Jesus and Gospel Traditions in Bilingual Context: A Study in the Interdirectionality of Language 3110266172, 9783110266177

Table of contents :
Acknowledgements
Abbreviations
1 The Directionality of the Transmission of the Jesus and Gospel Traditions: A History of Research
1.1 Herman Gunkel
1.2 Sitz im Leben Unidirectionality from Judaeo-Palestinian into Hellenistic Tradition
1.2.1 Martin Dibelius
1.2.2 Rudolf Bultmann
1.2.3 Gerd Theissen
1.3 Modal Unidirectionality from Oral into Written Tradition
1.3.1 Scandinavian School
1.3.1.1 Harald Riesenfeld
1.3.1.2 Birger Gerhardsson
1.3.2 Orality Theory
1.3.2.1 Werner Kelber
1.3.2.2 Joanna Dewey
1.3.3 The Earliest Written Tradition: Earle Ellis
1.4 Linguistic Unidirectionality from Aramaic into Greek Tradition
1.4.1 Aramaic Language as a Criterion
1.4.1.1 Charles C. Torrey
1.4.1.2 Matthew Black
1.4.1.3 Joachim Jeremias
1.4.1.4 Joseph Fitzmyer
1.4.2 Greek Language as a Criterion
1.4.2.1 Alexander Roberts
1.4.2.2 Aubrey Argyle
1.4.2.3 Nigel Turner
1.4.2.4 Robert Gundry
1.4.2.5 Stanley Porter
1.5 Interdirectionality Hypothesis
1.5.1 Sitz im Leben Interdirectionality: Martin Hengel
1.5.2 Modal Interdirectionality
1.5.2.1 Form Criticism: Helmut Koester
1.5.2.2 Textual Criticism: David Parker
1.5.2.3 Scandinavian School: Samuel Byrskog
1.5.2.4 Liturgical Use and Canon Formation: Harry Gamble
1.5.2.5 Notebook Theory: Graham Stanton and Richard Bauckham
1.5.3 Linguistic Interdirectionality: E. P. Sanders
1.5.4 Interdirectionality Hypotheses and Gospel Studies
Part I: Bilingualism of First-Century Palestine and the Roman Near East
2 Bilingualism and Diglossia
2.1 Definitions
2.1.1 Bilingualism
2.1.2 Bilingual Community
2.1.3 Bilingualism and Multilingualism
2.1.4 Early Bilingualism vs. Late Bilingualism
2.1.5 Primary Bilingualism vs. Acquired Bilingualism
2.1.6 Oral Bilingualism vs. Literate Bilingualism
2.1.7 Stages of Bilingualism
2.1.8 Balanced Bilingualism vs. Dominant Bilingualism
2.1.9 Matrix Language vs. Embedded Language
2.1.10 Substratum, Adstratum and Superstratum
2.1.11 Ancestor, Daughter and Sister Language
2.2 Diglossia in First Century Palestine?
2.2.1 Charles Ferguson
2.2.2 Joshua Fishman
2.2.3 Pinchas Lapide
2.2.4 Chaim Rabin
2.2.5 Bernard Spolsky
2.2.6 Jonathan Watt
2.3 Bilingualism and Diglossia Models
2.3.1 Characteristics of the Four Languages in First-Century Palestine
2.3.2 Language Preference Theory
2.3.3 Maximalism and Minimalism
2.3.4 Biliteracy of First-Century Palestine
3 Bilingualism of Jews in First-Century Palestine
3.1 Regional Bilingualism
3.1.1 Inscriptional Evidence
3.1.2 Papyrological Evidence
3.1.3 Population Geographical Evidence
3.2 Personal Bilingualism
3.2.1 Military Invasion and Colonization
3.2.2 Migration or Inflow
3.2.3 Ethnic Awareness
3.2.4 Bilingual Parents or Region
3.2.5 Formal Education
3.2.6 Occupation
3.3 Bilingualism of First-Century Palestine and Interdirectionality
4 Bilingualism of Jews in the First-Century Diaspora
4.1 Grounds for Bilingualism in the Jewish Diaspora
4.1.1 Bilingualism of the Roman Empire
4.1.2 Successive Immigration
4.1.3 Periodic Connection with Jerusalem
4.1.4 Learning Aramaic in Jerusalem
4.2 Alexandria
4.2.1 Bilingualism of Ptolemaic and Early Roman Alexandria
4.2.2 Archaeological Evidence
4.2.3 Successive Immigration
4.2.4 Periodic Connection with Jerusalem
4.3 Antioch
4.3.1 Bilingualism of Roman Antioch
4.3.2 Bilingualism of the Antiochene Jews
4.3.3 Bilingualism of the Antiochene Christians
4.4 Bilingualism of the Jewish Diaspora and Interdirectionality
5 The Bilingualism of the Earliest Christian Church in Jerusalem
5.1 The State of Affairs
5.2 Bilingualism of the Hebrews and Hellenists
5.2.1 Hebrews: Aramaic-Matrix Christians
5.2.2 Hellenists: Greek-Matrix Christians
5.3 The Bilingual Seven
5.3.1 Onomastica
5.3.2 Geographical Evidence
5.3.2.1 Caesarea Maritima
5.3.2.2 Samaria
5.3.2.3 Ethiopia
5.3.2.4 Antioch
5.3.3 Performance of Their Duty
5.4 Bilingualism of the Jerusalem Church and the Interdirectionality of the Jesus Tradition
Part II: Interdirectional Transmission of the Jesus and Gospel Traditions in Bilingual Contexts at the Levels of Syntax, Phonology, and Semantics
6 Syntax
6.1 Contact-Induced Syntactic Change Hypothesis in New Testament Scholarship
6.1.1 Adolf Deissmann
6.1.2 Linguistic Relatedness between Syntactic Changes
6.1.3 Bilingual Jewish Greek Theory
6.1.4 Hellenistic Greek Theory
6.1.5 Semitisms and Septuagintalisms
6.2 Internal-Induced Syntactic Change Hypothesis
6.2.1 Comparisons between Cognitive Linguistics and Other Linguistic Theories
6.2.1.1 Theoretical and Empirical
6.2.1.2 Diachronic and Synchronic
6.2.1.3 Langue and Parole
6.2.1.4 Significance of Semantics
6.2.2 Some Basic Concepts
6.2.2.1 Grammaticalization
6.2.2.2 Unidirectionality Hypothesis
6.2.2.3 Five Principles of Grammaticalization
6.2.2.4 Metaphor and Metonymy
6.2.2.5 Abstractness
6.2.2.6 Grammatical Polysemy
6.2.3 Grammaticalization and Language Contact
6.3 Syntactic Change of New Testament Greek as Grammatical Polysemy
6.3.1 Verbs
6.3.1.1 SAY Verb: λέγω
6.3.1.2 Movement Verbs: ἔρχομαι and πορεύομαι
6.3.1.3 Posture Verbs
6.3.1.3.1 στήκω and ἵστημι
6.3.1.3.2 ἀνίστημι and ἐγείρω
6.3.1.3.3 καθίζω and κάθημαι
6.3.2 Conjunctions
6.3.2.1 Ἵνα
6.3.2.2 Ὅτι
6.3.3 Adverbs: Matthean τότε, Marcan εὐθύς, and Johannine οὖν
6.4 Syntactic Change as Grammatical Polysemy and Interdirectionality
7 Phonology
7.1 Linguistic Factors of Transliterated Variants in Bilingual Contexts
7.1.1 Different Phonetic System
7.1.2 Representation
7.1.3 Phonetic Change
7.1.4 Dialects
7.2 Three Views of Variant Spellings in Transliteration
7.2.1 Orthographical View
7.2.2 Variational View
7.2.2.1 Ephraim Speiser
7.2.2.2 James Barr
7.2.2.3 Alan Millard
7.2.2.4 Some Scholars of New Testament Greek
7.2.3 Bilingual View
7.3 Variant Spellings as Transliterated Allolexemes in Bilingual Contexts
7.3.1 Transliterated Allolexemes
7.3.2 Consonants
7.3.2.1 Plosives and Fricatives
7.3.2.2 Sibilants
7.3.2.3 Liquids
7.3.2.4 Gutturals
7.3.2.5 Additions of Consonants
7.3.3 Vowels
7.4 Variant Spellings as Transliterated Allolexemes in the Four Gospels and Acts
7.4.1 Local Proper Nouns
7.4.1.1Ἰερουσαλήμ and Ἱεροσόλυμα
7.4.1.2 Ναζαρά and Ναζαρέθ
7.4.1.3 Γεννησάρ and Γεννησαρέτ
7.4.2 Personal Proper Nouns
7.4.2.1 Σίμων/Συμεών and Πέτρος/Κηϕᾶς
7.4.2.2 Ἐλισάβη and Ἐλισάβετ
7.4.3 Other Transliterated Words
7.4.3.1 The Cry from the Cross
7.4.3.2 Μεσσίας/Χριστός
7.5 Variant Spellings as Allolexemes and Interdirectionality
8 Semantics
8.1 Aramaic Embedded Words as Codeswitching
8.1.1 Codeswitching, Interference, and Borrowing in a Bilingual Society
8.1.2 Codeswitching in New Testament Scholarship
8.1.3 Types of Codeswitching
8.1.4 Pragmatic Functions of Codeswitching
8.1.4.1 Vividness
8.1.4.2 Emphasis
8.1.4.3 Politeness
8.1.4.4 Solidarity
8.2 ἀμήν Found in All Four Gospels
8.3 Codeswitchings in the Gospel of Mark
8.3.1 Ἀββα ὁ πατήρ
8.3.2 Ταλιθα κουμ and εφφαθα in the Miracle Stories
8.4 Codeswitchings in the Gospel of Luke and Acts
8.4.1 Language Change
8.4.2 Personal Names
8.4.2.1 Σιμών/Συμεών and Πέτρος/Κηφᾶς
8.4.2.2 Σαῦλος/Παῦλος and Σαούλ
8.4.3 Names of Places
8.4.3.1 Ἰερουσαλήμ and Ἱεροσόλυμα
8.4.3.2 Ναζαρά and Ναζαρέθ
8.5 Embedded Aramaic Words as Codeswitchings and Interdirectionality
9 Summary and Suggestions for Further Study
9.1 Summary of Results
9.2 The Gospel Tradition as Jesus Tradition
9.3 The Relationship between Historical Criticism and Literary Criticism
9.4 The Synoptic Problem
9.5 The Historical Jesus
9.6 Textual Criticism
9.7 Jewish Christianity and Gentile Christianity
9.8 Bilingual Christology
Bibliography
Index of Ancient Sources
Index of Languages and Place Names
Index of Modern Authors
Index of Subjects

Citation preview

Sang-Il Lee Jesus and Gospel Traditions in Bilingual Context

Beihefte zur Zeitschrift für die neutestamentliche Wissenschaft und die Kunde der älteren Kirche

Herausgegeben von

James D. G. Dunn · Carl R. Holladay Hermann Lichtenberger · Jens Schröter Gregory E. Sterling · Michael Wolter

Band 186

De Gruyter

Sang-Il Lee

Jesus and Gospel Traditions in Bilingual Context A Study in the Interdirectionality of Language

De Gruyter

ISBN 978-3-11-026617-7 e-ISBN 978-3-11-026714-3 ISSN 0171-6441 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internetat http://dnb.dnb.de. 쑔 2012 Walter de Gruyter GmbH & Co. KG, Berlin/Boston Printing: Hubert & Co. GmbH & Co. KG, Göttingen ⬁ Printed on acid-free paper Printed in Germany www.degruyter.com

For my family, Lee, Seong-Hwa (㢨㉥䞈) Noh, Geong-Ae (⊬ᷱ㚔) Chae, Soo-Jung (㵸㍌㥉) Lee, Un (㢨㡴) Lee, Sol (㢨㋈)

Acknowledgements Bilingualism is everywhere in my life. My daughters who were born in UK and grew up in USA are early bilinguals in English and Korean. My parents under Japanese colonial domination became late bilinguals in Korean and Japanese. They had to use both Korean and Japanese names under the Nation Obliteration Policy of Korean culture and language. Both Durham Korean Church in Durham, UK and Korean Presbyterian Church of Purdue in West Lafayette, IN, USA where I have served are bilingual Christian communities like Christian churches in Jerusalem and the Roman Near East in the first century AD. Such linguistic-related situations inspired me to make an attempt to write this topic. This book is a revision of my Ph.D thesis, originally submitted to Department of Theology and Religion, Durham University, under supervisions of Professor James D. G. Dunn and Professor Loren T. Stuckenbruck in 2008. First of all, I am extremely appreciative of Professor Dunn’s helpful insight, sharp criticism, and warmly encouragement for first two years. After his retirement, Professor Stuckenbruck supervised my research for four and a half years. I am immensely grateful for his supervision. He guided me with unceasingly encouragement and suggested invaluable comments and thoughtful criticisms. Whenever we discussed my arguments in his office, I was excited and joyful. Their enthusiasms for research and their attitudes to life were deeply inscribed in my mind with my respect. Also I thank the external examiner, Dr. James Aitken (Oriental Studies, Cambridge University) and the internal examiner, Professor Francis Watson. They gave me invaluable suggestions and criticisms. And during Professor Stuckenbruck’s two sabbatical leaves, Professor John Barclay and Dr. William Telford also kindly supervised my research for each term. I am extremely grateful for their invaluable comments and suggestions. Thankfully, Drs. Lionel & Wendy North read the first draft of chapter 6 and Dr. Ted Kaizer (Dept. of Classics and Ancient History, Durham University) read part I. They helped me greatly with insightful suggestions and encouragement. And I thank Professor Fergus Millar (Dept. of Oriental Studies, Oxford University) thankfully gave me invaluable suggestions. Also I wish to express my

viii

Acknowledgements

deep gratitude to my friend, Dr. Benjamin Wold (Trinity College, Dublin) who read Part I. I appreciate their very insightful and practical comments. Lastly, it was my privilege to meet and discuss with Dunelm living legends of New Testament studies such as Professor Kingsley Barrett and Professor C. E. B. Cranfield and to use an enormous amount of information from invaluable Durham Libraries. I am especially grateful to the editorial board of the BZNW series, James D. G. Dunn, Carl Holladay, Hermann Lichtenberger, Jens Schröter, Gregory Sterling, and Michael Wolter. And I also thank Dr. Albrecht Döhnert, Carsten Burfeind, and Sabina Dabrowski at Walter de Gruyter for their assistance and endurance to print my work. My manuscript was accepted in July 2009, my busy teaching schedule in Korea dragged me to publish it for many years. I also thank my teachers of New Testament studies and Linguistics in Korea. Professors, Hoon-Taek Chung, Seyoon Kim (now, Fuller Theological Seminary), Sung-Soo Kwon, and Han-Soo Lee (Chongshin University) and Chang-Ki Hwang (Kosin University). They led me to New Testament studies and opened my eyes to study the New Testament. I am also obliged to Professors, Yoon-Han Kim, Chung-Min Lee, Hyun-Bok Lee, Yang-Soo Moon, and Seung-Ho Nam who taught me Linguistics (Seoul National University). And I would like to express my gratitude with respect to many mentors, especially Professors Bong-Ho Son, Sang-Bok Kim, and Jung-Joo Kim for their encouragements and supports. I thank my friends, Professor Hyun-Woo Shin and Byeng-Ki Kim/Soo-Jean Ahn for financial support and encouragement. Thankfully, I had the privileges to present my papers; Society of Biblical Literature Annual Meeting in Philadelphia in 2005 entitled, “Syntactic Changes of Legon, Hina and Hoti” in the section of Biblical Greek Language and Linguistics (chapter 6), Society of Biblical Literature International Meeting in Edinburgh in 2006 entitled, “The Literary Functions of the Aramaic Words in the Markan Narrative” in the section of Synoptic Gospels (§8.3), British New Testament Conference in Sheffield University in 2006 entitled, “Bilingualism of the Hebrews and the Hellenists in the Jerusalem Church” in the section of Short Papers (chapter 5), and Society of Biblical Literature Annual Meeting in New Orleans, LA in 2009 entitled, “Semitic Spellings as Literary Device in the Lukan Narrative: in the Cases of Nazareth and Jerusalem” in the section of Synoptic Gospels (§8.4.3). I appreciate the audiences’ responses and encouragements. I have also enjoyed the opportunities to attend Durham-Tübingen Symposium in Durham in 2004 and the Studiorum Novum Testamentum Societas Meeting in Durham in 2002.

Acknowledgements

ix

I received scholarships by God’s grace. Without their financial support and prayer I could not complete this work: Ustinov College Travel Fund, Arts and Humanities Research Council (AHRC) Award (three times), The Whitecourt Charitable Trust (twice), The Mylne Trust, Doorae Institute Abroad Scholarship (Korea), Westminster Theological Seminary Scholarship (Korea), ISF, Lightfoot Scholarship of Department of Theology (Durham University), Sangjoon Scholarship Foundation in Seoul. Also I am indebted to many individuals in UK, USA, and Korea. Especially, the members of ISF (Korea), Durham Korean Church (UK), Korean Presbyterian Church at Purdue (USA), and Purdue Korean Bible Study (USA) gave me in countless ways through their love. My greatest debt, above all, is to my family. My wife, Soo-Jung Chae deserves my sincere admiration. My better half has supported me in every possible way with patience. This book is the product of her devotion. I thank my lovely daughters (Un and Sol) since their existence was always my joy and my strength. I am deeply grateful to my parents in-law, Hee-Joong Chae and Jung-Ae Byun for their love, support, and trust. Also I wish to express my gratitude to my grandmother (Sun Won) in memory and my grandmother in-law (Jeung-Chool Jung). And this book is dedicated to my parents (Seong-Hwa Lee and GeongAe Noh) who sacrificed themselves for me. Finally, I give my thanks to the Lord who is the ground, purpose, and hope of my existence. Only He is my all. evx auvtou/ kai. diV auvtou/ kai. eivj auvto.n ta. pa,nta\ auvtw/| h` do,xa eivj tou.j aivw/naj( avmh,n) Yong-In, Korea Dec. 2011 Sang-Il Lee.

Table of Contents Acknowledgements ................................................................................ vii Abbreviations .......................................................................................... xvii 1 1.1 1.2 1.2.1 1.2.2 1.2.3 1.3 1.3.1 1.3.1.1 1.3.1.2 1.3.2 1.3.2.1 1.3.2.2 1.3.3 1.4 1.4.1 1.4.1.1 1.4.1.2 1.4.1.3 1.4.1.4 1.4.2 1.4.2.1 1.4.2.2 1.4.2.3 1.4.2.4 1.4.2.5 1.5 1.5.1 1.5.2

The Directionality of the Transmission of the Jesus and Gospel Traditions: A History of Research ........ Herman Gunkel ...................................................................... Sitz im Leben Unidirectionality from Judaeo-Palestinian into Hellenistic Tradition ...................................................... Martin Dibelius ....................................................................... Rudolf Bultmann .................................................................... Gerd Theissen ......................................................................... Modal Unidirectionality from Oral into Written Tradition .................................................................... Scandinavian School .............................................................. Harald Riesenfeld ................................................................... Birger Gerhardsson ................................................................ Orality Theory ........................................................................ Werner Kelber ......................................................................... Joanna Dewey ......................................................................... The Earliest Written Tradition: Earle Ellis .......................... Linguistic Unidirectionality from Aramaic into Greek Tradition ...................................................................... Aramaic Language as a Criterion ......................................... Charles C. Torrey ................................................................... Matthew Black ........................................................................ Joachim Jeremias .................................................................... Joseph Fitzmyer ...................................................................... Greek Language as a Criterion ............................................. Alexander Roberts .................................................................. Aubrey Argyle ........................................................................ Nigel Turner ............................................................................ Robert Gundry ........................................................................ Stanley Porter .......................................................................... Interdirectionality Hypothesis .............................................. Sitz im Leben Interdirectionality: Martin Hengel .............. Modal Interdirectionality ......................................................

1 2 6 6 12 16 20 22 22 23 25 27 31 33 36 38 39 43 45 47 50 51 53 54 54 55 58 58 61

xii 1.5.2.1 1.5.2.2 1.5.2.3 1.5.2.4 1.5.2.5 1.5.3 1.5.4 Part I: 2 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.1.7 2.1.8 2.1.9 2.1.10 2.1.11 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.3 2.3.1 2.3.2 2.3.3 2.3.4 3 3.1 3.1.1 3.1.2 3.1.3

Table of Contents

Form Criticism: Helmut Koester .......................................... Textual Criticism: David Parker ........................................... Scandinavian School: Samuel Byrskog ................................ Liturgical Use and Canon Formation: Harry Gamble ....... Notebook Theory: Graham Stanton and Richard Bauckham ................................................................. Linguistic Interdirectionality: E. P. Sanders ....................... Interdirectionality Hypotheses and Gospel Studies .......... Bilingualism of First-Century Palestine and the Roman Near East ...................................................... Bilingualism and Diglossia ................................................... Definitions ............................................................................... Bilingualism ............................................................................ Bilingual Community ............................................................ Bilingualism and Multilingualism ....................................... Early Bilingualism vs. Late Bilingualism ............................ Primary Bilingualism vs. Acquired Bilingualism .............. Oral Bilingualism vs. Literate Bilingualism ........................ Stages of Bilingualism ............................................................ Balanced Bilingualism vs. Dominant Bilingualism ........... Matrix Language vs. Embedded Language ........................ Substratum, Adstratum and Superstratum ........................ Ancestor, Daughter and Sister Language ........................... Diglossia in First Century Palestine? ................................... Charles Ferguson .................................................................... Joshua Fishman ....................................................................... Pinchas Lapide ........................................................................ Chaim Rabin ........................................................................... Bernard Spolsky ..................................................................... Jonathan Watt ......................................................................... Bilingualism and Diglossia Models ..................................... Characteristics of the Four Languages in First-Century Palestine .......................................................... Language Preference Theory ................................................ Maximalism and Minimalism ............................................... Biliteracy of First-Century Palestine .................................... Bilingualism of Jews in First-Century Palestine ................. Regional Bilingualism ............................................................ Inscriptional Evidence ........................................................... Papyrological Evidence ......................................................... Population Geographical Evidence .....................................

61 63 64 66 67 70 72 75 77 79 79 81 81 81 82 83 84 84 84 85 85 86 87 88 88 90 92 94 96 96 97 99 103 105 105 106 108 110

Table of Contents

3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.2.5 3.2.6 3.3 4 4.1 4.1.1 4.1.2 4.1.3 4.1.4 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.3 4.3.1 4.3.2 4.3.3 4.4 5 5.1 5.2 5.2.1 5.2.2 5.3 5.3.1 5.3.2 5.3.2.1 5.3.2.2 5.3.2.3 5.3.2.4 5.3.3

xiii

Personal Bilingualism ............................................................ Military Invasion and Colonization ..................................... Migration or Inflow ................................................................ Ethnic Awareness ................................................................... Bilingual Parents or Region .................................................. Formal Education ................................................................... Occupation .............................................................................. Bilingualism of First-Century Palestine and Interdirectionality ...........................................................

112 113 114 114 115 120 130

Bilingualism of Jews in the First-Century Diaspora .......... Grounds for Bilingualism in the Jewish Diaspora ............. Bilingualism of the Roman Empire ...................................... Successive Immigration ......................................................... Periodic Connection with Jerusalem ................................... Learning Aramaic in Jerusalem ............................................ Alexandria ............................................................................... Bilingualism of Ptolemaic and Early Roman Alexandria ....................................................... Archaeological Evidence ....................................................... Successive Immigration ......................................................... Periodic Connection with Jerusalem ................................... Antioch .................................................................................... Bilingualism of Roman Antioch ........................................... Bilingualism of the Antiochene Jews ................................... Bilingualism of the Antiochene Christians ......................... Bilingualism of the Jewish Diaspora and Interdirectionality ...........................................................

135 137 138 141 145 147 150

The Bilingualism of the Earliest Christian Church in Jerusalem ............................................................................. The State of Affairs ................................................................. Bilingualism of the Hebrews and Hellenists ...................... Hebrews: Aramaic-Matrix Christians .................................. Hellenists: Greek-Matrix Christians .................................... The Bilingual Seven ................................................................ Onomastica .............................................................................. Geographical Evidence .......................................................... Caesarea Maritima ................................................................. Samaria .................................................................................... Ethiopia .................................................................................... Antioch .................................................................................... Performance of Their Duty ...................................................

133

152 156 160 162 162 163 167 170 172 175 176 182 184 189 197 197 198 199 200 201 203 204

xiv

Table of Contents

5.4

Bilingualism of the Jerusalem Church and the Interdirectionality of the Jesus Tradition ............................ 209

Part II:

Interdirectional Transmission of the Jesus and Gospel Traditions in Bilingual Contexts at the Levels of Syntax, Phonology, and Semantics .................................................... 213

6 6.1

Syntax ....................................................................................... Contact-Induced Syntactic Change Hypothesis in New Testament Scholarship ............................................. 6.1.1 Adolf Deissmann .................................................................... 6.1.2 Linguistic Relatedness between Syntactic Changes .......... 6.1.3 Bilingual Jewish Greek Theory ............................................. 6.1.4 Hellenistic Greek Theory ....................................................... 6.1.5 Semitisms and Septuagintalisms .......................................... 6.2 Internal-Induced Syntactic Change Hypothesis ................ 6.2.1 Comparisons between Cognitive Linguistics and Other Linguistic Theories .............................................. 6.2.1.1 Theoretical and Empirical ..................................................... 6.2.1.2 Diachronic and Synchronic ................................................... 6.2.1.3 Langue and Parole ................................................................. 6.2.1.4 Significance of Semantics ...................................................... 6.2.2 Some Basic Concepts .............................................................. 6.2.2.1 Grammaticalization ................................................................ 6.2.2.2 Unidirectionality Hypothesis ............................................... 6.2.2.3 Five Principles of Grammaticalization ................................ 6.2.2.4 Metaphor and Metonymy ..................................................... 6.2.2.5 Abstractness ............................................................................ 6.2.2.6 Grammatical Polysemy ......................................................... 6.2.3 Grammaticalization and Language Contact ....................... 6.3 Syntactic Change of New Testament Greek as Grammatical Polysemy ......................................................... 6.3.1 Verbs ........................................................................................ 6.3.1.1 SAY Verb: le,gw ...................................................................... 6.3.1.2 Movement Verbs: e;rcomai and poreu,omai ............................. 6.3.1.3 Posture Verbs .......................................................................... 6.3.1.3.1 sth,kw and i[sthmi ...................................................................... 6.3.1.3.2 avni,sthmi and evgei,rw ................................................................. 6.3.1.3.3 kaqi,zw and ka,qhmai .................................................................. 6.3.2 Conjunctions ........................................................................... 6.3.2.1 {Ina ............................................................................................ 6.3.2.2 [Oti ...........................................................................................

223 225 226 230 231 234 237 240 241 241 242 242 244 244 244 245 246 247 248 249 249 250 251 252 255 256 257 259 260 261 261 270

Table of Contents

6.3.3 6.4 7 7.1 7.1.1 7.1.2 7.1.3 7.1.4 7.2 7.2.1 7.2.2 7.2.2.1 7.2.2.2 7.2.2.3 7.2.2.4 7.2.3 7.3 7.3.1 7.3.2 7.3.2.1 7.3.2.2 7.3.2.3 7.3.2.4 7.3.2.5 7.3.3 7.4 7.4.1 7.4.1.1 7.4.1.2 7.4.1.3 7.4.2 7.4.2.1 7.4.2.2 7.4.3 7.4.3.1 7.4.3.2 7.5

xv

Adverbs: Matthean to,te, Marcan euvqu,j, and Johannine ou=n ................................................................... 274 Syntactic Change as Grammatical Polysemy and Interdirectionality ........................................................... 278 Phonology ............................................................................... Linguistic Factors of Transliterated Variants in Bilingual Contexts .............................................................. Different Phonetic System ..................................................... Representation ........................................................................ Phonetic Change ..................................................................... Dialects ..................................................................................... Three Views of Variant Spellings in Transliteration ......... Orthographical View ............................................................. Variational View ..................................................................... Ephraim Speiser ...................................................................... James Barr ................................................................................ Alan Millard ............................................................................ Some Scholars of New Testament Greek ............................ Bilingual View ........................................................................ Variant Spellings as Transliterated Allolexemes in Bilingual Contexts .............................................................. Transliterated Allolexemes ................................................... Consonants .............................................................................. Plosives and Fricatives ........................................................... Sibilants .................................................................................... Liquids ..................................................................................... Gutturals .................................................................................. Additions of Consonants ....................................................... Vowels ..................................................................................... Variant Spellings as Transliterated Allolexemes in the Four Gospels and Acts ................................................ Local Proper Nouns ............................................................... VIerousalh,m and ~Ieroso,luma ..................................................... Nazara, and Nazare,q ................................................................ Gennhsa,r and Gennhsare,t ........................................................ Personal Proper Nouns .......................................................... Si,mwn/Sumew,n and Pe,troj/Khfa/j ............................................ VElisa,bh and VElisa,bet ............................................................. Other Transliterated Words .................................................. The Cry from the Cross ......................................................... Messi,aj/Cristo,j ....................................................................... Variant Spellings as Allolexemes and Interdirectionality

281 283 283 288 292 294 295 295 298 298 299 299 300 301 302 302 304 305 305 306 306 307 308 310 310 311 311 314 317 317 319 320 320 325 328

xvi 8 8.1 8.1.1 8.1.2 8.1.3 8.1.4 8.1.4.1 8.1.4.2 8.1.4.3 8.1.4.4 8.2 8.3 8.3.1 8.3.2 8.4 8.4.1 8.4.2 8.4.2.1 8.4.2.2 8.4.3 8.4.3.1 8.4.3.2 8.5 9 9.1 9.2 9.3

Table of Contents

Semantics ................................................................................. Aramaic Embedded Words as Codeswitching .................. Codeswitching, Interference, and Borrowing in a Bilingual Society .............................................................. Codeswitching in New Testament Scholarship ................. Types of Codeswitching ........................................................ Pragmatic Functions of Codeswitching .............................. Vividness ................................................................................. Emphasis ................................................................................. Politeness ................................................................................. Solidarity ................................................................................. avmh,n Found in All Four Gospels ........................................... Codeswitchings in the Gospel of Mark ............................... A v bba o` path,r ........................................................................... Taliqa koum and effaqa in the Miracle Stories .................... Codeswitchings in the Gospel of Luke and Acts ............... Language Change ................................................................... Personal Names ...................................................................... Simw,n/Sumew,n and Pe,troj/Khfa/j ............................................ Sau/loj/Pau/loj and Saou,l ........................................................ Names of Places ...................................................................... VIerousalh,m and ~Ieroso,luma ..................................................... Nazara, and Nazare,q ................................................................ Embedded Aramaic Words as Codeswitchings and Interdirectionality ...........................................................

331 334 335 341 346 346 347 347 347 348 349 360 362 370 376 377 378 380 383 385 385 387 391

Summary and Suggestions for Further Study .................... Summary of Results ............................................................... The Gospel Tradition as Jesus Tradition ............................. The Relationship between Historical Criticism and Literary Criticism ............................................................ The Synoptic Problem ............................................................ The Historical Jesus ................................................................ Textual Criticism .................................................................... Jewish Christianity and Gentile Christianity ...................... Bilingual Christology .............................................................

410 410 414 416 417 417

Bibliography ............................................................................................ Index of Ancient Sources ........................................................................ Index of Languages and Place Names .................................................. Index of Modern Authors ...................................................................... Index of Subjects ......................................................................................

419 487 503 505 515

9.4 9.5 9.6 9.7 9.8

395 395 409

Abbreviations The following abbreviations are in addition to those found in Patrick H. Alexander et al., eds., The SBL Handbook of Style: For Ancient Near Eastern, Biblical, and Early Christian Studies (Peabody, Mass.: Hendrickson, 1999). Bap. Inst.

Beg.

CESG

CIJ

ESE

Fraser GLAJJ

GLICM

IGCH

Harkins, Paul W. (translated and annotated) St. John Chrysostom: Baptismal Instructions, Ancient Christian Writers (Westminster, Maryland: Newman Press, 1963) Foakes-Jackson, F. J. and Kirsopp Lake (eds.) The Beginnings of Christianity Part 1: The Acts of the Apostles 5 Vols. (London: Macmillan, 1920–1933). Kiraz, George Anton Comparative Edition of the Syriac Gospels: Aligning the Sinaiticus, Curetonianus, Peshitta and Harklean Versions 4 Vols. (NTTS 21/1-4; Leiden: Brill, 1996–2004) Frey, J.-B. Corpus Inscriptionum Judaicarum 2 Vols. (Roma: Pontificio Istituto di Archeologia Cristiana, 1936–1952) Lidzbarski, Mark Ephemeris für Semitische Epigraphik 3 Vols. (Giessen: Alfred Töpelmann, 1900–1915) Fraser, P. M. Ptolemaic Alexandria 3 Vols. (Oxford: Clarendon, 1972) Stern, Menahem Greek and Latin Authors on Jews and Judaism 3 vols. (Jerusalem: Publications of Israel Academy of Sciences and Humanities, 1974– 1984) Lehmann, Clayton Miles and Kenneth G. Holum The Greek and Latin Inscriptions of Caesarea Maritima (ASOR 5; Boston, MA: ASOR, 2000) Thompson, Margaret, Otto Mørkholm, and Colin Kraay (eds.) An Inventory of Greek Coin Hoards (New York: American Numismatic Society, 1973)

xviii JIGRE

K-G

Liddell-Scott LJNLA

P. Col.

Schürer

Vaganay-Amphoux

VGT

Wadd.

W-H

Zeraim

Abbreviations

Horbury, William and David Noy Jewish Inscriptions of Graeco-Roman Egypt: With an Index of the Jewish Inscriptions of Egypt and Cyrenaica (Cambridge: Cambridge University Press, 1992) Kühner, Raphael and Bernhard Gerth Ausführliche Grammatik der Griechischen Sprache 2 Vols. (Hannover: Verlag Hahnsche Buchhandlung, 1976) Liddell, H.G. and R. Scott Greek-English Lexicon (Oxford: Clarendon, 1940) Ilan, Tal Lexicon of Jewish Names in Late Antiquity Part I Palestine 330 BCE – 200 CE (TSAJ 91; Tübingen: Mohr Siebeck, 2002) Westermann, W. L., E. S. Hasenoehrl, C. W. Keyes, and H. Liebesny (ed.) Zenon Papri: Business Papers of the Third Century B.C. Dealing with Palestine and Egypt 2 Vols. (NY: Columbia University Press, 1934-1940) Schürer, Emil, Geza Vermes, Fergus Millar, and Matthew Black The History of the Jewish People in the Age of Jesus Christ 3 Vols. (Edinburgh: T&T Clark, 1973–1987) Vaganay, Leon and Christian-Bernard Amphoux An Introduction to New Testament Textual Criticism (Cambridge: Cambridge University Press, 1991) Moulton, James Hope and George Milligan The Vocabulary of the Greek Testament Illustrated from the Papyri and Other Non-Literary Sources (Grand Rapids, MI: Eerdmans, 1959) Waddington, W. H and Philippe Le Bas Inscriptions grecques et latines receuillies en Grèce et en Asie Mineure (Paris: Firmin-Didot frères, 1870) Westcott, Brooke Foss and Fenton John Anthony Hort The Greek New Testament with Comparative Apparatus Showing Variations from the Nestle-Aland and Robinson-Pierpont Editions with Greek Dictionary Revised and Expanded from a Pocket Lexicon to the Greek New Testament by Alexander Souter (Peabody, MA: Hendrickson, 2007 [1885]) Epstein, Isidore The Babylonian Talmud, Seder Zeraim (London: Soncino Press, 1948)

1. The Directionality of the Transmission of the Jesus and Gospel Traditions: A History of Research When biblical scholars investigate major issues of gospel studies, many of them assume three unidirectionality1 hypotheses of transmission of Jesus and gospel traditions. According to traditional wisdom, Jesus and gospel traditions were unidirectionally transmitted from JudaeoPalestinian tradition into Hellenistic tradition (i.e. geographical unidirectionality), from oral tradition into written tradition (i.e. modal unidirectionality), and from Aramaic tradition into Greek tradition (i.e. linguistic unidirectionality), and never vice versa.2 These three might be called the unidirectionality hypotheses of the transmission of the Jesus and gospel traditions.3 Scholars have supposed that the former three 1 2

3

The terms, “unidirectional” and “unidirectionality” are drawn from one of five principles of grammaticalization (§6.2.2.2). Kilpatrick’s remark, for instance, demonstrates this view. He (1970:170) mentions that “it is generally agreed that the tradition about Jesus made two migrations in this order: first there was the migration from Aramaic to Greek and, second, the migration from oral to written form“ [emphasis added]. Also, Cadbury (1958:62) writes that the Gospel of Luke “or at least much of it, has passed through two other processestransfer from oral to written form, translation from Aramaic to Greek…” Further, he (1958:27-8) suggests, “The gospel was transferred not only from Aramaic to Greek – but from Palestine to Europe, and from Jews to Gentiles.” Criticizing the “transfer” used by Cadbury, Sanders (1969:196) convincingly argues that “No doubt Cadbury actually does not think that the gospel left Palestine when it entered Europe, but that notion springs up here. It seems to lie hidden under the thinking of many scholars.” For Sanders’s detailed criticism, see §1.5.3. In addition, there is the unidirectionality hypothesis of the length of the Jesus and gospel traditions. It is usually assumed that shorter tradition developed into longer tradition. In other words, the shorter, the earlier. Bultmann (1961:41-2) considers “increasing” as a law governing the transmission of tradition. He explains, “Whenever narratives pass from mouth to mouth, the central point of the narrative and general structure are well preserved; but in the incidental details changes take place, for imagination paints such details with increasing distinctness.” Jeremias (1967:8990; cf. 1971:195; cf. §1.4.1.3), for instance, supposes that “according to all that we know about the tendency of liturgical texts to conform to certain laws in their transmission, in a case where the shorter version is contained in the longer one, the shorter text is to be regarded as original.” On the other hand, Vincent Taylor

2

The Directionality of the Transmission of the Jesus and Gospel Traditions

traditions (i.e. Judaeo-Palestinian, oral, and Aramaic) are earlier than the three later traditions (i.e. Hellenistic, written, and Greek). The geographical, the modal, and the linguistic unidirectionality hypotheses are based on the chronological unidirectionality hypothesis: the earlier, the more original. Contrary to this consensus, this study will expose the deficiencies of these presuppositions and, thus, rethink the chief principles of Gospel criticism. When the bilingualism of first-century Palestine and the Roman Near East is considered seriously, it must be granted that the Jesus and gospel traditions were interdirectionally transmitted. In other words, there was a complex and interactive relationship between JudaeoPalestinian and Hellenistic tradition, between oral and written tradition, and between Aramaic and Greek tradition. In this chapter, three hypotheses of unidirectionality will be examined. First of all, Gunkel, as the progenitor of unidirectionality, will be investigated (§1.1). I will then discuss the three hypotheses of unidirectionality: Sitz im Leben unidirectionality (§1.2), modal unidirectionality (§1.3), and linguistic unidirectionality (§1.4). Finally, I will propose an alternative to these by introducing a hypothesis of interdirectionality (§1.5).

1.1 Herman Gunkel Herman Gunkel was the first to introduce form criticism and the concepts of oral tradition and oral transmission in a full-dressed debate. Gunkel was the foremost scholar of the religionsgeschichtliche Schule. Although his major publications in his later years centered more on Old Testament Studies, 4 his methodology was highly influential in New Testament scholarship as well.

4

(1935:124) wrote, “The experiments show that the tendency of oral transmission is definitely in the direction of abbreviation. Additions are certainly made in all good faith through misunderstandings and efforts to picture the course of events, but almost always the stories become shorter and more conventional [emphasis in the original]. However, Charlesworth (1994:2-3; here 2) persuasively argues, “My study of the transmission of early Jewish writings … reveals no rule that texts, including liturgical documents, are expanded through transmission“ [emphasis in the original]. For a detailed criticism of Bultmann’s law of increasing distinctness, see Keylock 1975:193-210. Also, for criticisms of the unidirectionality hypothesis of length, see Farmer 1964:230; Sanders 1969:246; in relation to the Gospel of Thomas, see McArthur 1960:67; Frye 1978:263; Meier 1991:132. Known as an Old Testament scholar, Gunkel started his academic career as a New Testament scholar and continued to deal with New Testament literature throughout

Herman Gunkel

3

Above all, Gunkel detected that there was something “original” behind a text itself. He thought of the original as the oral tradition of the biblical narratives in the Sitz im Leben of ancient Israel. This attempt to locate the original diverted New Testament scholars’ concern from the texts themselves to something behind the texts. Form critics sought to find tradition that was “closer” to this oral tradition behind the Synoptic Gospels themselves (§1.2). Bultmann (1934:1) clearly states that the purpose of form criticism is to investigate the development of the oral tradition before the written documents. The Scandinavian school (which stood in opposition to form critics; cf. Sanders 1985:14) was thoroughly shaped by Gunkel’s method concerning oral transmission before the written tradition (§1.3.1; §1.5.2.3). As a consequence, the chief teachings of Gunkel are still reflected in contemporary New Testament publications regarding oral tradition and oral transmission. Despite his enormous contribution, many scholars discussing oral transmission and oral tradition neglect to mention his publications directly.5 Ironically, form critics soon dismissed the significance of the oral tradition when dealing with the transmission of the Jesus and gospel traditions. 6 Although New Testament scholars have disregarded Gunkel for approximately a hundred years, it is Gunkel who should occupy a prominent place when discussing the idea of something

5

6

his life. His contribution to New Testament studies is discussed in detail in Kümmel 1972:217-8, 230, 248-53, 257-9. Despite Gunkel’s tremendous contribution to oral tradition in relation to New Testament scholarship, New Testament scholars who deal with oral tradition have by no means held a debate on Gunkel’s works in their major publications on the oral tradition: Bryan 1993; Byrskog 2000; Millard 2000; Dewey 1989; 1994a; 2001; 2008; Dunn 2003; Fowler 1991; Henaut 1993; Horsley 1999; 2001; Kelber 1983; Riesenfeld 1970; Taylor 1956. Kelber notes Gunkel twice (1983:96, 107) in order to reinforce his argument, but does not pay careful attention to Gunkel’s view of oral tradition. Millard (2000:186, 197) and Horsley (2001:151) simply mention the point that form critics were influenced by Gunkel. Henaut (1993:28-9) also summarizes Gunkel’s theory briefly by citing McKnight (1969). It is regrettable that the rest of the above mentioned scholars have not been more attentive to Gunkel’s publications. Also, Gunkel’s proposition that “retelling generates variants” is repeated in modern scholars’ arguments; David Parker (§1.5.2.2), Samuel Byrskog (§1.5.2.3), and Harry Gamble (§1.5.2.4), however, do not cite Gunkel. However, credit is due to Culley and Güttgemanns. Culley (1986) summarizes Gunkel’s arguments in relation to the formgeschichtliche Schule. When Güttgemanns (1979:235-57) criticizes the arguments of the formgeschichtliche Schule in detail, he convincingly rebuts Gunkel’s arguments. Bultmann (1968:6) comments that “it is at this point a matter of indifference whether the tradition were oral or written.” Also Dibelius (1934:10) claimed that “it is of little value to distinguish these processes simply by the index word ‘oral tradition’.”

4

The Directionality of the Transmission of the Jesus and Gospel Traditions

original behind the texts. With this in mind, my history of research will begin with him. In the late nineteenth century, source critics like Wellhausen claimed that the Pentateuch developed from previous documents. JEDP became shorthand for the four major sources that were supposedly combined and edited by later writers.7 Gunkel, however, argued for a precompositional stage before the documents were fixed. In the introduction to his ground-breaking commentary on the book of Genesis (1997:vii-lxxxvi), he explained the major characteristics of the oral and written traditions of the Old Testament. Above all, Gunkel drew a sharp contrast between oral and written tradition. According to him (1997:viii), narratives of Genesis preserved a form called the “legend,” which was to be distinguished from “history” because the legend was usually derived from an oral tradition, whereas history is a written form. He (1997:viii) supposes that oral tradition was not a satisfactory vehicle for history since “oral tradition cannot remain pure over time.” The oral tradition was ultimately produced within the Sitz im Leben of the community of ancient Israel (1997:xxvi-xxvii). In this respect, oral tradition produced variants. He (1997:lvi) pointed out that oral retelling often generated narrative variations. When tradition was orally circulated in the community of Israel through various times, it would have changed in each retelling. He (1997:lvi) asserted that “the legend exists in variants, as do all oral traditions.”8 However, he (1997:lvi) argued that even though the oral tradition caused minor variations, the variations often still agreed in overall outline, even sometimes in wording. Regarding the proposed various solutions to the synoptic problem, he (1921:208) insightfully commented that scholars have paid more attention to the writings that fixed the tradition than to the oral tradition so that the synoptic problem became complicated in modern biblical exegesis (see n. 25). What is more, Gunkel was concerned about the relationship between oral and written tradition, which has had a significant influence upon gospel studies. He claimed that the oral tradition and the written tradition did not interact with each other. He (1997:lxix) presumed that during oral transmission, many collectors already had begun the very long process of the documentation of the oral legends, when they 7

8

Despite his rejection of the methodological presuppositions of the religionsgeschichtliche Schule, Wellhausen (1905:3, 43, 51-2, 94, 97-8, 107, 113-5) consented to Gunkel’s position; cf. Kümmel 1972:282, 445. Ironically, oral tradition was more original but more fluid, which leads to Bultmann’s skepticism of history including the historical Jesus (§1.2.2).

Herman Gunkel

5

feared that the oral tradition would disappear. When the writers as collectors assembled previously-existing oral traditions that were isolated and independent, they created only connecting pieces linking the oral traditions together. Furthermore, Gunkel (1997:lxix) assumed that the fixed written form of the oral tradition would have terminated “the other surviving remnants of oral tradition.” Thus, his assertions diverted scholarly attention away from literary sources and towards the oral forms behind the fixed written forms. Gunkel’s views of the oral tradition warrant several challenges. First of all, Gunkel classifies several oral genres within the Old Testament narrative and thus sets the oral traditions against history, as if there is a necessary dichotomy between them. However, why should the oral traditions not be regarded as “historical”? As to the “historicity” of oral tradition, Byrskog (§1.5.2.3) persuasively argues that “oral history” should be understood as oral stories which interact with history in first-century Palestine as well as in ancient times. Another of Gunkel’s weak points lies in his view of the relationship between the oral and the written tradition. He insisted that the oral traditions died out with the written documentation of the biblical narratives. In his view, textual tradition was intended to supplant oral tradition, not to supplement it. The relation between orality and textuality is not that of a partnership but rather of a competition. However, according to recent studies in orality and literacy in ancient times, Greek textuality played a role supportive of orality (cf. §1.5.2). Furthermore, as Gunkel assumed, once oral tradition is changed into written tradition, it is not necessarily so that the written tradition still remains in the written form without oral re-performance. Rather, it is natural that the written tradition is changed into another oral tradition through oral performance again. If we assume that the primary method of human communication in ancient times was orality rather than literacy, it should be thought that the directionality of the transmission of the biblical narratives is not “unidirectional” from oral to written, but “interdirectional” between the oral and the written tradition. Accordingly, there is a serious problem with Gunkel’s two assumptions, that is, with his concept of the oral tradition and his view of the relationship between the oral and the written traditions. This problem lies in the fact that the directionality of the transmission of the biblical narratives was seen as “unidirectional.” In other words, he saw it as moving only from oral tradition to written tradition and never vice versa. Consequently, the unidirectionality (that is, the unilinear and

6

The Directionality of the Transmission of the Jesus and Gospel Traditions

teleological view9) of the transmission causes both Old Testament and New Testament scholars to focus on the written traditions as their final fixed forms. In this sense, Gunkel is the progenitor of the unidirectionality hypothesis of the transmission of the biblical narratives. These chronic obstacles that Gunkel could not iron out in his publications left his successors to be heirs to the same problems repeatedly in their publications.

1.2 Sitz im Leben Unidirectionality from Judaeo-Palestinian into Hellenistic Tradition Form and redaction critics have emphasized Sitz im Leben unidirectionality. Furthermore, they presuppose two other kinds of unidirectionalities: modal unidirectionality and linguistic unidirectionality, although they have tended to assume that Sitz im Leben unidirectionality is more important than the other two. Here, three representative scholars will be discussed: Martin Dibelius (§1.2.1), Rudolf Bultmann (§1.2.2), and more recently, Gerd Theissen (§1.2.3).

1.2.1 Martin Dibelius Borrowing the overall conception of oral tradition from Gunkel, 10 Dibelius applied form criticism to analyzing “purer forms” behind the gospel tradition. In his From Tradition to Gospel he used three key dichotomies to sift single discrete “traditions” from the “Gospel.” He set oral tradition, Aramaic tradition, and eyewitness testimony in opposition to written tradition, Greek tradition, and eyewitness preaching. In his view, the first three traditions are transmitted to the latter three counterparts in a unidirectional way. Consequently, the first three traditions are more original. When dealing with sayings by Jesus and stories about Jesus, Dibelius argued that the first three manifestations represented purer forms than their counterparts. Dibelius (1936:28) assumed that single discrete traditions by and about Jesus were circulated in oral form before the Synoptic Gospels 9

10

Kelber (1983:8) considers Bultmann’s view “teleological.” Also criticizing Bultmann’s argument, G‫ܗ‬ttgemanns uses “linear.” He (1979:100; cf. Kelber 1983:8) says, “The evangelists are only the last link in the chain of a linear development over a longer period of time.” Dibelius (1934:3-7) admitted that he borrowed basic conceptions such as “smallest details,” “form criticism,” and “Sitz im Leben” from Gunkel.

Sitz im Leben

7

were written down in the fixed form unidirectionally. His argument of unidirectional transmission started with the illiteracy of Jesus’ disciples and eyewitnesses. (i) Because of their social status as indicated by their occupations, he supposed that Jesus’ disciples would not write down the Jesus tradition, but simply deliver it orally.11 (ii) He (1936:35) also considered that “the eye-witnesses of this life were not writers, and we cannot presume that they had any desire to hand down a history of what they had seen.” Interestingly, his assumption of their illiteracy might stem from his view of the Aramaic archaeological findings of his day. According to the evidence available to him, Dibelius (1934:32; cf. 1935:74) thought that the Aramaic tradition was not long in existence, since the Aramaic written tradition was not attested in the first or second century. The fact that the Nazarene Gospel for Aramaic speakers in Syria was translated from Greek implies that “there was no Aramaic source of the life of Jesus.” Dibelius’ presupposition about the Aramaic tradition might be well exposed when he mentions that “the formulation in Aramaic did not become very firm and was never ready for writing down.” Accordingly, it seems that his contemporary archaeological evidence caused him to make a sharper distinction between Aramaic and Greek traditions; Aramaic was associated with the oral tradition, whereas Greek was associated with the written tradition. In other words, he (1934:32-5; 1935:73-6) assumed that the oral tradition was circulated in Aramaic and then was written in Greek in a unidirectional way. As to the rendering of the Jesus tradition from Aramaic into Greek, he dealt with procedure, accuracy, and style in detail. As to procedure, interestingly, he (1949:25) observed that the translation is carried out not in “a single unified process” but in “a multiple process.” He (1934:34; 1949:25) illustrated two instances of the translation of Aramaic into Greek tradition. Jewish people who could understand stories about Jesus in Aramaic in the Jerusalem church most likely passed them on in Greek at their own home in Syria. Also, bilingual evangelists from Jerusalem might have delivered the stories in Greek to Phoenicia, Cyprus, and Antioch (Acts 11:19-20). Regarding the quality, he takes for granted that the translation was carried out with great fidelity to the testimony of living eyewitnesses. He (1949:32) wrote, “They, especially the personal disciples of Jesus, would have been in a position to correct any

11

He (1939:125) stated: “one must remind himself who these disciples were: for the most part Galilean fishermen, tax gatherers, perhaps also farm-labourers, unfamiliar with the literary practices of the world and probably not even accustomed to writing…” For discussion of bilingualism of disciples, see §3.2.6.

8

The Directionality of the Transmission of the Jesus and Gospel Traditions

egregious misrepresentation.” In terms of the style of the Greek traditions, he conceded that the Gospels have some solecisms. However, he (1934:34) did not regard Semitisms as defective translations except in incidental cases of wrong translations. Rather, he (1934:35; 1935:74-5) raised three possibilities for the Semitisms: unconscious dependence upon the diction of the Septuagint, conscious imitation of the style of the Old Testament, and the development of later Greek in general (see chapter 1 n.14; cf. §6.1.5). On this basis, he (1934:35) asserted that “it is impossible to solve the problem of the existence and nature of an Aramaic tradition by the evidence of Semitisms.” Accordingly, Dibelius (1934:32; 1935:73-4; 1936:46; 1949:31) concluded that the Aramaic tradition is of little theological significance. Also, he (1934:10) argued that “it is of little value to distinguish these processes simply by the index word ‘oral tradition.’” Then, he directed his attention from the linguistic and modal transmissions to the issue of Sitz im Leben, that is, the change from “testimony” to “preaching” by the eyewitnesses. Dibelius (1936:52) suggested that “unnamed and unknown” eyewitnesses played a significant role in the transmission of the Jesus tradition. He (1934:182-3; 1935:80; 1936:49-50) enumerated two instances of eyewitnesses from the Gospel of Mark. One is the youth who fled naked after leaving his linen cloth (Mk 14:51-52); the other is Simon of Cyrene who carried Jesus’ cross (Mk 15:21). He (1936:50; 1949:32) supposed that before the composition of the Gospel of Mark, a passion narrative originated among eyewitnesses who orally circulated this story in Christian communities when numerous witnesses of these events were still alive. According to Dibelius (1934:183-9; 1936:47-8), long before there was a written gospel, the earliest Christians began to read Old Testament passages into the passion of Jesus on the ground of their Easter faith, so that the story about his passion has been formed in the course of the “preaching” of public worship. He (1936:47-8) posited that these “preachings” of interpretations of passages in the Old Testament became more persuasive to the churches than the simple “testimonies” of the eyewitnesses, such as the women who were stunned or the disciples who were frightened away by their enemy. As a consequence of this, the testimonies of eyewitnesses have been interpreted in a different way in the process of the preaching by eyewitnesses. Dibelius took “preaching” seriously in his argument. On the basis of the Lukan prologue (Lk 1:1-2) he (1934:12; 1936:52) observed that, from the beginning, the unnamed and unknown eyewitnesses of Jesus were at the same time missionaries, preachers, and teachers who preached the gospel of the Christ. He (1934:14-5; cf. 1934:26-7) claimed that “preaching” should be regarded as “all possible forms of Christian

Sitz im Leben

9

propaganda” such as “mission preaching, preaching during worship, and catechumen instruction.” On this score, the preaching of eyewitnesses makes a significant contribution to the formation of the traditions of the Synoptic Gospels. This is because the evangelists were only collectors or redactors of the preaching.12 Accordingly, Dibelius suggested that the gospel tradition was formed in the process of the preaching in the Sitz im Leben of the earliest Christian communities. Dibelius, in large part, accepted Gunkel’s insight into the oral tradition and the oral transmission behind the fixed texts. In doing so, he made a great contribution to gospel studies but he also repeated Gunkel’s mistakes. In terms of the transmission of the synoptic tradition, he affirmed the unidirectional hypotheses of the Jesus traditions from oral into written tradition, from Aramaic into Greek tradition, and from Judaeo-Palestinian into Hellenistic tradition. Dibelius has some deficiencies due to his commitment to the unidirectional hypotheses. First of all, he failed to iron out Gunkel’s ambiguous understanding of “the oral tradition.” He assumed that the Jesus tradition was transmitted unidirectionally from oral into written tradition on the basis of apocalyptic considerations. 13 However, Dibelius deserved the same criticism that I leveled against Gunkel: Should we suppose that the written tradition fixed on the basis of the oral tradition also no longer continued to change into the oral tradition repeatedly? Furthermore, his concept of “preaching” was unclear. Stendahl (1968:14) persuasively argued that “the ambiguity of the word ‘Predigt’ has led to ambiguity in Dibelius’s train of argument.” In his view, missionaries, preachers, and teachers preached about Jesus and shaped the tradition accordingly. This preaching, such as missionary preaching, preaching during Christian worship, and catechetical instruction had to have been delivered in oral forms again. This implies that we cannot make a dis12

13

He (1936:52) mentioned, “Even before the Gospels in our sense arose, anonymous gatherers of tradition made beginnings of small collections which afterwards entered into the more inclusive works of the Evangelists and which, therefore, were no longer separately preserved. It was not the purpose of these collectors to write books, but to pass on tradition. Even the earliest Evangelists really intended nothing else. Thus the tradition of Jesus only gradually became literature, and this took place not on account of the literary ability of any author but by virtue of the significance of its content.” He (1934:9) remarked that “The company of unlettered people which expected the end of the world any day had neither the capacity nor the inclination for the production of books; we must not predicate a true literary activity in the Christian Church of the first two or three decades.” For a criticism of this apocalyptic view, see Ellis 1978:243; 1999b:53-4; Millard 2000:198-9. For more discussion of the earliest written tradition in parallel with oral tradition, see §1.3.3, §1.5.2.5.

10

The Directionality of the Transmission of the Jesus and Gospel Traditions

tinction between the oral tradition before the textual composition and the oral tradition after the textual composition. Rather, his concept of “preaching” lends support to the interdirectional hypothesis because, after the oral tradition was documented, the written materials most likely were used as “various kinds of preaching” in oral forms again. Second, as to the linguistic transmission from Aramaic to Greek,14 Dibelius (1934:34; 1949:25) argued that bilingual Christians took significant roles in linguistic transmission of the Jesus tradition in Phoenicia, Cyprus, and Syria (e.g. Antioch). There are two problems here. (i) He presupposed that Syrian Jews spoke Greek at home. However, although the linguistic milieu of Syria was bilingual Aramaic and Greek, it seems that Aramaic was still used dominantly in the home, whereas Greek was employed as a public language (§4.3). (ii) He (1949:25) argued that it was at Antioch in northern Syria that the Jesus tradition was translated from Aramaic into Greek because there were many bilingual members in Antioch. Also, he (1934:20) mentioned that bilingual Antioch “was obviously a much more significant starting-point for missionary activities than Jerusalem, which was relatively uni-lingual.” However, his argument would have been more persuasive if he had taken the bilingualism of first-century Palestine and the Roman Near East more seriously (cf. chapters 3 and 4). I will propose that the bilingual situations imply that the Jesus tradition was translated into Greek during Jesus’ ministry (chapter 3) as well as in the earliest Christian community in Jerusalem (chapter 5). Third, one of Dibelius’ faults is his assumption on the basis of his contemporary archaeological evidence that Aramaic is to be seen as the oral vehicle, whereas Greek is to be seen as the vehicle for literacy. On this account, he appears to claim that the transmission of the synoptic gospels was from oral into written and from Aramaic into Greek in a unidirectional fashion. In a series of his assumptions, he thought that since Jesus and his disciples used Aramaic, they were not literate, and the eyewitnesses who spoke Aramaic in Palestine would not even have thought about writing. Furthermore, because of his assumption that Q was a document, he postulated that it was highly probable that it was written in Greek rather than in Aramaic. Accordingly, his commitment to unidirectionality was in close connection with his presupposition of Aramaic as an exclusively oral language (for the same view of other scholars, see chapter 1 n.24). However, Dibelius neglected the pervasive

14

Dibelius interestingly suggested that some solecisms should be regarded as Septuagintalisms and that some solecisms could be due to later syntactic change of Greek language; for detailed discussion, see §6.1.5.

Sitz im Leben

11

biliteracy of first-century Palestine, as shown by Aramaic written materials like the Qumran documents (§2.3.4; cf. §1.5.2.5). Fitzmyer’s criticism (1979a:39) merits attention here: Since the discovery of the Qumran material, it is now evident that literature was indeed being composed in Aramaic in the last century B.C. and in the first century A.D. The number of extant Aramaic texts of a literary nature is not small, even though the fragments of them found in the various Qumran caves may be. … then one must beware of exaggerating theoretically the difference between the literary and spoken forms of the language.15

The fact that no Aramaic written Gospel has yet been found does not prove that it did not exist. Aramaic tradition continued in oral and written forms alongside Greek tradition in oral and written forms. It is more proper that the traditions should be circulated in four forms such as Aramaic oral tradition, Aramaic written tradition, Greek oral tradition, and Greek written tradition before the compositions of the four gospels (cf. §1.5.2.5). What is more, even after the composition of the four canonical Gospels, as archaeological evidence indicates, Aramaic composition was active. As Targumic tradition shows, the Greek Gospels could have been translated into Aramaic in places where Aramaic was used. So it is possible that four kinds of Gospels might have been circulated: Aramaic oral Gospel, Aramaic written Gospel, Greek oral Gospel, and Greek written Gospel. At this point, I will take the full bilingualism of Greek and Semitic seriously as shapers of the Jesus tradition during Jesus’ ministry and as shapers of the gospel tradition during the compositions of the four canonical gospels in the first century and continuing thereafter. On the basis of this analysis, it seems that the transmission of the Jesus tradition and the gospel tradition is not unilinear but hybrid, not teleological but circular. In addition, the transmission was not unidirectional but interdirectional. Therefore, the hybridity, the circularity, and the interdirectionality of the transmission of the tradition in the four gospels indicate that the traditions such as the oral tradition and the Aramaic tradition are not always earlier than the traditions such as the written tradition and the Greek tradition respectively. Further, the former two traditions are not always closer to the historical Jesus than the latter two traditions. Although Dibelius developed the discussion 15

Millar (1993:5) observes: “There is ample evidence that a number of Semitic languages were spoken, and mostly also written, in the Roman Near East: Aramaic, Hebrew, Phoenician, Nabataean, Palmyrene and Syriac, as well as the language known as ‘Safaitic’ attested mainly by large numbers of graffiti from the outer edges of the steppe region.” For detailed discussion, see §2.3.

12

The Directionality of the Transmission of the Jesus and Gospel Traditions

about something original behind the fixed Greek text in many parts to a great extent, he made some fatal mistakes. The contrasts he makes between the oral tradition and the written tradition, between the Aramaic tradition and the Greek tradition, and between the testimonies of eyewitnesses and the preachings of eyewitnesses are too sharp. Moreover, he relies too heavily upon assumptions of Aramaic illiteracy and a unilinear, unidirectional, and teleological view of transmission. Despite these serious drawbacks, his views continue to be highly influential upon New Testament scholars.

1.2.2 Rudolf Bultmann Bultmann maintained the three key unidirectional hypotheses of the transmission of the Jesus and gospel tradition. In his view, the traditions were transmitted from oral to written form, from Aramaic to Greek, and from Judaeo-Palestinian to Hellenistic tradition in a unidirectional way. He (1968:6) called this unidirectional concept “laws govern[ing] the development of material.” Like other form critics, he considered the Sitz im Leben transmission from Judaeo-Palestinian to Hellenistic tradition the essential factor in deciding upon modal or linguistic transmissions. As for the modal transmission, Bultmann using form criticism, granted, above all, the modal unidirectionality hypothesis that an oral tradition is unidirectionally transmitted into a written tradition. He (1934:1) asserted, “The purpose of Form Criticism is to study the history of the oral tradition behind the gospels.”16 However, he (1935:12-4) suggested that there were two layers of the oral traditions in the process of the transmission of the Synoptic traditions. One is the oral tradition of the historical Jesus; the other is created or modified in the Sitz im Leben of early Christian communities in and out of Palestine. Bultmann paid attention to a peculiar characteristic of oral tradition itself. Following Gunkel’s argument, he assumed that the oral tradition 16

Dibelius (1929:187) expressed earlier: “Die Formgeschichte hat es bekanntlich nicht mit den abgeschlossenen literarischen Werken zu tun, sondern mit den kleinen Einheiten, die in mündlicher oder schriftlicher Ueberlieferung weitergegeben werden, deren Kenntnis wir aber freilich aus Büchern schöpfen, in die sie Aufnahme gefunden haben.” Subscribing to Dibelius’ remark, Bultmann (1968:4) posits that the purpose of form criticism is “to rediscover the origin and the history of the particular units and thereby to throw some light on the history of the tradition before it took literary form.” Dodd (1938:78) introduced form criticism to English readers by explaining that form criticism is to seek “to reconstruct the oral tradition lying behind the proximate written sources.”

Sitz im Leben

13

was fluid, whereas the written tradition was fixed. He (1961:47; cf. Gunkel 1997:lvi) suggested that “whenever narratives pass from mouth to mouth the central point of the narrative and general structure are well preserved; but in the incidental details changes take place.” Also, the oral tradition was not a vehicle settled enough to preserve the original tradition of the historical Jesus. His skepticism in regard to the oral tradition of the historical Jesus is also shown by his separation of the kerygma from history.17 He proposed that the oral tradition of the historical Jesus was more original but can hardly be traced.18 Furthermore, he (1961:42) considered that the primitive Palestinian Christian church was already steadily forming new sayings of Jesus as oral traditions. He (1934:64-6; cf. 1968:60) supposed that the oral traditions, as faithful expressions of the early Christian community, were in circulation for cultic purposes such as public reading for worship, edification, or preaching in the Hellenistic world. Consequently, he maintained that the oral tradition produced in the Sitz im Leben of Palestinian and Hellenistic Christian communities should be distinguished from the oral tradition uttered by the historical Jesus. Adopting modal unidirectionality from oral to written forms, Bultmann (1968:239; cf. 1968:6, 87, 88, 321) argued that “this distinction [between oral and written tradition] is in my view relatively unimportant for the gospel tradition.” Instead, he continued, “Much more important is the distinction between the Palestinian and Hellenistic stages of the tradition” [emphasis in the original]. Accordingly, he proposed that the Sitz im Leben generally determines whether an oral tradition was produced by the historical Jesus or earliest Christian churches. Bultmann (1935:12-3; 1934:15) also endorsed the hypothesis of linguistic unidirectionality—that an Aramaic tradition was unidirectionally transmitted into a Greek tradition. As he insisted on two layers of the oral traditions in the process of the modal unidirectionality, so he divided the Aramaic tradition into two layers. One was the original tradition that may originate from the historical Jesus. The other is the secondary Aramaic tradition which betrays the specific interests of Palestinian or Hellenistic Christian churches. On account of this, the Aramaic tradition formed by Judaeo-Palestinian or Hellenistic Chris17

18

His view might be influenced by Gunkel’s sharp distinction between “legend” (i.e. the oral tradition) and “history” (i.e. the written tradition), as mentioned before (§1.1). Bultmann (1935:8; cf. 1935:12, 13; 1961:59; 1968:372) wrote: “We can now know almost nothing concerning the life and personality of Jesus, since the early Christian sources show no interest in either, are moreover fragmentary and often legendary; and other sources about Jesus do not exist.”

14

The Directionality of the Transmission of the Jesus and Gospel Traditions

tian communities should be distinguished from that formed by the historical Jesus. He posited three characteristics of the Aramaic “original” in the synoptic traditions: “linguistic peculiarity,” “the use of direct speech,” and “the Aramaic embedded words.” First, concerning linguistic peculiarity at the levels of syntax and semantics, he suggested that some Semitisms should be explained in a different way. He (1968:240) demonstrated that some usages called Semitisms 19 “could also have found their way into the koine Greek.” What is more, he (1968:240) regarded some phrases as Septuagintalisms, not as Semitisms (e.g. Lk 7:16) since “the Hellenistic-Christian linguistic usages were influenced by the LXX.” At the semantic level, he looked upon “direct speech” and “Aramaic embedded words” as insertions by Judaeo-Palestinian or Hellenistic Christian communities rather than sayings by the historical Jesus. He (1968:163) suggested that forms of direct speech like “I-sayings” might be inserted by Palestinian or Hellenistic churches. These speeches were placed on the lips of the risen Lord. With regard to Aramaic embedded words, he considered that they might have been created by Judaeo-Palestinian or Hellenistic churches. Concerning Aramaic embedded words in the miracle stories, he (1934:38; 1961:49; 1968:240) maintained that Aramaic embedded words such as taliqa koum (Mk 5:41) and Effaqa (Mk 7:34) imitate “wonder-working words” in a foreign tongue which was unknown to readers of the gospel. He (1968:222, cf. 1968:240; §8.3.2) suggested, “The miracle working word is frequently given in strange, incomprehensible sounds, or alternatively handed down in some foreign language.” As a result, he thought that the Aramaic embedded words were created in Hellenistic churches since they reflect “wonder-working words” of Hellenistic miracle workers. On this score, Bultmann hardly gives his attention to Aramaic embedded words because they are not always guarantees of the original sayings by the historical Jesus. Consequently, admitting the linguistic unidirectionality hypothesis from Aramaic to Greek, Bultmann argues that the Sitz im Leben generally determines whether Aramaic traditions are derived from the historical Jesus or Judaeo-Palestinian or Hellenistic Christian churches. Finally, Bultmann relied heavily on the concept of Sitz im Leben unidirectionality. He held that Jesus and gospel traditions were unidirectionally transmitted from Judaeo-Palestinian to Hellenistic churches.

19

He illustrates this with the phrases: evn pneu,mati avkaqa,rtw| (Mk 1:23, 5:2), fwnh/| mega,lh| (Mk 1:26, 5:7); evfobh,qhsan fo,bon me,gan (Mk 4:41); ei-j instead of tij (Mk 5:22); e;rcontai avpo. tou/ avrcisunagw,gou scil. tine,j (Mk 5:35); the distinctive sumpo,sia sumpo,sia and prasiai. prasiai, (Mk 6:39-40); the detached nominative h`me,rai trei/j (Mk 8:2), etc.

Sitz im Leben

15

This religionsgeschichtliche view, which was suggested by Bousset (1906) and Heitmüller (1912), is based on a sharp monolingual distinction between the Aramaic-speaking Judaeo-Palestinian church and Greekspeaking Hellenistic churches (cf. Bultmann 1934:17). Bultmann (1935:12-3) mentioned that “these gospels were composed in Greek within the Hellenistic Christianity community, while Jesus and the oldest Christian group lived in Palestine and spoke Aramaic.” Bultmann focuses on distinguishing Judaeo-Palestinian traditions from Hellenistic traditions, no matter which traditions are oral or written, or Aramaic or Greek. Furthermore, on the basis of the Sitz im Leben unidirectionality hypothesis, he unfolds the unidirectional development of christological titles. He (1934:17; 1952:51-52; 1956:175-8) advanced the idea that Jewish Palestinian Christians considered Jesus as the “Messiah” and “the Son of Man,” whereas Gentile Hellenistic Christians thought of Jesus as the “Lord” and “the Son of God” due to their linguistic difference. He (1956:176) supposed that Gentile Christians would not have understood “Christos” as a translation of “Messiah.” Accordingly, christological titles were unidirectionally developed from the “Messiah” and “the Son of Man” used by Aramaic-speaking Palestinian Christians to “Christos,” “Lord,” and “the Son of God” used by Greek-speaking Hellenistic Christians on the basis of the monolingualism of first-century Palestine and the Roman Near East. Despite his widespread influence upon New Testament scholarship, his arguments are open to criticism on several points. Three criticisms can be leveled. First of all, he holds the three unidirectionality hypotheses. However, it would be more persuasive if he took the possibility of the interdirectional transmission of the Jesus and gospel traditions into account more seriously, as we have discussed in detail (§1.2.1). Second, he does not consider embedded Aramaic words as sayings by the historical Jesus but as insertions by Palestinian or Hellenistic Christian churches. However, the study of the bilingualism of firstcentury Palestine and the Roman Near East shows that the embedded Aramaic words could have been used with the translated Greek words in and out of Palestine during Jesus’ ministry and continuing thereafter (cf. chapters 3 and 5). Still, the embedded Aramaic words in Greek texts are not strange, incomprehensible sounds but biblical authors’ literary device to stress their intention (§8.3.2). This means that the embedded Aramaic words could be sayings by the historical Jesus, as I will discuss (chapter 8).

16

The Directionality of the Transmission of the Jesus and Gospel Traditions

Third, Bultmann made a clear-cut distinction between the Aramaicspeaking Judaeo-Palestinian Christian and Greek-speaking Hellenistic Christian churches on the basis of their monolingualism. He (1925:145; 1967:102-3) placed a great deal of weight upon Syrian Christianity since the bilingual Antiochene Christian community bridges the linguistic gap between Judaeo-Palestinian Christianity and Hellenistic Christianity (cf. Introduction to chapter 4). In this respect, he persisted with his conception of the unidirectional development of christological titles. However, his argument would have been more persuasive if he took serious account of the bilingualism of first-century Palestine and the Roman Near East (cf. chapters 3-5). A better understanding of this bilingualism dilutes the sharp distinction between the Aramaic-speaking Judaeo-Palestinian church and Greek-speaking Hellenistic churches. This implies that the Aramaic christological titles were used with their translated christological titles in Greek in the first-century Palestinian Christian church as well as Hellenistic Christian churches. Christological titles were not developed from Aramaic into Greek titles but circulated in Aramaic as well as in Greek. I call this theory interdirectional Christology or bilingual Christology (cf. §5.4; §7.4.3.2; §9.8). Although Bultmann made an enormous and lasting contribution to the understanding of the transmission of the Jesus and gospel traditions, he neglected the bilingualism of first-century Palestine and the Roman Near East and was thus hindered by his commitment to the three unidirectional hypotheses.

1.2.3 Gerd Theissen Theissen devotes considerable attention to the history of the transmission of gospel tradition in earliest Christianity by using a sociological method. Consenting to the main ideas of form criticism,20 he criticizes form critics who disregarded sociological factors of earliest Christian communities and so could not explain by what mechanisms oral tradition was changed into written tradition. He (1992:33-7) proposes that the survival of the Jesus and gospel traditions depends on a specific “social” Sitz im Leben rather than on a “religious” Sitz im Leben. He demonstrates how economic, ecological, political, and cultural factors powerfully shaped the history of the synoptic tradition.

20

He admits some basic principles of form criticism such as small units, oral tradition behind texts, and a sharp distinction between Aramaic-speaking Palestine and Greek-speaking Hellenistic areas.

Sitz im Leben

17

Theissen (1991; summarized in 2003:34-46) explains how the oral tradition is changed into the written tradition by dividing the formation of the synoptic gospels into three stages. He (1991:25-122; 2003:46) states that at the first stage, the oral tradition of Jesus’ deeds and sayings was circulated as small units in three social contexts: (i) Disciples called itinerant charismatics spread the radical ethics of Jesus’ sayings. (ii) The oral tradition related to the apocalypse and the passion account was circulated in the local community in Jerusalem. (iii) The miracle stories about Jesus were handed down among people in Palestine or Galilee. He points out that at the first stage Jesus and his first followers were itinerant charismatics who were able to proclaim radical ethics concerning home and community in an oral form.21 In this respect, Theissen (1992:45) holds that the oral tradition is “authentic” and goes back to Jesus, the first wandering charismatic. At the second stage of the history of the Synoptic tradition, the oral tradition was turned into the written tradition. Theissen (1991:123) assumes that some large units like the great apocalyptic discourse in Mark 13 and the passion narrative might have been written. Analyzing the social and the political situations of the two narratives and the characteristics of the written traditions, he (2003:42-4) claims that the passion narrative was documented in Jerusalem in the 40s and 50s and that the synoptic apocalypse was written as early as 39/40. Moreover, he (1991:203, 233) presumes that the “Sayings Source” was also documented in Greek in the background of Palestine in 40-55. At the last stage, his sociological research (1991:257-281) leads him to the conclusion that the difference in ethical teachings between the Gospel of Mark and the other two synoptic gospels results from their social Sitz im Leben. The Gospel of Mark was written in Syria near Palestine and in close temporal proximity to the Jewish War. In contrast, the two other gospels were written after the Jewish War. By changing radical ethics in the Gospel of Mark into ethics of normalcy, the two later evangelists (i.e. Matthew and Luke) intended to give their local Christian communities strength for their lives continuing after the crisis had been overcome. Accordingly, Theissen (1991:287-9) contends that whereas the Gospel of Mark uses the oral and the written traditions of the ethical viewpoint of wandering charismatics, the other two gospels reshape Jesus’ ethics from the two written documents (i.e. the Gospel of Mark and Q). The synoptic gospels preserve both the oral tradition originally transmitted by itinerant charismatics and the written tradi21

For the characteristics of the radical ethics of wandering charismatics, see Theissen 1978; 1992:33-59.

18

The Directionality of the Transmission of the Jesus and Gospel Traditions

tion altered in the Sitz im Leben of their local communities. Accordingly, he insists that the social Sitz im Leben transmission decides on whether the Jesus tradition is oral or written. Although Theissen opens a new chapter on the sociological approach to the transmission of the synoptic gospels, he carries over the same chronic problems of form criticism. Above all, he (1991:2; cf 1991:2-13) defines the oral tradition as the “oral prehistory” of the synoptic gospels in the tradition of Gunkel and early form critics. His definition implies that oral tradition lacks historicity and, further, the oral tradition contrasts with written history. He (1992:35) does not trust the historicity of the oral tradition and, in comparison, is overconfident in regard to the written tradition. However, the transmission of the Jesus tradition, whether it is oral or written, depends on the transmitters’ interests; it is not because oral tradition itself is apt to change or lacks historicity.22 Secondly, his linguistic dichotomy between Aramaic and Greek depending on sociological-regional distribution is dubious. He (1992:54-5) writes: In the cities the common, everyday language was koine Greek, whereas in the country areas the original vernaculars survived – in Asia Minor until well into the sixth century. But in the Syro-Palestinian area, from the very beginning, the language of Christianity was Aramaic, which was the dialect of the country people; and it is Aramaic that clearly underlies the sayings tradition.

To put it another way, he assumes that the written tradition composed in cities was documented in Greek; on the contrary, the oral tradition was circulated in Aramaic in the Syro-Palestinian area. It seems that he follows the form critics’ presupposition that Aramaic exclusively represents spoken language, whereas Greek primarily represents written language. Unfortunately, Theissen, too, neglects a serious consideration of bilingualism in first-century Palestine and the Roman Near East. His argument might be more persuasive if he took into account the possibility of an Aramaic written gospel and a Greek oral gospel. Furthermore, his modal directionality is still based on the unidirectional hypothesis that the Jesus and gospel traditions were transmitted from oral into written tradition and never vice versa. For instance, he holds that the passion account was written in the 40s and 50s and that the synoptic apocalypse was composed as early as 39/40. If this is the case, as I mentioned before, it is hard to believe that the written tradi-

22

Also, for the discussion of reliability of oral tradition, see the Scandinavian school (§1.3.1; §1.5.2.3).

Sitz im Leben

19

tion remained the written tradition in spite of the gap of several decades, at its maximum thirty years between the written tradition (40s) and the Gospel of Mark (60s or 70s) without returning into oral tradition repeatedly. What is more, from the unidirectional viewpoint his sociological analysis of the different ethical teachings in the synoptic gospels shows that Mark collects both the oral and the written tradition, whereas the other two writers use the two written “documents.” However, did the written traditions (i.e. Sayings Source and the Gospel of Mark) remain the written traditions despite the gap of several decades, at its maximum sixty years between the written traditions (40s) and the Gospels of Matthew and Luke (80s and 90s)? Rather, it is more likely that, after the two documents were written, both were in circulation in oral form as well, and that the two writers (Matthew and Luke) both used their oral traditions and their written traditions, since orality was the major communicative vehicle in ancient times (cf. §1.3.2; §1.5.2). Because of this, his description of the modal unidirectionality was unconvincing. Although Theissen clothes the transmission of the gospel tradition with sociological methodology, he follows the main arguments of form criticism and his forerunner, Gunkel. Accordingly, Theissen should also be considered a unidirectionalist. To sum up, the three scholars surveyed have presupposed the three unidirectionality hypotheses. They supposed that the Jesus and gospel traditions were unidirectionally transmitted from oral into written forms, Aramaic into Greek, and Judaeo-Palestinian into Hellenistic tradition, and never vice versa. The three unidirectionality hypotheses are based on their hard and fast distinction between Aramaic-speaking Palestinian Jewish Christians and Greek-speaking Hellenistic Gentile Christians. This was based on their assumptions of monolingualism in first-century Palestine and the Roman Near East. To date, the three unidirectionality hypotheses have had strong and widespread influence upon New Testament scholarship and in discussions about the historical Jesus, the synoptic problem, textual-critical arguments, and development of christological titles, as I will discuss later. In what follows, I will argue that the linguistic milieus of firstcentury Palestine and the Roman Near East reflected high-levels of bilingualism, not diglossia or minimal bilingualism (chapters 2-5). Unfortunately, form criticism failed to accept that the linguistic milieus of first-century Palestine and the Roman Near East were predominantly bilingual. A better understanding of the bilingualism of first-century Palestine and the Roman Near East will support the three interdirectionality hypotheses. I will argue that the transmission of the Jesus and

20

The Directionality of the Transmission of the Jesus and Gospel Traditions

gospel traditions was not unidirectional but interdirectional between oral and written tradition and between Aramaic and Greek tradition. Accordingly, it is imperative to investigate the linguistic milieus of first-century Palestine and the Roman Near East with an eye toward bilingualism.

1.3 Modal Unidirectionality from Oral into Written Tradition As represented above (§1.2), many scholars have assumed the modal unidirectionality hypothesis. The general assumption is that the Jesus and gospel traditions were unidirectionally transmitted from oral into written tradition or oral tradition into written Gospels, and never vice versa. This unilinear concept is based on Gunkel’s three fundamental ideas and followed by early form criticism. Many still hold (i) that oral tradition was behind written tradition, (ii) that written tradition was the end product of oral tradition, and (iii) that oral tradition holds temporal priority over written tradition. This is well expressed by Robert Grant. He (1963:295) writes, “There was a period of oral tradition which preceded the writing of gospels, and the existence of this period, and of the traditions, can be proved from the New Testament itself.” Also, Kee (1989:245) remarks that “Our written Gospels were the end products of a process of oral transmission of Jesus tradition.”23 In this respect, scholars have sought to find oral features in the canonical and non-canonical gospels as an indication of modal unidirectionality. The presupposition of modal unidirectionality has been the basis of criteria in relation to major issues of textual criticism, the historical Jesus, and the synoptic problem. Oral tradition is generally considered earlier tradition. Also, it is generally considered closer to the original sayings of Jesus and stories about Jesus. Consequently, modal unidirectionality has functioned as a criterion for temporal priority: the more oral, the earlier.

23

In the Anglo-American tradition, modal unidirectionality was proposed by Westcott (1888:165-212). On the basis of form criticism, Dodd (1952:11) assumes that the Gospel of Mark “was written some thirty-five years, or a little more, after the events it records.” In relation to Jesus’ parables, Perrin (1976:8) notes that “the parables did not remain oral, immediate, and highly personal texts. They were remembered, retold, and finally written down…”

Modal Unidirectionality from Oral into Written Tradition

21

Interestingly, as many form critics supposed, both geographical (i.e. Sitz im Leben) and modal unidirectionalities were made plausible on the basis of the linguistic unidirectionality. The scholars examined here (§1.3) also assumed that the linguistic unidirectionality from Aramaic into Greek supports the modal unidirectionality from oral into written. 24 On this point, I will discuss their modal unidirectionality hypothesis, with special attention to the linguistic unidirectionality. Despite the fact that so many scholars have dealt with the relationship between oral and written tradition and between oral tradition and written Gospels in various unidirectional ways and that some of them, to some extent, have considered the interaction between oral and written tradition and oral tradition and written gospels, I cannot enter here into a detailed discussion of the divergent theories of relevant scholars.25 Two groups, first of all, will be surveyed. The Scandinavian school (§1.3.1) and recent orality theory (§1.3.2) will be observed, respectively. And then, Earle Ellis’ view of the relationship between oral and written tradition (§1.3.3) will be discussed. Although the three theories are quite different from each other, they have assumed the same hypothesis of modal unidirectionality.

24

25

Many scholars have thought that Aramaic was not used as a written communicative vehicle in first-century Palestine and the Roman Near East. Frommel (1908:18) supposed that the existence of a primitive gospel in Aramaic could not be found because it was oral. Also, Olmstead (1942) suggests that Aramaic tradition was necessarily oral. For more information, see Goodspeed 1937:127-68; 1942:315-40 (esp. 331-33). Despite the discovery of the Dead Sea Scrolls, some scholars still consider Aramaic tradition to be oral. Acknowledging bilingualism of first-century Palestine, Argyle disregards Aramaic written tradition and Greek oral tradition because he does not take biliteracy into account seriously. He (1974:89) presumes that “If Jesus and his disciples were as familiar with Greek as with Aramaic, the transition from the oral Aramaic stage to the Greek literary stage would have been natural and easy“ [emphasis added]. Although Williams (2004b:7) admits that the Dead Sea Scrolls prove that Aramaic was used as a written language, he plays down the literary function of Aramaic. He (2004b:7) notes that “the written forms of the language, rather than oral forms, are not seen as defining the norms of the language.” As a result, he proposes that it should be given “priority to spoken over written forms of the language.” Also, see Albright 1960:201-2; Black 1967:16. Remarkably, special attention should be paid to James Dunn’s insightful and influential theory. He (1987; 2003; 2005a; 2005b) has persistently considered some gospel tradition as oral tradition in order to solve major issues of the synoptic problem and the historical Jesus. For recent discussions of the relationship between oral and written tradition, especially see Horsley and Draper 1999; Horsley 2001; Horsley, Draper , and Foley 2006; Thatcher 2008.

22

The Directionality of the Transmission of the Jesus and Gospel Traditions

1.3.1 Scandinavian School Standing in the tradition of Gunkel’s modal unidirectionality hypothesis, the Scandinavian school considers that the Jesus and gospel traditions were unidirectionally transmitted from oral tradition into written tradition or written Gospels, and never vice versa. For Scandinavian scholars, the unidirectional concept of modal transmission is not much different from that of the form critics (§1.2), although it is well-known that they have views diametrically opposed to form critics at major points.26 Taking the process and techniques of the oral transmission of the Jesus tradition into serious consideration, Harald Riesenfeld raises major claims on the basis of the modal unidirectionality hypothesis. Birger Gerhardsson deepens this view while focusing upon rabbinic techniques of gospel transmission. These two scholars, Riesenfeld (§1.3.1.1) and Gerhardsson (§1.3.1.2) will be examined.

1.3.1.1 Harald Riesenfeld Riesenfeld pays attention to the oral transmission of the Jesus and gospel traditions in comparison with Jewish rabbinic techniques. Although many form critics regarded transmitters of the Jesus tradition as anonymous, Riesenfeld (1970:16, 18, 22; cf. 1970:54) argues that the transmitters are authorized and trained persons in early Christian churches, such as Paul and the twelve, and that the transmitters rigidly control the tradition by memorizing both form and content. He (1970:21-2) explains that the usages of para,dosij in the New Testament show that the Jesus tradition was kept strictly in the early Christian communities. Concerning the relationship between oral and written tradition, he (1970:52) supposes that the oral and the written tradition existed side by side. Since early Christian communities considered the para,dosij as the sacred word of the New Covenant in Christian worship, they documented the Jesus tradition in analogy with the Old Testament (1970:21-2). He (1970:4) suggests that the traditions were “originally oral, but gradually, as time went on, also written down.” He continues that “then [they were] transmitted, as individual fragments 26

Although Riesenfeld (1970:5-6, 10-1) criticizes the Sitz im Leben unidirectionality of form criticism, he does not deny form criticism’s contributions to the study of the transmission from oral into written tradition. He (1970:52) estimates that “the formcritical school has made a profitable and a lasting contribution.” Also, Gerhardsson (2001:1-3) refutes the Sitz im Leben unidirectional view of form criticism. However, he (2001:56-7) admits that the Sitz im Leben of Christian congregations influences Jesus traditions, in some respects.

Modal Unidirectionality from Oral into Written Tradition

23

or in small groups, until they found their final embodiment in the compilations known to us as our gospels“ [emphasis added]. In this way, he addresses the claim that the oral Jesus traditions were transmitted into written gospels in a unidirectional way. Consequently, his modal directionality seems unilinear, unidirectional, and teleological. As to the linguistic transmission, Riesenfeld seems to suppose that an Aramaic tradition was unidirectionally transmitted into a Greek tradition. He (1970:23) considers the Aramaic tradition to be originally uttered by Jesus. As evidence that sayings by Jesus were preserved in original Aramaic forms in the memories of Jesus’ disciples, he (1970:23) appeals to Aramaic embedded words such as Talitha qĀm. But in relation to Aramaic embedded words, he does not give any explanation to the question why the Greek texts preserve the foreign words (cf. §8.1). Moreover, he (1970:48) thinks that even though the earliest church in Jerusalem inaugurated translation from its beginning, one of Paul’s major contributions to Christianity was to translate Aramaic terms into Greek ones such as Messiah into Christos and maran/marana into kurios. However, he neglects consideration of bilingualism of first-century Palestine and the Roman Near East which might have led to another possibility: that the Jesus tradition was circulated in Aramaic as well as in Greek during Jesus’ ministry. Accordingly, Riesenfeld holds that the transmission of the Jesus tradition was unidirectional from oral into written and from Aramaic into Greek. It would be more convincing, however, if he considered the interdirectionality hypotheses of the modal and linguistic transmissions of the Jesus and gospel traditions.

1.3.1.2 Birger Gerhardsson In pursuing and augmenting earlier observations by Riesenfeld, Gerhardsson also maintains modal and linguistic unidirectionality. He assumes that the Jesus traditions were well-preserved in gospel traditions. This is because Jesus (as the only teacher) used rabbinic techniques of teaching and transmission and required his disciples to memorize the gospel tradition (2001:39, 47; cf. 1998:133). He (2001:49) says that “the continuity and reliability of the early Christian tradition have been preserved without interruption.” In this respect, he (2001:61 n.3) thinks of the Jesus tradition as gospel tradition so that the two terms “are used synonymously” in his publications. Some scholars like McArthur (1970:39-40), however, have criticized that his view of the rigid transmission of “a fixed ‘Holy Word’” could not explain the variations within the synoptic gospels and between the synoptic gospels and the fourth gospel. Davies (1963:467-8) also points out that the Fathers of

24

The Directionality of the Transmission of the Jesus and Gospel Traditions

the Church (e.g. Irenaeus and Papias) knew that the tradition was already multiform, and suggests, in contrast to Gerhardsson, that the tradition was “a more fluid, living tradition.” With respect to the relationship between oral and written tradition, Gerhardsson assumes that oral traditions interacted with written traditions before the written Gospels were fixed. He (2001:85) argues that “written records of varying length, like notes and memory aids (hypomn¾mata) were surely in circulation at a very early stage.” He (2001:85) further states, “It is not impossible, though perhaps not likely, that such records were kept by the disciples and others already in the period of Jesus’ activity.”27 Yet he (2001:50; cf. 2001:117) proposes that “the synoptic material clearly had this same association [with authoritative sacred scriptures] the whole time until it was finally written down“ [emphasis added]. He (2001:1) confines “a period of oral tradition” to “between Jesus’ ministry and the earliest written records.” This implies that he holds modal unidirectionality in that oral tradition ended after gospel traditions were literally fixed. However, if he (2001:117) subscribes to the interface between the oral and the written tradition before the four gospels were literally fixed, it should also have been suggested that the oral tradition interacted with the written gospels after their composition as well. What is more, if orality was accepted as a major communicative vehicle (Gerhardsson 2001:113-6), it is more likely that the oral traditions continued to be used after the literary compositions of the four gospels. The oral tradition and the reoralized tradition of written gospels were continually in circulation in the first century as well as in the second century even after parts of the oral tradition had been transmitted into the written gospels. Accordingly, it is more likely that the written synoptic gospels were transmitted into oral gospels again after their composition. In other words, after the synoptic gospels were documented, they were probably circulated orally as well as in written forms in an interdirectional way. When it comes to linguistic directionality, Gerhardsson also holds the unidirectionality hypothesis. He (2001:51) recognizes that the linguistic milieu of first-century Palestine was bilingual and that many 27

For his detailed discussion of the “use of written notes,” see 1998:157-63; 2001:71 n.25. In relation to Gerhardsson’s notes, Davids (1980:79) criticizes that Gerhardsson disregards “the role of written transmission”. He suggests, “Memory played a large part in the education and transmission of tradition in all of these groups, but it was the memorization of written texts” (e.g. 1Q Sa I, 4-8). In other words, written tradition could have been reoralized and then the reoralized memory may have been rewritten. Also, for other scholars’ arguments on the “notebook hypothesis,” see §1.3.3; §1.5.2.5.

Modal Unidirectionality from Oral into Written Tradition

25

bilingual persons in the Jerusalem church checked and corrected the Jesus tradition which had been translated into Greek. Despite their censorship, he (2001:51) supposes that translation itself caused changes in the tradition because “no translation can be completely identical with the original and that two or more translations will never be word for word the same.” However, he neglected two other possibilities created by bilingualism of first-century Palestine and the Roman Near East. (i) If the linguistic milieu of first-century Palestine was bilingual, the Jesus tradition in Aramaic could have been translated into Greek during Jesus’ ministry. As a result, the Greek tradition would have been circulated along with the Aramaic tradition during Jesus’ ministry. (ii) After the Aramaic Jesus tradition was translated into Greek, the Greek tradition should have been translated into Aramaic again among Aramaic-matrix Christians (for the definition of “matrix,” see §2.1.9). Accordingly, despite the fact that many Scandinavian scholars have tried to correct the form-critical method and suggested some valid alternatives, they could not escape from the modal and the linguistic unidirectionality hypotheses completely.

1.3.2 Orality Theory It has been widely presupposed that oral tradition has temporal priority over written tradition. This presupposition has led to two longstanding suppositions. First, it follows that oral characteristics show more affinity to original sayings by Jesus and stories about Jesus than written tradition. Second, oral features support the Markan priority hypothesis because the Gospel of Mark has been thought to have oral features more so than other synoptic gospels. In these respects, oral features have been used as a criterion for temporal priority. The oral tradition as a modal criterion is shown by Mournet’s expression (2005:156): “the more oral characteristics a text contains, the more likely that the tradition is derivative from the processes of oral communication“ [emphasis in the original]. Admitting that oral features cannot function as “independent criteria” but as “relative criteria,” Henderson (1992:294-5) also proposes that a modal criterion is “actually necessary.” With regard to the modal criterion, a large number of scholars have tried to associate the rough literary style of Mark’s Gospel with Markan priority. This was made popular by the form critics (§1.2).28 From the 28

E.g., Rawlinson 1925:xxxi-xxxii; Taylor 1952:52-3; Trocmé 1975:68-72. Conversely, I would regard them as Mark’s own literary style and deal with them sporadically as

26

The Directionality of the Transmission of the Jesus and Gospel Traditions

perspective of modal unidirectionality, they supposed that the Gospel of Mark was placed between oral tradition and the other written Gospels.29 Recently, the study of orality theory30 has encouraged New Testament scholars to separate oral characteristics from written texts. Under the aegis of the orality theory, they have tried to explain Mark’s own literary styles as oral features for oral circulation. In other words, the lack of literary skill in the Gospel of Mark is replaced by oral features in the Gospel of Mark. Many scholars have singled out oral features from the Gospel of Mark31 or Q.32 This implies that Mark and/or Q are considered to bridge the gaps between oral tradition and the other Synoptic Gospels.33 They have assumed that the Gospel of Mark or Q must have included the legacies of oral features (e.g. oral syntax) more than other materials on the basis of Markan priority hypothesis.34 As a consequence, although they apply oral theory to New Testament writings, they still insist on the modal unidirectionality hypothesis that says that transmission of Jesus and gospel traditions is unilinear, teleological, and unidirectional from oral into written. Although many scholars have dealt with the relationship between oral and written tradition and although their arguments are quite significant, a full exploration of their discussions exceeds the scope of my investigation. Two representative scholars will be mentioned: Werner Kelber and Joanna Dewey. Kelber (§1.3.2.1) argues for discontinuity between oral and written tradition, whereas Dewey (§1.3.2.2) supports partial discontinuity between oral tradition and written Gospels.

29 30

31

32 33 34

Mark’s own rhetoric (e.g. for the discussion of Markan euvqu,j, see §6.3.3; for the discussion of Markan codeswitching, see §8.3). As mentioned previously, Bultmann considered this unidirectionality hypothesis as “laws of development.” E.g. Ong 1967, 1977, 1982; Havelock 1963, 1982:3-38. For detailed discussion of the influence of oral theory upon Biblical Studies, see Horsley, Draper, & Foley 2006; Thatcher 2008. Kelber 1979, 1983; Williams 1985; Dewey 1989, 1994a, 2004, 2008; Pieter Botha 1991:304-31; Bryan 1993; Horsley 2001; Shiner 2003; Horsley, Draper, & Foley 2006. Criticizing the form-critical method, Bauckham (2006:242) persuasively points out that for form critics “Mark’s Gospel is to be seen as a kind of oral literature that incorporates the oral traditions much as they already existed.” Kelber 1983; Horsley & Draper 1999; Dunn 2005b; Horsley 2006. E.g., Kelber 1983:65-6; Dewey 1989; Bryan 1993:72-81. Kelber 1983:64-70; Dewey 1989:37 (cf. 2004); Bryan 1993:72-81. Kelber (1983:95) mentions that “The question is the more relevant if one subscribes to Markan priority, assuming the text is itself the primary record of the transition from orality to textuality.”

Modal Unidirectionality from Oral into Written Tradition

27

1.3.2.1 Werner Kelber Kelber applies recent orality theory to Gunkel’s view. He (1983:8, 15) rightly faults both form criticism and the Scandinavian school for hypothesizing that transmission of the synoptic gospels is “teleological,” “straight,” and “linear.” Nevertheless, on the basis of the modal and the linguistic unidirectionalities, he also emphasizes the difference between orality and literacy and the discontinuity between oral and written tradition. First of all, he (1983:65-66) pays attention to oral features of the Gospel of Mark and regards them as oral syntax and oral stylistic devices to show the oral tradition behind the written gospel. He illustrates the use of a;rcw with infinitive verbs (2:23; 6:7), the adverbial euvqu,j and kai. euvqu,j (1:29; 3:6), iterative pa,lin and kai. pa,lin (usually with verbs of movement, 2:1; 7:31; 14:40 or with verbs of speaking 4:1; 10:1, 10), the popular kai, (9:2, 11:20), kai. gi,netai or kai. evge,neto (1:9; 2:15), historic present, the frequent use of the third person plural instead of passive (5:14; 8:22), direct speech, etc. This indicates that he considers the Gospel of Mark, with its rich oral features, as a medium between oral tradition and the other synoptic gospels on the basis of Markan priority and from the perspective of the modal unidirectionality. Second, although Kelber only briefly mentions Gunkel (Kelber 1983:xiii, 87 n.96, 107), it seems that his overall arguments are heavily influenced by Gunkel (§1.1). He (1983:105) starts with the difference between oral and written tradition, that is, oral tradition is flexible and fragile, whereas written tradition is fixed and immortal. He (1983:30) asserts that “The concepts of original form and variants have no validity in oral life, nor does the one of ipsissima vox” [emphasis in the original]. For this reason, he (1983:94) suggests “oral pluralism,” which means that “each performance is ‘an’ original, if not ‘the’ original“ [emphasis in the original].35 Further, his overemphasis on the difference between orality and literacy led to a purported discontinuity between oral tradition and the Gospel of Mark. He (1983:95) argues that “the gospel [of Mark] arises not from orality per se, but out of the debris of deconstructed orality.” In this sense, he (1983:196) maintains that the Gospel of Mark was a text that was written “not to continue but to overcome oral mentality.” That is, according to Kelber (1983:185-6), the Gospel of Mark was written to supersede the oral tradition produced by “oral authorities” (i.e. disciples, Jesus’ family, false prophets, and Christs). 35

Kelber (1983:30) cites this from Lord’s statement (1960:101). However, this is not far from Gunkel’s original idea (§1.1). Kelber’s theory is partly endorsed by Dunn’s exhaustive series of New Perspective on Jesus (2003, 2005a, 2005b).

28

The Directionality of the Transmission of the Jesus and Gospel Traditions

Consequently, he proposes that the discontinuity between the oral tradition and the written gospel proves that the transmission of the gospel tradition was neither teleological nor linear. However, he still presumes the modal unidirectionality from oral into written tradition. He has three reasons for doing so. (i) Oral traditions lie behind the written gospels in fragments. (ii) After the Gospel of Mark (and gospels) was written, the oral tradition ended. (iii) The oral traditions could have temporal priority over the written tradition. Moreover, his modal unidirectionality seems to be based on linguistic unidirectionality. Kelber (1983:21) states that Jesus was only an oral performer who could not have left written materials. Moreover, he maintains that Jesus’ disciples displayed “only tenuous connections with literate culture.”36 He (1983:17-8) relates their illiteracy to the rural locations where the Jesus movement originated and where the Jesus tradition was delivered. Conversely, the written gospels were composed in the context of Hellenistic cities. He appeals to Theissen’s theory (1978; cf. §1.2.3) to explain the relations between orality and rural areas, and between literacy and cities. Following Theissen, Kelber associates Aramaic tradition with oral tradition. This is well expressed in this comment (1983:66): “Even if some of these features are traceable to Aramaisms or Semitisms, this does not preclude their oral propensity.” Reading between the lines, it seems that Kelber’s linguistic unidirectionality colored his understanding of the relationship between orality and literacy and the discontinuity between oral tradition and written gospels. Four shortcomings warrant critique. First, he tries to relate the oral features of the Gospel of Mark to Markan priority: that is, he places the Gospel of Mark between oral tradition and the other synoptic gospels. However, oral features should not always warrant temporal priority over written features. As is well known, when authors wrote their books in ancient times, they were intended for reading out loud and/or for oral performance (cf. §1.5.2.4, n.37). This means that oral features are not necessarily indicative of oral tradition and that temporal priority should not necessarily be granted to writings with oral features. Henaut (1993:113) rightly cautions that “Such often-cited criteria as alliteration, dualities, triads, chiasmus and inclusio are found in written texts.” He (1993:68) correctly points out that “An oral flavour to a nar-

36

As for the note-taking hypothesis, he (1983:25) mentions that “the taking of notes and the cultivation of writing was a world apart from the life style of these prophetic transmitters of Jesus’ sayings. They needed no aids in writing because they practiced the message they preached.” Contra his view, see §1.5.2.5.

