17: 3 17: 4 17: 7 17: 8

67, 67n, 71n, 132n 22 44, 71n, 128n 19 48n, 71n, 128n 21 128n 76, 118 81, 129n, 130n 130n 71n, 119n, 128n, 136 19 103n 61n 203 59, 59n, 68n, 70n, 84n, 119n 33n, 68n, 69n, 70n 128n 128n 51n, 119n 17n, 18n, 186n, 188, 218n 63, 134n 63n 87n, 134, 134n 65n, 219 134n 68, 68n, 70n, 71n, 72, 73n, 130n, 219 17: 9 59n, 90n, 91, 221n 17: 10 70n, 128n 17: 11–19: 1 25 17: 13 36n, 65n, 129n 17: 14 47, 129n, 133n 17: 16 115n, 132n 17: 17 59n, 129n 17: 17–54 22 17: 20 87n, 98n, 99, 128n, 134n 17: 23 67n, 128n, 219 17: 24 118n 17: 25 58n, 87n, 128n, 132n, 133n 17: 28 128n 17: 29 128n, 129n, 133n 17: 30 124n, 135n 17: 31–44 18, 19 17: 32 134n 17: 33 174 f, 187 17: 33–44 30, 174 17: 33–19: 5 19 17: 34 87n, 101n, 134n, 168n, 175, 187 17: 35 97, 128n, 168n, 176, 187

Index of biblical references 17: 36

68n, 70n, 87n, 101n, 107n, 135n, 177, 187 17: 37 128n, 178 f, 187, 188, 224 17: 38 179, 187, 188 17: 39 115, 180 f, 187, 188 17: 40 182 f, 187, 188 17: 41 49n, 50n, 114, 115n, 183, 187, 188 17: 42 184, 188 17: 43 108n, 128n, 184 f, 187, 188, 189 17: 44 185, 186, 189 17: 45 190n 17: 46 178 17: 49 128n, 135n 17: 51 129n 17: 52 135n 17: 53 115n 17: 55 95n 17: 55–18: 6 25 18: 6 203 18: 6–7 196 18: 6–9 22 18: 9 61n, 133n 18: 10 60n 18: 11 135n 18: 12 190n 18: 15 109, 133n 18: 17 60n, 61 18: 19 124n, 126n 18: 20 128n 18: 22 128n 18: 23 96n, 128n, 221 18: 25 124n, 128n 18: 28 124n 18: 28–22: 7 18 18: 29 50n 19: 2 54n, 83, 84n 19: 3 87n, 88n, 115, 115n, 133n 19: 4 132n, 133n, 135n 19: 5 128n, 129n, 132n, 135n 19: 6 95n, 133n 19: 7 133n 19: 9 132n 19: 10 48n, 132n, 133n 19: 14 128n 19: 15 129n 19: 16 124n 19: 17 133n 19: 20 40n, 43, 55n, 115n 19: 21 40n, 55n, 126n, 128n 19: 22 40, 55n, 71n, 74, 128n

19: 23 20: 1 20: 2 20: 3 20: 5 20: 6 20: 7 20: 8 20: 9 20: 10 20: 11 20: 12 20: 13 20: 14 20: 15 20: 16 20: 17 20: 19 20: 21 20: 22 20: 24 20: 25 20: 26 20: 27 20: 29 20: 30 20: 31 20: 32 20: 33 20: 34 20: 35 20: 37 20: 38 20: 39 20: 41 20: 42 21: 1 21: 3 21: 4 21: 5 21: 6 21: 7 21: 8 21: 9 21: 10 21: 12 21: 13 21: 13–14 21: 15 21: 15–16 21: 16 22: 1–6

251 48n, 119n, 130n 124n, 129n, 133 65, 70n, 71n, 74, 219 48, 70n, 98n, 115, 115n, 130n 38, 115n, 135n 114, 124n, 128n, 129n 48n, 68n, 70n, 90n, 124n 96n 90n, 93, 124n 128n 36n 109n 70n, 98n, 128n 90n 118 135n 50n, 102n, 104 f, 219 36n, 88n 68n, 70n, 132n, 135n 46n, 68, 68n, 70n, 129n 48n, 49, 51n 124n 102n, 105, 132n 129n, 136 95, 129n, 132n 79 54n, 84, 87n, 89, 115n, 220 125n 221 129n, 130n 98n, 132, 132n 87, 87n 113, 113n, 129n 125n 48n, 129n 135n 49n, 50, 129n 88, 88n 102n, 132n 36, 101n, 107n 102n, 106, 118, 134n, 219 113 135n 88n, 96n, 124n, 128n 70n 72n 48n 19 71n 72 71n 19