Modal Unidirectionality from Oral into Written Tradition

29

rative … does not prove an oral tradition.” For this reason, he (1993:113) insists that the stylistic devices “belong to Mark’s overall narrative structure, and are clearly a part of his literary technique” [emphasis in the original]. Consequently, they should be considered as Markan literary styles and do not necessarily bespeak oral tradition or temporal priority. Second, perhaps Kelber’s most problematic argument is that he overemphasizes the contrast between the oral tradition and the written gospels based on his sharp distinction between orality and literacy. In the same way that I criticized Gunkel’s arguments (§1.1), it can be said that the relationship between orality and literacy is not mutually exclusive but interwoven and complementary. Recently, the interaction between orality and literacy has been suggested by many scholars. 37 Moreover, if he (1983:23) thinks that the major communicative vehicle of first-century Palestine and the Roman Near East was orality, it is more likely that the Gospel of Mark was written for oral performance and then circulated in oral forms simultaneously with the circulation of written forms (cf. Kelber 1983:94).38 Accordingly, his emphasis on the stark contrast between orality and literacy is not persuasive. Third, Kelber does not hold his unidirectionality very firmly at times. On the one hand, he asserts that the Gospel of Mark put an end to oral tradition on the basis of discontinuity between oral tradition and

37

38

For instance, concerning the Hebrew Bible, Niditch (1996:1) argues that literacy in ancient Israel should be considered in relation to “its continuity and interaction with the oral world.” She (1996:5) assumes that an oral tradition should be documented and then, the written tradition may be reoralized again. Van der Toorn (2007:12) notes, “In Babylonia and Israel, writing was mostly used to support an oral performance.” Lentz (1989) and Thomas (1989, 1992) propose that orality and literacy are not exclusive but coexisted and interacted in ancient Greece. Gamble (1995:30) also criticizes Kelber’s thesis, saying “a too sharp theoretical differentiation [between oral and written modes] misconceives the situation.” He suggests that the two modes “were far closer and interactive.” Rebutting Kelber’s difference between orality and literacy Gerhardsson (2001:116-7; here 117; §1.5.2.4) suggests that “The two media stood in very intimate interaction.” Halverson (1994) persuasively refutes Kelber’s overemphasis on the contrast between orality and literacy. For more information, see Achtemeier 1990:3-27; Henderson 1992:283-306; Robbins 1993; Horsley, Draper & Foleys 2006. Furthermore, Fox (1994:127) rightly points out that “oral tradition continued to carry special authority well into the second century” after the written Gospels were composed beginning in the 60s. Fox (1994:127) also assumes that “Various Gospels soon circulated, but by the third quarter of the second century the four which we now recognize were widely, though not universally, regarded as special. As texts took over and the length of the oral tradition grew, respect for orally transmitted sayings diminished.”

30

The Directionality of the Transmission of the Jesus and Gospel Traditions

the written gospel. On the other hand, he (1983:17, 94) he presumes that the characteristics of the oral medium are “tenacious” and “dominated long after introduction of writing,” and that the written gospel was designed “to be recycled into the oral medium.” He (1983:93) subscribes to Koester’s assumption that the oral tradition coexisted with the written Gospels up to the middle of the second century. He (1983:23) states, “The concept of a predominantly oral phase is not meant to dispense with the existence of notes and textual aids altogether.” He (1983:93) proposes, “The transposition of oral forms of speech into gospel textuality did not put an end to speaking.” Consequently, his tendency to dichotomize orality and literacy is not helpful. As a matter of fact, he (1983:32) asserts that “Media differences are blurred or belittled [by Gerhardsson and Bultmann].” Yet, he (1983:23) suggests that “The lines of orality and textuality [i.e. textual aids] were indeed blurred in those instances.” His work (1983) tended toward the default of unidirectionality. Accordingly, a better explanation is that the written forms were not the end of the oral tradition but reoralized in an oral-centered society. Last, it seems that the discontinuity between the oral tradition performed in rural areas and the gospels written in cities stems from his linguistic bias. He seems to inherit this tendency from form criticism, of which he is overtly critical, as indicated by his appeal to Theissen’s stance that Aramaic was oral, whereas Greek was written. Along the same lines as my criticism of Theissen (§1.2.3), Kelber would have done better to consider the bilingualism of first-century Palestine and the Roman Near East. In relation to this, Niditch (1996:3) persuasively argues, “Orally composed and oral-style works can be rural or urban.” The study of bilingualism supports the modal and the linguistic interdirectionality hypotheses in that both written tradition in Aramaic and oral tradition in Greek would have been circulated in parallel with written tradition in Greek and oral tradition in Aramaic in rural areas as well as cities (see Part I). Kelber deserves credit for applying oral theory to gospel transmission in various ways. His innovative ideas have had wide influence upon New Testament studies and have attracted lively discussion among eminent scholars.39 However, he could not overcome the modal and the linguistic unidirectionality hypotheses completely and thus committed the same mistakes as Gunkel. Although Kelber (1983:32)

39

Especially, Horsley, Draper, & Foley (2006) edited a book dedicated to Werner Kelber in which they appropriately eulogized him (vii-xvi). Also, see Thatcher 2008:143.

Modal Unidirectionality from Oral into Written Tradition

31

partly escaped from the concept of unilinear transmission that form critics and the Scandinavian scholars held, he still failed to take the necessary step toward interdirectional transmission.

1.3.2.2 Joanna Dewey In many respects, Joanna Dewey’s work has exemplified modal interdirectionality between oral and written tradition. While she should be commended for this, I will argue that she still exhibits some modal unidirectionality tendencies. To her credit, she (1994b:37-8) convincingly critiques the form critics’ modal unidirectional assumptions40 and Kelber’s discontinuity between orality and literacy and his rapid supersession of oral tradition with written gospels. Instead, she proposes that orality influenced the Gospel of Mark both in the process of its composition and during its performance in Christian worship. Above all, Dewey (1989:33; cf. 1989:44, 1994a:145-63) stresses the interaction between oral tradition and Mark’s Gospel within the process of the gospel’s composition. She emphasizes this for two reasons. (i) The Gospel of Mark was written for listening audiences since orality was the central medium of first-century Christianity. (ii) Oral techniques influenced writing when the evangelist composed the gospel. She (2004:503-4) insists that the Gospel of Mark was “building on, refining, and developing an oral tradition that had already created a continuous, more-or-less coherent narrative” and that the written Gospel of Mark did not compete with, but coexisted with, oral circulation. Her view can be seen as a correction for Kelber’s extreme discontinuity between oral and written tradition. Taking a further step, she intriguingly pays attention to oral re-performance of the written text of the Gospel of Mark. She (2004:496; cf. 1994a:151-9) suggests that after the Gospel of Mark was composed around 70 AD, “it continued to be performed orally.” She (2004:498) considers that the boundaries between oral re-performance and a re-written Gospel of Mark were fluid. She (2008:82) convincingly concludes, “In the ancient media world, the Gospel of Mark was likely to have been oral, written, recycled into orality, rewritten, and so on.” With this argument in place, she adheres to the modal interdirectionality hypothesis.

40

She rebuts their three assumptions, saying that “the progression from oral performance to written text was a continuous linear development, with writing rapidly becoming the primary medium for Christians, and with written texts supplanting oral tradition as soon as they were composed.”

32

The Directionality of the Transmission of the Jesus and Gospel Traditions

Like many scholars (e.g. Kelber), Dewey (1994a:149-57; 1989:37) suggests some oral features in the Gospel of Mark on the basis of the Markan priority hypothesis: additive and aggregative composition (e.g. the use of paratactic kai,), agonistic tone, and participatory character. She is unclear about whether or not these oral features are indicative of Mark’s compositional placement. One might infer from her work that the Gospel of Mark was a medium between oral tradition and the other written synoptic gospels. This view, of course, is unconvincing for the same reasons that Kelber’s argument was unconvincing (§1.3.2.1). The oral features of the Gospel of Mark do not necessarily prove that the Gospel of Mark is more oral than other gospels and that the Gospel of Mark has temporal priority over other gospels. Moreover, it seems that she does not break with Kelber completely since she still suggests a dichotomy between nonliterate and literate carriers of New Testament tradition(s). When it comes to the transmission of New Testament manuscripts in the second and following centuries, she (1994b:59) holds a contrastive position in regard to the relationship between oral tradition and written Gospels. It seems, in a sense, that she replaces Kelber’s contrast between orality and literacy with her own contrast between nonliterate and literate. She (2008:86) argues that “the New Testament writings have ultimately silenced and largely controlled [Jesus]” [emphasis added]. According to Dewey (2008:86-7), only the literate (i.e. small educated male elite) can handle and transmit the written manuscripts, whereas the oral tradition delivered and circulated among the nonliterate (i.e. the largely uneducated, female non-elite) came to an end.41 In this respect, the written tradition triumphed over the oral tradition which was “the living, speaking Jesus” in circulation. As a result, she (1994b:59-60; cf. 1998) insists that the New Testament writings were distorted by the interests of the small educated male elite from the second century onward. Her view may reflect form criticism’s explanation for oral tradition. She (1994a:145; 2008:86-7) considers oral tradition as “nonliterate” tradition.42 Consequently, she holds that the written gospels are the end products of oral traditions. As such, while she does well to move beyond unidirectionality at points, this stance betrays the tendencies of the modal unidirectionality.

41

She (2001:242) describes, “The Gospel of Mark and John are at the oral end of the continuum.” 42 Henderson (1992:294) also describes oral tradition as “pre-history” tradition, like form critics defined.

Modal Unidirectionality from Oral into Written Tradition

33

This theory is not persuasive if orality was the major communicative vehicle not only in the first-century AD, but even in the fourth century AD and beyond. Also, she (2004:505) admits that “the oral tradition [of traditional European fairy tales] continued alive and largely unaffected by the printed versions until well into the twentieth century.” She also remarks that the oral tradition ended due to “increasing literacy of the population and the availability of cheap books, hardly factors in antiquity.” This means that oral tradition continued to be used together with written tradition. Second, the literate did not refer to only the educated male elite. Some lower-status Christians including slaves would have been literate (§3.2.6). Third, as Gamble (§1.5.2.4) persuasively argues, the fact that the same texts were read aloud repeatedly shows that reading written tradition did not provide new information but reminded Christians what they had already known. It also shows that 90% of the Christian people who were illiterate could share the knowledge of the gospel traditions. We cannot make a clear distinction between “orality for the illiterate” versus “literacy for the literate.” Consequently, her dichotomy is untenable since it leads back to the modal unidirectionality hypothesis: (i) oral tradition circulated among the masses who were uneducated, female, or non-elite behind written gospels which were distorted by people who were educated, male, and elite, (ii) the written tradition as the end product of the oral tradition, and (iii) temporal priority of the oral tradition over the written tradition. Accordingly, although Dewey persuasively suggests modal interdirectionality, it is regrettable that her argument is partly based on the modal unidirectionality hypothesis.

1.3.3 The Earliest Written Tradition: Earle Ellis Another point to be noted here is Earle Ellis’ argument. Contrary to the form critics’ view of “laws of development” (1978:240), he has persistently argued that some Jesus traditions were written during Jesus’ ministry. He (1999a:352) argues that “individual episodes (and groups of them) were put in writing very early, some perhaps during Jesus’ ministry.” Classical form criticism supposed that the Jesus tradition was orally circulated without a written form before the composition of the Gospel of Mark because Jesus’ followers expected the imminent end of the world (cf. §1.2), as Dibelius proposed. 43 Ellis (1978:243; cf. 1999b:53; 43

He (1934:9) says that “The company of unlettered people which expected the end of the world any day had neither the capacity nor the inclination for the production of

34

The Directionality of the Transmission of the Jesus and Gospel Traditions

2001:12) criticizes this apocalyptic view, saying that this view “foundered with the discovery of the Dead Sea Scrolls.” He continues, the Qumran community “viewed itself to be in the ‘last generation’ (1QpHab 2,7; 7,2), expected an imminent end but, nevertheless, produced a large body of literature.”44 As a consequence, although first followers of Jesus believed in the imminent end of the world, their position did not keep them from writing about Jesus. Ellis (1978:243-7; cf. 1999a:22) makes five suggestions to support his view. First, some disciples and followers of Jesus were literate, since Jewish Law commanded that Jews should teach their children to read and write, according to Josephus (C. Ap. 2.204; Ant. 4.211) and Philo (Leg. 115, 210). Second, Jesus and his disciples held a prophetic view concerning biblical tradition, which means that they, like the Qumran community, were not banned from writing sacred texts. Third, form critics presupposed that oral transmission led to variants of tradition which were quite different from the original tradition. This view has become passé. Recent studies of oral folk traditions show that oral transmission does not necessarily lead to variants. Fourth, when Jesus’ disciples were sent off on their brief mission, they might have left some literary forms for their hearers. Fifth, he considers some disciples and converts to be bilingual. Bilinguals both transmitted and translated the Jesus tradition in written forms as well as orally for purposes of wider transmission. Later, he adds one more reason to the list. He proposes

44

books, and we must not predicate a true literary activity in the Christian Church of the first two or three decades.” Following the apocalyptic view, Hengel (2005:73) states, “This absence of early literary witnesses is all too easy to understand: one who awaits the end of the ‘old, evil world’ in the near future is not at first interested in a literary consolidation of history for posterity. It is enough to proclaim orally what the disciples and he himself have from their experience with Jesus.” Davids (1980:79) argues that “some groups contemporary with the New Testament had no hesitation in composing their works as written documents (e.g. the apocalypticists and the sectaries of Qumran).” Millard (2000:198-9) also criticizes the apocalyptic view. He argues, “Nothing in the history of Greek and Roman literature or the distribution of literary texts among the papyri from Egypt supports the idea, either.” Taking a further step, he suggests an instructional view that the Jesus tradition was written for edificatory purposes, that is, for teaching in detail the earliest Christians in the churches. Furthermore, form critics assumed that the earliest kerygma did not include the gospel traditions related to church life and behavior and that they were added later by the Christian church because, although Jesus stressed his imminent return, his second coming was delayed. Blomberg (1992:246) criticizes this assumption and argues that Jews knew that “imminent” did not mean “right soon” because they knew the meaning of “the Day of the Lord was ‘at hand’” (e.g. Ps 90:4; Joel 2:1; Obad 15; Hab 2:3). Thus, Blomberg (1992:294) asserts that Jesus’ emphasis on his imminent return has been exaggerated.

Modal Unidirectionality from Oral into Written Tradition

35

that the prolific writing of first-century Palestine upholds the possibility that some Jesus traditions were written during Jesus’ ministry. He (1999b:54) suggests that “the widespread literacy in first-century Palestinian Judaism … would have facilitated the rapid written formulations and transmission of at least some of Jesus’ teaching.” Ellis deserves credit for having raised awareness about the possibility of written materials before the composition of the Gospel of Mark and even during Jesus’ ministry. Although he seems to push some of his assumptions too hard, his basic tenets are difficult to dispute. His argument is endorsed by Stanton and Bauckham, who proposed a further elaboration of notetaking theory (§1.5.2.5). Unfortunately, it seems that Ellis’ theory is still placed within the modal unidirectionality hypothesis. He seems to suppose that initial written tradition during Jesus’ ministry remained without reoralization. However, the written tradition is not the end of the oral tradition. If he assumes that the Jesus tradition was written during Jesus’ ministry, it is more likely that the written tradition was orally circulated again. Second, he (1978:246-7) does well to assert that the bilingualism of first-century Palestine undermines both the assumptions of the linguistic and the geographical unidirectionalities. In this way, his work stands out from his field. He believes that some disciples, hearers, and converts were bilingual. This implies that the Jesus tradition was translated from Aramaic into Greek during his ministry. Furthermore, contra form criticism’s Sitz im Leben unidirectionality hypothesis from Judaeo-Palestinian into Hellenistic tradition, he (1998:76) proposes that “good Greek in a Christian document was no sign of a later date or diaspora origin, although Semitic Greek might be a sign of Palestinian or Syrian origin.” For these reasons, he (1998:76) rightly suggests that bilingualism discloses “how false was the dichotomy drawn by the ‘history of religions’ school between Palestinian and diaspora Judaism.” However, some points related to the bilingualism of first-century Palestine and the Roman Near East need to be reconsidered.45 Although he (1998:76) mentions that “Palestine was largely bilingual,” his arguments seem to betray the fact that he (1978:245-7) holds to minimalism 45

He makes the linguistic distinction based on the ethnic division of Palestinian Jews and Gentiles from the Judaeo-Palestinian region. He (1999b:54 n.19) shows that some converts from Syria, the Decapolis, Tyre and Sidon could speak Greek. However, some converts from these regions must have spoken Aramaic as their matrix language (§5.2.1). Also, although he (1978:246-7) assumes that the Hellenists in Acts 6:1 were Greek speakers who could not speak Aramaic, it seems that the Hellenists were Greek-matrix speakers who were bilinguals in Greek and Aramaic (see chapter 5).

36

The Directionality of the Transmission of the Jesus and Gospel Traditions

(i.e. partial bilingualism; cf. §2.3.3). It is also regrettable that he did not further develop his valuable contribution. He neglects the implication that the bilingualism of this period showed the linguistic interdirectionality between Semitic and Greek traditions. Accordingly, although Ellis persuasively points out that the Jesus tradition was written and translated into Greek during his ministry, he seems to stand in the classic form criticism tradition that written Greek tradition is the end of oral Aramaic tradition in a unidirectional way. Nevertheless, his investigation shows that the study of bilingualism is urgently required when it comes to transmission of Jesus and gospel traditions. As surveyed (§1.3), many scholars hold the modal unidirectionality hypothesis: (i) oral tradition stands behind written tradition, (ii) written tradition is the end product of oral tradition, and (iii) oral tradition has temporal priority over written tradition. However, as we discussed, it is more persuasive to consider the modal interdirectionality hypothesis. Furthermore, many scholars have recently suggested the modal interdirectionality hypothesis, which will be investigated in detail (§1.5.2). Accordingly, one cannot make a clear distinction between oral and written tradition and insist on the priority of oral tradition over written tradition. The transmission of Jesus and gospel traditions is not unilinear, teleological, or unidirectional but hybrid, circular, and interdirectional.

1.4 Linguistic Unidirectionality from Aramaic into Greek Tradition There has been a general consensus that Jesus and his disciples usually spoke Aramaic, whereas the synoptic gospels we have were written in Greek. Although the linguistic gap between Aramaic and Greek has been explained in various ways, the arguments have been based upon the linguistic unidirectionality hypothesis that Aramaic tradition was transmitted into Greek tradition in a unidirectional way. Moreover, as we have surveyed, both the Sitz im Leben unidirectionality hypothesis (§1.2) and the modal unidirectionality hypothesis (§1.3) are heavily based upon the linguistic unidirectionality hypothesis. The linguistic unidirectionality hypothesis is closely related to viewing the linguistic milieu of first-century Palestine and the Roman Near East as monolingualism (§1.4; cf. §1.2, §1.3), diglossia (§2.2, §2.3), or partial bilingualism (§2.3.3). I will discuss these, respectively. The linguistic unidirectionality hypothesis has been maintained by two major groups, whether ipsissima verba Jesu graeca are accepted or

Linguistic Unidirectionality from Aramaic into Greek Tradition

37

not. Some scholars who consider the Aramaic language as a criterion for determining the authenticity of the Jesus tradition have supposed that the Jesus tradition was spoken (and written) in Aramaic and later translated into Greek without exception. They have concentrated on singling out Aramaic traditions from the Greek texts we have because they have considered that an Aramaic tradition is more original and earlier than a Greek tradition. Their default supposed the monolingualism of first-century Palestine, even though the linguistic milieu should be considered to be bilingual (chapters 2-4). On the other hand, others who take the view of a Greek language criterion for the authenticity of the Jesus tradition have taken account of bilingualism of first-century Palestine. Proponents of this view posit that Jesus, who was a bilingual, spoke Greek as well. This has led to an assumption that all Jesus tradition was not transmitted from Aramaic into Greek. Some exceptional Jesus traditions in Greek might have been spoken by Jesus. However, a problem stems from the fact that even the scholars who hold ipsissima verba Jesu graeca did not seriously consider the bilingualism of firstcentury Palestine. The ipsissima verba Jesu graeca are not the end product. If they had taken bilingualism into account more seriously, they might have concluded that the Greek Jesus traditions were translated into Aramaic, as Aramaic traditions were translated into Greek. Consequently, both approaches have been based on the unidirectional linguistic transmission of the Jesus and gospel traditions. If the bilingualism of first-century Palestine and the Roman Near East was considered more seriously, the history of the linguistic transmission of the Jesus and gospel traditions would have been totally different. The sayings by Jesus and the stories about Jesus were circulated in Aramaic as well as in Greek during Jesus’ historical ministry and continuing thereafter. Furthermore, the bilingualism of first-century Palestine and the Roman Near East implies that gospel tradition in Greek could have been translated into other languages. As a matter of fact, we know that the Gospel of Thomas, the Gospel of Mary, and the Gospel of Judas in Greek were circulated along with Coptic versions among Coptic-matrix speakers.46 The synoptic gospels in Greek were most likely translated into Aramaic in oral and written form among Aramaic-matrix Christians, which is reflected in the later Syriac versions (i.e. Sinaitic, Curetonian, Peshitta, and Harklean). Accordingly, the bilingualism of Palestine and the Near East occupied by the Roman Empire leads to the conclusion that the linguistic transmission of the Jesus and gospel tradi-

46

For detailed discussion, see §2.3.4. For the definition of “matrix speaker,” see §2.1.9.

38

The Directionality of the Transmission of the Jesus and Gospel Traditions

tions was not unilinear, teleological, and unidirectional but hybrid, circular, and interdirectional. The linguistic transmission of Jesus and gospel traditions has been discussed in relation to major issues of gospel studies such as the linguistic milieu of first-century Palestine, the historical Jesus, New Testament Greek, Semitisms and Septuagintalisms, the synoptic problem, the development of christological titles, etc, which will be discussed over the course of this work (cf. n.48). Here, the two approaches to the linguistic transmission of Jesus and gospel traditions will be reviewed: Aramaic language as a criterion (§1.4.1) and Greek language as a criterion (§1.4.2).

1.4.1 Aramaic Language as a Criterion Most scholars have assumed that Jesus and gospel traditions were transmitted from Aramaic into Greek without exception. They have usually focused greatly upon Semitisms as well as the phonological, syntactic, and semantic levels on the basis of the linguistic unidirectionality hypothesis.47 Aramaisms have been used as a criterion in this way: The more Aramaized, the earlier, which has been accepted widely as a criterion for deciding the temporal priority of the Jesus and gospel traditions.48 In this chapter, four major Semitic scholars will be discussed: 47

48

For an historical survey of the Semitisms of the New Testament, see Maloney 1981:734; Voelz 1984:894-930; NewDocs 5.5-40; Black 1998:v-xxv; Jung 2004:5-61. For the discussion of Semitisms on the basis of the linguistic unidirectionality hypothesis, see Syntax (chapter 6), Phonology (chapter 7), and Semantics (chapter 8). Regarding the unidirectional influence from one language to another, Thackeray (1909:21) posited that “the Greek language was at all times the giver rather than the receiver.” In this respect, Downes (1984:67) mentions that in bilingual contexts “one might expect the Low to be more open to borrowing terms from the High than vice versa, because of the prestige and standardization.” That is, it is assumed that Greek was prestigious and standardized, whereas Semitic languages were not. Silva (1980:207) also proposes that “the native language of a bilingual is not affected in the same way as his or her second language” [emphasis in the original]. However, illustrating that Greek was a low-prestigious language at Rome, Horsley (NewDocs 5.7) persuasively claims that “it is not true that the lower language is always the borrower and never the lender.” Furthermore, unfortunately, both Thackeray and Downes consider the linguistic milieu of first-century Palestine and the Roman Near East to be diglossia. However, the linguistic situation should be regarded as bilingualism rather than as diglossia; for detailed discussion, see chapter 2. Aramaisms have been used as a criterion in major issues of gospel studies. For instance, Kilpatrick (1963:126-7) suggests that the more Semitizing the text, the more original it is when he points out that avpokriqei.j ei=pen is more original because it is not a Greek expression. He (1970:170) also suggests that “we assume that where we

Linguistic Unidirectionality from Aramaic into Greek Tradition

39

Charles Torrey (§1.4.1.1), Matthew Black (§1.4.1.2), Joachim Jeremias (§1.4.1.3), and Joseph Fitzmyer (§1.4.1.4).

1.4.1.1 Charles C. Torrey Unlike form critics who made a sharp distinction between oral Aramaic and written Greek, Torrey took seriously the possibility of a written Aramaic Gospel.49 Although many scholars had previously considered the Aramaic approach, he offered a full-dress discussion of the possibility that the four gospels were a translation of Aramaic written Gospels.50 In terms of the Greek of the four gospels and the linguistic milieu of first-century Palestine, he suggested two points: (i) The four gospels are in unpolished Greek and involve mistranslations (1933:267-73). His view of the Greek is simply shown by his remark (1936:lviii): “Any have two forms of the Greek text, one of which is closer to Semitic idiom than the other, that the form nearer to Semitic idiom is more likely to be right, other things being equal. This is probably correct.” In terms of Markan Semitisms, Taylor (1952:56) thinks that “If Semitisms … appear in some narratives or types of narratives more than in others, the evidence may have an important bearing on the historical character and origins of Mark.” Wilcox (1965:184) asserts that Semitisms in the Apostles of Acts “are also a sign to us of the authenticity and antiquity of the material enshrined.” Quispel (1966:378) holds the same assumption in discussing the source of the Gospel of Thomas. As for textual criticism, Black (1998:31) asserts that “Codex D is “the more primitive type of text” because it “stands nearer the underlying Aramaic tradition.” Concerning the synoptic problem, Kümmel (1975:56-60) supposes that Markan Semitisms are one evidence in support of Markan priority. Taylor (1952:65) also mentions that “its sayings and many of its narratives [in the Gospel of Mark] stand near Semitic tradition.” Similarly, in relation to Jesus’ language Dalman (1909:42) says that “the fewer the Hebraisms, the greater the originality.” Henry Turner (1963:77-8) suggests Aramaic as a criterion for authenticating gospel traditions, saying that “the closer the approximation of a passage in the Gospels to the style and idiom of contemporary Aramaic, the greater the presumption of authenticity.” James Edwards (2009) states that the Hebrew gospel was widespread and widely known in early Christian communities. He advocates the view that Luke used both the Gospel of Mark and a Greek version of the Hebrew Gospel instead of Q. However, his argument is still based on the linguistic unidirectionality from Hebrew into Greek. 49 For this reason, Cadbury (1958:31) mentions: “If they [Torrey and Burney] should be ultimately accepted, this would add a whole new factor in our scheme of transmission of gospel history-a layer of Aramaic writings. It would mean that the transfer from Aramaic to Greek occurred not in the oral stage as we used to think, but in the literary stage.” 50 His argument is known as translation theory, which had been suggested by Torrey’s predecessors (e.g. Wellhausen and Dalman). Burney (1922, 1925), one of the scholars who made the largest contribution in this direction, also sought to achieve Aramaic gospel reconstruction. However, his drawback is similar to Torrey.

40

The Directionality of the Transmission of the Jesus and Gospel Traditions

high school boy fairly well trained in Greek composition would be ashamed of such sentences as 10:38a or 22:54a [of the Gospel of Luke].” (ii) Surprisingly, he supposed, even before the discovery of the Dead Sea Scrolls, that there was a large number of Aramaic written materials in Palestine in the first centuries BC and AD. He (1933:252; cf. 1912:289; 1942:71-85) speculated that “The pre-Christian Aramaic literature, which must have been very extensive, rich in every field, shared the general fate of other ancient literatures.” In regard to the linguistic transmission of the four gospels, Torrey (1912:289-307; 1942:84) argued that Mark, Q, Matthew, and John were composed and circulated in Aramaic and were translated into Greek. According to him (1912:288-97), Luke considered that Aramaic sources are more authentic than Greek so that he translated the original Aramaic recensions (i.e. Mk, Q, and Mt) into Greek in comparison with their Greek recensions. Torrey (1912:289-90, 308-9) also assumed that Aramaic documents were more authentic than Greek. He stated three reasons. (i) The earliest tradition in Palestine was documented and circulated in Aramaic (1912:289). (ii) Contemporary authors like Luke held the same view of the Aramaic original as the earliest tradition, mentioned before. He (1912:289) suggested that “to any collector of traditions, discourses, and other historic material, these Semitic documents would appeal as especially ‘authentic.’” (iii) Translations including the four gospels involve many mistranslations whether it is intentional 51 or not. 52 Consequently, he (1912:305) proposed that “Semitic documents would be valued higher than Greek.” Thus, he implied that the Aramaic tradition has temporal priority and authenticity more than Greek does. Taking a further step, his argument that the Jesus and gospel traditions were circulated in Semitic as well as in Greek between the first century and the first decades of the second century merits our attention. He (1912:296) posited that “In the Oriental church, both Semitic

51

52

Torrey (1912:279; 1933:ix-x; 1936:lviii) suggests his view that the authors of the four gospels intentionally preferred literary translation in order to transmit the Semitic original. He (1936) illustrated ten unintentional factors; ambiguity of the Aramaic text, questions misunderstood as declarations, the redundant “and,” the reflexive pronoun and its substitute, Aramaic dialect, wrong vocalization of the Aramaic, confusion of some similar Aramaic spellings, slight corruption of the Aramaic text, aleph inserted or omitted, and waw or yodh inserted or omitted.

Linguistic Unidirectionality from Aramaic into Greek Tradition

41

and Greek recensions were in circulation.”53 And when the Jesus traditions including the Gospel of Mark and Q were handed over to the great cities of the neighboring lands where Greek was used as their primary language, they were translated into Greek (1912:296-7). He (1936:li) intriguingly argued that “It must be in both languages, Aramaic and Greek, for both were used in all the Diaspora.” Consequently, he (1912:297) maintained that even though the Aramaic documents originally were translated into Greek, an original Aramaic recension was used until the first decades of the second century. His arguments have left three major points of debate. The first point relates to two languages (i.e. Aramaic and Greek). Concerning the Aramaic language, he reconstructs his Aramaic gospels on his assumption that literary Aramaic was uniform for centuries.54 However, he has been criticized for this. The Aramaic sources Torrey used were anachronistic because he used the Aramaic of the Targums of Onkelos and the Prophets.55 In the case of Greek, Torrey argued that the syntax of the Four Gospels and the first half of Acts was translated into Greek that was full of mistranslations.56 Black (1967:5) criticizes Torrey’s new translation, saying, “Most of his examples of mistranslation, however, and several of Burney’s are open to grave objection.”

53

54

55

56

Concerning Acts 1-15, he (1912:296) added, “The document which he [Luke] translated appears to have belonged to that earliest stratum of Palestinian Christian literature which was written and circulated in the Aramaic tongue.” Torrey (1933:249-50) wrote that “the mother-tongue of Jesus of Nazareth, of the Galilean villagers, of the people of Jerusalem and Judea, in short, of all Palestine, was a fairly homogeneous dialect of Western Aramaic. This had been the language of the land for centuries, and it continued to be the vernacular until long after the apostolic age. The speech of the common people of Galilee differed in some noticeable respects from the Judean dialect; the literary language however was uniform, not only throughout Palestine, but also in the Jewish Dispersion.” Black (1967:5) argues, “Both Burney and Torrey approach the study of the Aramaic of Jesus on the same linguistic assumptions as Dalman, that the Aramaic of the Targums of Onkelos and the Prophets is the best representative of Aramaic of Jesus.” Fitzmyer (1998:82) also criticizes Torrey’s source, saying that “This kind of Aramaic reveals part of the problematic character of Torrey’s thesis, because he had to rely not only on earlier Biblical Aramaic, but mostly on that of the later targums and rabbinic writings.” As a matter fact, a considerable number of Aramaic documents in Palestine in the first-century BC and AD were unearthed after Torrey’s publications (Fitzmyer 1979:39). Although he knew about the new arguments of New Testament Greek provoked by the excavation of Egyptian Greek papyri, he disagreed with Deissmann and Moulton’s views since he thinks Semitisms occur throughout the four gospels; see Torrey 1912:270-5. For detailed discussion of syntax among biblical scholars, see §6.1.

42

The Directionality of the Transmission of the Jesus and Gospel Traditions

The second frequent criticism is that Aramaic written Gospels did not exist because there was no Aramaic literature around during the first century. From the perspective of form criticism57, some scholars (e.g. Goodspeed and Olmstead) asserted with confidence that there were no Aramaic written materials in first-century Palestine so that Aramaic written Gospels could not exist.58 However, not until the past decade have Aramaic texts been excavated from the Qumran caves (cf. Fitzmyer 1979a:39). As a matter of fact, the Qumran texts provide us with voluminous Aramaic documents from the first-century BC and AD, which supports Torrey’s insightful proposal. Therefore, it is highly possible that Aramaic written Gospels were circulated in that time period. Third, his intuitive argument concerning the bilingual circulation of the four gospels in Palestine and all the Diaspora (1912:296-7; 1936:li) warrants highlighting. Nevertheless, it is regrettable that his bilingual circulation hypothesis was incomplete and inconsistent since he still held the linguistic unidirectionality hypothesis. If he was privy to the bilingual circulation hypothesis, he would have better recognized another possibility, namely, that some Greek traditions translated from Aramaic might have been translated into Aramaic again among Aramaic-matrix speakers in Roman Palestine and the Roman Near East (cf. Torrey 1936:li). Also, some Aramaic traditions retranslated from Greek might have been retranslated into Greek again in an interdirectional way. This implies that Aramaic traditions were not always more “original” or necessarily earlier than Greek traditions because Aramaic traditions were circulated with Greek traditions together. Moreover, we cannot make a distinction between an Aramaic tradition retranslated from Greek and an Aramaic tradition “original” to Jesus. Torrey also maintained modal unidirectionality. He (1936:xlv) mentioned that “the time [of oral circulation] was too short; there must also be teaching through the circulation of written material.” However, the fact that the 57 58

Most of the early form critics assumed that there were no Aramaic written materials. For their arguments and my criticism, see §1.2. Olmstead (1942:41-75; here 47) declared that “there never could have been any Aramaic literature“ [emphasis in the original]. Also Goodspeed (1942:315-40; here 331; cf. 1937:127-68) asserted that there was no “creative literary activity in written classical Aramaic by the year 50 after Christ.” Furthermore, he (1942:315) confidently added, “I know of no work written in Palestine between A.D. 1 and 50 – in Greek or in any other tongue.” The two critics seemed to prefer taking the position of form criticism to that of the Aramaic school, as Olmstead (1942:46) wrote, “New Testament critics … invoke Form Criticism to explain the long stretch of oral transmission demanded by their hypothesis that the Gospels were first written in Greek“ [emphasis in the original].

Linguistic Unidirectionality from Aramaic into Greek Tradition

43

primary communicative vehicle was oral communication supports the modal interdirectionality between oral and written tradition, mentioned before (§1.3). Consequently, Torrey’s biggest drawback was that he speculated that the linguistic and the modal transmission of the Jesus and gospel traditions were unidirectional from Aramaic into Greek and from oral into written.

1.4.1.2 Matthew Black Like the form critics, Black ascribes the Semitisms of the Gospels to the linguistic gap between the Aramaic that Jesus spoke and the gospels written in Greek. He (1965:23) admits the bilingual Jewish Greek Theory (§6.1.3) and asserts, “Biblical Greek is a peculiar language, the language of a peculiar people.” For this reason, he (1967:16) tries to find “a Palestinian Aramaic tradition” in the Greek gospels. He (1965:21) posits that “the bulk of Semitisms are translation phenomena … in the process of translating and paraphrasing the verba ipsissima of Jesus.”59 Consequently, he (1967:275) thinks that the Greek gospels were translated from an Aramaic original as “the end-product in Greek” in a unilinear, teleological, and unidirectional way. Furthermore, conceding Wensinck’s argument,60 Black seeks to connect Semitisms with textual problems. He (1967:277) assumes that Bezan texts are “more primitive type of texts in earliest circulation” and nearer to the underlying Aramaic tradition since they are “more frequently stained with Aramaic constructions and idiom than the Ba text.” He even adds some Syriacisms to Aramaisms. He says that “many of the alleged ‘Syriacisms’ in D … may be Aramaisms and come from the Aramaic sources and back-

59

60

Black (1967:7) is critical of Torrey’s emphasis on mistranslation of Aramaic tradition, complaining that “Both Torrey and Burney attach much importance to conjectural mistranslations of Aramaic as proof of source.” Wensinck (1937:47) argues that the Bezan text precedes the non-Western group because of the frequency of Aramaisms. However, criticizing Wensinck’s argument, Yoder (1959:317) persuasively suggests two points; (i) “when one takes into account not only the instances of Semitic phenomena in Codex Bezae, but also the Bezan variants which abandon Semitisms found in other MSS, the net increase of Semitisms is sometimes inconsequential, while in other respects this MS actually reveals fewer Semitisms than found in the B Aleph text.” (ii) “Ofttimes the data are concentrated in limited areas of the text, thus detracting from the supposed homogeneity of the Bezan text.” Klijn (1959:161) also proposes, “The question, however, is whether we are dealing with Aramaisms or syriasms. As long as the relation between D–old syr–old lat–diat has not been solved it is rather hazardous to base too many conclusions on semitisms in D.”

44

The Directionality of the Transmission of the Jesus and Gospel Traditions

ground of the Gospels.”61 This shows his presupposition that the more Semitized, the earlier. Accordingly, his arguments concerning the Jesus tradition and textual criticisms are based on the linguistic unidirectionality hypothesis that Semitisms represent earlier traditions and earlier texts. Three criticisms against Black’s arguments can be made. First of all, rejecting Dalman and Torrey’s source materials (i.e. Targum Onkelos, Targum Pseudo-Jonathan, and biblical Aramaic) Black (1967:22) considers the Aramaic of the Samaritan Targum and the Palestinian Pentateuch Targum and Christian Palestinian Aramaic to be the most reliable sources. He (1967:40) posits that Qumran Aramaic is similar to “old Aramaic.” However, reviewing Black’s publication, Fitzmyer (1968:420) convincingly argues that “this Qumran Aramaic … must be the latest Aramaic that should be used for philological comparisons of the Aramaic substratum of the Gospels and Acts.” Furthermore, Fitzmyer (1968:421; cf. 1968:417-28; Stuckenbruck 1991:3-29) insists that “the third edition of B.’s [Black’s] book no longer copes with the present-day problem of the Aramaic substratum of the Gospels and Acts.” Second, he makes a sharp linguistic distinction on the basis of the unidirectional transmission hypothesis of the Jesus and gospel traditions, as form critics have maintained. He (1967:16) contrasts Palestinian Aramaic tradition with Greek Gospels written in a predominantly Hellenistic environment. The linguistic transmission is described as an “end-product in Greek,” which is unilinear and teleological. However, the linguistic milieu of both first-century Palestine and the Roman Near East should be considered as bilingualism rather than monolingualism, as will be discussed (chapters 2-5). Bilingualism dilutes his sharp linguistic distinction between Aramaic tradition and Greek tradition. The linguistic interdirectionality hypothesis suggests that Greek tradition was translated into Aramaic among Aramaic-matrix speakers. Consequently, Greek traditions or Greek Gospels are not the “end-product.” Lastly, Black assumes that Aramaisms take temporal precedence so that Aramaic traditions and Bezan texts including Aramaisms should be considered to be earlier than Greek traditions or Western texts. However, when we take the bilingual approach to the linguistic transmission of the Jesus and gospel traditions, an Aramaic tradition is not always earlier than Greek tradition, as mentioned before (§1.4.1.1). This 61

He (1967:33) adds, “Syriac and West Aramaic can be clearly distinguished as different dialects of Aramaic, but they have also so much in common (the language is, after all, Aramaic) that what may be explained in Greek texts, where dialectical distinctions of Aramaic cannot always be detected, as a Syriacism may in fact prove to be an Aramaism.”

Linguistic Unidirectionality from Aramaic into Greek Tradition

45

implies that the frequency of Aramaisms does not guarantee the priority of the Bezan texts over the Western texts.

1.4.1.3 Joachim Jeremias Jeremias’ proposal warrants attention. First of all, he (1971:1) assumes that the Jesus tradition must have been changed in the course of transmission due to two reasons, oral tradition’s variability and translation from Aramaic into Greek. (i) He (1971:1) presumes that “it was inevitable that during this long period [more than thirty years] of oral transmission alterations took place in the tradition.” (ii) He (1971:1) submits that “by that time [after more than thirty years] his sayings had long been translated into Greek.” He seems to pay much attention to the latter. As for Semitisms caused by translation, he (1971:3) states that “The sayings of Jesus handed down by the synoptic gospels are clad in the garb of a koine Greek with a number of Semitic characteristics.” Dealing with the eucharistic words of Jesus, he (1966:173) claims that “We take a further step in the quest for the oldest tradition when we give attention to semitisms and to Palestinian idioms.” In this respect, he (1971:3-29) asserts that Aramaic original tradition can be traced, in case they are preserved in two types. One is the logia of Jesus62 and the other is Jesus’ manner of speaking in the Gospels. 63 He is also concerned with characteristics of ipsissima vox (i.e. parables of Jesus, riddles, the reign of God, amen, and abba). 64 Furthermore, he (1971:1; 1972:25) argues that translation from Aramaic into Greek caused Greek variants. He (1972:25) considers that retranslation into Palestinian Aramaic is “the most important aid to the recovery of their original meaning.” 65 Jeremias (1971:8) confidently concludes, “This linguistic

62

63 64

65

Jeremias (1971:5-6) cites as illustrations taliqa koum, hli hli lema sabacqani, Effaqa, Aramaic place-names, personal names, designations of descent or of groups, and etc. However, proper nouns and borrowings should not be regarded as Semitisms except in some cases where the authors’ intentions are to be shown (§8.4.2; §8.4.3). For the characteristics of borrowings, see §8.1.1. He (1971:8-29) illustrates divine passive, antithetic parallelism, rhythm, alliteration, assonance and paronomasia. Jeremias (1971:37) is cautious to distinguish ipsissima vox Jesu from ipsissima verba Jesu in that each individual instance of ipsissima vox is not always ipsissima verba. The Aramaic embedded words as ipsissima verba will be dealt in detail (chapter 8). He (1972:26) also suggests that “we possess an additional means of controlling the process of retranslation – the impression created by all this must be that there is now a greatly increased probability of getting back to the original text by the aid of translation variants.”

46

The Directionality of the Transmission of the Jesus and Gospel Traditions

evidence [Semitisms] takes us back into the realm of Aramaic oral tradition” [emphasis added]. Accordingly, Jeremias’ arguments (cf. 1972:236) are based on the Sitz im Leben, the modal, and the linguistic unidirectionality hypotheses. However, Jeremias does not seriously consider the bilingual situations of first-century Palestine and the Roman Near East.66 This bilingualism implies that Aramaic traditions were circulated with Greek traditions during and after Jesus’ ministry. This means that Aramaic traditions that retain Aramaisms do not always refer to traditions earlier than Greek traditions. Moreover, Jesus who was bilingual in Aramaic and Greek, could speak Greek when he conversed with Greekmatrix speakers. This means that Aramaic tradition does not always lie behind all Greek gospel tradition, as proponents of the Greek approach argued (§1.4.2). Second, Jeremias also holds the modal unidirectionality hypothesis. He assumes that oral tradition is variable and that written tradition is the end product of oral tradition. However, as the Scandinavian school (§1.3.1) rightly points out, oral tradition itself is not always variable. Third, Jeremias believes that Aramaic tradition was original and that variants in the Greek Gospels might have been caused by translation from Aramaic into Greek. He (1972:25) tries to make two different Greek versions “one and the same Aramaic tradition”67 by the process of his retranslation. However, it is hard to say that retranslation leads to the original words in Aramaic.68 Accordingly, Jeremias should have considered the modal and the linguistic interdirectionality hypotheses that make clear that the transmission of the Jesus and gospel traditions could have been hybrid, circular, and interdirectional.

66

67

68

It is regrettable that although Jeremias (1969) dealt with various social aspects in Jerusalem in the Time of Jesus in detail (i.e. economic condition, economic status, social status, and maintenance of racial purity), he neglected a linguistic study of Jerusalem. In terms of Luke 14:8-10 and Matthew 20:28 D it Syc, he supposes that “The parable of the Places at the Banquet may serve as an example showing how two widely differing Greek versions lead us back to one and the same Aramaic tradition.” Riddle (1932:17-21) persuasively points out that the failure of objectivity among Semitists (e.g. Burney, Torrey, and Burrows) appears when their lists of alleged mistranslated Semitisms are dissimilar to each other. He (1932:23) clearly demonstrates that, “The two versions of Daniel, in their differences from one another, further illustrate the point that the Semitic cannot be predicted from the Greek.” In terms of Aramaic retranslations, Grant (1963:41) remarks that “experts in Aramaic have a tendency to disagree as to what the original was.” It is also worth mentioning Hurst’s criticism (1986:63-80) against a theory seeking underlying Aramaic in Jesus’s sayings. Nevertheless, it is unfortunate that they are linguistic unidirectionalists.

Linguistic Unidirectionality from Aramaic into Greek Tradition

47

1.4.1.4 Joseph Fitzmyer69 On the basis of a maximal view of the use of Greek in first-century Palestine, 70 Fitzmyer’s moderate position 71 has been welcomed widely among biblical scholars.72 Despite his view of bilingualism, he seems to adhere to the unidirectionality hypothesis when it comes to the linguistic transmission of Jesus and gospel traditions. First of all, Fitzmyer (1992:63) suggests three stages in the transmission of the Jesus tradition. This has been widely accepted, but is based on linguistic unidirectionality. The first stage is what Jesus said and did during his ministry (1-33 AD). Second, his disciples and apostles taught and proclaimed sayings by Jesus and stories about Jesus (33-66 AD). Last, the evangelists selected homilies and instructions and redacted them according to their purpose and style (66-95 AD). He warns that some scholars confused “the Greek of stage III” with “Aramaic stage I.” This implies that he makes a sharp distinction between the Aramaic tradition stage and the Greek gospel stage, as made popular by form criticism. Second, his argument of the linguistic gap leads to the issues of Semitisms and Septuagintalisms. Fitzmyer tries to reduce the number of alleged Semitisms and accordingly proposes three other possibilities. (i) In relation to Lukan improvement of Markan Greek, he (1981-5:107) suggests that “Part of the improvement is the result of the use of fewer Semitisms, and part of it the use of more resources of the Greek lan69

70

71

72

Although Fitzmyer admits bilingualism in first-century Palestine and proposes balanced views, he denies ipsissima verba Jesu graeca. On this account, I tentatively put him into the group that appeals to “Aramaic language as a criterion.” Concerning the use of Greek in first-century Palestine, Fitzmyer (1992:59-60) supposes that “A good number of these indicate that Greek was used for public announcements. Others found on ossuaries--inscribed in Greek and Hebrew (or Aramaic), or in Greek alone--from the vicinity of Jerusalem testify to the widespread use of Greek among first-century Palestinian Jews at all levels of society.” For a fulldressed discussion of maximal bilingualism in first-century Palestine, see §2.3 & chapter 3. Fitzmyer (1979a:29-56; here 46) describes the linguistic milieu of first-century Palestine as quadrilingualism, which is well-balanced: “The most commonly used language of Palestine in the first century A.D. was Aramaic, but that many Palestinian Jews, not only those in Hellenistic towns, but farmers and craftsmen of less obviously Hellenized areas used Greek, at least as a second language. … But pockets of Palestinian Jews also used Hebrew, even though its use was not widespread. … Latin was really a negligible factor in the language-situation of first-century Palestine, since it was confined for the most part to the Roman occupiers.” He makes a huge contribution to the study of the Aramaic language of first-century Palestine. He opposes the view of the Kahlean School, including Black (cf. Kahle 1959; Black & Fohrer 1968), and contends that Qumran Aramaic is Jesus’ Aramaic; see Fitzmyer 1974; 1980; followed by Stuckenbruck 1991.

48

The Directionality of the Transmission of the Jesus and Gospel Traditions

guage.” This implies that he is also of the opinion that Semitic traditions are earlier than Greek traditions. (ii) Interestingly, Fitzmyer sometimes ascribes Lukan Aramaisms to Luke’s personal bilingualism in Syriac and Greek, not to more original Semitic sources. He (1981-5:116) says that “the source of such interference [Aramaisms] could be Luke’s origin in Syrian Antioch, where he lived as an incola, speaking the Aramaic dialect of the indigenous natives of that country,” although he accepts that Luke “was also educated in the good Hellenistic culture of that town.” Given this, it seems that he should have called his Lukan Semitisms Syriacisms instead of Aramaisms because of the abuse of the term “Semitism”, as he pointed out.73 (iii) He (1981:107-25) thinks of some alleged Aramaisms as Septuagintalisms which mean the wider and indirect interference of Semitisms on the gospels through the Septuagint.74 He (1998:116) confidently concludes, “Most of the instances cited as alleged Aramaisms are to be explained more correctly as Septuagintalisms.”75 Accordingly, although Fitzmyer accepts Semitic interference at the lexical and syntactic levels, he does not agree that the alleged Semitisms do not always guarantee more original sayings by Jesus and stories about Jesus because the usages appeared in the Septuagint. He (1998:91) is of the opinion that the “fact that they are drawn from the LXX in most instances reveals that they cannot be ascribed to a source that would have been written in Aramaic.” Consequently, in spite of his bilingual view of first-century Palestine, he also maintains

73

74

75

He (1979:5) mentioned: “the way in which claims are sometimes made for the Aramaic substratum of the sayings of Jesus, when the evidence is merely ‘Semitic’ in general, or, worse still, derived from some other Semitic language, e.g., Hebrew, should no longer be countenanced. In other words, the discussion of the Aramaic background of the NT should be limited to Aramaic evidence, and to Aramaic evidence of the period contemporary with or slightly prior to the composition of the Greek New Testament writings themselves.” Fitzmyer (1981-5:114) suggests that “the Semitisms of Lucan Greek which are found in the LXX should be frankly labeled as ‘Septuagintalisms,’ and only those that are not should be sorted out as true Aramaisms or Hebraisms.” Moreover, he (1981:125) also posits that “Whatever one wants to say about the alleged Semitisms in Luke’s Greek, one has in the long run to reckon with a great deal of influence from the LXX.” He (1979:40) admits Semitic interference. In the case of verbs, he (1981:114-25) regards avpokriqei.j ei=pen, poreuqe,ntej avpaggei,late, and avpekri,nato le,gwn as Septuagintalisms and h;rxato le,gein as an Aramaism because the former expressions occur in the Septuagint, whereas the latter style does not. Furthermore, to the list of Lukan Septuagintalisms he (1998:115) adds some instances of Acts such as avnasta,j ei=pen, avpokriqe,ntej ei=pon, and hvrw,twn auvto.n le,gontej. However, these verbs will be considered as grammatical polysemies due to internal-induced syntactic change (chapter 6).

Linguistic Unidirectionality from Aramaic into Greek Tradition

49

the linguistic unidirectionality hypothesis that Aramaic traditions were earlier than Greek traditions. Third, he takes for granted that the issue of ipsissima verba Jesu graeca is one of the most significant factors. He (1979a:10) insists upon a caveat, since the argument “affects the whole question of the origin of the saying-tradition.”76 Admitting that Jesus spoke Greek with Greek speakers, 77 he (1979a:10, 37) denies ipsissima verba Jesu graeca. 78 He (1979a:10) bluntly affirms that “I am extremely skeptical about the sayings-tradition in the New Testament being rooted, even in part, in Greek sayings of Jesus himself.” 79 Accordingly, he believes that the linguistic transmission of the Jesus and gospel traditions is unilinear and unidirectional from Aramaic into Greek, even though Jesus himself spoke Greek in some cases. Despite Fitzmyer’s tremendous contributions to our understanding of the languages of New Testament literature and first-century Palestine, three drawbacks in relation to directionality of transmission of the Jesus and gospel traditions should be mentioned. First, his view of contact-induced syntactic change of New Testament Greek (i.e. Semitisms or Septuagintalisms) seems to exclude the possibility of internalinduced syntactic change (see n.75; cf. chapter 6). Second, although Fitzmyer is attentive to the bilingualism of first-century Palestine, his view of linguistic transmission is still unilinear and unidirectional from Aramaic into Greek. If he grants the maximal view of the bilingualism of first-century Palestine (i.e. both lower-status and rural bilingualism; §2.3 & chapter 3), his view would be more persuasive if he also had considered the other three interdirectional possibilities of the linguistic transmission of the Jesus and gospel traditions. (i) As “an extensive Aramaic literary activity and an Aramaic literature” indicate (Fitzmyer 1979:39), sayings by Jesus and stories about Jesus were circulated in 76

77

78

79

Further, he (1979a:10) mentions that “The caveat is derived ultimately from other aspects of modern Gospel study, for no discussion of the Aramaic problem of the Gospels can prescind from the issues raised by Synoptic source criticism, form criticism, and redaction criticism” [emphasis in the original]. Fitzmyer (1992:60-1) cites as illustrations Pilate (Mk 15:2-5; Mt 27:11-14; Lk 23:3; Jn 18:33-8), a Roman centurion (Mt 8:5-13; Lk 7:2-10; Jn 4:46-53), a Syro-Phoenician woman (Mk 7:25-30), and Greeks (Jn 12:20-22, 7:35). He (1979a:10) also remarks that “The possibility of it is always there, but when other factors are taken into consideration, it seems rather unlikely.” For a detailed discussion of ipsissima verba Iesu graeca, see §1.4.2. He (1979a:37) suggests that “if it be used to insist that we might even have in the Gospels some of the ipsissima verba Iesu graeca, actually uttered by him as he addressed his bilingual Galilean compatriots, then the evidence is being pressed beyond legitimate bounds.”

50

The Directionality of the Transmission of the Jesus and Gospel Traditions

Greek among Greek-matrix speakers (e.g. individuals, families, or communities) as well as in Aramaic among Aramaic-matrix speakers during and after Jesus’ ministry. (ii) The Aramaic Jesus tradition could have been translated into Greek among Greek-matrix speakers; at the same time, the Greek Jesus tradition could have been translated into Aramaic among Aramaic-matrix speakers. (iii) As the Greek tradition which had been translated from Aramaic could have been retranslated into Aramaic, so the Aramaic tradition which had been translated from Greek could have been retranslated into Greek. Third, the bilingualism of the first-century Roman Near East (cf. chapter 4) also implies the linguistic interdirectional possibility that the synoptic gospels could have been translated into other languages, as Syriac versions indicate. Despite his maximal view of bilingualism, his argument is still based on the unidirectionality hypothesis. This implies that the linguistic unidirectionality hypothesis spread among New Testament scholars. Accordingly, the proponents of the Aramaic language as a criterion should have taken into consideration more seriously the bilingualism of first-century Palestine and the Roman Near East. The bilingualism of these regions lends support to the notion that linguistic transmission is not unilinear, teleological, and unidirectional but hybrid, circular, and interdirectional.

1.4.2 Greek Language as a Criterion The study of the bilingualism of first-century Palestine has promoted another possible view. It causes some scholars to pay attention to a “different original language” in relation to the authenticity of sayings by Jesus and stories about Jesus. Proponents have argued that if Jesus was bilingual and the linguistic milieu of first-century Palestine was bilingualism in Aramaic and Greek, the original language of some Jesus traditions was Greek. As a consequence of this, they propose that one of the criteria of authenticity (i.e. Semitisms) should be reexamined. Nevertheless, they also hold the unilinear, teleological, and unidirectional view that the Greek tradition Jesus spoke is the last stage. The bilingualism of first-century Palestine leads to the linguistic interdirectionality hypothesis. As the Aramaic Jesus tradition was translated into Greek, so the Greek Jesus tradition would likely have been translated into Aramaic simultaneously. In relation to Greek language as a criterion, I will survey five specialists: Alexander Roberts (§1.4.2.1), Aubrey Argyle (§1.4.2.2), Nigel Turner (§1.4.2.3), Robert Gundry (§1.4.2.4), and Stanley Porter (§1.4.2.5).

Linguistic Unidirectionality from Aramaic into Greek Tradition

51

1.4.2.1 Alexander Roberts Accepting the bilingualism of first-century Palestine, Roberts (1888) left behind a hefty book dealing with the language of Jesus and his disciples. He (1888:105, 400) posited that although Aramaic was the domestic language, Jesus and his apostles knew Greek as much as their contemporaries in Judaeo-Palestine. In this way, he seemed to assume that Jesus and his contemporaries were nearly balanced bilinguals in Aramaic and Greek (cf. §2.1.8). With regard to Jesus’ audience, he (1888:144) supposed that Greek was what a public teacher typically used and what audiences habitually expected. Roberts (1888:145-73) demonstrated six special proofs from the gospels that Greek was the prevailing language of first-century JudaeoPalestine, which have, to date, recurred in arguments along the same lines. (i) In the case of the Sermon of the Mount, the crowd who came from Galilee, the Decapolis, Jerusalem and Judea and from beyond the Jordan (Mt 4:25) spoke Greek. (ii) As regards the conversation of Christ with the woman of Samaria, the residents of Samaria spoke Greek. (iii) The Greeks who asked to see Jesus (Jn 12:20-21) spoke Greek and Philip, Andrew, and Peter must have spoken Greek. (iv) The situation of Pilate’s intercourse with Christ (Mt 27:11-14) implies that Pilate spoke Greek to Jesus without interpreters. Also the picture implies that Pilate understands what the people of the Jewish crowd cry and has a conversation with the Jewish multitude without interpreters (Mk 15:8; Jn 18:38-40, 19:4-7; cf. Lk 23:13-14). (v) The incident connected with the crucifixion demonstrates that Jews around the cross confused the cry on the cross with the calling of Elias (Mt 27:47-49; Mk 15:35-36). This implies that the Jews did not know Hebrew or Aramaic but were familiar with Greek. (vi) In the conversation between Jesus and Mary Magdalene (Jn 20:16), Jesus probably called Mary Mariam (Hebrew name) because she called Jesus Rabboni. Special attention, in this regard, should be paid to the point that he applied the concept of codeswitching to the interpretation of these two Semitic names (for codeswitching, see §8.1.2). Roberts (1888:171) noted that “We see at once a beauty and significance in the employment and preservation of these Hebrew terms if the rest of the conversation was in Greek.” He continued, “If it be supposed that the language used by Christ and Mary throughout was Hebrew, the meaning of these isolated expressions being retained in that tongue entirely disappears.”80 Based on the six episodes from

80

This concept of intentional language change (i.e. codeswitching) is also found in many scholars’ publications such as Abbott, Bultmann, and Birkeland (§8.1.2).

52

The Directionality of the Transmission of the Jesus and Gospel Traditions

the gospels, Roberts insisted that Jesus spoke Greek because his conversational partners spoke Greek. In relation to Aramaic tradition Roberts assumed that there were no Aramaic written materials since there were no Aramaic written versions in first-century Palestine. He (1888:143-4) raised a question: “Is it granted that neither could a written Aramaic version have constituted these Scriptures, since there is no reason to believe that any such version ever existed?” This implies that the sayings by Jesus and the stories about Jesus have been preserved in Greek without translation from the start.81 Consequently, with regard to the transmission of the Jesus tradition, he (1888:480-1) proposed that the gospels in Greek reflect ipsissima verba Jesu graeca as a whole except for some Semitic embedded words used as codeswitchings. However, and above all, it has been widely assumed that Jesus was a dominant bilingual (§2.1.8) and that Jesus and his contemporaries usually spoke Aramaic as their matrix language and Greek as their embedded language (chapter 3). The extreme view of Jesus’ matrix language as Greek has not been supported among recent scholars. Second, although it is of significance to introduce the concept of codeswitching (e.g. Hebrew words), Aramaic embedded words employed in the Gospels (i.e. codeswitchings) imply that Jesus spoke Aramaic as his matrix language because they are everyday expressions (cf. §8.3.2). Last, the biggest drawback is related to transmission of the Jesus tradition. If the linguistic milieu of the first-century Palestine was bilingualism, the Greek Jesus tradition was most likely translated into Aramaic during Jesus’ historical ministry and continuing thereafter. It is hard to distinguish ipsissima verba Jesu graeca from the gospel tradition, no mat81

He (1888:483-4) even mentioned that “instead of having to study a comparatively poor and unattractive language like the Syriac, in order to have the satisfaction of becoming acquainted with something like the expressions employed by our Redeemer, we have only to open our Greek New Testament to find still preserved to us, in living reality, the very words which issued from His lips.” He (1888:485-6), furthermore, alluded that if Jesus spoke Aramaic, he could not have made a distinction between file,w (Jn 11:3, 5, 21:15-17) and avgapa,w and between pe,troj and pe,tra. However, admittedly, the linguistic view of “a comparatively poor and unattractive language like the Syriac” has already been discarded by de Saussure (1959; originally 1916) of General Linguistics. Following Barr (1961:216-7), Carson (1996:30) may argue that “although it is doubtless true that the entire range of avgapa,w (to love) and the entire range of file,w (to love) are not exactly the same, nevertheless they enjoy substantial overlap; and where they overlap, appeal to a ‘root meaning’ in order to discern a difference is fallacious.” Carson gives some instances; avgapa,w refers to Amnon’s incestuous rape of his half-sister Tamar (u`pe.r th.n avga,phn h]n hvga,phsen auvth,n 2 Sam 13:15 LXX); Demas loved (avgaph,saj) this present world (2 Tim 4:10); the Father loved (filei/) the Son (Jn 5:20).

Linguistic Unidirectionality from Aramaic into Greek Tradition

53

ter which language Jesus spoke, since the bilingual context(s) of firstcentury Palestine implies that linguistic transmission was hybrid, circular, and interdirectional during Jesus’ ministry. 82 This drawback has been repeated in the arguments by the scholars who hold the Greek language as a criterion for the authenticity of the Jesus and gospel traditions.

1.4.2.2 Aubrey Argyle Unlike Roberts’s minority view of Greek as Jesus’ matrix language, Argyle (1955-56:92) contends that Jesus, who was reared in Galilee, spoke Greek as his embedded language. 83 His Greek competence is adduced by his quotations from the Septuagint (Mt 4:1-11, Lk 4:1-13, and Mt 5:39-40 / Isa 50:6-8), his teaching (iota at Mt 5:18), and conversation with Greek speakers (Mt 4:25, Mk 3:7-8, 7:26, Lk 6:17, Jn 7:35, 12:21; cf. Chapter 3 n.70). At least five out of his twelve disciples (e.g. Peter, Andrew, James, John, and Levi) could speak Greek, too.84 He (195556:93) insists that “He and His disciples may often have conversed in Greek.” Concerning linguistic transmission, he (1955-56:93) proposes two major points. As for ipsissima verba Jesu graeca, he says, “If in places we are reading the language He spoke, it is possible that we are reading His actual words.” Moreover, he relates ipsissima verba Jesu graeca to a Greek criterion for authenticating the original sayings by Jesus, saying that “it is a mistake to assume that there must be Aramaic behind all His words” [emphasis in the original]. In this sense, he (1955-56:93) suggests, “In some cases we may have direct access to the original utterances of our Lord.” However, it would have been more plausible if he had been guided by the probability that the Greek tradition Jesus spoke could have been translated into Aramaic since the linguistic milieu of first-century Palestine was bilingual, as he concedes. Consequently, it is pointless to seek to single out ipsissima verba, no matter

82

83 84

Fitzmyer (1992:63) convincingly suggests that “there is no way to sort out what he might have taught in Greek from what we have inherited in the Greek tradition of the Gospels.” For his view of New Testament Greek as Hellenistic Greek with some Semitisms and Septuagintalisms, see chapter 6 n.40. He (1955-56:93) suggests that many inhabitants of Galilee would speak Greek. In order that the four disciples could sell their fish in Gentile markets and that Levi could be engaged in government employ, they needed to speak Greek (cf. §3.2.4; §3.2.6).

54

The Directionality of the Transmission of the Jesus and Gospel Traditions

what language Jesus spoke. The ipsissima verba Jesu graeca simply cannot be used as a criterion for authenticity of the historical Jesus.85

1.4.2.3 Nigel Turner Turner (1965) contends that the linguistic milieu of first-century Palestine was quadrilingual. Quadrilingualism implies that the original sayings by Jesus were multilingual and originated from four different languages (1965:176-85). According to Turner, Jesus spoke Aramaic as his first language and normally conversed with his disciples in Aramaic. Some original sayings by Jesus were delivered in biblical Greek (§6.1.3) when he met Greek-speakers (e.g. Syro-phoenician woman, the Roman centurion and the procurator Pontius Pilate in the synoptic gospels and Nicodemus in John; also see 1965:181-3). Jesus at least knew Hebrew; he read Hebrew in the synagogue and spoke Hebrew as codeswitchings perhaps on solemn occasions (see chapter 8) such as his arguments with the Pharisees and the Last Supper (1965:182). Lastly, Turner suggests that Jesus also spoke Latin, as is shown by slight traces in his sayings (Mk 9:50). However, if Jesus spoke Greek and “Greek was probably the regular language of Palestine (Turner 1965:183),” he should also have considered interdirectional possibility of the linguistic transmission of the Jesus tradition. Although Turner concerned himself only with “multilingual original sayings by Jesus,” multilingual transmission should have been reflected as well.

1.4.2.4 Robert Gundry Gundry (1964:404-8) seeks to relate the trilingualism (i.e. Aramaic, Greek, and Hebrew) of first-century Palestine to ipsissima verba of Jesus and transmission of the Jesus tradition in more detail. According to him (1964:405-7), archaeological data support the view of trilingualism in first-century Palestine, that is, all three languages were commonly used by Palestinian Jews, especially Galileans. He (1964:407) notes that “Je85

Interestingly, Ross (1990:41-7) argues that Jesus spoke Greek and Aramaic in his teachings. Some constructions Jesus spoke in Greek (evpiou,sion and o` ui`o.j tou/ avnqrw,pou) were preserved because Jesus chose the words. However, he also maintains the linguistic unidirectionality of the Jesus tradition. If the linguistic milieu of first-century Palestine was bilingual, as he holds, the Greek tradition Jesus spoke must have been translated into Aramaic, and the Aramaic tradition which had been translated from Greek tradition must have been circulated as well. This means that to single out ipsissima verba Jesu graeca is meaningless.

Linguistic Unidirectionality from Aramaic into Greek Tradition

55

sus the Galilean and the apostles, who were predominantly if not exclusively Galilean, commonly used Greek in addition to the Semitic tongues.” In regard to sayings by Jesus, he (1964:408) argues that “Many may, in fact, be identical with dominical sayings originally spoken in Greek.” This means that “many of the dominical sayings in the present Greek text of the gospels may be closer to the ipsissima verba of Jesus than has been supposed” (1964:408). Going a step further, Gundry (1964:408) intriguingly proposes that the Jesus tradition was transmitted in the three languages in Palestine: “The tradition about Jesus was expressed from the very first in Hebrew, Aramaic and Greek ” [emphasis in the original]. Consequently, he (1964:404; 408) persuasively concludes that “The absence of Aramaisms (or more broadly, Semitisms) does not militate against authenticity” and that “even the presence of Semitisms does not necessarily indicate an Aramaic (or Hebrew) substratum.” However, much like Torrey, his view of multilingual transmission is incomplete (§1.4.1.1). Gundry does not mention the possibility that the Greek Jesus tradition was translated into Aramaic. He seems to suppose that the synoptic gospels in Greek represented the final form of the process. In this respect, his view is still unidirectional. If the linguistic milieu of first-century Palestine was trilingual, even ipsissima verba Jesu graeca would most likely have been translated into Aramaic as well during Jesus’ ministry. And the Aramaic tradition and Aramaic Gospels translated from Greek could have been retranslated to Greek. Accordingly, the idea that trilingualism implies that Semitisms in Gospels could not be “an Aramaic (or Hebrew) substratum” is warranted. However, it is regrettable that he did not further develop his view; nor has any other scholar paid attention to this multilingual transmission in detail.

1.4.2.5 Stanley Porter Describing the linguistic milieu of first-century Palestine as multilingualism (i.e. Aramaic, Hebrew, Greek, and Latin), Porter (2000a:134) assumes that Jesus was bilingual, with Aramaic as his first language and Greek (or Hebrew) as his second language. Citing five criteria for the authenticity of the historical Jesus,86 Porter follows Argyle’s view (§1.4.2.2). He (2000a:89-100, 126-180) is skeptical about the Aramaic language as a criterion for the idea that Semitic tradition lies behind all 86

The five criteria are double dissimilarity, least distinctiveness, coherence or consistency, multiple attestation, and Semitic language phenomena.

56

The Directionality of the Transmission of the Jesus and Gospel Traditions

gospels in Greek. Instead, he (2000a:142-3) tries to add Greek language as a criterion for authenticity. In relation to Greek language as a criterion, he (2000a:143-63; summarized on 158) proposes that seven episodes87 might have been ipsissima verba Jesu graeca on the basis of three caveats: participants and their backgrounds, context and theme of discussion, and determination of the words of Jesus. He (2000a:159) assumes that “all seven of these episodes are seen to be historically plausible as authentic to the Jesus tradition, at least in so far as the conversation taking place in Greek is concerned.” He considers that the relationship between the preserved Greek traditions Jesus spoke and linguistic transmission of the Jesus tradition is of significance. He (2000a:164) proposes: The corrective value of this criterion for historical-Jesus research is that we should not reject any words or episode as inauthentic, or as the creation of a later Greek-speaking Church, simply because they appear to have been spoken in Greek or were spoken in a Greek-speaking environment, or were spoken to those who appear to have been themselves Greek-speaking.

The hypothesis of Greek language as a criterion might be used as a corrective method of Aramaic language as a criterion. 88 However, like Roberts, Porter is inconsistent. He supposes that Greek tradition is the last stage despite the fact that he takes the multilingual context into consideration. The assumed ipsissima verba Jesu graeca (a la Porter) were probably translated into Aramaic among Aramaic-matrix speakers in Roman Palestine and the Roman Near East. This is a better solution than the idea that they were preserved without translation into Aramaic 87

88

Porter uses seven passages as illustrations: (i) Mt 8:5-13 = Lk 7:1-10 = Jn 4:46-54: Jesus’ conversation with the centurion or commander (but the Johannine account diverges in terms of wording) (ii) Jn 4:4-26: Jesus’ conversation with a Samaritan woman (iii) Mk 2:13-14 = Mt 9:9 = Lk 5:27-28 Jesus’ calling of Levi/Matthew (iv) Mk 7:24-30 = Mt 15:21-28 Jesus’ conversion with the Syrophoenician or Cannaanite woman (v) Mk 12:13-17 = Mt 22:16-22 = Lk 20:20-26 Jesus’ conversation with the Pharisees and Herodians over the Roman coin of Caesar (vi) Mk 8:27-30 = Mt 16:1320 = Lk 9:18-21 Jesus’ conversation with his disciples at Caesarea Philippi (vii) Mk 15:2-5 = Mt 27:11-14 = Lk 23:2-4 = Jn 18:29-38 Jesus’ trial before Pilate. The number of Greek traditions spoken by Jesus increases; he has cited five more passages; (i) Mk 7:25-30 (ii) Jn 12:20-28 (iii) Mt 8:5-13 = Lk 7:2-10 (iv) Mk 15:2-5 = Mt 27:11-14 = Lk 23:2-5 = Jn 18:29-38 (1994:149-53; cf. 1997:111-2) and Lk 17:11-19 (2003:201-24). The argument that Jesus used Greek when he spoke to his Greek-speaking partners has already been discussed in detail by Roberts (§1.4.2.1). Reviewing Porter’s publication (2000a), Fitzmyer (2001:412) persuasively criticizes a limitation of the hypothesis of Greek language as a criterion and raises a question: “the limited number of sayings, a mere seven, to which the criterion of the Greek language can be applied leaves a huge question: What about all the others preserved in Greek in the four Gospels?”

Linguistic Unidirectionality from Aramaic into Greek Tradition

57

and retranslated into Greek until they were delivered to the evangelists. Bird (2006:60) rightly criticizes Porter’s linguistic unidirectionality, saying, “If Aramaic sayings were translated into Greek for Greekspeaking audiences then, by analogy, it remains possible that Greek sayings (if they existed) were translated into Aramaic for Aramaicspeaking audiences.” As Porter (2006:69-74) correctly refutes, Bird’s minimalistic view of Greek in first-century Palestine is ambivalent and problematic. However, Porter cannot explain how “only the seven Greek traditions” were preserved in Greek without translation into Aramaic and then retranslation into Greek, despite his maximalistic view of Greek in the first-century Palestine. It is uncertain whether the assumed seven Greek traditions were directly delivered from the very Greek tradition Jesus spoke (i.e. ipsissima verba Jesu graeca) or retranslated from Aramaic traditions which had already been translated from Greek. As a result, he (2000a:93-4) would be faced with the same criticism that he levels; namely that Aramaic tradition cannot be reconstructed from Greek tradition. This is so because “much of this examination has neglected a number of crucial linguistic factors regarding translation between Greek and Aramaic.” Hence, one cannot guarantee that the seven Greek traditions we have are ipsissima verba Jesu graeca when the bilingualism of first-century Palestine and the Roman Near East is taken into consideration seriously. Consequently, Porter also takes the unilinear, teleological, and unidirectional view of the linguistic transmission. Accordingly, in terms of the transmission of the Jesus and gospel traditions, the proponents of “Greek language as a criterion” have not been able to overcome the pitfalls of linguistic unidirectionality. In other words, their arguments are inconsistent because it is a teleological view that Aramaic tradition was translated into Greek, whereas Greek tradition was preserved in bilingual contexts. In sum, scholars have discussed the linguistic transmission of the Jesus and gospel traditions from the two perspectives of Aramaic language as a criterion and Greek language as a criterion. Unfortunately, the former (except Fitzmyer) is based on a form of monolingualism of first-century Palestine and the Roman Near East. Although the latter (including Fitzmyer) took bilingualism into serious consideration, these scholars took little account of bilingual transmission in an interdirectional way. As a result, both proponents presuppose that Jesus and gospel traditions were transmitted from Aramaic into Greek in a unilinear, teleological, and unidirectional way. Both groups can be called linguistic unidirectionalists. However, when we take the bilingualism of first-century Palestine and the Roman Near East into serious consid-

58

The Directionality of the Transmission of the Jesus and Gospel Traditions

eration, it is more probable that the linguistic transmission was hybrid, circular, and interdirectional.

1.5 Interdirectionality Hypothesis As surveyed from §1.1 to §1.4, a large number of scholars have manifested three tendencies of unidirectionality. Most New Testament scholars have assumed that the Jesus and gospel traditions were unidirectionally transmitted from Judaeo-Palestinian into Hellenistic tradition, oral into written form, and Aramaic into Greek and never vice versa. In other words, (i) the first three traditions lie behind the latter three traditions, (ii) the latter three traditions are the end products of the first three traditions, and (iii) the first three traditions have temporal priority over the latter three traditions. Not all scholars, however, have assumed these three unidirectionality hypotheses. It is intriguing that some scholars have raised the possibility of the interdirectional transmission, although their concepts have tended to be under-developed. Accepting Sanders’ criticism (§ 1.5.3) of the unidirectionality hypothesis, Marshall (1972-73:281 n.5) argues that “while a strict unilinear view is not stated in so many words by some scholars, they nevertheless give the impression of being unconsciously influenced by such a postulate” [emphasis added]. As mentioned before (§1.4.2.4), Gundry suggests interdirectional transmission to some extent. In this chapter, I will review some scholars’ arguments in relation to the three interdirectionality hypotheses. The possibility of the interdirectional transmission between Judaeo-Palestinian and Hellenistic tradition has been consistently raised by Martin Hengel (§1.5.1). The modal interdirectionality between oral and written tradition has been suggested by many scholars (§1.5.2). Lastly, Sanders suggests the relationship between Semitisms and temporal priority (§1.5.3).

1.5.1 Sitz im Leben Interdirectionality: Martin Hengel In regard to the Sitz im Leben interdirectionality between JudaeoPalestinian and Hellenistic tradition, Martin Hengel warrants mention.89 Criticizing the sharp contrast between Palestinian and Diaspora Judaism, he (1995:381 n.52) argues that the religionsgeschichtliche Schule

89

I will deal with Hengel’s arguments more fully below (chapter 5).

Interdirectionality Hypothesis

59

missed three points: (i) “the richness and creative power of Palestinian Judaism,” (ii) “the intermingling of Jewish and Hellenistic thought in Jewish Palestine,” and (iii) “the importance of bilingualism in the Palestine Jesus Movement in Galilee and Jerusalem.” For these reasons, he contends, “The old assertions of the religionsgeschichtliche Schule, still given in some circles today, are no longer tenable.” Remarkably, Hengel (1974:104-5; 1983:37) takes the bilingualism of first-century Palestine into account more seriously than most.90 The probability of bilingualism leads him continuously and persuasively to propose that Palestinian Judaism must be designated Hellenistic Judaism (1974, 1983, 1986, 1989a, 2001). He (1974:104) affirms: Hellenism also gained ground as an intellectual power in Jewish Palestine early and tenaciously. From this perspective, the usual distinction between Palestinian and Hellenistic Judaism needs to be corrected. … from about the middle of the third century BC all Judaism must really be designated “Hellenistic Judaism” in the strict sense.

Much of Hengel’s work has successfully diluted the borderline between Palestinian Judaism and Hellenistic Judaism, thereby making it difficult to distinguish clearly between Judaeo-Palestinian tradition and Hellenistic tradition. His conception of geographical interdirectionality is supported by the bilingualism of first-century Palestine and the Roman Near East. Bilingualism dilutes the Sitz im Leben boundary between JudaeoPalestinian and Hellenistic traditions since one cannot make a clear distinction between Semitic and Greek tradition: Semitic traditions could have included both Judaeo-Palestinian Semitic traditions and Hellenistic Semitic traditions, and Greek traditions could have included both Hellenistic Greek tradition and Judaeo-Palestinian Greek tradition (cf. §5.4). As a consequence, it is hard to say that the Jesus and gospel traditions were transmitted from Judaeo-Palestinian traditions into Hellenistic traditions in a unidirectional way. In other words, there is not a clear distinction between Judaeo-Palestinian and Hellenistic tradition. Unfortunately, Hengel still betrays tendencies of the linguistic unidirectionality. This assumption results from his strict linguistic division between Aramaic-speaking Judaism and Greek-speaking Judaism. He (1974:104) states that “a better differentiation could be made between the Greek-speaking Judaism of the Western Diaspora and the Ara90

Some scholars have already proposed the assimilation of Palestinian and Diaspora Judaism; W. D. Davies 1948:1-16. Also, Sanders (1969:190-7) seeks to dilute the difference between Hellenism and Judaism in a similar way. Recently, Collins (2000:18) adduces Hengel’s argument. Hengel (1969) deals with the issue in a full discussion.

60

The Directionality of the Transmission of the Jesus and Gospel Traditions

maic/Hebrew-speaking Judaism of Palestine and Babylonia.” Although he takes the bilingualism of first-century Palestine into account, his work has not ultimately shed the linguistic distinction which has been proposed by the religionsgeschichtliche Schule since F.C. Baur.91 Furthermore, it is noteworthy that Hengel advocates a relatively early translation transmission hypothesis. Explaining the translation of the Jesus tradition from Aramaic into Greek, he (1983:27) proposes that at all events “the rendering of parts of the Jesus tradition into Greek and the development of a distinctive theological terminology”92 must have been made in the Hellenistic Jewish Christian community in Jerusalem, when they had vivid memories of the activities, death and resurrection of Jesus. In this respect, he (1983:27) asserts that “the translation into Greek did not begin in Antioch, Ephesus or Rome, but at a very early stage in Palestine itself.” Hengel deserves credit for antedating the translation of significant Christian terminologies while early form criticism supposed that these translations were made in Diaspora regions (see §1.2, introduction to chapter 4, chapter 5).93 However, his Jerusalem church translation theory is not satisfactory. He disregards the possibility that the bilingualism of first-century Palestine enabled Jesus’ bilingual audience to translate the Jesus tradition during Jesus’ ministry simultaneously. As I have proposed, it is more probable that the translation of the Jesus tradition already had started with Jesus’ ministry in Galilee. I call this the Galilee translation theory or the simultaneous translation theory (see §3.3, §5.1). Although Hengel has successfully diverted scholars’ attention from the bilinguals of Antioch to the bilinguals of the Jerusalem church, he neglected the bilinguals of Galilee (§5.4). Moreover, he does not seem to consider the interdirectional possibility that Greek traditions could 91 92 93

Baur 1873-5: especially, see 1.38-60; Bousset 1970:120. n.3; followed by form critics, especially, Bultmann, see §1.2. For full discussion of Hengel’s argument, see §5.1. For the list of the terms, see chapter 5 n.8; refer to Hengel 1989a:18; 1983:26-8. On the basis of epigraphical evidence, van der Horst agrees with the early translation transmission hypothesis. He (1991:130) considers that “The translation of the early Jesus tradition from Aramaic into Greek took place early, within one generation after the life of Jesus, and this Greek translation turned out to be of so much greater use that nobody bothered to preserve the Aramaic original.” He (1991:130) admits the widespread bilingual milieu of the first-century Palestine when he notes that “a majority of the Jews in Palestine and the western diaspora spoke Greek.” Nevertheless, it is regrettable that he (1991:130) also holds the linguistic unidirectionality hypothesis, when he mentions that “Early Christianity started as a Jewish sect in an Aramaic-speaking milieu. But even before it emancipated itself from its Jewish matrix so as to become an independent religious movement, it dropped Aramaic and adopted Greek as its linguistic vehicle.”

Interdirectionality Hypothesis

61

have been retranslated into Aramaic again in Palestine as well as in the Roman Near East. Indeed, bilingualism supports the linguistic interdirectionality hypothesis (cf. §1.5.3). Consequently, Hengel remains a linguistic unidirectionalist. Accordingly, his point that the bilingualism of first-century Palestine leads to his conclusion that first-century Palestinian Judaism can be called Hellenistic Judaism is remarkable. It implies that it is hard to distinguish Hellenistic tradition from Judaeo-Palestinian tradition. However, if he subscribes to the bilingualism of first-century Palestine and the Roman Near East, it might be more persuasive to consider both the simultaneous translation theory and the linguistic interdirectionality hypothesis. In other words, the Jesus tradition was translated into Greek during Jesus’ ministry and it was retranslated into Aramaic again in a hybrid, circular, and interdirectional way.

1.5.2 Modal Interdirectionality As discussed (§1.3), many scholars have held the modal unidirectionality hypothesis. Some scholars have recently proposed the possibility that the modal transmission was not unidirectional but interdirectional. In other words, Jesus and gospel traditions were transmitted in a hybrid, circular, and interdirectional way, not in a unilinear, teleological, and unidirectional way. I will survey six representative scholars from five groups: Helmut Koester from form criticism (§1.5.2.1), David Parker from textual criticism (§1.5.2.2), Samuel Byrskog from the Scandinavian school (§1.5.2.3), Harry Gamble from liturgical use and canon formation (§1.5.2.4), and Graham Stanton and Richard Bauckham from notebook theory (§1.5.2.5).

1.5.2.1 Form Criticism: Helmut Koester Koester stands in the tradition of the modal unidirectionality hypothesis on the basis of form criticism. Basically, he supposes that the Jesus and gospel traditions are transmitted from oral into written form in a unidirectional way. He (1971:164; cf. 1990:31) says that “It is beyond doubt that earlier written documents were an intermediate stage between free transmission and written gospels.” Nevertheless, he maintains that the oral tradition was continuously circulated in parallel with

62

The Directionality of the Transmission of the Jesus and Gospel Traditions

written tradition up to 150 AD.94 He (1971:165; cf. 1990:31) suggests that “the synoptic tradition continued to develop in its free oral form after the composition of written gospels.” Furthermore, he (2007:25) submits that even most of the gospel materials in the writings of the Apostolic Fathers were derived from the continuing oral tradition rather than the written synoptic tradition. Going a step further, he suggests an interdirectional transmission between the oral and written traditions. He (2007:27) asserts that “free oral tradition continued well into later centuries and influenced both apocryphal and canonical gospel manuscripts.” He illustrates the interdirectional transmission with the modal transmission of 2 Clement. He (1994:297) argues that after 2 Clement had been written, the document was used orally once again in churches. On this account, he (2007:23) concludes that modal interdirectionality implies that some apocryphal gospels (e.g. Gospel of Thomas, the Unknown Gospel, the Gospel of Peter, etc; see 2007:3-23) are “at least as old and as valuable as the canonical gospels as sources for the earliest developments of the traditions about Jesus.” However, his movement toward modal interdirectionality is incomplete. He maintains that some apocryphal gospels were performed orally, whereas written canonical gospels remained without oral performance. Although he (2007:38) acknowledges that “All these gospels were not primarily produced as ‘literature’ but as writings destined for oral performances,” 95 he would have done better to realize that the canonical gospels which had already been documented were still circulated orally as well as in their written forms before and after 150 AD. After the written four gospels were accepted among churches around 150 AD, the oral traditions were circulated and would have interacted with written gospels. The written gospels are not the end products of the oral traditions. Consequently, despite his assumption of the modal interdirectionality of some documents, including 2 Clement, Koester cannot completely break himself from the traditional view of form criticism. Although modal interdirectionality was one of the significant presuppositions to support his whole argument, regrettably he does not explain in detail nor develop the modal interdirectionality.

94

95

This view has been assumed by other scholars. For instance, over a century before, Renan (1890:5.49-50) mentioned that “until the middle of the second century the words of Jesus continued to be cited from memory often with considerable variations.” Also, he (1994:294) mentions, “All three [written gospels] were functional; that is, they were optional and convenient aids designed to strengthen the role that the tradition about Jesus played in the churches.”

Interdirectionality Hypothesis

63

1.5.2.2 Textual Criticism: David Parker Parker holds to the modal interaction between oral tradition and manuscripts in the field of textual criticism. He (1997:179) suggests that the Jesus tradition was “told and retold, written and rewritten, in oral tradition and in successive versions of texts.” First of all, he (1997:205) objects to the predominant view that the written gospels have put an end to the fluidity of the Jesus tradition. Instead, he (1997:210) claims that “Sometimes oral tradition has been influenced by the written tradition, and sometimes the influence has been in the opposite direction.” He continues, “The written and oral tradition have accompanied, affected and followed one another.” In this respect, he (1997:204) suggests that “there was a continuing interplay between the Scripture – the text copied- and the tradition – the person engaged in the process of copying in and for the church.” Taking a further step, he (1997:102) convincingly argues that “The oral tradition is thus not something which ended at some point in the second or third or fourth century.” “The way,” he continues, “in which we read the written text is a part of the whole tradition which has been passed on from generation to generation.” He calls gospel tradition “living text.” However, it is regrettable that he sometimes assumes the modal unidirectionality hypothesis. He (1997:199-200) writes, “The variation between the gospels answers the question with regard to the earliest oral tradition, and the variation between the manuscripts with regard the written tradition.” This implies that he holds modal unidirectionality. However, for the most part, he persuasively suggests modal interdirectionality. Also, he takes little account of linguistic transmission, even though it is not his emphasis. Variants, as a matter of fact, could result from the linguistic interdirectionality between Aramaic and Greek (or Greek and other vernacular languages) as well as the modal interdirectionality between oral and written tradition. He (1997:13-5) suggests that the New Testament was translated from Greek into other vernacular languages in a linear way. Also, he seems to exclude interaction between versions, when he (1997:14) says that “the versions all had their own separate history once they came into being.” In discussion with Matthean “debts” and Lukan “trespasses” as alternative translations to Aramaic hôb¬, he (1997:72) may assume the possibility that the two Greek words were translated from the Aramaic word in a linguistically unidirectional way. However, bilingualism of first-century Palestine and the Roman Near East (see Part I) might have enabled the gospels to be translated from Greek into other vernacular languages as well as from other vernacular languages into Greek in an interdirectional way.

64

The Directionality of the Transmission of the Jesus and Gospel Traditions

This implies that it is uncertain that the vernacular texts were translated from the Greek text the evangelists wrote or from Greek text which had been translated from other vernacular languages. Accordingly, although he (1997:121) rightly suggests modal interdirectionality, the modal transmission of the Jesus tradition “is not at all [an] orderly business;” the linguistic interdirectionality hypothesis is demanded as well.

1.5.2.3 Scandinavian School: Samuel Byrskog Following the traditional Scandinavian school’s emphasis on eyewitnesses (§1.3.1), Byrskog (2000:129) rejects the discontinuity between orality and literacy with the result that written tradition put an end to oral tradition and that the oral tradition behind the written tradition was forever lost, as Kelber (1983) and Henaut (1993) argued. Instead, through Byrskog’s exhaustive study of the oral history of eyewitnesses in late antiquity, he proposes a complicated interplay between historical truth and interpreted truth, tradition and transmission, the present of the evangelist and the past of the story of Jesus, and orality and literacy. In regard to the relationship between oral and written tradition, he repeatedly stresses that orality inseparably intermingled with literacy. He (2000:138-44; here 142) states, “The written texts … entered regularly into the oral currencies of the communities.” He (2000:127) considers that our written Gospels are “‘memorative literature,’ written from memory to memory” and that the written texts “presupposed and supplemented oral modes of communication, regularly returning to oral modalities.” This refers to, as he (2000:127) asserts, “a high degree of continuity between the spoken and the written word on several stages of the gospel tradition.” He (2000:139) expresses interaction between orality and literacy as “re-oralization.” By the same token, Byrskog thinks that orality does not always have temporal priority over literacy due to their intertwinement. As he (2000:139-40) mentions, “Oral and written transmission do not follow the logic of first oral then written.” He (2000:143) convincingly proposes that various versions of oral traditions were circulated together and some of them were documented. As a consequence, he (2000:143) concludes, “This constant interaction of written and oral material in a process of re-oralisation is thus an essential ingredient of the Gospel tradition during all stages of its formation.” In relation to the modal interdirectionality of Q, his grasp of modal interdirectionality is shown expressively (2000:140):

Interdirectionality Hypothesis

65

The written material of Q was regularly employed in oral and aural activities; these activities fused the written and the oral traditions into a new synthesis; further material from the oral tradition entered into the written body of material; that material was again performed orally and aurally, etc., etc. – a constant process of reoralization and feedback in early communities of interpretation [emphasis added].

When it comes to composition of the Gospels of Mark and Matthew, he shows how Petrine tradition was preserved through the modal interdirectional transmission in their gospels. He (2000:292) explains that Mark “incorporated Peter’s oral history into his story.” Petrine tradition in the gospel indicates a subtle interchange between orality and literacy. In this respect, he (2000:304) suggests that Mark “found the living voice of the apostle partly in a textualized and re-oralized form.” The hybrid interchange between the two media is also shown in the formation of the Gospel of Matthew. Subscribing to the Markan Priority hypothesis, he (2000:292-7; here 296) proposes that “the oral history of the primary eyewitness [i.e. Peter] of Jesus’ active ministry … had developed into an oral and re-oralized tradition of decisive importance in the Matthean community.” Accordingly, he (2000:299) concludes that the gospel (e.g. Mark or Matthew) is a synthesis “of history and story, of the oral history of an eyewitness and the interpretative and narrativizing procedures of an author.” His argument that oral tradition was transmitted into written tradition and the written tradition transmitted into oral tradition in a hybrid, circular, and interdirectional way is beyond reproach.96 However, he does not take seriously the consequences of the constant interaction

96

When he (2000:274-97) deals with the trajectory of transmission from [Peter as an] eyewitness to the Gospel of Matthew via the Gospel of Mark on the basis of Markan priority, it seems that he overemphasizes Peter as a significant eyewitness, although he (2000:294, 305) admits Matthew’s own “interpretative and narrativizing procedures.” It is appropriate for his view of strong relationship between Mark and Peter. According to him (2000:295-6), Matthew used oral and reoralized tradition from Peter, who was held in high respect in Antioch. He (2000:295) insists upon “even the special Matthean material associated with Peter, that is, in addition to 16:17-19, his walking on the water (14:28-31) and his discussion of the temple tax (17:24-27).” However, it would be more plausible if he took Matthew himself as another significant eyewitness into consideration. In the case of 14:28-31, for instance, Peter could have told what just happened to the other disciples, including Matthew, who were eyewitnesses on the spot. One is itching to talk to others, when s/he is nearly dead. Peter was almost drowned. It was critical enough to surprise the other disciples, too. It is unlikely that Peter kept his lips tight until he openly told his accident in Antioch after around fifty years. If we take notice of Matthew and John as eyewitnesses, there must be more something complicated to be said for it. Nevertheless, the basic tenets of his modal interdirectionality remain intact.

66

The Directionality of the Transmission of the Jesus and Gospel Traditions

between oral and written tradition. Oral features do not always guarantee earlier tradition in relation to major issues of the historical Jesus, the synoptic problem, and unilinear development of christological titles. What is worse, despite his confidence in modal interdirectionality, he also holds linguistic unidirectionality. He (2000:294; cf. 1994:359-60) supposes that Semitic features in the Gospel of Matthew decide which tradition is earlier. For this reason, he proposes that Mt 16:17-19 is earlier than Mt 18:18. Therefore, whereas he rightly elucidated his view of modal interdirectionality, he maintains linguistic unidirectionality.

1.5.2.4 Liturgical Use and Canon Formation: Harry Gamble Harry Gamble takes the interplay between orality and literacy in the Christian worship service into serious consideration. He proposes that the canon of the New Testament took shape in the course of the interaction between oral reproduction and written texts in the service. First, he criticizes Kelber’s modal contrast between orality and textuality, for two reasons. Kelber (1983:210-1) suggested that the destruction of Jerusalem catalyzed modal change of the synoptic tradition from orality into textuality. However, Gamble (1995:29) insists that “Evidence for the production, use, and appreciation of texts in early Christianity, even before the composition of the Gospels, is too strong to allow oral tradition and literary activity to be set off against one another.” His second criticism of Kelber is due to the interaction between oral reproduction and texts. Gamble (1995:30, 32) suggests that “texts were routinely converted into the oral mode” and that “ancient authors were all well aware that what they wrote would normally be not so much seen as heard in public reading.” Regarding the interaction between oral and written modes he (2004:30; cf. 1995:28-32) suggests: The literate and oral modes … intersected and overlapped, frequently and effectively, with the result that the illiterate majority had the opportunity to participate in literacy. … There the oral and the written modes were certainly not incompatible or mutually exclusive, but co-existed in a complex synergy. Furthermore, he persuasively dilutes the perceived borderline between the oral and the written media in ancient times. He (2004:30) argues that interplay was at work both in the composition of texts and in their uses. When a text was produced, “an author normally dictated (orally) to a scribe, who wrote what he heard and thus produced a text from speech, converting the oral into the written.” Also, when it is reproduced, the interaction between oral and written modes occurs. He (2004:30-1) mentions that “a scribe normally dictated to himself, that is,

Interdirectionality Hypothesis

67

read aloud the text he was copying, and transcribed into a new copy what he heard himself reading.” Accordingly, he (1995:30) concludes that the oral and the written modes “were far closer and interactive in antiquity than today.” Taking a further step, Gamble (2004:29-30) pays attention to the discrepancy in early Christianity between a text-centred community and a more than 90% Christian illiterate community. In reply to this, he (2004:32-4) posits that production of texts for the early Christianity were aimed at pragmatic use in the service.97 He (2004:38) proposes that the same texts (i.e. the New Testament) were read over and over again. This implies that the role of liturgical reading is not to provide new information but to call to mind what their audience had already known.98 Moreover, reading texts incorporated Christian identity. Oral reproduction in the Christian worship and reading codices aloud make it possible to share the Christian knowledge with the Christian illiterate. Hence, he (2004:36) argues that the canon of the New Testament took shape “as a function of the history of the public reading and interpretation of scripture.” The canon, in other words, is “as a consequence not of the history of doctrine, but of the history of liturgy.” Accordingly, Gamble persuasively affirms that the interplay between oral and written media was active even before the composition of the gospels.

1.5.2.5 Notebook Theory: Graham Stanton and Richard Bauckham Stanton (2004:186) revisited the notion that “Jesus traditions were always transmitted orally, and only orally” up to the composition of the Gospel of Mark. He suggests that the use of notebooks was widespread. This implies that Paul (and his co-workers) and Christian writers made use of “notebooks for preliminary notes and drafts and for copies of letters, as well as for collections of scriptural texts.” In the light of this, he (2004:185) rejects “pan-oral” theory because it cannot explain “the prevalence of written excerpts of ‘classical’ writings in the Graeco-Roman, Jewish, and early Christian worlds.” In the case of the 97

98

He (2004:34-5) calls it “textual pragmatism.” He presents four reasons. (i) The reason that the early Christian manuscripts are small and compact codices is that these were more appropriate for everyday use in Christian communities. (ii) The way manuscripts were written aimed at greater legibility. (iii) The use of nomina sacra implies that the abbreviations would have been readily intelligible only to Christians. (iv) The manuscripts include better reading aids. This view is of significance in that the audience who did not know Aramaic at all must have known the meanings of Aramaic embedded words in the New Testament; for more discussion, see my introduction to chapter 8 and §8.3.2.

68

The Directionality of the Transmission of the Jesus and Gospel Traditions

Jesus tradition, Stanton follows Lieberman’s assertion that written notes were used by Jesus’ disciples. Lieberman (1962:205) argues: The Jewish disciples of Jesus, in accordance with the general rabbinic practice, wrote the sayings which their master pronounced not in form of a book to be published, but as notes in their pinaces, codices, in their notebooks (or in private small rolls). … We would naturally expect the logia of Jesus to be originally copied in codices. 99

Stanton (2004:188-9) strongly opposes the “widely held view that the followers of Jesus were either illiterate or deliberately spurned the use of notes and notebooks for recording and transmitting.” He argues that this view “needs to be abandoned.” Instead, he (2004:189) insists that several forms of notebooks were used by Jesus’ disciples. Second, he holds a balanced position concerning the relationship between oral and written tradition. He (2004:85) states that “Acceptance of the fourfold gospel did not mean the end of oral traditions.” At the same time, “continuing use of oral traditions,” he suggests, “did not necessarily mean that written gospels were unknown or of marginal importance.” In this regard, Stanton (2004:189) persuasively concludes that oral and written traditions “could exist side by side; orally transmitted traditions could be written down by the recipients – and written traditions could be memorized and passed on orally.” Unfortunately, Stanton does not develop his valuable theory further. If notebooks were used by Jesus’ disciples, the oral tradition is not always earlier than the written tradition. Bauckham’s argument warrants closer attention. Above all, dealing with the relationship between oral tradition and writing, he (2006:2889) suggests that some notes of the Jesus tradition could have been written, for four reasons: (i) Use of notebooks was prevalent in the ancient world. (ii) Not all of Jesus’ disciples were illiterate. Matthew the tax collector might take some notes. (iii) Some first Christians were literate, as the authenticity of James shows (cf. Bauckham 1999). (iv) Greekspeaking Christians in the earliest Jerusalem church could write when they translated the Jesus tradition from Aramaic into Greek.100 As a 99

Davids (1980:82) argues that “Certainly there was a wide use of notebooks and perhaps even written mishnah collections during the New Testament period” and he (1980:87) rightly concludes that “there is no reason to assume that the early transmission was exclusively oral.” Millard (2000:204; cf. n.13) rejects the argument that the scriptures were prohibited to be written. Citing Lieberman’s suggestion, he (2000:205) insists, “Within ‘orthodox’ or ‘rabbinic’ circles, therefore, the possibility of unofficial written records of religious teaching was recognized and allowed.” 100 I disagree on this point. The Jesus tradition was translated by some bilinguals (e.g. the twelve such as Peter, Andrew, and Philip and eyewitnesses) during his ministry. I will discuss this fully (chapter 5).

Interdirectionality Hypothesis

69

consequence, he (2006:289) affirms that “it does seem unlikely that no one would have even noted down the Jesus traditions in notebooks for the private use of Christian teachers.” In terms of the relationship between oral and written tradition, he (2006:289) notes that these notebooks “would simply have reinforced the capacity of oral transmission itself to preserve the traditions faithfully.” Moreover, he (2006:287) suggests that “in the predominantly oral culture of the ancient world, including the early Christian movement, writing and orality were not alternatives but complementary.” He (2006:288) continues that “writing existed to supplement and to support oral forms of remembering and teaching.” Bauckham (2006:289) convincingly concludes that “their closeness [notebooks] to orality must make it virtually impossible for us to distinguish them from oral sources.” Accordingly, it is worth observing their arguments that supplementary function of note-taking in the ancient world implies that there was not a clear-cut distinction between oral tradition and written tradition. In other words, the diluted distinction between oral and written tradition shows that oral tradition is not always earlier than written tradition. To sum up, many scholars (§1.5.2) have explained the directionality between oral and written tradition as “told and retold, written and rewritten,” “living text,” “intertwinement,” “a new synthesis,” and “reoralization and rewriting.” Their concepts support the modal interdirectionality hypothesis that the modal transmission of Jesus and gospel tradition is not unilinear, teleological, and unidirectional but hybrid, circular, and interdirectional. As a consequence, it is hard to distinguish oral tradition from written tradition. For the purpose of dating the tradition and assessing “authenticity”, it is pointless to attempt to identify oral features or oral traditions in isolation from the gospels in their final written forms. Although one might find oral features and/or oral traditions, it is not certain whether these oral traditions are the original sayings of Jesus and stories about Jesus, reoralized traditions, or the intentional oral traditions for aurality, which the evangelists (or notes-takers) intended. Making that distinction is always a dubious move. Accordingly, (i) the oral tradition does not always have temporal priority over written tradition. Also, oral tradition does not always reflect the more “original” sayings of Jesus and stories about Jesus than written tradition. (ii) The oral tradition or oral feature is not always to be associated with earlier tradition over and against written tradition. The oral features could be literary characteristics of the Gospel of Mark or Q be-

70

The Directionality of the Transmission of the Jesus and Gospel Traditions

cause writing was intended for oral performance. This means that the oral features do not necessarily support the Markan priority hypothesis or the Q hypothesis. (iii) The modal interdirectionality hypothesis rejects the slogan: “the earlier, the more original.”101 (iv) We cannot deny the possibility of the earliest written tradition. Some disciples and followers could have written sayings and deeds of their teacher. (v) The oral-centered society and the role of oral performance within ancient society required a grammar of sound rather than a grammar of writing when the Jesus and gospel traditions were written. This means that we should deal with the Jesus and gospel traditions from the perspective of the grammar of orality at the levels of syntax (chapter 6), phonology (chapter 7), and semantics (chapter 8) (for more discussion, see introduction to Part II). We have examined how many scholars criticized modal unidirectionality and proposed modal interdirectionality (§1.5.2). However, there are two limitations. First, they did not develop their views further in relation to the transmission of the Jesus and gospel traditions. I will deal with the issue in detail from the perspective of modal interdirectionality. Second, it is regrettable that even these scholars accepted linguistic unidirectionality and disregarded linguistic interdirectionality because of the bilingualism of first-century Palestine and the Roman Near East.

1.5.3 Linguistic Interdirectionality: E. P. Sanders As mentioned before, Torrey (§1.4.1.1) and Gundry (§1.4.2.4) may assume, to some extent, linguistic interdirectional transmission. Scrutinizing the relationship between Semitisms and temporal priority, Sanders (1969:193-207) confirms that Semitisms do not indicate antiquity. To prove this, he discusses five arguments. (i) He dilutes the borderline between Judaism and Hellenism (cf. chapters 3-5) and insists on continuity between Judaism and Christianity. (ii) He proposes that some Semitisms might have been reintroduced later, which means that Semitisms do not refer to antiquity. He (1969:204) says that “many of the 101 Following Koester’s lead, Perrin (1967:34) suggests that oral tradition was still circulated after the canonical gospels were composed. Some scholars consider that some tradition of the Gospel of John is authentic because the fourth gospel includes oral tradition of the historical Jesus; see Gardner-Smith 1938:96-7; Dodd 1963:424. However, although their view is plausible, oral tradition is not always earlier than the written tradition. Discussing the relationship between the Synoptics and the Gospel of John, Sanders (1993:66-8) suggests that the fourth gospel describes some Jesus traditions in a more detailed way than the synoptic gospels.

Interdirectionality Hypothesis

71

Christians who handed down traditions in such churches as the one in Syrian Antioch were bilingual and may have introduced Aramaisms into the Greek traditions.” (iii) He raises the possibility that some usages of supposed Semitisms coincide with those of other non-literary Greek papyri (cf. §6.1.2). This implies that such usages do not demonstrate the antiquity of a given tradition. (iv) Agreeing with Jewish Greek theory (cf. §6.1.3), he insists that some Semitisms do not indicate antiquity. (v) Some assumed Semitisms should be regarded as Septuagintalisms. And Septuagintalisms do not indicate antiquity (cf. §6.1.5, §6.4). As a consequence, he (1969:199) persuasively argues, “There is not necessarily a positive correlation between Semitisms and antiquity. An early tradition could have Hellenistic coloring while a late one could have Semitic coloring.” With this observation, Sanders (1969:20755) analyzes the frequency of syntactic Semitisms in the Synoptic tradition.102 Following Beyer (1962), he (1969:254) demonstrates that Matthew keeps the highest overall percentage of Semitisms, Luke next and Mark least, unlike the general assumption that scholars have held before. In this sense, he (1969:254) suggests, “It is thus very difficult to argue that Matthew and Luke had the redactional tendency of avoiding Semitisms.” Also, he (1969:232) warns that “The tendency which exists today to attribute any passage which contains a few grammatical or syntactical Semitisms to Jesus or even to the most primitive Palestinian Church is clearly misguided.” Consequently, he (1969:232; 1985:15) convincingly concludes that supposed Semitisms prove neither antiquity nor authenticity. Although his theory is solid, it would be more persuasive if Sanders took the bilingualism of first-century Palestine into serious consideration. Indeed, when he properly takes account of the bilingualism of Syrian Antioch, he suggests that Semitisms were reintroduced among Syrian bilinguals. Since he takes so little account of the bilingualism of first-century Palestine, his linguistic transmission still seems to be unidirectional from Aramaic into Greek. Sanders (1969:203-4; here 203) accepts that “at least a significant proportion of the earliest Christian traditions was first formulated in Aramaic“ [emphasis added]. This, he continues, guarantees “a search for the Aramaic background of the Gospel materials.” Accordingly, Sanders also holds the linguistic unidirectionality hypothesis of the synoptic tradition, although he convincingly points out that Semitisms do not refer to temporal priority from

102 He examines kai, parataxis, historic present, the use of ei`j for tij, and wording made more Semitic. However, I doubt that kai, parataxis is a Semitism, as Deissmann suggested (chapter 6 n.73).

72

The Directionality of the Transmission of the Jesus and Gospel Traditions

the perspective of tendency of the synoptic tradition. If he took bilingualism into account seriously, his argument would have been more consistent. This means that the study of the bilingualism of firstcentury Palestine and the Roman Near East is urgently demanded in relation to the transmission of the Jesus and gospel traditions.

1.5.4 Interdirectionality Hypotheses and Gospel Studies The scholars mentioned (§1.1-§1.4) have presupposed the three unidirectionality hypotheses in relation to gospel studies whether they have recognized them or not. The Jesus and gospel traditions were transmitted from Judaeo-Palestinian to Hellenistic, from oral to written, and from Aramaic to Greek. The directionalities are unidirectional and never vice versa. The three unidirectionalities have functioned as criteria for deciding temporal priority in gospel studies. The former three traditions have been regarded as earlier than the latter three. Furthermore, notions of unidirectionality have undergirded several major issues of gospel studies: the synoptic problem, the historical Jesus, geographical provenance of the gospels, textual criticism of the gospels and Acts, and unidirectional development of christological titles. However, the corresponding interdirectionality hypotheses offer significant caveats to the discussions of these issues based the unidirectionality hypothesis. The adoption of an interdirectional model has the potential to affect these major issues of gospel studies (cf. chapter 1 ns. 48, 76). Fortunately, some scholars have doubted these three presuppositions (§1.5.1-§1.5.3). Hengel has cogently and persuasively argued that we cannot make a sharp distinction between Judaeo-Palestinian and Hellenistic tradition, since there was no sharp distinction between Palestinian Judaism and Diaspora Judaism. His view has been approved by many scholars (§5.1). He suggested the strong possibility of Sitz im Leben interdirectionality. Also, many scholars have admitted that written traditions were not the end-product of oral traditions and that the Jesus and gospel traditions were interdirectionally transmitted between oral and written traditions. Last, Sanders proposed that Semitisms do not prove temporal priority. Nevertheless, even these scholars have not adopted all three interdirectionality hypotheses in full. While they have made major strides toward interdirectionality, they have not tried to relate interdirectionality to the transmission of the Jesus and gospel traditions in a full discussion. Moreover, except for Hengel and Fitzmyer, they have considered the linguistic milieus of first-century Palestine and the Roman Near East as partial bilingualism, monolingualism, or diglossia. Indeed, the transmission of the Jesus and gospel traditions

Interdirectionality Hypothesis

73

is better understood in relation to the bilingualism of first-century Palestine and the Roman Near East (chapter 2). The study of bilingualism (chapters 3-5) dilutes the borderlines between Judaeo-Palestinian and Hellenistic tradition, oral and written tradition, and Aramaic and Greek tradition. Furthermore, it supports interdirectionality rather than unidirectionality. The interdirectionality hypothesis undermines the chronological rule: “The earlier, the more original.” In other words, the Jesus and gospel traditions were interdirectionally transmitted between Judaeo-Palestinian and Hellenistic tradition, between oral and written tradition, and between Semitic and Greek tradition. In this respect, focusing on the linguistic interdirectionality hypothesis, I will investigate fully the bilingualism of first-century Palestine and the Roman Near East. There are two reasons for this. First, as discussed previously (§1.5.1-§1.5.3), many scholars have raised the possibilities of the geographical and the modal interdirectionalities and have approved the two interdirectionality hypotheses. However, the notion of linguistic unidirectionality from Aramaic to Greek remains unchallenged. Second, the linguistic unidirectionality hypothesis has functioned as a foundation for both geographical and modal unidirectionalities.103 Accordingly, the study of bilingualism related to directionality of transmission of the Jesus and gospel traditions is urgently required. In Part I, the bilingualism of first-century Palestine and the Roman Near East will be described (chapters 2-4). Also, the bilingualism of the earliest church in Jerusalem will be investigated (chapter 5). In Part II, I will discuss the arguments of the Jesus and gospel traditions at the levels of syntax (chapter 6), phonology (chapter 7), and semantics (chapter 8).

103 As mentioned before, the criterion for distinguishing Judaeo-Palestinian from Hellenistic tradition depends on the linguistic distinction between Aramaic and Greek tradition (§1.2). Many scholars assume that oral traditions were Aramaic, whereas written traditions were Greek. In this respect, the modal distinction depends on the linguistic distinction (§1.3).

ȱ

ȱ

ȱ

Part I: Bilingualism of First-Century Palestine and the Roman Near East As surveyed before (chapter 1), most scholars presuppose the three manifestations of unidirectionality. Namely, the Jesus and gospel traditions were unidirectionally transmitted from Judaeo-Palestinian into Hellenistic tradition, oral into written form, and Semitic into Greek and never vice versa. However, when the study of bilingualism is considered in depth, these unidirectional hypotheses should be reconsidered. In this regard, bilingualism in relation to the interdirectionality of the transmission of the Jesus tradition will be explored in Part I. First of all, by comparing bilingualism with diglossia, I will suggest that the linguistic milieu of first-century Palestine should be considered as bilingualism, and not as diglossia (chapter 2). Then, regional bilingualism of first-century Palestine and the Diaspora will be examined (chapters 3 & 4, respectively). This will be followed by an examination of the communal bilingualism of the earliest Christian church in Jerusalem (chapter 5).

2. Bilingualism and Diglossia Although many New Testament scholars have frequently acknowledged the bilingual milieu of first-century Palestine, as investigated in chapter 1, some have tried to relate the linguistic circumstances to the transmission of sayings by Jesus and stories about Jesus. However, a concerted investigation of the relationship between bilingualism and the transmission of Jesus and gospel tradition, in terms of linguistics, has not yet been attempted. Indeed, it was not until the early 1960s that linguistic scholars moved beyond general linguistics1 to focus on the sociolinguistic issue of bilingualism itself.2 Furthermore, biblical scholars who consider these issues often have eschewed issues of bilingualism and thought of it as an issue subsidiary to the larger question of the authorship of certain New Testament writings or the language(s) of Jesus.3 The study of ancient bilingualism, especially bilingualism of firstcentury Palestine, is subject to three difficulties. First, although it is difficult to define precisely what a “language” is in modern times, it is even more challenging to do so in the ancient world. For instance, no matter how some modern biblical scholars have drawn such a distinction, how one differentiates between a “language” and a “dialect” in antiquity is ambiguous and cannot be located with any precision on a scale (Rabin 1976:1008). Taylor (2002:303) comments on this in regard to Aramaic dialects: “studies of language contact suggest that at the boundaries between different dialects there will have been intermediate forms containing some features of both.” Some care should be taken, however, to form distinctions between language and dialect. Second, different designations used in ancient texts to refer to a language have

1

2

3

Chambers (1995:25-33) suggests two linguistic approaches to the science of language; one is an empirical method focusing on parole; the other is a rational method stressing on langue according to Saussure’s term. Generally speaking, the former is called “general linguistics,” whereas the latter is called “sociolinguistics,” “anthropological linguistics,” “cognitive linguistics,” or “neurolinguistics”; cf. Reed 2000:121-9. Coulmas 1997:1. Mackey (1968:554) posits that “bilingualism is not a phenomenon of language; it is a characteristic of its use. It is not a feature of the code but of the message. It does not belong to the domain of ‘langue’ but of ‘parole’.” For instance, see §2.2, §2.3. Acceptance to the term (i.e. bilingualism) among biblical scholars also took time; see the changes of Hengel’s designations (chapter 5, n.12).

78

Bilingualism and Diglossia

caused serious confusion for modern scholars. In antiquity, it was possible for the same language to be referred to by different designations in different regions, a tendency noted by Rabin (1976:1008-9). Butcher (2003:270-2) too, for instance, points out that some ancient authors indicate different meanings when they use “Syriac” and “Arabic.”4 Noteworthy as well, one finds the same sort of discrepancy when ancient writers refer to ancient cities. The Ptolemaic city of Alexandria was occasionally called Babylonia (Neusner 1984:39). Third, a paucity of extant language data from the period of concern is problematic. When these data are applied to spoken languages in ancient times, one can hardly detect any literary trace (Cotton 2005:151-2). The survival of extant materials sometimes may reflect speakers’ ideology (Cotton 2005:152). It is necessary to compare the linguistic circumstances of analogous modern linguistic situations. Contemporary language settings clearly provide data that one may attempt to use in order to create theoretical models. These models may then assist in the evaluation of ancient linguistic contexts and allow a more confident approach to linguistics in the past (Paulston 2000:83). Recently, unlike general linguists, specialists in the areas of sociolinguistics, cognitive linguistics, anthropological linguistics, and neurolinguistics have produced a significant number of studies related to bilingualism. Other areas of research such as archaeology, sociology, neuroscience and classical studies have helped New Testament scholars to gain easier access to information about bilingualism in first-century Palestine and the Roman Empire. Significant for the present conversation, several classicists have recently investigated the linguistic phenomenon of bilingualism5 and the consequences of bilingualism.6 Fewster (2002:220) discusses the significance of the subject for the field, stating that “bilingualism is a lively subject, both in linguistics and, particularly in recent times, in classics.” Horsley (NewDocs 5.19) also proposes that bilingual theory has a lot to offer to our understanding of language in antiquity. Just as in the field of classics, the study of the bilingualism of ancient texts would provide New Testament scholars

4 5 6

Tcherikover (CPJ 1.5) also mentions that “Syrians” sometimes refers to “Jews.” For various references to Phoenicia, see Muhly 1970:19-64. See, especially, the bibliographies of the following two books: Adams 2003 and Adams et al. 2002. Language contact results in some linguistic phenomena; phonological level (transliteration), syntactic level (leveling, interference), and semantic level (interference, borrowing and codeswitching); see Adams 2003 (passim). I will deal with these points in chapters 6, 7, and 8.

Definitions

79

with a significant tool to research biblical texts. To address the question of bilingualism in antiquity is, therefore, an interdisciplinary enterprise.

2.1 Definitions Some sociolinguistic terms related to bilingualism may be used in a variety of ways and are often ill-defined. Thus, it is necessary to make clear how certain linguistic terms are used in this thesis. Furthermore, there are difficulties in applying modern definitions to linguistic situations of first-century Palestine and the Roman Near East. Large gaps between these periods relate especially to issues such as language education, literacy, and communication vehicles. With these points in mind, caution will be taken when describing how terms and definitions will be used and applied in the chapters to follow.

2.1.1 Bilingualism Just as bilingualism exists in the present, so too did it exist in ancient times. Also, bilingualism can be called a cross-linguistic phenomenon insofar as it is found in most areas of the world. As linguistic statistics prove,7 bilingualism is a natural phenomenon in the world (Hoffmann 1991:1). As Mackey (1967:11) rightly points out, bilingualism is “far from being exceptional, [and] is a problem which affects the majority of the world’s population.” Although bilingualism appears self-evident, definitions of bilingualism vary depending on each scholar’s criteria of what constitutes “proficiency” in a second language.8 Generally speaking, at one end of the definition spectrum is one of the greatest American linguists, Leonard Bloomfield. He (1933:56) considers bilingualism to be a “native-like control of two languages.” He thinks of bilingualism as a high level of proficiency in two languages (i.e. a maximalist position). Hakuta (1986:4) properly contends that “native-like control is difficult to define, and very few people who would generally be considered bilingual have anything resembling native-like control of both 7

8

Hoffmann 1991:2-3. Explaining the reason of bilingualism, she mentions that “there are an estimated 5,000 tongues in the world, but only some 190 states, so it follows that many countries must contain many different languages, i.e. be multilingual. Approximately, 95 percent of the world’s population are speakers of the 100 most frequently used languages… for many countries bilingualism is a normal requirement for daily communication and not a sign of any particular achievement.” It is very difficult to define “bilingualism” in many aspects; see Hamers 1989:6-7.

80

Bilingualism and Diglossia

languages.” Diebold (1961:99) offers a minimalist definition of bilingualism: a person can be called a bilingual even if they have no efficient control over two languages, but do have “a highly atomistic knowledge of the second language.” This view he calls “incipient bilingualism,” which is also generally known as “semibilingualism.”9 However, one disadvantage of the minimal proficiency definition is that practically everyone in UK, USA, Canada, and many countries can be considered bilingual, because everyone can know a few words in another language (Romaine 1995:11). Mackey’s definition is more attractive for the present discussion. He (1968:554-84) regards bilingualism simply as the alternate use of two or more languages (cf. Weinreich 1963:1). Most specialists agree that this is a well-balanced view (Romaine 1995:12). The definition of bilingualism faces similar problems among New Testament scholars. The extent of bilinguality10 and who can be called a bilingual is an obstacle to biblical scholars and appears to cause hesitation in taking part in a more full-blown discussion of bilingualism of the New Testament (cf. NewDocs 5.24). For most biblical scholars, it seems that bilingualism is generally accepted as native fluency to such a high degree that one can “equally” control two or more languages. The maximalist position is accepted virtually without challenge. If we follow this Bloomfield-like definition, very few bilinguals are found in the New Testament at all. If we take a Diebold-like definition, almost all Jews of first-century Palestine are to be considered bilinguals. If we take either extreme definition, the study of bilingualism in the firstcentury Palestine would become, in large part, a pointless endeavor. In the pages that follow, I will follow Mackey’s definition of bilingualism.

9

10

“Semibilingualism” is also called “passive bilingualism,” or “receptive bilingualism.” For the sake of easy understanding among New Testament scholars, I will use “semibilingualism” in this publication. Furthermore, Horsley (NewDocs 5.24-5) makes a distinction between “receptive” and “productive” bilingualism. However, it would be more plausible if he discussed the terms in relation to ancient literacy and orality, for the basic concepts of the terms in modern times are different from those of ancient times. In other words, it is well-known that the distinction between Greeks and non-Greeks depends on speaking Greek, not on writing Greek in ancient times. In this respect, bilingualism in ancient times does not mean biliteracy; refer to Spolsky 1983:100-7. For the same reason, Jonathan Watt’s argument is not persuasive (§2.2.6). The term “bilingualism” is also called “bilinguality” carrying the sense of proficiency. When proficiency is stressed, bilinguality will be used but it will be used as almost same meaning with bilingualism here.

Definitions

81

2.1.2 Bilingual Community It is necessary also to define what is meant when speaking of a “bilingual community.” For instance, is it only when every member of a community is bilingual that one should label it a bilingual community? If we follow this definition, there may be very few communities that could be labeled bilingual. If only a certain percent need to be bilingual, then what percentage point tips the balance, letting the entire society be called bilingual? On this point, the definition of Hamers and Blanc (1989:12) will help to bring greater precision to the discussion of bilingual communities: Every bilingual community is situated between the two poles of a continuum, ranging from a set made up of two unilingual groups each containing a small number of bilinguals, to a single group with a more or less large number of members using a second language for specific purposes. At one pole most speakers in each group use only one language for all functions, whereas at the other a varying number of speakers use both languages but for different purposes.

This definition will be applied to the discussion of the linguistic communities of the twelve disciples of Jesus, the earliest Christian community in Jerusalem, and the Christian communities in the Roman Empire.

2.1.3 Bilingualism and Multilingualism Generally speaking, scholars do not distinguish between “bilingualism” and “multilingualism,”11 because the definition of “bilingualism” itself is generally referred to as the “linguistic competence of two or more languages.”12 In this sense, “bilingualism” will be used when referring to bilingualism or multilingualism, and specific instances of multilinguality will be emphasized (with italics).

2.1.4 Early Bilingualism vs. Late Bilingualism The age at which a person acquires the second language results in a striking difference in the level of proficiency. This is referred to as the “critical period hypothesis” (Hoffmann 1991:18, 36-7; Paradis 2004:59-

11

12

Janse (2002:334) supposes that the term “multilingualism” was derived from Galen’s use of di,glwttoj and polu,glwttoj rather than from Plutarch’s use of di,glwttoj in the sense of “interpreter”; for Galen’s view of language, see Edlow 1977. Weinreich 1963:1 [emphasis in the original]; also refer to Beardsmore 1982:4; Romaine 1995:12.

82

Bilingualism and Diglossia

60). A recent neuroscientific experiment in the journal Nature explores this idea: We applied functional magnetic resonance imaging (fMRI) to determine the spatial relationship between native and second languages in the human cortex, and show that within the frontal-lobe language-sensitive regions (Broca’s area), second languages acquired in adulthood (‘late’ bilingual subjects) are spatially separated from native languages. However, when acquired during the early language acquisition stage of development (‘early’ bilingual subjects), native and second languages tend to be represented in common frontal cortical areas.13

These researchers (Kim et al. 1997:173) also propose that the “age of language acquisition may be a significant factor in determining the functional organization of this area in the human brain.” It is generally known that an “early” bilingual means that someone acquired the second language between the ages of 3 to 4, whereas a “late” bilingual after the ages of 11 or 12.14 The result of the above experiment indicates that “early” bilinguals use their second and foreign languages like their first language, whereas “late” bilinguals cannot use their second language like their first language at the levels of phonology and syntax (Hoffmann 1991:36-7), although the phonological and syntactical competence had developed in close proximity, as if their extent were limited by some factor like an inhibitory radius. Accordingly, a distinction between “early” and “late” bilinguals should be made.

2.1.5 Primary Bilingualism vs. Acquired Bilingualism To be a bilingual in terms of “context,” according to Hoffmann (1991:18), “the infant/child who acquires two languages from the speakers around him/her in an unstructured way” can be called a “primary” bilingual or a “natural” bilingual. 15 On the other hand, an 13

14 15

Kim, Relkin, Lee, and Hirsch 1997:171. The subjects who participated in the study are explained as follows (p. 174): “The mean age of subjects was 29.3 (±4.2) years. Six subjects (‘early’ bilinguals) were exposed to two languages during infancy, and six subjects (‘late’ bilinguals) were exposed to a second language in early adulthood. The mean age of initial exposure to the second language was 11.2 (±1.5) years and the mean age that conversational fluency was achieved was 19.2 (±4.1) years. Each of the ‘late’ bilingual subjects had lived in the country of the second language, which assured a high standard for fluency.” Kim, et al. 1997:171-4; Perani, et al. 1998:1841-1852; Evans, et al. 2002:292. Furthermore, although it is well-known, Horsley’s definition of a “primary bilingualism” is not satisfactory. He defines a “primary bilingual” as “a speaker who has picked up a second language by force of circumstances (e.g., from the work environment), without any formal instruction,” whereas a “secondary bilingual” refers

Definitions

83

“acquired” bilingual or a “secondary” bilingual refers to “the person who becomes bilingual through systematic or structured instruction, that is, undergoing some kind of training (Hoffmann 1991:19). This bilingual theory will be developed further in Chapter 3.

2.1.6 Oral Bilingualism vs. Literate Bilingualism It is generally assumed that modern bilinguals refer to those who read/write and speak two or more languages. However, not all bilinguals in ancient times could write or read their matrix language and/or their embedded language.16 Only those who learn reading and writing

16

to “a speaker who has learned a second language via systematic instruction” (NewDocs 5.24). On the basis of his definition, he considers that upper-status or urban Jews in the first century could be secondary bilinguals, whereas lower-status or rural ones could be primary bilinguals due to less access to formal education. However, the situation is more complicated than Horsley suggests; first of all, his definition omits “age,” which is the most significant factor of primary bilinguality (see §2.1.4), despite that, as a matter of fact, even Beardsmore [whom Horsley himself quotes (NewDocs 5.24)] mentions a child (Beardsmore 1982:8). In fact, a primary bilingual of first-century Palestine in Aramaic and Greek means three possibilities. (i) When he/she was an infant or a child, he/she learned two languages in a linguistic situation where both Aramaic and Greek were used (see §3.1). (ii) When he/she was an infant or a child he/she learned two languages at home from his/her parents. In this case, his/her parent(s) was/were bilingual(s) in Aramaic and Greek or one parent used one language and the other another (see §3.2.4). (iii) When he/she was an infant or a child he/she learned one language at home from his/her parents and other from people in public without formal education (see §3.2.4). In this respect, these three possibilities of primary bilingualism refute Horsley’s suggestion that “there may be some appropriateness in seeing upper-status, urban Jews as those more likely to be secondary bilinguals, primary bilinguals being those with less access to formal education or who lived in rural areas.” In other words, since Horsley mentions that secondary bilinguals (in Aramaic and Greek) belonged to upper-status and urban groups, primary bilinguals in Aramaic and Greek could be upper-status and urban Jews as well (see chapter 3). We cannot say that secondary bilinguals always have better proficiency of language performance than primary bilinguals do, as Horsley assumes. For, the point of the distinction between “primary” and “secondary” lies in that primary bilinguals use their two languages as mother tongues since they are early bilinguals, whereas secondary bilinguals could use the second language only as school language or cultural language. It is noteworthy to mention that Hoffmann (1991:19) points out that secondary bilingualism “involved with formal language teaching at school, during which the learner does not normally have much opportunity to practice the language outside the classroom environment.” Modern sociolinguists are not interested in the distinction between oral and literate bilingualism. However, the distinction is of importance to biblical and classical scholars. For more discussion of the difference of bilingualism between in modern times and in ancient times, see Adams 2002:2, 3-7.

84

Bilingualism and Diglossia

could read and write the languages. This divides ancient bilinguals into oral bilinguals and literate bilinguals.

2.1.7 Stages of Bilingualism Bilingualism itself is not stable.17 In the progress of time, one of two languages in a bilingual situation is always inclined to break the balance of the bilingualism. Younger speakers are inclined to learn and perfect the dominant language, whereas older speakers tend to maintain the balance of the bilingualism with full proficiency in the dominant and recessive languages. In this respect, Sommer (1997:67-8) suggests four phases of gradual development of bilingualism: (i) incipient; (ii) integral; (iii) progressive; and (iv) regressive bilingualism. In this regard, it can be said that in the case of the linguistic milieu of firstcentury Palestine, regional bilingualism was “progressive bilingualism” in Aramaic and Greek.18

2.1.8 Balanced Bilingualism vs. Dominant Bilingualism When considering equivalence, according to Hoffmann (1991:22), “balanced” bilingualism refers to using roughly equal proficiency in both languages, whereas “dominant” bilingualism means that a bilingual has one stronger language and a weaker one. Hoffmann (1991:22) goes on to comment that “the language a bilingual feels more at home in, the ‘preferred language,’ may coincide with the dominant one, but this will not necessarily happen in every case.”

2.1.9 Matrix Language vs. Embedded Language In regard to dominant bilingualism, it is important to recognize the one-sidedness of bilingualism, for a bilingual is often inclined to one language among the two or more languages. For bilinguals, the dominant language is called the “matrix language” and the less dominant language the “embedded language.”19 The matrix language plays the 17 18 19

Concerning unstableness of bilingualism in detail, see Winford 2003:99. Before “progressive bilingualism” in Aramaic and Greek, it can be called “regressive bilingualism” in Hebrew and Aramaic. Myers-Scotton 2002:16. Some sociolinguistic scholars may call matrix language “base language” or “L1.” In this respect, Aramaic embedded words in Greek gospels will be considered (chapter 8).

Definitions

85

greater part in determining grammaticality such as word order or grammatical morphemes (for a more detailed discussion, see Field 2002:16). Additionally, the matrix language of an individual does not always correspond to the dominant language of their bilingual community. For instance, although the expressions, “Aramaic-speakers” and “Greek-speakers” reflect a monolingual context, “Aramaic-matrix speakers” and “Greek-matrix speakers” take bilinguality into account. Aramaic-matrix speakers refer to those who speak Aramaic as their matrix language and their linguistic competencies open the possibility that they can speak other language(s) as their embedded language(s).

2.1.10 Substratum, Adstratum and Superstratum Also of concern are terms that show the relationship between languages in bilingual society. Language contact causes language strata to occur in communities and is referred to with words such as substratum, adstratum, and superstratum.20 In bilingual societies, if one language is dominant culturally, politically, economically, and/or socially, it is called “superstratum.” “Substratum” refers to the non-dominant language of the two languages and is the counterpart to a superstratum. If neither of the languages is clearly dominant then the languages are called “adstrata.”

2.1.11 Ancestor, Daughter and Sister Language When considering the ancestry of language, terms are introduced by the family tree hypothesis in order to show the nature of the relatedness of languages. On the basis of evolutionary theory, the hypothesis assumes that “after an ancestor language has split into two or more daughter languages, the speakers of the daughter languages go their 20

Jeffers & Lehiste 1979:141-3, 173-86. In its original use, the three terms are used by historical linguists to explain how language contact causes language change diachronically. They suggest three definitions of them as follows: “adstratum means that one of two (or more) languages spoken within an area by people who maintain their primary language while receiving influences from the other languages involved in the contact situation; substratum means that former primary language of a group of speakers who have shifted to their formerly secondary language; and superstratum means that former primary language of a group of speakers who have entered a linguistic community and have been absorbed by that community, giving up their former primary language.” However, in this publication, sociolinguistically, the three terms will be used in the sense of the synchronic milieu of first-century Palestine.

86

Bilingualism and Diglossia

separate ways, linguistically and often physically” (Jeffers & Lehiste 1979:27). For instance, Aramaic can be called an ancestor language and daughter languages would then be Palmyrene, Hatran, Syriac, Samaritan, and Christian Palestinian. The relationship between the daughter languages of Aramaic is referred to as sister languages (e.g. Palmyrene is a sister language to Hatran; see Table 1 p. 188).

2.2 Diglossia in First Century Palestine? First of all, the history of research needs to be considered. Many biblical scholars have applied the concept of “diglossia” to a description of the linguistic milieu without regard to bilingualism.21 If this were not problematic enough, scholars like Spolsky (§2.2.5) and Greg Horsley (chapter 2 n.27) make an obscure distinction between bilingualism and diglossia due to their vague definitions. In this respect, before making

21

Certain scholars have applied bilingualism to the biblical languages, even though their approaches are not fully satisfactory; Hengel (§1.5.1; chapter 5), Gundry (§1.4.2.4), Porter (§1.4.2.5), Greg Horsley (passim), Silva (passim), Schwartz (§2.2.4; n.409; passim), Casey (§8.1.2; passim); Vorster (§488); Richard Horsley 1996:158-62 (He applied both diglossia and bilingualism to the linguistic milieu of first-century Palestine. This confusion of categories made his treatment unconvincing. It should be regarded as bilingualism rather than diglossia); Groom 2003:28-41, and Hezser (passim). It is interesting that before discussion about bilingualism was full-fledged among sociolinguistic scholars, Gehman (1951:90) applied bilingualism to an analysis of the linguistic milieu of Alexandrian Jews. Especially, it is striking that Dalman had a recent concept of bilingualism; see §3.2.4. Also it is noteworthy to mention that Dibelius has a view of bilingualism of the Roman Near East at that time and he tries to relate the bilingualism of the Roman Near East to gospel transmission (§1.2.1). Most recently, Loveday Alexander (2005:231-52) applies a Furgusonian definition of diglossia to Albert Wifstrand’s view of the Greek of the New Testament (Wifstrand 2005 chapters 1, 2, 3, 4, and 5) in order to explain the difference between Lukan and Markan languages on the basis of dialectic difference; for detailed discussion of Wifstrand, see §6.1.5. When Alexander (2005:232-42) interprets Wifstrand’s approach to the relation between Lukan and Markan language, she considers Lukan Greek as Hcode, whereas Markan Greek as L-code, for Lukan Greek is more literary than Markan Greek. Moreover, she (2005:242-52) regards Lukan Semitisms as H-code, whereas Markan Semitisms as L-code. However, it is dubious whether the distinction of the features of the H-code (Lukan) and L-code (Markan) is clearly made, as will be discussed in detail (§2.2; §2.3). And if only one function (e.g. literacy) out of Ferguson’s nine functions is necessary for the diglossic situation to exist, then it seems inappropriate for her to have applied the concept of diglossia to compare the two languages, as will be mentioned (§2.2.5). In addition, there are some scholars who may apply diglossia theory to the similar linguistic situation; Goshen-Gottstein 1978:169-80; Harviainen 1984:95-113.

Diglossia in First Century Palestine

87

observations on the bilingualism of the first-century Roman Empire and Palestine, applications of the diglossia model to the linguistic milieu scholars have discussed will be considered.

2.2.1 Charles Ferguson There are many biblical scholars who apply the term diglossia to the linguistic milieu of first-century Palestine. Ferguson originally gave inspiration to both biblical and sociolinguistic scholars and will be considered first. He (1959:336) defines diglossia as follows: A relatively stable language situation in which, in addition to the primary dialects of the language (which may include a standard or regional standards), there is a very divergent, highly codified (often grammatically more complex) superposed variety, the vehicle of a large and respected body of written literature, either of an earlier period or in another speech community, which is learned largely by formal education and is used for most written and formal spoken purposes but is not used by any sector of the community for ordinary conversation. Following this definition of diglossia, Ferguson posits that there are two prerequisites for his application of diglossia: (i) genetic relationship between the two languages on some level (i.e. form) and (ii) complementary functional distributions between H-language and L-language (i.e. function). In other words, in relation to “form,” he suggests that the languages should be two dialects of the same language or at least genetically related languages. Ferguson (1959:325-340) provides four pairs of examples: spoken Arabic and classical Arabic; Swiss German and standard German; Haitian Creole and standard French; and spoken Greek and literary Greek. As to “function,” in diglossic speechcommunities he distinguishes a H (high) variety from a L (low) variety22 for two languages that are used for different complementary functions. He (1959:328-36) argues that the H variety learned in the context of formal education has more prestige (religious), more literary heritage, a strict standardization, a higher developed syntax, and more complicated morphophonemic phonology than the L variety, which is acquired at home. Problematic is that there are few examples to support Ferguson’s pure conditions of form and function. However, despite this, these distinctions between the two languages of diglossia impelled biblical

22

Sociolinguistic scholars mix “H-language and H-variety” and “L-language and Lvariety.”

88

Bilingualism and Diglossia

scholars to begin applying the sociolinguistic tool of the study of diglossia to the linguistic milieu of Palestine in the first century.

2.2.2 Joshua Fishman In comparison with Ferguson, Fishman should be mentioned because there are two sociolinguistic models for the analysis of a diglossic situation. One is the Fergusonian model, while the other is the Fishmanian one. Fishman (1967:29-38) extends the definition of diglossia to bilingualism and suggests the relationship between them by means of a fourfold table as follows:23 [Figure 1] The Relationship between Bilingualism and Diglossia

DIGLOSSIA BILINGUALISM +

+ Both diglossia and bilingualism Diglossia without bilingualism

Bilingualism without diglossia Neither diglossia nor bilingualism

As discussed, scholars such as Porter (§1.4.2.5) and Spolsky (§2.2.5), who apply diglossia theory to the linguistic milieu of first-century Palestine should have taken the Fishmanian model rather than the Fergusonian model into account since their theories are closer to the former model. In light of Fishman, the linguistic milieu of first-century Palestine should be regarded as “bilingualism without diglossia,” that is, just “bilingualism,” as will be demonstrated (§2.3).

2.2.3 Pinchas Lapide Lapide (1975:483-501) first applied diglossia theory to the linguistic milieu of first-century Palestine; however, his argument has not attracted much scholarly attention.24 Following the Fergusonian model, 23 24

It is also published in an expanded version (Fishman 1971:73-89). The scholars mentioned (§2.2) paid no attention to Lapide’s argument at all. (i) Rabin (1976) seems to overlook Lapide’s article, despite its similar methodology and consequence. (ii) There are no scholars who discussed Lapide’s argument in the mentioned book, Diglossia and Other Topics in New Testament Linguistics (Porter ed.). (iii) As a consequence, it is not proper that Spolsky and Watt consider that Rabin coined

Diglossia in First Century Palestine

89

he suggests that the linguistic situation should be regarded as diglossia, that is, Hebrew is H-language, whereas Aramaic is L-language. In his final conclusion, he (1975:498) adds Greek to the list so that he calls the linguistic milieu “triglossia.” Furthermore, in his view, Jesus could speak all three languages (i.e. Hebrew [H language], Aramaic [L language] and Greek). As to function, Lapide maintains that Aramaic (L-language) is a secular and spoken language, whereas Hebrew (H-language) is a sacred, written, and religious language. Aramaic as an L-language, he argues, is supported by the fact that, above all, there are not many literary works in Aramaic in first-century Palestine in comparison with Hebrew works.25 He divides New Testament Semitisms into Hebraisms as H-language (i.e. H Logia) and Aramaisms as L-language (i.e. L Logia) depending on diglossic functions. Interestingly, he suggests one criterion for determining the authenticity of the Jesus tradition: whether it is composed of H Logia and L Logia. However, his proposal is not convincing at some points. (i) In terms of form, there is no consistency. He (1975:485-6) mentions that both Aramaic and Hebrew are sister languages. Thus, the two languages fit into the form in comparison with the four pairs of languages suggested by Ferguson. However, Greek is not genetically related to the other two Semitic languages. The addition of Greek thus renders this situation more complex than Ferguson’s model can accommodate (§2.2.1). (ii) When it comes to function, many Aramaic writings have been recovered.26 This indicates that it is inappropriate to view Aramaic simply as an oral vehicle, while viewing Hebrew as a literary vehicle. Accordingly, his diglossic distinction between Hebrew as H-language and Aramaic as L-language is unconvincing. Nevertheless, it should be stressed that it is Lapide who first applied the sociolingistic theory (i.e. the diglossia theory) to the linguistic situation, although it is regrettable that he (1975:485-6) mentions that “The gospel evidence confirms his (Jesus’) bilingualism, which upon closer scrutiny, turns out to be Diglossia“ [emphasis in the original]. Closer scrutiny, conversely, shows that the linguistic milieu of the first-century Palestine requires a bilin-

25 26

“triglossia”, since Lapide has already used it; see chapter 2 n.33. Fortunately, Wise (1994:118) mentions the title of Lapide’s article and Voelz (1984:924) simply introduced his view. For my criticism against this argument, see §2.3.4. Fitzmyer (1979a:39) suggests that “it is now evident that literature was indeed being composed in Aramaic in the last century B.C. and in the first century A.D.” For the existence of plentiful Aramaic literatures of the first-century Roman Near East, refer to §2.3.4; §3.1.

90

Bilingualism and Diglossia

gual model rather than a diglossic model (§2.3). Lapide’s drawbacks still appear when the views of scholars discussed below are taken into account.

2.2.4 Chaim Rabin Rabin (1976:1007-39) describes the linguistic milieu of first-century Palestine from the perspective of diglossia. He (1976:1007-1039) applies the term diglossia when considering late biblical Hebrew, Mishnaic Hebrew and Aramaic. The significance of his proposal is that his sociolinguistic approach provides a general framework for later biblical scholars.27 His argument will be discussed in more detail. Analyzing the linguistic milieu according to the diglossic scheme, Rabin (1976:1008-9) offers three functional distributions of the functions of H- and L-varieties which Ferguson categorized: (i) formal (religious) / ordinary, (ii) prestigious / non-prestigious, and (iii) written / spoken. Rabin proposes three pairs of diglossic relationships between late biblical Hebrew, Mishnaic Hebrew and Aramaic. When Rabin (1976:1008 n.1, 1015-6) proposes that late biblical Hebrew was an H-language and Mishnaic Hebrew was an L-language, this logically leads to the assumption that late biblical Hebrew would have been used as a formal (religious) and written language with prestige, whereas Mishnaic Hebrew would have been used as an ordinary and spoken language without prestige. The problem, however, lies in the fact that his argument is not consistent with his theory of functional distributions. For instance, in terms of “religious” versus “ordinary” function, Rabin proposes that

27

See Wise 1992a:434-444. It is especially noteworthy that Horsley (NewDocs 5.5-48) applies sociolinguistic theories to the linguistic milieu of first-century Palestine in more detail. He regards diglossia as a societal phenomenon, whereas bilingualism is an individual one, when he (NewDocs 5.7) explains that “bilingualism is often viewed as an individual, not a group, phenomenon.” On this account, he (NewDocs 5.7) argues that an issue such as the bilingual community of first-century Palestine “brings us to the subject of diglossia.” In that sense, he (NewDocs 5.8) also proposes that the diglossic scheme fits the linguistic milieu well: “Do both [H-language and Llanguage] have a sufficiently clear, differentiated function that they can coexist peacefully within the society? All these points are of relevance to the question of language use in Palestine in I AD as, of course, elsewhere.” As a consequence, in terms of the relationship between Aramaic and Hebrew, Horsley agrees with Rabin’s diglossic scheme, when he (NewDocs 5.22) proposes that “the high/low relative status of Hebrew/Aramaic was clear-cut.” What is more, although he (NewDocs 5.22) also tries to apply the diglossic scheme to the relationship between Aramaic and Greek, it seems fruitless.

Diglossia in First Century Palestine

91

Mishnaic Hebrew was used as an everyday language,28 one of the characteristics of an L-language. Furthermore, in the case of imperial Aramaic-Mishnaic Hebrew diglossia, he (1976:1027) suggests that “Aramaic played the role of the upper, and Mishnaic Hebrew that of the lower language.” However, he (1976:1008-9) suggests at the same time that the sect of the Pharisees as well as the Qumranites used Mishnaic Hebrew as a language for religious teaching, which is one of the features of an H-language.29 As a consequence, Rabin’s functional distributions between H-varieties and L-varieties do not fit well. In this regard, the suggestion of Schwartz is particularly relevant, although I disagree that Hebrew was used in Palestine between 300 BC – 70 AD. He (1995:17) suggests that Hebrew from 300 BC – 70 AD in Palestine was both “classical” and “vernacular.” In other words, Hebrew functioned both as a literary language of the elite (H-language) and as a vernacular language of the peasantry (L-language). In this sense, in relation to the discrepancy of Rabin’s functional distribution theory, one should keep in mind a comment by Versteegh (2002:68): It has become abundantly clear that the view of diglossia as a linguistic situation, in which one language variety is used exclusively in formal writing and speaking, whereas the other is used for informal speaking, is too imprecise. In a diglossic speech community there are no discrete varieties, but linguistic variation is organized along a continuum between the standard language and the vernacular.

Schwartz (1995:17) does well to conclude that “The strict ‘diglossia’ model thus fails to explain the ancient Palestinian evidence.”30 Not only is there a difficulty regarding function, but Rabin’s analysis is also problematic when it comes to form. His discussion is connected with the genetic relation between the concerned languages. According to Ferguson’s definition, upon which Rabin’s is based, “diglossia” refers to a linguistic phenomenon that envisages two varieties of the same language or at least two genetically related languages (Ferguson 1959:336). On this point, Rabin’s diglossic scheme could not explain the other two important languages, such as Greek and Latin as 28

29 30

Rabin (1976:1015-6) considers that Mishnaic Hebrew “was the spoken language of the Judaean population. The mixed style was, therefore, a far-reaching concession to the language habits of the less educated reader, but enriched his limited everyday vocabulary from the rich reservoir of the Bible.” Rabin (1976:1019) suggests that “the Qumran sect did not use mishnaic Hebrew, but conducted also their oral teaching in late biblical Hebrew.” Ten years later, he still holds the same opinion in his revised article. In the paper, he is concerned with the concept of bilingualism rather than the concept of diglossia, which has been followed by some scholars; especially, Hezser (2001:246) agrees with his argument.

92

Bilingualism and Diglossia

a whole, despite the fact that Greek formed a significant part of the linguistic repertoire of first-century Palestine (1976:1011, 1032, 1036). In a sense, although the treatment of the Greek in his article seems to digress from the main point of his discussion (1976:1011, 1036),31 his diglossic scheme does not permit him to deal with other languages, especially when they are genetically unrelated languages. As a consequence, this diglossic model does not provide a more complete picture of the linguistic milieu of first-century Palestine. In summary, although Rabin’s discussion has had an enormous impact on subsequent biblical scholarship, his diglossic approach to first-century Palestine displays two problems. First, his functional allocations are inconsistent. Second, in terms of linguistic form, his diglossic scheme cannot reflect the linguistic milieu of first-century Palestine as a whole. Silva (1980:217 n.55) rightly objects that Rabin’s use “of the term ‘diglossia’ (Hebrew and Aramaic) to describe the status of Greek in Palestine seems to me unfortunate.”32 However, it is unfortunate that most subsequent scholars have simply concerned themselves with Rabin’s notion of the form of diglossia rather than with his inconsistent understanding of diglossic function.

2.2.5 Bernard Spolsky The works of two scholars, Spolsky and Jonathan Watt, pay close attention to form rather than function. First, Spolsky generally follows Rabin’s theory.33 In relation to form, he disagrees with Rabin because 31

32 33

Rabin, as a matter of fact, wrote chapter twenty-one ‘Hebrew and Aramaic in the First Century’ and subsequently, Mussies wrote chapter twenty-two ‘Greek in Palestine and the Diaspora’ in the same publiation. Silva’s view of relationship between bilingualism and Semitisms will be considered below (§8.1.2). Spolsky calls Rabin’s theory “triglossia”; however, it does not seem that Rabin himself regarded the linguistic milieu as triglossia; what is worse, it is not appropriate that Spolsky says that Rabin coined the term “triglossia,” when Spolsky (1991:85) mentions: “Rabin (1976) has suggested the term triglossia… Rabin’s terminological innovation is a useful one.” However, Rabin himself did not call his position “triglossia” at all in the article of 1976 that Spolsky cites. The reason Spolsky projected the term onto Rabin is that Spolsky himself intends to consider the linguistic milieu as triglossia in Aramaic, Hebrew, and Greek. Without a simple check of Rabin’s article, Jonathan Watt (2000b:27) also mentions that “Rabin (1976) coined the neologism ‘triglossia.’” However, no one will be able to find the term “triglossia” that Rabin supposedly coined in the article. Rather, as mentioned before (chapter 2 n.24), Spolsky and Watt should have cited Lapide’s article (1975:498) where “triglossia” had already been coined.

Diglossia in First Century Palestine

93

Rabin omits Greek. Spolsky (1983:95-109; 1985:35-50; 1991:85-104) adds Greek to his repertoire of languages and describes the situation as “triglossia” (i.e. Aramaic, Hebrew, and Greek). In relation to this, Porter (2000b:57) rightly objects that he should have taken the Fishmanian model rather than the Fergusonian model as his theoretical ground for diglossia in order to portray the diglossic relation among the three languages. As mentioned before (§2.2.2), the problem is that in Fishman’s term the linguistic phenomenon of ‘bilingualism without diglossia’ is the same as that of bilingualism. Simply put, Spolsky should have applied the model of bilingualism to first-century Palestine without the distraction of the diglossic model. As to function, although criticizing Rabin’s inappropriateness of functional distributions (Spolsky 1991:85), Spolsky (1991:95) also observes that only certain functional allocations correspond to the Fergusonian model. He (1983:99-100; 1985:40-1; 1991:94-5) suggests that the three languages functioned in a complementary way until the end of the Bar Kokhba revolt in 135 AD: Hebrew was used as a language of written sacred texts, Aramaic for legal contracts and commerce, and Greek for governmental language. Furthermore, Spolsky does not distinguish between the concepts of diglossia and bilingualism, mixing triglossia and multilingualism in many cases.34 However, in terms of function, three points should be considered. (i) If only one or two functions out of Ferguson’s nine functions are necessary for a diglossic situation to exist, then it seems inappropriate for him to have introduced the model of diglossia to the linguistic situation. Furthermore, it is hard to make a sharp distinction between the spoken and written forms of ancient languages (Porter 2000b:55-6). (ii) It is well-known that Aramaic and Greek were also used as languages of prayer and study; Greek was used as a trade language as well. Moreover, the functions such as the certain functions (i.e. written sacred texts, legal contracts and commerce, and governmental language) do not fit into the diglossic dichotomy of H-language or L-language. In other words, these particular functions belong to H-varieties, not Lvarieties. Rather, Spolsky’s distribution fits into the model of bilingual34

Describing the same linguistic situation, Spolsky uses multilingualism together with triglossia, as we can easily see from the titles of his three articles; ‘Triglossia and Literacy in Jewish Palestine of the First Century’ (1983), ‘Jewish Multilingualism in the First Century’ (1985), and ‘Diglossia in Hebrew in the Late Second Temple Period’ (1991); and elsewhere in the three articles. Unfortunately, explaining the linguistic milieu of first-century Palestine as multilingualism (i.e. Hebrew, Aramaic, Greek, and Latin), Spolsky and Cooper (1991:25-30) also apply Fergusonian diglossia theory to the analysis of the linguistic situation.

94

Bilingualism and Diglossia

ism for the bilingualism model offers coexistence of languages and language choice, as will be discussed.35 (iii) In this regard, it seems that he has confused the two concepts. However, the two models (i.e. bilingualism and diglossia) are totally different, as will be argued (§2.3.1). It is not proper that Spolsky mixes the two concepts.36 Accordingly, he should have applied the model of bilingualism to the linguistic milieu.

2.2.6 Jonathan Watt Watt pays significant attention to discussing form on the basis of the Fergusonian definition of diglossia. Disputing Spolsky’s triglossia theory, Watt (2000b:24) supposes that, according to Ferguson’s definition, one of two prerequisites for diglossia is the genetic relationship between the concerned languages, and he (2000b:32-33) stresses that “diglossia can apply only to the Semitic languages (or, theoretically, to the Greek alone), but not to Hebrew (or Aramaic) and Greek simultaneously“ [emphasis in the original]. On this account, he (2000b:32) considers Hebrew and Aramaic as essentially the same language; instead, he inserts Greek and Latin as a T (tertiary) category into the linguistic repertoire of first-century Palestine. In addition, in contrast to the bipolar model (i.e. H-variants vs. L-variants), he borrows the concept of “continuum” from John Platt (1977:361-378) and posits Aramaic and Hebrew within a diglossic continuum as represented below (Watt 2000b:34): High1 High2 Low1 Low2 T1 T2

= = = = = =

biblical Hebrew (written) Mishnaic Hebrew (spoken, written) Judaean Aramaic (spoken, written) Galilean Aramaic (distinguishable in speech only) Koine Greek (spoken and written) Latin (spoken? written?)

Three criticisms may be levelled at Watt’s thesis. (i) With respect to function, Platt (1977:364) analyses six functions of the different languages used by the English-educated Chinese communities of Singapore and Malaysia. He considers the functions of language in contexts such as family, friendship, religion, education, employment, or transaction in order to distinguish H- from L-varieties. However, Watt’s theory is weak in that, as his diagram shows, he seems to make three distinctions (i.e. H- from L-varieties, High 1 from High 2, or Low 1 from 35 36

For relationship between language choice and bilingualism, see §2.3.2. Paulston 2000:87-8. Paulston suggests six different points between diglossia and bilingualism. For detail discussion, see §2.3.1.

Diglossia in First Century Palestine

95

Low 2) on the basis of only a pair of functions, such as literacy versus orality. (ii) Spolsky (1983:100-107) seeks to demonstrate that in firstcentury Palestine, “oral” is superior to “written” and that “oral” is more prestigious than “written.” If Spolsky is indeed correct, the “spoken” function of language in ancient times could stand in a different relationship to the “written” language than in modern times. If “spoken” is more prestigious than “written,” Watt’s analysis of H- and Llanguages would reverse the nature of the relationship. (iii) Finally, in relation to form, Watt’s dependence on Ferguson’s definition of diglossia leads him to exclude Greek and Latin from his diglossic scheme and, consequently, to introduce his “T” category (the Tertiary category refers to the addition of Latin and Greek to his linguistic inventory of first-century Palestine). However, in contrast to Watt’s narrow definition of diglossia, some sociolinguistic scholars have applied the concept of diglossia to the bilingual situation. 37 In this respect, Paulston (2000:87; cf. Porter 2000b:58) rightly points out that “it is quite clear that the notion of functional complementary distribution can include separate languages (and language families) as well as varieties of the same language.” Furthermore, Ferguson (1991:215-6) himself later admitted the disadvantages of his narrow definition and would now consider a multilingual community such as Switzerland (where German, French, Italian, and Romansch are used) to be diglossic. To sum up, in terms of form, the application of the diglossia model inhibits Rabin from explaining the use of other languages such as Greek and Latin. Spolsky is inconsistent when he refers to bilingualism with diglossia, and would likely benefit from adopting Fishman’s model. Watt introduces the tertiary category into his diglossic scheme because he knows that the model of diglossia cannot make a place for Greek and Latin. Accordingly, various scholarly views indicate that the model of diglossia, in terms of form, is not in itself a proper tool to describe the linguistic milieu of first-century Palestine. Much more important are arguments related to function. Rabin’s proposal of the function is inconsistent in his argument. Further, Spolsky and Watt uncritically accepted it. As noted above, the inadequacies of the diglossic scheme of “functional distribution” applied to firstcentury Palestine can be summarized into three points. (i) The conditional lists of functional distributions suggested by Rabin, Spolsky, and Watt do not result in much agreement. (ii) The functional distributions

37

Since Fishman’s article, scholars have extended the definition of “diglossia”; see Fishman 1967:29. For detailed criticism of this extended definition, see Paulston 2000:85-87.

96

Bilingualism and Diglossia

of the H- and L-language in first-century Palestine served as only one or two levels of the functions, although Ferguson listed nine levels of functions. (iii) Some functions in antiquity, such as orality versus literacy, might be different from modern times. As a consequence, it is inappropriate to apply the diglossic scheme to the linguistic milieu of first-century Palestine because this model requires a clear distinction between H- and L-language. In terms of function, this will be discussed below (§2.3).

2.3 Bilingualism and Diglossia Models It is important to recognize that the linguistic milieu of first-century Palestine does not fit into the model of diglossia, but rather that of bilingualism. An unfortunate consequence of (the application of) the diglossia model is the obstruction of the application of the model of bilingualism and, moreover, the maximalist view of bilingualism in first-century Palestine. Indeed, focusing on the model of diglossia has diverted scholarly attention from the model of bilingualism because of the methodological overlap of the two sociolinguistic concepts. However, I propose that the linguistic milieu of first-century Palestine requires the model of bilingualism rather than the model of diglossia. The bilingualism model is supported by four points: (i) the features of the four languages of the first-century Palestine (§2.3.1), (ii) language preference theory (§2.3.2), (iii) the maximal use of Greek in first-century Palestine (§2.3.3) and (iv) the biliteracy of first-century Palestine (§2.3.4). Using these points, four comparisons between the bilingualism model and the diglossia model will be suggested.

2.3.1 Characteristics of the Four Languages in First-Century Palestine The functional distribution of diglossia does not fit into the linguistic milieu of first-century Palestine (§2.2). Among the nine diglossic functions Ferguson (1959:328-36) suggests, 38 Paulston (2000:87-8) adopts three different points between bilingualism and diglossia.39 (i) Whereas 38

39

In order to explain the diglossic situation, he enumerates nine points: function, prestige, literary heritage, acquisition, standardization, stability, grammar, lexicon and phonology (§2.2.1). In total, Paulston suggests five points. However it seems that two points, such as direction of language shift and ethnicity, require more discussion in the case of the linguistic milieu of first-century Palestine.

Bilingualism and Diglossia Models

97

L-language in diglossia is not standardized and often not written, the L-language in bilingualism can be standardized. (ii) The L-language in diglossia is oral, while in bilingualism both the H-language and the Llanguage can be oral as well as written. (iii) There are no native speakers of the H-language in the diglossic scheme, whereas there are native speakers of both languages in the bilingual scheme. In addition, two more points between these models, which Ferguson mentioned but Paulston did not, are in regard to grammar and morphophonemic phonology. (iv) In diglossia, the grammar of the H-language is more developed than that of the L-language, whereas in bilingualism the grammars of the two languages both could be highly developed. (v) In diglossia, the phonology of the H-language is more complicated that of the L-language, whereas in bilingualism the phonologies of the two languages both could be highly complicated. When the characteristics of the four languages (i.e. Hebrew, Aramaic, Greek and Latin) of first-century Palestine are taken into account individually, it is evident that the features of the four languages require the model of bilingualism rather than diglossia. Consider the following points: (i) Aramaic, Hebrew, Greek and Latin were standardized; (ii) their grammars were highly developed; (iii) they were used as spoken as well as written vehicles of communication;40 (iv) there have been native speakers of the four languages in Palestine; and (v) the morphophonemics of all four languages are complicated. For these reasons, Paulston is justified in stating: “I find very little explanatory power in the concept of diglossia when applied to first-century Palestine” (2000:88-9). Accordingly, the features of the four languages of firstcentury Palestine do not require the model of diglossia, but rather bilingualism.

2.3.2 Language Preference Theory According to recent studies of bilingualism in ancient societies, some linguistic situations also show that they cannot be explained by the simple diglossic scheme of H- and L-varieties (Adams 2002:9). The theory of “language preference” should first of all be explained. Language preference theory corresponds to the concept of bilingualism rather than to that of diglossia. Making a comparison between “bilingualism (on the part of psychologists) and diglossia (on the part of sociologists) (1967:29),” Fishman (1967:34) rightly points out that “bilingualism is 40

Aramaic and Hebrew were used as written languages as well (§2.3.4; §4.3.1). Hebrew could have been spoken in religious situations, to some extent.

98

Bilingualism and Diglossia

essentially a characterization of individual linguistic behavior, whereas diglossia is a characterization of linguistic organization at the sociocultural level” [emphasis added]. In other words, it may be said, in Chomskyan terms, that bilingualism is concerned with “language performance,” whereas diglossia is concerned with “language competence.”41 In bilingualism, a bilingual person and/or a bilingual community can choose a language among their repertoire. Romaine (1995:8) states, “Bilingualism exists within cognitive systems of individuals, as well in as families and communities.” Contrary to this, in the diglossic scheme, there is a certain degree of determination for using a particular language (i.e. H-language or L-language) at the socio-cultural level. David Taylor questions the diglossic dichotomy through his study of the linguistic situation of Palmyra. According to Taylor (2002:318320), the use of Greek and Aramaic in the Hellenistic city of Palmyra cannot be categorized into a simple diglossic scheme because both languages were used as H-varieties, thereby showing that both had equal public status. Furthermore, it is intriguing that although Greek was used in the context of public activities, Aramaic almost always was the language of funerary inscriptions.42 Taylor (2002:319) assumes that “this phenomenon is related to the psychology of identity … at Palmyra Aramaic was considered the appropriate language for the linguistic domain of religion” [emphasis added]. Thus, the distribution of language at Palmyra indicates social language preferences rather than a uniform diglossic opposition between H- and L-languages. In relation to Greek dialects, it was said among ancient Greeks that Ionic is for history, Doric for the choral lyric, and Attic for tragedy.43 This does not mean that Ionic is an H-variety, whereas Doric is an Lvariety, but that each dialect is used in a different (literary) context. This linguistic distribution shows the contemporary Greek impressions 41

42

43

In the same vein, as Taylor (2002:300) mentions, in Saussurean terms, it can be said that “parole” is related to the concept of bilingualism, whereas “langue” is related to that of diglossia. Taylor 2002:319; he adds, “I know of only one funerary relief with Greek descriptions of the deceased (Colledge 1976:pl.105), and its provenance is uncertain.” However, there are several exceptions. Ingholt (1935:57-120) suggests that four dated Palmyrene tombs of ‘Atenatan (98 AD), Malk‫( ܖ‬115 AD), Nasrallat (141 AD), and Bar‘â (186 AD) are engraved with bilingual inscriptions in Greek and Palmyrene. Ingholt (1938:114, 121) also suggests that the tomb of Lisams (186 AD) has two lines Greek inscriptions and the tomb of ‘Abd‘astôr (98 AD) is engraved with bilingual inscriptions in Palmyrene and Greek. Haugen 1972:98. Also refer to Charles V’s linguistic competence concerning his language preference and language choice between bilinguals in the Jerusalem Talmud (§8.1.1).

Bilingualism and Diglossia Models

99

of those dialects. In other words, this also indicates that the Greek dialects were related to the domain of language preference rather than to a uniformly diglossic dichotomy. Language contact brings about some linguistic phenomena such as borrowing, interference, and codeswitching, as will be discussed below (§8.1.1). The diglossic model can explain borrowing and interference but not codeswitching. Codeswitching in the New Testament literature can be explained most easily on the basis of the bilingualism model. One of the reasons that the study of codeswitching has not been discussed among biblical scholars results from the application of the model of diglossia. This will be argued comprehensively (chapter 8).

2.3.3 Maximalism and Minimalism In regard to lower-status and rural bilingualism, first-century Palestine has been viewed on a scale ranging from maximalism to minimalism.44 Minimalism is often related to the model of diglossia. This association warrants examination alongside the general association of the maximal view with bilingualism. Maximalists have proposed that first-century Palestine was largely bilingual.45 They generally grant both lower-status bilingualism and rural bilingualism. Regarding lower-status bilingualism, a maximalist takes Josephus’ claim (Ant. 20.264) seriously that it is easy for even slaves to learn Greek.46 Sevenster (1968:70) points out that

44

45

46

Kokkinos (1998:79-84) uses the two terms, maximalist and minimalist. Furthermore, van der Horst (2001:166) also calls Feldman, Rajak, Grabbe and the new Schürer minimalist interpreters, when he discusses influence of Greek over first-century Palestine. “Largely” is cited from Hengel 1989a:14-5; van der Horst 2001:166. The relationship between some terms should be mentioned in discussing the bilingualism of Palestine, for biblical scholars had discussed the linguistic milieu of first-century Palestine for quite a long time before the term, bilingualism, was established by sociolinguists in 1960s. Linguistically speaking, some concepts such as “Hellenization of Palestine” and “the use of Greek in Palestine”, which scholars have employed, can be transposed into “bilingualism of Palestine” due to conceptual continuities. For, it is generally said that, as the verb hellenizein originally means “to speak Greek,” the use of Greek is regarded as one of major characteristics of Hellenization. Goldstein (1981:67-9, 318-9) enumerates six characteristics of Hellenization and suggests that “our six traits are peculiarly Greek.” In this sense, the use of Greek in Palestine where Aramaic had been used as the primary language refers to being bilingual by and large; refer to Tcherikover 1959:344-57; Smith 1971:43-61; Barclay 1996:88-91. I will deal with this in more detail in §3.2.6. Also refer to Sevenster 1968:69-70. Even Feldman (1986:91-2), who is one of the representative minimalists, consents that many slaves were bilinguals because it was slaves rather than their masters who

100

Bilingualism and Diglossia

“everyone in the Jewish country had the chance of speaking [Greek, and it] ...could evidently be heard in all circles of Jewish society.” Lieberman (1965:39) too asserts, “The Greek language took hold of all classes among all the nations in the Mediterranean world. The Jews were no exception in this respect. We have already seen how deeply Greek penetrated into all the classes of Jewish society in Palestine.” In terms of rural bilingualism, Morton Smith (1971:67), for example, proposes that “The contrast commonly drawn between the Greek cities and the Semitic countryside has been exaggerated. The countryside was permeated by Greek elements and influences.” Van der Horst (2001:166; 1991:130) also asserts, “For most, or at least many, of the Jews in Palestine, Greek most probably remained a second language, certainly outside the urban areas.” Along these same lines, Hezser (2001:231) writes, “Greek was the language in which the Romans communicated with and issued decrees concerning the local Jewish population, not only in the Diaspora but in Palestine as well.” Maximalists hold to both “urban and upper-status” and “rural and lower-status” bilingualism. Minimalists propose partial bilingualism, thus confining the bilingualism of Palestine to urban and upper-status settings. They have explored the possibility of lower-status bilingualism. Schürer (2.74), for example, contends, “It is probable that the lower levels of Palestinian society had either no more than a limited acquaintance with it (Greek language) or none at all.” Tessa Rajak (2002: chapters 1, 2, and 7) also proposes that Greek was used among some upper-status Jews. More recently, Lester Grabbe (1992:1.158) concludes that “the use of Greek seems to have been confined to a particular segment of the population, namely, the educated upper-class.” Minimalist views on rural bilingualism are expressed by other scholars. Eric Meyers raises the idea of “Galilean Regionalism.” 47

47

dealt with business transactions and would have even served as their interpreters. However, it is intriguing that he denies upper-status bilingualism; see chapter 2 n.47. The proponents divide the linguistic situation in Galilee into Upper Galilee and Lower Galilee on the basis of Josephus’s evidence (Bell. 3.38-40). While Upper Galilee was rural, less Hellenized and Aramaic and Hebrew were used, Lower Galilee was urbanized, much Hellenized and bilingual in Greek and Aramaic. See Meyers 1976:93-101; Meyers & Strange 1981:91; Meyers 1985:5.115-131. Recently, many scholars have refuted this hypothesis; Freyne 1980:138-145; Richard Horsley 1996:88106; Crossan 1991:15-19. Ironically, Feldman (1986:93) does not follow the prototype of minimalism, when he remarks that “Greek was the language of the upper class, … whereas Aramaic was spoken by the uneducated, particularly in the rural areas. But … no such clear-cut distinction is defensible.” In this respect, he asserts lower-status bilingualism (see chapter 2 n.46). However, he (1986:94-5) conforms to the Galilean

Bilingualism and Diglossia Models

101

Goodman (1983:67) accepts the Galilean Regionalism hypothesis and adds, “Greek is found in use in some parts of Galilee.”48 In the case of Phoenician cities like Tyre and Sidon, Grainger (1991:77-83, 108-11) also thinks that the use of Greek does not necessarily apply to many small towns and villages. Returning to Grabbe, we note that he (1991:1.158, 1.170) suggests that “The number of Jews outside the Greek cities who were fluent in Greek seems small.” Minimalism can be summarized by Fergus Millar’s expression in his influential article: “‘Hellenization’ might, as is often supposed, have extended very little outside the towns or the upper classes.”49 The diglossic model is related to the minimalist view of Greek influence over first-century Palestine. Applying diglossia to the linguistic milieu of first-century Palestine, the diglossic model presupposes that the linguistic milieu can be categorized into H- and L-languages. In other words, minimalists assume that only upper-status and urban Jews used Greek (i.e. H-language), whereas lower-status and rural Jews used Aramaic (i.e. L-language). 50 Furthermore, the minimalist view seems to hold one criterion, that is, formal education. Since lowerstatus or rural Palestinians could not receive Greek formal education, they spoke Aramaic. However, formal education is only one out of six factors in bilingualism (§3.2). Although formal education is a significant factor, others like bilingual parents, bilingual region, and occupation could have contributed to the creation of Palestinian bilinguals (§3.2). In this respect, recent evidence strongly supports the maximalist view that both “upper and lower-status” Jews and “urban and rural” Jews used Greek. Regional bilingualism and personal bilingualism support

48 49

50

Regionalism Hypothesis. From the perspective of personal bilingualism (§3.2), it is persuasive that Feldman approves lower-status bilingualism, although it is problematic that he denies upper-status bilingualism. Furthermore, in relation to the Hellenization of Palestine, Feldman retorts that various aspects indicate that Palestinian Jews were not as much influenced by Hellenization as Hengel argued; Feldman 1977:371-82; 1986:83-111; 1993:3-44. However, it seems proper that Grabbe (1991:1.151), another minimalist, rightly criticizes that Feldman has two major flaws; “first, he seems to make a strong, underlying assumption that being Hellenized means ceasing to be a proper Jew. Second, his arguments against Hengel often depend on interpretations that would not be accepted by the majority of specialists.” He (1983:68) also mentions that “in Upper Galilee and probably in the area around Lake Tiberias, Greek was only a thin strand in the linguistic cloth.” Millar 1987b:132. Originally, although Millar was a minimalist, since his publication of 1993 he has changed his position from minimalism into maximalism; see chapter 3 n.12. This is related to Fishmanian model and Ferguson’s new definition; see Fishman 1967; Ferguson 1991:215-6.

102

Bilingualism and Diglossia

maximalism, as will be discussed below (chapter 3). Accordingly, it will be argued that the diglossic model does not fit with the description of the linguistic milieu in question. At this point, it is necessary to further explore the relationship between bilingualism and the discussion of maximalism and minimalism. Recently, minimalists have considered that bilingualism actually supports minimalism. As will be discussed (§4.1.1), most regions of the first-century Roman Empire were bilingual and used vernacular language(s) in addition to Greek. In relation to this, some scholars argue that the survival of vernacular languages indicates that the linguistic milieu of Palestine was partially bilingual. Millar (1983:55-71; 1987a: 143-64; 1987b:110-33) makes a strong argument that the survival of vernacular languages is evidence for partial bilingualism. Grabbe (1991:1.150-8; here, 1.156) also proposes that “although it is often asserted that Greek became the official language of the conquered territories, this seems mistaken: the Seleucid Empire was multilingual, and local languages continued to be used in official documents.” Feldman likewise argues: In language, as in culture generally, the degree to which Hellenism spread after Alexander has been much exaggerated. Thus, even in the most heavily Hellenized portions of Syria, Phoenicia, and Cyprus, bilingual inscriptions and coins for this period are common. In Antioch, the capital of the Seleucid Empire, for example, Aramaic remained as the second language and continued thus even after the Roman conquest. Hence, if the Greek language emerged clearly triumphant in the Land of Israel, this would be the exception to the general pattern.51

Indeed, minimalists consider bilingualism to be strong evidence that the influence of Greek over first-century Palestine is not hugely significant. It is ironic then that bilingualism, which resulted from the survival of vernacular languages, as suggested in Millar’s well-known article (1987b), is used as the evidence for both maximalists and minimalists. Indeed, Hengel (1989a:64 n.23) cites Millar’s argument when he proposes a maximalist view, while Grabbe (1991:1.156-8) employs the same argument to support his minimalist view! As discussed in detail below (§4.1.1), bilingualism is not a zero-sum game. Schwartz’s remark that the rise of vernacular languages does not generate a decline of Greek seems likely (§4.1.1). Bilingualism means that two or more languages coexisted. Therefore, the bilingualism of first-century Palestine is more compatible with the maximalist view than with the minimalist view. Furthermore, maximalism would also 51

Feldman 1993:14. Here, Feldman cites the bilingualism presented by Avi-Yonah (1978:182) and by Green (1990:313).

Bilingualism and Diglossia Models

103

have been supported by regional (§3.1) and personal bilingualism (§3.2) in first-century Palestine.

2.3.4 Biliteracy of First-Century Palestine As mentioned (§1.2), many scholars including the religionsgeschichtliche Schule consider that the Aramaic tradition was oral, whereas the Greek tradition was written, so that there were no Aramaic Gospels (see chapter 1 n.24). In other words, the Jesus tradition was transmitted from oral into written form, that is, from Aramaic into Greek. As mentioned (§2.2.3), Lapide applies the diglossic model to this linguistic milieu. He (1975:486-92) assumes that there were not many Aramaic writings around first-century Palestine because Aramaic was an L-language (i.e. spoken language), but Hebrew was an H-language (written language). In this respect, the diglossic model coheres well from the perspective of unidirectionality hypothesis (from Aramaic into Greek and from oral into written). However, as is well known, Aramaic literatures of first-century Palestine continue to be excavated, and these will be investigated in detail (§4.3.1).52 The existence of the Aramaic literature would seem to support the biliteracy model. This indicates that the Aramaic Jesus tradition was usually translated into Greek among Greek-matrix speakers and the Greek tradition was usually retranslated into Aramaic among Aramaic-matrix speakers. 53 Consequently, the Jesus tradition could have been circulated both in Aramaic and in Greek as well as both in oral and written form during and after Jesus’ ministry. A bilingual model supports the view that the transmission of the Jesus tradition was not unidirectional but interdirectional.

52

53

Butcher 2003:283-9; Gamble 1995:3; Millar 1993:5; Fitzmyer 1979a:39; Hitti 1951:257. For Aramaic literature excavated from Dura-Europos, see chapter 3 n.2 and §4.3.1. Also, attention should be paid to the testimony of Papias. He (HE III.39) mentioned that “Matthew compiled the oracles in the Hebrew [Aramaic] language.” Whether it refers to a note or his gospel, Papias attests that Aramaic could be used as a written language. Also, for detailed discussion of the ancient literacy of first-century Palestine and the Roman Near East, see Millard 2000:154-84, 204-9. Interestingly, James Edwards (2009) proposes that the apostle Matthew wrote the Hebrew Gospel, which was widespread and widely known in Early Christianity. It is plausible that Hebrew Gospels could have existed, but I regret that his argument is still based on the linguistic unidirectionality. Interestingly, Zuntz (1972:212) insists that Codex Bezae was of Syrian origin and, then, the Greek text was rewritten in Syriac at Edessa in early second century. This view is followed by Ellis (1998:91).

3. Bilingualism of Jews in First-Century Palestine As noted in chapter 1, most New Testament scholars have adopted three unidirectional views with regard to the transmission of Jesus and gospel traditions. The direction is from Judaeo-Palestinian to Hellenistic tradition, from oral into written tradition, and from Aramaic into Greek tradition. Such unidirectionality hypotheses, however, fail to consider the local impact of bilingualism. If we seriously consider bilingualism in first-century Palestine (chapter 3) and the Roman Empire (chapter 4), we are no longer in a position to draw sharp distinctions between Judaeo-Palestinian and Hellenistic traditions, between oral and written traditions, and between Semitic and Greek traditions in the course of the transmission of Jesus and gospel traditions. The bilingualism of Palestine (chapter 3), Alexandria (§4.2), and Antioch (§4.3) is mainly related to transmission of gospel tradition. The bilingualism of Palestine (chapter 3) and the Jerusalem church in the first century (chapter 5) is mainly related to the transmission of the Jesus tradition. Although the linguistic milieu of first-century Palestine has been considered complicated, there is no doubt that the linguistic milieu was largely bilingual. Concerning the bilingualism of first-century Palestine, regional bilingualism will be discussed (§3.1) as well as personal bilingualism (§3.2). Last, the relation between the bilingualism of firstcentury Palestine and the transmission of Jesus and gospel traditions will be considered (§3.3).

3.1 Regional Bilingualism Recent evidence concerning bilingualism in first-century Palestine supports the maximalist position. Three points will be considered: inscriptions (§3.1.1), papyri (§3.1.2) and population geography (§3.1.3). Whereas inscriptional and papyrological testimonies usually show some literate bilingualism, population geographical evidence exposes the linguistic milieu of the whole community, including non-literate bilinguals.

106

Bilingualism of Jews in First-Century Palestine

3.1.1 Inscriptional Evidence The excavation of Greek inscriptions from first-century Palestine is strong evidence that there were many Greek-speakers.1 The use of the language was not due to a diglossic phenomenon but to the choice of inscriptional language (Taylor 2002:319).2 According to Frey (CIJ 2.113– 339), in relation to Jewish inscriptions in Palestine, approximately 52 percent are in Greek, 40 percent in Hebrew/Aramaic, and 8 percent are bilingual, written in both Greek and Hebrew/Aramaic. On the basis of Frey’s CIJ (326) and later publications, SEG xi – xiii (114), Mussies (1976:1042–3) suggests that 440 Jewish inscriptions were found in Palestine proper: 69 in the cemeteries of Joppa (2nd –3rd century), 196 in BethShearim (1st–4th century) and 90 on the ossuaries of Jerusalem (2nd century BC–2nd century AD). On the basis of these data, van der Horst suggests that “The overall average of Greek inscriptions is slightly more than 53 percent.” It is striking, he (2001:157; cf. 168 n.16) adds, that “even though in the past 65 years the material has more than tripled, the numerical ratio of Greek and non-Greek material has not changed at all!” Hengel suggests that of 23 ossuaries of Akeldama tombs excavated in 1989, 13 have Greek and 5 Hebrew and 5 are bilingual.3 This indicates that around 80 percent of ossuaries were Greek or bilingual. 1

2

3

In relation to major objections to the validity of the study of epitaphs, Clauss (1973:411) raises a question: how can less than 1 percent of the epitaphs of the population represent the linguistic milieu of the whole society? Van der Horst (2001:15965) persuasively replies that the epitaphs from Palestine produce better statistical data because they were excavated from different regions (urban and rural areas), from different social status (upper and lower) and from various places. The language choice of inscriptional language can be related to its religious language. Kaizer (2004:180-1) mentions that although Palmyra was a bilingual city, the funerary context is dominated by Aramaic and that many particular aspects of the Palmyrene cults were originally expressed in Palmyrenean.” Taylor (2002:319) also points out that the reason that Palmyrene was chosen as a funerary language is that “at Palmyra Aramaic was considered the appropriate language for the linguistic domain of religion.” As will be mentioned (§4.3.1), it is interesting that although parchments and papyri from Dura-Europos (ca 165 – 255 AD) were written in Greek, Latin, Aramaic, Parthian, Middle Persian, and Syriac, liturgical text is in Hebrew. Furthermore, many papyri excavated from Jewish synagogues were written in Aramaic, although Greek was the predominant language in Dura as a whole. Kilpatrick (1964:218) concludes that “the simplest and most probable answer is that the Greek element in the congregation of the synagogue was so much the smaller and weaker.” Hengel 2001:27 (cf. Avni and Greenhut 1996:57–72; esp. 66, n.19). Intriguingly, Hengel adds that “the most interesting, number 19, mentions an Ariston from Apamea in Syria, a proselyte with the Hebrew name Juda, who is probably mentioned in the Mishna Halla 4:11 bringing first fruits from Apamea, which was ac-

Regional Bilingualism

107

Moreover, according to Rahmani (1994:12), 233 inscribed ossuaries from the environs of Jerusalem, between the time of Herod the Great and 70 AD, have been excavated. These include 143 Jewish scripts only, 73 Greek only, 14 or 15 bilingual, 2 Latin and 1 Palmyrene. Bilingual ossuaries are inscribed in Greek and Jewish scripts. Main texts are in Greek with a repeated summary in Hebrew. This means that 87 or 88 out of 233 are Greek. From these inscriptions, Rahmani (1994:13) also proposes that “in and around Jerusalem and Jericho even the lower classes of the Jewish population knew some Greek.” A more recent and fascinating piece of evidence is the Ordinance of Caesar, which is from first-century Nazareth and presently housed in the Bibliothèque Nationale, Paris. This inscription contains an imperial edict inscribed in Greek (cf. Van der Horst 1991:159-162; 2001:162). Van der Horst (2001:162) suggests that the Ordinance clearly indicates that “A good many of the local Galileans were expected to be able to read it.” What is striking is that numerous Greek funeral epitaphs in Jaffa and Beth Shearim prove that Greek was used by lower-status groups as well as by upper-status ones. It seems that the Greek epitaphs excavated from Beth Shearim were not influenced by school education. Rather, Schwabe and Lifshitz (1974:220) suggest that “It does seem as though the authors of the inscriptions learned their Greek from their pagan neighbors and knew how to speak it, but only seldom did they have a broader educational background.” As Sevenster (1968:182; cf. Lieberman 1965:30) points out, “the simultaneous occurrence of tidy, correct and clumsy, primitive inscriptions in Greek proves that this language was used in widely divergent layers of the Jewish population in Palestine.” Indeed, Hengel (1989a:10) is more than justified in asserting that Greek was used “by lower classes of the local Jewish population.”4 Fitzmyer (1992:60; cf. 1992:77 n.21) suggests that ossuaries “testify to the widespread use of Greek among first-century Palestine Jews at all levels of society.” Van der Horst (2001:163) also asserts that this shows that Greek was not restricted to the upper-status population in Palestine. Van der Horst (2002:15-16) writes: There is ample evidence that the epitaphs in Greek represent a wide stratum of the population. There is a great difference between a metrical epitaph in Homeric hexameters engraved upon luxurious and expensive sarcophagi with elaborate decorations on the one hand and poorly scratched

4

cepted by the priests for they said: He that owns [land] in Syria is as one that owns [land] in the outskirts of Jerusalem.” Hengel (1980:53) also asserts that “a more thorough ‘Hellenization,’ which also included the lower classes, only became a complete reality in Syria and Palestine under the protection of Rome …”

108

Bilingualism of Jews in First-Century Palestine

names on potsherds or wallplaster that marked the graves of deceased on the other (and we have many of the latter sort). ... To be sure, ... there are numerous very simple and poorly executed tombstones with inscriptions in poor Greek that undeniably stem from these lower strata of Jewish society.

Accordingly, inscriptional evidence proves that the bilingualism of first-century Palestine is rural bilingualism and lower-status bilingualism.5

3.1.2 Papyrological evidence Greek papyri excavated from rural areas in Palestine also strongly support the view of maximalism. One should consider, first, the Babatha archive to indicate village life in En-Gedi in Judea and Maoza, Zoara district in Arabia. These manuscripts, hidden at the time of the Bar Kokhba Revolt (ca. 132), date between c. 94 and 132 AD. Yadin (1971:229) mentions that the thirty-five documents were written in Greek, Aramaic and Nabataean: “Six in Nabataean, three in Aramaic, seventeen in Greek, and nine in Greek with subscriptions and signatures in Aramaic or Nabataean.” These documents can be considered to be typical of the bilingualism in those times. The Babatha archive displays the interaction between villagers and Roman officials within a rural context. Sevenster (1968:185) rightly points out that “It should not be forgotten that the inhabitants everywhere came into contact more or less frequently with government officials. In order to understand legal regulations pertaining to contracts, taxes, census, a knowledge of Greek was necessary, or in any case expedient.” Secondly, caches of Greek documents found in Waddis Murabaat, Seiyal and Nahal Hever indicate rural bilingualism, especially the Greek Minor Prophets Scroll from Nahal Hever (cf. Tov 1990). Tov (2001:5) suggests that “all the Greek texts (and in Wadi Murabaat and Masada, the great majority) are documentary, showing that Greek was actively used among the persons who deposited the texts.” 6 Special

5

6

Furthermore, numismatics shows that the influence of Greek was large in Palestine. Kokkinos (1998:81-2) proposes that “the numismatic evidence suggests that in the late Persian period Jewish society was exposed to Hellenic pressure.” Cf. Mussies 1976:1044-5. Concerning Greek documentary texts from Nahal Hever, Cotton (DJD 27.154) notes that “these non-Hellenized or semi-Hellenized Jews chose to write their contracts in Greek either because of the absence of Jewish courts and archives using Aramaic as the official language in the places mentioned in the papyri, or out of the desire to

Regional Bilingualism

109

attention should be paid to the relation between bilingualism and administration. It seems that Greek was known to residents in rural areas as the language of administration. Concerning the Greek documents in rural areas, Freyne (1980:139) explains that the new administrative and business personnel “were not confined to the cities but were distributed throughout the villages and estates in charge of the affairs of the government.” 7 Elsewhere Freyne (1980:139) writes that “the frequent journeys of these officials, some of higher, others of lesser rank, ensured a network of communication that tied village life to the various cities and touched everybody from the poorest peasant to the various village officials.” Bickerman (1962:59) suggests that “The Jewish territory itself was crowded with Greek officers, civil agents and traders, as the papyri show.” Hezser (2001:231) observes that “Greek was the language in which the Romans communicated with and issued decrees concerning the local Jewish population, not only in the Diaspora but in Palestine as well.” Butcher (2003:277, 284-5) also proposes that “The Greek language and Greek cultural forms and practices penetrated beyond the high culture of the cities into rural environments as well.” Accordingly, the documents found in rural areas support rural bilingualism. Third, it has been noted that twenty-five Greek papyri were found among the DSS, which also supports the maximalist view. Lim (2000:69) suggests that most of the Greek texts from caves 4 and 7 appear to be written in Septuagintal Greek.8 Greek texts from Qumran and the Judaean wilderness evidently show that Greek works were read in Judaea and elsewhere.9 Above all, the fact that a sectarian community which was xenophobic and highly wary of Gentile culture used

7

8

9

leave open their option to go to the court of their choice.” These texts support the contention that Greek was used in rural areas. In addition, Lewis shows the demise of the Demotic is due to the change of administration language (chapter 3, n.71). Polak (2006) also suggests that the main reason for language change from Hebrew to Aramaic in the Achaemenid Empire can be explained by the fact that Aramaic was used as an administration language. This shows that the change of the administration language promotes the substitution of a living language. In relation to this, following Jose O’Callaghan, some scholars may assert that the papyri are the NT passages; see Thiede 1992. On the other hand, some scholars consider them as fragments from the books of 1 Enoch; see Nebe 1988; Puech 1996; Muro 1997-1998. However, Nickelsburg (2004:634) persuasively concludes that “The identification of these fragments as the remnants of the Epistle of Enoch is as unproven as are previous attempts to assign them to the New Testament.” Hengel 1974:102; Fitzmyer 1990:11-3; Porter 1997b:293-316.

110

Bilingualism of Jews in First-Century Palestine

the Septuagint for religious purposes10 implies that Greek was broadly used in first-century Palestine. A colophon of Greek Esther (10:3l), the Book of Esther that was translated into Greek in Jerusalem in firstcentury BC, indicates that Greek literature was read among some Jews.11 Hengel (1974:100-2) states that some parts of I and II Maccabees were translated into Greek in Jerusalem. Also Lieberman (1965:30) suggests that “In the Hellenized town of Caesarea there were Jews who read the Shema in Greek.” After surveying an extensive collection of papyri from the Roman Near East, Cotton, Cockle, and Millar conclude that “a considerable proportion of the documents listed here do not emanate from cities, but from country districts characterized by villages or small towns.”12 Accordingly, papyrological evidence also suggests that the linguistic situation of first-century Palestine was largely bilingual.

3.1.3 Population Geographical Evidence In terms of rural bilingualism, a look at the population geography of first-century Palestine is also a valuable tool in describing the bilingual situation from a different angle. Whereas inscriptional and papyrological evidence usually imply the bilingualism of the literate of firstcentury Palestine, population geographical evidence indicates the bi10

11

12

Tov (2001:5) argues that “all of these [papyri of Cave 7] were brought directly to the cave from an archive outside Qumran or from a specific spot within the Qumran compound. … There is no reason to believe that any of these texts was penned down in Qumran.” However, topographically speaking, especially Cave 7 can be thought to have belonged to the Qumran site itself because geographical features of the district show its seclusion. A protective Wall runs south to the edge of the marl terrace from the southeast corner of the building complex at the base of high limestone cliffs. Bickerman 1944:339-62. Bickerman (1976:1.246-74; here, 258) mentions that “The book gives a quite favorable impression of the Greek used in Jerusalem in the time of Alexander Jannaeus.” Also he notes that, after the Greek version of the Torah, other Jewish books were translated into Greek; for more information, see Bickerman 1976:1.137-66, especially, 147-50. Cotton, Cockle, and Millar 1995:214-235; here 235. In his personal email communication of 20th Feb. 2006, Millar suggests that “the evidence both of inscriptions and of papyri and parchments (Cotton, Cockle, and Millar in JRS 1995) shows that everywhere in the Near East, where writing was in use at all, Greek was also used.” I appreciate his thoughtful reply and his permission to use our email correspondence. He replied to me that “I could certainly count myself as a ‘maximalist’ and rather more so even than in 1993 (his emphasis).” Although Millar had become one of the representative minimalists through his influential publications, he has changed his position since his publication in 1993.

Regional Bilingualism

111

lingualism of all social statuses (i.e. both the literate and the illiterate) due to the language contact ensuing through living in a border area, as will be discussed (§3.2.4). One may point to the suggestion of Millar that we do not draw a clear-cut distinction between Hellenistic cities and Jewish cities in the first-century Palestine. 13 Furthermore, even Jews and Greek-speaking Gentiles mingled in the cities of Palestine.14 Mussies (1976:1057-8), too, makes the relevant comment that “the Jewish area is seen to be discontinuous. … a number of Hellenistic towns were scattered over the remaining territory.” The discontinuity of population geography could cause Jews in rural areas near Hellenistic cities to learn Greek regardless of their formal education. Mussies names the following Hellenistic cities: the entire coastal strip from Raphiah to Ptolemais, Decapolis, Gadara, Pella, Scythopolis/Beth Shean,15 Gerasa, Philadelphia, Phasaelis in Judaea, Sepphoris and Tiberias in Galilee, Caesarea-Philippi and Bethsaida-Julias 16 in Batanaea, Heshbon and a second Julias in Peraea and Sebaste in Samaria.17 In this respect, Jews who lived in a border area would have obtained bilingual competence (§3.2.4). Mussies (1976:1058) proposes that “If we may so call the remaining Jewish country, which is so strongly fragmentated and has so many enclaves, the knowledge of Greek must have been much more widespread than the mere presence of some Greek schools and synagogues in Jerusalem might suggest.” The surrounding rural areas of the cities mentioned may be regionally 13 14

15 16 17

Millar 1993:343. For the interaction between Jews and Gentiles of Roman Galilee in the second century, see Goodman 1983:27-89. Millar 1993:343. Hengel and Schwemer (1997:281) suggest that “in Palestine and adjacent regions there are numerous significant ‘Hellenistic’ cities with a mixed population.” In addition, Hengel (1991:55) calculates that “around 10-15 % of the then inhabitants of Jerusalem spoke Greek as their mother tongue.” And he adds that “in a population of about 100,000 in Jerusalem and its surroundings that would give a population group of around 10,000 to 15,000.” If this is the case, Greek-matrix speakers should have made a contribution to the bilingualism of Jerusalem. Furthermore, he (1991:55) mentions that tens, indeed hundreds, of thousands of Greekspeaking Diaspora Jews streamed into, and some stayed in, Jerusalem during the festivals. Pilgrimage also made Palestinian Jews bilinguals to some extent. cf. Millar 1993:378. cf. Dalman 1935:165. Mussies 1976:1058. Concerning the bilingualism of Samaritans both inside and outside of the Samaritan homeland, van der Horst (1998:49-58; here 58) concludes that “This would make it exactly parallel to the Jewish situation in the later Roman Empire, where there was a Hebrew-speaking minority in the land of Israel, and a Greek speaking majority in the diaspora, as well as a large number of Greek-speakers in the homeland itself.” For Bickerman’s list of the Hellenized cities of Palestine, see 1962:58; for Hengel’s list, see 1989a:14-5; for Millar’s list, see 1993:374-386.

112

Bilingualism of Jews in First-Century Palestine

bilingual due to the influences of nearby Hellenistic cities without reference to upper-status or lower-status. Accordingly, population geographical evidence in the first-century Palestine supports lower-status and rural bilingualism. In summary, the epigraphical, papyrological and population geographical evidence shows both “urban and upperstatus” and “rural and lower-status” bilingualism.

3.2 Personal Bilingualism Fishman (1967:34) supports the individualistic nature of bilingualism, which is “essentially a characterization of individual linguistic behavior“ [emphasis added] and stands in contrast to diglossia which is “a characterization of linguistic organization at the socio-cultural level.” In this respect, regional bilingualism does not always correspond to personal bilingualism (Beardsmore 1982:30). Although the linguistic milieu of first-century Palestine should be viewed from a maximalist position, it is necessary to consider personal bilingualism in first-century Palestine with regard to the proper individual characteristics of bilingualism. Above all, in relation to personal bilingualism, several points should be kept in mind. First, personal bilingualism is clearly a matter of degree (Beardsmore 1982:10). Its spectrum is broad; some individuals have more or less full command of two languages, whereas others speak them at varying levels of competency (cf. §2.1.2). Second, those who acquired a bilingual ability within the Roman Empire may have used Greek for business, but their mother tongue at home. Third, even primary bilinguals favor one language over another. Even though a primary bilingual obtains the mastery of two languages fully, only one language functions as a matrix language, while the other acts as an embedded language (cf. §2.1.9). Fourth, an acquired bilingual’s linguistic competence in the embedded language can nevertheless be developed in close proximity to the matrix language. Let us turn our attention to factors of bilingualism in relation to personal bilingualism of first-century Palestine. A study of bilingualism should be undertaken alongside the circumstance of bilingualism. Hakuta (1986:4) makes the point that “the study of bilingualism should include not only the study of the bilingual person but also the circumstances surrounding the creation of bilingualism and its maintenance or attrition.” Myers-Scotton provides six factors that cause bilingualism,

Personal Bilingualism

113

which merits attention here.18 Adapting, and adding to, these six categories, we can enumerate the following causes of bilingualism in order to apply them to personal bilingualism at the turn of the eras. Factors One and Two. (1) “Military invasion and following colonization” and (2) “migration and inflow.” These two factors may be applied to every social status without regard to the level of social status (cf. §3.1.3). Furthermore, when we see language contact through the history of Greeks, Romans, and Jews, these two factors account for it. Various kinds of bilinguals such as “early,” “late,” “primary” and “acquired” are thus produced by the two factors. Factor Three. (3) “Ethnic awareness.” This factor is related to the loyalty to one’s nation. It depends on an individual’s devotion to Judaism. Factor Four. (4) “Bilingual parents and bilingual region.” These are the typical factors that cause “early” and “primary” bilinguals. Bilingual parents largely contribute to personal bilingualism, whereas the bilingual region is mostly related to regional bilingualism. For instance, if anyone is born and/or grows up in a bilingual circumstance, whether his/her parent is a bilingual or not, the individual would most likely be an early bilingual or a primary bilingual. Factors Five and Six. (5) “Formal education” and (6) “occupation.” These produce late bilingualism, secondary bilingualism, or semibilingualism. Formal education and occupation are usually associated with oral and written bilingualism.

3.2.1 Military Invasion and Colonization The most far-reaching language contact was caused by military invasion and the colonization that followed, as noted in detail (chapter 4). One may observe the report of the return from exile in Nehemiah (13:24) where their children could not speak the Jews’ language, but rather acquired Ashdod’s language. This indicates that their children were bilingual. Both Alexander the Great and the Roman Empire spread Greek and Latin throughout Palestine and the Middle East. As a result, Greek gradually took over from Aramaic in Palestine as the common language. Having been used from approximately the eighth century BC to the time of Alexander the Great, Aramaic also sup18

Myers-Scotton 2002:31-3; military invasion and subsequent colonization, living in a border area or an ethnolinguistic enclave, education as a factor in bilingualism, spread of international languages, ethnic awareness, and migration for social and economic reasons.

114

Bilingualism of Jews in First-Century Palestine

planted Hebrew as the ordinary language of most of the Jews in Palestine and Babylonia by the beginning of the Common Era (Sawyer 1999:14).

3.2.2 Migration or Inflow Migration or inflow for social and economic reasons would have been conducive to bilingualism. The military conquests of Alexander the Great gave rise to thousands of Greek immigrants to Egypt, Syria, Phoenicia, Palestine, Asia Minor and the region of Persia.19 Most conquests consisted of thousands of soldiers and civilians of various occupations, status and nationalities (Bubenik 1989:54-5; Tcherikover 1959:115). This Hellenistic Diaspora spread Greek to areas where it previously had little or no status. Jewish immigrants from Palestine into the regions of the Roman Empire caused Jews to become bilinguals in Aramaic and Greek in the Greek-speaking world (chapter 4). Furthermore, bilingualism in Jerusalem during the Second Temple period was also accelerated by the Jewish Diaspora (§4.1.3, §4.2.4). Diaspora Jews first spread to most countries in the Near East and the Mediterranean and then, in turn, returned to visit their ancestral land for such reasons as cult and religious pilgrimage (Rabin 1976:1007). Hence, one of the direct causes of widespread bilingualism in communities was migration or inflow.

3.2.3 Ethnic Awareness Using a mother tongue is often regarded as an important sign of one’s ethnic identity. Ethnic awareness purposefully attempts to preserve its language, thus creating a bilingual ethnic group. 20 In modern times, Welsh, Gaelic and Breton are good examples of this (Myers-Scotton 2002:33). Schwartz (2005:53) mentions that immigrant groups like most Amish or Hasidim may preserve their ancestral languages far longer due to a rigorously separatist religious ideology. For instance, the inscriptional language of Palestinian coins may reflect the Jewish attitude to Hellenism, since, as we can learn from them, the independent Jewish state in the first Jewish war minted coins in the Palaeo-Hebrew lan-

19 20

Bubenik 1989:54-6. He exemplifies their occupations as teachers, doctors, lawyers, actors, architects, engineers, painters, sculptors, merchants, etc. For the relationship between ethnic awareness and language, see Gumperz & CookGumperz 1982:1-21; Appel & Muysken 1987:12-6.

Personal Bilingualism

115

guage with non-Greek designs (Butcher 2003:275). In the early second century AD, the nationalistic use of Hebrew/Aramaic, with writings in Greek characters, and its significance for Bar Kokhba and his followers is well-known (cf. Yadin 1971:124, 181). 21 Cotton (1999:225) suggests that “Hebrew became the symbol of Jewish nationalism, of the independent Jewish state.” Bilingualism due to national ideology may explain how Hebrew was revived in the legal documents during the two Jewish revolts to become the language of the legal documents alongside with Aramaic.22 Fitzmyer also asserts that the majority of Qumran texts were written in Hebrew rather than in Aramaic because of national ideology.23 Hence, we can conclude that ethnic awareness may lead to bilingualism.

3.2.4 Bilingual Parents or Region A bilingual parent(s) or bilingual region (i.e. birthplaces or growthplaces) produces “early,” “late,” or “primary” bilinguals. In terms of bilingualism brought about by parent(s), intermarriage is usually the reason that produces this environment.24 One may note that the prac21

22

23

24

Incidentally, in regard to four dated legal documents, Yadin (1971:181) suggests that “It is interesting that the earlier documents [two of year one] are written in Aramaic while the later ones [two of year three] are in Hebrew. Possibly the change was made by a special decree of Bar-Kokhba who wanted to restore Hebrew as the official language of the state.” In relation to this, Rosén (1980:223:6) convincingly argues that Hebrew was used as a legal language only during the second revolt, which is supported by Mussies (1983:362-4). Cotton 1999:219-31; here, 220-5; 2005:153-6. She states that “the inner text of DJD XXVII no. 8, that is the part which is hidden, was written in Aramaic, whereas the outer text was written in Hebrew. In other words, the legally binding text, the inner one, is written in the normal language of legal documents at the time, whereas the Hebrew, displayed on the outside, advertises the ideology of the now independent Jewish state.” Fitzmyer (1979a:30) notes, “If the origins of the Qumran Essene community are rightly related to the aftermath of that revolt, this may explain why the majority of the Qumran texts discovered to date were written in Hebrew and composed at a time when most Palestinian Jews were thought to be speaking Aramaic.” Precisely speaking, the reason why Qumran community used Hebrew was related to national ideology and it was determined by language preference. (cf. §2.3.2) Mélèze-Modrzejewski (1983:241-68; esp. 248) argues that although some Hellenized native elites and some orientalized Greeks are of importance, they do not reflect Ptolemaic society as a whole, but are a marginal phenomenon of the society (cf. chapter 4, n.64). Nevertheless, bilingual parents or region are some of the factors contributing to a bilingual circumstance, because a bilingual community does not mean that all members of a community should be bilingual (see §2.1.2).

116

Bilingualism of Jews in First-Century Palestine

tice of intermarriage was actually encouraged in the Vienna papyrus which preserves “some decrees of Ptolemy II, Philadelphus, [and] seems to explicitly encourage intermarriage between natives and foreigners, by declaring that wives of such marriages were to be regarded as free women.”25 As a consequence, in Roman Egypt, native Greeks and Hellenized Egyptians in the first-century often married Egyptians and would have produced children who were early or primary bilinguals.26 Furthermore, some Greek and Macedonian soldiers and mercenaries in Syria married wives who were supplied from the native population (Hitti 1951:251). In the case of intermarriage, children could express themselves in both parental languages. This being the case, the reference in Acts 16:1 to Timothy, “whose mother was a Jewess...but whose father was a Greek,” would likely mean that he would have been an early bilingual. Anyone growing up in a bilingual region also would have become bilingual, even if their parents were monolingual (Hengel 1974:105). Living in a border area would have especially brought about bilingualism. As noted previously (§3.1.3), the population geography of firstcentury Palestine shows an interesting result. Millar (1993:343) shows that “there was no clear boundary between Jewish and Gentile settlement either eastwards, in Peraea and Gaulanitis, or in the territory of Caesarea Panias, founded by the tetrarch Philip; or north-westwards into the territory of Tyre and Sidon.” Furthermore, Scythopolis (Plain of Jezreel south of Galilee), Sebaste (site of ancient capital of Samaria), Tiberias (shore of the Sea of Galilee), and Sepphoris (near Nazareth) are all known to have had a mixed population of Gentiles and Jews (Millar 1993:343; cf. §3.1.3). As a matter of fact, there is no reason to assume that Jews and Gentiles represented separate communities in any formal or constitutional sense. Consequently, mixed populations created regional bilingualism. Dalman (1935:165) proposes that “Anyone brought up in Bethsaida would not only have understood Greek, but would have been polished by intercourse with foreigners and have had some Greek culture.” According to Hengel (1989a:16; cf. Bockmuehl 2005:81), first-century Bethsaida was “more markedly ‘Hellenized’ than the surrounding villages.” Moreover, the residents at Bethsaida were a mixture of Jews, Syrians, 25 26

Freyne 1980:152 n.94; Riggs 2005:19; Clarysse 1979:2.731-43; 1992:51-6; 1995:1-19. Fewster 2002:242; in terms of factors encouraging bilingualism in the second-century BC Egypt, Horsley (NewDocs 5.12) summarizes Antidoron Peremans’s arguments (1983:253-80) in four points: intermarriage, billeting civil servants in villagers’ homes, intermingling of Egyptians and foreigners in the same corps, and middleechelon officials in the chora.