252 22: 2 22: 3 22: 4 22: 5 22: 6 22: 6–13 22: 7 22: 8 22: 12 22: 14 22: 15 22: 17 22: 18 22: 21–23 22: 22 22: 23 23: 1 23: 2 23: 3 23: 3–4 23: 4 23: 5 23: 6 23: 7 23: 9 23: 10 23: 11 23: 13 23: 14 23: 15 23: 16 23: 17 23: 18 23: 19 23: 20 23: 21 23: 22 23: 23 23: 25 23: 26 23: 26–28 23: 27 23: 28 24: 1–23 24: 3 24: 5 24: 6 24: 7 24: 8 24: 10 24: 11

Index of biblical references 132n 98n, 129n 117 129n 124n 20 48n, 62, 135n 36n, 82n, 118n 128n, 133n, 135n 130n 69n, 70n, 73n, 124n, 134n 44, 124n 133n, 220 18 115n, 128n 67, 67n, 70n, 84n, 87n, 129n, 132n 112, 128n 36n 132n 50 49n, 169n, 195n 51, 51n, 130n, 132n 52n, 118n 124n 48n, 132n, 190n 84n 54, 98n, 132n 87n, 125n, 132n, 135n 84n 84n, 126n 132n 55, 56n, 102n, 132n 132n 128n, 134n 129n, 135n 133n 33n, 68n, 69n, 70n, 87n, 99 58n 84n, 129n 130n 22 36n 43, 128n, 130n 22 84n, 87n, 133n, 134n 124n 102n, 103 f 97, 102n 115n 84n, 128n, 129n 56n, 68n, 70n, 108

24: 12 24: 14 24: 15 24: 16 24: 17 24: 17–20 24: 18 24: 19 24: 20 24: 21–25: 28 25: 2 25: 3 25: 5 25: 6 25: 7 25: 8 25: 10 25: 11 25: 12 25: 13 25: 14 25: 15 25: 16 25: 19 25: 20 25: 21 25: 22 25: 23 25: 25 25: 26 25: 29 25: 30 25: 31–34 25: 32 25: 33–34 25: 34 25: 35 25: 36 25: 37 25: 38 25: 41 25: 42 25: 44 26: 1 26: 2 26: 5 26: 6 26: 7 26: 7–25 26: 8 26: 10

19, 44, 82n, 128n 98n 19, 71n, 124n 148n 52n, 59n, 71n, 73n, 107 19 81, 135n 109, 129n 96 f, 98n, 128n 18 51n, 118n, 133n 48n, 124n, 126n, 133n 133n 124n 57n, 82, 87n, 89 33n, 54n, 70n, 88n, 128n, 134n 71, 169n, 195n 71n, 87n, 135n 60n, 61, 128n 128n, 129n 128n 81n, 118n 87n 128n 51n, 59n, 210n 130n, 132n 69n, 70n, 95n 62 36n, 88n, 132n 84n, 98n, 135n 84n, 113, 129n 58n, 102n, 129n 23 134n 86 39, 94 87, 103n 124n 52n, 53, 107n, 108, 129n 52n, 59n 129n 130, 130n 135n 125n, 128n 84n 87n 128, 128n 222 19 129n 82, 95n, 135n, 219 f

Index of biblical references 26: 11 26: 12 26: 14 26: 15 26: 16 26: 17 26: 18 26: 19 26: 20 26: 21 26: 22 26: 23 26: 24 26: 25 27: 1 27: 4 27: 5 27: 8 27: 8–9 27: 10 27: 11 28: 2 28: 3 28: 3–5 28: 3–25 28: 4 28: 7 28: 10 28: 11 28: 12 28: 14 28: 15