Personal Bilingualism

117

and Greeks (1989a:16). On this account, five disciples (e.g. Peter,27 Andrew, Philip, James, and John) came from Bethsaida, as John 1:44, 12:21 indicate, and could have been primary bilinguals in spoken language.28 In other words, they could speak Greek as their embedded language in close proximity to Aramaic as their matrix language in varying degrees. Indeed, it appears from the New Testament account that Peter preached to congregations in territories where Greek was the dominant language.29 27

28

29

Following Pesch (1980:11-2), Bockmuehl (2005:60) may reasonably suggest that Peter had his own boat (Lk 5:3), which means that “Peter would have received the customary elementary education…” Cf. Wilkinson 1977:63. Bockmuehl (2005:82) convincingly elucidates that “there is therefore a very strong likelihood that Peter grew up fully bilingual in a Jewish minority setting. That his family and their friends were at ease with their Greekspeaking environment and its Herodian tetrarch seems reflected in the names they gave to their children.” And he (2005:64, passim) reasonably suggests that “Peter’s origin in Bethsaida does indeed make him bilingual.” This means that Peter was bilingual due to regional bilingualism. Although most of what he maintained is persuasive, it is regretful that he tries to connect bilingualism and “in a Jewish minority setting” or their Greek names. Two points should be mentioned. (i) It seems that “in a Jewish minority setting” implies that Jews could speak Aramaic, whereas Gentiles could speak Greek in Peter’s home town. He seems to mean that Peter could speak Greek well because he grew up “in a Jewish minority setting.” However, that Bethsaida was largely bilingual means that Greek was usually used as a matrix language, whereas Aramaic as an embedded language without regard to ethnological classification (Jews or Gentiles). Bilingualism means the coexistence of two or more languages (cf. §4.1.1). Regional bilingualism could provide bilinguality without regard to Peter’s race. (ii) It is well-known that name does not always indicate bilinguality (§5.3.1). Instead, it should be mentioned that the regional bilingualism of the firstcentury Bethsaida implies that what Peter experienced in childhood in Bethsaida means that he must have been an early or primary bilingual. Jesus’ disciples, Peter, Andrew, Philip, James and John who came from Bethsaida might have been cases similar to that of Peter. Cf. 1Cor. 1:12, 9:5. Hengel (1974:105) explains that “Simon Cephas-Peter … later undertook extensive missionary journeys among the Jewish Diaspora of the West, which spoke only Greek.” For Peter’s bilinguality, see Hengel 1983:11, 37, 162 n.39, 170 n.26; 1986:92-8; Mussies 1976:1056; 1983:364. Peter’s bilinguality is supported by four points: (i) In legend, according to the mission narrative in Acts 10, Peter was known as the founder of the Gentile mission in predominantly Gentile Caesarea as a result of the conversion of Cornelius; see Cullmann 1962:37-9; Mussies 1983:364. (ii) Personally, after deliverance from prison, when Herod sought for Peter, he went down from Judea to the Hellenistic city of Caesarea that Herod the Great built, and remained there (Acts 12:19). (iii) Social-economically, the fact that he owns a fishing boat might indicate that he may be employed in trade with Phoenician cities, because economically Galilee was largely dependent on Hellenized cities (Hengel 1989a:15). (iv) Peter, Andrew, Philip, James, and John, Jesus’ disciples, came from Bethsaida, which was Hellenized (Hengel 1989a:16). Regional bilingualism caused

118

Bilingualism of Jews in First-Century Palestine

In relation to regional bilingualism, it seems that Jews of firstcentury Palestine might have accepted non-elite bilinguality in the narrative of John 7:14-36. Two major lines of interpretation of who the {Ellhnej are (Jn 7:35) have been suggested. One interpretation proposes that they are Greek-speaking Jews30 who were not born in Palestine (i.e. native Greek speakers). The other view, proposed by a large number of interpreters, is that they are Greek-speaking Gentiles.31 It is not easy to say whether {Ellhn refers to Greek-speaking Jews or Greek-speaking Gentiles, or whether Jesus intended for a Gentile mission or this is rather John’s message (Lindars 1972:296). Regardless of disagreements on how one should understand {Ellhn, everyone agrees that the term refers to people whose first language was Greek. Despite the fact that Jesus is portrayed at one point as having never (formally) studied (Jn 7:15), there is an indirect assumption about Jesus’ ability to speak Greek in John 7:35-36. It seems evident that either the author/redactor of John’s gospel or the Jews portrayed here presupposed that Jesus could actually address Greek-speaking persons. This is adduced by John 12:20-23. When Greeks visited Jesus, Philip was sent to meet them. This implies that because Philip came from Bethsaida he could speak Greek (cf. Bockmuehl 2005:61). Philip asked Jesus to see some Greeks from Bethsaida of Galilee.32 This combination in John’s gospel may be a hint that some non-elite Galileans, including Jesus, would have become bilinguals without formal education. Noteworthy is that Dalman (1929:4) persuasively argues that Jesus was a bilingual and gives three reasons for this. First, his mother belongs to a priestly family in Judaea (Lk 1:36), implying she is an upper-

30 31 32

Peter to be a primary bilingual. For information on regional bilingualism and occupational bilingualism, refer to §3.2.4 and §3.2.6. Concerning the interpretation of Acts 4:13, Cullmann (1962:24) asserts that Peter’s bilinguality “does not prevent Acts 4:13 from characterizing him and his companion John as ‘uneducated’. He had not ‘studied’ either by Jewish or by Greek standards.” Also, Ellis (1978:243 n.24; cf. §1.3.3) mentions that “Doubts about the learning (gra,mma, Joh 7, 15; Acts 4, 13) of Jesus and his disciples do not refer to their literacy, a meaningless observation in the contexts, or to their biblical knowledge but to their ‘unprofessional’ status and the lack of sophistication associated with it.” Although Gamble (1995:9) assumes that “the meaning [of “uneducated”] is probably only that they were illiterate in Greek and had had no Greek schooling,” it seems that Peter as a fish-trader with Phoenician cities might have known Greek. Furthermore, some disciples such as Matthew and Peter and James, Jesus’ brother could read and write Greek; see Bauckham 2006:288-9. Robinson 1959-60:120, 124; Barrett 1975:11-19. Brown 1970:29, 314; Hengel 1989b:122; Fortna 1989:308 n.159; Lindars 1972:296. Argyle 1955-56:93; Ross 1990:41-3; interestingly, Porter may illustrate more cases in which Jesus spoke with his conversational partners in Greek (§1.4.2.5).

Personal Bilingualism

119

status Jewess. 33 Mary as an elite bilingual could have spoken Greek and, if this is indeed the case, it raises the possibility that her bilingualism would have led to Jesus’ early bilingualism. Second, Nazareth was a bilingual town, and Jesus “having been brought up there, could not have lived in isolation from the influence of Greek. (Dalman 1929:4)” Pratscher (1987:210-3) argues that James, the brother of the Lord, used Greek in his letter as his mother tongue. Hengel (1989a:17; cf. Mussies 1983:359; Fitzmyer 1992:61) also points out that “the situation of his [Jesus’] native Nazareth on the border of Galilee and five kilometers from Sepphoris, the old capital of the region, offered a variety of possibilities of contacts with non-Jews.” In light of these observations, it seems one may venture the assertion that regional bilingualism would have affected Jesus in this regard. Third, Jesus might have visited surrounding Hellenistic cities (Philadelphia, Hippo, Gadara, Pella, and Scythopolis) in his youth (Dalman 1929:4). On the basis of these three reasons, Dalman (1929:5) claims that “knowledge of Greek was common in Galilee,” and he cannot exclude Jesus, “who was a Galilean, from this common knowledge.” He (1929:5) continues that Mark “is aware that familiarity with several languages is not necessarily proof of higher education, but is rather a state of things arising out of the conditions of intercourse between the different populations.” 34 Speaking sociolinguistically, this means that Jesus’ bilinguality was due to regional bilingualism (cf. §3.2.4). Furthermore, Finkel illustrates an interesting case. He (1981:146) states: “In a Galilee town, Greek and Aramaic words were common speech forms.” He (1981:146) states that Jesus could speak Greek and his audience could understand the language,35 as Jesus’ saying implies: any father in a Galilee town would not confuse Aramaic patira (flat bread) and Greek petra (rock) when his son asked 33 34

35

For more information on Mary’s social status, see Dalman 1935:52. In regard to the linguistic competence of Jesus, Fitzmyer (1979a:37) proposes that “the general evidence that we have been considering would suggest the likelihood that Jesus did speak Greek.” Interestingly, Eshel and Edwards (2004:49-55) suggest that a potter’s abecedary from Khirbet Qana indicates the evidence that a craftsman tried to learn to read and write Aramaic in a small village in Galilee in the Early Roman period. This may make it probable that an artisan like Jesus, who belonged on a social level similar to that of the craftsman had the possibility of learning to read and write. Dibelius (1949:40) mentions that “Naturally a much greater Greek influence is to be assumed in Galilee, where the Jewish people bordered, so to speak, on the Hellenistic world, than in Judea. It is quite possible that Jesus and his disciples understood Greek, perhaps even spoke it.” Hengel (1989a:17) suggests that “he was capable of carrying on a conversation in Greek.” Van der Horst (1991:131) also notes that Jesus “understood and probably was even able to speak Greek.”

120

Bilingualism of Jews in First-Century Palestine

his father for food and pronounced PTR’. Dalman’s suggestion can be paraphrased in modern sociolinguistic terms. Jesus was an early bilingual due to parental bilingualism, a primary bilingual owing to geographical bilingualism, or at least a late bilingual because of the geographical bilingual factor.36 Consequently, it seems that Jesus could be an early bilingual due to regional bilingualism or at least a late bilingual.

3.2.5 Formal Education Formal education produces “acquired” bilingualism. One of the characteristics of acquired bilingualism is that one speaks the mother tongue at home and an acquired language for business. A language acquired through formal education never replaces the mother tongue. In this respect, we have much of evidence that there was a considerable number of “elite bilinguals” or “sub-elite bilinguals” in the first-century Roman Empire (Adams 2003:9; Hoffmann 1991:46). Instances of this are readily available. Zoskales, a king of first-century Ethiopia, was Hellenized enough to read and write Greek at home (Casson 1989:45, 52-3; cf. Huntingford 1980), thus displaying a prototype of elite bilingualism which is prevalent at the time in the general region. Furthermore, the leading Roman figures are no exception,37 and the examples are: Crassus (Cic. De orat. 2.2; Quint. 11.2.50), T. Albucius (Cic. Brut. 131), Augustus (Suet. Tib.21.4-6), Tiberius (Suet. Tib. 71.1), Claudius (Suet. Claud. 42.1) and Titus.38 Some kings of the Roman Near East were also bilinguals. Reporting the death of Crassus (BC 53), Plutarch (Crass. 32, 33) mentions that Orodes II (Hyrodes) of Parthia, who Crassus defeated, was well acquainted with the Greek language and literature and that 36

37 38

Taking a further step, it is interesting that Longenecker (1999:41-50) raises the possibility that Jesus spoke Aramaic, Hebrew, and Greek because the biblical quotations he used are strongly Septuagintal, whereas the evangelists’ quotations are Semitic. He (1999:50) raises the possibility that “in his use of the Old Testament, Jesus, who normally spoke Aramaic but could also speak Greek and Mishnaic Hebrew (at least to some extent), at times himself engaged in textual selection among the various Aramaic, Hebrew, and Greek versions then current, and that some of the Septuagintal features in the text forms attributed to him actually are to be credited to him.” This means that Septuagintalisms can be considered as the original sayings because Jesus cites the LXX. For more information, see §6.1.5. However, Goldstein (1981:69-70, 319-20) mentions that some conservative upperstatus Romans rejected Hellenism. For the argument of Titus’ Greek competence, also see Sevenster 1968:62-65; he (1968:65) concludes that “it is very possible that Titus was anything but a fluent speaker of Greek.”

Personal Bilingualism

121

the contemporary king, Artavasdes of Armenia, composed tragedies and wrote orations and histories. It is striking that “upper-class Romans who could not speak Greek (whether genuinely or allegedly) are sometimes disparaged (Adams 2003:9; cf. Cic. Verr. 4.127).” Most Jewish kings including Herod, Antipas, Philip, Agrippa I and Agrippa II were bilinguals. Herod seems to have attended a Greek elementary school in Jerusalem (Hengel 1974:77). Josephus (Ant. 16.6; cf. Ant. 18.143, 19.360, Vita. 359, C. Ap. 1.51) states that Herod sent his sons to study in Rome. This implies that his sons were bilinguals who have mastered two or more languages, such as Aramaic, Greek and Latin. Both in the Roman Near East and in Rome, powerful rulers were bilingual and sought Greek education. It is possible that this brought about a chain reaction. The Roman Empire employed bilingualism as a means to rule over other countries. Fewster (2002:220) maintains that “The fact that Roman administrators could virtually act as though there were only two languages is a sign of their ability to integrate provincial elites.” In all likelihood, bilingualism was promoted within the upper echelons of Jewish society. Those of a more mid-range social status also were enthusiastic for Greek learning, as Lieberman (1965:21; cf. 27) contends: “we know how eagerly the middle class imitates the upper class, and how readily the lower strata follow the example of the middle groups.”39 Greek was the official language of the Roman government and aristocracy in first-century Palestine.40 A language used in government administration promotes the learning of the language. One of the political purposes for encouraging Greek education in Palestine is “to gain greater influence over the Greek-speaking Diaspora” (Hengel 1974:77). Spolsky (1983:98) suggests that “A young Jew who wanted to rise in the world would have to learn Greek.”41 Additionally, 39

40

41

Sevenster (1968:60-1) also states that “the middle classes all too eagerly imitated the customs of the upper ones in this respect and that a knowledge of Greek and an ability to participate in conversation on matters concerning Greek culture became a sort of criterion of social status.” Mussies 1976:1058. Mussies (1983:358) suggests that “The official use of Greek for this period [first centuries] is attested by coins and inscriptions.” For instance, Eupolemus (1 Macc. 8:17-8) was probably a bilingual priest of second-century BC Jerusalem. He wrote a history about contemporary Judaism in Greek and used Greek and Hebrew texts; for detailed discussion, see Collins 2000:46-50. Furthermore, Hengel (1974:103) asserts that the Zenon papyri indicate that “the Greek language was known in aristocratic and military circles of Judaism between 260 and 250 BC in Palestine.” Interestingly, it seems that the role of Greek in Judaeo-Palestine is similar to that of English in many countries in modern times. Platt, Weber, and Lian (1984:29) consider English to be a status language in some African countries. They mention that

122

Bilingualism of Jews in First-Century Palestine

Hengel (1989a:17; 1974:76) mentions that “the better the knowledge of language a Palestinian Jew acquired, the more easily he could rise in the social scale.” Among the leading families of Jerusalem and the families of aristocrats, Greek was spoken (Hengel 1989a:14-17). The use of Greek in families implies that some upper-status Jews may be early bilinguals in Aramaic and Greek. The children of the leading families became early bilinguals who were taught by their bilingual father42 or private tutors43 and, at least, late bilinguals in elementary school.44 With these observations in mind, Goodman’s minimalistic proposal that “Aramaic would have been quite sufficient to make contact, both economic and cultural,” and that there was “no need to learn another lingua franca” (1983:67 [emphasis in the original]) is unconvincing. It is noteworthy that the definition of Hellene has changed from genealogy into education.45 Hellene refers to a Greek “who had a Greek

42

43

44

45

“English ... has connotations of education, socio-economic status and often also of power. ... People who went to school and speak English consider themselves ‘elite’ upper class.” This shows the relationship between a language and power. It seems that many young Jews tried to learn Greek as a language of upper class status. In the case of Egypt, it was the same situation; see Tcherikover 1963:15; Fewster 2002:242-3. The Torah (m. Qidd. 4:14 and t. Hag. 1:2) admonishes fathers to educate their sons. Josephus mentions that Jewish children in first-century Judaism should be taught to read (C. Ap. 2.204; Ant. 4.211; also see T. Levi 13.2; Philo, Leg. 115, 210). Hezser (2003:192) notes that in tannaitic times the duty to educate children rested almost exclusively on parents, with few opportunities provided outside the home.” Also see Goodman 1983:72. Rajak (2002:26) assumes that Josephus was educated by his parents. Hezser (2001:91) mentions that it is usual that upper-status Jews learned their second language from private tutors. In terms of domestic education by private tutors, Gamble (1995:6-7; here 6) asserts that “Domestic instruction in reading and writing under the guidance of a tutor was common among wealthy households, and home study was possible in Christian households too when they were well-to-do or when there was a literate family member.” Hengel (1989a:17) suggests that “The larger cities, primarily Jerusalem, but also Sepphoris and Tiberias, had Greek schools which presumably went as far as an elementary training in rhetoric.” Hengel (1974:76) also summarizes: “All these points suggest that even from the Ptolemean period the sons of the Jewish aristocracy in other words, a long time before the establishment of the gymnasium and the ephebate there was something like a Greek elementary school.” Also see Hezser 2001:912; Hezser (2001:91) mentions that Aramaic-speaking upper-status Jews learnt Greek as their second language from Greek-speaking non-Jews or Diaspora Jews in Palestine. On Jewish schools in Palestine, see Gamble 1995:7, 249 n. 22; Hengel 1974:1.7883. Grabbe 1991:1.166; Grabbe states that “the criterion soon became one not of genealogy but of education: a Greek was one who had a Greek education – a Greek was one who had a command of the niceties of the Greek language. This concept was ex-

Personal Bilingualism

123

education,” which means he “was one who had a command of the niceties of the Greek language.”46 Tcherikover (1963:14) points out that “a gymnasium education had served as an entrance-card into the privileged circles which, in the eyes of the Romans, were considered as ‘Greeks’ and sometimes even called ‘gymnasium-graduates’.” 47 For these reasons, change in definition of Hellene could also have accelerated Jewish learning of Greek. Most of the elite in first-century Palestine may have been primary or acquired bilinguals, although this does not mean that all urban elites such as ordines, rulers, the rich, and retainers at that time were in fact bilinguals.48 Personal bilingualism is individualistic and personal proficiency levels in Greek are varied. Nevertheless, elite bilingualism can be assumed. Elite individuals among the high priests, Sadducees, Herodians, scribes, Pharisees and Rabbis, in all likelihood, were bilingual.49

46

47

48

49

tremely important in breaking down the barriers between the settlers and the natives, as the natives began to acquire a Greek education.” Grabbe 1991:1.166. In this respect, the verb e`llhni,zw means “to speak Greek”; LSJ 536; refer to Hengel 1974:74. The educated persons were called Hellenes, who were defined in terms not of origin but of either their posts in the administration or their education. See Hengel 1974:73; Thompson 1992:324. She (1992:326) illustrates that “some (Hellenes), of course, were cleruchs, the Greek soldiers settled in the Egyptian countryside; others are specified as doctors, teachers, fullers, priests, and so on.” He (1963:5) mentions that the two principles for distinguishing Greeks from nonGreeks are permanent residence in the metropoleis and a Greek education in the gymnasium. It seems very difficult to define “elite” in ancient times. For the definition of the “elite,” it is noteworthy to mention that Stegemann and Stegemann (1999:53-95; cf. 303-16) suggest three criteria such as power, privilege, and prestige for distinguishing between elite (upper-stratum groups) and non-elite (lower-stratum groups). The former includes (i) ordines and ruling families in vassal states and provinces, (ii) the rich, and (iii) retainers, whereas the latter includes the (i) relatively poor, (ii) minimum existence, and (iii) absolutely poor. On the basis of Gerhard Lenski’s model (1984), Neyrey (1996:255-67) notes that the social stratification of an advanced agrarian society is composed of ruler, governing class, retainer class, merchants, priests, peasants, artisans, and unclean, degraded and expendables. Esler (2001:32-37) considers that the elite of first-century Palestine refers to members of the elite (high priests, Sadducees, and Herodians) and retainers of the elite (scribes and Pharisees). Richard Rohrbaugh (2000:204-5) provides useful sources; he identifies urban elites in the Gospel of Mark: Caesar (12:14, 17), Pontius Pilate (15:2, 8, 15), Herod (6:14; 8:15), Herodias (6:17), Herodias’ daughter (6:22), Philip (6:17), High Priest (2:26, 14:47), chief priests (8:31, 10:33), scribes (1:22, 2:6), elders (8:31), rich man (10:22), Sadducees (12:18), owner of upper room (14:14), Joseph of Arimathea (15:43), Jairus and his family (5:22), such retainers as Pharisees (2:16), people from Jairus’ house (5:35), men arresting John the Baptist (6:17), soldier of the guard (6:27), Levi (2:14), those selling in the temple (11:15), servant-girl of high priest (14:66), the

124

Bilingualism of Jews in First-Century Palestine

The most common way to become bilingual would have been the educational system of first-century Palestine,50 which enabled upperstatus Jews to become elite bilinguals (Hengel 1974:65-83; Sevenster 1968:183-4). Jews could receive Greek education in Jerusalem as well as in Greek Palestinian cities such as Sepphoris, Tiberias, and Julias Bethsaida (Hengel 1991:55; Hezser 2001:90-4). In first-century Palestine, “it was impossible to found a new ‘Hellenistic’ city without a Greek school” (Hengel 1991:55; cf. Hezser 2001:91-2). Hence, the educational system in first-century Palestine produced acquired bilinguals. The main purpose of formal education is to offer children access to political and social power by means of learning Greek language and literary skills.51 Children seem to desire to learn a refined style from model literary texts rather than a perceived vulgar style used by uneducated people (Butcher 2003:273). Welles, Fink, and Gilliam (P. Dura. V. I. 47) argue that the language of Greek parchments and papyri (the third century BC – the third century AD) excavated from Dura does not deviate from the standard Attic dialect of the fourth century BC except for a very few cases. They (P. Dura. V. I. 47) conclude that “It is clear that the clerks of Dura and the vicinity were carefully trained in the classical language.” It is quite intriguing that, when Egyptians learned Greek as a second language, they used a Ptolemaic Greek papyrus roll (ca. end of third or second century BC).52 The roll seems to be a Greek teacher’s manual, which includes a Greek education curriculum: reading, arithmetic and dating systems.53 Furthermore, the most remarkable feature of the curriculum is the introduction to Greek literature such as Homer -- taught at all levels of education54 -- the tragedians, New Com-

50

51

52 53 54

crowd sent from the chief priests, scribes and elders (14:43), physicians (2:17), Galilean priest (1:44), couriers, officers (6:21), Judas Iscariot (14:11), tax collectors (2:15), moneychangers (11:15), doorkeeper (13:34), soldiers (15:16), and the centurion (15:39). It seems that they could have been bilinguals. There were other ways to be bilinguals. Learning foreign languages is not special at that time. Josephus (Ant. 20.264) reports, “Jews look upon this sort of accomplishment as common not only to all sorts of freemen, but to as many of the servants as please to learn them.” Hengel (1974:60; see §4.2.4) mentions that Jews could have learned Greek from Diaspora Jews who visited from Egypt, Asia Minor and the Aegean to observe their holidays. Swain 2002:131. In the case of first-century Roman Egypt, Fewster (2002:242-3) supposes that “to be bilingual would have been a great advantage for these children (bilinguals).” Clarysse 1983:43-61; Thompson 1994:75-7; 1987:105-21. Clarysse 1983:44, 47-8; Thompson 1987:112-3; 1994:76. Hock 2001:56-77. The important role of Homer in Greek education was expressed by Heraclitus, Dio Chrysostom, and Quintilian; see Sandnes 2005:716-7. Morgan

Personal Bilingualism

125

edy, and other Greek writers.55 Greek literature was also known in the Roman Near East in the first-century BC. Plutarch (Crass. 32, 33) mentions that the people of Seleucia knew the wisdom of Aesop and that a passage of Bacchae of Euripides was recited and many Greek compositions were introduced during a celebration banquet for the death of Crassus in Parthia in 53 BC. According to Josephus56 and Philo57, even upper-status Jews in Antioch and Alexandria would be educated at a gymnasium, 58 which means that they may have known Homer and Euripides.59 In the case of Palestine, Hezser (2001:70) proposes that “this focus on Homer was not limited to Egypt but can be considered typical of the Graeco-Roman world from Hellenistic times onwards.” The curriculum of Greek education at the gymnasium in Jerusalem and other Greek Palestinian cities seems to include texts of famous Greek writers such

55 56 57 58

59

(1998:69) notices that “of the roughly 150 other texts by known authors, 97 are extracts from the Iliad and the Odyssey. The next most popular author is Euripides, with twenty texts.” Concerning the three stages of the Greek school, see Hengel 1974:66-7. For the detailed study of the education of children from birth to age fourteen in Athenian society (500-300 BC), see Alexandrakis 1987. Also, for the detailed study of the education in Ancient Rome from the elder Cato to the younger Pliny, see Bonner 1977. Clarysse 1983:44, 47-8; Thompson 1987:112-3; 1992:325; 1994:76. Ant. 12.119, 120, cf. C. Ap. 2.39 and Bell. 7.43, 44; cf. Tcherikover 1959:516 n.91; Bickerman 1962:53, 89 n.47; Hengel 1974:68. Spec. leg. 2. 230 (M 2, 298); Somn. 69 (M 2. 631), 129ff. (M 2, 640); Louis Feldman 1960:224-6. Hengel 1974:68; Tcherikover 1959:516 n.91; Bickerman 1962:53. In this regard, Nörden (1958:2.496-7) raises the possibility that Paul could have read some of Plato. On the other hand, Hengel (1991:37) supposes that Paul “probably had not read any classical Greek literature worth mentioning apart from the Greek (and Hebrew) Bible and Jewish pseudepigrapha like the Wisdom of Solomon.” In more recent years, some scholars have proposed that some biblical writers knew Homer, Pindar and the tragedians; Andersen & Robbins (1993). In terms of Homer, see Glockmann 1968:57; MacDonald 1999:88-107; 2000; 2003a; 2003b:189-203. It seems that the Greek writers can be known to writers and readers of the four gospels, if they had formal education. For, even upper-status Jews who can be readers of the gospels would be educated at gymnasium in Antioch or Alexandria where most scholars assert that the gospels would be written. However, Mitchell (2003:254) points out that their arguments showing Homeric influence still “are based on unconvincing and unexamined assumptions about ancient authorial practice and procedures.” In the case of Lucian’s time, Homer and Euripides in Antioch were well known to the public. Weitzmann (1941:233-47; here 233) suggests: “No wonder, that in Lucian’s time, when quotations from these two classics overcrowded literature, mosaicists depicted themes from Homer and Euripides and could be sure that these would be understood by a great public, even when the figures of the scenes were not accompanied by inscriptions.”

126

Bilingualism of Jews in First-Century Palestine

as Homer. Hengel (1974:103, 65-83) suggests that “A Greek School must have existed in Jerusalem… There are signs of a continuation of Greek education in the Jewish capital, which even included the knowledge of Homer.” Furthermore, it is striking that two epitaphs found in Beth Shearim were written in Homeric language both in form and content, although Beth Shearim was of high renown as a rabbinic center and was known for the epitaphs of great rabbis (Van der Horst 2001:164-5). The expression, ‘my Beth Shearim’ indicates that he was a native of Beth Shearim (Van der Horst 2001:163). The epitaph raises the possibility that a Palestinian native might learn Homeric Greek in Greek school. Accordingly, when upper-status and middle-status Jews learned Greek, they tried to imitate cultured Greek style, Atticism, not the vulgar style of street language. Using Atticism might be one of the characteristics of acquired bilinguals who received Greek education in first-century Palestine. When acquired bilinguals learned Greek, how much Greek did they actually learn? Two cases in this regard may be considered. First is the bilingualism of Lucian of Samosata, a Syrian Greek orator in the early second century. It seems that the linguistic milieu of second-century Samosata is similar to some extent to that of first-century Palestine.60 According to MacLeod, in the first half of the second century AD the inhabitants used Syrian Aramaic; some of them spoke Greek, and educated persons especially were bilinguals in Syriac and Greek. The headquarters of a Roman legion in Samosata may indicate that a few must have known Latin as well. Lucian spoke Syrian Aramaic as his native tongue during his boyhood and youth.61 His parents were poor and he could not formally learn Greek. Instead, his parents apprenticed him to his uncle as a stone-cutter and statuary (Allinson 1905:ix-x). After he went to Ionia to get formal education in Greek, he at last became one of the “high-priced sophists” in Gaul (Allinson 1905:xi) which “was well known as a home of rhetoric, Greek as well as Latin” (Jones 1986:12). The late bilingualism of Lucian of Samosata indicates that this acquired bilingual became the best Greek orator even though he acquired Greek as a second language. 60

61

The political and sociological situation of Samosata is similar to that of Jerusalem; according to Jones (1986:6), Samosata on the middle Euphrates “had once been the capital of the small kingdom of Commagene, founded in the third century BC and incorporated in the Roman Empire in 72 A.D.” MacLeod 1991:1. That Lucian was monolingual as a child indicates that his linguistic environments (e.g. parents and neighbors) were predominantly monolingual. For the relationship between the age of second language acquisition and the levels of proficiency, see Critical Period Hypothesis (§2.1.4).

Personal Bilingualism

127

Our second contributor to the question about the extent to which Greek was learned is Josephus. As to his acquisition of Greek as a second language, Josephus (Ant. 20.263-4; Steve Mason’s translation [2001:3]) states: For among my compatriots I am admitted to have an education in our country’s customs that far surpasses theirs. And once I had consolidated my knowledge of Greek grammar, I worked very hard also to share in the learning of Greek letters and poetry, though my traditional habit has frustrated precision with respect to pronunciation. Among us: they do not favor those who have mastered the accent of many nations and made their speech frilly with elegance of diction, because they consider such a pursuit to be common--not only among those who happen to be free citizens, but even among domestics if they desire it.

The above statement indicates four points at the levels of phonology, syntax, and stylistics. (i) His pronunciation was not precise. Naturally, it is difficult for acquired and/or late bilinguals to pronounce their second language precisely (§2.1.4). Because the phonetic inventory of Aramaic is different from that of Greek, it is quite common that phonological rules of Aramaic could intervene in Greek pronunciation.62 Although Josephus confesses his deficient competence in terms of Greek pronunciation, this is not particularly problematic as many others encountered the same issue. Rather, he ascribes his imprecise pronunciation to a negative view that the Jews had towards foreign language acquisition (Rajak 2002:50-1). Indeed, he intends to brag that he composes excellent Attic Greek at the levels of syntax and stylistics,63 although it is hard to acquire literate bilingualism enough to compose a polished-style Greek. (ii) Josephus was probably not an early or pri62

63

It is well-known among neurolinguistic scholars that when anyone learns the second language, there is a clear distinction in the outcome depending on his/her age (see §2.1.4). If he or she is over twelve years old, he or she will face some difficulties in conversation at two levels: phonologically, he or she often violates complicated phonological rules for the inventory of consonants and vowels which are different from that of the second language. Syntactically, he frequently breaks intricate syntactic rules. For instance, ‘He go to school’. However, in writing, he or she does not make such errors. Accordingly, Josephus’s violation of pronunciation in Greek shows that he learned Greek as an acquired language when he was older than twelve . Nevertheless, there is no doubt that he was known as one of the best writers in Greek. Similarly, although German speakers know sound values such as /Ό/ or /ð/, when they speak English, they often utter them in an imprecise way. However, their imprecise pronunciation does not prove that they write English in an imprecise way. Josephus was not a primary bilingual or an early bilingual, but a late or acquired bilingual. Sevenster (1968:70) states that “he asserts with pride and satisfaction that he has achieved a very great deal regarding the writing of Greek.”

128

Bilingualism of Jews in First-Century Palestine

mary bilingual but a late or acquired bilingual. If he were an early or primary bilingual, his pronunciation would have been precise and he would not have tried to consolidate his grammar. He was an Aramaicmatrix bilingual and Greek was his embedded language. This also indicates that his parents might have been Aramaic-matrix speakers and might not have used Greek at least at home (i.e. as is the case for most acquired bilinguals), although Josephus holds that his lineage is that of a royal-priestly family (Rajak 2002:14-22). (iii) Josephus was also an oral and literate bilingual. He could speak Aramaic and Greek and write his publications in Aramaic and Greek. Despite the fact that Greek was foreign and unfamiliar to Josephus (Ant. 1.7),64 when he tried to learn it in order to become bilingual, he acquired excellent competence in written composition. 65 Bilingualism of first-century Palestine persuaded Alexander Roberts (1888:444), a Scotch-English writer, to delineate the language of Josephus as below: But his object was very different. It was his ambition to write in the style of the classical writers, and that he found a matter most difficult of attainment. He had then to deal with a xe,nh kai. avllodaph. dia,lektoj, just as we formerly saw was the case with Scotch authors of the last century in their attempts at writing correct and idiomatic English. To them English was substantially “a foreign tongue,” as Josephus felt it to be with himself in regard to classical Greek. Yet his desire was to write in the style of the best authors, and his assiduous efforts to this effect were not in vain.

His attitude toward the style of Greek composition “could be highly significant” for answering the question of “how much knowledge of Greek a Jew in the first-century A. D. could acquire if he took serious pains about it” (Sevenster 1968:74; cf. Rajak 2002:50-1). In terms of Josephus’ acquisition of Greek in Jerusalem, it is noteworthy that Rajak (2002:46-64) convincingly considers that the bilingualism of Jerusalem enabled Josephus to learn Greek in Palestine. Josephus’ remark indicates that a bilingual Jew, whose matrix language was Aramaic and whose embedded language was Greek, could employ a good style in 64

65

Rajak (2002:47-8) mentions that Roman Greek writers (i.e. A. Postumius Albinus and Aulus Gellius) who used Greek as their second language confess that it is difficult for them to use Greek. Josephus, in this sense, “had good reason, then, to overstress rather conceal his linguistic inadequacies.” Many scholars accept the “assistant theory,” that is, that Josephus was helped by Greek assistants when he composed his books (Thackeray 1929:100-124). However, others like Rajak admit Josephus’s competence of Greek; she (2002:236) argues that “second-rate authors imitated their models” and proposes that “it would be natural for Josephus to try out a number of different styles in his work.” When we take the bilingualism of first-century Palestine into account seriously, it is highly possible that Josephus could have written in the styles of famous Greek writers.

Personal Bilingualism

129

Greek composition. 66 (iv) The contemporary Jewish view of second language acquisition indicates that there were many early and/or primary bilinguals with precise pronunciation because of regional or parental bilingualism. That is, there were many early bilinguals and primary bilinguals in first-century Palestine without regard to their social status. Paul seems to have been bilingual, also being very capable in Aramaic and Greek.67 Regarding his matrix language, there is some disagreement. Van Unnik (1973:304, 321-7), for example, suggests Paul’s matrix language was Aramaic, 68 whereas Hengel (1991:34-7, 125 n.188; 1983:37) thinks it was Greek. Becker (1993:34-6, 53) also argues that Paul’s matrix language was Greek and bases this on his splendid Greek composition. However, compositional aptitude in a particular language does not prove that it is one’s matrix language. Although Lucian and Josephus were acquired bilinguals both of whom used Aramaic as their matrix language, they became famous Greek orators or writers. Rather, we should pay attention to Paul’s boasting. The terms, ~Ebrai/oi (2 Cor 11:22) and ~Ebrai/oj evx ~Ebrai,wn (Phil 3:5) imply that Paul used Aramaic as his matrix language and Greek as his embedded language (see ch. 8 n.106; cf. §5.2.1). Paul considered his Aramaic competence (i.e. ~Ebrai/oj) to be one of his credentials in 2 Cor 11:22 and Phil 3:5. Although Paul’s matrix language was Aramaic, there is no doubt that he had a full mastery of Greek. Regional bilingualism of Tarsus of Cilicia made Paul an early or primary bilingual. Accordingly, from the perspective of bilingualism, it seems that although Paul’s matrix language was Aramaic, Paul could use his embedded language as proficiently as his matrix language (for instance, for Paul’s codeswitching, see §8.4.1).

66 67 68

Josephus (Vita 9.40) mentions another acquired bilingual, Justus of Tiberias; refer to Rajak 1973:345-68; Freyne 1992:79. Concerning Paul’s bilingualism, Millar (1993:364-5) draws a vivid picture of Paul in the bilingual situation of Jerusalem. It is interesting that Chrysostom regards Paul’s matrix language to be Aramaic. When Chrysostom (in 2 Tim 4, NPNF, 13.490) made a comparison between the emperor Nero and Aramaic-speaking Paul, he asserted that Paul “was a tent-maker, a poor man, unskilled in the wisdom of those without, knowing only the Hebrew tongue, a language despised by all, especially by the Italians.” Bruce (1977:43) also notes that “while Paul was born into a Jewish family which enjoyed citizen rights in a Greek-speaking city, Aramaic and not Greek was the language spoken in the home and perhaps also in the synagogue which they attended.”

130

Bilingualism of Jews in First-Century Palestine

3.2.6 Occupation Occupation can sometimes lead to individuals becoming acquired bilinguals. Professional bilingualism was required in a bilingual region or a bilingual society. Some occupations would have given preference to bilinguals (Hezser 2001:240). Some professions such as that of the scribe demand biliteracy and others like traders require bi-orality. For instance, it is well-known that an Egyptian who intended to work anywhere under the Ptolemies (256-255 BC) should speak and preferably write Greek as well (P. Col. 66): “I therefore request you, if you please, to order them to let me have what is owed to me / and in future to pay me regularly, so that I do not die of hunger because I do not know how to speak Greek” (cf. Austin 1981:418 no.245; Green 1990:313). This shows that getting a job in Egypt requires speaking Greek. Despite the fact that Phrygian is quite similar to Greek, it is interesting that Brixhe (2002:256) suggests who the bilinguals in Phrygian and Greek would have been: Those who, in the course of their professional activities, dealt directly with civic officials (e.g. contractors), or who in one way or another came into contact with the Roman authorities (e.g. the army); medium or large landowners who produced more than they required to satisfy the needs of their family and sold some of the excess on the market; those who managed, or who were involved in the management of, large estates owned by Greek speakers, etc.

Similarly, Jews of first-century Palestine also tried to learn Greek for the purpose of employment. Hezser (2001:231; cf. Sevenster 1968:183) suggests that for two types of professions Greek would have been a mandatory prerequisite. The first comprises Jews who intended to work at public offices at the municipal level because of contact with non-Jewish Greek-speaking officials. The other consists of Jews who were employed in trade relationships with Greek-speaking Gentiles or the Jewish Diaspora. Economically, Galilee depended by and large on completely Hellenized cities in Phoenicia (Hengel 1989a:15). Galilean residents who were employed in a trade, like Peter, would have been required to speak Greek.69 Those who worked as tax collectors, building craftsmen, and trade fishermen were considered to belong to the middle class and may have been bilingual (Hengel 1989a:17).70 In this respect, Argyle assumes that Jesus as a carpenter who grew up in Gali-

69 70

For Peter’s bilinguality, see chapter 3 ns. 28, 29, 72. For this reason, he assumes that Jesus, who was a building craftsman, might have been a bilingual. And Mark 2:15 shows that many tax collectors followed Jesus.

Personal Bilingualism

131

lee would speak Greek. 71 Similarly, he adds Simon, Andrew, James, John, and Levi to the list of contemporary bilinguals.72 They, at the very least, should be regarded as occupational bilinguals because they would have been in contact with Greek-speaking people to maintain their jobs, although, more importantly, they were regional bilinguals because they grew up in bilingual regions (see §3.2.4). Sevenster (1968:38-61; here, 44; cf. Hengel 1989a:16) asserts that “many Rabbis were completely familiar with Greek.” Palestinian rabbis generally spoke the Greek of the middle-class man of Palestine (Lieberman 1965:2). 73 This raises the possibility that the LXX was read elsewhere in Palestine as well as in the Diaspora.74 Moreover, scribes in first-century Palestine may have functioned as copyists like the scribes in Ptolemaic Egypt.75 A bilingual community requires bilingual copyists. For instance, the fact that Babatha’s legal documents were written

71

72

73

74

75

Argyle (1974:88) states, “Any Jewish tradesman who wished his business to prosper would be eager to make his range of customers as large as possible and so would welcome Greek-speaking Gentile customers as well as Jews. This would apply especially in Galilee of the Gentiles where the majority of the population was Gentile and Greek-speaking. … If Joseph and Jesus wanted their carpentry business to prosper, they would be happy to welcome Gentile as well as Jewish customers. They would therefore need to speak Greek as well as Aramaic if they were to converse with all their customers.” Fitzmyer (1992:61) also mentions that Jesus “was a skilled craftsman. … He would naturally have conducted business in Greek with Gentiles in Nazareth and neighboring Sepphoris.” Also see chapter 3 n.34; cf. §3.2.4. Argyle (1974:88; cf. 1955-56:92) mentions that “Simon and Andrew, James and John would need to know Greek if they were to sell their fish in Gentile markets. So would Levi, the inland revenue officer, the civil servant, engaged in government employ.” Cf. Lee 1985:6; Fitzmyer 1992:62; Porter 1994:136. For a rabbi’s Greek competence, see Lieberman 1965:27, 66; 1962:100-114. Philip Alexander (2001:74) also considers that the elites of Palestine were bilinguals: “there is surely a strong presumption that the rabbis, who were among the more educated members of Jewish society, would have known the language (Greek),” although he holds a different view from Lieberman’s argument. In relation to the LXX found in Qumran, it seems reasonable that Millar (1993:352) proposes that “if the texts found at Qumran included, as they did, at least some of the books of the Septuagint, we can be confident that the Bible could be read in Greek translation elsewhere in Judaea also.” For instance, concerning the colophon of Greek Esther, see chapter 4 n.79. Saldarini 1988:241-276; Goodman 1994:102-8; Esler 2001:34-5. Moreover, Hengel delineates that “whereas Greek education was designed to produce gentlemen amateurs, Eastern education was designed to perpetuate a guild of professional scribes.” For more information, see Hadas 1959:68.

132

Bilingualism of Jews in First-Century Palestine

in Greek, Aramaic and Nabataean (Yadin 1971:230) indicates that the scribes in the bilingual society must have been bilinguals.76 Furthermore, some lower-status Jews were bilinguals as well (see §2.3.3). Slavery due to involuntary migration sometimes caused servants to become bilingual (Myers-Scotton 2002:32). We know that some slaves in Rome could be bilinguals who spoke Greek as well.77 In addition, although slaves were from various countries, households preferred Greek-speaking slaves and, thus, forced them to learn Greek. Josephus (Ant, 20.264) mentions that even slaves could be bilinguals in Aramaic and Greek in the first-century Palestine.78 Feldman also allows the bilingualism of slaves when he (1993:19) mentions that “to be sure…slaves and freedmen knew and used Greek.”79 Gamble (2004:31) suggests that “Even the literate commonly preferred not to exert themselves in private reading, but to be read to by a literate household slave or freedman.” One may say that bilingual competence in first-century Palestine does not show social status because slaves, too, could learn and speak Greek. Sevenster (1968:70) suggests that “This demonstrates that Greek was not only spoken in a few groups and classes, but that everyone in the Jewish country had the chance of speaking it. Greek could evidently be heard in all circles of Jewish society.” As discussed, some professions prefer bilinguality while certain jobs require it. In sum, the epigraphical, papyrological and population geographical evidence indicates that first-century Palestine was largely bilingual. In other words, they show both “urban and upper-status” and the “rural and lower-status” bilingualism. Political, social, economical, cultural, religious, and geographical factors enabled a considerable number of upper-status and lower-status Jews to be bilinguals.

76

77

78 79

For instance, Dura Parchment 28 and P. Mesopotamia found in Osrhoene show that professional scribes were bilinguals in Syriac and Greek; See Brock 1994:151-2; Butcher 2003:143. The parchments and papyri found in Osrhoene were written in Greek and Syriac and were written by professional scribes. This also indicates that professional scribes in Edessa were bilinguals. See Brock 1994:151-2. Adams 2003:15, 761-2. Adams proves that some freedmen and slaves could speak Greek as well. Moreover, some of them might be literate, as epitaphs at Delos, Petronius, Plautus, and Rome show (p. 762). Sevenster 1968:69-70. For the information of literacy of slave in ancient times, see Gamble 2004:31; Starr 1991:337-43; esp. n.12. However, conversely, he denies that middle and upper status Jews of first-century Palestine could speak Greek. Nevertheless, if middle and upper status Jews could not speak Greek, “this would be the exception to the general pattern” of the phenomenon of bilingualism, as he expresses it (1993:14).

Bilingualism of First-Century Palestine and Interdirectionality

133

3.3 Bilingualism of First-Century Palestine and Interdirectionality Bilingualism means coexistence. The maximalist view of the bilingualism of first-century Palestine helps us to draw some pictures about both the historical Jesus and the Jesus tradition more completely. When Jesus spoke to his audiences, there were many bilinguals who spoke and understood Greek among both upper-status and lower-status Jews. The evangelists inform us that the report about Jesus spread throughout the region of Galilee. Matthew 9:26 and Luke 4:14 say that a report (fh,mh) concerning him went out through all that district / all the surrounding country in Galilee. After describing the recovery of a man with an unclean demon in Capernaum, Luke 4:37 also states that a report about him (h=coj peri. auvtou/) spread throughout the surrounding area (cf. Luke 7:17; Acts 26:26). The fh,mh and h=coj could have bilingually circulated in Galilee. The sayings by Jesus and stories about Jesus in Aramaic could have been translated into Greek among Greek-matrix eyewitnesses in Palestine during Jesus’ ministry. At the same time, the Jesus traditions in Greek were also translated into Aramaic among Aramaic-matrix eyewitnesses in Palestine during Jesus’ ministry. By the same token, some Aramaic traditions which had been translated from Greek could have been re-translated into Greek and some Greek Jesus traditions which had been translated from Aramaic could also have been re-translated into Aramaic during Jesus’ ministry. As a result, the Jesus tradition should have circulated in Aramaic as well as in Greek in Galilee during Jesus’ ministry. I call this the “Galilee translation theory” or “simultaneous translation theory” in that translation of the Jesus tradition started from Galilee (see §1.5.1, §5.1). Accordingly, the transmission of the Jesus tradition was interdirectional, circular and hybrid rather than unidirectional, teleological and unilinear.

4. Bilingualism of Jews in the First-Century Diaspora We have seen in chapter 1 that many scholars have considered that the Syrian Christian community made a significant contribution to the transmission of the Jesus and gospel tradition, although this firstcentury community has attracted some scholars’ attention without strong evidence. According to the religionsgeschichtliche Schule, one of the most important reasons for highlighting the Syrian community is to bridge a linguistic gap between the Aramaic language used by Jesus and his disciples and the Greek of the four gospels, and to span the distance between the Aramaic-based Palestinian Christianity and the Greek-based Hellenistic Christianity.1 Bultmann (1925:145; 1967:102-3; cf. Bousset 1970:119) suggests that Syrian Christianity should be urgently investigated for coherence between Palestinian Christianity and Hellenistic Christianity. He (1925:145) states that the problem of the Hellenization of earliest Christianity is closely related to that of Syrification. On this account, it seems that, for the religionsgeschichtliche Schule, the linguistic distinction is of significance in terms of the directionality of the transmission of the early Jesus tradition. Fitzmyer (1979a:123) questions in this context whether by “Hellenistic Christianity” Greek-speaking communities in a pagan environment (such as Syria, Egypt, or Asia Minor) are meant. Additionally, he (1979a:123) considers whether one can rule out that Palestinian Christian Hellenists had a great part in the formulation of the primitive kerygma. Hengel (1989a:1-6) has persuasively and exhaustively asserted that those who were “bilinguals in Aramaic and Greek” in the earliest Jerusalem church took a significant part in the transmission of the Jesus tradition.2 In other words, in regards to the linguistic transmission of 1 2

Heitmüller 1912:333-4; Bousset 1970:119-20; Dibelius 1949:25; Bultmann 1935:12-3; Mack 1988:101-2; Goulder 1994; also, see §1.2, §4.3.3, §5.3. Furthermore, he (1989a:1-6) reveals that one of his reasons for writing the publication (i.e. The ‘Hellenization’ of Judaea in the First Century After Christ) is to criticize the Syrian Christianity hypothesis by Bultmann. Hengel properly criticizes the Syrification hypothesis in several points. Hengel and Schwemer (1997:279-86; here 281) argue that the so-called “Hellenism” of the earliest Christian community in Jerusalem

136

Bilingualism of Jews in the First-Century Diaspora

the Jesus tradition, Bultmann focuses on the Syrian church, whereas Hengel focuses on the Jerusalem church because both take the bilinguals into serious consideration. More precisely, Hengel concentrates on “bilingual Palestinian Greeks” [emphasis added].3 However, it seems that most scholars, including Bultmann and Hengel, neglect the bilingualism of the Jewish Diaspora. One might consider, for example, that the Roman Province Syria was completely bilingual. Moreover, neither Bultmann nor Hengel takes seriously into account the Aramaic-matrix Diaspora Jews and Greek-matrix Palestinian Jews, as will be considered below. In relation to this concern, Millar (1993:21) questions: Had there been in this region Gentile Christian communities using Aramaic or a dialect of it? The Jewish community of Dura-Europos had certainly used Aramaic, as well as Greek. But had the Christian community nearby? Even if we do not treat the question as a purely linguistic one, was there, as so many modern books presume, a distinctive “Syrian” Christianity using Greek, which owed its character to its regional environment? If so, what would that mean? Of which of the many sub-regions of the Near East are we are talking, and what are the criteria of “Syrian” Christianity?

These questions raise the issue of whether or not there were a significant number of bilinguals in Syria alone. In more general terms: were the linguistic situations of Palestine as well as other regions of the Roman Near East bilingual or not?4 Did Diaspora Jews of Syria and Egypt speak only Greek? Should all Gentile Christians be regarded as Greekspeakers? To highlight the importance of these questions, we must keep in mind that, if the linguistic situations of Syria, Egypt, and even Jerusalem were not monolingual but rather bilingual (i.e. Aramaic and Greek), it is inappropriate to view the unidirectionality of the transmission of the Jesus and gospel traditions as being from Semitic to Greek and from Judaeo-Palestinian Christianity to Hellenistic Christianity. A study of bilingualism in the first-century Roman Near East and Palestine may indicate that the Jesus and gospel traditions are not to be viewed straightforwardly as having developed unidirectionally from Semitic tradition to Greek tradition. Indeed, both the Semitic and Greek traditions probably circulated simultaneously in oral and written forms. The phenomenon of bilingualism will also make more fluid the borders between the Judaeo-Palestinian Christian church and the Hel-

3 4

should “go back to Jerusalem itself” because Jerusalem was “far more Hellenistic than the history-of-religions school.” I will discuss with Hengel’s argument in detail in chapter 5. Millar (1993:564-76) suggests that the Roman Near East covers main sites such as Antioch, Cilicia, and Commagene (north), Phoenician Coast (west), Arabia, Sinai, the Red Sea, and the Hedjaz (south), and Mesopotamia, Dura, and Nisibis (east). I add Alexandria to the list of the regions in this chapter.

Grounds for Bilingualism in the Jewish Diaspora

137

lenistic church.5 Furthermore, under the presupposition of many scholars that the Semitic tradition was oral, whereas the Greek tradition was written (chapter 1), the study of bilingualism will seek to demonstrate that, just as oral tradition developed into a written tradition, so may the written tradition have been changed into oral tradition. This may be a much more frequent practice than has been recognized. Oral transmission was a major communicative vehicle in ancient times and was much more complex than accounted for by a simple oral-to-written directional paradigm. Accordingly, it may be assumed that the directionality of the transmission of the gospel tradition in first-century Palestine and the Roman Near East was more likely a hybrid than unilinear.6 It is fitting that the transmission of the gospel tradition should be considered as circular and interdirectional rather than teleological and unidirectional. In this chapter, the bilingualism of the Jewish Diaspora of the firstcentury Roman Near East will be observed. This chapter is concerned with three points: the grounds for the bilingualism of first-century Jewish Diaspora (§4.1); the bilingualism of Alexandrian Jews (§4.2); and also that of Antiochene Jews (§4.3).

4.1 Grounds for Bilingualism in the Jewish Diaspora It is generally said that the Jewish Diaspora of the first-century Roman Empire was composed of Greek-speaking communities. However, recent evidence seems to support a bilingual view of the Jewish Diaspora. Sociolinguists have discussed the factors influencing the maintenance of a minority language. Giles, Bourhis and Taylor (1977), whose study is expanded by Appel and Muysken (1987:33-8), consider that this maintenance depends on “ethnolinguistic vitality,” which is a combination of three main factors: (i) economic, social, sociohistorical and language status, (ii) demographic support (e.g. distribution and numbers), and (iii) formal and informal institutional support. They (1977:308) propose that “ethnolinguistic minorities that have little or no group vitality would eventually cease to exist as distinctive groups.” On the contrary, they continue, “the more vitality a linguistic group has, the more likely it will survive and thrive as a collective entity in an inter5

6

Furthermore, Hengel (1974; 1980; 1983; 1986; 1991) also points out in his major publications that there was no sharp distinction between Palestinian Judaism and Hellenistic Judaism. For detailed discussion, see §1.5.

138

Bilingualism of Jews in the First-Century Diaspora

group context.” It seems that the Jewish Diaspora had a high enough ethnolinguistic vitality to preserve its mother tongue in Antioch and Alexandria when their social status, their population, and educational support are taken into consideration. With this in mind, I will suggest four reasons for the possible existence of bilingualism in the Jewish Diaspora: (i) the bilingualism of the Roman Empire; (ii) successive immigration; (iii) periodic connection with Jerusalem; and (iv) learning Aramaic in Jerusalem.

4.1.1 Bilingualism of the Roman Empire It is said that one of the reasons for the success of the Roman Empire was the Roman policy of bilingualism (Fewster 2002:220). However, biblical scholars have generally regarded the linguistic milieu of the first-century Roman Empire as Greek monolingualism, which is one of the characteristics of Hellenization. This assumption results in three linguistic presuppositions in relation to the spread of earliest Christianity: (i) the Gentiles spoke Greek, whereas Jews spoke Aramaic; (ii) Diaspora Jews spoke Greek, whereas Palestinian Jews spoke Aramaic; and (iii) consequently Gentile Christianity comprises Greek-speaking communities, whereas Palestinian Christianity comprises Aramaicspeaking communities.7 However, classical historians have recently considered Hellenism as “coexistence” not as “fusion.” In regard to the origin of the fusion hypothesis, Johann Gustav Droysen (1836-43), who studied under Hegel, applied Hegelian dialectics (i.e. thesis – antithesis – synthesis) when Hegelian philosophy was widely accepted to Hellenism: Greek + Oriental = fusion of the two. Criticizing Droysen’s theory, Naphtali Lewis (1986:4) argues that an accumulation of archaeological evidence for the last half-century, which was not previously available, has demonstrated “the fallacy of the fusion hypothesis as a basis for understanding the history of Hellenistic times.”8 Rather, mutual influences were fringe phenomena (Lewis 1986:4). In terms of the ethnic relationship between Greeks and Egyptians, Maehler (2004:3-4) argues: But should this relationship be described as interaction, or as coexistence? Until relatively recently, most scholars thought that the two cultures began

7 8

For more arguments regarding the three presuppositions, see §5.2. In terms of onomastics, Mélèze-Modrzejewski (1983:248) suggests that Ptolemaic onomastics does not prove the linguistic milieu because the inscriptions show marginal group of the society so that the “fusion” hypothesis is based on the fragile foundation of special cases of limited scope.

Grounds for Bilingualism in the Jewish Diaspora

139

to merge early on, developing into a “mixed” Graeco-Egyptian civilization. Today, the opposite view, that Greeks and Egyptians lived their separate lives without taking much notice of each other, seems to be generally accepted.

As to languages, the consensus has been reached recently that although Greek was commonly employed in the first-century AD, the vernacular languages of most areas of the Roman Empire still persisted.9 For instance, Acts 14:11 indicates that even the urban inhabitants of Lystra of Galatia used Lycaonian, their own native language, when they spoke to each other. However, these people changed their language codes from Lycaonian into Greek when they communicated with Barnabas and Paul. Jones (1940:289) advocates that “this piece of evidence is of particular value since it shows that the urban proletariat there still spoke Lycaonian: a fortiori the peasantry must have continued to speak it far later.” The linguistic milieu of Lystra of Galatia seems to show a prototype of bilingualism in the Roman Empire in that Lycaonian was used as the primary language and Greek as the common language was second. Above all, first-century Rome was known as a bilingual city.10 Most areas where a Latin family of languages was used were also bilingual.11 It is striking that in the case of the regions of the Greek language family, vernacular languages survived and did not disappear. Strabo mentions that Cappadocia, as e`tero,glwttoij e;qnesi indicates, was a multilingual region (Janse 2002:347-9).12 In regard to Punic,13 Millar (1983:57) 9

10 11

12

13

Selective publications are as follows: Sartre 2005:274-96; Adams 2003; Butcher 2003; Adams 2002; Taylor 2002:298-331; Janse 2002:349; Fewster 2002:220-45; Brixhe 2002:246-66; Healey 2001:18-21; Porter 2000a:136-7; Cotton, Cockle and Millar 1995:214-235; Brock 1994:149-60; Millar 1993:xiv, 232-3, 503-4; 1992:97-123; Bubenik 1989:276-81; Sherwin-White 1987:1-31; Millar 1987a:143-64; 1987b:110-133; Drijvers 1984:1-27; Goodman 1983:64-5; Millar 1983:55-71; Mussies 1983:356-69; Rosén 1980:215-39; Mussies 1976:1040-64; Millar 1971:1-17; Brunt 1971:170-72; MacMullen 1966:1-17; Hitti 1951:256-7; Jones 1940:32-3, 288-93; Dibelius 1935:75; Roberts 1888:180-4; Bernhardy 1861:1.493. Furthermore, in terms of bilingualism in GreekRoman period, Horsley (NewDocs 5.11-2) offers more publications. Adams 2003 passim; in the case of Roman Jews, refers to Noy 1997:300-11; Rutgers 1995:176-209; Angus 1914:123. Adams extensively deals with the bilingualism of the Latin language family, that is, its sister languages and its daughter languages, in his monumental book (2003), mentioned above. Janse (2002:347-59) also mentions that, although what Cappadocian is like is under dispute, it seems that it was distinct from Greek and was a daughter language of Anatolian; cf. Jones 1940:288-90. Acts 28:2 indicates that the residents of Malta, called ba,rbaroi, spoke Punic, Greek and Latin; see Barrett 1994-1998:1220-1.

140

Bilingualism of Jews in the First-Century Diaspora

suggests that even “Punic culture survived not only in Sardinia; in Africa, Punic could be written until the second or third centuries, in neoPunic, Latin or Greek characters, and was spoken, at any rate as a peasant language, at least until the end of the fourth century [AD].” Phrygian people were also bilingual in Phrygian and Greek. Phrygian inscriptions were found until the second–fourth centuries AD (Janse 2002:350; Brixhe 2002:246-66; Bubenik 1989:277; Jones 1940:290). Patristic literature of Socrates and Sozomenos shows that Phrygian continued to be used up to the fifth century (Bubenik 1989:277; Janse 2002:350). The linguistic milieus of Ptolemaic and early Roman Egypt were bilingual in Egyptian and Greek, as will be mentioned later in detail (§4.2). In the case of the Roman Near East, even though Greek was used as the common language, Aramaic was also used as a primary language (Janse 2002:357; Millar 1993:503-4). The variety of the bilingual situation depends on the regions, as will be suggested. The survival of vernacular languages reveals that Greek did not supplant vernacular languages completely, but rather coexisted with them. In other words, the persistence of the languages brings about bilingualism. Emphasis should be placed on the fact that bilingualism is not a “zero-sum game.” As the definition indicates, bilingualism refers to the alternative use of “two or more” languages. To illustrate the point, if a city where people use Aramaic is influenced by Greek, generally speaking, Aramaic will be the primary language, whereas Greek will be a second language. This does not mean that residents speak only Greek, but rather that, as bilinguals, they speak Aramaic and Greek.14 In this respect, although unsatisfactorily, Schwartz considers the relationship between Greek and Hebrew. 15 His comment (2005:54) on the linguistic phenomenon of bilingualism merits attention: It would, for example, be simply false to say that the rise of Greek in the Hellenistic period generated a decline in Hebrew, or that the revival of Hebrew for some purposes in late antiquity implied a decline in Greek. Hebrew usage was only obliquely related to vicissitudes in the usage of Greek, if it was related to them at all. 14

15

Similar to this, the acculturation of Diaspora Jews does not mean assimilation. Although some Diaspora Jews were acculturated in Hellenism, this does not mean that they were assimilated to Hellenism. For detailed discussion, see Barclay 1996:92-8. Sterling (2001:274) also mentions that “no one seems to have thought that the use of the Greek language compromised Judaism. It is worth noting that Bar-Kochba carried on correspondence in Greek, even if the Greek letters were written by members of his staff. The use of Greek was therefore a matter of acculturation not assimilation.” Unfortunately, he assumes that the primary language was Hebrew.

Grounds for Bilingualism in the Jewish Diaspora

141

Almost all areas of the first-century Roman Near East were bilingual in Greek and their own vernacular languages. Bilingualism occurred in different degrees and under different situations. The bilingualism of the first-century Roman Near East should be understood from the perspective of the coexistence of two or more languages. As Green (1990:313; cf. Millar 1971:1-17) asserts, “even in the heavily Hellenized areas of Syria, Phoenicia and Cyprus bilingual inscriptions are common.” Along these same lines, Jones (1940:290) points out that “more significant is the evidence for the survival of Aramaic in apparently Hellenized regions.”16 Millar (1987a:144), as well, proposes that “various dialects of the Semitic language which we call ‘Aramaic’ were spoken all the way round the Fertile Crescent, from Babylonia to Arabia” [emphasis in the original]. As a consequence, the survival of Aramaic means the coexistence of both Greek and Aramaic. There were many bilinguals in Greek and Aramaic in the regions of the first-century Roman Near East. Some upper-status inhabitants17 in the regions of the Roman Near East spoke only Greek, whereas some lower-status inhabitants spoke only Aramaic (and/or a vernacular language). Many inhabitants could have been early or late bilinguals. In the same vein, many Diaspora Jews of the first-century Roman Near East could have been bilinguals. Some upper-status Diaspora Jews of the first-century Roman Near East spoke only Greek, whereas some lower-status Jews only Aramaic (and/or in vernacular language). What is more, many Jewish and Gentile Christians of the Roman Near East could have been bilinguals. Some upperstatus Jewish and Gentile Christians spoke only Greek, whereas some lower-status Jewish and Gentile Christians only Aramaic.

4.1.2 Successive Immigration It has been said that Palestinian Jews were scattered into regions of the Roman Empire for several reasons. Most obvious are: expulsion, political difficulties, religious persecution, internal conflicts, and tempting economic prospects in other countries. 18 After Palestinian Jews had 16 17

18

Cf. Hitti 1951:256-7; Healey 2001:18-21. Yadin (1971:234) also suggests that while Nabataeans must have spoken an Arabic dialect, they used Aramaic as well. In terms of measuring social stratification, I will use the term, “status,” one of three different kinds of ranking in Roman times: class, ordo, and status, as Meeks (1983:5173) suggests. Stern 1974:1.117; Kasher 1987:47. In this respect, Kasher (1987:47) mentions that the migration of Jews is different from that of Greeks, which was voluntary and encouraged by Hellenistic kings of those countries.

142

Bilingualism of Jews in the First-Century Diaspora

immigrated to the regions where Greek was employed as a primary language, they should have been acculturated and especially should have begun to speak Greek;19 in the process of acculturation, there is no doubt that the first generation of the Jews in the Diaspora must have tried to learn Greek. Roughly speaking, for first-generation Diaspora Jews, Aramaic was still their matrix language, while Greek was their embedded language. From the second generation, Diaspora Jews in Hellenistic circumstances might have chosen Greek as their matrix language, if they could choose their language; as a result, they would be early or late bilinguals in Greek as their matrix language and Aramaic as their embedded language if upper-status younger Jews could acquire Greek education or live in regions where Greek was the primary language (§3.2.4). Furthermore, over the course of time, the next generations of Diaspora Jews probably preferred Greek rather than Aramaic more and more. As a consequence of this, scholars have suggested that the Diaspora Jews might have been Greek-speakers. However, language maintenance theory in a bilingual setting should be considered. It has usually been reported that in the case of minority languages in multilingual settings, a language shift to the dominant language takes three or four generations. 20 Moreover, language shift in ancient times could be slower than in modern times.21 When Diaspora Jews spread into the Greek-speaking regions, language shift from Aramaic into Greek also might have taken three or four generations. The second or third generation of Diaspora Jews who were bilinguals 22 still could have spoken Aramaic in the Greek-speaking areas. It is interesting that Henry Cadbury (Beg. 5.59-74, esp. 62 n.4.) considers ~Ebrai/oi to be Aramaic-speaking Diaspora Jews.23 He (Beg. 5:62 n.4.) mentions that inscriptions of “synagogue of Hebrews” were found

19

20 21

22 23

In this respect, Hengel puts forward the idea that “in terms of language, the ‘Hellenization’ of these Jewish garrison troops or cleruchs must have made fairly rapid progress. Only about twenty-five percent of the names of Jewish military settlers mentioned in third-century papyri are Semitic; all the rest are already Greek.” Paulston 1994: chapter 2. Paulston (1994:17) suggests six general factors of opportunity of access to the second language for language shift: participation in social institutions, mass-media, transportation, travel, occupations and demographic factors. These factors hint that modern people more easily have the opportunity to learn the second language than ancient people. It is noteworthy that Paulston (1994:13) states that “The mechanism of language shift is bilingualism.” For full discussion of the Hebrews and the Hellenists, see chapter 5.

Grounds for Bilingualism in the Jewish Diaspora

143

in Corinth ([suna]gwgh. ~Ebr[ai,wn]) dated between 100 BC – 200 AD,24 in Rome (sunagwgh. Aivbre,wn),25 and in Philadelphia in Lydia (t]h/| a`giot[a,th| s]unagwgh/| tw/n ~Ebrai,wn).26 He assumes that the designation “Hebrews” refers to Aramaic-speaking Jews of the synagogue and distinguishes more recent Jewish emigrants from older emigrants.27 The former are Aramaic-speaking Diaspora Jews and the latter are Greek-speaking Diaspora Jews. Cadbury grants that some Diaspora Jews spoke Aramaic because they were newer Jewish emigrants from Palestine. Although it is not clear that the inscriptions are related to the newer Diaspora Jews, there is no doubt that incessant emigration shaped the Jewish Diaspora into two linguistic groups, Aramaic-matrix Jews and Greek-matrix Jews within the dispersion. Furthermore, attention should be drawn to the fact that Jews did not stop being war slaves, mercenary soldiers, traders, deportees, or refugees even after the fourth century BC.28 Tcherikover declares that the emigration of Syrians from Palestine still continued during the third and second centuries BC in considerable numbers, and that the immigrants included Jews who spoke Aramaic and “were engaged in various trades and belonged to all classes of society: Syrian merchants, officials, soldiers, field-hands, slaves, etc.”29 One of the major factors of Jewish immigration was related to war.30 There can be two ways one could become a war immigrant. The first

24 25 26

27

28 29 30

CIJ 718; for discussion of the inscription at Corinth, see Deissmann 1927:16-17 n.7; Deissmann 1957:86; cf. Müller 1919:24, 58, 72, 98, 106-7, 112, 173. CIJ 510; CIG 9909. For more inscriptions in Rome, see CIJ nos. 291 (~Ebr[ai,]wn), 317 (~Ebr[ai,]wn), and 535 (~Ebr[ai,]wn). CIJ 754. Keil and von Premerstein 1914:32-4, no.42. There were 2,000 Jewish families in Lydia and Phrygia who had been mass-transported from Mesopotamia and Babylonia (210-205 BC). If the Jews reside together, they might have spoken Aramaic to some extent. Beg. 5.62 n.4; van der Horst 1991:87. Schürer (2.2.248) states that “we also hear of a sunagwgh. Aivbre,wn, probably that of such of the Jews as spoke Hebrew, in contradistinction to those of them who had ceased to speak it.” Cf. Schürer 1879:35; for various interpretations of this “Hebrew,” see Leon 1995:147-9; Noy 1998:111-3. It seems that they were Aramaic-matrix speakers rather than Aramaic speakers, as many scholars suggest; see §5.2.1. Barclay (1996:21-2) also suggests these types of Jewish immigration. CPJ 1.5. Tcherikover (CPJ p.5 n.13) mentions that the “Syrians” sometimes refers to Jews; cf. chapter 4 n.101. Kasher (1987:48-53) deals fully with the reasons for Jewish immigration. For the case of mass-transportation, Josephus (Ant. 12:147-53) preserves Antiochus III’s letter ordering Zeuxis to transport 2,000 Jewish families in Mesopotamia and Babylonia into Phrygia and Lydia around 210-205 BC (Tcherikover 1959:287-8; Schalit 1960:289-318).

144

Bilingualism of Jews in the First-Century Diaspora

way was as a slave.31 The Letter of Aristeas (chapters 12-27) details that Ptolemy I (305-282 BC) took 100,000 captives to Egypt (cf. Ant. 12.11-33; Barclay 1996:21). Although this number is almost certainly exaggerated, it preserves the historical information that many Jews were caught by the Ptolemaic army (Tcherikover 1959:50-2, 425 n.45). During the Battle of Raphia (217 BC), many Jews became captives and were sold as slaves (Kasher 1987:49).32 Josephus (Ant. 12. 251; cf. C. Ap. 1.43) claims that Antiochus IV carried away about 10,000 Jewish captives.33 In the case of the Battle of Emmaus, the defeat of Judas Maccabeus resulted in a huge number of Jewish people being traded as slaves (Ant. 12.296, 299).34 Ptolemy IX Lathyrus (ca. 103-2 BC) conquered Asochis, a city of Galilee, and 10,000 inhabitants were sold into slavery (Ant. 13.337). Philo (Leg. 155) notes that after the fall of Jerusalem (63 BC) Pompey sent many Jews to Rome. Cassius Longinus invaded Taricheae and, according to Josephus (Bell. 1.180; Ant. 14.120), 30,000 inhabitants were sent into slavery.35 Herod made it a policy to sell brigands into slavery (Ant. 16.14).36 During Varus’ War (4 BC), Varus, the Roman governor of Syria, suppressed insurgents at Sepphoris and made its inhabitants slaves (Ant. 17.289, 298). Titus sent 700 Jewish men who were tall and handsome to Rome (Bell. 7.118)37 and also sent 97,000 captives to the East (Bell. 6.420). Vespasian in 66-70 AD sent 6,000 young men to Nero to dig the Corinthian Canal and 30,400 were sold as slaves (Bell. 3.540). As a consequence, many Jewish “immigrants” actually may have been slaves of war (cf. also Daniel). The second way Jews could emigrate was as mercenaries. Josephus (C. Ap. 1.192, 200) informs us that many Jews worked as mercenary soldiers in the fourth century BC and went to different areas of the Roman Empire (cf. Kraeling 1932:131; Borgen 1996:75-6; Barclay 1996:20). It has been suggested that Paul’s father or grandfather may 31 32 33 34 35 36 37

For more detail discussion, see Noy 1998:108-9. Kasher assumes that Antiochus’s order for the liberation of the descendants of Jewish slaves indicates that many Jews were enslaved during the Fourth Syrian War. Cf. 1 Macc 1:32; 2 Macc 5:14; Dan 11:33. The number of Jewish war prisoners is not certain; for more discussion of the number, see Kasher 1987:70 n.12. See 1 Macc. 3:41; 2 Macc. 8:11-14, 16, 25; 34:36. Cf. Ant.14.304, 313, 319-23. The number may be exaggerated; see Smallwood 1976:36 n.58, 131. For more reference, see Kasher 1987:51, 71 n.28; Herod (37 BC) might have sent Jewish captives to Rome. In the case of the Roman Diaspora, Rutgers (1995:168) notes that “it is true that part of the Jewish community in Rome originated among Jewish slaves who had been shipped there on various occasions during the first centuries BCE and CE.” Philo (Leg. 155) mentions that Roman Jews bought some Jewish slaves and set them free.

Grounds for Bilingualism in the Jewish Diaspora

145

have migrated as war immigrants from Judaea to Tarsus in the first century BC (Ramsay 1930:31-2; Hengel 1991:4-17).38 Josephus becomes a first-generation Jew in Rome in the first century AD (Shutt 1961:1-7).39 As an acquired bilingual, he tells us explicitly that his matrix language was Aramaic, having written his works in Greek. Political reasons prompted a massive number of Palestinian Jews to escape from their land.40 Furthermore, some upper-status Jews emigrated for economic gains (Kasher 1987:57-63). For these reasons, the theory that the next generations of Jewish immigrants in the Diaspora spoke Greek as their primary language is not persuasive, because who first-generation Diaspora Jews really were cannot be defined. First-generation Diaspora Jews used Aramaic, since their matrix language immigrated with them. Accordingly, the continual emigration from Palestine infuses Aramaic-speaking natives into the Jewish communities in the Diaspora, so that the Jewish communities there could have maintained their linguistic competence in Aramaic to some degree.

4.1.3 Periodic Connection with Jerusalem According to Paulston (1994:15-6), it is interesting that the language shift among Greek Americans into English took place within four generations, whereas the language shift among Italian Americans took three generations in Pittsburgh. She (1994:16) argues that some significant features supporting the prolongation of language maintenance include “marriage patterns of endogamy, a prestigious language with literary tradition, and access to a social institution with formal instruction, i.e. literacy in the original mother-tongue.” Jews were encouraged to practice endogamy and were educated in the Torah in their synagogues. These imply a possibility that Diaspora Jews could have used Aramaic longer than scholars might have thought. Although immigrants usually forget their ancestral language within three generations at modern rates, some linguistic communities which are geographically isolated or religiously separated can preserve their ancestral language far longer (Schwartz 2005:53). It is well-known that Jews of the Diaspora observed Jewish laws, to some degree, and con38 39 40

Deissmann (1957:90) assumes that Paul’s father or grandfather immigrated from Gishala in Galilee to Tarsus. Cf. Mommsen 1901:83-4 n.4. See §3.2.5. It is interesting that Rajak (2002:11, chapter 7) stresses that Josephus was a Diaspora Jew. Kasher (1987:54-7) illustrates political banishments and escapes in detail.

146

Bilingualism of Jews in the First-Century Diaspora

tinually kept in touch with their homeland. 41 Tax collection among Diaspora Jews indicates interaction between Diaspora Jews and leaders in the Levant (Barclay 1996:418-21). Josephus (Bell. 7.218; 16.162-7) mentions that Diaspora Jews paid two drachmas every year to the Jerusalem temple.42 According to Cicero (Pro Flacc. 28.67-9; cf. Josephus, Bell. 16.171), when Flaccus was a Roman governor of Asia (62-61 BC), he confiscated the gold that Jews sent from all the provinces to Jerusalem. It seems indisputable that the great majority of Diaspora Jews paid their temple tax.43 It is interesting that Jewish peasants grew rice in the plain near Antioch in the late second century (Tosefta Demai 2.1).44 Rabbis regularly visited even rural areas outside of Palestine in order to collect the tax from them (Wilken 1983:37). In regard to the transportation of the temple tax, Philo (Spec. 1.77-8; Leg. 156, 316) remarks that groups of Jews delivered the sacred money to Jerusalem. Josephus (Ant. 18.310-3; here 313) also reports, “Many tens of thousands of men undertook the transportation of those donations.” As a result, Philo (Spec., 1.69, 76-78; cf. Acts 2:5-11) notes that tens of thousands of Diaspora Jews flocked to Jerusalem at the time of the three pilgrimage festivals (i.e. Passover, Shevuot and Sukkot). Jeremias (1969:83) suggests that 125,000 pilgrims visited Jerusalem on average at Passover. Sanders (1992:127-8) estimates the figure at 300,000 – 500,000. When even Greekspeaking Jews of the Diaspora visited Jerusalem, where Aramaic was spoken as matrix language, they would have been exposed to Aramaicspeaking persons. Furthermore, in relation to the strong connection between Babylonia and Jerusalem, Stern (1974:117) supposes that the political and religious interactions between Aramaic-speaking regions and Palestine in the first century were very strong. For instance, Herod, a king of Palestinian Jews, promotes respect toward and friendship with Babylonian Jews, and, further, appoints Hananel, a Babylonian Jew of a sacerdotal family as high priest.45 Furthermore, that even Hillel, one of the two 41

42 43

44 45

To explain the presence of Semitisms of the synoptic gospels, Dalman (1909:16) supposes that “The spiritual intercourse also which Jewish Hellenists continuously had with Hebraists in Palestine implied a constant interchange between Greek and Aramaic (but not Hebrew) modes of expression.” Cf. Bell. 16.27-28, 169-73; on the basis of Ex30:14-15. Mandell (1984:223-32) confines the temple tax to Pharisees. However, many scholars agree that Diaspora Jews paid the temple tax; see Hengel 1991:55; Barclay 1996:41824; Wilken 1983:38-42; Safrai 1974:1.186-93; Lietzmann 1953a:77; CPJ 1.80. For more discussion of the Jewish peasants near Antioch, see §4.3.2. Neusner 1984:34-9; Schürer 1.400-67. For the information of the political-religious relationship between Herod and Hananel and Hyrcanus II from Parthia see Richardson 1999:162, 242-4.

Grounds for Bilingualism in the Jewish Diaspora

147

leaders of the Pharisees, came from Babylonia might indicate the religious relationship between Palestine and Babylonia.46 The strong connections between Palestine and Alexandria and Syria will be discussed further below (§4.2.4, §4.3.2).

4.1.4 Learning Aramaic in Jerusalem Some Greek-speaking Diaspora Jews who had extended stays in Jerusalem (e.g. Acts 2:5) may have acquired Aramaic in various ways (cf. §3.2), because even Greek immigrants to eastern countries in the Hellenistic-Roman era assimilated to eastern culture. Kasher (1987:46; cf. Clarysse 1992:51-2) suggests three reasons for the orientalization of the Greeks in the Mediterranean countries: intermarriage, coexistence far from the motherland, and pan-Greek solidarity among the Greek elite group. The strong interaction between Hellenism and Semitism in the Roman Near East is well established.47 For instance, Livy (History 38.17) was afraid of the orientalization of Roman soldiers and complained that the “Macedonians who occupy Alexandria, Seleucia, Babylonia and their other colonies throughout the world, have degenerated into Syrians and Parthians and Egyptians.” Hitti (1951:256-7) proposes that, in the case of Greek colonists “in Aramaic Syria and Jewish Palestine, the native culture in general more than held its own; it gave more than it received.” 48 In relationship to the Greeks of Ancient Egypt, D. O’Connor (1993:83) asserts that Greek residents there “became Egytianized to a significant degree while yet maintaining a distinctively ‘Greek’ culture.”49 Investigating the tombs of Roman Alexandria, Venit (2002:7-21) suggests that the ethnic boundaries between Greek and Egyptian became fluid. Another issue to consider is that some of the late Seleucid kings were given Aramaic nicknames. Alexander I was called Balas. Alexander II, who was not of Seleucid lineage, was called Zabinas, which means the “gifted one” (Hitti 1951:256). Butcher (2003:284) considers 46 47 48

49

For more information, see Neusner 1984:39-41. Fitzmyer (1979a:38-43) also mentions the interaction between Aramaic and Greek. Hitti (1951:256) also asserts that “the Greek colonists gradually became more affected by Semitic life than the natives by Greek life.” Tcherikover (1963:13) properly claims that “the East exerted more powerful influence upon the Greeks than they themselves were able to exert upon it.” It is interesting that even the cult of Isis in the Egyptian religion became widespread in Italy and elsewhere in Roman times (O’Connor 1993:83). Green (1990:315) also points out that “it was the alien Greek, who, by intermarriage and religious syncretism, slowly became Egyptianized.”

148

Bilingualism of Jews in the First-Century Diaspora

this practice and says it is striking that “in some regions we find people with Greek names giving their children Semitic names.” Greek inscriptions were unearthed from hill-top temples of the limestone massif, an area surrounded by long-established Greek cities in Northern Syria in 131 AD. The inscriptions show that the Greek-speaking builders may have had at least some knowledge of Aramaic because of the use of the term “Madbachos.” The Madbachos certainly originated from the Hebrew $bz and Aramaic root $bd and means “altar” (Millar 1987a:162-3). As an explanation of such practices, Tcherikover (1963:12-3) proposes that “Under the influence of the East, the Greeks themselves started to sink to the level of the natives among whom they lived. … even the eastern languages did not remain alien to them.” Furthermore, some individuals of two families among Greek immigrants (near Pathyris the second century BC) are identified in Demotic as Greeks born in Egypt (Riggs 2005:19).50 It is interesting that a famous letter in the second century BC depicts a Greek mother offering her Greek son congratulations for learning Egyptian and getting a job as tutor in an Egyptian family (Rostovtzeff 1998:2.883, 3.1545). In light of this letter, it is likely that some Greek-matrix Jewish and Gentile Christians who stayed in Jerusalem may have tried to learn or relearn Aramaic.51 Four points can be made by way of summary. First, the general linguistic milieu of the Jewish Diaspora of the Roman Near East can be regarded as bilingual in Aramaic and Greek, rather than monolingual in Greek. Many in the Jewish Diaspora could have spoken both Aramaic and Greek, 52 though some upper-status Jews could speak only Greek while some lower-status Jews only Aramaic. Second, successive immigrations transfuse Aramaic-matrix Jews and Aramaic-matrix speakers (e.g. Syrians) into the Jewish Diaspora communities so that many Jews could preserve their linguistic competence in Aramaic. Third, periodic connections with Jerusalem could have helped Jews to maintain their ability to speak or perhaps even to learn Aramaic. Fourth, even Greek-matrix Jews or Greek monolinguals who stayed for

50 51

52

For interconnections and ethnic designations, see Goudriaan 1988:58-69. It is interesting that Bengel (1859:2.525) proposed that some learned Aramaic in Jerusalem during their stay in Jerusalem. Schnabel (2004:654) suggests that the Hellenists who resided in Jerusalem spoke Aramaic as well. He explains, “They hardly would have settled in the Jewish capital if they were unable or unwilling to communicate in Aramaic.” However, it is unfortunate that he disregards Jerusalem as a Hellenized city. Torrey (1933:250) intriguingly notes that “Not only in western Asia, however, but also in Egypt and the Mediterranean lands, Aramaic was the language commonly spoken and written by the Jewish colonists.”

Grounds for Bilingualism in the Jewish Diaspora

149

long periods in Jerusalem might have tried to learn or re-learn Aramaic as their embedded language. In light of these considerations, I propose four points: (i) All Gentiles in the New Testament literature did not speak Greek as their matrix language and all Jews did not speak Aramaic as their matrix language; (ii) all Diaspora Jews did not speak Greek as their matrix language and all Palestinian Jews did not speak Aramaic as their matrix language; (iii) all Gentile Christians did not speak Greek as their matrix language and all Palestinian Christians did not speak Aramaic as their matrix language; and (iv) Gentile Christian communities in the first-century Diaspora were not simply Greekspeaking communities and the earliest Christian community in Jerusalem was not simply an Aramaic-speaking community. Both of them were bilingual. Although the bilingualism of Diaspora Jews is of significance in terms of the transmission of the Jesus tradition, this issue will not be treated here.53 Instead, three cases of regional bilingualism will be illustrated: Alexandria (§4.2), Antioch (§4.3), and Jerusalem (chapters 3, 5). These communities have been selected for three reasons: (i) Relatively speaking, a number of written sources are preserved and available to us from the three regions the cities represent.54 (ii) The Diaspora communities in Alexandria and Antioch were among the largest in ancient times (Stern 1974:1.138). 55 The regional congregations in the Jewish Diaspora in the empire played a significant role in the rapid transmission of the Jesus traditions. (iii) Last, these three cities are often mentioned in scholarship, which considers the provenance and the Sitz im Leben of the gospels. In terms of the provenances of the four gospels, the provenance of the Gospel of Matthew has been suggested as An-

53

54

55

In this regard, Taylor (2002:301) confesses that, in terms of the languages of only Syria and Mesopotamia, that “although it is probably beyond any one scholar (it is certainly beyond me) to control all of these languages, one should be alert to the possibility that they may also have had a significant impact on the other languages spoken.” As a matter of fact, when we deal with the bilingualism of the Jewish Diaspora of the first-century Roman Empire, we find some difficulties, for there are not many first and secondary materials to enable us to deal with the topic. On this score, reviewing the history of research of Early Judaism, Kraft and Nickelsburg (1986:13) remarks that “little work has been done on the question of the use of Semitic in nonPalestinian Hellenistic areas (e.g. Alexandria, Antioch).” It is said that there were one million Jews in Syria, Egypt, Babylonia and Asia Minor, who thus outnumbered the Palestinian Jews by three to one in the time of Philo of Alexandria. See Baron 1952:170-1. However, Russell (1967:103-106) persuasively suggests that this figure was much exaggerated.

150

Bilingualism of Jews in the First-Century Diaspora

tioch, Syria, Jerusalem, or Alexandria.56 The written place of the Gospel of Mark has been argued to be Rome, Galilee, Antioch, or small and rural Syria (Telford 1999:13-4). When it comes to the provenance of the Gospel of Luke, although Kümmel (1975:151) remarks that “There is no convincing argument for any of these conjectures, and we can say for certain only that Luke was written outside Palestine,” Fitzmyer (19815:116) suggests that the gospel was written in Syrian Antioch, which was bilingual. Finally, the fourth gospel is assumed to have been written in Syria (Antioch), Alexandria, or Asia Minor (Ephesus). Consequently, Antioch is the strongest candidate for the origins of the four gospels. One of the most significant reasons that these places of composition of the four gospels have been suggested is that they were known as bilingual regions. In this respect, the bilingualism of the three cities will provide fresh insights into issues of the directionality of the transmission of the gospel tradition in relation to the Sitz im Leben of the four gospels.

4.2 Alexandria It has been said that the Egyptian Jews were considered to be Greekspeakers. In relation to the linguistic milieu of Ptolemaic and early Roman Egypt, Cowley (CAP xiv, 191, 200) assumes that “it is unlikely that Aramaic continued in popular use in Egypt long after the time of Alexander the Great.”57 And he also states that in Egypt “Aramaic probably gave way to Greek by about 300 BCE.”58 Margolis and Marx (1927:129) likewise assume that “the younger generation spoke Greek, casting behind them the Hebrew language, or the Aramaic, which then had begun to displace Hebrew at home, at least in the rural districts.” As for the linguistic milieu of Alexandrian Jews, Silva (1980:209-213) applies bilingualism theory to the linguistic milieu of the Alexandrian Jews

56 57

58

Davies and Allison (2004:138-9) summarize the views of the origins which scholars have suggested. When he dates Aramaic papyri nos. 81, 82, and 83 (1923:190-1), his argument seems to be premised on the assumption that “No date is given, but the many Greek names suggest the Ptolemaic period, and this is corroborated by the character of the writing, which shows a much later stage of development than that of the Elephantine documents. It is unlikely, however, that Aramaic survived, even in individual cases, long after the time of Alexander, and we shall perhaps not be far wrong in assuming a date about 300 B.C.” CAP xv. For the information of replacement of Aramaic by Greek in the period of Ptolemaic kings, see Thompson 1994:67-83.

Alexandria

151

and, surprisingly, concludes that Jewish immigrants from Palestine into Alexandria quickly and thoroughly changed their language from Aramaic to Greek without a transitory period of bilingualism in Aramaic and Greek. 59 Along similar lines, Sterling (2001:273-4; cf. Feldman 1993:51-6) effectively summarizes five reasons why the Jewish Diaspora in Alexandria should be referred to as a Greek-speaking community. Sterling’s reasons and assumptions are: (i) the epigraphic excavations in Alexandria, which indicate the interval between the third century BC and the second century AD; (ii) all literature found from Alexandria was written in Greek; (iii) Greek was the norm for Jews in Alexandria; (iv) the populace used the Septuagint, and (v) Philo did not know any Semitic language. Another consideration that may be added to these five points is that Jews did not know or learn Egyptian.60 For these reasons given by a variety of scholars, it is supposed that the Alexandrian Jewish community in the first century was a Greek-speaking community. However, a sociolinguistic approach to the study of bilingualism can shed insight into the linguistic milieu of the Alexandrian Jewish community of the Ptolemaic and early Roman periods. Literary evidence now suggests that Alexandria was a bilingual city in Greek and Egyptian and there were many bilinguals in Greek and Egyptian.61 This indicates that the Jewish community in Alexandria seems to have been bilingual in Aramaic and Greek (or Egyptian) during the Ptolemaic as well as the early Roman periods. Bilingualism among Alexandrian Jews may be proposed based upon four basic reasons. As noted previously (§4.1), these reasons will be explored in more detail.

4.2.1 Bilingualism of Ptolemaic and Early Roman Alexandria Recently, scholars have reached a consensus that the linguistic milieu of early Roman Egypt was bilingual in both Greek and Egyptian since the

59

60

61

If Jewish immigration into Alexandria is unique, restricted, and small, Silva’s idea would be right. However, it seems that he disregards successive and massive immigrations, and periodic connection with Jerusalem. Hengel (1980:93) denies the possibility that Egyptian Jews spoke Demotic Egyptian; Griggs (1990:15) also proposes that “it is certain that the Jews did not use Egyptian as their spoken language.” However, some Jews spoke the Demotic, as will be suggested later. The linguistic milieu of Roman Egypt should be called bilingualism rather than diglossia; see Adams 2003:754.

152

Bilingualism of Jews in the First-Century Diaspora

Ptolemaic period. 62 This is based on a large collection of bilingual documents, many of which were relatively recently discovered (e.g. decrees, contracts, receipts, mummy labels, and ostraca) from both urban and rural areas.63 Jews inhabited all of Ptolemaic-Roman Egypt, in cities as well as in the countryside, since about the fourth century BC.64 Furthermore, it is intriguing that four types of writings in hieroglyphic, hieratic, demotic, and Coptic forms are still to be found even from the Roman Egypt period.65 Hieroglyphics or Demotics were still

62

63 64

65

Missiou 2007:336; Fewster 2002:228-236; Taylor 2000:335; Thompson 1994:73; 1992:324; Muhs 1992:249-51; Youtie 1975:201-221; Gignac 1976:46-8; MacMullen 1966:1, 7; Jones 1940:32-3. In terms of the linguistic situation, Fewster (2002:224-5) suggests that Greek was used as an administration language and the majority language was Egyptian; in addition, other minor languages such as Aramaic, Carian, and Nubian were also employed in restricted areas. However, generally speaking, it can be said that it was bilingual in Greek and Egyptian. Riggs 2005:100, 119-22, 203-5, 250-1; Fewster 2002:228-236; Midgley 1992:237-41. Papyri demonstrate that Diaspora Jews in Egypt were dispersed in cities as well as villages; refer to Mélèze-Modrzejewski 1993:72-6; Hegermann 1989:2.132; Tcherikover 1963:10. Tcherikover (CPJ 1.8) reports a Jewish synagogue list in Egypt in which includes cities as well as villages. Papyri indicate that Egyptian Jews were in rural areas; see CPJ no. 409 (Euhemereia in the Fayûm, 3 BC), no. 411 (Philadelpheia in the Fayûm, 13th Nov. 3 AD), no. 413 (Tebtynis, 16 AD), no. 416 (Philadelpheia, 25 AD), no. 420 (420a; Euhemereia, 26 AD, 420b; Euhemereia, 28-9 AD), no. 421 (Arsinoe in the Fayûm, 16th May, 73 AD), no. 427 (Apollonias in the Fayûm, 10th Feb, 101 AD), no. 430 (Arsinoe in the Fayûm, 105 AD), no. 432 (Arsinoe, 113 AD), and no. 433 (Philadelpheia, 1st century AD). One of the reasons that Jews spread to rural areas in Egypt is the slave trade. Tcherikover (1959:68-9) suggests that merchants bought slaves in Syria and sold them in Egypt; for instance, Johanna, probably a Jewish slave-girl, belonged to Apollonius in one of the villages of the Fayûm. Riggs 2005:35, 119-22, 203-5; Fewster 2002:225; Mertens 1992:233. The four Egyptian languages are as follows: (i) Hieroglyphic script is the oldest and archaic form. It was still used in Roman Egypt as a religious monumental script. (ii) Hieratic is simply hieroglyphs written quickly and cursively. It was a script for non-monumental writing. (iii) Demotic refers to both a script and to a particular stage in the Egyptian language. It is a cursive script and is derived from hieroglyphs. Demotic was used as the standard documentary script in the Persian and Ptolemaic periods. Since the Roman period in Egypt, it was used in a limited fashion and by the second century AD it was used only in tax receipts and the temple. During much of the Roman period in Egypt, most Egyptians could not write in their own languages (§1.5.2.4). Instead, when they wrote, they had to translate into Greek. (iv) Coptic refers to the Egyptian language written in Greek script with a few additional letters. Egyptians could get access to learning Coptic easier than Demotic. The earliest Coptic materials were used by Christians from the fourth century AD. For a detailed explanation of the four languages, see Fewster 2002:225-6; Ray 1994:51-66.

Alexandria

153

used in urban and rural areas of the early second century.66 Egyptian scribes in Ptolemaic and Roman Egypt also should have had to learn to write Greek as well as Demotic (Riggs 2005:18-9; Tait 1994:208).67 During the Ptolemaic and Roman periods, there were both Greek and Demotic school-exercises including word-lists and grammatical exercises (Tassier 1992:311-5). A variety of literatures and para-literatures in Demotic in the Ptolemaic and Roman periods68 were written or rewritten in Demotic and read among Egyptians (Riggs 2005:34-5). Alexandria in the Ptolemaic and early Roman periods was a bilingual city as well, although it was the most Hellenized city of the Roman Egypt. There are four reasons for this: (i) Various Demotic writings mentioned above would have been read among Alexandrian Egyptians. Egyptian fiction in Demotic, for instance, was supposed to have been written by priests and widely circulated among them (Tait 1994:203-8). Tait (1994:207; cf. MacMullen 1966:8) says that “this is hardly surprising, as in the Ptolemaic and Roman periods the priesthood roughly constituted that portion of the population that was literate in Demotic.” The upper-status Egyptians, such as priests and scribes, were bilingual in Demotic and Greek. The circulation of Demotic literature in the Ptolemaic and early Roman periods assumes that Demotic literatures in oral form were circulated among upper-status as well as, at least, lower-status Alexandrians. (ii) One of the reasons that Demotic was used in the Ptolemaic and Roman periods is that priests and scribes performed their rituals in their vernacular languages (Riggs 66

67

68

Riggs 2005:62-4, 2000:136; Grajetzki 2003:128. Jasnow (1990:89-96) also suggests that four demotic texts were excavated from Thebe and Hawara (?) in the Late Ptolemaic and early Roman period. On the other hand, Youtie (1975:205) suggests that “most Egyptians were so far removed from Greek as a language to be read or written that they had never even learned to speak it, at least sufficiently to maintain their side of an intelligible dialogue.” In the judicial cases, he (1975:205) also states that Egyptians often enlisted the aid of official interpreters even in fourth century AD. Tcherikover (1959:21) also notes that “the Egyptian language prevailed officially, assuming the form of a holy tongue. As a spoken language, too, Egyptian did not give way to the official language; not only in Upper Egypt (the Thebais), whither Hellenism hardly penetrated, but even in the Hellenized Fayûm, the natives frequently resorted to the assistance of interpreters when they had need of contact with the authorities.” Mertens (1992:234) illustrates a variety of the genres of Demotic literatures. The numbers of the genre of the Demotic literatures are as follows: narrative (136), poetical (2), wisdom (39), historical/prophetical (12), religious (24), mythological (12), funerary/mortuary (36), magical (18), omen literature (27), juridical (8), medical (11), mathematical (7), astronomical/astrological (44), school exercises (38), onomastica and word-lists (28), scientific and others or not specified (80) and possible literary texts (13).

154

Bilingualism of Jews in the First-Century Diaspora

2005:26). It is interesting that a Demotic version of The Book of the Dead was written as late as 64 AD (Riggs 2005:29).69 (iii) The royal library in Alexandria included Demotic bilingual literatures and some of them were translated into Greek (Fraser 1.681-7). (iv) Although inscriptions in Demotic continued to the fifth century AD and non-literary ostraca in Demotic to the third century AD,70 Demotic documents disappear in the second century AD. 71 This means that although documents (e.g. business documents) were written in Greek from the first-century AD, one can say at least that Demotic was still spoken all over the country.72 The possibility that Egyptians used Egyptian as their spoken language is also proved by the appearance of Coptic later, as Coptic is the “Egyptian language” written in Greek.73 In summary, the linguistic milieu of Alexandria of the Ptolemaic and early Roman periods was bilingualism. Many Alexandrians would have been bilingual in Greek and Egyptian. It seems that upper-status inhabitants used Greek as their matrix language, whereas lower-status inhabitants in the cities and most inhabitants in rural areas employed Egyptian as their matrix language. There would have been monolinguals also. Some upper-status inhabitants might have spoken only Greek, whereas some lower-status inhabitants only Egyptian depending on their regional distribution or educational level. Egyptian bilingualism hints at a linguistic distinction in the Jewish community in Alexandria between upper-status and lower-status speakers. Although the Alexandrian Jews comprised the poorest people 69

70 71

72

73

The cultural bilingualism of Ptolemaic Alexandria is supported by bilingualism in the tombs of Ptolemaic Alexandria; see Venit (forthcoming); I appreciate her sending this article along with permission to cite from it, although the book will be published later. Youtie (1975:203) mentions that Egyptians spoke Egyptian up to 7 century AD. Youtie (1975:203) suggests that the demise of Demotic stemmed from the difficulty of learning Demotic script. However, Lewis (1993:276) persuasively criticizes this view and proposes that the decline of Demotic documents was due to Roman policy. The Roman government forced the Egyptian administration to refuse Demotic documents. Maehler 2004:11. Thompson (1994:73) also asserts that although Greek had become the main language of administration, “at the same time surviving ostraca show the continued, but by no means exclusive, use of demotic for tax receipts in the towns and villages of Egypt – an indication of the spoken language of the majority of the population, including local officials.” See chapter 4 n.64. Youtie (1975:203) also claims that “The Egyptian language maintained itself as a medium of communication, both spoken and written, for almost a millennium under Greek, then Roman domination, with no break in the spoken tradition. There was, however, a late shift from the indigenous script known as Demotic to a Greek script.”

Alexandria

155

to the richest occupied in a variety of professions (Barclay 1996:115-6),74 it is known that there were two Jewish groups in Roman Alexandria under the influence of two opposite trends of Jewish tradition and assimilation to Hellenism. 75 One group consisted of upper-status Jews who tried to maintain their positions in the Greek society and to become acculturated in Greek society (Tcherikover 1963:20-5). Philo (Flacc, 57) enumerates five upper-status professions among Alexandrian Jews around 38 AD: shipping traders, merchants, artisans, landowners and bankers (e.g. Philo’s brother Alexander; Ant. 18.159).76 Upper-status Jewish boys, like Egyptian boys, were usually educated to learn the works of a wide range of Greek authors and the grammar and meaning of classical texts (cf. §3.2.5). Goodman (1983:73) considers this as well, suggesting that “in the first century Jews in Egypt may well have undergone the same curriculum (as Egyptians).” Although the cost of education was very high (Goodman 1983:72-3; cf. see Doran 2001:94-115), upper-status Jews wanted their children to get a Greek education in the gymnasium (Sterling 2001:277). Bilinguality would have offered a great advantage for any member of the society.77 Generally speaking, most upper-status Alexandrian Jews could have been bilingual in Aramaic and Greek. The other group of lower-status Jews constituted the majority of Alexandrian Jews and comprised simple peasants, craftsmen, shepherds, and small merchants who rejected Hellenism and were closer to Jewish tradition and Jerusalem.78 Most lower-status Alexandrian Jews who could not learn Greek because of economic or geographic reasons, may have been bilingual in Aramaic and Egyptian. It is highly possible that some of them might have spoken Egyptian as their embedded language instead of Greek.79 Consider the following: (i) in a situation of language contact, people typically learn their second language from those with whom they have (i.e. neighbors) unless they receive formal 74 75

76 77 78 79

Barclay (1996:103-24) specifically suggests the levels of assimilation among Egyptian Jews as high, medium, and low assimilation. Also see Paget 2004:146. In this sense, Tcherikover (1963:24) proposes that, faced with persecution in 38 AD, “Alexandrian Jewry was divided into two parts: on the one hand, the rich and the intellectuals defending their positions in Greek society and their civic rights by means of negotiation and explanation, and on the other the common people devoted to their ancestral tradition and ready to defend it by force of arms.” For detailed information on the five occupations, see CPJ 1.48-50, no. 152. Moneylenders are also mentioned by Philo (Spec. 2.75; Virt. 82-3). Fewster 2002:242-3; Tomson 2001:34; see §3.2.5. Tcherikover 1963:20-5; CPJ 1.50. Tcherikover (1963:10) mentions that “a Jew who lived in a village or a provincial town was far from the Hellenic spirit.”

156

Bilingualism of Jews in the First-Century Diaspora

education. (ii) In a bilingual circumstance, the language shift is related to adstratum or superstratum. Lower-status Jews who lived among lower-status Egyptians, whether in cities or rural regions, could have regarded Egyptian as their adstratum or superstratum. Accordingly, most upper-status Alexandrian Jews may have used Aramaic or Greek as their matrix language, whereas most lower-status Alexandrian Jews, who were the majority of Alexandrian Jews, may have used Aramaic or Egyptian as their matrix language.

4.2.2 Archaeological Evidence All this is also borne out by archaeological evidence: Inscriptions of Ptolemaic and early Roman Egypt demonstrate that Jews still used Aramaic and Demotic. Cowley (CAP 190-203) explains this by describing three Aramaic fragments from Elephantine in the third century and fourth centuries BC.80 (i) No. 81 (about 300 BC) is a 10-column papyrus of a business document (= Torrey 1937:4). (ii) No. 82 (early third century BC or much after 300 BC) is a 14-line papyrus of a Jewish legal document (Torrey 1937:4). (iii) No. 83 (between 400 and 300 BC) is 30 lines of names and accounts. Concerning ostraca of Ptolemaic Egypt, Mark Lidzbarski reports that three ostraca from the Greek period were written in Aramaic by Egyptian Jews. The first one (ESE II. 243-8) is a memorandum, a list of names and accounts in the second or third century BC.81 The other two ostraca (ESE, III.22-6) in the third century BC are at the Library at Strasbourg. One is a private letter concerning the sending of products. The other ostracon is also a memorandum of a list of names and accounts. Torrey (1937:5) points out that “it is natural to suppose that the language of the memorandum was that of the community where it was written.” Attention should also be paid to the Nash Papyrus. The papyrus, written in Hebrew, belongs to the Maccabean age (i.e. 165 and 37 BC) and might come from Fayûm (Albright 1937:143-76).82 It contains the Ten Commandments (Ex. 20:2-17 or Deut. 5:6-21) and the Shema (Deut. 6:4-5). According to Albright (1937:175), what is interesting is that the Nash Papyrus might have been used as a lectionary or for teaching purposes among Egyptian Jews. This oldest Hebrew biblical papyrus might indicate that Egyptian Jews knew a Semitic language (JIGRE 5). 80 81 82

These papyri do not refer to Elephantine documents of the fifth century BC. Lidzbarski (ESE 2.243-8) judges that the date of the ostracon was the second century BC, whereas Torrey (1937:5) decides that the date is the third century BC. For a long discussion of the date, see Albright 1937:150-72.

Alexandria

157

Torrey also suggests three pieces of evidence: (i) Nine of ten Alexander tetradrachms (318 BC) bear Aramaic graffiti from the Demanhur hoard (IGCH n.1664).83 These seem to be marks of identification. The graffiti indicate that the ordinary written language of those who inscribed these graffiti was Aramaic. For these reasons, he (1937:7) concludes that the “Jews of Egypt continued, among themselves, to use their own language.” (ii) The prefixed letters to 2 Maccabees were originally written in Aramaic in the second century BC, and translated into Greek (Torrey1937:5-6; 1900:225-7). This shows that Aramaic was commonly used among Egyptian Jews. (iii) A Roman tribune (Acts 21:37-8) considered Paul to be an Aramaic-speaking Egyptian Jew because he was surprised that Paul spoke Greek to him. This means that the Roman tribune thought that the ordinary language of the Egyptian Jews was not Greek but Aramaic (Torrey 1937:6).84 Furthermore, there are three epitaphs excavated from Roman Egypt. (i) Taken from an area around Al-Minya on the Nile south of Oxyrhynchus, an epitaph (?) was written in Hebrew of the second century AD or later (CIJ II no. 1533; CPJ III no. 1533; JIGRE no. 118). (ii) Excavated from Antinoopolis, an epitaph was written in Hebrew of the second century AD or later (CIJ II no. 1534; CPJ III no. 1534; JIGRE no. 119). (iii) Although its origin in Egypt is not clear, there is a mummy label in Hebrew of the second century AD or later (JIGRE no. 133; CIJ II no. 1536; CPJ III no. 1536). These epitaphs indicate that some Egyptian Jews knew Aramaic or Hebrew up to the second century AD. There is literary and archaeological evidence to show that some Alexandrian Jews in the Ptolemaic and early Roman periods knew Aramaic. Blau (1913:11; cf. Fuchs 1924:114-122) suggests that baraitha in Tosefta Kethuboth 4.9 (cf. b. Baba MeziȆa 104a, j. Yebamoth 15a) indicates that Aramaic was used in Jewish legal documents of Alexandria because the quoted text of the Alexandrian kethuboth is Aramaic. However, Tcherikover considers that the text does not prove the language of the community. He (CPJ 1.30 n.77) argues that “Since the Palestinian kethuboth were usually drawn up in Aramaic, it was natural to quote them in this language.” Nevertheless, the existence of the text indicates that the Alexandrian Jewish community read the Aramaic text. One should also note with interest that Horbury and Noy present five Jewish inscriptions (JIGRE nos. 3, 4, 5, 15, 17) from Ptolemaic and Roman Alexandria, which indicates that a Semitic language was used 83 84

Its modern name is Damanhûr. For two kinds of possibilities for the reason for the tribune’s surprise, see Barrett 1994-1998:1024-5.

158

Bilingualism of Jews in the First-Century Diaspora

among Alexandrian Jews until the second century AD and later. (i) No. 3 excavated from El-Ibrahimiya, Alexandria, is an epitaph in Aramaic with Hebrew proper names from the early Ptolemaic period.85 (ii) No. 4 excavated from El-Ibrahimiya, Alexandria, is also an epitaph in Aramaic or Hebrew of the early Ptolemaic period.86 (iii) No. 5 excavated from El-Ibrahimiya, Alexandria, contains Aramaic letters and numerals of the early Ptolemaic period.87 (iv) No. 15 excavated from the Roman Tower, Alexandria, is a votive column base written in Greek (dedication) and Hebrew (shalom) of the late Roman period.88 (v) No. 17 excavated from Alexandria is a part of a lintel written in Greek (dedication) and Hebrew (shalom) from the Roman period.89 Horbury (1994:13; JIGRE 4-5) suggests that “Some Egyptian Jews knew Aramaic under the Ptolemies, and Hebrew in the later Roman empire.” Horbury also adds some points to the evidence for the Jewish use of Aramaic in Roman Alexandria. (i) Philo (Flacc. 39) records that the Aramaic word Ma,rin was mimicked when Gentiles insinuated that some Jews use Aramaic in Alexandria (38 AD). This implies that some Jews used Aramaic. (ii) Paul’s use of mara,na qa, without any explanation in the letter sent to the Gentile church (1 Cor. 16:22) presumes that Jews of the Diaspora in the Hellenistic city might know some Aramaic (Horbury 1994:14).90 Furthermore, an Aramaic poem preserved in the fragmentary Targum on Exodus in Alexandria (fourth or fifth c.) as well as the Antinoe Marriage Contract written in both Aramaic and Greek (fifth c.) confirm that Aramaic was used among some Jews in Egypt (Horbury 1994:16). Horbury (1994:17) states, “The surviving Greek literature of Egyptian Jews is not fully representative of Jewish speech, which in Ptolemaic and early Roman Egypt also embraced Aramaic.” Based on the above evidence, he (1994:18) persuasively insists that “we 85

86 87 88 89 90

No.3 = CIJ II no.1424; ESE II.49; CPJ III no.1424; in relation to this epitaph, Hengel (1980:91) states, “The fact that in the early Ptolemaic necropolis of Alexandria we find Aramaic and Greek inscriptions on Jewish tombs among Gentile graves accords with the cohabitation of Jews and Macedonians or Greeks in mixed military units and military settlements.” Barclay (1996:29) also suggests that “Certainly, individual Jews, probably military settlers, are well attested in early Ptolemaic inscriptions found in an Alexandrian necropolis, written in both Aramaic and Greek.” However, Barclay (1996:30) denies the continuous bilingualism in Aramaic and Greek/Egyptian after the Hebrew Scriptures were translated into Greek. No.4 = ESE III. p.50; CIJ II no.1425; CPJ III no.1425. No.5 = ESE III. p.50; CIJ II no.1426; CPJ III no.1426. No.15 = CIJ II no.1438; CPJ III no.1438. No.17 = CIJ II no.1437; CPJ III no.1437. Sociolinguistically, the Aramaic term seems to be codeswitching. This will be explained in detail (chapter 8).

Alexandria

159

should probably allow for some circulation of Aramaic writings, as the Aramaic epitaphs suggested.” In terms of the Greek Old Testament, many scholars including Sterling, as mentioned before, assume that the Alexandrian Jews’ use of the Septuagint strongly supports the view that the Alexandrian Jews spoke Greek. However, Millar (1992:99) proposes that the publication of an Aramaic ketubah from Antinoopolis in Egypt (417 AD) reopens the question about Diaspora Jews and the use of Semitic languages. While acknowledging that the language of the Jewish Diaspora was undoubtedly Greek and that the LXX was used, Millar (1992:99) persuasively argues that this “does not prove that the Hebrew Bible was unknown in the Diaspora; nor does it prove that Hebrew or Aramaic was nowhere spoken or used.” The Greek translation of the Pentateuch might be strongly influenced by the Alexandrian atmosphere, in which books of eastern religions were translated into Greek. 91 It seems unlikely that the translation was made because Alexandrian Jews were not able to read Hebrew at all. Many Alexandrian Jews were bilingual and Aramaicmatrix speakers. There is also inscriptional evidence to support Demotic-speaking Jews. Intriguingly, one of the ten Aramaic graffiti, as Torrey (1937:7) mentions, was incised in Demotic Egyptian. The inscription is hp-hp p nf (?), that is, “Hphp the sailor (?).” He (1937:7) proposes that the ten graffiti are of Jewish origin of the fourth or the early third century BC. Even though it is only one out of the ten graffiti, palaeographic proof is excellent in quality. Unfortunately, Torrey pays no attention to the Demotic 91

As to the origin of the Septuagint, various explanations have been suggested, such as Palestinian, liturgical, educational, or interlinear origins; for theories of the origin in detail, see Marcos 2000:53-65; McLay 2003:101-103. However, in relation to the origin, one should also pay attention to the relationship between the origin and the Alexandria Library, the first universal library in history. Irenaeus (Haer. 3.21.2 in Eusebius HE 5.8.11-15) reports that Ptolemaic I ordered Demetrius of Phalerum to found a research center (i.e. the Mouseion) and the Royal library (for Mouseion, see Maehler 2004:1-14; for the Alexandria Library in detail, see Fraser 1.305-35; ElAbbadi 1990:73-141; MacLeod 2004). The royal library collected Greek literature and translated literature from other languages such as Egyptian and Persian (Fraser 1.330). The library also includes literature about Zoroastrianism and even Buddhism (El-Abbadi 2004:170). El-Abbadi (1990:170) mentions, “Intellectual curiosity and academic interest no doubt motivated scholars to study, translate and write about eastern religious and cultures.” Watts (2006:148) also notes that “its [the library’s] curators aspired to acquire a copy of every work in the Greek language.” This background of the Alexandria library, to some extent, corresponds to a dramatic story according to the Letter of Aristeas, according to which King Ptolemy II Philadelphus, on the advice of Demetrius of Phalerum, ordered a translation of the Pentateuch for his library.

160

Bilingualism of Jews in the First-Century Diaspora

graffito in terms of the Jewish language in Egypt.92 The existence of the Demotic graffito strongly suggests that some Egyptian Jews knew Demotic Egyptian in the fourth or the early third century BC. As some Alexandrian Greeks tried to learn Egyptian (§4.1.4), so did some Alexandrian Jews. Accordingly, the archaeological evidence proves that some Alexandrian Jews spoke Aramaic as well as Greek/Demotic in the Ptolemaic and early Roman periods.

4.2.3 Successive Immigration Palestinian Jews constantly immigrated to Alexandria. This would make it likely that Alexandrian Jews continued to speak Aramaic (Hengel 1980:93). Tcherikover demonstrates that Aramaic was spoken until the Graeco-Roman period. He (CPJ 1.30) states that “there was a continuous stream of Syrian immigrants (including Jews) during the entire Hellenistic and Roman periods.” Horbury (1994:15) suggests that both continuous immigration to and temporary sojourn in Egypt caused Jews there to become bilingual (Greek and Aramaic). Josephus (C. Ap. 194) mentions that a large number of Palestinian Jews were moved into Egypt after the death of Alexander. Josephus (C. Ap. 1.1869; cf. C. Ap. 2.33; Ant. 12.9, Aristeas 12-4) also records that after the battle of Gaza (312 BC) 100,000 Jews, including Hezekiah, a Jewish high priest, accompanied Ptolemy I Soter to Egypt (Tcherikover 1959:56). Among them, 30,000 were housed in fortresses and the rest (i.e. old men and children) were given to Ptolemy’s soldiers as slaves (Borgen 1996:75-6).93 This means that some Jewish captives could learn Greek as a second language from upper-status solders, whereas others might learn Demotic from lower-status Egyptians. In the second part of the second century BC, the immigration reaches its highest point. Papyri excavated from Egypt show that the proportion of Semitic names increased again.94 The unexpected increase of Semitic names seems to be 92 93 94

Albright (1937:156 n.29a) also has no comment on the No 1 coin written in Demotic, when he reviews Torrey’s the publication. Borgen mentions that later, Ptolemy II Philadelphus (284-246 BC) gave freedom to the slaves. CPJ 1.28, pp. 194-226. Kasher 1985:9. Hengel (1980:86) suggests two reasons for the increase of Semitic names; one is that “Jewish mercenaries were organized into independent units and gained considerable political significance under their own troop leaders.” The other is that “this increase in Jewish names is also a sign that Jewish national self-awareness had also increased as a result of their stronger political position in Egypt and in the home country of Palestine.” However, Honigman (1993:93-127) rebuts Hengel’s position due to chronological and methodological

Alexandria

161

the result of massive immigration to Egypt (ca. 165 BC). Josephus (Ant. 12. 387; cf. Bell. 7.423) reports that Onias IV fled to Alexandria (ca 162160 BC) with his Jewish community, and was warmly welcomed in the Ptolemaic court by Philometor (180-145 BC).95 The community settled in Leontopolis as a military force (Ant. 13.62-73; Bell. 7.421-36). In addition, Jewish immigration continued in the first century AD, as Claudius’s letter (41 c. AD) hints, “Jews are forbidden to bring into the city other Jews from Egypt or from Syria” (Horbury 1994:15 [emphasis in the original]).96 Furthermore, Horbury discusses that, except for mercenary soldiers, the reasons for Jewish immigration were various. He (1994:16) notes that visits for trade or other purposes made it possible for Alexandrian Jews to continue to speak Aramaic, such the cases of the tax-farmer Joseph and his son Hyrcanus (Ant. 12.154-236) and stories about the political refuge of Onias, the Sicarii’s attempt to find its refuge in Alexandria, or Jesus’ flight to Egypt.97 The incessant immigration does not permit us to decide who the second-generation immigrants into Alexandria were, although admittedly the immigrants spoke Greek as their matrix language from the second-generation (§4.1.2). As a consequence of these factors, some Alexandrian Jews and particularly first-generation immigrants from Aramaic-speaking regions spoke Aramaic (Horbury 1994:16).

4.2.4 Periodic Connection with Jerusalem Alexandrian Jews had continuous contact with Jerusalem (Tcherikover 1963:23-7).98 In the words of Pearson (1986:210): “traffic between Jerusalem and Alexandria was extensive.” By all accounts, the interrelation between Jerusalem and Alexandria was strong (Tcherikover 1963:20-7). There were two major reasons to visit Jerusalem. (i) Philo (Spec. 1.68) problems. Despite all this, whether the Semitic names were Jewish or Syrian, we cannot deny the fact that the Semitic names increased in the second century again, which proves that the first-generation immigrants used Aramaic as their matrix language. 95 For detailed discussion of the date, see Barclay 1996:35-7. 96 Cf. CPJ no.153 p.41 lines 95-6. However, Tcherikover (CPJ 1.5 n.13, 2.54) notes that the “Syria” refers to Palestine. 97 However, in terms of the Jesus narrative, refer to Allison 1993:140-65; he argues that the story of Jesus is often viewed as a purposeful and therefore non-historical account that casts Jesus as a Moses-like character. 98 Egyptian Jews had a continuous relationship with Jews in Jerusalem; see, Kasher 1985:1-28. For instance, when the Seleucids were defeated by the Hasmonaeans, Jews from Egypt and Cyrene made an active contribution to the Hasmonaean revolt. See Bickerman 1933:233-4; refer to Ant. 13.74-79.

162

Bilingualism of Jews in the First-Century Diaspora

mentions that Alexandrian Jews took part in Jewish pilgrimage holidays in Jerusalem. (ii) Alexandrian Jews paid the temple tax. Although concurring with Tcherikover’s observation that “a number of Alexandrian Jews lived in Jerusalem and enjoyed an independent organization, holding services in their own synagogues,”99 this does not mean that they were not exposed to Aramaic-speaking contexts or that they did not know Aramaic. It is probable that constant visits to Jerusalem would have provided some Alexandrian Jews with opportunities to speak Aramaic, thereby enabling them to sustain it. The Jewish community in Alexandria can be called a bilingual community. From the perspective of Hamers and Blanc’s definition (§2.1.2) there were various kinds of bilinguals among the Jewish community in Alexandria based on their social status. These groups were: monolingual Greek speakers, early or late bilinguals of Aramaic and Greek, early or late bilinguals of Aramaic and Egyptian, and only Aramaic or Egyptian monolinguals.

4.3 Antioch Many scholars have placed the Antiochene Christian community between Palestinian Jewish Christianity and Hellenistic Gentile Christianity. As previously noted, it has been suggested that Antioch could bridge the gap between Aramaic-speaking Palestinian Christianity and Greek-speaking Hellenistic Christianity.100 One assumption is that the Jewish Diaspora in Antioch used Greek because of Hellenizing influences. As Jerome Crowe (1997:xiii) mentions: “in Antioch...both Jews and pagans spoke Greek.” However, it seems that there were many bilinguals in Greek and a vernacular Aramaic dialect among Antiochene inhabitants and that most Antiochene residents spoke Aramaic as their matrix language, as will be discussed. This suggests that most Antiochene Jews might have spoken Aramaic as their matrix language and that most Antiochene Christians might have spoken Aramaic as their matrix languages.

99 Tcherikover 1963:24; cf. Tosefta, Megillah 3 (2), 6; Acts 6:9. For the religionsgeschichtliche Schule, see §1.2. For detailed discussion, see introduction to this chapter; Hengel and Schwemer 1997:279-286.

100

Antioch

163

4.3.1 Bilingualism of Roman Antioch The linguistic situation in the Roman province of Syria101 was multilingual. While Greek was employed as the common language, vernacular languages in each region were still used for everyday purposes.102 The bilingual circumstances of the Roman province of Syria cannot be summarized simplistically because the contact of Greek with the Aramaic language family resulted in different consequences depending upon dialects, regions, and times (Brock 1994:149). Moreover, the bilingual distribution in the regions of Syria varied. First of all, Palmyra was a publicly and officially bilingual society in Greek and a local Aramaic dialect (i.e. Palmyrenean) as a whole (Kaizer 2002:27-34). Kaizer (2002:27-34; here 27) mentions, “Out of ca. 3000 inscriptions from Palmyra, mainly written in the local dialect of Aramaic, over 200 are bilingual, consisting of a Greek and a Palmyrene text.”103 He (2004:166; cf. Gzella 2005:445-58) states that “the local community expressed itself both in Greek and in a local Aramaic dialect (Palmyrenean).” In Palmyra, Palmyrene, Greek, Aramaic, Latin, and Arabic were used, and numerous bilingual documents are preserved written in Greek and Aramaic, Latin and Aramaic, or even in these three languages com101 Admittedly, the term “Syria” is too broad to define at that time (cf. chapter 5 n.34); according to Strabo (Geog. 16.1.1-2) Syrians extended “as far as the Cilicians and the Phoenicians and the Judaeans and the sea that is opposite the Egyptian Sea and the Gulf of Issis. It seems that the name of the Syrians extended not only from Babylonia to the Gulf of Issus, but also in ancient times from this gulf to the Euxine. At any rate, both tribes of the Cappadocians, both those near the Taurus and those near the Pontus, have to the present time been called ‘White Syrians,’ as though some Syrians were black, these being the Syrians who live outside the Taurus.” As Butcher (2003:271) demonstrates, several ancient writers (e.g. Lucian of Samosata, Philostratus, and Strabo) mixed “Syrian” with “Assyrian.” I will follow Millar’s “Roman Province Syria” (1993:236). He suggests that the Roman province of Syria refers to three different areas: (i) Northern Syria, “stretching across from the Mediterranean coast, and the two ports of Laodicea and Seleucia, through Antioch to the Euphrates,” (ii) the Phoenician coast, “in its northern part backed into the mountain-chain now called the Jebel Ansariyeh and in the south onto Mount Lebanon and then the hills of Galilee. Provincial territory seems indeed to have extended south not only to Ptolemais-Akko but beyond Mount Carmel to the small town of Dora,” (iii) “isolated beyond Anti-Lebanon was the ancient city of Damascus, and further south at least some of the cities of the Decapolis.” 102 Sartre 2005:291-5; Brock 1994:149-160; Millar 1993:xiv, 232-3, 1987a:143-64, 1971:7-8; Drijvers 1992:124-146; McCullough 1982:9; Hitti 1951:256; Bernhardy 1861:1.493. 103 As to the sources of bilingualism of Palmyra, Dr. Ted Kaiser of the Department of Classics and Ancient History, Durham University, one of the leading specialists in Palmyra, provided me with many valuable materials related to the bilingualism of the Roman Near East. I really appreciate his kindness in providing materials, and his careful reading and invaluable suggestions about chapter 4.

164

Bilingualism of Jews in the First-Century Diaspora

bined.104 Based upon parchments and papyri excavated, the language repertoire of Dura Europos is Greek, Latin, Hebrew, Jewish Aramaic, Syriac, Parthian, Palmyrene Aramaic, Safaitic, and Middle Persian.105 Aramaic was not the common language of the area any longer (ca 165255 AD), but it was still used with Greek (Kilpatrick 1964:217). In the case of Samosata, the capital of Commagene, Aramaic and Greek were used as well (Taylor 2002:305; Jones 1940:291). As mentioned before (§3.2.5), although Lucian became the best orator in Greek and Latin, he could speak only Aramaic during his youthful years in Samosata. This indicates, according to Jones (1940:291), that “even in Samosata, the former royal capital and the metropolis of the district, the lower classes spoke Aramaic.” 106 At Osrhoene, the capital of Edessa, Syriac and Greek were used, and many in Syria in the Late Antiquity were bilingual in Greek and Syriac (Sartre 2005:293; Brock 1994:149-160). According to Egeria’s pilgrimage, Syriac was continuously used in a Christian community at Edessa.107 Although Greek was used in Phoenicia, the majority used Aramaic as their spoken language (Baslez 2007; Sartre 2005:292-3; see §5.3.2.1). According to Millar (1987a:152-5; cf. Grushevoi 1985:51-4), the Nabataean kingdom used Nabataean from the end of the fourth century BC, which continued to be used until the fourth century AD; inscriptions in Nabataean and Greek have also been discovered. Furthermore, surprisingly, there might have been educational systems for learning the Aramaic language in the Roman province of Syria, and written Aramaic dialects such as Nabataean, Palmyrene, Syriac and Hatrene “were presumably transmitted by some system of training or education” (Butcher 2003:285). There is papyrological evidence for bilingualism in rural regions of Syria. Such rural bilingualism indicates that Greek was commonly used even in rural areas; vernacular languages were also employed by the inhabitants. One bilingual ostracon in Aramaic and Greek (25 July 277 104 Adams 2003:248-64; Taylor 2002:317; Drijvers 1995:31-42; Davis and Stuckenbruck 1992:265-83; Millar 1993:233, 319-336, 470; Hitti 1951:256; Ingholt 1938:93-140, 1935:57-120. What is more, as Millar (1993:327) and Hitti (1951:256) mention, public use of Palmyrene Aramaic was employed as an official language until the late in the third century. 105 Cf. P. Dura V.I; Kilpatrick 1964:217; Hitti 1951:256. 106 Millar (1993:456) takes a cautious attitude in regard to Lucian’s language; “we still do not know whether Lucian, or anyone else in Samosata, or in Commagene generally, spoke a dialect of Aramaic“ [emphasis in the original]. 107 She (Gingras 1970:125-6; chapter 47) remarks: “A portion of the population in this province knows both Greek and Syriac; another segment knows only Greek; and still another, only Syriac. Even though the bishop may know Syriac, he always speaks Greek and never Syriac.”

Antioch

165

BC) was found in a small town in Idumaea called Khirbet el-Kôm, which lies along Wadi es-Saffar between Hebron and Lachish (see Geraty 1975:55-62). Grabbe (1991:1.185) writes that “the ostraca show that Greek was already well established alongside the local language.” The second, from the Euphrates archive, is known as the Mesopotamia papyri from the mid-third century and was written in Greek and Syriac. Although the exact location is not certain, the papyri appear to be associated with a village called Beth Phouraia because one petition relates to a property dispute between two groups of villagers (Butcher 2003:143; Brock 1994:151-2). Thirdly, as mentioned before, parchments and papyri found at Dura Europos provide information about village life in relation to the military and civilian contexts (see chapter 4 n.105). The documents were written in Greek, Latin, Syriac, Palmyrene, Parthian, and Middle Persian and the location might be villages along the Khabur River, Ossa, near the confluence of the Khabur and Euphrates, and Qatna on the Khabur (Butcher 2003:144). Lastly, Macadam (1983:106) also cites a funerary dedication (unfortunately undated) to prove that Hellenization was limited among villagers of southern Syria. The inscription from the village of Shaqq¬ (Wadd. 2143) notes that Alexander was a pious high priest as well as a dragoman for the procurators. When the Roman officials came down to the villages, he traveled among the villages with them to offer his translation wherever needed. It is intriguing that a bilingual inscription in Aramaic and Greek excavated from Armazi, the ancient capital of Iberia (Georgia), makes it possible to assume that Syrians who resided in Armazi in the midsecond century were bilingual.108 Accordingly, even though Greek was the official language of the Roman Near East (Jones 1940:288), vernacular languages in Syria survived. Antioch-on-the-Orontes, the capital of Syria and the third largest city of the Roman Empire,109 was bilingual in Greek and Aramaic as well. Green (1990:313) surmises that “Aramaic remained the second language of Antioch, and was spoken as the vernacular throughout

108 Metzger (1956:25) suggests that the terminus ad quem of the inscription can be shown by the assumption that the Pharsman of the inscription might have referred to Pharsman III (145-185 AD), although Mussies (1983:357) supposes that the inscription dated from as late as 75 AD. 109 Hengel and Schwemer (1997:186, 196) assume that the population of Antioch was approximately 300,000, and that the Jewish population might have been between 30,000 and 50,000. According to Meeks and Wilken (1978:8), the population of Antiochene Jews was about 22,000. For more discussion of the population, see Zetterholm 2003:37-8.

166

Bilingualism of Jews in the First-Century Diaspora

Syria long after the Roman conquest.”110 Even some upper-status Syrians used Aramaic as well. Taylor (2002:314-5) also argues that “in Northern Syria, even among the educated elites...there were many who also spoke the local dialects of Aramaic.” He (2002:305; cf. Jones 1940:290) further demonstrates that “there are several references to higher-status primary bilinguals [in Antioch].” We know an upperstatus Christian who spoke only Aramaic even in the third century; Eusebius (HE, 4.30) indicates that Bardesanes was one of the most able Christians and wrote a great deal of literature in Syriac Aramaic, including poems, even in the early third century and that his pupils translated them from Aramaic into Greek (cf. Drijvers 1984:109-22, 190210). Macedonius, who came from outside Antioch, needed interpreters because he spoke only a local vernacular language (Theodoret, HR 13.7). Abraham of Cyrrhus, a bishop of Harran, spoke Aramaic (Theodoret, HR 17.9). Jerome (Vita Malchi 2; Vita Hilar. 22 cf. Ep., 17.2) states that some monks spoke Aramaic, and Theodoret (HR 5; cf. 14, 17) also mentions that monks knew only Syriac at Zeugma in the fourth century. On the other hand, Severian of Gabala could speak Greek with a strong Syriac accent (Socrates, HE 6.1). Eusebius (HE, 1.13) also quotes extensive documents written in “the language of Suroi” from the archives of Edessa. In light of these findings, a significant number of upper-status inhabitants of northern Syria in the first five centuries might have spoken Aramaic as their matrix languages.111 In this respect, there is no question that lower-status or rural inhabitants112 in Antioch spoke Aramaic as their matrix languages.113 This is because lower-status Syrians could not acquire Greek due to geographical reasons (spoken Greek) or economic reasons (spoken and written Greek) (cf. §3.2). Even in the fourth century, Syriac was used in 110 Also refer to Sartre 2005:291-6; Feldman 1993:14; Bickerman 1979:97; Avi-Yonah 1978:182; Mørkholm 1966:18; Lietzmann 1953b:2.258; Rénan 1869:182. 111 Taylor 2002:305, passim; Butcher 2003:285; Brock (1994:150) states that large numbers of the residents in towns were Aramaic-matrix bilinguals. 112 The term “inhabitant” is different from “citizen.” Although a high proportion of the population of a Graeco-Roman city was composed of free foreigners and slaves rather than citizens, the concept of inhabitants comprises all of them (Zetterholm 2003:24). In Antioch, there were three kinds of inhabitants such as AthenianMacedonian soldiers and colonists, indigenous Semitic-speaking people and Jews; see Downey 1962:12-3. 113 Taylor 2002:305; Brock 1994:150. This does not deny that many of lower-status Syrians spoke Greek as their embedded languages. Taylor (2002:314) states, “Quite apart from soldiers, veterans, and government officials, it seems clear that significant numbers of individuals in northern Syria, even outside the cities, had productive control of Greek.”

Antioch

167

Antioch. For instance, when neighboring Syrians from the countryside came to the assembly, Chrysostom (Cat. ill. 8.2) recorded: “Let us not look to the fact that their speech is different from ours. Let us note carefully the true doctrine of their soul and not their barbarous tongue.”114 Chrysostom (Cat. ill. 8.4) repeatedly stressed such issues as: “let us not look simply at their appearance and the language they speak.” Elsewhere, and with the same concern, he (Stat. 19.2) makes remarks about “a people foreign to us in language, but in harmony with us concerning the faith” when rural Syrians come to Antioch on Sunday. On market days and festivals, Syriac-speaking peasants flocked to Antioch, which indicates that there was lively interaction between Syriac-speaking and Greek-speaking Syrians, thus allowing Antiochene inhabitants to continue to hear Syriac. Furthermore, adding to the general picture, it is thought that the Syriac Peshitta might have been used by Chrysostom (Krupp 1991:75). Based on this, three points can be summarized. (i) The linguistic milieu of Antioch was bilingualism in Aramaic and Greek. (ii) There may have been many bilinguals among both upper-status and lower-status Syrians. (iii) The inhabitants’ competence in speaking Greek depends on their social status and, by and large, it is appropriate to assume that upper-status inhabitants spoke Greek as their matrix languages while lower-status inhabitants spoke Aramaic as their matrix languages.

4.3.2 Bilingualism of the Antiochene Jews Although there are limited sources of information about Antiochene Jews in the time of the Hellenistic and Roman era (Barclay 1996:243), it seems evident that large numbers of Jews resided in Antioch from the founding of the city (Kraeling 1932:131). Josephus (Bell. 7.43) suggests two reasons, namely geographic proximity and the favor of the Antiochene. Antioch’s proximity to Palestine as well as to Persia might have allowed Jews to live together at Antioch (Wilken 1983:35; Stern 1974:137-8). Josephus (C. Ap. 2.39; Ant. 12.119) claims that Seleucus Nicanor settled a large number of Jewish mercenaries in Antioch and granted them privileges equal to those of the Macedonians and Greeks who were the inhabitants then.115 Furthermore, Syria could have been 114 Ad pop Anti 19.2 in Cat ill 8.2; Bap. Inst. 120 (Cat ill is well documented in Bap. Inst., which I cite). 115 Josephus notes that Seleucus Nicator granted civil rights (politei,a) to the Antiochene Jews and that they still held this citizenship in Josephus’ own day. The issue of the citizenship is still in dispute; Meeks and Wilken (1978:2) regard Josephus’ reference

168

Bilingualism of Jews in the First-Century Diaspora

considered as an extension of Palestine (Wilken 1983:35). 116 Barclay (1996:242) proposes that “Jews from Judaea who settled … in Damascus, Ascalon, Tyre or Antioch were not considered by most of their contemporaries to have ‘entered’ Syria, simply to have moved from one part of it to another” [emphasis in the original]. It is of interest that rabbis debated whether to tithe produce in Syria.117 This implies that Syria was considered as an extension of Palestine. Jewish leaders— Onias III, Jason, and Menelaus—frequently visited Antioch (Stern 1974:138). That Onias III took refuge at Daphne near Antioch from his opponents (2 Macc. 4.33) indicates that there were considerable numbers of Jews in Antioch (Barclay 1996:245; Stern 1974:138). Herod the Great also visited Antioch (Ant. 15.218, Bell. 1.512). For these reasons, Antioch became one of the largest Jewish communities.118 Antiochene Jewish linguistic competence seems to be deeply related to social status as well. In Jewish communities of the Roman Near East, including Antioch, there was a social distinction between upperstatus and lower-status Jews. It is well-known that during a JewishGentile conflict (66-67 AD), upper-status Jews had different positions from those of lower-status Jews. As Josephus (Bell 2.267, 287; Ant 20.178) writes, upper-status Jews in Caesarea were keen to maintain peace between Jews and Gentiles, as upper-status Jews in Egypt did. One could speculate that this Jewish-Gentile conflict may have been an explosion of antagonism ignited by lower-status Jews and Gentiles (Barclay 1996:254). In regard to upper-status Jews,119 most in Antioch would likely have been retired soldiers and their descendants (Downey 1961:79-80). Josephus (C. Ap. 2.39; Ant. 12.119) notes that Jewish mercenaries fought in Seleucus’s army and that some of them settled in Antioch. There

116

117 118

119

as formal citizenship, whereas Barclay (1996:244-5) does not agree with it. For the debate, see Zetterholm 2003:32-7. Nevertheless, the dispute does not deny the fact that there were upper-status Jews in Antioch due to Jewish mercenaries; Kraeling 1932:137-8; Kasher 1985:9-10. As to “land of Damascus” in CAIRO DAMASCUS, although this issue is under dispute, Murphy-O’Connor (1974:221, 1996:69) argues that Damascus is a symbolic name for Babylon not for Qumran and that “land of Damascus” refers to Syria too. According to M. Halah 4.11, “he that owns [land] in Syria is as one that owns [land] in the outskirts of Jerusalem.” See §4.1.3. Kraeling remarks that “it would appear to follow that as early as the first century A.D. there were no less than three Jewish communities in and about Antioch, one west of the city near Daphne, one in the city proper, and one east of the city in the ‘plain of Antioch.’” Barclay 1996:21-2; Hengel 1980:85-92. For instance, Roman society also has few upper-status inhabitants; see Meeks 2003:51-73; Gill 1994:105-118; Clarke 1993.

Antioch

169

were many Antiochene Jewish artisans who were sent to Jerusalem to design and work at the temple (Bell. 7.45). We also know of some upper-status Jewish Christians like Manaen and Barnabas (Acts 13:1; cf. Acts 4:36-7). Like many upper-status Syrians, many upper-status Antiochene Jews might have been primary bilinguals or acquired bilinguals through intermarriage (i.e. veterans and soldiers), regional bilingualism, or formal education (§3.2.5). Usually, upper-status Jews might have spoken Greek as their matrix language. Moreover, in the case of acquired bilinguals, Hitti (1951:257) notices that “The educated Syrians, no doubt, studied Greek and wrote in it. But there is no reason to believe that they used it at home.” From what is known of the function of bilingualism, it would be appropriate to surmise that Jews who were late or acquired bilinguals used Greek for business while speaking Aramaic in the home. Another likely scenario is that most Antiochene Jews were lowerstatus inhabitants at Antioch. Many Jewish war prisoners taken in Palestine were brought to Antioch (see §4.1.2). They would have spoken Aramaic as their matrix languages like lower-status Syrians. Moreover, Jews also lived outside of Antioch. Malalas (Chronographia, 10.45; p. 261) reports that the Jewish community in Daphne was only seven kilometers south of Antioch and had been there since the time of Emperor Tiberius (14-31 AD). Second, Jews lived in rural areas near Antioch in the late second century (Tosefta Demai 2.1).120 Jewish peasants grew rice in the plain of Antioch,121 and agricultural affairs clearly require permanent residence for a sustained period of time. Jewish farmers may have lived in the countryside of Antioch since at least the early second century. Jewish peasants could have been numerous. It is said that rabbis visited them to tithe produce. Libanius (Or. 47.13) also reports that some Jews, his tenants, lived in the countryside of Antioch for four generations. Libanius of the fourth century knew that the Jews had lived there as his tenants for at least four generations and they may have been there even longer. These two cases indicate that Jewish peasants lived in the rural areas of Antioch around the first and second century.122 The Jews in the countryside of Antioch may have spoken Aramaic as their matrix language. 120 Although the law of Demai is traditionally attributed to the Hasmonean High Priest Johanan Hyrcanus (135-104 BC), this tractate seems to indicate the situation in the latter half of the second century; see Zeraim, 2.52. 121 Although there are several opinions about the location, most scholars think of the place as the plain of Antioch; see Sarason 2005:2.889. 122 As mentioned (§4.3.1), Chrysostom notes that lower-status Syrians from rural areas around Antioch flocked together into Antioch on feast days and Sundays.

170

Bilingualism of Jews in the First-Century Diaspora

Accordingly, upper-status Antiochene Jews spoke Greek as their matrix language. Acquired or late bilinguals, unlike early or primary bilinguals, may have used Greek in business and Aramaic at home. And every community is composed of a majority of lower-status inhabitants and a minority of upper-status. Generally speaking, lowerstatus Antiochene Jews spoke Aramaic as their matrix language. This indicates that most Diaspora Jews of first-century Antioch spoke Aramaic as their matrix language.

4.3.3 Bilingualism of the Antiochene Christians The early Christian mission made great strides among lower-status Jews and Gentiles. Celsus (C. Cels. 3.44), writing against Christianity, disparages Christianity because it attracted “only the foolish, dishonourable and stupid...slaves, women, and little children.” He (C. Cels. 3.55) mentions that Christians are “wool-workers, cobblers, laundryworkers, and the most illiterate and bucolic yokels.” Furthermore, he (C. Cels. 1.62) considers Christianity to be a movement of people from the lowest status. From the start “Jesus collected round him ten or eleven infamous men, the most wicked tax-collectors and sailors…collecting a means of livelihood in a disgraceful and importunate way.” Although Celsus’ intention is to insult and depreciate Christianity, his information may be correct and the majority of Christians could actually have been among the lowerclass of the society, as 1 Corinthians 1:26-31 (cf. Mt 11:25-6) indicates. However, Stark (1996:33) suggests that the early Christian communities had a number of upper-class Christians.123 He (1996:33) concludes, “If the early church was like all the other cult movements for which good data exist, it was not a proletarian movement but was based on the more privileged classes.” Witherington (1995:114) suggests that the Corinthian Christians who were 123 He (1996:29-33) cites the list of biblical scholars who hold this view. Especially, Theissen (1982:73) calls them “a dominant minority.” Refer to the definitions of upper-status and lower-status, see chapter 4 n.17. As a matter of fact, fishermen and tax-collectors do not always belong to the lowerclass. Furthermore, in terms of members of early Christianity, Gamble (2004:30) mentions that “Modern studies of the social make-up of early Christian groups indicate that, especially in its urban settings, Christianity attracted a socially diverse membership representing a rough cross-section of Roman society, though lacking much representation from both its lowest and the highest strata. Since members of the upper classes were less numerous, high levels of literacy would have been somewhat rarer in Christian circles, but moderate levels of literacy were probably about the same in Christian groups as outside them.” For Gamble’s modern social description of early Christian communities, see 2004:30 n.10; 1995:248 n.12.

Antioch

171

wise, wealthy, and wellborn are around ten percent of about sixty. Nevertheless, it is reasonable to consider that around 90% of the members of an ancient community belonged to lower-status groups. Stegemann’s suggestion is noteworthy (1999:77): Estimates of the percentage of the upper stratum in the overall population depend on which total population figure one uses for the Roman Empire. If we also include the families, we can set the share of the overall population between 1 and 5 percent. In any case, it is clear that in ancient societies there was only a thin upper stratum.

In comparison with other communities or countries in ancient times, it is not unusual that over 90% of Christians belonged to a low-status group.124 In this regard, most Antiochene Christians might have been composed of lower-class Jews and Gentiles. Foakes-Jackson (1931:100) argues that: “it is by no means certain that the language of the majority of the Syrian Jews was Greek.” Most Antiochene Christians spoke Aramaic as their matrix language, whereas Greek was their embedded language. The linguistic milieu of the Syrians of Antioch reflects a clear cross-section of Antiochene Christians. Since many upper-status Syrians and Jews were bilinguals (§4.3.1, §4.3.2), many Gentile and Jewish Christians might have been bilinguals. Most upper-status Christians spoke Greek as their matrix language, whereas most lower-status Christians spoke Aramaic as their matrix language. It is of interest that Dibelius (1935:75) suggests that preachers who were bilinguals in Greek and Aramaic delivered the Gospel in Greek because “there were at that time men of all classes in Jerusalem, Antioch, and Damascus, who could speak both languages.”125 He proposes that the bilingualism of the Antiochene church created a bridge between the Palestinian Christianity and the Gentile Christianity (1949:25): Since, however, Jesus spoke Aramaic, though the tradition that has come down to us in the Gospels is framed entirely in Greek, the words of Jesus must have been translated. But since the earliest Christian communities on the language frontier in northern Syria – in Antioch, for example – contained many bilingual members, the translation will have been very easily effected through the simple process of repeating in the one language what had been heard in the other. 124 Gamble (1995:5-6) also writes that “In terms of social status, Christian communities had a pyramidal shape rather like that of society at large. … We must assume, then, that the large majority of Christians in the early centuries of the church was illiterate, not because they were unique but because they were in this respect typical.” 125 Although his view of the “bilingualism” of the Roman Near East is vague and exaggerated, it is noteworthy that he notices that there were many bilinguals in terms of gospel transmission.

172

Bilingualism of Jews in the First-Century Diaspora

Although Dibelius’ view remains problematic, 126 many Antiochene Christians were indeed bilinguals. Accordingly, it is safe to suggest that the Antiochene Christian community was bilingual. Many upper-status Gentile and Jewish Christians would have been primary or acquired bilinguals who spoke Greek as their matrix language. Some of them could certainly speak only Greek. Moreover, the Antiochene Jewish community by and large was composed of lower-status Jews who used Aramaic as their matrix language and many of them would have been acquired bilinguals in Greek and Aramaic.

4.4 Bilingualism of the Jewish Diaspora and Interdirectionality As investigated above, it can be said that the linguistic milieus of the Roman Near East, especially Alexandria and Antioch were bilingual. The bilingualism of the Jewish Diaspora suggests two possibilities related to the transmission of the Jesus tradition and the four gospels: (i) The Jesus tradition could have been circulated in Greek and in vernacular languages in the regions of the Roman Near East and (ii) the four gospels could have been circulated in Greek and in vernacular languages in the regions of the Roman Near East. The Jesus tradition and the four gospels very likely were transmitted from oral to written media, as well as from written into oral media, and from Aramaic into Greek as well as from Greek into vernacular languages in and out of Palestine. Furthermore, in parallel with the circulation of the Jesus traditions, after the canonical and non-canonical gospels had been documented in Greek or in a vernacular language, they were probably translated into other languages. This is adduced by the extant Coptic gospels (cf. introduction to §1.4; §2.3.4) and the extant Syriac versions (e.g. Sinaitic, Curetonian, Peshitta, and Harklean). It is of interest that, concerning the origin of the Western texts of the New Testament, many scholars, including Chase (1893:137-49; cf. Epp 1966:3), propose that Syriac texts lie behind the Western texts. This means that the Greek Western texts were written under the influence of Syriac texts which might have been translated from Greek texts. Consequently, the gospels in Greek were possibly translated into other language families, including the Aramaic 126 If the bilinguality of the Antiochene Christians made it possible to translate the Jesus tradition from Aramaic into Greek, it would be more convincing if the bilinguality of first-century Palestine already had allowed translation of the Jesus tradition. In other words, Jesus traditions in Aramaic and Greek were circulated in Galilee during Jesus’ ministry (see chapter 5).

Antioch

173

language family, and the translated versions were likely retranslated into Greek. That Greek versions survived does not deny that the Jesus and gospel traditions were translated and retranslated into other languages. Therefore, the bilingualism of the first three centuries in the Roman Near East makes it possible to consider that the transmission of the Jesus tradition and the gospels was not unilinear, teleological, and unidirectional, but hybrid, circular, and interdirectional.

5. The Bilingualism of the Earliest Christian Church in Jerusalem Bilingualism in first-century Palestine (chapter 3) and the Jewish Diaspora (chapter 4) leads to questions about bilingualism in the earliest Christian church in Jerusalem and, especially, to issues of bilingualism in regard to “the Hebrews and the Hellenists” and of “the seven.” As have been investigated, many Jews of first-century Palestine were bilinguals (chapter 3) and most Jews and Gentiles of the Roman Near East spoke Aramaic as their matrix language (chapter 4). If this is indeed the case, the Jerusalem church should be designated as a bilingual community1 and should be the subject of this chapter. Bilingualism in the Jerusalem community will be discussed in four parts. (1) The current discussion of the two particular designations in the Acts of the Apostles, with focus primarily on Hengel’s contribution, will be reviewed (§5.1). (2) Bilingualism of the Hebrews and the Hellenists, often considered to be monolinguals, will be suggested (§5.2). (3) The bilingual Seven who have been considered Greek-monolinguals will be reconsidered (§5.3). (4) Finally, the relationship between the bilingualism of the Jerusalem church and the interdirectionality of the Jesus tradition will be discussed (§5.4).

1

If we apply the definition to the linguistic situation (§2.1.2), it can be said that the linguistic milieu of the Jerusalem church can be situated between the two poles of a continuum. It ranges from a set made up of two unilingual groups of Aramaic and Greek each containing a small number of bilinguals in Aramaic and Greek, to a single group with a more or less large number of members using a second language for specific purposes. At one pole, most speakers in each group use only one language for all functions, whereas at the other, a varying number of speakers use both languages but for different purpose.

176

The Bilingualism of the Earliest Christian Church in Jerusalem

5.1 The State of Affairs Since Baur (1873:1.38-62; especially, 39-42),2 the identities of the ~Ebrai/oi and the ~Ellhnistai, (Acts 6:1) have provoked endless controversies. Subsequently, scholars have suggested various views of the designations, which include such views as: the linguistic distinction between Aramaic-speaking and Greek-speaking Jews, the geographical distinction between Palestinian and Diaspora Jews, the theological distinction between conservative and radical Jews, or the ethnic distinction between Jews and Gentiles.3 At present, it would appear that scholars have almost reached a consensus that the two terms stem from the linguistic-geographical distinction on the basis of an assumption that both of them were Jews. That is, the ~Ebrai/oi refer to Aramaic-speaking Palestinian Jews and the ~Ellhnistai, refer to Greek-speaking Diaspora Jews. There have been two approaches to the ~Ellhnistai,. The first is a bilingual approach and the other a monolingual approach. More than 250 years ago, Johann Bengel (in the original German edition in 1759) proposed that the ~Ellhnistai, refers to bilinguals in Greek and Hebrew (i.e. Aramaic). He (1859:2.564) insisted that the Hellenists “were Jews born outside of Palestine, to whom it seems the Greek tongue, besides the Hebrew, was vernacular.” In modern linguistic terms, this means that the Hellenists were primary bilinguals. Bengel (1859:2.525) also mentioned that “there is no doubt but that these Jews of all nations, who moreover were dwelling at Jerusalem, knew Hebrew” [emphasis in the original]. This means that the Hellenists acquired Hebrew (i.e. Aramaic) while living in Jerusalem. His suggestion that there were two kinds of bilingual Hellenists who used Greek and Hebrew (i.e. Aramaic) respectively, remained convincing for two reasons: (i) the Hellenists from all nations were primary bilinguals in Greek and Aramaic because Jews who spoke Aramaic were born in Greek-speaking regions. (ii) Some were acquired bilinguals acquiring Aramaic when they lived in Jerusalem or other Aramaic-speaking regions. Furthermore, in terms of the Seven, Bengel (1859:2.566) proposed that “the Seven were in part He-

2

3

According to Baur (1873:39), there were two groups in the Jerusalem church; one is the Hellenists whose leader was Stephen; the other is the Hebraists including the Apostles of Jesus. The former rejects Judaism and is in opposition to the existing Temple worship, whereas the latter more closely adheres to Temple worship. For the history of the debate, see Penner 2004:69-72; Hill 1992:5-17; Hengel 1983:1-11. Concerning the discussion over one hundred years ago, Roberts (1888:187-9 n.1) epitomized the history of research of the Hellenists and the Hebrews in detail.

The State of Affairs

177

brews, in part Hellenists.” It is regrettable that Bengel’s suggestions in this regard have not continued to be debated recently in academia. A number of scholars have given consideration to a monolingualgeographical view of the ~Ellhnistai,, as suggested by Moule (§5.3.3) and Hengel. Hill’s opinion is justified when he (1992:15) writes that “the most influential contemporary advocate of a view of the Hellenists consistent with that of F. C. Baur is Martin Hengel.”4 In light of the position of Hengel’s argument in the current debate, he serves as the primary counterpart below. Concerning the transmission of the Jesus tradition, the religionsgeschichtliche Schule, including Bultmann, devoted too much attention to the Hellenistic Christians of the Syrian Church (see introduction to chapter 4). On the contrary, Hengel (2001:6-37) has diverted scholars’ interest to the bilinguals of the Jerusalem Church.5 He argues that the Hellenists and the Seven function as pivotal parts in the formation of earliest Christianity in terms of their transmission of the Jesus tradition from Aramaic into Greek and Gentile mission. Acts 6 indicates that the early Christian church consisted mainly of the Hebrews and the Hellenists. Hengel suggests that the ~Ebrai/oi refer to Palestinian Jewish Christians who spoke Aramaic and some Greek, while the ~Ellhnistai, refer to Diaspora Jewish Christians who spoke only Greek, having returned from the Diaspora.6 The Hellenists’ monolingualism quickly led them to separate their community from worship with the Aramaic-speaking Jerusalem church (Hengel 1983:14-15; 4

5

6

Penner (2004:23) also notes that “the trajectory of scholarship on the Hellenists that begins with Baur culminates in Hengel’s work on Acts 6:1-8:1. Hengel not only provides the most systematic interpretation of the Hellenists in the tradition of Baur; he also fuels the debate over the reconstruction of early Christian origins.” To date, Hengel (2005:83-90) has persistently held his position. As a result, Penner (2004:25-6) properly points out that “through the Hellenists, Hengel is able to move the most significant developments of early Christianity into Jerusalem in the formative period of Christian history.” Hengel and Schwemer 1997:32-5; Hengel 1986:79, 1983:8-11. In fact, Hengel and Schwemer rejected the sharp distinction between Judaism and Hellenism. Hengel (1974:104) upholds the view that “a better differentiation could be made between the Greek-speaking Judaism of the Western Diaspora and the Aramaic/Hebrew speaking Judaism of Palestine and Babylonia.” He (1989a:7) also argues that “‘Hellenistic’ Jews and Jewish Christians are … those whose mother tongue was Greek, in contrast to the Jews in Palestine and in the Babylonian Diaspora who originally spoke Aramaic.” Hengel’s arguments are not persuasive in two points: (i) his definition of the “Western Diaspora” is not clear. Does, for instance, the Antiochene Christian community, one of the most important Christian communities, belong to the Western Diaspora? (ii) Diaspora Jews do not always refer to Greek-speaking Jews, as mentioned before (chapter 4).

178

The Bilingualism of the Earliest Christian Church in Jerusalem

1991:56, 68). 7 The formation of the Greek-speaking Jewish Christian community is noteworthy because the Hellenists made an essential contribution to the translation of the Jesus tradition from Aramaic to Greek. In other words, some parts of the Jesus tradition, including at least the passion narrative and the development of significant theological terms, 8 were most likely translated by the Hellenists. They may have had vivid memoirs of the activities, death and the resurrection of Jesus on which the Greek-speaking Christian community in the Jerusalem church was based (Hengel 1983:27). The Greek tradition of Jesus could have given impetus to the propagation of the message of Jesus to non-Jews by the earliest missionaries such as the Hellenists, including the Seven (Hengel 1983:26-7). On this account, Hengel (1983:29; cf. Hengel and Schwemer 1997:34) asserts that “we owe the real bridge between Jesus and Paul to those almost unknown Jewish-Christian ‘Hellenists’ of the group around Stephen and the first Greek-speaking community in Jerusalem which they founded; this was the first to translate the Jesus tradition into Greek.” But how was it possible for the Hellenists, who spoke only Greek, to translate the tradition of Jesus from Aramaic to Greek? Hengel’s argument appears to quarantine the Hellenists, influenced by Hellenistic syncretism, from having any active participation in the process of translating the earliest Jesus tradition from Aramaic to Greek (Hengel 1983:24). For Hengel, although the Hellenists were significant in the transmission process (1983:29), their monolingual handicap indicates that their roles were limited and passive in the formation of basic Christian terms and traditions (1983:27). 9 Their monolingualism makes it possible for Hengel to presuppose that the earliest Jesus tradition was

7

8

9

Hengel (1983:55) notes, “The separation of two groups in Jerusalem had become necessary because of the language of their liturgy.” This view is further developed by Esler (1987:135-159; esp. 143, 159). Hengel (1989a:18; 1983:26-8) enumerates the terms such as abba (transliterated), avgaph,, avpo,stoloj, evkklhsi,a, evlpi,j, euvagge,lion, koinwni,a, parousi,a, parrhsi,a, pisteu,ein, pi,stij, ca,rij, ca,risma, o` ui`o.j tou/ avnqrw,pou, h` basilei,a tou/ qeou/, etc. Hengel and Schwemer (1997:33-35) suggest that although the Hellenists had independent worship in Greek in the Jerusalem church, they were “more dependent on the formation of a fixed and firm tradition than the Hebrews...” because of their monolingual handicap [emphasis added]. Additionally, Hengel (1983:27) asserts that even Paul, who can be called a Hellenist, only receives (paralamba,nw) and delivers (paradi,dwmi) “confessional formulae” in 1 Cor 15:3-4 and Rom 1:3-4. In relation to this, Hengel considers that Paul’s mother tongue is Greek; see §3.2.5.

The State of Affairs

179

unidirectionally transmitted from the ~Ebrai/oi to the ~Ellhnistai, in the community in Jerusalem (1983:37; 1986:79).10 Hengel assigns the translation duty to bilinguals. The bilinguals played the role of translators-transmitters of the Jesus tradition between Aramaic-speaking Palestinian Jewish Christians (i.e. the ~Ebrai/oi) and Greek-speaking Diaspora Jewish Christians (i.e. the ~Ellhnistai,) in the Jerusalem church (1983:14). 11 At points he calls the bilinguals “jüdische ‘Graekopalästiner’” (1974:104-5; 1969 in German) and at a later point develops the designation “zweisprachige ‘Graekopalästiner’” (1983:11; 1975 in German).12 He provides examples of bilinguals13 such as: Mary and her son John Mark (Acts 12:12, 25; 13:5, 13; 15:37), SilasSilvanus (Acts 15:22, 27, 32, etc.; 1 Thess 1:1; 2 Cor 1:19, etc.), Judas Barsabbas (Acts 15:22, 32), Menahem (i.e. the younger contemporary of Herod Antipas; Acts 13:1), Matthew (Hengel 1974:105 n.359; cf. Jeremias 1959:270-4), and Peter (Hengel 1983:37, 162 n.39, 170 n.26). Hengel (1974:105) includes in this list “Jews who themselves came from the Diaspora but whose families were closely associated with Palestine and spent a great part of their life there.” These were bilinguals such as Jason of Cyrene and later the Levite Joseph Barnabas from Cyprus, the “cousin” of John Mark (Acts 4:36 etc; cf. Col 4:10), and the Pharisee Saul-Paul from Tarsus (Hengel 1974:105).14 In addition, he (1989a:17-8) adds more names to the list of the bilinguals later: Johanna (the wife of Chuza, evpi,tropoj of Herod Antipas, i.e. his steward, Lk 8:3), the tax farmers (e.g. the avrcitelw,nhj Zacchaeus in Jericho, Lk 19:2), and then men like Nicodemus (Jn 3:1) and Joseph of Arimathaea (Mt 27:57). Hengel (1983:11, 29) also believes that “these bilingual ‘Palestinian Greeks’ are of decisive significance for the development of early Christianity.” According to him, they take an active and important role in the translation of the Jesus tradition, thereby bridging the linguistic gap between the groups of the Hebrews and the Hellenists. This can be 10

11 12

13 14

It is regrettable that in terms of the directionality of the Jesus tradition, Hengel still holds to a linguistic unidirectionality hypothesis (§1.5.1), although he successfully dilutes the borderline between Judaism and Hellenism in his major publications. He (1983:14) presumes that “The Hellenists may have been converted by the preaching of bilingual disciples and conversation with them.” First, Hengel calls them “jüdische ‘Graekopalästiner’” in his Judentum und Hellenismus (in 1969:193; in English 1974:1-4-5); later, he takes up “zweisprachige ‘Graekopalästiner’” in his ‘Zwischen Jesus und Paulus,’ (ZTK 72 [1975] p .171). He mixes the two terms. However, the problem lies in that he does not define the terms clearly, as will be discussed (§5.2). The lists of bilinguals of the earliest Christian community in Jerusalem are as follows: 1974:104-5; 1983:10-11; 1989a:14-8. For Paul’s bilinguality, see §3.2.5.

180

The Bilingualism of the Earliest Christian Church in Jerusalem

called the Jerusalem church translation theory (§1.5.1). As mentioned earlier (§1.5.1, §3.3), however, the bilingualism of first-century Palestine could have led to the Galilee translation theory (i.e. the simultaneous translation theory) rather than the Jerusalem church translation theory. In his discussion of the “Seven,” Hengel goes even further in relating the far-reaching program of the linguistic transmission of the earliest Jesus tradition. His position (1983:12; 2001:28) is that “the ‘Seven’ were apparently all ‘Hellenists.’” For him, this means that the Seven could speak only Greek.15 Hengel provides three basic reasons for this conclusion: (i) They had Greek names (1991:68; 1983:12, 145 n.90); (ii) the resolution of the Hellenists’ complaint resulted in the appointment of the Seven (1983:13); and (iii) after the martyrdom of Stephen, the Hellenists were scattered and became active missionaries outside Judaea where the primary language was Greek (1983:13). These three points will be examined and subsequently challenged (§5.3). Many scholars have criticized Hengel’s arguments. However, those who disagree with him have also based their comments upon a monolingual approach. 16 Furthermore, it is interesting that when Hengel 15

16

As a matter of fact, it does not seem that Hengel is confident whether the Hellenists spoke only Greek. For, he (1974:105) mentions that “one might also point to the seven ‘Hellenists’ of Acts 6, though it is uncertain whether they knew Aramaic.” After two decades he opens the possibility that the Hellenists might be bilinguals; Hengel (1989a:18) adds the Seven to the list of bilinguals in Jerusalem, when he proposes that “not least, mention should of course be made here (the list of bilinguals) of the ‘Seven’ as the spokesmen of the Hellenist community (Acts 6:5)…” Furthermore, he considers that although Paul was a Hellenist, he was bilingual. This shows that Hengel’s concept of the linguistic competence of the Hellenists is obscure. Their debates will not be dealt with in this publication. For information on the debates, see the references in chapter 5 n.3. However, it is noteworthy that Craig Hill criticizes the traditional position raised by Baur and Hengel on the basis of the theological distinction in his Hellenists and Hebrews. (i) Aramaic-speaking Jews were also persecuted. (ii) Stephen’s speech does not criticize the Law and the Temple; his argument is also summarized in Hill 1992:462-9; 1996:129-153. Hill’s view is supported by some scholars including Hurtado (2003:207-14) and Witherington (1998:243-7), whereas others like Donald Hagner disagree with his view (1997:580-1). However, it is regrettable that Hill’s view is also based on a monolingual-geographical approach (Hill 1992:23-24). Concerning the existence of the Hellenists, Watson (1986:26) also argues that “indeed, it seems very doubtful if there ever were two groups in the Jerusalem church, the Hellenists and the Hebrews.” His view is followed by Räisänen (1992:149-58). Unfortunately, they are not interested in the bilingualism of the community as well. However, if the possibility of the bilingualism of first-century Jerusalem church is not considered, many complicated linguistic problems in narratives of the Acts will occur. Richard (1978:341) also assumes that “the author says very little concerning linguistic and theological factors. Stephen debated with diaspora Jews living in Jerusalem. From this it would be possible to consider him a Hellenist; however, Paul does the same and is not for that matter considered as such. Furthermore,

The State of Affairs

181

criticizes other scholars’ positions, he also proposes that they fail to consider the bilinguals who took a significant part in the transmission of the Jesus tradition.17 For nearly thirty years, Hengel has been discussing this issue from the perspective of bilingualism.18 More recently, he (2001:7) has emphasized that “we should not forget, especially in Jerusalem but also in Galilee or Syria and Babylonia, the bilingual Jews, who are so important for the beginnings of Christianity.” Despite these remarks, a fundamental problem remains in that his own work does not clearly or comprehensively engage with the phenomena of bilingualism. The fact that Hengel does not take the bilingualism of Diaspora Jews and Gentile Christians of the Roman Empire seriously is apparent, although he accepts the following three points; (i) the bilingualism of first-century Jerusalem (1983:10), (ii) the bilingualism of Jewish Graeco-Palestinians, including some of Jesus’ disciples in the Jerusalem church (1974:104-5; 1983:11; see chapter 5 n.13), and (iii) the bilingual “Jews who themselves came from the Diaspora but whose families were closely associated with Palestine and spent a great part of their life there” (1974:105). His suggestion that the Hellenists spoke only Greek and the Hebrews spoke Aramaic and some Greek seems to be an important starting point for the interpretation of the linguistic transmission of the Jesus tradition in earliest Christianity. At the same time, he misinterpreted the entire situation. The two terms (i.e. ~Ellhnistai, and ~Ebrai/oi) are used together only by Luke (in the case of the ~Ebrai/oi used by Paul, see chapter 8 n.106, §3.2.5). If Luke designates the Hebrews as those who speak Aramaic and some Greek, why do we not think that he uses the Hellenists as those who speak Greek and some Aramaic? The counterpart of the two-language speakers (i.e. the Hebrews) should be the two-language speakers (i.e. the Hellenists). Therefore, the bilingualism of the Hebrews and the Hellenists of the commu-

17

18

no thought is given in Acts 6:9-15 to language problems since the “Greek” opponents stir up the populace, the elders, and the scribes… and accuse Stephen before an Aramaic Sanhedrin.” On this account, Richard is dubious concerning its historicity. However, it would be more plausible that he opened the possibility of the bilingualism of first-century Palestine and Roman Near East. Criticizing Heitmüller’s (1912) and Kramer’s (1966) arguments of the Greekspeaking Jewish-Christian community and the Aramaic-speaking primitive community, Hengel (1983:32-7; 1995:381 n.52) criticizes that they fail to take account of bilingual Palestinian Greeks. See his Judentum und Hellenismus (1974:104-5; 1969 in German) and recently in his ‘Judaism and Hellenism Revisited’ (2001:6-37).

182

The Bilingualism of the Earliest Christian Church in Jerusalem

nity in Jerusalem (§5.2) as well as the bilingual Seven (§5.3) needs to be considered from the perspective of bilingualism theory.

5.2 Bilingualism of the Hebrews and Hellenists Hengel designates the ~Ebrai/oi as Aramaic-speaking Palestinian Jewish Christians and the ~Ellhnistai, as Greek-speaking Diaspora Jewish Christians; that is, he provides a monolingual, geographical, and ethnic period distinction. It seems that his geographical and ethnic designations are based on his monolingual approach. Hengel’s conception of bilinguals (§5.1) is not clear for five reasons.19 (i) He neither adequately defines “jüdischen ‘Graekopalätiner’” and “zweisprachige ‘Graekopalästiner’” nor explains the difference between them, although these two designations can be used differently to some degree. The relevance of this becomes clear when one asks whether Matthew or Peter is to be designated a bilingual Palestinian Greek. (ii) It is not certain that bilinguals belong to either the Hebrews or the Hellenists. Also not clear is whether the bilingual group should be considered as “third parties” (Hengel and Schwemer 1997:35). (iii) If the bilinguals belong to the Hellenists, how can “bilinguals like Paul from Tarsus and Joseph Barnabas from Cyprus” be distinguished from the Hellenists who spoke only Greek? (iv) If the bilinguals belong to the Hebrews, how can “Jews who themselves came from the Diaspora but whose families were closely associated with Palestine and spent a great part of their life there” (1974:105) be labeled as the Hebrews who spoke Aramaic as their matrix language? Some of them appear to be the Hellenists who spoke Greek as their matrix language. As discussed previously (§2.1.4), 19

As a matter of fact, many scholars do not articulate the concept of the bilinguals of the Jerusalem church. It is interesting, for instance, that although Johannes Weiss recognized that most of the Hellenists in Jerusalem were bilinguals and many of them were Aramaic-matrix speakers, he did not overcome Baur’s framework. He (1937:1.166) states, “The fact that these groups [the Hellenists] assembled in separate synagogues for worship indicates that the reading of Scripture and the prayers were conducted there in Greek. Many of them may also have still spoken their Aramaic mother-tongue, but we must assume that the permanent population of Jerusalem could also make itself understood by those who spoke only Greek. … Finally, unity with such Hellenists in the Christian church was possible only if the large majority of the members were to some extent at home in both languages.” In other words, those who spoke “Aramaic mother-tongue” refer to Aramaic-matrix speakers. If most of the Hellenists were bilinguals and many of them were Aramaic-matrix Jews, why did they separate themselves from the Aramaic-matrix Christians in the Jerusalem church?

Bilingualism of the Hebrews and Hellenists

183

if anyone learns a second language after the ages of 11-12, they cannot use the second language with the same adeptness as their primary language. This means that unless we have their personal language acquisition details, it is difficult to classify them as the Hebrews. (v) Can bilinguals like Peter from Bethsaida (an Aramaic-matrix bilingual) and Joseph Barnabas from Cyprus (a Greek-matrix bilingual) be categorized as belonging to the same group of “zweisprachigen ‘Graekopalästinern’”?20 Hengel seeks to introduce bilinguals as a bridge between two linguistic communities in the Jerusalem church; however, there is no place for bilinguals in his hypothesis because he (1991:56, 68; 1983:14) recognizes a strict dichotomy between the two groups on the basis of monolingual-geographical distinction. However, it is noteworthy that Marshall (1972:277-8) argues that “members of each group might know the other’s language and both linguistic groups were to be found in Palestine and in the Diaspora.”21 He (1972:280) convincingly insists that “In Jerusalem and elsewhere, the early Christian communities included bilingual members,” and that “this makes the thesis of a completely separate Aramaic-speaking church in Jerusalem all the less likely.”22 His suggestion leads us to

20

21

22

It has been discussed among some scholars that the monolingual approach to the Hebrews and the Hellenists is faced with the difficulty to which category Paul and Barnabas as bilinguals belong. For instance, in terms of Paul and Barnabas, Haenchen (1971:365-6) assumes that Paul and Barnabas belong to the Hellenists. However, Barrett (1994-1998:548) mentions that “though the witness of Acts is complicated by the fact that Luke seems to regard Paul and Barnabas as standing in succession from Stephen, Barnabas seems on the whole to belong to ‘the Hebrews’ rather than to ‘the Hellenists’.” Rather, the fact that the monolingualists have different views hints that the bilingual approach to them should be required. R. E. Brown (1983:74-79) argues that it is hard to make a simple distinction between Jewish Christianity and Gentile Christianity in the first century AD. Instead, he suggests four groups of Jewish Christians and Gentile converts depending on the observance of Jewish laws. However, it is unfortunate that he does not consider the bilinguality of the Christian communities in the first century AD. He (1972:280) notes that “it was once suggested by W. Bousset that the use of the term ‘Lord’ may have arisen in bilingual Christian circles; it was then possible to retain a Hellenistic origin for the phrase ‘Maranatha’ by ascribing it to the bilingual church in Antioch. The argument has generally been rejected, but the premise remains a sound one.” For the discussion of Maranatha, see chapter 5 n.92. However, as mentioned before (§4.1.2), it is noteworthy to consider Cadbury’s argument again. Cadbury regards the ~Ebrai/oi as Aramaic-speaking Diaspora Jews. Citing the inscriptions of “synagogue of Hebrews” found at Corinth, Rome, and Philadelphia in Lydia, he (Beg. 5.59-74; esp. 62 n.4) asserts that the “Jews of the dispersion, whether at Rome, Corinth, Tarsus, or even at Jerusalem, might distinguish from members of the older dispersion the more recent Jewish emigrants from Pales-

184

The Bilingualism of the Earliest Christian Church in Jerusalem

consider the bilingualism of the earliest Christian community in Jerusalem. If the bilingualism of Jerusalem and the Jewish Diaspora is taken seriously into account, the “bilingual distinction” of the two designations should be investigated for three reasons. (i) Above all, the bilingual distinction can better explain the situation of the Jerusalem church. Whereas, in a sense, Aramaic-speakers and Greek-speakers are associated with monolingual terminology, Aramaic-matrix speakers and Greek-matrix speakers are associated with bilingual terminology. Aramaic-matrix speakers refer to those who speak Aramaic as their matrix language and their linguistic competencies open the possibility that they can speak another language(s) as their embedded language(s) (cf. §2.1.9). (ii) The monolingual distinction does not cover all members of the Jerusalem church, especially bilinguals. However, the bilingual distinction offers a place for bilinguals. Aramaic-matrix Christians in the Jerusalem church mean those who spoke Aramaic as their matrix language and Greek as their embedded language. Greek-matrix Christians mean those who spoke Greek as their matrix language and Aramaic as their embedded language. (iii) As will be discussed, Luke seems to find it difficult to group the members of the Jerusalem church into Aramaic-speakers and Greek-speakers (monolingual distinction), Palestinian and Diaspora persons (geographical distinction) or Jews and Gentiles (ethnic distinction) in a clear-cut way. It is because the members of the earliest Christian community in Jerusalem were mixed linguistically, geographically, and ethnically. Accordingly, Luke seems to divide the church members into the two groups, the ~Ebrai/oi and the ~Ellhnistai, on the basis of the bilingual context of the Jerusalem church. On the basis of this bilingual context, it will be proposed that the distinction between the ~Ebrai/oi and the ~Ellhnistai, does not refer to a monolingual, geographical, or ethnic distinction but a bilingual distinction in Aramaic-matrix Christians (§5.2.1) and in Greek-matrix Christians (§5.2.2).

5.2.1 Hebrews: Aramaic-Matrix Christians Hengel considers the ~Ebrai/oi as Aramaic-speaking Palestinian Jews. However, bilingualism in the Jewish Diaspora and the Roman Near East in the first century suggests that the ~Ebrai/oi refer to Aramaicmatrix Christians in the earliest Christian church in Jerusalem without tine…” The former is Greek-speaking Diaspora Jews and the latter is Aramaicspeaking Diaspora Jews without regard to geographical distinction.

Bilingualism of the Hebrews and Hellenists

185

regard to geographical or ethnic distinction. As noted previously (chapter 4),23 most Diaspora Jews and Gentiles in the Jerusalem church who were from Aramaic-speaking areas could have spoken Aramaic as their matrix language.24 If this is the correct model, the Greek-speaking Diaspora Jews hypothesis, that is, that the Diaspora Jews who spread throughout the Roman Near East spoke only Greek, would have been the exception to the norm. And the Greek-speaking Gentile Christians hypothesis, that is, that Gentile Christians spoke Greek, should be reexamined because not all Gentiles do speak Greek as their matrix language. Accordingly, although some Christians are Gentiles or come from outside of Palestine, they should be called the ~Ebrai/oi because they spoke Aramaic. Hengel’s monolingual-geographical distinction has also been doubted by some scholars, most often in connection to Paul’s competence in Aramaic.25 The occurrences of the term ~Ebrai/oj in 2 Cor. 11:22 23

24

25

Four points are as follows: (i) bilingualism of the Roman Empire; (ii) successive immigration; (iii) periodic connection with Jerusalem; (iv) learning Aramaic in Jerusalem. Especially, in relation to “learning Aramaic in Jerusalem,” Hengel’s position is not consistent. Hengel (1983:11) suggests that it is less likely that one could expect Diaspora Jews returning to Jerusalem to have learned Aramaic, although he (1983:143 n.75; 1980:76) does not deny the possibility completely. However, Hengel (1974:105) also assumes that Jews who came from the Diaspora are to be regarded as bilingual Palestinian Greeks because their families were closely associated with Palestine, having lived for much of their life there. In light of such reasoning, it seems difficult at best to deny the very real possibility that they learned Aramaic while in Palestine. Indeed, as he later admits, some Jewish bilinguals who had been Diaspora Jews in Jerusalem may have relearned Aramaic. An awareness of this possibility arises almost unintentionally when he (1989a:18) writes: “[those] whose mother tongue was already Greek, even if they still understood Aramaic or had relearned it” [mphasis added]. On the one hand, Hengel denies that Hellenists ever relearned Aramaic while, on the other, he inadvertently allows for this possibility. In light of the strong interaction between Hellenism and Semitism in the Roman Near East (§4.1.4), it is likely that some Greek-matrix Jewish Christians in the Jerusalem church may have tried to learn or relearn Aramaic because the Jerusalem church was a bilingual community. Millar (1987a:144) puts forward that “various dialects of the Semitic language which we call ‘Aramaic’ were spoken all the way round the Fertile Crescent, from Babylonia to Arabia“ [emphasis in the original]. Also refer to Millar 1993:503-4; Taylor 2002:298-331. For Paul’s bilinguality, see §3.2.5. Michael (1928:142) mentions that “the latter (Hebrew) are Jews who, whether dwelling in Palestine or not, continue to use Hebrew (or Aramaic).” Beare (1959:107) notes that “a family which continued to speak Aramaic… even after long years of settlement in a Greek city; in some parts of Canada, Scottish families still pride themselves on preserving the Gaelic speech of their forebears who emigrated from the Western Highlands three or four generations ago.” Bruce (1971:240) suggests that “‘Hebrews’ … denotes Jews whose family ties were

186

The Bilingualism of the Earliest Christian Church in Jerusalem

and Phil. 3:5 point to Paul’s linguistic competence despite the fact that he is a Diaspora Jew. Scholars argue that the ~Ebrai/oi refer to Aramaicspeaking Jews both in Palestine and in the Diaspora without respect to geographical distinction. Marshall (1972:278) concludes that “even, then, if there was a linguistic difference between ‘Hebrews’ and ‘Hellenists’, this was not necessarily the same thing as a difference between Palestinian and Diaspora Jews or between Palestinian and Hellenistic Jews.” The fact that Paul, who is a bilingual, calls himself ~Ebrai/oj strongly indicates that the ~Ebrai/oi refer to a bilingual term deserves serious consideration. Accordingly, if Paul’s usage is the same as Luke’s, the ~Ebrai/oi of the Jerusalem church refer to Aramaic-matrix Christians both in Palestine and from the Diaspora without geographical distinction. The language lists of Josephus and Acts 2:9-11 support the view that the designations ~Ebrai/oi and ~Ellhnistai, refer to the bilingual distinction and not the monolingual-geographical-ethnic distinction. Josephus implies that Aramaic was the matrix language for some Diaspora Jews. In Bellum Judaicum he states in the introduction (1.3-6) that this was originally an Aramaic work translated into Greek at a later time. Josephus explains that it was intended for Diaspora Jews who lived among “the Parthians, and the Babylonians, and the remotest Arabians, and those of our nation beyond the Euphrates, with the Adiabeni.” Josephus implies that his Palestinian Aramaic work could be understood well by Diaspora Jews spread throughout these regions.26 Comparisons should be made with the Lucan linguistic list of Diaspora Jews who visited Jerusalem (Acts 2:9-11): “Parthians and Medes and Elamites and residents of Mesopotamia, Judea 27 and Cappadocia, 28

26

27

Palestinian, if they were not wholly resident in Palestine.” Bruce (1952:151) also proposes that “the word (~Ebrai,ouj) here stands in opposition to ~Ellhnistw/n, and apparently means Hebrew or Aramaic-speaking Jews, whether of Palestine or, like Paul (Phil.iii.5, ~Ebrai/oj evx ~Ebrai,wn), of the Dispersion” [emphasis added]. Fitzmyer (1998:347) also writes that it “is not so simple, because this distinction does not explain the Pauline use of Hebraios...[and Paul]...never calls himself Hellenistes, but rather Hebraios.” Richardson (1969:118 n.2) submits that “it is possible that ~Ebrai/oi could refer to Diaspora Jews speaking Hebrew as well as Greek.” Sevenster (1968:61) also mentions that “the first impression given by the scattered comments made by Josephus on the language problems is that Aramaic was the language most familiar to the Jewish people in the first century CE” In terms of the Lucan geographical horizon, there has been no consensus about why the Lukan list includes “Judea” in the list of the nations. However, it seems that the problem results from a different view of the distinction between dialect and language. There was not a clearer distinction between dialect and language in ancient times than in modern times (cf. introduction to chapter 2). In other words, Luke seems to distinguish “Judaean Aramaic” from “Galilean Aramaic” (cf. Janse

Bilingualism of the Hebrews and Hellenists

187

Pontus and Asia, Phrygia29 and Pamphylia, Egypt (see §4.2) and the parts of Libya belonging to Cyrene, and visitors from Rome, both Jews and proselytes, Cretans and Arabians” (italics added).30 It is noteworthy that most nations from Luke’s linguistic list are those of known bilingual regions.31 Janse (2002:349) observes that “what is interesting about the e;qnh ‘nations’ (Acts 2:5)...is that most of them are known to be bilingual in the first century AD, speaking either Greek or Aramaic as a second language (as opposed to their ‘own native language’).” In comparison with these two ethnic lists, it seems that most Diaspora Jews in Jerusalem who came from the overlapped countries (see italicized names) spoke Aramaic as their matrix language (cf. Millar 1993:503-4; Bruce 1965:61-2). Acts 2:8 explains that the reason they were bewildered and marveled lies in the fact that they heard the Galileans employ “our own native languages” (th/| ivdi,a| diale,ktw| h`mw/n evn h-| evgennh,qhmen). This means that the Diaspora Jews discerned their own native language and that they distinguished their languages from the Galilean Aramaic, which Diaspora Jews had expected to hear from the Galileans. Even though the apostles spoke Galilean Aramaic, most Diaspora Jews who used a language from the Aramaic family32 could have understood what the

28 29 30

31 32

2002:349). In fact, most Jews who came from Parthia, Media, Elam, Mesopotamia and Arabia also used Aramaic dialects that are sister languages of Galilean Aramaic. Luke seems to classify “Judean Aramaic” as a dialect of the Aramaic language family like Galilean, Syriac, Palmyrene or Nabatean. Accordingly, in comparison with Galilean Aramaic, Luke seems to add Judea to the list. Cappadocia was multilingual; see §4.1.1. Phrygians used Phrygian and Greek; see §4.1.1. Quoting a letter from Agrippa I to Caligula, Philo (Leg. ad Cai. 281-2) mentions that Jerusalem is the mother city not only for Judaean Jews, but also those of Egypt, Phoenicia, Syria, Coele-Syria, Pamphylia, Cilicia, Asia, Bithynia, Pontus, Europe, Thessaly, Boeotia, Macedonia, Aeolia, Attica, Argos, Corinth, the Peloponnese, the isles of Euboea, Cyprus, Crete, the lands beyond the Euphrates, Babylonia and its neighboring satrapies. For the history of Cretan Jewry in antiquity, see Van der Horst 2006:12-27 = 1990:148-65; for the history of Jews in Cyprus, see Van der Horst 2006:28-36. Janse (2002:357) also suggests that “in the easternmost parts of Asia Minor a number of non-indigenous languages coexisted with Greek.” As mentioned earlier (§2.1.11), Aramaic can be called an ancestor language. Galilean, Judean, Syriac, Nabatean and Palmyrene can be called daughter languages of Aramaic, and those languages can be called sister languages to each other. Taylor (2002:302) also suggests that the Late Aramaic dialects include Middle Aramaic dialects of the Nabataeans (Petra, Bostra, and Hatra), Edessa, Syriac, Mandaic, Aramaic of the Babylonian Talmud, Jewish Palestinian Aramaic, Christian Palestinian Aramaic and Samaritan Aramaic. It is interesting that Chrysostom (in 2 Tim; NPNF 13.490) mentions that Syriac “has an affinity with the Hebrew.”

188

The Bilingualism of the Earliest Christian Church in Jerusalem

disciples were speaking about to a great extent.33 Taylor tabulates morphological and lexical differences and agreements among Aramaic sister languages (see §2.1.11). Table 1 is helpful in the present discussion (2002:303):34 [Table 1] Differences and Agreements among Aramaic Sister Languages

yktb (=yiktçb?) lktb (=liktçb?) nktwb (nekôb)a

’nhnw ---’nhnn/hnn

SM demonst. Pronoun: ‘this’ dnh hdyn/Ȇdyn hn’

yktwb (=yiktôb)

’nh/’nn

hdn/hdyn

hm’

yktb (=yiktab)

’nn

hdn/’hn

hm’, Ȇm’, hz’

Aramaic Dia- S3M imperfect: ‘he will write’ lects Palmyrene Hatran Syriac Christian Palestinian Samaritan a

P1 pronoun: ‘we’

(Vocabulary): ‘he saw’ hz’ hz’ hz

Early Syriac inscriptions have the form yktb.

As Table 1 demonstrates, although the Aramaic family consists of diverse dialects both within and outside of Palestine, Aramaic language family speakers would have understood the Aramaic dialects of others.35 Clearly, most Diaspora Jews in Jerusalem who came from Aramaic-speaking countries spoke Aramaic as their matrix language just like Palestinian Jews. The fact that many Aramaic-matrix Diaspora Jews resided in first century Jerusalem (Acts 2:9-11) leads to the conclusion 33

34

35

Cf. chapter 4 n.101. Taylor 2002:302-3; Cotton 2005:162. Tcherikover (CPJ 1.5) states that “though differing from Jews in their religion, they (Syrians and Jews in Egypt, my insertion) had a common language with them (Aramaic). … No wonder, then, that the Egyptian population confused all peoples coming from ‘Syria’ and called them all ‘Syrians’; even the Hebrew language was sometimes mistaken for ‘Syrian’, i.e. Aramaic.” In terms of similarity between the Cursive Jewish Script and the Nabataean Script, Yardeni (2000:11) investigates non-literary Aramaic, Hebrew and Nabataean documents from the Judaean desert (in the caves of Wadi Murabba‘at and Nahal Hever and nearby) dating from the late Second Temple period to the end of the Bar Kokhba revolt in 135 AD and (2000:216-7) concludes that “the examination of the letter-forms in the Jewish cursive script and their comparison with the Nabataean script showed a considerable number of similar types.” In relation to this, it seems persuasive that Naveh (1982:112-4, 162-4, followed by Rahmani 1994:12 n.6) uses “Jewish script” as a local variant of Aramaic script developed in the Hellenistic age. Millar (1993:9) rightly suggests that “the use of the term ‘Semitic’ is harmless and unavoidable; the mutual resemblances in vocabulary and grammatical form, the identity of alphabet and the similarities of script of all the languages concerned are unmistakable.”

Bilingualism of the Hebrews and Hellenists

189

that there were many Aramaic-matrix Diaspora Christians in the earliest Christian church in Jerusalem. Furthermore, as will be discussed in detail (§5.2.2), most proselytes (ethnically, Gentiles; Acts 2:5, 11, 6:5) who came from Aramaic-speaking regions would have spoken Aramaic as their matrix language. In other words, most Gentiles who were from Aramaic-speaking regions should have spoken Aramaic as their matrix language, as mentioned before (chapter 4). That most Gentiles of the Roman Near East spoke Aramaic as their matrix language means that many Gentile Christians in the Jerusalem church and churches in the Roman Near East spoke Aramaic as their matrix language. Accordingly, Aramaic-matrix Diaspora Jewish Christians and Gentile Christians should be classified as the ~Ebrai/oi as well in spite of their name tags such as “Diaspora” or “Gentile.” It is evident that the distinction between the ~Ebrai/oi and the ~Ellhnistai, is not made through monolingual, geographical or ethnic distinctions but by bilingual distinction. The ~Ebrai/oi at Acts 6:1 refers to Aramaic-matrix Christians, that is, Aramaic-matrix Jewish and Gentile Christians, and Aramaic-matrix Palestinian and Diaspora Christians.

5.2.2 Hellenists: Greek-Matrix Christians As mentioned before, Hengel holds that the ~Ellhnistai, refer to Greekspeaking Diaspora Jews, that is, he makes a monolingual, geographical and ethnic distinction. However, when the bilingualism of the firstcentury Roman Near East is taken into account seriously, the ~Ellhnistai, (Acts 6:1) mean Greek-matrix Christians. The bilingual approach will be made by considering three points: textual, ethnic, and geographical arguments. When we deal with the referent of ~Ellhnistai, (Acts 6:1), special attention should be given to Luke because the ~Ellhnistai, occurs on only three occasions (Acts 6:1; 9:29; 11:20) in the New Testament literature. In the case of the ~Ellhnistai, (9:29), there is a general consensus that the ~Ellhnistai, in 9:29 refer to Greek-speaking Jewish non-Christians (Foakes-Jackson 1931:99; Simon 1958:14-5; Fitzmyer 1998:440). Furthermore, there is no textual problem because the ~Ellhnista,j has no variants except for Codex A (see chapter 5 n.36). The term ~Ellhnista,j on 11:20, on the other hand, is very complicated because some manuscripts like î74, ʠ2, A, and D* read {Ellhnaj.36

36

Codex A shows the same tendency to alter ~Ellhnista,j into {Ellhnaj at Acts 9:29.

190

The Bilingualism of the Earliest Christian Church in Jerusalem

Some scholars prefer to read {Ellhnaj rather than ~Ellhnista,j.37 They suggest three main reasons for this: (i) Luke makes an ethnic contrast between VIoudai,oij in v. 19 and {Ellhnaj in v. 20. (ii) The Lucan focus at 11:19-20 moves toward a Gentile mission. (iii) The strongest argument is the assumption that Luke uses the ~Ellhnistai, to designate Greekspeaking Jews, as the other two occurrences (6:1; 9:29) show. For these reasons, it is assumed that {Ellhnaj (i.e. Gentiles) is a better reading. Other scholars consider that the {Ellhnista,j is the proper reading.38 Two major reasons have been suggested. (i) {Ellhn occurs in Acts ten times. Whenever an ethnic contrast is made, {Ellhn is coupled with VIoudai/oj. 39 It is obvious that Luke makes an ethnic contrast between VIoudai/oj and {Ellhn in Acts. 40 Rather, in this sense, the reading ~Ellhnista,j is better because it is lectio difficilior. (ii) Although the ~Ellhnistai, (6:1, 9:29) refer to Greek-speaking Jews, ~Ellhnistai, (11:20) refer to Greek-speaking persons. For instance, Barrett (1994-1998:550) claims that “It is thus by no means impossible that the word should have a third meaning here: at 6.1, Greek-speaking Jewish Christians; at 9.29, Greek-speaking Jews; at 11.20, Greek-speaking Gentiles.” 41 Accordingly, there is a scholarly consensus that (i) ~Ellhnista,j (11:20) is improperly used and that (ii) ~Ellhnistai, at 6:1 and 9:29 refer to Greekspeaking Jews. However, it seems that Luke intentionally uses ~Ellhnisth,j three times in his narrative because the ~Ellhnistai, at 6:1 and 9:29 as well as at 11:19 refer to Greek-speaking persons. In other words, like the term ~Ebrai/oi, ~Ellhnistai, refer to Greek-speaking persons who were Jews and Gentiles without regard to the ethnic distinction. As will be discussed, Luke recognizes that the earliest church in Jerusalem includes Greek-speaking Gentiles from the start. He intentionally intimates the existence of Gentiles in the Jerusalem church from the beginning. This 37

38

39

40

41

Bengel 1859:2.611; Hort 1896:94; Warfield 1883:120-5; Rackham 1951:166; FoakesJackson 1931:99-100; Moule 1958-59:100; Haenchen 1971:365; Neil 1973:144; Conzelmann 1987:87; Johnson 1992:203; Dunn 1996:154; Fitzmyer 1998:476. Ropes Beg. 3.106; Cadbury Beg. 5.71; Williams 1964:142; Bruce 1952:151, 235; 1965:237; Martin ABD 136; Barrett 1994-1998:550-1; Metzger 1994:340-2; Witherington 1998:242. Acts 14:1; 18:4; 19:10, 17; and 20:21. All the occurrences in Acts are related to the dissemination of the gospel. Luke states that the gospel was preached to Jews and Greeks, that is, to all people. In this regard, Johnson (1992:105) argues that when Luke intends to make an ethnic distinction between Jews and Greeks, he employs VIoudai/oj and {Ellhn. Johnson prefers to read {Ellhnaj. Bruce (1952:151) also proposes: “here (vi. 1), Greek-speaking Jewish Christians; in ix. 29 probably Greek-speaking Jews in the synagogues; in xi. 20, probably Gentiles.”

Bilingualism of the Hebrews and Hellenists

191

argument leads us into the issue of the ethnicity of the members of the Jerusalem church at 2:5, 11 and 6:5. The term VIoudai/oi (Acts 2:5) has variants. The fact that VIoudai/oi was absent in the majority of texts may indicate that it was added later.42 Metzger (1994:251) points out that “Since Jews were already an e;qnoj, to say that these were from another e;qnoj is tantamount to a contradiction of terms.”43 If the VIoudai/oi is inserted later, it can be said that the residents of Jerusalem are composed of both Jews and proselytes from every nation under heaven, not only Jews. This view is strongly supported by internal evidence, the phrase VIoudai/oi, te kai. prosh,lutoi (Acts 2:11). 44 Johnson (1992:44; cf. Fitzmyer 1998:243) proposes that “this (both Jews and proselytes) is a summary rather than a separate ethnic entry.” This means that the residers in Jerusalem refer to Diaspora Jews and proselytes. On the other hand, Lake (Beg. 5.113-4) submits that “In vs. 10 ‘Jews and proselytes’ are treated as one of the component parts of the crowd. If so, obviously the rest of the crowd was not composed of Jews.”45 This means that most of the residents in Jerusalem are Gentiles and some Diaspora Jews and proselytes. Whether Johnson or Lake might be right, Luke considers some residents in Jerusalem as Gentiles because proselytes mean ethnically Gentiles. Luke (2:41, 4:4) reports that Peter’s sermon converted 3,000 and 5,000 persons to Christianity. The Christians should have included the Jews as well as the proselytes in Jerusalem. In this sense, Luke seems to have in mind that the earliest Christian church in Jerusalem includes Jews and Gentiles from the start. Furthermore, the possibility that Luke insinuates that there were some Gentiles in the Jerusalem church can be shown at Acts 6:5. It is of interest that the origin of Nicolas, one of the Seven (Acts 6:5), was a

42

43

44

45

Ropes (Beg. 3.12) suggests that the original text including ʠ omits VIoudai/oi, that the Western text holds VIoudai/oi, and that the Old Uncials inserts it again; also see Lake, Beg. 5.113-4; Williams 1964:64. He raises two more textual problems; (i) why should Luke think it necessary to mention that Jews were dwelling in Jerusalem? (ii) Why should it be said that they were devout men; would not this be taken for granted from the fact that they were Jews? Most scholars, including Hengel, consider that the distinction between the ~Ebrai/oi and the ~Ellhnistai, is linguistic. However, linguistically speaking, there is no difference between proselytes and Gentiles because the difference is religious. Lake (Beg. 5.113-4) assumes that “Probably VIoudai/oi in vs. 5 was originally a mistaken gloss on euvlabei/j.” He (114) argues that “The desire of the writer was … to show that from the beginning the Gentiles heard and the Jews refused the testimony of the Spirit.”

192

The Bilingualism of the Earliest Christian Church in Jerusalem

Gentile who came from Antioch.46 The introduction of the Seven resulted from the practical purpose to solve the problem of the alienation of Greek widows. The Seven served tables although they, of course, were leaders of the Jerusalem church. Why was an Antiochene Gentile chosen as one of the seven representatives by the church members? Why, in an ethnic sense, does the Jerusalem church need a Gentile leader to serve tables? The fact that a Gentile leader was selected implies that there were some Gentile Christians in the earliest Christian community in Jerusalem.47 In this sense, some scholars consider that the ~Ellhnistai, (6:1) refer to proselytes 48 or Gentiles. 49 Others, including, Hengel rebut the ethnic distinction and propose that the ~Ellhnistai, (6:1) refer to only Jews, especially Diaspora Jews. The most significant reason is that Luke intends to unfold his story toward the Gentile mission step by step. In this respect, they consider that the appearance of Gentiles in Acts 6:1 is a gross anachronism, let alone in 2:5 (Witherington 1998:241). Fitzmyer (1998:239) mentions that “the Lucan story disregards any others in Jerusalem who might not be Jews.” Johnson (1992:105) also asserts that “Luke takes such pains to show the gradual development of the Gentile mission after the close of the Jerusalem narrative” [emphasis added].50 Interestingly, although many of the scholars mentioned above consider that the residents in Jerusalem (2:5) were ethnically composed of 46

47

48 49

50

Only the origin of Nicolas is mentioned in the present verse. This implies that the other six were born Jews; see Haenchen 1971:264; Barrett 1994-1998:315. Even Kraeling (1932:147) considers Nicolas as a Greek. However, there is no information that he is a Greek. Some scholars, in the same vein, consider that the Seven were composed of the ~Ebrai/oi and the ~Ellhnistai,, lest they disregard the Hebrew widows; for detail discussion, see §5.3.3. Blackman 1936-7:524-5; Grundmann 1939:45-73; Reicke 1957:116-7, 121; Schwartz 1963:146-7; Schmithals 1965:34 n.71; Bauer 1967:107-8. Warfield 1883:113-127; Wetter 1922:411-2, 404-5; Cadbury Beg. 5.68; Windisch, TDNT 2.512. As a consequence, Cadbury (Beg. 5.65, 71-2) assumes that the ~Ellhnista,j of 11:20 refer to Greek-speaking Gentiles because he supposes that the ~Ellhnisth,j is synonymous with {Ellhn ,such as the relationship between the two Greek spellings of Jerusalem. Barrett (1994-1998:309) also proposes that “If from 6.1 (or from the day of Pentecost…) there had been Jewish and Gentile elements in the Jerusalem church, or if Luke had believed that it was so, could he have made so much of Peter’s preaching to Cornelius, of the founding of the mixed church in Antioch, of Paul’s break with the synagogue at Pisidian Antioch and of the Council of Ch. 15?” Haenchen (1971:260) mentions that “these (proselytes) are no pagans, for the Gentile mission did not (according to Luke) begin until Peter baptized Cornelius (Chapter 10).” Also see Bruce 1952:83; 1965:61; Simon 1958:15; Moule 1958-9:100-2; Güting 1975:149-69.

Bilingualism of the Hebrews and Hellenists

193

Jews and Gentiles, they suggest that the ~Ellhnistai, (6:1) refer to only Diaspora Jews.51 Given this fact, where are the Christian Gentiles prior to the time at which Peter appears for the Gentile missions? In reply to this, two points mentioned above can be suggested: (i) Luke recognizes that the Jerusalem church includes Gentiles from the start. (ii) Luke intends to unfold his story in the direction of the Gentile mission. Barrett (1994-1998:118) persuasively concludes that Luke regards the Jerusalem church as a universal church including Gentiles from the start: “from the beginning the Christian church was an inspired community and a universal community. It therefore included both Jews and ‘pious men’ of every kind.” It seems that Luke depicts his narrative about the Gentile mission not in a unilinear but in a spiral way. The literary style is supported by the fact that he repeatedly introduces the Gentile mission.52 As a consequence, it seems that Luke intimates to his readers that the Jerusalem church includes Gentile Christians from the start, as Acts 2:5, 11 and 6:5 imply. The use of the terms ~Ebrai/oi and ~Ellhnistai, (i.e. Aramaic-matrix speakers and Greek-matrix speakers, including both Jews and Gentiles) seems to be placed in the medial stage between VIoudai/oj (i.e. Jews) and {Ellhn (i.e. Gentiles). Concerning the expansion of the gospel, Luke intends to dilute the ethnic border line between Jews and Gentiles by means of using the linguistic distinction as a bridge. Accordingly, it can be concluded that (i) Luke uses ~Ellhnistai, for Greek-speaking persons at Acts 6:1, 9:29 and 11:20 without regard to ethnic distinction (cf. Barrett 1994-1998:550; Metzger 1994:342; Witherington 1998:242), and (ii) the reading {Ellhnaj (11:20) does not seem appropriate because the ethnic distinction that the other two occurrences (6:5, 9:29) refer to only Jews is not persuasive. Here, it is noteworthy to mention Craig Hill’s argument. He insists that those who were persecuted were both the Hebrews and the Hellenists.53 If those who were scattered were composed of Greek-matrix as well as Aramaic-matrix Christians, then the fact that those who were scattered travelled and preached the word only to Jews (mo,non VIou-

51 52

53

Bruce 1965:60, 128; Haenchen 1971:171, 260; Marshall 1980:70, 125-6; cf. Fitzmyer 1998:239, 347; Schnabel 2004:654-5. See Hengel’s view (§5.1). Cadbury (Beg. 5.66-8) considers that Luke repeats the beginnings of Gentile Christianity several times. As such repetition, he cites the missionary journey of Paul and Barnabas (chapters 13- 14), the first conversion of Gentiles (11:19-20), the story of Cornelius (chapters 10 – 11), Philip and the Ethiopian (chapter 8), and the story of Pentecost (chapter 2). In relation to the Lucan focus on Gentile mission, some scholars suggest that Acts is a response concerning the identity issue of Gentile Christianity; see Jervell 1972:41-69; Franklin 1975:116-44; Maddox 1982:183-6. See chapter 5 n.16; also see Witherington 1998:243-7.

194

The Bilingualism of the Earliest Christian Church in Jerusalem

dai,oij) in Phoenicia, Cyprus and Antioch is more understandable. It seems that the term VIoudai,oij at 11:19 means an ethnic distinction (Johnson 1992:105). On the other hand, Luke seems to draw a linguistic distinction by using ~Ellhnistai, at 11:20. “Some of them” who were from Cyprus and Cyrene preached even to ~Ellhnista,j, that is, Greekmatrix speakers. Did the people referred to as “some of them” intentionally preach to only Greek-speaking Gentiles? Rather, it seems that “some of them” preached the word to Greek-matrix speakers, Jews as well as Gentiles in Antioch. The implied audience entails two ethnic groups, Jews and Gentiles. There are two main reasons: (i) the population of the Antiochene Jews may be between 30,000 and 50,000 out of 300,000, the total population of Antioch (Hengel 1997:186, 196).54 It is unlikely that the Christians intentionally avoided preaching the word to Antiochene Jews. (ii) Not all Antiochene Gentiles spoke Greek as their matrix language. Rather, most Antiochene Gentiles in the first century spoke Aramaic as their matrix language. 55 “Some of them” from Cyprus and Cyrene preached to ethnically mixed Greek-matrix inhabitants, both Greek-matrix Jews and Gentiles. Metzger (1994:342) convincingly suggests that “the word (~Ellhnista,j) is to be understood in the broad sense of ‘Greek-speaking persons,’ meaning thereby the mixed population of Antioch in contrast to the VIoudai/oi of ver. 19.” In this sense, it seems that Luke intentionally chooses to use the linguistic term ~Ellhnista,j rather than the ethnic term {Ellhnaj at 11:20. Luke finds it difficult to group the inhabitants of Antioch, to whom the Christians preached, into Gentiles on the basis of the ethnic distinction like the case of the ~Ebrai/oi. Accordingly, Luke seems to select the linguistic term (~Ellhnista,j) rather than the ethnic term ({Ellhnaj). Hengel’s position of the geographical distinction in relation to the ~Ellhnisth,j will now be challenged. It seems that Hengel’s position on the geographical distinction does not seem persuasive. Some scholars, including C. S. Mann,56 J. P. Meier,57 L. T. Johnson,58 and Barrett59 throw

54 55

56 57

For a detail discussion, see §4.3.1. It is interesting that Pierson Parker (1964:168) suggests that “~Ellhnista,j is correct. Luke probably meant to indicate that some Jews at Antioch spoke Greek, others Aramaic-which was the case-and that the wandering missionaries addressed themselves to both of the Jewish groups in that city.” Mann (1967:301) mentions that “many Jews in Palestine, unless living far from cities, would be more or less at home in the vernacular Greek of the time (koine).” Meier (1991:267) asserts that “Probably some of these Diaspora Jews were bilingual, but others (e.g. recent émigrés from the Diaspora) may have spoken nothing but Greek.