69, 70n, 132n 58, 128n 72 f, 128n 71n, 72n 129n 71n, 73n 71n, 74n 124n, 128n, 132n 84n, 98, 98n, 135n, 219 87n, 128n 132n 109 85, 98n 115n 84n, 91, 129n, 132n 54n, 55, 84n, 103, 129n 133n, 166n 134n 22 124n, 130n 70n 128n 129n, 134n 22 19 55, 63, 203 84n 95n 128n, 129n 36n, 39 124n 54, 56n, 70, 70n, 71n, 74, 128n 28: 16 37 f, 72n, 128n 28: 16–30: 1 18 28: 16–30: 5 191n 28: 17 98n, 133n 28: 17–18 106 f 28: 18 68n, 70n, 101n 28: 19 109n, 132n 28: 19–24 18n 28: 20 62n, 80 28: 21 128n, 130n, 133n 28: 22 102n, 106, 129n 28: 23 46 28: 24 62n 28: 25 129n 29: 3 61n, 129n 29: 3–9 20 29: 4 87n, 129n, 134n 29: 5 88 f, 88n, 189 f, 196, 219, 224 29: 5–9 21, 30, 189

29: 6

253

61n, 130n, 190 f, 195, 196, 224 29: 7 37, 193, 195 29: 8 61n, 71, 125n, 130n, 193 f, 195 29: 8–11 18n 29: 9 169n, 192n, 194 f, 196, 219, 224 29: 10 63, 87n, 219 30: 1 18n, 36n 30: 2 76 30: 3 81n 30: 3–5 18 30: 3–11 20 30: 4 98, 98n 30: 5–24 18 30: 6 135n 30: 7 132n 30: 8 114, 128n, 129n 30: 9 48n, 129n, 132n, 197 30: 10 197 30: 12 57, 57n, 82 30: 13 71n 30: 15 96, 128n, 130n 30: 16 51n 30: 18 51n, 88n 30: 19 67, 67n, 88n, 132, 132n 30: 20 128n 30: 21 48n, 56, 56n, 190n, 196 f, 202, 224 30: 21–24 21, 30, 196 30: 22 78, 79, 88n, 129n, 198 f, 201, 202 30: 23 128n, 134n, 200, 202 30: 24 101n, 107n, 201, 202 30: 24–31: 13 18 30: 25 51n, 59n, 60n, 61, 130n 30: 26 124n 31 30, 66n 31: 1 202 f, 216, 217, 224 31: 1–13 19, 202 31: 2 203 f, 216, 217 31: 3 204 f, 216 31: 4 43, 67n, 129n, 205 f, 216, 217, 218 31: 5 40n, 55n, 207, 216, 217 31: 6 129n, 204, 208, 210, 216, 217 31: 7 208 f, 216 31: 8 204, 209 f, 216, 217, 218 31: 9 210 f, 216, 217, 218, 224 31: 10 212 f, 216, 217, 218 31: 11 88n, 213 f, 216

254 31: 12 31: 13

Index of biblical references 213, 214 f, 216, 217, 218, 224 60n, 215, 216, 217, 218

2 Samuel 1: 1–11: 1 11: 2 15: 20–30 17: 13 17: 26 18: 12

Jonah 1: 14

70n

Hab 17n 17n 18n 100n 149n 91

1: 13 3: 1–7

142n 19n

Matthew 6: 18

99

Mark 1 Kings

2: 26

2: 11 17n 2: 12–21: 43 17n 22 17n

Luke

2 Kings 2: 13

153n

2 Chron 35: 7

10: 34 12: 58 13: 13 14: 18 20: 37 24: 32

106n

186n 109n 84 115n 109n 109n

153n

Acts Ps 73: 4 148: 14

114 153n

20: 16 23: 24

99 186n

1 Cor Jer 9: 9 28: 62

186n 186n

Dan 12: 7

153n

1: 29 15: 39

99 186n

De Septuaginta Investigationes (DSI) Herausgegeben von Anneli Aejmelaeus, Kristin De Troyer, Wolfgang Kraus, Emanuel Tov.

Vol 9 / Band 9: Tuukka Kauhanen / Anneli Aejmelaeus (ed.) The Legacy of Barthélemy 50 Years after Les Devanciers d’Aquila 2017. Ca. 232 pages, hardcover ISBN 978-3-525-54062-6 eBook: ISBN 978-3-647-54062-7

Les Devanciers d’Aquila by Dominique Barthélemy (1963) is perhaps the most influential work on the textual history of the Septuagint during the recent decades. The present volume mainly consists of papers presented at the 50th anniversary symposium of Devanciers in St Andrews, Scotland, in 2013.