Bilingualism of the Hebrews and Hellenists

195

doubt upon the strict geographical distinction between Aramaicspeaking Palestinian Jews and Greek-speaking Diaspora Jews. They have proposed the possibility that some Hellenists were Greek-speaking Diaspora Jews and others Greek-speaking Palestinian Jews. Recently, Schnabel criticizes Baur and Hengel’s view that there was an early schism in the Jerusalem church. He (2004:654; cf. §4.1.4) rightly suggests that the Hellenists who resided in Jerusalem spoke Aramaic as well because “They hardly would have settled in the Jewish capital if they were unable or unwilling to communicate in Aramaic.” For this reason, he concludes that “we must remember that presumably the majority of Palestinian Jewish Christians were bilingual, speaking both Aramaic and Greek.”60 Some Palestinian Jews would have used Greek as their matrix language and Aramaic as their embedded language from the perspective of regional and personal bilingualism. As for the regional bilingualism of first-century Palestine (§3.1), it can be summarized in three points: (i) In relation to language choice (§2.3.2) and inscriptions (§3.1.1), the statistical data about the inscriptional languages from first-century Palestine leads to the conclusion that some Palestinian Jews used Greek as their matrix language, at least to the extent that they chose Greek as the appropriate language for the linguistic domain of sublime inscriptional language. (ii) Greek papyri excavated from Palestine such as the Babatha archive, Waddis Murabaat and Seiyal, Nahal Hever, and DSS suggest that the linguistic milieu of first-century Palestine is largely bilingual (§3.1.2). The papyrological evidence also persuasively supports the assertion that some Palestinian Jews could speak Greek as their matrix language. (iii) Population geographical evidence indicates that Jews and Gentiles mixed in many cities throughout the Levant (§3.1.3). Living in a border area or among different language groups would have brought about bilingualism (§3.2.4). These regional bilingualism factors lead to the causes of the personal bilingualism such as bilingual parents and region (§3.2.4), formal education (§3.2.5) and occupation (§3.2.6). Personal bilingualism factors of first-century Pales58

59 60

Johnson (1992:105) also states that “hellenistes refers to a Jew who predominantly speaks Greek. Some of them were probably from the Diaspora (see Acts 2:7-12; 4:36; 6:9), although Greek was widely spoken in Palestine as well” [emphasis added]. Barrett (1994-1998:315) suggests that “Luke’s usage of the word Hellenist must not be taken to mean that they were necessarily Diaspora Jews.” Frommel (1908:18) also considers the Jerusalem church as a bilingual community from the beginning “so that in public assemblies the Words of Jesus may have been communicated in Greek as well as in Aramaic.” However, he assumes that an Aramaic Gospel was an oral gospel on the basis of unidirectionality (§1.3).

196

The Bilingualism of the Earliest Christian Church in Jerusalem

tine strongly support the idea that some Palestinian Jews in early Christianity used Greek as their matrix language. Therefore, it can be concluded that Greek-matrix Palestinian Jews and Greek-matrix Palestinian Jewish Christians should be classified as ~Ellhnistai,. The ~Ellhnistai, refer to Greek-matrix persons, whether they were Jews or not. In summary, Hengel suggests that the ~Ebrai/oi refer to Aramaicspeaking Palestinian Jewish Christians and the ~Ellhnistai, Greekspeaking Diaspora Jewish Christians in the Jerusalem church on the basis of a monolingual, geographical, and ethnic distinction. This suggestion then becomes the starting point for a far-reaching theological program about how the Jesus tradition was transmitted. Contrary to Hengel’s opinion, an evaluation of bilingualism within the Roman Near East and Judaeo-Palestine in the first century results in a much more complex picture than what he conceives. The most significant aspects of this linguistic view may be summarized. (i) It is clear that Luke’s designations, the ~Ebrai/oi and the ~Ellhnistai, , are based not upon monolingual-geographical-ethnic but bilingual distinctions. Both groups could have included Jews and Gentiles and residents in and out of JudaeoPalestine. (ii) The bilingualism of first-century Jews in both Palestine and the Diaspora implies that the Jerusalem church was a bilingual community and that many bilingual Christians were in the Jerusalem church. (iii) The ~Ebrai/oi were Aramaic-matrix Christians (both Jews and Gentiles) who were in Palestine or came from the Diaspora. Aramaic-matrix Christians (i.e. dominant bilinguals in Aramaic) used Aramaic as their matrix language and Greek as their embedded language. One can easily support the notion that early or primary bilinguals spoke Greek as much as they would have spoken Aramaic. On the contrary, late or acquired bilinguals spoke Greek less than they spoke Aramaic and to varying degrees. Aramaic-matrix Christian semibilinguals used Greek for special purposes. (iv) The ~Ellhnistai, were Greek-matrix Christians (both Jews and Gentiles) who were in Palestine or came from the Diaspora. Greek-matrix Christians (i.e. dominant bilinguals in Greek) used Greek as their matrix language and Aramaic as their embedded language. For instance, early or primary bilinguals spoke Aramaic just as much as Greek. On the contrary, late or acquired bilinguals spoke Aramaic less than they spoke Greek and to varying degrees. Greek-matrix Christian semibilinguals used Aramaic for special purposes. (v) Many monolinguals may have tried to acquire Aramaic or Greek as their secondary language in the bilingual Jerusalem church. (vi) The bilingualism of the first-century Jerusalem church indicates that the Jesus tradition “coexisted” in Aramaic as well as in Greek in the Jerusalem church.

The Bilingual Seven

197

5.3 The Bilingual Seven Hengel assumes that the Seven mentioned in Acts were Greekspeaking Hellenists based upon three points, as mentioned before (§5.1):61 (i) The Seven have Greek names; (ii) the resolution of the neglected Hellenists is establishment of the Greek-speaking Seven and (iii) the Hellenists, including Philip, one of the Seven, preached outside of Palestine, where the primary languages were not Aramaic. Consequently, the linguistic limitation of the Hellenists resulted in an early schism between the Hebrews and the Hellenists, and further, between Jewish Christianity and Hellenistic Christianity (for a detailed discussion regarding the difference between Petrine Christianity and Pauline Christianity due to the early schism, see Goulder 1994). However, Hengel’s arguments are not persuasive from the perspective of the bilingualism of first-century Jerusalem and Roman Near East. It should be considered that the Seven, as leaders of the whole bilingual church in Jerusalem, would themselves be bilingual Christians. Either Jews or Gentiles could have spoken Greek, Aramaic, or another language as their matrix language and also spoken one more language as their embedded language. Hengel’s three points will be reexamined in the next chapter .

5.3.1 Onomastica One of Hengel’s primary considerations is the issue of Greek names. He assumes that their Greek names indicate that the Seven were the Hellenists. However, Fitzmyer (1998:350; cf. Barrett 1994-1998:314) has revealed that even though all Seven have good Greek names, this does not show whether they should be located among the Hellenists or the Hebrews, as many Jews of that period bore Greek names.62 Conversely, 61

62

Some scholars (for instance, Penner 2004:69) question whether the Seven belong to the Hellenistic group. It is not persuasive when some scholars (for instance, Spiro 1967:285-300; Scharlemann 1968:17-22) propose that Stephen was a Samaritan. On the other hand, it is interesting that Cullmann (1955) tries to make a strong connection among ~Ellhnistai,, Essenes and the fourth gospel, and that one of the ~Ellhnistai,, Philip, was a missionary to Samaria. However, this argument also is not persuasive. Furthermore, it is appropriate that Richard Horsley (1996:159) mentions that, without regard to their identity, people in a bilingual situation “may adopt a name or a second name in the politically dominant or more prestigious language.” Bruce (1979:50) also considers that their Greek names do not show their identities because some of Jesus’ disciples (e.g. Andrew and Philip) had good Greek names. However,

198

The Bilingualism of the Earliest Christian Church in Jerusalem

there were a number of examples where residents with Greek names in the Near East gave Semitic names to their children (Butcher 2003:284).63 A Greek name itself should not be considered indicative of someone’s language.64

5.3.2 Geographical Evidence Hengel makes another assumption that should not go unchallenged. He (1991:56) is convinced that the Hellenists from the Greek-speaking Diaspora would not have understood Hebrew/Aramaic worship in the synagogues of Jerusalem. However, as discussed before (chapter 4; §5.2), the monolingual-geographical distinction of “Greek-speaking Diaspora Jews” is problematic because the existence of Aramaic-matrix Diaspora Jews and Gentiles simply cannot be denied. The linguistic milieus of the four regions where the episode of the Seven is said to have occurred need further consideration. Philip appears in three locations: (i) Caesarea Maritima, (ii) Samaria, and (iii) Ethiopia.65 Furthermore, Nicolas’ origin is said to be (iv) Antioch. The linguistic environ-

63

64

65

it is regrettable that Bruce holds the traditional view: Following the monolingual designation of the Hellenists, Moule and Hengel insist on a linguistic distinction (cf. §5.3.3) and Bruce (1979:50) observes that the Seven “were probably leaders of the Hellenistic group in the church. This group was the foremost in propagating the Christian message throughout Judaea and the neighbouring regions; it eventually launched the Gentile mission, and in particular was responsible for founding the church of Syrian Antioch.” For my criticism of Moule and Hengel, see §5.3.3. He (2003:284) convincingly states that “it is often unclear whether the use of a Greek personal name means that the individual was a Hellenized Greek speaker, or that a Semitic personal name signifies someone who was less Hellenized.” Bauckham (2006 chapter 4) proposes that the relative frequency of names in the Gospels corresponds to that of the names of Palestinian Jews listed by Tal Ilan (LJNLA). This correspondence shows the Palestinian origins of names in gospel traditions. When he (2006:70; cf. 289) counts the Jewish names, he excludes the names of the Seven because they were born in the Diaspora but lived in Palestine on a monolingual basis. He seems to make a distinction between Aramaic-speaking Palestinian Jews and Greek-speaking Diaspora Jews. As a result, he (2006:86) submits that “these features of the New Testament data would be difficult to explain as the result of random invention of names within Palestinian Jewish Christianity and impossible to explain as the result of such invention outside Jewish Palestine.” However, names themselves do not always indicate ethnicity or languages because the linguistic milieu of first-century Palestine and the Roman Near East was bilingualism and the earliest Christian community in Jerusalem was a bilingual community. Philip preached in Azotus (Acts 8:40). However, this will not be dealt here for the linguistic situation of Azotus is similar to that of Caesarea. For more information, see Hengel 1983:110-5.

The Bilingual Seven

199

ment of these places is significant because they may well indicate the linguistic competence of the evangelists.

5.3.2.1 Caesarea Maritima It has been suggested that after baptizing the Ethiopian eunuch, Philip is depicted as going to Caesarea (Acts 8:40), where he remains until Paul visits (Acts 21:8). The linguistic milieu of Caesarea could understandably lead to the view that Philip spoke Greek because the city was well-known as a Graeco-Roman city. This is expressed by Downey’s remark (1975:23): If, as seems likely, Philip was a Greek-speaking Jew, it would be natural for him to settle in the provincial capital and cosmopolitan seaport; and this apparently is why we next find him in Caesarea about twenty years later...Philip…represents the first continuous leadership in the Christian community at Caesarea.

However, the linguistic milieu of Caesarea seems to be different from what has been considered. Caesarea’s population in the mid-first century, according to Josephus (Bell. 2.266-8), consisted mainly of Jews, Syrians, and Roman soldiers.66 Caesarea was likely a bilingual city with residents speaking both Aramaic and Greek. This is supported by the many bilingual inscriptions excavated there. 67 Later, the Jerusalem Talmud (Sota. 7.1.21b) has a revealing comment attributed to Rabbi Levi that he heard some Jews recite Shema in Greek and was surprised not to hear Hebrew/Aramaic. This indicates that there were some Greek-speaking Jews in the Jewish community in Caesarea. The linguistic milieu of Caesarea in the mid-first century seems to be similar to that of other Hellenistic cities of the Roman Near East such as Antioch or Alexandria (cf. chapter 4). Josephus (Ant. 20.175-7; Bell. 2.268, 2.287) suggests that there were more rich Jewish residents than Syrians. Upper-status Jews were primary or acquired bilinguals 68 and acquired bilinguals spoke Greek for business but Aramaic at home. Most lowerstatus Jews may have spoken Aramaic. Most of them could have been bilinguals due to geographical or occupational bilingualism. All in all, 66

67 68

Cf. Levine 1975:16, 22; Goodman 1987:64. Josephus mentions that although the Jewish population is at least 20,000 (Bell 2.457), it was for the greatest part inhabited by Greeks (Bell 3.409). Baslez 2007; Sartre 2005:292-3; Van der Horst 2001:161; Levine 1998:160-1; 1975:15-33; Millar 1993:377-8; Hengel 1983:110-115; Spolsky 1983:99; Lieberman 1965:29-67. Synagogue inscriptions from Caesarea (Roth-Gerson 1987:111-24) imply that all donors of the inscriptions could be Greek-speaking Jews, as their Greek names show.

200

The Bilingualism of the Earliest Christian Church in Jerusalem

most Caesarean Jews can be called Aramaic-matrix speakers. Taking the linguistic environment of Caesarean Jews into consideration, the episode in Acts about Philip does not suggest that he spoke only Greek. Moreover, there is no evidence that he had stayed and worked continuously as a Christian leader in Caesarea until Paul visited him in 58. (i) Cornelius became a Gentile Christian in Caesarea. Before his conversion, there is no mention that Philip preached to Gentiles or Jews and baptized them when he had worked at other places (Acts 8:12, 38). Although Cornelius was a pious God-fearer, he never heard about the gospel of Jesus before Peter started teaching about the total life of Jesus, including death and resurrection (Acts 10:34-43), as Levine (1975:24-5) points out. (ii) Philip did not appear in relation to Cornelius’s conversion. Because Philip might have not been in Caesarea, it seems that Cornelius called Peter in Joppa (Acts10:5). Krentz (1992:262) argues that while Peter preached the gospel and baptized him, “Philip is not mentioned; he does not function as leader of the church at Caesarea.”69 As a result, it is uncertain that he stayed continuously between his arrival in Caesarea (8:40) and Paul’s visit in 58 (21:8). Accordingly, his residence in Caesarea should not be taken as proof that Philip was a monolingual Greek, but rather that he could easily have been like other Jews in Caesarea: bilingual.

5.3.2.2 Samaria Philip preached the Christian message, baptized, and performed great miracles in the city of Samaria (Acts 8:5-13).70 When Philip along with the two apostles Peter and John returned to Jerusalem they preached the gospel to many Samaritan villages (Acts 8:25).71 Philip preached his message both in the city as well as in villages. The linguistic situation of Samaria was likely similar to that of Judean bilingualism (i.e. Samaritan Aramaic and Greek) (Hengel 1989a:8). Most inhabitants of rural regions may have spoken Aramaic as their matrix language, although some of them spoke Greek as their embedded language. Many of the inhabitants of urban regions could have been bilinguals. As seen in the exam-

69

70 71

Krentz (1992:262 n.13) also argues that “there is no evidence that Philip spent the intervening years there, or that any ‘evangelist’ ever served as a resident leader of a local Christian community.” Scholars assume that the city may be Sebaste or Shechem; see Haenchen 1971:301-2; for information of Sebaste, refer to Bell. 1.403; Strabo, Geog. 16.2.34. Marshall (1980:160) persuasively points out that “the subject of the verse is vaguely expressed, but no doubt includes Philip.”

The Bilingual Seven

201

ple of Caesarea Maritima, the linguistic situation of Samaria does not prove Philip spoke only Greek. Rather, one could just as easily infer that Philip preached the gospel bilingually in the city and in the countryside.

5.3.2.3 Ethiopia Acts 8:26-39 tells of the episode of Philip teaching an Ethiopian eunuch, who is said to be a minister of Candace72 and is presently on his way back home after worshiping in Jerusalem. This eunuch’s ethnicity and language background merits attention. The language spoken by the eunuch may indicate Philip’s linguistic competence. As to his ethnicity, Rénan (1869:145, 145 n.4) suggests that the eunuch was a Falasha.73 Also, Fitzmyer (1998:410) argues that the eunuch “is to be understood as a Jew, or possibly a Jewish proselyte, who comes from a distant land.” However, many scholars have convincingly suggested that the eunuch was a Gentile from Meroe.74 Furthermore, it seems reasonable when Haenchen (1971:314) notes that “Philip would have forestalled Peter, the legitimate founder of the Gentile mission!”75 Although the reference to “Ethiopia” is vague,76 for Luke, Aivqi,oy is likely Meroe, the capital of the Nubian kingdom in what is today part of Sudan. Unfortunately, little is known of the people and language of first-century Meroe, as O’Connor (1993:72) states that “all scholars agree that a much more representative range of sites needs to be exca72

73

74

75 76

As to Candace, Meinardus (1970:369) mentions that “Candace is not the name of a queen, but the royal title of several queens who ruled over Ethiopia just before and after the Christian era.” For more information, see Budge 1928:62, 111-3. The Amharic term, “Falashas,” refers to the Jews of Ethiopia. It is interesting that the term, “Cush” or “Cushites” occurs approximately fifty times in the Old Testament and most of them are translated into “Ethiopia.” The Jews of Ethiopia were well known to people of the Old Testament; for more information, see Spencer 1992:14950. While Falasha is a term used by outsiders, they call themselves Beta Israel. For this reason, Kaplan (1992:8-9) may prefer to call them Beta Israel. It has been suggested that the Falashas emigrated from the north (e.g. Jewish garrison at Elephantine) or via South Arabia (Ullendorff 1968:15-25). Eusebius, HE 2.1.13; Munck 1967:78; Haenchen 1971:314-5; Schneider 1982:1.498; Tannehill 1986:2.110; Conzelmann 1987:67; Bruce 1965:186, 1989:379; Spencer 1992:129, 146-8; Barrett 1994-1998:425-6; Witherington 1998:297. The Seven performed the missionary work that Jesus commanded. For a detailed discussion, see §5.3.3. The term Aivqi,oy or Aivqiopi,a is ambiguous in race and geography. Even Herodotus (Herodotus 7.70) mentions that it refers to all peoples of dark skin, from the country south of Egypt, Nubia, to India.

202

The Bilingualism of the Earliest Christian Church in Jerusalem

vated.” He (1993:74) asserts that although there is not a lot of clear information about ethnicity in the Napatan-Merotic kingdom, the little we do know would allow cautious conjecture that the society was not homogeneous. In regard to language, there are a number of inscriptions; however, the longest and most significant are not yet translated due partly to the inability to decipher the language (O’Connor 1993:71).77 As in other regions of the first-century Roman Empire, many Nubians may have been bilinguals in Egyptian and Meroitic (O’Connor 1993:83). The official language was Egyptian during Napatan times and Meroitic would likely have been the spoken language (Shinnie 1967:134).78 In relation to the linguistic competence of the eunuch, other evidence may be considered. There are ethnographic and linguistic materials from first-century Aksum, adjoining Meroe, which is one of two ancient kingdoms called Ethiopia. Kaplan’s research (1992:17) on the Falashas informs us that some Ethiopian traditions indicate that half the population of Aksum was Jewish before the growth of Christianity. Although these sources likely exaggerate the percentage of Jews, Kaplan (1992:17) notes that there is an overwhelming influence of Hebraic patterns on early Ethiopian culture. Despite the syncretism of beliefs and ceremonies, evidence supports the assertion that Falashas in the pre-Christian Aksumite Kingdom adhered to Jewish customs such as holidays, Sabbath, circumcision, kashrut; and prescriptions of ritual cleanness were observed with considerable strictness (Ullendorff 1990:93-107; esp. 106-7; 1956:247-50). The linguistic milieu of Aksum was bilingual, the two major languages being Greek and Ge~ez, the latter belonging to the same Semitic family as Hebrew and Aramaic (Pankhurst 1998:24-5). Ullendorff (1990:106) considers that most Falashas could speak Ge~ez, but were ignorant of Hebrew. However, there is a strong case to be made that first-generation immigrants used Aramaic (Kaplan 1992:19). Consistent Jewish influence on Ethiopia is apparent on a number of levels (Kaplan 1992:26-32). For instance, there were lively exchanges between Ethiopia and Elephantine Jews and/or Jews in South Arabia, which indicates that some Ethiopians would have known Aramaic or Hebrew (Kaplan 1992:26-32; cf. 17-9; Ullendorff 1990:95-6). 77

78

He (1993:71) adds that “on the linguistic side we know that even in Napatan times the elite probably spoke Meroitic, a language not written down until the second century BCE. Yet, by Christian times…Nubian, a language different from Meroitic, was spoken throughout Nubia.” For more information of the relationship between Egyptian and Meroitic, see Shinnie 1967:132-40.

The Bilingual Seven

203

The eunuch’s visit to Jerusalem to worship may presuppose that he was a devoted Gentile, observant in Jewish practices, and visited Jerusalem with religious enthusiasm (Fitzmyer 1998:412). This implies that he could have been deeply influenced by Jews in Meroe or Aksum. If he often had contact with Falashas in Meroe or Aksum, he might have learnt Hebrew, which was a Semitic language identical with Meroitic and was sufficient for reading Isaiah. On the other hand, the eunuch as a court official would have been a bilingual in Greek and Meroitic. It is not certain which language Philip spoke to the chamberlain. Accordingly, conclusions about the eunuch’s ethnicity and linguistic background need to be reserved until such time when more excavations can be carried out. The eunuch episode itself does not deny that Philip was a bilingual.

5.3.2.4 Antioch There is only one geographical provenance known from the Seven and that is from Nicolas, who is known as a proselyte from Antioch. Many scholars have interpreted Nicolas as a Gentile who was originally from Antioch.79 It is generally hypothesized that Antiochene Gentiles must have spoken Greek. A Gentile derivation of Nicolas seems to reinforce the assumption that at least one out of the Seven who we can know would have spoken Greek. However, the description of him as an “Antiochene Gentile” itself does not prove he was a monolingual Greek. Antioch was a fully bilingual city, in which Greek and Aramaic were used (cf. §4.3). Many upper-status Syrians (Gentiles) were bilingual and most Syrians spoke Aramaic. That Nicolas was of Gentile origins does not suggest whether he spoke Aramaic or Greek. The episodes of these Seven take place in four regions: Philip’s three regions (Caesarea Maritima, Samaria and Meroe of Ethiopia) and Nicolas’ region (Antioch). Many have presupposed that these four regions were straightforwardly Greek-speaking regions. Nicolas of Antioch was just a Greek-speaking Gentile. The Seven, therefore, were Greek-speakers. That the Seven were Greek-speakers supports the assumption that the Hellenists spoke only Greek. However, Caesarea Maritima, Samaria, Meroe, and Antioch were bilingual regions. An analysis of the regions gives no credence to the notion that the Seven and the Hellenists spoke only Greek. Indeed, the nature of these cities and region give every reason to consider that the Hellenists could have been bilingual. 79

Barrett 1994-1998:315; Haenchen 1971:264; Simon 1958:12; Blackman 1936-7:524-5.

204

The Bilingualism of the Earliest Christian Church in Jerusalem

5.3.3 Performance of Their Duty C. F. D. Moule’s approach is that the ~Ebrai/oi were at least semibilinguals while the ~Ellhnistai, were Greek monolinguals (1958-9:100-2).80 He (1958-9:101) argues that despite the fact that all the Seven were the Hellenists who spoke only Greek, “the care of the ‘Hebrew’ needed no special attention and would go on smoothly as before.” Accepting Moule’s linguistic distinction, Hengel (1983:11) stands in the traditional view of Baur’s early schism theory that the linguistic difference resulted in two different linguistic groups having their own services (chapter 5 n.2). He (1983:14-5; 1991:56, 68) argues that the Palestinian Jewish Christians and Hellenistic Jewish Christians would have met separately because the Hellenists would not have understood Aramaic/Hebrew. And he (1983:14-5; 1986:78) views the Seven as the leaders of only the ~Ellhnistai, in the earliest church who were in subordination to the twelve apostles. As a result, he holds the early schism theory in the Jerusalem church due to the linguistic difference. There has been criticism of the traditional view. Some believe too much has been made of the ~Ellhnistai,. If their appointment is related to duties, including the distribution of alms, the ~Ebrai/oi should also have been taken into account seriously. Johnson (1992:110) raises a question: “if we are to conclude from the Greek names of the seven that they were to take over the care of the ‘Hellenist’ widows, who would look after the ‘Hebrew’ widows once the apostles left their station?” Harvey (1996:136) is right to doubt that “the dominant (at least in this situation) Hebrew or Aramaic speaking group did not understand (or want to understand) the needs of the Greek-speaking group (the ‘Seven’).” 81 In relation to these questions, Robert Murray (1982:204) insists that “the committee of seven would have been chosen to represent both sides.”82 Furthermore, Munck (1967:57) suggests: Surely, to assume that the primitive church would choose a committee for social services in which only one of the feuding parties was represented 80 81

82

It is well-known that John Chrysostom (Homilies 14 on Acts 6:1 and 21 on 9:29) first advocated the linguistic approach. In this regard, he suggests that the “Hebrew” implies “good Jews” who were traditional and conservative. However, Penner (2004:70) criticizes two points: (i) The meaning of the term is different from text to text. (ii) There is no evidence that “Hebrews” of Acts 6:1 is indicative of “good Jews’ in 15:5 and 21:20. Unfortunately, however, both Harvey’s and Penner’s arguments are also based on the monolingual approach. Similarly, Simon (1958:5-6) considers that, although the Jerusalem church intended to choose the Seven from the two groups, the result turned out that the seven were all Hellenists.

The Bilingual Seven

205

would be to underestimate its efficiency in practical matters. Such procedure would probably have given rise to complaints from the Hebrews. There were presumably representatives of both groups among the seven, of which the best-known members, Stephen and Philip, may very well have been Hebrews.

The problem, however, lies in that both sides in this discussion have presupposed that the Seven were monolinguals in Greek or Aramaic. Solutions to the problem have been attempted based solely on linguistic differences; the Seven were either Greek-speakers or Aramaicspeakers. Both sides could not overcome their monolingual limitations. The linguistic milieus of most regions of the first-century Roman Near East were bilingual (chapter 4). Many Palestinian Jews were also bilinguals (chapter 3). The Jerusalem church was a bilingual community as well (§5.2). If the Jerusalem church was a bilingual community which mingled Aramaic-speakers with Greek-speakers (cf. Simon 1958:4-5), it makes a great deal of sense that church members would have chosen bilingual rather than monolingual leaders for performance of their duty. The Seven already may have been known as bilingual representatives of the Jerusalem church before the dispute between the Hebrews and the Hellenists broke out.83 Richard (1978:342, 346) suggests that if the Seven were chosen from Jewish and Gentile converts like Nicolas, they may be distinguished and reputable members of the community, unlike the twelve apostles, who were mostly common fishermen.84 It is highly possible that the Seven were upper-status Jews or Gentiles. In relation to the social power of bilingualism in the Roman Empire at this time, bilinguality could have been one of their credentials. As Sterling (2001:274) appropriately points out, “the use of Aramaic seems to have been a point of pride for some first-century Jews; at least Saul of Tarsus thought that being a ‘Hebrew of the Hebrews’ was worthy of inclusion in a credentials list.” Moreover, as I mentioned before (§3.2.5), Hengel (1989a:17) asserts that “the better the knowledge of language a Palestinian Jew acquired, the more easily he could rise in the social scale.” If the church members who were composed of Aramaic-speakers and Greekspeakers picked the Seven as their leaders, they would have chosen bilinguals for harmonious and efficient performance of their duties.

83 84

Simon (1958:8) mentions that “the college of the Seven, far from being constituted by apostolic initiative, already existed before the conflict broke out.” Hengel (1983:26) also puts forward that “earliest Christianity was changed from a basically rural and rustic sect whose founders were Galilean ‘backwoodsmen’ into an active and successful city religion.”

206

The Bilingualism of the Earliest Christian Church in Jerusalem

With regard to their duties, Luke seems to consider the bilingual Seven to be important characters as the leaders of the entire earliest church in Jerusalem. In Acts 6:3, first of all, the Seven as leaders must meet three basic qualifications: (i) A good reputation (marturou,menoi); (ii) be filled with the spirit (plh,reij pneu,matoj); and (iii) wisdom (sofi,a). The two qualifications, pneu,matoj and sofi,a, remind us of Jesus’ promise in Luke 21:15: I will give you a mouth and a wisdom, which none of your adversaries will be able to withstand or contradict. Plummer (1922:479) suggests that a mouth in Luke 21:15 refers to “the power of speech.” The word sofi,a in Luke-Acts plays a significant part in its theology. The term sofi,a in Luke-Acts is used nine times. Those who are filled with wisdom are Jesus (Lk 2:40, 52), Solomon (Lk 11:31), the Twelve (Lk 21:15), the Seven (Acts 6:3), Stephen (Acts 6:10), Joseph (Acts 7:10), and Moses (Acts 7:22).85 The line-up of the wisdom-figures in the New Testament is Jesus, the Twelve, and the Seven. Moreover, the sofi,a in Luke 21:15 does not occur in parallel passages of the synoptic gospels (Mt 10:17-20 or Mk 13:11), which means that Luke intends to emphasize the line-up of Jesus–the Twelve–the Seven. 86 Despite the fact that Jesus gives the promise to the twelve disciples, Luke explains that the promise is fulfilled in the Seven (Barrett 1994-1998: 325; Fitzmyer 1998:358; Conzelmann 1987:47; Haenchen 1971:271). The adversaries could not withstand the wisdom and the spirit with which he spoke (Acts 6:10). Comparing Luke 21:15 with Acts 6:10, Barrett (19941998:325) rightly states that “the coincidence in language can hardly be accidental.” Fitzmyer (1998:358) suggests that “the promise [of Luke 21:15] made there is now fulfilled” [emphasis added]. The Seven fulfilled Jesus’ promise and stood in the line of Jesus and his twelve apostles. Consequently, as the Twelve were the leaders of the whole community of the Jerusalem church, so the Seven were the leaders of the whole community in the Jerusalem church. Second, this line-up is also supported by the fact that Luke transposes the apostles dw,deka for the first time when he denominates the church leaders e`pta, in Acts 6:2-3. This means the two numeral usages result from his literary intention, not from his source, contrary to some scholars who argue that the occurrence of dw,deka here is due to a new source (Barrett 1994-1998:311; Conzelmann 1987:44; Bruce 1979:50). Luke intentionally uses the two numbers dw,deka and e`pta, to show their 85 86

Although the term was also used in Lk 7:35, it is uncertain to whom it refers. However, it is associated with Jesus. Johnson (1992:112) also regards sofi,a as a fulfillment of the promise by Jesus. Witherington III (1998:257) submits that “we see in him [Stephen] the fulfillment of what Jesus promised his disciples he would equip them with for their witness.”

The Bilingual Seven

207

similar and successive leadership. When Jesus called together the apostles, Luke named them the dw,deka in Lk 6:13 (cf. Lk9:1, 12; 18:31). He uses dw,deka only once in Acts. When the new leaders were selected among Jesus’ followers, Luke designated them the e`pta,. Regarding the number, the e`pta,, Simon (1958:118 n.13) mentions that “The number seven need not make us suspicious; it was quite common in the Jewish congregation.” There was a Jewish custom of appointing “seven” to discharge some task in ancient Jewish literature.87 Pearce (1995:477-92; esp. 481-4) argues that the Seven of Ant. 4.214 functioned as rulers in first-century Galilee and suggests that the Seven of Acts 6:3 could be rulers of the new community. As the Twelve represented the twelve tribes of Israel in Lk 22:30 (Jervell 1972:75-112), so the Seven were depicted as the new leaders of the whole Jerusalem church. Hence, Luke intentionally employs the two numbers to show their successive roles as leaders of the whole community. Third, Jesus commanded his twelve disciples to undertake the mission to the world. However, the Seven also were given the charge of mission. Marshall (1980:125) points out that “although Luke depicts them formally as being in charge of the poor relief, he does not disguise the fact that they were spiritual leaders and evangelists.” As Acts 21:8 indicates, some of the Seven play a role not only in distributing alms, but also in administering the message (Fitzmyer 1998:355; Barrett 19941998:306; Tannehill 1990:104; Bruce 1979:50; Simon 1958:6-8). In Acts 1:8, Jesus tells the apostles that they shall receive power when the Holy Spirit has come upon them; and they shall be his witnesses in “Jerusalem” and in “all Judea” and “Samaria” and to “the end of the earth.” According to Luke, the charge of Jesus to the Twelve was also fulfilled by the Seven. Luke shows that the Seven performed the missionary work that Jesus commanded to the Twelve. The Seven were full with the Spirit (Acts 6:3, 10); Stephen in Jerusalem (6:8-7:60), Philip in Samaria (8:1-25), and Philip with the Ethiopian eunuch who comes from the end 87

Josephus mentions “seven” twice. In Ant. 4.214, “Let there be seven men to judge in every city.” In relation to Deut 16:18, he (Bell. 2.571) also states that “as he chose seven judges in every city to hear the lesser quarrels; for as to the greater causes, and those wherein life and death were concerned, he enjoined that they should be brought to him and the seventy elders.” Also, he (Ant. 4.287) notes, “if he in whom the trust was reposed, without any deceit of his own, lose what he was intrusted with, let him come before the seven judges.” Barrett (1994-1998:312) enumerates two cases; one is that Josephus mentioned that “appointed Seven judges in every city in Galilee.” The other is to be found in Megillah 26a and j. Megillah 3.74a.16, the “seven best men in a city.” Citing Billerbeck, Haenchen (1971:263) mentions that “in Jewish communities the local council usually consisted of seven men known as the ‘Seven of the Town’ or ‘Seven Best of the Town.”

208

The Bilingualism of the Earliest Christian Church in Jerusalem

of the earth (8:26-39). Concerning the end of the earth, Ethiopia was quite often identified in the mythological geography of the ancient Greek historians with the southern end of the earth (for more information of the end of the earth of ancient times, see Herodotus, History 3.25; Homer, Odyssey 1.22-4; Strabo, Geog. 1.1.6; 1.2.24; Philostratus, Vita Apoll. 6.1; for more detailed discussion of the end of the earth, Ellis 2001:53-58). The mission to the ends of the world in Acts 1:8 was accomplished by Philip, one of the Seven (Witherington 1998:290; Bruce 1989:380). Luke gradually discloses his narratives of the Gentile mission from Jerusalem to the ends of the earth in a spiral way (§5.2.2): The Twelve converted many Gentiles to Christianity in Jerusalem from the start. The number of Gentile Christians in the Jerusalem church was high enough for the Antiochene Nicolas’ selection as one of the seven leaders. Luke delineates that the Seven succeeded to the missionary work from Jerusalem to the end of the earth that Jesus commanded the Twelve to undertake. Accordingly, the bilingual Christian community in Jerusalem would have chosen the bilingual Seven as their leaders based on the performance of their duty. Luke stresses the line-up: Jesus–the Twelve– the Seven. The Seven fulfilled the promise Jesus gave the Twelve (Lk21:15) and performed the mission Jesus commanded the Twelve (Acts 1:8). The Seven were in succession to the Twelve. The bilinguality of the Hebrews and the Hellenists and bilingual circulation of the Jesus tradition in the Jerusalem church indicate that there was no early schism between the Hebrews and the Hellenists in the Jerusalem church. Rather, the Hellenists inherited the charge of the mission from Jerusalem to the ends of the earth that Jesus commanded the Twelve. Moreover, the bilingualism of the first-century Jerusalem church and Gentile churches and bilingual circulation of the Jesus tradition in the Jerusalem church and the Gentile churches makes unlikely the possibility that the Jesus tradition was misunderstood or mistranslated due to a linguistic gap in relation to translation of the Jesus tradition. Accordingly, the missionary works of the Seven did not result in differences between Jewish Christianity and Gentile Christianity, or between Petrine Christianity and Pauline Christianity.

Bilingualism of the Jerusalem Church

209

5.4 Bilingualism of the Jerusalem Church and the Interdirectionality of the Jesus Tradition The bilingualism of Christian churches of first-century Jerusalem88 and the Roman Near East casts doubt on long-held prejudices that the language of the Diaspora Jews and Gentiles (e;qnh) in the New Testament and other literature was Greek. Many scholars, including the religionsgeschichtliche Schule, assume that Diaspora Jewish Christians and Gentile Christians should be viewed as Greek-speakers when they make a distinction between “Aramaic-speaking Palestinian Jewish Christians and Greek-speaking Diaspora Jewish Christians” or the “Aramaicspeaking Palestinian Christianity and Greek-speaking gentile Christianity.”89 These sharp distinctions between the Aramaic-speaking and Greek-speaking Christian communities are deeply related to the unidirectional development of christological titles because there is an assumption that Greek Christian communities could not understand Aramaic terms. 90 For instance, Messi,aj 91 circulated in the Palestinian setting and was developed into Cristo,j in the Hellenistic setting. The title, o` ui`o.j tou/ avnqrw,pou used in the Palestinian setting was developed into o` ui`o.j tou/ qeou/ used in the Hellenistic setting. Mare in the Palestin88

89

90

91

Some scholars assume that there were many bilinguals in the Jerusalem church, although they do not try to relate the bilingualism to directionality of transmission of the Jesus and gospel traditions. For instance, Hurtado (2005:86) mentions that “we should remind ourselves that from its earliest moments the young Christian movement (at least in Jerusalem and other urban settings) was a bilingual entity, comprising Greek-speaking and Aramaic-speaking believers, and a good many others who were effectively speakers of both languages.” Cf. Heitmüller 1912; Bultmann 1952:1.48-53; 1956; Fuller 1965; Kramer 1966; Conzelmann 1969:72-86; Hahn 1969; Mack 1988; Casey 1991, 1999. For criticism against the clear distinction of the two groups, see Cullmann 1959:323; Kümmel 1973:105-6, 118-9. Conzelmann (1969:72; cf. 1973:69) supposes that “The development of the significance of Cristo,j in the direction of becoming a name took place naturally in the Greek-speaking world. The Greeks do not understand the title Messiah.” However, it is highly possible that the translated form (Cristo,j) was circulated with the transliterated form (Messi,aj) in the first-century Jerusalem church (see §7.4.3.2). It is usual that significant religious terminologies are not translated but transliterated. In modern Bible, it is natural that Cristo,j is transliterated not translated in almost versions in the world like “Christ” in English versions. It is interesting that only the Gospel of John, which is considered as the latest Gospel, twice includes the transliterated title (Messi,aj) rather than the translated title (Cristo,j) . It hints that the title was used both in the Palestinian and Hellenistic settings, when the gospel was written. Redating the Gospel of John, in a sense, Robinson (1976:293) persuasively insists that “The gospel shows the marks of being both Palestinian and Greek.” For the detailed discussion of Messi,aj, see §7.4.3.2.

210

The Bilingualism of the Earliest Christian Church in Jerusalem

ian setting was developed into ku,rioj in the Hellenistic setting.92 It has been assumed that the christological titles correspond to the unidirectionality hypothesis that the Jesus tradition was transmitted from the Aramaic-speaking setting into the Greek-speaking setting. As mentioned before (§5.1), Hengel has diverted scholars’ attention to the translation of the Jesus tradition from the Gentile Christian community in Antioch into Greek-speaking Jewish Christianity in the earliest church in Jerusalem. It was a great turning point. He considers that the Jesus tradition in the Jerusalem church was transmitted from the ~Ebrai/oi who were Aramaic-speaking Palestinian Jews into the ~Ellhnistai, who were Greek-speaking Diaspora Jews. It was the bilinguals in the Jerusalem church who played an important role in bridging the gap between the two different linguistic communities. Although the Hellenists, because of their monolingual handicap, have limited and passive participation in the transmission of the Jesus tradition, the translation-transmission was made by the Hellenists including the Seven in Jerusalem, not “decades later outside Palestine in Antioch or elsewhere” (Hengel 1989a:18). Hengel (1989a:18; 1983:26-8) suggests that some parts of the Jesus tradition, such as the passion narrative and significant theological terms, were translated into Greek (chapter 5 n.8). Although Hengel quite often takes bilingualism into account, his methodology still stays within a monolingual framework. His notions of bilinguals are inconsistent and vague. As a consequence, he also assumes the linguistic unidirectionality hypothesis that the translation92

In the case of mara,na qa, of 1Cor.16:22, the religionsgeschichtliche Schule considers that the Maranatha formula was derived from Hellenistic churches (especially, Antioch), since rm could not have been used in Palestine; Heitmüller 1912:333-4; Bousset 1970:129; cf. 134; Bultmann 1952:51-2; Schulz 1962:126-7; Kramer 1966:99-107, 175. Others have maintained that the formula originated from the earliest Palestinian church; Moule 1977:35-46; Horbury 1998:116; Vermes 2000:201; Fitzmyer 1974:386; 2000:31; Hurtado 2003:172-5. Whether it originated from Jerusalem or Antioch, both positions hold unidirectional Christology based on the linguistic unidirectionality hypothesis. However, the bilingualism of first-century Palestine and the Roman Near East implies that both the Aramaic transliterated form (maranaqa; cf. Didache10:6) and Greek translated equivalents (e;rcou ku,rie VIhsou/ Rev.22:20 and o` ku,rioj evggu,j Phil.4:5) could have been circulated together in the earliest church in Jerusalem as well as in Hellenistic churches (cf. Marshall 1972-3:280; Hurtado 1998:100-111). The Aramaic formula in the Greek epistle was used as codeswitching to stress their Christian identity [SOLIDARITY] (cf. §8.1.4.4). Furthermore, many scholars have suggested the Jewish origins of christological titles (Hengel 1976; Vermes 1983a:83-222; 2000:201; Newman, Davila, & Lewis 1999; Fitzmyer 1974:386; 2000:31; 2007; Chester 2007; Collins & Collins 2008). This supports the bilingual circulation of christological titles during Jesus’ ministry. That is, Jesus was called Mare as well as Ku,rie during Jesus’ ministry (cf. §9.8).

Bilingualism of the Jerusalem Church

211

transmission of the Jesus tradition is unidirectional from Aramaicspeaking into Greek-speaking community in Jerusalem. Contrary to such assumptions, it seems that the Jesus tradition was translated into Greek earlier than the linguistic unidirectionalists (i.e. the old religionsgeschichtliche Schule including Baur or the new religionsgeschichtliche Schule including Hengel) thought. If Hengel subscribes to “a large bilingualism of first-century Palestine,” the Jesus traditions including the passion narrative, significant theological terms, and even some christological titles could have started to be translated from Aramaic into Greek and retranslated from Greek into Aramaic, and circulated in Aramaic as well as in Greek in Galilee during Jesus’ public ministry rather than in the Jerusalem church or in the Antioch church. Consequently, translation-transmission started from Galilee not from the Jerusalem church or the Antioch church. Aramaic was the matrix language of many regions of the firstcentury Roman Near East. This implies that a great number of Gentiles in the Roman Near East spoke Aramaic, too, and that some Gentiles would have spoken only Aramaic. A large population of Aramaicmatrix Gentiles and Greek-matrix Jews suggests that many Aramaicmatrix Gentile Christians and many Greek-matrix Jewish Christians were in the Jerusalem church and in Christian communities of the Roman Empire. The bilingualism of first-century Jerusalem and the Roman Near East indicates that many Gentile and Jewish Christians could be bilinguals and that the Jerusalem church was a bilingual community. In this regard, the Jesus tradition was circulated and performed in Aramaic as well as in Greek in the Jerusalem church and in the churches of the Roman Empire. The bilingualism of first-century Palestine and the Roman Near East throws doubt on the traditional views of Hebrews as Aramaicspeaking Palestinian Jewish Christians and Hellenists as the Greekspeaking Diapora Jewish Christians. Instead, I argue that Hebrews refer to the Aramaic-matrix Christians and that Hellenists refer to the Greekmatrix Christians without respect to monolingual, geographical, and ethnic distinctions. I also propose that in comparing the Twelve with the Seven, Luke expresses the idea that the Seven were successors to the Twelve. Luke shows that the Seven inherited the promise Jesus gave to the Twelve and accomplished the Gentile mission to the ends of the earth commanded by Jesus to the twelve. The line-up of Jesus–the Twelve–the Seven is emphasized in the Lucan narratives. The two points, (i) the bilinguality of Hebrews and Hellenists and (ii) bilingual circulation of the Jesus tradition in the Jerusalem church and the Greek-matrix churches in the first-century Roman

212

The Bilingualism of the Earliest Christian Church in Jerusalem

Empire, support rejection of the theory of the early schism between Hellenists and Hebrews, and the sharp distinction between the Jewish Chrisitanity and the Gentile Chrisitanity. Furthermore, it seems that the monolingual view of Palestinian Christians allows Baur and Bultmann to presuppose that the Palestinian Christians, including the disciples of Jesus, had a linguistic reason to limit the course of transmission of the translation from Aramaic into Greek in the Jerusalem church. However, the translation-transmission of the Jesus tradition in the Jerusalem church was performed strictly under the supervision of bilingual disciples93 and followers of Jesus94 who were eyewitnesses to the sayings by Jesus and participants in the stories about Jesus.95 Bilinguality of the disciples and eyewitnesses of Jesus’ acts makes it possible to impose censorship upon the translationtransmission of the Jesus tradition during Jesus’ ministry and the Jesus and gospel tradition in the earliest churches in Jerusalem and the Roman Empire in the first century. Accordingly, the Jesus and gospel tradition was translated under controlled circumstances from Aramaic into Greek and Greek into Aramaic, and it circulated in Greek and vernacular languages including Aramaic in the earliest Jerusalem church as well as in churches in the Roman Empire, including Antioch. Furthermore, we cannot deny the possibility that the synoptic gospels written in Greek were retranslated into Aramaic and circulated in Aramaic-matrix Christian communities orally as well as in written form. Moreover, the synoptic gospels in Greek could have been translated into the Aramaic language family for the most part. The transmission of the Jesus and gospel tradition is not unilinear but hybrid, not teleological but circular, not unidirectional but interdirectional during Jesus’ ministry, in the Jerusalem church, and in Gentile Christian communities.

93

94

95

Most disciples of Jesus could be bilinguals in the Jerusalem church (§3.2.4): Peter, Andrew, Philip, James, John, and Matthew. And Jesus and his mother could be bilinguals (§3.2.4). Many followers of Jesus should be bilinguals in the Jerusalem church (cf. §3.2): at least, Nicodemus (Jn 3:1), a centurion near the Cross (Mk 15:39), a centurion and his servant in Capernaum (Mt 8:5; Lk 7:1), Joseph of Arimathea (Mk15:43), Jairus and his family (Mk5:22), Alexander and Rufus (Mk 15:21), and Zacchaeus (Lk19:2). Dibelius (1949:31-2) persuasively suggests that “The Greek narratives arose at a period when many eyewitnesses of Jesus’ ministry were still living. They, especially the personal disciples of Jesus, would have been in a position to correct any egregious misrepresentation.” On the eyewitnesses to Jesus and their functions in the transmission of gospel tradition, see Byrskog 2000:65-91, chapters 4 & 6 (§1.4.1.3) and Baukham 2006.

Part II: Interdirectional Transmission of the Jesus and Gospel Traditions in Bilingual Contexts at the Levels of Syntax, Phonology, and Semantics Bilingualism of first-century Palestine and the Roman Near East was previously investigated (Part I). As chapters one and two showed, the topic of bilingualism in Palestine has been widely considered among biblical scholars even since the nineteenth century. However, their approaches had two drawbacks. First, it seems that their concepts and methodology of bilingualism have been obscure and immature. This is because the study of bilingualism as a branch of sociolinguistics had not been taken seriously until the 1960s (cf. introduction to chapter two).1 Second, scholars’ applications of bilingualism to the four gospels and Acts are either excessive or insufficient. In relation to the syntactic level of New Testament Greek, scholars have applied the concepts of language contact to the analysis of New Testament Greek. They commonly assumed that language contact between Semitic and Greek resulted in Semitisms or Septuagintalisms. This would imply that syntactic Semitisms provide us with temporal priority. That is, gospel traditions which include syntactic Semitisms could be older traditions. However, it seems that most syntactic shifts called Semitisms or Septuagintalisms occur due to internal-induced syntactic change rather than contact-induced syntactic change (chapter six). Biblical scholars, in this respect, have overemphasized the bilingual approach to the syntax of New Testament Greek. On the other hand, most scholars have not taken bilingualism into account seriously at the levels of phonology and semantics. In terms of transliterated variants, they have tried to single out the original Semitic spelling from variant spellings of transliterated words on the basis of orthography. Three assumptions have been made: (i) The original spelling is the correct spelling and the correct spelling is the original spell-

1

Many biblical scholars, for instance, used to think of “bilinguals” as “early bilinguals,” “primary bilinguals,” or “balanced bilinguals” regardless of “late bilinguals,” “acquired bilinguals,” or “dominant bilinguals,” as mentioned before (§2.1).

214

Interdirectional Transmission of the Jesus and Gospel Traditions

ing. (ii) A Semitic type spelling is more original than a Greek type spelling. (iii) A Semitic transliterated word is closer to the original sayings of Jesus and stories about Jesus, than a Greek translated word. However, it is normal that the transliteration of proper nouns from a source language to a target language leads to phonological variants due to various factors so that it is meaningless to seek the original spelling from among variant spellings (chapter 7). Furthermore, the bilingualism of the first-century Palestine and Roman Near East made it possible to consider that Semitic-type spellings and Semitic transliterated words were circulated with their Greek equivalents (i.e. Greek type spellings and Greek translated words) from the start. This means that Semitic spellings are not always earlier than Greek spellings (chapter 7). Bilingualism in first-century Palestine and Roman Near East should be taken into consideration seriously at the semantic level. Bilingualism brings about interference, borrowing, and codeswitching. In terms of Aramaic embedded words and language changes from Greek into Aramaic, most scholars have regarded such occurrences as interferences or borrowings. Speakers and writers, however, could use Aramaic embedded words in their Greek texts intentionally in a bilingual context (chapter 8). When we take seriously the bilingualism of firstcentury Palestine and the Roman Empire, we find that biblical writers could employ the Aramaic embedded words as an intentional literary device called “codeswitching.” This means that Aramaic embedded words themselves do not always indicate temporal priority. Accordingly, it seems that most biblical scholars have held a one-sided view of bilingualism at the syntactic, phonological, and semantic levels. I would contend that their improper understanding of bilingualism is deeply related to their views of language change at the syntactic, phonological, and semantic levels. First of all, two approaches to language change need to be investigated. The study of langue, with a theoretical inclination, has been in the mainstream of modern linguistics. It was propounded by Ferdinand de Saussure, who opened a new chapter of modern linguistics called “general linguistics,” which was developed by Noam Chomsky. Criticizing historical linguistics in the nineteenth century, de Saussure did not take language variation or language change into consideration seriously. As is well-known, his train of thought can be summarized like this (1959:26-58): (i) Langue is a system, whereas parole is its utterance. (ii) Langue is homogeneous, whereas parole is heterogeneous. (iii) Language shifts in parole, not in langue. (iv) The shifts of parole cannot change the langue. (v) The mission of linguists is to analyze langue, not parole.

Interdirectional Transmission of the Jesus and Gospel Traditions

215

For de Saussure, language variation or language change is considered an abnormal linguistic phenomenon. Reemphasizing the Saussurian dichotomy, Chomsky substitutes language competence and language performance for langue and parole. He (1965:3) also excludes language change or language variation from the subject of linguistics and defines the scope of the study of linguistics as follows: Linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech-community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of the language in actual performance.

His concept of a speaker or a linguistic community has been criticized, however, because the ideal model of an ideal speaker-listener in an ideal linguistic community could not be found in the world. Consequently, it has been objected that the structuralism of SaussureChomsky failed to explain language variation or language change adequately, because they do not take it as a part of linguistics. The pendulum swings between rationalism and empiricism in the history of research. The study of parole, with an empirical inclination, has been investigated in sociolinguistics, cognitive linguistics, cultural linguistics, and anthropological linguistics. Since the 1960s, scholars using these approaches have taken seriously language variation caused by diachronic (i.e. historical factors) and synchronic (i.e. social factors) change. 2 Unlike Saussure-Chomsky, Labov (1972:188-91; originally 1960s) suggests that there are four distinct difficulties that have had a significant effect on linguistic practice: (i) ungrammaticality of speech, (ii) variation in speech and in the speech community, (iii) difficulties of hearing and recording, and (iv) rarity of syntactic forms. Labov (1966) considered linguistic variables as structural units, a view that has been supported by Chambers (1995:12-32) and Chambers and Trudgill (1998:127). This means, as Chambers (1995:25) mentions, that “hitherto, all linguistic units--phones, phonemes, morphemes, phrases, clauses-had been invariant, discrete, and qualitative. The variable is none of these.” He continues, “Instead, it is variant, continuous, and quantitative.” Going a step further, Chambers (1995:26-32) calls the study of 2

Although Sapir (1921:147) confidently observed that “everyone knows that language is variable,” Chambers (1995:12) states, “Though linguistic variation may be obvious, no linguists analyzed it systematically until the inception of sociolinguistics in the 1960s.” Around twenty years ago, Labov (1972:186) also mentions, “The science of parole never developed, but this approach to the science of langue has been extremely successful over the past half-century.”

216

Interdirectional Transmission of the Jesus and Gospel Traditions

langue categoricity theory and the study of parole variation theory. He (1995:25) strongly insists that “The variable can exist only in a theory that abandons the axiom of categoricity.” 3 In sum, categoricity theory (i.e. the study of langue) is not concerned with language variation, language change, or variants. In contrast, variation theory (i.e. the study of parole) regards language as variation, change, or variants. Many biblical scholars have considered Hellenistic Greek as langue on the basis of categoricity theory (§6.1).4 In other words, they have considered language variation, language change, or variants found in New Testament Greek to be abnormal. However, the fact that Hellenistic Greek itself changed synchronically and diachronically means that Hellenistic Greek should be considered as parole, as will be fully discussed below (§6.1). As investigated in Part I, the linguistic milieus of first-century Palestine and the Roman Near East were bilingualism in Greek and regional vernacular languages. I have also argued that the earliest Christian community in Jerusalem was a bilingual community. If the Jesus tradition was formed in bilingual contexts and the synoptic gospels were circulated in bilingual contexts, which approach can explain their linguistic situations more appropriately? Milroy and Muysken (1995:3) observe: Noam Chomsky’s meta-theoretical focus on the ideal native speaker in the ideal speech community is perhaps the most famous modern embodiment of this monolingual and non-variationist focus. While generative grammar has flourished by focusing on simple cases and ignoring more complex situations such as bilingualism, generativists are not alone amongst modern linguists in reflecting such a traditional orientation.

Mackey (1968:554) also suggests that “bilingualism is not a phenomenon of language; it is a characteristic of its use. It does not belong to the domain of ‘langue’ but of ‘parole’.” Although some biblical scholars accept the idea that the linguistic milieus of first-century Palestine and the Roman Near East were bilingualism, it is regrettable that their arguments are still based on Saussure-Chomskyan theory. Consequently,

3

4

Incidentally, examining the languages of Aristophanes from the perspective of sociolinguistics, Colvin (1999:307) concludes that “The Greeks did not think of other dialects of Greek as imperfect approximations of their own dialect; they did not suppose their own dialect to be more correct than others“ [emphasis in the original]. Fortunately, Reed (2000:121-9; cf. 1995:222-265) encourages New Testament scholars to discuss New Testament Greek on the basis of the science of parole not that of langue. Also, Silva tries to investigate New Testament Greek from the perspective of style (i.e. parole), although he considers Hellenistic Greek langue and New Testament Greek parole. See my criticism of his argument (§6.1.4).

Interdirectional Transmission of the Jesus and Gospel Traditions

217

the approach to the language of the New Testament should be investigated from the perspective of variation theory. Moreover, if Jesus usually spoke Aramaic and the surviving gospels were written in Greek, this example of language contact warrants an application of translation theory to this field. Koller (1979:183) calls translation theory “a science of parole.” Fawcett (1997:4) also posits that “the view that translation theory must be studied as parole (a communicative event) rather than langue (an abstract system) is now widely accepted….” He explains that “translation is a fact of parole that there is no such thing as the one ‘right’ translation of a message.” The Jesus and gospel traditions were translated from Aramaic to Greek and later, vice versa. As such, our analysis of Jesus and gospel traditions in bilingual contexts should be examined from the view of the study of parole rather than langue, language performance rather than language competence, and variation theory rather than categoricity theory. In relation to categoricity theory, for instance, James Barr made an enormous impact when he encouraged biblical scholars to consider general linguistics (i.e. the science of langue), especially lexical semantics seriously. 5 Barr (1961) makes two major arguments: (i) biblical scholarship has drawn a sharp contrast between Hebrew and Greek thought derived from linguistic relativism (e.g. Boman 1960), and (ii) the etymological approach to biblical studies6 is flawed. “Etymology,” he (1961:107) proposes, “is not, and does not profess to be, a guide to the semantic value of words in their current usage, and such value has to be determined from the current usage and not from the derivation.” Barr’s application of general linguistics should be used as a corrective to help biblical scholars who formerly have been inclined toward only a diachronic approach, without regard to synchronic research.7 And yet,

5

6

7

While Barr holds the categoricity theory at the lexical semantic level, he takes variation theory at the phonological level (§7.2.2.2). Although Barr (1961:294) acknowledged the contemporary empirical approach (e.g. Humboltian heritage, Greenberg’s arguments, and Sapir-Whorf hypothesis), it seems that he might not have expected that the empirical approaches would flourish since the 1960s. He disregards the science of parole. The contributions of Boman to biblical studies (1960) should be reconsidered, to some extent. The science of langue has disregarded etymology. In this respect, Amsler (1989:8) reasonably complains, “Etymology…has been systematically excluded from most contemporary discourses of language science.” Barr’s two arguments have been welcomed by many scholars: Louw 1982; Silva 1983; Thiselton 1985:75-104; Carson 1996; Black 2001:249; Groom 2003:61-71; Turner, 2010:189-217; etc.

218

Interdirectional Transmission of the Jesus and Gospel Traditions

Barr’s influence was so overrated that some biblical scholars became overly sensitive to his views.8 However, we must reconsider the criticism of linguistic relativism (e.g. the Sapir-Whorf hypothesis). First, despite Saussure and Barr’s warning, linguistic relativism has revived and flourished in contemporary anthropological, psychological, and cognitive-linguistic researches.9 Bonvillain (2007:44-49) suggests that Sapir and Whorf had great influence on linguistic anthropology and he re-examines linguistic relativity. Second, Barr (1961:107) considers that etymology “is a historical study” so that “the past of a word” cannot contribute to the determination of its current meaning. However, the etymology of a word may influence its current meaning. Etymologization is caused by a human cognitive process. When a person hears a word, its etymology may occur to him or her on the level of analogy between the word and its etymology. There is no reason to think that the first-century Roman Near East was an exception. Before Barr’s famous publication (1961), Curtius (1953:495-500) defines etymology as “a category of thought.” He (1953:495-6) shows that etymology was used by Homer, Pindar, Aeschylus, Plato, Aristotle, Cicero, Quintilian, Theon, Virgil, Jerome,

8

9

Their sensitiveness to Barr is shown that when biblical scholars deal with arguments in relation to linguistic relativism or etymology, they usually have mentioned that they fully acknowledged Barr’s criticisms (e.g. Grabbe 1988:3). Meadowcroft (1995:153-4 n.60) also mentions that “I am conscious of the caution from J. Barr, … for care to be taken in etymological matters. In this instance, however, there does seem to be evidence of historical development of meaning in the directions I have indicated.” However, he would have done better to recognize that etymology was and is a useful method without this excuse. Running to excess, Brown (1989:132) recognizes that the “style” of the authors of the New Testament literature belongs to parole. However, he tries to squeeze the concept of “style” into the scope of langue because Saussure considers parole outside the scope of linguistics. For von Humbolt’s relativism, see von Humbolt 1988. Sapir and Whorf developed the Humboltian linguistic relativism; see Sapir 1929:207-214; Carroll 1964:207-219; Kay & Kempton 1984:65-79. For a detailed discussion of the revival of linguistic relativism, see Cooper and Spolsky 1991; Gumperz and Levinson 1996; Duranti 1997; Foley 1997. Lucy (1992) also performed a case study of the linguistic relativity hypothesis by means of making a comparison of a language category and a cognitive category between Yucatec Maya and American English. Recently, it is interesting that cognitive science supports linguistic relativism. According to Aubrey Gilbert, it is shown that color discrimination is affected by language (Gilbert, et al 2006:489494; cf. Gilbert, et. al. 2008:91-98). Also, Franklin suggests that lateralization of categorical perception of color is affected by color language (Franklin, A. et al. 2008:18221-18225).

ȱ

Interdirectional Transmission of the Jesus and Gospel Traditions

219

Augustine, and Isidore of Seville.10 Especially, both Cicero and Quintilian believed that etymology “is useful to the orator” (Curtius 1953:495). Fortunately, Sawyer (1999:116), among biblical scholars, observes that “Monolingual etymological or pseudo-etymological explanations of the meaning of words can be and have been used by a great many authors, ancient and modern.” He (1999:112-28) illustrates instances of etymological interpretations of biblical names. Grabbe (1988) points out that Philo used etymological interpretation. Contemporary etymology also was reflected in the Greek spelling of Jerusalem (~Ieroso,luma) (§7.1.2). With this in mind, it is no surprise that Jesus might have made use of etymological interpretation. The use of an Aramaic nickname Boanhrge,j (Mk 3:17) was deeply related to its folk etymology (§7.1.2). When Jesus mentions that evpi. tau,th| th/| pe,tra| oivkodomh,sw mou th.n evkklhsi,an (Mt 16:18), this is related to Peter’s Greek nickname Pe,troj.11 When apk is transliterated into its equivalent Greek name (i.e. Khfa/j), it seems that etymology has influence upon the transliteration. Fitzmyer (1979b:123) reasonably presumes that “Today we smile at the relation seen between Greek Khfa/j and Latin caput by some patristic writers, who assumed a connection between Khfa/j and kefalh,.” Paul also uses etymology when he uses the pun that VOnh,simoj is not a;crhstoj but eu;crhstoj at Philemon 1:11 (for more Pauline examples, see Leary 1992:468 n.6). Etymology was and is used as a valid device of human communication (cf. Sawyer 1999:116-7). Consequently, there is no question that the writers as well as readers of the New Testament literature take etymological interpretation into consideration. Moreover, orality theory recently has shed fresh insight into the study of the transmission of the Jesus and gospel traditions, as we dis10

11

Isidor of Seville (c. 560-636) systematizes etymology as grammatical science and gave etymological interpretations of the most significant figures of the Old Testament. For the detailed discussion of his publications, see Curtius 1953:496-7; Amsler 1989:133-72. Also, Sluiter (1990:12) suggests that etymologies are found in the Greek poets, Aristotle, and the Stoics. He (1990:12) mentions that “In Aristotle, etymologies serve the purpose of a posteriori confirming Aristotle’s interpretation of a given word” and that “the most famous etymologists in Antiquity were the Stoics.” For the discussion of ancient etymology, see Woodhead 1928; Amsler 1989:15-56; Sluiter 1990:5-37. Of course, Barr also acknowledged that etymological interpretation was used in the first century. He (1961:109) states, “Place names were also etymologized, but less often in the OT period; in the NT we find a case like the etymologizing of Salem as ‘peace’ (along with that of Melchizedek as ‘king of righteousness’) in Hebr. 7:2. A more systematic etymologization of the biblical place-names was carried out by Philo, Origen, Jerome and others.” Whether its origin was in Aramaic (apk) or Greek (Pe,troj) or whether it was said by Jesus or not, the fact remains that etymology was indeed used. For detailed discussion, see §7.4.2.1; §8.4.2.1.

220

Interdirectional Transmission of the Jesus and Gospel Traditions

cussed in detail (§1.3.2; §1.5.2). Orality theory requires us to take a serious account of the grammar of sound.12 In antiquity, orality was heavily emphasized. Furthermore, writing was intended for reading. Dewey (2008:85) properly points out that “All ancient literature was designed for oral performance and would be composed with a listening audience in mind.” Achtemeier (1990:15; cf. Gilliard 1993:689-94) states that “the oral environment was so pervasive that no writing occurred that was not vocalized.” He (1990:19) proposes that “organization of written materials will depend on sound rather than sight for its effectiveness.” In this sense, he (1990:19) stresses the grammar of sound rather than the grammar of writing when he states, “Similarities of sound will be more important than similarities of visual appearance, and sound patterns will provide the clues rather than the visual patterns created by similar or identical phrases.” On this account, when we deal with our gospels, which were intended for reading aloud, the grammar of sound should be taken into consideration rather than the grammar of writing. The concept of the grammar of sound supports variation theory, the science of parole and language performance at the syntactic, phonological and semantic levels (for detailed discussion, see §7.1.2). As a matter of fact, the concepts of the categoricity theory, the science of langue, the grammar of writing, and language competence are deeply related to the unidirectionality hypotheses from JudaeoPalestinian to Hellenistic tradition, from oral into written form, and from Semitic into Greek. From the perspective of historical criticism (e.g. source criticism, textual criticism, form criticism, and redaction criticism), some observers have sought to reconstruct the “original traditions” of Jesus and the gospels. Powell (1990:8) states that historical criticism aims at “a reconstruction of something to which the text attests.” He (1990:9) also suggests that New Testament scholars who use historical criticism regard texts of the New Testament as “the end product of a process of development.” Historical criticism is based on the three unidirectionality hypotheses. Even some scholars of literary criticism, as mentioned before (§1.3.2), have tried to find more original traditions of Jesus and the gospels by means of screening oral features out of the written four gospels. New Testament scholars who use either 12

Dean (1996:53-70) employs “grammar of sound” which is accepted by Hearon (2006:4-6; originally, 2004). For a detailed discussion of the grammar of sound, see Dean 1996. Scott and Dean (1996) also try to explain the Sermon on the Mount from the perspective of the grammar of sound. It is noteworthy to mention that Shiner (2003, cf. 2009) shows the Gospel of Mark to be oral performance. Also, Rhoads (2009:83-100) suggests that performance studies should be a method of biblical interpretation.

Interdirectional Transmission of the Jesus and Gospel Traditions

221

historical criticism or literary criticism have tried to single out oral features or Semitisms (or Septuagintalisms) at the phonological, syntactic, and semantic levels from the four gospels and Acts. In doing so, they assume that sayings and stories which betray oral features or Semitic influence/interference have a greater claim to “authenticity.” This view considers such tradition to be closer to the historical Jesus according to the unilinear transmission hypothesis. The bilingualism of first-century Palestine and the Roman Near East, however, requires another view of the transmission of the Jesus and gospel traditions. From the perspectives of variation theory, the science of parole, the grammar of sound, and language performance, Part II will be devoted to the transmission of the Jesus and gospel traditions at the levels of syntax, phonology, and semantics. With respect to the syntax of New Testament Greek, many scholars have assumed that Hellenistic Greek is langue, whereas New Testament Greek is parole due to Semitisms or Septuagintalisms. However, both Hellenistic Greek and New Testament Greek should be thought of as parole. Furthermore, grammaticalization theory from the perspective of cognitive linguistics shows that syntactic changes are not Semitic interference due to contact-induced syntactic shifts, but grammatical polysemies due to internal-induced syntactic shifts (chapter 6). At the phonological level, when many scholars deal with phonological variants, they have sought to single out only the original Semitic spelling from transliterated variants on the basis of categoricity theory, the science of langue, language competence, and the grammar of writing. In other words, they do not seriously consider that transliteration due to language contact tends to result in phonological variants and that bilingual circumstance does not give temporal priority to Semitic spelling over Greek spelling. Moreover, since they neglected the possibility that the oral dimension of the New Testament literature requires the grammar of sound, their discussions might have been made on the categoricity theory. In this respect, I would propose that transliterated variants will be considered to be transliterated allolexemes (§7.3.1) from the view of variation theory (chapter 7). Finally, Semitic words embedded in the synoptic gospels and Acts have been investigated from the perspective of language competence. Many New Testament scholars have considered Semitic embedded words as interference (or borrowing). In this view, Semitic words have temporal priority because Jesus and his disciples used Aramaic as their matrix language. Such semantic interference, however, should be investigated from the view of language performance in the bilingualism in first-century Palestine and the Roman Near East. The bilingual authors

222

Interdirectional Transmission of the Jesus and Gospel Traditions

intentionally change language codes in bilingual contexts for pragmatic purposes; this is otherwise known as codeswitching (chapter 8).

ȱ

6. Syntax Syntactic changes13 are often found in the language of the synoptic gospels and Acts. Most biblical scholars have assumed that these syntactic shifts have resulted from language contact. In fact, language contact at the syntactic level may cause syntactic interference. It has been taken for granted that the bilingualism of first-century Palestine and the Roman Near East made it possible for Semitic syntax to interfere with the Greek syntax of the synoptic gospels and Acts. Consequently, the syntactic changes of New Testament Greek have been viewed as Semitic interference (Semitisms or Septuagintalisms). This can be called the contact-induced syntactic change hypothesis. The contact-induced syntactic change hypothesis implies that syntactic changes in New Testament Greek are abnormal linguistic phenomena in comparison with classical Greek or Hellenistic Greek. This implies that Semitic interference hypothesis (i.e. Semitisms or Septuagintalisms) is based on the concept of categoricity theory. On the basis of the unidirectionality hypothesis, many scholars have used Semitic coloring as criteria for determining the temporal priority of Jesus and gospel traditions, but not so for the Pauline epistles.14 It has been presupposed that Semitic tradition is more original than Greek tradition and that Semitic tradition is closer to the original sayings of Jesus and stories about Jesus. In this respect, the Semitic interference hypothesis (on the assumption of the unidirectionality

13

14

Semitic interference, Semitic coloring, or syntactic solecism, which is a well-known term, has usually been used on the presupposition that the syntactic change is due to language contact. I will use syntactic change because syntactic change or shift is a more neutral term. The Semitisms of the synoptic gospels and Acts can be compared with the Pauline epistles. Despite the fact that Paul was bilingual and also inserted two Aramaic embedded words in his epistles (i.e. abba and Marana,qa), Pauline Semitisms usually have been neglected. In this respect, Fitzmyer (1979a:4) remarks that “most discussions of the Aramaic problem have been limited to the Gospels and Acts.” It may be (i) because Paul is considered to be an early or primary bilingual (§3.2.5) and (ii) because Semitic interference hypothesis plays in a significant role of the gospel studies in relation to temporal priority.

224

Syntax

hypothesis) is deeply connected with major issues of gospel studies such as the historical Jesus, the synoptic problem, the provenance of the gospels and Acts, textual-criticism of the gospels and Acts, etc. I would suggest, however, that the Semitic interference hypothesis based on contact-induced syntactic change is too hasty a conclusion. It disregards the other possibility of syntactic shift of New Testament Greek, an internal-induced syntactic change hypothesis. Cognitive linguists have argued that syntactic change occurs in a language itself, in the grammaticalization process caused by human cognition (§6.2.2.1) regardless of language contact. In relation to this, they call syntactic changes syntactic polysemies. If syntactic shifts of the four gospels and Acts that have been called Semitisms or Septuagintalisms correspond to the paradigm of the grammaticalization process and the so-called syntactic solecisms are found in Hellenistic Greek literature, the syntactic changes should be considered to be syntactic polysemies due to internal-induced syntactic changes, not Semitisms or Septuagintalisms due to contact-induced syntactic changes. Also, if syntactic shifts observe the typical pattern of grammaticalization process, although the usages are not found in Hellenistic Greek literature except for the New Testament literature, then it is highly probable that the syntactic changes can be ascribed to internal-induced changes because the vocabulary of the New Testament is a tiny fraction of Hellenistic Greek, as Caragounis (2004:123) mentions: As such it [NT Greek] is part of the ‘Koine.’ But the Koine is a much larger entity. Its vocabulary must have run up to hundreds of thousands of words, whereas the New Testament contains only about 4,900 of these. In other words, the New Testament represents only a tiny fraction of Koine Greek.15

Consequently, in what follows, the unidirectionality hypothesis at the syntactic level will be examined from the perspective of the internalinduced syntactic change hypothesis. First, I will discuss the history on the research of New Testament Greek with regard to the commonly 15

As for emphasis on heavy Semitic interference in New Testament Greek, Caragounis (2004:40-1; 121-3) considers Koine Greek as “a sub-standard language” and New Testament Greek as non-standard Koine Greek used by non-Greeks, although he admits that New Testament Greek is “one branch of ‘Koine’ Greek.” Unfortunately, his view of New Testament Greek is also based on the contact-induced syntactic change hypothesis and the categoricity theory. One more mention should be added. Despite Caragounis’s protest (2005) against Silva’s review, Silva (2005) reasonably criticizes that he disregards some basic concepts of modern linguistics and general linguistics. It would be more persuasive if he employed modern linguistic methodology in order to deal with Greek language properly. At the least, he should consult modern historical linguistics to support his position.

Contact-Induced Syntactic Change Hypothesis

225

held presuppositions of the contact-induced syntactic shift hypothesis (§6.1). Second, internal-induced syntactic change hypothesis as an alternative to the traditional view will then be introduced (§6.2). Some cases of syntactic changes which have been known as Semitisms or Septuagintalisms will be regarded as grammatical polysemies on the basis of the internal-induced syntactic change hypothesis (§6.3). Finally, a relationship between internal-induced syntactic change hypothesis and interdirectionality hypothesis will be suggested (§6.4).

6.1 Contact-Induced Syntactic Change Hypothesis in New Testament Scholarship Most New Testament scholars have considered the syntactic changes of the four gospels and Acts to be Semitisms or Septuagintalisms on the basis of the contact-induced syntactic change hypothesis. However, as mentioned before, both approaches (i.e. the contact-induced syntactic change hypothesis and the internal-induced syntactic change hypothesis) should be examined whenever the syntactic changes are found in New Testament Greek. It is regrettable that New Testament scholarship by and large has adopted the contact-induced syntactic change hypothesis on the basis of the categoricity theory without the necessary consideration of internal-induced syntactic change. For a long time, the language of the New Testament was called “Holy Greek,” “Biblical Greek,” “sacred Greek,” or “Jewish Greek.”16 This means that syntactic changes in New Testament Greek have peculiar characteristics that are not found in Hellenistic Greek. This opinion was fiercely and convincingly attacked by Adolf Deissmann’s vernacular Greek theory. However, on account of Deissmann’s inconsistency, the concept of the peculiar Greek has been unceasingly maintained in New Testament scholarship. As a turning point in this debate,17 Deissmann will be considered first (§6.1.1). Then, four responses to Deissmann will be investigated respectively: the linguistic relationship between syntactic changes (§6.1.2), bilingual Jewish Greek theory (§6.1.3), Hellenistic Greek theory (§6.1.4), and Semitisms and Septuagintalisms (§6.1.5).

16 17

For the detailed history of Jewish Greek hypothesis, see Voelz 1984:894-906; Marcos 2000:3-6; Jung 2004:7-8. Many scholars consider Deissmann’s opinion to be a watershed; Voelz 1984:906-10; Gamble 1995:32; Caragounis 2004:124 n. 120.

226

Syntax

6.1.1 Adolf Deissmann Although Deissmann convincingly limited the alleged Semitisms of the Septuagint and the New Testament to a very small number at the lexical-semantic levels, he still accepted Semitisms and Septuagintalisms at the syntactic level. His vague view of the language of the Septuagint and the New Testament seems to make his argument abstruse18 and even incoherent. It seems that he considers the syntactic shift as a contact-induced change, but the lexical-semantic shift as an internalinduced change. Deissmann defines the language of the translators of the Septuagint and the authors of the New Testament as “Hellenistic Greek.”19 This is adduced from his archaeological evidence. In the case of Septuagintal Greek, he (1903:70, 72) suggests that Egyptian papyri of the Ptolemaic period are of the highest significance in understanding the Greek of the Septuagint because the spoken and written Greek of the translators of the Septuagint is the vernacular language of contemporary Egyptian Greek. He also considers the Greek of the authors of the New Testament as Hellenistic Greek. He (1903:84; 1991:53) argues that the inscriptions of Asia Minor make a huge contribution to the study of the Greek of the New Testament. Moreover, he (1903:83-4) rejected the view that the authors brought lexical-semantic borrowings of theological terms from the usages in the Septuagint (i.e. Septuagintalisms). He (1903:83) also mentions that “the mere use of LXX-words on the part of an inhabitant of Asia Minor is no guarantee that he is using the corresponding LXX-conceptions.” Simply speaking, he (1903:81) proposes that “just as we must set our printed Septuagint side by side with the Ptolemaic Papyri, so must we read the New Testament in the light of the opened folios of the Inscriptions.” Consequently, Deissmann asserts that the authors of the Septuagint and the New Testament used Helle18

19

Silva (1980:199) mentions that “the standard (and almost wearisome) characterization nowadays is that Deissmann made a major contribution to the field but that he took his views to an extreme and so they need considerable modification; yet when concrete instances of modifications are given, they often turn out to be items that had been readily admitted by Deissmann himself“ [emphasis in the original]. The term is also called “koine Greek,” “Greek world language,” “middle Greek,” “late Greek,” or “old Greek koinh,.” It seems that the koine Greek enjoys common popularity among scholars (Deissmann 1991:42). Although he (1991:43) suggests that “Hellenistic world Greek” is the clearest term, he (also Moulton 1919a:2 n.1) uses the briefer term, “Hellenistic Greek.” So I use “Hellenistic Greek.” And I follow Deissmann’s broader definition of “Hellenistic Greek”: “the entire Greek language development from Alexander the Great until into the sixth century CE, i.e. until the main characteristics of the new Greek had formed, and not only the written but also the colloquial language” (1991:42-3).

Contact-Induced Syntactic Change Hypothesis

227

nistic Greek as their spoken and written language and that the Greek of the Septuagint and the New Testament should be considered as the vernacular Hellenistic Greek. Meanwhile, Deissmann (1991:55-6) thinks of the language changes in the Greek of the Septuagint and the New Testament as the influences of translation Greek at the syntactic level. He assumes that Semitic syntax (i.e. Hebrew and Aramaic) interferes with the Greek of the Septuagint and the New Testament in the process of translation respectively. In the case of the Greek syntax of the Septuagint, he (1903:69) claims that “the fact that the Alexandrian Old Testament is a translation is of fundamental importance for an all-round criticism of its syntax.” He (1903:67) proposes that the translators of the Septuagint “frequently stumbled at the syntax of the Hebrew text.” Surprisingly, he suggests that the Greek of the Septuagint might aptly be called “JudaeoGreek,”20 although he (1991:56) prefers to call it “translators’ Greek.” As to New Testament syntax, he regards some syntactic interference as an intentional imitation of the Septuagint (i.e. Septuagintalisms).21 Deissmann (1908:49) also regards Jesus’s sayings as examples of “translators’ Greek” (i.e. Semitisms).22 Whether he calls it “Judaeo Greek” or “translators’ Greek,” he clearly admits that Semitic languages interfere with the Greek of the Septuagint and the New Testament in translation at the level of syntax.23 His acceptance of Semitic Greek at the syntactic 20

21

22

23

He (1903:69) mentions that “it is, of course, possible that the continually repeated reading of the written Judaeo-Greek [of Septuagint] may have operated upon and transformed the ‘feeling for language’ of the later Jews and of the early Christians“ [emphasis added]. He (1903:70; cf. 76) proposes that “the occasional literary impressiveness is an intentional imitation of the austere and unfamiliar solemnity of that mode of speech which was deemed to be biblical“ [emphasis in the original]. He (1908:49) suggests that “the sayings of Jesus in the Synoptic Gospels and perhaps more of the New Testament must be counted with the examples of translators’ Greek.” While Deissmann (1903:74) considers that the translators were bilinguals who were well enough trained to compare the Hebrew version with the corresponding Greek and to recognize the semantic difference between the two versions due to their syntactic and semantic difference, he (1929:88-9) suggests six reasons for the Semitisms and Septuagintalisms of New Testament Greek. (i) Native Jews, for the most part, wrote the New Testament. (ii) The sayings of Jesus were translated from Aramaic into Greek. (iii) Syntactic changes which are not found in the most vulgar papyri (e.g. examples of word order) should be Semitic interference (1991:55-6). (iv) Some peculiarities of the New Testament resulted from the intentional imitation of the usages of the Septuagint (e.g. for endowing Jesus’ sayings with authority from God; 1903:76). (v) Septuagintalisms gradually became common usage especially where the Septuagint was read and heard (1903:69; 1991:57). (vi) He (1991:57) mentions that “The so-called Jewish Greek was … merely caused … by a wrong method of transla-

228

Syntax

level is evidently presented in his remark that New Testament Greek at the syntactical level “contains, naturally, a number of Semitisms. I have never denied that” (1929:89; cf. 1927:69 n.1).24 Through extensive investigation Deissmann convincingly proposed the vernacular language theory at the lexical-semantic level, which makes a trenchant, polemical attack on Jewish Greek theory. He still leaves a backdoor open, however, when he argues that the syntactic changes in the Septuagint and the New Testament (i.e. Semitisms and Septuagintalisms) are caused by external interference from other languages, not by internal change within Hellenistic Greek itself. 25 His

24

25

tion.” Ironically, this has been regarded as one of the definitions of a Semitism. Torrey assumes that Aramaisms are due to mistranslation (§1.4.1.1). Burney (1922:19) also states that “What is needed to substantiate the theory of an Aramaic original for Mark is some cogent evidence of mistranslation.” For a detailed reference of Semitisms as mistranslation, see Porter 2000a:93 n.78. This shows that Deissmann was confused and that his argument is inconsistent. He (1929:89) adds, “What I do deny is this: that the number of these Semitisms is enough to disprove the general view that New Testament Greek is the colloquial Greek of the period, and to revive the view that it is Jewish Greek, and to call it Greek-Yiddish.” Similarly, Moulton (1919b:127), one of founders of the Hellenistic Greek theory, also submits that “What I have just been describing goes to show that we must expect to find our lights upon Biblical Greek, not as hitherto mainly in the narrow circle of Jewish language and ideas, but in the immense field of Greek as spoken and written throughout the Gentile world. I do not, of course, mean to say that Semitic influences are not to be found in the Greek of the New Testament. But we are able to delimit them much more closely and hold the balance more evenly between the Greek and the Semitic.” Noteworthy is Thackeray’s argument: Accepting Deissmann’s argument, he goes a step further. As for the nature of Hellenistic Greek, he (1909:26) argues that koinh, Greek “had little or no dialectic differentiation and was proof against the intrusion of foreign elements to any considerable extent …” Furthermore, in terms of the Semitisms of the Septuagint, he (1909:27) states that “the extent of Semitic influence on the Greek language appears to have been limited to a small vocabulary of words expressing peculiarly Semitic ideas or institutions. The influence of Semitisms on the syntax of the Jewish section of the Greek-speaking world was probably almost as inappreciable as its syntactical influence on the koinh, as a whole, an influence which may be rated at zero.” It seems that his position is more Deissmanian than Deissmann’s own. However, his view of syntactic change is still based on the same concept of the categoricity theory. Following Thackeray, Horrocks (1997:57) mentions that Hellenistic Greek’s “general grammatical and lexical make-up is that of the ordinary, everyday written Greek of the times. It therefore constitutes an important source of information for the development of the language in the Hellenistic period, with the translation of the Pentateuch, for example, reflecting a very natural contemporary Koine.” Although Horrocks (1997:92) also admits Septuagintalisms and Semitisms of New Testament Greek, he (1997:95) concludes that “this [New Testament Greek] is in fact good, basic Koine Greek of its time.”

Contact-Induced Syntactic Change Hypothesis

229

acceptance of Semitisms and Septuagintalisms at the syntactic level undermines the consistency of his vernacular language theory. His acceptance of Semitisms and Septuagintalisms at the syntactic level is attributed to his view of language change based on the same concept of the categoricity theory. His argument can be explained in Saussurean terms; langue (the language of the translators and writers as Hellenistic Greek) does not change, whereas parole (the language of the Septuagint and the New Testament as a transitory translation language) changes.26 Still, he has the same assumption that parole cannot change langue, as de Saussure and later, Chomsky in general linguistics assumed. Although he (1903:69) argues that the Septuagint did not influence the readers of the next period, he (1903:70; cf. 76) admits that there are Septuagintalisms in the New Testament.27 In this respect, he seems to accept Semitisms and Septuagintalisms as parole (§6.1.5). However, it is regrettable that Deissmann does not view Hellenistic Greek (including New Testament writers’ language) as part of the concept of parole. When he deals with the syntax of the Septuagint and the New Testament, he should have taken the variation position rather than the categoricity position because syntactic changes in New Testament Greek are syntactic polysemies within Hellenistic Greek itself rather than abnormal linguistic phenomena (§6.3). These weak points also recur in other scholars’ later publications whether these arguments have been supported or not. Deissmann’s inconsistency in his view of the vernacular language theory at the lexical-semantic level and his acceptance of Semitisms and Septuagintalisms at the syntactic level have evoked four major reactions with regard to syntactic change. Above all, two different positions of linguistic relationship between syntactic changes will be discussed (§6.1.2). The reintroduction of the Jewish Greek theory as the basis of 26 27

Silva interprets Deissmann’s concepts as langue and parole, and Porter regards them as a language and a dialect; for more discussion, see §6.1.4. He (1903:69) suggests that “it is, of course, possible that the continually repeated reading of the written Judaeo-Greek may have operated upon and transformed the ‘feeling for language’ of the later Jews and of the early Christians.” In order to escape from this discrepancy, he tries to distinguish a written Greek from a spoken Greek. He (1903:76; 1991:56-7) mentions that the so-called Jewish Greek [Greek of the Septuagint and the New Testament] was not a spoken language Paul or the authors of the Gospels used. Instead, they spoke Hellenistic Greek. Moreover, he (1991:56) says that “the translator’s Greek is an artificial, paper Greek, and not a spoken Greek; its numerous, specifically syntactical, Semitisms are occasional.” However, what is important is that Semitisms or Septuagintalisms with which scholars concern themselves are related to the extant written Greek, not to the Greek language which the authors spoke.

230

Syntax

the study of the bilingualism of first-century Palestine and the Roman Near East (i.e. bilingual Jewish Greek theory) will also be considered (§6.1.3). The third response to Deissmann, the Hellenistic Greek theory to which scholars applied linguistic theories, will then be presented (§6.1.4). Last, Semitisms and Septuagintalisms at the syntactic level will be investigated (§6.1.5).

6.1.2 Linguistic Relatedness between Syntactic Changes Deissmann’s argument that the syntax of the New Testament is similar to that of the papyri in Egypt and inscriptions in Asia Minor stimulated scholars to investigate similar linguistic phenomena found in other regions. The similarity of syntactic changes in the biblical languages (i.e. the Septuagint and the New Testament) and the papyri due to bilingualism has been explained by two major suppositions. Some scholars have maintained that the similarity is due to the same influence of the common language (i.e. Aramaic).28 This means that the syntactic anomalies found in the Septuagint and in the New Testament should be considered as Aramaic interference. As with Jewish Diaspora communities, Aramaic had been used from Syria to Egypt (e.g. Alexandria) until it was substituted for Greek in the same regions. The language contact of the two languages has led to the assumption that Aramaic influenced the Greek used in Palestine as well as in Jewish Diaspora. As a result, similar syntactic changes are shown in the papyri as well as in the Septuagint and the New Testament. Consequently, the language of the New Testament was influenced by Aramaisms, which was the language of the papyri. Other scholars have argued that different languages (Coptic or Semitic), due to bilingualism, intervened in the Greek of the papyri or the New Testament respectively.29 The Greek of the papyri was influenced by their vernacular language, Egyptian Coptic (i.e. Copticisms), whereas the Greek of the Septuagint and the New Testament was influenced by Hebrew or Aramaic (i.e. Semitisms). Admittedly, Coptic is similar to Semitic at the syntactic level.30 Despite the different syntactic 28

29

30

Redpath 1903:10-1; Moulton 1906:242; Swete 1909:CXXV n.1; Dalman 1909:17; James 1920:57-75, esp. 70-1; Ottley 1920:164-5; Burney 1922:4-7; Robertson 1934:91; Turner 1965:184 (§1.3.2.4). Lefort 1928:152-60; Vergote 1938:3.1353-60; Gignac 1976:1.46-8; 1970; 1985:154-65; Marcos 2000:15. The history of research on the grammar of the Egyptian papyri is well summarized by Gignac (1976:41-3). Gignac (1985:165) suggests that syntactical anomalies due to bilingual interference “illustrate the rich possibilities a bilingual approach to Koine Greek offers, not only

Contact-Induced Syntactic Change Hypothesis

231

interference, the close relationship between Hamitic syntax and Semitic syntax resulted in similar linguistic changes. Hence, the syntax of the papyri is similar to that of the New Testament on account of respective different interferences. Scholars have noted that similar syntactic changes are commonly found in different regions (i.e. Jewish Diaspora in the Roman Near East) and in different language families (i.e. Hamitic). It is unfortunate, however, that both groups still understand syntactic changes as syntactic interference on the basis of the categoricity theory. 31 Syntactic changes found in the Septuagint, the New Testament, or the papyri are considered to be anomalies caused by interference of other linguistic syntax, which means that these languages were peculiar Greek, not Hellenistic Greek. Decisions on syntactic interference due to contactinduced change should not be made before the possibility of syntactic change within Hellenistic Greek itself is conclusively excluded. Furthermore, they should have considered that the similar patterns of the syntactic changes called Copticisms or Semitisms are found not only in the two languages (Coptic and Aramaic) but also in other ancient and modern languages (§6.3). Accordingly, they disregarded the other possibility of syntactic changes due to internal-induced syntactic change.

6.1.3 Bilingual Jewish Greek Theory Although Deissmann’s strong contention seemed to bring the Jewish Greek hypothesis to an end, some scholars reintroduced the Jewish Greek hypothesis. The reason why the theory revived after Deissmann can be summarized in three points in relation to bilingualism. (i) As mentioned before (§6.1.1), Deissmann himself has called New Testament Greek “Judaeo Greek” or “translation Greek.” His ambiguous stance still envisioned the possibility that the syntax of the Septuagint and the New Testament was influenced by Semitisms and Septuagintalisms (§6.1.1). (ii) Some scholars consider that the language of the Egyptian papyri was influenced by bilingual Alexandrian Jews (cf. §6.1.2). This view holds that Semitisms interfered with the syntax of the New Testament (e.g. Turner 1965:183-4). (iii) The study of bilingualism confirmed the Jewish Greek theory because language contact due to

31

in Egypt, but wherever this common language of everyday commerce was adopted by speakers of other linguistic affinities.” Furthermore, the argument of similarity between New Testament Greek and Egyptian Coptic has been criticized as an anachronism because Coptic was usually used later than the Greek papyri in Egypt.

232

Syntax

bilingualism may cause syntactic change. This view is adduced from recent studies and excavations. For these reasons, the bilingual Jewish Greek hypothesis is different, to some extent, from the classical Jewish Greek hypothesis.32 Gehman considers the Alexandrian Jewish community to be bilingual. He (1951:90) refers to the language of the Septuagint as Jewish Greek at the transitional stage from an older generation who spoke Aramaic as their matrix language to a younger generation who spoke Greek as their matrix language. He (1951:90) states that “The language of the LXX certainly would have caused trouble to a Greek who was not acquainted with the psychology of the Hebrew language, its idioms, and its construction.” Instead, he (1951:90) suggests that “its language made sense to Greek-speaking Jews.” Examining syntax (e.g. conjunctions, prepositions, and pronouns) as well as the vocabulary of the Septuagint, he (1953:148, cf. 1951:81) concludes that the Greek of the Septuagint was influenced by Hebraisms so that it was different from Hellenistic Greek.33 Taking up Gehman’s Jewish Greek hypothesis of the Septuagint, Turner extends the view to the language of the New Testament. In terms of the relationship between the Greek of the New Testament and that of Egyptian papyri, he concludes that the Greek of those papyri was influenced by Semitisms used by Alexandrian Jews (§6.1.2). Turner also seriously takes bilingualism into account. According to Turner, the linguistic milieu of first-century Palestine was quadrilingual, and Jesus could have spoken four languages (§3.2.4). He assumes that the language contact due to bilingualism in Aramaic and Greek led into Semitized Greek (i.e. Jewish Greek). The “Greek Jesus” who was bilingual (i.e. Greek and Aramaic) spoke Jewish Greek (1965:182). Concern32

33

Vorster assumes that the author of the Gospel of Mark and the earliest Jewish Christians are bilinguals and admits “Jewish Greek theory.” He (1999:36) proposes that “Semitic interference in the Gospel of Mark was not caused by translation of original Aramaic documents (whether his Gospel or assumed sources which our author might have used), by translation Greek or the influence of the Septuagint, or the fact that he used ‘Jewish Greek.’ … Both the fact that Greek was the second language of many early Christians, including our author, and the nature of ‘New Testament Greek’ make it necessary that we study the Greek of a document such as the Gospel of Mark from the perspective of bilingualism.” Silva (1980:210-12) rebuts Gehman’s proposal in four points. (i) Although Gehman hesitates to use Jewish Greek jargon, he still uses it. (ii) The Septuagintal Greek itself is not a well-defined entity. (iii) Gehman’s argument that it is difficult for someone who is unfamiliar with Hebrew or Aramaic to understand idioms of the Septuagint is inconsistent because he suggests that the idioms can be understood by Hellenistic Jews who do not know those languages. (iv) It is not proven that a transitional period between generations can cause Semitic interference.

Contact-Induced Syntactic Change Hypothesis

233

ing Jewish Greek, he (1965:183) explains, “It was biblical Greek, of a kind not very different from the Septuagint–a branch of the Koine, but very different from what we read in the Egyptian rubbish heaps or on the papyrus of more literate people.” Furthermore, complaining about Gehman’s argument that the Jewish Greek of the Septuagint is a transitional language, 34 Turner (1965:183) proposes that “Biblical Greek was…the normal language of Jesus, rather a separate dialect of Greek than a form of the Koine, and distinguishable as something parallel to classical, Hellenistic, Koine and Imperial Greek.” Moreover, New Testament Greek is different from Hellenistic Greek but shared by the writers of the Septuagint and pseudepigraphic and apocalyptic works (Turner 1963:4-5; 1965:182-5). 35 Consequently, Turner concludes that Jewish Greek is “a unique language with a unity and character of its own” (1963:4). The bilingual Jewish Greek theorists have a strong point in that they apply the concept of bilingualism to analysis of the language of the Septuagint and the New Testament, because language contact may cause syntactic change. However, some weaknesses have been pointed out. First, the proponents seem to exaggerate the effect of language contact because bilingualism does not always generate “a unique language” colored with syntactic interference. In this respect, it seems that Gehman’s transitional Jewish Greek is closer to pidgin, and Turner’s unique Jewish Greek is closer to Creole. Second, it seems that Turner’s concept of Jewish Greek based on “frequency” runs to excess. Turner (1955b:208-10; 1955c:252-3) contends that if constructions in the New Testament are repeated more frequently than in papyri, they can be manifestations of biblical language even though the constructions occur in other literatures. However, Turner’s frequency resulted from the collection from various sources, as McKnight (1966:39-41) criticizes.36 Third, their definitions of a language are unacceptable. Turner seems to consider “style” or “dialect” (i.e. the New Testament and the Septuagint) as a distinct language. However, “style” or “dialect” cannot be regarded as “a language.” 37 Fourth, their concepts of Jewish/Biblical 34 35 36

37

Turner (1965:183) argues that “I do not think it was merely transitional. Certainly it was not artificial.” He (1954-55:220) considers the language of Testament of Abraham as “an excellent example of the ‘Jewish’ Greek language of the early Christian centuries.” McKnight (1966:40) also argues that Turner’s “statistics show that only Matthew, Mark, and Luke-Acts are comparable. Paul and John show the normal usage of the papyri and the other books are too evenly balanced to allow any conclusion.” Concerning the act of making an ambiguous distinction between dialect and language, Porter (1989:595) points out that “Turner, Gehman, and their followers, seem

234

Syntax

Greek are overly delimited by ethnic or religious categories and terminology. Why should Semitized Greek always be characterized only Jewish or biblical language? It must be emphasized that the linguistic milieus of first-century Palestine and the Roman Near East was bilingualism in Aramaic and Greek (chapters 3 and 4). If Aramaic influenced the Greek of Jews in the first-century and the New Testament, then most Aramaic language families likely interfered with Greek in the Roman Near East. We cannot state with certainty that only Jewish Greek is distinct from Hellenistic Greek in the Roman Near East, where there were bilingual regions of Aramaic and Greek.38 Fifth, it seems that the proponents of the bilingual Jewish Greek theory still consider only one possibility of contact-induced language shift hypothesis. Language change in the Septuagint and the New Testament resulted from the syntactic interference of language contact, as others do. Accordingly, although both Gehman and Turner tried to differentiate their positions from classical Jewish Greek theory,39 the view that syntactic change is regarded as syntactic interference cannot avoid their returning to a position predating Deissmann.

6.1.4 Hellenistic Greek Theory The third response to Deissmann is related to the exponents of the Hellenistic Greek theory. In fact, Deissmann (1991:57) requires a further comparative study between New Testament Greek and Hellenistic Greek to prove a more syntactic relationship between translation Greek and Hellenistic Greek: A study of Semitisms from this point of view, for example in the early Christian texts, is a pressing necessity. A comparative consideration of examples of Hellenistic colloquial language would not infrequently result in the realization that an apparent Semitism must rather be defined as characteristic of popular Greek.

38

39

to be confused about the terms ‘dialect’ and ‘language,’ … so that for them, if there are irregularities between the NT and papyri, the result must be a new ‘language.’” McKnight (1966:40) proposes that Turner “assumes that any influence from Semitic sources supports a religious language unrelated to wider influences. But were there not non-religious Jews who spoke Greek? Was the use of Aramaic necessarily limited to religious gatherings?” Gehman (1951:81) mentions, “It certainly would be too bold to speak of a ‘JewishGreek jargon,’ and yet we can hardly avoid speaking of a Jewish-Greek….” Also Turner (1955b:208) states that “This is not the position which used to be adopted when the papyri were first investigated by such scholars as Deissmann and Moulton, but neither is it to return to the theological assumptions implicit in a ‘Holy Ghost language.’”

Contact-Induced Syntactic Change Hypothesis

235

In reply to this, scholars have applied linguistics to explanations of the relationship between New Testament Greek and Hellenistic Greek. Here, two representative scholars (i.e. Silva and Porter) will be presented. 40 Silva (1980) points out that the entangled knot of the languages of the Septuagint and the New Testament in relation to especially Deissmann’s argument results from biblical scholars’ misunderstanding of a distinction between langue and parole. Silva makes a comparison between Deissmann and Turner. He (1980:217) mentions that “Deissmann, concerned with grammatical rules (langue), insisted rightly that New Testament Greek cannot be isolated from the Hellenistic form.” He says, conversely, “Turner, who has devoted his efforts to syntactical phenomena—an area of grammar that constantly ‘infringes’ on stylistics (parole)—sees an undeniable distinctiveness in the Biblical language.” 41 In this respect, Silva (1980:216) makes a distinction be40

41

This view has been supported by many scholars. Among them, Thumb (1901; 1914) extends Deissmann’s view to Hellenistic Greek syntax. Mentioning his purpose for writing A Grammar of New Testament Greek, Moulton (1919:4) starts with the following remark: “Professor Deissmann dealt but briefly with the grammatical features of this newly-discovered Greek; but no one charged with the duty of editing a Grammar of NT Greek could read his work without seeing that a systematic grammatical study in this field was the indispensable equipment for such a task.” Also Moulton and Milligan (1914-29) suggest that the grammar of the Egyptian papyri is significant for elucidating the syntax of the New Testament (cf. Milligan 1922). Argyle is also one of the proponents of the Hellenistic Greek theory. He (1955, 1956, 1969) asserts that, although Black and Turner assume that the repetition of a preposition before every noun of a series which it governs is a Semitic usage, he resists the alleged Semitisms because the usages are also found in classical Greek as well as in Egyptian papyri. However, although he (1956:247) proposes that New Testament Greek “is also to a considerable extent colloquial Greek,” he also admits Semitisms and Septuagintalisms of the New Testament as external influence of Semitic languages. It is of value to pay attention to Mealand’s argument. Comparing so-called Semitisms with Hellenistic writers including Polybius, he (1991:51-2) mentions, “I could not altogether refuse the term ‘coincidental Semitisms,’ but I would be inclined to retort that the phrase ‘Hellenism’ might often be just as appropriate to describe the same phenomena. Where such a word or phrase can be found not only in Hellenistic but also in Classical Greek I think it is most inadvisable to describe it as any kind of Semitism. It is simply normal Greek and it is most misleading to describe it as anything else.” On the basis of the concept of variation theory, Mealand (1991:52) persuasively concludes that “The Greek language developed continuously from Homer to the present day and Classical, Hellenistic, and Byzantine Greek are stages in that development. There is also variety due to dialect and locality, and to genre.” However, it is regrettable that he did not develop this further and did not consult other sciences (i.e. cognitive linguistics). Similarly, he (1980:209) mentions that “whereas Thumb’s views referred to langue, they appeared absurd to Vergote only because the latter was concerned with parole.” However, Silva fails to recognize that Deissmann himself has already made a

236

Syntax

tween Hellenistic Greek as langue and biblical Greek as parole. Consequently, he (1980:216-9) concludes that the Semitisms of the New Testament belong to parole and that Semitized Greek is a style of Hellenistic Greek. Porter has applied linguistic methods to analysis of New Testament Greek continuously and extensively (cf. §1.4.2.5). Accepting Deissmann’s vernacular Greek theory as a principle, he criticizes Deissmann’s vague position of New Testament syntax. He (1989:593) says: “Much of the difficulty stems from stereotypes of the DeissmannMoulton position.” He (1989:594) continues that Deissmann’s position “that the LXX is written in ‘Semitic-Greek’ and that the ‘Hebraisms’ (his term) possibly influenced the language of early Christianity…is often overlooked.”42 However, he still accepts that Semitisms interfere with New Testament Greek. He (1989:596-7) contends that “Virtually every scholar of Hellenistic Greek recognizes the presence of Semitisms in the New Testament, but they are seen as not affecting the essential grammatical structure or code of the language….” In other words, he (1989:597) interprets the relationship between Hellenistic Greek and New Testament Greek as “the distinction between code and text, or grammar and style,” as Silva (1980:216-9) has already emphasized in the distinction between langue and parole or grammar and style before. Consequently, he also concludes that New Testament Greek is a form of Hellenistic Greek with Semitic interference due to contact-induced syntactic change. It seems that both scholars’ views of New Testament Greek are also based on categoricity theory (i.e. the study of langue). Silva considers Hellenistic Greek to be langue, not parole, and Porter adds that New Testament Greek can be explained as code. As a result, they accept the Semitisms of the New Testament at the syntactic and semantic levels as abnormal Greek usages due to external interference.43 However, they create too sharp a distinction between langue and parole, as Deissmann did. The problem lies in the fact that if langue refers to a so-called abstract language system, can we really call Hellenistic Greek an abstract language system, and not a concrete substance of a particular period? In other words, is it plausible to consider that Hellenistic Greek (i.e.

42 43

distinction between the concept of langue and the concept of parole. Deissmann regards the Greek of the New Testament as langue and Semitisms and Septuagintalisms due to translation as parole at the syntactic level (§6.1.1). It is inappropriate when Jung (2004:19-20 n.48) mentions that Porter misunderstands Deissmann. Silva, moreover, is confused between borrowing and interference at the semantic level (§8.1.2).

Contact-Induced Syntactic Change Hypothesis

237

langue) did not change? They should have considered that every grammar is changing (i.e. parole) in the metaphorical or metonymic process of human cognition (i.e. langue) so that we cannot deny the interaction between langue and parole (§6.2.1.3). Furthermore, both of them assume that the linguistic milieu(s) of the first-century Palestine and Roman Near East was bilingualism. Given that this is the case, the bilingual linguistic situation of first-century Palestine and the Roman Near East should be investigated from the perspective of parole rather than langue (see the introduction to Part II).

6.1.5 Semitisms and Septuagintalisms Deissmann denied Semitisms and Septuagintalisms at the lexicalsemantic level but admitted them at the syntactic level, a distinction criticized by many scholars.44 This view has permitted biblical scholars to maintain that New Testament Greek is a Hellenistic Greek colored with Semitisms at the syntactic level. In fact, whatever syntactic changes of New Testament Greek were caused by Aramaic or Hebrew (§6.1.2), or whenever they can be explained by bilingual Jewish Greek theory or Hellenistic Greek theory (§§6.1.3, 6.1.4), the four theories accept Semitisms of the New Testament due to a contact-induced syntactic change. At the same time, Deissmann’s argument allowed biblical scholars to hold the view that Greek anomalies of the synoptic gospels and Acts can be ascribed to Septuagintalisms as a literary device. Cadbury (1920b:452) mentions that the Semitisms of Acts “can easily be explained as due to the extensive influence of the LXX.” Dibelius also regards some Semitisms as Septuagintalisms (§1.2.1). He (1934:35; 1935:74-75; 1949:25) suggests two possibilities. First, the writers try to imitate the style from the sacred narratives of the OT. Second, the writ-

44

Barr (1961:240 n.3; cf. 213-4, 239, 267) criticizes Deissmann, saying “Even Deissmann admitted the existence of Semitisms in the NT, though he tended to minimize them, and in this the trend of modern scholarship is contrary to his judgment.” Maloney also suggests that Deissmann considers that Semitisms interfered with the Greek of the New Testament when he (1981:8) mentions that Deissmann “never denied that some influence of the Semitic languages existed in New Testament Greek…” Marcos (2000:9) rightly questions Deissmann’s work, saying, “In spite of Deissmann’s results, biblical philologists continued to look for Semitisms in the New Testament. For even though the data from the papyri had been decisive in the area of the lexicon, the constructions that diverged from classical Greek were so important that the explanation of a few sporadic agreements with the Greek of the papyri was not explanation enough.”

238

Syntax

ers can be influenced by the diction of the Septuagint so that they unconsciously use the Septuagintalisms. The list of Septuagintalisms is well summarized by Fitzmyer (1981-5:114-6, 1998:114-6). As to Septuagintalisms as a literary device, special attention should be paid to Wifstrand. 45 He (2005:57) follows the Hellenistic Greek theory when he suggests that Hellenized Jews spoke ordinary Greek, not any special Jewish Greek.46 He also regards Septuagintalisms in the Gospel of Luke as an intentional literary device.47 That is, the alleged Semitisms in Luke are deliberately borrowed from the Septuagint in order to imitate the style of the Septuagint.48 Wifstrand (2005:31) concludes that the Semitisms in Luke “are wholly intentional, that they are directly patterned on the LXX and thus should not be regarded as Semitisms in any real sense, whether ‘primary’ or ‘secondary’.” Accordingly, Wifstrand develops Deissmann’s view of Septuagintalisms of the New Testament and argues that Luke uses Septuagintalisms as his literary device at the level of syntax.49 Whether the syntactic changes of the Septuagint and the New Testament are explained as Semitisms or Septuagintalisms,50 the two theo45

46

47

48

49

50

Loveday Alexander (2005:231-52) tries to apply diglossia theory to interpretation of Wifstrand’s theory. However, the model of diglossia does not fit into the linguistic milieu; for my criticism in detail, see chapter 2 n.21. Also, Aejmelaeus merits our attention in that he considers o[ti causale an intentional Septuagintalism (§6.3.2.2). Wifstrand (2005:57) insists that “we cannot discover any special Greek dialect spoken by the Hellenized Jews; in phonology, accidence, syntax, word formation and many significations of words their language was ordinary koine“ [emphasis in the original]. He makes a distinction between intentional and accidental Semitisms. Intentional Semitism (i.e. imitation theory) suggests three possibilities (cf. Wifstrand 2005:30-1): (i) Jesus may have echoed the usages of the Old Testament (e.g. the prefatory avmh,n– formula; see 8.2). (ii) The authors of the Gospels may have imitated the usages of the historical Jesus with reverence to show that Jesus’ words are authoritative from God. (iii) The authors of the gospels may have echoed the usages of the Old Testament to add authority to their narratives. He (2005:36) observes that “quite a number of other expressions within Luke’s narratives and speeches that agree closely with the LXX but are otherwise unusual in Greek are likewise deliberate and intended imitation.” As to the Lucan literary device at the lexical-semantic level, it will be stated that Luke uses Semitic spellings as his literary device at the level of Semantics (§8.4). It is noteworthy to mention Guenther. He (1992:441-62) holds the view that Palestinian coloring resulted from deliberate redaction. However, one of the biggest problems Septuagintalism has faced is the heterogeneity of the Septuagint. A heterogeneous character of the Septuagint has come into question in discussing Septuagintalisms of the synoptic gospels and Acts. Torrey (1912:276) complained that “many Old Testament scholars have failed to recognize the fact that these renderings are not all alike; and most of them still continue to op-

Contact-Induced Syntactic Change Hypothesis

239

ries are based on the common foundation of a contact-induced syntactic change. In other words, most scholars, including Deissmann and the four major groups as mentioned before, have considered that the syntactic changes resulted from external interference (i.e. Semitic languages) on the basis of the categoricity theory and the science of langue, and the study of language competence. Accordingly, the contactinduced syntactic change hypothesis, based on the categoricity hypothesis, makes it possible that Semitic colorings are more original Jesus and gospel tradition and closer to the historical Jesus. The problem, however, is that the syntactic changes, called Semitisms or Septuagintalisms, are also found in other Hellenistic Greek literature as well as in modern languages in spite of a lack of linguistic, geographical, ethnological, or political relation to the Semitic languages (§6.3). This universal occurrence raises another possibility that syntactic change could be caused by a human cognitive system because human cognition is universal. In this respect, special attention should be paid to cognitive linguistics. According to cognitive linguistics, syntactic variants due to syntactic changes can be considered to be syntactic polysemies caused by the grammaticalization process due to human cognition (§6.2.2.1). This means that syntactic changes of the Septuagint and the New Testament could occur due to internal-induced syntactic change.51

51

erate with ‘the LXX’ (meaning the text printed by Swete or Tischendorf) as though it were homogeneous, and to use it in Kings, or Ezekiel, or Koheleth, in exactly the same way in which they had used it in the Pentateuch.” Criticizing Gehman’s Jewish Greek theory, Silva (1980:210) says that the Septuagintal Greek is not a well defined entity because of “the wide divergences in translation technique – and therefore in the character of the resulting language – among the books of in the Greek OT.” Arguing that the Septuagint itself is heterogeneous in external and internal aspect, Tov (1988:169) mentions that “The heterogeneous character of the LXX derived not only from the aforementioned external conditions, but also from differences in translation techniques between the various books.” Also Wilcox (1994:366) suggests that “to speak of the style of the LXX is to be confusing, since the style varies considerably from book to book, and from section to section.” He continues, “The work is, after all, not that of one individual, but seems to have been done by a largish number of translators over a period of centuries.” Furthermore, the fact that many scholars have recently applied sociolinguistic methods to analyzing classical Greek is encouraging. Bubenik (1989) examines Hellenistic Greek from the perspective of sociolinguistics. Colvin (1999) applies sociolinguistic method to study dialect in the works of Aristophanes. Willi (2003) investigates the languages of Aristophanes on the basis of language variation. Also, see Adams, Janse, & Swain (2002) noted before.

240

Syntax

To date, some scholars have applied the grammaticalization theory to analyzing Semitic languages.52 On the other hand, many cognitive linguists have applied the grammaticalization theory to Greek. It is taken for granted that their arguments are not related to Semitic interference (i.e. Semitisms or Septuagintalisms) or New Testament studies. Unfortunately, biblical scholars have disregarded the grammaticalization theory of cognitive linguistics because they assume that syntactic changes are caused by contact-induced syntactic shift and because cognitive linguistics is a young science in a sense.53 Accordingly, cognitive linguistics will be introduced to address the intricate issues of the syntactic changes of the Septuagint and the New Testament from the perspective of internal-induced syntactic change hypothesis.

6.2 Internal-Induced Syntactic Change Hypothesis As mentioned before (§6.1), most New Testament scholars, on the basis of the categoricity theory, have presupposed that if a grammatical phenomenon of New Testament Greek is a solecism and the same linguistic phenomenon occurs in Semitic languages, the syntactic solecism is regarded as interference by the Semitic languages. This is a contactinduced syntactic change hypothesis. The linguistic view easily leads to the assumption that syntactic anomalies of New Testament Greek are Semitisms or Septuagintalisms, that is, an external interference due to language contact. However, if the common shift occurs in many languages as well as Semitic languages, the syntactic change should be considered as a cross-linguistic phenomenon rather than a peculiarity of biblical Greek. Interestingly, Deissmann (1908:127) mentions that “many phrases…which [classical Greek scholars and scholars of sacred

52

53

Rubin (2005), for instance, applies grammaticalization theory to dealing with grammars of Semitic language family. Also as to the grammaticalization of Old Akkadian, see Deutscher 2000 (cf. §6.3.1.1). Nevertheless, it is very welcomed that the section “Use of Cognitive Linguistics in Biblical Interpretation” of the SBL Annual Meeting has started from 2006, although it seems that it is concerned with a semantic level more than a syntactic level.

Internal-Induced Syntactic Change Hypothesis

241

Hebrew] triumphantly brand as Semiticisms, are not always Semiticisms, but often international vulgarisms” [emphasis added]. Why did the vulgarisms occur internationally? Why does a similar pattern of Greek syntactic change occur in a cross-linguistic way? Regarding syntactic change, cognitive grammar is in line with variation theory in that cognitive grammarians argue that grammar is not static but unceasingly changing in a unidirectional way. Bybee, Perkins, and Pagliuca (1994:4) assert that “language does not provide a static organization of meaning. Grammatical meaning is changing constantly.” This raises the possibility that the syntactic change in New Testament Greek (i.e. so-called Semitisms or Septuagintalisms) can be explained by internal change due to human cognition rather than by external interference of Semitic languages. Before applying cognitive grammar to the syntactic change of New Testament Greek, some preliminaries will be mentioned briefly because cognitive linguistics is little known to New Testament scholars. First of all, the difference in syntactic change between cognitive linguistics and other linguistics will be suggested (§6.2.1). Second, some basic concepts for understanding cognitive linguistics will be explained (§6.2.2). And last, the concept of grammaticalization theory, which is a representative theory of the cognitive grammatical approach to syntactic change, will be presented (§6.2.3).

6.2.1 Comparisons between Cognitive Linguistics and Other Linguistic Theories Cognitive linguistics understands a language in relation to human cognition in general. Although it includes various approaches, methodologies, and emphases, Taylor (2002:4) submits that it “aims for a cognitively plausible account of what it means to know a language, how languages are acquired, and how they are used.” The primary methodological differences between cognitive linguistics and non-cognitive linguistics can be explained in four points, which help biblical scholars to understand where cognitive linguistics is placed.

6.2.1.1 Theoretical and Empirical While Saussurean structuralism and Chomskyan generativism hold a theoretical approach to language, cognitive linguistics takes an empirical approach (Geeraerts 2006b:25-31). In other words, cognitive linguistics is grounded in a “usage-based” conception of language (cf.

242

Syntax

Tomasello 2000:61-82; Taylor 2002:27-8). Dirven (2005:39-40) suggests that a usage-based grammar refers to a “bottom-up” attempt to describe language acquisition and grammar, which is in contrast with Chomsky’s generative grammar which supports “top-down” rule deduction processes. A usage-based model takes serious account of the “actual use of the linguistic system and a speaker’s knowledge of this use.” In this respect, the cognitive linguistics approach shares common concerns with the sociolinguistic approach (Dirven 2005:39-45).54 Accordingly, cognitive linguistics is concerned with parole rather than langue, variation theory rather than categoricity theory, language performance rather than language competence, and language use rather than the language system.

6.2.1.2 Diachronic and Synchronic While historical linguistics in the nineteenth century takes a diachronic approach, two major linguistic approaches in the twentieth century, namely Saussurean structuralism and Chomskyan generativism, have focused on a synchronic approach (see the introduction to chapter 6). Also, sociolinguistics in the second half of the twentieth century considers only the synchronic dimension. However, cognitive linguistics takes both the diachronic and the synchronic dimension into account. According to Heine, Claudi, and Hünnemeyer (1991:12), “Dissatisfaction with the restrictions of structuralism and other rigidly synchronic models led to an awareness of the need.” The consideration of both approaches is shown by Givón’s well-known slogan (1971:413): “today’s morphology is yesterday’s syntax.” Layering theory also indicates that the grammaticalization theory takes account of both diachronic and synchronic dimensions (§6.2.2.3).

6.2.1.3 Langue and Parole De Saussure makes a sharp distinction between langue (language system) and parole (language use) and bestows priority to langue. Chomsky tries to bridge the gap between the system and its use and between the community and the individual, where de Saussure left off. Although he argues that an individual’s linguistic knowledge is based on a language 54

Also Geeraerts (2006b:23-5) suggests five characteristics of empirical research such as a data-driven approach, quantitative methods, formulation of hypotheses, operationalization of hypotheses, and the empirical circle on which cognitive linguistics are based.

Internal-Induced Syntactic Change Hypothesis

243

system,55 he could not overcome his linguistic dichotomy (i.e. competence and performance) because he also excluded language performance from the object of linguistic studies (Introduction to Part II; cf. Geeraerts 2006a:6, 26). On the other hand, sociolinguistics based on variation theory has placed emphasis on parole. Variationists explain that the study of parole offers an understanding of the other side of a language. Languages change. As mentioned before (§6.1.4), when New Testament scholars deal with the syntax of New Testament Greek, they make a sharp distinction between langue and parole (de Saussure), language competence and performance (Chomsky), text and code (Porter), or grammar and style (Sylva). These distinctions have thrown scholars into confusion. However, cognitive linguistics can bridge the gap. Traugott and König (1991:1.189) assert that “The study of grammaticalization challenges the concept of a sharp divide between langue and parole, and focuses on the interaction of the two.” Geeraerts (2006a:27) also suggests that “the link between linguistic performance and grammar is reestablished by the view that language is usage-based, i.e. that there is a dialectic relationship between langue and parole.”56 In other words, cognitive grammarians consider that every grammar is changing (i.e. parole) in the metaphorical or metonymic process of human cognition (i.e. langue), which means that the syntactic changes occur in a similar pattern. In this respect, while sociolinguistics helps us to understand the syntactic variants of New Testament Greek, cognitive linguistics explains why the syntactic variants occur.

55

56

Regarding the Chomskyan concept, Geeraerts (2006a:26) explains that “competence interiorizes the notion of the linguistic system.” He goes on to say that “competence is the internal grammar of the language user, the knowledge that the language user has of the linguistic system and that he or she puts to use in actual performance.” Croft (2003:287-8) also states that “The typological approach to the relationship between language use, language change and the linguistic system implies a different conception of the nature of the linguistic system than either the Saussurean structuralist view or the generative approach. This view of the linguistic system accounts for the nature of the ‘biological’ explanation for language in the typological approach, in contrast to that of generative grammar. In generative grammar, the biological explanation is a literal one: language universals are ultimately innate properties of human beings. In the typological approach, the biological explanation is not literal. It is an evolutionary explanation, but of the evolution of language structures in language use, not of biological genetic evolution of the human language capacity.”

244

Syntax

6.2.1.4 Significance of Semantics One of the major differences between cognitive linguistics and noncognitive linguistics is where semantics is placed in their linguistic theories. Traditionally, semantics has been neglected in favor of syntax in linguistic theories. Geeraerts (2006a:3) remarks that whereas generative linguistics is mainly concerned with syntax, cognitive linguistics primarily focuses on the meaning of language.57 Concerning the relationship between semantics and syntax, Taylor (2002:29) mentions that “syntactic (and morphological) facts of a language will be motivated by semantic aspects and that they can be exhaustively described by means of symbolic structures“ [emphasis in the original]. Accordingly, semantics plays a significant role in cognitive linguistics.

6.2.2 Some Basic Concepts Cognitive linguistics is quite a new science that is little known to biblical scholars. Brief explanations for some basic concepts related to cognitive linguistics will be mentioned.

6.2.2.1 Grammaticalization The fact that the syntactic changes similar to so-called Semitisms and Septuagintalisms of New Testament Greek occur in a cross-linguistic way pushes us to examine the possibility that the syntactic changes are not Semitic interference due to contact-induced change but to internalinduced change in Hellenistic Greek itself. Internal-induced syntactic change includes some cognitive linguistic processes such as “grammaticalization” 58 and “lexicalization.” 59 If some syntactic changes in 57

58

59

As a matter of fact, Geeraerts (2006a:3) distinguishes “uncapitalized cognitive” from “Cognitive” because generative grammar (i.e. Chomskyan grammar) “is definitely also a ‘cognitive’ conception of language.” This is also called “grammaticization” (Givón 1975:49) or “syntacticization” (Givón 1979:208-10). For the introductory concept of “grammaticalization theory” in detail, see Sweetser 1990; Heine, Claudi, and Hünnemeyer 1991; Traugott & Heine 1991; Hopper & Traugott 2003; For the historical research of the grammaticalization theory in detail, see Heine, Claudi, & Hünnemeyer 1991:5-22. According to Brinton & Traugott (2005:18), the term, “lexicalization” has been used in two ways. Synchronically, it refers to the “coding of conceptual categories.” On the other hand, diachronically it has been used for the “adoption into the lexicon” or as “falling outside the productive rules of grammar.” For a discussion of lexicalization in detail, see Brinton and Traugott 2005. For brief explanation of grammaticalization and lexicalization, see Traugott 1999.

Internal-Induced Syntactic Change Hypothesis

245

New Testament Greek (i.e. so-called Semitisms and Septuagintalisms) observe the typical patterns of grammaticalization processes, the grammaticalization theory should be applied to the analysis of New Testament Greek. Above all, grammaticalization is usually defined as: The process whereby lexical items and constructions come in certain linguistic contexts to serve grammatical functions, and, once grammaticalized, continue to develop new grammatical functions (Hopper & Traugott 2003:xv).

Abraham (1991:337) explains that the process of grammaticalization should involve at least one out of seven linguistic phenomena: (i) the increase of the morphosyntactic range of a morpheme, (ii) a shift from lexical to grammatical (or from less grammatical to more grammatical) status (e.g. from a derivative formant to an inflectional one), (iii) the loss of semantic complexity, (iv) the loss of pragmatic significance, (v) the loss of syntactic freedom, (vi) the loss of phonetic substance, and (vii) increased constrainedness by grammatical rules. Simply speaking, grammaticalization means that a lexical item is changed into a grammatical item or a less grammatical item into a more grammatical item as a derivative gram60 is changed into an inflectional gram (cf. KuryÙowicz 1975:52).

6.2.2.2 Unidirectionality Hypothesis It has been known that when a syntax changes, the shift has a directionality. Bybee, Perkins, and Pagliuca (1994:13) write that “The crosslinguistic consistency of such results encourages us to see the creation of grammatical material as evolution of substance from the more specific to the more general and abstract.” Regarding the directionality of language change, Givón (1979:209) shows a grammaticalization schema like this: discourse > syntax > morphology > morphophonemics > zero. This can be paraphrased into his famous slogan, “Today’s syntax is yesterday’s pragmatic discourse,” as he (1971:413) asserted just before he stated that “today’s morphology is yesterday’s syntax.” The directionality of grammaticalization is unilinear. Heine and Kuteva (2005:8) argue that it is common in many languages that “to go” is grammaticalized into a tense marker FUTURITY. However, to date, the case that a future tense marker is changed into the lexical item “to

60

A “gram” refers to a grammatical morpheme (Bybee, Perkins and Pagliuca 1994:2).

246

Syntax

go” is not found in any language.61 Heine and Kuteva (2005:17) conclude that “more recent research has shown that there are some examples contradicting the unidirectionality hypothesis; however…such examples are few, accounting for less than one tenth of all cases of grammatical change.” Thus, the grammaticalization process is unidirectional.

6.2.2.3 Five Principles of Grammaticalization Hopper (1991:22-31) proposes five principles of grammaticalization which are suggestive of an understanding of syntactic change in New Testament Greek. He (1991:22) summarizes them as: (1) Five Principles of Grammaticalization (a) Layering. Within a broad functional domain, new layers are continually emerging. As this happens, the older layers are not necessarily discarded, but may remain to coexist with and interact with the newer layers.62 (b) Divergence. When a lexical form undergoes grammaticization to a clitic or affix, the original lexical form may remain as an autonomous element and undergo the same changes as ordinary lexical items. (c) Specialization. Within a functional domain, at one stage a variety of forms with different semantic nuances may be possible; as grammaticization takes place, this variety of formal choices narrows and the smaller number of forms selected assume more general grammatical meanings. (d) Persistence. When a form undergoes grammaticization from a lexical to a grammatical function, so long as it is grammatically viable, some traces of its original lexical meanings tend to adhere to it, and details of its lexical history may be reflected in constraints on its grammatical distribution.63 61

62

63

Heine and Kuteva (2005:108) observe that “Grammaticalization is essentially unidirectional: At least 90 percent of all instances of grammatical change can be assumed to be in accordance with principles of grammaticalization.” Hopper (2003:125) adds, “Layering is the synchronic result of successive grammaticalization of forms which contribute to the same domain.” In English, several grams to denote FUTURE coexist (e.g. “will”, “be going to”, “be + -ing”, and “be + to”). In this respect, these grams can be called “allograms.” Bybee, Perkins, and Pagliuca (1994:16) illustrate that “although will has largely displaced shall as a future, contexts remain in which only shall and not will may occur.” For instance, make a comparison of (a) with (b): (a) Shall I call you a cab? (b) Will I call you a cab?

Internal-Induced Syntactic Change Hypothesis

247

(e) De-categorialization. Forms undergoing grammaticization tend to lose or neutralize the morphological markers and syntactic privileges characteristic of the full categories Noun and Verb, and to assume attributes characteristic of secondary categories such as Adjective, Participle, Preposition, etc.64

6.2.2.4 Metaphor and Metonymy It is suggested that when a lexical source is grammaticalized, one finds a metaphorical and/or metonymic relation between the source item and the target item.65 Metaphor is generally defined as a cognitive mapping from the source domain66 to the target domain by means of the “similarity” of two concepts. Basic concepts like “love,” “anger,” “life,” “death,” and “time” are metaphorically mapped in mental spaces. One conceives of “love” as a physical force, a patient, madness, magic, or war (Lakoff & Johnson 1980:49). Also “ideas” are thought of as food, people, plants, products, commodities, resources, money, cutting instruments, or fashions (Lakoff and Johnson 1980:46-8). In this sense, Lakoff and Johnson (1980:3) advocate that “metaphor is pervasive in everyday life, not just in language but in thought and action.” Johnson (1987:137) proposes that metaphorical systems “constrain our understanding and generate definite patterns of inference.” On the other hand, metonymy is usually defined as “a cognitive process in which one conceptual entity, the vehicle, provides mental access to another conceptual entity, the target, within the same domain, or idealized cognitive model” (Kövecses 2002:145).67 In other words, metonymy refers 64

65

66

67

Traugott and König (1991:1.189) suggest that the study of grammaticalization “challenges the concept of categoriality, and takes as central the concept of a continuum of bondedness from independent units … to dependent units….” In relation to “decategorialization”, Muraoka (1973:205-19; esp. 206) rightly points out that biblical scholars are used to making a sharp division between grammatical categories (e.g. RESULT and PURPOSE). Howe (2006) applies conceptual metaphor of cognitive linguistics to the interpretation of moral meaning of 1 Peter. She explains some basic concepts of cognitive linguistics in detail. Croft (1993:339) mentions that the “domain” can be defined as “a semantic structure that functions as the base for at least one concept profile (typically, many profiles).” He (1993:340) prefers to use “domain matrix” because “a concept may presuppose several different domains.” The relationship between metaphor and metonymy is under debate. Criticizing a sharp distinction of metaphor and metonymy on the basis of traditional definitions of the two concepts, Croft (1993:335-70) supposes that the domain should be replaced by the domain matrix and that it occurs in a domain matrix. Other scholars

248

Syntax

to a cognitive mapping from a whole to a part in a single domain by means of the “contiguity” of two concepts. 68 Lakoff and Johnson (1980:39) also submit that “metonymic concepts structure not just our language but our thoughts, attitudes, and actions.” The concepts of metaphors and metonymies are built on our experience.69 The cognitive processes (i.e. metaphorical and metonymic mappings in mental space) are one of the major reasons for syntactic change. What is important here is that grammaticalization is construed as metaphors and metonymies in the process of human cognition. Furthermore, metaphorical and metonymic mechanisms are not random or arbitrary, but systematic (Lakoff and Johnson 1980:10-3, 35-40). In this respect, the syntactic change due to metaphorical and metonymic mappings can be explained in a systematic way. Consequently, the metaphorical and metonymic processes as systematic grounds of syntactic change can help explain the syntactic changes of New Testament Greek (so-called Semitisms or Septuagintalisms) as syntactic polysemies due to cognitive process.

6.2.2.5 Abstractness A grammaticalization process refers to “a shift from a more concrete to a more abstract domain” (Bybee, Perkins, and Pagliuca 1994:283). A grammaticalized item is more abstract than a non-grammatical item. Heine, Claudi, and Hünnemeyer (1991:42) suggest that “the distinction between nongrammatical and grammatical ‘meaning’ is frequently described as one of ‘concrete’ versus ‘abstract’ meaning.” As the grammaticalization of “to go” in “I am going to go to school” shows, the grammaticalized item GOING is more abstract than the lexical item GO (see (8)). Bybee and Pagliuca (1985:72) submit that “A concrete lexical item is recruited to express a more abstract concept.” They continue that “this emptying of lexical content is a prerequisite to grammaticiza-

68 69

(Geeraerts 2002:454-64; Ruiz de Mendoza and Díez:2002:489-532) also take seriously the interaction between metaphor and metonymy. For the recent debate on the metonymy, see Peirsman and Geeraerts 2006a:269-316, 2006b:327-35 and Croft 2006:31726. For instance, WALL STREET IS IN PANIC (Lakoff and Johnson 1980). WALL STREET refers to “the place for the institution.” Lakoff and Johnson 1980:39. It is noteworthy to mention that Lakoff and Johnson (1980:39-40) stress that “Cultural and religious symbolism are special cases of metonymy. […] DOVE FOR HOLY SPIRIT. As is typical with metonymies, this symbolism is not arbitrary. It is grounded in the conception of the dove in Western culture and the conception of the Holy Spirit in Christian theology.”

Internal-Induced Syntactic Change Hypothesis

249

tion because grammatical functions in themselves are necessarily abstract.” In this respect, Heine, Claudi, and Hünnemeyer (1991:48) suggest a chain of metaphorical abstraction such as (2): (2) PERSON > OBJECT > ACTIVITY > SPACE > TIME > QUALITY.

6.2.2.6 Grammatical Polysemy The layering theory, which means that older meanings coexist with newer meanings (§6.2.2.3a), implies that the process of grammaticalization results in lexical polysemies or grammatical polysemies. Syntactic arguments are frequently mixed with Semantic arguments because cognitive syntax starts from lexical semantics and depends on lexical semantics.70 The process of grammaticalization by metaphorical and/or metonymic mapping can result in a lexical-semantic or a grammatical shift. The former is called a lexical polysemy and the latter a grammatical polysemy. Consequently, according to cognitive linguistics, some cases of syntactic changes (§6.3) found in the synoptic gospels and Acts which have been known as Semitisms or Septuagintalisms can be called lexical polysemies or grammatical polysemies as the result of grammaticalization process.

6.2.3 Grammaticalization and Language Contact One might raise the possibility that language contact has an influence upon the grammaticalization process of New Testament Greek. Although grammaticalization is known as syntactic change due to an internally cognitive process within a language itself, it can be also influenced by other languages. Bisang (1998:13-58, here 13) remarks, “The sociolinguistic factor of language contact enforces processes of grammaticalization because it supports the exchange of mechanisms such as reanalysis and analogy which both can be determined by constructions.” Going a step further, Heine and Kuteva (2005:80; esp. 13-21) coined the term “contact-induced grammaticalization,” by which they mean “a grammaticalization process that is due to the influence of one language on another.” This implies that it cannot be excluded that Semitic languages could influence the grammaticalization process of syntactic changes in New Testament Greek. However, if syntactic changes which have been known as Semitisms or Septuagintalisms (i) show a typical paradigm of grammaticali70

Cognitive linguists consider Semantics a starting-point for syntax (§6.2.1.4).

250

Syntax

zation process found in many languages and (ii) their interpretations support grammaticalization process, it is highly possible that the syntactic changes are syntactic polysemies due to internal-induced syntactic change, although the usages are not found in other Hellenistic Greek literature except for the New Testament. Furthermore, if (i) syntactic changes which have been known as Semitisms or Septuagintalisms observe a typical paradigm of the grammaticalization process found in many languages, (ii) their interpretations support the grammaticalization process, or (iii) if the usages occur in other Hellenistic literature as well as the New Testament, it is suggests that the syntactic changes are syntactic polysemies which have resulted from internal-induced syntactic change(s).

6.3 Syntactic Change of New Testament Greek as Grammatical Polysemy The difference between Semitisms and Septuagintalisms depends on occurrence, whether the syntactic changes occur in New Testament literature (i.e. Semitisms) or in the Septuagint (i.e. Septuagintalisms). Given this, if the syntactic changes which were known as Semitisms or Septuagintalisms are also found in classical Greek or Hellenistic Greek, the syntactic changes should be considered a syntactic shift in Hellenistic Greek itself. Furthermore, if the same cases are found in other languages which are not linguistically, geographically, chronologically, or politically related, they should be thought of as common linguistic phenomena in human languages and, thus, grammaticalization. In this respect, some well-known alleged Semitisms or Septuagintalisms of the three parts of speech (verbs, conjunctions, and adverb)71 will be considered from the perspective of grammaticalization theory to show that the syntactic changes can occur within a language itself due to universal human cognition. Since grammaticalization of New Testament Greek is beyond the scope of the present study, we will take the time to investigate some representative cases of Semitisms or Septuagintalisms. Grammaticalization of verbs will be examined first (§6.3.1). Then, grammaticalization of conjunctions (§6.3.2) and adverbs (§6.3.3) will be investigated respectively.

71

A grammatical item does not belong to only one part of speech but extends over two or more parts of speech because most grammatical items may have their grammatical polysemies (cf. §6.2.2.3(e)).

Syntactic Change of New Testament Greek as Grammatical Polysemy

251

6.3.1 Verbs Some hypotactic and paratactic verbal expressions 72 in the synoptic gospels and Acts have been called Semitisms or Septuagintalisms.73 The frequent verbs cited are a saying verb (le,gw), movement verbs (e;rcomai and poreu,omai), and posture verbs (sth,kw, i[sthmi, avni,sthmi, evgei,rw, kaqi,zw, and ka,qhmai). There are two reasons why hypotactic expressions of these verbs are called Semitisms or Septuagintalisms. Above all, the participial expressions are redundant. Second, these redundant expressions also occur in Semitic languages. However, it seems that the participial expressions could have resulted from syntactic change due to internal-induced syntactic change in the development of Hellenistic Greek by the process of grammaticalization regardless of contactinduced syntactic change. The grammaticalization of the verbs means that the grammaticalized verbs have lost their lexical meanings and gained grammatical meanings. In other words, while an agent performs the action indicated by the main verb, the grammaticalized verbs are freely employed in contexts where the agent should not necessarily do the actions of the grammaticalized verbs.

72

73

It is generally said in linguistic fields that “hypotaxis” is a kind of sentence type used to connect A with B through a participle or a subordinate conjunction. A hypotactic construction forms a complex sentence. On the other hand, it is generally said that “parataxis” is a kind of sentence type that connects A with B through a coordinate conjunction such as kai,. A paratactic sentence forms a compound sentence. For an explanation of development from parataxis into hypotaxis, see Harris and Campbell 1995:282-313; Deutscher 2000:13-4. In the case of the paratactic kai,, Deissmann (1927:132) argues, “Parataxis appears to be not Greek only from the orthodox point of view of the Atticists, … As a matter of fact, parataxis is the original form of every primitive speech, including the Greek; it survived continuously in the language of the people, and even found its way into literature when the ordinary conversation of the people was imitated.” Further, he (1927:132-3) cites from Brugmann: “it is beyond doubt that the language of Homer exhibits on the whole far more of the original paratactic structure than the language of Herodotus and the Attic prose writers, such as Thucydides, Plato, Demosthenes. … There is in fact no very great difference to be detected between Homeric Greek and the Modern Greek dialects in this particular. … we find authors using paratactic constructions where they might have employed hypotactic forms, such being in general use in the cultivated language, we may generally assume that there has been an upward borrowing from the forms of the language of every-day life.” However, Sanders (1969:250) persuasively raises a question: do Markan paratactic constructions show that the Gospel of Mark has temporal priority over the other Synoptic Gospels? He says that “We can hardly think, then, that parataxis can really serve to prove what Black claims for it.”

252

Syntax

Concerning grammaticalization of verbs it is worthwhile to mention Dalman’s list of Hebraisms and Aramaisms to mind (1909:22-8).74 Although I suspect that all verbs on the list can be syntactic polysemies due to grammaticalization process, this is not the place to deal with all of them. Some verbs and their instances suggested by Dalman will be discussed.

6.3.1.1 SAY Verb: le,gw Hypotactic le,gontej / le,gwn / le,gousa are accompanied with “action,” “thinking,” “feeling” and “saying” verb groups in the synoptic gospels.75 The hypotactic participle of le,gw accompanied with verba dicendi has especially attracted scholars’ attention due to its tautology and semantic incongruity. Also, le,gwn RECITATIVUM has been considered as a Semitism or Septuagintalism. (3) e;kraxan le,gontej\ ti, h`mi/n kai. soi,( ui`e. tou/ qeou//. (Mt 8:29) RECITATIVUM (4) pw/j ei=pen auvtw/| o` qeo.j le,gwn\ evgw. o` qeo.j VAbraa.m. (Mk 12:26) RECITATIVUM It seems that le,gwn/le,gontej in (3) and (4) is used redundantly, which means that the lexical relationship between le,gwn/le,gontej and the main verbs is hyponymy or total synonymy. Many biblical scholars consider 74

75

He made the list as follows: evlqw,n, evrco,menoj, avfei,j, katalipw,n, kaqi,saj, avnasta,j, evgerqei,j, avpokriqei,j ei=pen, evla,lhsen (ei=pen)-le,gwn, h;rxato, h;rxanto, etc. It seems that the syntactic changes of h;rxato and h;rxanto can be called grammaticalization (cf. Moulton 1906:14; Cadbury 1920a:162-3). Some instances of le,gwn recitativum accompanied with some verb groups are as follows: with ACTION verbs avposte,llw (Mt 21:1-2, Mk 8:26; 12:6, Lk 19:29-30), gra,fw (Lk 1:63), diarrh,gnumi (Mt 26:65), di,dwmi (Mt 26:27, Lk 22:19), di,dwmi su,sshmon (Mk 14:44), evktei,nw (Mt 8:3), e;rcomai (Mt 24:5, Mk 5:35, Lk 19:18), r`api,zw (Mt 26:67-8), paragi,nomai (Lk 19:16), poti,zw (Mk 15:36), prose,rcomai (Mt 26:17, Mk 14:45), and prosporeu,omai (Mk 10:35); with THINKING or FEELING verbs dialogi,zomai (Lk 5:21), avganakte,w (Mt 26:8), evkplh,ssw (Mk 6:2), and qauma,zw (Mt 8:27); with SAYING verbs aivte,w (Mk 6:25), avnaboa,w / boa,w (Mt 27:46, Lk 9:38), avpokri,nomai (Mt 3:15, Mk 15:2, Lk 5:22; 13:2; 23:3 (fhmi,)), avrne,omai (Mt 26:70, Mk 14:68, Lk 22:57), blasfhme,w (Mk 15:29), dialogi,zomai (Mt 16:7, Mk 11:31), diaste,llw (Mk 8:15), doxa,zw (Mk 2:12), evkpeira,zw (Lk 10:25), evnte,llomai (Mt 17:9), evpifwne,w (Lk 23:21), evpiscu,w (Lk 23:5), evpitima,w (Mt 20:22, Mk 1:25; 9:25, Lk 4:35), evrwta,w / evperwta,w (Mt 15:23, Mk 8:27, Lk 9:18), khru,ssw (Mk 1:7), kra,zw / avnakra,zw (Mt 8:29, Mk 3:11), lale,w (Mt 23:1, 2), le,gw / fhmi, (Mt 8:8, Mk 8:28, Lk 7:39), parakale,w (Mk 5:12), fwne,w (Mk 15:34), yeudomarture,w (Mk 14:57), and suzhte,w (Mk 1:27). In the case of avpokriqei.j ei=pen, as shown above, avpokri,nomai is grammaticalized into avpekriqei,j. The lexical meaning of avpekriqei,j was added to that of le,gw. This case is also common cross-linguistically.

Syntactic Change of New Testament Greek as Grammatical Polysemy

253

the two cases as Semitisms or Septuagintalisms due to contact-induced syntactic change76 because the syntactic changes are also found in the Semitic sources. However, it seems that the usage of le,gwn can be considered to be grammatical polysemy of le,gw in the development of Hellenistic Greek. Attention should be paid to the fact that the biblical scholars recognize that le,gwn RECITATIVUM also occur in other Hellenistic Greek literature, although they traditionally played down their significance. BDF (§420) mentions that “Herodotus has a comparable usage e;fh le,gwn, eivrw,ta le,gwn, e;lege fa,j and the like.” Turner (1976b:397; cf. 1963:155-6) presents the cases, e;fh le,gwn (Sophocles and Herodotus), e;faske le,gwn (Aristophanes), and le,gwn ei=pen ou[tw (Demosthenes). Porter (1989:590) illustrates other extra-biblical Greek usages (e.g. Thuc. 2.65.1; Plato Apol. 29C; Gorg. 512C1; P.Fay. 123.15-16 [AD 100]; Hdt. 1.88.2; 1.118.1; 1.125; Epict. 4.8.26; BGU 624.15 [CE 284-305]). Moreover, the syntactic change from verba dicendi into LEXICAL ZERO/RECITATIVUM is a typical pattern of grammaticalization process in world languages. As (3) and (4) indicate, the second SAY has lost its lexical meanings and gained the grammatical feature RECITATIVUM, which can be aptly identified as a grammaticalization process. As a result, (3) can be translated: “they shouted: …” and (4): “how God said to him: …”

76

Bousset 1906:243; Burney 1922:52-5; Moulton and Howard 1929:453; BDF §§397(3), 420; Thompson 1985:70; David Black 1995:216, 219; Chapman 1993:161; Wilcox 1994:368. As for avpokriqe,ntej ei=pon Dalman (1909:24-5) mentions that the usage “is a well-known peculiarity of Hebrew narrative style” found in 1 Macc., Tobit, Book of Enoch (1:2; 6:4; 15:1; 21:9; 22:7, 9; 24:6; 25:1, 3; 27:1), Apocalypse of Baruch, 2 Esdras, Assumptio Mosis, Book of Jubilees, Judith, 2 Macc., and also in Ech. R. i. 4, j. Erub. 18d. Wilcox (1965:124) argues that “It is a well-attested fact that Aramaic and Syriac not infrequently employ certain verbs in a kind of auxiliary sense, to introduce the action of the main verbal idea. As a result, their own specific meaning either becomes redundant or is otherwise virtually lost” Especially, exploring Babatha archive, Lewis (1998:14) notes that “Direct discourse is regularly introduced by a main verb plus le,gwn or le,gousa. This addition of the participle is one of the most familiar Hebraisms of the Greek Old and New Testament.” Fitzmyer (1981:114-5) also regards avpokriqei.j ei=pen, poreuqe,ntej avpaggei,late, and avpekri,nato le,gwn as Septuagintalisms and h;rxato le,gein as an Aramaism because the former expressions occur in the Septuagint, whereas the latter style does not. Furthermore, to the list of Lucan Septuagintalisms, he (1998:115) adds some instances from Acts such as avnasta,j ei=pen, avpokriqe,ntej ei=pon, and hvrw,twn auvto.n le,gontej. In addition, another interesting Semitism is paratactic and hypotactic constructions of le,gw; for more detail instances, see Dalman 1909:24-5. For the grammaticalization from paratactic (e.g. avpekri,qh kai. le,gei) into hypotactic construction (e.g. avpokriqei.j ei=pen), see the grammaticalization from BICLAUSAL into MONOCLAUSAL (cf. chapter 6 n.97).

254

Syntax

The grammaticalization from SAY verbs into LEXICAL ZERO/RECITATIVUM is a widespread linguistic phenomenon in other languages. Above all, it is worthwhile to mention Old Akkadian RECITATIVUM. Deutscher (2000:66) says that the Akkadian RECITATIVUM is a “model textbook example of a process of grammaticalization.” He (2000:66-75) suggests that enma RECITATIVUM (Old Akkadian c. 25002000 BC) and later, umma X-ma RECITATIVUM (Early Old Babylonian c. 2000-1800 BC) are grammaticalized from “saying.” He (2000:74) notes that umma X-ma RECITATIVUM is accompanied with verba dicendi despite its redundancy. The Akkadian grammaticalization processes are similar to le,gwn RECITATIVUM. The same grammaticalization process is not limited to Semitic. This is also adduced by Klamer 2000 (Kambera, Buru, and Tukang Besi) and Heine and Kuteva (2002:267-8; Nama, Twi, Kusasi, Cahuilla, Saramaccan, Sranan, West African PE, Thai, Khmer, Vai, Lezgian, and Buru).77 Accordingly, the linguistic phenomena of le,gontej / le,gwn / le,gousa LEXICAL ZERO/RECITATIVUM occur in Hellenistic Greek as well as in other languages in the past and the present. This means that the syntactic change should not be regarded as a peculiar characteristic of New Testament Greek, nor as a Semitism or a Septuagintalism.78 The syntac77

78

The grammaticalization from SAYING verbs into COMPLEMENTIZER is also common in other languages; see Lord 1976:187-9; Hopper and Traugott 2003:187, 196; Frajzyngier 1984; 1991; Klamer 2000; Heine and Kuteva 2002:261-5; especially, Frajzyngier (1996:200) who cites from forty languages to show the grammaticalization process from verbs of saying into complementizers. For the grammaticalization process from COMPLEMENTIZER into RECITATIVUM, see the grammaticalization process from o[ti COMPLEMENTIZER into o[ti RECITATIVUM (§6.3.2.2). In the cases of the grammaticalized le,gw participles of New Testament Greek, there are four interpretative possibilities. The first case is that the semantic domain of the main verb in a sentence does not overlap with that of the SAY verb. The SAY verb preserves its lexical meaning with a grammatical function of RECITATIVUM. For instance, avpe,steilen auvto.n eivj oi=kon auvtou/ le,gwn of Mk 8:26 can be interpreted as “Jesus sent him home, saying, ~.” And e;dwken auvtoi/j le,gwn of Mt 26:27 can be interpreted as “he gave (it) to them, saying, ~.” The second case is that the semantic field of the main verb in a sentence is contradictory to that of the SAY verb. The SAY verb lost its lexical meaning, but gained a grammatical function (i.e. RECITATIVUM). For instance, e;grayen le,gwn of Lk 1:63 can be interpreted as “he wrote, ~” without the meaning of le,gw. The meaning of gra,fw is utterly different from that of le,gw. The actions of the two verbs cannot occur at the same time. Zacharias could not both “write” and “say” simultaneously. Hence, it can be interpreted as “he wrote, ~”. The third case is that the meaning of the main verb in a sentence is a hyponym of a SAY verb. The participle of a SAY verb lost its lexical meaning, but it gained a grammatical function. For instance, h;rxanto dialogi,zesqai oi` grammatei/j kai. oi` Farisai/oi le,gontej of Lk 5:21 can be interpreted as “the scribes and the Pharisees began to question, ~.” And avpokriqh,sontai

Syntactic Change of New Testament Greek as Grammatical Polysemy

255

tic change from this “saying” verb into RECITATIVUM is an internalinduced syntactic change in the metaphorical process, and le,gwn LEXICAL ZERO/RECITATIVUM is a grammatical polysemy of le,gw.

6.3.1.2 Movement Verbs: e;rcomai and poreu,omai The verbs, COME and GO (e;rcomai and poreu,omai) are representative motion verbs. It is usually said that the COME form is represented as a “motion-towards-speaker,” whereas the GO form is characterized as a “motion-away-from-speaker” (Wilkins and Hill 1995:209). The hypotactic forms of e;rcomai and poreu,omai are accompanied with other verbs and even a synonymic verb (e.g. “go” and “come”) like (5) to (7). (5) evgw. evlqw.n qerapeu,sw auvto,n. (Mt 8:7) FUTURITY (6) evlqw.n katw,|khsen eivj Kafarnaou.m. (Mt 4:13) PERFECTIVE (7) poreuqe,ntej eivsh/lqon eivj kw,mhn Samaritw/n. (Lk 9:52) PERFECTIVE It has been pointed out that in (5) to (7) the lexical meanings of GO and COME are redundant. These two verbs have traditionally been regarded as Semitisms79 or Septuagintalisms.80 However, it seems that the syntactic changes of GO and COME are due to internal-induced syntactic change rather than contact-induced syntactic change. Above all, it is frequently used in classical and Hellenistic Greek.81 Furthermore, it is

79

80

81

auvtw/| oi` di,kaioi le,gontej of Mt 25:37 can be interpreted as “the righteous will answer him, ~”. The SAY verb (le,gw) should not be translated literally because both dialogi,zomai and avpokri,nomai are hyponyms of le,gw. The last case is a tautology. The main verb and the participle of a SAY verb in a sentence are the same verbs. In this case, the participle of a SAY verb also lost its lexical meaning, but it functions only grammatically. For instance, o` VIhsou/j evla,lhsen auvtoi/j le,gwn of Mt 28:18 can be translated as “Jesus said to them, ~”. And oi` de. ei=pan auvtw/| le,gontej Îo[tiÐ of Mt 8:28 can be translated as “and they told him, ~”. Dalman 1909:20-1; Moulton and Howard 1929:452-3; Zerwick 1963:127. Black (1967:126) suggests that evlqw,n is an Aramaism because it occurs in Rabbinic literature. Criticizing Dalman’s “superfluous usage,” Gathercole (2004:86 n.14) mentions that “it is not at all clear that there is evidence of superfluous reference to ‘coming.’” Wifstrand (2005:33), for instance, considers that Lk 9:52 poreuqe,ntej eivsh/lqon is a Septuagintalism because “a participle like that is at times found in the Pentateuch with h=lqon or similar verbs to represent one of two combined Hebrew finite verbs (Num13:26 kai. poreuqe,ntej h=lqon, cf. Jos. 8:11 kai. poreuo,menoi h=lqon). Also Fitzmyer 1981:115; Davies and Allison 1988-97:1.274 Mealand (1991:54-7) suggests that poreu,omai is frequently found in Polybius, Herodotus Hist. 7.196.3, Thucydides Hist. 1.137.3, 4, Xenophon, Hell 3.2.9.11-2, Hell 6.4.3.14, Anab 4.4.16.1, Plato, Phaedo 113d4, Tim 42b4, Xenophon Hell 4.7.2.12, 7.5.14.5, Mem 3.11.2.1, 3.13.6.1, Demosthenes Or 23.167.2.

256

Syntax

well-known that MOVEMENT VERBS (i.e. GO and COME) are sources for TENSE / ASPECT MARKERS in many languages. The two most basic motion verbs are grammaticalized into different grammatical functions. It is usually noted that the “motion-towards-speaker” verb is grammaticalized into a PERFECTIVE MARKER, 82 whereas the “motion-away-from-speaker” verb into a FUTURITY MARKER. Grammaticalization of GO into FUTURITY is the most typical case such as (8): (8) Bill is going to go to college after all. (Hopper & Traugott 2003:1) The second GO preserves its lexical meaning “go,” whereas the first GO has lost its lexical meaning and gained a grammatical feature FUTURITY. This can be represented as [-LEXICAL] [+TEMPORAL].83 The grammaticalization from MOVEMENT VERBS (i.e. GO and COME) into TENSE / ASPECT MARKERS are found in many languages.84 Consequently, evlqw,n and poreuqe,ntej have lost their lexical meanings and gained grammatical meanings.ȱThis means that the participles evlqw,n andȱ poreuqe,ntejȱ are not Semitisms or Septuagintalisms due to contact-induced syntactic change but grammatical polysemies due to grammaticalization process. (5) can be translated: “I will heal him.” (6) can be translated: “He has dwelt in Capernaum.” Lastly, (7) can be translated: “They haveȱentered a village of the Samaritans.”85

6.3.1.3 Posture Verbs The grammaticalization of POSTURE VERBS (STAND and SIT) is a very widespread linguistic phenomenon. In the case of the grammati82 83

84

85

For the definitions of meaning labels, see Bybee, Perkins and Pagliuca 1994:316-24. Sweetser (1988:392) persuasively explains that “we lose the sense of physical motion (together with all its likely background inferences). We gain, however, a new meaning of future prediction or intention—together with its likely background inferences. We thus cannot be said to have merely “lost” meaning; we have, rather, exchanged the embedding of this image-schema in a concrete, spatial domain of meaning for its embedding in a more abstract and possibly more subjective domain.” Motion verbs such as “to go” or “to come” are main sources for the grammaticalization of the “future tense” in the world languages; see Bybee and Pagliuca 1987; Heine, Güldeman, Kilian-Hatz, Lessau, Roberg, Schladt, and Stolz 1993; Lichtenberk 1991; Emanatian 1992; Bybee, Perkins & Pagliuca 1994; Wilkins and Hill 1995; Giacalone Ramat 2000; Heine and Kuteva 2002:161-3. More instances can be suggested. The evlqw,n FUTURITY MARKER includes Mt 2:8, Lk 7:3; 14:9; 18:8; etc., whereas the evlqw,n PERFECTIVE MARKER includes Mt 2:23; 9:18; 25:27; 26:43, Mk 14:40, 45; 15:43, Lk 15:17; 19:23; etc. The poreuqe,ntej PERFECTIVE MARKER includes Mt 22:15; 27:66; etc.

Syntactic Change of New Testament Greek as Grammatical Polysemy

257

calizations of POSTURE VERBS (STAND and SIT), the POSTURE VERBS lost their lexical meanings and gained their grammatical features (e.g. ASPECT MARKER). It is well-known that the STATIVE SENSE of the POSTURE VERBS is grammaticalized into PROGRESSIVE or HABITUAL.86 On the other hand, the DYNAMIC SENSE of the POSTURE VERBS (STAND) is grammaticalized into a PERFECTIVE MARKER. Consequently, it should be remembered that the POSTURE VERBS do not restrict the meanings of main verbs, even though the two lexical meanings tend to overlap or to polarize because their lexical meanings of the POSTURE VERBS have become ZERO LEXICAL.

6.3.1.3.1 sth,kw and i[sthmi A STATIVE SENSE of sth,kw or i[sthmi in the Synoptic Gospels and Acts is grammaticalized into an auxiliary marker: PROGRESSIVE or HABITUAL.87 (9) o[tan sth,kete proseuco,menoi( avfi,ete ei; ti e;cete kata, tinoj. (Mk 11:25) PROGRESSIVE (10) ei`sth,keisan de. oi` avrcierei/j … euvto,nwj kathgorou/ntej auvtou/. (Lk 23:10) PROGRESSIVE (11) filou/sin evn tai/j sunagwgai/j … e`stw/tej proseu,cesqai. (Mt 6:5) HABITUAL (12) o` Farisai/oj staqei.j pro.j e`auto.n tau/ta proshu,ceto\ o` qeo,j( euvcaristw/ soi o[ti ouvk eivmi. w[sper oi` loipoi. tw/n avnqrw,pwn … (Lk 18:11) HABITUAL It has been pointed out that the participle form of sth,kw or i[sthmi is redundant and an Aramaism.88 On the other hand, refuting the REDUNDANCY of the verbs, Moulton (1919a:16; cf. 1906:15-6) suggests that STAND preserves its lexical touch, that is, STAND depicts the bodily posture of praying. However, the linguistic phenomenon where the STATIVE SENSE of the POSTURE VERBS is grammaticalized into PROGRESSIVE or HABITUAL is regarded as a cross-linguistically internal-induced syn-

86

87 88

The grammaticalization of the STATIVE SENSE (SIT, STAND, and LIE) into the PROGRESSIVE has been well proved by many publications. See Kuteva 1999:191213; 2001:43-74; Heine and Kuteva 2002:193-4, 276-9, 282; Newman 2002; Newman and Rice 2004. Newman and Rice (2004:354-5) provides more publications. Bybee, Perkins and Pagliuca (1994:127) suggest that HABITUAL ASPECT refers to “customarily repeated on different occasions.” Dalman 1909:23; Wilcox 1965:125; Voelz 1984:962.

258

Syntax

tactic change.89 As (9) to (12) show, STAND has lost its lexical meaning and gained a grammatical feature, PROGRESSIVE or HABITUAL. In this respect, the verb sth,kw of (9) is represented as [ZERO LEXICAL] and [PROGRESSIVE]. The phrase sth,kete proseuco,menoi does not always refer to standing prayer.90 It can be interpreted like this: “whenever you are praying [no matter how he or she kneels or stands], forgive if you have anything against any one.” The verb i[sthmi of (10) can be interpreted as PROGRESSIVE like this: “the chief priests…were vehemently accusing him.” Interestingly, in relation to Acts 1:11, Wilcox (1965:125) correctly suggests that the lexical meaning of i[sthmi is “virtually lost,” as is shown by the fact that the Aramaic ~wq can be interpreted as “to be.” As the Aramaic ~wq is grammaticalized into a PROGRESSIVE MARKER, so the Greek i[sthmi is also grammaticalzed into a PROGRESSIVE MARKER. Likewise, (10) i[sthmi of Acts 1:11 (ti, e`sth,kate ÎevmÐble,pontej eivj to.n ouvrano,n) can be interpreted as the PROGRESSIVE, “Why are you looking into heaven?”91 On the other hand, the verbs sth,kw of (11) and i[sthmi of (12) function as another auxiliary marker, HABITUAL. The decision of PROGRESSIVE or HABITUAL depends on their contexts, although it might be unclear in some cases. The STATIVE sth,kw of (11) can be interpreted: “they (habitually) love to pray in the synagogues …” Lastly, (12) can be interpreted: “The Pharisee (habitually) prayed thus to himself, ‘I thank thee that I am not like other men’ …” Consequently, the participle forms of sth,kw and i[sthmi are not redundant nor Aramaisms/Septuagintalisms due to contactinduced syntactic change but function as the auxiliary marker (PROGRESSIVE or HABITUAL) due to the grammaticalization process.

89 90

91

Bybee, Perkins, and Pagliuca 1994:125-75; Heine and Kuteva 2002:280-2. Comparing yLEc:m. ~yaeq" (“he stood praying” j. R. H. S. 58b) and ha'L"c:l. ~yaeq" (Est. R. 34), Dalman (1909:23) states that “it is also implied that standing was the usual attitude at prayer; it is not, however, a regular phrase to say, ‘he stood and prayed.’” In fact, the other custom of prayer (e.g. kneeling prayer) was practiced among the Jews (1 Kgs 8:54, Ezra 9:5, Dan 6:10, Mt 26:39, Acts 7:60; 20:36; 21:5, and Eph 3:14). Swete (1898:245) also suggests that “kneeling seems to have been preferred on occasions of great solemnity or of distress.” Wilcox (1965:125) mentions, “It could be that the writer really did intend to refer to the posture of the disciples, and perhaps we should look no further.” Unfortunately, he should have considered that “stand” can refer to the auxiliary marker of the PROGRESSIVE. In addition, i[sthmi of Acts 5:25 (e`stw/tej kai. dida,skontej) and Acts 16:9 (e`stw.j kai. parakalw/n) can be considered as a PROGRESSIVE MARKER.

Syntactic Change of New Testament Greek as Grammatical Polysemy

259

6.3.1.3.2 avni,sthmi and evgei,rw The participle forms of DYNAMIC SENSE of POSTURE VERBS (i.e. avni,sthmi and evgei,rw) have also attracted scholars’ attention on account of their redundancy. (13) avnasta.j h=lqen pro.j to.n pate,ra e`autou/. (Lk 15:20) PERFECTIVE (14) VAnasta/sa de. Maria.m … evporeu,qh eivj th.n ovreinh.n. (Lk 1:39) PERFECTIVE (15) evgerqei.j o` VIhsou/j hvkolou,qhsen auvtw/.| (Mt 9:19) PERFECTIVE The words avnasta,j and evgerqei,j have been considered as Semitisms92 or Septuagintalisms.93 However, Mealand (1991:48-51) shows that avnasta,j occurs frequently in Polybius. For this reason, he (1991:50-1) convincingly concludes, “There is absolutely no need to think that this is rare in Greek or that it is peculiarly Semitic.” Furthermore, the grammaticalization process of the avnasta,j or evgerqei,j from POSTURE VERBS into PERFECTIVE is a common linguistic phenomenon.94 While the STATIVE VERBS, sth,kw and i;s, thmi of (9) to (12) function as PROGRESSIVE or HABITUAL, the DYNAMIC VERBS avni,sthmi and evgei,rw of (13) to (15) function as PERFECTIVE. The verbs avnasta,j and evgerqei,j of (13) to (15) have lost their lexical meanings and gained the PERFECTIVE. It is interesting that despite his inconsistent translation, Fitzmyer points out that avnasta,j is “used inchoatively” (1981-5:114) and that “the combination connotes inception” (1981-5:362; cf. 590).95 The two verbs, avnasta,j and evgerqei,j of (13) to (15) can be interpreted in the PERFECTIVE ASPECT. The sentence (13) can be interpreted: “he [has] started on ….” (14) can be translated: “Mary has gone ….” Last, (15) can be interpreted: “Jesus has followed him.” More instances of avni,sthmi in the PERFECTIVE can be suggested (e.g. Mt 9:9, Mk 2:14; 7:24; 10:1; Lk 1:39;

92 93

94

95

Moulton and Howard 1929:453; Turner 1976a:51; 1976b:393; BDF (§419(2)) suggests that it is pleonastic. Fitzmyer (1981-5:362) posits, “Luke uses here the pleonastic or redundant ptc. anastas(a) with another verb … The usage is a Septuagintalism.” He (1981-5:912) also seems to consider that evgerqei.j is a Septuagintalism. In fact, the usage occurs in the Old Testament (Num 22:14, Judg 13:11, 1 Kgs 19:21, Ezra5:2), 1 Macc 9:44, Enoch 89:47, 48, and Josephus (Ant. 9.108). For the same pattern of grammaticalization in other languages, see Austin 1998:1936; Kuteva 1999:191-213; Bybee, Perkins, and Pagliuca 1994:125-75; Newman 2002; Heine and Kuteva 2002:276-8; John and Rice 2004:351-96. Fitzmyer’s translations are inconsistent. He sometimes interprets avnasta,j in Lk 6:8 literally (He got up and stood [there]), whereas he freely translates avnasta,j at Lk 1:39 (Mary set out and went in haste).

260

Syntax

4:29, 38; 5:28; 6:8; 15:18, 20; 17:19; 23:1; 24:12, 33, Acts 1:15; 5:6, 34; 8:27; 9:11; etc.)

6.3.1.3.3 kaqi,zw and ka,qhmai SIT VERBS (kaqi,zw and ka,qhmai) have been considered as Semitisms due to the redundant mention of bodily posture (cf. Dalman 1909:22). (16) pa/j o` lao.j h;rceto pro.j auvto,n( kai. kaqi,saj evdi,dasken auvtou,j. (Jn 8:2) PROGRESSIVE (17) kaqi,santej evlalou/men tai/j sunelqou,saij gunaixi,n. (Acts 16:13) PROGRESSIVE (18) pa,lai a'n evn sa,kkw| kai. spodw/| kaqh,menoi meteno,hsan. (Lk 10:13) PROGRESSIVE The SIT of (16) to (18) seems to observe the grammaticalization pattern of POSTURE VERBS (STAND and SIT) into TENSE/ASPECTUAL MARKER, as mentioned before. These participles of SIT have lost their lexical meanings and gained the PROGRESSIVE function as in many other languages. 96 Accordingly, the sentence (16) can be interpreted: “… he was teaching them.” Sentence (17) can be interpreted: “We were speaking to the women who had come together.” Sentence (18) can be translated: “They would have been repenting ….” I would suggest that the SIT PROGRESSIVE ASPECT includes Mt 13:48; 27:36, Mk 2:6;97 5:15; 9:35;98 12:41,99 Lk 5:3; 7:32; 8:35; 14:28,100 31; 16:6, Jn 9:8, Acts 12:21,101 Rev 11:16; etc.

96

97

98

99

For the grammaticalization of STATIVE SIT VERB into PROGRESSIVE, see Bybee, Perkins, and Pagliuca 1994:125-75; Austin 1998:19-36; Kuteva 1999:191-213; Kuteva 2001:43-74; Newman 2002; Heine and Kuteva 2002:276-8; John and Rice 2004:351-96. Although this verse is a paratactic construction at the syntactic level (h=san de, tinej tw/n grammate,wn evkei/ kaqh,menoi kai. dialogizo,menoi evn tai/j kardi,aij auvtw/n), it is interpreted as a monoclause at the semantic level; for biclausal syntax and monoclausal interpretation, see Kuteva 2001:72. In terms of Mk 9:35, there are two kinds of interpretations of SIT: (a) Jesus was tired of his journey so that he sat down (Rawlinson 1925:127). (b) Jesus’s teaching style as formal teaching like a Rabbi (Swete 1898:193; Taylor 1952:404; Cranfield 1959:307; Hooker 1991:227; Evans 2001:61). However, this POSTURE VERB does not refer to lexical meaning but is an ASPECTUAL MARKER. It is interesting that Mk 12:41 has another variant (e`stw,j W Q f1 f13 28 565). Metzger (1994:94) explains that “Those responsible W Q f1 f13 28 565 al obviously thought that it was more appropriate for Jesus to stand (e`stw,j) than to sit in the temple.” However, there is no semantic difference whether it is kaqi,saj or e`stw,j since they are grammaticalized into AUXILIARY MARKER.

Syntactic Change of New Testament Greek as Grammatical Polysemy

261

In conclusion, the hypotactic constructions of a saying verb (le,gw), movement verbs (e;rcomai and poreu,omai), and posture verbs (sth,kw, i[sthmi, avni,sthmi, evgei,rw, kaqi,zw, and ka,qhmai) have been viewed as Semitisms or Septuagintalisms due to contact-induced syntactic change. However, it is more appropriate to consider the syntactic changes as grammatical polysemies of the verbs caused by the grammaticalization process due to internal-induced syntactic change.

6.3.2 Conjunctions Usages of i[na CONSECUTIVE and o[ti CAUSALE and RECITATIVUM have been generally considered typical Semitisms or Septuagintalisms on the basis of contact-induced syntactic change. Furthermore, the linguistic phenomenon that w[ste overlaps with the grammatical domain i[na FINAL is also called a Semitism. However, from the perspective of cognitive linguistics, the usages can be regarded as grammatical polysemies of the conjunctions due to internal-induced syntactic changes because they are also found in classical Greek literature and the paradigms of the syntactic change are common in the world grammatical lexicon. First of all, i[na CONSECUTIVE as grammatical polysemy will be investigated with w[ste FINAL (§6.3.2.1). Next, o[ti CAUSALE and RECITATIVUM will be presented as grammatical polysemies of o[ti (§6.3.2.2).

6.3.2.1 {Ina Based on contact-induced syntactic change, many biblical scholars have suggested that i[na CONSECUTIVE in New Testament Greek is influenced by the Aramaic yD.102 They explain that the Aramaic ambiguous yD 100 As to Luke 14:28, although Fitzmyer (1981-5:1065; cf. 114) also suggests that the kaqi,saj functions as INCHOATIVE, it seems to function as PROGRESSIVE. 101 It is uncertain whether the agent sits or stands. The prepositional phrase evpi. tou/ bh,matoj of Acts 12:21; 25:6, 17 is not always accompanied by kaqi,zw; it also occurs with i;,sthmi. Josephus uses two terms: kaqi,zw (Bell.2.172 Pila/toj kaqi,saj evpi. bh,matoj) and i;,sthmi (Ant. 4.209 o` avrciereu.j evpi. bh,matoj u`yhlou/ staqei,j; Ant 7.370 sta.j evfV u`yhlota,tou bh,matoj o` basileu.j e;lexe pro.j to. plh/qoj). Neh 8:4 uses i;,sthmi (e;sth Esdraj o` grammateu.j evpi. bh,matoj xuli,nou). At any rate, it seems that the SIT of Acts 12:21 (also Bell.2.172, Ant 7.370) denotes only an ASPECTUAL MARKER without lexical meanings. 102 Burney (1922:75) assumes that i[na or o[ti constructions were frequently used and mistranslated in the Gospels, especially in the Gospel of John. Manson (1935:76) argues that “The stumbling-block here is the i[na.” “The Marcan version,” he (1935:78)

262

Syntax

caused i[na FINAL to extend to i[na CONSECUTIVE. In parallel with this, w[ste FINAL has been considered as a Semitism. Consequently, the syntactic change of i[na from FINAL into CONSECUTIVE or w[ste CONSECUTIVE into FINAL has been submitted as an evidence for Semitic interference.103 However, it seems that the syntactic change of i[na from FINAL into CONSECUTIVE and the syntactic overlapping between i[na and w[ste are a typical grammaticalization process found in other languages.104 This means that these manifestations should be understood as grammatical polysemies due to internal-induced syntactic change rather than Semitisms or Septuagintalisms due to contactinduced syntactic change. Interestingly, in order to explain the syntactic shift from FINAL into CONSECUTIVE or the overlapping of the two conjunctions i[na and w[ste, some biblical scholars have tried to make a cognitive connection between “purpose” and “result.” Moule (1959:142) mentions that “the Semitic mind was notoriously unwilling to draw a sharp dividingline between purpose and consequence.” 105 However, the cognitive continues, “rests on a misunderstanding of the Aramaic due mainly to the ambiguity of the particle d¼.” Explaining overlapping of i[na with w[ste Moule (1959:142) also suggests, “It may be for this reason (or, at least, Semitic influence may be a contributory cause) that i[na with Subj. sometimes occurs in contexts which seem to impose a consecutive, instead of final, sense upon it.” Black (1967:76-81) proposes that the Greek i[na was a mistranslation of the Aramaic yD and presented i[na as evidence of Aramaic rendering. In addition, for translations of Aramaic yD in the Septuagint, see Andrews 1947:15-51. Interestingly, Turner (Elliott 1993:132) argues that i[na after verbs is influenced by Latin grammar. 103 Nevertheless, some scholars are doubtful that i[na is a mistranslation of yD. In relation to Mark 4:12 and its parallels (Mt 13:13 and Luke 8:10), Fitzmyer (1981:708-9) suggests that “there is no need to invoke a ‘mistranslation’ of an underlying Aramaic d¼, which should have been understood as a relative pron. but is taken in the sense of causal hoti by Matthew and final hina by Mark and Luke.” It is noteworthy that Ullendorff (1955:50-2) suggests that yD can be translated into i[na. He argues that yD was used as FINAL because its counterparts in other languages (Hebrew, Aramaic dialects, Akkadian, and Ethiopic) function as conjunctions and relative pronouns. However, Ullendorff would have done better to consider the other possibility, that is, that the syntactic change is due to an internal-induced shift. It seems that grammatical functions of yD are its grammatical polysemies because grammaticalization processes of yD are similar to those of yk; see chapter 6 n.131. 104 For a detailed study of grammaticalization of i[na, see my MA thesis (Lee 2000); it, however, is written in Korean. 105 Some scholars also think that the semantic relationship between “purpose” and “result” is ascribed to Semitic thought. BAG (378) assumes that “In many cases purpose and result cannot be clearly differentiated, and hence i[na is used for the result which follows according to the purpose of the subj. or of God. As in Jewish and pagan thought, purpose and result are identical in declarations of the divine will.” Following Burton (1900:§375) Zerwick (1963:122) states: “In Biblical use a confusion of

Syntactic Change of New Testament Greek as Grammatical Polysemy

263

connection should not be limited to Semitic language because the cognitive process is a common linguistic phenomenon due to human cognition. In fact, a syntactic change of i[na has been familiar to linguists for some time. In 1912, Antoine Meillet (1912:398-400), a student of de Saussure and one of the most eminent linguists, proposed that qe,lw i[na in classical Greek became (diachronically) qa in modern Greek and first coined grammaticalisation to refer to the syntactic change. From the perspective of cognitive syntax, various grammatical functions of i[na in classical and Hellenistic Greek, including i[na CONSECUTIVE of New Testament Greek, are grammatical polysemies of i[na which are caused by the cognitive process. Metaphorical or metonymic processes cause grammatical functions of i[na to extend to other grammatical functions (e.g. CAUSAL or TEMPORAL). Grammatical shifts of i[na will be presented in a diachronic and a synchronic way in comparison with grammatical shifts of other conjunctions, which show similar patterns of grammaticalization found in other languages. In what follows, I will show that the grammatical shift from FINAL into CONSECUTIVE and the overlapping of i[na can be also explained by internal-induced syntactic change. The syntactic functions of i[na could be roughly divided into three periods. In classical Greek, the grammatical polysemies of i[na are DEMONSTRATIVE ADVERB (there), LOCATIVE RELATIVE ADVERB (where), TEMPORAL RELATIVE ADVERB (when), and FINAL (purpose). (19) i[na ga,r sfin evpe,fradon hvgere,qesqai. (Iliad 10.127) DEMONSTRATIVE ADVERB (20) limh.n eu;ormoj, i[n’ ouv crew. pei,smato,j evstin. (Odyssey 9.136) LOCATIVE RELATIVE ADVERB (21) soi. de. ga,moj scedo,n evstin, i[na crh. kala. me.n auvth.n e[nnusqai. (Odyssey 6.27) TEMPORAL RELATIVE ADVERB (22) VAcaiw/n calkocitw,nwn evj pedi,on katabh/nai, i[n’ o[rkia pista. ta,mhte. (Iliad 3.252) FINAL {Ina of (19) functions as a demonstrative adverb “there” (Monro 1891:260).106 Apollonius Dyscolus (coni. 243.11-3), a Greek grammarian the two notions [CONSECUTIVE and FINAL] is perhaps helped by the Hebrew idea that God is the principal and universal cause of all that happens.” Hooker (1991:128) maintains that “Jewish thought tended to blur the distinction between purpose and result; if God was sovereign, then of course what happened must be his will, however strange this appeared.” 106 Similarly, the original meaning of yk is “here”; see chapter 6 n.131.

264

Syntax

of the second century AD argued that i[na was used as a local adverbial. {Ina of (20) functions as a relative adverb. The “there” (i.e. de re) had lost its lexical meaning, gained a new grammatical feature CONNECTIVE, and became a relative adverb (i.e. de dicto).107 The grammaticalization from DEMONSTRATIVE into CONNECTIVE is found in many languages. 108 {Ina of (21) shows that i[na became more grammaticalized from “where” to “when” (cf. Traugott 1978:3.369-400). The grammaticalization process from SPATIAL into TEMPORAL is the process of abstraction (§6.2.2.5). It is known that TEMPORAL is more abstract than SPATIAL (Hopper and Traugott 2003:85). 109 In fact, people see TEMPORAL through SPATIAL, as “in the weeks ahead of us” and “that is all behind us now” show (Sweetser 1988:389-405).110 Finally, i[na of (22) gains a feature of FINAL. A similar pattern is found in the English preposition “for.” Heine, Claudi, and Hünnemeyer argue that the “for” serves LOCATIVE, FINAL, and CAUSAL purposes, as in (23):111 107 The shift from DEMONSTRATIVE into CONNECTIVE or RELATIVE (i.e. gaining connectivity or relativeness) is also explained by the grammaticalization from de re (i.e. “reference-based” or “domain of reality”) into de dicto (i.e. “text-based” or “domain of speech”) in a metaphorical process. In other words, the reference-based “there” is grammaticalized into a text-based “there” (i.e. a connective or a relative) in the process of abstraction. For discussion of the shift of de re into de dicto, see Frajzyngier 1991:219-51; 1996:169-203. 108 Diessel (1999:125) mentions that sentence connectives “are frequently formed from a pronominal demonstrative and some other element (e.g. an adverb or adposition) that indicates the semantic relationship between the conjoined propositions.” Illustrating the grammaticalization from œfter as an Old English preposition into the subordinator “after” in Middle English, Traugott and König (1991:208) suggest that “examples of metaphoric changes in the grammaticalization of spatial terms include the development of adverbs and prepositions into clause connectives.…” For instances of grammaticalization from DEMONSTRATIVE into RELATIVE or SUBORDINATOR, see Heine and Kuteva 2002:113-116. Rubin (2005:49) states that “A few Semitic languages exhibit a relative pronoun which derives from a noun *’aqar- ‘place’ (cf. Akkadian asrum, Aramaic ’atra ‘idem’). This development is best known from Biblical Hebrew, where ‘aser has become the normal relative pronoun (and later, complementizer), and the lexical form has been lost.” For more information on grammaticalization from DEMONSTRATIVE into CONNECTIVE, see Diessel 1999. 109 What is more grammaticalized? Heine, Claudi, and Hünnemeyer (1991:156-61) give an answer with the degrees of grammaticalization: PERSON > OBJECT > ACTIVITY > SPACE > TIME > QUALITY. 110 Heine (1990:131, 133) suggests that the grammaticalization of “from” indicates the process from SPATIAL to TEMPORAL and from TEMPORAL to CAUSAL as below: a. He returned late from the pub. SPATIAL b. He drank from morning to evening. TEMPORAL c. He died from excessive drinking. CAUSAL 111 Originally, they (1991:150-3) suggest that ALLATIVE in many languages is grammaticalized into BENEFACTIVE, PURPOSE, and CAUSE in this order. They used

Syntactic Change of New Testament Greek as Grammatical Polysemy

265

(23) a. Mary left for Paris. LOCATIVE b. Mary worked hard for her exam. FINAL c. Mary couldn’t sleep for pain. CAUSAL Consequently, in classical Greek, major grammatical polysemies of i[na are DEMONSTRATIVE, LOCATIVE RELATIVE, TEMPORAL RELATIVE, and FINAL. In New Testament Greek, i[na also has grammatical polysemies as followings: FINAL, TEMPORAL, CAUSAL, CONSECUTIVE, COMPLEMENTIZER, and INFINITIVE. The major usages of i[na and w[ste FINAL of New Testament Greek are shown in (24) to (30):112 (24) prose,feron auvtw/| paidi,a i[na auvtw/n a[yhtai. (Mk 10:13) FINAL (25) evlh,luqen h` w[ra i[na doxasqh/| o` ui`o.j tou/ avnqrw,pou. (Jn 12:23) TEMPORAL RELATIVE (26) VAbraa.m … hvgallia,sato i[na i;dh| th.n h`me,ran th.n evmh,n. (Jn 8:56) CAUSAL, COMPLEMENTIZER (27) poreuqei.j avna,pese eivj to.n e;scaton to,pon( i[na o[tan e;lqh| o` keklhkw,j se evrei/ soi\ fi,le( prosana,bhqi avnw,teron. (Lk 14:10) CONSECUTIVE (28) h;gagon auvto.n e[wj ovfru,oj tou/ o;rouj … w[ste katakrhmni,sai auvto,n. (Lk 4:29) FINAL (29) evxelqo,ntej evkh,ruxan i[na metanow/sin. (Mk 6:12) COMPLEMENTIZER (30) ti,j evstin( ku,rie( i[na pisteu,sw eivj auvto,nÈ (Jn 9:36) INFINITIVE {Ina of (24) or (25) shows that i[na still preserves FINAL and TEMPORAL respectively, like polysemies of classical Greek. {Ina of (26) is syntactically ambiguous in the CAUSAL 113 and COMPLEMENTIZER. 114 The grammaticalization from PURPOSE into CAUSE 115 is found in

112

113

114 115

three examples above. I changed ALLATIVE into LOCATIVE because LOCATIVE includes ALLATIVE. One of syntactic polysemies of i[na (e.g. Mk 5:23, Jn 9:3 etc.) was IMPERATIVAL, but the usage was also used in Hellenistic Greek (Moule 1959:144-5; Doudna 1961:49-51; Zerwick 1963:139). As for the complexity of i[na CAUSAL, Sluiter (1992:39) mentions that “To all appearances causal i[na was a ghost-construction, unknown from any extant Greek text. Grammars of New Testament Greek every now and then revived the concept in order to explain some problematic instances of the conjunction.” For an explanation of COMPLEMENTIZER, see i[na COMPLEMENTIZER below. It is also possible that i[na TEMPORAL is grammaticalized into i[na CAUSAL. Traugott and König (1991:195) present the instances of temporal-causal polysemies in many languages. They (1991:194-5) cite “since” as one of the examples of temporal-causal polysemies as below: a. I have done quite a bit of writing since we last met. TEMPORAL

266

Syntax

many languages. 116 In other words, PURPOSE is reinterpreted into CAUSE in the metonymic process. The cognitive relationship is adduced by their use of the same morphology. Thompson and Longacre (1985:185; cf. Cristofaro 2003:161-2) posit that “Many languages use the same morphology for purpose and reason clauses.” Heine and Kuteva (2002:247) also state that “Purpose and cause are … the same polysemy set.” In modern English, the semantic domain of CAUSAL overlaps with that of FINAL. Heine (1990:133) asserts that the English preposition “for” can be interpreted into both CAUSAL and FINAL because of syntactic ambiguity, as in (31). (31) Mary is working for her exam. CAUSAL, FINAL Furthermore, the fact that i[na CAUSAL is not a Semitism or a Septuagintalism is supported by a remark of Apollonius Dyscolus. He bore witness to the circumstance that i[na functioned as CAUSAL in Hellenistic Greek. 117 This is adduced by many occurrences of the i[na CAUSAL in Hellenistic Greek.118 On this score, Sluiter (1992) strongly suggests that the i[na CAUSAL is “sound Greek,” and he (1992:39-53; here 53, cf. 1990:143-71) considers i[na CAUSAL as “colloquial language of the educated Greek”.119 He considers that the syntactic shift is due to confusion of a grammatical terminology and shows that the i[na

116

117

118

119

b. Since you are so angry, there is no point in talking with you. CAUSAL c. Since you are not coming with me, I will have to go alone. CAUSAL d. Since Susan left him, John has been very miserable. TEMPORAL CAUSAL Heine 1990:133; Heine and Kuteva 2002:246-7. Whether the syntactic change from PURPOSE into CAUSE is grammaticalization is in question. Heine and Kuteva (2002:247) suggest that “we argue that the former PURPOSE precedes the latter CAUSE in time; so far, however, there is no conclusive historical evidence to support this hypothesis.” However, the syntactic change of i[na may support the assumption because i[na CAUSE is not found in classical Greek but found in Hellenistic Greek. Apollonius dealt with i[na in his works (coni. 243.11-3, esp. 244.24-6; synt. 381.10-2, 388.9-11). He (coni 243.21) mentioned two instances of CAUSE: i[na avnagnw/, evtimh,qen and i[na loidorh,sw evpeplh,cqhn. Sluiter (1992:43-8) suggests nine cases: Anth. Pal IX 169; Johannes Chrysostomus, de Sacerdotio SC 272 (ed. Malingrey) I 4.33-5; Basilius of Caesarea, Regulae brevius tractatae PG 31, 1237-40: Question 233; Pseudo-Johannes Chrysostomus, Oratio catechetica in dictum evangelii [Matt 20:1] PG 59, 582; Pseudo-Johannes Chrysostomus, de Sacerdotio 1. VII, PG 48, 1069-70; Johannes Chrysostomus, de poenitentia hom. VIII, PG 49, 339; Johannes Chrysostomus, de Virginitate, SC 125 (ed. Musurillo), 22, 1; Johannes Chrysostomus, in Acta Apostol. homil. III, PG 60, 40.; Severianus of Gabala, homil. in qua potestate, PG 56, 419. Sluiter (1990:158) points out that i[na CAUSAL occurs in John 9:2 and Revelation 22:14.

Syntactic Change of New Testament Greek as Grammatical Polysemy

267

CAUSAL is a syntactic shift in Hellenistic Greek itself.120 Accordingly, the syntactic shift of i[na from FINAL into CAUSAL does not result from contact-induced syntactic change but from internal-induced syntactic change. {Ina of (27) is used as a CONSECUTIVE. In parallel with i[na CONSECUTIVE, it should be mentioned that w[ste frequently functions as FINAL instead of CONSECUTIVE, as (28) shows. Interestingly, i[na extends to CONSECUTIVE (i.e. the primary usage of w[ste), whereas w[ste overlaps with FINAL (i.e. the primary usage of i[na). {Ina CONSECUTIVE and w[ste FINAL have been considered as Semitisms in that Aramaic yD is mistranslated on the basis of contact-induced syntactic change, as mentioned above. However, the syntactic changes of i[na CONSECUTIVE and w[ste FINAL should have been considered on the basis of internal-induced syntactic change. Above all, i[na CONSECUTIVE is found in Hellenistic Greek. It is adduced by VGT 305 (P. Lond 96413 [ii/iii CE], Epict. iv. 8. 21) and by Jannaris (1897:§§1758, 1951). Also w[ste denotes pure PURPOSE regardless of Semitic influence in Hellenistic Greek, and Muraoka (1973:210-3) suggests three cases from the Hellenistic Greek literature (purely final, consecutive-final and pseudo-final). Furthermore, there is a cognitive relation between FINAL and CONSECUTIVE. Their grammatical overlaps (i.e. i[na CONSECUTIVE and w[ste FINAL) are also shown by their grammatical ambiguities. {Ina is sometimes ambiguous in FINAL and CONSECUTIVE while w[ste is ambiguous in CONSECUTIVE and FINAL.121 It is also interesting that biblical scholars pay attention to the cognitive relation between FINAL and CONSECUTIVE. Burton (1900:§370(d)) explains FINAL as “intended” CONSECUTIVE. Radermacher (1925:188) mentions that the difference between FINAL and CONSECUTIVE in relation to infinitive is decided by subjective or objective coloring of the thought. BDF (§391) contends that “the line dividing it from purpose clauses is hardly distinguishable.” Even Moule compares FINAL with CONSECUTIVE

120 He (1992:41) considers that the origin of the usage is derived from terminological confusion between FINAL and effective CAUSE in relation to interpretation of su,ndesmoi aivtiologikoi,. 121 Moule (1959:140-6) observes that i[na overlaps with w[ste in relation to their ambiguity. Some i[na constructions are syntactically ambiguous in FINAL and CONSECUTIVE (Mt 1:22; 23:26, Mk 1:38; 4:12, Lk 9:45, Rom 5:20f; 6:1, Gal 5:17, 1Thess 5:4, 1 Jn 1:9, Gen 22:14). The syntactic ambiguity of w[ste in CONSECUTIVE and FINAL occurs in Mt 10:1; 27:1, Lk 4:29; 9:52; 20:20, 1 Cor 5:1.

268

Syntax

from the perspective of the human mind. He (1959:138) states that “so as to” is used as both FINAL and CONSECUTIVE as in (32):122 (32) He hit the ball hard so as to score. FINAL, CONSECUTIVE Modern English uses the same syntax to refer to FINAL and CONSECUTIVE as (32) illustrates. The grammatical decision whether they denote FINAL or CONSECUTIVE depends on its context as in (27) and (28). Zerwick (1963:122) observes that “it is finally only from the context that the distinction between final and consecutive sense can be gathered in Hellenistic usage.” Similarly, Sluiter (1990:157) suggests that the decision of the grammatical function of i[na CAUSAL cannot be made by grammatical analysis but by theological argument. The fact that i[na and w[ste are syntactically ambiguous in relation to human cognition implies that the syntactic changes are not due to contact-induced change but internal-induced change. In this respect, a grammatical polysemy depends on its context, as lexical polysemy does. Also, the grammaticalization process of FINAL into CONSECUTIVE is found in other languages.123 This grammaticalization process enables i[na to overlap with CONSECUTIVE and w[ste to overlap with FINAL respectively in New Testament Greek. Accordingly, the linguistic phenomena (i[na CONSECUTIVE and w[ste FINAL) are not Semitic interference but grammatical polysemies due to cognitive mapping of the human mind based on internal-induced syntactic change. {Ina of (29) is used as a COMPLEMENTIZER. The grammaticalization from DEMONSTRATIVE to NOMINAL COMPLEMENT is a typical pattern of the grammaticalization process (Heine and Kuteva 2002:106-7). The lexical meaning of i[na DEMONSTRATIVE is lost, but the grammatical function CONNECTIVE is gained. That is, i[na became more grammatical. The syntactic ambiguity in CAUSAL and COMPLEMENTIZER is found at John 8:56 and Revelation 14:13. Interestingly, the syntactic ambiguity of i[na in ADVERBIAL CONJUNCTION and NOMINAL COMPLEMENT is the same with o[ti. Finally, i[na of (30) functions as an INFINITIVE. This results in infinitive na, of modern Greek in the course of the development of the Greek infinitive.124 The 122 Zerwick (1963:122) mentioned that “There is thus a considerable affinity between the two types of clause, and it is not astonishing that there should be a tendency to confuse the two kinds of clause (cf. in English ‘so as to’ is used frequently in a final sense though primarily consecutive). 123 Rubin (2005:40) suggests that hatt¬ is used as “so that” and “in order that” in literary Arabic and in some colloquial dialects. 124 For the grammatical change of qe,lw i[na, see Browning 1983:8, 43, 79; Joseph 1990:113-59.

Syntactic Change of New Testament Greek as Grammatical Polysemy

269

grammaticalization from PURPOSE to INFINITIVE is found in many languages (Heine and Kuteva 2002:247-8). Hence, the grammatical functions of i[na in Hellenistic Greek have extended to LOCATIVE RELATIVE (o[pwj), TEMPORAL RELATIVE (o[te), FINAL (o[pwj), CAUSAL (o[ti), CONSECUTIVE (w[ste), COMPLEMENTIZER (o[ti), and INFINITIVE. Ironically, syntactic functions of the Greek conjunctions concerned have also been changed and their grammatical functions tend to overlap with each other.125 In modern Greek, the form of i[na is changed to gia ná, na, or qa and it performs three grammatical functions (i.e. ADVERBIAL CONJUNCTION, FUTURITY AFFIX, and INFINITIVE). Gia na, ADVERBIAL CONJUNCTION preserves its grammatical function from classical Greek. Na starts to function as an INFINITIVE since Hellenistic Greek.126 Qa derived from qe,lw i[na functions as a FUTURITY AFFIX.127 Rubin (2005:3940) mentions that a future tense construction is derived from a purpose clause in Tigrinya, Tigré, and qljltu-Arabic as well. To sum up, (i) {Ina has continuously been changing from classical Greek to modern Greek. It seems that the syntactic change of i[na FINAL into CONSECUTIVE or the syntactic overlapping between i[na and w[ste is one of their syntactic shifts in Hellenistic Greek itself. (ii) The syntactic changes of i[na have similar paradigms to those of other languages including Semitic in the world syntactic lexicon. Accordingly, i[na CONSECUTIVE and the syntactic overlapping between i[na and w[ste are grammatical polysemies of the two conjunctions respectively due to the grammaticalization process.

125 BAG notes that Greek conjunctions overlap with each other: o[pwj (“how,” “in order that,” and “that”), o[te (“when,” “while,” and “as long as”), o[ti (“that,” “so that,” “because,” and Ø), and w[ste (“for this reason,” “therefore,” “so,” “so that”). 126 In modern Greek, na INFINITIVE can be traced back to the Hellenistic Greek verbs (e.g. avfi,hmi, zhte,w, and aivte,omai). vAfi,hmi takes an infinitive to complete its construction in classical Greek but takes the i[na clause in Hellenistic Greek like Mk 11:16. Zhte,w takes an infinitive in classical Greek but takes the i[na clause (e.g. 1 Cor 4:2; 14:12) in Hellenistic Greek. According to BAG (577), aivte,omai takes o[pwj instead of the infinitive more and more in New Testament Greek (e.g. Mt 9:38, Lk 10:2, Ac 8:24). The change from taking the INFINITIVE to using a i[na or o[pwj clause implies that the INFINITIVE was replaced by the i[na clause in Hellenistic Greek. 127 For the development of future tense from classical Greek to Modern Greek, see Roberts and Roussou 2003:58-71.

270

Syntax

6.3.2.2 [Oti Regarding the syntactic shift of o[ti from SUBSTANTIVAL to CAUSAL or to RECITATIVUM, biblical scholars have presupposed that the syntactic change results from language contact between Greek and Semitic. The contact-induced syntactic change hypothesis has been suggested in two opinions. Some scholars suggest that the grammatical functions of o[ti of the New Testament were influenced by the Aramaic D, which are called Aramaisms.128 Others propose that Hebrew yk interferes with the Greek o[ti, because yk is rendered by o[ti in the Septuagint. This means that the o[ti CAUSAL of the New Testament seems to be a Septuagintalism.129 Furthermore, o[ti RECITATIVUM of New Testament Greek also has been explained as a Semitism (Zerwick 1963:142 n.9) or a Septuagintalism (Aejmelaeus 1993:37-48). It seems, however, that the o[ti CAUSALE or RECITATIVUM results from an internal-induced syntactic shift because the o[ti CAUSALE or RECITATIVUM is also found in classical Greek literature and because the cognitive process from SUBSTANTIVAL to CAUSALE or to RECITATIVUM is a common linguistic shift from the perspective of cognitive grammar. Above all, o[ti causale occurs in classical Greek as well as in koine Greek, regardless of Semitic interference. Cristofaro (1998:67) contends that “after Homer, hoti starts to be used to introduce reason clause.” In order to explain the grammaticalization process of o[ti from SUBSTANTIVAL to CAUSALE, she gives the following examples: (33) cai,rwn vAntilocw| o[ti oi` fi,loj h-en evtai/roj. (Iliad 23.556) SUBSTANTIVAL / CAUSALE

128 Burney 1922:76-7; Zerwick 1963:145; Wilcox 1965:115-121 Black 1967:70-79. 129 Thackeray 1909:54. It is worthwhile to mention Aejmelaeus’s argument, which is similar to Wifstrand’s (§6.1.5). He (1993:34; followed by Casey 1998:56, 95) claims that although Luke was a competent Greek writer, he tried to imitate “biblical” Greek. This, he argues, is adduced by his use of o[ti CAUSAL in the first half part of the Acts. He (1993:36) concludes that “o[ti causale in place of ga,r must be considered a Septuagintalism.” Applying bilingualism theory to o[ti CAUSAL, Casey (1998:94) presupposes that “All bilinguals suffer from interference.” He (1998:95) suggests that “this is also the correct pattern for Aejmelaeus’ demonstration that o[ti is used incorrectly and too often in the LXX as a rendering of yk precisely because it is so often used correctly.” However, syntactic imitation is not syntactic interference. It seems that syntactic imitation of the Septuagint should be regarded as syntactic codeswitching. For codeswitching as intentional device, see chapter 8. Explaining Gen 4:12 and 2 Sam 4:10, Gehman (1951:82-3) also suggests that TEMPORAL CONJUNCTION of o[ti is a Hebraism due to literal translation of yk into o[ti.

Syntactic Change of New Testament Greek as Grammatical Polysemy

271

(34) pollw/n de. ei[neka ouv foneu,sw min kai. o[ti auvtw/| moi suggenh,j evsti. o` pai/j. (Herodotus 1.109.9) CAUSALE (35) ouvk i;sasin o[ti ~Ippi,j (…) me.n h=rce. (Thucydides 1.20.2.4) SUBSTANTIVAL The passage from Homer (33) shows that the o[ti of the Iliad was syntactically ambiguous and can be interpreted both as “that” and “because.”130 This means that o[ti is used as SUBSTANTIVAL and sometimes extends to CAUSALE in classical Greek. (34) indicates that o[ti functions as an unequivocal CAUSALE. In (35) o[ti still preserves the SUBSTANTIVAL. The above cases of Homer and Herodotus demonstrate that o[ti already had been used as “because.” This implies that the authors of the New Testament might have known the usages as grammatical polysemies, so that they might have used the two grammatical functions of o[ti (i.e. “that” and “because”). Moreover, as (33) to (35) show, when the translators of the Septuagint rendered the Hebrew Bible into a Greek version, it seems that o[ti was considered to be equivalent to yk.131 So yk is mostly translated by o[ti in the Septuagint. 130 Another Greek conjunction, w`j presents a similar case: cai,rei de, moi h`tor w`j meu avei. me,mnhsai (Iliad 23.648). This construction shows that w`j is also as syntactically ambiguous as “that” and “because.” The grammaticalization process of w`j from TEMPORAL or SUBSTANTIVAL into CAUSAL is similar to that of i[na or o[ti. The syntactic ambiguity of a conjunction in TEMPORAL and CAUSAL is shown in modern English. Traugott and König (1991:195) give an example of the grammaticalization of TEMPORAL and CAUSAL like “since” (see chapter 6 n. 115). Also they (1991:194-5) present more instances of the grammaticalization process in other languages; German (infolgedessen), French (puisque), Spanish (pues), Swedish (eftersom and emedan), Rumanian (din moment ce), Dutch (want), Estonian (paräst), and Finnish (koska). Concerning grammaticalization of w`j, see Cristofaro 1998. 131 Concerning yk RECITATIVUM, some scholars assume that the function is derived from an emphatic function so that they disregard the existence of yk RECITATIVUM as a syntactic category. Muilenburg (1961) tries to explain the syntactic shift of yk from DEMONSTRATIVE or EMPHATIC into DEITIC in rhetorical contexts. Schoors (1981:256-9; here 259) argues that yk recitativum as a syntactic category, “should be deleted from grammars and dictionaries”; for the list of more supporters, see Aejmelaeus 1993:38 n.3. However, Aejmelaeus (1993:37-48) convincingly suggests that yk RECITATIVUM is derived from SUBSTANTIVAL, not from EMPHATIC. His argument is adduced from grammaticalization theory in that the grammaticalization process from SUBSTANTIVAL to RECITATIVUM is a typical linguistic phenomenon. The grammaticalization process of Hebrew yk is similar to that of i[na (§6.3.2.1). The grammaticalization processes of yk are as follows: “here” DEMONSTRATIVE > “when” TEMPORAL CONJUNCTION > “because” CAUSALE > “that” SUBSTANTIVAL > RECITATIVUM. These grammaticalization processes are very common (cf. §6.3.2.1; §6.3.2.2). Especially, Heine and Kuteva (2002:171-2) illustrate that the languages in which “here” is grammaticalized into TEMPORAL CONJUNCTION and CAUSAL CONJUNCTION are Lingala (áwa “here”) and Albanian (ke “here”).

272

Syntax

In parallel with this, o[ti of New Testament Greek also covers two major functions (i.e. SUBSTANTIVAL and CAUSALE), as (36) to (38) show: (36) … oi;damen o[ti avlhqh.j ei= … (Mk 12:14) SUBSTANTIVAL (37) u`mw/n de. maka,rioi oi` ovfqalmoi. o[ti ble,pousin kai. ta. w=ta u`mw/n o[ti avkou,ousin. (Mt 13:16) SUBSTANTIVAL / CAUSALE (38) evkwlu,omen auvto,n( o[ti ouvk hvkolou,qei h`mi/n (Mk 9:38) CAUSALE In (36), o[ti functions as a conjunction to introduce a substantival clause. (37) indicates that the o[ti clause shows syntactic ambiguity (i.e. SUBSTANTIVAL and CAUSALE). Many o[ti clauses are ambiguous in SUBSTANTIVAL and CAUSALE at the syntactic level.132 In (38) o[ti obviously means CAUSALE. According to cognitive linguistics, the shift from SUBSTANTIVAL to CAUSALE results from re-analysis in the metonymic process. Interestingly, explaining that the syntactic function of o[ti CAUSALE was influenced by Hebrew yk, Casey (1998:56) argues that “this set up too close an association between two words in the minds of translators who were suffering the double level of interference which is inevitable when translators translate texts.” Although he assumes the linguistic shift on the basis of contact-induced change (i.e. interference), it should be stressed that Casey understands the syntactic shift as a cognitive process. From the perspective of cognitive linguistics, the substantival clause is paraphrased by a causal clause, as English does.133 Explaining the ambiguous o[ti in SUBSTANTIVAL and CAUSALE, Cristofaro (1998:67) convincingly explains that “these clauses refer to an event which represents the ground for the occurrence of the event coded by the main predicate, so that both a complement and a causal reading are possible.” Consequently, re-analysis due to metonymic process makes the grammatical function of o[ti extend to CAUSALE. Furthermore, o[ti RECITATIVUM in New Testament Greek has been regarded as a Semitism or a Septuagintalism, as mentioned before. It seems, however, that the syntactic shift of o[ti from SUBSTANTIVAL to RECITATIVUM is an example of grammaticalization. First, o[ti SUBSTANTIVAL is grammaticalized into o[ti RECITATIVUM. The shift

“Here” is also grammaticalized into RELATIVE in the cases of Tondano and Buang (Heine and Kuteva 2002:174). 132 For the same cases, see Mt 5:45; 6:5, Mk 1:34; 4:41, Lk 2:10-11, Jn 5:39; 8:45; 9:16, 17; etc.; see Zerwick 1963:146. 133 In English, “I am pleased that he comes” can be paraphrased into “I am pleased because he comes.”

Syntactic Change of New Testament Greek as Grammatical Polysemy

273

pattern of the grammaticalization process is typical.134 Second, o[ti RECITATIVUM is also found in classical Greek literature (e.g. Homer, Plato, Xenophon, etc).135 These points suggest that o[ti RECITATIVUM is a grammatical polysemy of o[ti due to the grammaticalization process. Many biblical scholars have considered that the ambiguous Aramaic D or Hebrew yk caused synoptic variants or syntactic variants when it was translated into an equivalent Greek conjunction, i[na or o[ti.136 It seems, however, that some grammatical functions of Aramaic D and Hebrew yk show syntactic ambiguity because their syntactic functions extend to other functions and overlap with them due to grammaticalization processes. As layering theory explains (§6.2.2.3 (1)), their old layers coexist and interact with new layers. In this respect, the syntactic changes of Aramaic D or Hebrew yk are their grammatical polysemies. Likewise, some grammatical functions of i[na and o[ti show their syntactic ambiguity because their syntactic functions extend to other functions and overlap with them due to typical grammatical processes found in classical or Hellenistic Greek as well as in other languages. So i[na CONSECUTIVE, w[ste FINAL, o[ti CAUSALE and RECITATIVUM, which have been identified as Semitisms or Septuagintalisms, are grammatical polysemies of i[na, w[ste, and o[ti due to internal-induced syntactic change.

134 Lord (1976:179-80) suggests that the Ewe-saying verb is grammaticalized into RECITATIVUM and COMPLEMENTIZER; see §6.3.1.1. Many languages have the same pattern of grammaticalization from “saying” into COMPLEMENTIZER and RECITATIVUM. Traugott (1999:180) observes that “In the earliest stages, the verb introduces a direct quotation” in the process from saying into COMPLEMENTIZER. 135 Kieckers 1915:51-3, 1921:183; K-G 2.551.4; Schwyzer 1950:2.637-8; Mayer 2.3§155. 136 Black (1967:70) mentions that “The translation and mistranslation of the ambiguous Aramaic particle D is one of the best known of Gospel Aramaisms.” Also Burney 1922:69-78. As a matter fact, Andrews (1947:37-8) suggests that the yD are translated in New Testament Greek in various ways: (i) article with attributive phrases, with attributive participles, and with infinitives (ii) relative pronoun in attributive clauses, in relative clauses of cause, in relative clauses of purpose and result, in conditional relative clauses, with prepositions, with an adverbial accusative as antecedent, and with relative adverb of place, (iii) conjunctions as o[ti, i[na, o[pwj, and w`j, (iv) miscellaneous renderings as o[te or o[tan, relative adverbs, conjunctions used in translating compound conjunctions, and prepositions, (v) omission from Greek translations in genitive phrases, in expressions of substance, in direct discourse when replaced by kai, before anaphoric suffixes, and in idiomatic translations.