Band 7 / Vol 7: Raimund Wirth Die Septuaginta der Samuelbücher Untersucht unter Einbeziehung ihrer Rezensionen 2016. 271 Seiten mit 6 Tabellen, gebunden ISBN 978-3-525-53694-0 eBook: ISBN 978-3-647-53694-1

Raimund Wirth rückt die umfassende Untersuchung der Übersetzungsweise der Samuel-Septuaginta ins Zentrum seiner Arbeit. Hinzu kommen Beobachtungen zu den späteren Rezensionen und eine Skizze der Textgeschichte bis hin zum Masoretischen Text.

Band 6 / Vol 6: Sven Lesemann „Und Gideon starb in einem guten Greisenalter“ Untersuchungen zu den hebräischen und griechischen Texttraditionen in Ri 6–8 unter Einbeziehung des jüdisch-hellenistischen und frühen rabbinischen Schrifttums 2016. 411 Seiten mit einer Abb., gebunden ISBN 978-3-525-53681-0 eBook: ISBN 978-3-647-53681-1

Sven Lesemann untersucht die Texttradition eines der textgeschichtlich komplexesten Bücher der Septuaginta – das Richterbuch – am Beispiel des Gideon-/Jerubbaal-Zyklus in Ri, 6,18,35.

Vol 5 / Band 5: Mariachiara Fincati The Medieval Revision of the Ambrosian Hexateuch Critical Editing between Septuaginta and Hebraica Veritas in Ms. Ambrosianus A 147 inf. 2016. 456 pages with 3 figures, hardcover ISBN 978-3-525-53618-6 eBook: ISBN 978-3-647-53618-7

At the end of the 11th century the Greek text of an ancient unical Hexateuch was thoroughly revised in order to conform to the Hebraica Veritas. Mariachiara Fincati analyses such a revision by comparisons with extant Jewish and Christian Greek translations of the Bible.

www.v-r.de

De Septuaginta Investigationes (DSI) Herausgegeben von Anneli Aejmelaeus, Kristin De Troyer, Wolfgang Kraus, Emanuel Tov.

Band 4 / Vol 4: Siegfried Kreuzer / Marcus Sigismund (Hg.) Der Antiochenische Text der Septuaginta in seiner Bezeugung und seiner Bedeutung 2013. 284 Seiten, gebunden ISBN 978-3-525-53608-7 eBook: ISBN 978-3-647-53608-8

Der antiochenische Text ist eine der wichtigsten Textformen der Septuaginta, deren Alter und Bedeutung allerdings umstritten ist. Der vorliegende Band dokumentiert die aktuelle Diskussion und neue Perspektiven.

Vol 3 / Band 3: Tuukka Kauhanen The Proto-Lucianic Problem in 1 Samuel 2012. 240 pages, hardcover ISBN 978-3-525-53459-5 eBook: ISBN 978-3-647-53459-6

The study assesses the agreements between the Lucianic text and the earliest witnesses of 1 Samuel with a special emphasis on methodology. It also provides much background information on the use of the Septuagint among early Christian writers.

Vol 2 / Band 2: Timothy Michael Law Origenes Orientalis The Preservation of Origen‘s Hexapla in the Syrohexapla of 3 Kingdoms 2011. 383 pages, hardcover ISBN 978-3-525-53405-2 eBook: ISBN 978-3-647-53405-3

T.M. Law evaluates the worth of the Syrohexapla as a witness to the Hexapla of Origen in 3 Kingdoms (1 Kings) and thereby prepares the way for a new critical edition of these hexaplaric fragments.

Vol 1 / Band 1: Staffan Olofsson As a Deer Longs for Flowing Streams A Study of the Septuagint Version of Psalm 42-43 in its Relation to the Hebrew Text 2011. 235 pages, hardcover ISBN 978-3-525-53383-3 eBook: ISBN 978-3-647-53383-4

This volume of the new DSI series offers a detailed and comprehensive investigation of Ps 42-43 in Hebrew and Greek from a translation technical and text critical perspective.

www.v-r.de