274

Syntax

6.3.3 Adverbs: Matthean to,te, Marcan euvqu,j, and Johannine ou=n Despite the fact that a syntactic change in New Testament Greek is similar to the general paradigm of the grammaticalization process in many languages, the same shift, to date, may not be found in Hellenistic Greek literature except for biblical language literature. If it corresponds to the typical pattern of the grammaticalization process, however, it is highly probable that the syntactic change can be regarded as grammaticalization. The three DISCOURSE MARKERS (Matthean to,te, Markan euvqu,j, and Johannine ou=n), which are grammaticalized from adverbs, are the relevant cases. I will now begin by focusing on Matthean to,te. The syntactic polysemy of the Matthean to,te includes three major functions. It is used as a (i) DISTAL TEMPORAL ADVERB “at that time,” even in the future tense; (ii) LOGICAL CONNECTIVE,137 though pleonastically; and (iii) as a DISCOURSE MARKER as shown below. (39) to,te hvge,rqhsan pa/sai ai` parqe,noi evkei/nai… (Mt 25:7) DISTAL PAST TEMPORAL At that time all those maidens rose… to,te evrei/ o` basileu.j toi/j evk dexiw/n auvtou/… (Mt 25:34) DISTAL FUTURE TEMPORAL At that time the King will say to those at his right hand… (40) h' pw/j du,natai, tij eivselqei/n eivj th.n oivki,an tou/ ivscurou/ kai. ta. skeu,h auvtou/ a`rpa,sai( eva.n mh. prw/ton dh,sh| to.n ivscuro,nÈ kai. to,te th.n oivki,an auvtou/ diarpa,sei. (Mt 12:29) LOGICAL CONNECTIVE Or how can one enter a strong man's house and plunder his goods, unless he first binds the strong man? Then indeed he may plunder his house. (1) To,te o`moiwqh,setai h` basilei,a tw/n ouvranw/n de,ka parqe,noij… (Mt 25:1) DISCOURSE MARKER GÞG The kingdom of heaven shall be compared to ten maidens… Scholarly attention has been given to the third function (i.e. DISCOURSE MARKER), known as a characteristic of the Gospel of Matthew. 138 In regard to the unique Matthean usage of to,te, McNeile

137 Robertson (1934:429) mentions that to,te “in the apodosis accents the logical connection of thought” at Mt 12:29, Mk 13:14, Jn 7:10, 20, 21, 1 Cor 15:54, 2 Cor 7:12; etc. 138 To,te as a DISCOURSE MARKER occurs 61 times out of 91 times in the Gospel of Matthew (cf. McNeile 1910:127; Robertson 1934:119). However, the same usage oc-

Syntactic Change of New Testament Greek as Grammatical Polysemy

275

(1915:16-7) points out that “to,te … carries the reader to the next event in the narrative, often with no strict historical sequence.” Furthermore, it does not function as a logical connection. In this respect, the Matthean to,te has been known as an Aramaism.139 Going a further step, Buth (1990) considers to,te as a diagnostic word in order to make a distinction between the Aramaic and Hebrew sources of the four gospels, because to,te corresponds to !yda(b) and because it “is virtually nonexistent in LXX translation from Hebrew narrative.” Regarding to,te as an Aramaism, Buth (1990:47) suggests that the absence of a narrative to,te in Mark and Luke indicates that “any major Semitic sources for Mark and Luke should be Hebrew” and that “Mark and Luke do not reflect reliance on Aramaic narratives.” Accordingly, Buth suggests that Matthew uses Aramaic sources, whereas Mark and Luke use Hebrew sources depending on using to,te as a DISCOURSE MARKER. Buth, however, seems to disregard three points. Of primary importance, !yda(b) might be translated into kai, (Dan 5:6, 8; 6:6, 12, 15) or de, (Dan 4:16; 6:5) as well as to,te. The Aramaic !yda(b) is not the only equivalent to to,te. Second, if we should consider to,te as Semitic interference, to,te should be regarded as a Septuagintalism rather than an Aramaism because za' as a DISCOURSE MARKER is translated into to,te as a DISCOURSE MARKER in the Septuagint (Exod 15:1, Lev 26:34, Num 21:17, Deut 4:41, cf. Josh 22:1, 1 Kgs 3:16).140 In addition, the three major functions of to,te are similar to those of za', as will be mentioned below. Last, it is highly probable that to,te employed as a DISCOURSE MARKER in the Gospel of Matthew is a grammatical polysemy of to,te due to the grammaticalization process, rather than an Aramaism or a Septuagintalism due to a contact-induced syntactic change. It is common that a TEMPORAL ADVERB is grammaticalized into a LOGICAL CONNECTIVE and a LOGICAL CONNECTIVE into a DISCOURSE MARKER. Above all, the syntactic change of to,te shows a typical paradigm of the grammaticalization process from TEMPORAL ADVERB “at that time” into the LOGICAL CONNECTIVE “then.” The grammaticalization of a DISTAL TEMPORAL ADVERB into a CONNECTIVE can be explained by the metaphorical process from de re into de dicto (§6.3.2.1). Rosén (1961:189-90) also considers that the adverb “may acquire a relacurs in Acts 2:14, 37 in the Bezan text where it is used as a DISCOURSE MARKER and used solely without co-occurrence with other conjunctions (e.g. kai, or de,). 139 McNeile 1910:128; Torrey 1933:290; Turner 1963:341, 1976b:397; Buth 1990. cf. Gundry 1994:30; cf. Allen 1912:13; Davies and Allison 1988:1.264. 140 The use of to,te in the FUTURE TENSE is regarded as one of characteristics of the Matthean to,te. As za' as a TEMPORAL ADVERB is used in past time as well as future time, the Matthean to,te is used in both tenses.

276

Syntax

tive function” and states that !yda(b) “in that case” also refers to “when” or “as” and “because”.141 Lastly, the LOGICAL CONNECTIVE “then” of to,te is grammaticalized into a DISCOURSE MARKER. The lexical meaning is lost to ZERO (Þ) and to,te still functions as a discourse marker CONTINUATIVE like “and.”142 The grammaticalization process of to,te is like (42): (42) TEMPORAL ADVERB

> LOGICAL CONNECTIVE > DISCOURSE MARKER

-TEMPORAL +CONNECTIVE

-LOGICAL + CONTINUATIVE

The pattern of the grammaticalization process in (42) is found in other languages. Explaining the grammaticalization process from TEMPORAL to LOGICAL, Heine and Kuteva (2002:291) suggest that “spatial and temporal markers are grammaticalized in specific contexts to markers of ‘logical’ grammatical relations ….” In the case of Semitic, it seems that the Hebrew za' might observe the general pattern of the grammaticalization process. It functions as (i) TEMPORAL ADVERB “at that time” (past time at Gen 12:6 and future time at Isa 35:5, 6); (ii) LOGICAL SEQUENCE like “then,” introducing apodosis after conditional clause at Isa 58:14; (iii) DISCOURSE MARKER at Exod 15:1, etc., as mentioned above.143 Similarly, the Aramaic !yda(b), which includes grammatical functions (TEMPORAL ADVERB, LOGICAL CONNECTIVE, and DISCOURSE MARKER) does as well. The pattern is also found outside Semitic languages. Moine (1996) suggests that (a)lors in French, which is equivalent to “then,” has the same pattern of grammaticalization process. The function of DISTAL TEMPORAL is grammaticalized into ANAPHORIC, and then into a DISCOURSE MARKER to refer to consequence and mere discourse continuity. Abraham (1991:373) suggests that denn in German was also grammaticalized from LOCALISTIC into DISCOURSE FUNCTIONAL and proposes the following paradigm: LOCALISTIC > TEMPORAL > LOGICAL > ILLOCUTIVE > DISCOURSE FUNCTIONAL. Heine and Kuteva (2002:291-3)

141 Although Buth (1990:37) criticizes that Rosen’s proposal “should be rejected,” his argument is unconvincing. 142 Schiffrin (1987:152) considers conjunctions as discourse markers and proposes that “and is a structural coordinator of ideas which has pragmatic effect as a marker of speaker continuation“ [emphasis in the original]. 143 McNeile (1910:128) mentions that za' as a narrative marker in the Hebrew Old Testament occurs 20 times.

Syntactic Change of New Testament Greek as Grammatical Polysemy

277

provide many more instances of this and show that it is a widely distributed phenomenon in several languages worldwide. Accordingly, it is highly probable that the to,te DISCOURSE MARKER is not an Aramaism or a Septuagintalism due to contactinduced syntactic change but a grammatical polysemy of to,te due to internal-induced syntactic change. Also, it seems that its equivalents, Hebrew za' and Aramaic !yda(b), follow the general pattern of grammaticalization process.144 The similarities of grammatical functions between Greek to,te and the two equivalents suggest that the syntactic changes which show the general pattern in many languages are due to internalinduced change because human cognition is common. When it comes to the Matthean to,te as a DISCOURSE MARKER which has zero lexical meaning, the Marcan euvqu,j and Johannine ou=n (which are similar to the Matthean to,te) catch our attention because these three particles are known as peculiar characteristics of the three gospels respectively (cf. McNeile 1910:127). Burkitt (1904:2.89) surprisingly suggests that Mark’s euvqu,j and John’s ou=n “mean the same thing, and that they really correspond to the Hebrew ‘w¬w consecutive’.”145 In fact, Mark does not refer to the TEMPORAL as in “immediately” when he employs euvqu,j in Mark 1:21, 23, 29, 30, etc. Citing Mark 1:10, Marcus (1934:224-5) supposes that “euvqu,j in the koiné has lost some of its original force and come to mean ‘thereupon.’” BDF (§102 (2)) mentions that “Mark uses euvqu,j with two different functions, temporal and stylistic.” BAG (321) also suggests that euvqu,j “is weakened to then, so then.” Still, it is of interest that John does not use ou=n as “therefore” (e.g. Jn 3:25; 4:1, 46; 6:52; 8:21; etc.), as Burkitt mentions (1904:2.89). In relation to semantic bleaching, Burkitt (1904:2.89) convincingly states that the fact “That this translation so often omits euvqu,j in S. Mark, and so often omits ou=n in S. John or translates it by a simple ‘and’ is strong evidence that these particles are in their essential meaning nothing more than a copula.” I propose that the Matthean to,te should be added to the list of zero lexical-semantic adverbs which function as a DISCOURSE MARKER. In 144 To,te as a DISCOURSE MARKER is coherently translated into hydyn in early Syriac versions. Chase (1893:23) mentions that hydyn “seems to be a favorite particle of connexion in Syriac.” Williams (2004:170-3) mentions that to,te is sometimes omitted in Syriac versions (Mt 9:37 P, 13:26 S, 15:28 C, 21:1 P, 27:16 SP, 27:38 P), when it is considered to be redundant. 145 Burkitt (1904:89) explains, “Not, of course, that either of these Gospels is a translation from the Hebrew; but if the authors of these Gospels were familiar with the Old Testament otherwise than through the awkward medium of the LXX, they might well have felt themselves in need of something to correspond to the Hebrew idiom.” However, Burkitt still assumes the contact-induced shift on the basis of categoricity theory.

278

Syntax

this respect, it seems that the three adverbs also show the same patterns of the grammaticalization process from ADVERB into DISCOURSE MARKER. Accordingly, it seems that the syntactic changes of the Matthean DISCOURSE MARKER to,te, Mark’s euvqu,j and John’s ou=n can be considered to be grammatical polysemies of the ADVERB to,te, euvqu,j, and ou=n due to the grammaticalization process rather than Semitic interference due to contact-induced syntactic change.

6.4 Syntactic Change as Grammatical Polysemy and Interdirectionality The bilingualism of first-century Palestine and the Roman Near East at the syntactic level has been discussed with primary emphasis on Semitic interference (i.e. Semitisms and Septuagintalisms) among New Testament scholars. The Semitic interference hypothesis is based on contact-induced syntactic change hypothesis, the categoricity theory, the study of langue, the grammar of writing, and language competence. The discussion of Semitisms or Septuagintalisms at the level of syntax has caught scholars’ attention because they have assumed that Semitic traditions are earlier and closer to sayings by Jesus and the stories about Jesus than are Greek traditions. For this reason, the Semitic interference hypothesis has been used as a criterion in relation to major issues of gospel studies such as the synoptic problem, the historical Jesus, provenances of the gospels and Acts, source criticism, textual criticism, form criticism, redaction criticism, literary criticism, the unidirectional development of christological titles and the development of early Christianity. The hypothesis that Semitic tradition is more original than Greek tradition has supported the unidirectionality hypothesis that the Jesus tradition is transmitted from Semitic into the Greek language, from the Judaeo-Palestinian into the Hellenistic Sitz im Leben, from the oral into the written communicative vehicle in a unilinear way. From the perspective of cognitive linguistics, however, it seems that the syntactic changes which have been known as Semitisms or Septuagintalisms are grammatical polysemies. I have illustrated three parts of speech such as verbs, conjunctions and adverbs. In the case of verbs, I have exemplified three groups of verbs. They were grammaticalized and have lost their lexical meanings and gained grammatical functions; for instance, a saying verb (le,gw), movement verbs (e;rcomai and poreu,omai), and posture verbs (sth,kw, i[sthmi, avni,sthmi, evgei,rw, kaqi,zw, and ka,qhmai). The SAY verb (le,gw) was grammaticalized into RECITA-

Syntactic Change as Grammatical Polysemy and Interdirectionality

279

TIVUM. In case of movement verbs (e;rcomai and poreu,omai), a motiontoward-speaker verb was grammaticalized into PERFECTIVE, but a motion-away-from-speaker verb was grammaticalized into FUTURITY. As for grammaticalization of STAND verbs (sth,kw, i[sthmi, avni,sthmi, evgei,rw), a stative sense of STAND verbs was grammaticalized into PROGRESSIVE or HABITUAL, whereas a dynamic sense of STAND verbs was grammaticalized into PERFECTIVE. And SIT verbs (kaqi,zw, ka,qhmai) were grammaticalized into PROGRESSIVE. Second, I submitted that i[na CONSECUTIVE, w[ste FINAL, and o[ti CAUSALE and RECITATIVUM were not Semitisms or Septuagintalisms due to contactinduced syntactic changes but syntactic polysemies as a result of the process of grammaticalization due to internal-induced syntactic changes. I have pointed out that Greek conjunctions of the New Testament as syntactic polysemies could extend their grammatical functions without the syntactic interference. Last, I have shown how Matthean to,te, Markan euvqu,j, and Johannine ou=n, the three adverbs, have lost their lexical meanings and were grammaticalized into DISCOURSE MARKERS. These usages are found in classical or Hellenistic Greek in most cases. Moreover, the patterns of the syntactic changes are universally found in other languages which are not related linguistically, culturally, politically, or geographically. For these reasons, the syntactic changes are grammaticalization, universal human cognitive process due to internal-induced change in Hellenistic Greek itself rather than Semitisms or Septuagintalisms due to contact-induced syntactic change. Accordingly, the syntactic Semitisms and the syntactic Septuagintalisms discussed before do not support the unidirectionality hypothesis. The bilingualism of first-century Palestine and the Roman Near East implies that the linguistic transmission of the Jesus and gospel traditions was hybrid, circular and interdirectional. The Jesus and gospel traditions were most likely circulated in Greek as well as in Aramaic during Jesus’s ministry and thereafter. Although one finds obvious Semitisms, the assumed syntactic Semitisms do not always refer to temporal priority. The bilingualism of first-century Palestine and the Roman Near East dilutes the boundary between the Judaeo-Palestinian and Hellenistic traditions. It also dilutes the boundary between the Semitic and Greek traditions. Semitic traditions are not always from Judaeo-Palestine, whereas Greek traditions are not always from Hellenistic regions. The Semitic tradition includes both Hellenistic Semitic tradition and Judaeo-Palestinian Semitic tradition. Also, the Greek tradition includes both Judaeo-Palestinian Greek tradition and Hellenistic

280

Syntax

Greek tradition. It is useless to look for a Semitic tradition as more “original” at the syntactic level. Accordingly, the Semitisms or Septuagintalisms at the syntactic level do not support the unidirectionality hypothesis. The transmission of the Jesus and gospel traditions at the level of syntax is hybrid rather than unilinear, circular rather than teleological, and interdirectional rather than unidirectional.

7. Phonology It usually has been thought that while Aramaic was used by Jesus as his primary language in first-century Palestine, the four gospels and Acts were written in Greek. This language contact between Aramaic and Greek leads to copious amounts of transliterated words in the four gospels and Acts. Generally speaking, when a story is translated, proper nouns of the story are transliterated. Likewise, local proper nouns in the settings of Jesus’ ministry were transliterated. Second, personal proper nouns as characters in the four gospels and Acts were transliterated as well. Third, it is interesting that some Aramaic words were transliterated and preserved in the Greek narratives. The problem is that the three kinds of Aramaic transliterated words have their variant spellings. There are two approaches to variant spellings of transliterated words at the phonological level. One is an orthographical view. Many biblical scholars have investigated variant spellings of transliterated words on the basis of the orthographical view. This view stems from a particular understanding of language change. That is, phonological variation or language change found in the four gospels and Acts are understood as abnormal. The presupposition here is based on the categoricity theory, the study of langue, the grammar of sound, and language competence. In terms of transliterated words, this approach has tried to single out the “correct” spelling from the “corrupted” spellings and the “right” spelling from the “erroneous” spellings. This is because they have assumed that the “correct” transliterated spelling is the “original” transliterated spelling which was used by the original writers. Furthermore, scholars have assumed that Semitic type spellings (e.g. VIerousalh,m, Sumew,n) precede Greek type spellings (e.g. ~Ieroso,luma, Si,mwn). Also, transliterated words (e.g. Khfa/j, Messi,aj, or avmh,n) precede translated words (e.g. Pe,troj, Cristo,j, or Greek equivalents of avmh,n). In other words, Semitic spellings and words have temporal priority over Greek spellings and words. This assumption has been used as a criterion for deciding which sayings by Jesus and stories about Jesus are the earlier tradition in relation to source criticism, textual criticism, the

282

Phonology

synoptic problem, the historical Jesus, and the unidirectional development of christological titles. This approach is based on the monolingual view of first-century Palestine and the Roman Near East and on the linguistic unidirectionality hypothesis that the Jesus and Gospel traditions are linguistically transmitted from Semitic to Greek. Also, this view is related to the grammar of writing rather than the grammar of sound (cf. the introduction to Part II). Variant spellings of transliterated words at the phonological level, however, should be observed from the perspective of a variational view, the study of parole, the grammar of sound, or language performance. The phonological variability of transliterated words and the bilingualism of first-century Palestine and the Roman Near East should be taken into consideration. This is because transliteration itself leads to variant spellings due to four major factors (§7.1). Variability of transliterated spellings does not permit us to exclude variant spellings from orthography. Furthermore, the bilingual view implies that transliterated spellings, Semitic words, and Semitic type words were circulated with their equivalents such as translated spellings, Greek words, and Greek type words during Jesus’ ministry and later. This runs against the idea that Semitic traditions ought to have a temporal priority over Greek traditions. Consequently, Semitic traditions in the four gospels and Acts at the phonological level do not support the unidirectionality hypothesis but the interdirectionality hypothesis. In chapter 7, variant spellings of transliterated words in the four gospels and Acts will be considered to be transliterated allolexemes (§7.3.1) from the perspective of the variational view, the study of parole, the grammar of sound, and language performance. Above all, four linguistic factors that cause variant spellings will be investigated (§7.1). Second, three approaches to the variant spellings (i.e. orthographical, variational, and bilingual views) will be illustrated (§7.2). Third, from the perspective of the variational view, variant spellings as transliterated allolexemes will be discussed (§7.3). Then, important variant spellings in the four gospels and Acts will be considered as transliterated allolexemes (§7.4). Last, the relationship between variant spellings as transliterated allolexemes and interdirectionality will be observed (§7.5).

Linguistic Factors of Transliterated Variants in Bilingual Contexts

283

7.1 Linguistic Factors of Transliterated Variants in Bilingual Contexts The bilingualism of first-century Palestine and the Roman Near East brought about variant spellings in the four gospels and Acts when Semitic words or sentences were transliterated into their Greek equivalents. Phonologically speaking, there are four major factors causing phonological variants. A different phonetic system (§7.1.1), representation (§7.1.2), phonetic change (§7.1.3), and dialects (§7.1.4) can cause transliterated spellings to be varied. Since this is not the place to enter into detailed discussions, the four factors will be considered briefly.

7.1.1 Different Phonetic System In treating the transliteration of foreign words in the four gospels and Acts, many scholars have not seriously considered that transliteration from a source language into a target language produces phonological variants. When a foreign word of a source language is introduced into a target language for the first time, the transliterated words are adapted to the phonetic system of the target language. The problem is that the phonological correspondences between the source language and the target language are not a one-to-one function. On this score, different phonetic systems in the source and the target languages lead to phonological variants of transliterated words. When Tübingen, a German city, is transliterated into English, it may be transcribed as Tubingen or Tuebingen. Köster, a German name, is a similar case. Despite the fact that English is one of the languages closest to German (i.e. Germanic family), the English phonetic inventory does not have the two vowel values (i.e. ü and ö) so that the two proper nouns in German are transcribed as the closest available sounds of the English phonetic system. On the contrary, the German phonetic inventory does not have a consonant equivalent to /q/ in the English name “Thatcher.” If a German person who does not know English hears the name in English, he may transcribe it as /Tatcher/ instead of /Thatcher/. Although English belongs to the Germanic family, there are many differences in phonetic values between the languages when a word is transliterated from German to English or vice versa. If both a source language and a target language do not belong to one language family, the phonological correspondences between them will be much more complicated. When English words as a source language are transliterated into Korean as a target language, different

284

Phonology

phonological correspondences occur. Some examples of the phonological correspondences between the English language of the IndoEuropean family and the Korean language of the Altaic family are shown below. (See Table 2) [Table 2] Phonological Correspondence between English and Korean

PHONETIC

ENGLISH

VALUE

WORD

[f]

FLASH

[q]

THINK

[ Ţ]

TOP

[ ƚ]

MONEY

KOREAN WORD KOREAN TRANSLIT-

KOREAN PRONUN-

ERATION

CIATION

[䝢⧮㓂]

[plæǏi]

[䤚⧮㓂]

[huræǏi]

[⧋䋂]

[’tiĎk]

[㞓䋂]

[’siĎk]

[䐇]

[tob]

[䌧]

[tab]

[Ⲏ┞]

[mˬni]

[Ⱎ┞]

[mani]

As [Table 2] demonstrates, [f] of FLASH, an English consonant, is transliterated into [p] or [h] of a Korean transliterated word because the Korean consonant inventory does not have the sound [f]. This means that [f] sounds [p] or [h] to Koreans.1 Even though the [p] is completely different from the [h] in the Korean phonetic system, the sounds /p/ and /h/ are interchangeable in transliteration from English into Korean. The two Korean transliterated sounds can be called “transliterated allolexemes” of the English [f]. In the case of THINK, [q] is transcribed as a double [d] or a double [s] because Korean phonetic system does not have the sound [q] either. This means that [q] sounds like a double [d] or a double [s] to Koreans.2 Even though the phonetic value of a double 1

2

When there is no exact phonological correspondence between a source language and a target language, a transcribed sound takes over one or two phonological features of its source sound. The Korean phonological feature does not include the articulation place, [LABIODENTAL], whereas it includes the articulation manner, [FRICATIVE]. The English sound [f] is [LABIODENTAL] [FRICATIVE]. When the English sound [f] is transliterated into the [p] or [h] of Korean sounds, [p] is [BILABIAL] [PLOSIVE] and [h] is [GLOTTAL] [FRICATIVE]. The [BILABIAL] of the sound [p] is the closest one to articulation place of [f] in Korean consonant inventory. The [FRICATIVE] of the sound [h] is taken from [FRICATIVE] of [f]. In the case of [q], the phonological features of the sound [q] are [INTERDENTAL] [FRICATIVE]. The Korean phonological feature does not include the articulation place, [INTERDENTAL], whereas it includes the articulation manner, [FRICATIVE]. When the English sound [q] is transliterated into a double [t] or a double [s] of Ko-

Linguistic Factors of Transliterated Variants in Bilingual Contexts

285

[d], of course, is also totally different from a double [s] in the Korean phonetic system, the sound [q] is transliterated into the sound a double [d] or a double [s] in transliteration from English into Korean. This means that the two pairs of the two Korean sounds are “transliterated allolexemes” of the English sound [q]. Similarly, in cases of vowel transliteration, the Korean vowel system does not have the English vowel values [ Ţ] and [ ƚ], so the [ Ţ] is transcribed as /o/ or /a/ and [ ƚ] is transcribed as /a/ or /ˬ/. In transliteration from English into Korean what is surprising is that when English consonants and vowels are transliterated into Korean, most of their phonological correspondences do not function in a one-to-one relationship. Nevertheless, what is important is that the target language users, both writers/speakers and readers/listeners, understand the phonological variants of the transliterated words without any confusion. Furthermore, these variant spellings are officially written in the newspapers, journals, and academic books without any confusion. These variant spellings are also formally used in broadcast programs, governmental addresses, and academic discussions without any confusion. This suggests that the variability of variant spellings in transliteration from one language to another in ancient times was predicted and accepted without any confusion. If so, transliterated variants in the first century were much more fluid than previously thought. The phonological variants in transliteration from Semitic into Greek can be produced by two major mechanisms.3 (1) UNDERDIFFERENTIATION is the process in which phonological variants are caused by confused Greek sounds when two or more Semitic sound values in the Semitic sound inventory are distinguished but their Greek equivalents in the Greek sound inventory correspond to one Greek sound value. (2) OVERDIFFERENTIATION is the process in which phonological variants are caused by confused Greek sounds when two or more Greek sound values in the Greek sound inventory are distinguished but their Semitic equivalents in the Semitic sound inventory have only one Semitic sound value. The Hebrew sibilants s, v, f, c usually correspond to only one Greek sibilant s/j. When Hebrew sibilants are transliterated into Greek, s/j is underdifferentiated (cf. Sperber 1937-8:115). On the contrary, s, v, f, and c are overdifferentiated when a Greek sound s/j is transliterated into Hebrew sibilants.

3

rean consonants, a double [t] is [ALVEOLAR] [PLOSIVE] and a double [s] is [ALVEOLAR] [FRICATIVE]. The [ALVEOLAR] of the sound [t] is the closest one to articulation place of [q] in the Korean consonant inventory. The [FRICATIVE] of the sound [s] is taken from [FRICATIVE] of [q]. Cf. Jeffers and Lehiste 1979:139-40, 180-7.

286

Phonology

In this regard, Leclercq (1972-3:188) argued persuasively that “Transliteration cannot presume a uniform pronunciation of Greek in any part of the Greek world at a given period of time.”4 The differences in phonetic systems between Greek and other languages imply that when Greek proper nouns are transliterated into other languages (or vice versa), phonological variants are produced. Because this topic is beyond the scope of the present study, I will limit my illustration of these phenomena to four transliterations between Ancient Greek and other languages.5 As far as irregular phonological correspondence between Greek and Egyptian is concerned, when ancient Egyptians of Roman Egypt wrote / spoke Greek, they would have been confused about voiceless consonants and voiced consonants. Thackeray (1903:100-1 n.1) suggests that “It appears probable that Egyptians, in the early centuries of our era, could not pronounce Greek g and d.”6 In this case, Egyptians would have pronounced the closest sounds in Egyptian to Greek g and d when they transliterated the Greek sounds into Egyptian. Hence, the transliterated spellings of g and d could be varied. Vergote (1984:1387) also points out the confusion of occlusives and fricatives between (b), g, d and (p), t, c and s and /z/ among bilingual Egyptians (cf. Gignac 1976:85, 124). Following Vergote, Fewster (2002:235) mentions, “The t/d

4

5

6

He suggests two reasons: (i) there is no consensus of pronunciation of ancient Greek among modern scholars; (ii) many modern languages pronounce the same Greek letters in different ways. Transliteration from a source language into a target language has been studied in many publications in regard to phonological reconstruction. Generally, Greek pronunciation (Allen 1987); Greek euphony (Stanford 1967); transliteration of Greek NT words into Latin characters (Leclercq 1972-3:187-90); transliteration of Hebrew terms in the Septuagint (Tov 1973); Latin personal proper nouns in NT Greek (Bauckham 2002); transliteration of NT proper names from Greek into Syriac (Burkitt 1912); Greek words transliterated into Hebrew (Rosén 1963); phonological correspondence between Egyptian and Semitic (Muchiki 1999); transliteration of NT proper names from Semitic or Latin into Greek (Williams 1995) have been addressed. Incidentally, according to historical linguistics, there are two methods of reconstruction on the basis of relatedness and regularity hypotheses: comparative reconstruction and internal reconstruction (Jeffers and Lehiste 1979:17-54). Aramaic embedded words in the New Testament literature can be also reconstructed by both methods to a large extent. This is adduced from three bits of evidence. (i) The Greek k is used as equivalents to both the g and k of Demotic in demotic papyri of the second century AD. In other words, Greek g and k is underdifferentiated to Egyptians. (ii) In Sahidic, the equivalent consonants are rarely employed except in Greek words. (iii) Greek papyri show interchanges between k and g, and t and d.

Linguistic Factors of Transliterated Variants in Bilingual Contexts

287

confusion is common in the Greek of Egypt as is that between g/k, p/b, and s/z.”7 In the case of Phrygian Greek, there is a different phonetic system as well in Phrygian as compared with Greek. Brixhe (2002:262) mentions that Phrygians were confused between k and c, p and f, and t and q. They also found it difficult to pronounce the Greek cluster st wherever the cluster is placed. This difficulty results in two variants (Brixhe 2002:263). One is that speakers of Phrygian used prothesis (e.g. ivsth,lhn, eivqh,l[h], Eivstratonikwj, or VHstefa,nw|). The other way is to reduce st to t. The difference in the phonetic systems of Greek and Scythian, which is an Iranian language, is implied by Aristophanes. He (Thesm 1001-1230) described a Scythian archer who spoke broken Greek. Some scholars have suggested that the phonological variants are due to phonological interference.8 Brixhe (1988; cf. Stanford 1967:140-4) presents two points: (i) the Scythians substituted unaspirated p, t, k for aspirated f, q, c. (ii) Ei, h, and i are confused. Austin and Olson (2004:308) add two points: the Scythians neglected (i) initial aspiration and (ii) final –n and –j. This means that the Scythian phonological system caused them to underdifferentiate Greek plosives (p/f, t/q, or k/c). A case of transliteration between Greek and Latin indicates that some Greeks could not pronounce the Latin f due to a different phonetic system. This is mentioned in Quintilian 1.4.14: “The Greeks aspirate the letter f like their phi; for example, Cicero in Pro Fundanio makes fun of a witness who is unable to pronounce the first letter of this name.” 9 Also, Quintilian reports that the different consonant system in Greek (Amphionem) and Latin (Ampionem) might have caused a case of contempt of court (12.10.57): “When the judge asked the uneducated witness whether he knew Amphionem, and the witness said that he did not, the judge repeated the name without an aspirate and with a shortened second syllable and then he knew him well.”10 Consequently, the different phonetic system possibly caused phonological variants in transliteration of Latin to Greek. 7 8

9 10

The phonological change in Egyptian Greek has been vividly discussed, a discussion which is summarized by Horrocks (1997:61-3). For the detailed discussion of the phonological reconstruction of Scythian, see Willi 2003:202-11. Concerning the phonology and phonetics of Aristophanes’ Attic, see Willi 2003:233-41. This translation is Biville’s (2002:83) “Graeci aspirare ‘f’ ut phi solent, ut pro Fundanio Cicero testem qui primam eius litteram dicere non posit irridet” This translation is Biville’s (2002:83-4) “cum interrogasset rusticum testem an ‘Amphionem’ nosset, negante eo, detraxit aspirationem breuitauitque secundam eius nominis syllabam, et eum sic optime norat.”

288

Phonology

In sum, there can be little doubt that transliterations of Semitic words in the New Testament literature into Greek are no exception. First of all, the sound values of both Semitic (i.e. Aramaic and Hebrew) and Greek are not established completely. This issue is still under debate.11 Second, the degree of transliterated differentiation could have been larger than others because Semitic and Greek come from different language families. Most Semitic consonants did not exactly correspond to their Greek equivalents in transliteration, as will be discussed (§7.3.2). When Semitic vowels are transliterated into their Greek equivalents, many more variants would be produced because the boundaries between the vowel sound values were more obscure than those between the consonant sound values. This will be further investigated later.

7.1.2 Representation When a word is transliterated from a source language into a target language, the process of representation may cause variants at phonological, morphological, and semantic levels. Phones at the phonetic level are translated into phonemes by means of phonological rules12 at the phonological levels.13 As is well known in English, although [n] and [m] are totally different at the phonetic level, [n] sounds like /m/ by mediating an English phonological rule. In the case of “input”, unlike “inset,” [n] is realized as /m/ in the process of consonant assimilation on the influence of the bilabial sound [p].14 Similarly, euphony has been said to be one of the factors of sound change (Stanford 1967:74-98, 149-50; Jeffers and Lehiste 1979:89-93). Buck (1965:55-85) also suggests dialectal variants (i.e. sound changes) such as rhotacism, assimilation, dissimila-

11

12 13

14

See Stanford 1967; Leclercq 1972-3:188; Allen 1987; Caragounis 2004. Weinberg (1969-70:1-32) suggests that it is not easy to transcribe and transliterate the sound values of Hebrew as a source language in a unified way. For the discussion of Hebrew sound values and their Greek equivalents, see Sperber 1937-8:103-274; esp. 127132. Hyman 1975:12-5; one of the well-known English phonological rules is this: /s, z, t/ becomes changed into [š, ž, ²] before /y/ when it is pronounced. According to Hyman (1975:2), “A phonetic study tells how the sounds of a language are made and what their acoustic properties are. A phonological study tells how these sounds are used to convey meaning.” The units of phonological representation are “phonemes” which are transcribed between diagonal bars //, whereas the units of phonetic representation are “phones” (i.e. sound segments) between square brackets [ ]; for detailed discussion, see Hyman 1975:8-9. Cf. Hyman 1975:1-15; Carr 1993:13-32.

Linguistic Factors of Transliterated Variants in Bilingual Contexts

289

tion, transposition, doubling, simplification, elision, aphaeresis, crasis, or apocope due to the ease of pronunciation. The primary method of communication in first-century Palestine and the Roman Near East was not written but oral language. Oral transmission of the Jesus and gospel traditions caused variant spellings in transliteration.15 This is so because Greek phones may have become phonemes differently from the written spellings on the basis of some phonological rules. When one investigates the spellings in New Testament literature, s/he should take account of the grammar of sound rather than the grammar of writing. When the audience heard the Jesus and gospel traditions, they did not hear the spellings written in the texts but sounds represented by their phonological rules. In the case of Greek, as Hort (1896:156-7) mentions, when su,n- is pronounced, it is changed into su,m- before bilabial consonants (e.g. b, p, f, and m) in the phonological representation.16 The euphony results in two phonological variants (i.e. su,n- and su,m-). BDF (§19) illustrates another assimilation; su.n Maria,m and su.m Maria,m of Luke 2:5 (complete assimilation); evn Kana, and evg Kana of John 2:11 (partial assimilation). Caragounis (2004:379) proposes that “The k of the preposition evk before B, G, and D as well as before L, M, and N is regularly changed to G for euphonic reasons.” When these Greek phrases are transliterated to Semitic, they would have been transcribed into two Semitic words. Also, Beth Shearim is transliterated into Bhsa,ra, Besara, or Bisara (Mazar 1973:1-2). It seems that the consonant t is absorbed into s for convenient pronunciation because both of them are alveolars, as shown by Bêsân for Beth Sheˬan and Benafsha for Beth Nafsha (Mazar 1973:1-2). A variation occurs due to the secondary articulation (Sloat, Taylor, and Hoard 1978:44-6).17 When the sound [r] is produced in the initial position in English, the lips are rounded. The /w/ is not realized in its written form, but when it is pronounced, it is unconditionally represented in pronunciation. And when an American-English native speaker quickly pronounces WRITING, s/he may pronounce [r] instead of [t] of WRITING because [t] or [d] could be modified into [r] in the 15 16

17

Metzger 1992:190-2; Vaganay-Amphoux 56; BDF §14. For detailed discussion of the oral circulation of the Jesus and gospel traditions, see §1.3; §1.5.2. This phonological representation implies that b, p, f, and m were bilabials, although f may have been regarded as a labiodental among many English scholars. Ironically, as Quintilian mentioned, some Greeks could not speak a Latin consonant [f]. Instead, they pronounced it like a Greek consonant [f]. This means that Greek [f] was different from a Latin [f]. The “secondary articulation” refers to “modifications which are imposed on the primary articulation of a sound” (Sloat, Taylor, and Hoard 1978:41).

290

Phonology

particular context. Consequently, when WRITING is transliterated into Korean, the two secondary articulations could make four transliterated Korean spellings from WRITING, such as in [Table 3]. [Table 3] Four Transliterated Korean spellings of ‘Writing’ + ROUNDED

- ROUNDED

[t]

[rwaitiĎ] ⫎㧊䕛

[raitiĎ] ⧒㧊䕛

[t] > /r/

[rwairiĎ] ⫎㧊Ⱇ

[rairiĎ] ⧒㧊Ⱇ

The four transliterated Korean spellings are predictable, explainable, and acceptable. Although [raitiĎ] is common, the other three spellings are acceptable to the learned who know especially the phonological realizations of English [r] and [t] in the process of phonological representation. The secondary articulation causes the transliterated word to be varied. This seems to be related to the variant spellings of Nazareth; Nazara, can be changed to Nazwrai/oj due to secondary articulation (§7.4.1.2; §8.4.3.2). In the case of the differences in Jesus’ cry of despair, Williams (2004b:1-12) persuasively argues that the different spellings in Mt 27:46 and Mark 15:34 are caused by the phonological representation of oral sounds (§7.4.3.1). Consequently, it should be considered that many phonological rules in Semitic and Greek languages caused spellings to be varied in the phonological representations in transliteration from Semitic into Greek. Morphological representation18 is related to variant spellings. When a word is transliterated, the spellings can be different whether morphemes are transliterated or not. Ilan (LJNLA 22-8) presents detailed instances of transliterated variations from Hebrew or Aramaic into Greek in the process of morphological representation (e.g. declensions, feminine suffixes, or Semitic suffixes).19 Rahmani (1994:13) points out that Goliath (nos 799, 800) from Jericho is not inflected as in the Septuagint, which is different from Goliathos or Goliathes in Josephus (Ant. 6:171, 177). The two variant spellings of Nazareth (i.e. Nazara, and Nazare,q) may be explained due to morphological representation of 18

19

Morphology is usually defined as a “branch of linguistics that is concerned with the relation between meaning and form, within words and between words (Lardiere 2006:59).” She (LJNLA 18) suggests the tendencies of suffix transliteration of personal proper nouns found in the Septuagint, Josephus, and the NT. In the Septuagint, when the personal proper nouns were transliterated into Greek, they were not declined. In the case of Josephus’ works, the transliterations of the personal proper nouns are more Hellenized and their suffixes tend to decline. The authors of the NT tend to be closer to the Septuagintal style, but when it has a Hellenized form, it declines.

Linguistic Factors of Transliterated Variants in Bilingual Contexts

291

Hebrew itself because both –eth and –ah are feminine endings (§7.4.1.2; §8.4.3.2). It seems that i of Taliqa koumi (muted case ending) might be added when the scribe took its morpheme into consideration (§8.3.2). In the case of representation at the semantic level, the proper nouns of the synoptic gospels and Acts are represented in two ways. Most proper nouns are usually transliterated from Semitic words into Greek words. Personal proper nouns, on the other hand, may also be translated. Mussies (1994:249) assumes that two spellings, Mna,swn (Acts 21:16; CPJ28 1.17) and Mn[i]ase,aj (CIJ 508) could be regarded as a translation of Zakaryah and as a transliteration of Manasseh, and that Justus (Acts 1:23) may be derived from translating Zadok or from transliterating Joseph. Messiah and Christ (§7.4.3.2), and Cephas and Peter (§7.4.3.2) have two forms (i.e. a transliterated and a translated form). The two styles lead to two spellings for one proper noun. It is interesting that when a word from a source language is transliterated into a target language, transliteration is sometimes influenced by etymology. Etymology (e.g. semantic interference or onomatopoeia) may cause variants. As is well known, the hiero- of the Greek spelling of Jerusalem is relevant to “Holy” (Holy Solyma or Temple of Solyma) among Greek writers (cf. Smith 1907-8:1.261-2; cf. Sylva 1983:214). In fact, -ry should have been transliterated into the light breathing (VIerou-) rather than the rough breathing (~Iero-). The popular etymology may cause two spellings of Jerusalem. In terms of transliteration of Boanhrge,j (Mk 3:17), the oa remained unsolved. The shewa can be transliterated into a or o. This means that both Bane- and Bone- are possible transliterations, whereas oa cannot be represented. However, Zimmermann (1979:10) suggests that oa in Boanhrge,j was added by a Greek scribe because he considered the word (i.e. sons of thunder) as boao, boa, or boe which is related to great noises (e.g. war, wind, and waves). Taking a further step, Buth connects this transliterated term with popular etymology. He (1981:29) suggests that “Such a play on words could have influenced someone to break with conventional transliteration and to derive the ‘unexplainable’ form Boan- from -ynb.”20 Taking a further step, in relation to the Greek transliteration (i.e. ~Ieroso,luma) of Jerusalem (§8.4.3.1), Sylva (1983:216) convincingly maintains that ancient etymology does not show the exact phonological correspondence between a 20

Buth (1981:30) argues that the author of Boan- might be a Greek copyist rather than Mark because “Mark would be more conscious of the Semitic source than a Greek copyist.” It seems, however, that the inventor of the term could be Mark himself, because a bilingual makes a pun here by means of transliteration. More possible is that Jesus, who might be a bilingual, could have used Boan- as codeswitching; cf. the bilinguality of Jesus (§3.2.4) and codeswitchings in the NT (chapter 8).

292

Phonology

source word and a target word. Popular etymology may cause phonological variants. Moreover, phonological representation in ancient times made spellings more frequently varied. When texts were copied, they were read aloud and copyists transcribed what they heard (Gamble 2004:30-1). In other words, the copyists did not reproduce a phonetic representation but a phonological representation, to which phonological rules already had applied. In this regard, modal interdirectionality between oral and written traditions promoted various spellings. Accordingly, variant spellings did not result from orthographic errors but from phonological representation due to modal interdirectionality between oral and written tradition. The study of spellings of the New Testament literature should be investigated from the perspective of the grammar of sound rather than the grammar of writing.

7.1.3 Phonetic Change Phonetic changes always occur in any language, which makes it difficult to reconstruct the original sound values of an ancient language. As Voelz (1984:939) points out, although Gospels were written in the first century, the authoritative manuscripts were written at least several centuries later. This gap makes it more difficult to trace the sound values of Hellenistic Greek. Concerning phonetic change, we should mention Gignac’s suggestions (1976:58-9): (i) Although there is no exact way to reconstruct the precise phonetic values of the original or transitional sounds, the study of comparative phonology of cognate sounds of related languages can provide us with reconstructed sound values. (ii) When a phonetic change is not reflected in writing, it is difficult to trace the changed sound. When it is transliterated, it may produce variant spellings. (iii) It is hard to decide the precise time limits of sound changes. (iv) Sound changes do not occur extensively and simultaneously but in localized areas and for a particular period. Generalizations should not be made. (v) Phonological variations may occur due to mistakes or slips of the pen. Concerning phonetic change, some scholars have suggested that the pronunciation of Hellenistic Greek might be different from that of classical Greek.21 Concerning the sound change in Rabbinic Greek, Rosén 21

Although this is a knotty problem, what is certain is that phonetic values are changing. For detailed discussion of changing pronunciation of Hellenistic Greek, see Caragounis 2004; Bubenik 1989; BDF §§17-35; Robertson 1934:236-41; Forster 1922:108-15.

Linguistic Factors of Transliterated Variants in Bilingual Contexts

293

(1963:68) suggests that there were consonant changes from nd or nt into d; mb or mp to b; Ďg or Ďk into g. He shows that the Septuagint transliterates Hebrew b into mb in (e.g. qwqbx = VAmbakoum) and that this is the same with Talmudic Greek (e.g. tlbs = sum-bolh,; twnwlbs = su,m-bola; alwba = e;m-boloj; yjbma = evm-bath,). Moreover, the sound interchanges in Hellenistic Greek can cause phonological variants in New Testament Greek.22 In relation to Egyptian Greek, Gignac (1976-81) suggests that many sounds were used interchangeably. He (summarized by van der Horst 1991:25-7) illustrates the interchanges of ai and e, ei and i, h and i/ei, h and e/ai, oi and u, u and ei/i/h, ei/i and e/ai, w and o, a and e, a and e, a and o, a and au, and g and i, etc. According to GLICM (27-8), Greek inscriptions from the Roman and Late Antique periods excavated from Caesarea Maritima also show similar phonetic changes: (i) i, h, u, and ei sound alike and were confused in spelling. (ii) Itacisms were found.23 (iii) There were interchanges of i and h, i and e, i and ei, h and ei, i and oi, and ei and oi. (iv) Diphthongs had a tendency to change into monophthongs between o and ou, u and oi, u and ui, u and h, e and ai, and e and eu. (v) The diphthong au was expanded into auou. (vi) Some vowel quantities were interchanged between e and h and o and w. In the case of consonants, there were some phonetic changes: failure of assimilation of n to g, losses of liquid l, nasal n, and sibilant s, simplification of doubled nasal n, and germination of s. Caragounis (2004:365-382) also suggests interchanges between ei and i, h and e/i, g and i, u and i, oi and i, hi and i, o/ou and w, oi and wi, ai and e, au/eu and hu, and etc. Also, Greek inscriptions at Beth Shearim (about 10 miles from Nazareth) show phonetic changes. Schwabe and Lifshitz (1974:201-3) maintain that: (i) Iotacisms are found. Vowels ei, h, oi and u are exchanged for i and vice versa. (ii) Conversely, i disappears. (iii) Some vowels replaced others: oiu was used for eu, o for ou, o for u, o for a, ou for w, a for e, e for a, and w for e. In the case of consonants, (i) haplography is common. (ii) There are interchanges between q and t and between c and k. (iii) P is confused with b. (iv) K is weakened into g and g into i. (v) C is replaced by ks. (vi) Dissimilation at a distance is shown. Consequently, the phonetic changes of a source language and a target language occur in various ways. When proper nouns are transliterated, the phonetic changes may cause transliterated words to produce 22 23

Hort (1896:157-62) shows interchanges between consonants and vowels of the New Testament texts. On the basis of Jewish names revealed by excavations of Palestinian sites dating from 330 BC to 200AD, Ilan (LJNLA 21) states that “One of the most common features of Koine Greek is itacism. This phenomenon is already evident in many of the name variations found.”

294

Phonology

various spellings. Accordingly, Schwabe and Lifshitz (1974:203) reasonably conclude that “the transcription in the New Testament is by no means consistent.”

7.1.4 Dialects All languages have their dialects.24 Both the source languages (i.e. Semitic) and the target language (i.e. Hellenistic Greek)25 in the first centuries had their dialects as well.26 It is well-known that B.Erub. 53b provides evidence that Galileans were notorious for mispronunciation of gutturals and vowel sounds, which is adduced from Peter’s mispronunciation (Mark 14:70; cf. Mt 26:73): Now, as for that Galilean who said: ‘Who has ´amar? Who has ´amar?’ They said to him: ‘Galilean fool! (Do you mean) an ass [ham¬r] to ride on? Or wine [hamar] to drink? Wool [ˬamar] for clothing? Or a sheepskin [´ômar] for a covering?’ (Taylor 2002:303)27

This means that the pronunciations of x, [, h, and a were not differentiated among Galileans, with the result that they could understand them but not articulate them. Interestingly, Speiser (1933b:238) mentions that “It is, doubtless, in self-defense that the Galilean scholars chide the Babylonian Hiyya for his failure to differentiate between x, [, and h.” Dalman (1905:57-61) suggests that the inhabitants of Beth Shearim could not articulate gutturals. Mazar (1973:1-2) also states that the original spelling of Beth Shearim included [, although the inhabitants could not speak the guttural. The Egyptian dialect can cause Greek variant spellings. Taylor (2002:312) proposes that “while some of these forms are attested in other Aramaic dialects, they are best explained not as examples of interference but as defective spellings, since some are

24

25

26

27

In reply to the question “Did dialectic differences persist?,” Moulton (1919:39) maintains that “we may reply in the negative. Dialectic differences there must have been in a language spoken over so large an area. But they need not theoretically be greater than those between British and American English.” For the discussion of Greek dialects, see Buck 1965; Barton¼k 1972; Smyth 1974. Sperber (1937-8:149-53) discusses Hebrew dialects. Also, Lehmann and Holum (2000:27-8) mention some phonological peculiarities of Greek inscriptions of Caesarea Maritima. Marcos (2000:14) may observes that “there were greater degrees of dialectal differentiation than we know through the process of linguistic uniformity imposed by a great section of literary koiné and the way of speaking well and writing well spread by the Atticist movement.” For more illustrations of Galilean mispronunciation, see Bockmuehl 2005:63-4 n.56.

Three Views of Variant Spellings in Transliteration

295

also to be found in the early Syriac inscriptions of Edessa”28 It is well known that the transliterated difference between taliqa koum and taliqa koumi of Mark 5:41 is due to dialectic differences (§8.3.2). Reed (2000:124) persuasively suggests that “Variant spelling practices probably due to pronunciation differences, as found in ancient Greek papyri, reveal such varieties of dialect in Egypt.” Even Robertson (1934:178) proposes that the Greek dialects “explain some variations in orthography,” although his view is based on an orthographical approach. Accordingly, dialects produce transliterated variations.

To sum up, it is normal that transliterated spellings have their variant spellings at the level of phonology. The variant spellings are caused by four major factors: different phonetic systems, representation, phonetic change, and dialects. In these cases, the variability in transliteration is predictable, explainable, and acceptable.

7.2 Three Views of Variant Spellings in Transliteration The fact that Aramaic was used as a matrix language in first-century Palestine and that the four gospels were written in Greek produced many Semitic transliterated words in the Greek gospels. The Semitic transliterated words have their variant spellings, which have attracted New Testament scholars’ attention. Scholars have considered the variable spellings of Semitic transliterated words three ways. Traditionally, variant spellings have been investigated from the perspective of the orthographical view on the basis of categoricity theory (§7.2.1). On the other hand, the variational view on the basis of variationism has been discussed by some scholars (§7.2.2). Last, the bilingual view will be suggested (§7.2.3).

7.2.1 Orthographical View Many scholars have assumed that the single original spelling, the single original reading, and the single original text of the New Testament can be reconstructed by means of external and internal methods (see chap-

28

Taylor (2002:308) mentions that “This suggestion that the failure to write accurate Greek does not necessarily imply a lack of ability in spoken Greek is reinforced by a short inscription on a lintel in Bashakuh, Jebel Barisha, where the wordplay in part depends upon the local pronunciation/orthography.”

296

Phonology

ter 7 n.5).29 As Barr (1989:187) points out, “this century the American school in the tradition of Albright” has held the orthographical view. This view is supported by Cross (1952, 1964, 1975) and Freedman (1962, 1969). They have posited that official spellings of epigraphic evidence can provide us with historical reconstruction and that the evidence of a chronology of orthographical developments can confirm the date of the original text. Selecting the “original” from the “secondary” has also been applied to the investigation of Semitic transliterated words without exception. Scholars usually would have ascribed the phonological variants of Semitic transliterated words to “phonological errors.” These errors have been caused either by evangelists in the transmission of the Jesus tradition or by copyists in the transmission of the gospel tradition. This orthographical approach to the variant spellings of Semitic transliterated words has been unceasingly supported by authoritative publications. Many scholars have dealt with the phonological variants of Semitic transliterated words under the title of “orthography,” “errors,” or “corruption” (Hort 1896:148-79, Thackeray 1909:71-139; cf. 160-71, Robertson 1934:177-245, BDF §§36-42, Vaganay-Amphoux 52-61, and Metzger 1992:186-206). From the perspective of the orthographical view, Leon compares Roman Jewish epitaphs in Greek and Latin. He (1960:87) notes that “we find far fewer examples of aberrations from correct Latin pronunciation and grammar“ [emphasis added]. Analyzing Jewish names in Late Antiquity, Ilan (LJNLA 1.16) also holds the orthographical view, when she states that “If Greek orthography corresponds to the official orthography of the name as recorded in the LXX translation, this is noted in the footnotes“ [emphasis added]. She (LJNLA 1.16) complains that “scribes were very free in their transliteration of biblical names,” but it is normal that transliteration itself from a source language into a target language is free to a great extent. Consequently, the scholars who hold the orthographical view assume that the correct spelling should be singled out from corrupted 29

Westcott and Hort (1896:303) submit that “We have therefore thought it best to aim at approximating as nearly as we could to the spelling of the autographs by means of documentary evidence.” Souter (1954:3) states that “Textual criticism seeks, by the exercise of knowledge and trained judgment, to restore the very words of some original document which has perished.” Aland and Aland (1989:214) resolutely propose that “Only one reading can be original.” Klijn (1966:103) claims that “Textual criticism takes into account external evidence and internal evidence. At present the significance of internal evidence for restoring the original text is emphasized.” Recently, Greenlee (1995:1) considered that “Textual criticism is the study of copies of any written work of which the autograph (the original) is unknown, with the purpose of ascertaining the original text.”

Three Views of Variant Spellings in Transliteration

297

spellings, since they assume that the correct spelling is the original spelling used by the four evangelists. The original spelling is the only single correct spelling. Furthermore, scholars have made efforts to single out a Semitic spelling from variant spellings of transliterated words in the four gospels and Acts. They have done this because they presumed (i) that a Semitic spelling is an earlier spelling than its equivalent Greek spelling and (ii) that a Semitic tradition is closer to the original sayings by Jesus and the original stories about Jesus than Greek tradition. These assumptions are based on the presupposition that the linguistic transmission of the Jesus and gospel traditions is unidirectional from Semitic to Greek. Many scholars, on the one hand, have considered the Semitic spelling to be the original from the perspective of the monolingual view. It has been suggested that “final consonant mute forms” of three proper nouns (i.e. Nazara (§7.4.1.2), Gennhsar (§7.4.1.3), and Elisabh (§7.4.2.2) reflect Semitic spellings. This means that the Semitic spellings are the more original and more correct spellings. Fitzmyer (1981-5:530) suggests that Nazara “may reflect a more ancient Semitic form of the name.” BAG (156) and BDF (§39(2)) suppose that Gennhsar should be regarded as the more correct form. BDF (§39(2)) also mentions that Elisabee/Elisabh is the original. On the other hand, it has been presumed that the Semitic type spellings (e.g. Sumew,n, Khfa/j, and VIerousalh,m) are earlier spellings than their equivalent Greek type spellings (e.g. Si,mwn, Pe,troj, and ~Ieroso,luma) respectively (§7.4.2.1; §7.4.1.1). The orthographical view is deeply related to the Sitz im Leben and the linguistic unidirectionality hypotheses of the transmission of the Jesus and gospel traditions. Scholars have supposed that the authors and the readers of the gospels were Greek speakers. They have supposed that the synoptic gospels were written in Greek monolingual contexts (e.g. Alexandria, Antioch, Galilee, Jerusalem, Rome, or Syria; see §4.1). The Matthean community, for instance, is said to be located in Syria. Kilpatrick (1946:104) proposes that “All this evidence taken together implies that the community was Greek-speaking rather than bilingual.” The monolingual view of linguistic provenance leads to the assumption that Semitic spellings are older than Greek spellings at the level of phonology. Furthermore, if one takes the grammar of sound instead of the grammar of writing, s/he will take it for granted that most transliterated words could have their variants. Accordingly, the orthographical view is based on the categoricity theory, the study of langue, the grammar of writing, and language competence. Spellings are transmitted from Semitic into Greek in a unidirectional way.

298

Phonology

7.2.2 Variational View Some scholars, however, have taken the phonological variants of Semitic transliterated words into consideration on the basis of variationism. Hebrew Bible scholars rather than New Testament scholars have made the major contribution to the variational approach because they have dealt more with issues related to Semitic transliteration in the Septuagint. Three scholars will be mentioned briefly.

7.2.2.1 Ephraim Speiser Speiser (1926; 1933a; 1933b) showed his pivotal insight into the study of transliteration from Hebrew into Greek from the variational angle. Regarding the transliteration of the Hexapla, he (1926:360) observed that “There is (sic!) a number of variations in the transliteration of the same word in different passages.” He proposes that transliterations of the Hexapla are inconsistent. What is significant is his variational view of transliteration. He posits that the transliteration from one language into another naturally leads to variants. He suggests three reasons for this: phonological representation, disagreement of phonetic correspondence, and sudden readjustment. First, he states that the variant spellings result from phonological representation in the process of transliteration. Mentioning the statement of Jespersen, who was one of the most renowned linguists of his time, he (1926:360) maintains that “the pronunciation of a given word depends as much upon the position of the latter within the sentence as the exact sound of a given vowel or consonant varies with the character of the neighboring sounds.” Linguistically speaking, phones at the phonetic level are realized into phonemes in a given word in the process of phonological rules so that the phonemes are varied (cf. §7.1.2). Second, the phonetic correspondences from Hebrew to Greek do not function in a one-to-one relationship, which leads to phonological variants (cf. §7.1.1). He (1926:361) contends that “it is self-evident that no two corresponding sounds of any two languages are, strictly speaking, identical.” What is more, he (1926:361) claims that even “similar sounds differ nevertheless in every language.” This means that correct transliteration itself is impossible from the outset. Last, Speiser (1926:362) surprisingly submits that “This constantly increasing disparity results in sudden readjustments.” This opens the possibility that scribes would have readjusted the transliterated spellings according to their own habits or principles without hesitation. Accordingly, he convincingly concludes that it is normal that transliteration from Hebrew into Greek causes variant spellings be-

Three Views of Variant Spellings in Transliteration

299

cause of phonological representation, phonetic correspondence, or readjustment. Unfortunately, little scholarly attention has been given to Speiser, despite the fact that his variational view of transliteration from Semitic to Greek was so vital. His view should have been developed further.

7.2.2.2 James Barr Although Barr holds the categoricity theory at the semantic level (1961), he (1989; cf. 1985) takes the variational view at the phonological level. He persists in the view that there is no orthographical pattern or rigid laws of spellings in the Hebrew Bible. Instead of this, the spellings are random because of scribal style (1989:194). Consequently, Barr (1989:204) argues that variable spellings should not be decided from the perspective of orthography (i.e. right or wrong spellings). He demonstrates this with many cases at the lexical level. In the case of toledot (1989:2), it occurs irregularly in Genesis: twdlwt (2:4), tdlwt (in the next five occurrences, 5:1, 6:9, 10:1, 11:10, 27), tdlt (25:12), tdlwt (25:19), and twdlt (next three occurrences, 36:1, 9, 37:2). In terms of proper nouns, Barr (1989:161-7) states that spellings of personal and local names are inconsistent and unpredictable. 30 As for Absalom with and without waw, it changes seven times within the eight verses 2 Sam 16:16-23 (Barr 1989:23). Many personal and local proper nouns preserve the variant spellings (Bar 1989:161-7). Even the spelling of David has variants (with and without yod; cf. 1989:166, 197). As a consequence, he (1989:24) strongly insists that “Surely no ‘official’ person or body of officials or official decision could have opted for the sort of alternation we have seen to exist, within one book or passage.” Taking a further step, Barr (1989:187) proposes that variable spellings do not provide temporal priority for any text.

7.2.2.3 Alan Millard Accepting Barr’s proposal, Millard raises the question whether the irregular occurrences of spellings are common phenomena in ancient times, as spellings of the Hebrew Bible were. He (1991) replies that the unconscious irregularities of spellings are also found in early Hebrew inscriptions, early Aramaic texts, and other ancient writing systems 30

I will suggest that variant spellings of transliterated words are predictable to some extent (§7.3.1).

300

Phonology

(e.g. West Semitic languages, Babylonian cuneiforms, or Egyptian hieroglyphs). He (1991:114-5) decisively concludes that arbitrary spellings “can now be seen to be part of a common feature of ancient near eastern scribal practice,” and that where “there were no rigid rules, variable spelling was permissible.”

7.2.2.4 Some Scholars of New Testament Greek The variational view of Hebrew Bible scholars is also of significance to New Testament literature. In regard to the New Testament literature, not all scholars have disregarded the variational approach to variant transliterated spellings of the literature. Raymond Brown (1993:209) convincingly suggests the variability of transliterated spellings in two points. First, although scholars have considered that the Aramaic embedded spellings are transliterated by “strict phonological rules,” their derivations are “rarely accurate by scientific criteria.” Second, scholars have insisted that transliterated words occur “with exclusivity,” that is, that one variant is right but another is wrong. Transliterated spellings, however, should not be judged from the view of “either … or” because the spellings are “often a ‘both…and,’ rather than an ‘either … or.” As previously mentioned (§7.1.1), phonological correspondences do not have a one-to-one function, so that the transliterated spellings are variable. Brown (1994:1052) maintains that “transliteration of Semitic vowels and consonants was not an exact procedure.” He applies the variational view to interpretation of “the cry of the cross” (§7.4.3.1). In relation to transliteration between Semitic and Greek/Latin, Bauckham (2002) interestingly proposes the sound equivalency theory that when a Semitic name was transliterated, one selected a Greek/Latin name which was the closest sound to a Semitic pronunciation of the Semitic names. He illustrates Paul’s three names (i.e. Saou/loj, Sau/loj, and Pau/loj; cf. chapter 8 n.124 and Hengel 1991:9). To his credit, Bauckham takes into consideration a grammar of sound. His argument, however, is based partly on the variational view. He does not explain the phonological representation for why the sounds were considered the closest pronunciations on the basis of comparative phonetics between Semitic and Greek/Latin languages. Williams (2004b) also holds the variational view. He tries to apply it to analysis of variant spellings of the cry of despair on the cross (§7.4.3.1). Concerning the examination of Aramaic transliterated words in the New Testament literature, he (2004b:8) observes: The phenomenon of alternative written realizations of the same sound is nothing like as common in English (which has largely standardized spell-

Three Views of Variant Spellings in Transliteration

301

ings) as it was in Koine Greek. Moreover, an even greater number of possible realizations would have existed for representing non-Greek words in Greek. … To insist that one must be the better or worse transcription may be rather beside the point.

Consequently, (i) a Semitic spelling of a source language (i.e. Semitic) may not be the single original Semitic spelling, but the original Semitic spellings because the Semitic spelling itself can be variable. (ii) Transliterated spellings of Semitic words in the four gospels and Acts can be variable. (iii) The original spelling may not be the correct spelling, and the secondary spelling may not be a corrupted spelling. This is so because transliterated spellings as such cannot be judged by exclusivity. These views deny the traditional orthographical view that we can single out the single original spelling, the single original reading, and the single original text of the New Testament. When it comes to the investigation of transliterated words, Leiwo (2002:182) rightly contends that “A descriptive analysis is far more important than a prescriptive one.”31 Variant spellings of transliterated terms of the four gospels and Acts should not be investigated from the perspective of the orthographical approach, the study of langue, the grammar of sound, and language competence but from the variational approach, the study of parole, the grammar of writing, and language performance. However, if scholars do not take the bilingual view below into consideration, despite their variational view of the New Testament, they cannot escape from the orthographical view completely.

7.2.3 Bilingual View The bilingualism of first-century Palestine and the Roman Near East should be taken into consideration seriously in dealing with transliterated words in relation to the linguistic transmission of the Jesus and gospel traditions. Unfortunately, scholars largely have been unaware of the bilingual view and its logical implications. The bilingual view does not seem to support the orthographical view based on the linguistic unidirectionality hypothesis that a Semitic spelling has temporal priority over a Greek spelling. The variational-bilingual view implies two points. First, Semitic words may have been translated into Greek words. The Semitic spell31

Leiwo (2002:182) also posits that “The main point … is not the ‘correctness’ of the Latin or the Greek, but the forms of deviation from the language generally used in these funerary inscriptions, which can tell us something about the varieties used within these communities.”

302

Phonology

ings (e.g. Messi,aj) could have been circulated with their equivalent Greek spellings (e.g. Cristo,j) in first-century Palestine and the Roman Near East. This means that Semitic spellings were not always earlier than Greek spellings. Second, bilingual authors and readers could have recognized both Semitic and Greek spellings without any difficulties because any bilingual can transliterate a word of a source language to a target language. Also, authors and readers in all likelihood knew the variability of transliterated spellings without any confusion because it is known that phonological correspondence between Semitic and Greek is not a one-to-one function (§7.1.1). Going a step further, whenever authors and scribes met unfamiliar transliterated spellings, they might have preserved, adjusted, or readjusted the transliterated spellings in their own styles (cf. Speiser 1926:362), which could have caused a new variant. In this respect, it is necessary to consider the variant spellings in the New Testament literature to be transliterated allolexemes.

7.3 Variant Spellings as Transliterated Allolexemes in Bilingual Contexts Here, I will take both the variational and bilingual approaches to the analysis of transliterated spellings in the four gospels and Acts. First of all, variant spellings will be considered as transliterated allolexemes (§7.3.1). Then, consonants of variant spellings as transliterated allolexemes will be discussed (§7.3.2). Lastly, vowels of variant spellings as transliterated allolexemes will be investigated (§7.3.3).

7.3.1 Transliterated Allolexemes In regard to the transliteration from a source language into a target language, a “variant” can be defined as a word which includes one or more different “distinctive features.” 32 Nazaret and Nazared are two variants of the six Greek spellings of Nazareth. The difference in the two spellings is only the one single distinctive feature, [sVOICED]. Their distinctive features can be analyzed: the t is [-VOICED] [ASPIRATED] [+PLOSIVE] [+ALVEOLAR], whereas the d is [+VOICED] [-ASPIRATED] [+PLOSIVE] [+ALVEOLAR].33 This means that the [s 32 33

A “distinctive feature” is usually defined as the most basic unit of phonological structure. It seems that Greek alveolar plosives (d, t, and q) can be classified by two distinctive features, [ASPIRATED] and [VOICED]; d is [‫ڈ‬ASPIRATED] [+VOICED]. T is [‫ڈ‬

Variant Spellings as Transliterated Allolexemes in Bilingual Contexts

303

VOICED] makes their spellings different. The different spellings of Nazaret and Nazared can be called “minimal pairs.”34 In the case of the two Greek common nouns, te,rma and de,rma, this minimal pair has different meanings: te,rma means “end,” whereas de,rma means “skin.” 35 However, the two variant spellings of “Nazareth” have the same meanings despite different consonants (i.e. t and d), unlike the two common nouns. When trcn is transliterated, the two consonants, t of Nazaret and d of Nazared are varied but predictable without any semantic change. These can be called “transliterated allophones.”36 Furthermore, the two transliterated spellings of trcn can be called “transliterated allolexemes.”37 The two spellings are varied but predictable in a transliterated context without any semantic change. The variant spellings of Nazareth (i.e. Nazared, Nazaret, and Nazared) show a variability that can be predictable and explainable without any semantic change. The three consonants (d, t, q) share the two distinctive features, [+PLOSIVE] [+ALVEOLAR]. The consonants z or x never occur as a transliterated allophone of t of trcn because the distinctive features of z or x are totally different from those of d, t, or q. The selection of d, t, or q rather than z or x is decided by phonological correspondence rules according to the phonetic system of the source language (i.e. Hebrew) and the target language (i.e. Greek). Consequently, the variant spellings, d of Nazared, t of Nazaret, and q of Nazareq, can be called transliterated allophones of t of trcn and the three variant spellings can be called transliterated allolexemes of trcn, not corrupted spellings, or erroneous spellings. The three transliterated variants are acceptable. From the perspective of the variational theory, the study of parole, the grammar of sound, and language performance, variant spellings in transliteration should be considered to be transliterated allolexemes

34 35

36 37

ASPIRATED] [-VOICED]. And q is [‫چ‬ASPIRATED] [-VOICED]. The sound values of the Greek phonetic system are still under debate. For recent other opinions, see Allen 1987:12-88; Caragounis 2004:352-83. “Minimal pairs” are defined as “Pairs of words that differ in only a single sound in the same position (Zsiga 2006:38). Two words, “try” and “dry” are a minimal pair. Allen makes a distinction between voiceless unaspirated (p, t, k) and voiceless aspirated (f, q, c), when he (1987:14) states that “Their distinctiveness is demonstrated by minimally different pairs such as po,roj/fo,roj, pa,toj/pa,qoj, and le,koj/le,coj. Allophones refer to “contextual variants,” which means that allophones are varied but predictable without any semantic change (Hyman 1975:7-8). I have coined the term “transliterated allolexemes” in analogy to allophones. Analogies of allophones are used in other sciences. Salzmann (1998:81-3) introduces the coinages in analogy to allophones: forms versus alloforms and facts versus allofacts in archaeology; motif versus allomotif in folklore.

304

Phonology

when they are predictable, explainable, and acceptable.38 This means that the original spelling cannot be the correct spelling but one of the many conventional correct spellings. The corrupted or erroneous spellings could be one of the correct spellings when they are predictable, explainable, and acceptable. When we investigate transliterated variants in New Testament literature, we should not take the orthographical view but the variational-bilingual view. Some instances of transliterated allolexemes of phonological correspondence between Semitic and Greek will be illustrated in order to show variability and predictability. Since this chapter is not the place to become involved in the phonological correspondence between Semitic and Greek for its own sake, some cases of consonants and vowels from BDF, LJNLA (330 BC -200 AD), and NA27 will be considered briefly.39

7.3.2 Consonants Consonants are classified according to three factors, namely, articulatory manner, articulatory place, or voicing. When consonants are transliterated from a source language into a target language, authors and scribes choose the closest sound of the phonetic system of the target language. However, they cannot but prefer one or two factors of them since all three factors of the consonant of the target language do not completely correspond to those of the consonant of the source language. As a result, the transliterated spellings depend on the authors or scribes’ preference for the factors to some degree. For instance, in the case of the transliteration of Nazareq, the scribe (e.g. D*) of Luke 4:16 took d, the distinctive features of which are [PLOSIVE] [ALVEOLAR] [+VOICED] instead of t or q [-VOICED]. Depending on consonantal classification by articulatory manners, transliterated variations of consonants will be considered.

38

39

Also, abbreviated forms of a name could result in two spellings (e.g. VIwsh.f of Mt 13:55 and VIwsh/j of Mk 6:3). Concerning the change of the name from VIwsh.f into VIwsh/j, Bauckham (2006:78) explains that the writer intends to distinguish him (VIwsh.f) from his father (VIwsh.f). For more information of transliteration from Hebrew to Greek or Latin on the basis of Tiberian vocalization, see Sperber 1937-8:103-274. He illustrates the variant spellings caused by transliteration with the cases from the Septuagint, the Second Column of Origen’s Hexapla, and St. Jerome.

ȱ

Variant Spellings as Transliterated Allolexemes in Bilingual Contexts

305

7.3.2.1 Plosives and Fricatives Semitic plosives and fricatives (e.g. b, p, d, t, j, g, k, q) are transliterated into Greek plosives and fricatives (e.g. f, q, c, p, t, k, b, d, g).40 Bilabial interchanges between f and p occur. BDF (§39(2)) presents Kafarnaou,mKapernaou,m, pa,sca-faska-pascw,r-fassou,r, and Sarefqa-Sarepta. Interestingly, when a word is transliterated from Greek into Hebrew, the plosive b occasionally transliterates into p or f. In LJNLA (20), two cases are given as examples: swlqypa for :Amfikloj and swblq for Kleopa/j. Following N. Cohen’s suggestion that usually t was transliterated into q, whereas j into t, Ilan (LJNLA 1.19) also demonstrates alveolar interchanges by using the following five variations: Ma,rqa-Marat, Naqana,hloj-Natana,hlou, Sabaqeo,n-Sabbatai/oj, Seqi,-Shtou, Mattaqi,ajMatqi,aj-Maqqai/oj-Maqeq[oj]. BDF (§39(2)) also suggests the case that t transliterates t (e.g. sa,bbata). Furthermore, BDF (§39(2)) states consonant interchanges between q and t, Elisabeq-Elisabet, Sarefqa-Sarepta, Nazareq-Nazaret, and Gennhsareq-Gennhsaret. On the other hand, in transliteration from a Greek word into Hebrew, it is known that t is generally transliterated into j, whereas q is translated into t. However, the similar phonetic values of both the source language and the target language make them mixed. LJNLA (21) cites some cases: !yjXd from Dosi,qeoj, swjwn- swtwn from No,qoj, hrjnp from Panqh,raj, ymlt from Ptolemai/oj, and symljpa-symltpa from Eu;tolmoj. As for interchanges of velar sounds, the interchanges between k, c, and kc occur. It is usual that c transliterates k (e.g. Kafarnaou,m). LJNLA (19) shows that $wrb is transliterated into Barou/coj, hyrkz into Zacari,aj, abswk into Cwsiba/, abkwk into Cwcebaj, atwkm into Macouqaj, and hyklm into Ma,lcoj. However, it still has variants, as shown by Cananai,a / Kananai,a. Interestingly, according to (LJNLA 19), Josephus and the Septuagint usually use kc for k with a dagesh (Zakcou for yK'êz:, Zakcour for rwkuz, and zakcw for $zng). Also BDF (§39(2)-(3)) asserts that c may translate q (e.g. sabacqa,ni for yntqbv), a (~Akeldama,c, Sira,c), or x (as will be mentioned).

7.3.2.2 Sibilants Hebrew sibilants (s, v, f, z, and c) are usually transliterated into Greek sibilants. It is usual that z transliterates into z. However, z may be used for intervocalic c and s may translate to z. BDF (§39(4)) may cite Nazare,q 40

The reason for classifying plosives and fricatives into one group is that when Hebrew plosives and fricatives are transliterated to Greek, plosives are mixed with fricatives. Sound values of consonants and vowels of Semitic and Greek are under debate, as mentioned before.

306

Phonology

as z used for c. LJNLA (19) demonstrates the following cases; Sakcai/oj (yakz), VEski,aj (hyqzx), and VIw,asdroj (rz[wy). BDF (§39(4)) suggests Bo,ej from z[b. Ilan (LJNLA 20) suggests that the Latin s is sometimes translated into z (i.e. Zene,kaj) and that the double s sometimes was used in the transliteration of X such as VAbessalw,m (~wlXba), VElissai/oj ([Xyla), Manassh/j (hXnm), and VIassou,ou ([wXy). In transliteration from a Greek word into Hebrew, the Hebrew sibilant c was sometimes used to transliterate s because the c is similar to s (LJNLA 20). LJNLA (20) illustrates by citing three cases: !yrdcqla from VAle,xandroj, !wjcra from VAri,stwn, and ylybwrjca from VAristo,bouloj. Although the Greek letter s/j is usually transliterated into s, X also transliterates into s/j. LJNLA (20) suggests some cases: Xrgna from VAqhnago,raj, aXkla from VAlexa/j, !yjXd from Dosi,qeoj, !wyXqpa from Euvxa, Xynrpa from Euvfro,nioj, and Xmwn from No,moj. Ilan (LJNLA 20) also mentions that some Semitic names mix a X with a s; abX-abs, abXk-abswk, and hlbXm-hlbsm deserve mention.

7.3.2.3 Liquids Hebrew liquids (l and r) are transliterated into Greek liquids. In transliteration the Hebrew letters, m and n may be interchangeable at the phonological level as LJNLA (29) cites in two cases: ~yky-!yky and ~wlX!wlX (Salwm -Salwn). And m and n are interchangeable in Marian/Mariam/Maria.

7.3.2.4 Gutturals The fact that gutturals (x, [, h, a) do not have their equivalents in the Greek phonetic system implies that when they are transliterated, they must have produced phonological variants. 41 The Patriarchs’ names, ~h'Þr"b.a,; qx'²c.y,I and bqoø[]y: have gutturals. When they are transliterated into Greek, the gutturals are shorn, as shown by VAbraa,m, VIsaa,k, and VIakw,b. In relation to this, Burkitt (1912:378) stated that “without private information, the retranslator from Greek into a Semitic language would not know where to put the gutturals in.”42 Lightfoot (1872:154-5) noted that, 41

42

Speiser (1933b:234) also states that “It is well-known that an extensive system of laryngals is one of the outstanding characteristics of the Semitic languages… It follows that speakers of other languages will find the laryngals difficult to pronounce.” For Origenic transliterations of gutturals in detail, see Speiser 1933b:233-41. Ropes (Beg 3.270) insists that gutturals could produce variants in relation to the names of Si,mwn and Sume,wn. Also, he (Beg 3.269-270, here 270) claims that “It would

Variant Spellings as Transliterated Allolexemes in Bilingual Contexts

307

“For the Hebrew gutturals again the Greeks had no equivalent, and were obliged either to omit them or to substitute the nearest sound which their language afforded.” He (1872:155) demonstrated that the final x is represented by an e; often in the middle by c. Although the gutturals were pronounced, they could never be written. In this respect, Lightfoot (1872:155) proposed that “any divergence from the Hebrew form which can be traced to this cause might be neglected.” Westcott and Hort (1896:313) also showed that when the gutturals such as a, h, [, x are transliterated into Greek, the transliterations among manuscripts are inconsistent, although it is usual that a and [ are represented by smooth breathing while h and x by rough breathing. 43 Westcott and Hort, followed by Robertson (1934:225), cite the following cases: a may be transliterated into VAbraa,m, VElmada,m, VHlei,aj, or w`sanna,; h into {Agar, ~Errw,m, ~Hlei,, or w`sanna,; [ into :Agaboj, :Eber, or :Hr; x into ~Akeldama,c, ~Enw,c, or Eu[a.44 Further, initial [ in some cases is rendered by g (e.g. Gomo,rra and Gaza,; BDF §39 (3)). The guttural letter x is usually translated into a vowel. Sometime c translates it, as shown by ~Raa,b-~Raca,b from bx'r." BDF (§39 (3)) suggests ~Rach,l, VAca,z, Carra,n, and pa,sca. Ilan (LJNLA 19) also mentions the following cases: VAci,aboj for bayxa, Cagei,raj for hrygx, Celki,aj for hyqlx, :Isacoj for xqcy, and Calfei, for yplx. Ironically, she (LJNLA 20) suggests that when a Greek word is transliterated into Hebrew, only k (never x) transliterates into c, as VIsto,macoj is transliterated into swkmjsya, Ni,karcoj into skrqyn, and Su,mmacoj into swkmws. This implies that the transliteration correspondence between Hebrew and Greek does not function in a one-to-one relationship.

7.3.2.5 Additions of Consonants Some consonants can be added for their pronunciation. BDF (§39(5)) suggests some cases: t or d of VIsrah,l-VIstrah,l (Mt 19:28 W, Mk 12:29

43

44

be natural to suppose ‘Sileas’ due to an adaptation to the form of a Semitic name containing a guttural (cf. Si,mwn, Sume,wn), but the names alyv (Talmudic), alyav (Palmyrene), do not exactly correspond to the variation in the Greek and Latin texts of Acts.” However, he also stands in the tradition of the orthographical view. Following Westcott and Hort, BDF (§39(3)) observes that “Yet it is to be hoped that future editions will follow Lagarde, Rahlfs, and the Göttingen editions of the LXX which omit both accents and breathing in proper names and other transliterations wherever absence of terminations and inflection indicate that no Grecizing was intended.” Kahle (1959:164-71) also mentions that the Hebrew gutturals are transliterated into Greek a in the same way (e.g. Aernwn [!wnra] Jer 31:20 A; Aermwn [[!wmrx] Deut 3:8, Ahlam [~ly[] 1Kings 9:1A; Aendwr [rad !y[] 1Chron 8:24A).

308

Phonology

DW, Lk 2:32 D, Acts B several times)-VIsdrah,l (occasionally t, S always d in Acts), t of Mestrem, b of Nebrwd, b of VIambrh/j (2Tim 3:8), p of Samyw/n (Heb 11:32), and p of Salamyiwn. The insertion of consonants is used to articulate the r or s clearly, although the consonant may be mute. It is interesting that the insertion of consonants may betray the phonological values of Hellenistic Greek and, further, help to reconstruct the ancient Greek phonetic system. When the consonants were chosen for articulate pronunciation, the decision was made according to the distinctive features of the previous sound of the inserted consonants. In the case of t or d of VIstrah,l-VIsdrah,l and t of Mestrem (BDF §39; p.21), the fact that t/d is inserted after s indicates that the articulatory place of t/d is the same as that of s since the same articulatory places are required for articulatory convenience. It seems that the three consonants (i.e. t, d, and s) were alveolars.45 As for b of VIambrh/j, p of Samyw/n, and p of Salamyiwn, the fact that b or p is inserted after m means that the articulatory place of b or p is the same as that of m because the same articulatory places are required for convenient articulation. It seems that the three consonants (b, p, and m) must be bilabials.46 The insertions of consonants cause variant spellings. Accordingly, when Semitic words in the New Testament literature are transliterated into Greek in terms of consonantal variations, the variant spellings occur. The variant spellings should be considered as transliterated allolexemes.

7.3.3 Vowels As for vowel transliteration, it is generally agreed that the variational gap between vowels is much wider than that between consonants. When the Semitic vowel letters are transliterated into Greek, they must have produced many transliterated variants. As mentioned before, vowels are varied due to vowel changes (§7.1.3) or dialects (§7.1.4). Above all, the vocalizations of Semitic language make transliterated spellings more varied. Furthermore, concerning irregular correspondence between Semitic and Greek vowels, Rosén (1963:64) persuasively maintains: We are much worse off for the vowels, and the reason is obvious. There is the problem of deciding whether or not, in each particular case, the nonnotation of a Greek vowel by a mater lectionis is just a case of defective spelling or an indication that the Greek word contained no i or no u, or no vowel at all. 45 46

Caragounis (2004:352, 378) supposes that d is pronounced as th of “then” in English. Caragounis (2004:352, 380) supposes that b is pronounced as v of “van” in English.

ȱ

Variant Spellings as Transliterated Allolexemes in Bilingual Contexts

309

The shewa creates many Greek vowel variants. Schmidt (1894:51-2) pointed out that vowel transliteration from Hebrew to Greek in the LXX and the New Testament literature is not a one-to-one correspondence. He showed that the initial shewa, medial shewa, hateph segol, segol, or sere is usually rendered into alpha in an irregular way. Rehm (1958:275 n.2) considers that shewa could be transcribed to e or a (§7.4.3.1). Buth (1981:29) also mentions that shewa can be transliterated into a or o. In the transliteration of a Greek word into Hebrew, LJNLA (22) points out the interchanges between a and h; VIassou,ou-VIhssou,ou, VIwna,qajVIwna,qhj, VIou,dhj-VIou,daj, VIhsiaj-VIaesaiou, and Base,aj-Bhsa/j. The interchange between o and w in transliteration from Hebrew into Greek occurs; LJNLA (22) shows the examples for this: Neika,nor-Neika,nwr, Filo-Filw, Dws[iqeoj]-Dosi,qeoj, Qeodwsiou-Qeodo,sioj, and Kleopa/jKlwpa/j; in the case of interchanges between h and e: Iesou/j-VIhsou/j, VIwse,VIwsh,, Mna,seou-Mna,shou, Sel[m],e-Selame, Fa,beij-Fa,bhj, Lhouei,j-Leui,j, and Seqi-Shtou/. The interchange between a and e also occurs; KafarnaoumKapernaum, Nazara(q)-Nazareq, and Zara-Zare. BDF (§38) shows the interchange between i and ei; Beniamin-Beniamein (î46 Phil 3:5, Rev11:1), Dauid-Daui?d (î45 Mt 20:31)-Daueid (î46 Mt 20:31), VElisabet-VElisabeit (B always, S mostly, CD sporadically), VIericw-VIereicw (Mt 20:29 BCLZ), and Leui(j)-Leuei(j) (î46 Heb 7:5), and Sa,pfira (Acts 5:1; MSS ei, i, u). The example of the interchange between a and e is like these: Dalmati,a and Delmati,a (2Tim 4:10; cf. Deissmann 1903:182). Buck (1927:23) suggests that “the special Attic-Ionic h from a (long vowel) are both seen in Attic-Ionic mh,thr = ma,thr of other dialects.” Furthermore, Gignac (1989:38) mentions that the vowel h in the first-century AD was bivalent. Consequently, when a word is transliterated from Semitic to Greek, it is normal that vowels are frequently varied. On the other hand, embedded Aramaic words in Syriac texts which were re-transliterated from Greek imply that it is more difficult to reconstruct the phonetic values in bilingual situations. Burkitt (1912:378) ascribed the intricateness of transliteration to the different phonetic systems in the source language and the target language. His suggestion is worth quoting in full: Like Hebrew, many of the vowels [of Syriac] do not appear in writing, and those that are written are given in a notation that, according to our ideas, is singularly imperfect. On the other hand, many distinctions are made, especially in the sibilants, which disappear in the Greek, and (as in Hebrew) there are four true guttural sounds which are not represented in Greek at all. … The real difficulty and the real interest arises (sic!) when, as so often

310

Phonology

in the New Testament, the Proper Name in the Greek is itself a transliteration or adaption of a Semitic word.

Consequently, when a Semitic word was transliterated into Greek or vice versa in terms of vowel variations, the variant spellings occur. Accordingly, it is normal that when Semitic words in the four gospels and Acts were transliterated, variant spellings occurred. The transliterated variants should be called transliterated allolexemes rather than erroneous spellings or corrupted spellings.

7.4 Variant Spellings as Transliterated Allolexemes in the Four Gospels and Acts There are three kinds of transliterated spellings in the four gospels and Acts. Many scholars have investigated the spellings from the perspective of the orthographical view on the basis of the linguistic unidirectionality hypothesis from Semitic to Greek. In this respect, when the spellings are discussed, scholars have taken the “source” into consideration seriously. However, the transliterated spellings will be considered as transliterated allolexemes from the perspective of variational theory on the basis of the linguistic interdirectionality hypothesis between Semitic and Greek. I will illustrate some instances. First of all, three local proper nouns will be discussed (§7.4.1). Then, two personal proper nouns will be investigated (§7.4.2). Last, two other transliterated words will be studied (§7.4.3).

7.4.1 Local Proper Nouns As for transliterations of local proper nouns, a distinction in local proper nouns could be made between very well-known words and little known words. Jerusalem was a very well-known local proper noun in the first centuries. This means that it has two fixed Greek spellings without any variant; a Semitic type and a Greek type. On the other hand, most local proper nouns have their variant spellings in transliteration. The two spellings for Jerusalem and the six spellings for Nazareth will be discussed respectively (§7.4.1.1; §7.4.1.2). Then, the spellings of Gennesaret will be observed (§7.4.1.3).

Variant Spellings as Transliterated Allolexemes in the Four Gospels and Acts

311

7.4.1.1 VIerousalh,m and ~Ieroso,luma In the New Testament literature, Jerusalem has two fixed spellings. One is VIerousalh,m, which is a transliterated form. The other is ~Ieroso,luma, a more morphologically adjusted form with declined spelling. Hort (1881:313) assumes that “All names beginning with y have received the smooth breathing” and he reconstructs VIeroso,luma rather than ~Ieroso,luma. However, many scholars reconstruct ~Ieroso,luma. Bruce (1952:68-9) posits that “As ~Ieroso,luma is a conscious formation from i`ero,j, it should have a rough breathing, pace WH [Westcott-Hort].” The spelling of ~Ieroso,luma was influenced by etymology, as mentioned before (§7.1.2). Concerning the relationship between the two spellings of Jerusalem and their sources, it has been assumed that the Semitic type spelling is related to a Palestinian source, whereas the Greek type spelling is related to a Hellenistic source. Recently, Kilpatrick made a distinction between them. He (1981:353) insists that VIerousalh,m of Mt 23:37 and Luke 13:34 is derived from a common source and that ~Ieroso,luma of Luke 2:22 originates from another source. However, Cadbury (1966:91) persuasively points out that although scholars have tried to connect the two spellings of Jerusalem with the “source,” “the attempts have failed, and probably the two forms owe their adoption to the changing fancy of the writer in each several instance.”47 Winter (1956-57:141) also suggests that “The reading i`erousalh,m is no proof of a Hebrew or Aramaic source for any section of the Lucan writing.” The two variant spellings of Jerusalem are predictable, explainable, and acceptable. The two spellings are transliterated allolexemes of Jerusalem.48

7.4.1.2 Nazara,, and Nazare,q The spelling of Nazareth in the synoptic gospels occurs nine times (Mt 2:23, 4:13, 21:11, Mark 1:9, Luke 1:26, 2:4, 39, 51, 4:16) and is transliterated into six spellings (i.e. Nazara,, Nazara,t, Nazara,q, Nazare,d, Nazare,t, 47

48

Cadbury (1966:91) also states that “If this tendency to vary is a trait of Luke, these variations must not be used as some of them often have been, as marks of written sources slavishly followed and worked up into a patchwork like the Hexateuch in the Old Testament. For instance, the shifting use of and in Acts has been observed, and attempts made to use it as a criterion for the analysis into sources. In this particular case the attempts have failed, and probably the two forms owe their adoption to the changing fancy of the writer in each several instance.” Luke intentionally uses the two allolexemes of Jerusalem as codeswitching at the semantic level (§8.4.3.1).

312

Phonology

and Nazare,q). The variant spellings of Nazareth have attracted scholars’ attention and raised three philological questions. The first question is: which spelling is the original spelling out of the six transliterated spellings? Many scholars have assumed that Nazara, is the original Semitic spelling.49 Under this assumption, scholars have suggested two views of the synoptic problem (Q hypothesis and Markan priority without Q; to be discussed later §8.4.3.2). Second, Nazwrai/oj cannot be identical with Nazarhno,j derived from Nazara, because there is a difference between w and a. Third, c is usually transliterated into s, not z. One of the solutions some scholars have suggested is that Nazwrai/oj is related to a pre-Christian sect, “Nazarenes.” However, whichever position scholars may hold, it seems that the issues have been discussed on the basis of the orthographical view and the unidirectionality hypothesis.50 From the perspective of variation theory, it seems that the six spellings are predictable, explainable, and acceptable, and that the two gentilics (Nazwrai/oj and Nazarhno,j) are identical. (i) In terms of the interchange between a and w, it seems that the secondary articulation of [r] in the process of phonological representation (§7.1.2) enables Nazara, to be Nazwrai/oj. Comparing 1QIsaa with the MT, Wise (1992b:573-4) suggests that when schewa mobile is followed by a plosive b or a trilled r, it is realized as o/u vowel.51 When the Aramaic spelling of Nazareth is transliterated into Greek, the ordinary transliteration (i.e. written transliteration) is Nazarhno,j, whereas the pronounced transliteration (i.e. oral transliteration), which is adapted by the secondary articulation

49 50

51

Weiss 1900:30, 303; Burkitt 1912:392; Kilpatrick 1946:50; Fitzmyer 1981-5:530; Tuckett 1996:228 n.64; Goulder 2003:365. Burkitt (1912:392) implies unidirectional transmission when he mentions that if the two spellings (Nazare,t and Nazara,) are Semitic, “our Greek Gospels are some two generations earlier than any surviving monument of Semitic Christianity.” In other words, he posits that the Semitic spelling is earlier than Greek spelling. Schaeder (TDNT 4.875) also remarks that “n¬ór¬j¬ derives directly from the usage of the Aram. speaking disciples of Jesus and the primitive Jerusalem community.” Many scholars have discussed the origin of Nazara, on the basis of the unidirectionality hypothesis; see §8.4.3.2. Wise (1992b:573) reasonably argues that “He [the scribe of 1QIsaa] did not make this substitution in every case, but only when the schwa mobile was followed by the bilabial b or an r … and even then not always.” It seems that “in every case” demonstrates that he seems to regard the phonological change as an allophone which always occurs in the particular condition. “Even then not always” means that the spellings can be varied depending on the application of the phonological representation (i.e. [+ROUNDED]) to the spelling.

Variant Spellings as Transliterated Allolexemes in the Four Gospels and Acts

313

[+ROUNDED], is Nazwrai/oj. 52 Consequently, the two spellings, Nazarhno,j and Nazwrai/oj, are determined by whether the secondary articulation [+ROUNDED] was adapted or not. From the perspective of the grammar of sound, the two spellings are predictable, explainable, and acceptable. This means that the two spellings are allolexemes. (ii) The intervocalic sibilant c could render z as well as s (§7.3.2.2). This means that the transliteration of c of hrcn/trcn to z is predictable, explainable, and acceptable.53 Moore (Beg. 1.427) persuasively suggests that the fact that z of Nazwrai/oj and Nazarhno,j is uniformly transliterated into c in the Old Syriac and Peshitta versions indicates that c of the Aramaic spelling also transliterates z of Nazarhno,j and Nazwrai/oj in New Testament times. Rüger (1981:262) also suggests that Hebrew rcn could have been pronounced as Na,zar in New Testament times. Brown (1993:207-8) asserts that one of the peculiarities of the Palestinian Aramaic dialect is that an intervocalic c could be transliterated into z due to partial assimilation.54 The other five spellings (i.e. Nazara,q, Nazara,t, Nazare,d, Nazare,t, and Nazare,q) can be explained as transliterated allolexemes more easily because interchanges of d/t/q in transliteration are acceptable, as Zenner (1894:745) has suggested (also, see §7.3.2.1). The variant vowels (i.e. e and a) are interchangeable (§7.1.3; §7.3.3).55 The z as a transliterated allophone of c of hrcn/trcn shows that Nazwrai/oj is derived from Nazara,.56 52

53

54

55 56

This is supported by two earlier scholars. Klein (1923:202-5; followed by Moore [Beg1.427-9]) admits metathesis between a and o and asserts that Aramaic Nazareth could be vocalized into Nazare,q (aq)-et(at), Nazara,, as hbjy and tbjy show. Another example is the transliteration of [r] of “WRITING” into Korean variant spellings (Table 3). For the detailed discussion of the etymological origins of the two spellings (i.e. Nazarhno,j and Nazwrai/oj), see Brown 1993:209-213; 223-5; Davies and Allison 19881997:275-81. Regarding the transliteration of c into z Brown (1993:207-8) illustrates çz of Job 1:1 (some LXX mss.) and Zogora in Greek copies of Jer 48:34 (LXX31:34). Burkitt (1912:404-6) presents ten cases from the OT, although he denies that the two spellings are not identical. It is adduced from the fact that the partial assimilation, the change of an intervocalic sibilant [-VOICED] into [+VOICED], is also found in Greek as well as German. It is well-known that the interchange between © and ¼ in transliteration occurs; Sperber 1937-8:180; Albright 1946:398. Brown (1993:209) criticizes the orthographical view that “the biblical attitude is often a ‘both … and,’ rather than an ‘either…or’” and suggests that both Nazwrai/oj and Nazarhno,j are acceptable (cf. §7.2.2.4). Denying the similarity between Nazwrai/oj and a pre-Christian sect, the “Nazarenes,” Davies and Allison (1988-1997:281) state that “So it seems more prudent to accept the simplest solution: Nazwrai/oj = o` avpo. Nazare,q. This entails further that any connexion with n¬zîr or n¾óer should be re-

314

Phonology

(iii) It is not certain whether Nazara, is the earlier spelling compared to Nazare,q because both –ah and -eth are feminine endings in Hebrew. Zenner (1894:744) suggests that Nazara, and Nazare,q are variants which refer to the same place because of the two accepted Hebrew endings. Dalman (1935:59) enumerates many proper nouns which preserve the t ending; the place names are Daberath (tr:bÞ D. ("/Dabirwq Josh 19:12; 21:28; 1Chr 6:57) and Tsarephath (tp;rê c> /å" Sa,refqa Obad 20; cf. 1Kings17:9). The personal names are Basemath (tm;îf.B' /Basemmaq Gen 36:3; 1Kings 4:15) and Asenath (tn:©s.a'(/Asenneq Gen 41:45).57 This implies that it is possible that the Aramaic original spelling is both hrcn and trcn. In other words, it is hard to insist that Nazara, is earlier than Nazare,q. Accordingly, all six Greek spellings, Nazara,, Nazara,t, Nazara,q, Nazare,d, Nazare,t, and Nazare,q, are transliterated allolexemes which are equivalent to the two Aramaic spellings, hrcn and trcn. The two variants, Nazwrai/oj and Nazarhno,j, were used identically for a man of Nazareth. Furthermore, the bilingualism of first-century Palestine and the Roman Near East implies that both the Aramaic spellings and Greek spellings were circulated among Judaeo-Palestinian residents. Writers/speakers and readers/listeners could understand the variant spellings without any confusion.

7.4.1.3 Gennhsa,r and Gennhsare,t The transliterated spellings of Gennesaret will be investigated from the perspective of the same variation pattern of Nazareth (§7.4.1.2) and Elizabeth (§7.4.2.2). It seems that the spellings of Gennesaret have two kinds of Semitic feminine spellings. Above all, there is no doubt that

57

garded not as primary but as secondary, the result of homeophony noticed once Nazwrai/oj and Nazarhno,j had already come into existence.” Fitzmyer puts forward a well-balanced conclusion. He (1981-5:1215) claims that “Probably the best explanation of Nazçraios at the moment is to regard it as a gentilic adj. meaning ‘a person from Nazara/Nazareth,’ but with the possible added nuance of either n¬zîr, ‘consecrated one,’ or n¾óer, ‘sprout, scion’ of Davidic lineage.” For more cases, see Zenner 1894:744 n.1; Dalman 1935:59. Zenner (1894:744) suggests that Nazare,q is the original ending out of the two Hebrew spellings. Moore (Beg1.429) maintains that “there is no philological obstacle to deriving Nazwrai/oj, Nazarhno,j, from the name of a town, Nazareth.” Criticizing Kennard (1946:131-41; 1947:79-81), Albright (1946:398) proposes that “All three forms [Daberath, Anaharath, and Nazareth] are normal Aramaic.” “Our three names,” he (1946:398 n.2) continues, “were pronounced in the first century A.D. with final et.” He concludes that Nazwrai/oj is a demonym of Nazare,q and the two spellings are Aramaic. Rüger (1981:262) also upholds that transliterated variants, Dabwr and Debhrwq, of rwbd are a similar case to Nazwrai/oj and Nazare,q.

Variant Spellings as Transliterated Allolexemes in the Four Gospels and Acts

315

Gennhsa,r was commonly used in the first centuries. 58 However, Gennhsare,t appears in the authoritative manuscripts of the synoptic gospels (Mt 14:34, Mark 6:53, Luke 5:1).59 This implies that Gennesaret in Semitic could have had two feminine-ending spellings which are equivalent to Gennhsa,r and Gennhsare,t.60 From the perspective of the orthographical view, some scholars have assumed that Gennhsa,r is more correct than Gennhsare,t and that the former has a temporal priority over the latter in the synoptic gospels. Even Dalman (1935:121) states that “The lengthened form of Gennesar, Gennesaret … is foreign to Aramaic-speaking people, and is most probably a wrong formation, one the model of Nazareth.” Following Bauer (BAG 156), BDF (§39 (2)) insists that “Gennhsareq, -ret are incorrect, Gennhsar is correct in D, LXX, Jos. and elsewhere.” On the other hand, others have admitted that Gennhsa,r is the “correct” spelling, although it is not the “original” spelling in the synoptic gospels. Burkitt (1912:391) is convinced that although the correct spelling of Gennesaret was Gennhsa,r, the original spelling in the Greek Gospels could be Gennhsare,t rather than Gennhsa,r because the former occurred in non-Western texts. Matthew and Luke replaced Gennhsare,t with Gennhsa,r. 61 And, later, Gennhsare,t “was re-introduced into the Greek text of all three Gospels.” McNeile (1915:221) also asserts that Gennhsa,r is “probably more correct … But it is not necessarily the true Gk. reading in the gospels.” Taylor (1952:332) suggests that “Gennhsa,r is probably the more correct form … it is less certain that is the original 58

59

60

61

The spelling occurs in the Talmud (rswnyg), Targums (rsyng Onk. Jer 2:5 on Dt 33:23; Onk. Jer 1 on Num 34:11), Syriac versions (rsng) including vocalized G¾n¾sar in the Peshitta, Josephus (Gennhsa,r; Bell.2.573, 3.463, 506, 515, 516. Ant.13.158; Gennhsari,tidi Ant.18.28, 36; Genhsari,doj Ant.5.84), and 1 Macc 11:67 (Gennhsa,r). Strabo (Geogr. 16.2.16 = GLAJJ1.288) called it Gennhsari/tij. Pliny the Elder (Nat. Hist. 5.15.71 = GLAJJ1.469) named Gennesaret Genesaram. Luke preserves Gennhsare,t without any variant. Along a similar line, the spelling Nazare,q in Luke is used on four occasions without variants (Lk 1:26; 2:4, 39, 51; 4:16), although Nazara, is used as a codeswitching at 4:16 (§8.4.3.2). Matthew and Mark show its variant spellings. Matthew 14:34 has four variant spellings: Gennhsare,t (a B C (N) W G 0106 f 1 33. 579. 892. (1241. 1424). l 844. l 2211 al f (syh) mae), Gennhsa,r (D* 700 lat sys.c.p), Gennhsara,t (Dc), and Gennhsare,q ((L) Q f 13 Û q (sa bo)). Mark 6:53 has Gennhsare,t with two variant spellings: Gennhsa,r (D it vgmss sys.p boms) and Gen(n)hsare,q (B* K N Q f 1.13 565 pm lat co). The variability of the two transliterated spellings can be explained. The interchange between t and q is predictable and acceptable (§7.3.2.1). For the vowel interchanges, see §7.1.3; §7.3.3. Burkitt (1912:391) explains the reason for its change “by the more literary Evangelists Luke and Matthew. Harmonistic corruption would then cause the rarer form ‘Gennesaret’ to drop out of Mark.”

316

Phonology

Marcan reading.” However, this discord between “correct” and “original” is due to the orthographical view of variant spellings. From the perspective of the variational view, the two transliterated spellings are predictable, explainable, and acceptable. I would contend that the two spellings (Gennhsa,r and Gennhsare,t) are transliterated allolexemes. As mentioned before, Luke preserves the -re,t ending spelling without any variation and the authoritative texts of Matthew and Mark have Gennhsare,t as well. This fact leads to the assumption that Gennesaret could have had two Semitic feminine spellings due to morphological leveling62 analogous to the spellings of Nazareth (§7.4.1.2) and Elizabeth (§7.4.2.2). Also, the morphological leveling could be caused by the connection between Gennesaret of the synoptic gospels and Chinnereth of the OT. This is adduced from the Targums. The Targums indicate the connection between Gennhsare,t and Chinnereth. Hebrew tr,N